Fusion of a scid Pre-B Cell with a Wild Type (Myeloma) B Cell Results in Correct Rearrangement of a V(D)J Recombination Substrate

Mice with the scid mutation have a defect in the V(D)J recombinase. In order to determine whether the SCID product is normally present in mature B cells that do not have the recombinase activity, scid pre-B cells were fused with myeloma cells. It was found that in the hybrid cells, a rearrangement test gene was correctly joined immediately after fusion. The same test gene was aberrantly rearranged in the scid pre-B cells. Stable hybrids between the scid pre-B and the myeloma cells had lost the expression of RAG-1 and RAG-2 genes, supporting the previous finding of an inhibitor of rearrangement in myeloma cells that acts shortly after fusion. Thus, mature B cells apparently contain the SCID product, the wild type SCID function is not competitively interfered with by products present in scid pre-B cells, and the SCID product seems not to be a target for the recombinase inhibitor.


INTRODUCTION
Mice homozygons for the scid mutation do not produce B or T lymphocytes (Bosma et al., 1983) because of a defect in V(D)J recombination (Schuler et al., 1986). Rearrangement of immunoglobulin (Ig) genes in scid pre-B cells results in nonfunctional Ig genes due to large deletions (Hendrickson et al., 1988;Kim et al., 1988;Lieber et al., 1988;Malynn et al., 1988;Okazaki et al., 1988;Blackwell et al., 1989;Bosma and Carroll, 1991). The rearrangement defect suggested that the SCID product may be an essential component of the V(D)J recombinase. Our studies were undertaken in order to determine whether mature wild type B cells express the SCID product. We fused scid pre-B cells with mature B (myeloma) cells and determined the rearrangement status of a test gene that had previously been stably transfected into the myeloma cells. It was found that the test gene was rearranged correctly in the hybrids immediately after fusion. The same gene was rearran'ged aberrantly in the scid pre-B cells. *Corresponding author. It had been found previously that fusion of normal pre-B cells with myeloma cells abolishes the rearrangement activity of the pre-B cells due to suppression of the expression of the recombinase activating genes RAG-1 and RAG-2 (Engler et al., 1991b). Thus, mature B cells contain an inhibitor of the recombinase; Apparently, the inhibition is at the transcriptional level, because RAG-1 and RAG-2 genes under the transcriptional control of a ubiquitously expressed promoter/enhancer combination are expressed in myeloma cells (J. Zhao, P. Roth, and U. Storb, unpublished). Thus, during a short time after fusion before the existing RAG-1 and RAG-2 mRNAs and proteins are degraded, the recombinase is active. We utilized this window to determine whether normal SCID function was present in the myeloma cells.

RESULTS
A recombination test gene, pHRD-neo (Fig. 1B), was transfected into the normal pre-B cells 38B9 and stable transfectants were selected with the drug G418. As shown in Fig aThe number of independent rearrangement events. In parentheses is shown the presence (+) absence (-) of RAG-1 and RAG-2 mRNAs. bSee text. CPresumably  only expressed immediately after fusion (see Engler et al., 1991b;and text). The difference in the frequency of correct rearrangements between the scid pre-B cells (0/34) and the hybrids (2/2) is highly significant, p<-0.002 (Fisher's Exact Test, Kendall and Stuart, 1973 3B). The rearrangement substrate is partially unrearranged in six of the scid transfectants shown (#14, 15, 18, 19, 20, and 21). When rearranged, it shows mostly bands of the wrong size (#14, 15, 16, 20, 22, and 26) or is completely deleted (#17, 23, 24, and 25). All stable transfectants have retained the neo gene, presumably because of the selection with G418. Thus, deletions larger than 6.6 kb would not be scored, because they would invade the neo gene. These results confirm previous findings that this scid pre-B cell has recombinase activity, but that its function is defective (Schuler et al., 1986), because none of the rearranged genes was correctly rearranged (Table 1). Presumably, scid pre-B cells produce all the components required for rearrangement, except for the SCID product.
To determine whether the wild type SCID product is active in mature B cells that have lost the ability for V(D)J rearrangement, we fused $33 cells with the myeloma X63-Ag8.653 (hereafter referred to as Ag8). The,myeloma cells had been stably cotransfected with the pHRD test gene and the neo gene before fusion (Ag8HRD7; Engler et al., 1991b;see Fig. 1C). The myeloma transfec-tants Ag8HRD7 had been kept in culture for months and did not show evidence of rearrangement of the test gene (Fig. 2D). However, two of the .hybrids between Ag8HRD7 and $33 did rearrange pHRD, apparently correctly (Fig. 2C).
No rearranged bands of incorrect length were found after fusion of $33 with Ag8.
In order to determine if indeed fusion with the myeloma cells had resulted in correcting the scid rearrangement defect, the VJ joints of the pHRD test gene in the two scid pre-B x myeloma hybrids were sequenced. This confirmed that they were rearranged correctly with small deletions (3 and 5 bp) and the addition of one uncoded nucleotide (N) (Fig. 3A). Eleven independent DNA clones of the rearranged pHRD were sequenced from hybrid #7 and five from hybrid #5 and shown to be identical.
Some of the joints formed in the scid cells were also sequenced. We concentrated on those joints that gave near normal size rearrangement fragments on Southern blots. The smallest deletion was 23 nucleotides, but most were larger than 100 bp ( Fig. 3B and Table 1). In other VJ joints of the pHRD test gene in wild-type pre-B cells (Engler and Storb, 1987) and in B and T cells of transgenic mice (Engler et al., 1991a;and  pko-neo (6.2 kb) FIGURE 1. Maps of rearrangement substrates. (A) pHRD rearrangement test gene (redrawn after Engler et al., 1991b). The individual components are mouse Ig heavy-chain enhancer, mouse metallothionein-1 promoter, 7mer-spacer-9mer recombinase recognition sequences from an Ig Vie region, rat preproinsulin initiation codon and surrounding sequences, 9mer-spacer-7mer recognition sequences from a Jr region, E. coli xanthine-guanine phosphoribosyl transferase coding sequence, mRNA splicing and polyadenylation signals from SV40; the pUC13 vector is not shown. Before rearrangement, this plasmid, when digested with PstI and probed with gpt, results in a 2.6-kb fragment; after rearrangement between the V and recognition sequences, a new PstI fragment of 3.0 kb is seen (these sizes were measured as 2.5 and 2.7 kb before the gpt sequence was available; Engler and Storb, 1987 ). (C) Relationship of the two copies of the rearrangement test gene pHRD and the two copies of the selectable marker gene pko-neo stably cointegrated in the myeloma Ag8HRD7. The arrows indicate the left-to-right orientation of pHRD (see A) and the 5' to 3' direction of the neo genes. Restriction enzymes: E=EcoRI, P=PstI. tional form that can complement the scid defect.
The SCID product is presumably a mRNA/protein. The scid pre-B cells may produce a nonfunctional product that has a lower affinity for the substrate or for other recombinase components and thus does not compete with the wild-type product in V(D)J recombination. Alternatively, they may pro.duce no product due to a nonsense mutation or a large deletion in the scid locus. The correct rearrangement in $33x myeloma hybrids suggests that in pre-B cells of F1 hybrids between scid and normal mice, where correct Ig gene rearrangement is observed (Lieber et al., 1988), the pre-B cells are not selected for suppression of the mutated scid allele.
Most of the scid pre-B cell transfectants show multiple rearrangements of pHRD, some of which can be easily scored as different events because of their different sizes on a Southern blot (Fig. 2B). Others can be distinguished by DNA sequencing (Fig. 3B, #20.1 and 20.2). Thus, in the scid pre-B cells RAG-1 and RAG-2 mRNAs con-290 J. ZHAO AND U. STORB  (Table 1). This low rate is presumably due to an inhibitor of rearrangement that we pre-viously identified in the myeloma cells (Engler et al., 1991b). When Ag8HRD7 or other Ag8 cells stably transfected with pHRD are fused with normal pre-B cells, a few rearrangements also occur early, but most hybrids also do not rearrange the test gene (Engler et al., 1991b). In this previous study, the pre-BxAg8 stable hybrids had lost the   Fig. 2B. The top line is the unrearranged sequence, including the recognition sequence 7mers (lower case letters). The single nucleotides in the center presumably represent N regions. The indicates a changed base (A>G); we have not determined whether this is a Taq error.
expression of the RAG-1 and RAG-2 genes. Similarly, no RAG-1 or RAG-2 mRNA can be detected in the scid pre-Bxmyeloma hybrids, whereas the RAG mRNAs are present in the scid $33 cells before fusion (Table 1) and proteins encoded by them presumably cooperate in V(D)J joining reactions immediately after fusion. In hybrids between scid and normal pre-B cells, RAG-1 and RAG-2 mRNAs continue to be present (not shown), suggesting that the fusion event itself is not responsible for the loss of RAG gene expression in the S33xAg8 hybrids. Preliminary results have suggested that the recombinase inhibitor interferes with transcription of the RAG genes (J. Zhao, P. Roth, and U. Storb, unpublished). Thus, early after fusion, the V(D)J recombinase mRNAs and translated proteins may act for a period of time defined by their halflives. These half-lives have not been determined, but must be relatively short. Because only one type of rearranged joint was found in each of the hybrids ( Fig. 3A; 11 and 5 independent DNA clones, respectively, were found to have the same VJ-joints) and because the intensities of the rearranged and unrearranged pHRD bands on Southern blots are the same (Fig. 2C), it appears that all cells of one hybrid, clone have the same one of the two copies of'pHRD rearranged in the same way. Thus, VJ joining presumably occurred before the first cell division after cell fusion. This means that the V(D)J recombinase provided by the scid pre-B cell and the SCID product provided by the myeloma cell must have assembled soon after nuclear fusion on one of the two copies of pHRD integrated in the Ag8 genome. Further rearrangements were then apparently prevented by the action of the inhibitor. It appears unlikely that the SCID product was induced by the fusion, because the rearrangement seems to have taken place immediately after fusion, before the first cell division.  (Swan et al., 1979) of the Ag8 cell (Figs. 2E and 2F). These Ag8 chromosomes were lost in three hybrids that had lost the inhibitor and were able to rearrange Ig test genes (Engler et al., 1991a).
The inhibitor may thus be encoded by genes on these chromosomes or another chromosome that happened to be lost at the same time in the hybrids that continuously produced V(D)J recombinase (Engler et al., 1991a). If the inhibitor operates on the transcriptional level, the expression of its target genes would continue to be inhibited after fusion. Because mature B cells can complement the scid defect despite the continuous presence of the inhibitor of rearrangement, apparently the SCID gene or its product is not a target of the inhibitor.
While this work was in progress, the finding of a general X-ray hypersensitivity (Fulop and Phillips, 1990;Biederman et al., 1991;Hendrickson et al., 1991) due to a defect in double-strand DNA repair in scid fibroblasts was reported. However, there was some ambiguity as to whether scid pre-B cells have this defect (Weaver and Hendrickson, 1989). The finding of SCID product in mature B cells that lack the V(D)J recombinase further supports the idea that the SCID product may be ubiquitous and that it may function in both DNA repair and Ig gene rearrangement.

DNA Amplification by PCR and Sequencing
The DNAs of (S33xAg8HRD) hybrid cells or of $33 cells with rearranged pHRD bands of nearly correct size were digested with Pstl to eliminate the unrearranged DNA and amplified by PCR using primers PE1 (5'AGACCTCTcTAGAGG-ATCCCcGGGC3') and PE2 (5'CCAGTAACGC-ACCCGGTACCAGACC3'). After electrophoresis of the PCR product in a 2% agarose gel, a band corresponding to about 349 bp (which would be the correctly rearranged size) was isolated. In case of the hybrids between SCID pre-B cells and Ag8, the rearranged DNA was directly cloned into Bluescript IIKS+ (strata-gene) (both the amplified DNA and the vector were cut with BamHI/Sstl). In the case of $33, the isolated rearranged DNA was reamplified by PCR using primers PE1 and OGPT2 (5'CGCTCATGTG-AAGTGTCCCAG3') and the near correctly rearranged band was isolated from an agarose gel and cloned as before. The DNAs were sequenced using the T3 primer of BSK. Some clones were also sequenced in the other direction using a metallothionein primer OPmt21 (5'CCTCACTTACTCCGTAGCTCC3').