Gene 334 (2004) 105 – 111 www.elsevier.com/locate/gene

The goat as1-casein gene: gene structure and promoter analysis Luigi Ramunno a,*, Gianfranco Cosenza a, Andrea Rando b, Rosa Illario a, Daniela Gallo a, Dino Di Berardino a, Piero Masina b a

Dipartimento di Scienze Zootecniche e Ispezione degli Alimenti, Universita` degli Studi di Napoli ‘‘Federico II’’, Via Universita` 133, 80055 Portici (Na), Italy b Dipartimento di Scienze delle Produzioni Animali, Universita` degli Studi della Basilicata, Potenza, Italy Received 11 November 2003; received in revised form 24 February 2004; accepted 5 March 2004 Available online 5 May 2004 Received by A. Sippel

Abstract The level of as1-casein in goat milk shows strong variations determined by at least 15 alleles associated with four different efficiencies of protein synthesis. The nucleotide sequence of the whole goat as1-casein-encoding gene (CSN1S1) plus 1973 nucleotides at the 5Vflanking region and 610 nucleotides at the 3V flanking region was determined and aligned with its bovine counterpart. The gene is spread over 16.7 kb and consists of 19 exons varying in length from 24 bp (exons 5, 6, 7, 8, 10, 13 and 16) to 385 bp (exon 19) and 18 introns from 90 bp of intron 10 to 1685 bp of intron 2. Furthermore, highly conserved sequences, mainly located in the 5V flanking region, were found between this gene and other casein-encoding genes. Finally, seven interspersed repeated elements (10 in the bovine CSN1S1 gene) were also identified at four different locations of the sequenced region: 5V untranscribed region and introns 2, 8 and 11. D 2004 Elsevier B.V. All rights reserved. Keywords: Capra hircus; Casein; CSN1S1 gene; Nucleotide sequence

1. Introduction In ruminants, the four caseins (as1, h, as2 and n) represent about 80% of milk proteins. They are characterised by specific properties such as a low solubility at pH 4.6 and an organisation in clusters of protein chains, called micelles. Furthermore, three (as1, h and as2) of the four caseins are sensitive to calcium precipitation, show similar molecular weights (around 24 kDa), promoter regions, leader peptide sequences and locations of the major phosphorylation site. These data support the hypothesis of a common evolutionary origin of these genes from the duplications of a unique ancestral gene (Rijnkels, 2002). Abbreviations: CSN1S1, as1 casein encoding gene; cDNA, DNA complementary to RNA; mrNA, messenger RNA; PCR, polymerase chain reaction; kDa, kilodalton; SINE, short interspersed nuclear element; LINE, long interspersed nuclear element; Bov-B, Bov-A2, Bov-tRNA, Bovidae SINEs; Tris – HCl, tris(hydroxymethyl)aminomethane HCl; Na2EDTA, disodium ethylenediamine tetraacetate; TBE, Tris – Borate – EDTA; dNTPs, deoxynucleotide triphosphates. * Corresponding author. Tel.: +39-812-539-004; fax: +39-817-762-886. E-mail address: [email protected] (L. Ramunno). 0378-1119/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.gene.2004.03.006

The nucleotide sequences of mRNAs produced by as1, h, as2 and n-casein genes (CSN1S1, CSN2, CSN1S2 and CSN3, respectively) of several species are available. These genes have been mapped in the order CSN1S1, CSN2, CSN1S2 and CSN3 in a 250 kb (kilobase) DNA region of chromosome 6 (6q31) in Bos taurus. Furthermore, in this species the genomic sequences of the four casein genes have been identified. In both cattle and goat the CSN1S1 and CSN2 genes are convergently transcribed and only 20 and 12 kb apart, respectively (Rijnkels, 2002). The calcium-sensitive casein genes show similar gene structures, with several small exons and a low exon/ intron ratio. Both CSN1S1 and CSN1S2 genes show a relatively large transcriptional unit, about 17.5 and 18.5 kb, respectively, with a similar number of exons. CSN2 is approximately half the size of the other calcium-sensitive casein genes and contains half the number of exons (Rijnkels, 2002). At present, in goat the cDNA sequences of the four casein genes (Rijnkels, 2002) and the whole genomic sequence of two alleles of the CSN2 gene (EMBL accession no. AJ011018 and AJ011019) are available. Only parts, in some cases few exons with flanking regions, of the genomic

106

L. Ramunno et al. / Gene 334 (2004) 105–111

sequences of the goat CSN1S1, CSN1S2 and CSN3 genes are available (Leroux et al., 1992; Ramunno et al., 2001; Lagonigro et al., 2001; Coll et al., 1995; Ward et al., 1997; Cosenza et al., 2003). In this paper, we report the complete nucleotide sequence of the gene encoding the goat as1casein (CSN1S1) and the comparison of its 5V region with the bovine counterpart. In goat, this locus shows a high level of polymorphism determined by at least 15 alleles associated with both qualitative and quantitative differences. In particular, at least four different synthesis efficiencies (3.5, 1.1, 0.45 and 0.0 g/l per allele) (Bevilacqua et al., 2002; Chianese et al., 1997; Martin et al., 1999) are observed and the mutations characterising the different alleles are quite different, from single nucleotide substitutions/deletions to large insertions/deletions (Cosenza et al., 2003; Bevilacqua et al., 2002; Chianese et al., 1997; Martin et al., 1999).

the amplified fragments were analysed by electrophoresis on 1% to 2% agarose gels (Biorad) in TBE buffer and stained with ethidium bromide. 2.4. DNA sequencing Before nucleotide sequencing, PCR products were purified with QIAquick columns (Qiagen). Nucleotide sequencing was carried out according to the dideoxynucleotide chain-termination technique (Sanger et al., 1977) by using a BigDyek Terminator cycle sequencing kit (Applied Biosystems, Warrington, UK) and an ABI PRISM 377-18 (Applied Biosystems, Foster City, CA) nucleotide sequencer.

3. Results 3.1. Structure of goat CSN1S1 gene

2. Materials and methods 2.1. DNA samples Genomic DNA of a goat homozygous for the CSN1S1A allele was extracted from leukocytes (Gossens and Kan, 1981) obtained from a blood sample collected using Na2EDTA as anticoagulant. 2.2. Primer design Primers for amplification and sequencing were designed by means of DNASIS-Pro software (Hitachi), using the cDNA sequence of the goat CSN1S1A allele (Leroux et al., 1992) and the complete sequence of the bovine CSN1S1A allele (Koczan et al., 1991) as templates. Additional primers, designed on newly determined intron sequences, were also used for sequencing. 2.3. PCR conditions DNA regions spanning from nucleotide 1973 to + 17411 of the CSN1S1 gene of a goat with A/A genotype were amplified using a Gene Amp PCR System 2400 (Perkin Elmer). A typical 50-Al reaction mix comprised: 100 ng of genomic DNA, 50 mM KCl, 10 mM Tris –HCl (pH 9.0), 0.1% Triton X-100, 3 mM MgCl2, 200 nmol of each primer, dNTPs each at 400 AM, 2.5 U of Taq DNA Polymerase (Promega, Madison, WI), 0.04% BSA. The amplification programs consisted of 31 cycles. The first one characterised by denaturation at 97 jC for 2 min, annealing of primers at 46– 62 jC for 45 s and an extension step at 72 jC for 2 min. The next 30 cycles involved a denaturation step at 94 jC for 45 s, annealing at 46– 62 jC for 45 s and extension at 72 jC for 2 min with the exception that in the last cycle the extension time was 10-min long. All

By using genomic DNA as template, we sequenced the whole gene encoding the as1-casein (CSN1S1) plus 1973 nucleotides at the 5Vflanking region and 610 nucleotides at the 3V flanking region of a goat homozygous for the CSN1S1A allele (GenBank accession no. AJ504710). The goat CSN1S1 gene extends over 16 785 bp (base pairs) including 1138 bp of exonic regions and 15 647 bp of intronic regions with a total similarity with the corresponding bovine sequence of about 57%. The main feature of the goat CSN1S1 gene is the extremely split architecture. It contains 19 exons ranging in size from 24 (exons 5, 6, 7, 8, 10, 13 and 16) to 385 bp (exon 19) and 18 introns from 90 bp (intron 10) to 1685 bp (intron 2). The first exon (53 bp) is not coding at all. The whole leader peptide as well as the first two amino acids of the mature protein are encoded by exon 2 (63 bp) and the translation stop codon TGA is created by coupling the final nucleotides TG of exon 17 with the first nucleotide A of exon 18. All splice junctions follow the 5V GT/3VAG splice rule. A comparison of the sequence of the exons of the gene that we sequenced with the published sequence of the goat CSN1S1A allele cDNA (Leroux et al., 1992) shows five nucleotide differences, none of which results in an amino acid substitution (Fig. 1). 3.2. Goat CSN1S1 promoter analysis The organisation at the 5V and 3V ends is very similar to that of the other calcium-sensitive caseins. In particular, the goat CSN1S1 gene shows a GATA factor, a TATA Box and a polyadenylation signal (AATAAA) located, with reference to the first nucleotide of the first exon, at nucleotides 1217/ 1212, 23/ 29 and + 16656/ + 16661, respectively. In addition, the 5V end of the gene shows a 16-bp Milk Box motif (CTCCTCAGAATTTCTT, 142/ 157) (SchmittNey et al., 1991), the reverse complement (RC) of a Progesteron Receptor (PR, 125/ 132) (Bailly et al., 1986), two CCAAT/enhancer binding protein (C/EBP,

L. Ramunno et al. / Gene 334 (2004) 105–111

107

Fig. 1. Comparison between exons of the sequenced gene (upper line) with the published sequence of the goat CSN1S1A allele cDNA (Leroux et al., 1992). Numbers in vertical framed arrows indicate the beginning of exons. Dashes indicate nucleotides identical between the two sequences. The stop codon is symbolized by ***.

872/ 880 and 1306/ 1314) sites (Raught et al., 1995), a Mammary Cell-Activating Factor (MAF, 745/ 753), (Welte et al., 1994) and a putative binding site (GAATTCTTAGAATT) for Signal Transducer and Activator of Transcription 5 (STAT5) which mediates prolactin signal transduction in lactating mammary gland (Wakao et al., 1994). Other DNA cis-acting elements, such as the Simian Virus 40 (SV40)-type enhancer (Weiher et al., 1983), a Nuclear Factor Octamer-1 (NF Oct-1) site (Bohmann et al., 1987) and a YYI (Yin and Yang factor 1) common factor (Seto et al., 1991), are spread over the large sequence

GAAACCACARAATTAGCATNTT ( 43/ 64), which is conserved in all calcium sensitive casein genes and represents one of the motifs typical for milk protein gene promoters (Malewski, 1998). As already reported for the bovine CSN1S1 gene (Schild and Gendermann, 1996), two binding sites for the pregnancy-specific mammary nuclear factor (PMF) (Lee and Oka, 1992) (nucleotides 1132/ 1142 and + 8/ + 16) and three for the Activator Protein (AP-1) (nucleotides 1333/ 1339, 849/ 855 and 175/ 181) (Lee et al., 1987) have been observed. At least other 18 possible YY1 and two NF Oct-1 binding sites

108

L. Ramunno et al. / Gene 334 (2004) 105–111

can be observed. All these sequences are double-underlined and displayed by shaded bold letters in Fig. 2. 3.3. Artidactyla retroposons The goat CSN1S1 gene sequence is characterised by seven DNA elements showing similarity to artiodactyla

retroposons (Fig 3). Three of them are located in the 5V untranscribed region. In particular, the first (A) is located in the distal promoter region (from 1646 to 1976) and appears to be a ART-2 retroposon (Duncan, 1987), whereas the two others (B, from 1415 to 1638 and C, from 632 to 821) show strong similarities (79% and 80%, respectively) with full-length Bov-tA elements (Len-

Fig. 2. Homology between the nucleotide (nt) sequences of the 5V flanking region and exon 1 of goat (upper line) and cattle CSN1S1 gene. Numbering is relative to the first nucleotide of the first exon ( + 1) and dashes represent nt identical to those in upper lines. RC: Reverse Complement. Congruent and putative factors are double-underlined and in shaded bold letters, respectively. Motifs typical for milk protein gene promoters (designed A1 – A6) are in italic letters. Abbreviations: YY1, Yin and Yang common factor 1; PMF, Pregnancy-specific Mammary nuclear Factor; MPBF/STAT5, Milk Protein Binding Factor; MAF, Mammary Cell-Activating Factor; C/EBP, CCAAT/enhancer-binding protein; NF Oct-1, Nuclear Factor Octamer-1; PR, Progesteron Receptor; AP-1, Activator Protein.

L. Ramunno et al. / Gene 334 (2004) 105–111

109

Fig. 3. Schematic representation of the CSN1S1 gene and of the artiodactyla retroposons observed in cattle and goat.

stra et al., 1993) and imperfect tandem repeats at their ends. The D element, flanked by 12-bp direct repeats (ACCAAATATACT), is located in the second intron (from + 1888 to + 2157) and shows a similarity of 97.2% with the published Bov-A2 sequence (Lenstra et al., 1993). The G element is located in intron 8 (from + 6754 to + 7189) and is flanked by 11-bp direct repeats (TGGGCTATGCT). It shows a low level of similarity (about 68%) with the Art2 elements described by Duncan (1987) (EMBL accession no. Z25529). The last two artiodactyla retroposons, I (from + 9350 to + 9568) and L (from + 9521 to + 9839), are located in intron 11. The first shows a similarity of 81% with a Bov-tA element and the second a similarity of 75% with the aforementioned G element (Lenstra et al., 1993). The seven repetitive elements observed at the 5V and inside the goat CSNISI gene represent the 11.4% of the sequence deposited in the EMBL database. This figure raises to 14.7% in the bovine counterpart because of three extra repetitive elements, for a total of 687 bp, observed in this species (Fig. 3). As observed in the bovine CSNISI gene (Koczan et al., 1991), the regions containing the 10th (from + 8575 to 8711) and the 13th (from + 10677 to 10812) exons and flanking introns show a similarity of 96.2%. Therefore, these regions appear to be the result of a recent duplication supporting the hypothesis that exon duplication is a tool of casein gene assembly during evolution. Finally, different microsatellite sequences are present in the goat CSNISI gene. In particular, according to a prelim-

inary population analysis accomplished on 20 DNA samples belonging to two different breeds, differences in the number of the GT repeats located in the 18th intron, between nucleotides + 15937 and + 15960, are responsible for at least a diallelic polymorphism (Ramunno L., personal communication).

4. Discussion In this paper we provide the entire sequence of the goat CSN1S1 gene plus 5Vand 3Vflanking regions. Analyses of 5V- and 3V-flanking regions of the goat CSN1S1 gene provide an important contribution in order to evaluate the role and the importance of factors involved in the regulation of milk protein gene expression and the transcriptional effects of polymorphisms located in such regions (Ramunno L., manuscript in preparation). Knowledge of the transcriptional and translational effects of CSN1S1 gene polymorphisms will provide new opportunities to select the best dairy goats for the preferred milk protein producing genotype. Comparative analysis of the first 200 bp of the CSN1S1 promoter regions of different species shows, as expected, a homology between goat and other ruminants (similarities of about 96% with cattle, sheep and yak) stronger than that observed with non-ruminants (similarities of about 88% with rabbit, 80.5% with human and 77% with rat). The existence of six superfamily-specific motifs typical of milk protein gene promoters (designed A1 –A6) (Fig. 2), characterized by a full homology with the consensus

110

L. Ramunno et al. / Gene 334 (2004) 105–111

sequences reported by Malewski (1998), was confirmed. Motifs A5 and A6 are located in a distal region apart from A1 to A4 motifs which are located near the start of transcription. The region spanning A1 to A4 motifs contains binding sites for at least six transcription factors which could affect expression and regulation of goat as1-casein in the lactating mammary gland. In fact, progesterone receptor (PR), activating protein-1 (AP-1), CCAAT/enhancer-binding protein (C/EBP), signal transducer and activator of transcription 5 (STAT5, originally identified as milk protein binding factor, MPBF), and pregnancy-specific mammary nuclear factor (PMF) are involved in transcriptional activation, whereas Yin Yang (YY1) is involved in repression. Regarding the 3Vend of the goat CSN1S1 gene, at least two cis acting factors can be observed. The first element is the polyadenylation signal (AATAAA) found 18 bp upstream from the end, which is essential for the cleavage of the message (polyadenylation site: between a pyrimidine and an A residue, 17 F 3 bp from the polyadenylation signal). The other is a putative cis-acting element (a GT/Trich sequence) located further downstream (3V) from the cleavage. This sequence appears to be critical for the efficient cleavage of the hnRNA at the 3V processing site (Leroux et al., 1992; Christofori and Keller, 1988). On the whole, the goat CSN1S1 gene shares a similar organization with the bovine counterpart (Koczan et al., 1991), with some differences in intronic size. In fact, the analysis of the goat CSN1S1 gene evidenced an exon/intron size ratio higher (1:13.74) than that observed in cattle (1:14.38). The different ratio observed between the two species is mainly consequence of the three extra artiodactyla retroposons located in introns of cattle CSN1S1 gene (Fig. 3). All vertebrate genomes contain repetitive sequences. In particular, according to the nomenclature of Lenstra et al. (1993), the bovine genome contains three different kinds of dispersed repeats, namely Bov-B, Bov-A2 and Bov-tA, accounting for 0.5%, 1.8%, and 1.6% of the whole bovine genomic DNA, respectively. The Bov-B has been originally described as a SINE (Art2, PstI repeat) of around 560 bp in length characterised by a PstI repeat and a 78-bp segment homologous to Bov-A at the 3Vend (Duncan, 1987). Characterisation of the first full-length (3.1 kb) Bov-B (bovine dimer-driven family, BDDF) leads to its reclassification as a LINE (Szemraj et al., 1995). Malik and Eickbush (1998) confirmed that Art2 and PstI repeats are deletions of full-length LINEs and not target sites for Bov-B insertions as proposed by Szemraj et al. (1995). The Bov-tA (a heterodimer of a 73-bp tRNAGly pseudogene for a tRNA plus a Bov-A) and Bov-A2 (arose from duplication of Bov-A) repeats are the two most common SINEs in Bovidae (Lenstra et al., 1993). In particular, they are characterised by the common presence of the Bov-A repeat (115 bp in length), which is never found alone and was probably generated by the deletion of the central part of

the Bov-B element, so that the right and the left sequences became linked together (Okada et al., 1997). The extra elements that characterize the bovine sequence correspond to a Bov-A2 in intron 2 (E), a Bov-tA in intron 11 (H) and a possible truncated long interspersed repeat element (F) in intron 5. The first two elements probably arose from internal duplication phenomena of the adjacent D and I elements. As far as the third element (F) is concerned, its absence has been underlined also in six different goats and six sheep (Bevilacqua et al., 2002), confirming the phylogenetic proximity between these two species. On the whole, it appears that elements E, F and H are rather young insertions and their presence/absence can be considered a powerful phylogenetic marker for the clustering study of ruminants. The main differences observed between bovine and goat species are clustered in the central part of the gene (between introns 2 and 12). This observation supports the hypothesis of the definition in three parts of the CSN1S1 and CSN2 genes, which could have been assembled prior to evolutionary diversification of both genes (Koczan et al., 1991). A general consideration is that genomic expansion (the increase in size of introns and intergenic regions as a consequence of retroposons insertion) could be interpreted as an evolutionary strategy adopted in order to reduce the availability of own ‘‘vulnerable’’ targets (exons, promoters or other regulatory elements) to the effects of mutations and retronuon insertions (Brosius, 1999). Acknowledgements This work was supported by Cofinanziamento Programmi di Rilevanza Nazionale (MIUR). References Bailly, A., Le Page, C., Rauch, M., Milgrom, E., 1986. Sequence-specific DNA binding of the progesterone receptor to the uteroglobin gene: effects of hormone, antihormone and receptor phosphorylation. EMBO J. 5 (12), 3235 – 3241. Bevilacqua, C., Ferranti, P., Garro, G., Veltri, C., Lagonigro, R., Leroux, C., Pietrola`, E., Addeo, F., Pilla, F., Chianese, L., Martin, P., 2002. Interallelic recombination is probably responsible for the occurrence of a new alpha(s1)-casein variant found in the goat species. Eur. J. Biochem. 269 (4), 1293 – 1303. Bohmann, D., Keller, W., Dale, T., Scholer, H.R., Tebb, G., Mattaj, I.W., 1987. A trascription factor which binds to the enhancers of SV40, immunoglobulin heavy chain and U2 snRNA genes. Nature (Lond.) 325, 268 – 273. Brosius, J., 1999. Genomes were forged by massive bombardments with retroelements and retrosequences. Genetica 107, 209 – 238. Chianese, L., Ferranti, P., Garro, G., Mauriello, R., Addeo, F., 1997. Occurrence of three novel alpha s1-casein variants in goat milk. Milk Protein Polymorphism FIL-IDF Palmerston North, N. Z., 259 – 267. Christofori, G., Keller, W., 1988. 3Vcleavage and polyadenylation of mRNA precursors in vitro requires a poly(A) polymerase, a cleavage factor, and a snRNP. Cell 54, 875 – 889. Coll, A., Folch, J.M., Sanchez, A., 1995. Structural features of the 5V

L. Ramunno et al. / Gene 334 (2004) 105–111 flanking region of the caprine kappa-casein gene. J. Dairy Sci. 78 (5), 973 – 977. Cosenza, G., Illario, R., Rando, A., Di Gregorio, P., Masina, P., Ramunno, L., 2003. Molecular characterization of the goat CSN1S101 allele. J. Dairy Res. 70, 237 – 240. Duncan, C.H., 1987. Novel Alu-type repeats in artiodactyls. Nucleic Acids Res. 15, 1340. Gossens, M., Kan, Y.W., 1981. DNA analysis in the diagnosis of hemoglobin disorders. Methods Enzymol. 76, 805 – 817. Koczan, D., Hobom, G., Seyfert, H.M., 1991. Genomic organization of the bovine as1-casein gene. Nucleic Acids Res. 19, 5591 – 5596. Lagonigro, R., Pietrolla`, E., D’Andrea, C., Veltri, C., Pilla, F., 2001. Molecular genetic characterization of the goat as2-casein E allele. Anim. Genet. 23, 391 – 393. Lee, C.S., Oka, T., 1992. A pregnancy-specific mammary nuclear factor involved in the repression of the mouse beta-casein gene transcription by progesterone. J. Biol. Chem. 267 (9), 5797 – 5801. Lee, W., Mitchell, P., Tjian, R., 1987. Purified transcription factor AP-1 interacts with TPA-inducible enhancer elements. Cell 49, 741 – 752. Lenstra, J.A., Van Boxtel, A.F., Zwacgstra, K.A., Schwerin, M., 1993. Short interspersed nuclear element (SINE) sequences of the Bovidae. Anim. Genet. 24, 33 – 39. Leroux, C., Mazure, N., Martin, P., 1992. Mutations away from splice site recognition sequences might cis-modulate alternative splicing of goat as1-casein transcripts. Structural organization of the relevant gene. J. Biol. Chem. 267, 6147 – 6157. Malewski, T., 1998. Computer analysis of distribution of putative cis- and trans-regulatory elements in milk protein gene promoters. Biosystems 45, 29 – 44. Malik, H.S., Eickbush, T.H., 1998. The RTE class of non-LTR retrotransposons is widely distributed in animals and is the origin of many SINEs. Mol. Biol. Evol. 15, 1123 – 1134. Martin, P., Ollivier-Bousquet, M., Grosclaude, F., 1999. Genetic polymorphism of casein; a tool too investigate casein micelle organization. Int. Dairy J. 9, 163 – 171. Okada, N., Hamada, M., Ogiwara, I., Ohshima, K., 1997. SINEs and LINEs share common 3Vsequences: a review. Gene 205, 229 – 243.

111

Ramunno, L., Longobardi, E., Pappalardo, M., Rando, A., Di Gregorio, P., Cosenza, G., Mariani, P., Pastore, N., Masina, P., 2001. An allele associated with a non-detectable amount of as2 casein in goat milk. Anim. Genet. 32 (1), 19 – 26. Raught, B., Liao, W.S.L., Rosen, J.M., 1995. Developmentally and hormonally regulated CCAAT/enhancer-binding protein isoforms influence h-casein gene expression. Mol. Endocrinol. 9, 1223 – 1232. Rijnkels, M., 2002. Multispecie comparison of the casein gene loci and evolution of casein gene family. J. Mammary Gland Biol. 7, 327 – 345. Sanger, F., Nicklen, S., Coulson, A.R., 1977. DNA sequencing with chainterminating inhibitors. Proc. Natl. Acad. Sci. U. S. A. 74, 5463 – 5467. Schild, T.A., Gendermann, H., 1996. Variants within the 5V-flanking regions of bovine milk-protein-encoding genes. III. Genes encoding the Ca-sensitive caseins as1, as2 and h. Theor. Appl. Genet. 93, 887 – 893. Schmitt-Ney, M., Doppler, W., Ball, R.K., Groner, B., 1991. h-Casein gene promoter activity is regulated by the hormone-mediate relief of transcriptional repression and a mammary gland-specific nuclear factor. Mol. Cell. Biol. 11, 3745 – 3755. Seto, E., Shi, Y., Shenk, T., 1991. YY1 is an initiator sequence-binding protein that directs and activates transcription in vitro. Nature 354, 241 – 245. Szemraj, J., Plucienniczak, G., Jaworski, J., Plucienniczak, A., 1995. Bovine Alu-like sequences mediate transposition of a new site-specific retroelement. Gene 152, 261 – 264. Wakao, H., Gouilleux, F., Groenen, B., 1995. Mammary gland factor (MGF) is a novel member of the cytokine regulated transcription factor gene family and confers the prolactin response. EMBO J. 13, 2182 – 2191. Erratumu. EMBO J. 14 (4), 854 – 855. Ward, T.J., Honeycutt, R.L., Derr, J.N., 1997. Nucleotide sequence evolution at the kappa-casein locus: evidence for positive selection within the family Bovidae. Genetics 147 (4), 1863 – 1872. Weiher, H., Koning, M., Gruss, P., 1983. Multiple point mutation affecting the simian virus 40 enhancer. Science (Wash. D.C.) 219, 626 – 631. Welte, T., Philipp, S., Cairns, C., Gustafson, J.A., Doppler, T., 1994. Involvement of Ets-related proteins in hormone-indipendent mammary cell-specific gene expression. Eur. J. Biochem. 223, 997 – 1006.