| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gene MAR Regulates the Probability of Premature V-J Joining and Somatic Hypermutation1
*
Department of Molecular Biology and Oncology, University of Texas Southwestern Medical Center, Dallas, TX 75235; and
Department of Microbiology and Immunology, University of Michigan Medical School, Ann Arbor, MI 48109
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene contains an evolutionarily conserved nuclear matrix
association region (MAR) adjacent to the intronic enhancer. To test for
the function of this MAR, we created mouse lines with a targeted MAR
deletion. In MAR knockout animals, the immune system was normal in
nearly all respects, including the distributions of various B cell
populations and Ab levels. However, in pro-B cells, enhanced
rearrangement was noted on the MAR- allele in
heterozygotes. In addition, the efficiencies for targeting and
generating somatic mutations were reduced on MAR-deleted alleles. These
results provide evidence for the MAR negatively regulating the
probability of premature rearrangement and positively regulating the
probability of somatic hypermutation. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Previously, we identified a MAR adjacent to the intronic enhancer of
the mouse Ig gene (5). Several lines of evidence suggest that this
gene MAR serves significant biological functions. The juxtaposition
of MARs with transcriptional enhancer elements has been evolutionarily
conserved within the Ig genes of the mouse, rabbit, and human (5, 9, 11). Mouse gene constructs lacking the MAR exhibit lower and
erratic expression in stable ectopic integration experiments, both in
cultured cells and transgenic mice (12, 13, 14). In addition, expression is
much lower in rabbit genes that bear natural mutations in the MAR
(9, 15, 16). Interestingly, the gene MAR has been shown to be
required for triggering demethylation of methylated DNA constructs
stably introduced into cultured cells (17, 18). Furthermore, the
gene MAR has been shown to be necessary for achieving high levels of
somatic hypermutation in rearranged Ig genes introduced into mice (14).
Because nearly all of the functions attributed to the gene MAR have
been derived from the results of experiments employing ectopically
integrated reporter genes, we sought to investigate the function of
this MAR in its natural chromosomal environment by creating a MAR
knockout mouse. This approach also allows addressing the role of the
gene MAR in recombination, which is an open and important issue. In
previous studies, the gene intronic enhancer along with a segment
of the flanking MAR were deleted in embryonic stem (ES) cells by
Cre-LoxP targeting (19). The deletion, studied by the
recombination-activating gene 2-deficient blastocyst complementation
assay, severely affected, but did not fully abolish, gene
rearrangement and expression (19). To study the function of the MAR per
se, while still maintaining an intact intronic enhancer, we have made a
full-length MAR deletion and created mice with only 8 bp of foreign DNA
in place of the MAR. Our analysis reveals unexpectedly that no
significant defect exists in the levels of gene rearrangement, B
cell populations, or Ab production. However, in a fraction of B cells,
V-J joining occurred earlier during development at MAR-deleted
alleles, and somatic hypermutation in germinal centers was reduced. Our
results not only are suggestive of a new MAR function in recombination
timing, but also indicate that only the somatic hypermutation function
previously described in the ectopic integration experiments, and not
transcription level or demethylation, is obeyed in the animal model.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A 4.8-kb XhoI/BamHI fragment
carrying both the pol II neor gene and pgk HSVtk
gene from plasmid pAD72 (gift of Alan Bradley, Baylor College of
Medicine, Houston, TX) was inserted into a 5.8-kb YCp LEU2
yeast/Escherichia coli shuttle vector, termed pRS315 (20),
forming a 10.6-kb plasmid containing selectable markers for yeast,
E. coli, and mammalian cells, termed pLAD. A 9-kb
NotI/BamHI fragment containing the germline mouse
J-C region cloned from an lFIX II genomic library of 129/SV mouse
DNA was inserted into pLAD forming pLADK2. To ensure a full-length
deletion of the MAR, we also deleted flanking sequences surrounding the
core element that had an average adenine and thymine content of 74%.
The region deleted corresponded to a 420-bp sequence from coordinates
3251–3670 bp (21) (GenBank accession number V00777). We utilized the
recombinatorial machinery of yeast to delete the MAR, as follows. Two
gene fragments, corresponding to 154 bp immediately 5' of the
420-bp MAR and 329 bp 3' of this MAR, were generated using PCR from
129/SV DNA; these were joined together using PCR SOEing (combining and
extending overlapping sequences) such that between the two fragments a
unique PmeI restriction site was inserted (22). This was
subcloned into pBluescript and sequenced to assure fidelity of the PCR
reactions. Finally, a URA3 gene flanked by PmeI
site linkers was introduced into the plasmid at the PmeI
site generating pURA3-T. This plasmid was amplified in E.
coli, and the -URA3- fragment was excised with
PvuII. The fragment was then introduced into strain W303-1B
of Saccharomyces cerevisiae carrying the corresponding
pLADK2 plasmid (23). Leu+, Ura+ yeast colonies
were screened by Southern blot analysis to identify one-step gene
replacement integrants that had substituted the URA3 gene
for the 420-bp MAR (24). Four different positive clones were introduced
into E. coli DH5
. The URA3 gene was then
excised from these plasmids using PmeI digestion, followed
by religation. The net result was a precise deletion of the MAR
with the introduction of only an 8-bp PmeI restriction site,
unique to the final constructs, which has been named p
MAR,
confirmed by DNA sequencing of the pertinent regions. The final
construct was linearized at a unique AvrII site and used for
electroporation of mouse ES cells.
Site-directed integration in ES cells
The procedures used have been described in detail elsewhere (25). Briefly, ES cells derived at this institution, termed KG1, were at passage 4 from subcloning. Cells were grown in DMEM supplemented with nonessential amino acids, glutamine, penicillin/streptomycin (Life Technologies, Grand Island, NY), 2-ME (Sigma, St. Louis, MO), and 15% FBS (HyClone, Logan, UT). Ten million ES cells were electroporated with 50 µg DNA at 330 microfarads at 275 V low impedance using a Life Technologies Cell Porator. Cells were plated 24 h after electroporation onto SNL 6-76 feeder layers, which were mitotically inactivated by irradiation @10,000 rad, under G418 selection (300 µg/ml active compound) (Geneticin; Life Technologies), and continued for 9 days. A total of 288 individual clones was picked, trypsinized (0.05% trypsin/EDTA; Life Technologies), and plated onto feeder layers on 24-well plates. After 4 days, clones were trypsinized and stored frozen at -80°C. An aliquot of cells from each well was grown up on gelatinized 24-well plates for DNA preparation. After two hit clones were identified, the cells were thawed and plated onto feeder layers grown in T25 flasks for expansion. Cells were passaged once onto 10-cm feeder layers and grown for 5 days before FIAU [1-(2’-deoxy-2’-fluoro-ß-D-arabinofuranosyl)-5-iodouracil] (0.2 µM) selection was begun and continued for 9 days. After that time, 288 individual clones were picked and treated as before. Five recombinant clones were identified, expanded, and stored frozen before microinjection.
Generation of mouse lines
Approximately 12 cells were microinjected into C57BL/6 blastocysts and reimplanted into the uterus of 2.5-day pseudopregnant Institute for Cancer Research (Harland) embryo transfer recipients (8–15 embryos/mouse). A total of 68 embryos was reimplanted into 5 embryo transfer recipients, and a total of 37 pups was born. Coat color chimerism was assessed on day 7, and male chimeras that were greater than 50% chimeric were bred to C57BL/6 females for germline transmission of the mutation.
Hybridization probes
Probes were created by subcloning sequences from the Ig locus
(26). Probes A-C are shown in Fig. 1
A. The A probe is the
1.1-kb SalI Ig fragment, 5' of J region, from the
plasmid pSPIg8/B(f). The B probe is the 2.8-kb BglII Ig
fragment encompassing J and C region from the plasmid pJC6.8.
The C probe is the 1.2-kb BamHI/BglII Ig
fragment, 3' of C region, from the plasmid pSP64E. The probe in Fig. 2B is the 650-bp HhaI/SacII Ig
fragment, 5' of J region, from plasmid pSP64B. Sequences
complementary to GAPDH were obtained from Ambion (Austin, TX).
|
|
Genomic DNA was isolated from cells after lysis in 0.1 M EDTA,
0.5% N-lauroyl sarcosine, and 100 µg/ml proteinase K with
no less than 3-h incubation at 55°C. Lysates were extracted once with
equilibrated phenol:chloroform:isoamyl alcohol (25:24:1), 1x
chloroform:isoamyl alcohol (24:1), and the DNA was then precipitated
from the aqueous phase by the addition of 2 vol of ethanol. The final
precipitates were washed in 75% ethanol, briefly air dried, and
resuspended in TE (10 mM Tris, 1 mM EDTA, pH 8). Samples were digested
with the indicated restriction enzymes following the manufacturer’s
recommendations and resolved by electrophoresis in 0.95% agarose gels
using 1x TAE (0.04 M Tris-acetate, 1 mM EDTA) as the running buffer.
Resolved DNA samples were transferred to Zeta-Probe GT membranes
(Bio-Rad, Richmond, CA) by capillary transfer using alkaline transfer
buffer (0.4 N NaOH, 0.2 M NaCl). Membranes were then briefly rinsed in
2x SSC and allowed to air dry. DNA was UV cross-linked to the
membranes using a UV Stratalinker (Stratagene, La Jolla, CA) set on
auto cross-link. Prehybridization and hybridization were performed at
65°C with Church-Gilbert buffer (250 mM
Na2HPO4 buffer, pH 7.4, 1 mM EDTA, and 1% w/v
BSA) (27). The membranes were hybridized to
[-32P]dCTP-labeled DNA probes (Pharmacia Oligolabeling
Kit; Piscataway, NJ). Following hybridization, membranes were rinsed
once in 0.5x SSC, 20 mM Na2HPO4 buffer, pH
7.4, and 2% SDS at room temperature, followed by two or three washes
in 0.1x SSC, 20 mM Na2HPO4 buffer, pH 7.4, and
2% SDS at 60–65°C for 20–30 min each. Membranes were imaged and
quantitated using a PhosphorImager and ImageQuant software (Molecular
Dynamics, Sunnyvale, CA). Membranes were stripped for rehybridization
by washing two to three times for 20 min in 0.1x SSC and 0.5% SDS at
95°C.
Cell fractionation
Single cell suspensions were prepared from spleen, bone marrow, or thymus, as described (28). A total of 4 x 107 cells was incubated with an optimal concentration of biotinylated Abs in 100 µl PBS/0.1% BSA for 10 min on ice. After washing with PBS/0.1% BSA, cells were incubated with 20 µl of streptavidin microbeads (Miltenyi Biotec, Auburn, CA) for 30 min at 4°C. After removal of unbound microbeads, the cells were loaded into a mini MACS separation unit (Miltenyi Biotec). Columns were washed three times with 200 µl PBS/0.1% BSA before the positive fraction was eluted by using the supplied plunger with 1 ml of PBS/0.1% BSA after the column was removed from the magnet following the manufacturer’s instructions. During the procedure, aliquots of fractions were stained with streptavidin-PE for FACS analysis to monitor the quality of the fractionation. The fractionated cells were lysed with 0.5% N-lauroyl-sarcosine (Sigma), 100 mM EDTA, and 100 µg/ml proteinase K for genomic DNA isolation.
V-J rearrangement PCR assay
A total of 100 ng genomic DNA samples isolated from
-producing B cells of mice was used as templates for PCR
amplification of products of VJ rearrangement in linear ranges
(data not shown). PCR amplification reactions were performed using
the Expand High Fidelity PCR system (Boehringer Mannheim) supplemented
with 50 nM each oligonucleotide, 3 mM MgCl2, 200 µM each
deoxynucleotide, and 4 U Expand enzyme mix in a 100 µl reaction
volume. PCR cycles were as follows: 15 cycles at 94°C for 1 min,
69°C for 2 min, and 72°C for 3 min, followed by 14 similar cycles,
except for a 10-s automatic extension at 72°C segment in each cycle.
To eliminate heteroduplexes in analyzed reaction products (29), for
last round extension [
-32P]ATP end-labeled V primer
was added to each reaction to an approximate concentration of 10 nM,
followed by one cycle of 94°C for 1 min, 69°C for 2 min, and 72°C
for 3 min, finished by 10-min extension at 72°C. A total of 10 µl
of each reaction was digested with PmeI restriction
endonuclease (New England Biolabs, Beverly, MA). PCR products and
their digestion products were resolved at 4°C on 0.95% agarose
gels with 1x TAE running buffer. Gels were dried onto Whatman blotting
paper with a slab gel dryer (Savant SGD 2000), and subsequently
visualized by PhosphorImager (Molecular Dynamics).
Oligonucleotides used as primers were as follows: 5' primer: V,
GTCCCTGCCAGGTTC/TAGTGGCAGTGGA/GTCT/AA/GGGAC; V2,
GTCAAGTCAGAGCCTCTTAGATAGTGATGGAAAG; V21G,
GAGCCAGTGAAAGTGTTGATAGTTATGGCAATAG; 3' primer: R3-1,
CAGACCCTGGTCTAATGGTTTGTAACCACATGGG.
Flow-cytometric analysis
Single cell suspensions were prepared from bone marrow and
spleen, as described above. Cells (106) were stained with
an optimal concentration of monoclonal fluorescence-conjugated or
biotinylated Abs in 120 µl PBS/1% FCS for 30 min on ice. The biotin
conjugates were further stained with fluorescence-conjugated
streptavidin. After washing with 10 ml of PBS/1% FCS, cells were
resuspended in 400 µl of PBS plus 1% paraformaldehyde and analyzed
with a CellQuest program on a FACScan (Becton Dickinson, Mountain View,
CA). Only cells residing in the lymphocyte gate as defined by light
scattering were analyzed quantitatively. R-PE-conjugated anti-B220,
FITC-conjugated anti-Ig
1 and Ig2, biotin-conjugated
anti-CD43, and FITC-conjugated anti-CD25 were purchased from
PharMingen (San Diego, CA). FITC-conjugated anti-Ig was
purchased from Southern Biotechnology (Birmingham, AL).
RNA analysis
Total RNA was isolated from mouse spleens or bone marrow by
extraction with RNA STAT-60 isolation reagent (Tel-Test, Friendswood,
TX), according to the manufacturer’s instructions, except that the RNA
samples were solubilized in diethyl pyrocarbonate-treated
solubilization buffer consisting of 0.1% SDS, 25 mM
Na2HPO4 (pH 7.4), and 0.5 mM EDTA. Northern
blots were performed using 10 µg RNA from each spleen sample, as
described (30), except that the membranes were prehybridized,
hybridized, washed, and imaged, as described above. For RT-PCR, 1 µg
total RNA from bone marrow was reverse transcribed with Superscript II
reverse transcriptase (Life Technologies, Gaithersburg, MD) at 42°C
for 50 min with 250 ng random primers (Life Technologies) in a 20 µl
reaction mixture according to the manufacturer’s recommendations.
After cDNA synthesis reactions were diluted by the addition of an equal
volume of distilled H2O, PCR amplification of spliced
germline transcripts and porphobilinogen deaminase (PBGD) gene
transcripts as an internal control were performed as described
previously (31). The PCR amplification conditions of primary germline
transcripts were essentially the same as that of the spliced one,
except that the primers used were: 5' primer,
GAGGGGGTTAAGCTTTCGCCTACCCAC; 3' (J1) primer,
CTGTATCTTTGCCTTGGAGAGTGCCAGAATCTGG.
Cell sorting and PCR-based Southern blotting
Starting with 2 x 108 bone marrow single cells
from four to five MAR+/- littermates, IgM-
B220+ cells were obtained by cell fractionation using mini
MACS separation units, according to manufacturer’s instructions.
Briefly, the IgM+ population was eliminated by reacting
cells with biotinylated anti-IgM Abs (PharMingen) and subsequently
with streptavidin microbeads (Miltenyi Biotec), followed by passing the
cells through mini MACS separation units. Similarly, the
IgM- population was reacted with biotinylated
anti-B220 Abs (PharMingen) and streptavidin microbeads, followed by
passing the cells through mini MACS separation units. The positive
IgM- B220+ cell fractions were stained with
FITC-conjugated anti-B220 Abs and PE-conjugated anti-CD43 Abs
(PharMingen). The B220+ CD43- and
B220+ CD43+ cells present in the lymphocyte
gate were analyzed and sorted using a dual laser flow cytometer
(FACStarPlus; Becton Dickinson). These populations were
found to be more than 96% pure (data not shown). The cells were lysed
with a buffer consisting of 10 mM Tris (pH 8), 2.5 mM
MgCl2, 50 mM KCl, 200 µg/ml gelatin, 0.45% Nonidet P-40,
0.45% Tween-20, and 60 mg/ml proteinase K (32). The lysates were
directly subjected to PCR reactions similar to the V-J
rearrangement PCR assay with V and R3-1 primers. One-fifth of the
PCR products were digested with PmeI before gel
electrophoresis and Southern blot-hybridization analysis with the probe
in Fig. 2
B.
Somatic hypermutation analyses
B200+ PNAhigh and B220+
PNAlow B cells were isolated by flow cytometry from
Peyer’s patches, and DNA was obtained from cell pellets, as described
elsewhere (33, 34). DNA segments from rearranged genes were
amplified with the Expand Long Template PCR system (Boehringer
Mannheim, Indianapolis, IN) using a degenerate V primer (32) and a
3' primer located approximately 200 bp downstream of the MAR deletion
(AACAATAGAATTATGAGCAGCC). The resulting size differences of the PCR
products allowed us to verify independently that DNA was derived from
either wild-type or MAR knockout mice. PCR products were cloned by the
TA Cloning Kit (Invitrogen, San Diego, CA). J5 clones were
identified and sequenced by use of a degenerate V primer, as
reported previously (34).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We adopted the "HIT and RUN" procedure of Hasty et al.
(35) to create embryonic stem cell lines that possessed a targeted MAR
deletion for subsequent blastocyst fusions (Fig. 1). To generate a MAR deletion in the
targeting construct, we utilized the recombination machinery in yeast
to create a site-directed 420-bp MAR deletion in a 9-kb gene
insert, leaving only 8 bp of foreign DNA as a PmeI site in
place of the deletion (see Materials and Methods). After
linearization of the targeting construct at a single AvrII
site to direct the position of site-directed integration (Fig. 1A), a targeted locus integrant (HIT) results in a
partial tandem duplication (Fig. 1B), which upon loop-out
recombination (RUN) can give rise to retention of the desired mutation
(Fig. 1C). We found that the frequency of HIT clones was
1/25, whereas that of RUN clones retaining the desired mutation was
about 1/500. Two independent RUN clones were isolated and successfully
used to generate founder mice lines that germline transmitted the MAR
deletion in a Mendelian fashion upon breeding (data not shown), first
creating heterozygotes (M+/-) and then subsequent
homozygotes (M-/-). The results of extensive genomic
Southern analyses using probes external (probe A) or internal (probe B)
to the targeting construct substantiate the integrity of targeted loci
at each step of these experimental manipulations (Fig. 1, D
and E, and data not shown).
The MAR deletion does not affect the efficiency, quality, or tissue specificity of gene rearrangement or DNA methylation status
To assess allele usage for productive V-J joining in
MAR+/- heterozygotes, we isolated -producing B
lymphocytes from control and heterozygote splenic cells utilizing
anti- biotinylated Abs and streptavidin-coated magnetic beads.
FACS analysis revealed that the +-bound cell population
was >95% pure (Fig. 2A). To
evaluate allele usage in the + cell population, we
performed Southern analysis to specifically detect the unrearranged
3.5-kb DraI fragment of the wild-type allele and the
corresponding 3.1-kb DraI fragment of the MAR-
allele in heterozygotes; because the 5' DraI site is
upstream of J1 and lost after V-J joining (Fig. 2B),
rearranged genes only create a diffuse background surrounding the
germline unrearranged bands. We found that the ratio of these germline
bands appeared very similar (Fig. 2C, lane 4).
Furthermore, by PhosphorImager analysis, we found that from 70–80% of
these germline bands were accessible to cleavage by the
methylation-sensitive enzymes HhaI and AvaI (Fig. 2C, lanes 5 and 6), in contrast to
mouse tail DNA, which was more fully methylated at these index
restriction sites (Fig. 2C, lanes 11 and
12). The observed hypomethylation of unrearranged genes in
+ B cells is unexpected based on another report (36);
possibly this discrepancy is because of the different genetic
backgrounds of the mouse strains studied. In conclusion, both
MAR+ and MAR- alleles are equally
hypomethylated and used in productive V-J joining events.
To more carefully evaluate potential preferences between wild-type and
mutant alleles for V or J region selections during V-J joining
events, we utilized a modification of the PCR assay of Schlissel and
Baltimore (32) to assay for gene rearrangement. Use of a degenerate V
region primer (VD) along with a primer downstream of the
MAR deletion results in eight different bands as the combined PCR
products from the MAR+ and MAR- alleles of
splenic DNA from heterozygotes, because V regions can be joined to four
different functional J regions (J1, J2, J4, or J5) for each
allele (Fig. 3A shows
rearrangements only to J1). These products, whose initial lengths range
from 3.4–1.7 kb, can be most readily size fractionated after
PmeI digestion to preferentially shift the size of the
MAR- allele’s PCR products (before PmeI
digestion MVJ1 and MVJ4 have lengths close to WTVJ2 and WTVJ5,
respectively). While this assay may be subject to length effects on the
efficiencies of PCR amplifications, and hence not useful for precise
estimates of J region usage, quantitative comparisons between
MAR+ and MAR- alleles for usage of a specific
J region are valid, because differences in initial amplified product
lengths between these alleles are only from 12–20%. As shown in Fig. 3B (lane 4), for each specific J region
utilized in recombination in + B cells, the ratios of
PCR products from MAR+ and MAR- alleles are
very similar, indicating that the efficiency for V-J joining and choice
of J regions are similar between the alleles. Similar results were
obtained using a primer for the most distal V region, V2(70/3)
(Fig. 3C), which resides about 3.5 Mb from J1 (37, 38),
or a primer for the most proximal V region, V21G (Fig. 3D), residing 18 kb from J1 (George-Raizen et al., our
unpublished results). We conclude that the efficiency and quality of
gene rearrangement are not affected by the MAR deletion.
|
locus is not rearranged in T cells of normal mice, although
sometimes corresponding transgenes disobey this tissue-specificity rule
(39, 40, 41). To determine whether the MAR deletion led to gene
rearrangement in T cells, we utilized anti-CD3
biotinylated Abs
and streptavidin-coated magnetic beads to isolate T cells from thymus
tissues of homozygous wild-type or MAR- animals and
heterozygotes. FACS analysis revealed that the bound cell population
consisted of CD3low and CD3high subpopulations
that were >95% pure (Fig. 4A). Rearrangement PCR assays
on the total bound cell population revealed no significant V-J
joining in these T cells (Fig. 4B, lanes 1–6).
Furthermore, an analysis of methylation status revealed that unlike
+ B cells (Fig. 2C), index HhaI
and AvaI sites in the locus are heavily methylated in T
cell DNA, independent of the presence of the MAR (Fig. 4C).
In conclusion, deletion of the MAR did not abolish tissue specificity
in rearrangement or methylation status.
|
We used FACS to assay for the distributions of B cell
subpopulations in relevant spleen and bone marrow samples. As shown in
Fig. 5, similar proportions of cells from
wild-type, heterozygous, and homozygous MAR- animals
expressed similar levels of B220 and (upper panels), or
B220 and (middle panels) on their cell surfaces. In
addition, analysis of early developing cell populations by B220 and
either CD25 or CD43 markers (42) again revealed no significant
differences between animals (Fig. 5, lower panels). We
conclude that the MAR deletion has little effect on the development of
B cell populations or on the expression levels of relevant surface Ags.
|
mRNA transcripts or secreted Abs
To determine the consequences of the MAR deletion on gene
transcription, we assayed for germline transcripts by RT-PCR and for
mature mRNA by Northern analysis in bone marrow and splenic RNA
samples, respectively, from homozygous wild-type or MAR-
animals and heterozygotes. As shown in Fig. 6, there are no significant differences
in the levels of these transcripts between these animal populations. We
also used ELISA to assay for the serum levels of IgM and IgG in
relevant samples. Using the Mann-Whitney rank order test, we found no
statistically significant differences between animal groups
(n = 8/group); [ light chain-bearing IgM as µg/ml
(152 ± 60; 198 ± 143; 114 ± 51) and IgG as µg/ml
(495 ± 155; 555 ± 258; 350 ± 152)] were found in
sera from wild-type, heterozygous, and homozygous MAR-
littermates, respectively. In addition, total IgM or light
chain-containing IgM levels were also similar among these animal
populations (data not shown). We conclude that the MAR does not
quantitatively regulate RNA levels or secreted Ab production.
|
The experiments described above examined the later stages of B
cell development and did not reveal a phenotype for the MAR deletion.
To determine whether the MAR deletion affected earlier events, we first
assayed for the presence of N regions in V-J joints. Previous
studies have shown that TdT is expressed during the pro-B stage of
development when heavy chain gene segments undergo rearrangement, and
that the enzyme is involved in inserting nucleotides (N regions)
between V-D-J junctions (43, 44). When genes undergo rearrangement
later during B cell development, TdT activity is low and N regions at
V-J junctions are rare (41, 45, 46, 47). To determine whether
differences existed in N regions for V2(70/3)-J1 junctions, we
cloned and sequenced the PCR products amplified from bone marrow DNA
samples of wild-type and MAR- homozygotes. However, the
results of this analysis only revealed minor differences between these
allelic sources (data not shown).
To more carefully evaluate the role of the MAR in regulating the
developmental timing of V-J joining, we used the combination of
affinity chromatography and flow cytometry sorting to isolate
CD43+, IgM-, B220+ pro-B cells
(42) from bone marrow cells of MAR+/- heterozygotes for
gene rearrangement PCR assays, followed by Southern blot analysis.
PhosphorImager analysis of the data shown in Fig. 7 (lane 1) reveals that
the MAR- allele from CD43+ sorted cells was
preferred almost 6-fold for usage in V-J joining over the
wild-type allele, and only MAR-deleted alleles were used for J1 and
J2 rearrangement, whereas similar usage between alleles was observed
in CD43- sorted cells, total bone marrow, and spleen cells
(Fig. 7, lanes 2–4). Length effects on PCR amplification
efficiencies are unlikely explanations for these observed differences.
The corresponding J5-rearranged PCR-amplified products from
wild-type and MAR-deleted alleles only have a 20% length difference
before PmeI digestion (2.1 and 1.7 kb, respectively), and
even the longer PCR products of J1 (3 kb) and J2 (2.6 kb)
rearrangement derived from MAR- alleles were detected in a
comparable intensity with J5 (2.1-kb) wild-type rearranged products.
Whether prematurely rearranged Ig alleles express Ig proteins
could be addressed by assaying for the presence of the protein in the
cytoplasm, since only 6% of the heavy chain genes have undergone V(D)J
rearrangement in this cell population (42), but previous studies have
shown that rearranged Ig genes that have been introduced into the
endogenous Ig locus in transgenic animals are poorly expressed in
pro-B cells (48). In conclusion, we have evidence that MAR-deleted
alleles are used earlier than wild type for V-J joining during B
cell development.
|
Because previous studies on rearranged transgenes have suggested a
requirement for the Ig MAR in increasing the proportion of B cells
undergoing somatic hypermutation (14), we investigated this issue in
our knockout mice. We isolated germinal center B cells from Peyer’s
patches of three wild-type and three homozygous knockout mice by flow
cytometry sorting of B220+ PNAhigh cells (33).
After PCR amplification, cloning, and screening, we selected for
analysis about a 180-bp region immediately 3' of V-J5
recombination products. We have shown previously that this region is as
highly mutated as the V-J region itself (49, 50). We found that
the percentage of alleles exhibiting no mutations was 48% for
MAR-deleted alleles, but only 17% for the wild-type alleles (Fig. 8). If one averages the overall mutation
frequencies for these animals, the 27.6 mutations/kb observed for
wild-type alleles are in excellent agreement with published results for
both V(D)J transgenes and endogenous sequences (34, 49, 50, 51, 52). The
overall mutation frequency of 11.3 mutations/kb for MAR knockout
animals is only 40% of that exhibited by wild type (Fig. 8). Similar
reductions were also observed when data from individual
MAR-/- and wild-type mice were compared. Statistical
analysis of the pooled data reveals that this difference is significant
(p < 0.01). However, among clones bearing
mutations, the range of mutations (2–14 versus 1–11) and extent of
transitions versus transversions (0.9 versus 1.3) were similar for
wild-type and knockout mice, respectively. Among clones bearing
mutations, the mutation frequency was significantly altered (34.1
mutations/kb versus 20.8 mutations/kb) for wild-type and knockout
animals, respectively (p < 0.01). The
reason for this difference is that 7 of 31 clones from
MAR- alleles exhibited only one mutation (Fig. 8). In
conclusion, we have evidence that the MAR is necessary for ensuring
that a high proportion of alleles receives mutations as well as
determining that significant numbers of mutations are generated once
mutations begin to be introduced.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Ectopic integration experiments have demonstrated
functions for the mouse Ig MAR in mediating demethylation (17, 18),
yet we found that germline alleles possessing or lacking the MAR
were equally undermethylated in -producing B cells, or equally
hypermethylated in thymus and tail DNA. In addition, other ectopic
integration experiments have revealed that the MAR is required for high
level expression of mRNA levels (12, 13, 14), yet we demonstrate in
this work no differences in the levels of various RNA transcripts or
sera Abs arising from the expression of alleles possessing or lacking
this MAR. Taken together, redundant elements present in the normal
locus must compensate for the MAR requirement observed in these ectopic
integration experiments. Although we have demonstrated previously
that only a single major MAR exists within a 16-kb EcoRI
fragment encompassing the germline Ig gene sequence (5), it is
possible that weaker, secondary MARs known to be localized nearby or
stronger unidentified MARs located further away participate in such
compensation. A lack of concordance has also been noted between
results for deletion of the mouse Ig gene 3' enhancer from ectopic
constructs versus the native locus (53).
The MAR does not regulate the tissue specificity of recombination
A rigorous test for maintenance of tissue specificity in
recombination is to assay for V-J joining in T cells. The same
recombinase machinery is involved in TCR rearrangement, but V-J
joining does not occur in normal mouse T cells (54). Because certain
germline Ig transgenes have been observed to undergo rearrangement
in T cells (39, 41, 55), it can be concluded that elements were missing
from these constructs that specified tissue specificity to the
rearrangement process, and that V-J recombination signal
sequences can be recognized in T cells under certain conditions.
However, our present study demonstrates that V-J joining does not
occur in T cells after deleting the MAR. Thus, our evidence indicates
that the MAR does not confer tissue specificity to the rearrangement
process.
The Ig gene intronic enhancer is a pivotal positive
regulatory element for V-J joining
Previous studies in which the mouse Ig gene intronic enhancer
and a portion of the MAR were deleted from the native germline locus
revealed a marked alteration in the population of splenic B cells
expressing Ig on the cell surface (19). This observation, together
with our results, in which we fully deleted the MAR but left the
intronic enhancer intact and saw no effect on splenic or bone marrow B
lymphocyte lineage cell populations, point to the importance of the
intronic enhancer as a pivotal element in Ig regulation.
The MAR deletion of 420 bp probably does not generate the observed
phenotypes simply because the intronic enhancer is moved closer to J
regions
Deletion mutations have the caveat of altering spacing distances,
but unfortunately, neutral DNA sequence controls are not well
developed for creating large substitution mutations while
maintaining spacing. Nevertheless, it is unlikely that moving the
enhancer 12–20% closer to J regions contributes to our observed
phenotypes, which exhibit opposing effects with respect to somatic
hypermutation and gene rearrangement (discussed below). These spacing
changes are far less than those that normally result from V
rearrangement in the wild-type locus to different J regions. In
addition, it has been shown that the deletion of a 7.34-kb fragment
between C and the 3' enhancer in a transgene did not affect
somatic hypermutation (51). Furthermore, J1 is normally the
preferred substrate for rearrangement (56), yet this element resides
the furthest away from the intronic enhancer.
The Ig gene MAR is required for efficient targeting and
production of somatic hypermutations
There is evidence for the importance of the Ig gene MAR in
somatic hypermutation from transgenic mice experiments (14). Deletion
of the MAR was found to markedly reduce the proportion of transgenes
that underwent detectable somatic hypermutations, and for those that
did exhibit mutations, the mutation frequency was about 50% of that
exhibited by MAR-containing constructs. Our results for MAR knockout
animals are similar in that the probability for observing any mutations
in the Ig gene locus was reduced after deleting the MAR, and we also
found that the efficiency for generating mutations was also reduced in
MAR knockout mice. Although we still see mutations arising in alleles
lacking the Ig gene MAR, transgenic mice experiments have already
shown that multiple elements are required for high efficiency
somatic hypermutation (14, 51, 57). While it is possible that the
MAR optimizes the timing of entry of B cells into germinal centers with
respect to mutual expression of mutator components (58, 59), the MAR
may increase the efficiency of targeting and action of the mutator
apparatus by altering chromatin accessibility (14, 60, 61).
Evidence that the MAR negatively regulates V-J joining
During B cell development, the heavy chain locus normally
rearranges before the Ig locus, allowing µ-chains to be evaluated
for proper pairing with surrogate light chains, thereby shaping the
VH repertoire (62). We observed a 6-fold preferential usage
of the MAR- allele for V-J joining in flow
cytometry-sorted pro-B cells. Although it is often commonly believed
that the heavy chain gene locus must undergo productive V-D-J joining
before recombination can occur at the Ig locus, studies with normal
and knockout mice have shown that an alternative order for
recombination occurs in about 3% of the V-J joints in bone
marrow (63), consistent with our observation that the wild-type allele
in pro-B cells also exhibited detectable recombination. The Ig gene
intronic and 3' enhancers are required for efficient rearrangement of
the locus (19, 64). The MAR, which resides between these enhancers and
the J regions, may act to insulate, or silence enhancer activities
in pro-B cells, by analogy to the negative regulatory functions of MARs
in other systems (65, 66, 67, 68, 69). We have evidence from experiments in a
cultured pre-B cell line, in which a MAR deletion was targeted in a
tandem integrant, that also supports a role for the MAR in negative
regulation of V-J joining during early B cell development (70).
Interestingly, in both these cell and animal experiments, we have found
preferential usage of J5, which we have interpreted to reflect
repeated recombination on the MAR- hyperrecombinogenic
allele (70). In the future, creation of immortalized B cell lines
arrested in early development from the MAR knockout mice might allow a
more definitive examination of this regulation. In addition,
quantitative assessment of this premature rearrangement phenotype in
animals may be possible by single cell PCR; however, the length of the
amplification products required for allele comparisons is beyond our
current technical ability.
In conclusion, our results support the idea that MARs play divergent roles in gene regulation dependent upon chromatin context and cell developmental stages, and offer alternative models for future mechanistic studies.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. William T. Garrard, Department of Molecular Biology and Oncology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75235-9148. E-mail address:
3 Abbreviations used in this paper: MAR, matrix association region; ES, embryonic stem.
Received for publication November 9, 1998. Accepted for publication February 26, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
immunoglobulin gene occurs next to the enhancer in a region containing topoisomerase II sites. Cell 44:273.[Medline]
immunoglobulin gene connecting the J and C regions reveals a matrix association region (MAR) next to the enhancer. Nucleic Acids Res. 20:4929. gene expression after stable integration I: role of the intronic MAR and enhancer in plasmacytoma cells. J. Biol. Chem. 264:21183. gene expression after stable integration: role of the intronic MAR and enhancer in transgenic mice. J. Biol. Chem. 264:21190. transgenes show clonal recruitment of hypermutation: a role for both the MAR and the enhancers. EMBO J. 16:3987.[Medline]
light chain locus in the rabbit: evidence for differential gene conversion events. Proc. Natl. Acad. Sci. USA 83:5180.2 J-C locus reveals multiple deletions. Nucleic Acids Res. 11:8877. chain enhancer sequence. Cell 76:913.[Medline]
B in B-cell-specific demethylation of the Ig locus. Nat. Genet. 13:435.[Medline]
light chain intronic enhancer/matrix attachment region impairs but does not abolish V-J rearrangement. Immunity 4:377.[Medline]
immunoglobulin J and C region genes. J. Biol. Chem. 256:5116. gene locus. J. Biol. Chem. 261:3838. immunoglobulin light-chain locus: NF-B-dependent and independent pathways of activation. Mol. Cell. Biol. 17:3477.[Abstract]
gene rearrangement correlates with induction of germline gene transcription. Cell 58:1001.[Medline]
light-chain transgenes. Proc. Natl. Acad. Sci. USA 90:9862. chain monoallelic demethylation and the establishment of allelic exclusion. Genes Dev. 12:1801. locus. Proc. Natl. Acad. Sci. USA 92:12421. light chain locus. Immunogenetics 45:180.[Medline]
light chain gene in transgenic mice. Proc. Natl. Acad. Sci. USA 84:4229.-J joining. Cell 83:1113.[Medline]
light chain rearrangement in mouse fetal liver. J. Immunol. 153:1150.[Abstract]
light chains. J. Immunol. 152:3467.[Abstract]
transgene, but not a VJ gene segment targeted into the Ig locus, can 
the
error rate of the polymerases (<1 mutation/kb).