PROCESS FOR HYDROLYSING FISH BONE, PRODUCT THEREFROM AND ITS USE FOR IMPROVING FLESH PIGMENTATION IN A SALMONID FISH

The present invention relates to the use of fish bone hydrolysate as a feed ingredient for fish to improve the uptake and/or retention of dietary astaxanthin. In particular, it relates to the use of fish bone hydrolysate to enhance flesh

pigmentation in farmed fish, especially in salmonids such as Atlantic salmon. The invention further relates to novel processes for preparing fish bone hydrolysate materials.

Astaxanthin is the major carotenoid pigment responsible for the pink-red flesh colour of certain fish such as salmon. It is also present in the external skeleton of other aquatic species such as shrimp, krill, lobster and crayfish. Aquatic animals cannot synthesise astaxanthin. In the aquatic environment, astaxanthin is produced in microalgae which are consumed by zooplankton, insects or crustaceans which accumulate astaxanthin, and which in turn are ingested by fish (e.g. salmon and trout). In fish farming astaxanthin must be supplemented in the diet to achieve the desired flesh colour. Products containing astaxanthin are commercially available as feed additives for the pigmentation of farmed fish and are widely used in the aquaculture industry. One approved feed ingredient is CAROPHYLL ^® Pink (available from DSM Nutritional Products, France); this contains astaxanthin encapsulated in a starch-coated matrix of carbohydrate and gelatin.

Astaxanthin may be extracted from krill and shrimp processing waste, however almost all astaxanthin for use in aquaculture is chemically synthesised from petrochemical sources. Free (i.e. non-esterified) astaxanthin is available in the form of products such as Lucanthin Pink (BASF) and Carophyll Pink (DSM), and in the form of a Chinese Pink product (Zhejiang Nhu Import & Export Company). These all provide 10 % free astaxanthin and are similar in price, stability and availability. The cost of astaxanthin has declined during recent years and today it accounts for about 3 % of the dietary cost.

The market value of various fish products, in particular those from Atlantic salmon, is closely related to the pink-red flesh pigmentation, and in real choice experiments it has been shown that consumers are willing to pay significantly more for fillets with normal or above normal redness compared to paler fillets. As the pink-red colour of salmon fillet is an important quality criterion, any feed ingredient which is capable of improving uptake and/or incorporation of astaxanthin in the flesh would be of significant commercial value.

Although adequate flesh pigmentation is essential for the salmon farming industry it can be difficult to obtain satisfactory fillet pigmentation (industrial goal: 6-7 mg carotenoid kg ^"1 ). There may be several reasons for this. For example, digestive and absorptive processes and metabolic turnover have been shown to influence the utilisation and flesh deposition of carotenoids in salmonid fishes. In salmonid fishes, the concentration of idoxanthin (3,3',4'-trihydroxy-P,P-carotene-4- one), a reductive metabolite of astaxanthin, has been found to decrease with fish size which implies that the rate of metabolic transformation of astaxanthin is higher in smaller than in larger fish. Astaxanthin is also a powerful antioxidant and in addition acts as a precursor of vitamin A in salmonid fishes. Carotenoid retention also varies considerably during the seawater production phase. A low astaxanthin deposition rate in the flesh may be a result of reduced uptake from the intestine and/or an increase in metabolic turnover of the digested astaxanthin, and there are also indications of a higher turnover of astaxanthin during periods of increased oxidative stress.

A need exists for methods of enhancing the deposition and/or retention of astaxanthin in farmed fish, in particular in salmonids, in order to improve pigmentation and the commercial value of the fish in the food industry.

The inventor has now found that the addition of a hydrolysate of fish bone (also referred to herein as "fish bone hydrolysate" or simply "FBH"), e.g. a spray dried hydrolysate of fish bone, to the diet of farmed fish (e.g. Atlantic salmon) can influence the uptake and/or incorporation of dietary astaxanthin into the flesh, thus improving its pigmentation and enhancing its appeal to consumers. More specifically, as shown in the studies described herein, the inventor has found that the concentration of astaxanthin increases in different tissues such as plasma, liver and muscle of Atlantic salmon by up to 30-40 % when fed a diet containing about 4 % (w/w) fish bone hydrolysate, thus demonstrating a significant and pronounced biological impact on pigment utilisation. Methods for the hydrolysis of fish bone to solubilise minerals, in particular phosphorus, and make them more available to fish have previously been proposed. WO 03/056936, for example, describes a process for the production of a feed for fish and animals in which a raw material comprising a fraction of bone tissue is subjected to an acid or alkaline treatment in order to solubilise the minerals in the bone tissue. Incorporation of the resulting material in a feed product is proposed for use in improving the biological digestibility of minerals, in particular the

digestibility of phosphorus. In WO 03/056936 the use of 3 parts water and ½ part 36% hydrochloric acid is proposed for the treatment of fish bones from salmon.

Fish bone hydrolysate is a relatively complex substance containing several macro and micro minerals, in addition to bone collagen proteins and other water soluble compounds from fish bones. It should not be confused with fish protein hydrolysate (FPH) which is produced by enzymatic hydrolysis of fish protein. FPH production is used to convert the muscle protein in fish by-products into acceptable protein ingredients for the fish feed or food industry. Such FPH materials are described, for example, in EP 0951837 and in Hevray et al. (Aquaculture Nutrition 11 : 301-313, 2005).

In a previous study with Atlantic salmon, FBH produced from blue whiting improved growth by about 5% (Albrektsen et al., (2013) - Improved phosphorus utilization in Atlantic salmon (Salmo salar L.) by acid hydrolysis of bone minerals in fish meal - abstract (poster) at the Interdisciplinary Approaches in Fish Skeletal Biology (IAFSB), Tavira, Portugal). Significantly increased protein, lipid and total energy digestibility due to FBH inclusion was also found in this study, to some extent explaining the improved growth performance. Details of the inventor's earlier work relating to the use of herring filleting waste as a source of phosphorus can be found in popularized form at http://nofima.no/en/nyhet/2014/04/from-granite- to-gold-in-phosphorus-production/.

Fish bone processing (e.g. hydrolysis) and methods to produce feed ingredients from fish bone are generally known, but to date it has not been recognised that these ingredients might have any effect on flesh pigmentation in fish whose diet includes astaxanthin. The inventor's finding that such materials are capable of enhancing flesh pigmentation arising from dietary astaxanthin is contrary to expectation; it is well known that certain minerals in common use in fish feeds (e.g. those present in FBH) may act as pro-oxidants, potentially depressing pigment utilisation due to increased oxidation of astaxanthin.

Thus, in a first aspect the invention provides the use of a fish bone hydrolysate material as a fish feed or component of a fish feed for improving flesh pigmentation in a salmonid fish.

In a further aspect the invention provides a method of improving flesh pigmentation in a salmonid fish, said method comprising the step of feeding to said fish a fish bone hydrolysate material or feed composition containing said material.

As used herein the term "fish bone hydrolysate material" encompasses any product which contains fish bone hydrolysate or which is derived (e.g. extracted) therefrom. "Fish bone hydrolysate" refers to any product produced from a material which contains a major proportion of fish bone by a process which involves acid or alkaline hydrolysis. As a result of the hydrolysis process, proteins and minerals in the bone are solubilised and so the resulting product contains at least a proportion of solubilised bone minerals, solubilised bone proteins (e.g. bone collagen protein), and typically also organic sulphur-containing compounds such as glucosaminoglycans (GAGs).

Fish bone hydrolysate materials and methods for their preparation are generally known in the art and include those described in WO 03/056936, the contents of which are hereby incorporated by reference. Any of the hydrolysed materials described in this earlier reference may be used in the invention. As described herein, the inventor has also developed novel processes for the production of fish bone hydrolysate materials. Any of the products resulting from these processes may also be employed in any aspect of the invention relating to enhancement of flesh pigmentation.

Starting materials from which fish bone hydrolysate may be produced include whole fish, by-products (i.e. waste material) from the fish industry or intermediate products produced in fish processing and fish meal production.

By-products from the fish industry will typically comprise fish skin and bones which are produced following skinning and filleting. Intermediate products produced in fish processing include, for example, particles of bone from fish meal preparation. Such particles may be collected by sieving before milling of the fish meal.

Fish by-products and intermediate products from fish processing will generally be used in the production of fish bone hydrolysate. The species of fish from which these are obtained is not critical, although for large scale production of the hydrolysate material those fish which are readily available as a source of raw material will generally be chosen. This includes fish which are extensively fished in large quantities, for example in Scandinavia, Iceland and the Faroe Islands, such as cod, herring and blue whiting, and which are traditionally used in the production of fish meal and/or for human consumption. In Chile and Peru, fishery by-products from other species such as Anchoveta, menhaden and horse mackerel are produced in large volumes and these may also be used. Generally, any locally produced fish waste or bone material can be used as the source of the fish material for the hydrolysis process.

Fish bone hydrolysate for use in the invention may be prepared using methods known in the art or using any of the new methods herein described. In an initial step, the bone fraction will typically be separated from any remaining flesh attached to the bones, for example mechanically in a bone separator or by sifting of the bone fraction from a fish meal. Enzymatic methods may also be employed to remove flesh from the bones and thereby result in the production of another added value by-product, i.e. fish protein hydrolysate (FPH). The separated bone material may be washed (e.g. using water) prior to treatment to produce the hydrolysate material. The bone fraction may then be further processed by acid or alkaline treatment in order to solubilise and release the minerals. Acid hydrolysis is generally preferred. Typically, the raw bone material and water are provided in a ratio by weight ranging from 1 :3 to 1 :7, preferably about 1 :5, prior to addition of the acid.

Hydrolysis may be carried out using an aqueous acid to produce the hydrolysate. The acid used may be any strong acid, but will generally be an inorganic acid such as sulphuric acid, hydrochloric acid, or phosphoric acid. It is possible to use a weak acid, such as formic acid, but this is considered to be less efficient for the release of minerals. Mixtures of acids may also be used. The use of sulphuric acid (H ₂ S0 ₄ ) is especially preferred in preparing the hydrolysate material - this has been found to be more efficient than hydrochloric acid (HC1). The amount of sulphuric acid to be added may range from about 1 to 10 wt%, preferably from about 1 to 8 wt% (based on the total weight of raw bone material and water), and will typically be about 1.5-5 wt%. Alternatively, the amount of sulphuric acid may range from about 1 to 10 vol.%, preferably from about 1 to 8 vol.% (based on the volume of sulphuric acid added to the water), and will typically be about 1.5-5 vol.%. As evidenced herein, there may be advantages associated with the use of lower concentrations of sulphuric acid, e.g. concentrations less than 5 wt.%, e.g. about 2-2.5 wt.%, or concentrations less than 5 vol.%, e.g. about 2-2.5 vol.%. Although HC1 is efficient in solubilising minerals in the bone material, results obtained by the inventor suggest this may result in a high proportion of Ca in the product which can have a negative impact on nutrient digestibility. However, when using H ₂ S0 ₄ , any excess sulphate ions rapidly precipitate Ca by producing a low soluble salt, CaS0 ₄ . This salt can easily be separated and, if desired, provides a useful by-product. Where HC1 is used, a further processing step may be performed in order to separate Ca from the bone hydrolysate, thereby eliminating the anti-nutrient properties of Ca. The use of sulphuric acid in a method for the preparation of a fish bone hydrolysate material forms a further aspect of the invention.

When performing hydrolysis using an acid, the precise hydrolysis conditions may be readily determined according to the raw material to be treated and the nature of the acid used (e.g. type of acid, acid strength, etc.). When using, for example, concentrated sulphuric acid (e.g. 95-98% H ₂ S0 ₄ ), this may preferably be added to the bone material in an amount of about 2.5 to 5 % w/w, e.g. about 5 % w/w or about 2.0 to 2.5 % w/w (based on the total weight of raw bone material and water). Alternatively, this may be employed in an amount of about 2.5 to 5 % vol., e.g. about 5 % vol. or about 2.0 to 2.5 % vol. (based on the volume of sulphuric acid added to the water). Other acids, such as hydrochloric acid, may also be used in a similar amount. When using hydrochloric acid this will also typically be used in concentrated form, e.g. about 37% HC1. Other acids which may be used include phosphoric acid and formic acid. Acid hydrolysis may be performed in a conventional fashion. Typically, hydrolysis, which is exothermic, will be carried out without cooling in order to maintain the temperature of the hydrolysis mixture in the range from 20 to 40°C, preferably 25 to 30°C. Duration of the hydrolysis process will depend on the volume of raw material to be treated and the nature of the acid used, but generally this will be carried out for a period of at least 4 hours, typically up to 24 hours, e.g. 4 to 20 hours. Hydrolysis overnight with continuous stirring is generally preferred.

Following hydrolysis, the resulting product will comprise a liquid (i.e.

aqueous phase) which contains solubilised components including minerals, and a solid phase containing non-solubilised bone residue. Whilst this product may be used directly in the methods of the invention, i.e. without further processing, it is preferred that only the aqueous (i.e. soluble) fraction is used. In a preferred embodiment, the liquid and solid phases may therefore be separated using conventional separation techniques, e.g. by sieving and/or filtration. To further enhance the recovery of solubilised materials resulting from the hydrolysis process, the solid residue may be washed (e.g. with water and/or acid) one or more times and the resulting washings combined with the separated liquid phase. The solubilized bone minerals may be returned to a fish meal to improve the mineral digestibility as described, for example, in WO 03/056936. However, the use of the soluble fraction alone, i.e. without returning this to the raw fish material or combining this with any intermediate products in the fish meal processing, is preferred. A method for the production of a fish bone hydrolysate material which involves separation and recovery of the aqueous (i.e. soluble) fraction following acid treatment represents a further aspect of the invention.

Addition of a base may be carried out to neutralise the acid, for example by adjusting the pH of the liquid fraction to a pH range of about 2 to 3.5, preferably 2.5 to 3.2, e.g. about 2.8 or about 3.1. Any suitable base may be used, for example ammonia solution, potassium hydroxide, sodium hydroxide, etc. In some embodiments, use of potassium hydroxide as a neutralising agent may be preferred. Standing for a period of up to 24 hours, preferably from 18 to 24 hours, e.g.

overnight, may be done to allow for the precipitation of any remaining solids. Any oil components (e.g. lipids) which float on the top of the liquid following standing may be removed, for example by gravity separation or decanting.

Following removal of any remaining solids and/or oil components, a clear liquid remains. This contains the solubilised mineral components from the fish bone, and other water-soluble components. The resulting clear liquid may be used as it is as a fish bone hydrolysate material, however this will typically be further treated prior to use. For example, this may be sterilised by methods known in the art, such as heating to a temperature of up to 100°C, preferably up to 95°C, e.g. to about 90°C. Heating may be carried out for a period of up to 10 minutes, preferably up to 5 minutes. Following heating, the sterilised liquid may be filtered to remove any remaining traces of solid material, e.g. by sieving through a fine sieve (for example, a sieve having a pore size in the range of 50 to 100 μπι, e.g. about 70 μπι).

The liquid hydrolysate may be used directly (i.e. without further processing) in the products and methods herein described. Alternatively, this may be subjected to further processing steps before use. Further processing steps may include concentration to remove excess water from the product. The choice of any additional drying step or steps will depend on the water content of the material and the desired moisture content of the final product and can be appropriately selected by those skilled in the art. The degree of concentration of the product may be measured using the refractive index of sugar as an indirect measure of its dry matter content. For example, concentration may be performed to yield a product with a dissolved sugar content in the range of 15 to 35 °Brix, preferably about 33 °Brix. Removal of some or all of the water may be carried out by spray drying in order to provide a free flowing particulate product. Typical particle sizes for the dried product may range from 10 to 100 μπι, preferably from 20 to 60 μπι, e.g. about 40 μηι.

Whilst other methods may be used to carry out the step of hydrolysis of the bone material, such as alkaline treatment, acid treatment as herein described is generally preferred.

As mentioned above, certain processes for the production of the fish bone hydrolysate materials as herein described are in themselves novel and these represent further aspects of the invention. In a further aspect the invention thus provides a process for hydrolysing fish bone comprising at least the following steps:

(a) subjecting a raw material which comprises at least a proportion of fish bone to acid hydrolysis in the presence of sulphuric acid;

(b) optionally processing the resulting hydrolysis product whereby to

separate the solid phase and soluble fraction; and

(c) optionally returning said soluble fraction (or a proportion thereof) to the raw material or to any other intermediate product produced during the process.

The resulting hydrolysis product may be used directly as a fish feed or feed ingredient or, more preferably, further processed as herein described to produce a fish feed or feed ingredient.

Preferred embodiments of this process are as described above in respect of the general hydrolysis method. Step (a) may be carried out by mixing the raw bone- containing material with water and concentrated sulphuric acid (e.g. 95-98%

H ₂ SO ₄ ). Typically, the raw material (e.g. the separated fish bone) and water are provided in a ratio by weight ranging from 1 :3 to 1 :7, preferably about 1 :5. The amount of sulphuric acid to be added may range from about 1 to 10 wt.%, preferably from about 1 to 8 wt% (based on the total weight of raw material and water), and will typically be about 1.5 to 5 wt.%. Alternatively, the amount of sulphuric acid may range from about 1 to 10 vol.%, preferably from about 1 to 8 vol.% (based on the volume of sulphuric acid added to the water), and will typically be about 1.5 to 5 vol.%). Hydrolysis may be carried out at a temperature of from 10 to 40°C, e.g. for a time of at least 4 hours, typically up to 24 hours, e.g. 4 to 20 hours.

Following hydrolysis with sulphuric acid, the resulting product will comprise a liquid (i.e. aqueous phase) which contains solubilised components including minerals, and a solid phase containing non-solubilised bone residue. The solid phase may also contain precipitated CaS0 ₄ . The resulting product may be separated and the aqueous phase further processed as described above. The separated soluble fraction may be treated with a base (e.g. potassium hydroxide or ammonia solution) whereby to neutralise the acid, e.g. to adjust the pH to between 2.0 and 3.1. The separated soluble fraction may be substantially free from Ca due to its precipitation as CaS0 ₄ and removal as part of the solid phase.

Thereafter this soluble fraction may be returned to the raw fish material (e.g. the raw fish bone) or any intermediate products produced in the production method. In a preferred embodiment, however, the processed soluble fraction is not returned to the raw material or to any intermediate products in order to produce the final hydrolysed product. The final product is thus preferably substantially free from any non-solubilised bone residues, e.g. this contains less than 10 wt.% of such residues, preferably less than 5 wt.%. In producing the final product, the soluble fraction will typically be concentrated and dried as herein described.

In this aspect of the invention hydrolysis is effected by use of the sulphuric acid alone, e.g. in the absence of other known hydrolysis agents such as enzymes.

In another aspect the invention provides a process for hydrolysing fish bone comprising at least the following steps:

(a) subjecting a raw material which comprises at least a proportion of fish bone to acid hydrolysis;

(b) processing the resulting hydrolysis product whereby to separate the solid phase and soluble fraction; and

(c) recovering said soluble fraction.

The resulting soluble fraction may be used directly as a fish feed or feed ingredient or, more preferably, further processed as herein described to produce a fish feed or feed ingredient.

In this aspect of the invention any acid may be used, including both sulphuric and hydrochloric acids, however the use of sulphuric acid is generally preferred.

Preferred embodiments of this process are as described above in respect of the general hydrolysis method. Step (a) may be carried out by mixing the raw bone- containing material with water and acid. Typically, the raw material (e.g. the separated fish bone) and water are provided in a ratio by weight ranging from 1 :3 to 1 :7, preferably about 1 :5. The amount of acid may range from about 1 to 10 wt.%, preferably from about 1 to 8 wt% (based on the total weight of raw material and water), and will typically be about 1.5 to 5 wt.%. Alternatively, the amount of acid may range from about 1 to 10 vol.%, preferably from about 1 to 8 vol.% (based on the volume of sulphuric acid added to the water), and will typically be about 1.5 to 5 vol.%. Hydrolysis may be carried out at a temperature of from 10 to 40°C, e.g. for a time of at least 4 hours, typically up to 24 hours, e.g. 4 to 20 hours. Where HC1 is used as the acid, for example, hydrolysis may be performed for about 2 to 6 hours, e.g. about 4 hours.

In this aspect of the invention the solid and liquid phases are separated using conventional separation techniques (e.g. by sieving and/or filtration) and only the liquid phase forms the final product or is further processed to form the final hydrolysis product. The final product is thus substantially free from any non- solubilised bone residues, e.g. this contains less than 10 wt.% of such residues, preferably less than 5 wt.%. The separated soluble fraction may be treated with a base (e.g. potassium hydroxide or ammonia solution) whereby to neutralise the acid, e.g. to adjust the pH to between 2.0 and 3.1. Where H ₂ S0 ₄ is used for hydrolysis, the separated soluble fraction may be substantially free from Ca due to its precipitation as CaS0 ₄ and removal as part of the solid phase. Thereafter, the product may be concentrated and dried as herein described to produce a final hydrolysis product.

The hydrolysis products formed by the new processes herein described also form a further aspect of the invention.

For use in the invention, typically the fish bone hydrolysate will be incorporated into a conventional fish feed, for example a formulated fish feed, which will also comprise known feed ingredients such as fish meal, plant ingredients, etc. This feed material may additionally include astaxanthin, although this is not essential. In the case where the fish feed is not formulated to contain astaxanthin, this component of the diet would be provided as a separate product to be fed to the fish in conjunction (e.g. separately or simultaneously) with the hydrolysate- containing product. Generally, however, it is envisaged that all required feed ingredients, including astaxanthin, would be presented in a single product.

In a further aspect the invention thus provides a fish feed comprising a fish bone hydrolysate material as herein described, together with astaxanthin. As used herein, the term "fish feed" will generally be understood to be one which provides all the necessary components of a fish diet, including protein, lipids, carbohydrates, etc.

In a related aspect the invention provides the use of a fish feed comprising a fish bone hydrolysate as herein described and astaxanthin for improving flesh pigmentation in a salmonid fish.

In a yet further aspect the invention provides a method of improving flesh pigmentation in a salmonid fish, said method comprising the step of feeding to said fish a fish feed as herein described.

Typically the fish bone hydrolysate material and astaxanthin will be directly incorporated into a conventional fish feed, for example, a formulated fish feed containing other known fish feed components such as fish meal, plant ingredients, etc.

When provided as a component of a fish feed, the fish bone hydrolysate may be provided, for example, as an admixture with other conventional ingredients such as any or all of the following: fish meal; plant materials such as soybean meal, maize gluten, wheat gluten, pea protein concentrate, and sunflower meal; oils such as fish oil and vegetable oils; binders such as wheat and horse beans; and additives such as vitamins and minerals. Methods for mixing feed components and providing fish feeds, for example in extruded or pellet form, are known in the art. Preferred forms for the fish feed include dry pelleted, expanded and extruded forms and also moist and semi-moist forms. In many cases, it is envisaged that the admixed product will be produced at the point of use by the fish farmer. Typical processes include mixing and co-extrusion of the various components in order to provide an extruded product in pellet form.

Astaxanthin may be provided in any of the feed products herein described by admixture of the feed ingredients with a known commercial source of astaxanthin, for example Carophyll ^® Pink (available from DSM), Lucanthin Pink (available from BASF), etc. Such products are typically supplied in protected form in which the astaxanthin is embedded in a matrix to prevent its degradation during handling and storage. These products also typically include an antioxidant to protect the sensitive astaxanthin molecules. During the manufacture of the fish feed (e.g. by admixing and extrusion of the various components), some of the astaxanthin may be liberated from any protected (e.g. encapsulated) form. Free astaxanthin is unstable and to a large extent this may be oxidised during high temperature extrusion.

The amount of astaxanthin present in the feed product will be dependent on factors such as the type and size of fish to which this is to be fed, the maturity of the fish, etc. but can readily be determined by those skilled in the art. Suitable levels of incorporation of astaxanthin may range from 10 to 60 mg per kg diet, preferably from 20 to 50 mg per kg diet. Commercial astaxanthin sources in common use, such as Carophyll Pink, typically provide 10 % astaxanthin.

Although the invention is described primarily in the context of improving the utilisation of astaxanthin in order to enhance flesh pigmentation, other pigment sources may also benefit from the methods herein described. Other pigment sources include synthetic pigments such as canthaxanthin, and natural pigments derived from sources such as yeast Phaffia sp., green algae Haematococcus sp., and crustaceans. Though not widely used in Atlantic salmon, these other pigments find use in feeds for other salmonids such as trout.

Typical levels of inclusion of the fish bone hydrolysate in the diet will be dependent on several factors, including its phosphorus content, dietary P

requirement, the species of fish, time of year, etc., but may readily be determined by those skilled in the art. Typically the fish bone hydrolysate may be incorporated at a level of 1 to 5% by weight, preferably 2 to 4% by weight (based on the total weight of the feed).

A typical feed for a salmonid may comprise: 2 to 4 % by weight fish bone hydrolysate, 20 to 50 mg per kg astaxanthin, 10 to 30 % by weight fish meal, 5 to 15 % by weight fish oil, 15 to 30 % by weight plant oils, 10 to 50 % by weight plant proteins, 8 to 16 % by weight starch components (e.g. wheat), and 2.5 to 5 % by weight of other components (e.g. other minerals, vitamins, etc.). Examples of such feeds are provided in the Examples presented herein.

In one embodiment the fish feed for use in the methods herein described may comprise (by total weight of feed): 15 % by weight fish meal, 6 % by weight fish oil, 24 % by weight plant oils, 37 % by weight plant proteins, 14 % by weight starch components (e.g. faba beans), and 4 % by weight of other components (e.g. minerals, vitamins, amino acids, etc.). The fish feed of the invention will typically comprise other components for the health and nutrition of fish, such as are listed above. The quantities of these additional components may be determined by those skilled in the art.

Fish which may benefit from the invention are those which deposit axtaxanthin in their flesh and which are naturally red / pinkish in colour. Primarily the invention is applicable to salmonids (of the family Salmonidae) and particularly salmon, e.g. Atlantic salmon, Pink salmon, Coho salmon, Chum salmon; and trout, for example rainbow trout, etc.

Although the invention is described herein primarily with reference to the use of hydrolysate materials produced from fish bone, it extends also to the use of hydrolysis products from other animal bones, e.g. bones from poultry (chickens, turkeys, etc.), pigs, cattle and sheep. Methods for producing such products may be analogous to those described herein for the production of fish bone hydrolysate.

In a broader aspect the invention thus provides the use of an animal bone hydrolysate material as a fish feed or component of a fish feed for improving flesh pigmentation in a salmonid fish.

In a further aspect the invention also provides a method of improving flesh pigmentation in a salmonid fish, said method comprising the step of feeding to said fish an animal bone hydrolysate material or feed composition containing said material.

In a yet further aspect the invention provides a fish feed comprising an animal bone hydrolysate material, together with astaxanthin. The use of such a fish feed in a method of improving flesh pigmentation in a salmonid fish also forms part of the invention.

The invention will now be described in more detail in the following non- limiting Examples and with reference to the accompanying Figures in which:

Fig. 1 shows daily growth (SGR, %) and feed conversion rate (FCR) in Atlantic salmon fed the control diet (Dl) with added inorganic P and with FBH added at 4.2 % (D2) for a feeding period of 11 weeks according to Example 1;

Fig. 2 shows apparent digestibility (ADC) of astaxanthin in Atlantic salmon fed the control diet (Dl) with added inorganic P and with FBH added at 4.2 % (D2) for a feeding period of 1 1 weeks according to Example 1; Fig. 3 shows muscle Ax retention (% of dietary Ax) in Atlantic salmon fed the control diet (Dl) with added inorganic P and with FBH added at 4.2 % (D2) for a feeding period of 11 weeks according to Example 1;

Fig. 4 shows muscle Ax retention (mg/kg weight gain) in Atlantic salmon fed the control diet (Dl) with added inorganic P and with FBH added at 4.2 % (D2) for a feeding period of 11 weeks according to Example 1;

Fig. 5 shows apparent mineral digestibility (ADC) in fish fed solubilized P from HC1 and H ₂ SO ₄ treated fish bone meal relative to fish fed added inorganic P (the 0-baseline) according to Example 2; and

Figs. 6 to 8 show total astaxanthin (Ax) and Ax isomers in plasma, liver and whole body of fish, respectively, fed a P non-supplemented control diet (Dl), and P supplemented as inorganic P (D2), HC1 treated fish bone meal (D3) and H ₂ SO ₄ treated fish bone meal (D4) according to Example 2.

Example 1 - Feeding trial with 1.7 kg Atlantic salmon

Production of fish bone hydrolysate (FBH):

Commercially produced and separated fish bone meal from blue whiting (Karmsund Fiskemel AS, Norway) was sifted (4.6 mm) and used as raw material for hydrolysis in a strong acid (H ₂ SO ₄ ). The sifted fish bone meal contained 41.5 % ash and 11 kg batches were mixed with water (1 :5) in 6 x 100 L plastic drums.

Concentrated H ₂ S0 ₄ (95 %) was added in an amount of 5 % w/w and hydrolysis was performed overnight (18 hrs) to solubilise the minerals from the hydroxy apatite skeleton.

The water fraction was manually separated from the solid phase by filtering over a fine mesh cloth (70 μπι). The solid phase was washed twice in cold water to collect all water soluble compounds released during hydrolysis. The liquid phase was adjusted to pH = 2.1 with 25 % NH ₃ solution, and to a final pH = 2.8 with KOH (33.3 % solution). The six drums were allowed to stand overnight to precipitate more solids.

Floating lipid on top was removed, and the clear liquid was carefully pumped into a boiling pan, heated to 90°C (5 min) and separated over a Jesma sieve (80 μηι cloth) prior to separation of the lipid (GEA Westfalia separator, Germany). The liquid phase was concentrated (33° Brix) and spray dried (Niro Atomizer, Denmark).

About 20.5 kg fish bone hydrolysate from blue whiting (FBH blue whiting) was produced. The P content in the final spray dried FBH was 9.5 % total P and 9.5 % soluble P. Protein content in the final product was 14.1%, lipid content 1.1 %, ash 61.5 %, and water content 2.1 %.

Experimental diets:

Two experimental diets with similar levels of soluble (non-bone P)

(8 gkg ^"1 ) were produced by extrusion at Nofima's Feed Technology Centre, Bergen, Norway. A practical diet with 36% protein and 34% lipid that reflected a standard commercial diet for Atlantic salmon of size > 1.5 kg was produced. The basal diet contained 15% fish meal (21.8 gkg ^"1 total P, 10.9 gkg ^"1 soluble P), 10% pea protein concentrate (8.0 gkg ^"1 total P, 7.8 gkg ^"1 soluble P), 19% soy protein concentrate (SPC) (6.9 gkg ^"1 total P, 6.9 gkg ^"1 soluble P) and 6 - 6.5% wheat gluten (2.3 gkg ^"1 total P, 2.3 gkg ^"1 soluble P). Astaxanthin was supplemented as free astaxanthin (80% All-E, 5% 9Z, 15% 13Z) in Carophyll Pink (CP 10% - CWS) from DSM (DSM Nutritional Products, France SAS) at 55 mgkg ^"1 . The feeds were stored frozen at -20°C from arrival at the research site until loading on the feeding automats to avoid any loss of astaxanthin.

The diets were balanced to meet the requirement of available P (8 gkg ^"1 ), NRC (2011), and essential amino acids (Lys, Thr) for normal growth and

development. The control diet (Dl) contained about 4 gkg ^"1 soluble P from the basal ingredients and further contained 4.0 g kg ^"1 P from CaP0 ₄ . The FBH diet (D2) was supplemented with 4.2 % FBH to provide the additional 4 gkg ^"1 soluble P. The vegetable proteins provided about 28% of the total level of soluble P (2.3%), of which 60 - 80 % is expected to be phytic acid. Diet composition and chemical contents are presented in Table 1 below: Table 1 : Diet ingredients and composition

Dl D2

SoluWe P (g P kg;^) 8 8

ingredients, gkg ^'1

Fish meal ³ 150 150

FBH Blue whiting ^b 42

SPC 190 190

Wheat gluten 67 63

Pea protein concentrate 100 100

Lysine-HCl 8 8

Thr 4 4

Fish oil ⁰ 66 66

Canola oil 236 236

Soya lecithin 5 5

Wheat 119 105

Betafin 5 5

Vitamin mix ^d 20 20

Mineral mix ⁶ 5.2 5.2

CaP0 ₄ - MCP (T29/10) 2.4 0

Carophyll Pink, 10% 0.55 0.55

Y ₂ 0 ₃ 0.2 0.2

Diet composition

Protein, g kg ^"1 366 366

Lipid, g kg ^"1 334 319

Water, g kg ^"1 64 70

Ash, g kg ^"1 62 73

Total energy ¹ , MJ kg ^"1 24.8 24.2

Total P, % 11.9 10.8

Soluble P, % 8.6 7.6

Free Astaxanthin, mgkg ^"1 56 57

pi I 5.6 5.8 ^a LT fish meal, SILFAS, N-5892, Bergen, Norway. Protein: 708 gkg ^"1 , Lipid: 110 gkg ^"1 , Ash: 108 gkg ^"1 , Water: 87 gkg ^"1 .

FBH Blue whiting: Protein: 89 gkg ^"1 , Lipid: 16 gkg ^"1 , Ash: 677 gkg ^"1 , Water: 38 gkg ^"1 .

cNorsalmoil, Norsildmel AL, N-5141 Fyllingsdalen, Norway

d Added per kg feed: vitamin D ₃ 3000 I.E., 160 mg; vitamin E, 136 mg; thiamin, 20 mg; riboflavin, 30 mg; pyrodoxine-HCl, 25 mg; vitamin C, 200 mg; calcium pantothenate, 60 mg; biotin. 1 mg; folic acid, 10 mg; niacin, 200 mg; vitamin Bi _2> 0,05 mg; menadion bisulphite, 20 mg

e Added per kg feed: magnesium 500 mg; potassium 600 mg; zinc 120 mg; iron 60 mg; manganese 30 mg; copper 6 mg. Experimental fish and handling:

The experiment was carried out in 6 experimental sea cages at the research station of Nofima AS, Averay, Norway (63° 03 ΤΝΓ, 7° 35Έ) and run for 11 weeks. At the start of the experiment, the salmon (1.7 kg ± 0.01 kg) were randomly distributed to six 5 x 5 x 5 m (125 m ³ ) cages, each holding 55 fish. The fish were fed one of the two experimental diets in triplicate sea cages for a feeding period of 79 days. Prior to the trial, the fish were stocked in a 7 x 7 m cage and fed a pigment-free experimental diet (~ 2.4 mg astaxanthin kg ^"1 ) during a 4 week acclimatisation period. The fish had been fed commercial Skretting feeds during transfer at Sunndals0ra and prior to the acclimatisation period. The salmon were from the Salmobreed strain delivered as 0 ⁺ smolts from Nofima' s Aquaculture Station, Sunndals0ra.

The fish were fed three 30 min meals per day by automatic feeders (Betten Maskinstasjon AS, Norway). Waste feed pellets were collected by means of a lift- up system and quantified daily. Recovery tests were done with each diet in empty cages to adjust for water loss (Helland et al., Aquaculture 261 : 603-614, 2006). Pumping of waste feed from the bottom cone in each cage started 4 min after termination of the meal and the collection lasted for 4 min. Based on the feed intake from the previous 2-3 days, feeding levels were adjusted in an attempt to maintain approximately 10-15 % excess feeding to ensure feeding to satiation. The water temperature was measured daily and decreased with the ambient temperature throughout the trial. Average temperatures measured at 2 week intervals were 9.1, 8.1, 7.4, 5.8, 4.7 and 4.6°C, respectively, and the average temperature for the total feeding period was 6.8 ± 1.7°C. The fish were exposed to ambient temperature and light regimes. As the ambient temperature decreased to less than 5°C in the last 4 weeks of the trial, the feed intake was very low in the last weeks of the trial.

Collection of faeces was difficult because only small remains of faeces were found in the hind gut of several fish.

At the start of the experiment, the salmon were dip-netted from the 7 x 7 cages, anaesthetised in Finquel (Western Chemical Inc., Scottsdale, USA, 0.35 gL ^"1 ) and individually weighed (nearest 10 g) and length measured (fork length, nearest 0.5 cm). Fish weighing from 1.5 to 1.9 kg were selected for the experiment and fish were distributed to the six experimental cages. Three samples of 10 fish were taken for analyses of muscle pigmentation and lipid analyses at the start of the trial.

Another two samples of 10 fish were sampled for analyses prior to the 4 week depigmentation period.

At week 11, all cages were emptied and the continuously fed fish from each cage were transferred to and anaesthetised in Finquel (Western Chemical Inc., Scottsdale, USA, O SgL ¹ ) in large tanks supplied with running seawater. Pooled samples of approximately 20 - 30 g of faeces (wet weight) were collected and immediately frozen at -20°C following manual stripping of faeces of fish from each respective tank (n = 20 - 30 fish). Ten immature fish from each cage were transported to the processing room and blood withdrawn from Vena caudalis (5 ml heparanized syringes, 5000 IE heparin, n = 10 fish) before stunning the fish with a blow to the head. The fish were gutted, the sex registered and fish measured for weight, fork length, gutted weight and liver weight. Individual plasma samples were centrifuged at 4000 rpm, 6 min, and stored in dry ice (-80°C) before transport and analyses of plasma astaxanthin at the laboratory at Sunndals0ra. Both sides of the muscle were collected, the left half side of muscle was used for visual (Salmofan), colorimetric (Photobox) and chemical (astaxanthin) analyses and for calculation of muscle astaxanthin retention. The other half side of the muscle (n = 10), as well as liver (n = 10) were immediately frozen at -20°C without homogenisation for later transport and analyses at laboratories at Sunndals0ra and Bergen where they were kept at -30°C. Pooled samples of liver and muscle were analysed for lipid, dry matter, astaxanthin, vitamin E and lipid peroxidation (TBARS).

Chemical analysis:

All chemical analyses were carried out in duplicate by laboratories accredited by the Norwegian National Accreditation body. In feed ingredients, diets, faeces and fish tissue, crude protein (N x 6.25) was determined by the Kjeldahl method (ISO 5983-1997), moisture gravimetrically after drying for 4 hrs at 105°C (ISO 6496-1999), and ash after combustion for 16 hrs at 550°C (ISO 5984-2002). Lipid contents in feed and faeces were determined by acid hydrolysis («EU-lipid»), (Commission Directive 98/64/EC, Part B) and by Folch extraction with acid hydrolysis (SSF-report: A-102, 1978) respectively, and the lipid content in liver and muscle following acidic extraction (Folch). Astaxanthin was analysed according to an HPLC (high performance liquid chromatography) method developed by Hoffman La Roche (Analytical Methods for Vitamins and Carotenoids in feed, Revised supplement: Determination of stabilized astaxanthin in Carophyll Pink, premixes and fish feeds, pages 59-61, 1994) following ethanol and di chloride methane extraction of astaxanthin from the experimental diets and fillets. Carophyll Pink in the experimental diets was enzymatically treated in hot water prior to the extraction procedure. Crude protein in the experimental diets was enzymatically treated in hot water prior to the extraction procedure. Yttrium was determined in feed and faeces by inductively coupled plasma atomic emission spectroscopy (ISO 11885-1996). Muscle and liver vitamin E was determined by HPLC and fluorescence detection after saponification and extraction according to Lie et al. (Transport of alpha- tocopherol in Atlantic salmon (Salmo salar) during vitellogenesis, Fish Physiology and Biochemistry 13 : 241-247, 1994). Certified reference materials, in house control materials and control chart were used for quality assurance. Total P in feed ingredients and feeds were determined by spectrophotometry (430nm) following ashing and acid digestion in 6 M HCl (ISO 6491-1998). Soluble P (NaOH extracted P) was also determined according to the same method following incubation of duplicate ingredient samples (0.8 g) in 80 mL of 1 N NaOH for 16 hrs prior to acid digestion, according to a procedure described in Ruban et al. (J. Anal. Chem. 370: 224-228, 2001), modified by Hua et al. (Journal of Nutrition 02; 142(4): 668-74, 2005) and later modified and validated by Hovde (Validation of a method for analysis of soluble phosphorus by use of alkaline extraction and spectrophotometric determination, Nofima report 17, ISBN: 978-82-8296-077-9 (pdf), 2013).

Pigment analysis (Salmofan andphotobox):

The visual colour of the fillets was measured using the Roche SalmoFan™ (Hoffman La-Roche, Basel, Switzerland) score (Scale 20-34) and by digital image analysis using a standardised photo box, a digital camera and prediction equations for astaxanthin (mg/kg), Salmofan score and fat (%) (Folkestad et al., Aquaculture 280: 129-135, 2008). The right fillet was placed in a Salmon Colour Box with standardised light conditions and SalmoFan readings were performed immediately after slaughter in 3 positions (upper dorsal, NQC and tail) in the upper region of the fillet by two judges. The fillets were then photographed by a digital camera (Dolphin F145C, Allied Vision Technologies, Stadtroda, Germany) in a light proof aluminium box (800 mm x 830 mm x 955 mm) with standardised illumination provided by four fluorescent bulbs (OSRAM Lumilux 55 W, OSRAM, Augsburg, Germany) placed along each of the walls in the upper part of the box. The camera was placed on the top of the box, facing perpendicularly downwards. A reference image of a white plate covering the entire bottom plate was used to correct for shading, i.e. spatially inhomogeneous illumination of the bottom plate and/or response of the camera lens. A calibration card, QPcard 101, (QPcard AB,

Gothenburg, Sweden) with standardised white and grey patches was placed on the bottom plate and included in each image. It was used for calibration of lightness and white balance. Any variation in illuminance (lux) of the photographed area was thereby corrected for.

Digestibility:

Apparent digestibility coefficient (ADC) of lipid, protein and astaxanthin (Ax) were determined. Yttrium, lipid, protein and total free Ax were determined in the four experimental feeds and in pooled samples of faeces from each replicate cage. Samples of frozen faeces were lyophilised (final plate temperature 24°C) and homogenised prior to chemical analyses. ADC of nutrients in the experimental diets was calculated according to the formula:

ADC =100 - 100 x -

CX _d x Y.

Where d is diet, f is faeces, Y is yttrium concentration and CX is nutrient concentration.

Statistical analyses and calculations:

Biological and analytical data were subjected to Student's T-test analysis by using STATISTICA (Ver 7.1, StatSoft, Tulsa, OK, USA) and differences between means were tested using Tukey HSD test (Sokal and Rohlf (1981) Biometry. W.H. Freeman and Co., NY, 859 pp). Effects with a probability P < 0.05 were considered significant. Calculations:

- Growth, feed intake and feed conversion ratio were determined according to the following formula, where BW ₂ = final body weight and BWi = initial body weight;

- Specific growth rate, SGR = (In BW ₂ - In BWi) / feeding days.

- Thermal growth coefficient, TGC = (BW ₂ ^1/3 - BW ^/3 ) * 1000 /∑ (temp.(°C) * feeding days) according to Cho (Feeding systems for rainbow trout and other salmonids with reference to current energy and protein requirements. Aquaculture 100: 107-123, 1992).

- Daily feed intake per fish = g feed intake / days / number of fish.

- Daily feed intake, % of mean bodyweight = g feed intake / days /((BW ₂ + BWi)/2 * 100) / fish no.

- Feed conversion ratio = g dry feed eaten / g live weight gain. Results:

Feed composition

Proximate composition of experimental diets showed that the intended levels of nutrients, energy, astaxanthin and soluble P was achieved (Table 1). The diets contained average 36.6 ± 0.0 g kg ^"1 protein, 32.7 ± 1.1 g kg ^"1 lipid and 24.5 ± 0.4 KJ g ^"1 gross energy (94 % dm). Dietary astaxanthin was similar in both diets and averaged 56.5 ± 0.7 mg/kg following extrusion, while dietary level of soluble P was 8.1 ± 0.7 %. The low basal level of free astaxanthin found in both diets (2.4 ± 0.2 mg kg ^"1 ) and in the depigmentation diet fed during the acclimatisation period prior to the trial, resulted from a small contribution of natural astaxanthin in the fish oil.

Daily growth and feed conversion rate

During the feeding trial the fish showed high growth rate relative to the low seawater temperature, with an average SGR of 0.52 ± 0.03 % and TGC of 3.23 ± 0.22 for all diets (Table 2). The fish increased the average weight from 1.7 to 2.5 kg during the trial. The FBM diet D2 slightly improved SGR by nearly 4 % and reduced the variation between replicates by about 50 %, but was not significantly different from the control diet (Dl), p > 0.05 (Table 2, Fig. 1). Table 2: Daily specific growth rate (SGR, %), feed intake and feed conversion rate (FCR) in Atlantic salmon fed 8 g kg ^'] P from a basal control diet (Dl) or from FBH (D2) for a feeding period of 11 weeks. Mean values and STD (x) is given for each respective diet, n=3.

Dl D2 Student's T-test

Soluble P (g P kg ^"1 ) 8 8 P^ O.05

Initial weight, g 1707 (7) 1694 (13) ns

Final weight, g 2543 (93) 2565 (55) ns

Weight gain, g 836 (89) 871 (44) ns

SGR, % 0.51 (0.04) 0.53 (0.02) ns

TGC 3.16 (0.29) 3.29 (0.13) ns

FCR 1.01 (0.02) 1.01 (0.03) ns

Feed intake, %/bw/d 0.57 (0.06) ^a 0.56 (0.01) ^a ns

Feed intake, mg/fish/d 12.0 (1.5) 11.9 (0.3) ns

Significant differences within rows are shown with different letters, P < 0.05, ns = not significant.

Nutrient digestibility and Astaxanthin (Ax) retention

Digestibility of protein and lipid was not significantly affected by the diets (P > 0.05) - while the apparent Ax digestibility was increased by 18 % in fish fed 4.2 % FBH (D2) as compared to the control diet (Dl) - see Table 3 and Fig. 2. Ax retention values in the muscle showed significant dietary differences and fish fed 4.2 % FBH (D2) showed significantly higher retention values than fish fed Dl (P < 0.05), Table 3. Muscle retention values were increased by respectively 40, 35 and 35% (Table 3) in fish fed FBH 4.2% (D2) when measured as mg Ax retained, % of dietary Ax retained ( Fig. 3) and mg Ax kg ^"1 weight gain (Fig. 4). Muscle Ax retention calculated as % of Ax absorbed, was increased by 11 % in fish fed D2.

Table 3 : Digestibility of protein, lipid and astaxanthin (Ax), and Ax utilisation measured by muscle Ax retention in Atlantic salmon fed 8 g kg ^'] P in the control diet (Dl) and in the FBH diet (D2) for a feeding period of 11 weeks. Mean values and STD (x) is given for each respective diet, n=3.

1)1 D2 Student ^' s ^: r-test

Soluble P (g P kg ^"1 ) 8 8 P ¹ < 0.05

Apparent digestibility, %

Protein 86.3 (1.0) 86.2 (1.2) ns

Lipid 96.6 (0.7) 96.6 (0.4) ns

Ax 24.4 (1.4) ^a 28.7 (0.5) ^b 0.01

Ax utilization

Muscle Ax ret, mg 113.6 (6.5)' 159.3 (19.1) 0.05

Muscle Ax ret, % of 4.62 (0.18)' 6.24 (0.69) 0.05

dietary Ax

Muscle Ax ret, mg/kg 2.56 (0.14) ^a 3.46 (0.29) ^b 0.01

weight gain

Muscle Ax ret, % of Ax 19.0 (1.4) 21.7 (2.1) ns

absorbed

¹ Significant differences within rows are shown by different letters (P < 0.05), ns = not significant.

Ax concentration in liver, plasma and muscle reflected the increased digestibility in D2, and were increased by 30, 22 and 7 % in the respective tissues, whereas not showing significant dietary differences (P > 0.05) - see Table 4.

Plasma and muscle Ax correlated and showed general high levels of trans Ax (92 - 93 %) - see Table 4. In liver, trans Ax was lower and averaged 58% for both diets. However, fish fed FBH 4.2% (D2) showed 10% higher % trans Ax as compared to fish fed the control diet (Dl). The metabolite of Ax, Idoxanthine (Idox) in the tissues was low for both diets when measured as % of total carotenoid content (3 - 6 %), but tended to be slightly lower in D2 versus Dl . Table 4: Muscle carotenoids measured chemical (Astaxanthin), visually (Salmofan) and colorimetric (Photobox) in Atlantic salmon fed 8 g kg ^'] P from a basal control diet (Dl) or from FBH (D2) for a feeding period of 11 weeks. Muscle Ax is also presented during the 4 week Ax depletion period (2.4 mg Ax/kg feed) prior to start of the trial. Liver and plasma carotenoids in Atlantic salmon are presented at the end of the 11 week feeding trial. Mean values and STD (x) is given for each respective diet, n=3.

Dl D2 Student's T-test

Soluble P (g P kg ^"1 ) 8.0 8.0 P^ O.05

Muscle Astaxanthin, mgkg ^"1

Tot Ax at start 4.87 (0.08)

Tot Ax after depletion 4.03 (0,30)

Muscle carotenoids, mgkg ^'1

Tot Ax 4.63 (0.16) 4.97 (0.30) ns

Ax Trans 4,25 (0.16) 4.58 (0.29) ns

Ax Cis 0.38 (0.01) 0.39 (0.02) ns

Idox (mg/kg) 0.25 (0.04) 0.21 (0.03) ns

Idox( % of total carotenoids) 5.1 4.1

Salmofan

Upper dorsal 28.2 (0.4) 28.1 (0.5) ns

NQC 28.3 (0.4) 28.2 (0.5) ns

Tail 28.2 (0.5) 28.0 (0.6) ns

Photobox

Salmofan 26.0 (0.4) 26.5 (0.4) ns

Ax, mgkg ^"1 6.1 (0.4) 6.7 (0.5) ns

Liver carotenoids, mgkg ^'1

Tot Ax 1.32 (0.34) 1.71 (0.35) ns

Ax Trans 0.76 (0.26) 1.01 (0.21) ns

Ax Cis 0.56 (0.10) 0.65 (0.14) ns

Idox (mg/kg) 0.08 (0.05) 0.08 (0.03) ns

Idox (% of total carotenoids) 5.7 4.5

Plasma carotenoids, mgkg ^'1

Tot Ax 1.38 (0.35) 1.69 (0.20) ns

Ax Trans 1.28 (0.33) 1.57 (0.21) ns

Ax Cis 0.10 (0.02) 0.12 (0.01) ns

Idox (mg/kg) 0.08 (0.03) 0.06 (0.02) ns

Idox (% of total carotenoids) 5.5 3.4 ns

¹ Significant difference is shown by different letters (P < 0.05), ns = not significant. The metabolite level tended to be lower in fish fed FBH at 4.2 % inclusion (D2). The specific Idox concentrations (mg/kg) were reduced by 26, 16 and 3% respectively in plasma, muscle and liver of fish fed D2. Although the change in metabolite level was consistent in all tissues, no significant dietary differences were found (P > 0.05).

Other pigmentation analyses (Salmofan and photobox)

Muscle pigmentation evaluated by Salmofan in the upper dorsal, NQC and tail region of the muscle, correlated to muscle Ax concentration (r = 0.7, P < 0.05), while apparently it did not reflect the variation in the colorimetric and chemical measurements (Table 4). Salmofan was on average 28.2 ± 0.4 and similar for all diets (P > 0.05). Pigmentation measured colorimetric by Salmofan were 2 units lower; 26.2 ± 0.7 and correlated to the observed dietary variation in muscle Ax concentration, showing marginal higher value in fish fed FBH 4.2 % (D2).

Pigmentation evaluated colorimetric (Photobox) showed 9.8 % increased Ax in the muscle of fish fed D2 as compared to Dl, reflecting the increased Ax concentration analysed. Pigmentation measured by colorimetric analysis was 30 - 35 % higher than Ax measured by chemical analyses in the muscle homogenates.

Discussion:

The growth conditions in this winter trial were very good (SGR: 0.52 %; TGC: 3.23) and close to what can be expected in commercial salmon production under such conditions. Due to the extremely low temperature at the end of the trial, declining below 5°C, the feed intake was very low in the last 3 - 4 weeks of feeding and no significant difference in growth between the diets was found. Dietary P supported from the two different P sources was similar and adequate to cover the dietary P requirement established in cold water species (National Research Council, NRC (2011). Nutrient requirements of fish. Washington D.C., National Academic Press). Fish fed FBH at 4.2 % (D2) still showed 4 % higher daily specific growth (SGR) and 50 % less variation between triplicate cages. Similar results have been shown in Atlantic salmon smolt (Albrektsen et al., 2013, supra) and apparently cannot be explained by dietary P provided by the different P sources. Absorptive processes in the gut constitute a major limitation for effective absorption of carotenoids from the gastrointestinal tract in salmonid fishes. The apparent digestibility coefficients (ADC) of Ax reported in the literature vary considerably, but typically range between 20 to 60 %. The digestibility of Ax is negatively correlated with water temperature, feed intake and diet concentration. The ADC of Ax ranged from about 24 to 28 % in the present study. There was an 18 % increase in ADC of Ax in fish fed diet D2 as compared to the control diet, Dl, which indicates a positive effect on gut Ax absorption. The increased ADC of Ax calculated was confirmed by higher Ax concentrations in plasma, muscle and liver, and by higher flesh Ax retention in fish fed D2. Muscle retention of Ax ranged from 4-6 % of dietary Ax and 21.7 % of the digested or absorbed Ax was retained in the flesh of fish fed D2. Considering the amount of absorbed Ax calculated, about 14 % more of the absorbed Ax seems to be deposited in the muscle of fish fed D2 as compared to fish fed the control diet Dl .

The plasma concentration of carotenoids is considered to be a good indicator of bioavailability of carotenoids in salmonid fishes. In this trial, the concentration of Ax in plasma, liver and muscle were all increased in fish fed diet D2, and the dietary impacts on pigmentation was confirmed by the colorimetric (Photobox) and the visual (Salmofan) evaluation, although the dietary differences with respect to Salmofan was negligible.

A low Ax deposition rate in the flesh may be a result of reduced uptake from the intestine and/or an increase in metabolic turnover of the digested Ax. Salmonid fishes are capable of metabolically transforming 4(4 ^' )-ketocarotenoids, and metabolic transformation of the ingested carotenoids plays an important role for accumulation of carotenoids in various body compartments. Idoxanthin is the first reductive product of Ax, and accumulates in plasma and tissues of Atlantic salmon. Plasma idoxanthin concentration is therefore an indication of the metabolism at the time of sampling, whereas the deposition in flesh and liver reflects the metabolism of Ax over a longer period of time. In general, about 60 to 80% of the absorbed Ax is metabolically transformed into other compounds. In this trial there was a slight tendency towards reduced metabolism of Ax into idoxanthin in plasma, liver and muscle by FBH, however the difference was small and significant changes were not present. The idoxanthin content in plasma (in % of total carotenoids in plasma) was lower in salmon fed FBH (3.3-3.4 %) as compared to the control diet (5.4-5.5 %). This slight reduction in formation of idoxanthin indicates positive impacts of FBH on Ax utilization, but cannot explain the 40 % increase in Ax retention observed in fish fed FBH at 4.2 % inclusion (D2). As most of the ingested Ax is metabolized further into colorless compounds, the amount of idoxanthin deposited in muscle and other organs represents only a minor fraction of the total metabolized Ax. This means that potential effects of FBH on Ax metabolism cannot be ruled out on the basis of the results from the present study.

Conclusion:

In this low temperature winter trial the Ax retention was low in fish fed the control diet (Dl), but was increased by 40 % in Atlantic salmon fed FBH at 4.2 % inclusion (D2). The results on Ax uptake and retention from this study suggest that the most probable explanation for the increased pigment utilization efficiency in fish fed FBH is related to improved digestibility of Ax, increasing the intestinal uptake and muscle deposition of Ax. This study confirms that FBH improves pigment utilisation efficiency in Atlantic salmon significantly. The study also confirms that dietary P level cannot explain any of the added value properties observed in this trial, as dietary P was provided at similar levels.

Example 2 - Feeding trial with 170 g Atlantic salmon

The dietary impact of FBH on pigmentation of Atlantic salmon was assessed in a trial with postsmolt Atlantic salmon. In this trial, Atlantic salmon (170 g) were fed one of the four experimental diets in triplicate tanks (n = 60 fish) for a feeding period of 89 days (10°C). Results were evaluated by growth, feed conversion, nutrient and mineral digestibility, P retention and specific mineral status in whole body, scale, vertebrae and plasma of Atlantic salmon (Albrektsen et al., 2013, supra). Based on visual pigmentation observations at the end of the trial, the dietary impact of FBH was also evaluated by whole body, liver and plasma total astaxanthin (Ax) and its free isomers (All-E, 13 Z, 9Z). The FBH were produced by hydrolysis of a fish bone meal separated from blue whiting fish meal, in two different strong acids (HCl, H ₂ S0 ₄ ) according to the following procedure.

Production of fish bone hydrolysates:

Commercially produced and separated fish bone meal from blue whiting (Karmsund fiskemel AS, Norway) was sifted (4.6 mm) and used as raw material for hydrolysis by using two different strong acids (HCl and H ₂ S0 ₄ ). The sifted fish bone meal contained 41.5 % ash, and 10 kg batches were mixed with acid/water (1 :5 w/w) in 4 x 100 L plastic drums, of which 2 drums were used for hydrolysis in each respective acid. Concentrated HCl (37 %) and H ₂ S0 ₄ (98 %) was added (5 % w/w) and a 4 h (HCl) and 24 h (H ₂ S0 ₄ ) hydrolysis process was performed to solubilise the minerals from the hydroxyapatite skeleton. The water fraction was manually separated from the solid phase by filtering over a fine mesh cloth (70 μπι). The solid phase was washed twice in cold water to collect all water soluble compounds released during hydrolysis. pH of the liquid phase was adjusted to pH = 2.6 (HCl) and pH= 3.1 (H ₂ S0 ₄ ) with KOH (33.3 % solution) and the drums were allowed to stand overnight to precipitate the solids. Floating lipid on top was removed and the clear liquid was carefully pumped into a boiling pan, heated to 90°C (5 min) and thereafter separated over a Jesma sieve (80 μπι cloth). The liquid phase was directly concentrated (33° Brix) and spray dried (Niro Atomizer, Denmark) without separation of the lipid. About 7.5 kg (HCl) and 4.5 kg (H ₂ S0 ₄ ) fish bone hydrolysate (FBH) was produced from blue whiting fish bone meal (FBH blue whiting). The total ash content of the ingredients was respectively 63 and 73.6 %, protein content was 4.7 and 9.2 %, lipid content was 1.5 and 1.8 %, total P was 7.9 and 11.7 % and soluble P was 0.4 % and 10.3 % for FBH treated with HCl and H ₂ S0 ₄ , respectively. The marine P ingredients were both evaluated as alternative P sources in the diets.

Experimental diets:

To a low P basal diet (Dl : 6.6 g total P kg ^"1 ) was added 4 g kg ^"1 soluble P as inorganic P (D2: 10.8 g total P kg ^"1 ), as P solubilized by HCl hydrolysis of fish bone meal (D3 : 11.5 g total P kg ^"1 ), or as P solubilized by H ₂ S0 ₄ hydrolysis of fish bone meal (D4: 11.5 g total P kg ^" ' . The inorganic P source was a 1 : 1 mix of KH ₂ P0 ₄ and NaH ₂ P0 ₄ . The basal diet contained 30% fish muscle meal (15.3 g total P kg ^"1 ) separated from blue whiting meal, as the main protein and P source, providing about 4.6 g total P kg ^"1 diet, while the vegetable proteins were selected to provide low additional P to avoid interaction with phytate-P that may inhibit uptake of P and other minerals. Diet composition and chemical contents are presented in Table 5 and Table 6, respectively, and show that the dietary level of soluble (highly available) P (Hovde, 2013, supra) was 2.8, 6.8, 4.9 and 6.9 gkg ^"1 , respectively, in Dl, D2, D3 and D4. In all diets, 64 mg kg ^"1 Ax was supplemented as Carophyll Pink (10%) before extrusion of the feeds.

Table 5: Diet composition

Dl D2 D3 D4

Added inorganic P (g soluble P kg ^"1 ) 0 4 4 4

Ingredient content, gkg ^'

Fish muscle meal, < 1 mm ^a 30 30 30 30

HC1 treated FBH ^b 6.6

H ₂ S0 ₄ treated FBH ^C 4.3

SPC 10 10 10 10

Wheat gluten meal 13.5 13.9 14 13.5

Wheat meal 16.24 14.16 10.59 13.39

Fish oil ^d 25 25 25 25

Carophyll Pink (CP 10%) 0.064 0.064 0.064 0.064

Thr 0.2 0.2 0.2 0.2

Inositol 0.03 0.03 0.03 0.03

Betafin 0.5 0.5 0.5 0.5

Vitamin mixture ⁶ 2.0 2.0 2.0 2.0

Mineral mixture ¹ 0.5 0.5 0.5 0.5

KH ₂ P0 ₄ (22.5 % P) 1.12

NaH ₂ P0 ₄ (22.5 % P) 1.12

K ₂ C0 ₃ (56%) 1.45 0.89

Y ₂ 0 ₃ 0.02 0.02 0.02 0.02 ^a Separated from LT fishmeal , Norway. Protein: 794 gkg ^"1 , Lipid: 63 gkg ^"1 , ash: 103 gkg ^"1 , moisture: 55 gkg ^"1 .

b HCl hydrolysed fish bone meal. Protein: 47 gkg ^"1 , Lipid: 15 gkg ^"1 , ash: 630 gkg ^"1 , moisture: 87 gkg ^"1 .

c H ₂ S0 ₄ hydrolysed fish bone meal. Protein: 92 gkg ^"1 , Lipid: 18 gkg ^"1 , ash: 736 gkg ^"1 , moisture: 24 gkg ^"1 .

dNorsalmoil, Norsildmel AL, N-5141 Fyllingsdalen, Norway

e Provided per kg feed: vitamin D ₃ 3000 I.E., 160 mg; vitamin E, 136 mg; thiamin, 20 mg; riboflavin, 30 mg; pyrodoxine-HCl, 25 mg; vitamin C, 200 mg; calcium pantothenate, 60 mg; biotin. 1 mg; folic acid, 10 mg; niacin, 200 mg; vitamin B _12> 0,05 mg; menadion bisulphite, 20 mg

f Provided per kg of feed: magnesium 500 mg; potassium, 600 mg; zinc, 120 mg; iron, 60 mg; manganese, 30 mg; copper, 6 mg.

Table 6: Proximate composition, energy and mineral contents of

experimental diets

Dl D2 D3 D4

Added inorganic P (g soluble P 0 4 4 4 kg ^"1 )

Ingredient content

Protein, g kg ^"1 443 434 436 436

Lipid, g kg ^"1 279 275 276 277

Moisture, g kg ^"1 56 70 53 65

Ash, g kg ^"1 60 67 85 75

Gross energy ¹ , MJ kg ^"1

Minerals

Ca, g kg ^"1 diet 9.4 8.4 18.9 10.4

Total P, g kg ^"1 diet 6.6 10.6 11.5 11.5

Soluble P 2.8 6.8 4.9 6.9

Dietary Ca/P -ratio 1.4 0.8 1.6 0.9

¹ Gross energy values are calculated according to the following caloric values (MJ/kg): Protein 23.6, lipid 39.5 and carbohydrate 17.1. All values are corrected for indigestible energy from cellulose.

Experimental fish and handling:

Underyearling (0 ⁺ ) Atlantic salmon smolt (Salmo breed) were obtained from a commercial hatchery (Sj0troll, Norway) and transported to NOFIMA's aquaculture research station in Austevoll, Bergen. The fish were fed a commercial diet (Skretting Spirit 75) throughout the 6 week acclimatization period and with a P non-supplemented diet (2.8 g soluble P kg ^"1 ) for another 3 weeks in order to reduce the body stores of P prior to start of the experiment. The salmon (-170 g) were randomly distributed to 12, 1.5x1.5 m glass fibre tanks (2.1 m ³ ), each holding 60 fish. The fish were fed one of the four experimental diets (5 mm pellet), in triplicate tanks, for a feeding period of 89 days. All tanks were equipped for continuous monitoring of feed refusals. Fish were fed to satiety by automatic feeders, and the daily feed rations were adjusted according to assumed fish biomass and feed intake, and with 10 % feed in excess. The fish were continuously fed for 16 h at two daily feeding periods, 1400 - 0300 and 0500 - 0800, each of 10 seconds duration intervened by 145 seconds. The collected feeds were dried in an oven at 40°C for 24 h, and the calculated amount of feed eaten was used for determination of daily feed intake and feed conversion. All tanks were supplied with running seawater taken from 50 m depth. Average water temperature was 10.1 ± 0.4 °C during the trial, kept constant by a heat exchanger. Salinity was 3.1 - 3.2 % throughout the feeding trial, water flow 50-55 L min ^"1 and the oxygen content in the outlet water not lower than 7.5 mgL ^"1 (85 % saturation). The fish were exposed to 24 h light during the experimental period of 89 feeding days.

Weight and length were recorded in individual fish from each tank at the start and end of the experiment. At the end of the trial, fish from each tank (Fish no 1 - 15) were collected for X-ray analyses and for carcass proximate composition and mineral analyses (Fish no 1- 5). Weight and length was recorded in 5 fish (Fish no. 6 - 10) and pooled samples of plasma (200 μΐ from each fish, n = 5), scale and vertebra were collected for mineral analysis, and liver for proximate analyses.

Blood samples were withdrawn from vena caudalis in 5 ml heparanized blood syringes (5000 IE heparin). The blood samples were kept refrigerated (4°C) and plasma collected following centrifugation at 4000 rpm for 6 min. Faeces were collected by stripping from 25-30 fish (about 30 g w/w.). Before handling, the fish was sedated in Eugenol (50:50 Eugenol/Ethanol). Sampled fish were transferred to smaller tanks with about 15 - 20 L seawater added 20 mL Metacain (Finquell) to anesthetize the fish. All tissue homogenates were frozen at minimum -20°C, and samples of whole body, liver and plasma were later analysed for total and free Ax, including the Ax isomers (All-E, 13 Z and 9 Z Ax).

Chemical analysis:

All chemical analyses were carried out in duplicate by a laboratory accredited by the Norwegian National Accreditation body. In feed ingredients, diets, feces and fish tissue, crude protein (N x 6.25) was determined by the Kjeldahl method (ISO 5983-1997), moisture gravimetrically after drying for 4 h at 105°C (ISO 6496-1999), and ash after combustion for 16 h at 550°C (ISO 5984-2002). Lipid contents in feed and feces were determined by acid hydrolysis («EU-lipid»), (Commission Directive 98/64/EC, Part B) and by Folch extraction with acid hydrolysis (SSF-report: A- 102, 1978) respectively, and tissue lipid content following acidic extraction (Folch). Total P in feed ingredients and feeds were determined by spectrophotometry (430nm) following ashing and acid digestion in 6 M HC1 (ISO 6491-1998). Soluble P (NaOH extracted P) was determined according to the same method following incubation of duplicate ingredient samples (0.8 g) in 80 mL of 1 N NaOH for 16 h prior to acid digestion, according to a procedure modified and validated by Hovde (2013, supra). Yttrium was determined in feed and feces by inductively coupled plasma atomic emission spectroscopy (ISO 11885-1996). Freeze-dried samples of salmon tissues and feces were subjected to analysis of multiple elements by ICP-AES (EPA method 200.7 by Analytica AB, Lulea, Sweden) after acid digestion with a mixture of nitric, hydrochloric and hydrofluoric acid in sealed teflon containers in a microwave system.

Astaxanthin was analyzed according to a HPLC (high performance liquid chromatography) method developed by Hoffman La. Roche (1994) following ethanol and dichloride methane extraction of Ax from the experimental diets, in the whole body, liver and plasma. Carophyll Pink in the experimental diets was enzymatically treated in hot water prior to the extraction procedure. Certified reference materials, in house control materials and control chart were used for quality assurance.

Calculations:

Growth, feed intake and feed conversion ratio were determined according to the following formula, where BW2 = final body weight and BW1 = initial body weight: Specific growth rate, % SGR = (In BW ₂ - In BWi) * 100 / feeding days. Thermal growth coefficient, TGC = (BW ₂ ^1/3 - BW ^/3 ) * 1000 /∑ (temp.(°C) * feeding days) according to Cho (1992, supra). Daily feed intake, % of mean bodyweight = g feed intake / days / ((BW ₂ + BWi)/2 * 100) / fish no. Feed conversion ratio = g dry feed eaten / g live weight gain. CF (condition factor) = (g live weight / L ³ )* 100. HSI (hepatosomatic index) = (g liver/g live weight)* 100. DOP (Dressing out percentage) = (g dress-out weight /g live weight)* 100 Statistical methods:

Dietary impacts on biological and analytical data were statistically evaluated according to one-way analysis of variance (ANOVA) and Tukey HSD test (Sokal and Rohlf, 1981, supra). All data were statistically treated according to

STATISTICA 7.0 for Windows version 5.02 (Wilkinson et al., 1992. Statistics. SYSTAT Inc, Evanston, Illinois. 750 pp.). Pearson correlation analysis was used to examine possible relationships between feed parameters and dietary responses. Effects with a probability P < 0.05 were considered significant.

Results:

Dietary impacts on growth FCR and mineral utilization

During the feeding experiment the best performing fish quadrupled their body weight, and showed high growth rate (SGR > 1.6 %), Table 7. Fish fed the P non- supplemented control diet Dl showed significant lower growth and feed efficiency (P < 0.05) compared to the other P supplemented diets. Fish bone meal hydrolyzed with H ₂ S0 ₄ (D4) improved growth by about 5% (ns) while fish bone meal hydrolyzed with HC1 (D3) reduced growth significantly by 10% (P < 0.05) as compared to fish fed added inorganic P in the diet (D2). The daily feed intake was not affected by diet when measured as % of mean bodyweight per day (ns), although total feed intake was significantly reduced in fish fed Dl (P < 0.05). Feed conversion rate (FCR) was increased by 8 % (P < 0.05) in fish fed D3 (HC1 treated FBH) and by 20 % in fish fed Dl as compared to fish fed D2 added inorganic P. In fish fed D4 (H ₂ S0 ₄ treated FBH), FCR was similar to fish fed D2.

Table 7: Growth, feed intake and feed conversion rate (FCR) of Atlantic salmon fed diets supplemented with 0 (Dl) or 4 g soluble P kg ^'1 , provided as inorganic P (D2), HCl (D3) and H ₂ SO ₄ (D4) treated FBH, all values given as means and SEM (n=3).

Dl D2 D3 D4 ANOVA

Added inorganic P 0 4 4 4 P*<

(g soluble P kg ^"1 )

Body weight at start, a 172 (2) 172 (1) 172 (<1) 171 (1) ns

Body weight at end, g 550 (8) ^a 724 644 762 (8) ^c 0.01

(17) ^c (12) ^a

Weight gain, g 358 (9) ^a 552 (16) ^b 472 (12) ^c 591 (8) ^b 0.01

SGR, % 1.27 1.62 1.48 1.68 0.01

(0.02) ^a (0.02) (0.02) ^c (0.01)

TGC 2.81 3.79 3.41 3.97 0.01

(0.05) ^a (0.07) (0.06) ^c (0.03)

Feed intake, %/bw/d 1.04 1.04 1.05 1.09 ns

(0.01) (0.01) (0.01) (0.01)

FCR 0.91 0.75 0.81 0.76 0.01

(0.01) ^a (0.01) (0.01) ^c (0.01)

¹ Significant difference within rows is shown with different superscript letters, P < 0.05. ns = not significant. ² p < 0.06

Whole body chemical composition and nutrient digestibility

Whole body protein was unaffected by diet (P > 0.05), while whole body lipid and ash was significantly increased and reduced, respectively, in fish fed the P non-supplemented diet Dl as compared to fish fed the P supplemented diets (P < 0.05). In fish fed HCl treated fish bone meal (D3), whole body ash was significantly reduced as compared to fish fed D2 and D4 (P < 0.05), consistent with the lower growth in this group. For all diets, an inverse relation between whole body lipid and ash appeared (P < 0.05). In fish fed H ₂ SO ₄ hydrolyzed fish bone meal (D4), the apparent digestibility (ADC) of protein, lipid and energy (Table 8), and a range of macro- and micro minerals (Fig. 5) was significantly improved (P < 0.05). Table 8: Whole body chemical composition and apparent digestibility coefficients (ADC) of nutrients in Atlantic salmon fed diets supplemented with 0 (Dl) or 4 gP kg ^'1 , provided as inorganic P (D2), HCl treated FBM (D3) and H ₂ SO ₄ treated FBM (D4), all values given as means and SEM (n=3).

Dl D2 D3 D4 ANOVA

Added inorganic P 0 4 4 4 P*<

(g soluble P kg ^"1 )

Whole body, l OOg ^{" 1}

Protein 17.2 17.0 17.2 17.1

(0.3) (0.1) (0.1) (0.2) ns

Lipid 17.8 15.4 15.8 16.3

(0.5) ^a (0.1) 0.2) 0.2) 0.01

Ash 1.20 1.73 1.57 1.73

(<0.01) ^a (0.03) (0.03) ^c (0.03) 0.01

Phosphorus, P 2373 3490 3297 3251

(36) ^a (28) (69) (133) 0.01

ADC, %

Protein 86.9 85.7 85.7 87.9

(o.i) ^a (0.2) (0.2) (0.3) ^c 0.01

Lipid 92.1 91.8 90.9 94.8

(0.8) ^a (0.9) ^a (0.5) ^a (0.2) 0.01

Energy 81.1 80.2 79.8 83.9

(0.4) ^a (0.4) ^a (0.2) ^a (0.4) 0.01

¹ Significant difference within rows is shown with different superscript letters, P < 0.05. ns = not significant.

In fish fed HCl treated FBH (D3), the digestibility of protein, lipid and energy was not affected (P > 0.05), Table 8, while ADC of ash, Cu, Zn, Se, Mo and Co was increased and ADC of (P), Mn, Fe, Cr, and Li was reduced relative to fish fed added inorganic P (D2), Fig. 5.

Whole body, liver and plasma Ax and TBARS

Total free Ax in plasma, liver and whole body of fish fed H ₂ S0 ₄ treated fish bone meal (D4) were significantly increased by 55, 29 and 37 %, respectively, as compared to fish fed inorganic P (D2), P < 0.05, Figs. 6-8. Similar changes were observed for All-E Ax, while 13Z Ax was increased mainly in plasma, and not in whole body (ns).

Whole body, liver and plasma total free Ax and All-E Ax of fish fed HC1 treated fish bone meal (D3) were significantly lower than fish fed D2 and D4 (P < 0.05), consistent with the lower growth in this group. Whole body total Ax and All- E Ax was significantly higher as compared to fish fed Dl (P < 0.05), while similar for liver and plasma Ax (P > 0.05).

In the P non-supplemented diet (Dl), whole body, liver and plasma total free Ax and All-E Ax were all significantly lower than in fish fed D2 and D4, P < 0.05. Plasma 13Z Ax was not detectable in fish fed Dl and D3 while 9Z Ax was below detection limits in all the tissues.

The relative amount of All-E Ax compared to total Ax ranged from 85-90% in whole body (P > 0.05), 54 - 80 % in liver (P < 0.05) and 85-100% in plasma (P < 0.05). In liver, the highest amount of All-E Ax was found in the two best performing diets D2 and D4 (-80 %), while in plasma, these diets showed the lowest relative All-E Ax (85-90%) although the specific Ax values were higher.

Apparently, some of the free Ax is converted into 13Z Ax when plasma Ax is high, as no detectable level of 13Z Ax was found in fish fed Dl and D3.

Whole body TBARS was significantly increased in fish fed HC1 treated FBH (D3) as compared to any of the other diets (p < 0.05), while liver TBARS was unaffected by diet (p > 0.05). The variation in liver TBARS was high, possibly indicating some problem with the reliability of the method. Besides from

demonstrating similar and adequate P utilization as compared to commercial inorganic P salt added at the same level, the FBH produced by H ₂ S0 ₄ hydrolysis strongly affected uptake and/or incorporation of Ax in Atlantic salmon. Ax concentration in the tissues of salmon was 30-50 % increased by the FBH feed ingredient, demonstrating a very significant and pronounced biological impact on pigment utilization. Inclusion of H ₂ S0 ₄ treated fish bone meal also increased digestibility of protein, lipid, total energy and a range of minerals significantly, possibly explaining the slightly improved growth performance and accompanying lower variation in growth observed in fish fed D4 relative to D2. In conclusion, improved pigmentation was observed by feeding the H ₂ S0 ₄ treated FBH.

Conclusions:

• Fish bone meal hydrolyzed with H ₂ S0 ₄ significantly increased digestibility of protein, lipid and energy, and of macro- and micro minerals (average increase 58%), explaining the 5 % higher growth performance of salmon as compared to fish fed added inorganic P in the diet at the same level.

• The H ₂ S0 ₄ treated fish bone meal showed 30 - 50 % increased Ax levels in plasma, liver and whole body. Results suggest either improved intestinal uptake of Ax, less metabolization of digested Ax or lower Ax turnover, or any combinations thereof.

Example 3 - Production of fish bone hydrolysate

Commercially produced and separated fish bone meal from blue whiting was sifted (4.6 mm) and used as raw material for hydrolysis in a strong acid (H ₂ S0 ₄ ). The sifted fish bone meal contained 41.5% ash and was mixed with water (1 :5) in 2 x 100 L plastic drums. Either 95% H ₂ S0 ₄ (5% w/w) or 37% HC1 (5% w/w) was added to solubilize the minerals from the hydroxyapatite skeleton during continuous hydrolysis with either HC1 (4 hours) or H ₂ S0 ₄ (24 hours).

The water fraction was manually separated from the solid phase by filtering over a fine mesh cloth (70 μπι). The solid phase was washed twice in cold water to collect all water soluble compounds released during hydrolysis. The liquid phase was adjusted to pH = 3.1 with 33% KOH and thereafter stored at ambient temperature (10-15°C) overnight to precipitate solids. Floating lipid on top was removed, and the clear liquid was carefully pumped into a boiling pan, heated to 90°C (5 min) and separated over a Jesma sieve (80 μπι cloth). The liquid phase was concentrated (33°Brix) and spray dried (Niro Atomizer, Denmark) to a final dry ingredient (fish bone hydrolysate, FBH). The chemical composition and mineral content of the spray dried FBH were determined as follows:

HC1 hydrolysed fish bone hydrolysate - Protein: 47 gkg ^"1 , Lipid: 15 gkg ^"1 , ash: 630 gkg ^"1 , moisture: 87 gkg ^"1 .

H ₂ S0 ₄ hydrolysed fish bone hydrolysate - Protein: 92 gkg ^"1 , Lipid: 18 gkg ^"1 , ash: 736 gkg ^"1 , moisture: 24 gkg ^"1 .

Example 4 - Production of fish bone hydrolysate

Fish bone hydrolysate (FBH) was produced from herring by-products (head and backbone) collected at Norway Pelagic AS, Liavag, Norway). Bone was separated from fish muscle prior to freezing at -20°C. The raw bone material was thawed before hydrolysis in a process where fish bone was mixed with water (1 :5 w/w) in 100 L plastic barrels. 95% H ₂ SO ₄ (5% w/w) was added and the bone material was hydrolysed for 18 hours to dissolve the minerals from hydroxyapatite.

The water soluble fraction was separated from the solids by sieving (100 μπι), and the solid phase was washed twice with cold water to collect the water soluble material. Thereafter, pH was adjusted with 25% NH ₃ solution (pH 2.8) and the water fraction precipitated overnight. Floating lipids were removed and the clear liquid was pumped into a kitchen kettle and sterilized by heating to 90°C for 5 min before separation and removal of the solid matter by sieving (Jesma sieve, 80 μπι cloths). The water fraction was concentrated by evaporation (33°Brix), and spray dried (Niro Atomizer, Denmark) into a fine powder (20-60 μπι particle size).

The fish bone hydrolysate (FBH) produced from herring contained 10.9% total P and 9.2% soluble P. The product contained 33.9% protein, 0.8% lipid, 31.8% ash and 97.3%) dry matter. The product further contained 32.5%> (NH^SC^ originating from the sulphuric acid and NH ₃ -solution used for acid hydrolysis and pH adjustment, respectively, during processing of the fish bones. Example 5 - Production of fish bone hydrolysate

About 884 kg of frozen head and backbone of herring was bought from Norway Pelagic AS, Kaldvagen, Norway. Fish raw material was thawed (15°C) and the bones (142 kg) were separated from the muscle by physical separation in a bone separator (UF2 DX Meat Separator, BIBUN Machine Construction, CO., LTD., Hiroshima, Japan).

The herring bones were mixed with H ₂ S0 ₄ (5% w/w) and hydrolysed (18 hours) to solubilize the minerals from hydroxyapatite. The water soluble fraction was titrated to pH = 2.8 with NH ₃ solution (25% w/w), and thereafter up-concentrated by evaporation (33°Brix) and spray dried (Niro Atomizer, Denmark) to a small particulate size ingredient (20-60 μπι).

Another fish bone hydrolysate (FBH) was produced from blue whiting fish bone meal according to the same procedure, except by using KOH as the neutralizing agent.

The FBH ingredients from herring and blue whiting contained similar P levels; 10.4% total P and 9.1% soluble P.

Example 6 - Effects of fish bone hydrolysate produced using different

concentrations of H ₂ S0 ₄

The objective of the experiment was to study if fish bone hydrolysates (FBH) of cod, blue whiting and herring affect the cell's ability to handle oxidative stress, and if different processing conditions of FBH may result in different responses.

Hepatic cells harvested from salmon were used as an in vitro model system to examine the bioactive functions of FBH produced from herring, blue whiting and cod using different concentrations of sulphuric acid. The initial in vitro trial was a screening trial where we tested FBH produced from herring and blue whiting using 5 vol.% sulphuric acid for hydrolysis. In another in vitro trial with FBH produced from cod we used 2.5 vol.% sulphuric acid for hydrolysis of the bones, and measured the effects of different neutralizing agents (KOH and NH ₃ solution).

The primary goal of this study was to examine potential impacts of FBH on the cell's natural defence system against oxidative stress, in normal control cells and in cells induced for oxidative stress by using either rancid fish oils or a chemical stress inducer such as buthionine sulfoximine (BSO). Rancid fish oil contains variable oxidation products that may stimulate the cell's antioxidant defense. The three most important intracellular antioxidants in the oxidative defense system are superoxide dismutase (SOD) that converts superoxide to hydrogen peroxide, and catalase and glutathione peroxidase (GPX) which removes hydrogen peroxide. It has been reported that oxidative stress in fish can lead to inhibition of mitochondrial β-oxidation and thus reduced metabolism, induce apoptosis and alter structural compounds in muscle.

Materials and methods:

Production of fish bone hydrolysate (FBH):

FBHs were produced from different species and using different neutralizing agents (Table 9). Process conditions for acid hydrolysis of FBH produced from herring and blue whiting are described in Albrektsen et al. (Aquaculture, 2016, 459, 173). The process conditions for FBH production from cod were similar but using a lower (2.5 vol.%) concentration of acid for hydrolysis of the cod bones. FBH cod was neutralized respectively with KOH or ammonia solution (25%).

Table 9: Chemical composition and neutralizing agent used for production of

FBH

*the protein content is not corrected for non-protein N (NH ₃ - ) added to FBH

As seen from these results, there is a marked increase in mineral (P) content in respect of the FBH produced from cod using 2.5 vol.% sulphuric acid. This suggests that changing the process parameters alters the composition of the FBH by increasing the concentration of bioavailable minerals by more than 60%.

Isolation of hepatic cells for the in vitro trial:

Hepatocytes were isolated from Atlantic salmon according to the method of Dannevig and Berg (1985). After isolation, the cells were filtrated through a 100 μΜ nylon filter. The cells were washed three times in L-15 medium (Sigma- Aldrich, St. Louis, USA) and centnfuged for 2 minutes at lOOg between each wash. Finally, the cells were added to Glutamax growth media containing 10 % fetal calf serum, 1 % bicarbonate, 5mM Hepes and 1 % PenStrep (Sigma-Aldrich, St. Louis, USA). Approximately 4 x 10 ⁷ cells were seeded in 9.6 cm ² wells coated with laminin (Merck Millipore, Darmstadt, Germany). Cells were cultured at 13°C overnight and then added to the experimental growth media. The experimental growth media consisted of 2 % foetal calf serum, 1 % bicarbonate, 5mM Hepes and 1 % PenStrep, and were added either to FBH produced from blue whiting, herring or cod, and sulphate (ammonium-sulphate), or phosphorus (P). In vitro trial with FBH from herring and blue whiting:

Hepatic cells were cultured in experimental growth media for 48 hours added to four different concentrations of FBH produced from herring and blue whiting before harvesting of the cells to different analyses. The concentrations of FBH were calculated from soluble phosphorus (P) and adjusted to reflect physiological P levels in plasma. The concentrations of FBH were 300, 400, 500 and 600 μg/mL soluble P. After 24 hours, stress was induced in about 50 % of the cells by adding an oxidized fish oil (0.5 mg/mL; PV: 46, AV 26). Cells were harvested after incubation in 48 timer at 13°C, for measuring effects on the redox enzymes SOD, GPX and catalase.

In vitro trial with FBH from cod:

Cells were cultured in experimental growth media for 48 hours added to four different concentrations of FBH before harvesting of the cells to different analyses. The concentrations of FBH were calculated from soluble phosphorus (P) and adjusted to reflect physiological P levels in plasma. The concentrations of FBH were 50, 150, 450 and 600 μg/mL soluble P. After 24 hours, about 50 % of the cells were added to 2 mM buthionine sulfoximine (BSO) in order to induce a chemical stress reaction. Cells were harvested for measuring effects on the redox enzymes SOD, GPX and catalase, all measured in triplicates.

Measurement of antioxidant enzymes:

Superoxide dismutase (SOD)

SOD catalyses the reduction of superoxide to oxygen and hydrogen peroxide. SOD enzyme activity was measured using Superoxide Dismutase Assay Kit (Cayman Chemicals, Ann Arbor, MI, USA) according to the manufacturer's protocol. The enzyme reaction was measured at 450 nm in a Spectrostar Nano plate reader (BMG LABTECH GmbH, Ortenberg). Catalase

Catalase enzyme activity was measured according to a method described in

Baudhuin et al. (1964). Hydrogen peroxide produced in the peroxisomes is decomposed by catalase into oxygen and water. The reaction is stopped by addition of a saturated solution (0.45 %) of titanium oxysulfate in 2 M sulfuric acid.

Titanium oxysulfate reacts with hydrogen peroxide and produces a yellow solution of peroxy titanium sulfate. This was measured spectrophotometrically at 405 nm in a Spectrostar Nano microplate reader from BMG labtech GmbH (Ortenberg).

Glutathione peroxidase (GPX)

GPX catalyses the reduction of hydroperoxides including hydrogenperoxide.

Enzyme activity of GPX was measured using a commercial kit (Cayman Chemicals, Ann Arbor, MI, USA), following the manufacturers' protocol. The reaction was read at 340 nm in a Spectrostar Nano plate reader from BMG LABTECH GmbH (Ortenberg).

Statistical method:

The data from the in vitro trial with FBH from cod was evaluated by ANOVA- oneway analysis, at a significance level of P < 0.05.

In vitro trials with blue whiting and herring:

FBH produced from blue whiting and herring was added to cultured hepatic cells stressed with oxidised fish oil or unstressed without addition of fish oil (control), and their impacts on the oxidative stress enzymes was measured.

In vitro trials with cod:

FBH produced from cod and neutralized with either KOH or NH ₃ solution was added to cultured hepatic cells stressed with BSO or unstressed without BSO

(control). By reducing H ₂ SO ₄ concentration from 5 vol.% to 2.5 vol.%, less neutralizing agent was also required. The changes in processing conditions were reflected by a change in the ingredient composition, significantly increasing the P concentration in the FBH cod ingredients by more than 60% (Table 9). Results and Discussion:

Cell culture trial with FBH produced from blue whiting and herring:

FBH from blue whiting weakly reduced the enzyme activity of the intracellular antioxidants GPX and SOD in both stressed and non-stressed cells, and that of catalase in stress-induced cells. FBH herring gave decreased GPX activity, and showed minor effects on SOD and catalase. Blue whiting and herring bones were both hydrolysed in 5 vol.% H ₂ SO ₄ for efficient release of P, and in general, there were small differences in the activity of SOD, GPX and catalase between these two raw materials.

Cell culture trial with FBH produced from cod:

Our results show that oxidative stress induced by BSO leads to a tendency to increased activity of all three protective enzymes GPX, SOD and catalase. In stressed cells, our results further show that there is a tendency that FBH produced from cod increases the activity of the protective enzymes catalase and GPX. In cells stressed with BSO our results also show a stimulation of the important antioxidant SOD. FBH, especially FBH cod- H ₄ , leads to a slight inhibition of SOD in non- stressed cells, but the SOD activity is maintained (corresponding activity in the control cells) in cells stressed with BSO.

Conclusions:

Addition of FBH from blue whiting and herring (produced using 5 vol.% H ₂ S0 ₄ ) to Atlantic salmon liver cells led to reduced activity of the antioxidant enzymes, even in cells induced by oxidative stress, indicating that these FBH partially inhibit cell defense mechanisms against oxidative stress. In the tests with cod FBH (produced using 2.5 vol.% H ₂ S0 ₄ ), and where a powerful chemical inducer of oxidative stress was added to the cells, on the contrary a positive effect of FBH was seen with an upregulation of enzymes involved in the oxidative defense system. This is indicative of a normal and healthy response in the cells. The neutralizing agent used during production of FBH may influence the cellular responses. FBH cod-KOH showed a trend towards a better effect on SOD than FBH cod-NH ₄ .