AND OTHER ANIMALS

Government Rights This invention was made with government support under grant number NA16RG1609, project number R/SAq-04-NSI-NMAI, awarded by NOAA. The government has certain rights in the invention.

Background and Summary This invention concerns the delivery of bioactive materials, such as nutrients and vaccines, to animals.

Various microparticle types and preparation methods have been developed as potential means of nutrient delivery to suspension feeders, such as bivalve mollusks and the larvae of many species of crustacean and fish larvae, in attempts to overcome problems associated with acceptability, digestibility and rapid leaching of water-soluble nutrients.

Acceptability of microparticulate diets may be improved by manipulating the particle's intrinsic characteristics such as buoyancy, size, color, smell, hardness and taste and by manipulating culture conditions such as lighting, color of tank walls and agitation to affect movement of particles in the water column (Backhurst JR, Harker, JH (1988) "The settling rates of larval feeds." Aquacultural Engineering 1:363-366; Ostrowski AC (1989) "Effect of rearing tank background color on early survival of dolphin larvae." Prog. Fish-Cult. 51:161-163; Fernandez-Diaz C, Pascual E, Yύfera M (1994) "Feeding behavior and prey size selection of gilthead seabream, Spares aurata, larvae fed on inert and live food." Mar. Biol. 118:323-328; Kolkovski S, Koen W, Tandler A (1997) "The mode of action of Artemia in enhancing utilization of microdiet by gilthead seabream, Spams aurata, larvae." Aquaculture 155:193- 205; Cahu CL, Zambonino Infante, JL, Peres A, Quazuguel P, Le Gall MM (1998) "Algal addition in seabass, Dicentrarchus labrax, larvae rearing: effect on digestive enzymes." Aquaculture 161:109-119; Koven W, Kolkovski S, Hadas E, Gamsiz K, Tandler A (2001) "Advances in the development of microdiets for gilthead seabream, Sparus aurata: a review." Aquaculture 194:107-121).

Microbound particles are commonly used in weaning fish and crustacean larvae onto artificial diets. Various types of binders including carrageenan, alginate, gelatin and zein have

been described. Although microbound particles are relatively easy and economic to manufacture, the lack of a distinct wall surrounding the particle results in high leaching rates of water-soluble nutrients; for example, Lόpez-Alvarado et al., (Lόpez-Alvarado J, Langdon CJ, Teshima S, Kanazawa A (1994) "Effects of coating and encapsulation of crystalline amino acids on leaching in larval feeds." Aquaeulture 122:335-346), reported that 81, 85 and 91% of amino acids were lost from alginate, carrageenan and zein microbound particles (ZMP), respectively, following 2 min aqueous suspension. As much as 60% of dietary free amino acids was reported to be lost from carrageenan-bound and zein-coated, gelatin-bound diets within 1 min of suspension (Baskerville-Bridges B, Kling, LJ (2000) "Development and evaluation of microparticulate diets for early weaning of Atlantic cod, Gadus morh.ua, larvae." Aquacult. Nutr. 6:171-182).

US 6,180,614 describes the use of transfection agents, particularly liposomes (artificial microscopic vesicles consisting of an aqueous core enclosed in one or more phospholipid layers), fluorocarbon emulsions, cochleates, tubules, gold particles, biodegradable microspheres, and cationic polymers, to convey vaccines to an aquaeulture species.

However, there remains a need to develop new particle types to facilitate manufacturing and to overcome digestibility, high leaching rate, and other problems of currently available particles for delivery of water soluble materials to aquatic and marine animals, and to provide new delivery vehicles for other animals.

Brief Description of the Drawings

In the drawings:

FIG. 1 is a schematic diagram of a system used to prepare lipid spray beads. FIG. 2 is a schematic diagram of a system used to prepare complex particles.

Detailed Description Definitions

To facilitate review, the following definitions of terms and explanations of abbreviations are provided solely for the proposes of this disclosure. As used herein, the singular forms "a" or "an" or "the" include plural referents unless the context clearly indicates otherwise. Thus, for example, reference to a "nutrient" includes a

plurality of such nutrients and equivalents thereof known to those skilled in the art, unless it is clear from the context that a single nutrient is intended.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood to one of ordinary skill in the art to which this invention belongs.

Alcohol solvent: This term is used inclusively to refer to absolute alcohols, blends of alcohols, and aqueous solutions of one or more alcohols that are capable of dissolving and/or suspending a coating material.

Atomize: This refers to reduction of a liquid into small droplets. Complex microparticle (CP): This refers to a microparticle that comprises one or more lipid beads within a body of coating material. The body of coating material may additionally contain substances other than lipid beads.

Inclusion particle: This refers to a particle of a solid material that is contained within a lipid bead. An inclusion particle typically comprises a substance that is bioactive with regard to an animal, such as a nutrient, hormone, digestion aid, probiotic bacterium, therapeutic substance, vaccine, toxin, or biocide.

Liquid inclusion material: This refers to a liquid material, encapsulated as one or more droplets within a lipid bead. A liquid inclusion material may be bioactive with regard to an animal, such as a nutrient, digestion aid, or therapeutic agent, or may be bioinactive. A liquid inclusion material may comprise a bioactive agent dissolved or suspended in a liquid carrier such as water or alcohol.

Lipid bead (LB): This refers to a bead of lipid material containing one or more trapped inclusion particles and/or droplets of a liquid inclusion material. In the case of inclusion particles, there need not be a continuous wall of lipid material surrounding each trapped inclusion particle; one or more trapped particles may protrude through the surface of the lipid material.

Lipid extruded bead (LEB): This refers to a lipid bead made by passing a mixture of one or more inclusion materials and a molten lipid material through an extrusion die into absolute alcohol or an alcohol solution at a temperature below the melting point of the lipid material.

Lipid spray bead (LSB): This refers to a lipid bead made by spraying a mixture of one or more inclusion materials and a molten lipid material into an environment at a temperature below the melting point of the lipid material.

Low molecular weight: This refers to a molecular weight of less than 10,000 Daltons. Microparticle: This refers to a particle typically having a diameter less than 5 mm.

Volatile: This refers to a substance that evaporates readily at a temperature below the temperature at which lipid material melts.

Water-soluble: This refers to a substance having a solubility of at least 0.01 g/1 in water at 25 ⁰ C. Zein-bound complex microparticle: This refers to a microparticle that comprises zein and at least one lipid bead that is at least partially coated by the zein.

Zein microbound particle (ZMP) / zein-bound microparticle (ZBP): These terms refer to a microparticle comprising a zein matrix that contains nutrients and/or some other material to be delivered to an aquatic or marine animal and that does not contain a lipid bead. Complex microparticles are formed by mixing lipid beads containing a substance to be delivered to a target organism and a coating agent dissolved or suspended in an alcohol solvent. The mixture is atomized to form complex microparticles that comprise lipid beads that are at least partially coated with the coating agent. Materials and processes are selected so as to complete the formation of the complex microparticles at a temperature below the melting temperature of the lipid material of the lipid beads.

More specifically, complex microparticles may be formed by making lipid beads that contain particles and/or droplets of a substance to be delivered to an animal and by making an alcohol solution or suspension containing a coating agent. A volume of the lipid beads and a volume of the alcohol solution or suspension are combined at a temperature below the melting point of the lipid material. The resulting mixture is atomized at a temperature below the melting point of the lipid material to form complex microparticles that typically have a maximum cross-sectional dimension of 5 mm and that comprise lipid beads at least partially coated with the coating agent. Alcohol is removed at a temperature below the melting point of the lipid material to complete the formation of the complex microparticles. If made using an aqueous alcohol solution, the complex microparticles also may be dried to remove water, if desired, at a temperature below the melting point of the lipid material to produce a dry complex microparticle product.

The coating agent may be any coating material that can be dissolved or suspended in a liquid alcohol solvent and that will deposit as a solid coating upon vaporization of alcohol from the alcohol solvent at a temperature that does not melt the lipid material. The coating agent should be selected to provide a CP having a surface that makes it easy to wet and disburse the CP in water and that keeps the CPs from adhering together during storage. The coating should be a solid, water insoluble mass at the conditions of the water in which the CPs will be provided to the target organism. The coating also should be digestible by the target organism if the CP is to be delivered orally. It is sometimes also helpful for the coating agent to be selected so that the CP is palatable to the target organism. Zein is a superior coating agent because it has the properties mentioned above. In particular, zein is soluble in aqueous solutions of certain alcohols that can be easily removed during microparticle preparation. Zein is a high molecular weight (38,000 Daltons), non-toxic, edible protein with amphoteric and thermoplastic properties. It is soluble in 60-95 per cent ethanol but insoluble in water and absolute alcohol. Through the use of zein, it is possible to prepare a protective coating without the use of more toxic solvents in the complex particle preparation process. Other coating agents that sometimes may be used include such non-protein polymers as cellulose esters and starch esters. Mixtures of appropriate coating agents also may be used.

The coating agent/alcohol solvent solution or suspension may contain additional materials that are bioactive or non-bioactive. Examples include materials that determine physical properties of the CPs, such as buoyancy-adjusting clay particles, coloring agents, and/or binding facilitators such as oils (e.g., menhaden oil), triglycerides, starches (e.g., rice starch), and/or proteins (e.g., casein). Insoluble nutritional materials may be directly mixed with the coating agent/alcohol solvent solution or suspension to supplement one or more substances contained within LBs. Advantageously, any included binding facilitators are selected to serve also as nutrients for the target organism. The coating agent/alcohol solvent solution or suspension also may contain a digestion aid, such as one or more digestive enzymes.

For coating lipid beads, best results are achieved using an alcohol solvent that has a high concentration of a volatile alcohol. For example, if any of the included nutrients and/or other bioactive agents are water soluble but only slightly soluble in absolute ethanol, it is best to use a high concentration (at least 90%) ethanol solution with zein and to maintain a relatively low temperature, e.g. of 15 ⁰ C, to reduce dissolution of the nutrients and/or bioactive agents

during manufacture of the microparticles. Excellent results are achieved with a 90-95% aqueous ethanol solution to which zein is added at about 1% to about 10% weight by volume.

The alcohol solvent should contain an alcohol that is volatile, having a boiling point that is below the melting point of the lipid beads and that is below the boiling point of the coating agent. Best results are achieved with alcohols having boiling points below 20 ⁰ C, as such facilitates vaporization. The alcohol solvent should not dissolve the lipid material of the lipid beads, but should be capable of dissolving or suspending at least 1 g/1 of the coating agent. The alcohol solvent is best selected to be substantially nontoxic to the target animals, although this may not be important when the complex particles are used to deliver biocides or other toxins. Examples of suitable alcohols for inclusion in alcohol solvents are methanol, ethanol, propanol, butanol, and mixtures thereof.

The lipid beads may be formed by mixing one or more inclusion materials with a molten lipid material to form a mixture or emulsion. The lipid material should be edible and digestible by the target organism. At least 50% of the lipid material composition should comprise a triglyceride, fatty acid, wax, or mixture of such materials that is substantially insoluble in the alcohol solvent to be used for CP formation and that has a melting point of 5 to 90 ⁰ C. Such materials may include tripalmitin, methyl palmitate, stearic acid, and/or menhaden stearine. The lipid material may include one or more emulsifiers to adjust the stability and/or physical characteristics of the lipid material/inclusion material emulsion and/or the performance of resulting LB. Example emulsifiers include liquid emulsifiers such as sorbitan sesquioleate (Arlacel 83; SSO) and solid emulsifiers such as sorbitan monopalmitate (Span 40; SMP). The lipid material should not include a high concentration of phospholipids, as such are soluble in alcohol solvents. The inclusion materials) may be inclusion particles and/or droplets of one or more liquid inclusion materials. The inclusion material may be any one or more of numerous substances to be delivered to target animals, such as nutrients (e.g., amino acids, vitamins, minerals, and sugars), digestion aids such as digestive enzymes, probiotic bacteria, hormones, therapeutic substances such as antibiotics, vaccines, toxins, and biocides. Toxins or biocides may be included when the goal is to kill or to limit or stop the growth of target animals, such as within pipes of industrial installations that are prone to become clogged by the growth of aquatic or marine animals.

To form lipid spray beads, the mixture is sprayed into an environment at a temperature below the melting point of the lipid material to solidify the lipid material, thereby entrapping

one or more inclusion particles and/or droplets of liquid inclusion material and forming lipid spray beads that typically have a particle size from 2 to 500 μm. The environment can be air, water, liquid nitrogen, alcohol, or an alcohol solution. Spraying into an alcohol solvent is particularly advantageous if the alcohol solvent is appropriately selected so that it may subsequently be used for dissolving or suspending the coating agent to be used for forming complex particles. When such an alcohol solvent is used, it is not necessary to separate the lipid spray beads from the alcohol solvent before mixing with the coating agent. It is also possible to spray into a mixture comprising an alcohol solvent and a coating material, for example into a mixture comprising zein dissolved in an aqueous ethanol solution. To form lipid extruded beads, the mixture of molten lipid and inclusion material can be extruded under pressure through a single or multiple orifices into liquid nitrogen or cooled water or alcohol solvent. The lipid walls of the beads are hardened to form the individual lipid extruded beads. Extruding into an alcohol solvent is particularly advantageous if the alcohol solvent is appropriately selected so that it may subsequently be used for dissolving or suspending the coating agent to be used for forming complex particles. When such an alcohol solvent is used, it is not necessary to separate the lipid extruded beads from the alcohol solvent before mixing with the coating agent. It is also possible to extrude into a mixture comprising an alcohol solvent and a coating material, for example into a mixture comprising zein dissolved in an aqueous ethanol solution. Because LEBs are manufactured by an extrusion technique, they can be larger than LSBs, and typically are in a size range of 2 μm to 2 mm.

Complex microparticles described herein may be disbursed directly into the water environment of aquatic or marine animals to which inclusion material is to be supplied. Or such complex microparticles may be incorporated into delivery systems for larger animals, such as orally ingestible feed pellets, which are of a larger size than CPs and which may contain more than one CP. Complex microparticles also may be administered by forming an aqueous suspension of the complex microparticles and directly injecting the suspension into animals to which inclusion material is to be supplied.

Example 1 A first example concerns a method that results in the production of LSBs containing particulate riboflavin (vitamin B2) as an inclusion particle material. The performance of LSBs made with a lipid material primarily comprising a low-melting point wax (methyl palmitate

(MP); melting point 27-28 ⁰ C) was determined to evaluate their potential usefulness in delivering water-soluble nutrients to fish larvae.

Addition of emulsifiers can greatly affect the stability and physical characteristics of lipid/core emulsions as well as the performance of resulting LSB; therefore, the effects of additions of various emulsifiers at different concentrations to LSB were determined. Initially, the leaching characteristics of LSB containing a liquid emulsifier, sorbitan sesquioleate (Arlacel 83; SSO) and a solid emulsifier, sorbitan monopalmitate (Span 40; SMP) are described. The effects of additions of various concentrations of SMP on the retention of riboflavin by LSB composed of MP were then determined. Finally, retention of riboflavin by LSB composed of MP+10% SMP+ 10% ethyl cellulose (EC) was described and compared with that of LSB composed of 60% tripalmitin + 40% menhaden oil (60% T+ 40% M). EC is a non-gelling polysaccharide with good film forming, coating and binding characteristics and when mixed with low melting point lipids may result in a matrix with good stability and digestibility properties. For core materials encapsulated within microparticles to be assimilated by fish larvae, the core must be released either by enzymatic digestion or physical breakdown of the particle or by simple diffusion of the core material from the particle into the gut lumen. Feeding experiments were carried out using first-feeding larvae of zebrafish, Brachydanio rerio and glowlight tetra, Hemigrammus erythrozonus to determine whether riboflavin was released from LSB into the gut lumens of these species. LSB were coated with zein, forming a complex particle to reduce particle hydrophobicity and increase acceptability.

MATERIALS AND METHODS LSB preparation LSB were prepared by a modification of the method described by Buchal and Langdon

(Buchal M.A.; Langdon CJ. (1998) "Evaluation of lipid spray beads for the delivery of water- soluble materials to a marine suspension-feeder, the Manila clam Tapes philippinarum (Deshayes 1853)" Aquaculture Nutrition 4:263-274), incorporated herein by reference. A total of 10 g of lipid material were placed in a water bath (65 ⁰ C). The ratio of emulsifier to lipid was varied depending on the formulation to be tested and ranged between 0-25 % w/w of lipid. Two grams of finely ground riboflavin powder (<10 μm particles; McCrone micronizing mill, McCrone Scientific Ltd.) were mixed with 1O g of the molten (60-65 ⁰ C) lipid/emulsifier

mixture and sonicated at half power (B. Braun Labsonic L, B. Braun Biotech Inc.). The suspension of particulate core material in lipid/emulsifier mixture was then poured into a heated (65 ⁰ C) aluminum container. The temperature of the container was controlled by a temperature controller (model CN9000A, Omega Engineering Inc.) in combination with a heating coil and a thermocouple. The core/lipid suspension was then sprayed through an atomizing nozzle (1/4 JBCJ, Spraying Systems Co.) supplied with pressurized nitrogen into a steel conical chamber that was chilled to -2O ⁰ C using liquid nitrogen. Following atomization, lipid droplets solidified in the steel chamber and the resulting beads were collected, freeze-dried and stored under refrigeration until use. LSB containing EC were prepared by first dissolving Ig EC in 10 ml of methylene dichloride. Two grams of finely ground crystalline riboflavin were then added to the EC solution and the suspension was sonicated for 30 seconds at half power to provide thorough mixing. The temperature of the riboflavin/EC suspension was then raised to 65 ⁰ C and ten grams of molten lipid (65 ⁰ C) were added gradually. The resulting molten lipid/EC/riboflavin mixture was then sonicated for 30 sec at half power and placed in a water bath at 65 ⁰ C until excess methylene dichloride was removed by bubbling with N ₂ gas. After removal of methylene dichloride, the suspension was sprayed to form LSB, as described above. LSB used in feeding experiments were fed to fish larvae in the form of complex particles in order to reduce their hydrophobicity and improve their acceptability. CP with a size range of 45-106 μm were stored for use in feeding experiments.

Measures of LSB Performance Inclusion efficiency ( ^" JE)

Inclusion efficiency (JE) was expressed as the percentage of core material originally present in the lipid mixture that was successfully encapsulated. IE was determined as follows: triplicate samples of 25 mg from each batch of LSB were first dissolved in 5 ml chloroform and the core material extracted by successive additions of 25 ml aliquots of distilled water. The aqueous supernatant was removed and the extraction was repeated until no absorbance reading was obtained in the supernatant compared to a blank. Aqueous extractions were then combined and absorbance was compared to that of a blank at the core material's maximum absorbance peak (riboflavin 267nm). Absorbance was converted to core concentration using regression equations derived from standard curves. Triplicate subsamples of 1 ml of the chloroform phase

was removed from the capsule extraction and transferred to dry, tarred aluminum weighing boats. The chloroform was removed by heating for 24h at 5O ⁰ C. Lipid contents of the extracted microparticles were calculated based upon the amount of lipid recovered after evaporation of chloroform. IE was calculated as: IE= [(w/w ratio of riboflavin to lipid in LSB) / (w/w ratio of riboflavin to lipid in the pre-spray formula)] X 100

Encapsulation efficiency (EE)

Encapsulation efficiencies of LSB were expressed as the percentage of core material weight to lipid weight after preparation of LSB (Langdon C. J. ; Buchal M. A. (1998) "Comparison of lipid- walled microcapsules and lipid spray beads for the delivery of water-soluble, low-molecular- weight materials to aquatic animals" Aquaculture Nutrition 4:275-284).

Retention efficiency (RE)

RE of LSB were expressed as the percentage of initial core material retained after suspension of LSB in water. Samples of suspended LSB were taken over a given time period to determine changes in retention of riboflavin over time. Triplicate samples of 10 mg LSB for each time interval were added to glass vials containing 10 ml water. The concentration of LSB added to vials was chosen so that the quantity of core material contained within LSB could potentially completely dissolve in the water volume. The vials were sealed and placed on an orbital rotator and the experiments were carried out in a temperature-controlled incubator at 25±1°C, in darkness. At each time interval, the contents of each triplicate set of vials were filtered onto a glass-fiber filter (Whatman 934-AH GFIC; Whatman) using a separatory funnel attached to a vacuum pump and the filtrate was collected in a test tube to determine the concentration of leached riboflavin as described above. Retention efficiency (RE) was then calculated as:

RE = {[(initial encapsulated riboflavin weight)-(riboflavin weight in filtrate)] / (initial encapsulated riboflavin weight) }xl 00

Delivery efficiency (DE) DE were calculated to allow comparisons of LSB preparations in terms of the amount of core material that could be potentially delivered to a larva. DE was defined as the amount of

core material remaining in microparticles (mg core material g ^'1 lipid) after a given period of time in which LSB were suspended in water.

Breakdown of LSB and release of riboflavin by larvae Feeding experiments were carried out to determine if fish larvae could ingest and break down LSB, thereby releasing riboflavin into the digestive system. Broodstock of glowlight tetra and zebrafish were spawned in aquaria and the resulting first-feeding larvae were transferred to 11 Imhoff settling cones receiving carbon-filtered, up-flowing water (26±1°C) from the base, to promote suspension of CP (Onal U, Langdon CJ (2000) "Characterization of two microparticle types for delivery of food to altriciai fish larvae." Aquacult. Nutr. 6:159-170). Fish larvae were observed to contain CP within Ih after the start of the experiment but were allowed to continue feeding until they started defecating 4-5h later. Sampled larvae were then examined for evidence of broken or mis-shapen microparticles and released core material under an epifluorescent microscope (Zeiss reflected light microscope; Carl Zeiss Inc.) at xlOO magnification. The larva's digestive system was photographed using a photo-microscopic camera (Carl Zeiss MC63). Additional background lighting, without obscuring the fluorescence emitted by CP, reduced exposure times to within a couple of seconds resulting in sharp images of microparticles together with the overall outline of larvae. Released riboflavin could be observed in the digestive tract of larvae as a diffuse yellow fluorescence. In addition, change in the shape of CP in the digestive tract and feces indicated breakdown of the microparticles.

Results

Inclusion efficiencies (IE) and encapsulation efficiencies (EE) of riboflavin by LSB

The mean IE, EE and associated standard deviations for LSB formulations are given in Table 1.1:

Exρ # Lipid matrix composition IE w/w (%) EE w/w (%)

1 MP alone 93.98±0.88 15.69±0.62

MP+5% SSO 97.58±0.96 16.30±0.73

MP+5% SMP 97.5±0.89 16.28+-0.69

2 MP alone 92.86±0.74 15.34+-0.55

MP+1% SMP 98.79±1.04 16.50±0.53

MP+5% SMP 98.13±0.56 16.39±0.33

MP+10% SMP 98.70+-0.61 16.48±0.68

MP+25% SMP 97.38+-0.58 16.26±0.41

3 MP+10% SMP 98.31±0.46 16.98±0.29

MP+25% SMP 97.39±0.61 17.89±0.46

60%T+40%M 97.50±0.96 16.28±0.56

MP+10%SMP+l 0%EC 91.25±0.71 15.24±0.43

Table 1.1: Inclusion and encapsulation efficiencies (mg core 100 mg ^'1 lipid) ± 1 SD of lipid spray beads (LSB) containing riboflavin. MP: methyl palmitate, SSO: sorbitan sesquioleate, SMP: sorbitan monopalmiate and EC: ethyl cellulose.

Concentrations of 167 mg particulate riboflavin g ^"1 lipid present in the initial spray mixture were not equally encapsulated in LSB. The IE of LSB composed of methyl palmitate (MP) alone and MP+ 10% ethyl cellulose (EC) were significantly less than those containing either sorbitan sesquioleate (SSO) or sorbitan monopalmitate (SMP; p<0.005; Tukey's HSD). In addition, the IE and EE of LSB composed of MP+10% EC were significantly less than LSB composed of MP only (p<0.005; Tukey's HSD). IE and EE of LSB containing either SSO or SMP did not differ (p>0.05; Tukey's HSD). Overall, LSB containing either SSO or SMP had a mean IE of 97.94% and a mean EE of 16.36 mg riboflavin 100 mg ^'1 lipid.

Experiment 1: Effect of liquid vs. solid emulsifier on retention and delivery of riboflavin Suspension time and treatment had a significant effect on retention of riboflavin (p<0.001 for main effects; df=62; two-way ANOVA). Suspension time x treatment interaction effect also had a significant effect on retention of riboflavin such that percent retention of riboflavin at each time interval depended on lipid composition (pO.OOl; two-way ANOVA). There were significant differences among treatments after only 2 min of aqueous suspension. LSB without any emulsifier had a RE of 95.5% which was significantly higher than those of the

other two treatments (pO.OOl; Tukey's HSD). Furthermore, at 2 min, LSB composed of MP + 5% SMP retained a significantly higher percentage of riboflavin (92.3%) compared to 91.6% for LSB composed of MP + 5% SSO. After 60 min, LSB containing no emulsifϊer had a significantly higher (pO.OOl; Tukey's HSD) RE (80% of the initial concentration) compared with those of the other two LSB types.

Regression analysis indicated that there was a significant relationship between the fraction of riboflavin retained and the duration that LSB were suspended in water for each treatment. Table 1.2 summarizes the regression equations fitted to the observed data, R ² , the time to 75% RE and the associated standard deviations for each treatment. A straight line relationship was obtained for all treatments when percent retention rates were plotted against log time:

Exp # LSB Composition Regression equation T75 (min)

1 MP alone -5.82 x (/nthne) + 101.6 97.3 95.9±0.37

MP+5% SSO -6.53 x (/ntime) + 97.10 98.6 29.5±0.39

MP+5% SMP -6.99 x (/ntime)+ 100.25 97.3 37.1±0.57

2 MP alone -5.42 x (Mme) + 99.51 97.8 92.0i0.39

MP+1% SMP -5.73 x (toime) + 99.81 95.7 75.9±0.56 MP+5% SMP -6.86 x (/ntime) + 99.86 97.3 37.5±0.57 MP+10% SMP -0.71 x (Intimef- 0.73 x (/ntime) + 96.43 99.7 148.9±0.33 MP+25% SMP -0.99 x (time) ¹ ' ² + 97.65 97.3 523.4±0.29

MP+10% SMP -0.71 x (/ntime) ² - 0.73 x (/ntime) + 96.43 99.4 148.9±0.33

MP+25% SMP -0.99 x (time) ¹ ' ² + 97.65 96.7 523.4±0.28

60% T+40% M -1.05 x (time) ¹ ' ² + 100.38 97.7 584.3±0.58

MP+10%SMP+10%EC -0.031 x (time) + 97.38 99.1 722+.0.45

Table 1.2: Regression equations fitted to describe change in riboflavin retention efficiency (RE) over a 24h period, associated R ² and T75 (time to 75% retention ±1 SE) values for each LSB composition. MP: methyl palmitate, SSO: sorbitan sesquioleate, SMP: sorbitan monopalmitate and EC: ethyl cellulose.

DE of LSB followed a different pattern. LSB containing no emulsifier had significantly higher DE throughout the experiment except at 2 min and 24 h. At 2 min of suspension, DE did not differ among treatments (p>0.05; Tukey's HSD). At min 60, LSB containing no emulsifier

had a DE of 1.26 mg riboflavin 10 mg ^"1 lipid which was significantly higher (p<0.05; Tukey's HSD) than the DE of LSB composed of MP + 5% SSO or MP+5% SMP (1.12 vs. 1.16 mg riboflavin 10 mg ^"1 , respectively). At the end of 24h, LSB composed of MP alone had a DE of 0.83 mg riboflavin 10 mg ^"1 which was significantly higher than those of other treatments (p<0.05; Tukey's HSD).

Experiment 2: Effect of different concentrations of SMP on retention and delivery of riboflavin Suspension time and concentration of SMP in LSB had significant effects on retention of riboflavin (p<0.001 for main effects; df=89; two-way ANOVA). Suspension time x treatment interaction effect also had a significant effect on retention of riboflavin such that percent retention of riboflavin at each time interval depended on concentration of SMP (p<0.001; two-way ANOVA). Riboflavin was retained better in LSB when the concentration of SMP was greater than >10% (w/w). LSB composed of MP+25% SMP had significantly higher retention efficiencies compared to the other treatments at all time intervals tested except at 24h (Fig 2.3). Retention of riboflavin after 2 min of suspension by LSB composed of MP+25% SMP was 97.93 % which was significantly higher than for all other treatments (p<0.005; Tukey's HSD). After 60 min of aqueous suspension, LSB made up with 25% SMP still contained 89.33% of the initial riboflavin concentration which was significantly higher than for other treatments (p<0.005; Tukey's HSD). Since the release patterns of riboflavin from LSB were not identical, different models were used to describe release kinetics (see Table 1.2). Best linear relationships for LSB composed of MP+0, 1 and 5% SMP were obtained using a semilogarithmic (log time vs. percent retention) plot. The release of riboflavin from LSB composed of MP+10% SMP showed a biphasic pattern and was best described by a bi-exponential equation. On the other hand, the release of riboflavin from LSB composed of MP+25% SMP was best described by the Higuchi membrane diffusion-controlled model.

DE followed a similar pattern to RE with highest concentrations of riboflavin delivered by LSB composed of MP+25%SMP followed by LSB composed of MP+10%SMP (Fig 2.4). At min 2 and 20, LSB composed of MP+10% SMP and MP+ 25% SMP had similar DE which were significantly higher than LSB composed of either MP alone, MP+1% SMP or MP+ 5% SMP (p<0.005; Tukey's HSD). After 60 min of suspension, LSB composed of MP+25% SMP had significantly higher DE than the other treatments (p<0.005; Tukey's HSD). At the end of

24h, LSB composed of MP+25% SMP and MP+1% SMP had the highest DE. LSB composed ofMP+ 25% SMP had the highest retention efficiency as indicated by its T75 of 342 min, further improving the retention of riboflavin by 193 min compared to that of LSB composed of MP + 10% SMP.

Experiment 3: Effect of ethyl cellulose (EC) on retention and delivery of riboflavin

The retention of LSB composed of MP with either 10% w/w SMP or 25% w/w SMP or 10% SMP+10% w/w ethyl cellulose (EC) were compared to determine if addition of EC to MP further improved RE for riboflavin. In addition, the performance of LSB composed of 60% tripahnitin+ 40% menhaden oil (60% T+40% M) was included because this LSB type had been studied previously (Langdon and Buchal, 1998). Suspension time, treatment and time x treatment interaction all had significant effects on retention of riboflavin (p<0.001 for all effects; df=71; two-way ANOVA).

After 2 min of aqueous suspension, LSB composed of MP+10% SMP+10% EC and 60% T+ 40% M had similar RE (98.86 vs. 99.38%; p>0.05; Tukey's HSD) and they were significantly higher than those of LSB composed of MP+25% SMP and MP+10% SMP (97.81 and 95.38%, respectively). Similarly, 60 min RE of LSB composed of MP+10% SMP+10% EC and 60% T+ 40% M did not differ from each other (94.17 vs. 93.95%; p>0.05; Tukey's HSD) and they were significantly higher (p<0.05) than those of LSB composed of MP+25% SMP and MP+10% SMP (88.28 and 81.61%, respectively). Regression analysis indicated that for each treatment, there was a significant relationship between the fraction of riboflavin retained and the duration that LSB were suspended in water. The time to 75% retention efficiency, R ² and the associated standard deviations for each treatment. Release from MP+10% SMP+10% EC followed a zero-order model. The release pattern of LSB composed of MP+25% SMP and 60% T+40% M was proportional to square root of time indicating that the overall release mechanism followed the Higuchi kinetic model.

DE of LSB followed a different pattern compared to RE. LSB composed of 60% T+40% M delivered the highest riboflavin concentrations throughout the experimental period. MP+10% SMP+10% EC showed significantly less DE than other treatments at min 2 and 20, however, there were no significant differences between LSB composed of MP+10% SMP+10% EC and MP+ 25% SMP after 60 min (p<0.005; Tukey's HSD). A comparison of T75 values

showed that LSB composed of MP+10% SMP+10% EC prolonged the release of riboflavin up to 722 min.

Breakdown of LSB and release of riboflavin by fish larvae First-feeding larvae of glowlight tetra and zebrafish ingested LSB coated with zein as complex particles and their gut contents were easily observed through transparent tissue surrounding the larval gut. Ingested LSB with different wall compositions contained intensely fluorescent riboflavin and no differences in the intensity of florescence among different LSB formulations could be detected. LSB within complex particles were visible as compressed pellets in the gut and released riboflavin was obvious in all cases as indicated by intense diffuse fluorescence observed throughout the lumen. Release of dissolved riboflavin from the gut continued even after defecation of particulate material, giving rise to a plume of riboflavin originating from the larva's anus.

Example 2

LSB with different lipid compositions containing either glycine or tyrosine were prepared and their performances were compared using short-term experiments. Determination of retention patterns of individual amino acids is essential for identification of potential limitations of LSB with regard to the delivery of different amino acids to fish larvae. Retention patterns of LSB containing different physical forms of core materials (particulate or aqueous core) were also compared as different physical forms of core materials can be used to affect core retention in microencapsulated delivery systems (Buchal and Langdon, 1998). Short-term leaching experiments allow for quick identification of promising lipid matrixes, leading to efficient development of LSB for delivery of water-soluble nutrients. In order for core materials encapsulated within microparticles to be assimilated by fish larvae, the core must be released either by enzymatic digestion or physical breakdown of the particle or by simple diffusion of the core material from the particle into the gut lumen. Feeding experiments were carried out using 3 day-old clownfish larvae, Amphiprion percula, in order to determine whether LSB were broken down in the gut lumen of this species. For this purpose, LSB containing a non-toxic dye (Poly-red 478) were coated with zein forming a complex particle to reduce particle hyrophobiciry and increase acceptability. Complex particles have

been used previously in delivering nutrients to penaeid shrimp larvae and larvae of hybrid striped bass.

MATERIALS AND METHODS

Experimental approach

Initially, glycine was encapsulated within various LSB as it represents a low molecular weight (75.07 Daltons), highly water-soluble material. Glycine is very hydrophilic and its solubility in water is very high (250g/l at 25 ⁰ C) contributing to its fast leaching from inicroparticles in water. LSB composed of 80% MP+10% SMP+10% EC were very promising in delivering riboflavin and, therefore, retention and delivery efficiencies of glycine by LSB composed of 80% MP+10% SMP+10% EC were compared to those of LSB composed of menhaden stearine (MS).

Menhaden stearine has the consistency of peanut butter at room temperature, and was substituted with spermaceti (Sp) in order to increase the melting point of the lipid matrix and its physical stability. Sp is a marine wax with a melting point of 55 ⁰ C, and LSB made from 75% MS+25% Sp (w/w) were harder at room temperature. In order to increase digestibility by fish larvae, MS was also substituted with coconut oil (CO) (50% w/w; melting point 28 ⁰ C) that resulted in softer LSB. After identification of the best lipid matrix, retention and delivery effciencies of LSB containing either particulate glycine or aqueous glycine solution were compared.

Finally, the retention of LSB composed of 100% MS containing either 7% particulate glycine or 7% particulate tyrosine were compared to determine if they differed in leaching rates. Tyrosine is a low-molecular weight (181.19 Daltons), poorly water-soluble amino acid (0.453g/l at 25 ⁰ C), exhibiting a different leaching pattern compared to glycine. In addition, the performance of LSB composed of 100% MS containing 23% tyrosine was included in order to determine if higher tyrosine concentrations resulted in higher leaching rates.

LSB preparation

LSB were prepared by a modified melt-spray method using an apparatus of the type shown in FIG. 1. In general, the apparatus of FIG. 1 is used by adding a molten lipid/core

mixture in the form of an emulsion of aqueous core material and molten lipid or a suspension of dry particulate core material in molten lipid to a heated metal container. The metal container is heated using a heating coil (not shown). The lipid core mixture is atomized using pressurized nitrogen (N2) into a steel conical chamber. The apparatus shown in FIG. 1 includes a sonicator A, a pressurized inlet N2 for continuous flow of a molten lipid/core mixture B 1 , a pressurized inlet N2 for atomization B2, a metal container C, metal tubing D, a valve E, a nozzle F, a steel container G, a collection cup H, and an inlet I for liquid nitrogen vapor.

Initially, lipids were melted in a water bath at 7O ⁰ C. In order to obtain suspensions of lipid and amino acid, SMP (sorbitan monopalmitate) was incorporated into lipid matrix formulations containing either MP or trilaurine (T). No SMP was added to lipid formulations containing menhaden stearine (MS) as suspensions of MS and glycine were very stable. The ratio of emulsifler to lipid was varied depending on the formulation to be tested and ranged between 0-10 % w/w of lipid. Two grams of finely ground glycine or tyrosine powder (<10 μm particles; McCrone micronizing mill, McCrone Scientific Ltd.) were mixed with 8 g of a molten (60-65° C) lipid/emulsifier mixture and sonicated (B. Braun Labsonic L, B. Braun Biotech Inc.). For encapsulating aqueous solutions of amino acids, glycine was dissolved in water and aqueous core materials were added to molten lipid followed by sonication. The suspension of core material and lipid/emulsifier mixture was then poured into a heated (65 ⁰ C) aluminum container. The temperature of the container was controlled by a temperature controller (model CN9000A, Omega Engineering Inc.) in combination with a heating coil and a thermocouple. In order to prevent settlement of core material, the suspension was continuously sonicated (Braun- sonic 2000, B. Braun Instruments,) during spraying with a sonicator. The core/lipid suspension was then sprayed through an atomizing nozzle (JBCJ, Spraying systems) supplied with pressurized nitrogen, into a steel conical chamber chilled to -2O ⁰ C using liquid nitrogen using the apparatus shown in FIG. 1. Following atomization, lipid droplets solidified in the steel chamber and the resulting beads were collected, freeze-dried and stored under refrigeration until use.

LSB containing ethyl cellulose (EC) were prepared by first dissolving Ig EC in 10 ml of methylene dichloride. Two grams of finely ground glycine were then added to the EC solution and the suspension was sonicated for 30 seconds at half power to provide through mixing. The temperature of the ribofiavin/EC suspension was then raised to 6O ⁰ C and ten grams of molten lipid (6O ⁰ C) was added gradually. The resulting molten lipid/EC/amino acid

mixture was then sonicated for 30 sec at half power and placed in a water bath at 6O ⁰ C until excess methylene dichloride was removed by bubbling. After complete removal of methylene dichloride, the suspension was sprayed to form LSB, as described above.

Measures of LSB Performance

Amino acid analysis

Amino acid concentrations were determined using ninhydrin reagent (Sigma Chemicals, Inc). One ml samples containing glycine or tyrosine were acidified by 0.05% acetic acid. One ml of ninhdyrin reagent was added to each sample and the samples were placed in a boiling water bath (100 ⁰ C) for 10 min for color development. After 10 min, the samples were transferred to a cold water bath and 5 ml of 95% ethanol was added to each sample to stabilize color development. Absorbance was determined spectrophotometrically at 570nm. Absorbance was converted to core concentration using regression equations derived from standard curves.

Breakdown of complex particles (CPI by clownfish larvae

LSB composed of various lipid formulations containing 10% w/w solution of Poly-red (PoIy-R 478; Sigma Chemical Inc.) were prepared and embedded within a mixture of zein/dietary ingredients forming CP, in order to reduce their hydrophobicity and improve their acceptability. CP were prepared by coating LSB with a zein-bound dietary mixture. CP with a size range of 45-106 μm were stored for use in feeding experiments. CP were fed to three day- old clownfish, Amphiprion percula, larvae in order to determine whether CP were broken down in the larval gut. Larvae were obtained from three pairs of wild-caught broodstock kept in 50 1 fiberglass aquaria. An indoor recirculating system, with a total volume of 12001, provided a constant temperature year around of 26±1°C, and 30±2 ppt salinity, with no detectable ammonia and nitrite. A 14L/10D photoperiod was maintained by fluorescent lighting. Broodstock were fed a variety of feeds, 3-4 times a day. Broodstock fish spawned every 12-20 days, with approximately 250-500 eggs per spawn. Eggs were transferred to 121, cylindrical larval rearing tanks with black walls, 3-4h prior to hatching and provided with gentle aeration. After hatching, clownfish larvae were maintained in a static system, at 24-25 ⁰ C, and a salinity of 30 ppt. Larvae were fed on rotifers, Brachionus plicatilis, at a density of 5/ml, in combination with Tetraselmis chui, at 5,000-10,000 cells/ml for two days. Prior to the experiment, 20 larvae were

randomly selected and placed in a 1 1 glass beaker and starved overnight (14-16h). Fifty mg of diet were then fed to larvae with gentle aeration that helped particles stay in suspension. Larvae were allowed to feed for 2-4h and individuals were sampled to examine fecal strands and gut contents under microscope (Nikon Optiphot-2) and photographed (Nikon N6000 camera).

Results

Inclusion efficiency (IE) and encapsulation efficiency (EE) of glycine and tyrosine by LSB

The mean EE, EE and associated standard deviations for LSB formulations are given in Table 2.1:

Exp # LSB composition Core Material IE w/w (%) EE w/w (%)

1 90% T+10°/o SMP part glycine 80.97±0.65 12.01±0.38

80% MP+10% SMP+10% EC part glycine 81.38±0.84 12.04±0.41

75% MS+25% Sp part glycine 86.26±0.54 12.77±0.62

2 50% MS+50% CO part glycine 85.90±1.12 16.98±0.69 ^•

75% MS+25% Sp part glycine 89.45±0.84 17.89±0.43

100% MS part glycine 86.75±0.79 17.35±0.68

3 75% MS+25% Sp part glycine 88.97±0.66 6.21±0.45

75% MS+25% Sp aq glycine 89.56±0.45 4.95±0.38

4 100% MS part glycine (7%) 90.54±1.01 6.32±0.56

100% MS part tyrosine (7%) 89.25±0.63 6.23±0.51

100% MS part tyrosine (23%) 91.08±0.60 21.02±0.83

Table 2.1: Inclusion (IE) and encapsulation efficiencies (EE; see materials and methods) ± 1 SD of lipid spray beads (LSB) containing glycine. MP: methyl palmitate,

MS: menhaden stearine, Sp: spermaceti, CO: coconut oil, T: trilaurin, SMP: sorbitan monopalmitate and EC: ethyl cellulose, part gly: particulate glycine, aq gly: aqueous glycine, tyr: tyrosine.

Overall, IE ranged between 80-91% with significant differences due to lipid matrix composition. IE of LSB composed of either trilaurin or MP were significantly less for glycine than that of LSB composed of MS (p<0.001; t-test). EE values indicated that up to 21% tyrosine could be incorporated within LSB.

Experiments 1 and 2: Effect of wall composition on retention of glycine

Suspension time and treatment had significant effects on retention of glycine (pO.OOl; df=35; two-way ANOVA). The time x treatment interaction also had a significant effect on glycine retention such that the rate of loss of glycine depended on lipid matrix composition (pO.OOl). LSB composed of 90% trilaurin+10% SMP (90% T+10% SMP) had significantly higher RE compared to those of other treatments throughout the experimental period. There were significant differences among treatments after only 2 min of aqueous suspension. LSB composed of 90% T+10% SMP had 83.2% RE which was significantly higher than those of the other two treatments. LSB composed of 75% MS+25% Sp had significantly higher RE compared to LSB composed of 80% MP +10% EC+10% SMP until 10 min, with no significant differences after 10 min. At min 60, there were no significant differences between the RE of LSB composed of 80% MP+10% EC+10% SMP and 75% MS+25% Sp (14.12 vs. 13.72%; p=0.769) which were both significantly less (p<0.05) than that of LSB composed of 90%T+10%SMP.

Regression analysis indicated that there was a significant relationship between the fraction of glycine retained and the duration that LSB were suspended in water for each treatment. Table 2.2 summarizes the regression equations fitted to the observed data, R ² , the time to 50% retention efficiency (T50) and the associated standard errors for each treatment.

Exp LSB Composition Regression equation R ² T50 (min)

#

1 90%T+10% SMP -0.065 x (Zntime)''- 0.128 x (Zntime) + 4.58 98.9 10.76±0.40

80%SMP+10%SMP+10%EC -0.438 x (Zntime) + 4.44 98.6 3.34±0.39

75% MS+25% Sp -0.463 x (Zntime)+ 4.56 98.5 4.01±0.57

2 50% MS+50% CO -0.108 x (Intimef- 0.094 x (Zntime) + 4.52 95.4 9.97±0.33

75% MS+25% Sp -0.374 x (Zntime) + 4.58 98.6 5.59±0.54

100% MS -0.237 x (Mme)+ 4.59 97.7 17.18±0.56

3 75% MS+25% Sp -0.492 x (Zntime) + 4.59 98.5 3.98±0.35

75% MS+25% -0.022 x (Zntime) ² - 0.047 x (Zntime) + 4.56 97.5 90.38+-0.51

4 100% MS -0.272 x (Zntime)+ 4.57 99.7 11.18±0.32

100% MS -0.014 x (Zntime) ² - 0.018 x (Zntime) + 4.60 98.5 2280±0.46

100% MS -0.014 x (Zntime) ² - 0.015 x (Zntime) + 4.60 98.3 1935±0.58

Table 2.2: Regression equations fitted to describe change in glycine and tyrosine retention efficiency over a Ih period of aqueous suspension, associated R ² and T50 (time to 50% retention ±ISE) values for each LSB composition. MP: methyl palmitate,

MS: menhaden stearine, S: spermaceti, T: trilaurin, SMP: sorbitan monopalmitate and EC: ethyl cellulose.

While a linear regression equation best described the leaching pattern of LSB composed of 80% MP+10% EC+10% SMP and 75% MS+25% Sp on a logarithmic scale, a polynomial equation best described the leaching pattern of LSB composed of 90% T+ 10% SMP. T50 values showed that leakage rates of glycine from all treatments were rapid.

Although the T50 value for LSB composed of 90% T+10% SMP was greater than those of other treatments, trilaurin has a melting point of 46 ⁰ C and was, therefore, less likely to be broken down by early fish larvae. In contrast, menhaden stearine has a melting point of 35 ⁰ C and LSB composed of 75% MS+25% Sp performed better than LSB made up of 80% MP+10% EC+10% SMP. Based on the results of this experiment and the potential for breakdown by fish larvae, LSB composed of MS were further evaluated in this study.

LSB composed of trilaurin delivered significantly higher glycine concentrations throughout the experimental period. After 1 h of aqueous suspension, LSB composed of 90% T+10% SMP delivered 0.236 mg glycine lOmg ^"1 lipid compared to 0.175 and 0.170 mg glycine lOmg ^"1 lipid delivered by LSB composed of 75% MS+25% Sp and 80% MP+10% SMP+10% EC, respectively (pθ.05; Tukey's HSD).

Experiment 2

Suspension time and wall composition of LSB had significant effects on retention of glycine (p<0.001; df=35; two-way ANOVA). Time x treatment interaction effect also had a significant effect on glycine retention such that the rate of loss of glycine depended on wall composition (pO.OOl). There were significant differences among treatments after 2 min suspension in water. LSB composed of 100% MS had significantly higher RE compared to those of other treatments (pO.OOl). After 60 min of aqueous suspension, LSB composed of 100% MS had a significantly higher RE (54%; p<0.001) compared to those of LSB composed of either 75% MS+25% Sp (34%) or 50% MS+50% CO (19%).

Regression analysis indicated that for each treatment, there was a significant relationship between the fraction of glycine retained and the duration that LSB were suspended in water. Table 2.2 summarizes the regression equations fitted to the observed data, the time to 50% retention efficiency (T50), R ² and the associated standard errors for each treatment. LSB composed of 100% MS had a T50 of 17.8 min compared to T50 values of 9.97 and 5.59 min for LSB composed of 50% MS+50% CO and 75% MS+25% Sp, respectively.

There were no significant differences between the DE of LSB composed of 100% MS and 50% MS+50% CO after 2 min of suspension (p>0.05; Tukey's HSD). However, LSB composed of 100% MS had significantly higher DE during the rest of the experimental period. At the end of Ih, LSB composed of 100% MS delivered 0.523 mg glycine lOmg ¹ lipid which was 0.212 and 0.305 mg more than those of LSB composed of 75% MS+25% Sp and 50% MS+50% CO, respectively.

Experiment 3: Effect of liquid vs. solid core Suspension time and treatment had a significant effect on retention of glycine (p<0.001 ; df=29; two-way ANOVA). Time x treatment interaction effect also had a significant effect on glycine retention such that the rate of loss of glycine depended on the nature of the core material (pO.OOl). There were significant differences in retention among treatments after only 2 min of aqueous suspension, with LSB containing aqueous glycine showing an RE of 86.41%, which was significantly higher than that of LSB containing particulate glycine. A similar trend was

observed throughout the experiment with significantly higher RE for LSB composed of 75% MS+25% Sp containing aqueous glycine.

Table 2.2 summarizes the regression equations fitted to retention of glycine over time, the time to 50% RE (T50), R ² and the associated standard errors for each treatment. While a linear regression equation best described the leaching pattern of LSB composed of 75% MS+25% Sp with particulate glycine on a logarithmic scale, a polynomial equation best described the leaching pattern of LSB with an aqueous core. LSB composed of 75% MS+25% Sp with an aqueous core had a T50 value of 90.38 min that was promising for prolonging glycine release. DE of LSB followed a similar pattern to RE with higher concentrations of glycine delivered by LSB containing aqueous glycine throughout the experimental period. At the end of 2 min, LSB composed of 75% MS+25% Sp with aqueous core had a DE of 0.428 mg glycine lOmg ^"1 lipid which was significantly higher than that of LSB containing particulate glycine (p<0.05; Tukey's HSD). After Ih, LSB with an aqueous core delivered a significantly higher concentration of 0.290 mg glycinelO mg ^'1 lipid compared to 0.090 mg glycine 10 mg ^"1 lipid delivered by LSB containing particulate glycine (p<0.05; Tukey's HSD).

Experiment 4: Retention of glycine vs. tyrosine

Suspension time, treatment and time x treatment interaction effects all had significant effects on retention of amino acids (p<0.001; df=44; two-way ANOVA). Throughout the period of aqueous suspension, LSB containing tyrosine had RE that were significantly higher than that for glycine. After 2 min of aqueous suspension, LSB containing 7% and 23% particulate tyrosine had similar RE (98.26 vs. 98.84%, respectively; p>0.05) but they were significantly higher than that of LSB containing 7% particulate glycine (73.21%; p<0.001). Similarly, 60 min RE for LSB containing either 7 or 23% tyrosine were not significantly different from each other (74.81 and 73.54%, respectively; p>0.05) but they were significantly higher (p<0.05) than that of LSB containing 7% particulate glycine (31.55%, p<0.001).

Regression analysis indicated that for each treatment, there was a significant relationship between the fraction of amino acid retained and the duration that LSB were suspended in water. Table 2.2 summarizes the regression equations fitted to the observed data, the time to 50% RE (T50), R ² and the associated standard errors for each treatment. While a linear regression equation best described the retention pattern of glycine on a log-scale,

polynomial equations best described the retention of tyrosine by LSB. Although retention profiles and T50 values for glycine and tyrosine differed considerably, different concentrations of tyrosine were retained similarly by LSB.

Although RE were identical, there were significant differences in DE of LSB containing tyrosine due to differences in their EE. After 2 min, LSB containing 23% tyrosine delivered significantly more tyrosine (2.174 mg 10 mg ^"1 lipid) compared to 0.591 mg 10 mg ^'1 delivered by LSB containing 7% tyrosine (p<0.05; Tukey's HSD). DE for glycine were significantly less than that for tyrosine throughout the experiment (p<0.05; Tukey's HSD). After 1 h of suspension, LSB initially containing 7% glycine delivered only 0.206 mg 10 mg ^"1 lipid compared to 0.507 mg tyrosine 10 mg ^"1 lipid delivered by LSB containing 7% tyrosine. After 1 h, DE of LSB containing 23% tyrosine, on the other hand, was 1.819 mg 10 mg ^"1 lipid, which was significantly higher than those of both the other treatments (p<0.05; Tukey's HSD).

Breakdown of CP by larvae Direct observation of the digestive tract was not possible due to heavy pigmentation of clownfish larvae. Therefore, visual observations were carried out on fecal strands and alimentary tracts that are isolated from the larvae. A compacted, pink-colored, single mass containing digested LSB and dietary ingredients was evident and indicated that CP were broken down by the larva.

Example 3

Zein-bound complex particles were compared with zein microbound particles and lipid spray beads for delivery of amino acids to marine fish larvae (Table 3.1). In these experiments, the CP comprised a body of coating material containing a minimum of two different particle types (one or more LSBs containing amino acids to be delivered to early marine fish larvae and, not contained within LSBs, a dietary mixture to be delivered to the early marine fish larvae).

The potential of CP for delivering a mixture of free amino acids (FAA) (glycine, alanine, serine, leucine and tyrosine) to early marine fish larvae was evaluated. Release patterns of individual FAA from CP were characterized in order to identify amino acids with higher retention and delivery efficiencies.

MATERIALS AND METHODS

LSB preparation

LSB were prepared using the melt-spray method as described above and by Onal and Langdon (Onal U, Langdon CJ (2004) "Lipid spray beads for delivery of riboflavin to first- feeding fish larvae" Aquaculture, 233:477-493; Onal U, Langdon CJ (2004) "Characterization of lipid spray beads for delivery of glycine and tyrosine to early marine fish larvae" Aquaculture, 233:495-511, which are incorporated herein by reference).

Briefly, a FAA mixture (30% w/w; alanine, glycine, leucine, serine and tyrosine; Sigma Chemicals, Inc.) or glycine alone (25% w/w) was encapsulated within LSB composed of 100% menhaden stearine (Omega Protein, Inc). The FAA mix was prepared by first dissolving equal molar concentrations (5 mM) of each amino acid in 1 1 of distilled water. The dissolved amino acid mixture was then spray-dried (Buchi Mini Spray Dryer, Switzerland) at 100 ⁰ C. Spray- drying facilitated the preparation of a finer FAA powder (particles 2-10 μm in size) compared with grinding particles in a ball mill.

CP preparation

CP were prepared by a spray-air method. Two grams of zein (Pfaltz and Bauer, Inc.) were dissolved in 50 ml of 90% aqueous ethanol solution using a homogenizer (PowerGen 700, Fisher Scientific). Menhaden oil (10% w/w) in combination with 10-20% w/w (based on formulation) rice starch (Sigma Chemicals, Inc.) and 10% w/w casein (Sigma Chemicals, Inc.) were added and homogenized until the ingredients were bound by zein. The dietary mixture was then cooled to 15 ⁰ C in an ice bath and LSB (40-50% w/w based on formulation) containing 25% w/w particulate glycine or 30% particulate FAA mix (glycine, alanine, serine, leucine, and tyrosine) were added to the dietary mixture.

The zein/LSB/dietary mixture was then atomized into a conical-bottomed, fiberglass cylinder (50x120cm) (FIG. 2). The diet was first placed in a pressurized chamber made from 5 cm diameter PVC tubing and connected to the fluid inlet port of a spray nozzle system (1/4 JBCJ; Spraying Systems Co.). High-pressure nitrogen was delivered via pressure regulators to both the gas inlet port of the spray nozzle and to the top of the chamber. Nitrogen gas pressure (5-10 psi) forced the dietary mixture downward inside the chamber, extruding the mixture through the fluid orifice of the spray nozzle. Use of a 90% alcohol solution to dissolve the zein

allowed air-drying of atomized particles at room temperature, avoiding loss of LSB by melting. Particles were collected using a soft brush, sieved through successive mesh sieves, the 45-106 μm size fraction was collected and refrigerated under nitrogen until use. CP were examined to determine size and structure, using both optical and scanning electron microscopes (SEM).

ZMP preparation

Two grams of zein (Pfaltz and Bauer, Inc.) were dissolved in 50 ml of 60% alcohol solution using a homogenizer (PowerGen 700, Fisher Scientific). Menhaden oil (10% w/w) in combination with 10% w/w of casein, 40% w/w rice starch and 20% w/w FAA were added and the dietary mixture was spray dried (Mini Spray Dryer B-191, Bϋchi, Switzerland) using the following parameters: flow rate of dietary mixture at 5 ml min ^"1 , inlet temperature at 100 ⁰ C, outlet temperature at 65 ⁰ C, atomizing air pressure at 40 psi, atomizing air flow rate at 500 cc sec ^"1 . The resulting ZMP were sieved through mesh sieves, the 45-106 μm size fraction was collected and refrigerated under nitrogen until use.

Measures of microparticle performance for delivery of amino acids

Measures of microparticle performance included inclusion (IE), encapsulation (EE), retention (RE) and delivery efficiencies (DE) in addition to T50 values, as determined according to methods described in Example 1. Briefly, IE was expressed as the percentage (w/w) of core material in the dietary mixture that was incorporated in the prepared particle. EE was expressed as the percent of the total particle weight made up of core material. RE was expressed as the percentage of initial core material retained after suspension of particles in water. DE was defined as the weight of core material remaining (mg core material per g of particles) after a given amount of time in which particles were suspended in water. T50 values for each suspended particle type were expressed as estimated times at which 50% retention occurred and were calculated using regression equations derived for each treatment.

RE and DE of microparticle types were determined according to methods described in Example 1. Briefly, triplicate samples of 10 mg particles were added to glass vials containing 10 ml water. Samples of suspended particles were taken over a given period to determine changes in retention and delivery of amino acids over time. Experiments were carried out at 25 °C±1, except when noted otherwise. In cases of potential interference from dietary ingredients, a control treatment was included, in which core material (FAA mix or glycine) was excluded

from the diet and absorbance due to dietary ingredients was subtracted from absorbance values of each treatment containing the core material. Glycine concentrations were determined according to the method described in Example 2. Concentrations of FAA lost from CP and ZMP were determined using HPLC analysis (AAA Services Lab, Boring, OR, USA).

Breakdown of CP by clownflsh larvae

CP within a size range of 45-106 μm were fed to three day-old clownfish, Amphiprion percula, larvae in order to determine if they were broken down and their contents released into the larval gut, according to methods described in Example 2. In order to help visualize LSB digestion, LSB were prepared containing a 10% w/w solution of Poly-red (PoIy-R 478; Sigma Chemical Inc.) and release of free Poly-red into the larval gut was observed.

After hatching, clownfish larvae were maintained in a static system, at 24-25 ⁰ C, and 30 ppt salinity. Larvae were fed on rotifers, Brachionus plicatilis. at a density of 5 ml ^"1 , in combination with Tetraselmis chui. at 5,000-10,000 cells ml ^"1 for two days. Prior to feeding experiments, 20 larvae were randomly selected and placed in a 1 1 glass beaker and starved overnight (14-16 h). Fifty mg of diet were then fed to larvae with gentle aeration that helped particles stay in suspension. Larvae were allowed to feed for 2-4 h and individuals were sampled to examine fecal strands and gut contents under a microscope (Nikon Optiphot-2) and photographed (Nikon N6000 camera).

Results

Physical properties of CP

CP prepared by the spray method typically ranged in size from 20 to 500 μm. Particle size could be manipulated by varying atomizer orifice size and nitrogen flow rates. SEM images of fractured CP indicated a distinct wall, possibly consisting of zein, surrounding core materials, including putative spherical LSB. CP were non-sticky, could be stored for many months in sealed, dry vials without observable physical deterioration and were readily dispersible in seawater.

Comparison of CP and LSB for delivery of glycine

Mean glycine IE for CP and LSB were high at 91.88% and 94.02% while mean EE were 9.19% and 18.80%, respectively as shown in Table 3.1.

Exp # Microparticle type and Core IE w/w (%) EE w/w (%) preparation method

1 LSB (melt-spray) glycine 94.02+0.64 18.80+0.71 CP (spray-air) glycine 91.88+ 1.25 9.19+ 0.42

2 ZMP (spray-air) FAA mix 89.70+0.82 17.94+0.63 CP (spray-air) FAA mix 89.73+1.11 13.46+0.49

Table 3.1: Inclusion (IE) and encapsulation efficiencies (EE) of LSB, ZMP and CP containing either glycine or FAA mix. LSB: lipid spray beads; ZMP: zein microbound particles; CP: complex particles. Means + 1. s.d.

Suspension time, particle type and time x particle type interaction all had significant effects on glycine retention (pO.OOl; two-way ANOVA). Leaching patterns of both particle types were best described by linear regression equations on a log-log scale (Table 3.2).

Exρ # Microparticle type Core Regression equation R ² T50 (min)

1 LSB (melt-spray) glycine -0.258x (/ntime) + 4.50 98.0 9.81

CP(spray-air) glycine -0.468 x (/ntime) + 4.42 98.4 2.97

2 ZMP (spray-air) FAA mix ND ND ND

CP (spray-air) FAA mix alanine/glycine/serine -0.635 x (Mime) + 4.46 93.8 2.38 leucine -0.514 x (/ntime) + 4.52 95.4 3.23 tyrosine -0.184 x (/ntime) + 4.58 98.9 36.7

Table 3.2: Regression equations fitted to describe change in retention efficiencies for glycine and FAA mix, associated R ² and T50 (time to 50% retention) values for each microparticle type. ZMP: zein microbound particles; CP: complex particles; ND: Not determined.

Throughout the experimental period, LSB had a significantly higher RE for glycine compared to that of CP (p<0.01; Tukey's HSD; Fig 3 a); for example, after 2 min of aqueous suspension, LSB had a RE of 69.0% which was significantly higher than 44.1 % RE for CP (p<0.05; Tukey's HSD) and after 1 h, LSB had a significantly greater (p<0.05; Tukey's HSD) RE of 31.6% compared to 12.7% for CP. The suspension period for 50% retention of glycine (T50) by LSB was 9.81 min, compared to 2.97 min for CP (Table 3.2).

DE followed a similar pattern to RE, with higher concentrations of glycine delivered by LSB throughout the experimental period; for example, after 2 min of suspension, LSB delivered a significantly higher amount of glycine (129.7 mg glycine g ^'1 ) compared to 48.1 mg g ^"1 delivered by CP. (p<0.05; Tukey's HSD) and after 1 h, DE of LSB was significantly higher (59.4 mg g ¹ ) than that of CP (11.6 mg g ¹ ; ρ<0.05; Tukey's HSD).

Comparison of CP and ZMP for delivery of a FAA mixture

Mean IE for the FAA mixture for CP and ZMP were high at 89.75% and 89.70%, respectively, while mean EE were 13.46% and 17.94%, respectively (Table 3.1). Suspension time, particle type and time x particle type interaction all had significant effects on retention of FAA mix (pO.OOl; two-way ANOVA). Due to very high initial leaching rates, retention

patterns of FAA in ZMP could not be described by simple equations (Table 3.2). ZMP initially had very high leaching rates and retained only 12.7% of the initial FAA after 2 min of aqueous suspension in contrast to CP that retained 44.3%. Throughout the experimental period, CP had significantly higher RE compared to ZMP (Fig 4 a; p<0.05; Tukey's HSD). At the end of 1 h, CP showed a RE of 14.5% that was significantly higher than a 1% RE for ZMP (p<0.001 ; Tukey's HSD).

The leaching patterns of alanine, glycine and serine from CP did not differ from each other (ANCOVA, p=0.371) whereas separate regression equations were necessary to describe leaching patterns for tyrosine and leucine (p<0.001; ANCOVA). After 1 h of suspension, RE for tyrosine was 44.9% and this was significantly higher than for other FAA. T50 values indicated that CP retained 50% of the initial tyrosine concentration after 36.7 min of suspension.

DE for all FAA were always higher for CP than for ZMP throughout the experimental period (p<0.001; two-way ANOVA). DE of serine and alanine for CP did not differ from each other (p=0.614; ANCOVA) whereas DE of tyrosine, leucine and glycine were significantly different (p<0.001; ANCOVA; Fig 4 c). DE for tyrosine was 20.3 mg g ^"1 CP particle and this was significantly higher compared to those for other amino acids throughout the experimental period (p<0.05; Tukey's HSD). Regression analysis indicated a significant relationship between retention of FAA by CP and their solubilities in water (pO.OOl).

Breakdown of CP by larvae

Observation of the digestive tract of intact clownfish larvae was not possible due to pigmentation of the body wall. Therefore, visual observations were carried out on fecal strands and alimentary tracts that were dissected from larvae. Release of Poly-red dye from CP into the guts of 3 -day old larvae was observed.

Discussion

This example demonstrates that, using a spray-air method, that LSB containing a FAA mixture were successfully incorporated in zein-bound CP. Since FAA are only slightly soluble in absolute alcohol, using an alcohol concentration of 90% for the zein solution and lowering its temperature to 15 ⁰ C likely reduced dissolution of amino acids during manufacture and resulted in high IE (Table 3.1). High alcohol concentration also allowed air-drying of the .atomized droplets without the need for additional heat that could melt LSB during CP preparation. Using

this modification resulted in high EE (17.9%) of FAA in CP. This high EE is a major improvement compared to earlier methods for preparing CP that involved the use of aquaeous solutions that promoted loss of amino acids during the preparation process.

ZMP had a very low RE and DE and lost 88% of FAA after only 2 min of aqueous suspension. These results indicated that ZMP do not efficiently retain low-molecular weight, water-soluble nutrients, such as amino acids. Use of CP, on the other hand, substantially reduced FAA leaching rates.

RE and DE patterns differed for individual FAA incorporated in CP. Tyrosine was retained and delivered with the highest efficiencies throughout the experimental period. Despite differences in RE for individual amino acids, retention patterns were linear when plotted on a log-log scale. Retention patterns of alanine, glycine and serine were identical and could be defined by a single regression equation. A significant correlation was found between RE and water solubility of FAA.

Although CP were larger than LSB, with lower surface area to volume ratios and greater diffusion distances, coating LSB with zein did not reduce leaching rates. Higher RE and DE of LSB may have been due to differences in suspension characteristics as LSB were hydrophobic, sticky, and less dispersed in aqueous suspension compared to CP. Clumping in aqueous suspension makes LSB unsuitable for use as feeds for fish larvae; therefore, LSB are better used as inclusion particles rather than as free particles for this application. The development of CP that provide amino acids and other low molecular weight, water-soluble nutrients to marine fish larvae allows diets that are more readily utilized than existing diets prepared with complex, high molecular weight nutrients. Free amino acids are present in food items commonly reported in diets of the early stages offish larvae, such as Artemia sp. and other zooplankton and phytoplankton. Dietary free amino acids have been reported to be preferentially assimilated by the early stages of marine fish larvae compared with polypeptides and protein. And free amino acids released from the cells of ingested prey may also increase the feeding response and release of digestive enzymes in marine fish larvae by stimulating olfactory and gustatory receptors.

In summary, zein-bound CP containing LSB can deliver low molecular-weight, water- soluble nutrients, such as amino acids, to marine fish larvae. Such a particle type may allow early weaning offish larvae onto artificial diets.

Example 4

This example concerns the development of a microparticle type that effectively retains riboflavin when suspended in water. Initially, the effect of adding different lipid levels and types on retention efficiencies of zein-bound microparticles (ZBP) was tested. Next, performances ZBP prepared using different particle-forming methods were compared. Finally, performances of ZBP and complex particles were compared.

MATERIALS AND METHODS

Experimental approach

Riboflavin was chosen as the core material because it represented a water-soluble (0.33g/l at 25 ⁰ C) vitamin, a common ingredient of artificial diets for aquatic animals. Crystalline riboflavin was incorporated into zein-bound particles (ZBP). Zein is soluble in 60- 95% aqueous alcohol solutions making it very suitable for a variety of manufacture processes. Initially, leaching rates of ZBP containing different levels of menhaden oil (MO;

Omega Protein Inc., USA) were compared to determine if lipid level affected riboflavin leaching rates (treatments T1-T3, Experiment 1). After determination of the lipid content that resulted in the lowest leaching rates, ZBP alone and ZBP containing either methyl palmitate (MP) or trilaurin (T) were compared for retention and delivery of riboflavin (treatments T1-T3, Experiment 2). MP is a low-melting point wax (melting point 28-29 ⁰ C) whereas trilaurin is a high-melting point triacylglycerol (melting point 55 ⁰ C).

ZBP were prepared using different particle-forming methods in order to determine the best method of preparation (treatments Tl - T3, Experiment 3). Finally, experiments were carried out to compare leaching rates of ZBP and CP.

Microparticle Preparation

Preparation of ZBP containing riboflavin

A mixture of dietary ingredients and riboflavin (see Table 4.1a and 4.1b) were prepared using a method similar to that described by Kanazawa et al. (Kanazawa, A.; Teshima, S.;

Inamori, S.; Sumida, S.; Iwashita, T., "Rearing larval red sea bream and ayu with artificial diets" Mem. Fac. Fish., Kagoshima Univ. 1982 pp. 185-193).

g/lOOg

Diet formulation Exp l Exp 2 Exp 3

Tl T2 T3 Tl T2 T3 Tl T2 T3

ZBP

Krill meal" 24.75 21.95 17.56 18.48 18.48 18.48 23.36 23.36 23.36

Egg solid ^b 15.00 13.30 10.64 11.20 11.20 11.20 13.30 13.30 13.30

Fish meaf 14.10 12.50 10.00 10.53 10.53 10.53 14.16 14.16 14.16

Artemia meal ⁰ 6.00 5.32 4.26 4.48 4.48 4.48 5.66 5.66 5.66

Liver meal ^d 3.75 3.33 2.66 2.80 2.80 2.80 3.54 3.54 3.54

Wheat gluten ^b 3.75 3.33 2.66 2.80 2.80 2.80 3.54 3.54 3.54

Vitamin premix ^b 2.25 2.00 1.60 1.68 1.68 1.68 2.12 2.12 2.12

Ascorbic acid ^b 0.75 0.67 0.53 0.56 0.56 0.56 0.71 0.71 0.71

Trace min premix ^b 0.075 0.067 0.05 0.056 0.056 0.056 0.071 0.071 0.071

Inositol" 0.075 0.067 0.05 0.056 0.056 0.056 0.071 0.071 0.071

Menhaden oil" 4.50 12.5 25.0 12.50 0.00 0.00 12.50 12.50 12.50

Methyl palmitate ^f 0.00 0.00 0.00 0.00 12.50 0.00 0.00 0.00 0.00

Trilaurin ⁶ 0.00 0.00 0.00 0.00 0.00 12.50 0.00 0.00 0.00

Riboflavin ⁸ 5.00 5.00 5.00 15.00 15.00 15.00 1.00 1.00 1.00

Zein ^f 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00 20.00

Table 4.1a: Composition of experimental diets used in exp 1, 2 and 3. ^a International Proteins Corp., USA, ^b International Ingredient Corp., USA, ^c American Protein Corp., USA, ^d ESTVE Aquaculture, Belgium, Omega Protein Inc, USA, ^f Pfaltz and Bauer, Inc., ⁸ Sigma-Aldrich, USA

Diet formulation

ZBP g/100g LSB g/100g CP g/100g

Krill tneaT 19.50 Methyl palmitate* 78 Krill meal" 14.06

Egg solid ^b 11.82 SMP ^g 10 Egg solid ⁶ 8.52

Fish meaf 11.11 Riboflavin 12 Fish meaf 8.01

Artemia meal ^c 4.73 Artemia meal ⁰ 3.41

Liver meal ^d 2.96 Liver meal ^d 2.13

Wheat gluten ^b 2.96 Wheat gluten" 2.13

Vitamin premix ^b 1.77 Vitamin premix ^b 1.28

Ascorbic acid ^b 0.59 Ascorbic acid ^b 0.43

Trace min premix ^b 0.059 Trace min ρremix ^b 0.043

Inositol" 0.059 Inositol" 0.043

Menhaden oil ⁶ 12.50 LSB 28.00

Riboflavin ⁶ 12.00 Riboflavin ⁶ 12.00

Zein ^f 20.00 Zein ^f 20.00

Table 4.1b: Composition of experimental diets used in exp 4. "International Proteins Corp., USA, ^b Mernational Ingredient Corp., USA, ^c American Protein Corp., USA, ^d INVE Aquaculture, Belgium, Omega Protein Lie, USA, ^f Pfaltz and Bauer, USA, ^g Sigma-Aldrich, USA.

Two grams of zein (Pfaltz and Bauer, USA) were dissolved in 50 ml of 60-90% aqueous alcohol solution using a honiogenizer (PowerGen 700, Fisher Scientific, USA). Preliminary observations indicated a minimum of 4-5% w/w lipid was necessary for effective binding of ZBP to prevent crumbling following manufacture. Menhaden oil (4.5-25% w/w) and dietary mixture (52.5-70.5% w/w) in combination with riboflavin (1-15% w/w) and other ingredients were added and homogenized until the ingredients were bound by zein. In Experiments 1 and 2, ZBP were prepared by the spray-water method whereas in Experiment 3, ZBP were prepared by three different methods as described below. In Experiment 4, ZBP were prepared by the spray-air method.

Spray-water method: The dietary mixture was atomized as described by Onal and Langdon (2000) for the preparation of gelatin-alginate beads. In order to optimize delivery of the dietary mixture, a pressurized chamber was made from 5 cm diameter PVC tubing and connected to the fluid inlet port of a spray nozzle system (1/4 JBCJ; Spraying Systems Co.).

High-pressure nitrogen was delivered via pressure regulators to both the gas inlet port of the spray nozzle and to the chamber top. Nitrogen gas pressure (5-10 psi) forced the dietary mixture downward inside the chamber, extruding the mixture through the fluid orifice of the spray nozzle where it was atomized into a PVC cylinder as described by Villamar and Langdon (Villamar DF, Langdon CJ (1993) "Delivery of dietary components to larval shrimp, Panaeus vannamei, by means of complex microcapsules" Mar. Biol. 115: 635-642). When the atomized droplets came into contact with chilled water, particles were formed ranging from 20-500 μm in diameter. ZBP were then rinsed with distilled water and sieved through 45-106 μm mesh sieves. In Experiment 2, the amount of water that was used to rinse ZBP particles were limited to 300 ml to reduce losses due to rinsing. The particle suspension was sieved through 45-106 μm mesh sieves and different size fractions were freeze-dried and stored at 5 ⁰ C until use.

Sprav-drv method: The dietary mixture, prepared as described above, was spray dried (Mini Spray Dryer B-191, Bϋchi, Switzerland) using the following parameters: flow rate of coating mixture: 5 ml/min; inlet temperature: 100 ⁰ C, outlet temperature: 65 ⁰ C, atomizing air pressure: 40 psi, atomizing air flow rate-. 500 cc/sec. The resulting ZBP were sieved through 45- 106 μm mesh sieves and stored at 5 ⁰ C until use.

Freeze-dry method: ZBP were prepared according to the method described by Kanazawa et a (1982). The dietary mixture was freeze-dried for 72 h, ground with a pestle and mortar and sieved through 45-106 μm mesh sieves and stored under refrigeration until use. Spray-air method: In Experiment 4, ZBP were prepared as described in Example 3 using a spray-air method that was originally developed for preparation of CP (see below).

Preparation of CP containing riboflavin

CP were prepared using a spray-air method as described in Example 3. Two grams of zein (Pfaltz and Bauer, USA) were dissolved in 50 ml of 90% aqueous ethanol solution using a homogenizer (PowerGen 700, Fisher Scientific, USA). The dietary mixture (40% w/w; see Table 4. Ib) was added and homogenized until the ingredients were bound by zein. The dietary mixture was then cooled to 15 ⁰ C in an ice bath and LSB were added that were composed of a lipid wall of 90% MP and 10% emulsifϊer, sorbitan monopalmitate (SMP) with a core of 30% w/w ground riboflavin. The riboflavin added in LSB resulted in a final riboflavin concentration of 12% in CP.

The zein/LSB/dietary mixture was then atomized into a conical-bottomed, fiberglass cylinder (50x120cm) using a pressurized chamber and spray nozzle (described above). Use of 90% alcohol to dissolve zein allowed air-drying of atomized particles at room temperature, avoiding loss of LSB by melting at high temperatures associated with the spray-dryer method and eliminating loss of riboflavin due to contact with water associated with the spray-water method. Particles were sieved through successive mesh sieves and the 45-106 μm size fraction collected and refrigerated under nitrogen until use.

LSB Preparation LSB were prepared by the melt-spray method as described above in Examples 1 and 2.

In order to obtain stable suspensions of lipid and riboflavin, 10% SMP (sorbitan monopalmitate) was incorporated in MP. Free LSB prepared for Experiment 4 contained 12% w/w riboflavin. The proportion of riboflavin in LSB incorporated into CP was increased to 30% w/w so that the total concentration of riboflavin in CP was 12% (Table 4.1b). LSB were prepared by mixing with sonication (B. Braun Biotech Inc., USA) finely ground riboflavin powder (<10 μm particles; McCrone micronizing mill, McCrone Scientific Ltd., UK) with molten (60-65 ⁰ C) lipid/emulsifier mixture. The core/lipid mixture was sprayed into a stainless steel cylinder that was cooled with vapor from liquid nitrogen. The hardened LSB were collected and stored in the dark under nitrogen at -2O ⁰ C until use.

Measures ofMicroparticle Performance

Inclusion efficiency (IE), encapsulation efficiency (EE), retention efficiency (RE) and delivery efficiency (DE) of LSB were determined by methods described in Example 1. Briefly, IE was expressed as the percentage (w/w) of core material in the dietary mixture that was incorporated in prepared particles. EE was expressed as the percent of total particle weight made up of core material. RE was expressed as the percent of initial core material retained after suspension of particles in water. DE was defined as the weight of core material remaining (mg core material per 10 mg of particles) after a given amount of time in which particles were suspended in water. T50 values were expressed as the time in minutes for 50% retention to occur for microparticles suspended in water and were calculated using regression equations derived for each LSB type.

Results

Inclusion efficiencies (IE) and encapsulation efficiencies (EE) of riboflavin by zein microparticles The proportion of riboflavin present in the initial spray mixture that was incorporated in the particles depended on the preparation method (Table 4.2).

Exp # Particle type IE EE

1 ZBP w/ 4.5% MO 4.80 0.24

ZBP w/ 12.5% MO 4.20 0.21

ZBP w/ 25% MO 3.40 0.17

2 ZBP w/ MO 30.5 4.57

ZBP w/ MP 31.5 4.72

ZBP w/ T 30.0 4.50

3 ZBP (freeze-dry) 96.10 0.96

ZBP (spτay-dτy) 107.0 1.07

ZBP (spray-water) 17.20 0.17

4 LSB (melt-spray) 91.4 10.97

CP (spray-air) 84.5 10.14

ZBP (spray-air) 88.3 10.60

Table 4.2: Inclusion (IE) and encapsulation (EE) efficiencies (mg core 100 mg ^"1 lipid) of microparticles. ZBP: zein bound particles; CP: complex particles; LSB: lipid spray beads; MO: menhaden oil; MP: methyl palmitate; T: trilaurin.

Preparation method had a significant effect on IE; ZBP prepared by the spray-water method resulted in significantly lower IE (P<0.005; Tukey's HSD), ranging between 3.40 and 31.5%, compared to IE of ZBP prepared by either the freeze-dry (96.0%) or spray-dry method (107%). ZBP containing MP had significantly higher IE and EE than ZBP containing either

MO or T. CP prepared by the spray-air method resulted in 84.5% IE and up to 10.1 % EE for riboflavin.

Experiment 1. Comparison of different lipid levels on performance of ZBP containing riboflavin Suspension time, lipid level and interaction (suspension time x lipid level) all had significant effects on retention of riboflavin (PO.001; two-way ANOVA). There were significant differences among treatments after 2 min suspension in water. ZBP with 12.5% MO had a significantly higher RE compared to those of other treatments (P<0.001; Tukey's HSD). A similar trend was observed throughout the experiment. Regression analysis indicated that for each treatment, there was a significant relationship between the fraction of riboflavin retained and the duration that ZBP were suspended in water. A biphasic pattern was observed for all types when log percent retention rate was plotted against log suspension time. Table 3 summarizes the regression equations fitted to the observed data, T50 and R ² for each treatment. ZBP containing 12.5% MO had a T50 of 7.05 min compared to T50 values of 5.22 and 3.39 min for ZBP containing 25% and 4.5% MO, respectively.

Exp # Particle type Regression equation R ² T50 (min)

1 ZBP w/ 4.5% MO 0.047 t ^J - 0.532 / + 4.491 97.2 3.39

ZBP w/ 12.5% MO 0.035 / ² + 0.405 / + 4.570 98.3 7.05

ZBP w/ 25% MO 0.026 Z ² + 0.393 / + 4.467 96.4 5.22

2 ZBP (common slope) -0.404 / + 4.602 96.3 5.52

3 ZBP (freeze-dry) -0.605 / + 4.758 97.4 4.04

ZBP (spray-dry) -0.393 / + 4.645 99.1 6.45

ZBP (spray-water) -0.058 t ² - 0.130 / + 4.423 94.8 5.08

4 LSB (melt-spray) -7.678 /+ 102.29 98.5 907.32

CP (spray-air) -11.677 / + 98.01 95.8 60.10

ZBP (spray-air) -1.757 f - 24.328 / + 90.968 95.8 7.21

Table 4.3: Regression equations fitted to describe change in riboflavin retention efficiency (RE), associated R ² and T50 (time to 50% retention) values for each microparticle type. ZBP: zein bound particles; CP: complex particles; LSB: lipid spray beads; MO: menhaden oil.

There were significant differences between the DE of ZBP after 2 min of suspension (P<0.05; Tukey's HSD). At 2 min, ZBP containing 12.5% MO had a DE of 14.8 μg riboflavin 10 mg ^"1 particle which was significantly higher than those of ZBP prepared with 4.5% and 25% MO (13.3 and 13.5 μg 10 mg ^"1 particle, respectively). After 1 h, ZBP containing 12.5 and 25% MO delivered 7.1 and 6.7 μg riboflavin 10 mg ^"1 particle, respectively, compared to significantly less (5.9 μg riboflavin; PO.05; Tukey's HSD) delivered by ZBP containing 4.5% MO. At 540 min (9 h), ZBP containing 12.5% MO had a DE of 5.8 μg riboflavin 10 mg ^"1 particle which was significantly higher than those of other particles (PO.05; Tukey's HSD).

Experiment 2. Comparison of lipid type on performance of ZBP containing riboflavin Regression analysis indicated that for each treatment, there was a significant relationship between the fraction of riboflavin retained and the duration that ZBP were suspended in water. However, the leaching patterns of ZBP containing different lipid types did not differ from each other (ANCOVA, p=0.115), and a single, linear regression equation was used to describe leaching patterns after log percent retention rates were plotted against log of time. A common T50 value for all particles was estimated to be 5.52 min.

There were significant differences between the DE of particles; ZBP containing 12.5% MP had significantly higher DE throughout the experimental period (PO.05; Tukey's HSD; Fig 2b). At 1 h, ZBP containing 12.5% MP delivered 0.127 mg riboflavin 10 mg ^"1 particle which was significantly higher than for ZBP containing 12.5% MO and 12.5% T, respectively.

Experiment 3. Comparison of different preparation methods on performance of 7BP containing riboflavin

Suspension time, preparation method and their interaction all had significant effects on retention of riboflavin (PO.001; two-way ANOVA). Throughout the experimental period, ZBP prepared by using a spray-dryer had significantly higher RE compared to those of other treatments (PO.001; Tukey's HSD) except at 60 and 180 min. At the end of 540 min, ZBP prepared by the spray-dry method had a RE of 8.33% that was significantly higher than those of ZBP prepared by the freeze-dry and spray-water methods (2.46 and 2.87%, respectively; PO.05; Tukey's HSD).

While a linear regression equation best described the leaching pattern of riboflavin from ZBP prepared by the freeze-dry and spray-dry methods on a log-log plot, a polynomial equation best described the leaching pattern of ZBP prepared by the spray-water method (Table 3). ZBP prepared by spray-dry method had a T50 of 6.45 min compared to T50 values of 4.04 and 5.08 min for ZBP prepared by freeze-dry and spray-water methods, respectively.

Throughout the experimental period, ZBP prepared by the spray-dry method had significantly higher DE compared with those of particles prepared by the freeze-dry and spray- water techniques. DE of ZBP prepared by the spray-dry and freeze-dry methods were significantly higher than those using the spray-water method throughout the experimental period due to significantly greater IE and EE. After 60 min, ZBP prepared by the spray-dry method delivered a significantly higher proportion of riboflavin (24.1 μg lOmg ^"1 particle) compared to 11.5 and 3.5 μg 10 mg ^'1 delivered by ZBP prepared by the freeze-dry and spray/water methods, respectively (P<0.05; Tukey's HSD).

Experiment 4: Comparison of performance of different particle types containing riboflavin

Suspension time, particle type and their interaction all had significant effects on retention of riboflavin (PO.001; two-way ANOVA). There were significant differences among treatments after 2 min suspension in water. LSB composed of MP+10% SMP had a significantly higher RE compared to those of other treatments throughout the experiment (PO.001; Tukey's HSD). CP had significantly higher RE compared to those of ZBP throughout the experiment.

Regression analysis indicated that for each treatment, there was a significant relationship between the fraction of riboflavin retained and the duration that particles were suspended in water. A straight line relationship was obtained for LSB and CP when percent retention rates were plotted against log of time. However, a biphasic pattern, described by a polynomial equation, was observed for ZBP when percent retention rates were plotted against log time. This was due to high initial leaching rates of riboflavin from ZBP followed by slower leaching rates observed after Ih suspension. LSB had a T50 of 907.3 min compared to T50 values of 7.2 and 60.1 min for CP and ZBP, respectively. LSB had significantly higher DE throughout the experimental period. After 60 min,

LSB delivered 0.810 mg riboflavin 10 mg ^"1 particle, compared to significantly less delivered by CP and ZBP (0.524 and 0.215 mg riboflavin, respectively; P<0.05; Tukey's HSD).

Discussion

The spray-dry method overcomes a major limitation in the preparation of ZBP by eliminating water from the manufacture process, resulting in higher IE and EE compared to ZBP prepared by the spray-water method. Although ZBP prepared by the spray-dry and freeze- dry methods had high IE and EE, elimination of water from the manufacture process did not reduce leaching rates of particles suspended in water. Overall, our findings show that ZBP are not effective for delivering low molecular weight, water-soluble nutrients.

Lower RE of ZBP containing 25% MO may have been due to higher lipid levels causing softer and more unstable particles compared to ZBP containing 12.5% MO. Biphasic retention patterns of these ZBP indicates an initial rapid release phase followed by slower second phase. The initial rapid release may indicate dissolution of unencapsulated or poorly encapsulated riboflavin and the second phase may have represented matrix-controlled release of riboflavin. ZBP containing 12.5% T did not show lower leaching rates compared to those of ZBP containing either 12.5% MO or 12.5% MP. Higher leaching rates of riboflavin from ZBP containing 12.5% T may have been due to incomplete melting of T during the preparation process, resulting in poor coating of riboflavin crystals. In contrast, addition of MP may have resulted in better coating of riboflavin crystals than with MO and, consequently, reduced leaching rates.

Higher RE and DE of ZBP prepared by the spray-dry method compared with the freeze- dry method may have been due to the presence of more rounded and smooth shaped particles, resulting in lower surface to volume ratios. In addition, grinding of freeze-dried particles may have caused cracks in the zein/diet matrix, increasing leaching rates. CP containing LSB were prepared by the spray-air method to prevent melting of LSB and contact with water during manufacture. In the present study, CP had lower riboflavin leaching rates compared to ZBP. LSB performed better than CP, possibly due to the higher hydrophobicity of LSB and greater difficulty in suspending them in water. Consequently, LSB should be used as inclusion particles rather than as free particles for delivery of water-soluble substances to fish larvae. Use of CP containing LSB should allow delivery of complete diets to early fish larvae, including water-soluble vitamins and other nutrients such as amino acids and trace minerals.

Example 5

Lipid extruded beads (LEBs) could be substituted for LSBs in the manufacture of CPs as described above. LEBs could, for example, be made by placing a total of 1Og of menhaden stearine in a water bath (75 ⁰ C). Two grams of finely ground riboflavin powder (<10 μm particles; McCrone micronizing mill, McCrone Scientific Ltd.) could then be mixed with 1O g of the molten (60-65 ⁰ C) lipid and sonicated at half power (B. Braun Labsonic L, B. Braun Biotech Inc.). The suspension of particulate inclusion material in molten lipid could then poured into a heated (65 ⁰ C) aluminum container with a perforated die attached to the bottom that is perforated with orifices 250 microns in diameter. The temperature of the container and die could be controlled by a temperature controller (model CN9000A, Omega Engineering Inc.) in combination with a heating coil and a thermocouple. The lipid/inclusion material suspension then could be extruded under pressure through the perforated heated die submerged in chilled ethanol solution contained within a vessel. A rotating mixing bar placed near to the surface of the die plate could be used to disrupt the flow of the lipid/inclusion material suspension extruded through the orifices to form LEBs.

Variations within the scope and spirit of the disclosure above will be apparent to those of ordinary skill in the art. The scope of coverage is accordingly defined not by the particular examples and variations explicitly described above, but by the claims below.