of April 13, 2018.This information is current as

-Like Germline TranscriptsµMultiple I Exon Splice Donor Site: Evidence forµthe I

Normal Isotype Switching in B Cells Lacking

Young, Jianzhu Chen and Andrea BottaroIgor I. Kuzin, Gregory D. Ugine, Dongming Wu, Fay

http://www.jimmunol.org/content/164/3/1451doi: 10.4049/jimmunol.164.3.1451

2000; 164:1451-1457; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/164/3/1451.full#ref-list-1

, 18 of which you can access for free at: cites 41 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2000 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Normal Isotype Switching in B Cells Lacking the Im ExonSplice Donor Site: Evidence for Multiple Im-Like GermlineTranscripts

Igor I. Kuzin, 1* Gregory D. Ugine,1* Dongming Wu,* Fay Young,* Jianzhu Chen,† andAndrea Bottaro2*

Ig class switch recombination (CSR) in activated B cells is preceded by the generation of “switch” transcripts from the heavy chainconstant region (CH) genes targeted for rearrangement. Switch transcripts include a sterile “I” exon spliced onto the first CH exon.Targeted mutations disrupting the expression or splicing of I exons severely hamper CSR to all tested CH loci, exceptm. However,all m switch transcript mutations tested so far have left the Im exon splice donor site intact. To test the possibility that the residualCSR activity in I m mutants could be due to splicing of a truncated Im exon, we generated new mutants specifically lacking the Imsplice donor site. Surprisingly, normal CSR was observed in the Im splice donor mutants even in the absence of detectable splicedIm transcripts. In a search for potential alternative sources of switch-like transcripts in them locus, we identified two novel exonswhich map just upstream of the Sm region and splice onto the Cm1 exon. Their expression is detectable from early B celldevelopmental stages, and, at least in hybridomas, it does not require the Em enhancer. These studies highlight a unique structurefor the m locus I exon region, with multiple nested switch transcript-like exons mapping upstream of Sm. We propose that all ofthese transcripts directly contribute to m class switching activity. The Journal of Immunology,2000, 164: 1451–1457.

I mmunoglobulin class switching occurs during B lymphocyteactivation as a result of a DNA recombination event involv-ing long repetitive sequences (S regions)3 located 59of every

heavy chain constant region (CH) gene except Cd (1).Both in vivo and in vitro, class switch recombination (CSR) can

be directed toward specific CH genes depending on the signalsreceived by the B cell from activators and cytokines, and is pre-ceded by transcriptional activation of the induced genes (1, 2). Thetranscripts generated at this stage (called CH germline, or “switch”transcripts) derive from promoter elements located directly up-stream of the S regions targeted for rearrangement, and in theirmature form contain a noncoding “I” exon spliced onto exon 1 ofthe CH gene. Targeted mutagenesis experiments have indicatedthat integrity of the germline transcription units (promoters and/orexons), rather than transcription through the S regions, is essentialfor CSR (3–9). More recently, splicing of germline transcript pre-cursor has been shown to probably constitute an essential step inCSR activation (7, 10, 11). Thus, a 59-truncated or spliced-outRNA species containing the S region may be an integral part of theCSR enzymatic machinery, or perhaps the same factors that effectRNA splicing may also be involved in DNA recombination pro-

cesses. In support of the hypothesis of a functional connectionbetween RNA splicing and CSR is the identification of proteinsinvolved in RNA processing as components of switch regionDNA-binding complexes (12–14) and the precedent of yeast fac-tors involved in both splicing and meiotic recombination (15).

The IgHm gene, like the other CSR-capable CH genes, harborsan I exon, Im, located immediately downstream of the IgH intronicenhancer (Em), which serves as its promoter (16, 17). Interestingly,the E4 box and the octamer elements within Em seem to be spe-cifically required for Im transcription, rather than Em enhancerfunction (17), and ectopic expression of the E4-binding E47 tran-scription factor has been shown to be sufficient to induce Im tran-scripts in non-B lymphoid cells (18, 19). Thus, Im transcriptionappears to be a very specific and regulated aspect of Em activity,supporting the hypothesis of an important role for Im transcripts inB cell function. However, Im expression is quite constant through-out B cell development and does not increase upon CSR activation(16, 20–22). This observation, along with the finding that, in themajority of LPS-activated B cells, Sm regions remain remarkablystable and show no signs of recombination activity (23–25), sug-gests that other layers of control, in addition to the production ofIm transcripts, may be present form CSR.

Deletions encompassing Em and part of Im significantly reduce,but do not completely block,m CSR (24–26). This is in contrastwith the other CH genes, which are entirely dependent on theexpression and splicing of their I region transcripts for efficientCSR (3, 4, 6, 7, 9). The analyzed mutations however involved onlythe 59-most 500 bp of the 700-bp Im exon, leaving the splice donorsite intact. Therefore, those experiments cannot rule out the pos-sibility that transcripts generated from the more upstream VH or Dregion promoters could be aberrantly spliced at the remaining Im

splice donor, allowing for at least some Im function.To specifically test whether splicing of the Im exon is indeed

necessary form CSR, we therefore generated a mutation that spe-cifically deletes the Im splice donor.

*Departments of Medicine and Microbiology and Immunology, and Cancer Center,University of Rochester School of Medicine and Dentistry, Rochester, NY 14642;and †Department of Biology, Massachusetts Institute of Technology, Cambridge,MA 021139

Received for publication July 2, 1999. Accepted for publication November 19, 1999.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby markedadvertisementin accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 I.I.K. and G.D.U. contributed equally to this work.2 Address correspondence and reprint requests to Dr. Andrea Bottaro, ImmunologyUnit, University of Rochester Medical Center, Box 695, 601 Elmwood Avenue,Rochester, NY 14642. E-mail address: andrea_bottaro@urmc.rochester.edu3 Abbreviations used in this paper: S, switch; CH, heavy chain constant region; CSR,class switch recombination; ES, embryonic stem; RPA, RNase protection assay;RACE, rapid amplification of cDNA end.

Copyright © 2000 by The American Association of Immunologists 0022-1767/00/$02.00

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Materials and MethodsGeneration of targeted embryonic stem (ES) cell clones andmutant mice

TheDIm-s construct was generated by replacing the 236-bpBglI-SpeI frag-ment spanning the Im splice donor site with a LoxP-flankedpgk-neor gene(Fig. 1). The two flanking arms are an 8.7-kbEcoRI-BglI fragment at the59, and a 632-bpSpeI-HindIII fragment at the 39; a pgkpromoter-driventkgene was inserted at the 39end of the construct for double G418/gancy-clovir selection of homologous recombinants (27).DIm-s transfected G418/gancyclovir-resistant ES cells clones were screened for homologous re-combination by Southern blotting. Four targeted clones were injected intoC57BL/6 blastocysts, and resulting chimeric mice bred for germline trans-mission with C57BL/6 mice. Germline heterozygous progeny of the orig-inal chimeras were then bred with the adenovirus EIIA promoter-drivenCre transgenicdeletermouse (28) for LoxP-mediated deletion of theneor

gene, and deleted progeny were bred back into C57BL/6 and to each otherto obtain homozygotes.

Ig production analysis

Serum from.6- to 16-wk-old homozygous and heterozygous mutant miceas well as normal littermates and C57BL/6 controls was analyzed byELISA using mouse isotype-specific goat polyclonal Abs (Southern Bio-technology Associates, Birmingham, AL). Splenocytes from 6- to 16-wk-old mutants and controls were cultured for 5 days in complete RPMI 1640medium with 10% FCS and 20mg/ml LPS, 10mg/ml dextran sulfate withor without 25 ng/ml recombinant mouse IL-4 (R&D Systems, Minneapolis,MN). At day 5, supernatants were collected and analyzed by ELISA, andB cell blasts were stained for surface Ig heavy chain isotype expressionusing fluorochrome-conjugated monoclonal rat Abs (PharMingen, San Di-ego, CA, and Southern Biotechnology). Stained cells were analyzed byflow cytometry on a Becton Dickinson FACScalibur instrument usingCellQuest software (Mountain View, CA).

RNA analysis

Total RNA was extracted from splenocytes, LPS-activated blasts, and in-dicated cell lines using the Trizol reagent (Life Technologies, Rockville,MD). Northern blot analysis was performed after electrophoresis of 15mgof total RNA on 2% formaldehyde and 1% agarose gels using random-primed32P-labeled probes. Antisense probes for RNase protection assays(RPAs) were labeled with [32P]UTP using the T3/T7 Maxiscript kit (Am-bion, Austin, TX). A total of 15mg of total RNA/sample was hybridizedto the radiolabeled probes and subjected to RPA using the RPAII kit (Am-bion). The RPA products were separated by 6 M urea/PAGE and detectedby autoradiography for 2–16 h.

59 rapid amplification of cDNA end (RACE) cloning of newm-containing transcripts

Cm-containing transcripts were cloned from total RNA from the Em-de-leted hybridoma FSKO-LPS7 (25) using the Boehringer-Mannheim 59/39RACE kit (Indianapolis, IN) with the following nested Cm antisense oli-gonucleotides: AS-Cm1 (59-TTCTGGTAGTTCCAGGTGAA-39), 39mCH1(59-ACCAGATTCTTATCAGACAG-39), and Cm1AS29–47 (59-CTCTCGCAGGAGACGAGG-39). The resulting mixture of PCR products was clonedinto a PCR2.1 vector (Invitrogen, San Diego, CA), and individual recombinantplasmids were directly analyzed by sequencing.

ResultsGeneration of targeted ES cell clones and of knockout mice

J1 ES cells were transfected with theDIm-s construct and grownunder double G418/gancyclovir selection (27). Resistant colonieswere picked, expanded, and analyzed by Southern blotting. In thegermline configuration, the CmXH probe detects 2 bands onBamHI-digested DNA because of the presence of an internalBamHI site (Fig. 1A). The larger 12-kb band corresponds to the 39portion of them gene and its 39flanking sequences, which wouldnot be affected by targeting of the Im exon. The smaller 9.5-kbband spans the JH-Cm intron, and uponDIm-s homologous recom-bination is reduced to about 7 kb. This novel pattern was detectedin 14 of 168 analyzed clones (Fig. 1B). Homologous integration ofthe 59arm of the construct was confirmed using the 59 arm-internalEm probe (Fig. 1A; data not shown). Four targeted clones were

injected into C57BL/6 blastocysts, chimeric mice were generated,and germline transmission of the mutation was obtained from twoindependent clones (Fig. 1C). Mice bearing the targeted mutation(DIm-sneo/neomice) were then bred to thedeletertransgenic mouse(28), which induces LoxP-Cre-recombinase-mediated deletion oftheneor gene at the flanking LoxP sites. In theDIm-sneomice, theCre-deletion event causes the replacement of theneor-containing4.2-kb BamHI band with a shorter 2.7-kb band (DIm-s2/2 mice,Fig. 1C). The Cre-deleted mice were then bred back in theC57BL/6 background; the genotype of the line is therefore mixed(mostly C57BL/6, with FVB contribution from thedeleterstrainand 129 from the original ES cells).

Normal CSR in Im splice donor mutants

As a first approach to assessing the effects of deletion of the Imsplice donor on CSR, we tested serum Ig levels inDIm-s2/2 miceand controls and found no significant differences between the co-horts (Fig. 2A). Similarly, supernatants from 5-day LPS/dextransplenocyte cultures with or without IL-4 had similar levels of all

FIGURE 1. Generation of Im splice donor mutants.A (top), Scheme ofthem locus showing the DQ52 and JH segments, Em enhancer, Im exon, Smregion, and part of the Cm gene. Relevant restriction sites are indicated (R,EcoRI; B,BamHI; H,HindIII; Bg, BglI; Sp,SpeI; Sa,SacI). DNA probesused in this work are shown above the map, as follows:mo probe, a 380-bpSacI-ApaI fragment; Em probe, a 1.6-kbHindIII-EcoRI fragment; Improbe, a 0.7-kbEcoRI-HindIII fragment; 39Im, a 0.7-kbHindIII fragment;CmXB probe, a 0.9-kbXbaI-BamHI fragment; and CmXH, a 1.2-kbXbaI-HindIII fragment. Middle, Structure of theDIm-s targeting construct, inwhich a LoxP site (Œ)-flankedpgk-neor gene replaces a 236-bpBglI-SpeIfragment spanning the Im splice donor site.Bottom, Structure of the tar-geted locus after Cre-mediated deletion.B, Southern blot analysis ofDIm-s-transfected J1 cell DNAs cut withBamHI and hybridized with the CmXHprobe. The top band represents the 39portion of the Cm gene, the lower9.5-kb band spans the JH-Cm intron. This band is reduced to about 7 kbupon homologous recombination of theDIm-s construct, as in the twotargeted clones shown (right lanes).C, Southern blot analysis of tail DNAfrom a normal (1/1), heterozygous (1/neo), and homozygous (neo/neo)mouse bearing theDIm-sneor replacement. The DNA was cut withBamHIand hybridized with an Em probe, which detects a 9.5-kb band for thegermline locus and a 4.2-kb band for theneor-replaced locus.D, Southernblot analysis of genomic tail DNA from a normal (1/1) and twoCre-deletedDIm-s homozygous (2/2) mice.BamHI-cut DNA hybridized withthe Em probe reveals now a 2.7-kb band in the Cre-deleted mice, markingthe deletion of theneor gene.

1452 NEW Im-LIKE GERMLINE TRANSCRIPTS

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

tested Ig isotypes (IgM, IgG1, IgG3, and IgG2b) inDIm-s2/2 andcontrol mice (Fig. 2B). Homozygous mutant B cells also displayeda pattern of expression of surface IgM, IgG1, IgG2b, IgG3 (Fig. 3),and IgE (data not shown) indistinguishable from that of controls.

Therefore, we conclude that mutation of the Im splice donor,unlike that of the Ig1 exon (7, 10), does not affect CSR.

Lack of steady-state Im transcripts in Im splice donor mutants

To confirm that replacement of the Im splice donor site with aLoxP site did not lead to the production of aberrantly spliced Im-like transcripts, we performed Northern blot analysis of total RNAfrom mutant and control LPS blasts using an Im-specific probe.Unlike control samples, which display a prominent 2.5-kb Im-hy-bridizing band, mutants show only weak hybridizing signals withthis probe (Fig. 4A). These weak bands are however most likelyderived from cross-hybridizing transcripts, since they are also ob-served in RNA from LPS blasts in which the entire region corre-sponding to the Im probe (EcoRI-HindIII fragment) has been re-placed by aneor gene (Fig. 4A) (29). Thus, in the absence of theIm splice donor site, there is no accumulation of steady-state Imtranscripts. Although this is presumably due to instability of thelong unspliced precursor, we cannot rule out the possibility thattranscriptional regulatory elements relevant for Im expression havebeen affected by theDIm-s mutation.

Novel transcripts are expressed in them locus of Em/Im-deletedhybridomas

If the widely accepted model of a role for I exon splicing in CSR(1, 7) is correct, the results mentioned above raise the question of

FIGURE 2. Ig isotype levels in serum and LPS supernatants ofDIm-smutant mice. IgM, IgG3, IgG2b, and IgG1 levels were measured in theserum (A) and in day 5 LPS and LPS/IL-4 culture supernatants (B) fromDIm-s2/2 and control (heterozygous,1/2, normal,1/1) mice. The data inB are pooled from two independent representative experiments. No signif-icant differences are observed between mutant and normal mice in eitherset of data.

FIGURE 3. FACS analysis of in vitro Ig class switching inDIm-s mu-tant mice. B cell blasts from day 5 LPS and LPS/IL-4 cultures ofDIm-s2/2

and normal splenocytes were stained for surface expression of IgG3, IgG1,IgG2b, and IgM (horizontal axes) and B220 (vertical axes). The surfaceexpression profiles are essentially identical for all isotypes in normal(1/1) and mutant (2/2) mice. Similar results were also observed for IgEexpression (data not shown).

FIGURE 4. Northern blot analysis ofm transcripts inDIm-s mutants andEm/Im-deleted hybridomas.A, Total RNAs from splenocytes and day 5LPS and LPS/IL-4 cultures of normal (1/1) andDIm-s2/2 mice, as wellas from a LPS culture of B cells homozygous for a deletion of theEcoRI-HindIII fragment spanning most of Im (ImKO LPS lane) (29), were hy-bridized sequentially with the Im probe (top panel) and the CmXB probe(bottom panel). Unlike normal controls, only weak hybridizing bands areobserved with the Im probe in theDIm-s2/2 samples. Similar bands arehowever also seen in the ImKO sample, which lacks the entire probe re-gion, suggesting that they are likely due to cross-hybridization.B, TotalRNAs from Em/Im-deleted hybridomas (FSKO-32 and -131 with germlineJH regions and FSKO-144 with a rearranged JH region on the targeted loci)(25) were hybridized sequentially with the CmXB probe (left panel) and themo probe (right panel). Only a subset of them-containing transcripts hy-bridize with themo probe. Specificity of themo probe is confirmed by thelack of hybridization in the FSKO-144 hybridoma, in which the proberegion has been deleted by the JH rearrangement.

1453The Journal of Immunology

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

how m CSR is accomplished in the absence of detectable Im splic-ing. We hypothesized that other previously unrecognized tran-scripts from them locus could provide the required splicing event.Normal cells express a number of differentm-containing tran-scripts (20, 21) hampering the identification of potential new tran-scripts. As a simplified model,we took advantage of hybridomalines expressing only a limited subset ofm transcripts. The hybrid-omas were generated from B cells heterozygous for a deletionextending from the Em enhancer to the 59portion of Im, up to 320bp upstream of the splice donor (Em-del hybridomas) (25). The Cmgene has been lost on the normal allele of these IgG-secretinghybridomas because of CSR, while it is retained on the targetedallele. In addition, the JH region of the targeted locus is in thegermline configuration because of a significant defect in VDJ re-combination (29). Finally, because of the Em deletion, these hy-bridomas cannot express any Im-initiated transcripts. Thus, theEm-del- hybridomas represent a simplified system to look for Em-independent transcripts in them locus. The only known Cm tran-scripts that can be expressed from the targeted loci are those orig-inating from the DQ52 promoter, which proceed through the JHregion and splice the JH1 segment onto the Cm1 exon (mo tran-scripts) (30, 31). These transcripts can be detected with amo-spe-cific probe mapping between DQ52 and JH1. Surprisingly, how-ever, Northern blot analysis showed that only a subset of Cm-containing transcripts from the hybridomas also hybridizes withthe mo probe (Fig. 4B). Initial RPA with Cm1-specific probesshowed that these transcripts were in the sense orientation andincluded unknown sequences spliced onto the Cm1 acceptor site(data not shown). These data show that a previously unrecognizedclass of transcripts is expressed in them locus of Em-del-deletedhybridomas, and that these transcripts include the Cm gene splicedonto novel sequences.

Cloning and characterization of two classes ofmx transcripts

The detection of a novel type of transcripts in Em/Im mutantsstrengthened our hypothesis that other transcripts may be respon-sible for the CSR activity in Im splice donor mutants. If this is thecase, however, these transcripts should originate in a region rele-vant for CSR (possibly upstream of Sm), and they should be ex-pressed in normal cells, in particular in B cells undergoing CSR.We therefore utilized 59RACE to identify Cm containing tran-scripts from one of the Em-del hybridomas described above. Fromthe pool of clones obtained by 59 RACE, we sequenced nine clonesselected for hybridizing with a Cm-specific oligonucleotide but notthe mo probe. Two of these clones showedm locus sequencesspliced onto the Cm1 exon; both sequences mapped to the regionbetween Im and Sm. In particular, the two clones,mx18 andmx31,correspond to almost contiguous stretches located immediately up-stream of the Sm repeats (Fig. 5). A consensus splice donor ispresent in both genomic sequences immediately 39of the site ofjunction of the novel exon with the Cm1 exon, suggesting that thecloned products resulted from a bona fide spliced RNA transcriptrather than artifactual PCR products. Both clones, although prob-ably not complete (see below), have multiple stop codons in allthree reading frames, a feature common to all I region germlinetranscripts (reviewed in Ref. 2).

To confirm thatmx18 andmx31 transcripts are not an artifact ofPCR cloning, we generated RPA probes with the clonedmx18-Cmandmx31-Cm sequences and tested the original Em-del hybridomasfor their expression. Indeed, fully protectedmx18-Cm andmx31-Cm fragments can be detected in the original clone and otherEm-del hybridomas along withmx18- and mx31-only protected

fragments (Fig. 6A). The latter probably represent precursor RNAsor the same exons spliced onto other sequences, such as the Cg

genes expressed by the normal allele and the fusion partnerIgH loci.

We then used the same probes to assess the pattern of expressionof the mx18 andmx31 transcripts in other B cells. RPA analysisshowed that both types of transcripts can be detected in B lym-phoid cell lines of as early a stage as pro-/pre-B cells (such asRag22/2 Abelson murine leukemia virus-transformed lines), andmost importantly in normal spleen and LPS-activated B cell blasts(Fig. 6B) as well as inDIm-s2/2 samples (Fig. 6C). Using RPAprobes formx and Cm exon sequences in combination, we couldestimate the amount ofmx-Cm transcripts to be about 1–5% of thetotal m transcripts in splenocytes and activated B cells (data notshown). Thus, the steady-state level ofmx transcripts is rather low,although clearly detectable, and quite constant in lines at differentdevelopmental stages, a pattern that is very similar to that of Im.While our data suggest that, like Im, the novel transcripts are notup-regulated by activation, specific experiments comparing pureresting B lymphocytes (rather than total splenocytes) and B cellblasts are required to conclusively establish this point. However,the expression ofmx transcripts from loci bearing a deletion of theEm enhancer and Im 59 region (such as Em-del hybridomas) indi-cates that these novel transcripts are under different regulatory el-ements than Im.

FIGURE 5. Structure of themx31 andmx18 transcripts.Top,Sequencesof the mx31 andmx18 clones identified by 59RACE in an Em-deletedhybridoma. In each sequence, uppercase letters denote Cm1 exon se-quences, and lowercase letters denote the sequences of the putative novelexons. Minor differences from published sequences (GenBank accessionnumber J0040) are marked by italics, and they are most likely due tosequencing errors or polymorphism between 129 and BALB/c strain al-leles. The two novel exon sequences map to a region just upstream of Sm(center, not in scale), and are separated by just 4 bp.Bottom, The sameregion in expanded view (in scale), highlighting the genomic sequences atthe exon/intron boundaries, with the canonical splice donor sitesunderlined.

1454 NEW Im-LIKE GERMLINE TRANSCRIPTS

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

One of the novel Im-like transcript likely initiates within the Im-Sm region

An important aspect regarding themx transcripts is whether theyindeed initiate independently of otherm locus transcripts from spe-cific promoter elements, or whether they represent alternativelyspliced versions of VDJ- or Im-initiated transcripts.

As discussed above, themx transcripts are expressed at very lowlevels, and they are nested within a large intron of two mRNAs,VDJ-Cm and Im-Cm, which are transcribed at a much higher rate.These two factors complicate the search for initiation sites usingconventional techniques, generating high lane background in RPA,primer extension, and S1 nuclease protection experiments.

Nevertheless, RPA analysis using a genomic probe spanning the59 portion of the clonedmx18 sequence and its 59flanking regionsrevealed three predominant protected bands located within a 80-bprange (Fig. 7A). These bands may represent either bona fide initi-ation sites for themx18 transcript or splice acceptor sites from a yetunidentified upstream exon. However, no splice acceptor consen-sus sequences are detected in the relevant regions by either thehsplor nnsspsplice site search programs available at the Baylor Col-lege of Medicine Search Launcher web site [http://dot.imgen.bc-m.tmc.edu:9331/seq-search/gene-search.html) (32)]. Moreover,the region immediately preceding the putative initiation sites con-tains a number of potential transcription factor DNA binding sites,including an octamer and an E47 consensus binding site (Fig. 7B),both of which have been shown to be critical for Im expression(17). A more detailed structural and functional analysis of thisputative promoter region to establish whether elements in this re-gion are involved inmx18 transcription is in progress.

DiscussionTargeting experiments have shown that CSR of Cg, Ce, and Cagenes is dependent on the expression of I region germline tran-scripts, and that splicing of the germline transcript precursors isprobably an essential step for CSR activation (Refs. 7, 10, and 11;reviewed in Ref. 1). It is therefore rather surprising that analogous

experiments in them locus show that only 50–70% of the CSRactivity at them locus is eliminated by deletion of the 59half of theIm exon and its promoter elements located within the Em intronicenhancer (24–26). However,m locus transcription is still clearlydetectable in these mutants, and since the Im splice donor site wasnot deleted by the mutation, this site could still have served asdonor for splicing events of large precursor transcripts generatedupstream of the JH segments. Therefore, those experiments werenot conclusive in terms of the role of Im splicing in CSR.

In the present report, we show that a specific mutation of the Im

splice donor does not abolish CSR, even in the absence of detect-able splicing of the residual Im sequences. We therefore reasonedthat either the CSR/germline transcript “splicing model” had to berevised, at least for them locus, or that other transcripts wereproviding the necessary splicing event form CSR in these mutants.Indeed, we were able to show that two novel exon sequences,mx18andmx31, mapping to the region between Im and Sm, are splicedonto the Cm1 exon in B cells from different stages of development,including switching LPS-activated B cell blasts. Although tran-scripts for both sequences are expressed from mutants lacking theEm enhancer, we were able to identify potential initiation sites inthe Im-Sm region only formx18, which may therefore represent abona fide second class of germline transcript. As discussed inRe-sults, the low level of expression of themx transcripts does notallow us to completely rule out the possibility that they representalternative splicing products of largerm locus precursor tran-scripts. Thus, a definitive proof of the origin ofmx18 andmx31transcripts will require additional experiments.

The mx18 transcripts share a number of similarities with the Imtranscripts. Like Im, the mx18 exon is spliced onto the Cm1 exonand does not possess any significant open reading frame, although,like Im, it could potentially be translated starting at a Cm1 ATGcodon (33). Also similar to Im, mx18 is expressed at relativelyconstant levels throughout B cell development and possiblythrough activation. Finally, if our RPA initiation site data are cor-rect, mx18 may initiate in proximity to putative binding sites for

FIGURE 6. RNase protection analysis of themx31 andmx18 transcripts.A, 32P-labeled antisense RNA probes, containing 257 bp ofmx31 or 258 bp ofmx18 cDNA sequence and 150 bp of plasmid sequences (total of about 400 bp), were generated from the two 59 RACE-clonedmx-Cm sequences andhybridized to total RNA from yeast, the NS1 fusion partner and Em/Im-deleted hybridomas (one representative sample, FSG38, shown here). Ab-actinprobe was added to one set of digestions as an internal control (right panel). Labeled RNA m.w. marker bands are shown on theleft. Both probes detecteda fully protected band of about 255 bp (mx-Cm) and a shorter band of about 210 bp corresponding to themx exons alone, either from the precursor RNAsor from mature transcripts with themx exons spliced onto a different CH gene.B, The same probes as inA were hybridized to RNA from a RAG22/2

Abelson virus-transformed cell line (pro-/pre-B cell stage), am-chain-positive Abelson line (18.8, pre-B cell stage), normal spleen, and LPS-activated blastsfrom a 129 mouse, as well a GAPDH-specific probe as an internal control (39).mx-Cm transcripts were detected in all analyzed samples at relatively constantlevels.C, The same probes as inA were hybridized to RNA from LPS cultures of normal 129 mouse andDIm-s2/2 mouse splenocytes. Comparable levelsof mx18 andmx31 transcripts were observed in both sets of samples.

1455The Journal of Immunology

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Oct and E47, which are the principle effectors of Im transcription(17). It should be noted, however, thatmx18 andmx31 representonly a small fraction (a few percent) of the totalm-containingtranscripts, in contrast to the high levels of expression of Im, whoselevels are comparable to those of VDJ-m transcripts (16, 21).

Detection ofmx31-Cm- and mx18-Cm-containing transcripts inEm-del hybridomas strongly suggests that their expression is in-dependent of Em. Transgenic studies from Sigurdadottir et al. (34)have shown that a potential novel enhancer element involved inIgH transcription in activated B cells may map to the region be-tween Im and Sm. In addition, DNase-hypersensitive sites havebeen detected in themx exon regions in human B cells, and at leastin T cells in the mouse (35, 36). It is intriguing to speculate thatsuch potential enhancer element(s) may be involved inmx tran-script expression, just as Em acts as the Im promoter.

In mice bearing the JHT mutation (24), which spans the entireDQ52-JH-Em region but not the putative Sigurdadottir enhancer(34), only traces ofm-containing transcripts are detected. It wouldtherefore be of interest to establish whether these remainingmtranscripts in JHT mutants are indeed themx18 andmx31 tran-scripts at their normal steady-state rate (which as we discussed ismuch lower than otherm transcripts), or whether the presence ofthe larger deletion does in fact affect their expression, either bydeleting some unknown regulatory element or by significantly al-tering the structure of the locus. Since the JHT heterozygotes, likethe other Em mutants, are still able to switch at low levels, wewould predict that they would expressmx18 andmx31 transcripts.

Also notable is the fact that DNase I-hypersensitive sites wererecently identified about 1–2 kb 39of the Ig1 exon (37), raising thepossibility that other germline transcription units may possess astructure similar to the Im/mx region. However, whether additionaltranscripts are indeed generated from theg1 region remains to beshown.

Finally, why should them locus present with such a complex,apparently redundant transcriptional activity? One obvious but te-leological argument is that multiple transcription units may protectagainst the effects of mutations eliminating one or the other. SinceIm mutations could be predicted to block CSR altogether, theywould result in extremely severe immunodeficiency comparable tothat observed in diseases such as hyper-IgM syndrome or commonvariable immunodeficiency.

A more likely explanation is that duplication of the Im/mx tran-scription units may have resulted from the inherent instability of Sregions, which often undergo germline deletion/duplication events(38); this would explain the presence of octamer and E47 bindingsites upstream of both Im and mx18 as remnants of the originalduplicated unit. We should however point out that both the E47and Oct sites in themx18 putative initiation region, although con-forming to published consensus sequences, are not closely relatedto those in the Em enhancer. Finally, as discussed above, an in-herent promoter activity may be associated with the enhancer el-ement identified by Sigurdadottir et al. (34).

In summary, our results highlight a surprisingly complex andunique structure for them locus germline transcript region, withtwo, possibly three Im-like exons. These findings suggest a likelyexplanation for the puzzling result of Em/Im-independent switch-ing, and most importantly provide a new framework for the char-acterization of the elements involved in the control of CSR at them locus.

AcknowledgmentsWe are grateful to Gail Ackermann and Tara Schmidt for microinjectionand maintenance of theDIm-s mutant mouse strain, and to Dick Insel, John

FIGURE 7. Putativemx18 initiation sites.A, A 32P-labeled antisenseRNA probe was generated from the genomic 325-bpHindIII-SacI fragmentthat spans the 59-most 185 bp of themx18-cloned sequence and 140 bp ofits 59 flanking sequence. The location of the probe (indicated by the thickline) compared with themx18 exon sequence is also shown. The probe washybridized to total RNA from normal spleen and LPS blasts. In each lane,beside the completely protected probe, three major bands are detected atlengths of about 300 bp, 275 bp, and 220 bp (arrows).B, Sequence of theregion upstream of themx18 exon. The 59end of the cloned cDNA (3mx18) and the approximate position of the putative major initiation sites (p)are indicated. Underlined sequences mark some potential binding sites forvarious transcription factors, identified from the TRANSFAC database oftranscription factor binding sites (40) using the Genomatics MatInspectorProfessional program (at http://genomatix.gsf.de) and the TFSEARCH pro-gram (Yutaka Akijima, http://www.rwcp.or.jp/papia/). The binding sitescorrespond to the following TRANSFAC matrices: ets-1, M0032; C/EBPb,M00109; c-myb, M0004; oct, M00137; and Th1/E47, M00222. Note thatthe Th1/E47 site corresponds to the binding consensus of heterodimers ofE47 and Thing1, an HLH transcription factor which is probably not ex-pressed in lymphocytes (41). The E47 hemi-site in this sequence is how-ever almost perfectly conserved [AA(T/A)(G/T)CCAG consensus vsAAGGCCAG], suggesting that E47 could bind this site in association withsome other factor.

1456 NEW Im-LIKE GERMLINE TRANSCRIPTS

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Manis, Kathy Seidl, and Eiko Sakai for careful review of this manuscript.A.B. also thanks Bortolo Nardini for inspiration.

References1. Stavnezer, J. 1996. Immunoglobulin class switching.Curr. Opin. Immunol.

8:199.2. Coffman, R. L., D. A. Lebman, and P. Rothman. 1993. The mechanism and

regulation of immunoglobulin isotype switching.Adv. Immunol. 54:229.3. Jung, S., K. Rajewsky, and A. Radbruch. 1993. Shutdown of class switch re-

combination by deletion of a switch region control element.Science 259:984.4. Zhang, J., A. Bottaro, S. Li, V. Stewart, and F. W. Alt. 1993. A selective defect

in IgG2b switching as a result of targeted mutation of the Ig2b promoter andexon.EMBO J. 12:3529.

5. Xu, L., B. Gorham, S. C. Li, A. Bottaro, F. W. Alt, and P. Rothman. 1993.Replacement of germlinee promoter by gene targeting alters control of immu-noglobulin heavy chain class switching.Proc. Natl. Acad. Sci. USA 90:3705.

6. Bottaro, A., R. Lansford, L. Xu, J. Zhang, P. Rothman, and F. W. Alt. 1994. Sregion transcription per se promotes basal IgE class switch recombination butadditional factors regulate the efficiency of the process.EMBO J. 13:665.

7. Lorenz, M., S. Jung, and A. Radbruch. 1995. Switch transcripts in immunoglob-ulin class switching.Science 267:1825.

8. Harriman, G. R., A. Bradley, S. Das, P. Rogers-Fani, and A. C. Davis. 1996. IgAclass switch in Ia exon-deficient mice: role of germline transcription in classswitch recombination.J. Clin. Invest. 97:477.

9. Seidl, K. J., A. Bottaro, A. Vo, J. Zhang, L. Davidson, and F. W. Alt. 1998. Anexpressed neor cassette provides required functions of the Ig2b exon for classswitching.Int. Immunol. 10:1683.

10. Hein, K., M. G. O. Lorenz, G. Siebenkotten, K. Petry, R. Christine, andA. Radbruch. 1998. Processing of switch transcripts is required for targeting ofantibody class switch recombination.J. Exp. Med. 188:2369.

11. Qiu, G., G. Harriman, and J. Stavnezer. 1999. Ia exon-replacement mice syn-thesize a spliced HPRT-Ca transcript which may explain their ability to switchto IgA: inhibition of switching to IgG in these mice.Int. Immunol. 11:37.

12. Hanakahi, L. A., L. A. Dempsey, M. J. Li, and N. Maizels. 1997. Nucleolin is onecomponent of the B cell-specific transcription factor and switch region bindingprotein, LR1.Proc. Natl. Acad. Sci. USA 94:3605.

13. Borggrefe, T., M. Wabl, A. T. Akhmedov, and R. Jessberger. 1998. A B-cell-specific DNA recombination complex.J. Biol. Chem. 273:17025.

14. Dempsey, L. A., H. Sun, L. A. Hanakahi, and N. Maizels. 1999. G4 DNA bindingby LR1 and its subunits, nucleolin and hnRNP D: a role for G-G pairing inimmunoglobulin switch recombination.J. Biol. Chem. 274:1066.

15. Ogawa, H., K. Johzuka, T. Nakagawa, S. H. Leem, and A. H. Hagihara. 1995.Functions of the yeast meiotic recombination genes, MRE11 and MRE2.Adv.Biophys. 31:67.

16. Lennon, G. G., and R. P. Perry. 1985. Cm-containing transcripts initiate hetero-geneously within the IgH enhancer region and contain a novel 59-nontranslatableexon.Nature 318:475.

17. Su, L.-K., and T. Kadesch. 1990. The immunoglobulin heavy-chain enhancerfunctions as the promoter for Im sterile transcription.Mol. Cell. Biol. 10:2619.

18. Schlissel, M., A. Voronova, and D. Baltimore. 1991. Helix-loop-helix transcrip-tion factor E47 activates germ-line immunoglobulin heavy-chain gene transcrip-tion and rearrangement in a pre-T-cell line.Genes Dev. 4:1367.

19. Choi, J. K., C.-P. Shen, H. S. Radomska, L. A. Eckhardt, and T. Kadesch. 1996.E47 activates the Ig-heavy chain and TdT loci in non-B cells.EMBO J. 15:5014.

20. Alt, F. W., N. Rosenberg, V. Enea, E. Siden, and D. Baltimore. 1982. Multipleimmunoglobulin heavy chain gene transcripts in Abelson murine leukemia virus-transformed lymphoid cell lines.Mol. Cell. Biol. 2:386.

21. Nelson, K. J., J. Haimovich, and R. P. Perry. 1983. Characterization of productiveand sterile transcripts from the immunoglobulin heavy chain locus: processing ofmm andms mRNA.Mol. Cell. Biol. 3:1317.

22. Li, S. C., P. B. Rothman, J. Zhang, C. Chan, D. Hirsh, and F. W. Alt. 1994.Expression of Im-Cg hybrid germline transcripts subsequent to immunoglobulinheavy chain class switching.Int. Immunol. 6:491.

23. Winter, E., U. Krawinkel, and A. Radbruch. 1987. Directed Ig class switch re-combination in activated murine B cells.EMBO J. 6:1663.

24. Gu, H., Y. R. Zou, and K. Rajewsky. 1993. Independent control of immunoglob-ulin switch recombination at individual switch regions evidenced through Cre-loxP-mediated gene targeting.Cell 73:1155.

25. Bottaro, A., F. Young, J. Cheng, M. Serwe, F. Sabltzky, and F. W. Alt. 1998.Deletion of the immunoglobulin intronic enhancer and associated matrix-attach-ment regions decreases, but does not abolish, class switching at them locus.Int.Immunol. 10:799.

26. Sakai, E., A. Bottaro, and F. Alt. 1999. The immunoglobulin heavy chain intronicenhancer core region is necessary and sufficient to promote efficient class switchrecombination.Int. Immunol. 11:1709.

27. Mansour, S. L., K. R. Thomas, and M. R. Capecchi. 1988. Disruption of theproto-oncogene int-2 in mouse embryo-derived stem cells: a general strategy fortargeting mutations to non-selectable genes.Nature 336:348.

28. Lakso, M., J. G. Pichel, J. R. Gorman, B. Sauer, Y. Okamoto, E. Lee, F. W. Alt,and H. Westphal. 1996. Efficient in vivo manipulation of mouse genomic se-quences at the zygote stage.Proc. Natl. Acad. Sci. USA 93:5860.

29. Chen, J., F. Young, A. Bottaro, V. Stewart, R. K. Smith, and F. W. Alt. 1993.Mutations of the intronic IgH enhancer and its flanking sequences differentiallyaffect accessibility of the JH locus.EMBO J. 12:4635.

30. Alessandrini, A., and S. V. Desiderio. 1991. Coordination of immunoglobulinDJH transcription and D-to-JH rearrangement by promoter-enhancer approxima-tion. Mol. Cell. Biol. 11:2096.

31. Schlissel, M. S., L. M. Corcoran, and D. Baltimore. 1991. Virus-transformedpre-B cells show ordered activation but not inactivation of immunoglobulin generearrangement and transcription.J. Exp. Med. 173:711.

32. Smith, R. F., B. A. Wiese, M. K. Wojzynski, D. B. Davison, and K. C. Worley.1996. BCM search launcher: an integrated interface to molecular biology database search and analysis services available on the world wide web.Genome Res.6:454.

33. Bachl, J., C. W. Turck, and M. Wabl. 1996. Translatable immunoglobulin germ-line transcript.Eur. J. Immunol. 26:870.

34. Sigurdadottir, D., J. Sohn, J. Kass, and E. Selsing. 1995. Regulatory regions 39ofthe immunoglobulin heavy chain intronic enhancer differentially affect expressionof a heavy chain transgene in resting and activated B cells.J. Immunol. 154:2217.

35. Storb, U., B. Arp, and R. Wilson. 1981. The switch region associated with im-munoglobulin Cm genes is DNase I hypersensitive in T lymphocytes.Nature294:90.

36. Mills, F. C., L. M. Fisher, R. Kuroda, A. M. Ford, and H. J. Gould. 1983. DNaseI hypersensitive sites in the chromatin of humanm immunoglobulin heavy-chaingenes.Nature 306:809.

37. Cunningham, K., H. Ackerly, F. W. Alt, and W. Dunnick. 1998. Potential reg-ulatory elements for germline transcription in or near murine Sg1. Int. Immunol.10:527.

38. Migone, N., J. Feder, H. Cann, B. van West, J. Hwang, N. Takahashi, T. Honjo,A. Piazza, and L. L. Cavalli Sforza. 1983. Multiple DNA fragment polymor-phisms associated with immunoglobulinm chain switch-like regions in man.Proc. Natl. Acad. Sci. USA 80:467.

39. Okada, A., M. Mendelsohn, and F. W. Alt. 1994. Differential activation of tran-scription versus recombination of transgenic T cell receptorb variable regiongene segments in B and T lineage cells.J. Exp. Med. 180:261.

40. Heinemeyer, T., E. Wingender, I. Reuter, H. Hermjakob, A. E. Kel, O. V. Kel,E. V. Ignatieva, E. A. Ananko, O. A. Podkolodnaya, F. A. Kolpakov, et al. 1998.Databases on transcriptional regulation: TRANSFAC, TRRD and COMPEL.Nu-cleic Acids Res. 26:362.

41. Hollenberg, S. M., R. Sternglanz, P. F. Cheng, and H. Weintraub. 1995. Identi-fication of a new family of tissue-specific basic helix-loop-helix proteins with atwo-hybrid system.Mol. Cell. Biol. 15:3813.

1457The Journal of Immunology

by guest on April 13, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from