| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Institute of Virology and Immunobiology, University of Würzburg, Würzburg, Germany;
Lilly GmbH, Bad Homburg, Germany;
Center for Genomics Research, Karolinska Institute, Stockholm, Sweden;
§
National Veterinary and Food Research Institute, Regional Laboratory of Kuopio, Kuopio, Finland; and
¶
Department of Clinical Microbiology, University of Kuopio, and Department of Pediatrics, Kuopio University Hospital, Kuopio, Finland
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Based on the assumption that the V(D)J recombination machinery is active throughout early lymphocyte development, the availability of gene segments to the recombinase in various Ig loci, a concept referred to as accessibility, might be a key factor in controlling the rearrangement process (5, 6). Previous studies have suggested that regulation of rearrangement is affected by cis-acting DNA elements that target the recombinase activity to the gene segments. Among the candidate regulatory elements, promoters and enhancers involved in transcriptional control of the IgH locus are of particular interest (5, 6). For example, in mice lacking the IgH intron enhancer, the efficiency of the initial DH to JH rearrangement is only marginally affected, while the VH to DJ joining is practically absent. These data predict the existence of an additional cis-regulatory element(s), different from the IgH intron enhancer, which is particularly important for D to J rearrangement (7, 8)
A candidate element for such a regulatory role has been described by
Kottmann et al. (9). This gene element has been mapped in
the DQ52 locus, in the sequence flanking the most JH proximal D gene,
DQ52. It has been shown to be a complex genetic element that interacts
with several DNA binding proteins and is transcriptionally active in B
cells, as measured by transient transfection assays (10).
Additional experiments have revealed that one of the DNA binding
proteins interacting with the DQ52 locus is PAX-5 (T. Tallone et al.,
manuscript in preparation). Several other features also suggest a
special role for the DQ52 locus in the primary assembly of IgH genes.
1) The DQ52 element is preferentially used, together with DFL16.1, in
early or primary D-J rearrangements (11, 12, 13).
Interestingly, because the initial gene rearrangement events are not
entirely lineage restricted, inappropriate rearrangement of IgH gene in
thymocytes at an intermediate stage of maturation can occur
(
30–50% of peripheral T cells have DH to JH joins). Surprisingly,
DQ52 is involved in these inappropriate early rearrangements with
unexpectedly high frequency (14, 15, 16). 2) DQ52 is the only
DH element to be transcribed very early before the onset of the V(D)J
recombination (17, 18). These transcripts, called
µ0 germline transcripts (19) can
be considered one of the earliest indications of B-lineage commitment
(18) and appear to originate, in both mouse and human,
from a region immediately 5' to the DQ52 element (17, 20)
(T. Tallone et al., manuscript in preparation). Thus, it is possible
that the DQ52-JH region of the IgH locus might become preferentially
accessible as a consequence of early events underlying commitment to
the lymphoid/B cell lineages. 3) The genomic organization of the DQ52
locus is highly conserved between species (the position, immediately 5'
of the JH segments), and the sequence around the DQ52 element shows
high interspecies homology but low homology with other D elements
within the same species (21).
To test the role of the DQ52 locus in the regulation of VDJ recombination and B cell development, we have generated mutant mice lacking the DQ52 gene with its regulatory sequences. We demonstrate here that deletion of the DQ52 region results in selective reduction of VDJ recombination, incomplete processing of D-J joints and altered D gene usage.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The DNA containing the DQ52 element, the JH cluster, and
the intron enhancer was isolated from plasmids pME184-8
(22) and pSV2gptVDJ-Cµ (23). The
XhoI-SacI fragment containing the DQ52 element
with its regulatory sequences (10) (T. Tallone et al., in
preparation) was replaced by a cassette consisting of a neomycin
resistance gene and a HSV-tk gene flanked by loxP elements
(24). The wild-type sequence was further modified by
exchanging a HindIII site in the intron between JH3 and JH4
to an EcoRV site by the ligation of a short oligo-nucleotide
linker (Fig. 1
A). BALB/c
embryonic stem (ES)2
cells (25) were transfected with the linearized targeting
vector as previously described (26) and screened with
nested PCR using the primers 5'KO1 (5'-GATCTTCAGACCTAGACCATAG) and
OLNeo5 (5'-GCCGATTGTCT GTTGTGCCC; first round), and
5'KO2 (5'-GAGCAGAAGTGTAGATAG GTGG) and OLNeo2A
(5'-CGGAGAACCTGCGTGCAATC; second round). Targeted ES cell clones were
transiently transfected with 20 µg of supercoiled pMC-CreN, a
modified version of the pMC-Cre expression vector (27),
and selected with ganciclovir for loss of the selection cassettes by
site-specific recombination as previously described (28).
The DQ52-deleted ES cell clones were injected into C57BL/6 blastocysts
to obtain chimeric mice as previously described (25).
After germline transmission, heterozygous mice were intercrossed to
obtain DQ52-/- mice. For initial experiments,
targeted ES cell clones were also injected into recombinase-activating
gene-2 (RAG-2)-/- blastocysts.
|
For analysis of B cell numbers DQ52-/- mice (BALB/c background) were mated with C57BL/6 mice. The offspring F1 mice were examined in comparison with age-matched F1 controls (BALB/c x C57BL/6). Single-cell suspensions of spleen and bone marrow (BM) were prepared according to standard procedures and analyzed by flow cytometry. Cells were stained with the following mAb: anti-µa, RS3.1-FITC (29); anti-µb, AF6-78.25-biotin (30); and anti-B220, RA3-6B2-PE (PharMingen). The biotinylated Ab was visualized with streptavidin coupled to Red670 (Life Technologies, Gaithersburg, MD).
Sorting of BM cells and preparation of DNA samples for PCR analysis
Bone marrow cells were stained with anti-IgM, B7-6-FITC (31) and anti-B220, RA3-6B2-PE (PharMingen) and sorted with FACSort equipment (Becton Dickinson, Mountain View, CA) into an IgM-B220low population. The sorted IgM-B220low population did not contain any IgD+ cells, but was about 70% BP-1+ pre-B cells (data not shown). DNA from the sorted and total BM cells for the PCR analysis was prepared as previously described (26).
Analysis of the rearrangement frequency by PCR
Two microliters of the BM lysate or purified spleen DNA was
amplified by PCR using conditions previously described
(32). The following 5' primers were used: DSF
(5'-AGGGATCCTTGTGAAGGGATCTACTACTGTG), a universal DH primer detecting
DJ rearrangements of all other DH elements except DQ52; V7183
(5'-CGGTACCAAGAASAMCCTGTWCCT GCAAATGA), VQ52
(5'-CCAGACTGARCATCASCAAGGACAAYTCC), and V588
(5'-CGAGCTCTCCARCACAGCCTWCATGCARCTCA) primers, specifically amplified
V(D)J rearrangements of VH elements belonging to
VH7183, VHQ52, and VH588 gene families, respectively; and 3' primer,
JH/KO2 (5'-CCACATTCTGTTCCCTGCCTGTGC). Circular
excision products after DH to JH4 rearrangements (signal joints) were
amplified by nested PCR with two primer pairs (Fig. 2D): first pair,
DFL16.1/REC/EXT (5'-GCTCAAGAGTCT GCTGGGCACAATG) and 3'JH3/EXT
(5'-ACTGTCTCTGCAGGTGAGTC CTAAC); and second pair, DFL16.1/REC/INT
(5'-GTTTTTGCTGATG GATATAGCACTG) and 3'JH3/INT
(5'-TTCTCCCATTCTAAATGCATGTTGG). DFL16.1/REC primers are specific to
DFL16.1 but have only 2–3 bp mismatches to most members of the DSP2
family (33). PCR products were purified with Quick Spin
columns (Qiagen, Chatsworth, CA), and 1/10th of the purified reaction
was digested with either EcoRV or HindIII. The
digested PCR products were fractionated on a gel and blotted to a
filter, which was hybridized with the 32P-labeled
oligo-nucleotide probe JH3/S (5'-TCTCCCATTCTAAAT GCATGTTGG) by
standard procedures. To confirm that the PCR products were fully
digested with EcoRV, the filters were washed and
rehybridized with an oligo-nucleotide probe, EcoRV/KO
(5'-CAGAACTGAAGCTCCGAT), that is specific for the introduced
EcoRV modification on the targeted allele.
|
Bone marrow cell lysates or spleen DNA samples were used to amplify PCR products as described above with either DSF or V558 and 3'JH4 (5'-AACCTGCAGAGGCCATTCTTACCTGAGGAG) primers. Twenty-five microliters of each PCR was electrophoretically fractionated, bands corresponding to JH3 rearrangements were cut out from the gel, and the DNA was eluted. One microliter of eluted DNA was amplified with either DSF or V558, as in the primary reaction, and JH/KO2 primers. PCR products were cloned into plasmids, and nucleotide sequencing was performed with the automatic DNA Sequenator A.L.F. using an Auto Read kit supplied by Pharmacia Biotech (Piscataway, NJ).
Generation of Abelson murine leukemia virus (Ab-MuLV)-transformed pro-B cell lines
For generation of Ab-MuLV-transformed cell lines, DQ52-/- mice were crossed onto a RAG-2-/- background (34). Bone marrow cells were isolated from femurs of DQ52+/-/RAG-2-/- mice and transformed as previously described (35). The Ab-MuLV-transformed clones obtained were classified by flow cytometry as pro-B because they were surface positive for the B-lineage markers B220 and CD43 (data not shown).
Preparation of a probe of the mutated DQ52 locus and S1 nuclease protection assay
Genomic DNA from the liver of DQ52-/-
mice was prepared, and a PCR fragment was amplified with 5'KO1 and
DQ/KO3.1 (5'-TCCTGACCTC CAGAAAGCTAC) primers. A
KpnI/BglII subfragment from this PCR product was
cloned into a plasmid vector and sequenced. Subsequently, a
BglII/NheI fragment 3' of DQ52 was added to the
plasmid. A KpnI/PstI insert (containing the
mutated DQ52 locus) was purified and used for the generation of a
radiolabeled DNA probe (see Fig. 7A). The denatured plasmid
fragment was hybridized with the ST3 oligonucleotide and radiolabeled
with Klenow DNA polymerase (Pharmacia). The radiolabeled probe was then
purified on a polyacrylamide gel by excision of the appropriate sized
fragment. The labeled ssDNA was eluted with the StrataEluter
Electroelution Device (Stratagene, La Jolla, CA) and phenol/chloroform
extracted. Isolation of total cellular and cytoplasmic RNA from pro-B
cell lines was performed as previously described (36, 37).
Thirty micrograms of RNA and 3.5 x 105 dpm
of the probe were coprecipitated. The pellet was suspended in 15 µl
of aqueous hybridization solution consisting of 12 µl of 100%
formamide and 3 µl of hybridization buffer (200 mM PIPES (pH 6.4), 2
M NaCl, and 25 mM EDTA (pH 8)). After denaturing at 95°C, the
hybridization was performed at 50°C overnight. The sample was then
mixed with 300 µl of digestion solution (300 U of S1 nuclease (MBI
Fermentas), 10 µg/ml sonicated calf thymus DNA, 0.25 M NaCl, 0.03 M
NaAc (pH 4.5), and 3 mM ZnSO4) and incubated at
30°C for 80 min. After phenol/chloroform extraction and ethanol
precipitation, the products of the nuclease digest were separated on a
polyacrylamide gel and autoradiographed.
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To study the in vivo function of the murine DQ52 gene in
conjunction with its cis-acting regulatory elements, the
DQ52 region was deleted by gene targeting. The targeting vector was
designed so that a 240-bp XhoI-SacI fragment
containing the D segment as well as all the described flanking
sequences with promoter and/or protein binding activities
(10) (T. Tallone et al., manuscript in preparation) was
replaced by neomycin resistance and HSV-tk cassettes (Fig. 1
A). Two loxP sites from the bacteriophage P1 flanked the
selection cassettes (24). The loxP sites were added to
remove the promoter/enhancer-containing selection cassettes after
homologous recombination with the Cre enzyme to avoid any potential
compensatory or enhancing activity of these regulatory elements
(28). In addition, to carry out a quantitative V(D)J
recombination analysis in heterozygous mice, a wild-type
HindIII site between JH3 and JH4 was replaced by an
EcoRV site in the targeting vector to mark the targeted
chromosome (see below). The targeting vector was transfected into ES
cells of BALB/c origin (25). After selection, 20 PCR
positive clones were obtained and further analyzed by Southern blotting
to confirm the proper targeting event and the simultaneous presence of
the introduced EcoRV site (Fig. 1B). Of 20
clones, nine contained the EcoRV modification. Two of these
ES clones were transiently transfected with Cre, and after negative
selection with ganciclovir, all of the surviving clones carried the
expected loxP deletion (Fig. 1A) as analyzed by PCR (not
shown). By blastocyst injection, chimeric mice were generated that
transmitted the mutation into the germline, and the offspring of these
mice were analyzed by PCR (not shown). The correct Cre-mediated
deletion was confirmed by PCR cloning and sequencing of DNA from BM
cells of homozygous DQ52-/- mice. The first
experiments were performed with mice derived from two independent ES
cell clones (clones 5 and 24), which showed identical phenotypes.
Selective reduction of the frequency of rearrangements on the DQ52- allele
The effect of DQ52 deletion on the rearrangement frequency at the
IgH locus was first analyzed. This analysis was performed by PCR
amplification of rearranged genes from heterozygous
DQ52+/- mice with either DH or
VH primers in combination with a JH4 primer. The
principle and specificity of the method are presented in detail in Fig. 2A. DJ rearrangements were analyzed by amplifying DNA of
DQ52+/- mice from total spleen and BM cells as
well as from BM cells sorted to the pre-B cell fraction
(B220lowIgM-). PCR
products amplified with DSF- and JH4-specific primers were digested
with either EcoRV or HindIII and analyzed by
Southern blotting in combination with a densitometric analysis (for
details, see Materials and Methods). DSF primer amplifies
rearrangements of all other DH genes except DQ52. Rearrangements of
VH genes from three VH
families, VH7183, VHQ52, and VH558, were amplified with the
corresponding V7183, VQ52, and V558 primers and analyzed as described
above. Representative examples of the analysis are shown in Fig. 2B. Results from the densitometric analysis of several
experiments are presented as the mean ± SD in Fig. 2C.
In all three cell populations a strong reduction of both DJ and VDJ
rearrangements to JH3, but not to JH1 or JH2, elements of the
DQ52- allele was found. Reduction in the
rearrangement frequency to JH3 was between 50 and 70% for DH, VH7183,
and VH558 elements, whereas it was between 15 and 35% for the VHQ52
family members compared with the wild-type allele (Fig. 2C).
The selective reduction of JH3 rearrangements was rather surprising.
Therefore, it was important to know whether rearrangements to JH4 were
reduced as well. This could not be analyzed in our assay due to the
location of the introduced EcoRV site (see Fig. 1A). Rearrangements can also be analyzed by demonstrating
the presence of circular excision products in actively rearranging
tissues. Based on this fact we designed PCR analysis of excision
products that were produced after DH to JH4 rearrangements. Wild-type
and DQ52- alleles could be discriminated in a
similar manner as described above (Fig. 2D). PCR primers
used in the analysis detected rearrangements of DFL16.1 and most
members of the DSP2 family. DNA of DQ52+/- mice
from sorted BM B220lowIgM-
pre-B cells was amplified, and the PCR products were analyzed as
described above. DH to JH4 rearrangements were reduced by 61 and 76%
in two pre-B cell populations (Fig. 2E). The reduction was
even more dramatic (96%) when DQ52+/- pre-B
cells from complemented RAG-2-/- chimeras were
analyzed (Fig. 2E).
Cell number analysis
The reduced rearrangement frequency on the
DQ52- allele stimulated us to analyze whether it
would also result in a reduction of B cell numbers. To analyze this we
crossed DQ52-/- mice on the BALB/c
(IgMa) background to wild-type C57BL/6
(IgMb) mice and used IgM allele-specific Abs to
analyze the F1 offspring soon after birth. There
was no difference in the relative amounts of
IgMa- and IgMb-positive B
cells in control F1 mice, whereas there was a
significant reduction in the amount of IgMa B
cells (DQ52- B cells) in
DQ52+/- F1 mice (Fig. 3). The relative number of
IgMa B cells was reduced to a similar extent at
most analyzed time points after birth. The average ratio of
IgMa B cells to IgMb B
cells was 38 vs 62% for the BM cells and 43 vs 57% for splenocytes,
respectively. A similar reduction was detected among peritoneal B1
cells (CD5+, IgM+) of
3-wk-old DQ52+/- F1 mice
(data not shown). Compared with normal BALB/c mice, the reduction of B
cell numbers in homozygous DQ52-/- mice was not
evident. Also, no difference in the proportions of different B cell
precursor populations was found in the BM between
DQ52-/- and normal BALB/c mice (data not
shown).
|
After noticing that rearrangements to JH3 and JH4 were strongly
reduced on the DQ52- allele we decided to clone
and sequence JH3 rearrangements to determine whether the reduced
frequency of rearrangements was connected to changes in the potential
Ab repertoire (altered DH and VH usage). Due to
the vast number of VH members in the VH558 family
and their broad distribution within the VH locus,
only VH558-JH3 rearrangements were sequenced to avoid bias caused by a
limited number of VH elements and their strict
location in the VH locus. DH rearrangements were
isolated from sorted BM
B220lowIgM- pre-B cells of
homozygous DQ52-/- and normal BALB/c mice (for
details, see Materials and Methods).
VH rearrangements were isolated from sorted pre-B
cells of DQ52+/- mice and from spleen cells of
homozygous mice. D elements were recognized by comparing the sequences
obtained to the published germline sequences (38, 39, 40).
Some D genes have a minor nucleotide change at their 5' end or even
have an identical coding region when compared with the other D gene
sequences. Therefore, some D genes could not be discriminated from each
other and were grouped as a recognition family (Fig. 4).
|
A).
In the great majority (63%) of rearrangements of
DQ52-/- mice, DFL16.1 was used. The other D
elements used more than once were DFL16.2 (11%) and DSP2.2 (18%). DJ
rearrangements of pre-B cells from BALB/c mice showed a similar
preference for DFL16.1 (40%) and DFL16.2 (20%; Fig. 4A).
However, a shift toward the use of more 5' D elements was noticed in
DQ52-/- rearrangements (e.g., no usage of
DSP2.8, but more frequent usage of DFL16.1; see also Fig. 8), although
all differences were not significant.
|
B).
D elements of DFL16.1 and DSP2.3 families were not used at all on the
DQ52- allele, while they were used in 27 or
20%, respectively, of VDJ joints of the wild-type allele (both
p < 0.05, by Student’s t test). In
contrast, D elements of the DSP2.1 family were frequently used on the
DQ52- chromosome (31 vs 0% on the wild-type
allele; p < 0.05; Fig. 4B). Thus, on the
DQ52- allele, the DFL16.1 was not used at all in
VDJ joints, although it had been the most frequently used D family in
DJ joints (compare Fig. 4, A and B). It was also
interesting that the D usage in the nonselected repertoire of the
B220lowIgM- pre-B cells
was similar to that in the selected spleen cell repertoire (Fig. 4, B vs C). Also in the spleen, D elements of the
DFL16.1 family were only used in the wild-type VDJ joints (20 vs 0%;
p < 0.05). The reverse was true for D elements of the
DSP2.1 family (0 vs 28%; p < 0.01; Fig. 4C). Processing of the rearranging gene elements is altered
When the sequences of nonselected DJ rearrangements were analyzed
in detail, it was noticed that the average DJ length (nucleotide length
from the beginning of the DH coding sequence to the first codon in the
JH element) was only slightly longer on the
DQ52- than on the wild-type allele (10.4 vs 9.15
codons; not significant; Fig. 5A). The usages of the three
reading frames were similar in DQ52-/- and
wild-type joints. However, JH3 elements were much more frequently
intact, without trimming (47 vs 15%; p < 0.05, by
Student’s t test), and P nucleotides were added more
frequently to JH3 (36 vs 5%; p < 0.05) in
DQ52-/- than in BALB/c rearrangements (Fig. 5A). Interestingly, a similar difference in the processing
of JH3 was detected in VDJ joints isolated from spleen cells of
DQ52-/- and BALB/c mice despite the strong
cellular selection (Fig. 5C).
|
To understand the selection against the usage of DFL16.1 and
DFL16.2 on Igs derived from the DQ52- allele, we
analyzed the amino acid composition of the CDR3 region obtained from
functional VDJ rearrangements of both BALB/c and
DQ52-/- spleen cells. Amino acid sequences of
CDR3 regions (Fig. 6) were translated
from the corresponding VDJ sequences shown in Fig. 5C and
presented in a color code as in the report by Wilbert et al.
(41). It was noticed that the complete usage of the 5' end
of JH3 resulted in the incorporation of more hydrophobic amino acids
(shown in dark blue) in this region (mean number of hydrophobic amino
acids in BALB/c and DQ52-/- sequences, 2.0 and
2.9, respectively). In contrast, the amino acids encoded by the 3' end
of the VH gene and the DH element were more
charged (red and purple) and polar (green and light blue) in
DQ52-/- sequences than in the wild-type ones
(Fig. 6). This shift to more charged and polar amino acids seemed to be
caused by an altered usage of D elements in the selected sequences of
the DQ52- allele. A hydropathicity index was
calculated for CDR3 (42) on a scale with Ile, as the most
hydropathic amino acid of 4.5, the neutral amino acid Gly with an index
of -0.4, and Arg as the amino acid with the highest charge of -4.5. D
elements with a mean hydropathicity index of -1.0 (for reading frame
1), such as DFL16.1 and DFL16.2, that were frequently used in DJ
rearrangements (Fig. 4A) were replaced by D elements with a
higher negative hydropathicity index, i.e., with a more polar character
such as DSP2.1 (index, -1.56) or DSP2.2 (index, -2.18) on the
DQ52-/- CDR3s. This was in contrast to BALB/c
sequences (Fig. 4C). Both effects, more hydrophobic amino
acids in the 3' part and more polar/charged amino acids in the 5' part
of the CDR3s of the DQ52-/- mouse, seemed to
compensate each other, because a similar average hydropathicity was
calculated for the wild-type and DQ52-/-
sequences (hydropathicity index of -0.1 for wild-type and -0.38 for
DQ52-/- CDR3). Despite larger JH3 segments due
to reduced trimming, the CDR3 length was only marginally larger in
DQ52-/- sequences than in BALB/c sequences
(mean length, 10.8 vs 9.50 aa; not significant). Thus, despite the
similar hydropathicity index, the CDR3s of splenic B cells in the
DQ52-/- mouse had an altered amino acid
composition compared with the wild-type B cells.
|
The reduction of rearrangements to JH3 and JH4, but not to JH1 or
JH2, was an unexpected finding that indirectly suggested that the
phenotype might be due to a partial effect of the deletion to the
regulation of VDJ recombination. Therefore, it was analyzed whether the
µ0 transcript that originates from the DQ52
promoter (10) (T. Tallone et al., manuscript in
preparation) was completely absent in homozygous
DQ52-/- mice. cDNA from BM cells of BALB/c and
DQ52-/- mice was amplified by PCR with a sense
primer that hybridized 3' of the DQ52 deletion and an antisense Cµ
primer as described previously (43). Surprisingly, a
µ0 transcript could also be amplified with
these primers from DQ52-/- BM cells (data not
shown), although the correct deletion of the DQ52 locus including its
promoter region was confirmed by cloning and sequencing. Additional
experiments were performed to map start sites and quantify the aberrant
µ0 transcript from the targeted allele. For
this purpose, Abelson transformed pro-B cell lines were generated from
BM cells of mice that were heterozygous for the DQ52 deletion in the
RAG-2-/- background. Expression of
µ0 transcripts was analyzed simultaneously from
both alleles by S1 nuclease protection assay. A radiolabeled probe for
the assay was generated by primer extension from a plasmid containing
the DQ52 deletion cloned from the targeted allele (see Materials
and Methods and Fig. 7A).
Thus, the probe was specific for the detection of all potential
transcripts originating from the targeted locus upstream of the DQ52
deletion. However, transcripts arising from the wild-type allele also
generated S1 nuclease-resistant fragments of about 160 bp in length
(Fig. 7, A and B) that served as an internal
control for the normal level of µ0 transcript.
Transcripts of the wild-type allele started from several initiation
sites, as typical for TATA-less promoters. Interestingly, only two
minor transcripts (A and B) initiating around the remaining loxP site
were detected on the targeted allele, but their intensities were only
10 and 5%, respectively, that of the wild-type
µ0 transcript (Fig. 7B). Thus,
despite the complete deletion of the natural start sites of the
µ0 transcript, two very weak aberrant sterile
transcripts starting from the inserted modifications of the DQ52 locus
could be detected on the DQ52- allele.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Three previously described gene deletions, namely IgH enhancer (Eµ)
(8, 44), PAX-5 (45), and IL-7R
(46), have a selective negative effect on VDJ
recombination at the IgH locus. Deletion of any of these genes causes a
strong reduction in V to DJ joining, although DJ joining is practically
unaffected. However, the effect of the Eµ deletion on VDJ
recombination is far milder than that of the TCR
enhancer (E).
Deletion of E practically inactivates both D to J and V to DJ
joining at the TCR locus (47, 48), suggesting that
additional cis-acting elements have to be involved in the
activation of D to J recombination at the IgH locus (8).
Therefore, deletion of the DQ52 locus was a logical concept to test
this hypothesis. This is, in fact, the first gene deletion (except for
the RAG gene deletions) that significantly inhibits DJ rearrangements
at the IgH locus, although the effect is selective to JH3 and
JH4. One possible explanation is that there are several
elements controlling the activation of DJ recombination, and we have
deleted only one of them (see model discussed below). The other
alternative is that the DQ52 locus is the key element, but it has been
only partially inactivated by this particular deletion. All known start
sites of the µ0 transcript are in the deleted
region, and therefore, µ0 transcription should
have been eliminated in the DQ52- allele.
However, a weak µ0-like sterile transcript was
detected in the targeted allele. This transcript started in two
positions around the loxP site retained in the locus after successful
Cre-mediated deletion of DQ52. It remains obscure whether this weak
(only 1/10th of the wild type) transcriptional activity was able to
reverse the effect of the targeted deletion. A novel mutant mouse line
in which the rest of the D gene locus, except for DFL 16.1, has been
removed seems to shed additional light on this question. Preliminary
analysis of rearrangements in these mutant mice shows that the targeted
deletion of the DH locus has a very similar, although stronger,
selective effect on the usage of JH elements in rearrangements as the
deletion of DQ52 (H. W. Schroeder, Jr., personal communication).
This suggests that transcription from the cryptic sites does not
dramatically alter the effect of the DQ52 deletion. Formally, we cannot
absolutely exclude that insertion of the EcoRV linker
between JH3 and JH4 may also influence DJ rearrangement. We think this
is very unlikely, because we have inserted only a few-base pair-long
linker to exchange HindIII to EcoRV without
deleting DNA. In addition, this region is not highly conserved
(49). Allele-specific changes in the close proximity of
the wild-type HindIII site can be found in several mouse
strains. Previously, germline sterile transcripts, potentially involved
in the regulation of rearrangement, have been eliminated at TCR or Ig
loci by gene targeting in three mouse models. Deletion of TCR D1 and
J
KI-KII promoters leads to the significant reduction of
rearrangement at D1-J1 and J loci, respectively (50, 51). In contrast, deletion of the T early 
element
selectively down-regulates TCR V rearrangements to the most 5' J
genes (52). These data together with DQ52 deletion
indicate that the accessibility of each rearranging locus is
individually controlled, most likely by multiple gene elements.
Our finding that processing of JH3 elements was altered in
DQ52-deficient mice is rather surprising, but the present progress in
the knowledge of VDJ recombination might offer a proper explanation.
The VDJ recombination process is initiated by targeting of the RAG-1
and RAG-2 proteins to the conserved recombination signal sequences
(RSSs). Thereafter, double-strand cleavage of DNA occurs precisely at
the junction between an RSS and the adjacent coding sequence (for
review, see Ref. 53, 54). The coding ends are then
processed and ligated by additional activities of several factors
involved in general DNA double-stranded break repair to form a coding
end (for review, see Ref. 55, 56). Gene-targeting
experiments have confirmed that transcriptional enhancers play an
important role in the regulation of VDJ recombination at TCR and Ig
loci (6, 57). It has been generally postulated that these
enhancers control for recombinatorial accessibility. However, Hempel et
al. have demonstrated that deletion of the TCR E enhancer reduces
the accessibility of the locus for recombinase far less than further
processing of the coding ends (57). In addition, it has
been shown that enhancers can selectively regulate the rearranging
elements. By using a human TCR
minilocus system it was demonstrated
that inactivation of the E enhancer selectively reduces
formation of double-stranded breaks 5' of J1 but not 3' of
D3 (58). We have now demonstrated that deletion of the
DQ52 locus causes a selective alteration in the processing of JH3, but
not DH, elements in the DJ rearrangements of unselected BM preB cells.
This suggests that a cis-acting sequence flanking DQ52 plays
a similar role in the IgH locus to the one played by E and E
enhancers in the TCR loci, when one assumes that the action of DQ52 is
Eµ dependent, as shown previously (10).
The analysis of D usage in the sequences from DJ and VDJ rearrangements revealed several interesting observations that merit discussion. First, D usage in VDJ rearrangements from both spleen cells and BM B220lowIgM- pre-B cells was very similar, but markedly different from the primary DJ rearrangements. This suggests that both BM pre-B cells and peripheral B cells are similarly ligand selected. However, it has been demonstrated by Gu et al. that pre-B cells express a broad random repertoire of VH genes, whereas in mature peripheral B cells only some sets of VH genes are dominantly expressed and, therefore, are ligand selected (59). This obvious discrepancy might be explained by the fact that we have amplified VDJ rearrangements primarily from the small pre-B cell subpopulation (fraction D) that is expanded in response to the pre-B cell receptor, composed of a complete heavy chain and surrogate light chains. This expansion may already involve selection for optimal CDR3s. Alternatively, our results may indicate that only optimal CDR3s pair with the surrogate light chains. This idea is further supported by the finding that DFL16.1 and DFL16.2 gene elements were not used in the VDJ rearrangements from the DQ52- allele of pre-B cells or DQ52-/- spleen cells, although they were frequently detected in the corresponding wild-type populations. This was not due to the lack of DFL16 rearrangements, because both DFL16.1 and DFL16.2 genes were very frequently used in the primary DJ joints. Therefore, it is possible that DQ52-/- cells expressing a combination of VH558-DFL16.1/2-JH3 do not form a proper CDR3 that can interact with a putative ligand or pair with the surrogate light chains. Altered processing in DQ52-/- cells frequently leads to the complete use of JH3 and addition of P nucleotides. This can potentially result in binding sites that have longer CDR3s than the wild-type ones and have a stretch of hydrophopic amino acids at the carboxyl-terminal end of CDR3. Because the size of the CDR3 region is tightly controlled (60) and DFL16.1 is 6 bp longer than most other D elements, DQ52-/- B cells expressing an IgH chain with DFL16.1 on the cell surface could be counterselected. Counterselection against DFL16.2 is more difficult to explain, because it has the same germline length as all the members of the DSP family. The amino acid analysis of functional rearrangements from BALB/c and DQ52-/- spleen cells suggests that selection against DFL16.1 and DFL16.2 is due to the germline amino acid composition and altered processing of the these elements. Complete usage of the 5' end of JH3 and addition of P nucleotides on the DQ52- allele result in the incorporation of hydrophobic amino acids. This is compensated by selection of D elements such as DSP2.1 and DSP2.2 that have, on the average, higher polar/charged amino acids instead of D elements with lower average polar character, DFL16.1 and DFL16.2. In consequence, the higher number of polar/charged amino acids in the 5' part of the CDR3 seems to balance the longer stretch of hydrophobic amino acids in the 3' part, resulting in an overall similar mean hydropathicity index in selected DQ52-/- and wild-type CDR3s. This is in good agreement with the strict control of the hydropathicity of the CDR3 as presented by Schroeder et al. (61). Despite the similar mean hydropathicity, the amino acid composition of the CDR3 of the DQ52- allele is altered, which most likely influences the ligand recognition or association with the surrogate light chain of the resulting IgH chain.
To explain our results we propose the following model. JH1 and JH2
genes are accessible for rearrangement in the first wave of
recombination simultaneously with DQ52. Therefore, DQ52 elements
frequently rearrange to JH1 and JH2 when other DH elements are not yet
accessible, and DQ52 is found in early primary rearrangements
(11, 12, 13); thereafter, other elements
(VH, DH and JH) become accessible. However, the
accessibility of JH3 and JH4 genes is enhanced by the DQ52 element
already rearranged to JH1 or to JH2 (Fig. 8). Thus, the DQ52 promoter region,
perhaps in concert with the heavy chain intron enhancer, could act as a
trigger for secondary rearrangements, involving other Ds rearranging to
further 3' JH elements, i.e., primarily JH3 and JH4. In
DQ52-/- B cells the DQ52 gene is not available,
and other DH elements substitute it for the primary rearrangements. As
a consequence, secondary rearrangements are not initiated to the same
extent as in wild-type B cells, as reflected by a lower number of
rearrangements to 3' JH elements (Fig. 8). Assuming that DJ joints of
JH3 and JH4 are frequently the result of secondary rearrangements (as
indirectly indicated in the human and mouse by comparing fetal and
adult JH usages (59, 60, 61)), the
DQ52- allele should use more 5' positioned D
elements compared with the wild-type allele. This is what was found
(see the usage of D elements that have been mapped within the IgH
locus, Fig. 8). An alternative explanation is that the absolute number
of secondary rearrangements is maintained on the
DQ52- allele, but the timing is delayed. This
may be one reason why we were able to detect differences in
rearrangement frequencies in the competitive situation of wild-type and
DQ52- B cells of the +/-
F1 mouse, but not in the homozygous
DQ52-/- mouse, where such a kinetic difference
would be more difficult to identify. As a consequence of reduced
rearrangements to JH3 and JH4, the cells expressing Ag receptors
encoded by Ig genes of the targeted allele have a competitive
disadvantage in heterozygous mice and are hence found in lower numbers
in both BM and spleen. It is possible that the reduced number of B
cells derived from the targeted allele is due to the inefficient
expansion of pre-B cells in the BM or the lower environmental
stimulation of peripheral B cells because of their altered Ig
repertoire. Whether the altered repertoire has an effect on immune
responsiveness will be tested in future experiments by immunizations
with Ags resulting in canonical Ig responses involving DJH3 or
DJH4.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Abbreviations used in this paper: ES, embryonic stem; Ab-MuLV, Abelson murine leukemia virus; RAG, recombinase-activating gene; RSS, recombination signal sequence; BM, bone marrow.
Received for publication July 10, 2000. Accepted for publication November 30, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene rearrangement and selection during fetal thymocyte development. J. Immunol. 159:3338.[Abstract]
-chain enhancer: the role of enhancer elements in controlling V(D)J recombination accessibility. Proc. Natl. Acad. Sci. USA 93:7871. T-cell development. Proc. Natl. Acad. Sci. USA 93:7877.1 gene segment at the TCR locus by a germline promoter. Immunity 10:313.[Medline]
cluster impairs chain rearrangement in cis in mice and in the 103/bcl2 cell line. J. Exp. Med. 189:1443. following targeted deletion of T early (TEA): implications for TCR locus accessibility. Immunity 5:331.[Medline]
locus: differential effects on DNA cleavage and joining. Genes Dev. 12:2305.This article has been cited by other articles:
|
|
 |
|
  T. Chakraborty, T. Perlot, R. Subrahmanyam, A. Jani, P. H. Goff, Y. Zhang, I. Ivanova, F. W. Alt, and R. Sen A 220-nucleotide deletion of the intronic enhancer reveals an epigenetic hierarchy in immunoglobulin heavy chain locus activation J. Exp. Med., May 11, 2009; 206(5): 1019 - 1027. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Zemlin, R. L. Schelonka, G. C. Ippolito, C. Zemlin, Y. Zhuang, G. L. Gartland, L. Nitschke, J. Pelkonen, K. Rajewsky, and H. W. Schroeder Jr. Regulation of Repertoire Development through Genetic Control of DH Reading Frame Preference J. Immunol., December 15, 2008; 181(12): 8416 - 8424. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. G. Bates, D. Cado, H. Nolla, and M. S. Schlissel Chromosomal position of a VH gene segment determines its activation and inactivation as a substrate for V(D)J recombination J. Exp. Med., December 24, 2007; 204(13): 3247 - 3256. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Retter, C. Chevillard, M. Scharfe, A. Conrad, M. Hafner, T.-H. Im, M. Ludewig, G. Nordsiek, S. Severitt, S. Thies, et al. Sequence and Characterization of the Ig Heavy Chain Constant and Partial Variable Region of the Mouse Strain 129S1 J. Immunol., August 15, 2007; 179(4): 2419 - 2427. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. C. Ippolito, R. L. Schelonka, M. Zemlin, I. I. Ivanov, R. Kobayashi, C. Zemlin, G. L. Gartland, L. Nitschke, J. Pelkonen, K. Fujihashi, et al. Forced usage of positively charged amino acids in immunoglobulin CDR-H3 impairs B cell development and antibody production J. Exp. Med., June 12, 2006; 203(6): 1567 - 1578. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Pawlitzky, C. V. Angeles, A. M. Siegel, M. L. Stanton, R. Riblet, and P. H. Brodeur Identification of a Candidate Regulatory Element within the 5' Flanking Region of the Mouse Igh Locus Defined by Pro-B Cell-Specific Hypersensitivity Associated with Binding of PU.1, Pax5, and E2A. J. Immunol., June 1, 2006; 176(11): 6839 - 6851. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. L Cassady-Cain and A. K Kaushik Increased negative selection impairs neonatal B cell repertoire but does not directly lead to generation of disease-associated IgM auto-antibodies Int. Immunol., May 1, 2006; 18(5): 661 - 669. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Maes, S. Chappaz, P. Cavelier, L. O'Neill, B. Turner, F. Rougeon, and M. Goodhardt Activation of V(D)J Recombination at the IgH Chain JH Locus Occurs within a 6-Kilobase Chromatin Domain and Is Associated with Nucleosomal Remodeling J. Immunol., May 1, 2006; 176(9): 5409 - 5417. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Afshar, S. Pierce, D. J. Bolland, A. Corcoran, and E. M. Oltz Regulation of IgH Gene Assembly: Role of the Intronic Enhancer and 5'DQ52 Region in Targeting DHJH Recombination J. Immunol., February 15, 2006; 176(4): 2439 - 2447. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Schelonka, I. I. Ivanov, D. H. Jung, G. C. Ippolito, L. Nitschke, Y. Zhuang, G. L. Gartland, J. Pelkonen, F. W. Alt, K. Rajewsky, et al. A Single DH Gene Segment Creates Its Own Unique CDR-H3 Repertoire and Is Sufficient for B cell Development and Immune Function J. Immunol., November 15, 2005; 175(10): 6624 - 6632. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Kalmanovich and R. Mehr Models for Antigen Receptor Gene Rearrangement. III. Heavy and Light Chain Allelic Exclusion J. Immunol., January 1, 2003; 170(1): 182 - 193. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Delpy, C. Decourt, M. Le Bert, and M. Cogne B Cell Development Arrest Upon Insertion of a neo Gene Between JH and E{micro}: Promoter Competition Results in Transcriptional Silencing of Germline JH and Complete V(D)J Rearrangements J. Immunol., December 15, 2002; 169(12): 6875 - 6882. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
JH1, 
