Transgenic Research 11: 505–513, 2002.© 2002 Kluwer Academic Publishers. Printed in the Netherlands.

505

Targeted expression of the only zinc finger gene in transgenic mice isassociated with impaired mammary development

Nathalie Besnard1,∗, Marie-Annick Persuy∗,∗∗,∗∗∗, Marie-George Stinnakre1, LaurenceLepourry1, Jose Costa Da Silva1, Gerard Goubin2 & Jean-Luc Vilotte1

1Laboratoire de Génétique Biochimique et de Cytogénétique. I.N.R.A., 78352 Jouy-en-Josas Cedex, France2Laboratoire d’Oncologie, U.M.R. 147, C.N.R.S., Institut Curie, 75005 Paris, France.

Received 11 November 2001; revised 25 February 2002; accepted 4 March 2002

Key words: deregulation, development, mammary gland, only zinc finger, transgenic mice

Abstract

The only zinc finger (OZF) gene encodes a protein consisting mainly of 10 zinc finger motifs of the Krüppel type ofyet unknown function. To potentially assess its in vivo role, mammary targeted deregulation of the expression of themurine gene was performed in transgenic mice using a goat β-casein-based transgene. Mammary expression of thetransgene was observed in the 11 lines obtained. In three expressing lines, this expression was tissue-specific anddevelopmentally regulated. Further analysis of mice from two expressing lines revealed that transgene-homozygousfemales could not sustain full growth of their pups. This phenotype was associated with an impaired mammarygland development noticeable only after mid-gestation. It was characterised by an increase of the adipocyte toacini ratio and low or absence of fat globules within these acini compared to non-transgenic control animals. Thesetransgenic observations strongly suggest that OZF is active in the mammary gland, interfering with the lactationprocess and thus that the described transgenic mice could be useful models to search for the cellular partner(s) ofthis protein.

Introduction

The only zinc finger (OZF) gene encodes a proteinconsisting of 10 zinc finger motifs of the Krüppel type,preceded by 10 amino acids (Le Chalony et al., 1994).Like most of the Krüppel proteins, it is assumed to beinvolved in gene regulation (Rosenberg et al., 1986,Chowdhury et al., 1987; Bellefroid et al., 1989). Indomestic cattle, but not in closely related species, suchas Bison and Buffalo, OZF was translocated then amp-lified on chromosome X (Le Chalony et al., 1996). Inhumans, amplification of the OZF gene was observedin 15–25% of pancreatic carcinoma and overexpres-sion in more than 50% of these tumors (Ferbus et al.,1999).

∗These two authors contributed equally to the work.∗∗Author for correspondence: E-mail: map@diamant.jouy.inra.fr

∗∗∗Present address: Laboratoire de Biologie Cellulaire etMoleculaire, I.N.R.A., 78352 Jouy-en-Josas Cedex, France

This gene is expressed at various levels in normaltissues, including mammary gland, heart and liver. Inthe former, OZF expression was found to be devel-opmentally regulated during the pregnancy-lactation-weaning period with the highest level observed duringpregnancy and a decreased at parturition (Puissant etal., 1996). From this data, it was suggested that OZFmight be involved in the negative control of mammarycell differentiation. Analysis of OZF expression inmammary tumours did not reveal any clear association(Blottière, 1999).

Overall, the function, potential cellular partner(s)and/or target(s) of OZF remain mostly unknown. Inan attempt to assess its role, targeted deregulation ofOZF expression in the mammary gland of transgenicmice was achieved. In the two expressing lines furtheranalysed, an under development and differentiation ofthe mammary gland was observed that could lead tolactation failure.

506

Materials and methods

Construction of the transgene

The mammary-specific expression cassette was de-rived from a goat β-casein gene (Persuy et al.,1992). It comprises 3 kb of 5′ flanking region andthe start of the transcription unit leaving exon 1,intron 1, and the first 12 nucleotides of exon 2.This 5′ region was linked to the end of the genetranscription unit, 78 nucleotides before the 3′ endof exon 7, alongside 5.7 kb of 3′-flanking region,using a NotI oligonucleotide. The 848 bp murineOZF ORF was PCR-amplified using oligonucleotidesCTCTTGCGGCCGCATGTCACAGCTCAGCCAGC,GATATTCGCGGCCGCTTAATGAGTATGG (theNotI site is in italics) and its cDNA (EMBL AJ224763)as template. The amplified fragment was digested byNotI and inserted into the corresponding site of theexpression cassette. Its sequence was confirmed andits orientation was checked by RFLP analysis. The11.3 kb SalI insert was gel-purified from the vector aspreviously described (Vilotte et al., 1989).

Generation of transgenic mice

Transgenic founders were obtained by pronuclearmicro-injection of the construct into C57Bl/6xCBAF1 mouse eggs using standard procedures (Hoganet al., 1986). Identification of the transgenics wasperformed by PCR analyses of the tail-extracted ge-nomic DNA using the set of OZF primers describedabove (10 cycles: 1 min at 94◦C, 45 s at 64◦C, 45 sat 72◦C followed by 25 cycles: 1 min at 94◦C,45 s at 60◦C, 45 s at 72◦C) and, as a confirmation,a set of goat β-casein-specific primers (GTCTAA-GAGGATTTCAATGTGAATG and ATGCCTAAGG-GTTAATTTATT; 10 cycles: 1 min at 94◦C, 1 min at52◦C, 1 min at 72◦C followed by 30 cycles: 1 min at94◦C, 1 min at 42◦C, 1 min at 72◦C).

Establishment of OZF homozygous mice

Homozygous mice were established for lines 08,24, and 120. Suspected homozygous animals, werecrossed with F1 (C57Bl/6xCBA) wild-type mice. Amouse was considered to be homozygous at the trans-gene locus if all its progeny obtained with suchcrosses (n> 20) was transgenic. Two homozygoustransgenic lines (08 and 24) were retained for de-tailed mammary gland development analyses (seebelow).

Transgene expression analysis

Isolation of total RNA was performed as previouslydescribed (Persuy et al., 1992) with an extra RNAse-free DNAse I treatment to ensure the lack of any DNAcontamination. Oligo dT-primed reverse transcriptionwas performed using 1–2 µg of total RNA accord-ing to the manufacturer’s instructions (Gibco-BRL,Cergy Pontoise, France). Transgene-derived cDNAwas amplified by PCR using 10% of the first strandproduct and either the murine OZF set of primersdescribed above or goat β-casein-specific set ofprimers (ATCCATTCAGCTTCTCCTTCACTTC, AT-GCCTAAGGGTTAATTTATT) with 10 cycles (1 minat 94◦C, 45 s at 64◦C, 1 min at 72◦C) followed by25 cycles (1 min at 94◦C, 45 s at 60◦C, 45 s at 72◦C)of PCR. Absence of DNA contamination in the RNAsamples and quality of the first strand synthesis weretested by PCR using murine actin specific oligo-nucleotides, TTGAGACCTTCAACACCCC and AT-GTAGTTTCGTGGATGCC (30 cycles: 1 min at 94◦C,45 s at 60◦C, 45 s at 72◦C).

Overall assessment of the lactation performance bypup growth analysis

Assessment of the lactation performance of homozyg-ous females from lines 08 (n= 4) and 24 (n= 2) wasperformed comparatively with that of non-transgenicF1 (n= 4) mice. To this aim, at parturition, the lit-ter was removed and replaced by 7 F1 pups born thesame day. Pups were weighted 2, 4, and 6 days afterbirth.

Transgene expression effect on mammary glanddevelopment

The β-casein promoter is activated only at mid-pregnancy and fully active during lactation. To defineOZF transgene expression effects on mammary glanddevelopment, the structural and histological appear-ance of this tissue in F1 and homozygous mice fromlines 08 and 24 was comparatively studied at differentstages of pregnancy (early : day 8, middle : day 13and late : day 18) and at day 7 of lactation. Five con-trol F1 mice were slaughtered at 8 (n= 1), 13 (n= 1),18 (n= 1) days of pregnancy and 7 day of lactation(n= 2). Eighteen transgenic mice were slaughteredat 8 (n= 1 for each line), 13 and 18 days (n= 2 foreach line) and at 7 days of lactation (n= 4 for eachline).

507

Figure 1. Structure of OZF transgene and analysis of its expression in transgenic lines. (A) Schematic representation of the goat β-casein-basedOZF transgene (not at scale). White boxes: Goat β-casein exons. Black boxes: murine OZF ORF derived from its cDNA. (B) OZF-specificprimers RT-PCR products derived from transgene transcripts isolated from 7 days lactating mammary gland of mice from 10 transgenic lines1:1 kb ladder (Gibco-BRL, Cergy pontoise, France). 2–11: Sample from trangenic mouse of line 07, 08, 69, 73, 77, 82, 86, 93, 109, 120,respectively. 12 and 13: wild type F1 mice. 14: Genomic DNA from 08 line. The visible lower bands are the unincorporated primers.

Collection of mammary glandFor each animal, one single inguinal mammary glandwas removed and mounted on glass slide to performwhole mount experiments and one other piece ofmammary gland was taken and subsequently treatedfor histological analysis.

Whole mount analysisSurgically removed mammary gland samples werefixed overnight in ethanol:chloroform:acetic acid(6:4:1) Carnoy’s fixative, washed in 70% ethanol for15 min, and then stained overnight in carmine alumto visualise the ductal trees and alveolar buds. Afterwashing in ethanol, tissues were cleared and stored inxylene.

Histological analysisMammary glands were fixed in Bouin overnight, sub-sequently embedded in paraffin, sectioned (7 µm) andstained in Toluidine blue.

Results and discussion

Construct design and generation of transgenic mice

Goat β-casein based transgenes have been widely usedby us (Persuy et al., 1995) and others (Rudolph, 1999

for recent review) to target mammary-specific, de-velopmentally regulated expression of various foreigngenes in mice and other mammals. Aims include theuse of the mammary gland as a bioreactor but alsoassessment of the function of proteins taking advan-tages of the capacity of this organ to undergo cycles ofgrowth, differentiation and involution in adults.

To induce OZF deregulated expression from mid-gestation through lactation, the ORF of its cDNA wasinserted between exon 2 5′-UTR and exon 7 3′-UTRof a goat β-casein expression cassette (Figure 1). The11.3 kb insert was microinjected into C57Bl/6xCBAF1 eggs and 16 transgenic pups were detected out ofthe 145 screened. Five founders failed to transmit afteranalysis of more than 30 offspring and were discarded,whereas five others showed very low rate of transmis-sion, probably reflecting that they derived from mosaicembryos (Whitelaw et al., 1993).

OZF expression analysis

OZF is expressed in the lactating mammary gland ofall 11 transgenic lines

RT-PCR analysis of total RNA extracted from mam-mary gland of G1 mice at day 7 of lactation revealedthe occurrence of transgene-derived transcripts in all

508

Figure 2. Growth rate of wild-type pups fed by OZF 24 or wild-type females during the first days of lactation. The average weight of wild-typepups fed by OZF 24 (hatched squares) or wild-type (grey squares) females is shown. The number of pups per litter was normalised. Thedifference observed was found to be statistically significant at the three analysed stages (Student test: p < 0.0003). The given data correspondsto the analysis performed on one litter. Overall, two and four female lactation performances were analysed for transgenic lines 24 and 8,respectively, and statistically significant differences were consistently observed when compared with wild-type females (data not shown).

11 transgenic lines (Figure 1). Cloning and sequencinganalysis of some of the RT-PCR products obtainedconfirmed their origin (data not shown). It once againdemonstrates that goat β-casein-based expression cas-settes are efficient for targeting expression of foreigngenes in the lactating mammary gland.

In our conditions, OZF transcripts were never de-tected in lactating mammary gland of non transgenicmice. It was reported that mouse OZF expressionlevel declines after parturition, although low expres-sion could still be detected (Puissant et al., 1996).This apparent discrepancy with our own data maybe explained by the design of the primers and thePCR conditions. Indeed, the two OZF primer usedare complementary over their last 13 or 21 nt to theendogenous- or transgene-derived OZF transcripts, re-spectively. Since the hybridisation temperatures usedin the PCR steps were at least of 60◦C, these oligonuc-leotides might only hybridise to the transgene-derivedcDNAs.

OZF expression is tissue-specific anddevelopmentally regulatedAnimals from lines 8, 24, and 120 were further anal-ysed for tissue and stage specificity of the transgeneexpression because some heterozygous mice fromthese three lines failed to lactate or did not sustain afull lactation process (data not shown and see below).

Since this phenotype was observed in three independ-ent lines and at a heterozygous stage, we can excludethat it was related to the transgene integration site.

Transgene transcripts were not detected in eitherthe liver, heart, spleen, kidney, salivary glands, andthymus (data not shown). In the mammary gland, theiroccurrence could be detected only from day 13 ofpregnancy (data not shown), concomitantly with thatof the endogenous β-casein mRNAs (Persuy et al.,1992). Although this is just a qualitative analysis, itdemonstrates that the OZF-transgene is, at least par-tially, developmentally regulated. Homozygous micewere derived from lines 08 and 24 to potentiallyfurther increase the transgene expression level andtheir mammary gland development and lactation werefurther analysed.

OZF homozygous mice cannot sustain full growth ofthe pups

All transgenic females analysed (OZF 08, n= 4 andOZF 24, n= 1) showed a lactation failure (Figure 2and data not shown). The weight of F1 pups fed byOZF transgenic mice was smaller compared to that ofthe same number of littermates fed by F1 mice, anobservation already observed but not quantified withheterozygous transgenic females. In addition, 25% ofhomozygous mice from line 08 were unable to lactate

509

Figure 3. Whole mount analysis of mammary gland during pregnancy (day 13 and 18) and lactation (day 7) in OZF 08, 24, and wild-typemice. The inguinal mammary glands were dissected from female mice at the indicated stages and stained with alum carmin. (A), (B), and(C): Mammary gland from wild-type, OZF 08, and 24 mice, respectively, at day 13 of pregnancy. Magnification: 200-fold (D), (E), and(F): Mammary gland from wild-type, OZF 08, and 24 mice, respectively, at day 18 of pregnancy. Magnification: 400-fold (G), (H) and (I):Mammary gland from wild-type, OZF 08, and 24 mice, respectively at day 7 of lactation. The arrow indicate the alveoli filled of milk seen in(G). Magnification: 400-fold.

correctly and to keep their pups alive after one day oflactation.

A decrease in the growth rate of the pups could re-sult from either the secretion of milk with a very poornutritive value or a reduced overall milk production.Milk was collected from two mice from line 08 and itsoverall composition globally compared to that of F1animals. No significant difference could be observedin both the protein concentration and composition, asdeduced from Coomassie blue staining of samples fol-lowing SDS-PAGE (data not shown). These resultsstrongly suggested that the lactation defect observedin OZF transgenic mice more probably resulted froma reduced volume of milk production. Analysis of themammary gland development in these mice was thusundertaken.

Impaired mammary gland development in OZFhomozygous transgenic mice

Comparative mammary gland whole mount analysisbetween OZF 08, 24, and F1 mice was performed at

different times of pregnancy (day 8, 13, and 18) and atday 7 of lactation. Mice analysed were of similar ageand in their first gestation or lactation.

At day 8 of pregnancy, no significant differencein the mammary gland development was observedbetween OZF mice from the three genotypes (data notshown), which is consistent with the lack of detectabletransgene expression at this early gestation stage. Atday 13 of pregnancy, mammary gland developmentaldifferences could be detected between OZF 08, OZF24 mice, and F1 mice (Figure 3). A growth delayof the mammary gland with a less advanced devel-opment of alveolar buds in both OZF 08 mice butonly in 1 out of the 2 OZF 24 mice analysed wasnoticeable when compared to F1 controls. At day 18of pregnancy, a great delay in the development of themammary gland was observed in one of the two OZF08 and 24 mice compared to that of F1 control mice(Figure 3). Alveolar buds, which has already differen-tiated into individual alveoli in F1 mice at this stage ofpregnancy, are much less developed in OZF transgenic

510

Figure 4. Histological analysis of mammary gland during pregnancy (day 8, 13, and 18) in OZF 08, 24, and wild-type mice. All sectionshave been stained with Toluidine blue. (A), (B) and (C): Mammary gland from wild-type, OZF 08 and 24 mice, respectively, at day 8of pregnancy. Magnification: 30-fold. (D), (E) and (F): Mammary gland from wild-type, OZF 08 and 24 mice, respectively at day 13 ofpregnancy. Magnification: 200-fold. (G), (H), and (I): Mammary gland from wild-type, OZF 08 and 24 mice, respectively at day 18 ofpregnancy. Magnification: 200-fold.

mice (Figure 3). Such mammary gland developmentaldifferences were also observed at day 7 of lactationbetween OZF 08 and 24 mice and F1 mice (Figure 3).Indeed, two of the four OZF 08 and three out of thefour OZF 24 mice displayed a great delay in mam-mary gland differentiation compared to F1 mice at thesame stage (Figure 3). The size of the lobuloalveolarstructures were greatly reduced in the transgenic mice(Figure 3).

Taken together, these data indicate that OZF ex-pression during the second half of pregnancy andlactation is associated in the two analysed lines withan impaired mammary gland development. To fur-ther analyse this phenotype, histological analyses wereundertaken.

Histological examination of the mammary glandsfrom OZF homozygous transgenic mice revealedstriking differences with that of F1 animalsAs observed using whole mount analysis, no signif-icant difference in the histology of the mammary gland

was noted between F1 and OZF 08 and 24 mice at day8 of pregnancy, which is consistent with the lack ofexpression of the transgene at this developmental stage(Figure 4). The adipocyte to acini ratio was high inmice of the three genotypes.

At day 13 of pregnancy, a decrease in the num-ber and diameter of the acini and a higher adipocyteto acini ratio was observed in the mammary gland ofOZF 08 mice when compared to that of F1 mice (Fig-ure 4). In contrast, the mammary gland histology ofF1 and OZF 24 mice was similar (Figure 4). Althoughthe RT-PCR analysis of OZF transcripts was not quan-titative, the results obtained suggested that mice fromline 08 expressed the transgene more than those of line24. The present observation might be consistent with adose-dependent associated phenotype, at least duringthe early stages of the transgene expression. At day 18of pregnancy, the adipocyte to acini ratio was higher,the size of the acini and the number of fat globulesboth in the lumen and in the cytoplasm of the epithelialcells were greatly reduced in the mammary gland of

511

Figure 5. Histological analysis of mammary gland at day 7 of lactation in OZF 08, 24, and wild-type mice. All sections have been stained withToluidine blue. Mammary gland from wild-type (1), OZF 08 (2), and 24 (3) mice. Magnifications: 200-fold (A) or 300-fold (B). The presenceof fat globules located inside the acini lumen and epithelial cells is indicated (arrow in B1). Note the very low amount of fat globules in theacini lumen from OZF transgenic mice.

both analysed mice from both OZF lines compared towhat seen in F1 control animals (Figure 4). The lumenof the acini of the transgenic mice apparently failedto expand at day 18 of pregnancy. Strong mammarygland developmental differences were also observedat day 7 of lactation (Figure 5). The adipocyte to aciniratio was higher in OZF compared to F1 mice. Trans-genic mice had very few if any fat globules in thelumen of their acini whereas they are abundant in F1mice (Figure 5). Since the same phenotype was ob-served in both analysed independent transgenic lines,it is unlikely that it results from the transgene integra-

tion site but rather is the direct consequence of OZFderegulation.

Overall, all these data suggested an OZF dependentincrease of the mammary gland adipocytes to acini ra-tio in transgenics starting at mid-gestation and furtherincreasing at late gestation and early lactation. Thiswould argue either for an inhibitory effect of OZF onmammary epithelial cell proliferation and/or differen-tiation or for an OZF-induced apoptosis. Further ex-periments would be required to discriminate betweenthese different hypothesis. Within the acini, only lowamount or absence of fat globules were observed in

512

the lumen and in the cytoplasm of the epithelial cellsat day 18 of gestation and day 7 of lactation. It wassuggested that the expression of milk components washeterogeneous in the mammary gland, at least in ru-minants, and that small groups of epithelial cells couldbe specialised in the synthesis of some major milk pro-teins, such as α-lactalbumin, or fat (Molenaar et al.,1992). If a similar situation occurs in rodents, OZFinduced phenotype is apparently more pronounced forthe later group of cells.

OZF transgenics as a step forward towards theunderstanding of the protein function ?In humans, OZF was found to be overexpressed inbetween 15 and 25% of pancreatic tumors and in 85%of colon carcinoma (Ferbus et al., 1999). The situ-ation with regard to human mammary carcinoma isless clear (Blottière, 1999). During the time courseof this experiment, none of the transgenic mice ap-peared to have developed mammary carcinoma des-pite, for some of them, having experienced severalcycles of pregnancy/lactation. It would suggest thatthe deregulation of OZF expression per se is notenough to induce mammary tumorigenic changes asseveral other oncogenes did or that the β-casein pro-moter did not induce it at the critical physiologicaltime.

As a zinc finger protein, OZF is suspected to bindto DNA specific consensus sequences. In vitro bindingwith random oligonucleotides led to the identifica-tion of a 15 nucleotide motif (Pibouin & Goubin,unpublished). The absence of a clearly recognisabletransactivating or transinhibiting domain suggests thatOZF may bind as a competitor for other DNA-bindingtranscription factors. Alternatively, it may interactwith other proteins through its zinc finger domains toindirectly modulate gene transcription. So far, noneof the cellular partner and/or target of OZF has beenclearly identified. Deregulation of its expression in themammary gland induced a marked phenotype, sug-gesting that it affect the normal lactation physiologicalprocess. Comparative analysis of the mammary epi-thelial cell transcriptome in these transgenic and F1mice might allow the identification of genes that areaffected by OZF deregulation. Because this trans-gene transcription induction time is relatively welldefined, it may be possible to perform such analysis ata physiological stage where mostly only OZF dereg-ulation primary consequences will be observed. Thus,these mice could be a powerful tool to put in practicethe transcriptome methodology to better understand

the function of this protein. Experiments are underwayto test this hypothesis.

Acknowledgements

We are most grateful to Drs L Blottière and D Ferbus(UMR 147, Inst. Curie) for the murine OZF cDNA, toB Nicolas (INRA) for the photos and to Drs R Salesse(BCM, INRA) and EP Cribiu (LGBC, INRA) for theirsupport.

References

Bellefroid EJ, Lecocq PJ, Benhida A, Poncelet DA, Belayew Aand Martial JA (1989) The human genome contains hundred ofgenes coding for finger proteins of the Krüppel type. DNA 8:377–387.

Blottière L (1999) Etude des gènes OZF (Only Zinc Fingers) hu-main et murin: Incidence dans les cancers du sein. Thesis,University Paris VI, Pierre et Marie Curie, Paris, France.

Chowdhury K, Deutsch U and Gruss P (1987) A multigene familyencoding several ‘finger’ structures is present and differentiallyactive in mammalian genomes. Cell 48: 771–778.

Ferbus D, Flechon A, Muleris M, Li Y, Hanash S, Terris B et al.(1999) Amplification and over-expression of OZF, a gene encod-ing a zinc finger protein, in human pancreatic carcinomas. Int JCancer 80: 369–372.

Hogan B, Costantini F and Lacy E (1986) Manipulating the MouseEmbryo. A Laboratory Manual. Cold Spring Harbor Laboratory,Cold Spring Harbor NY.

Le Chalony C, Apiou F, Pibouin L, Dutrillaux B and Goubin G(1996) Constitutive amplification of a zinc finger protein genein cattle. DNA Cell Biol 15: 83–88.

Le Chalony C, Prosperi MT, Haluza R, Apiou F, Dutrillaux Band Goubin G (1994) The OZF gene encodes a protein con-sisting essentially of zinc finger motifs. J Mol Biol 236: 399–404.

Molenaar AJ, Davis SR and Wilkins RJ (1992) Expression of α-lactalbumin, αs1-casein and lactoferrin genes is heterogeneousin sheep and cattle mammary tissue. J Histochem Cytochem 40:611–618.

Persuy MA, Legrain S, Printz C, Stinnakre MG, Lepourry L,Brignon G et al. (1995) High-level, stage- and mammary-tissue-specific expression of a caprine κ-casein-encoding minigenedriven by a β-casein promoter in transgenic mice. Gene 165:291–296.

Persuy MA, Stinnakre MG, Printz C, Mahe MF and Mercier JC(1992) High expression of the caprine β-casein gene in trans-genic mice. Eur J Biochem 205: 887–893.

Puissant C, Lennaouar M, Le Chalony C, Goubin G and HoudebineLM (1996) OZF gene expression in growing mouse and rabbitmammary gland in rabbit mammary cells and Nb2 cells underprolactin action. Biochem Mol Biol Int 38: 543–552.

Rosenberg UB, Schröder C, Preiss A, Kienlein A, Coté S, RiedeI et al. (1986) Structural homology of the product of the Dro-sophila Krüppel gene with Xenopus transcription factor IIIA.Nature 319: 336–339.

Rudolph NS (1999) Biopharmaceutical production in transgeniclivestock. TibTech 17: 367–374.

513

Vilotte JL, Soulier S, Stinnakre MG, Massoud M and MercierJC (1989) Efficient tissue-specific expression of bovine α-lactalbumin in transgenic mice. Eur J Biochem 186: 43–48.

Whitelaw CBA, Springbett AJ, Webster J and Clark AJ (1993)The majority of Go transgenic mice are derived from mosaicembryos. Transg Res 2: 29–32.