A lacZ-based vector system for the rapid detection of V(D)J recombinase activity

JOURNALOF IMMUNOLOGICAL METHODS Journal of ImmunologicalMethods 175 (1994) 259-266

ELSEVIER

A lacZ-based vector system for the rapid detection of V(D)J

recombinase activity Bernd Biihler, Georges K6hler, Peter J. Nielsen

*

Max-Planck-lnstitut fiir Immunbiologie, Stiibeweg 51, D-79108 Freiburg, Germany

Received 6 April 1994; revised received 6 June 1994; accepted 15 June 1994

Abstract

Episomal vectors have been developed which are useful for studying V(D)J recombination both after transient transfections and in stably transfected cells. In contrast to recombination substrates previously described for transient assays, rearrangement of these vectors results in expression of fl-galactosidase which can be visualized directly in the transfected cell, shortening the time required for the assay to 1-2 days instead of 3-4 days. When these substrates are stably integrated into a preB cell line, subclones are found which show no /3-galactosidase staining, although the substrate is properly integrated, transcriptionally active and the transfectants still possess recombinase activity. This finding suggests that, at least in some chromosomal locations, transcription through a locus bearing recombination signal sequences is not sufficient for V(D)J recombination. Using these same vectors, we estimate that the frequency with which V(D)J recombination-negative preB variants arise is less than 10 -4 per generation. Keywords: V(D)J recombination; Recombinase activity; scid; fl-Galactosidase

1. Introduction

V(D)J recombination is the process that assembles immunoglobulin and T cell receptor genes from separate germline regions coding for variable (V), diversity (D) and joining (J) ele-

Abbreviations: FDG, fluorescein-di-fl-D-galactopyranoside; scid, severe combined immune deficiency; X-Gal, 5-bromo-4chloro-3-indolyl-fl-D-galactoside. *Corresponding author. Tel.: + +49-761-5108319; Fax: + + 49-761-5108221; electronic mail: [email protected].

ments. This assembly occurs only in developing lymphoid cells and depends on conserved recombination signal sequences (RSS) which flank the coding segments. The RSS consist of a palindromic h e p t a m e r and AT-rich n o n a m e r separated by a nonconserved spacer of 12 or 23 base pairs (Tonegawa, 1983). V(D)J recombination generates 2 types of junctions, coding joints (junction between coding elements) and signal joints (reciprocal junction usually with a precise fusion of the RSS) (Lewis and Gellert, 1989). A significant step in the study of V(D)J recombination was achieved with the demonstration that plasraids carrying RSS undergo recombination when

0022-1759/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0022-1759(94)00170-2

260

B. Biihler et al. /Journal of lrnmunological Methods 175 (1994) 259-266

transiently introduced into lymphoid cells. Typically, substrate plasmids are recovered 48 h after transfection and transformed into bacteria. There, recombined plasmids can be identified, either by selection with chloramphenicol (Hesse et al., 1987) or by staining with X-Gal for bacterial /3-galactosidase (Kallenbach et al., 1990). In another vector system, recombination can be detected by polymerase chain reaction (Abe et al., 1991). All these plasmid substrates require extraction from the transfected cell and either transformation into E. coli or PCR analysis. In contrast, other assays for recombination have utilized stably integrated vectors containing a marker gene that becomes transcriptionally active only when inverted by a site-specific recombination event between RSS. Detection of recombination involves drug selection and Southern blot analysis or enzymatic detection of Escherichia coli /3-galactosidase (Schatz and Baltimore, 1988; Yancopoulos et al., 1990, Matsuoka et al., 1991). Here, we describe new recombination substrates which permit a direct evaluation of recombination in transient assays within the eukaryotic cell. In addition, we show that these substrates are valuable for studies of V(D)J recombination in stably transfected cells.

2. Material and methods

2.1. Plasmid construction The rearrangement substrate pNZP4 was constructed by first ligating the 6.3Kb SalI fragment from pJH201 (Hesse et al., 1987) containing the polyoma TAg, the polyoma origin of replication and the pUC13 backbone with the 4.8 kb SalI fragment from pSVZt (a gift from G. McMaster, Ciba-Geigy, Basel) which contains the bacterial lacZ gene flanked on either end by the SV40 promoter and polyadenylation signal. This construct was termed p201-SVZ8 and, when transfected into eukaryotic cells, is able to replicate episomally and expresses /3-galactosidase constitutively. In a separate reaction, the oop-transcriptional termination element in pJH290 (Lieber et al., 1988) was replaced with the neomycin resis-

tance coding region and SV40 polyadenylation signal derived by EcoRV-BamHI digestion of p R c / C M V (Invitrogen). Then, the neo gene and flanking RSS derived from pJH290 was inserted into p201-SVZ8 between the SV40 promoter and the lacZ gene. To construct pNZP3, the same procedure was followed except that pJH288 was used instead of pJH290 and the RSS containing the 12 bp spacer in pJH288 was first inverted by SalI digestion and religation before inserting the neomycin gene. 2.2. Cell culture and transfections Cell lines were cultured in Iscove's MDM (Gibco-BRL) supplemented with 10% FCS (Seromed), 100 U / m l penicillin, 100 /zg/ml streptomycin, and 50 /xM /3-mercaptoethanol at 37°C and 7.5% CO 2. 33.1.1 + and 33.1.1- are sublines from 33.1 derived from a /z + K transgenic (Swiss albino X BALB/c) F1 mouse (Iglesias et al., 1990). The AMuLV-transformed Scid 1 line was a generous gift from U. Fritzsche. For electroporation, 1 x 107 washed cells in 250 /zl serum free culture medium were mixed with 50 /~1 PBS containing 10 /~g vector DNA and 10/.Lg pCMV-Luc to monitor the transfection efficiency. After 10 min at room temperature, electroporation was performed at 225 V and 950 /zF. Cells were diluted with 12 ml culture medium and cultured for 24-48 h before analysis. To obtain stable transfectants, a 5.6 kb AseI fragment of pNZP4 was isolated (Biotrap; Schleicher and Schiill). After electroporation of 10 /zg of this fragment, selection was started 24 h later in medium containing 2 mg/ml G418 (Gibco). 2.3. V(D)J recombination assay Cells were fixed 48 h after the transfection with 2% glutaraldehyde, 2% formaldehyde, 100 mM Pipes (pH 6.9), 2 mM MgCI2, 5 mM EGTA for 10 min at room temperature and washed two times with PBS/2 mM MgCI 2. The staining solution consisted of 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6, 2 mM MgC12, 0.01% NaDOC, 0.02% NP40 and 1 mg/ml X-Gal in PBS. Staining was performed in 100/zl for 1 h for transiently trans-

B. Biihler et aL /Journal of Immunological Methods 175 (1994) 259-266

fected cells and overnight with ceils having a stably integrated copy. Labelling for flow cytometry with fluoresceindi-/3-galactosidase-D-galactopyranoside (FDG) was done as previously described (Nolan et al., 1988). Briefly, cells were washed and resuspended at 2-10 × 106 cells/ml in PBS/4%FCS and 50/zl of this cell suspension were incubated for 10 min in 4 ml FACS tubes (Falcon) in a 37°C water bath. An equal volume of prewarmed 2 mM FDG dissolved in water was added, further incubated for 1 min and finally diluted with 1 ml cold PBS/4% FCS. Samples were kept on ice, until analysis was performed using a FACScan cytometer (Becton Dickinson) with LysislI software. 2.4. Southern blot analysis

Genomic DNA was isolated according to Miller et al. (1988), 10 /~g were digested with HindlII and fractionated on a 0.8% agarose gel. After blotting on Biodyne A, the filter was hybridized with a radio-labelled E c o R I - H i n d l I I fragment of lacZ (Feinberg and Vogelstein, 1984).

Hindlll \

A

~

Asel I JHindlll

261

3. Results 3.1. A new vector for a transient recombination assay

Self-replicating recombination substrates were constructed to measure recombinase activity directly in transiently transfected eukaryotic cells (Fig. 1A). The pUC13-based vectors pNZP3 and pNZP4 contain the polyoma early region for replication, the bacterial lacZ gene flanked by the SV40 promoter/enhancer and SV40 polyadenylation signal. In addition, a neomycin resistance gene flanked by V(D)J recombination signal sequences (RSS) is inserted between the promoter and the translational start of the lacZ gene. In both vectors, the RSS are in opposite orientation so that rearrangement results in deletion of the neomycin resistance gene. In pNZP3, successful recombination leads to a head to head ligation of the heptamers within the vector, forming a signal joint. In pNZP4, the RSS are both inverted and therefore not retained on the vector after recombination, a coding joint is formed. As a result, the formerly silent lacZ gene is expressed and rear-

l/Hindlll

B~OOOi Hindlll '9.

loo. x

E~RI

~ LL

Asel / ori ~ - - " - " / ~ \ E c o R I

nd 33.1.1+

33.1.1-

Scid 1

Fig. 1. A: recombination substrate pNZP4. In the Vector pNZP3 the RSS (represented by triangles) are both inverted. B: frequency of fl-galactosidase-positive preB cells. PreB cells were transfected with the appropriate vectors and stained with X-Gal 48 h post transfection. Frequency of positive cells are depicted as blue,/3-galactosidase-positive cells per 106 cells ( × 10-6). Values given are mean values of seven transfections, nd: not detectable.

B. Bi~hler et al. /Journal of lmmunological Methods 175 (1994) 259-266

262

rangement of the vector can be detected by a simple staining protocol for/3-galactosidase activity directly in the eukaryotic cell. Both vectors were tested by electroporation into 3 AMuLV-transformed preB cell lines (Fig. 1B). Three series of transfection were done with independent DNA preparations. Each electroporation was performed in duplicates or triplicates. The 33.1.1 + line is a late preB line which has lost recombinase activity (Iglesias et al., 1991). With both substrates, no blue cells could be detected in more than 107 transfected cells of this line. Therefore, the background of this assay must be below 10 -7. The recombinase active 33.1.1 - , in contrast, shows blue cells with a frequency between 4 and 120 X 10 -6, which is at least 2-3 orders of magnitude above background. In order to use the frequency of blue cells as an estimate for the recombination frequency, the fraction of cells which received the recombination substrate was determined. To do this, a parallel electroporation was performed with p201-SVZ8.

0 ec~,-~t",,I,--~ "g

0 1 e¢~ e¢~

This vector is a precursor of pNZP4, which lacks the neomycin gene and its' flanking RSS and therefore constitutively expresses the /3-galactosidase gene without recombination. Hence, the percentage of cells that were successfully transfected with p201-SVZ8 can be visualized by X-Gal staining. After this correction, the frequency of transfected cells which rearranged the substrate was between 0.1 to 1% (not shown). The ratio between the frequencies of blue cells with pNZP3 and pNZP4, respectively, was always below 1 (0.57 + 0.20). The reason for this is as yet unclear, but the presence of the RSS in the rearranged pNZP3 might negatively affect expression of the lacZ gene. Scid 1 is a preB cell line derived from the scid mouse which rearranges the pNZP3 substrate to a high frequency, comparable to the 33.1.1 - . In contrast, a 30-fold lower frequency of blue cells is observed with the coding joint substrate pNZP4. This difference between the efficiency of coding and signal joint is a feature of the scid mutation, already ob-

O 0 I 04

I

! ,--~ M" 04 0 4 0 4 I

+

+

I

y.~

r--4 r-4

+ !

r-'-4

"~ r-4 "~ O "1( 04 r-4 r--4 r--~ r-4 ,--4 C,,I tr~ ,..~ t',,I tt~ ~,D ,--~ e¢~ ,~1- ,...4 ,~. oO M- r..~ t'~ e,3 e¢~ e¢~ e¢~

23 9.6 6.6

3 ,3 k b 2.2 2.0

Fig. 2. A n a l y s i s of the i n t e g r a t e d r e c o m b i n a t i o n substrate. S o u t h e r n blot of HindIII-digested g e n o m i c D N A from 16 i n d e p e n d e n t c l o n e s of 33.1.1 - t r a n s f e c t e d w i t h the AseI f r a g m e n t of t h e r e a r r a n g e m e n t s u b s t r a t e p N Z P 4 . G e n o m i c D N A f r o m 33.1.1 - was i n c l u d e d as n e g a t i v e control. T h e t h r e e rows o n the right are also D N A from 33.1.1 - with one, t h r e e a n d five copies of p N Z P 4 p e r genorne a d d e d to the d i g e s t i o n as a positive control for hybridization. * i n d i c a t e s the /3-galactosidase-negative clones; v a l u e s in b r a c k e t s are copy n u m b e r s e s t i m a t e d f r o m a S o u t h e r n blot with EcoRI-digested D N A (not shown).

B. Biihler et al. /Journal of Immunological Methods 175 (1994) 259-266

served by others (Lieber et al., 1988; Harrington et al., 1992).

Table 1 Recombinase activities in transfectants of 33.1.1 Cell line

3.2. Recombination of a stably transfected substrate The neomycin gene in the vectors should permit the selection of stably integrated, unrearranged substrate in the genome of a cell line. To test this, we transfected a 5.6 kb AseI fragment of pNZP4 containing only the part of the construct necessary for expression and rearrangement into 33.1.1 - (see Fig. 1A). Recombination of the integrated copy was tested on 60 independent neomycin-resistant subclones by staining for/3-galactosidase activity with X-Gal. In contrast to the staining for transient transfection, this staining takes more than 6 h, because the /3-galactosidase activity per cell is derived from only 1 (to maybe a few) rearranged copy. Of these 60 clones, approximately one half contained blue cells with a frequency greater than 0.1%. For the other half of the clones, no blue cells were seen when 104-105 cells were examined. This absence of blue cells could be the result of damage to the vector sustained during integration. We therefore performed a Southern blot analysis of 16 clones, 4 of which belonged to the /3-galactosidase-negative group (Fig. 2). Because all subclones were selected for neomycin resistance, the 5' part of the cassette with the promoter and the neomycin gene can be regarded as intact. Hence, mutations are most likely in the lacZ part. However, the Southern blot demonstrated that all 4 /3-galactosidase-negative clones have only the expected 3.3 kb HindlII fragment, except G1 which also has additional bands. Concerning the /3-galactosidase-positive clones, 3 of the 12 lines show bands larger than the expected 3.3 kb HindlII fragment. This is probably due to loss of the HindlII site at the 3' end of the lacZ gene during integration, since the AseI site used to prepare the DNA for transfection is very close to the HindlII site. On a Southern blot from EcoRI-digested DNA (not shown), the same fragments hybridized with a lacZ- and a neo probe. This means that all neomycin genes are followed by an intact lacZ gene although not all neomycin genes are neces-

263

A1 Bll G1 A2 A3 B4 E12 33.1.1 -

Frequency of fl-gal-positive cells ( x 1 0 - 6 ) From integrated copy

From pNZP4 experiment 1

From p N Z P 4 experiment 2

< 1.6 * < 0.4 < 0.9 1.4- 3.6× 104 1.8- 4.0× 104 8 -16 x104 5 - 9 ×104 -

230 50 30 250 310 540 90 60

420 20 270 100 280 590 110 200

Cells were transfected with p N Z P 4 and 48 h later stained for

lacZ activity for 30-60 min. Untransfected cells do not show any staining derived from rearranged integrated copies after that time. T h e frequency of blue cells from the integrated copy is given as a comparison. Values from two independent experiments are shown and represent frequencies of blue cells in 106 cells ( × 10-6). * prolonged staining for 2 days sometimes gave a light blue granular staining localized to a small region in the cytoplasm in less than 0.01% of the cells.

sarily expressed. Therefore, the absence of /3galactosidase staining in these subclones cannot be explained by gross damage to the integrated copy (copies) of the rearrangement vector. However, minor mutations in the lacZ gene cannot be excluded. An alternative explanation would be that the /3-galactosidase-negative subclones have lost recombinase activity, which was shown to be unstable in some cell lines (Rathbun et al., 1993). To test this, we took advantage of the fact that the staining for/3-galactosidase activity derived from the endogenous integrated copy takes more than 6 h. Thus, 3 /3-galactosidase-negative and 4 /3galactosidase-positive clones were supertransfected with pNZP4 to check for recombinase activity. The results in Table 1 demonstrate that all clones including the /3-galactosidase-negative subclones possess recombinase activity. 3.3. Stability of recombinase activity in long term culture The recombination substrate described here was used to test whether recombinase activity is a stable property in these cells. If variant cells

264

B. Biihler et al. /Journal of lmmunological Methods 175 (1994) 259-266

15-

•

in G418

[]

4 days without G418

m

O3

= o >~

10-

.=,_

i

g O. u

C Q> n

I _hi ,i I

!

I

I

I

IJl

of G418. In addition, when A3 was subcloned, no G418-resistant cultures were obtained in which 100% of the cells were /3-galactosidase-positive (not shown). To select for recombinase-negative variants, A3 and B4 were maintained in medium containing G418 for more than 3 months (> 100 generations). Subclones established from this longterm culture were stained with X-Gal for/3-galactosidase-positive cells. Alternatively, cells were loaded with FDG by osmotic shock and analyzed by flow cytometry for/3-galactosidase activity. All of the subclones, 102 from A3, and 107 from B4, contained /3-galactosidase-positive cells, suggesting that the loss of recombinase activity is rare (see discussion section).

Cell lines Fig. 3. Effect of G418-selection on the number of 13-galactosidase-positive ceils. Cells from the preB line 33.1.1-and six subclones of B4 were stained with FDG and analyzed by flow cytomet~. Subclones were propagated for 10 days at low densities to ensure effective G418 selection. The percent of

/3-galactosidase-positive cells is depicted for G418-selected cells (filled bars) and for the same cells cultured for 4 days without G418 (open bars).

lacking a recombinase activity arise in the culture, they are expected to have a growth advantage, because their descendants do not die due to recombinational loss of the neomycin gene. Two transfectants carrying a stably integrated recombination substrate were chosen for this experiment (clones A3 and B4). B4 has only one copy of the substrate, so that a cell with a rearranged locus is no longer neomycin resistant and dies when grown in G418. This was shown by cultivating several subclones of B4 at low density (to ensure efficient selection) with and without the antibiotic G418 for 4 days before loading the cells with the/3-galactosidase substrate FDG and subsequent analysis by flow cytometry. Increased numbers of /3-galactosidase positive cells were observed in all 6 subclones after removal of the G418 selection (Fig. 3). Although A3 has 2 integrated copies of the rearrangement substrate, rearrangement in this clone leads to G418 sensitivity because higher levels of positive cells were also observed when it was cultured in the absence

4. Discussion

The transient V(D)J recombination assay we describe here has several advantages over previously described assays. No recovery of the transfected plasmid is required since rearrangement is visualized directly by a simple staining procedure of the transfected cells. We consistently found higher amounts of /3-galactosidase-positive cells when cells from the same transfection were stained 24 h instead of the standard 48 h post transfection (data not shown). Thus, this method is both easier and less time consuming. For the recombinase-negative 33.1.1 + , no background staining was observed. The recombination-positive 33.1.1 - line, in contrast, shows at least 2 to 3 orders of magnitude higher levels of /3-galactosidase-positive cells. After correcting for variations in transfection efficiencies, the relative recombination efficiency was estimated to be about 0.1-1%. This is similar to the efficiencies reported using other extrachromosomal substrates (Hesse et al., 1987; Kallenbach et al., 1990). The coding joint substrate pNZP4 always showed higher frequencies of blue cells in 33.1.1- than pNZP3 (0.57 + 0.20). In contrast, the frequency of signal to coding joints was reported to be between 1 and 2.5 in lymphoid cells from normal mice (Lieber et al., 1988) and between 5 and 6 in human preB cells (Gauss and

B. Biihler et al. /Journal of lmmunological Methods 175 (1994) 259-266

Lieber, 1993). Since we found a slight preference for the coding joint, we favour the idea that the remaining RSS in the rearranged pNZP3 negatively affects expression of the lacZ gene. Apart from this small difference, both vectors work equally well in a normal mouse cell line. In ceils carrying the scid mutation, coding joint formation is strongly affected. A difference of more than 360-fold or even more than 1000-fold was reported for similar substrates (Harrington et al., 1992; Lieber et al., 1988). We found only a 30-fold difference with our substrate in the Scid 1 cell line. The reason may be that we do not look at single rearranged plasmid molecules, but at a pool of unrearranged, incorrectly rearranged and correctly rearranged molecules in one cell. One single correct recombination in this pool can lead to /3-galactosidase activity, while several independent correct rearrangements in the pool would be scored as a single blue cell by the /3-galactosidase staining. The consequence of this is that high recombination frequencies might be underestimated, whereas low recombination frequencies nevertheless result in a substantial quantity of/3-galactosidase-positive cells. Transcription was reported to be necessary for accessibility of the recombinase machinery (Blackwell et al., 1986) and lymphoid-specific enhancer elements are suggested to regulate the accessibility of B and T cell receptor loci (Ferrier et al., 1990; Capone et al., 1993). From studies in cell lines, it is known that these enhancers can be replaced by the promiscuous, viral SV40 enhancer (Oltz et al,. 1993). The rearrangement substrate we describe here uses the SV40 enhancer/promoter element for expression of the neomycin gene and, following rearrangement, the /3-galactosidase gene. Surprisingly, when the neomycin gene in the unrearranged substrate was used for selection of stable integration in the genome, almost 50% of the independent transfectants in the preB cell line 33.1.1 - did not contain /3-galactosidase-positive cells. When 3 representative clones of these transfectants were analyzed, they did not show any gross rearrangements of the integrated substrates and still retained recombinase activity when tested with extrachromosomal substrates. This lack of rearrangement can-

265

not be explained by a general inaccessibility of the substrate resulting from integration within heterochromatin, for example, because the cell lines express the neomycin gene present between the recombination signal sequences. Furthermore, the same promotor driving neomycin transcription before rearrangement is used for transcription of the lacZ gene after recombination. Therefore, it seems unlikely that the lacZ gene would not be expressed after recombination. But this does not rule out the possibility, that the expression in some chromosomal positions is too low to be detectable by X-Gal staining. The presumed lack of rearrangement of transcriptionally active, integrated substrates may suggest that recombination requires a certain threshold level of transcription which is only ensured when integration has occurred into an appropriate chromosomal position, for example close to an active enhancer. Chromatin conformation, independent of transcription might also influence the efficiency of rearrangement. Alternatively, additional factors needed for rearrangement of chromosomal but not extrachromosomal loci might be absent in these cells. Interestingly, in another study with a similar substrate, two out of five stable transfectants were also devoid of rearrangements of the integrated vector (Yancopoulos et al., 1986). In two stable transfectants of 33.1.1-, no recombinase deficient variants were found during 3 months of culturing. Because cells were grown in neomycin-selective medium, recombinasenegative mutants have a growth advantage and should be enriched. Based on the observation that clones grown without G418 gain in average 8% more blue cells in 4 days (Fig. 3), we estimate the recombination frequency between 1 and 2% per generation. Assuming this frequency of rearrangement and an appearance of recombinasenegative variants with a frequency of 0.01%, we calculated that a culture should contain 1.7-3.7% recombinase-negative cells after 100 generations. We found none within more than 200 subclones (less than 0.5%) and hence the frequency of variants must be below 0.01%. Although unstable RAG expression has been reported in some preB lines (Rathbun et al.,

266

B. Biihler et al. /Journal of Immunological Methods 175 (1994) 259-266

1993), the 33.1.1- line used here maintains an extremely constant level of recombinase activity in long term cultures.

Acknowledgments We thank B. Groetschel for expert technical assistance, Dr U. Fritzsche for the Scidl preB cell line, Dr. G.K. McMaster for pSVZt and Dr. M. Gellert for the recombination substrates pJH 201, 288 and 290.

References Abe, M., Hayashida, K., Takayama, K. and Shiku, H. (1991) V(D)J recombinase activity in primary and secondary murine lymphoid organs: assessment by a PCR assay with extrachromosomal plasmids. Int. Immunol. 3, 1025-1033. Blackwell, T.K., Moore, M.W., Yancopoulos, G.D., Suh, H., Lutzker, S., Selsing, E. and Alt, F.W. (1986) Recombination between immunoglobulin variable region gene segments is enhanced by transcription. Nature 324, 585-589. Capone, M., Watrin, F., Fernex, C., Horvat, B., Krippl, B., Wu, L., Scollay, R. and Ferrier, P. (1993) TCR beta and TCR alpha gene enhancers confer tissue- and stagespecificity on V(D)J recombination events. EMBO J. 12, 4335-4346. Feinberg, A.P. and Vogelstein, B. (1984) A technique for radiolabelling DNA endonuclease restriction fragments to high specific activity. Anal. Biochem. 137, 266-267. Ferrier, P., Krippl, B., Blackwell, T.K., Furley, A.J.W., Suh, H., Winoto, A., Cook, W.D., Hood, L., Constantini, F. and Alt, F.W. (1990) Seperate elements control DJ and VDJ rearrangement in a transgenic recombination substrate. EMBO J. 9, 117-125. Gauss, G.H. and Lieber, M.R. (1993) Unequal signal and

coding joint formation in human V(D)J recombination. Mol. Cell. Biol. 13, 3900-3906. Harrington, J., Hsieh, C.-L., Gerton, J., Bosma, G. and Lieber, M. (1992) Analysis of the defect in DNA end joining in the murine SCID mutation. Mol. Cell. Biol. 12, 4758-4768. Hesse, J.E., Lieber. M.R., Gellert, M. and Mizuuchi, K. (1987) Extrachromosomal DNA substrates in pre-B cells undergo inversion or deletion at immunoglobulin V-(D)-J joining signals. Cell 49, 775-783. Iglesias, A., Kopf, M., Williams, G.S., Biihler, B. and K6hler, G. (1991) Molecular requirements for the mu-induced light chain gene rearrangement in pre-B cells. EMBO J. 10, 2147-55. Kallenbach, S., Goodhardt, M. and Rougeon, F. (1990) A rapid test for VDJ recombinase activity. Nucleic Acids Res. 18, 6730. Lewis, S. and Gellert, M.R. (1989) The mechainism of antigen receptor gene assembly. Cell 59, 585-588. Matsuoka, M., Nagawa, F., Okazaki, K., Kingsbury, L., Yoshida, K., Miiller, U., Larue, D.T., Winer, J.A. and Sakano, H. (1991) Detection of somatic DNA recombination in the transgenic mouse brain. Science 254, 81-86. Miller, S.A., Dykes, D.D. and Polesky, H.F. (1988) A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 16,1215. Oltz, E.M., Alt, F.W., Lin, W.-C., Chen, J., Taccioli, G., Desiderio, S. and Rathbun, G. (1993) A V(D)J recombinase-inducible B-cell line: Role of transcriptional enhancer elements in directing V(D)J recombination. Mol. Cell. Biol. 13, 6223-6230. Rathbun, G., Oltz, E.M. and Alt, F.W. (1993) Comparison of RAG gene expression in normal and transformed precursor lymphocytes. Int. Immunol. 5, 997-1000. Schatz, D.G. and Baltimore, D. (1988) Stable expression of immunoglobulin gene V(D)J recombinase activity by gene transfer into 3T3 fibroblasts. Cell 53, 107-115. Tonegawa, S. (1983) Somatic generation of antibody diversity. Nature 302, 575-581. Yancopoulos, G., Nolan, G., Pollock, R., Prockop, S., Li, S.C., Herzenberg, L.A. and Alt, F.W. (1990) A novel fluorescence-based substrate for assaying and separating live cells according to VDJ recombinase activity. Mol. Cell. Biol. 10, 1697-1704.