Localization of the casein gene family to a single mouse chromosome

Diminished milk synthesis in upstream stimulatory factor 2 null mice is associated with decreased circulating oxytocin and decreased mammary gland expression of eukaryotic initiation factors 4E and 4G

Previous studies have suggested that upstream stimulatory factors (USFs) regulate genes involved with cell cycle progression. Because of the relationship of USFs to an important oncogene in breast cancer, c-myc, we chose to determine the importance of USF to normal mammary gland development in the mouse. Expression of USF in the mammary gland throughout development demonstrated only modest changes. Mutation of the Usf2 gene was associated with reduced fertility in females, but had no effect on prepartum mammary gland development. However, lactation performance in Usf2-/- females was only half of that observed in Usf2+/+ females, and both lactose and nitrogen were decreased in milk from Usf2-/- dams. This decrease was associated with diminished mammary tissue wet weight and luminal area by d 9 of lactation and with a decreased protein-DNA ratio. This decrease was associated with reduced abundance of the eukaryotic initiation factors eIF4E and eIF4G. Blood oxytocin concentrations on d 9 postpartum were also lower in Usf2-/- mice than Usf2+/+ mice. In contrast, the mutation had no effect on blood prolactin concentrations, mammary cell proliferation or apoptosis, mammary tissue oxytocin receptors, or milk protein gene expression. The mutation had only modest effects on maternal behavior. These data support the idea that USF is important to physiological processes necessary for the establishment and maintenance of normal lactation and suggest that USF-2 may impact lactation through both systemic and mammary cell-specific mechanisms.

Perturbation of beta1-integrin function in involuting mammary gland results in premature dedifferentiation of secretory epithelial cells

To study the mechanism of beta1-integrin function in vivo, we have generated transgenic mouse expressing a dominant negative mutant of beta1-integrin under the control of mouse mammary tumor virus (MMTV) promoter (MMTV-beta1-cyto). Mammary glands from MMTV-beta1-cyto transgenic females present significant growth defects during pregnancy and lactation and impaired differentiation of secretory epithelial cells at the onset of lactation. We report herein that perturbation of beta1-integrin function in involuting mammary gland induced precocious dedifferentiation of the secretory epithelium, as shown by the premature decrease in beta-casein and whey acidic protein mRNA levels, accompanied by inactivation of STAT5, a transcription factor essential for mammary gland development and up-regulation of nuclear factor-kappaB, a negative regulator of STAT5 signaling. This is the first study demonstrating in vivo that cell-extracellular matrix interactions involving beta1-integrins play an important role in the control of milk gene transcription and in the maintenance of the mammary epithelial cell differentiated state.

The mineralocorticoid receptor may compensate for the loss of the glucocorticoid receptor at specific stages of mammary gland development

To study the role of glucocorticoid receptor (GR) at different stages of mammary gland development, mammary anlage were rescued from GR-/- mice by transplantation into the cleared fat pad of wild-type mice. In virgin mice, GR-/- outgrowths displayed abnormal ductal morphogenesis characterized by distended lumena, multiple layers of luminal epithelial cells in some regions along the ducts, and increased periductal stroma. In contrast, the loss of GR did not result in overt phenotypic changes in mammary gland development during pregnancy, lactation, and involution. Surprisingly, despite the known synergism between glucocorticoids and prolactin in the regulation of milk protein gene expression, whey acidic protein and beta-casein mRNA levels were unaffected in GR-/- transplants as compared with wild-type transplants. That mineralocorticoid receptor (MR) might compensate for the loss of GR was suggested by the detection of MR in the mammary gland at d 1 of lactation. This hypothesis was tested using explant cultures derived from the GR-/- transplants in which the mineralocorticoid fludrocortisone was able to synergistically induce beta-casein gene expression in the presence of prolactin and insulin. These studies suggest that MR may compensate for the absence of GR at some, but not at all stages of mammary gland development.

Structure of the rabbit alphas1- and beta-casein gene cluster, assignment to chromosome 15 and expression of the alphas1-casein gene in HC11 cells

Several casein (CSN) genes (CSN1, 2, 10 and alphas2-CSN) have been described and shown to be clustered in mouse, man and cattle. These genes are expressed simultaneously in the mammary gland during lactation, but they are silent in most mammary cell lines, even in the presence of lactogenic hormones. However, it has been shown that the CSN2 gene, and this gene only, can be induced in certain mammary cell lines, such as HC11. In the present paper, we describe three overlapping bacterial artificial chromosome (BAC) clones which harbor both the rabbit CSN1 and CSN2 genes. These two genes are in a convergent orientation, separated by an intergenic region of 15 kb. DNA from one of the CSN/BAC clones was used as a probe for in situ hybridization to show that the CSN1 and CSN2 gene cluster is located on chromosome 15 band q23 and not on chromosome 12 as had been previously reported. Each of the three CSN/BAC DNAs was transfected into HC11 cells. In the presence of lactogenic hormones, the rabbit CSN1 gene was clearly expressed from all three CSN/BAC DNAs, whereas the rabbit CSN2 gene, which at the most possesses a 1 kb upstream region in one of the CSN/BAC DNAs, was not expressed at detectable levels on Northern blots. The transfected HC11 cells now express both rabbit CSN1 and mouse CSN2 genes. These transfected cells will be used as a model to study the role of CSN1 in milk protein secretion.

The poly(A) tail length of casein mRNA in the lactating mammary gland changes depending upon the accumulation and removal of milk

The length of casein mRNA from the lactating mouse mammary gland, as assessed on Northern blots, is shorter after weaning, but is elongated following the removal of milk. In order to investigate this phenomenon, the molecular structures of beta- and gamma-casein mRNAs were analysed. The coding and non-coding regions of the two forms were the same length, but the long form of casein mRNA had a longer poly(A) tail than the short form (P<0.05). In order to examine the stability of casein mRNA under identical conditions, casein mRNAs with the long and short poly(A) tails were incubated in the rabbit reticulocyte lysate (RRL) cell-free translation system. Casein mRNA with the long poly(A) tail had a longer half-life than that with the short tail (P<0.05). The beta- and gamma-casein mRNAs were first degraded into 0.92 and 0.81 kb fragments respectively. With undegraded mRNA, the poly(A) tail shortening by exoribonuclease was not observed until the end of the incubation. Northern blot analysis showed that casein mRNA with the long poly(A) tail was protected efficiently from endoribonucleases. We conclude that the length of the poly(A) tail of casein mRNA in the lactating mammary gland changes depending upon the accumulation and removal of the gland's milk, and we show that the longer poly(A) tail potentially protects the mRNA from degradation by endoribonucleases.

Cell-extracellular matrix interactions and EGF are important regulators of the basal mammary epithelial cell phenotype

The mammary epithelium is composed of a luminal epithelium and a basal layer containing myoepithelial cells and undifferentiated precursors. Basal cells express specific protein markers, such as keratin 14 (K14) and P-cadherin. To study the factors that regulate the basal mammary epithelial cell phenotype, we have established two clonal derivatives of the mouse HC11 cell line, BC20 and BC44, expressing high levels of K14 and P-cadherin. Unlike the parental HC11 cells, these basal cells did not produce beta-casein in response to lactogenic hormone treatment; however their phenotype appeared to be plastic. Cultured in EGF-free medium, they exhibited enhanced cell-extracellular matrix adhesions and deficient cell-cell junctions, whereas long-term treatment with EGF induced a decrease of focal contact number and establishment of cell-cell junctions, resulting in downregulation of K14 and P-cadherin expression at the protein and mRNA levels. To determine whether cell-extracellular matrix interactions mediated by integrins have a role in the regulation of the expression of K14 and P-cadherin, the amounts of transcripts for the two proteins were analysed in the basal cells, which were plated on the function-blocking antibodies against beta1 and alpha6 integrin chains, on fibronectin and on laminin 5. The amount of P-cadherin transcript was 2- to 4-fold higher in cells plated on the function-blocking anti-integrin antibodies and on the extracellular matrix proteins, as compared to cells plated on poly-L-lysine, whereas the K14 transcript levels were not significantly modified in response to adhesion. The data demonstrate that integrin-mediated cell interaction with extracellular matrix is directly implicated in the control of P-cadherin expression, and that EGF and cell-extracellular matrix adhesion events are important regulators of the basal mammary epithelial cell phenotype.

Polymorphic insertions/deletions of both 1550nt and 100nt in two microsatellite-containing, LINE-related intronic regions of the rabbit kappa-casein gene

The most frequent allele of the rabbit kappa-casein (kappa-Cas)-encoding gene (A allele) has previously been shown to possess two sequences similar to those found in the 5' end of long interspersed repeated elements (LINE). Part of an inverted rabbit LINE is present in the first intron and part of a direct rabbit LINE in the fourth intron. We describe herewith a less frequent allele (B allele) that lacks both 100bp in the first intron and 1550bp in the fourth intron. It was not possible to identify any allele exhibiting only one of the deletions in a population of 55 rabbits. The 100bp present in the first intron of the A allele, but absent from the B allele, are located at the 5' end of the inverse complementary LINE and include the poly (T) track of the LINE. The 1550bp present in the fourth intron of the A allele, but absent from the B allele, include the entire direct LINE sequence. Therefore, the B allele only possesses one partial LINE sequence that is located in the first intron and is truncated when compared to the copy found in the first intron of the A allele. The B allele might thus be more recent than the A allele. Differences between the sequences of transcripts corresponding to each allele are limited to two silent mutations and three modifications in the 3' UTR. In the mammary glands of lactating rabbits, which are homozygous for both alleles, kappa-Cas mRNA accumulate to similar levels and are translated into identical kappa-Cas that are secreted at similar concentrations into milk.

Perturbation of beta1-integrin function alters the development of murine mammary gland

The expression of a transgene coding for a chimeric molecule, containing the cytoplasmic and transmembrane domains of the beta1-integrin chain and the extracellular domain of the T-cell differentiation antigen CD4, was targeted to the mouse mammary gland by the mouse mammary tumor virus (MMTV) promoter. The chimera does not interact with the extracellular ligands; however, its expression in cultured cells was shown to interfere with focal adhesion kinase (FAK) phosphorylation following ligation of endogenous beta1-integrin. Therefore, expression of the transgenic protein on the cell surface should uncouple adhesion from intracellular events associated with the beta1-cytoplasmic domain and thus perturb beta1-integrin functions. Although most of the transgenic females were able to lactate, their mammary glands had a phenotype clearly distinct from that of wild-type mice. At mid-pregnancy and the beginning of lactation, transgenic glands were underdeveloped and the epithelial cell proliferation rates were decreased, while the apoptosis levels were higher than in wild-type glands. In lactation, the amounts of the whey acidic protein (WAP) and beta-casein gene transcripts were diminished, and the basement membrane component, laminin and the beta4-integrin chain accumulated at the lateral surface of luminal epithelial cells, revealing defects in polarization. Our observations prove that in vivo, beta1-integrins are involved in control of proliferation, apoptosis, differentiation and maintenance of baso-apical polarity of mammary epithelial cells, and therefore are essential for normal mammary gland development and function.

High-level expression of bovine alpha s1-casein in milk of transgenic mice

The bovine alpha s1-casein gene, isolated from a cosmid library, was introduced into the murine germline. Transgene expression occurred in all transgenic mice, and was confined to the lactating mammary gland. Half of the mouse lines (five out of ten) expressed at relatively high expression levels (> 1 mg ml-1). The highest levels of expression were obtained with a transgene containing 14.2 kb of 5' flanking sequence, in two cases expression levels comparable to (10 mg ml-1) or well above (20 mg ml-1) alpha s1-casein levels in bovine milk were obtained. Transcription initiation occurred at the same site in the bovine alpha s1-casein gene in transgenic mouse as in the cow. A marked induction of expression occurred at parturition rather than at mid-pregnancy, and thus resembled the bovine rather than the murine developmental expression pattern. Bovine alpha s1-casein specific immunoblotting and RIA were developed for characterization and quantification of the recombinant protein. Using these assays, the properties of the recombinant protein could not be distinguished from those of the natural bovine protein. In spite of the high-level tissue-specific and correctly regulated developmental expression of the transgene, expression levels were integration-site dependent. This may indicate that not all cis-acting regulatory elements involved in bovine alpha s1-casein expression were included in the transgene.

Physical mapping of the murine casein locus reveals the gene order as alpha-beta-gamma-epsilon-kappa

The murine casein locus has been characterized by long-range restriction mapping and the analysis of large fragment genomic clones. Cloned sequences from five mouse casein genes (alpha, beta, gamma, epsilon, kappa) were used to screen a murine (strain 129) genomic library in a bacterial artificial chromosome vector (BAC). Of the nine clones isolated, two contained three casein genes alpha, beta, gamma and gamma, epsilon, kappa, respectively. The following combinations were found in other clones: alpha + beta, beta + gamma; and gamma + epsilon. Thus, the gene order in the locus can be deduced to be alpha-beta-gamma-epsilon-kappa. This order was confirmed by restriction analysis of the clones. A contig map of the clones and flanking sequences has been established by characterizing seven BAC clones, which together span approximately 470 kb. Long-range restriction analysis of genomic DNA indicates that the murine casein locus is confined to a 250-kb partial Xho I fragment. The alpha and beta casein genes were shown to be arranged in a tail-to-tail orientation.

Structure and expression of the mouse casein gene locus

The analysis of yeast artificial chromosomes (YACs) containing the complete mouse casein gene locus revealed the presence of five casein genes, alpha-, beta-, gamma-, delta-, and kappa-casein, in this order, in the locus. The alpha- and beta-casein genes are only 10 kb apart and have convergent transcriptional orientations. The distance between the beta-casein gene and the alpha s2-like gamma-casein gene is about 70 kb, and these genes have divergent transcriptional orientations. The gamma- and delta-casein genes, both encoding a alpha s2-like casein, are linked within 60 kb and convergently transcribed. The kappa-casein gene is located about 100 kb from the delta-gene. Except for the presence of the delta-casein gene, the organization of the mouse casein locus resembles that of the bovine locus, including the transcriptional orientation of the genes. In contrast to the other casein genes, which are strongly induced at mid-lactation, expression of the delta-casein gene is abruptly induced upon parturition. Comparative analysis of alpha s2-like sequences from various species suggests that the ancestral alpha s2-like gene duplicated around the time of radiation of the rodent and artiodactylid ancestors.

Variegated transgene expression in mouse mammary gland is determined by the transgene integration locus

Mice carrying an ovine beta-lactoglobulin (BLG) transgene secrete BLG protein into their milk. To explore transgene expression stability, we studied expression levels in three BLG transgenic mouse lines. Unexpectedly, two lines exhibited variable levels of transgene expression. Copy number within lines appeared to be stable and there was no evidence of transgene rearrangement. In the most variable line, BLG production levels were stable within individual mice in two successive lactations. Backcrossing demonstrated that genetic background did not contribute significantly to variable expression. Tissue in situ hybridization revealed mosaicism of transgene expression within individual mammary glands from the two variable lines; in low expressors, discrete patches of cells expressing the transgene were observed. Transgene protein concentrations in milk reflected the proportion of epithelial cells expressing BLG mRNA. Furthermore, chromosomal in situ hybridization revealed that transgene arrays in both lines are situated close to the centromere. We propose that mosaicism of transgene expression is a consequence of the chromosomal location and/or the nature of the primary transgene integration event.

Expression analysis of the individual bovine beta-, alpha s2- and kappa-casein genes in transgenic mice

To identify cis-acting regulatory elements involved in the regulation of expression of the casein genes, the bovine beta-, alpha s2- and kappa-casein genes were isolated from cosmid libraries and introduced into the murine germline. Bovine casein expression was analysed at the RNA and protein level. The bovine beta-casein gene, including 16 kb of 5'- and 8 kb of 3'-flanking region, appeared to be expressed in all 12 transgenic mouse lines analysed. In 50% of these lines expression levels in milk exceeded 1 mg/ml. Three lines displayed expression levels comparable with or well above (20 mg/ml) the beta-casein levels in bovine milk. Transgene expression was restricted to the mammary gland. Strong induction of expression occurred at parturition and thus resembled the bovine rather than the murine pattern. In spite of this high-level tissue-specific and developmentally regulated expression, beta-casein expression levels were integration-site-dependent, suggesting that not all elements involved in regulation of expression were included in this beta-casein clone. Neither the bovine alpha s2- nor the kappa-casein gene, including 8 kb and 5 kb of 5'- and 1.5 kb and 19 kb of 3'-flanking sequences respectively, were properly expressed in transgenic mice. However, they were transcribed in stably transfected mouse mammary epithelial cells. This indicates that regulatory elements required for high-level, mammary gland-specific expression are not present in the alpha s2- and kappa-casein clones used in this study and are probably located elsewhere in the casein gene locus.

Secretory proteins compete for production in the mammary gland of transgenic mice

To explore the possibility that genes might compete for expression, we have studied transgenic mice producing high levels of the sheep milk protein, beta-lactoglobulin (BLG), in the mammary gland. Mice carrying one or more transgene loci expressed BLG in milk at levels ranging from 7 to 33 mg/ml. The effects of BLG synthesis on the levels of endogenous milk gene expression were examined. No significant increase in total milk protein concentration was recorded even in mice expressing the largest amounts of BLG. Measurement of individual milk proteins showed that transgene protein was manufactured at the expense of host protein synthesized in the gland. Whey acidic protein production was more suppressed than casein production. Suppression of endogenous proteins was matched by a reduction in the corresponding steady-state mRNA levels; in double-transgenic mice, which expressed the largest amounts of BLG, beta-casein and whey acidic protein mRNA populations were reduced to 75 and 56% of control levels respectively. We demonstrate that an exogenous gene competes effectively for expression with endogenous genes. Possible mechanisms of competition are discussed.

Human alpha s1-casein: purification and characterization

The human counterpart of alpha s1-casein has been purified by a combination of gel-filtration and ion-exchange chromatography under denaturing conditions. SDS-PAGE analysis revealed the presence of a diffuse ladder with a high molecular mass which upon reduction was replaced by several closely spaced bands of lower molecular masses and a broad diffuse band corresponding to kappa-casein. Amino acid sequence analysis of the closely spaced bands all resulted in the same N-terminal sequence which was found to be homologous with alpha s1-casein from other species. Sequence analysis of a major radiolabelled tryptic peptide from purified 14C-carboxymethylated alpha s1-casein demonstrated that the protein contains at least two cysteine residues. As judged by SDS-PAGE in the presence or absence of a reducing agent, the molecular structure of the polymers constituting the ladder is composed of heteropolymers of alpha s1- and kappa-casein cross-linked by disulfide bonds.

Ectopic expression of beta-lactoglobulin transgenes

Genomic constructs comprising the ovine beta-lactoglobulin gene are expressed in a position-independent manner in the mammary gland of transgenic mice. In some lines however, constitutive low-level transgene expression was detected in all other tissues. This ectopic expression presumably represents a position-dependent phenomenon since it was observed in only a proportion (40%) of the lines generated. Different lines of BLG transgenic mice displayed similar temporal patterns of ectopic expression. This pattern differed from that of BLG in the mammary gland. These data imply that the DNA elements that direct position-independent expression of beta-lactoglobulin transgenes in the mammary gland do not have the ability to insulate them from position effects in other tissues. Furthermore, the relatively high frequency and constitutive nature of ectopic expression suggests that transgene integration may not be totally random.

Mouse rump-white mutation associated with an inversion of chromosome 5

The rump-white (Rw) mutation in the mouse was previously mapped as part of a cluster of spotting genes on Chromosome (Chr) 5 that includes the dominant spotting (W) and patch (Ph) loci. Recent studies have shown that the W locus encodes the KIT tyrosine kinase cell surface receptor and that Ph is a deletional mutation encompassing the platelet-derived growth factor receptor alpha subunit (Pdgfra) gene. However, the molecular basis of the Rw mutation remains to be established. We have analyzed an interspecific Mus spretus backcross segregating Rw and several loci proximal and distal to the W/Ph/Rw region to study the basis of this mutation. These studies indicated that loci within the En2 to Kit region of the chromosome do not recombine with one another even though they have been separated in other mapping studies presented here and elsewhere. We conducted a series of fluorescent in situ hybridization (FISH) studies with genomic probes to En2, Msx1, D5Buc1, and Kit to compare the physical order of these loci on the Rw and wild-type chromosomes. The Kit locus mapped to approximately the same region on both chromosomes of the Rw heterozygotes, while the positions of En2, Msx1, and D5Buc1 were reversed on the two chromosomes. Taken together, both the genetic and physical mapping data establish that the Rw mutation is associated with an inversion involving loci in the proximal region of Chromosome 5.

Expression of human lactoferrin in milk of transgenic mice

The expression of human lactoferrin (hLF) in the milk of transgenic mice is described. Regulatory sequences derived from the bovine alpha S1-casein gene were fused to the coding sequence of the hLF cDNA and several lines of transgenic mice were generated. Human LF RNA was detected exclusively in the mammary gland of lactating females and only after the onset of lactation. No aberrant RNA products could be detected using northern blotting and primer extension analysis. The hLF concentrations in the milk ranged from less than 0.1 to 36 micrograms ml-1. Human LF thus expressed did not differ from human milk derived LF, with respect to molecular mass and immunoreactivity with monoclonal and polyclonal antibodies.

Structure and function of milk protein genes

Interspecies comparisons of cDNA and mosaic milk protein genes have confirmed their high rate of evolution, but the overall gene organization has been conserved. The three Ca-sensitive casein genes, which share common motifs in the promoter region and contain similar sequences that encode signal peptide and multiple phosphorylation sites, probably derived from a common ancestor. alpha s1- and alpha s2-casein genes, divided into many small exons, undergo complex splicing, and the deleted caseins arise from exon skipping. The four bovine casein genes are clustered on 200 kb of chromosome 6. alpha-Lactalbumin and beta-lactoglobulin pseudogenes occur in ruminants. Study of the expression of native and modified milk protein genes in mammary cell lines and transgenic animals and DNA footprinting have shown the occurrence of important regulatory motifs in the proximal 5' flanking region, including one recognized by a specific mammary nuclear factor. Good stage- and tissue-specific expression has been obtained in transgenic animals with milk protein genes having less than a 3-kb 5' flanking region. Better knowledge of both the structure and function of milk protein genes, which has already allowed the use of powerful techniques for the rapid identification of alleles, offers the potential for the genetic modification of milk composition.

Nucleoproteins derived from subnuclear RNA polymerase complexes of metastatic large-cell lymphoma cells possess transcription activities and regulatory properties in vitro

Intact nuclei derived from poorly or highly liver-metastatic murine large-cell lymphoma cell line RAW117 were digested to discrete subchromatin deoxyribonucleoprotein/ribonucleoprotein (DNP/RNP) complexes with Msp-I. The DNP/RNP complexes were composed of DNP/RNPs which were derived from the DNP/RNP complexes by incubation in the presence or absence of DNase-I and subsequent isolation by two-dimensional [isoelectric focusing/sodium dodecylsulfate (SDS)] polyacrylamide gel electrophoresis (PAGE), electroelution from the gel, and removal of SDS. Approximately 450 DNP/RNPs in the two-dimensional gels corresponding to discrete spots or in some cases streaks were analyzed for the presence of v-abl, p53, c-neu, c-H-ras, beta-casein, 18s rDNA, and mu-chain immunoglobulin genes using a hybridization technique. Ten DNP/RNP complexes contained tightly associated p53 DNA, whereas six contained c- or v-abl, four contained mu-chain gene, two contained c-H-ras, one contained dot-blot beta-casein, two contained 18s rDNA, and c-neu was found in one of the DNP/RNPs. The DNP/RNPs were also analyzed for in vitro RNA polymerase and primase activities. To assess the potential transcription abilities of the isolated DNP/RNPs, individual DNP/RNPs or DNP/RNP mixtures (reconstituted after SDS-PAGE separation) were examined for RNA polymerase initiation and synthesis. When RNA products were formed, these were purified by extracellulose chromatography and used as back-hybridization probes for the genes of interest. The RNA products were also analyzed by RNA gel electrophoresis. RNA formation was inhibitable by actinomycin D, and the RNAs formed ranged in size from approximately 80 kbp to approximately 1 kbp. By mixing various DNP/RNP complexes together, different patterns of RNA synthesis were found. For example, one DNP/RNP of M(r) approximately 140,000, isoelectric point (pl) approximately 5.8 synthesized a high molecular weight RNA in vitro that hybridized with beta-casein cDNA, but beta-casein is not expressed in RAW117 cells, suggesting that the silencing of the beta-casein gene was negated by isolation of the DNP/RNP. Mixing this DNP/RNP with two other specific DNP/RNPs again inhibited the synthesis of beta-casein RNA, suggesting that interactions between DNP/RNP complexes can result in differential RNA expression or regulation of RNA polymerases in vitro.

Synteny mapping in the bovine: genes from human chromosome 4

Genes homologous to those located on human chromosome 4 (HSA4) were mapped in the bovine to determine regions of syntenic conservation among humans, mice, and cattle. Previous studies have shown that two homologs of genes on HSA4, PGM2 and PEPS, are located in bovine syntenic group U15 (chromosome 6). The homologous mouse genes, Pgm-1 and Pep-7, are on MMU5. Using a panel of bovine x hamster hybrid somatic cells, we have assigned homologs of 11 additional HSA4 loci to their respective bovine syntenic groups. D4S43, D4S10, QDPR, IGJ, ADH2, KIT, and IF were assigned to syntenic group U15. This syntenic arrangement is not conserved in the mouse, where D4s43, D4s10, Qdpr, and Igj are on MMU5 while Adh-2 is on MMU3. IL-2, FGB, FGG, and F11, which also reside on MMU3, were assigned to bovine syntenic group U23. These data suggest that breaks and/or fusions of ancestral chromosomes carrying these genes occurred at different places during the evolution of humans, cattle, and mice.

Position-independent expression of the ovine beta-lactoglobulin gene in transgenic mice

The major milk whey protein of sheep, beta-lactoglobulin (BLG), is expressed specifically in the mammary gland in a developmentally regulated pattern. To identify the cis-acting DNA regions involved in the regulation of BLG expression, resected gene constructs were analysed in transgenic mice. BLG transgenes which contain at least the proximal 406 bp of the 5' flanking region were expressed in all mice analysed, at levels related to transgene copy number, and thus were expressed in a position-independent manner. Expression was restricted to the mammary gland, except in a few lines where low-level expression was also detected in the salivary gland. In these mice, BLG transgenes were expressed during pregnancy and lactation in the appropriate temporal pattern. Further resection of the 5' proximal region to -146 bp resulted in a dramatically reduced frequency of expression, without affecting tissue specificity, while a construct which retained only 79 bp of 5' flanking region was not expressed. Chromatin analysis of isolated sheep nuclei showed that the promoter resides within a DNAaseI-hypersensitive region in the mammary gland but not in the liver. A BLG transgene displayed a similar tissue-specific pattern of DNAaseI hypersensitivity in mice. These data demonstrate an essential role of the proximal DNAaseI-hypersensitive sequences for position-independent expression of the BLG gene.

Casein, actin, and tubulin expression during early involution in bovine and murine mammary tissue

Hybridization methods and in vitro translation were used to examine the expression and functional condition of messenger RNA encoding caseins and cytoskeletal proteins in the mammary gland during early involution. In the mouse, steady state mRNA levels for alpha-, beta-, and gamma-caseins coordinately decreased to 20% of initial levels between 12 and 72 h after pup removal. In vitro translatability of mouse casein mRNA, as determined by immunoprecipitation, electrophoresis, and gel slice counting, revealed a pattern that closely paralleled mRNA expression. In contrast, bovine casein mRNA levels were only slightly reduced by 72 h postmilking, whereas in vitro translatability decreased by about one-half. Northern blot analysis of total mouse mammary RNA that were hybridized with probes to cytoskeletal proteins showed a gradual decrease of alpha-tubulin mRNA, but an increase in beta-actin mRNA during early involution. Two-dimensional gel analysis of in vitro translated products indicated a concordant increase in beta-gamma-actin. In the cow, beta-actin mRNA at 72 h of involution was equal to or greater than that during lactation. These results demonstrate the generally slower involution response in the cow and suggest that differing regulations are involved. Early events of cellular involution may be related to a reorganization of the cytoskeleton.

Differential regulation of the Wnt gene family during pregnancy and lactation suggests a role in postnatal development of the mammary gland

The mouse Wnt family comprises at least 10 members sharing substantial amino acid identity with the secreted glycoprotein Wnt-1/int-1. Two of these, Wnt-1 and Wnt-3, are implicated in mouse mammary tumor virus-associated adenocarcinomas, although neither member is normally expressed in the mammary gland. These results suggest the presence of active cellular pathways which mediate the action of Wnt-1 and Wnt-3 signals. An understanding of the normal role of these signalling pathways is clearly necessary to comprehend the involvement of Wnt-1 and Wnt-3 in mammary tumorigenesis. We demonstrate here that five Wnt family members are expressed and differentially regulated in the normal mouse mammary gland. In addition, some of these genes are also expressed in both Wnt-1-responsive and nonresponsive mammary epithelial cell lines. We propose that Wnt-mediated signalling is involved in normal regulation of mammary development and that inappropriate expression of Wnt-1, Wnt-3, and possibly other family members can interfere with these signalling pathways.

Differential regulation of the Wnt gene family during pregnancy and lactation suggests a role in postnatal development of the mammary gland

The mouse Wnt family comprises at least 10 members sharing substantial amino acid identity with the secreted glycoprotein Wnt-1/int-1. Two of these, Wnt-1 and Wnt-3, are implicated in mouse mammary tumor virus-associated adenocarcinomas, although neither member is normally expressed in the mammary gland. These results suggest the presence of active cellular pathways which mediate the action of Wnt-1 and Wnt-3 signals. An understanding of the normal role of these signalling pathways is clearly necessary to comprehend the involvement of Wnt-1 and Wnt-3 in mammary tumorigenesis. We demonstrate here that five Wnt family members are expressed and differentially regulated in the normal mouse mammary gland. In addition, some of these genes are also expressed in both Wnt-1-responsive and nonresponsive mammary epithelial cell lines. We propose that Wnt-mediated signalling is involved in normal regulation of mammary development and that inappropriate expression of Wnt-1, Wnt-3, and possibly other family members can interfere with these signalling pathways.

Regional localization of human beta-casein gene (CSN2) to 4pter-q21

Milk proteins are crucial for the development of all newborn mammals. Caseins that constitute the bulk of the protein in mammalian milk have been shown to be members of a multigene family in at least two species. They are among the most rapidly diverging groups of proteins, and their numbers vary widely among species. beta- and kappa-Caseins are the only caseins present in human milk. Using polymerase chain reaction on genomic DNA from somatic cell hybrids, we have localized the human beta-casein gene (CSN2) to 4pter----q21.

Transgenic mice carrying the guinea-pig alpha-lactalbumin gene transcribe milk protein genes in their sebaceous glands during lactation

We have generated transgenic mice carrying the entire guinea-pig alpha-lactalbumin gene. Lactating transgenic mice expressed high levels of correctly initiated and processed guinea-pig alpha-lactalbumin mRNA in the secretory epithelium of their mammary glands, and secreted guinea-pig alpha-lactalbumin in their milk. Transcripts were detectable after 7 days of pregnancy, indicating that the transgene was under correct hormonal control. Whereas no or negligible transcription was detectable in all other tissues tested, high levels of transcripts were found in the skin of lactating transgenic mice. Guinea-pig alpha-lactalbumin protein was undetectable in the skin, however. In situ hybridization analysis showed that expression was localized to the undifferentiated cells in the basal layer of the sebaceous glands. Further studies revealed high levels of endogenous beta-casein mRNA in normal lactating mouse skin, demonstrating that the transcription of milk protein genes in lactating mouse skin is a normal event, and is not peculiar to the transgene. This surprising finding highlights the developmental relationship of the mammary gland to other specialized structures of the skin, supports a role for epithelial-extracellular matrix interactions in the regulation of milk protein gene expression in vivo, and identifies the skin as a particularly accessible model system in which to study the regulation of milk protein gene expression. In addition, the guinea-pig alpha-lactalbumin gene will be a source of regulatory sequences with which to direct heterologous gene expression to the sebaceous glands of transgenic mice.

Genomic analysis of the major bovine milk protein genes

The genomic arrangement of the major bovine milk protein genes has been determined using a combination of physical mapping techniques. The major milk proteins consist of the four caseins, alpha s1 (CASAS1), alpha s2 (CASAS2), beta (CASB), and kappa (CASK), as well as the two major whey proteins, alpha-lactalbumin (LALBA) and beta-lactoglobulin (LGB). A panel of bovine X hamster hybrid somatic cells analyzed for the presence or absence of bovine specific restriction fragments revealed the genes coding for the major milk proteins to reside on three chromosomes. The four caseins were assigned to syntenic group U15 and localized to bovine chromosome 6 at q31-33 by in situ hybridization. LALBA segregated with syntenic group U3, while LGB segregated with U16. Pulsed-field gel electrophoresis confirmed genetic mapping results indicating tight linkage of the casein genes. The four genes reside on less than 200 kb of DNA in the order CASAS1-CASB-CASAS2-CASK. Multiple restriction fragment length polymorphisms were also found at the six loci in three breeds of cattle.

Casein expression in cytotoxic T lymphocytes

A cDNA that expresses a mRNA restricted to cytotoxic T lymphocytes (CTL) and mammary tissue has been isolated and characterized. The deduced amino acid sequence from this cDNA shows extensive homology with the previously reported amino acid sequence for rat alpha-casein. Indeed, the presence of a six-residue-repeated motif that is specific for rodent alpha-caseins strongly supports the identification of this cDNA as mouse alpha-casein. Northern (RNA) blot analysis of many hematopoietic cell types revealed that this gene is restricted to CTL, being expressed in four of six CTL lines examined. Furthermore, CTL that express this gene were also found to express other members of the casein gene family, such as beta- and kappa-casein. These results suggest that caseins may be important in CTL function, and their potential role in CTL-mediated lysis is discussed.

In vitro and in vivo growth and casein gene expression of mouse mammary tumor epithelial cells in response to hormones

Cells from autochthonous mouse mammary carcinomas which display estrogen-independent growth in vivo were studied for their hormonal responses in primary culture. A culture system employing insulin-supplemented, serum-free medium and basement membrane Matrigel as a substratum was used to cultivate tumor cells. The cells did not exhibit in vitro estrogen- or prolactin-dependent growth. Primary tumors still displayed a constitutional expression of alpha-, beta-, and gamma-casein mRNAs. These messages were dramatically reduced during the culture period. However, seven to eightfold increases in alpha- and beta-casein mRNAs were inducible in the 5-day cultures by treatment with prolactin and hydrocortisone. If the hormones were present through a 2-week culture period, the levels of alpha-, beta-, and gamma-casein mRNAs in the cells were maintained and displayed in a time-dependent increase with a peak at 10-14 days. The accumulation of beta-casein mRNA in vitro did not require DNA synthesis. Administration of prolactin directly into the growing tumors in vivo could also enhance beta-casein mRNA levels in the tumor cells. Morphological studies of the cells cultured in the presence of prolactin and hydrocortisone did not reveal visible changes compared with those without hormonal treatment. Transplantation of tumor cells cultured in the presence or absence of hormones resulted in the development of tumors in mice at approximately the same time. The current studies suggest that the autochthonous mammary tumor cells, independent of estrogen for cell growth, were still inducible for casein gene expression in vitro and in vivo by appropriate hormones. The induction and maintenance of casein messages by a single hormonal treatment did not appear to correlate with morphology and DNA synthesis of cells in vitro or with tumor-producing capacities in vivo.

Caseins of various origins and biologically active casein peptides and oligosaccharides: structural and physiological aspects

The first part of the present review is focused on structural aspects concerning the so far studied casein fractions of various origins: they are compared to the four classical major bovine caseins (alpha s1-, alpha s2-, beta- and kappa). The calcium-sensitive casein fractions are always phosphorylated whereas kappa-caseins are glycosylated. The study of the casein genes showed that the calcium-sensitive caseins diverged from a common ancestral gene and during the evolution, intergenic and intragenic duplications occurred. The considerable conservation of the phosphorylation sites emphasizes the importance of phosphorylated residues for the function of caseins, i.e. the formation of micelles and the binding of Ca2+. In kappa-caseins all the prosthetic sugar groups are linked by O-glycosidic linkages: their number varies from 0 to 5 in bovine kappa-casein and up to 10 in human kappa-casein. The structures of the known kappa-casein carbohydrate moieties are described. Finally the milk clotting process (interaction kappa-casein/chymosin) is compared to the blood clotting process (interaction fibrinogen/thrombin): a large number of similarities could be noted between both clotting phenomena. The second part of the review is devoted to the study of short casein peptides endowed with various biological activities. Some of them behaved as immunomodulators or casomorphins or angiotensin I converting enzyme inhibitors; others demonstrated an effect on platelet functions. A 'strategic zone' containing immunostimulating and opioid peptides could be located in cow and human beta-caseins. Furthermore bitter peptides, emulsifying peptides, calcium absorption enhancing peptides, chymosin-inhibiting peptides, have also been described and several further properties have been attributed to the kappa-caseinoglycopeptide; two tetrasaccharides isolated from the latter possess blood group activities. In conclusion caseins, the main milk proteins, should not only be considered as a nutriment but as a possible source of biologically active components. If, in the future, some of the discussed active peptides cannot be characterized in vivo, they can all, nevertheless, be synthesized and used either as food additives or in pharmacology.

Both cell substratum regulation and hormonal regulation of milk protein gene expression are exerted primarily at the posttranscriptional level

The mechanism by which individual peptide and steroid hormones and cell-substratum interactions regulate milk protein gene expression has been studied in the COMMA-D mammary epithelial cell line. In the presence of insulin, hydrocortisone, and prolactin, growth of COMMA-D cells on floating collagen gels in comparison with that on a plastic substratum resulted in a 2.5- to 3-fold increase in the relative rate of beta-casein gene transcription but a 37-fold increase in beta-casein mRNA accumulation. In contrast, whey acidic protein gene transcription was constitutive in COMMA-D cells grown on either substratum, but its mRNA was unstable and little intact mature mRNA was detected. Culturing COMMA-D cells on collagen also promoted increased expression of other genes expressed in differentiated mammary epithelial cells, including those encoding alpha- and gamma-casein, transferrin, malic enzyme, and phosphoenolpyruvate carboxykinase but decreased the expression of actin and histone genes. Using COMMA-D cells, we defined further the role of individual hormones in influencing beta-casein gene transcription. With insulin alone, a basal level of beta-casein gene transcription was detected in COMMA-D cells grown on floating collagen gels. Addition of prolactin but not hydrocortisone resulted in a 2.5- to 3.0-fold increase in beta-casein gene transcription, but both hormones were required to elicit the maximal 73-fold induction in mRNA accumulation. This posttranscriptional effect of hormones on casein mRNA accumulation preceded any detectable changes in the relative rate of transcription. Thus, regulation by both hormones and cell substratum of casein gene expression is exerted primarily at the post transcriptional level.

Genetic analysis of the dominant white-spotting (W) region on mouse chromosome 5: identification of cloned DNA markers near W

We have assigned several mouse cDNA and genomic clones to the W region of mouse chromosome 5, established their position with respect to various marker loci in the region, and provided molecular verification that the W19H mutation is a deletion. Meiotic recombination analysis of an interspecific mouse backcross indicated the following gene order and distances [in centimorgans (cM)]: centromere-Emv-1-(13 cM)-D4S76-(17 cM)-D5SC25-(5 cM)-alpha-casein-(1 cM)-beta- casein-(6 cM)-alpha-fetoprotein-(18 cM)-beta-glucuronidase. D5SC25, an anonymous locus defined by a mouse brain cDNA, maps near the map position of W and within the breakpoints of the presumed genetic deletion that causes the W19H phenotype. Southern analysis of DNAs of W19H/+ interspecific F1 hybrid mice and somatic cell hybrid lines carrying the W19H deletion chromosome showed the deletion of D5SC25. In fact, analysis of other mutations at or near the W locus, which had been transferred from the strain of origin for many backcross generations, revealed the retention of donor restriction fragment length polymorphisms at the D5SC25 locus. Such evidence confirms close linkage between D5SC25 and W (within 1 cM) and indicates that the D5SC25 cDNA clone could serve as a starting point in a chromosome "walk" to W and other closely linked loci that affect development.

Both cell substratum regulation and hormonal regulation of milk protein gene expression are exerted primarily at the posttranscriptional level

The mechanism by which individual peptide and steroid hormones and cell-substratum interactions regulate milk protein gene expression has been studied in the COMMA-D mammary epithelial cell line. In the presence of insulin, hydrocortisone, and prolactin, growth of COMMA-D cells on floating collagen gels in comparison with that on a plastic substratum resulted in a 2.5- to 3-fold increase in the relative rate of beta-casein gene transcription but a 37-fold increase in beta-casein mRNA accumulation. In contrast, whey acidic protein gene transcription was constitutive in COMMA-D cells grown on either substratum, but its mRNA was unstable and little intact mature mRNA was detected. Culturing COMMA-D cells on collagen also promoted increased expression of other genes expressed in differentiated mammary epithelial cells, including those encoding alpha- and gamma-casein, transferrin, malic enzyme, and phosphoenolpyruvate carboxykinase but decreased the expression of actin and histone genes. Using COMMA-D cells, we defined further the role of individual hormones in influencing beta-casein gene transcription. With insulin alone, a basal level of beta-casein gene transcription was detected in COMMA-D cells grown on floating collagen gels. Addition of prolactin but not hydrocortisone resulted in a 2.5- to 3.0-fold increase in beta-casein gene transcription, but both hormones were required to elicit the maximal 73-fold induction in mRNA accumulation. This posttranscriptional effect of hormones on casein mRNA accumulation preceded any detectable changes in the relative rate of transcription. Thus, regulation by both hormones and cell substratum of casein gene expression is exerted primarily at the post transcriptional level.

The gene for human alpha-lactalbumin is assigned to chromosome 12q13

A cDNA clone complementary to the mRNA encoding human alpha-lactalbumin (ALA) has been used as a probe in the analysis of DNA from panels of rodent/human somatic cell hybrids. The presence of the ALA gene correlates with the presence of chromosome 12. In situ hybridization localizes the ALA gene to 12q13.

Milk protein expression and ductal morphogenesis in the mammary gland in vitro: hormone-dependent and -independent phases of adipocyte-mammary epithelial cell interaction

Epithelial cell differentiation frequently occurs in situ in conjunction with supporting mesenchyme or connective tissue. In embryonic development the importance of the supporting mesenchyme for cytodifferentiation and morphogenesis has been demonstrated in several epithelial tissues, but the importance of epithelial-connective tissue interactions is less well studied in adult epithelial organs. We have investigated the interaction of adult mammary epithelial cells with adipocytes, which compose the normal supporting connective tissue in the mammary gland. Mammary epithelial cells from mice in various physiological states were cultured on cellular substrates of adipocytes formed from cells of the 3T3-L1 preadipocyte cell line. We found that there were two distinct phases to the interaction of epithelial cells with adipocytes. Cytodifferentiation of the epithelial cells and milk protein production were dependent on lactogenic hormones (insulin, hydrocortisone, and prolactin), whereas ductal morphogenesis was lactogenic hormone independent. When cultured on preadipocytes or adipocytes, mammary epithelial cells from never pregnant, pregnant, lactating, and involuting mice responded to lactogenic hormones rapidly by producing and secreting large amounts of alpha-, beta-, and gamma-casein and alpha-lactalbumin. This response was seen in individual as well as in clusters of epithelial cells, but was not seen if the same cells were cultured on tissue culture dishes without adipocytes, on fibroblasts (human newborn foreskin fibroblasts) or in the presence of adipocytes but in the absence of lactogenic hormones. Continued incubation of mammary epithelial cells on adipocytes in the presence or absence of lactogenic hormones resulted in the formation of a branching ductal system. Mammary epithelial cells in ducts that formed in the absence of lactogenic hormones produced no casein, but rapidly synthesized casein when subsequently exposed to these hormones. Ultrastructural studies revealed that the formation of a basement membrane occurs only in co-cultures of mammary epithelium with adipocytes or preadipocytes. Ultrastructural changes associated with secretion occurred only in the presence of lactogenic hormones. We propose that growth and formation of a ductal system in vitro can occur in the absence of lactogenic hormones, but that certain environment-associated events must occur if the epithelium is to become responsive to lactogenic hormones and undergo the cytodifferentiation associated with lactation.

Hormonally induced elevations of alpha- and beta-casein mRNA levels are blocked by cyclic adenine nucleotide and prostaglandins

Normal mammary gland development during pregnancy follows a coordinated program of morphological development (formation of lobuloalveoli) and biochemical differentiation (casein production). In culture, whole mammary glands of Balb/c mice can be similarly induced by application of a mixture of insulin, prolactin, aldosterone and hydrocortisone (IPAH) for 7 days. Our previous reports have shown that lobuloalveolar development, induced by IPAH, is competitively inhibited by the simultaneous presence of dibutyryl cyclic AMP (Bt2cAMP), prostaglandins (PGs) E1, E2, and B1, and papaverine (pap). However, if this mixture is not added until day 4, lobuloalveolar development is relatively unaffected but casein synthesis is repressed. This report explores the mechanism by which cyclic adenine nucleotides and prostaglandins interfere with the normal developmental pathway. The accumulation of alpha- and beta-casein mRNAs induced by prolactin, hydrocortisone and aldosterone is blocked by the combination of Bt2cAMP, PGs E1, E2, and B1, and pap added to the medium for the final 3 days (days 4-7). Under these conditions the glands retain their lobuloalveoli, and little squamous metaplasia can be discerned. Furthermore, de novo synthesis of both caseins is selectively inhibited, despite the continued presence of casein mRNAs in the glands and normal protein synthesis. In contrast, the synthesis of keratin is stimulated. Incomplete mixtures of Bt2cAMP and pap or the combination of PGs E1, E2, and B1, are only partly effective in preventing the accumulation of casein mRNAs. All three mixtures bring about similar effects on both alpha- and beta-casein mRNAs.(ABSTRACT TRUNCATED AT 250 WORDS)

Evolution of the casein multigene family: conserved sequences in the 5' flanking and exon regions

The rat alpha- and bovine alpha s1-casein genes have been isolated and their 5' sequences determined. The rat alpha-, beta-, gamma- and bovine alpha s1-casein genes contain similar 5' exon arrangements in which the 5' noncoding, signal peptide and casein kinase phosphorylation sequences are each encoded by separate exons. These findings support the hypothesis that during evolution, the family of casein genes arose by a process involving exon recruitment followed by intragenic and intergenic duplication of a primordial gene. Several highly conserved regions in the first 200 base pairs of the 5' flanking DNA have been identified. Additional sequence homology extending up to 550 base pairs upstream of the CAP site has been found between the rat alpha- and bovine alpha s1-casein sequences. Unexpectedly, the 5' flanking promoter regions are conserved to a greater extent than both the entire mature coding and intron regions of these genes. These conserved 5' flanking sequences may contain potential cis regulatory elements which are responsible for the coordinate expression of the functionally-related casein genes during mammary gland development.

Construction and identification of recombinant plasmids carrying cDNAs coding for ovine alpha S1-, alpha S2-, beta-, kappa-casein and beta-lactoglobulin. Nucleotide sequence of alpha S1-casein cDNA

An ovine mammary cDNA library has been constructed from total poly(A)+ RNA isolated from the mammary gland of a lactating ewe, using a classical procedure. Blunt-ended double-stranded cDNAs prepared with reverse transcriptase and nuclease S1 were tailed with dCTP, inserted into the dGMP-tailed PstI site of plasmid pBR322 and cloned in E. coli. Five series of homologous clones representing abundant messenger RNAs (strong hybridization with a single-stranded cDNA probe generated from total poly(A)+ RNA) were selected using each time a different predominant cloned ds-cDNA as probe, then identified by positive hybridization-translation of the cognate mRNA and subsequent immunoprecipitation and electrophoresis of the protein. The lengths of alpha s1-, alpha s2-, beta-, kappa-casein and beta-lactoglobulin mRNAs are in the range of 1.2, 1.1, 1.25, 1.0 and 0.85 kb, respectively, as determined by Northern blotting analysis. Five homologous mRNAs of similar sizes were identified in the porcine species by dot blot hybridization and Northern analyses. The nucleotide sequence of alpha s1-casein mRNA was determined by sequencing, according to Maxam and Gilbert, both a 1080 bp long cloned ds-cDNA and a ss-cDNA (268 nucleotides) generated by 5' extension of a 5' terminal truncated radiolabeled fragment (83 bp) of the relevant ds-cDNA, used as primer for reverse transcription. The 3' non coding region (431 nucleotides, excluding the poly(A) tail) represents 70% of the length of the coding region (618 nucleotides) flanked by a 61 nucleotide 5' region. Comparison of sequences of ovine and bovine, rat and guinea-pig alpha s1-casein mRNAs has revealed a greater homology in the 3' and especially 5' non coding regions. In the reading frame, the conserved regions are essentially those corresponding to the signal peptide and phosphopeptide domains. The derived 206 amino acid sequence of ovine pre-alpha s1-casein differs from that of its bovine counterpart (genetic variant B) by 24 amino acid substitutions and a deletion of 8 amino acid residues occurring in the polypeptide chain of the mature protein. Such a variation (84% homology only) in two phylogenetically closely related species indicates a high rate of evolution of alpha s1-casein.

Interaction of mouse mammary epithelial cells with collagen substrata: regulation of casein gene expression and secretion

Mouse mammary epithelial cells (MMEC) secrete certain milk proteins only when cultured on floating collagen gels. We demonstrate here that modulation of milk proteins by substrata is manifested at several regulatory levels; (i) Cells cultured on floating collagen gels have 3- to 10-fold more casein mRNA than cells cultured on plastic or attached collagen gels. (ii) Cells on the latter two "flat" substrata, nevertheless, synthesize a significant amount of caseins, indicating that the remaining mRNA is functional. (iii) Cells on all substrata are inducible for casein mRNA and casein proteins by prolactin, but the extent of induction is greater on collagen than that on plastic--i.e., the substratum confers an altered degree of inducibility. (iv) Cells on all substrata synthesize casein proteins at rates proportional to the amount of casein mRNA, but the newly synthesized caseins in cells on plastic are degraded intracellularly, whereas those synthesized by cells on floating gels are secreted into the medium. (v) Cells on all substrata examined lose virtually all mRNA for whey acidic protein despite the fact that this mRNA is abundant in the mammary gland itself; we conclude that additional, as-yet-unknown, factors are necessary for synthesis and secretion of whey acidic protein in culture.