- Email: [email protected]
Spontaneous germline amplification and translocation of a transgene array
Mutation Research 486 (2001) 125–136
Spontaneous germline amplification and translocation of a transgene array Margot Kearns a , Christine Morris b , Emma Whitelaw a,∗ a
b
Department of Biochemistry, G08, University of Syndey, Sydney NSW 2006, Australia Cytogenetic and Molecular Oncology Unit, Department of Pathology, Christchurch School of Medicine, P.O. Box 4345, Christchurch, New Zealand Received 8 November 2000; received in revised form 20 March 2001; accepted 9 April 2001
Abstract The majority of the mammalian genome is thought to be relatively stable throughout and between generations. There are no developmentally programmed gene amplifications as seen in lower eukaryotes and prokaryotes, however a number of unscheduled gene amplifications have been documented. Apart from expansion of trinucleotide repeats and minisatellite DNA, which involve small DNA elements, other cases of gene or DNA amplifications in mammalian systems have been reported in tumor samples or permanent cell lines. The mechanisms underlying these amplifications remain unknown. Here, we report a spontaneous transgene amplification through the male germline which resulted in silencing of transgene expression. During routine screening one mouse, phenotypically negative for transgene expression, was found to have a transgene copy number much greater than that of the transgenic parent. Analysis of the transgene expansion revealed that the amplification in the new high copy transgenic line resulted in a copy number approximately 40–60 times the primary transgenic line copy number of 5–8 copies per haploid genome. Genetic breeding analysis suggested that this amplification was the result of insertion at only one integration site, that it was stable for at least two generations and that the site of insertion was different from the site at which the original 5–8 copy array had integrated. FISH analysis revealed that the new high copy array was on chromosome 7 F3/4 whereas the original low copy transgene array had been localised to chromosome 3E3. DNA methylation analysis revealed that the high copy transgene array was heavily methylated. The amplification of transgenes, although a rare event, may give insight into amplification of endogenous genes which can be associated with human disease. © 2001 Elsevier Science B.V. All rights reserved. Keywords: DNA amplification; Transgene silencing; Translocation; Germline mutation
1. Introduction The majority of the mammalian genome is thought to be relatively stable throughout and between generations. For survival, the number of genes must ∗ Corresponding author. Tel.: +61-2-9351-2549; fax: +61-2-9351-4726. E-mail address: [email protected] (E. Whitelaw).
remain essentially the same, however, non-essential non-coding regions of the genome do exhibit a certain plasticity. Gene amplification is a common feature in prokaryotes [1]. The amplification often results in increased protein production and is thought to be an adaptive mechanism in response to a changing environment [2]. Specific regions of the genome are hotspots for amplification [1] and the unit of DNA can be amplified up to 500 times [3]. Gene amplification
0921-8777/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 8 7 7 7 ( 0 1 ) 0 0 0 8 4 - 2
126
M. Kearns et al. / Mutation Research 486 (2001) 125–136
is not restricted to prokaryotes but has also been observed in yeast, Drosophila and vertebrates (reviewed in [2]). Developmentally regulated gene amplification of specific genes occurs in some organisms. Amplification of rRNA genes occurs during oogenesis in many species but has been studied mainly in Xenopus laevis (reviewed in [4]). Development-specific gene amplification is also seen in Drosophila where the chorion genes which encode the major protein constituent in eggshells are amplified for a limited time [5]. In yeast the rRNA genes are also in a state of flux with increases and losses within the tandemly repeated units [6,7]. Other examples of gene duplication in Saccaromyces cerevisiae involve translocations in conjunction with the amplification process. In one case in yeast a gene, which normally lies on chromosome 2, has been reported to undergo amplification and translocation to the telomere of another chromosome [8]. In mammals there are no known developmentally regulated amplification events, however, numerous examples of amplification in cell lines and tumours have been reported [2,9]. The treatment of cell lines with certain cytotoxic agents can result in selection of mutants which in many cases have developed resistance to the offending agent by gene amplification. The amplification results in increased production of a protein that can neutralise the effects of the drug [10]. Increases in gene copy number of 10- to 50-fold have been observed [2] and often these selection-driven amplifications recede once the selection is removed [11]. Gene amplification has been linked to tumorigenesis following the observation that many tumor cells were found to contain chromosomal abnormalities, double minutes (DMs) and homogeneously staining regions (HSR) associated with amplified DNA (reviewed in [9]). Although, transgene integration is usually stable, of the thousands of transgenic mice produced, a few have been reported to undergo heritable amplification [12,13]. There has been one report of the amplification of a murine transgene that involved the germline expansion of a 4–5 copy integrant to one containing 20–21 copies. The expanded locus continued to express the transgene, however, reduced levels of RNA were seen in some offspring [14]. Here we report the spontaneous germline amplification of a murine transgene array that results in loss
of transgene expression. During routine screening of a previously reported transgenic line, one individual with an increase in copy number was identified. In this mouse the transgene copy number was 200–300 copies per mouse genome, which indicated a 40–60-fold amplification of the original array. This spontaneous amplification was investigated and it was found that the expansion resulted in some rearrangement of the transgene array and a loss of transgene expression. DNA methylation analysis of the transgene array revealed that the expanded array was heavily methylated. Breeding studies revealed that the initial amplification event had probably taken place in the germline and also suggested a translocation event was associated with the amplification. Translocation of the high copy array to a new chromosomal location was confirmed by FISH analysis. This event highlights the inherent instability of the mammalian genome. Although, a rare event, amplification and translocation of transgenes can occur and study of these events may help us to understand the mechanisms underlying the amplification of endogenous genes. 2. Materials and methods 2.1. Transgene construct and the generation of transgenic mice The transgene construct, DM2, contains the human -globin promoter from −127 to +6 linked to an oligonucleotide (SDK) containing the Shine-Dalgarno and Kozak sequences, the lacZ reporter gene with the SV40 polyadenylation signal and a 4.1 kb fragment containing the human ␣-globin locus enhancer region (HS-40) previously shown to direct high levels of expression in erythroid cells [15]. Three transgenic lines were produced by microinjection of DNA into the pronucleus of fertilised mouse oocytes from the Pathology Oxford (P.O.) mouse strain [16]. Initial studies of these lines have been previously reported [16,17]. Of these lines 239B is the founding line of this study. The copy number of the line 239B was estimated to be 5–8 [16,17]. 2.2. Southern blotting and hybridisation Southern blotting of genomic DNA was carried out following digestion with EcoR V. The membrane
M. Kearns et al. / Mutation Research 486 (2001) 125–136
was hybridised with a 4.1 kb fragment containing the ␣HS-40 region, a 3.5 kb fragment containing the lacZ gene or a 1.4 kb fragment containing the mouse ␣-globin gene. The probes were labelled by random priming using [␣-32 P]dCTP (Bresatec) and the Megaprime Labelling Kit (Amersham). When screening for the presence of the human -globin promoter the MC 286 oligo (5 TAG TGC TCA GCC CCC AGG GTC CCC TTA TAT ACA GG 3 ) a kind gift from Dr. Merlin Crossley, Department of Biochemistry, University of Sydney) was endlabelled with [␥-32 P]ATP (Bresatec) and used as a probe. 2.3. Methylation analysis of the transgene array DNA was purified from circulating erythroid cells of 12.5 d.p.c. transgenic embryos. Aliquots of 2 g DNA were digested with BamH I alone or BamH I with either Msp I, Hpa II or Hha I. The resulting fragments were analysed by Southern transfer and hybridisation following separation on a 1.2% agarose gel. The membrane was probed with either a 4.1 kb fragment containing the ␣HS-40 region or a 1.4 kb mouse ␣-globin fragment. 2.4. β-Galactosidase activity in whole cells Transgene expression was analysed at the single cell level in erythroid cells from mice at 3 weeks after birth as described previously [16]. Blood from postnatal mice was collected by tail bleeding, and a dilute suspension of cells was fixed and washed for 5 min each and then stained with 5 bromo-4chloro-3-indolyl--d-galactopyranoside (X-gal) at 37◦ C for up to 24 h. A minimum of 1 × 105 erythrocytes were inspected for detectable blue stain; non-expressors are those mice having no stained cells in greater than 105 erythrocytes. For positively staining mice, at least 500 erythroid cells were counted to determine the percentage of stained cells. 2.5. Culturing of mouse embryonic fibroblasts Embryos resulting from a cross between a homozygous transgenic male and a wildtype P.O. female were dissected at 12.5 d.p.c. and bled. After removing the head and liver, the remainder of two embryos were
127
pooled then homogenised using a 19G needle and syringe in AmnioMax C100 media (Life Technologies). The emulsion was treated with collagenase (2000 U) for 4 h then incubated at 37◦ C in 5% carbon dioxide. About 24 h prior to harvesting, the confluent cultures were passaged. Mouse metaphase cells were prepared using standard colcemid and harvesting procedures. 2.6. Fluorescent in situ hybridisation The plasmid, DM2, containing the entire transgene construct, was random-primer labelled with biotin (High Prime kit, Boehringer Mannheim), and hybridised without repeat suppression to denatured transgenic mouse metaphase chromosomes spread on microscope slides using methods previously described [18]. Slides were incubated overnight at 37◦ C, and washed to a final stringency of 0.1 × SSC at 60◦ C before signal amplification with FITC-avidin and biotinylated anti-avidin (Vector Laboratories). Slides were mounted in glycerol-containing anti-fade and DAPI as counterstain for G-band visualisation before analysis under a Leitz Aristoplan microscope. Selected fields were transferred for image-processing to a Power MacIntosh 8100–80 computer via a Photometrics cooled CCD camera using appropriate wavelength filters, IPLab Spectrum and Multiprobe extension software. G-band nomenclature for description of mouse karyotypes and probe localisation was according to Nesbitt and Francke [19]. 2.7. PCR analysis of transgene–transgene and transgene–genome junctions Genomic DNA was analysed by PCR using a combination of the following primers: HS-40 1906- 5 GCG AAG CTG TTG TGA GGA C 3 , HS-40 2012-5 ACA TCT GTA TCA TTA ACA CTG AAT GAC AGC 3 , HS-40 4014- 5 GCA AGT CTG ACT GCA CAA TG 3 , HS-40 4108- 5 ACC CAC TGC ACT AGC TGG G 3 , 239B Genome 3 (1)- 5 ATC TTG CTC AGC TGA GCT GG 3 , SDK- 5 ATG GTG GCC TCC GAC CTG C 3 . A typical PCR reaction contained 1 l undigested genomic DNA, 5 l 10 X reaction buffer (Life Technologies), 5 l DMSO, 0.4 mM each primer, 0.2 mM each dNTP (dATP, dCTP, dGTP, dTTP), 3.0 mM MgCl2 , 1 U Taq (Life Technologies), and MQW to
128
M. Kearns et al. / Mutation Research 486 (2001) 125–136
a final volume of 50 l. The following PCR cycling conditions were used, 1 cycle of 94◦ C for 5 min followed by 30 cycles of 94◦ C for 30 s, 55◦ C for 30 s then 72◦ C for 60 s. The PCR was completed with one cycle of 72◦ C for 10 min. PCR products were visualised with UV light following gel electrophoresis. 3. Results 3.1. Spontaneous transgene amplification The transgenic line 239B [17] was generated using a construct which contained 127 bp of the human -globin promoter linked to the E. coli lacZ reporter gene and the human ␣-globin HS-40 enhancer. The transgenic line was estimated to have a copy number of 5–8 transgenes per genome integrated into one chromosomal location. To maintain stocks of this mouse line, hemizygous transgenic males were mated to wildtype females and transgenic offspring were identified by Southern transfer of tail DNA and hybridisation with a transgene-specific probe. During routine screening it was noticed that the transgene DNA of one pup appeared to be at a much higher copy number than its littermates. This individual was a female and appeared physically no different from her siblings. Tail sampling was repeated and the same result was obtained (Fig. 1). The membrane was hybridised with both a probe from the lacZ gene in the transgene and a probe from the ␣HS-40 fragment. In both cases the increase in the intensity of the signal was similar suggesting that the amplification most likely involved the whole transgene. The membrane was stripped and rehybridised with a mouse ␣-globin gene to correct for sample loading on the gel. In order to overcome any possible saturation of the signal which could occur at such high copy numbers, dilutions of the DNA from the high copy mouse were prepared (Fig. 1). Using these dilutions, the degree of amplification was determined to be approximately 40 times that of the transgenic parent. To determine whether the expansion had occurred prior to the formation of germ cells, the high copy female was mated to a wildtype male and the transgene copy number of her offspring was determined. All 24 transgenic offspring tested carried the transgene expansion (an example is G1 in Fig. 1). The initial DNA
Fig. 1. Quantitation of the DNA amplification and comparison of copy number in successive generations. (I) Genomic DNA extracted from tails samples collected from the transgenic parent of the high copy mouse (P) and the founder high copy mouse (F) were digested with EcoR V. Equal concentrations of samples P and F were electrophoresed along with dilutions (1:5, 1:10, 1:50 and 1:100) of the high copy digest as well as digested DNA collected from a sibling (S) of the high copy founder, and offspring from the next 2 generations of the high copy line (G1 and G2). The membrane was hybridised with either the radiolabelled 3.5 kb lacZ probe, 4.1 kb HS-40 probe or the 2.2 kb mouse ␣-globin probe.
amplification, therefore, must have occurred either in the germline of the 239B transgenic parent or early in the development of the founder high copy female prior to determination and migration of primordial germ cells. The new high copy mouse line was designated 239BHC. The high copy line was bred a further generation and the expansion was, once again, stably inherited (G2 in Fig. 1). 3.2. Rearrangement of the transgene in the high copy expansion The arrangement of copies in the transgene array of the parent line, 239B, and the high copy line, 239BHC, was investigated using restriction enzyme digestion, Southern transfer and hybridisation. As there is only one EcoR V recognition sequence in the transgene, digestion with this enzyme should produce one full length transgene–transgene band of 8.1 kb if the array is in a purely head to tail orientation. DNA from the parental line, 239B, produced an intense band migrating at approximately 8 kb and a much
M. Kearns et al. / Mutation Research 486 (2001) 125–136
129
Fig. 2. Determination of transgene array in the 239BHC line. (I) Schematic representation of a transgene array depicting the EcoR V restriction sites. The probe is shown as a black bar. With a head to tail array a transgene–transgene band of 8.1 kb is generated; with head to head transgenes a 2.6 kb; band and with tail to tail transgenes a 13.6 kb band should be seen. (II) Genomic DNA from the transgenic parent of the high copy founder (P), the high copy founder mouse (F) and one of the offspring in the next generation (G1) was digested with EcoR V which cuts the 127/lacZ construct only once. Following electrophoresis and Southern blotting the membrane was hybridised with a radiolabelled 3.5 kb lacZ probe. The membrane was stripped and rehybridised with a probe to the mouse ␣-globin gene to check for equal loading.
fainter band of higher molecular weight that migrated more slowly (Fig. 2). This may represent the edge fragment: transgene DNA and the adjoining genomic DNA. The predominant band obtained for the high copy line, 239BHC, was also approximately 8 kb representing the head to tail arrangement. The presence of this band suggested that there was no substantial loss of sequence in the majority of the transgene copies. However, the banding pattern obtained for the 239BHC line was complex: the presence of multiple smaller and larger bands suggested that large deletions had occurred in a subpopulation of copies (Fig. 2). 3.3. The transgene amplification resulted in loss of transgene expression Previously it had been demonstrated that male transmission of the 239B transgene invariably resulted
in transgene expression in erythrocytes [17]. The number of erythrocytes expressing the transgene was affected by parent of origin. The percentage of erythroid cells expressing the transgene at 3 weeks of age in mice resulting from male transmission was 7.5%. Transgene expression was analysed in the founder 239BHC mouse and following passage through both germlines at 3 weeks of age. In all cases no transgene expression was detected. A total of 61 mice were analysed; 36 following male transmission and 25 following female transmission. 3.4. Transgene silencing correlates with hypermethylation As methylation has been shown to be associated with the lack of expression in repeat induced gene silencing in mice [20] and other organisms [21], the methylation state of the high copy array was
130
M. Kearns et al. / Mutation Research 486 (2001) 125–136
Fig. 3. Methylation state of the transgene in the line 239BHC. DNA extracted from the erythrocytes of 12.5 d.p.c. embryos (following male transmission) was digested with BamH I alone or in combination with Msp I (M) or the methylation sensitive restriction enzymes Hpa II (H) or Hha I (Hh). Following electrophoresis the membrane was hybridised with a radiolabelled 4.1 kb HS-40 probe, then stripped and rehybridised with 2.2 kb mouse ␣-globin probe. The DNA from the high copy line 239BHC is more heavily methylated than that from the low copy line 239B.
investigated. DNA was purified from erythrocytes collected from 12.5 d.p.c. embryos resulting from the cross between a transgenic male and a wildtype female. DNA was digested with BamH I or BamH I plus either Msp I, Hpa II or Hha I. Southern blotting and hybridisation was carried out as described previously and the methylation state of selected sites within the HS-40 region of the transgene was determined (Fig. 3). The presence of many larger bands in the Hpa II lane with the 239BHC DNA suggests that the transgene was more heavily methylated in the high copy embryos compared with the 239B controls. The transgene array however was not methylated in all cells as the 1.6 and 1.0 kb Msp I/Hpa II bands were still present. It is interesting to note that the BamH I
only digest produced a 4.1 kb band in the 239B line but not with the 239BHC line (Fig. 3). Instead a higher molecular weight band was present suggesting that one of the BamH I restriction sites was absent and digestion at the remaining site has generated a near full length transgene–transgene band (approximately 8 kb) (Fig. 3). 3.5. Loss of transgene sequence in the amplification process The loss of transgene expression seen with the amplification in the 239BHC sub-line could also result from the loss of important regulatory sequences. In this case any loss of sequence could not be extensive
M. Kearns et al. / Mutation Research 486 (2001) 125–136
131
Fig. 4. Loss of sequence at the ends of the transgene construct. (I) PCR analysis of transgene–transgene junctions. Schematic representation of a 239B head to tail transgene–transgene junction showing approximately position of primers used in various PCR reactions with outcome indicated. PCR was performed on genomic tail DNA collected from 239B (L) 239BHC (H) and compound hemizygotes 239B/239BHC mice (L/H). M represents a 1:1 mix of DNA from 239B and 239BHC samples and W represents wildtype DNA. (II) 239BHC DNA hybridises to a radiolabelled oligonucleotide containing human -globin promoter sequence. Southern blotting of genomic tail DNA, digested with Pst I, was followed by hybridisation using a radioactive endlabelled oligonucleotide MC 286 which contains human -globin promoter sequence from −35 to −2. Samples included the high copy line 239BHC and wildtype DNA.
since the EcoR V digests give a head to tail band approximately equivalent to the full length transgene (Fig. 2). As some degree of sequence loss was evident with BamH I digest described above, however, the situation was investigated further by PCR analysis. A number of different primers (three from 3 region of the transgene and two from the 5 region of the transgene) which should amplify the transgene–transgene junctions (Fig. 4(I)), failed to produce any product with the 239BHC line (data not shown). Using the SDK primer and two primers that anneal to sequences near the centre of the HS-40 element (HS-40 1906 and HS-40 2012) a product was obtained. The expected size of a 239B transgene–transgene band is approximately 2.5 kb. A band of this size was obtained for the 239B samples but a band of approximately 2.0 kb was obtained with the 239BHC samples (Fig. 4(I)). This would indicate that approximately 500 bp of
sequence at the transgene junction was deleted in the amplification of the process. As it was possible that the loss of sequence included the promoter sequence, Southern blotting was performed using a probe to the human -globin promoter, oligo MC 286 (from −35 to −2 of the human -globin promoter). This probe successfully hybridised to 239BHC DNA (Fig. 4(II)). Therefore, at least some of the promoter sequence, approximately 30 bp, was present in the high copy transgenic line suggesting that the absence of PCR product with the HS-40 4014 and HS-40 4108 primers and the -promoter and SDK primer pairs, was due to the loss of sequence at the 3 end of the HS-40 fragment. This is also consistent with the loss of the BamH I site which lies at the very end of the transgene construct. The precise -globin promoter sequence present in the 239BHC line is unknown.
132
M. Kearns et al. / Mutation Research 486 (2001) 125–136
The deletion has involved some sequence at the 3 end of the HS-40 fragment. The activity of the HS-40 element has been localised to a 350 bp core region containing multiple transcription factor binding sites, situated between positions 2991 and 3335 of the 4.1 kb element [22]. The deletion in the 239BHC array does not include this region. 3.6. The 239BHC transgene array is situated at a different chromosomal location To investigate whether the amplified DNA was at the same chromosomal location as the 239B site or at a new chromosomal location due to a translocation a number of approaches were employed: PCR analysis of the trangene–genome junction using the 3 integration site from 239B; systematic breeding of mice carrying both the 239B allele and the 239BHC allele; and FISH. As the 3 genomic flanking sequence at the site of 239B integration had been determined [17], the presence of this transgene–genome junction in the high copy line would suggest that transgenic DNA remained at this locus. Using a primer to HS-40 and a primer to the 3 flanking DNA, a PCR product was only obtained for the low copy controls and DNA samples which contained both high and low copy DNA (Fig. 5). This suggested that there was no resid-
ual transgene DNA at the previous site of integration on chromosome 3 band E3. However, this analysis does not rule out the possibility that the junction was lost in the amplification process. Genetic analysis was carried out to see if the 239B allele and the 239BHC allele were on separate chromosomes. This was achieved by breeding a homozygous 239B male mouse with a hemizygous 239BHC female (Fig. 6). All the offspring carried the low copy allele which conferred transgene expression. After screening for the presence of the high copy allele by Southern blotting and hybridisation, a positive high copy mouse was selected for further breeding. This compound hemizygous mouse now carried a low copy allele and a high copy allele (239B/239BHC). By breeding this male mouse to a wildtype female the genetics of segregation of alleles could be analysed. If the amplification was at the same chromosomal location or so close on the same chromosome that recombination did not occur, then segregation of the alleles would result in approximately 50% of the offspring carrying the low copy allele which would confer transgene expression and the remaining offspring would carry the silenced high copy allele. If the amplification was at a new chromosomal location which segregated independently from the low copy allele then 4 different groups of mice would be obtained in equal proportions: 239B/- (low copy hemizygotes);
Fig. 5. PCR analysis of transgene–genome junction. PCR was performed on genomic tail DNA using the HS-40 4014 primer and a primer to the 3 genomic flanking DNA at the 239B site of integration. PCR was performed on genomic tail DNA collected from 239B (L) 239BHC (H) and compound hemizygotes 239B/239BHC mice (L/H). M represents a 1:1 mix of DNA from 239B and 239BHC samples and W represents wildtype DNA. PCR product was only obtained in those samples containing the 239B allele and not in 239BHC hemizygote samples KSH and MK are markers, Bluescript KS cut with Hpa II and 100 bp ladder, respectively.
M. Kearns et al. / Mutation Research 486 (2001) 125–136
133
Fig. 6. Genetic analysis of segregation of two independent transgene arrays, 239B and 239BHC. Mating a homozygous low copy (239B) male to a high copy (239BHC) female produces mice carrying the low copy allele as well as the high copy allele. Breeding these compound hemizygous mice to wildtype mice will segregate the alleles. If the alleles are on different chromosomes that randomly assort during meiosis then four genotypes are possible in the next generation: low copy, high copy, compound heterozygous and wildtype. Phenotypes were determined by staining erythrocytes with X-gal and detecting the presence of blue stained cells under bright field microscopy. The presence of blue staining cells indicated that the mouse carried the 239B allele. 239BHC genotypes were determined by screening genomic DNA for the presence of the transgene by Southern transfer and hybridisation with a transgene-specific probe. A total of 101 mice were analysed. The 25 offspring that displayed transgene expression in the presence of the high copy array is indicative of reconstitution of the parental genotype — the compound hemizygote (239B/239BHC). Statistical analysis using χ 2 based on the assumption of segregation of independent chromosomes giving proportions of 1:1:1:1 gave a χ 2 value of 0.82: with 3 degrees of freedom, 0.9 > P > 0.7.
239B/239BHC (compound hemizygotes); 239BHC/(high copy hemizygotes); and non-transgenic. The 101 individual mice analysed did indeed fall into 4 groups supporting the hypothesis that the two transgene arrays were at different chromosomal locations. The genetic studies described above indicated that the 239BHC transgene array was at a different chromosomal location to the 239B integrant which had previously been shown to have localised to mouse chromosome 3 region E3 (Fig. 7) [17]. In order to localise the high copy array, primary fibroblasts were cultured from embryonic tissue collected from 239BHC transgenic embryos and FISH analysis was performed. A clear signal that hybridised to the proximal region of chromosome 7 band F3/4 was observed in all metaphase preparations examined. This band is situated near the telomere (Fig. 7) and as telomeric regions of chromosomes typically contain heterochromatin, this also could be a contributing factor to the transgene silencing observed with the 239BHC amplification.
4. Discussion Germline transformation of mice is generally thought to produce a relatively stable transgene array that can be inherited for many generations. This study identified a germline amplification of a transgene array from 5 to 8 copies to 200–300 copies. The amplification appears to have occurred in the germline. It is possible that the original event happened very early in embryogenesis of the founder high copy mouse before the primordial germ cells were determined. If, however, this had been the case then the founder would be mosaic with some cells being low copy and some cells high copy. Using tail DNA there was no evidence of the previous transgene–genome junction or head to tail transgene–transgene junctions in the founder. It is therefore likely that the amplification occurred during gametogenesis in the male germline of the low copy parent. Most studies of gene amplification in mammalian cells have used cultured cells or tumor tissue [9]. In
134
M. Kearns et al. / Mutation Research 486 (2001) 125–136
Fig. 7. FISH localising the chromosomal positions of the 239B and the 239BHC transgene arrays. FISH was performed using the DM2 plasmid containing the 8.1 kb 127/lacZ construct used for microinjection to produce the transgenic line 239B. (I) Localisation of the 239B transgene integrant to mouse chromosome 3 band E3: (A) a representative metaphase cell and chromosome under fluorescence showing the single signal on chromosome 3. The fluorescent signal is indicated by the orange arrow; (B) chromosome 3 from panel A aligned with ideogram showing precise location of the probe signal in band E3. (II) Localisation of the 239BHC transgene integrant to mouse chromosome 7 F3/4: (C) a representative metaphase cell and chromosome under fluorescence showing the single signal on chromosome 7; (D) chromosome 7 from panel C aligned with G-banding depicting that the signal has hybridised to region F3/4.
these cases the amplifications in somatic cells often occur under selection pressure and result in increased protein production [9]. The lack of reports of germline amplification of genes may indicate that although these events occur they are not often observed due to embryonic lethality. The germline amplification of a transgene carrying a reporter gene, however, may not be hazardous to the developing embryo enabling these events to be detected. The 239BHC array was found to be hypermethylated. Hypermethylation of silent high copy transgene arrays have been observed in plants [21,23] and mice [20]. However, as repeat induced gene silencing is
also found in Drosophila [24,25] where there is no methylation it has been suggested that the relationship between gene silencing and methylation is not causal but consequential. Other studies on high copy transgene arrays report that even constituitively expressed house keeping genes are hypermethylated when in multi copy arrays [26]. It has been suggested that the increase in methylation with increased copy number could be the result of depletion of a limited pool of proteins that bind to DNA protecting it from methylation [26,27]. Why the expansion of transgene arrays [20] results in silencing, when amplification of oncogenes in
M. Kearns et al. / Mutation Research 486 (2001) 125–136
tumours and amplification in mammalian cell lines results in overexpression of the protein product, is not clear. It could be argued that the size of the expanded array or the number of repeats could influence silencing but it is noteable that the v-erbA transgene expansion of up to 20–21 copies did not result in silencing and oncogenes can be amplified up to 500 copies and still maintain expression. 4.1. Possible mechanisms involved in the amplification process A number of models of DNA amplification have been proposed based on extensive study of gene amplification in prokaryotes. Large scale amplifications, like that seen with 239BHC, are not thought to be due to unequal crossover or sister chromatid exchange as these events produce a maximum two-fold expansion per generation [2]. PCR analysis detected a transgene deletion of about 500 bp with no evidence of any copies containing the original transgene–transgene junctions found in the 239B low copy line. This suggests that the amplification involved one incomplete transgene unit of approximately 7.6 kb. This unit was amplified up to 300 times with the majority of the copies being in a head to tail array. The process also involved a translocation event. So a possible model would involve some unequal recombination event between two adjacent transgenes which produced a circle that contained a 7.6 kb portion of transgene sequence and trapped a replication fork. The rolling circle model would account for the high number of copies being in a tandem repeat array [28]. The newly formed concatamer then translocated to a telomeric region as seen in yeast with the alkaline phosphatase gene [8]. As this event is thought to have occurred during meiosis, the original transgene integrant and the new high copy translocated array could segregate into different spermatids. Many of the proposed models of amplification link the process to replication and recombination [29]. Recombination is a normal event in meiosis where it occurs after the chromosomes have duplicated and involves the cellular repair enzymes. In general, recombination occurs more frequently in female meiosis, however, in certain regions of the genome recombination occurs predominantly during male meiosis. It has been suggested that these male–female
135
differences in recombinational activity during meiosis may be related to differences in gene activity in imprinted regions [30–32]. One such differential region exists on human chromosome 15q11 which is associated with Prader–Willi and Angelman syndromes, where a region with a very high male recombination frequency is adjacent to a region with a high female recombination frequency [33]. Interestingly the amplification event in 239B which may have involved recombination within the transgene has also occurred at a locus that exhibits parental imprinting [17]. Although, a rare event, amplification and translocation of transgenes can occur but whether these events are more common at sites which have been introduced into the genome, than at endogenous loci, is unclear. Amplification of endogenous loci may go undetected due to lethality of chromosomal imbalances or gene silencing. Recently a number of cases of Pelzaeus– Merzbacher disease (PMD) have been attributed to this type of genetic instability [34]. The proteolipid protein gene (PLP) is normally present at chromosome Xq22. Mutations and duplications of this gene are associated with PMD. Hodes et al. [34] report one case in which an additional copy of the gene is found at a novel site on the X chromosome and no gross chromosomal rearrangements could be detected by standard karyotype analysis.
References [1] D. Romero, R. Palacios, Gene amplification and genome plasticity in porkaryotes, Ann. Rev. Genet. 31 (1997) 91–111. [2] G.R. Stark, G.M. Wahl, Gene amplification, Ann. Rev. Biochem. 53 (1984) 447–491. [3] P. Leblond, P. Demuyter, L. Moutier, M. Laakel, B. Decaris, J.M. Simonat, Hypervariability: a new phenomenon of genetic instability related to DNA amplification in Streptomyces ambofaciens, J. Bacteriol. 71 (1994) 419–423. [4] K.D. Tartof, Redundant genes, Ann. Rev. Genet. 9 (1975) 355–385. [5] A.C. Spradling, The organisation and amplification of two chromosomal domains containing Drosophila follicle cells, Cell 27 (1981) 193–201. [6] H.L. Klein, T.D. Petes, Genetic mapping of Ty elements in Saccharomyces cerevisiae, Nature 289 (1981) 144–148. [7] T.D. Petes, Unequal meiotic recombination within tandem arrays of yeast ribosomal DNA genes, Cell 26 (1980) 765–774. [8] P.E. Hansche, V. Beres, P. Lange, Gene duplication in Saccharomyces cerevisiae, Genetics 88 (1978) 673–687.
136
M. Kearns et al. / Mutation Research 486 (2001) 125–136
[9] M. Schwab, Amplification of oncogenes in human cancer cells, BioEssays 20 (1998) 473–479. [10] C. Ramel, H. Cederberg, J. Magnusson, E. Vogel, A.T. Natarajan, L.H. Mullender, J.M. Nivard, J.M. Parry, A. Leysen, M.A. Comendador, L.M. Sierra, J.A. Ferreiro, J.A.S. Consuegra, Somatic recombination, gene amplification and cancer, Mutat. Res. 353 (1996) 85–107. [11] D.A. Haber, R.T. Schmidt, Unstable amplification of an altered dihydrofolate reductase gene associated with doubleminute chromosomes, Cell 26 (1981) 355–362. [12] M. Lacey, S. Alpert, D. Hanahan, Bovine papillomavirus genome elicits skin tumours in transgenic mice, Nature 322 (1986) 609–612. [13] R.L. Brinster, H.Y. Chen, A. Messing, T. van Dyke, A.J. Levine, R.D. Palmiter, Transgenic mice harboring SV40 T-antigen genes develop characteristic brain tumors, Cell 37 (1984) 367–379. [14] C. Barlow, B. Meister, U. Lendahl, B. Vennstrom, Altered expression after expansion of a v-erbA transgene in transgenic mice, Transgenic Res. 4 (1995) 378–387. [15] J.A. Sharpe, D.J. Wells, E. Whitelaw, P. Vyas, D.R. Higgs, W.G. Wood, Analysis of the human ␣-globin cluster in transgenic mice, Proc. Natl. Acad. Sci. U.S.A. 90 (1993) 11262– 11266. [16] G. Robertson, D. Garrick, W. Wu, M. Kearns, D. Martin, E. Whitelaw, Position-dependent variegation of globin transgene expression in mice, Proc. Natl. Acad. Sci. U.S.A. 92 (1995) 5371–5375. [17] M. Kearns, J. Preis, M. McDonald, C. Morris, E. Whitelaw, Complex patterns of inheritance of an imprinted transgene suggest incomplete germline erasure, Nucl. Acids Res. 28 (2000) 3301–3309. [18] C. Morris, C. Courtay, A. Guerts van Kessal, N. Heisterkamp, J. Groffen, Localization of a gamma-glutamyltransferase-related gene family on chromosome 22, Human Genet. 91 (1993) 31–36. [19] M.N. Nesbitt, U. Francke, A system of nomenclature for band patterns of mouse chromosomes, Chromosoma 41 (1973) 145–158. [20] D. Garrick, S. Fiering, D.I.K. Martin, E. Whitelaw, Repeat-induced gene silencing in mammals, Natl. Genet. 18 (1998) 56–59. [21] F.F. Assaad, K.L. Tucker, E.R. Signer, Epigenetic repeatinduced gene silencing (RIGS) in Arabidopsis, Plant Mol. Biol. 22 (1993) 1067–1085.
[22] A.P. Jarman, W.G. Wood, J.A. Sharpe, G. Gourdon, H. Ayyub, D.R. Higgs, Characterization of the major regulatory element upstream of the human ␣-globin gene cluster, Mol. Cell. Biol. 11 (1991) 4679–4689. [23] F. Linn, I. Heidmann, H. Saedler, P. Meyer, Epigenetic changes in the expression of the maize A1 gene in Petunia hybrida: role of numbers of integrated gene copies and state of methylation, Mol. Gen. Genet. 222 (1990) 329–336. [24] D.R. Dorer, S. Henikoff, Expansions of transgene repeats cause heterochromatin formation and gene silencing in Drosophila, Cell 77 (1994) 1–20. [25] J.F. Sabl, S. Henikoff, Copy number and orientation determine the susceptibilty of a gene to silencing by nearby heterochromatin in Drosophila, Genetics 142 (1996) 447–458. [26] M. Mehtali, M. LeMeur, R. Lathe, The methylation-free status of a housekeeping transgene is lost at high copy number, Gene 91 (1990) 179–184. [27] G. Gundersen, A.-B. Kolsto, H. Prydz, Differential methylation of a CpG-island concatemer in hemi- and homozygous transgenic mice, FEBS Lett. 295 (1991) 214–218. [28] M. Young, J. Cullum, A plausible mechanism for large-scale chromosomal DNA amplification in streptomycetes, FEBS Lett. 212 (1987) 10–14. [29] G.R. Stark, M. Debatisse, E. Giulotto, G.M. Wahl, Recent progress in understanding mechanisms of mammalian DNA amplification, Cell 57 (1989) 901–908. [30] A. Paldi, G. Gyapay, J. Jami, Imprinted chromosomal regions of the human genome display sex-specific meiotic recombination frequencies, Curr. Biol. 5 (1995) 1030–1035. [31] M.A. Hulten, J.C. Hall, Proposed meiotic mechanism of genomic imprinting, Chromosomes Today 10 (1989) 157–162. [32] B.J. Thomas, R. Rothstein, Sex, maps and imprinting, Cell 64 (1991) 1–4. [33] W.P. Robinson, M. Lalande, Sex-specific meoitic recombination in the Prader–Willi/Angelman syndrome imprinted region, Hum. Mol. Genet. 4 (1995) 801–806. [34] M.E. Hodes, K. Woodward, N.B. Spinner, B.S. Emanuel, A. Enrico-Simon, J. Kamholz, D. Stambolian, E.H. Zackai, V.M. Pratt, I.T. Thomas, K. Crandall, S.R. Dlouhy, S. Malcolm, Additional copies of proteolipid protein gene causing Pelizaeus–Merzbacher disease arise by separate integration into the X chromosome, Am. J. Hum. Genet. 67 (2000) 14–22.
Spontaneous germline amplification and translocation of a transgene array Margot Kearns a , Christine Morris b , Emma Whitelaw a,∗ a
b
Department of Biochemistry, G08, University of Syndey, Sydney NSW 2006, Australia Cytogenetic and Molecular Oncology Unit, Department of Pathology, Christchurch School of Medicine, P.O. Box 4345, Christchurch, New Zealand Received 8 November 2000; received in revised form 20 March 2001; accepted 9 April 2001
Abstract The majority of the mammalian genome is thought to be relatively stable throughout and between generations. There are no developmentally programmed gene amplifications as seen in lower eukaryotes and prokaryotes, however a number of unscheduled gene amplifications have been documented. Apart from expansion of trinucleotide repeats and minisatellite DNA, which involve small DNA elements, other cases of gene or DNA amplifications in mammalian systems have been reported in tumor samples or permanent cell lines. The mechanisms underlying these amplifications remain unknown. Here, we report a spontaneous transgene amplification through the male germline which resulted in silencing of transgene expression. During routine screening one mouse, phenotypically negative for transgene expression, was found to have a transgene copy number much greater than that of the transgenic parent. Analysis of the transgene expansion revealed that the amplification in the new high copy transgenic line resulted in a copy number approximately 40–60 times the primary transgenic line copy number of 5–8 copies per haploid genome. Genetic breeding analysis suggested that this amplification was the result of insertion at only one integration site, that it was stable for at least two generations and that the site of insertion was different from the site at which the original 5–8 copy array had integrated. FISH analysis revealed that the new high copy array was on chromosome 7 F3/4 whereas the original low copy transgene array had been localised to chromosome 3E3. DNA methylation analysis revealed that the high copy transgene array was heavily methylated. The amplification of transgenes, although a rare event, may give insight into amplification of endogenous genes which can be associated with human disease. © 2001 Elsevier Science B.V. All rights reserved. Keywords: DNA amplification; Transgene silencing; Translocation; Germline mutation
1. Introduction The majority of the mammalian genome is thought to be relatively stable throughout and between generations. For survival, the number of genes must ∗ Corresponding author. Tel.: +61-2-9351-2549; fax: +61-2-9351-4726. E-mail address: [email protected] (E. Whitelaw).
remain essentially the same, however, non-essential non-coding regions of the genome do exhibit a certain plasticity. Gene amplification is a common feature in prokaryotes [1]. The amplification often results in increased protein production and is thought to be an adaptive mechanism in response to a changing environment [2]. Specific regions of the genome are hotspots for amplification [1] and the unit of DNA can be amplified up to 500 times [3]. Gene amplification
0921-8777/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 8 7 7 7 ( 0 1 ) 0 0 0 8 4 - 2
126
M. Kearns et al. / Mutation Research 486 (2001) 125–136
is not restricted to prokaryotes but has also been observed in yeast, Drosophila and vertebrates (reviewed in [2]). Developmentally regulated gene amplification of specific genes occurs in some organisms. Amplification of rRNA genes occurs during oogenesis in many species but has been studied mainly in Xenopus laevis (reviewed in [4]). Development-specific gene amplification is also seen in Drosophila where the chorion genes which encode the major protein constituent in eggshells are amplified for a limited time [5]. In yeast the rRNA genes are also in a state of flux with increases and losses within the tandemly repeated units [6,7]. Other examples of gene duplication in Saccaromyces cerevisiae involve translocations in conjunction with the amplification process. In one case in yeast a gene, which normally lies on chromosome 2, has been reported to undergo amplification and translocation to the telomere of another chromosome [8]. In mammals there are no known developmentally regulated amplification events, however, numerous examples of amplification in cell lines and tumours have been reported [2,9]. The treatment of cell lines with certain cytotoxic agents can result in selection of mutants which in many cases have developed resistance to the offending agent by gene amplification. The amplification results in increased production of a protein that can neutralise the effects of the drug [10]. Increases in gene copy number of 10- to 50-fold have been observed [2] and often these selection-driven amplifications recede once the selection is removed [11]. Gene amplification has been linked to tumorigenesis following the observation that many tumor cells were found to contain chromosomal abnormalities, double minutes (DMs) and homogeneously staining regions (HSR) associated with amplified DNA (reviewed in [9]). Although, transgene integration is usually stable, of the thousands of transgenic mice produced, a few have been reported to undergo heritable amplification [12,13]. There has been one report of the amplification of a murine transgene that involved the germline expansion of a 4–5 copy integrant to one containing 20–21 copies. The expanded locus continued to express the transgene, however, reduced levels of RNA were seen in some offspring [14]. Here we report the spontaneous germline amplification of a murine transgene array that results in loss
of transgene expression. During routine screening of a previously reported transgenic line, one individual with an increase in copy number was identified. In this mouse the transgene copy number was 200–300 copies per mouse genome, which indicated a 40–60-fold amplification of the original array. This spontaneous amplification was investigated and it was found that the expansion resulted in some rearrangement of the transgene array and a loss of transgene expression. DNA methylation analysis of the transgene array revealed that the expanded array was heavily methylated. Breeding studies revealed that the initial amplification event had probably taken place in the germline and also suggested a translocation event was associated with the amplification. Translocation of the high copy array to a new chromosomal location was confirmed by FISH analysis. This event highlights the inherent instability of the mammalian genome. Although, a rare event, amplification and translocation of transgenes can occur and study of these events may help us to understand the mechanisms underlying the amplification of endogenous genes. 2. Materials and methods 2.1. Transgene construct and the generation of transgenic mice The transgene construct, DM2, contains the human -globin promoter from −127 to +6 linked to an oligonucleotide (SDK) containing the Shine-Dalgarno and Kozak sequences, the lacZ reporter gene with the SV40 polyadenylation signal and a 4.1 kb fragment containing the human ␣-globin locus enhancer region (HS-40) previously shown to direct high levels of expression in erythroid cells [15]. Three transgenic lines were produced by microinjection of DNA into the pronucleus of fertilised mouse oocytes from the Pathology Oxford (P.O.) mouse strain [16]. Initial studies of these lines have been previously reported [16,17]. Of these lines 239B is the founding line of this study. The copy number of the line 239B was estimated to be 5–8 [16,17]. 2.2. Southern blotting and hybridisation Southern blotting of genomic DNA was carried out following digestion with EcoR V. The membrane
M. Kearns et al. / Mutation Research 486 (2001) 125–136
was hybridised with a 4.1 kb fragment containing the ␣HS-40 region, a 3.5 kb fragment containing the lacZ gene or a 1.4 kb fragment containing the mouse ␣-globin gene. The probes were labelled by random priming using [␣-32 P]dCTP (Bresatec) and the Megaprime Labelling Kit (Amersham). When screening for the presence of the human -globin promoter the MC 286 oligo (5 TAG TGC TCA GCC CCC AGG GTC CCC TTA TAT ACA GG 3 ) a kind gift from Dr. Merlin Crossley, Department of Biochemistry, University of Sydney) was endlabelled with [␥-32 P]ATP (Bresatec) and used as a probe. 2.3. Methylation analysis of the transgene array DNA was purified from circulating erythroid cells of 12.5 d.p.c. transgenic embryos. Aliquots of 2 g DNA were digested with BamH I alone or BamH I with either Msp I, Hpa II or Hha I. The resulting fragments were analysed by Southern transfer and hybridisation following separation on a 1.2% agarose gel. The membrane was probed with either a 4.1 kb fragment containing the ␣HS-40 region or a 1.4 kb mouse ␣-globin fragment. 2.4. β-Galactosidase activity in whole cells Transgene expression was analysed at the single cell level in erythroid cells from mice at 3 weeks after birth as described previously [16]. Blood from postnatal mice was collected by tail bleeding, and a dilute suspension of cells was fixed and washed for 5 min each and then stained with 5 bromo-4chloro-3-indolyl--d-galactopyranoside (X-gal) at 37◦ C for up to 24 h. A minimum of 1 × 105 erythrocytes were inspected for detectable blue stain; non-expressors are those mice having no stained cells in greater than 105 erythrocytes. For positively staining mice, at least 500 erythroid cells were counted to determine the percentage of stained cells. 2.5. Culturing of mouse embryonic fibroblasts Embryos resulting from a cross between a homozygous transgenic male and a wildtype P.O. female were dissected at 12.5 d.p.c. and bled. After removing the head and liver, the remainder of two embryos were
127
pooled then homogenised using a 19G needle and syringe in AmnioMax C100 media (Life Technologies). The emulsion was treated with collagenase (2000 U) for 4 h then incubated at 37◦ C in 5% carbon dioxide. About 24 h prior to harvesting, the confluent cultures were passaged. Mouse metaphase cells were prepared using standard colcemid and harvesting procedures. 2.6. Fluorescent in situ hybridisation The plasmid, DM2, containing the entire transgene construct, was random-primer labelled with biotin (High Prime kit, Boehringer Mannheim), and hybridised without repeat suppression to denatured transgenic mouse metaphase chromosomes spread on microscope slides using methods previously described [18]. Slides were incubated overnight at 37◦ C, and washed to a final stringency of 0.1 × SSC at 60◦ C before signal amplification with FITC-avidin and biotinylated anti-avidin (Vector Laboratories). Slides were mounted in glycerol-containing anti-fade and DAPI as counterstain for G-band visualisation before analysis under a Leitz Aristoplan microscope. Selected fields were transferred for image-processing to a Power MacIntosh 8100–80 computer via a Photometrics cooled CCD camera using appropriate wavelength filters, IPLab Spectrum and Multiprobe extension software. G-band nomenclature for description of mouse karyotypes and probe localisation was according to Nesbitt and Francke [19]. 2.7. PCR analysis of transgene–transgene and transgene–genome junctions Genomic DNA was analysed by PCR using a combination of the following primers: HS-40 1906- 5 GCG AAG CTG TTG TGA GGA C 3 , HS-40 2012-5 ACA TCT GTA TCA TTA ACA CTG AAT GAC AGC 3 , HS-40 4014- 5 GCA AGT CTG ACT GCA CAA TG 3 , HS-40 4108- 5 ACC CAC TGC ACT AGC TGG G 3 , 239B Genome 3 (1)- 5 ATC TTG CTC AGC TGA GCT GG 3 , SDK- 5 ATG GTG GCC TCC GAC CTG C 3 . A typical PCR reaction contained 1 l undigested genomic DNA, 5 l 10 X reaction buffer (Life Technologies), 5 l DMSO, 0.4 mM each primer, 0.2 mM each dNTP (dATP, dCTP, dGTP, dTTP), 3.0 mM MgCl2 , 1 U Taq (Life Technologies), and MQW to
128
M. Kearns et al. / Mutation Research 486 (2001) 125–136
a final volume of 50 l. The following PCR cycling conditions were used, 1 cycle of 94◦ C for 5 min followed by 30 cycles of 94◦ C for 30 s, 55◦ C for 30 s then 72◦ C for 60 s. The PCR was completed with one cycle of 72◦ C for 10 min. PCR products were visualised with UV light following gel electrophoresis. 3. Results 3.1. Spontaneous transgene amplification The transgenic line 239B [17] was generated using a construct which contained 127 bp of the human -globin promoter linked to the E. coli lacZ reporter gene and the human ␣-globin HS-40 enhancer. The transgenic line was estimated to have a copy number of 5–8 transgenes per genome integrated into one chromosomal location. To maintain stocks of this mouse line, hemizygous transgenic males were mated to wildtype females and transgenic offspring were identified by Southern transfer of tail DNA and hybridisation with a transgene-specific probe. During routine screening it was noticed that the transgene DNA of one pup appeared to be at a much higher copy number than its littermates. This individual was a female and appeared physically no different from her siblings. Tail sampling was repeated and the same result was obtained (Fig. 1). The membrane was hybridised with both a probe from the lacZ gene in the transgene and a probe from the ␣HS-40 fragment. In both cases the increase in the intensity of the signal was similar suggesting that the amplification most likely involved the whole transgene. The membrane was stripped and rehybridised with a mouse ␣-globin gene to correct for sample loading on the gel. In order to overcome any possible saturation of the signal which could occur at such high copy numbers, dilutions of the DNA from the high copy mouse were prepared (Fig. 1). Using these dilutions, the degree of amplification was determined to be approximately 40 times that of the transgenic parent. To determine whether the expansion had occurred prior to the formation of germ cells, the high copy female was mated to a wildtype male and the transgene copy number of her offspring was determined. All 24 transgenic offspring tested carried the transgene expansion (an example is G1 in Fig. 1). The initial DNA
Fig. 1. Quantitation of the DNA amplification and comparison of copy number in successive generations. (I) Genomic DNA extracted from tails samples collected from the transgenic parent of the high copy mouse (P) and the founder high copy mouse (F) were digested with EcoR V. Equal concentrations of samples P and F were electrophoresed along with dilutions (1:5, 1:10, 1:50 and 1:100) of the high copy digest as well as digested DNA collected from a sibling (S) of the high copy founder, and offspring from the next 2 generations of the high copy line (G1 and G2). The membrane was hybridised with either the radiolabelled 3.5 kb lacZ probe, 4.1 kb HS-40 probe or the 2.2 kb mouse ␣-globin probe.
amplification, therefore, must have occurred either in the germline of the 239B transgenic parent or early in the development of the founder high copy female prior to determination and migration of primordial germ cells. The new high copy mouse line was designated 239BHC. The high copy line was bred a further generation and the expansion was, once again, stably inherited (G2 in Fig. 1). 3.2. Rearrangement of the transgene in the high copy expansion The arrangement of copies in the transgene array of the parent line, 239B, and the high copy line, 239BHC, was investigated using restriction enzyme digestion, Southern transfer and hybridisation. As there is only one EcoR V recognition sequence in the transgene, digestion with this enzyme should produce one full length transgene–transgene band of 8.1 kb if the array is in a purely head to tail orientation. DNA from the parental line, 239B, produced an intense band migrating at approximately 8 kb and a much
M. Kearns et al. / Mutation Research 486 (2001) 125–136
129
Fig. 2. Determination of transgene array in the 239BHC line. (I) Schematic representation of a transgene array depicting the EcoR V restriction sites. The probe is shown as a black bar. With a head to tail array a transgene–transgene band of 8.1 kb is generated; with head to head transgenes a 2.6 kb; band and with tail to tail transgenes a 13.6 kb band should be seen. (II) Genomic DNA from the transgenic parent of the high copy founder (P), the high copy founder mouse (F) and one of the offspring in the next generation (G1) was digested with EcoR V which cuts the 127/lacZ construct only once. Following electrophoresis and Southern blotting the membrane was hybridised with a radiolabelled 3.5 kb lacZ probe. The membrane was stripped and rehybridised with a probe to the mouse ␣-globin gene to check for equal loading.
fainter band of higher molecular weight that migrated more slowly (Fig. 2). This may represent the edge fragment: transgene DNA and the adjoining genomic DNA. The predominant band obtained for the high copy line, 239BHC, was also approximately 8 kb representing the head to tail arrangement. The presence of this band suggested that there was no substantial loss of sequence in the majority of the transgene copies. However, the banding pattern obtained for the 239BHC line was complex: the presence of multiple smaller and larger bands suggested that large deletions had occurred in a subpopulation of copies (Fig. 2). 3.3. The transgene amplification resulted in loss of transgene expression Previously it had been demonstrated that male transmission of the 239B transgene invariably resulted
in transgene expression in erythrocytes [17]. The number of erythrocytes expressing the transgene was affected by parent of origin. The percentage of erythroid cells expressing the transgene at 3 weeks of age in mice resulting from male transmission was 7.5%. Transgene expression was analysed in the founder 239BHC mouse and following passage through both germlines at 3 weeks of age. In all cases no transgene expression was detected. A total of 61 mice were analysed; 36 following male transmission and 25 following female transmission. 3.4. Transgene silencing correlates with hypermethylation As methylation has been shown to be associated with the lack of expression in repeat induced gene silencing in mice [20] and other organisms [21], the methylation state of the high copy array was
130
M. Kearns et al. / Mutation Research 486 (2001) 125–136
Fig. 3. Methylation state of the transgene in the line 239BHC. DNA extracted from the erythrocytes of 12.5 d.p.c. embryos (following male transmission) was digested with BamH I alone or in combination with Msp I (M) or the methylation sensitive restriction enzymes Hpa II (H) or Hha I (Hh). Following electrophoresis the membrane was hybridised with a radiolabelled 4.1 kb HS-40 probe, then stripped and rehybridised with 2.2 kb mouse ␣-globin probe. The DNA from the high copy line 239BHC is more heavily methylated than that from the low copy line 239B.
investigated. DNA was purified from erythrocytes collected from 12.5 d.p.c. embryos resulting from the cross between a transgenic male and a wildtype female. DNA was digested with BamH I or BamH I plus either Msp I, Hpa II or Hha I. Southern blotting and hybridisation was carried out as described previously and the methylation state of selected sites within the HS-40 region of the transgene was determined (Fig. 3). The presence of many larger bands in the Hpa II lane with the 239BHC DNA suggests that the transgene was more heavily methylated in the high copy embryos compared with the 239B controls. The transgene array however was not methylated in all cells as the 1.6 and 1.0 kb Msp I/Hpa II bands were still present. It is interesting to note that the BamH I
only digest produced a 4.1 kb band in the 239B line but not with the 239BHC line (Fig. 3). Instead a higher molecular weight band was present suggesting that one of the BamH I restriction sites was absent and digestion at the remaining site has generated a near full length transgene–transgene band (approximately 8 kb) (Fig. 3). 3.5. Loss of transgene sequence in the amplification process The loss of transgene expression seen with the amplification in the 239BHC sub-line could also result from the loss of important regulatory sequences. In this case any loss of sequence could not be extensive
M. Kearns et al. / Mutation Research 486 (2001) 125–136
131
Fig. 4. Loss of sequence at the ends of the transgene construct. (I) PCR analysis of transgene–transgene junctions. Schematic representation of a 239B head to tail transgene–transgene junction showing approximately position of primers used in various PCR reactions with outcome indicated. PCR was performed on genomic tail DNA collected from 239B (L) 239BHC (H) and compound hemizygotes 239B/239BHC mice (L/H). M represents a 1:1 mix of DNA from 239B and 239BHC samples and W represents wildtype DNA. (II) 239BHC DNA hybridises to a radiolabelled oligonucleotide containing human -globin promoter sequence. Southern blotting of genomic tail DNA, digested with Pst I, was followed by hybridisation using a radioactive endlabelled oligonucleotide MC 286 which contains human -globin promoter sequence from −35 to −2. Samples included the high copy line 239BHC and wildtype DNA.
since the EcoR V digests give a head to tail band approximately equivalent to the full length transgene (Fig. 2). As some degree of sequence loss was evident with BamH I digest described above, however, the situation was investigated further by PCR analysis. A number of different primers (three from 3 region of the transgene and two from the 5 region of the transgene) which should amplify the transgene–transgene junctions (Fig. 4(I)), failed to produce any product with the 239BHC line (data not shown). Using the SDK primer and two primers that anneal to sequences near the centre of the HS-40 element (HS-40 1906 and HS-40 2012) a product was obtained. The expected size of a 239B transgene–transgene band is approximately 2.5 kb. A band of this size was obtained for the 239B samples but a band of approximately 2.0 kb was obtained with the 239BHC samples (Fig. 4(I)). This would indicate that approximately 500 bp of
sequence at the transgene junction was deleted in the amplification of the process. As it was possible that the loss of sequence included the promoter sequence, Southern blotting was performed using a probe to the human -globin promoter, oligo MC 286 (from −35 to −2 of the human -globin promoter). This probe successfully hybridised to 239BHC DNA (Fig. 4(II)). Therefore, at least some of the promoter sequence, approximately 30 bp, was present in the high copy transgenic line suggesting that the absence of PCR product with the HS-40 4014 and HS-40 4108 primers and the -promoter and SDK primer pairs, was due to the loss of sequence at the 3 end of the HS-40 fragment. This is also consistent with the loss of the BamH I site which lies at the very end of the transgene construct. The precise -globin promoter sequence present in the 239BHC line is unknown.
132
M. Kearns et al. / Mutation Research 486 (2001) 125–136
The deletion has involved some sequence at the 3 end of the HS-40 fragment. The activity of the HS-40 element has been localised to a 350 bp core region containing multiple transcription factor binding sites, situated between positions 2991 and 3335 of the 4.1 kb element [22]. The deletion in the 239BHC array does not include this region. 3.6. The 239BHC transgene array is situated at a different chromosomal location To investigate whether the amplified DNA was at the same chromosomal location as the 239B site or at a new chromosomal location due to a translocation a number of approaches were employed: PCR analysis of the trangene–genome junction using the 3 integration site from 239B; systematic breeding of mice carrying both the 239B allele and the 239BHC allele; and FISH. As the 3 genomic flanking sequence at the site of 239B integration had been determined [17], the presence of this transgene–genome junction in the high copy line would suggest that transgenic DNA remained at this locus. Using a primer to HS-40 and a primer to the 3 flanking DNA, a PCR product was only obtained for the low copy controls and DNA samples which contained both high and low copy DNA (Fig. 5). This suggested that there was no resid-
ual transgene DNA at the previous site of integration on chromosome 3 band E3. However, this analysis does not rule out the possibility that the junction was lost in the amplification process. Genetic analysis was carried out to see if the 239B allele and the 239BHC allele were on separate chromosomes. This was achieved by breeding a homozygous 239B male mouse with a hemizygous 239BHC female (Fig. 6). All the offspring carried the low copy allele which conferred transgene expression. After screening for the presence of the high copy allele by Southern blotting and hybridisation, a positive high copy mouse was selected for further breeding. This compound hemizygous mouse now carried a low copy allele and a high copy allele (239B/239BHC). By breeding this male mouse to a wildtype female the genetics of segregation of alleles could be analysed. If the amplification was at the same chromosomal location or so close on the same chromosome that recombination did not occur, then segregation of the alleles would result in approximately 50% of the offspring carrying the low copy allele which would confer transgene expression and the remaining offspring would carry the silenced high copy allele. If the amplification was at a new chromosomal location which segregated independently from the low copy allele then 4 different groups of mice would be obtained in equal proportions: 239B/- (low copy hemizygotes);
Fig. 5. PCR analysis of transgene–genome junction. PCR was performed on genomic tail DNA using the HS-40 4014 primer and a primer to the 3 genomic flanking DNA at the 239B site of integration. PCR was performed on genomic tail DNA collected from 239B (L) 239BHC (H) and compound hemizygotes 239B/239BHC mice (L/H). M represents a 1:1 mix of DNA from 239B and 239BHC samples and W represents wildtype DNA. PCR product was only obtained in those samples containing the 239B allele and not in 239BHC hemizygote samples KSH and MK are markers, Bluescript KS cut with Hpa II and 100 bp ladder, respectively.
M. Kearns et al. / Mutation Research 486 (2001) 125–136
133
Fig. 6. Genetic analysis of segregation of two independent transgene arrays, 239B and 239BHC. Mating a homozygous low copy (239B) male to a high copy (239BHC) female produces mice carrying the low copy allele as well as the high copy allele. Breeding these compound hemizygous mice to wildtype mice will segregate the alleles. If the alleles are on different chromosomes that randomly assort during meiosis then four genotypes are possible in the next generation: low copy, high copy, compound heterozygous and wildtype. Phenotypes were determined by staining erythrocytes with X-gal and detecting the presence of blue stained cells under bright field microscopy. The presence of blue staining cells indicated that the mouse carried the 239B allele. 239BHC genotypes were determined by screening genomic DNA for the presence of the transgene by Southern transfer and hybridisation with a transgene-specific probe. A total of 101 mice were analysed. The 25 offspring that displayed transgene expression in the presence of the high copy array is indicative of reconstitution of the parental genotype — the compound hemizygote (239B/239BHC). Statistical analysis using χ 2 based on the assumption of segregation of independent chromosomes giving proportions of 1:1:1:1 gave a χ 2 value of 0.82: with 3 degrees of freedom, 0.9 > P > 0.7.
239B/239BHC (compound hemizygotes); 239BHC/(high copy hemizygotes); and non-transgenic. The 101 individual mice analysed did indeed fall into 4 groups supporting the hypothesis that the two transgene arrays were at different chromosomal locations. The genetic studies described above indicated that the 239BHC transgene array was at a different chromosomal location to the 239B integrant which had previously been shown to have localised to mouse chromosome 3 region E3 (Fig. 7) [17]. In order to localise the high copy array, primary fibroblasts were cultured from embryonic tissue collected from 239BHC transgenic embryos and FISH analysis was performed. A clear signal that hybridised to the proximal region of chromosome 7 band F3/4 was observed in all metaphase preparations examined. This band is situated near the telomere (Fig. 7) and as telomeric regions of chromosomes typically contain heterochromatin, this also could be a contributing factor to the transgene silencing observed with the 239BHC amplification.
4. Discussion Germline transformation of mice is generally thought to produce a relatively stable transgene array that can be inherited for many generations. This study identified a germline amplification of a transgene array from 5 to 8 copies to 200–300 copies. The amplification appears to have occurred in the germline. It is possible that the original event happened very early in embryogenesis of the founder high copy mouse before the primordial germ cells were determined. If, however, this had been the case then the founder would be mosaic with some cells being low copy and some cells high copy. Using tail DNA there was no evidence of the previous transgene–genome junction or head to tail transgene–transgene junctions in the founder. It is therefore likely that the amplification occurred during gametogenesis in the male germline of the low copy parent. Most studies of gene amplification in mammalian cells have used cultured cells or tumor tissue [9]. In
134
M. Kearns et al. / Mutation Research 486 (2001) 125–136
Fig. 7. FISH localising the chromosomal positions of the 239B and the 239BHC transgene arrays. FISH was performed using the DM2 plasmid containing the 8.1 kb 127/lacZ construct used for microinjection to produce the transgenic line 239B. (I) Localisation of the 239B transgene integrant to mouse chromosome 3 band E3: (A) a representative metaphase cell and chromosome under fluorescence showing the single signal on chromosome 3. The fluorescent signal is indicated by the orange arrow; (B) chromosome 3 from panel A aligned with ideogram showing precise location of the probe signal in band E3. (II) Localisation of the 239BHC transgene integrant to mouse chromosome 7 F3/4: (C) a representative metaphase cell and chromosome under fluorescence showing the single signal on chromosome 7; (D) chromosome 7 from panel C aligned with G-banding depicting that the signal has hybridised to region F3/4.
these cases the amplifications in somatic cells often occur under selection pressure and result in increased protein production [9]. The lack of reports of germline amplification of genes may indicate that although these events occur they are not often observed due to embryonic lethality. The germline amplification of a transgene carrying a reporter gene, however, may not be hazardous to the developing embryo enabling these events to be detected. The 239BHC array was found to be hypermethylated. Hypermethylation of silent high copy transgene arrays have been observed in plants [21,23] and mice [20]. However, as repeat induced gene silencing is
also found in Drosophila [24,25] where there is no methylation it has been suggested that the relationship between gene silencing and methylation is not causal but consequential. Other studies on high copy transgene arrays report that even constituitively expressed house keeping genes are hypermethylated when in multi copy arrays [26]. It has been suggested that the increase in methylation with increased copy number could be the result of depletion of a limited pool of proteins that bind to DNA protecting it from methylation [26,27]. Why the expansion of transgene arrays [20] results in silencing, when amplification of oncogenes in
M. Kearns et al. / Mutation Research 486 (2001) 125–136
tumours and amplification in mammalian cell lines results in overexpression of the protein product, is not clear. It could be argued that the size of the expanded array or the number of repeats could influence silencing but it is noteable that the v-erbA transgene expansion of up to 20–21 copies did not result in silencing and oncogenes can be amplified up to 500 copies and still maintain expression. 4.1. Possible mechanisms involved in the amplification process A number of models of DNA amplification have been proposed based on extensive study of gene amplification in prokaryotes. Large scale amplifications, like that seen with 239BHC, are not thought to be due to unequal crossover or sister chromatid exchange as these events produce a maximum two-fold expansion per generation [2]. PCR analysis detected a transgene deletion of about 500 bp with no evidence of any copies containing the original transgene–transgene junctions found in the 239B low copy line. This suggests that the amplification involved one incomplete transgene unit of approximately 7.6 kb. This unit was amplified up to 300 times with the majority of the copies being in a head to tail array. The process also involved a translocation event. So a possible model would involve some unequal recombination event between two adjacent transgenes which produced a circle that contained a 7.6 kb portion of transgene sequence and trapped a replication fork. The rolling circle model would account for the high number of copies being in a tandem repeat array [28]. The newly formed concatamer then translocated to a telomeric region as seen in yeast with the alkaline phosphatase gene [8]. As this event is thought to have occurred during meiosis, the original transgene integrant and the new high copy translocated array could segregate into different spermatids. Many of the proposed models of amplification link the process to replication and recombination [29]. Recombination is a normal event in meiosis where it occurs after the chromosomes have duplicated and involves the cellular repair enzymes. In general, recombination occurs more frequently in female meiosis, however, in certain regions of the genome recombination occurs predominantly during male meiosis. It has been suggested that these male–female
135
differences in recombinational activity during meiosis may be related to differences in gene activity in imprinted regions [30–32]. One such differential region exists on human chromosome 15q11 which is associated with Prader–Willi and Angelman syndromes, where a region with a very high male recombination frequency is adjacent to a region with a high female recombination frequency [33]. Interestingly the amplification event in 239B which may have involved recombination within the transgene has also occurred at a locus that exhibits parental imprinting [17]. Although, a rare event, amplification and translocation of transgenes can occur but whether these events are more common at sites which have been introduced into the genome, than at endogenous loci, is unclear. Amplification of endogenous loci may go undetected due to lethality of chromosomal imbalances or gene silencing. Recently a number of cases of Pelzaeus– Merzbacher disease (PMD) have been attributed to this type of genetic instability [34]. The proteolipid protein gene (PLP) is normally present at chromosome Xq22. Mutations and duplications of this gene are associated with PMD. Hodes et al. [34] report one case in which an additional copy of the gene is found at a novel site on the X chromosome and no gross chromosomal rearrangements could be detected by standard karyotype analysis.
References [1] D. Romero, R. Palacios, Gene amplification and genome plasticity in porkaryotes, Ann. Rev. Genet. 31 (1997) 91–111. [2] G.R. Stark, G.M. Wahl, Gene amplification, Ann. Rev. Biochem. 53 (1984) 447–491. [3] P. Leblond, P. Demuyter, L. Moutier, M. Laakel, B. Decaris, J.M. Simonat, Hypervariability: a new phenomenon of genetic instability related to DNA amplification in Streptomyces ambofaciens, J. Bacteriol. 71 (1994) 419–423. [4] K.D. Tartof, Redundant genes, Ann. Rev. Genet. 9 (1975) 355–385. [5] A.C. Spradling, The organisation and amplification of two chromosomal domains containing Drosophila follicle cells, Cell 27 (1981) 193–201. [6] H.L. Klein, T.D. Petes, Genetic mapping of Ty elements in Saccharomyces cerevisiae, Nature 289 (1981) 144–148. [7] T.D. Petes, Unequal meiotic recombination within tandem arrays of yeast ribosomal DNA genes, Cell 26 (1980) 765–774. [8] P.E. Hansche, V. Beres, P. Lange, Gene duplication in Saccharomyces cerevisiae, Genetics 88 (1978) 673–687.
136
M. Kearns et al. / Mutation Research 486 (2001) 125–136
[9] M. Schwab, Amplification of oncogenes in human cancer cells, BioEssays 20 (1998) 473–479. [10] C. Ramel, H. Cederberg, J. Magnusson, E. Vogel, A.T. Natarajan, L.H. Mullender, J.M. Nivard, J.M. Parry, A. Leysen, M.A. Comendador, L.M. Sierra, J.A. Ferreiro, J.A.S. Consuegra, Somatic recombination, gene amplification and cancer, Mutat. Res. 353 (1996) 85–107. [11] D.A. Haber, R.T. Schmidt, Unstable amplification of an altered dihydrofolate reductase gene associated with doubleminute chromosomes, Cell 26 (1981) 355–362. [12] M. Lacey, S. Alpert, D. Hanahan, Bovine papillomavirus genome elicits skin tumours in transgenic mice, Nature 322 (1986) 609–612. [13] R.L. Brinster, H.Y. Chen, A. Messing, T. van Dyke, A.J. Levine, R.D. Palmiter, Transgenic mice harboring SV40 T-antigen genes develop characteristic brain tumors, Cell 37 (1984) 367–379. [14] C. Barlow, B. Meister, U. Lendahl, B. Vennstrom, Altered expression after expansion of a v-erbA transgene in transgenic mice, Transgenic Res. 4 (1995) 378–387. [15] J.A. Sharpe, D.J. Wells, E. Whitelaw, P. Vyas, D.R. Higgs, W.G. Wood, Analysis of the human ␣-globin cluster in transgenic mice, Proc. Natl. Acad. Sci. U.S.A. 90 (1993) 11262– 11266. [16] G. Robertson, D. Garrick, W. Wu, M. Kearns, D. Martin, E. Whitelaw, Position-dependent variegation of globin transgene expression in mice, Proc. Natl. Acad. Sci. U.S.A. 92 (1995) 5371–5375. [17] M. Kearns, J. Preis, M. McDonald, C. Morris, E. Whitelaw, Complex patterns of inheritance of an imprinted transgene suggest incomplete germline erasure, Nucl. Acids Res. 28 (2000) 3301–3309. [18] C. Morris, C. Courtay, A. Guerts van Kessal, N. Heisterkamp, J. Groffen, Localization of a gamma-glutamyltransferase-related gene family on chromosome 22, Human Genet. 91 (1993) 31–36. [19] M.N. Nesbitt, U. Francke, A system of nomenclature for band patterns of mouse chromosomes, Chromosoma 41 (1973) 145–158. [20] D. Garrick, S. Fiering, D.I.K. Martin, E. Whitelaw, Repeat-induced gene silencing in mammals, Natl. Genet. 18 (1998) 56–59. [21] F.F. Assaad, K.L. Tucker, E.R. Signer, Epigenetic repeatinduced gene silencing (RIGS) in Arabidopsis, Plant Mol. Biol. 22 (1993) 1067–1085.
[22] A.P. Jarman, W.G. Wood, J.A. Sharpe, G. Gourdon, H. Ayyub, D.R. Higgs, Characterization of the major regulatory element upstream of the human ␣-globin gene cluster, Mol. Cell. Biol. 11 (1991) 4679–4689. [23] F. Linn, I. Heidmann, H. Saedler, P. Meyer, Epigenetic changes in the expression of the maize A1 gene in Petunia hybrida: role of numbers of integrated gene copies and state of methylation, Mol. Gen. Genet. 222 (1990) 329–336. [24] D.R. Dorer, S. Henikoff, Expansions of transgene repeats cause heterochromatin formation and gene silencing in Drosophila, Cell 77 (1994) 1–20. [25] J.F. Sabl, S. Henikoff, Copy number and orientation determine the susceptibilty of a gene to silencing by nearby heterochromatin in Drosophila, Genetics 142 (1996) 447–458. [26] M. Mehtali, M. LeMeur, R. Lathe, The methylation-free status of a housekeeping transgene is lost at high copy number, Gene 91 (1990) 179–184. [27] G. Gundersen, A.-B. Kolsto, H. Prydz, Differential methylation of a CpG-island concatemer in hemi- and homozygous transgenic mice, FEBS Lett. 295 (1991) 214–218. [28] M. Young, J. Cullum, A plausible mechanism for large-scale chromosomal DNA amplification in streptomycetes, FEBS Lett. 212 (1987) 10–14. [29] G.R. Stark, M. Debatisse, E. Giulotto, G.M. Wahl, Recent progress in understanding mechanisms of mammalian DNA amplification, Cell 57 (1989) 901–908. [30] A. Paldi, G. Gyapay, J. Jami, Imprinted chromosomal regions of the human genome display sex-specific meiotic recombination frequencies, Curr. Biol. 5 (1995) 1030–1035. [31] M.A. Hulten, J.C. Hall, Proposed meiotic mechanism of genomic imprinting, Chromosomes Today 10 (1989) 157–162. [32] B.J. Thomas, R. Rothstein, Sex, maps and imprinting, Cell 64 (1991) 1–4. [33] W.P. Robinson, M. Lalande, Sex-specific meoitic recombination in the Prader–Willi/Angelman syndrome imprinted region, Hum. Mol. Genet. 4 (1995) 801–806. [34] M.E. Hodes, K. Woodward, N.B. Spinner, B.S. Emanuel, A. Enrico-Simon, J. Kamholz, D. Stambolian, E.H. Zackai, V.M. Pratt, I.T. Thomas, K. Crandall, S.R. Dlouhy, S. Malcolm, Additional copies of proteolipid protein gene causing Pelizaeus–Merzbacher disease arise by separate integration into the X chromosome, Am. J. Hum. Genet. 67 (2000) 14–22.