microcell human-mouse hybrids - Europe PMC

The EMBO Journal Vol.2 No.9 pp. 1577 - 1584, 1983

Integration of Ecogpt and SV40 early region sequences into human chromosome 17: a dominant selection system in whole cell and microcell human-mouse hybrids A. Tunnacliffe1, M. Parkar3, S. Povey3, B.O. Bengtsson4, K. Stanley2, E. Solomon2 and P. Goodfellowl* 'Laboratory of Human Molecular Genetics, and 2Laboratory of Somatic Cell Genetics, Imperial Cancer Research Fund, 44 Lincoln's Inn Fields, London WC2A 3PX, 3MRC Human Biochemical Genetics Unit, Wolfson House, 4 Stephenson Way, London NWI 2HE, UK, and 41nstitute of Genetics, Solvegatan 29, S-22362 Lund, Sweden Communicated by W.F. Bodmer Received on 9 June 1983

The dominant selectable gene, Ecogpt, has been introduced, by the calcium phosphate precipitation technique, into normal human fibroblasts, along with the SV40 early region genes. In one transfectant clone, integration of these sequences into human chromosome 17 was demonstrated by the construction of human-mouse somatic cell hybrids, selected for by growth in medium containing mycophenolic acid and xanthine. A whole cell hybrid, made between the human transfectant and a mouse L cell, was used as donor of the Ecogpt-carrying human chromosome 17 to 'tribrids' growing in suspension, made by whole cell fusion between a mouse thymoma cell line, and to microcell hybrids made with a mouse teratocarcinoma cell line. Two tribrids contained karyotypically normal human chromosomes 17 and a small number of other human chromosomes, while a third tribrid had a portion of the long arm of chromosome 17 translocated to mouse as its only human genetic material. Two independent microcell hybrids contained a normal chromosome 17 and no other human chromosome on a mouse teratocarcinoma background. These experiments demonstrate the ability to construct human-mouse somatic cell hybrids using a dominant selecdon system. By applying this approach it should be possible to select for a wide range of different human chromosomes in whole cell and microcell hybrids. In particular, transfer of single human chromsomes to mouse teratocarcinoma cells will allow examination of developmentally regulated human gene sequences after differentiation of such hybrids. Key words: Ecogpt/hybrid selection/microcell fusion/human chromosome 17 Introduction Recently, a number of selection systems have become available for the stable introduction of cloned genes into mammalian cells. Several of these are 'dominant' selection systems in that they do not rely on the complementation of a geneteic defect in the recipient cell. For example, the bacterial gene Ecogpt, whose expression is controlled by SV40 sequences in recombinant plasmids (Mulligan and Berg, 1980), encodes an enzyme activity not found in eukaryotic cells which can be selected for in appropriate growth media (Mulligan and Berg, 1981). Other genes can be co-transferred into cells along with a selectable marker, such that, using present gene transfer techniques, it is possible to stably introduce cloned genes into a large number of different cell types. *To whom reprint requests should be sent. © IRL Press Limited, Oxford, England.

In contrast, systems for the selection of chromosomes in somatic cell hybrids are limited in that only a small number of chromosomes can be selected for in cell fusions. Thus, in human-mouse hybrids, human chromosomes X or 17 are selected for in HAT medium (Szybalski et al., 1962; Littlefield, 1964) and human chromosome 16 in AA medium (Kusano et al., 1971). This is dependent on the appropriate mouse fusion partner being available, which must be defective in hypoxanthine phosphoribosyl transferase (HPRT), thymidine kinase (TK) or adenine phosphoribosyl transferase (APRT), respectively. The presence of other human chromosomes in human-rodent hybrids is largely fortuitous since they are non-selectable. We have been interested in extending the selection procedures available and, in particular, would like to develop a dominant selection system for each human chromosome in human-rodent hybrids. There are several reasons for pursuing this. (i) It would cancel the need for mutants as rodent hybrid parents and would allow the construction of a panel of somatic cell hybrids containing single human chromosomes. If 24 such hybrids were available, one for each human chromosome, human gene mapping would be greatly facilitated. (ii) Particular human chromosomes could be introduced into mouse cells of different tissue origins, which would allow investigation of the regulation by differentiated mouse functions of human gene expression, where those genes are in a native chromosomal environment. The advantage over conventional hybrids would be that the selected chromosome is necessarily present in all hybrid cells. (iii) By transferring human chromosomes to mouse teratocarcinoma cells and then causing these hybrids to differentiate in vivo or in vitro, developmentally regulated human sequences may be identified. We have taken an approach to this problem which involves the integration of the dominant selectable gene Ecogpt into human chromosomes, which can then be selected for in human-mouse hybrids. We show here that a human chromosome 17 containing Ecogpt and co-transferred SV40 early region genes can be transferred to three different mouse cell types, both by whole cell and microcell fusion (Foumier and Ruddle, 1977).

Results Transfection of human fibroblasts with pSV3-gpt A normal diploid fibroblast, HFL121, from a human male, was transfected by the calcium phosphate precipitation technique (Graham and van der Eb, 1973; Wigler et al., 1977) with the plasmid pSV3-gpt (Mulligan and Berg, 1980), which contains the dominant selectable gene Ecogpt and a complete SV40 early region. By selecting for transfectants according to Mulligan and Berg (1981) in medium containing HAT, mycophenolic acid (MPA) and xanthine (BlO medium) and by picking colonies with a transformed phenotype, we were able to isolate cells expressing both xanthine phosphoribosyl transferase (XPRT; encoded by Ecogpt) and SV40 large T antigen. Expression of T ensures partial transformation of human fibroblasts (reviewed in Sack, 1981; Tooze, 1982) and enables large numbers of cells to be grown up for analysis and subse1577

A. Tunnadiffe et al.

quent fusions. It will also prove useful as a model system for a co-transferred gene in our subsequent hybrids (details of large T expression will be presented elsewhere). 20 Ag of plasmid, without carrier DNA, were used per 106 fibroblasts transfected, of which -1 in 105 gave foci in selective medium after 2 - 3 weeks. Fifteen independent clones were expanded into mass culture and one of these, Egpt2, is described here. Whole cell fusion of Egpt2 with mouse L cell If the transfectant Egpt2 contains copies of pSV3-gpt integrated into human chromosomal DNA, it should be possible to select for the presence of such chromosomes in hybrids made with mouse cells. Accordingly, we fused Egpt2 with the HPRT - mouse L cell derivative IR (Nabholz et al., 1969) using polyethylene glycol (PEG; Pontecorvo, 1975) as described in Materials and methods. Selection against the human parent 10 was with lOM ouabain (Kucherlapati et al., 1975) and against IR either by the requirement for XPRT activity (BlO medium) or by HAT alone. The initial reason for the use of two different selection media was to compare the efficiencies of hybrid formation in B1O plus ouabain and in HAT plus ouabain (where selection is for the human X chromosome in hybrids). It was found that, whilst hybrids appeared with normal efficiencies (10-5- 10-4) in HAT, no colonies were seen in BlO medium. The reasons for this are unclear, but might be due to the need to select for both XPRT and human HPRT simultaneously (since IR is HPRT -). Subsequent fusions with HPRT - cells were selected in medium with MPA and xanthine, but no HAT, which gave reasonable yields of hybrids. We were able to rescue hybrids expressing XPRT from the Egpt2 x IR fusion by changing the medium of HAT-selected hybrids to B1O after colonies had appeared. The majority of hybrids subsequently died, suggesting that if integrated Ecogpt sequences are present in Egpt2, they are not in the X chromosome, which should be present in all HAT-selected hybrids. However, some colonies survived and one such hybrid, 21R, was grown into mass culture and was maintained in BlO medium without methotrexate (B10' medium). Its hybrid nature was confirmed by the presence of H-2K and mouse isozymes (see below). Presence of XPRTand pS V3-gpt sequences in Egpt2 and 21R Growth of both Egpt2 and 21R in medium containing MPA and xanthine implies the presence of XPRT in these cells. This was tested qualitatively by starch gel electrophoresis of cell extracts followed by staining for HPRT activity. Since the bacterial enzyme XPRT possesses some HPRT activity (as well as XPRT activity; Miller et al., 1972), this gel system allows simultaneous inspection of both bacterial and mammalian enzyme activities. Figure 1 shows an HPRT gel where track a represents HFL121 (normal human fibroblast), track b Egpt2, track c 21R and track d IR. The lower series of bands represents human HPRT, the more anodal band being bacterial XPRT. It can be seen that XPRT is present in Egpt2 and the hybrid 21R, but not in HFL121 or IR, as expected. Human HPRT is present in both human cells and also in 21R, which suggests that the human X chromosome is in this hybrid. Mouse 1R is HPRT - and therefore gives no bands on this gel. DNA sequences complementary to pSV3-gpt were detected in transfectant and hybrid cells by Southern blotting and filter hybridisation (Southern, 1975). Figure 2a shows four tracks of high mol. wt. Egpt2 DNA digested with EcoRI, BamHI, PvuII and HindIII, respectively. Figure 2, panel b shows 21R

1578

Flg. 1. Analysis of HPRT activity in human, mouse and hybrid cells. Starch gel electrophoresis of cell extracts and HPRT staining was performed according to standard techniques. Tracks are as follows: (a) HFL121; (b) Egpt2; (c) 21R; (d) IR; (e) TRI C4; (f) TRI D60; (g) TRI D62; (h) BW5147.

Fig. 2. Sequences in Egpt2 and 21R hybridising to pSV3-gpt. Panel a is a filter hybridisation analysis of Egpt2 DNA cut with EcoRI, BamHl, PvuII and Hindlll, respectively. Panel b is 21R DNA and the first four tracks of panel c are HFL121 DNA plus approximately single copy amounts of pSV3-gpt DNA, cut with the same enzymes as Egpt2. The last two tracks in panel c are EcoRI digests of HFL121 and IR DNAs, respectively. Each track represents 20 Ag of DNA electrophoresed through a 1% agarose gel, transferred to nitrocellulose and hybridised to labelled pSV3-gpt. Numbers indicate positions of Hindlll-digested X marker fragments (kilobases).

DNA and panel c HFL 121 DNA mixed with roughiy single copy equivalent amounts of pSV3-gpt DNA, cut with the same enzymes in the same order. The last two tracks in panel

Dominant selection for human chromosome 17

c are controls, being EcoRI digestions of HFL 121 and IR DNAs. Nick-translated, 32P-labelled pSV3-gpt was used as a probe. There are several results of interest from this experiment. Firstly, as predicted, sequences complementary to pSV3-gpt are present in high mol. wt. DNA of both Egpt2 and 21R. Secondly, all the bands seen in 21R are present in Egpt2, indicating the transfer of this material from human to hybrid cell. However, there are also extra bands in Egpt2 not found in the hybrid. The fact that there are two EcoRI bands larger than pSV3-gpt in Egpt2, of which only one is present in 21R, suggests that integration of pSV3-gpt into at least two different human chromosomes may have occurred, only one of which has been transferred to the hybrid. Some bands are probably from free, unintegrated copies of plasmid-derived sequences reminiscent of those found in some human cells transformed by SV40 virus. We have noticed that different clones from the HFL121 transfection possess markedly variable amounts of unintegrated plasmid-like DNA (unpublished results). Egpt2 has of the order of one free copy per cell, on average, and this is represented by the lower hybridising band in the EcoRI digest of Figure 2a. The third point of interest is that the pattern of fragments of pSV3-gpt DNA in Egpt2 and 21 R is not consistent with a simple integration event and it is likely that deletion and rearrangement has occurred. Also, since the relative signal strengths of different bands vary, amplification of some fragments is implicated. However, the restriction pattern is consistent with the regions encoding both XPRT and SV40 large T antigen being intact. For example, the 3-kb BamHI fragment in all three panels of Figure 2 represents the complete SV40 early region. Chromosomal marker analysis of 2JR The presence of human chromosomes in the hybrid 21R was assayed by analysis for both marker human enzymes and cell surface antigens. Isoenzyme markers exist for each of the

human chromosomes except Y (McKusick and Ruddle, 1977) and we have also used monoclonal antibodies to eight chromosome-specific human antigenic determinants (reviewed in Goodfellow and Solomon, 1982; Tunnacliffe et al., 1983; Tunnacliffe and Goodfellow, 1983). Enzymes were assayed by a variety of gel electrophoresis methods (Harris and Hopkinson, 1976) and antigens by indirect radioimmunoassay (IRIA; Williams, 1977). Hybrid 21R contains the following human chromosomes: 4, 6, 7, 10, 11, 12, 13, 14, 15, 16, 17, 18, 21 and X (Table I). The Y chromosome was not specifically assayed, although there is a gene encoding the 12E7 antigenic determinant on the Y, a well as the X chromosome (Goodfellow et al., 1983). Enzyme and antigen analyses were in complete concordance. The presence of mouse isozymes and H2Kk (recognised by monoclonal antibody 114.1; Oi et al., 1979) confirm the hybrid nature of 21R. Whole cell fusion of 21R to mouse thymoma line BW5147 The hybrid 21R contains many human chromosomes: to determine which chromosome carries the integrated pSV3-gpt sequences, subcloning and back-selection experiments might be attempted. However, we decided instead to attempt to transfer the selectable chromosome to other mouse cell types, since this would have two effects. Firstly, we would expect resultant hybrids to have reduced numbers of human chromosomes compared with 21R, which might allow chromosomal assignment of the pSV3-gpt integration site. Secondly, it would demonstrate the applicability of the dominant selection system to other hybrid phenotypes. We chose the mouse thymoma, BW5147 (Hymans and Stallings, 1974), as a partner for fusion with 21R, partly because of its suspension growth properties and partly because it rapidly segregates non-selected human chromosomes after fusion with human cells (Goodfellow et al., 1980). Approximately 10'

Table I. Marker analysis of whole cell and microcell hybrids 1

2

3

4

5

6

m~

7

8

9 10

11

12

13 14

15

16

17

18 19 20 21 22

>

=

2~~~~~~~~~~~~~~~E

Egpt2

+

+

+ +

21R

-

-

- -

+ + + + + + + + + + + + + + + + + + + + + + + - + + + + - - + + + + + + + + + + + + + - -

IR

-

-

-

-

-

X

00

Cl

()U

W

x

WL

. >

+ + + + - NDNDND + ND + + + ND ND ND

TRI C4 --- ------+ + + + + - + - - + + NDND TRI D60 - - - - + - + + - ND ND + - + + + + - + - + + NDND TRI D62 - - - - --------+ +----+ NDND . . . . . . . BW5147 - - - . - . . ..+ ND ND GPT17/1 GPT17/2 GPT17/3

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

--

+ +

GPT17/4+ PCC4

-

+

-

-

-

-

-

ND +ND

+

-

+

-

+ND

+

-

ND + ND + -

Isozymes were assayed by standard techniques (Harris and Hopkinson, 1976). Antigens were recognised by monoclonal antibodies using indirect radioimmunoassy (Williams, 1977) and a positive reaction was scored as greater than three times background binding by myeloma antibody P3.X63.Ag8 (Kohler and Milstein, 1975). ND

=

not determined.

1579


:n .:z

i,

Fig. 4. Galactokinase activities in human, mouse and hybrid cells. Standard starch gel electrophoretic analysis of extracts of (a) HFL121; (b) Egpt2; (c) 21R; (d) IR; (e) TRI C4; (f) TRI D60; (g) TRI D62 and (h) BW5147, stained for GALK.

Fig. 3. Sequences in tribrids complementary to pSV3-gpt. Filter hybridisation as in Figure 2 with (a) Egpt2; (b) 21R; (c) TRI C4; (d) TRI D60; (e) TRI D62 and (f) BW5147. Numbers indicate fragment sizes in kilobases.

cells each of 21R and BW5147 were fused with PEG 1000 and plated in 48 2 ml wells in RPMI 1640 supplemented with 10% foetal calf serum. After 24 h, MPA and xanthine were added to the medium. Every 3 or 4 days, cells in suspension were transferred to fresh wells to remove attached 21R cells from the cultures. After 3-4 weeks, suspension cells growing in three wells were picked and grown accordingly. We have called these 'tribrids', since at least two of them show evidence of a genetic contribution from three different cells (see below), and named them TRI C4, TRI D60 and TRI D62. Analysis of tribrids Figure 1, tracks e h, shows staining for HPRT activity of tribrids and BW5147. TRI C4 and TRI D62 show the presence of only the bacterial enzyme, whilst TRI D60 also has human HPRT. BW5147 is HPRT - and therefore shows -

1580

no staining. In addition, Figure 3 shows, by filter hybridisation analysis, the presence of sequences complementary to pSV3-gpt in the tribrids, but not in BW5147. The BglI restriction fragments shown are the same as those in hybrid 21R and indicate the transfer of genetic material carrying pSV3-gpt sequences from hybrid to tribrid. Egpt2 has an additional band of 1.6 kb. Isozyme and antigenic marker analyses (Table I) suggest that the pSV3-gpt sequences are integrated into human chromosome 17 since this is the only chromosome for which all three tribrids clearly possess appropriate markers, namely galactokinase (GALK; Figure 4) and H207 antigen. Indeed, these are the only human markers found with TRI D62. There is one discrepancy in the marker analysis: TRI D60 is positive for the X chromosome markers glucose-6-phosphate dehydrogenase (G6PD) and HRPT (Figure 1), but negative for marker antigen 12E7. This anomaly might be explained by a chromosome deletion (see below). Karyotype analysis was performed on the tribrids: TRI C4 and TRI D60 had chromosome numbers in the ranges 75-79 and 67- 76, respectively, whilst TRI D62 had a low median number of between 41 and 44 chromosomes. A high chromosome number is expected for what is effectively an intraspecific hybrid (21R x BW5147), and thus TRI D62 exhibits a surprisingly low chromosome number. GIl staining (Bobrow and Cross, 1974) showed the presence of 3-4 human chromosomes per cell in TRI C4 and quinacrine banding (Caspersson et al., 1971) showed that human chromosome 17 was present in most cells (Figure 5). This confirms chromosome 17 as the pSV3-gpt integration site. Occasionally, other human chromosomes were seen: the TRI C4 spread shown had chromosome 13 in addition to two copies of 17. TRI D60 had intact chromosomes 17 in the majority of cells, as well as other human chromosomes indicated by the marker analysis, but also had a translocation of human to mouse which may have been part of the long arm of the X


*.

.-.

;n Q

Fig. 5. Karyotype analysis of TRI C4. GIl staining distinguished human and mouse chromosomes and Q-banding identified human chromosomes 13 and 17 in this spread of TRI C4.

chromosome. This would explain the presence of human G6PD and HPRT in TRI D60, while 12E7 antigen, encoded by the short arm of the X chromosome, is absent (Table I). TRI D62 was intriguing in that it contained, as its only detectable human genetic material, a translocation to mouse of the long arm of chromosome 17. In addition, in some cells, this was modified such that additional mouse material was added onto the terminus of human 17q. This further defines the integration site of pSV3-gpt to this region. Since TRI D62 is positive for H207 antigen, the gene, MIC6, controlling expression of the antigen (Bai et al., 1982), must also be in this region of chromosome 17. Sheer et al. (1983) have recently further localised MIC6 to 17q2. 1-qter. Both TRI C4 and TRI D60 have H-2Kk on the cell surface (as detected by monoclonal antibody 11-4.1), whereas H-2Kk is not detectable by IRIA on TRI D62 or BW5147. The lack, or low level, of H-2K seems to be a peculiarity of AKR thymomas (Schmidt et al., 1979; Bishop et al., 1982) and it is possible that the H-2Kk on TRI C4 and TRI D60 is as a result of the presence of chromosomes from IR in these cells. This would mean that the reduced expression effect is not dominant in these hybrids, in contrast to results obtained with lymphocyte-thymoma hybrids (Bishop et al., 1982). There are several isozyme differences between IR and BW5147, one of which is in malic enzyme (MES): IR has a slower migrating form than BW5147 (Henderson, 1966). The presence of both mouse forms of MES in TRI D60 (not shown) again suggests a genetic contribution from the IR cell in this tribrid, consistent with its chromosome number. Only the fast form of mouse MES was seen in TRI C4 and TRI

D62. All tribrids are positive for Thyl . as recognised by monoclonal antibody OX7 (Mason and Williams, 1980), confirming a contribution from BW5147. The tribrids have been subcloned, and subclones are largely indistinguishable from the original cultures. These experiments show that it is possible to transfer the human chromosome carrying Ecogpt from attached to suspension cells. This may be a general way of introducing genes into specific cell types which are difficult to transfect by direct methods, such as calcium phosphate co-precipitation. Thus, a two-step gene transfer can be performed by transfection of a cooperative cell, followed by cell fusion with the intractable cell.

Microcell transfer of integratedpS V3-gpt sequences to mouse teratocarcinoma line PCC4 The integration of Ecogpt into a single human chromosome should allow the construction of hybrids containing only that chromosome by the microcell fusion technique (Fournier and Ruddle, 1977) where single chromosomes are transferred from donor to recipient cell. It is generally found that using a human cell as a microcell donor is less efficient than a rodent cell, although it has been possible to make human-rodent microcell hybrids (McNeill and Brown, 1980). We have found it easier to produce an intermediate conventional human-mouse hybrid and then to use this as a microcell donor of human chromosomes to mouse cells. In this way, a human X-6 translocation chromosome has been transferred to mouse teratocarcinoma PCC4 (Jakob et al., 1973) to produce the microcell hybrid, 1581


the presence of bacterial XPRT (Figure 6) and human GALK in all four microcell hybrids (Table I). No other human enzymes were seen. Human antigenic marker analysis was in accordance with these results except in the case of GPT17/1 where the marker for chromosome 17, H207, was absent, whilst that for the X chromosome, 12E7, was present. This was almost certainly due to a chromosomal rearrangement which was detected by karyotype analysis (not shown). However, GPT17/2 and GPT17/3 had normal chromosomes 17 (Figure 7). GPT17/4 contained a fragment of human chromosomal material translocated to mouse. All four clones were positive for the cell surface antigen SSEA-1 (Table I), recognised by a monoclonal antibody, which confirms the teratocarcinoma-like nature of these hybrids, since this antigen is found primarily on undifferentiated teratocarcinomas and early mouse embryos (Solter and Knowles, 1978). PCC4, as a teratocarcinoma cell, does not show detectable surface H-2b, although derived from a b haplotype 129 mouse (Jakob et al., 1973). The monoclonal antibody 28-14-8S (Ozato and Sachs, 1981), which recognises H-2Db, does not bind to PCC4 or the microcell hybrids (Table I). Fig. 6. XPRT analysis of microcell hybrids. Tracks show (a) human trol; (b) PCC4; (c-f) GPT17/1-4, treated as in Figure 1.

con-

Fig. 7. Karyotype analysis of GPT17/3.

MCP-6 (Goodfellow et al., 1982). We therefore decided to the hybrid 21R as a donor of the Ecogpt-carrying human chromosome 17. By introducing human chromosomes into mouse teratocarcinomas, it should be possible to study the developmental regulation of human genes when these hybrids are made to differentiate. Thus, we adopted PCC4 as the recipient cell in this fusion. Hybrid 21R cells were micronucleated by sequential Colcemid and cytochalasin B treatments (adapted from Fournier, 1981) and microcells prepared on a Ficoll/cytochalasin B step gradient (Wigler and Weinstein, 1975). Microcells from -2 x 107 21R cells were fused to 107 PCC4 cells after the method of Mercer and Schlegel (1979), using phytohemagglutinin P to agglutinate microcells and whole cells, and PEG 1500 to fuse them. Fusion products were distributed in four 25 cm2 flasks and selection was imposed 16 h later by adding MPA and xanthine to the medium. After -2 weeks, colonies with a teratocarcinoma-like morphology were picked and one clone from each flask grown up. These were named GPT17/# (where # denotes clone number). Enzyme analysis showed use

1582

Discussion The experiments in this paper describe the integration of genes for the bacterial enzyme XPRT and for SV40 early region proteins into human chromosome 17. By selecting for XPRT activity using MPA and xanthine, we have been able to construct human-mouse somatic cell hybrids containing this human chromosome. Hybrids have been made with three different mouse lines, namely L cell, suspension thymoma and tetratocarcinoma, which suggests general applicability of this selection system in hybrid formation. In particular, it has been possible to construct single human chromosome hybrids, after microcell transfer, with the mouse teratocarcinoma cell PCC4. This is potentially of interest since among several genes mapped to human chromsome 17 are pro-a (I) collagen (Sundar Raj et al., 1977; Huerre et al., 1982) and the myosin heavy chain complex (L.A. Leinwand, personal communication). These represent developmentally-regulated genes whose expression it may be possible to examine in the microcell hybrids, since equivalent mouse genes are expressed in some differentiating teratocarcinomas. Using a microcell hybrid, MCP-6, containing a human X-6 translocation chromosome on a PCC4 background (Goodfellow et al., 1982), it has been possible to isolate differentiated clones, after retinoic acid treatment, with either endoderm or fibroblastoid morphology (Benham et al., 1983). These are being examined for regulated human sequences: a similar procedure will be possible with the microcell hybrids described here. Similar approaches to that used here for novel humanmouse hybrid construction have been employed where u.v.-inactivated herpes simplex virus (HSV) has been used to transform TK - HeLa cells to TK +, followed by whole cell fusion to TK-deficient mouse L cells, selecting hybrids in HAT plus ouabain. Different human chromosomes were identified as HSV integration sites in each case: two marker chromosomes related to 17 and the short arm of X, respectively, (Donner et al., 1977), and chromosome 18 (McKinlay et al., 1980). Smiley et al. (1978) also used u.v.-inactivated HSV to transform TK - mouse cells which were then used as microcell donors to TK - Chinese hamster recipients. We have used a plasmid containing a cloned HSV-TK gene to transfect HeLa TK - cells in analogous experiments. How-


ever, subsequent hybrids were found to contain rearranged human chromosomes (unpublished data) which are of limited value for gene mapping. The advantages of our present experiments are that a dominant selection system is used and that an initially normal human cell is used as the chromosome donor. This increases the likelihood of karyotypically normal human chromosomes in hybrids. In addition, we have been able to construct single human chromosome hybrids by the microcell technique. It is perhaps ironic that the first example of a human chromosome containing a dominant selectable gene should be 17, for which classical HAT selection utilising human TK is available. However, the dominant nature of this selection system should be stressed, which obviates the need for mutant mouse cell lines as hybrid parents. The fact that none of the three mouse parents described in this work is TK - means that the presence of human chromosome 17 in hybrids made with these cells could not normally be selected. Having two selectable genes on chromosome 17 may, indeed, be of use: transfer of this chromosome to a TK - mouse cell under MPA/xanthine selection, coupled with selection against TK with 2-bromodeoxyuridine (BrdU), may result in chromosome fragmentation useful for regional mapping. Other reports exist of insertion of exogenous genes into human chromosome 17, one of which was of SV40 viral DNA responsible for transformation of human fibroblasts (Croce, 1977). The plasmid pSV3-gpt could be regarded as a late-defective SV40 genome, since it contains an intact early region. Although SV40 integration into human chromosomes is believed not to be site specific, since insertion into chromosome 5 (Hwang and Kucherlapati, 1980), 7 (Croce et al., 1973), 8 (Kucherlapati et al., 1978), and 17 (Croce, 1977) has been observed, only examination of a large number of integration sites will indicate any chromosome preferences. Further hybrids made from different pSV3-gpt-transfected human cells will be useful in this respect. A possible objection to the strategy of these experiments is that we are selecting for expression of SV40 large T antigen in the initial transfections. This was adopted to partially transform the human fibroblast transfectants to enable large numbers of cells to be grown. However, it is known that SV40-transformed human cell lines display marked karyotypic rearrangements (reviewed in Harnden, 1974), a phenomenon which, if it occurred in our transfectants, would defeat the object of the experiment. Nevertheless, we were encouraged by the example of LNSV, an SV40-transformed human Lesch-Nyhan fibroblast (Croce et al., 1973) and a human-mouse hybrid, Clone 21, derived from it, which contains only chromosome 7 as its human genetic component (Strand et al. 1974). Several tandemly-arranged copies of the SV40 genome are integrated in this chromosome (Campo et al., 1978). Clone 21 was made presumably before any major rearrangements had occurred in LNSV, since the human chromosome 7 in this hybrid is karyotypically normal and has been so for many pasages in our hands. In contrast, LNSV has developed multiple chromosomal rearrangements on repeated passaging (Begovich and Francke, 1979). The greater stability of the SV40-containing chromosome in the hybrid is almost certainly due to the non-permissive nature of the mouse parent cell with respect to SV40 replication, compared with the semi-permissivity of a human cell. The fact that we have karyotypically normal human chromosomes 17 in all but two of the hybrids described here is consistent with

this hypothesis and validates our approach. We have not seen any evidence of chromosome breakage upon propagation of our hybrids. That the bacterial gene Ecogpt is expressed in a teratocarcinoma environment in the microcell hybrids deserves comment since this gene is under SV40 early region control: enhancer, promoter, splicing and polyadenylation sequences are all from the virus genome. It has previously been argued that SV40 early region proteins are not produced after introduction of corresponding gene sequences into a teratocarcinoma cell (Swartzendruber and Lehman, 1975), either due to post-transcriptional (Segal et al., 1979) or translational (Linnenbach et al., 1980) control. However, the GPT17 microcell hybrids clearly express XPRT, which is responsible for their survival in selective medium. Thus, the SV40 control sequences allow efficient production of this protein. This is in agreement with the results of Jami and colleagues, who have used pSV2-gpt (Mulligan and Berg, 1980) to transfect PCC4: clones able to grow in MPA/xanthine have been isolated, albeit at 100-fold lower frequencies than with L cells (Bucchini et al., 1983). A different mouse teratocarcinoma cell line was also successfully transfected with pSV2-gpt by Wagner and Mintz (1982). The expression of SV40 large T antigen in both whole cell and microcell hybrids and its possible influences on levels of the cellular protein p53, which complexes with T (Lane and Crawford, 1979), will be the subject of further experiments. Materials and methods Cells Attached cells were grown at 37°C in Dd-lbecco's modified Eagle's medium (DME) in a 100o C02:90% air atmosphere. Suspension cells were grown at 37°C in RPMI 1640 medium in a 5% C02:95% air atmosphere. Media were supplemented with 10% foetal calf serum (FCS) and, where applicable, with HAT (1.6 1zM hypoxanthine, 10 AM methotrexate, 100 jiM thymidine), 25 1,g/ml MPA and 250 jig/ml xanthine (B10 medium). BIO' medium lacks methotrexate. Transfection followed the method of Wigler et al. (1977) except that no carrier DNA was used and 20 jg pSV3-gpt was added per 106 cells. Fusions Whole cell fusions involved 10' cells of each parent, fused with PEG 1000 in suspension (Galfre et al., 1977). Selection was imposed 24 h after fusion. For the Egpt2 x IR fusion, human cells were selected against using 10 jiM ouabain (Kucherlapati et al., 1975) and mouse cells with B1O or HAT medium. For the tribrid fusion, BW5147 was killed by MPA and xanthine and 21R selected against by repeated transfer of suspension cells to fresh 2 ml wells. Microcell fusion was based on the methods of Fournier and Ruddle (1977), Fournier (1981) and Goodfellow et al. (1982): - 2 x 107 21R cells were plated at two-thirds confluence in 0.2 Ag/ml Colcemid for 48 h, when they were replated in 2jg/ml cytochalasin B overnight. Microcells were separated by Ficoll/cytochalasin B step gradient centrifugation (Wigler and Weinstein, 1975), fractions containing microcells pooled, washed and enriched by selection for cells not attached to tissue culture dishes after 90 niin. Microcells were mixed with 107 PCC4 cells in RPMI 1640 and PHA-P (Difco) at 100 jg/ml added for 10 min at 37°C. After spinning to a common pellet, PEG 1500 (BDH) was used as fusogen as in whole cell fusions. Fusion products were distributed in four 25 CM2 flasks in DME plus 10% FCS and medium changed 16 h later with addition of 10 jg/mi MPA and 250 jg/ml xanthine.

Hybrd analysis Isozymes were analysed as in Harris and Hopkinson (1976). Radioimmunoassay was based on Williams (1977) as outlined in Goodfellow et al. (1980). Monoclonal antibodies are listed in Tunnacliffe et al. (1983). Karyotypes were analysed by GIl staining (Bobrow and Cross, 1974) and quinacrine banding (Caspersson et al., 1971). Southern blotting was based on Southern (1975) as detailed in the Cold Spring Harbor 'Molecular Cloning' manual (CSH, 1982). Briefly, 20 -g of high mol. wt. DNA was digested with a 3-fold excess of restriction enzyme (purchased from Biolabs) overnight at 370C. DNA was ethanol-precipitated and electrophoresed through 1%o agarose gels in TBE

1583

A. Tunnaciffe et al.

(90 mM Tris-borate, 90 mM boric acid, 2 mM EDTA), denatured in situ and transferred to nitrocellulose. Baked filters were hybridised overnight in 6 x SSC, 0.01 M EDTA, 5 x Denhardt's, 0.5% SDS, 10% dextran sulphate, 100 ,g/ml salmon carrier DNA with 100 ng denatured pSV3-gpt, nicktranslated (Rigby et al., 1977) to a specific activity of at least 2 x 106 c.p.m./ug. Filters were washed to 0.1 x SSC and exposed to film.

Acknowledgements We would like to thank R.C. Mulligan and P. Berg for plasmids; Lilly Research for mycophenolic acid; E. Simpson for 28-14-8S; A. Williams for OX7; L.A. Leinwand and T.B. Shows for communicating unpublished results; L. Crawford for comments on the manuscript; the I.C.R.F. Photography Department for photographs and C. Furse for excellent typing and secretarial assistance. The visit of B.O.B. to the I.C.R.F. was supported, in part, by the Swedish Natural Science Research Council and the Nilsson-Ehle Fund.

References Bai,Y., Sheer,D., Hiorns,L., Knowles,R.W. and Tunnacliffe,A. (1982) Ann. Hum. Genet., 46, 337-347. Begovich,A. and Francke,U. (1979) Cytogenet. Cell Genet., 23, 3-11. Benham,F.J., Wiles,M.V., Banting,G., Andrews,P. and Goodfellow,P.N. (1983) in Damjanov,I., Knowles,B.B. and Solter,D. (eds.), Human Teratomas, Humana Press, NJ, pp. 313-341. Bishop,C.E., Lewis,C.M. and Festenstein,H. (1982) Somatic Cell Genet., 8, 623-634. Bobrow,M. and Cross,J. (1974) Nature, 251, 77-79. Bucchini,D., Lasserre,C., Kunst,F., Lovell-Badge,R., Pictet,R. and Jami,J. (1983) EMBO J., 2, 229-232. Campo,M.S., Cameron,I.R. and Rogers,M.E. (1978) Cell, 15, 1411-1426. Caspersson,T., Lomakka,G. and Zech,L. (1971) Hereditas, 67, 89-102. Croce,C.M. (1977) Proc. Nat!. Acad. Sci. USA, 74, 315-318. Croce,C.M., Girardi,A.J. and Koprowski,H. (1973) Proc. Nat!. Acad. Sci. USA, 70, 3617-3620. Donner,L., Dubbs,D.R. and Kit,S. (1977) Int. J. Cancer, 20, 256-267. Fournier,R.E.K. (1981) Proc. Natl. Acad. Sci. USA, 78, 6349-6353. Fournier,R.E.K. and Ruddle,F.H. (1977) Proc. Natl. Acad. Sci. USA, 74, 319-323. Galfre,G., Howe,S.C., Milstein,C., Butcher,G.W. and Howard,J.C. (1977) Nature, 266, 550-552. Goodfellow,P.N. and Solomon,E. (1982) in McMichael,A. and Fabre,J.W. (eds.), Monoclonal Antibodies in Clinical Medicine, Academic Press, NY, pp. 365-393. Goodfellow,P.N., Banting,G., Levy,R., Povey,S. and McMichael,A. (1980) Somatic Cell Genet., 6, 777-787. Goodfellow,P.N., Banting,G., Trowsdale,J., Chambers,S. and Solomon,E. (1982) Proc. Natl. Acad. Sci. USA, 79, 1190-1194. Goodfellow,P., Banting,G., Sheer,D., Ropers,H.H., Caine,A., FergusonSmith,M., Povey,S. and Voss,R. (1983) Nature, 302, 346-349. Graham,F.L. and van der Eb,A.J. (1973) Virology, 52, 456-467. Hamden,D.G. (1974) in German,J. (ed.), Chromosomes and Cancer, Wiley, London, pp. 151-190. Harris,H. and Hopkinson,D.A. (1976) Human Genetics, Suppl. 1977 and 1978, published by North-Holland, Amsterdam. Henderson,N.S. (1966) Arch. Biochem. Biophys., 117, 28-33. Huerre,C., Junien,C., Weil,D., Chu,M.-L., Morabito,M., Van Cong,N., Myers,J.C., Foubert,C., Gross,M.-S., Prockop,D.J., Boue,A., Kaplan, J.-C., de la Chapelle,A. and Ramirez,F. (1982) Proc. Natl. Acad. Sci. USA, 79, 6627-6630. Hwang,S.-P. and Kucherlapati,R. (1980) Virology, 105, 196-204. Hymans,R. and Stallings,V. (1974) J. Natl. Cancer Inst., 52, 429-436. Jakob,H., Boon,T., Gaillard,J., Nicolas,J.-F. and Jacob,F. (1973) Ann. Microbiol., 124B, 269-282. K6hler,G. and Milstein,C. (1975) Nature, 256, 495-497. Kucherlapati,R.S., Baker,R.M. and Ruddle,F.H. (1975) Cytogenet. Cell Genet., 14, 192-193. Kucherlapati,R., Hwang,S.P., Shimizu,N., McDougall,J.K. and Botcham, M.R. (1978) Proc. Nat!. Acad. Sci. USA, 75, 4460-4464. Kusano,T., Long,C. and Green,H. (1971) Proc. Natl. Acad. Sci. USA, 68, 82-86. Lane,D. and Crawford,L. (1979) Nature, 278, 261-263. Linnenbach,A., Huebner,K. and Croce,C.M. (1980) Proc. Nat!. Acad. Sci. USA, 77, 4875-4879. Littlefield,J.W. (1964) Science (Wash.), 145, 709-710. Mason,D.W. and Williams,A.F. (1980) Biochem. J., 187, 1-20. McKinlay,M.A., Wilson,D.E., Harrison,B. and Povey,S. (1980) J. Nat!. Cancer Inst., 64, 241-248.

1584

McKusick,V.A. and Ruddle,F.H. (1977) Science (Wash.), 196, 390405. McNeill,C.A. and Brown,R.L. (1980) Proc. Natl. Acad. Sci. USA, 77, 53945398. Mercer,W.E. and Schlegel,R.A. (1979) Exp. Cell Res., 120, 417-421. Miller,R.L., Ramsey,G.A., Krenitsky,T.A. and Elion,G.B. (1972) Biochemistry (Wash.), 11, 4723-4731. Mulligan,R.C. and Berg,P. (1980) Science (Wash.), 209, 1422-1427. Mulligan,R.C. and Berg,P. (1981) Proc. Natl. Acad. Sci. USA, 78, 20722076. Nabholz,M., Miggiano,V. and Bodmer,W.F. (1969) Nature, 223, 358-363. Oi,V.T., Jones,P.P., Goding,J.W. and Herzenberg,L.A. (1979) Curr. Top. Microbiol. Immunol., 81, 115-129. Ozato,K. and Sachs,D.H. (1981) J. Immunol., 126, 317-321. Pontecorvo,G. (1975) Somatic Cell Genet., 1, 397-400. Rigby,P.W.J., Dieckmann,M., Rhodes,C. and Berg,P. (1977) J. Mol. Biol., 113, 237-251. Sack,G.H. (1981) In Vitro, 17, 1-19. Schmidt,W., Atfield,G. and Festenstein,H. (1979) Immunogenetics, 8, 311321. Segal,S., Levine,A.J. and Khoury,G. (1979) Nature, 280, 335-338. Sheer,D., Hiorns,L.R., Stanley,K.F., Goodfellow,P.N., Swallow,D.M., Povey,S., Heisterkamp,N., Groffen,J., Stephenson,J.R. and Solomon,E. (1983) Proc. Nat!. Acad. Sci. USA, in press. Smiley,J.R., Steege,D.A., Juricek,D.K., Summers,W.P. and Ruddle,F.H. (1978) Cell, 15, 455-468. Solter,D. and Knowles,B.B. (1978) Proc. Natl. Acad. Sci. USA, 75, 55655569. Southern,E.M. (1975) J. Mol. Biol., 98, 503-515. Strand,M., August,J.T. and Croce,C.M. (1976) Virology, 70, 545-549. Sundar Raj,C.V., Church,R.L., Klobutcher,L.A. and Ruddle,F.H. (1977) Proc. Natl. Acad. Sci. USA, 74, 4444-4448. Swartzendruber,D.E. and Lehman,J.M. (1975) J. Cell Physiol., 85, 179-187. Szybalski,W., Szybalska,E.H. and Ragni,G. (1962) Natl. Cancer Inst. Monogr., 7, 75-89. Tooze,J., ed. (1982) DNA Tumor Viruses, Part 2, 2nd edn. revised, published by Cold Spring Harbor Laboratory Press, NY. Tunnacliffe,A., Jones,C. and Goodfellow,P. (1983) Immunol. Today, in press. Tunnacliffe,A. and Goodfellow,P. (1983) in Goodfellow,P. (ed.), Receptors and Recognition, B16, Chapman and Hall, London, in press. Wagner,E.F. and Mintz,B. (1982) Mol. Cell Biol., 2, 190-198. Wigler,M.H. and Weinstein,I.B. (1975) Biochem. Biophys. Res. Commun., 63, 669-674. Wigler,M., Silverstein,S., Lee,L.S., Peilicer,A., Cheng,Y.-C. and Axel,R. (1977) Cell, 11, 223-232. Williams,A.F. (1977) Contemp. Top. Mol. Immunol., 6, 83-116.