The elements which regulate gene expression have traditionally been identified by their effects on reporter genes which have been transfected into cell lines or animals. It is generally assumed that these elements have a comparable role in expression of the corresponding endogenous locus. Nevertheless, several studies of immunoglobulin heavy-chain (IgH) gene expression have reported that the requirements for expressing IgH-derived transgenes differ from the requirements for expression of the endogenous IgH locus. Thus, although expression of transgenes requires multiple elements from the J H -C intron-the E core enhancer, the matrix attachment regions (MARs) which flank E, and several switch-associated elements-B-cell lines in which expression of the endogenous heavy-chain gene is maintained at the normal level in the absence of these intronic elements have occasionally been reported. Gene targeting offers an alternative method for assessing regulatory elements, one in which the role of defined segments of endogenous genes can be evaluated in situ. We have applied this approach to the IgH locus of a hybridoma cell line, generating recombinants which bear predetermined modifications in the functional, endogenous heavy-chain gene. Our analysis indicates the following. Chromatin fibers appear to exist as loops which are anchored to the nuclear matrix, or scaffold (reviewed in reference 25). The loops are thought to correspond to functional domains in which the transcriptional state of one domain is largely independent of its neighbors. It is generally considered that the sites of anchorage in the DNA have a particular affinity for the matrix and can therefore be identified as matrix attachment regions (MARs) or scaffold attachment regions.As illustrated in Fig.

A general method of relating molecular function and structure is to examine the biological and chemical effects of defined mutations. In many cases, particularly those concerned with the rate or efficiency of gene expression, it is important to assess mutations in the normal chromosomal context. There are two methods of obtaining such mutants: (i) site-directed mutagenesis of the chromosomal locus, using homologous recombination to target defined mutations to the gene of interest, and (ii) phenotypic selection of mutant organisms. For most mammalian genes the rarity of targeted recombinants and phenotypically evident mutants impede the use of either of these approaches. However, various genetic and biochemical features render the Ig heavy chain locus in B cell lines amenable to both gene targeting and phenotypic selection of mutants. We describe here a replacement-type vector in which the selectable marker is an enhancerless gpt gene which is particularly suitable for targeting the IgH locus. Deletion of the enhancer greatly decreased the frequency of gpt+ random transformants while still allowing properly targeted transformants to be gpt+, such that transformants with the predicted recombinant structure comprised 25% of the gpt+ population. Thus, the labor involved in mutagenizing the chromosomal locus using this method is comparable to the usual method of isolating randomly inserted transformants, but offers the important advantages that the copy number and integration site are the same in independent transformants. In the hybridoma cell lines which we have tested, the consistent copy number and integration site are sufficient to yield a uniform level of recombinant gene expression.(ABSTRACT TRUNCATED AT 250 WORDS)

The switch regions adjoining the DNA encoding the Ig heavy chain constant regions have been implicated in gene expression as well as isotype switching, in that transgenic mice express switch-containing transgenes at a level 100- to 1000-fold higher than the corresponding switch-deleted transgenes. To test whether the switch region of the natural IgH locus is also required for high level expression we have used homologous recombination to generate targeted recombinant hybridoma cell lines that lack the switch region sequences from the major intron of the mu gene. The expression pattern of these switch knock-out cell lines was compared with that of the parental cell line as well as to that of control recombinants using both steady-state mRNA level and nuclear run-on activity to assess heavy chain gene expression. In striking contrast with the results reported for transgenic animals, we have found only a minimal effect, if any, of deleting the switch element from the natural chromosomal location.

No abstract