An element in the endogenous IgH locus stimulates gene targeting in hybridoma
cells
An element in the endogenous IgH locus stimulates gene targeting in hybridoma cells
Alla
Buzina*
and
Marc J.
Shulman
1
Departments of Immunology and
1
Molecular and Medical Genetics, Medical Sciences Building, University of
Toronto,
Toronto
M5S 1A8,
Canada
Received December 4, 1995;
Revised and Accepted February 29, 1996
ABSTRACT
Gene targeting of the immunoglobulin (Ig) heavy chain locus is the basis of
improved methods of investigating gene expression and of antibody engineering.
The VH-C
[mu]
intron is a convenient region for mediating homologous recombination events
which result in production of Ig bearing an altered heavy chain. Also, this
segment includes several elements which are important for gene expression,
replication and isotype switching: in some cases it will be advantageous to
alter these processes by modifying this intron. Considering that multiple
targeting steps might be needed to accomplish all the requisite changes, it is
important to know whether any of the anticipated modifications also alter the
recombinogenicity of the IgH locus. To test this possibility we have measured
the frequency at which a mutation in the C
[mu]
3 exon of the endogenous
[mu]
gene is corrected by homologous recombination with a transfected segment of C
[mu]
DNA. Comparison of recombination frequencies in several engineered hybridomas
indicates that deletion of a 7.1 kb segment from the VH-C
[mu]
intron depresses recombination by ~10-fold.
INTRODUCTION
It has long been observed that genetic recombination does not occur uniformly in
the genome. Numerous elements of different types and from widely divergent
organisms have been found to stimulate homologous recombination in nearby
genetic intervals. In some cases the molecular basis of such differences in
recombinogenicity are understood. Thus in
Escherichia coli
the
chi
sequence (GCTGGTGG) alters the activity of the RecBC(D) protein to reduce
exonuclease and enhance helicase activities, leading to an increased frequency
of homologous recombination in nearby intervals (
1
). The M26 site in
Schizosaccharomyces pombe
(ATGACGT) binds a specific heteromeric protein (
2
) and appears to be a preferred site of DNA cleavage (
3
). Homologous recombination is also stimulated in regions which are near an
origin of replication, perhaps because the nicks or single-stranded segments which are associated with replication initiate
recombination (
4
). Also, transcription has been seen to increase homologous recombination in
yeast (
5
), in mammalian cells (
6
,
7
) and inconsistently in bacteria (reviewed in
8
). The mechanism linking transcription and recombination is unknown, but it has
been proposed that transcription might render the DNA more accessible to
recombinases or that topoisomerase cleavages entrained by the unwinding
associated with transcription might create structures which promote
recombination.
Gene targeting of the immunoglobulin loci has been studied intensely, both
because this technique permits highly controlled analysis of gene function and
because modified immunoglobulins have diverse practical and medical
applications. The immunoglobulin heavy chain (IgH) locus contains several
elements which function in transcription, replication and DNA rearrangement. In
constructing recombinant hybridoma cells to analyse the role in gene expression
of elements lying in the intron between the VH and C[mu] exons it became important to know whether these elements also affect the
efficiency of gene targeting of the IgH locus. Here we report that deletion of
a major part of this intron depresses recombination in an adjoining interval by
~10-fold.
MATERIALS AND METHODS
Tissue culture
The Sp6 hybridoma cell line and its normal (Sp6/HL) and mutant (igm482 and
igm692) subclones have been described (
9
), as have the derivation of the recombinants 692R1 from igm 692 and 482R1 from
igm482 (
10
).
Construction of plasmids and targeted recombinants
Figure
1
shows construction of the vector pI[Delta]C[mu]482, which bears the frameshift mutation in C[mu]3 and the 7.1 kb intron deletion and was used to generate the I
+
and I[Delta] recombinant hybridomas. An analogous vector with a normal C[mu]3 exon was used to construct the ER50 cell line (A.E.Oancea and
M.J.Shulman, manuscript in preparation). The pC[mu](
Acc
I) plasmid was prepared by inserting the 2.4 kb
Acc
I fragment bearing the C[mu] exons into pTZ18. To control for efficiency of transfection we measured the
frequency of puromycin-resistant transformants conferred by pPur/sqs, a vector derived from
pBABEDpCMVpuro by A.Cochrane.
Measurement of recombination frequency
To measure the frequency of recombination a mixture of 70 [mu]g plasmid pC[mu](
Acc
I) DNA and 2 [mu]g pPur/sqs was transfected into 4 * 10
7
cells, grown as indicated for construction of targeted recombinants. Cells were
returned to normal medium for 2 days, after which they were plated at limiting
dilution in selective medium containing 10 [mu]g/ml puromycin to assess transfection efficiency. On day 7, as well as
several weeks thereafter, the transfected cells were plated to measure plaque-forming cells (PFC), and thus recombinants, as described (
14
).
Analysis of DNA structure
PCR analysis was performed using
Taq
DNA polymerase (Boehringer Mannheim) according to the following protocol for 30
cycles: denaturation, 1 min at 94oC; re-annealing, 2 min at 65oC; extension, 3 min at 72oC, which was increased by 3 s/cycle. Oligonucleotide
primers: 1, 5'-TTACCTGGGTCTATGGCAGT-3'; 2, 5'-GTCACTGTAAATGCTTCGGG-3'; 3, 5'-GGGCACATGCAGATCTCTGTTTTTGC-3' (
3
).
Genomic DNA for Southern blot analyses was prepared by the SDS/proteinase K
method (
15
). DNA (7 [mu]g) digested with
Bam
HI or
Bcl
I was electrophoresed in 0.8% agarose, transferred to Hybond N membrane
(Amersham) and hybridized with the
32
P-probes radiolabeled by random priming.
RESULTS
The system which we used to measure recombination frequency is based on the
hybridoma Sp6, which secretes IgM([kappa]) specific for trinitrophenyl (TNP) and forms plaques on TNP-coupled erythrocytes (
14
). This cell line bears a single copy of the [mu] heavy chain gene (
16
). To prepare cell lines with a convenient genetic marker for measuring
recombination frequency we introduced a 2 bp deletion into the C[mu]3 exon of the endogenous [mu] gene of the hybridoma cell line, thus causing production of a truncated
[mu] heavy chain (Fig.
1
). Because the resulting IgM is non-cytolytic, the recombinant hybridomas do not form plaques. Recombination
between a transfected normal C[mu] segment and the mutant endogenous [mu] gene can restore normal IgM production (Fig.
2
). Our previous analyses indicate that most if not all PFC arise by homologous
recombination (
11
,
14
), so recombination frequency can be measured by assaying the number of PFC. To
assess the importance of the VH-C[mu] intron segment we constructed cell lines in which the 2 bp deletion
was
in cis
with either a normal or a truncated intron and then measured the frequency of
PFC which arise after transfection of the normal C[mu] segment, as described below. By using a C[mu] segment which lies entirely outside the intron deletion, each cell line
will present the same target for homologous recombination and any differences
in recombination can then be ascribed to differences in recombination
efficiency.
DISCUSSION
Genetic recombination depends on both enzymatic factors and
cis
-acting elements. Because the cell lines used to measure recombination bear
only a single copy of the [mu] gene, the increased recombination which we observe to be associated with
the intact VH-C[mu] intron could in principle reflect an IgH-derived
trans
-acting factor, RNA or protein, as well as the action of a
cis
-acting element. However, only two transcripts have been detected for this
region, that which yields mRNA for the [mu] heavy chain and the apparently non-coding I[mu] transcript, which intiates in the vicinity of the E[mu] enhancer (
18
). It is therefore unlikely that this region encodes a protein which promotes
recombination.
As summarized in the Introduction, homologous recombination in other systems has
been found to be stimulated by transcription and replication, as well as by
seemingly recombination-specific elements. Neither the
chi
nor M26 sequences occur in the sequenced part of the deleted region. However,
the 7.1 kb deletion examined here has removed other elements of interest.
First, the deletion removed a segment which appears to function as a
replication origin in B cell lines (
19
). Second, the deletion removed the switch region (S[mu]), which is the preferred site for the rearrangements which cause isotype
switching (
20
). The switch region also promotes rearrangements involving nearby non-switch DNA (
21
). Third, the deleted interval contains several elements which greatly stimulate
transgene expression: the E[mu] enhancer (
22
), the flanking matrix attachment regions (MARs) (
23
), S[mu] and two other switch-associated elements, RegA and RegS (
24
,
25
).
It is still unclear whether these elements function in the recipient hybridoma
cell line and, if so, whether their identified activities contribute to the
higher recombination frequency associated with the intact intron. Inasmuch as
the E[mu]-associated replication origin was observed in B cells but not in
fibroblasts, we suppose that it functions in the I
+
recipient cell lines. Switch recombination in these hybridomas occurs very
rarely, if at all (
26
,
27
), and studies with artificial switch substrates suggest that similar cell lines
lack the switch recombinases (
28
). Nevertheless, the switch region in a closely related hybridoma cell line is a
preferred site for insertion of transfected DNA (
29
). Inasmuch as normal switch recombination occurs in adjoining non-switch DNA, we consider it possible that the switch region might also
stimulate homologous recombination in an adjoining interval.
The functional importance of the `transcriptional' elements in the endogenous
IgH locus is uncertain. Previous studies on myeloma and hybridoma cells imply
that deletion of E[mu], the MARs and the switch-associated elements does not depress expression of the natural IgH
locus in mature or pre-B cell lines (
17
,
29
-
33
). In the case of the particular hybridoma used in our present study the
deletion depressed transcription ~4-fold, an effect which correlates with the loss of MAR content
(A.E.Oancea and M.J.Shulman, manuscript in preparation). Further experiments
will be needed to ascertain which, if any, of these defined elements
contributes to recombination and, if so, whether their importance relates to a
role in transcription or in some other process. It is also possible that the
deletion did not remove a specific element which contributes to recombination.
For example, the deletion might have changed the interval which separates the C[mu] exons from an important stimulatory or inhibitory recombination element.
Our results also have practical implications for genetic engineering. In some
cases it will be advantageous to construct recombinant loci in multiple steps.
For example, it might be important that the modified loci lack the drug
resistance genes commonly used to enrich for targeted recombinants. In
principle, several methods using either homologous (hit-and-run or bait-and-switch) or site-specific (lox/Cre or frt/Flp) recombination can be
used for this purpose in constructing recombinant immunoglobulin loci (
10
,
16
,
34
). Because recombination frequency can be affected by the changes introduced in
a previous recombination step, the order of operations, as well as the
structure of the intermediate recombinants, can be important for success.
ACKNOWLEDGEMENT
This work was supported by grants from the Medical Research Council of Canada.
