AU698597B2 - Modified alpha-lactalbumin - Google Patents

Description

MODIFIED ALPHA- CTALBUMIN

This invention relates to modified proteins useful in the diet of patients with hyperphenylalaninemia.

Human hyperphenylalaninemia, due to genetic causes, is the result: of mutations at the phenylalanine hydroxylase (PAH; E.C. 1.99.1.2) locus, at loci encoding DHPR (dihydropteridine reductase) and at two loci in the BH₄ (tetrahydrobiopterin) synthesis pathway. Mutations at non-PAH loci account for fewer than 3% of all hyperphenylalaninemic patients and is probably closer to 1% in Caucasian patients. Over half the patients with hyperphenyl-alaninemia due to PAH mutations have severe PAH deficiency (classic phenyl etonuria or "PKU") . This hepatic enzyme is responsible for converting phenylalanine to tyrosine.

Phenylalanine <—(PAH) // >Tyrosine

Phenylpyruvic Acid

Phenyllactic Acid Phenylacetic Acid

As shown in the above scheme, a deficiency in PAH results in a build up of phenylalanine and its organic acid metabolites in tissues and the blood which, if left unchecked, results in severe and irreversible mental retardation in infants and young children.

The frequency of PKU varies worldwide, with a rate of 1 in 5,000 in Ireland to about 1 in 100,000 in Japan. A conservative average estimate worldwide for incidence of PKU is 50 cases per million live births. In the US the frequency of PKU is about 1 in 10,000 of all screened newborns.

Once affected infants are identified, early initiation of dietary treatment to lower the blood phenylalanine level is critical. The current goal of treatment is to maintain blood phenylalanine levels below 1000 μmol/1 (16.5mg/dl) and preferably below 600 μmol/1. Recent neurological and behavioural evidence points to an even lower treatment goal in the future: maintenance of phenylalanine values as near to normal as possible (<120 μmol/1, or <1.98 mg/dl) .

In theory, ideal biochemical control in PKU is achieved with frequent blood monitoring and adherence to a phenylalanine restricted diet (Trahms, "Nutritional Care in Metabolic Disorders" in Food, Nutrition and Diet Therapy, 8th Edition (1992), Mahan & Arlin, eds., W. B. Sanders Co., Philadelphia, Pennsylvania; Kitagawa et al , Enzyme 38321-327 (1987) ; and Owada et al , Acta Paediatr. Jpn . 30 405-409 (1988)), but in practice the ideal is not always achieved. Specifically, nutritional therapy is planned around the use of formulas with reduced phenylalanine content. Currently available formulas for infants and children with PKU are either 95 or 100% phenylalanine-free protein hydrolysates or mixtures of free amino acids with little or no additions of phenylalanine. Tyrosine is usually a supplemented amino acid, as there continues to be apprehension about allowing tyrosine levels in treated phenylketonuria patients to remain below the normal mean. Dietary compliance is critical, not only in early life to prevent mental retardation in affected infants and children, but also later, in adulthood and especially in pregnant women with hyperphenylalaninemia. Offspring of untreated pregnant women virtually all suffer from complications such as mental retardation and microcephaly. A minority also have congenital heart disease and low birth weight. These outcomes are extremely unfortunate since in most cases the infant is not hyperphenylalaninemic itself. Progressively decreasing IQs, learning difficulties and behavioural problems are also reported in hyperphenylalaninemic older children, adolescents and adults who have increasingly longer life spans but who nevertheless practice poor or no dietary phenylalanine control. It is apparent, therefore, especially with recent reports, that PKU represents a lifelong and transgenerational problem, requiring continuous and lifelong dietary control (Scriver et al , The Metabolic Basis of Inheri ted Disease, Scriver, CR. , Beaudet, A., Sly, . , Valle, D., (eds) , 7th edit. McGraw Hill, in press, Chapter 27; Joint PAO/ HO/UNU Expert Consultation, Energy and Protein Requirements, WHO Tech. Rept. Ser. No. 724, WHO, Geneva, Switzerland (1985) ; Levy H.L. Treatment of genetic diseases, Ed., R.J. Desnick, Churchill-Livingston, NY, (1991)) .

The feeding strategy of PKU patients is threefold. First, all essential amino acids except for phenylalanine are supplied to allow for proper growth and development of infants and children, or maintenance, in the case of adults. Vitamins and minerals are usually supplied in this amino acid component which is commonly in the form of a defined formula. For infants, the baseline amino acid profile should match that of human breast milk

(Nayman et al . 32 1279-1289 (1979) and Joint PAO/WHO/UNU

Expert Consultation, Energy and Protein Requirements, WHO

Tech. Rept. Ser. No. 724, WHO, Geneva, Switzerland (1985)), except for the phenylalanine content, which is low relative to human breast milk or absent altogether.

For children and adults, the minimum formula amino acid profile, except for phenylalanine, should meet the

FAO/WHO/UNU suggested pattern of requirement for these groups.

Second, foods with defined phenylalanine content are then added as natural supplements to the amino acid formula base diet. Since these foods contain varying levels of phenylalanine, what is chosen and how much is consumed gives more or less flexibility and palatability to the overall diet. The level of phenylalanine allowed is only that which is required to meet growth and/or maintenance needs. The maximum level of tolerance of phenylalanine intake throughout childhood, for example, is 500mg/day, with an acceptable range of 200-500mg/day.

Estimates of absolute phenylalanine requirement in humans are, according to Young & Pellett (Am. J. Clin . Nutr. 45 1323 (1987)) and the Committee on Nutrition of The

American Academy of Pediatrics (Pediatrics 57 783

(1976)), as follows in Table 1:

^"Estimates are from PKU subjects with near total deficiency of PAH activity The third component of the PKU diet is, after essential amino acid and phenylalanine needs are met, to ensure adequate calories. This is achieved with low protein, high carbohydrate foods such as fruits, vegetables, fats and oils.

While the current therapeutic formulas mentioned above are reasonably effective in supplying essential amino acids and in restricting phenylalanine intake and achieving adequate biochemical control, their nutritional adequacy and/or bioavailability may be questionable with regard to some vitamins, minerals and amino acids (Acosta et al . J. Inher. Metab. Dis . 5 107 (1982)). In addition, they suffer from poor taste, smell and high osmolar loads and, as a result, poor compliance. The need for dietary compliance and regulatory control throughout life underscore the need for better accepted food products that contain low or no phenylalanine.

Ideally, such products should taste good, smell good, and allow maximum flexibility of food choices and menu planning when using them. The ideal product or formula should be highly palatable and low, preferably free, in phenylalanine: these are key qualities encouraging compliance and allowing the maximum intake of low protein phenylalanine-containing foods and a more normal diet. A high biological value food which is low or totally free of phenylalanine would be an extremely valuable dietary component because a greater variety and volume of natural foods containing phenylalanine would be allowed.

The technology now exists to approach this problem in a new and more effective way. Using a combination of site- directed mutagenesis, protein engineering and recombinant DNA technology, a specific known protein can be altered to reduce or eliminate the phenylalanine residues, by mutating the relevant DNA codons.

In recent years, our understanding of recombinant DNA and genetic technologies, and protein engineering, has expanded considerably. This body of knowledge allows us to choose a specific protein for mutation, based on its known sequence, structure and function. The DNA coding for that protein can then be cloned from a suitable source, mutated to the desired content, sequenced to verify the changes, and cloned into a suitable vector for expression.

The DNA sequences encoding a variety of medically and nutritionally important proteins have been cloned in recent years. These include insulin, plasminogen activator, c.₁-antitrypsin, blood factors VIII and IX, lysozyme, αt-lactalbumin and lactoferrin. From these DNA sequences the protein sequences can be deduced and verified by peptide mapping and Edman degradation.

Protein X-ray crystallography and multidimensional NMR spectroscopy has enabled the three-dimensional structure of many proteins to be determined. This has been coupled with very powerful computers and advanced programs, which can model how changes in the amino acid residues may affect the structure and function of the proteins (Nilsson et al , Curr. Opin . in Struc. Biol . 2 569-575 (1992); Presta, L.G., Curr. Opin. in Struc. Biol . 2 593- 596 (1992); Cech, T.R., Curr. Opin . in Struc. Biol . 2 605-609 (1992) ; Pickersgill and Goodenough, Trends in Food Sci . and Tech . 5 122-126 (1991)) . Having established which protein is to be modified and what the modifications will be, strategies can be developed to isolate the DNA coding for that protein, and to generate a mutant DNA sequence which encodes the desired amino acid changes. Such strategies involve the subcloning of that region(s) of the gene which contains the amino acids which are to be changed. The sub-cloned fragments are then subjected to site directed mutagenesis. This technique can be used to target precisely the location and type of modification desired. At its most accurate, a single base in the DNA sequence can be changed to any of the other three bases, by means of oligonucleotide primers, which serve to mutagenise the DNA sequence. The DNA may then encode a totally different amino acid (McPherson, M.J., Directed Mutagenesis, Oxford Univ. Press, NY (1991); Carter, P., Biochem. J. 237 1-7 (1986); Nickoloff and Deng, Anal. Biochem. 200 81-88 (1992) ) .

There are several points which have to be considered when choosing a suitable protein for phenylalanine substitution. First, a protein which is to be used as a food must be available in large quantities. Therefore, a high level of expression must be achieved and it could be advantageous if the product were easily accessible and readily purified. Secondly, although protein modelling is becoming increasingly sophisticated, the structural and functional consequences of multiple amino acid changes remain conjectural. Since larger polypeptides contain more amino acids, increased size is usually reflected in a greater number of phenylalanine residues, with a corresponding need for greater mutagenesis to substitute them. Smaller polypeptides a priori are a better proposition for manipulation. In principle, an intracellular or secretory protein could be chosen. However, this choice will dictate the choice of expression system.

When considering which protein to modify, regard must be had to a variety of factors, including the structure of the protein, physiological and nutritional concerns, and the ease or otherwise of extracting the protein from a host synthesising cell.

As far as structure is concerned, protein characteristics such as activity, specificity and stability are controlled by the relative positions of amino acids in the protein and by the physical and chemical properties of the amino acid. The structure of the protein is closely related to its function. Generally, it is important when making amino acid, and thus conformational, changes to a protein, not to alter its function, and to make sure it is processed properly and, if necessary, secreted from the cell.

Physiological concerns depend on the nature of the chosen protein and whether it is an enzyme, a structural, regulatory, transport or food protein. One potential problem is that the mutation(s) affect structural and functional properties of a protein, such that the host cell or its function is compromised in some way.

Nutritionally, the protein needs to be as chemically well balanced as possible. Therefore, before any thought is given to possible mutations, the native protein must be analysed to assess its nutritional potential, in terms of amino acid content, digestibility and allergenicity. The ease of extraction will to some extent, be dependent on the method of production. This will dictate what contaminating substances are present and hence what method of extraction is used.

The present invention is based, at its broadest, on the realisation that all these various and sometimes conflicting considerations can be adequately addressed by the choice of α-lactalbumin as the candidate for modification.

According to a first aspect of the present invention, there is provided a modified α-lactalbumin comprising no phenylalanine residues or fewer phenylalanine residues than wild type α-lactalbumin, other than a variant of human α-lactalbumin in which Phe-80 is substituted by lysine.

a-Lactalbumin is the major whey protein in human milk, being present at levels of -2.5g/l (29% of total protein - Heine et al , J. Nutr. 121 277-283 (1991)). It is a globular, calcium binding protein, with a relative molecular mass of 14,200, and is found in almost all mammalian milks.

α-Lactalbumin has a high nutritional value with essential amino acids accounting for 63% of the total. Both bovine (Vilotte et al , Biochimie 69 609-620 (1987) and human α- lactalbumins (Hall et al , Biochem. J. 242 735-742 (1987) contain 123 amino acid residues, with only four phenylalanine, and four tyrosine residues per monomer, α-Lactalbumin has formed part of our diet for millennia in the form of human, cow, sheep, goat and camel milks. The function of α-lactalbumin is twofold. Its main role is as one component of the lactose synthase system. This enzyme synthesises lactose in the lactating mammary gland. α-Lactalbumin acts as a specifier protein, by modifying the substrate specificity of UDP _/3l,4- galactosyl transferase (E.C. 2.4.1.22), the second component of the enzyme system. Secondly, α-lactalbumin has a role as a source of nutrition.

Nine α-lactalbumin sequences have been completely determined (Mackenzie and White, Adv. in Prot . Chem. 41

173-315 (1991); Vilotte et al , Gene 112 251-255 (1992)), and the amino acid compositions of 13 α-lactalbumins have been published (Mackenzie and White, Adv. in Prot . Chem.

41 173-315 (1991)). The crystal structure at 1.7A of baboon α-lactalbumin was published in 1989 by Acharya et al (J. Mol . Biol . 20899-127 (1989)), and recently the human α-lactalbumin crystal structure at 1.7A has been published (Acharya et al , J. Mol . Biol . 221 571-581

(1991)) .

It is generally accepted that α-lactalbumin and c-type lysozymes have diverged from a common precursor (Hall and Campbell, Essays in Biochem. 22 1-26 (1986) The Biochemical Soc) .

α-Lactalbumin is a relatively small secretory protein which has only four phenylalanine residues. As it is a milk protein it is ideally suited for expression in the mammary gland of a transgenic mammal which, as will be discussed below, is the expression system of choice, α- Lactalbumin is well suited to meeting the structural, physiological, nutritional and harvestability considerations discussed above. Structurally, the four phenylalanine residues in the human and bovine mature proteins are at positions 3, 31, 53 and 80, and 9, 31, 53 and 80 respectively. A protein engineering study of these two proteins shows that almost any residues could in principle be accommodated in place of phenylalanine, on the basis of size and space. However, positions Phe-31 and 80 are involved with the lactose synthase reaction and calcium binding loop respectively, and so more care had to be taken when considering suitable substitutions for these positions.

It is known that α-lactalbumins can function with galactosyl transferases from other species (Khatra et al , Eur. J. Biochem. 44 537-560 (1974) ; Powell and Brew, Biochem. J. 142 203-209 (1974)).

Nutritionally, three aspects needed to be considered when choosing an appropriate protein for modifications: amino acid content, digestibility and antigenicity. As this food will be fed to infants as well as adults, the nutritional quality and level of essential amino acids must be as high as possible. In milk proteins generally, essential amino acids account for 51.4% of casein, 52.5% of whole milk proteins and 58.8% of bovine whey proteins. However, in α-lactalbumin the essential amino acids account for 63.2% of the total (Heine et al, J. Nutr. 121 277-283 (1991) ) . More importantly, α-lactalbumin has a high content of lysine and cysteine and a particularly high content of tryptophan (5.9% of total amino acids). There are problems associated with the measurement of true digestibility in vivo; however, the current available studies suggest that the digestibility of α- lactalbumin is as high as that of casein and better than that of .-lactoglobulin (BLG) . The major bovine whey protein, BLG, is not found in human milk and is thought to be a potent antigen in susceptible humans. Bovine α- lactalbumin is very similar to human α-lactalbumin and these similarities might be expected to minimise the antigenic properties of bovine α-lactalbumin. Compared to other milk proteins this seems to be true (Goldman, A.S., Pediatrics 32 425-443 (1963); Strobel, S., Hum. Nutr. :Appli . Nutr. 40A(1) 45-54 (1986)).

Phenylalanine-reduced, or phenylalanine-free, α- lactalbumin product will for preference be produced in the mammary glands of lactating transgenic animals, in which the native protein may still be present. This may make purification difficult unless the amino acid substitutions by themselves alter the protein's characteristics. However, a particular advantage of modifying α-lactalbumin and expressing it in the milk of a dairy animal is that purification away from the native α-lactalbumin may not be essential if the expression of the dairy animal's native α-lactalbumin is very low relative to expression of the modified α-lactalbumin, or if expression of the native or wild-type form is prevented or suppressed; this latter effect may be achieved by making the dairy animal transgenic through the use of embryonic stem cells. If purification is needed and conservative changes are necessary to maintain the protein's structure, then one possibility is to attach an additional amino acid sequence, such as a tail to the C-terminus, (Smith et al , Gene 32 321-327 (1984)) to facilitate purification. The additional amino acid sequence would have one or more characteristics which lead to easier separability of the protein; for example, the additional amino acid sequence may enable the protein to be separated on the basis of charge. A poly-Arg, or poly-Arg/Lys, tail would alter the charge sufficiently to allow more efficient and cheaper extraction.

In view of the phenylalanine content of other milk proteins, such as caseins, it is likely to be the case in practice that the modified α-lactalbumin should for preference be at least partially purified away from such proteins as was native α-lactalbumin.

The invention is particularly useful in improving the palatability of, and hence compliance for, the protein content in the diet of PKU sufferers. This is accomplished by choosing a suitable protein, namely α- lactalbumin, which fulfils the criteria of having a high nutritional value, and of being able to address all of the considerations discussed above. The DNA encoding this protein can then be cloned, mutated to remove at least some, but preferably all, of the phenylalanine residues, and expressed in a system which can produce a quantity large enough to supply the world market.

The α-lactalbumin which is used as the base for the modifications can be of any suitable animal origin. Humans and dairy animals such as cows, sheep, goats and camels are preferred. α-Lactalbumin haying either a human or a bovine origin is likely to be optimal. Both have similar nutritional value and humans have been consuming them for millennia.

Bovine or human α-lactalbumin would be particularly appropriate to use as the starting point for the mutational modifications as man is accustomed to eating both proteins; and they are very similar, having 72% homology at the amino acid level. Some, or all of the phenylalanine residues will be replaced with one or several amino acids depending on structural and/or nutritional considerations. The preferred amino acids from which the choice can be made are shown in Table 2. A combination of any of these amino acids, in any of the four phenylalanine positions, may be substituted. It is preferred to make substitutions at all these positions, but failing that at three or two of them, although the invention also covers substitutions at one position.

TABLE 2 Preferred Phenylalanine Substitutions Human Bovine

Tyr,Leu,Arg,Met,Trp Tyr,Ser,Leu Leu,Ile,Trp,Tyr Leu,Ile,Trp,Tyr

Tyr,Leu,Met,Trp Tyr,Leu,Met,Trp Tyr,Met,Trp,Leu Tyr,Met,Trp,Leu

If selection of the substituting amino acid is made solely on the basis of energy minimisation and structural considerations, then Tyr₃-Leu₃₁-Tyr₅₃-Tyr₈₀ human α- lactalbumin andTyr₉-Leu₃₁-Tyr₅₃-Tyr₈₀ bovine α-lactalbumin are preferred (or substitution mutants having fewer than those four substitutions) . If nutritional requirements are also taken into account, then Tyr_3/9-Tyr₃₁-Tyr₅₃-Tyr₈₀ human and bovine α-lactalbumin (or substitution mutants having fewer than those four substitutions) are the mutants of choice. A further way of determining the optimal substitution(s) is to look at the amino acids occupying corresponding positions in α-lactalbumins of other species and in lysozymes of vertebrates, since the α-lactalbumins are lysozymes from vertebrate species have, where examined on publicly accessible databases, very similar structures and are believed to have derived from the same ancestral gene; on this basis, Tyr₃-Tyr₃₁- Leu₅₃-Leu₈₀ human α-lactalbumin (or substitution mutants having fewer than those four substitutions) are preferred.

Some or all of the phenylalanine residues could be substituted. One or more of them may be preferred for structural purposes, and may therefore be conserved. In the absence of any structural considerations, at least one, two, three or four of the natural phenylalanine residues may be substituted, in increasing order of preference. It is known that a natural variant of human α-lactalbumin exists which has its Phe-80 substituted by lysine (Maynard, F., J^". Dairy Res . 59 425-429 (1992)); this variant per se is not part of the invention, although it may be used in the formulas and methods of the invention to be discussed below.

To assist purification, further modification of the α- lactalbumin can be made; for example, a poly-Arg or Arg/Lys tail may be added at the C-terminus of the protein. This will alter the charge of the protein and enable it to be purified relatively cheaply on a large scale. It could also have the advantage of improving the nutritional quality of the protein.

Modified α-lactalbumin of the invention may be produced by recombinant DNA technology, using a collection of techniques that have come to be known as protein engineering.

The co-ordinates for baboon α-lactalbumin are available from the Brookhaven database (Acharya et al , J. Mol . Biol . 208 99-127 (1989)). Those for the human protein and the cow protein are not presently available, but the three dimensional structure of the baboon protein is thought to provide an acceptable template for modelling the human and bovine wild type and mutant α-lactalbumins.

The effects of replacing an amino acid may be considered first at the level of amino acid type. Thus, aromatic should replace aromatic, charge charge, shape shape and so on.

The first stage is to assess the state of each of the four Phe residues in the native presumed three- dimensional structure. These are: Phe-3 (human) or Phe-9 (bovine) , which is on the surface; Phe-31, which is exposed on the surface; Phe-53, which is in an internal cavity; and Phe-80, which is largely buried. Conveniently, the four residues are fairly well separated such that it is reasonable to assume that, provided changes introduced are largely conservative, little interaction is likely between the sites.

Model building shows that essentially any residue could be accommodated at Phe-3/9 and 31, since both are on the surface. The buried Phe residues can both be accommodated with little extra movement of the surrounding protein. Further, Leu and Met can be put in, but polar or charged groups are likely to disrupt the structure and so are less preferred.

Some guidance, but not specific for α-lactalbumin or the purpose in hand, can be had from the literature (such as Bordo and Argos (J. Mol . Biol . 217 721-729 (1991)), Lesk and Boswell (Curr. Opin . Struct . Biol . 2 242 (1992)), Singh and Thornton ("Atlas of Protein Side-Chain Interactions", Vols 1 and 2, IRL Press, Oxford, England

(1992) ) ; and computer programs (for example the program

SYBYL from Tripos Associates) may be useful in assessing any other modifications which it may be desirable to make.

The following remarks summarise the position:

Posi tion 3 : Surface residue which can accommodate most residues: Tyr, Leu, Arg, Met, Trp in order of increasing relative energy, the most favoured being

Tyr.

Posi tion 9 : Surface residue which can accommodate most residues. Most α-lactalbumins have a Ser at this position: Tyr, Ser, lie.

Posi tion 31 : Exposed on the surface, so any residue will do. However, this region is important for the lactose synthase reaction and a hydrophobic interaction may be important. Therefore, Leu and possibly lie, then Trp, would be the most preferred here if retention of the activity is important.

Posi tion 53 : Largely buried. Tyr, Leu, Met, Trp is the order in increasing energy.

Position 80 : Totally buried. Tyr, Met, Trp, Leu is the order here. This is the calcium binding loop region and so changes here might be important. Tyr is the nearest isomorph.

C- terminus : Surface and therefore quite exposed.

Addition of several residues here should not have a marked effect on the structure but could improve the separation properties.

Any combination of these suggested substitutions in human, bovine or other α-lactalbumins may be used. While in principle modified α-lactalbumins in accordance with the invention may be made by any process, including chemical synthesis, the process of choice will generally involve the use of recombinant DNA technology in which the protein is expressed from a corresponding nucleic acid sequence.

According to a second aspect of the invention, there is provided a process for the preparation of a modified α- lactalbumin as described above, the process comprising coupling together successive amino acids, and/or ligating oligo- and/or poly-peptides. For preference, as foreshadowed above, coupling together successive amino acids will preferably be ribosomally mediated. Many expression methodologies are today well within the common general knowledge of those skilled in the art. In particular, reference is made to Sambrook et al , Molecular Cloning 2nd ed. , Cold Spring Harbor Laboratory Press (1989) .

Nucleic acid useful in such expression itself forms part of the invention. According to a third aspect of the invention, there is provided an isolated or recombinant nucleic acid encoding a modified α-lactalbumin comprising no phenylalanine residues or fewer phenylalanine residues than wild type α-lactalbumin.

In view of the degeneracy of the genetic code, many different nucleic acid sequences will encode any particular protein within the invention. All such sequences are encompassed within the invention, although in practical terms the preferred nucleic acid sequences will be those: (a) whose codons correspond where possible to those of the wild type gene;

(b) whose codons are selected in conformity with codon preferences of the expression host; or

(c) whose codons are selected on the basis of a compromise between criteria (a) and (b) .

Recombinant DNA encoding modified α-lactalbumin in accordance with the invention is not restricted either to cDNAs or to otherwise natural genomic sequences encoding the modified protein, although these sequences may be preferred. The "minigene" approach of WO-A-9005188, in which some, but not all, of the introns present in the natural gene are present in the construct introduced into the host may be used.

Recombinant DNA in accordance with this aspect of the invention will often be in the form of a vector. The invention is not limited to any particular type of vector, which may for example be a plasmid, cosmid or phage. Vectors containing sufficient regulatory sequences

(including a promoter) operatively coupled to the sequence encoding the modified α-lactalbumin may be useful as expression vectors. Vectors not including regulatory sequences operatively coupled may be useful as cloning vectors .

Recombinant DNA in accordance with the invention may also be in the form of a construct suitable for use in preparing a transgenic animal, for example by microinjection or by homologous recombination. Nucleic acid of the invention may be prepared, in a fourth aspect of the invention, by coupling together successive nucleotides and/or ligating oligo- and/or polynucleotides, as is well within the knowledge of those skilled in the art.

According to a fifth aspect of the invention, there is provided a host containing recombinant DNA as described above. The host may be a cloning host, in which case it will often be a bacterium such as Escherichia coli , or it may be an expression host. In an expression host, DNA encoding a modified α-lactalbumin will be operatively coupled to sufficient regulatory sequences (including a promoter) for expression and the host will permit expression to take place either constitutively or under regulated conditions. A wide range of expression hosts may be useful in the invention, including bacteria

(particularly E. coli) , yeasts (such as Saccharomyces cerevisiae, Hansenula polymorpha and Pischia pastoris) , insect cells (such as Spodoptera frugiperda) and mammalian cell lines (such as BHK and CHO cell lines) . By far the most preferred expression hosts, though, are animals, in particular placental non-human mammals whose germline includes DNA encoding modified α-lactalbumin, adult females of the line being capable of expressing modified α-lactalbumin in the mammary gland such that the modified α-lactalbumin accumulates in the milk. If the modified α-lactalbumin is derived from a different species from the host animal, the transformed host may properly be said to be transgenic; for example a cow or bull carrying DNA encoding modified human α-lactalbumin is transgenic for the human gene. Not all host animals in accordance with the invention are transgenic, though, as it is well within the possibilities of the invention, and it may be preferred in certain circumstances, to provide a host animal of a given species (for example cattle) carrying DNA encoding modified α-lactalbumin of the same species (in the example, bovine α-lactalbumin) . Such "quasi-transgenic" animals may be prepared by routine adaptations of the techniques which have been developed and will continue to be developed for the preparation of transgenic animals, or by the techniques of gene therapy, which are being developed for the replacement or supplementing of one version of a gene with another.

The mammary gland expression system has the advantages of high expression levels, low cost, correct processing and accessibility. Bovine and human α-lactalbumin have been produced in lactating transgenic animals by several workers (Vilotte et al , Eur. J. Biochem. 186 43-48

(1989) ; Hochi et al , Mol . Reprod. and Devel . 33 160-164

(1992); Soulier et al , FEBS Letters 297(1,2) 13-18 (1992)) and has been shown to produce high levels of protein. The α-lactalbumin promoter can direct the expression of heterologous genes (Stinnakre et al , FEBS Letters 284(1) 19-22 (1991)) to the lactating mammary gland (WO-A-8801648) .

Other transgenic animals, including transgenic mammals expressing heterologous protein in the milk of adult females, have been described in the literature. For example, WO-A-8800239 and WO-A-9005188 describe the production of transgenic mammals such as sheep which express pharmaceutically valuable proteins including factor IX and α-^antitrypsin. The methodology set out in all the above and other publications may be adapted to produce transgenic or quasi-transgenic animals in accordance with the present invention. It is preferred to use known dairy mammals, such as cows, sheep, goats or even camels or pigs, as hosts in the present invention, but the choice of host will be dictated primarily by convenience and not by any requirement of the invention itself.

Transgenic animals or other hosts according to the invention may be prepared by any convenient methodology. Accordingly, the invention is not limited to any particular method for their preparation. Transgenic animals, for example, may be prepared by microinjection, as described in WO-A-8800239 and WO-A-9005188 above, or they may be prepared by embryonic stem cell technology such as is described in WO-A-9003432.

Expression of the wild type α-lactalbumin gene of the host may be prevented or at least attenuated, so as to simplify purification.

According to a sixth aspect of the invention, there is provided a foodstuff suitable for sufferers of hyperphenylalaninemia, the foodstuff comprising at least one modified α-lactalbumin comprising no phenylalanine residues or fewer phenylalanine residues than wild type α-lactalbumin.

Preferred features of the modified α-lactalbumin(s) are as described above.

The foodstuff may, and often will, contain other components. The modified protein, alone or in combination with L-amino acids and/or other low or phenylalanine-free proteins or peptides, would be appropriate for use by infants, children and adults with PKU. For infants and very young children, a formula with low or phenylalanine-free α-lactalbumin as the protein base would look and taste more like normal whole protein infant formulas and would, as a result, be better tolerated than the current special dietary formulas of protein hydrolysate or amino acid mixtures.

A special dietary food or supplement for infants and very young children containing modified α-lactalbumin in accordance with the invention may contain other ingredients, such as a fat source, a carbohydrate source, and appropriate levels of vitamins, minerals and other nutritional factors. An additional nitrogen source besides the modified α-lactalbumin can also be any L- amino acid (such as L-tyrosine) , which would be added to improve the overall essential amino acid quality of the product to make it nutritionally adequate. In any case, the total phenylalanine content of the final product should generally be less than 80 mg, preferably less than 25 mg per 100 g powder equivalent, or phenylalanine free. Also any low-Phe or Phe-free peptide, such as casein- derived peptide (CDP) (derived from the cheese making process when the enzyme chymosin (rennet) splits the milk protein, α-casein, and releases CDP into the whey fraction) , would also be appropriate as an additional nitrogen source to enhance the overall quality of the resultant amino acid profile. L-tyrosine and non- essential amino acids may also be added in amounts up to and/or higher than that found in human milk and normal children's diets to ensure nutritional adequacy.

This special dietary food or supplement for infants and young children containing modified α-lactalbumin may be a ready-to-feed liquid, or in the form of a powder or concentrated liquid adapted to provide a ready-to-feed form by the additional of water and stirring. Each 100 ml (60-75 kcal/ml) of the ready-to-feed liquid formula or reconstituted powder will generally contain from 1.2 to 3.0 g, preferably about 1.3 to 1.5 g, of protein; from 2.2 to 4.0 g, preferably about 3.6 g, of a fat or mixture of fats; and from 6 to 9 g of carbohydrate.

The carbohydrate source lactose is generally preferred for most infants, but corn syrup solids, sucrose or other sugars can also be used in the special dietary food for infants and young children with PKU.

Additionally, special dietary food containing modified α- lactalbumin as low-Phe or Phe-free protein source would generally contain minerals to provide nutritionally acceptable quantities of calcium, phosphorus, potassium, sodium, chloride, magnesium, iron, copper, zinc, manganese, molybdenum, selenium, chromium, fluorine and iodine plus adequate quantities of vitamins such as vitamins A, D, E, B_1# B₂, B₆, B₁₂, C, nicotinamide, pantothenic acid, folic acid, vitamin K, biotin and choline. Other nutritional factors such as carotene, taurine, inositol, carnitine and nucleotides may also be added.

The amounts of each of the above nutrients and their acceptable chemical forms for inclusion into a special dietary food are provided in the U.S. Code of Federal Regulations (Part 181-184) and the U.S. Infant Formula Act of 1980, the U.S. Federal Food, Drug and Cosmetic Act, Section 2, Chapter IV, Section 192 Public Law 96- 359, September 26, 1980, Revised October 7, 1985. Similar but not exact recommendations are provided by the EC Commission Directive on Infant Formula and Follow-on Formula, Official Journal of the European Communities, No. L 175/35, 1991, and the Codex Alimentarius Commission, Recommended International Standards for Foods for Infants and Children. Joint FAO/WHO Food Standards Programme, CAC/RS, 72/74, 1967. Table 3 shows recommended minimum and maximum values and preferred values.

TABLE 3

"^•Calculated based on 66.67 kcal/100 m£

For infants and very young children, therefore, the ideal special dietary supplement should still resemble human breast milk, except for its phenylalanine content which should be less than 80 mg per 100 g formula or solids. It is imperative to continue low phenylalanine diet therapy throughout infancy, childhood and adolescence. Daily requirements are shown in Table 4 below: TABLE 4

All known essential amino acids (other than phenylalanine) , essential fatty acids, minerals, and vitamins should be provided in adequate amounts.

** Taken from: Modern Nutrition in Health and disease: Eds M. Shils, V. Young; 7th Ed. 1988, Lea and Febiger, Philadelphia.

The diets of older children, adolescents and pregnant women in particular with PKU might be comprised of a low -Phe or Phe-free formula plus more than 50% of total calories from low phenylalanine foods. Such a formula can be specifically designed for each age group, ie, with recommended daily intakes in mind, or one common formula can be developed with specific usage directions and dilution differences for each category. Such a formula for those other than infants and very young children would include (approximately) the following RDI levels of nutrients in about 500-1000 kcal or 500-1000 ml, as shown in Table 5. TABLE 5

For children and adults, including pregnant women, the formulation goal of a special dietary supplement would be that the formula is palatable and a high quality component to the remainder of the diet. Such a formula should contain all essential amino acids (other than phenylalanine) , and those vitamins and minerals in particular which are not typically present in high amounts in low protein foods, such as fruits and vegetables. Such nutrients, besides protein, are calcium, iron, folic acid, magnesium and trace minerals.

A typical formula for children, adults and pregnant women with PKU would contain but not be limited to the following preferred ranges of nutrients, as shown in Table 6.

TABLE 6

Low-phenylalanine or phenylalanine-free α-lactalbumin would also be highly appropriate for inclusion in milk drinks for older children, pregnant women, adolescents and adults with PAH deficiencies.

Alternately, such a modified α-lactalbumin or dietary food supplement containing modified α-lactalbumin could be used in foods, such as breads, ice cream and frostings, as sources of functional protein; they would still be free of phenylalanine, or at least of reduced phenylalanine content.

The milk of transgenic or quasi-transgenic host animals as described above may be suitable as foodstuffs in accordance with the invention with or without further modification, particularly if expression of the host's wild type α-lactalbumin gene has been attenuated or prevented. Further processing of the modified protein could involve removal of the amino acid tail for regulatory or other purposes, for example, prior to its incorporation into a formula.

According to a seventh aspect of the invention, there is provided a method of feeding a sufferer of hyperphenylalaninemia, the method comprising administering to a sufferer of hyperphenylalaninemia one or more foodstuffs containing modified α-lactalbumin comprising fewer phenylalanine residues than wild type α- lactalbumin.

Again, preferred features of the modified α-lactalbumin are as described in relation to the first aspect of the invention above.

Preferred features of the foodstuff are as described in relation to the second aspect of the invention above. More generally, features of each aspect of the invention are as for each of the other aspects, mutatis mutandis .

The amount and timing of the administration may be left to the discretion of the sufferer, but will generally be within the guidance of a physician, nutritionist or other medical adviser.

The invention will now be illustrated by the following non-limiting examples. The examples refer to the accompanying drawings, in which: Figure 1 is discussed in Example 1 and shows a restriction map of two overlapping genomic λ clones for the human α-lactalbumin gene;

Figure 2 is discussed in Example 1 and shows human α-lactalbumin transgene constructs;

Figure 3 is discussed in Example 1 and shows SDS-PAGE analysis of skimmed milk from human α-lactalbumin transgenic mice run against non transgenic mouse milk;

Figure 4 is discussed in Example 1 and shows a Western blot of the milk from human α-lactalbumin transgenic mice run against human α-lactalbumin standard;

Figure 5 is discussed in Example 2 and shows the sequence of bovine α-lactalbumin PCR primers;

Figure 6 is discussed in Example 2 and shows position of bovine α-lactalbumin PCR primers and products;

Figure 7 is discussed in Example 3 and shows the sequence of bovine α-lactalbumin PCR primers used for mutagenesis;

Figure 8 is discussed in Example 4 and 5 and shows PCR strategies for Phe substitutions and cloning strategy for transgene constructs PKU1 to 4;

Figure 9 is discussed in Example 4 and shows Western analysis of milk from PKU1 to 4 transgenic mice run against control mouse milk and bovine α-lactalbumin standard;

Figure 10 is discussed in Example 4 and shows quantitation by Western blot analysis of milk from two PKU1 transgenic mouse lines run against dilutions of bovine α-lactalbumin standard;

Figure 11 is discussed in Example 5 and shows PCR strategies for Phe substitutions and cloning strategy for transgene constructs PKU5 and PKU1H;

Figure 12 is discussed in Example 6 and shows cloning strategy for bovine α-lactalbumin and PKU5 to 10 constructs for expression in COS cells; and

Figure 13 is discussed in Example 6 and shows an autoradiograph of immunoprecipitated material from the supernatants of COS cells transfected with pC-BOVA and pC-PKU5 to 10 constructs.

EXAMPLE 1 - Cloning and Expression of Human α-lactalbumin in Transσenic Mice

Cloning of the Human α-lactalbumin Gene The DNA sequences of human α-lactalbumin has been published (Hall et al . , Biochem. J. 242 735-742 (1987)). Using the human sequence, PCR primers were designed to clone two small fragments from human genomic DNA, one at the 5' end of the gene and the other at the 3' end. These were subcloned into the pUC18 vector and used as probes to screen a commercial (Stratagene) λ genomic library. Two recombinant bacteriophages, 4a and 5b.1, which contained the α-lactalbumin gene, were isolated by established methods (Sambrook et a . , Molecular Cloning 2nd Ed., Cold Spring Harbor Laboratory (1989)). Restriction mapping demonstrated that both clones contained the complete coding sequence for the human α-lactalbumin gene but differed in the amount of 5' and 3' sequences present (Figure 1).

Sequence analysis of exons from clone 5b.1 and partial sequencing of clone 4a showed that these clones were identical to the published sequence.

Transgene Constructs (Figure 2) pHA-1 consists of the human a-lactalbumin coding region, -1.8kb of 5' flank and 3kb of 3' flank derived from 1 clone 5b.1 cloned as a 7kb EcoRI/Sall fragment into pUC18.

pOBHA (ovine S-lactoglobulin, human α-lactalbumin) was constructed from four DNA fragments:

(1) a 4.2kb Sall/EcόRV fragment containing the ovine jβ-lactoglobulin promoter (WO-A-9005188) ;

(2) a 74bp synthetic oligonucleotide corresponding to a 8bp Bell linker and bases 15-77 of the human α-lactalbumin sequence used as a blunt Bgll fragment;

(3) a 6.2kb Bgll/Pstl human α-lactalbumin fragment derived from 1 clone 4a comprising a region between a Bgll site at base 77 and a Xhol site in the 3' flank; and

(4) pSL1180 (Pharmacia) cut with Pstl and Sail. Human α-lactalbumin Expression in Transgenic Mice

Two constructs, namely pHA-1 (containing the human α-lactalbumin gene and flanking regions) and pOBHA (containing the human α-lactalbumin gene driven by the ovine .-lactoglobulin promoter) , were injected into mouse embryos and gave rise to transgenic animals.

The expression levels of human α-lactalbumin in the milk from these mice ranged from undetectable to -3mg/ml. Table 7 gives a summary of the relative amount of the transgenic protein. Skimmed milk from these animals was analysed by SDS-PAGE stained with Coomasie blue, isoelectric focusing, Western blots visualised with a commercial anti-human α-lactalbumin antibody (Sigma) and chromatofocusing. The results from these analyses showed that the transgenic protein was of the correct size, pi and antigenicity when compared to a human α-lactalbumin standard (Sigma) .

The results from several mice are shown in Figures 3 and

,4. Figure 3 shows an SDS-PAGE analysis of skimmed transgenic mouse milk run against a non-transgenic control mouse milk. Figure 4 shows a Western blot of these milks run against a human α-lactalbumin standard.

Table 7

Milk analyses from transgenic mice

(Table 7 shows the relative levels of human α-lactalbumin in transgenic mouse milk as estimated by comparison with protein standards on Coomassie gels and Western blots.

= <0.5mg/ml + = ~0.5-lmg/ml ++ = -l-2mg/ml +++ = ~2-3mg/ml n.d.= not determined)

EXAMPLE 2 - Cloning and Expression of Bovine α-lactalbumin in Mice Cloning bovine α-lactalbumin There are three known variants of bovine α-lactalbumin, of which the B form is the most common. The A variant from Bos (Bos) nomadicus f . d. indicus differs from the B variant at residue 10: Glu in A is substituted for Arg in B. The sequence difference for the C variant from Bos (Bibos) javanicus has not been established (McKenzie & White, Advances in Protein Chemistry 41 173-315 (1991) . The bovine α-lactalbumin gene (encoding the B form) was cloned from genomic DNA using the PCR primers indicated in Figure 5. The primers have been given the following SEQ ID NOs:

The source of DNA in all the PCR reactions was blood from a Holstein-Friesian cow.

The length of the amplified promoter region using primer

Ba-7 in combination with primer Ba-8 is 2.05kb. This BamHI/_-CθRI fragment was cloned into Bluescript (pBA-P2) .

The entire bovine α-lactalbumin gene including 300bp of 3'-flanking region was amplified using primer Ba-9 in combination with primer Ba-2. These primers include Bamϋl restriction enzyme recognition sites, which allowed direct subcloning of the amplified 3kb fragment into the BamHI site of pUC18, giving rise to construct pBA-G3.

Ligation of the BamΑI/EcoRV fragment from clone pBA-P2 to the EcoRV/BamHI fragment of BA-G3 gave rise to construct pBA (Figure 6) .

Since Taq polymerase lacks proof-reading activity, it was essential to ensure that the amplified bovine α- lactalbumin DNA was identical to the published bovine α- lactalbumin gene. Sequence analysis was carried out across all the exons and the two promoter fragments. Comparison of the bovine α-lactalbumin exons with those published by Vilotte showed 3 changes: (1) Exon I at +759: C to A; 5' -non-coding region. This change has also been reported by Hochi et al . , Mol . Reprod . and Devel . 33 160-164 (1992) .

(2) Exon I at +792: CTA to CTG; both code for Leucine.

(3) Exon II at +1231: GCG to ACG; Alanine to Threonine. This is indicative of the more common B form of the protein.

Although misreading of sequence during the PCR amplification can not be ruled out, the above mismatches were probably due to the difference in the source of bovine DNA.

Bovine α-lactalbumin Expression in Transgenic Mice

The construct pBA (Figure 6) was injected into mouse embryos and has given rise to 9 transgenic animals. Milk analysis was carried out as described in Example 1. Analysis by SDS-PAGE gel stained with Coomasie blue and comparison with standard amounts of α-lactalbumin showed expression levels varying from non-detectable to ~0.5-lmg/ml.

EXAMPLE 3 - Site-directed Mutagenesis

SEQ ID NOs for PCR Primers

The following SEQ ID NOs have been assigned to the PCR primers used in this and subsequent examples:

PKU-6 SEQ ID NO:11

PKU-7 SEQ ID NO:12

PKU-8 SEQ ID NO:13

PKU-9 SEQ ID NO:14 PKU-10 SEQ ID NO:15

Substituting the Phe residues

The four phenylalanine (Phe) residues in bovine α- lactalbumin occur within the first 3 exons of the gene.

A set of seven oligos (PKU-PCR primers 1, 2, 2L, 3, 4, 7 and 8) was designed for Phe substitutions based on protein modelling, nutritional aspects, and on amino acid variants present in either native α-lactalbumin or lysozyme genes from different species. The sequence of the primers (complementary to either the coding or non-coding strand) , the predicted change of amino acid and the mutagenised site of each are shown in Figure 7. These oligos act as both PCR and mutagenic primers. PKU primer 1 was used in combination with 2 or 2L and 7 in combination with 8, giving rise to a 435bp amplified fragment (A) containing the first two Phe substitutions. PKU primer 3 and 4 or 9 and 10 would produce a PCR product (B) of 601bp containing the second two Phe substitutions (see Figure 8) .

The PCR products were subcloned for sequencing using the sites EcoRI / BairSil and Bamϋl/Xbal respectively. The Pvul site, which had been engineered into PKU primers 2, 2L, 3, 7, and 8, created a unique restriction site between amino acid 31 and amino acid 53, without changing the encoded amino acids in that region.

Amino Acid Tail at C Terminus The addition of extra amino acids at the C-terminus of the α-lactalbumin gene can assist separation from endogenous α-lactalbumin protein. The purpose was to determine whether additional amino acids would influence expression of the transgene.

poly-Arg Tail

The primers PKU-5 and 6 (see Figures 7 and 8) were used to amplify up the 3' end of the bovine α-lactalbumin gene giving rise to the 0.44kb PCR product C. Primer PKU 6 contained the coding region for six Arg residues, immediately following the last natural Leu amino acid in the bovine α-lactalbumin sequence and created unique Sail and SnaBI sites. The unique Sall/SnaBI sites can be digested, and linkers inserted, to create any length of poly-Arg or poly-Arg/Lys tail that is required. The native stop codon and -30bps of 3' sequence followed, ending in a BspMI restriction enzyme site.

EXAMPLE 4 - Assembly and Expression of Bovine α-lactalbumin Constructs in vivo Constructs for Expression in the Mammary Gland of Transgenic Animals

Five constructs were made in all. The native bovine α- lactalbumin gene with a poly-Arg tail (pBARG-H) and two mutated bovine α-lactalbumin genes, with and without a poly-Arg tail (PKU-1 to PKU-4) . Phe mutations in these constructs were based on protein modelling and nutritional aspects. The design of the constructs is outlined in Figure 8.

pBARG-H was constructed from five DNA fragments:

(1) a 2.04kb ≤tøtl to Hpal fragment derived from pBA;

(2) a 1.25kb Hpal to HindiII fragment derived from pBA; (3) a 0.44kb HindiII to BspMI fragment derived from PCR product C;

(4) a 0.51kb BspMI to BgllI fragment derived from pB ; and (5) the vector pSL1180 digested with Sstl and

Bglll.

PKU-1 was constructed from six DNA fragments:

(1) a 2.04kb Sstl to Hpal fragment derived from pBA;

(2) a 0.46kb Hpal to Pvul fragment derived from PCR product A (PKU-primer pair 1 and 2) ;

(3) a 0.60kb Pvul to BsaBI fragment derived from PCR product B (primer pair 3 and 4) ; (4) a 0.22kb BsaBI to HindiII fragment derived from pBA;

(5) a 0.95kb Hindlll to Bglll fragment derived from pBA;

(6) the vector pSL1180 digested with Sstl and Bglll.

PKU-2 was constructed in the same way as PKU-1 with the exception of fragment (5) , which was derived from pBARG-H and contained the poly-Arg tail.

PKU-3 was constructed in the same way as PKU-1 with the exception of fragment (2) (PCR product A) , which had been amplified using PKU-primer 1 in combination with 2L.

PKU-4 was constructed in the same way as PKU-3 with the exception of fragment (5) , which was derived from pBARG-H and contained the poly-Arg tail. The amino acid substitutions resulting from the mutagenesis above are shown for plasmids PKU-1 to PKU-4 in Table 8.

Human α-lactalbumin 3'-Flanking Region

The 9.3kb BamHI fragment derived from λ clone 4a (Figure 1) contains 3' flanking region of the human α-lactalbumin gene and was included in all PKU constructs as a suitable target for preimplantation screening in transgenic embryos. The 9.3kb fragment was inserted into the unique Bglll site present in all the above constructs (see Figure 8) .

The DNA for microinjection was digested away from the vector sequences by cutting with BamHI. DNA was prepared for microinjection and transgenic mice generated as published (Hogan et al . , "Manipulating the Mouse Embryo:

A Laboratory Manual", Cold Spring Harbor Laboratory Press

(1986) ) .

Table 8

Amino Acid Substitutions present in Transgene Constructs n 3 ' flank Plasmid

pPKU-1 pPKU-2 pPKU-3 pPKϋ-4 pBARGh

Phenylalanine-free Bovine α-lactalbumin Expression in Transgenic Mice

Transgenic founder animals were milked at day 10 postpartum and the milk was analysed on Western Blots

(see Figure 9) . All constructs -- the wild type bovine α-lactalbumin pBARG-H and the four mutagenised bovine α-lactalbumin constructs (PKU1-4) -- were capable of expressing and secreting bovine α-lactalbumin protein in the milk. The strong upper band (see Figure 9) is probably due to a glycosylated variant; it is also present in pBARG-H (data not shown) . Table 9 shows the relative amount of phenyl-alanine free bovine α- lactalbumin protein in milk of these transgenic mice. For the protein quantitation serial dilutions of the milk samples were run on Western Blots and compared to bovine α-lactalbumin standards. Figure 10 shows this procedure as applied to samples from PKU-1 transgenic mice.

The protein is further analysed with a combination of established techniques such as ELISA, pi, sequencing and circular dichroism. These methods are intended to give information about the size and charge of the protein, and some idea of the folding.

Table 9

Expression of PKU 1-4 in milk of transgenic mice construct α-lac protein in transgenic milk pBARG-H 0-0.50mg/ml

PKU-1 0-0.25 mg/ml

PKϋ-2 0-0.50 mg/ml

PKU-3 0-0.10 mg/ml PKU-4 0-0.50 mg/ml

RNA Analysis

Total RΝA from transgenic mammary glands was analysed on Northern Blots using a bovine α-lactalbumin specific probe. As expected all samples from animals with detectable levels of recombinant protein showed expression of bovine α-lactalbumin mRNA. There was no clear relation between the amount of mRNA and transgenic protein present.

To verify that the expressed proteins were derived from mutant α-lactalbumin genes the bovine α-lactalbumin mRNA present in the extracted RNA was cloned by the method of RT-PCR and sequenced. The results proved that the mRNA was transcribed from the PKU1-4 constructs and that the translated protein should therefore be phenylalanine free.

EXAMPLE 5 - Expression of Mutagenised Bovine α-lactalbumin under the Control of the Human α-lactalbumin Promoter in vivo Data from Examples 1, 2 and 4 indicated that expression of the human α-lactalbumin transgene is considerably higher than that of the native bovine α-lactalbumin transgene, reflecting the difference in expression levels of the endogenous bovine and human genes. As this might contribute to differences in the 5' control region, the 5'region of the bovine α-lactalbumin transcriptional start site was substituted with sequences from the human α-lactalbumin gene.

Two constructs were made, namely PKU-5 and PKU-1H, which incorporate the amino acid substitutions shown in Table 10. The design of the constructs is outlined in Figure 11.

PKU-5: In a first cloning step three fragments were subcloned into the _?coRI/Ba_τ_HI site of pUC18:

(1) the EcoRI to Pvul fragment (fragment A, Fig. 11) derived by PCR amplification using PKU-primer 7 in combination with 8; (2) the Pvul to BsaBI fragment (fragment B, Fig. 11) derived by PCR amplification using PKU-primer 9 in combination with 10; and

(3) the BsaBI to Hindi11 fragment derived from pBA.

The final construct included 6 DNA fragments:

(1) the 3.7kb Sail to Kpnl fragment containing the human α-lactalbumin promoter derived from λ clone 4a (Figure 1) ;

(2) the 152bp synthetic oligonucleotide containing human α-lactalbumin sequences from the Kpnl site to the AUG and bovine α-lactalbumin sequences from the AUG to the Hapl site;

(3) the 1.25kb Hpal to Hindi11 fragment from the first subcloning step;

(4) the 0.95kb HinduI to Bglll fragment derived from pB ;

(5) the 3.7kb BamHI to Xhol fragment from the 3' flank of the human α-lactalbumin gene derived from λ clone 4a (Figure 1) used as a BarriHI fragment; and

(6) a Bluescript KS- vector cut with Sail and BaznHI .

PKU-IH was constructed in the same way as PKU-5 with the exception of fragment (3) , which was derived from PKU-1 (Example 4) . Table 10

Amino Acid Substitutions present in Transgene Constructs

Substitutions Human promoter Plasmid Human 3 ' flank pos ' n 9 30 53 80

Tyr , Tyr , Tyr , Tyr + pPKU- 1H

Ser, yr, eu, Leu + pPKU-5

Expression in transgenic mice

The two constructs PKU-IH and PKU-5 have been injected into mouse embryos . So far transgenic animals were derived for the PKU-5 construct . These animals are set up for breeding to allow milk analysis .

EXAMPLE 6 - Expression of Mutagenised Bovine α-lactalbumin in vi tro

Constructs for Transient Expression

To analyse a large range of mutations of the α- lactalbumin gene, an in vi tro expression assay was established. For this, the coding region of the native bovine and mutated bovine α-lactalbumin genes (Examples 2, 3, 4, and 5; sequence position 756 to 3030; Vilotte et al . , Biochemie 69 , 609-620 (1987)) were cloned into the COS cell expression vector pcDNA3 (Invitrogen) giving rise to pC-BOVA and pC-PKU-5 to -10.

The details of construction are outlined with reference to Figure 12, as follows:

(1) a synthetic oligonucleotide containing bovine α-lactalbumin sequences from position 756 to the Hpal site at position 831 was cloned into the Hindlll/BamHI site of the vector pCDNA3; (2) the coding region of the bovine α-lactalbumin gene was inserted as a Hpal to BarπHI fragment, giving rise to construct pC-BOVA;

(3) the 1.25kb Hpal to HindiII fragment in pC-BOVA was exchanged for the 1.25kb Hpal to Hindlll fragment containing the Phe substitutions (Figure 11) giving rise to constructs pC-PKU-5 to -10.

The presence of the restriction sites PpuMI between amino acids 9 and 30 and Avrll between amino acids 53 and 80 (Figure 11) allowed generation of PKU constructs which have either one or two of the four phenylalanines substituted as described in Table 11.

Table 11

Amino Acid Substitutions present in COS Cell Constructs

Substitutions Plasmid pos'n 9 30 53 80

Phe,Phe,Phe,Phe pC-BOVA

Ser,Tyr,Leu,Leu pC-PKU-5

Phe,Tyr,Leu, he pC-PKU-6

Tyr,Phe,Phe,Tyr pC-PKU-7 Phe,Phe,Phe,Tyr pC-PKU-8

Phe,Phe,Phe, eu pC-PKU-9

Ser,Phe,Phe,Phe pC-PKU-10

Expression in COS cells Cells were transfected with lOμg of plasmid; after 72hrs proteins were ³⁵S-methionine and -cysteine labelled. Supernatant was collected and α-lactalbumin was analysed on polyacrylamide gels after immunoprecipitation.

Figure 13 shows that transient expression in COS cells leads to secretion of native and mutated bovine α-lactalbumin. Amino acid substitution Phe₉ to Ser (pC-PKUlO) gave expression levels equivalent to wild type bovine α-lactalbumin. Substitution Phe₃₀ to Tyr in combination with substitution Phe₅₃ to Leu (pC-PKU6) , and substitution Phe₈₀ to Tyr (pC-PKU8) had reduced expression levels. Expression of pC-PKU5, 7, and 9 was below the level of detection in this assay system. This system will allow the definition of amino acid substitutions for maximum expression in transgenic animals.

EXAMPLE 7 - How to use Milk Containing Modified Protein For children with phenylketonuria, a special dietary supplement free of phenylalanine is required. Since dietary compliance is critical during the early years, any therapeutic formula which tastes good and enhances dietary compliance is preferred.

In planning a diet for the phenylketonic child, it is necessary to determine the daily protein, calorie and amino acid requirements (including phenylalanine) to support age and growth requirements, calculate the protein and calories provided by the special dietary supplement, and then determine what natural foods are needed to meet but not exceed the phenylalanine requirement. Normally such provision for phenylalanine is made by the addition of measured amounts of milk and other regular foods.

A typical daily menu for children with phenylketonuria would contain up to 3 servings/day of a special supplement containing modified α-lactalbumin and other nutrients in Table 3 described previously. The current recommendation is that the restricted phenylalanine diet, adjusted regularly to meet age, growth and maintenance requirements, be continued for life.

EXAMPLE 8 - Preparation/Isolation of Modified Protein from Milk For most if not all cases of special formula preparation it will be desirable to separate and purify regular α- lactalbumin from modified α-lactalbumin protein. For children, adolescents and pregnant women, for example, formulas zero or the least amount of Phe would allow for greatest dietary flexibility and make adherence to lifelong dietary management of PKU most likely.

The attachment of the poly-Arg or Arg/Lys tail to the C- terminus was therefore designed to allow more efficient and cost effective separation. Such a design should alter the charge on the modified protein so that it can be separated from regular α-lactalbumin. Such separation should be possible by laboratory and industrial scale whey fractionation and subsequent liquid chromatography techniques, the latter separation possible on the basis of molecular charge differences. Liquid chromatography have already shown promise in the separation and purification of proteins and other ingredients from dairy whey.

Other techniques may also be applicable particularly for laboratory or bench top characterisations. These could include but not be limited to SDS-PAGE chromatography and radioimmunoassay.

EXAMPLE 9 - Formulation of Modified Protein From this invention it will be possible to formulate several special dietary formulas for the treatment of hyperphenylalaninemia. These formulas are unique from all current formulas because a whole protein - modified α-lactalbumin - is the predominant protein (N) and Phe- free amino acid source, whereas, in current formulas, the N source is exclusively in the form of L-amino acids or extensively hydrolysed casein protein.

The predominant whole protein nature of these new therapeutic formulas will enhance dietary compliance because taste and palatability is greatly improved; pure mixtures of L-amino acids and peptides are typically bitter and poorly tolerated. It is therefore advantageous to produce a series of formulas for dietary treatment throughout the lifespan, starting from infancy and continuing through pregnancy and adulthood.

These formulas, while containing the same modified α- lactalbumin source, will vary in their total protein content, the addition of L-amino acids to improve nutritional score, non-protein sources of energy, such as sugars, corn syrups and/or fats, and vitamins, minerals and flavours, if appropriate. Obviously a special dietary formula for infants will attempt to match a human milk profile or its physiologic effects, and as a result, will contain a large percentage of fat calories. A formula for children and adults will contain more total protein and less fat, if any, plus differing ratios of vitamins and minerals than that required for infants and young children. Natural foods to be taken in combination with formula for these non-infant groups are typically arranged in food groups with similar phenylalanine levels and can therefore be exchanged one for the other to further enhance diet variety and, therefore, compliance. SEQUENCE LISTING

(1) GENERAL INFORMATION:

(i) APPLICANT FOR ALL DESIGNATED STATES EXCEPT US:

(A) NAME: PHARMACEUTICAL PROTEINS LIMITED

(B) STREET: ORCHARD BRAE HOUSE, 30 QUEENSFERRY ROAD

(C) CITY: EDINBURGH (E) COUNTRY: UNITED KINGDOM

(F) POSTAL CODE (ZIP) : EH4 2HG

(i) INVENTORS/APPLICANTS FOR US ONLY:

(A) NAME: COLMAN, Alan (B) STREET: 65 CROMWELL LANE

(C) CITY: COVENTRY

(E) COUNTRY: UNITED KINGDOM

(F) POSTAL CODE (ZIP) : CV4 8AQ (A) NAME: WRIGHT, Gordon

(B) STREET: 168 ESKHILL

(C) CITY: PENICUIK

(D) STATE: MIDLOTHIAN

(E) COUNTRY: UNITED KINGDOM (F) POSTAL CODE (ZIP) : EH26 8DQ

(A) NAME: SAWYER, Lindsey

(B) STREET: 14 MORNINGSIDE PARK

(C) CITY: EDINBURGH (E) COUNTRY: UNITED KINGDOM

(F) POSTAL CODE (ZIP) : EH10 5HB

(A) NAME: RIGDEN, Daniel John

(B) STREET: 38 DENVER AVENUE (C) CITY: CREWE

(D) STATE: CHESHIRE

(E) COUNTRY: UNITED KINGDOM

(F) POSTAL CODE (ZIP) : CW2 7PX (ii) TITLE OF INVENTION: MODIFIED ALPHA-LACTALBUMIN

(iii) NUMBER OF SEQUENCES: 15

(iv) COMPUTER READABLE FORM: (A) MEDIUM TYPE: Floppy disk

(B) COMPUTER: IBM PC compatible

(C) OPERATING SYSTEM: PC-DOS/MS-DOS

(D) SOFTWARE: Patentln Release #1.0, Version #1.25 (EPO)

(2) INFORMATION FOR SEQ ID NO: 1:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 27 base pairs (B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 1: GCGGATCCAC AACTGAAGTG ACTTAGC 27 (2) INFORMATION FOR SEQ ID NO: 2:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 35 base pairs

(B) TYPE: nucleic acid (C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 2: GATGGATCCT GGGTGGTCAT TGAAAGGACT GATGC 35

(2) INFORMATION FOR SEQ ID NO: 3:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 43 base pairs (B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 3: GCAGGCGAAT TCCTCAAGAT TCTGAAATGG GGTCACCACA CTG 43

(2) INFORMATION FOR SEQ ID NO: 4:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 33 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear (ii) MOLECULE TYPE: cDNA

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 4: GAGGATCCAA TGTGGTATCT GGCTATTTAG TGO 33

(2) INFORMATION FOR SEQ ID NO: 5: (i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 46 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 5:

GCTGAATTCG TTAACAAAAT GTGAGGTGTA TCGGGAGCTG AAAGAC 46

(2) INFORMATION FOR SEQ ID NO: 6:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 58 base pairs (B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 6: GCGGATCCGA TCGCTTGTGT GTCATAACCA CTGGTATGGT ACGCGGTACA GACCCCTG 58

(2) INFORMATION FOR SEQ ID NO: 7:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 58 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear (ii) MOLECULE TYPE: cDNA

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 7: GCGGATCCGA TCGCTTGTGT GTCATAACCA CTGGTATGGA GCGCGGTACA GACCCCTG 58 (2) INFORMATION FOR SEQ ID NO: 8: (i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 69 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 8:

GCGGATCCGA TCGTACAAAA CAATGACAGC ACAGAATATG GACTCTACCA GATAAATAAT 60 AAAATTTGG 69^' (2) INFORMATION FOR SEQ ID NO: 9:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 34 base pairs

(B) TYPE: nucleic acid (C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 9:

GCTCTAGATC ATCATCCAGG TACTCTGGCA GGAG 34

(2) INFORMATION FOR SEQ ID NO: 10:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 24 base pairs (B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA

( i) SEQUENCE DESCRIPTION: SEQ ID NO: 10: GCTGAAGCTT CACTTACTTC ACTC 24

(2) INFORMATION FOR SEQ ID NO: 11:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 65 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear (ii) MOLECULE TYPE: cDNA

(Xi) SEQUENCE DESCRIPTION: SEQ ID NO: 11:

GCGGATCCAA AGACAGCAGG TGTTCACCGT CGACGACGCC TACGTAACTT CTCACAGAGC 60 CACTG 65

(2) INFORMATION FOR SEQ ID NO: 12:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 46 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 12: GCTGAATTCG TTAACAAAAT GTGAGGTGAG CCGGGAGCTG AAAGAC 46 (2) INFORMATION FOR SEQ ID NO: 13:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 54 base pairs

(B) TYPE: nucleic acid (C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: CDNA (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 13:

GCGGATCCGA TCGCTTGTGT GTCATAACCA CTGGTATGAT ACGCGGTACA GACC 54

(2) INFORMATION FOR SEQ ID NO: 14:

(i) SEQUENCE CHARACTERISTICS: (A) LENGTH: S9 base pairs (B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: cDNA

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 14: GCGGATCCGA TCGTACAAAA CAATGACAGC ACAGAATATG GACTCCTCCA GATAAATAAT 60 AAAATTTGG 69

(2) INFORMATION FOR SEQ ID NO: 15: (i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 34 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: CDNA

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 15:

GCTCTAGATC ATCATCCAGC AGCTCTGGCA GGAG 34

Claims (28)

1. A modified α-lactalbumin comprising no phenylalanine residues or fewer phenylalanine residues than wild type α-lactalbumin, other than a variant of human α- lactalbumin in which Phe-80 is substituted by lysine.

2. A modified α-lactalbumin as claimed in claim 1, comprising an amino acid tail to the C-terminus to facilitate purification.

3. A modified α-lactalbumin as claimed in claim 2, wherein the C-terminal tail is poly-Arg or Arg/Lys.

4. A modified α-lactalbumin as claimed in claim 1, 2 or

3, which is modified bovine α-lactalbumin.

5. A modified bovine α-lactalbumin as claimed in claim

4, in which Phe₉ is substituted by Tyr, Ser or Leu.

6. A modified bovine α-lactalbumin as claimed in claim 4 or 5, in which Phe₃₁ is substituted by Leu, He, Trp or Tyr.

7. A modified bovine α-lactalbumin as claimed in claim 4, 5 or 6, in which Phe₅₃ is substituted by Tyr, Leu, Met or Trp.

8. A modified bovine α-lactalbumin as claimed in any one of claims 4 to 7, in which Phe 80 is substituted by

Tyr, Met, Trp or Leu.

9. A modified α-lactalbumin as claimed in claim 1, 2 or 3, which is modified human α-lactalbumin.

10. A modified human α-lactalbumin as claimed in claim 9, in which Phe₃ is substituted by Tyr, Leu, Arg, Met or Trp.

11. A modified human α-lactalbumin as claimed in claim 9 or 10, in which Phe₃₁ is substituted by Leu, He, Trp or Tyr.

12. A modified human α-lactalbumin as claimed in claim 9, 10 or 11, in which Phe₅₃ is substituted by Tyr, Leu,

Met or Trp.

13. A modified human α-lactalbumin as claimed in any one of claims 9 to 12, in which Phe₈₀ is substituted by Tyr, Met, Trp or Leu.

14. A process for the preparation of a modified α- lactalbumin as claimed in any one of claims 1 to 13, the process comprising coupling together successive amino acids, and/or ligating oligo- and/or poly-peptides.

15. An isolated or recombinant nucleic acid encoding a modified α-lactalbumin comprising no phenylalanine residues or fewer phenylalanine residues than wild type α-lactalbumin.

16. An isolated or recombinant nucleic acid encoding a modified α-lactalbumin as claimed in any one of claims 1 to 13.

17. Recombinant DNA as claimed in claim 15 or 16, which is in the form of a vector.

18. A process for the preparation of nucleic acid as claimed in claim 15 or 16, the process comprising coupling together successive nucleotides and/or ligating oligo- and/or polynucleotides.

19. A host containing recombinant DNA as claimed in claim 15, 16 or 17.

20. A host as claimed in claim 19, which is a placental non-human mammal whose germline includes DNA encoding modified α-lactalbumin, adult females of the line being capable of expressing modified α-lactalbumin in the mammary gland such that the modified α-lactalbumin accumulates in the milk.

21. A host mammal as claimed in claim 20, which is a cow (or bull) , sheep, goat, camel or pig.

22. A host as claimed in claim 19, 20 or 21, wherein expression of the wild type α-lactalbumin gene of the host is prevented or attenuated.

23. A foodstuff suitable for sufferers of hyperphenylalaninemia, the foodstuff comprising no phenylalanine residues or at least one modified α- lactalbumin comprising fewer phenylalanine residues than wild type α-lactalbumin.

24. A foodstuff as claimed in claim 23, wherein the modified α-lactalbumin, or at least one of the modified α-lactalbumins, is as claimed in any one of claims 2 to 13.

25. A foodstuff as claimed in claim 23 or 24, which is a dietary supplement or formula suitable for infants or children with hyperphenylalaninemia.

26. A foodstuff as claimed in claim 23 or 24, which is the milk of a female mammal as claimed in claim 19, 20 or 21.

27. A method of feeding an individual with phenylketonuria, the method comprising administering to the individual one or more foodstuffs containing modified α-lactalbumin comprising no phenylalanine residues or fewer phenylalanine residues than wild type α- lactalbumin.

28. A method as claimed in claim 27, wherein the foodstuff is as claimed in any one of claims 23 to 26.