Apoptosis Regulator Gimap4/IAN1 a Natural Hypomorphic Variant Of

A Natural Hypomorphic Variant of the Apoptosis Regulator Gimap4/IAN1 Christine Carter, Carine Dion, Silke Schnell, W. John Coadwell, Margaret Graham, Lucy Hepburn, Geoffrey This information is current as Morgan, Amanda Hutchings, John C. Pascall, Heinz Jacobs, of September 26, 2021. J. Ross Miller and Geoffrey W. Butcher J Immunol 2007; 179:1784-1795; ; doi: 10.4049/jimmunol.179.3.1784

http://www.jimmunol.org/content/179/3/1784 Downloaded from

References This article cites 36 articles, 18 of which you can access for free at: http://www.jimmunol.org/content/179/3/1784.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 26, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

A Natural Hypomorphic Variant of the Apoptosis Regulator Gimap4/IAN11

Christine Carter,* Carine Dion,2* Silke Schnell,2† W. John Coadwell,* Margaret Graham,* Lucy Hepburn,* Geoffrey Morgan,* Amanda Hutchings,* John C. Pascall,* Heinz Jacobs,† J. Ross Miller,* and Geoffrey W. Butcher3*

The Gimap/IAN family of GTPases has been implicated in the regulation of cell survival, particularly in lymphomyeloid cells. Prosurvival and prodeath properties have been described for different family members. We generated novel serological reagents to study the expression in rats of the prodeath family member Gimap4 (IAN1), which is sharply up-regulated at or soon after the stage of T cell-positive selection in the thymus. During these investigations we were surprised to discover a severe deﬁciency of Gimap4 expression in the inbred Brown Norway (BN) rat. Genetic analysis linked this trait to the Gimap gene cluster on rat chromosome 4, the probable cause being an AT dinucleotide insertion in the BN Gimap4 allele (AT(؉)). This allele encodes a Downloaded from truncated form of Gimap4 that is missing 21 carboxyl-terminal residues relative to wild type. The low protein expression associated with this allele appears to have a posttranscriptional cause, because mRNA expression was apparently normal. Spontaneous and induced apoptosis of BN and wild-type T cells was analyzed in vitro and compared with the recently described mouse Gimap4 (knockout. This revealed a “delayed” apoptosis phenotype similar to but less marked than that of the knockout. The Gimap4 AT(؉ (allele found in BN was shown to be rare in inbred rat strains. Nevertheless, when wild rat DNA samples were studied the AT(؉ http://www.jimmunol.org/ allele was found at a high overall frequency (ϳ30%). This suggests an adaptive signiﬁcance for this hypomorphic allele. The Journal of Immunology, 2007, 179: 1784–1795.

he Gimap/IAN proteins are a family of putative GTPases the anti-apoptotic wild-type Gimap5 protein. Consistent with this in- found in vertebrates. They have relatives in higher plants terpretation, prosurvival properties of human, rat, and mouse Gimap5 T but not in most other organisms studied (1). The impor- have been demonstrated in vitro (9–11). Gimap5 has a predicted tance of the Gimap/IAN GTPase family for T lymphocyte survival transmembrane domain that appears to direct its expression to the became evident from investigations of the rat lymphopenia (lyp) surface of mitochondria (a key site of apoptotic regulation) as well as gene. This recessive mutation causes severe peripheral T lym- to some other internal cell membranes (9, 10). by guest on September 26, 2021 phopenia when homozygous and is an essential susceptibility locus Rats, mice, and humans each express seven or eight Gimap in autoimmune models of both spontaneous diabetes mellitus (the genes clustered tightly on a single autosome. The predicted pro- BioBreeding diabetes prone (BB-DP)4 rat) and eosinophilic bowel teins that they encode share amino-terminal features consisting of disease (the PVG-RT1u,lyp/lyp rat) (2–4). Positional cloning of a GTPase domain and other sequence motifs leading to their clas- lymphopenia identified a single base pair deletion in the Gimap5 sification in the AIG1 family, a sequence-based category named gene. The predicted frameshift leads to a polypeptide product from after a protein involved in responses to bacterial pathogens in this gene that is truncated by about two-thirds compared with wild plants (12). The carboxyl-terminal features of the Gimap proteins type (5, 6). The lymphopenic phenotype observed in Gimap5-de- are more diverse, with some carrying obvious transmembrane do- ficient BB-DP rats is thought to result from the in vivo apoptosis mains. Given the profound impact of Gimap5-deficiency on T lym- of mature thymocytes and T lymphocytes (7, 8) in the absence of phocyte survival, it is important to discover whether all of the Gimap genes are engaged in the regulation of apoptosis as has been suggested in a recent publication by Nitta et al., who also *Babraham Institute, Cambridge, United Kingdom; and †The Netherlands Cancer reported physical interactions between Gimap proteins and mem- Institute, Amsterdam, The Netherlands bers of the Bcl-2 protein family (11). In addition, it will be of Received for publication July 13, 2006. Accepted for publication May 11, 2007. interest to ascertain whether these proteins act separately or in a The costs of publication of this article were defrayed in part by the payment of page coordinated fashion to regulate lymphocyte survival. charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. We previously demonstrated interesting changes in the expres- 1 This work was supported by Competitive Strategic Grant funding from the Bio- sion of the Gimap genes during the development of rat T lympho- technology and Biological Sciences Research Council (BBSRC), BBSRC GAIN cytes (13). Among the most dramatic of these were the changes in Grant 202/GAN13085 (to G.W.B. and J.R.M.), and The Netherlands Cancer Institute the expression of Gimap4 (IAN1), previously reported in mice Grant SFN SFR 2.1.29 (to H.J.). S.S. was supported by a travel allowance from Boehringer Ingelheim Fonds. (14), that shows a substantial rise in expression at the thymic ϩ ϩ 2 C.D. and S.S. contributed equally to this study and are listed alphabetically. CD4 CD8 double positive to mature single positive transition. 3 Address correspondence and reprint requests to Dr. Geoffrey W. Butcher, Babraham Gimap4 lacks an obvious transmembrane domain and appears to Institute, Cambridge, U.K. E-mail address: [email protected] lie downstream of the TCR in an as yet ill-defined signaling path- 4 Abbreviations used in this paper: BB-DP, BioBreeding diabetes prone; BN, Brown way (14–16). Norway; DN, double negative; LN, lymph node; ORF, open reading frame; SNP, We raised Abs against Gimap4 to study its protein expression in single nucleotide polymorphism; UTR, untranslated region. rats and mice (Ref. 13 and current report). Using these reagents, Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 we have encountered an unexpected deficiency in Gimap4 protein www.jimmunol.org The Journal of Immunology 1785 expression in the inbred Brown Norway (BN) rat, a strain used with monoclonal anti-actin (catalog no. A5441; Sigma-Aldrich) as a load- widely in immunological research as well as in other biomedical ing control. disciplines. We also report here functional comparisons of BN rat Real-time PCR T lymphocytes with those from mice carrying a targeted deletion in Gimap4, recently described by one of our laboratories (16). Quantitative differences in Gimap4 message levels were measured using comparative real-time PCR. Lymphocyte subpopulations were flow sorted as described above. RNA was prepared using TRIzol (Invitrogen Life Materials and Methods Technologies) and cDNA was made using SuperScript III reverse tran- Animals scriptase (Invitrogen Life Technologies). Methods are essentially as de-

n u b u scribed previously (13, 19), including the use of two control genes, 6-phos- Rats of the strains BN, PVG-RT1 (BN), PVG-RT1 ,RT7 , PVG-RT1 ,lyp/ phofructokinase C (6PFK) and cirhin. Exon-specific primer pairs were lyp, BB-disease resistant (BB-DR)/Ed (which is also genetically lyp/lyp; designed to span an intron in genomic DNA such that any products from Ref. 17), and DA, as well as some backcross populations (see Results), contaminating genomic DNA could be identified and excluded on the basis were bred and maintained in specific pathogen-free conditions at the of size. Primers for Gimap4 were: 5Ј-GAGCAGCCATGAGCTTGGAAT- Babraham Institute. In unpublished analyses we have confirmed that the Ј Ј Ј u 3 and 5 -TCAACAGGGAACAGCATCCTTG-3 . PCR was performed PVG-RT1 ,lyp/lyp strain, which derives its mutant lyp gene from the Ed- using a SYBR Green kit in accordance with the manufacturer’s instructions inburgh subline of BB-DP rats (8, 17), carries the frameshift mutation in (Applied Biosystems), and samples were amplified using the Chromo4 the Gimap5/Ian5 gene described previously (5, 6). Rats of the strains LEW, system (Bio-Rad). A two-way ANOVA was applied to the data to deter- LOU/C, WKY, and WAG were purchased from Harlan UK. C57BL/6 mice mine the relative expression between BN and wild-type rats for each of the used for Western blotting experiments were bred at the Babraham Institute. cell types tested. The analysis was performed in triplicate on three biolog- Gimap4 knockout mice on the C57BL/6 background (16) and their het- ical samples for each cell population. erozygous littermates (used as controls) were bred in specific pathogen-free animal facilities at The Netherlands Cancer Institute. Animals were used at Pulse-chase analysis of rat Gimap4 variants Downloaded from between 8 and 20 wk of age. All animal usage had been previously ap- proved by the respective ethical review committees. cDNA sequences encoding the rat Gimap4 proteins predicted to be expressed in PVG-RT1n (BN) or BN rats were cloned downstream of the Ab production CMV promoter in plasmid pcDNA3.1/myc-His(Ϫ)A in frame with the C-terminal myc-His tags. The resulting plasmid constructs were transfected Mouse and rat Gimap4 were PCR amplified from splenic or thymic cDNA into the HEK293T cell line using jetPEI transfection reagent (Polyplus and cloned in the GST fusion vector pGEX-4T-1 (Amersham Biosciences). Transfection) according to the manufacturer’s instructions (2 ␮g of plas- http://www.jimmunol.org/ Rats were immunized with purified GST fusion proteins of mouse and rat mid per 35-mm dish of cells). Cells were left overnight and then washed Gimap4 to produce both a polyclonal antiserum and mAbs reactive against with 2 ml of methionine- and cysteine-free DMEM medium supplemented both species. Custom production of a rabbit polyclonal antiserum against rat with 5% (v/v) dialyzed calf serum, 100 U/ml penicillin, and 100 ␮g/ml Gimap4-GST fusion protein was conducted by Harlan Sera-Lab. The rat mAb streptomycin and then incubated at 37°C for1hinthesame medium. The MAC 417 was derived from a fusion of immune splenocytes with the rat medium was replaced with 0.5 ml of the same medium supplemented with Y3Ag1.2.3 plasmacytoma (18) and is of the IgG2b subclass. a 3.7 MBq/ml mixture of [35S]methionine and [35S]cysteine (Tran35S-la- ELISAs bel; ICN Pharmaceuticals), and the cells were then incubated at 37°C for 15 min. After this time, the radioactive medium was replaced with 0.5 ml The wells of flat-bottom 96-well plates were incubated overnight at 4°C DMEM containing 10% FCS, 1 mM methionine, and 1 mM cysteine for with recombinant GST, rat Gimap4-GST, or mouse Gimap4-GST at 2.5 various times as indicated. At each time point, cells were washed with 1 ml

␮g/ml in 0.1 M sodium bicarbonate (pH 8.3). The wells were washed with of PBS and lysed into 1 ml of ice-cold CHAPS lysis buffer (10 mM by guest on September 26, 2021 water and then blocked with 2% FCS and 0.01% Tween 20 in PBS for 90 HEPES, 150 mM sodium chloride, and 1% (w/v) CHAPS) containing one min. at 37°C. After washing, the wells were incubated with MAC 417 tablet per 10 ml of complete Mini EDTA-free protease inhibitor mixture hybridoma supernatant for1hat37°C. Binding was detected using a bi- (Roche) for 10 min. The resultant cell lysates were centrifuged at 11,000 ϫ otin-rabbit anti-rat IgG (Dako E0468; mouse absorbed) followed by a g for 5 min and the supernatant was added to 1 ␮l of rabbit anti-rat streptavidin-HRP complex (catalog no. RPN 1051; Amersham Bio- Gimap4-GST fusion protein antiserum (see above) diluted in 100 ␮lof sciences), both diluted in blocking buffer. The HRP substrate 3,3Ј5,5Јtetra- CHAPS lysis buffer and 50 ␮l of 1:1 (v/v) suspension of protein A-Sepha- methylbenzidine (T2885; Sigma-Aldrich) was used for color development. rose beads, and the suspension was rotated overnight at 4°C. The beads The results were read at 450 nm on a Multiskan EX plate reader were then washed with 5 ϫ 1 ml CHAPS lysis buffer at 4°C and bound (Labsystems). protein was eluted into 100 ␮lof2ϫ Laemmli sample buffer by heating to 100°C for 5 min. Proteins were separated on 10% SDS polyacrylamide FACS sorting of lymphocyte subsets gels. The gels were then fluorographed, dried, and exposed to x-ray film. Single cell suspensions of rat thymocytes or lymph node (LN) cells were DNA sequencing washed three times with PBS containing 5% FCS and then incubated with primary mAbs for 30 min. at 4°C. Cells were washed twice and then incubated Genomic DNA samples from tissues of inbred rat strains or wild rats were with secondary reagents. After final washing, the cells were filtered through a either prepared at the Babraham Institute using the DNeasy tissue kit (Qia- 40-␮m filter immediately before cell sorting. The mAbs used were W3/25 gen) or generously supplied by the collaborators indicated in Tables III and (anti-CD4), 341 (anti-CD8␤), MARK1 (anti-Ig␬ L chain), and MARM4 (anti- IV. Gene-specific primers and PCR were used to amplify Gimap4 from IgM). These mAbs came either from Serotec or from in-house stocks and were genomic DNA. The resulting PCR products were treated with exonuclease either FITC- or biotin-conjugated. In the latter case, PE-Cy5.5-conjugated I and shrimp alkaline phosphatase (Amersham Biosciences) before being streptavidin (eBioscience) was used as a secondary fluorescent reagent. Cell sent for commercial sequencing (Lark Technologies). populations were sorted using FACSDiva or FACSAria from BD Biosciences. Purities were checked to be between 90 and 99.9%. 3Ј-RACE PCR Western blotting Total RNA was prepared using TRIzol reagent (Invitrogen Life Technol- ogies) from the LN cells of BN and PVG-RT1u, RT7b rats and from the Cell samples at a concentration of 2 ϫ 108/ml were incubated in lysis splenocytes of C57BL/6 mice. Complementary DNA was then synthesized buffer (2% Nonidet P-40, 20 mM Tris, 150 mM NaCl, and 1 mM MgCl2 using Superscript III (Invitrogen Life Technologies) and an oligo(dT) an- (pH 8) supplemented with proteolytic inhibitors) for 30 min at 4°C. Nuclei chor primer (5Ј-GACCACGCGTATCGATGTCGACTTTTTTTTTTTTTT were removed by centrifugation at 14,000 ϫ g. An equal volume of Lae- TTV-3Ј, where V represents A, C, or G). A 1-␮l aliquot of the cDNA mmli sample buffer (Bio-Rad) was added to each sample. Ten-microliter reaction (final volume 50 ␮l) was then used for RACE-PCR using a 3Ј PCR samples, equivalent to 1 ϫ 106 cells, were run on 15% SDS-polyacryl- anchor primer (5Ј-GACCACGCGTATCGATGTCGAC-3Ј) and an appro- amide gels and then blotted onto Immobilon P membranes (Millipore). priate 5Ј gene-specific primer for rat Gimap4 (5Ј-GAGAGAGAGAGAGC Blots were blocked overnight at 4°C with 4% milk powder in PBS with ACGGATAAG-3Ј) and mouse Gimap4 (5Ј-AGAGGGAATTCAAGGAGA 0.1% Tween and incubated with primary and secondary Abs for1hat GG-3Ј). After the completion of PCR, the products were treated with room temperature. Blots were developed with a SuperSignal West Pico shrimp alkaline phosphatase and exonuclease 1 to remove unincorporated chemiluminescent substrate (Pierce). After staining for Gimap4, blots were primers and dNTPs. A 1.5-␮l aliquot of reaction product was then used for stripped with Restore Western Blot stripping buffer (Pierce) and stained a second PCR using the 3Ј anchor primer (see above) and gene-specific 1786 CHARACTERIZATION OF A HYPOMORPHIC Gimap4 ALLELE

“nested” primers for rat Gimap4 (5Ј-CAAGATCTGAGAGATGAGCTGG- 3Ј) and mouse Gimap4 (5Ј-AAGGAAGTTGAGAACACAAGTATG-3Ј). Parallel reactions were conducted on genomic DNAs in which case no products were observed, indicating that the RACE-PCR products were derived from cDNA. Second-round RACE-PCR products were then treated, as above, with alkaline phosphatase and exonuclease 1. The products were then sequenced using the second-round gene-speciﬁc primers. Additional sequencing primers were subsequently used to extend the sequence; these were 5Ј-GCAACTCAAAGCAGGAACCTG-3Ј for rat and 5Ј-AGTTCC AATCTTTGCATGAG-3Ј for mouse. Bioinformatics DNA repeats were analyzed using RepeatMasker (http://repeatmasker. genome.washington.edu/). Other bioinformatic tools used were ClustalX (20) and Dotter (21). In vitro survival assays For the experiment shown in Fig. 8A, rat LN T cells were ﬂow sorted as described above and purities were checked to be between 91 and 99%. Triplicate samples of cells from each strain were incubated in 24-well plates at 5 ϫ 105 cells per milliliter in RPMI 1640 medium supplemented

with 20% FCS, 2-ME, and nonessential amino acids. In vitro cell survival Downloaded from was calculated from daily counts of viable cells in each sample. The data shown are representative of three separate experiments. For the experiment shown in Fig. 8B, mouse T cells were isolated from LNs using a BD Biosciences T lymphocyte enrichment set. For the isolation of rat T cells a mixture comprising the following biotin-conjugated mAbs was used: anti- rat B220 (clone MRC OX33), anti-rat NK cells (clone 3.2.3), anti-rat mac- rophage (MRC OX42) and anti-rat Ig␬ (RG7/9.1). Cells bearing bound biotinylated mAbs were retained on magnetic nanoparticles as for mouse T http://www.jimmunol.org/ cells. In this manner, cells were 90–96% pure as determined by flow cytometric analysis. T cells (0.2 ϫ 106 cells per well) were subjected to various apoptotic stimuli in 96-well plates for up to 3 days. The kinetics of apoptosis was determined by flow cytometry gating on T cells (anti-rat ␣␤TCR; clone R73). Apoptotic cells were identified by annexin V and FIGURE 1. Gimap4 expression. A, ELISA showing reactivity of mAb propidium iodide staining according to the apoptosis detection kit from BD MAC 417 with rat (r) and mouse (m) GST-Gimap4 fusion proteins. B, Biosciences. Analysis was performed on a FACSCalibur using CellQuest Western blot analysis of Gimap4 in splenocytes from 10 rat strains: Lane software. 1, LEW; lane 2, PVG-RT1n(BN); lane 3, PVG-RT1u,RT7b; lane 4, PVG- RT1u,lyp/lyp; lane 5, BB-DR/Ed; lane 6, WKY; lane 7, WAG; lane 8, BN; Subcellular fractionation lane 9, DA; and lane 10, LOU/C. C, Western blot analysis of BN Gimap4 by guest on September 26, 2021 PVG-RT1u,RT7b and BN LN cells were washed twice with PBS and lysed expression using different reagents. Rat antiserum (panel I), mAb MAC by sonication in 1 ml of ice-cold Break buffer (20 mM HEPES, 1 mM 417 (panel II), and rabbit antiserum (panel III) were used to probe lysates ␤ n EGTA, 0.5 mM MgCl2, 10 mM NaF, 10 mM -glycerophosphate, 0.13 M from unseparated LN cells from PVG-RT1 (BN) (lanes 1) and BN rats sucrose, 50 mM NaCl, and 1% protease inhibitor mixture (catalog no. (lanes 2) and a C57BL/6 mouse (lanes 3). The lysates in this blot were P8340; Sigma-Aldrich). An aliquot of the sample was reserved as the “total subjected to nuclear spinout, but a duplicate blot with total lysate gave com- lysate” and the remainder was centrifuged at 250 ϫ g for 5 min. at 4°C in parable results (data not shown). In blots B and C the top panel shows Gimap4 a swing-out rotor to pellet the nuclei. The supernatant was then ultracen- trifuged in a swing-out rotor (Beckman TLS-55) at 195,000 ϫ g for1h.at expression and the lower panel is stained with monoclonal anti-actin as a 6 4°C to obtain the cytosol and membrane fractions. The nuclei and membranes loading control. Lysate from 1 ϫ 10 cells was loaded in each lane. were washed twice with Break buffer. All fractions were resuspended in equivalent volumes of Break buffer and the nuclear and membrane pellets were lysed by sonication for analysis by Western blotting. Blots were probed with with Gimap4 being detected predominantly as a ϳ38-kDa band. anti-Gimap4 (MAC 417), and anti-Hsp90 (BD Biosciences) and anti-VDAC-1 (Calbiochem) served as fractionation controls. Some rat strains gave very weak signals. Two of these strains, PVG-RT1u, lyp/lyp and BB-DR/Ed, are inherently T lymphopenic Intracellular staining due to a mutation of their Gimap5 (lyp) gene and were previously Rat T cells were isolated by magnetic selection as described above. Cells noted to exhibit poor Gimap4 expression in peripheral lympho- were treated with fixation buffer (catalog no. 00-8222-49; eBioscience) for cytes (13). The poor expression of Gimap4 by the BN rat spleen, 15 min on ice and then incubated with permeabilization buffer (00-8333- however, was unexpected. To confirm this finding, similar tests 56;eBioscience). Intracellular staining was detected by using FITC-conjugated IgGs from normal rabbit antiserum as a negative control and from were conducted using two polyclonal antisera against Gimap4 anti-Gimap4 rabbit antiserum. Staining was detected by flow cytometry raised in rats and rabbits, respectively (Fig. 1C). These tests also using either a FACSCalibur or LSRII analyser from BD Biosciences. showed weak Gimap4 signals in BN rats compared with a representative “normal” rat (PVG-RT1n). The conclusion that these re- Results sults were due to a genuine deficiency of the Gimap4 protein in BN Strain variation in rat Gimap4 expression rats and not just the absence of a strain-dependent serological To investigate Gimap4 protein expression in rat lymphocytes, we epitope was strengthened by the fact that the two polyclonal reagents raised rat antisera and rat mAbs against recombinant (rat plus displayed different serological behavior, i.e., the rabbit anti-rat anti- mouse) Gimap4 and a rabbit antiserum against rat Gimap4 alone. serum cross-reacted much more weakly on mouse Gimap4, indicating The mAb used in this study is designated MAC 417. It reacts with that they possessed distinct epitope specificities (Fig. 1C). both rat and mouse Gimap4 as shown by ELISA (Fig. 1A) and Western blotting (Fig. 1C). Expression of Gimap4 in lymphocyte subsets MAC 417 was used to test a number of different inbred rat The deficiency in expression of Gimap4 by BN rats was investi- strains for the presence of Gimap4 in spleen cell lysates (Fig. 1B), gated further by analyzing subsets of purified lymphocytes. Flow The Journal of Immunology 1787

FIGURE 2. Protein expression of Gimap4 in FACS-sorted populations of thymocytes and lymphocytes in BN vs wild-type (WT) rats. Gimap4 expres- u b ϫ u b ϫ sion was compared in BN, PVG-RT1 ,RT7 (WT.), and (BN PVG-RT1 ,RT7 )F1 (WT BN)F1) rats using MAC 417 in Western blots. Thymocytes were sorted into CD4ϪCD8Ϫ DN, double-positive (DP), CD4ϩ and CD8ϩ single-positive populations (95–99.8% purity) and LN cells of similar purity were sorted as CD4ϩ and CD8ϩ T cells, and B cells (B). The equivalent of 1 ϫ 106 lysed cells was loaded per lane. The blots were stripped and stained with monoclonal anti-actin Ab as a loading control. The results shown are representative of four experiments. cytometry was used to generate these cells from the thymuses and acids of the wild type and that is 18 amino acids shorter than the u b Ϫ ϩ Downloaded from LNs of BN and PVG-RT1 , RT7 rats as well as F1 hybrids of these AT( ) form. The BN genome encodes the AT( ) form; the al- two strains. Western blots of the “wild-type” PVG-RT1u,RT7b ternative AT(Ϫ) form is similar to the Gimap4 sequences seen in cells showed the proﬁle predicted from earlier studies in mice and other species, including mouse and human. When Gimap4 was rats (14, 16, 13), namely the absence of Gimap4 expression in sequenced from three independent rat strains, BN, PVG, and DA, CD4ϩCD8ϩ double-positive thymocytes, strong expression in sin- no coding differences other than the AT dinucleotide insertion gle-positive thymocytes and peripheral T cells, and weaker but were encountered although a small number of single nucleotide

signiﬁcant expression in the double-negative (DN) subset (Fig. 2). polymorphisms (SNPs) were recorded. These are annotated in the http://www.jimmunol.org/ Expression by B lymphocytes was low. Similar analysis was per- European Molecular Biology Laboratory (EMBL) Nucleotide Se- formed on BN subsets. Very low levels of Gimap4 were detectable quence Database (www.ebi.ac.uk/embl/) under accession nos. in BN rats although the same subset differences were discernible as BC070952, AM285343, and AM285683, respectively. for the “wild type” strain, albeit at a lower level of signal. The Ј absence of a Gimap4 signal in DN thymocytes of BN shown in A deletion in the 3 -untranslated region (UTR) of rat Gimap4 Fig. 2 is presumably due to the overall weakness of signals in this Bioinformatic analysis suggested a second unusual feature of strain and does not indicate a difference in the pattern of expression the rat Gimap4 gene. When human and mouse sequences com- of this gene in BN rats. Indeed a weak band has been seen for BN prising the last exon of Gimap4 were compared, substantial DN thymocytes in some experiments using longer exposure. The sequence conservation extending through the 3Ј-UTR to the by guest on September 26, 2021

F1 hybrid rat samples gave strong Gimap4 signals, indicating that poly(A) signal was observed (Fig. 4). When rat Gimap4 was the deﬁciency observed in BN rats is not genetically dominant but entered into the same comparison, however, all sequence con- probably codominant. servation ended some 16 nucleotides 3Ј of the normal (non-BN) stop codon. A subsequent alignment of sequences from human, The BN rat has a C-terminal AT insertion chimpanzee, orangoutang, cow, dog, mouse, and rat indicated Bioinformatic searches revealed two forms of rat Gimap4 (Fig. 3). that the rat has acquired a deletion at this point relative to all the These differ with respect to an AT dinucleotide insertion toward other mammals analyzed. Table I presents the alignment of hu- the 3Ј end of the transcript but within the open reading frame man, mouse, and rat sequences. To conﬁrm this in silico infor- (ORF). As a result of this frameshift the form with the AT insertion mation, we undertook a 3Ј RACE-PCR on Gimap4 from mouse (AT(ϩ)) contains an earlier stop codon, thereby generating a pre- and rat using primers located ϳ150 bp 5Ј of the normal rat/ dicted polypeptide that has lost the 21 carboxyl-terminal amino mouse stop codon. This analysis was done on total RNA from

FIGURE 3. Different carboxyl termini of the two rat, the mouse, and the human Gimap4 variants. A, Structure of Gimap4. The Gimap4 polypeptide has a GTP binding domain (motifs G1–G5) with sequence features within the indicated region placing it in the AIG1 domain family. A calmodulin interaction domain (IQ) and consensus protein kinase C sites (✦) have been identiﬁed (16). B, A comparison of the predicted C-terminal amino acid sequences of mouse (National Center for Biotechnology Information GenBank accession no. NM_174990; www.ncbi.nih.gov), human (EMBL accession no. AK001972), and the two rat forms (EMBL accession nos. BC070952 (short form) and AM285343 and AM285683 (long form)) of Gimap4. 1788 CHARACTERIZATION OF A HYPOMORPHIC Gimap4 ALLELE

Linkage of low Gimap4 expression to the AT dinucleotide insertion PVG congenic rats were backcrossed on to the BN background to generate segregant animals homozygous and heterozygous for BN- derived alleles. These animals were typed for the presence of the Gimap4 AT dinucleotide insertion by combined genomic PCR and DNA sequencing, while tissues or blood leukocytes were collected to test for Gimap4 protein expression by Western blotting with mAb MAC 417. The results are summarized in Table II, and examples of the Western blotting analysis are shown in Fig. 5. A total of 49 rats was analyzed of which 48 showed a direct correlation of AT dinucleotide status with protein expression level, i.e., 27 AT(ϩ)/AT(Ϫ) heterozygotes showing high protein expression and 21 AT(ϩ)/AT(ϩ) homozygotes showing low protein expression. n ϫ ϫ Western blots from 18 of the (PVG-RT1 (BN) BN)F1 BN rats are shown in Fig. 5A. The AT(ϩ)/AT(Ϫ) heterozygotes for Gimap4 gave strong signals in their LN tissue and weaker but easily visible signals in thymocytes. By contrast, all but one of the AT(ϩ)/AT(ϩ) homozygotes gave either weak bands in LN only or Downloaded from an absence of signal from both the LN and the thymus. The data from a single exceptional rat, an AT(ϩ)/AT(ϩ) rat with high protein expression, are included in Fig. 5A (rat no. 10). Because this Western blotting was performed on tissues post mortem, it was not possible to progeny test this individual. The signiﬁcance of this ﬁnding is discussed later. Nevertheless, overall these data clearly http://www.jimmunol.org/ show that Gimap4 protein expression appears to behave as a Men- delian trait under the control of a gene or genes in, or close to, the FIGURE 4. Dotter plots showing sequence homology in rat, mouse, and Gimap complex. human Gimap4 genes. The rat and mouse forms of Gimap4 are highly conserved within the open reading frame, but in the 3Ј-UTR there is a BN T cells have a common low level of Gimap4 expression ϳ1800-bp deletion in the rat as compared with the mouse. The upper panel The observed very low level of Gimap4 in BN T cells could be due shows the rat vs mouse comparison, depicting the deletion in the rat gene either to all BN T cells expressing low levels of this protein or the starting in exon 3 and extending beyond it. The lower panel shows a com- presence of a small number of T cells expressing normal or high by guest on September 26, 2021 parison of human vs mouse. levels of Gimap4. We conducted subcellular fractionation on PVG-RT1u,RT7b thymocytes to determine the cellular localization of Gimap4 by Western blotting. Abs against Hsp90 and voltage-

u b dependent anion channel 1 (VDAC-1) were used as fractionation a C57BL/6 mouse and from both BN and PVG-RT1 ,RT7 rats. markers for cytoplasm and membranes, respectively (data not We studied two different rat strains as we wished to discover shown). The results in Fig. 6A show that rat Gimap4 is cytosolic in whether, other than the AT insertion, the BN rat had any addi- wild-type cells and, despite the much lower level of expression, the tional sequence differences in this region. same localization is observed in BN cells. Next, intracellular stain- Our results confirmed as correct the existing database sequences ing and flow cytometry were used to study Gimap4 expression in containing the Gimap4 3Ј-UTRs for both mouse and rat. Hence, purified LN T cells from BN and PVG-RT1n(BN) rats. The results we verify that the rat Gimap4 sequence has undergone a ϳ1800-bp in Fig. 6B show that all BN T cells express very low levels of deletion relative to other mammals immediately downstream of the Gimap4 (note that the BN profile is slightly shifted relative to the termination codon, with the consequence that a DNA repeat region control), whereas all PVG-RT1n(BN) T cells express a high level containing an MTC (mammalian apparent long-terminal repeat ret- of Gimap4. This excludes the possibility that there is a subpopu- rotransposon (MaLR)) repeat and a rat MER1B repeat is brought lation of high-expressing cells. in juxtaposition to the end of exon 3 of rat Gimap4. The distinct locations of the poly(A) addition sites in rat and mouse were as Transcript levels in BN vs wild type rats predicted by bioinformatics (Table I) and indicate different mes- Having defined large differences in Gimap4 protein expression in sage sizes for the two species, these being 1.4 kb in rat and 1.7 kb immune cells of BN compared with other rats, we wished to de- in mouse, the latter agreeing with previous Northern blot data (14, termine whether or not these were reflected at the level of mRNA Ј u b 16). When we compared the 3 -UTRs of BN and PVG-RT1 ,RT7 expression. Gimap4 message levels in flow-sorted populations of with each other, we found just three polymorphisms in addition to LN T and B cells and CD4ϩ and CD8ϩ single-positive thymocytes the AT insertion within the ORF. The three SNPs are located at from BN and PVG-RT1n(BN) rats were measured using com- positions 1052, 1134, and 1219 in the 3Ј-UTR of the PVG rat parative real-time PCR. Transcript expression was measured (GenBank accession no. AM285343). Because the genomic dele- relative to two control genes, cirhin and 6-phosphofructokinase tion observed is common to both AT(ϩ) and AT(Ϫ) rats, it cannot C (6PFKc), which we had previously selected for their stable ex- be directly responsible for the deficiency in Gimap4 at the protein pression in the tissues studied (13). In both rat strains the highest level in BN rats. Searching the rat genome database with the Gimap4 expression was seen in LN T cells, whereas thymic CD4 mouse sequence that represents the 1800-bp deletion revealed no and CD8 cells had lower and relatively similar expression. B cells significant hits. in both rats had significantly less Gimap4 message than lymphoid The Journal of Immunology 1789

Table I. Multiple sequence alignment of the third exon of Gimap4 (Ian01) from human, mouse, and rata Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

a Symbols are as follows: **, BN rat dinucleotide insertion; Ter BN, termination codon used in the BN rat; Ter, termination codon used in human, mouse, and other rats (all termination codons are shown in lower case); $, start of deletion in rat UTR; ******, poly(A) addition signals for rat (r) and mouse/human (m,h). 1790 CHARACTERIZATION OF A HYPOMORPHIC Gimap4 ALLELE

Table II. Summary of Gimap4 genotyping and protein phenotyping of 49 segregants from a number of backcross litters

No. of Phenotype Backcross Rats Genotype Tissues Blotted (Gimap4 Protein)

(PVG-RT1n (BN) ϫ BN)F1 ϫ BN 24 9 ϫ AT(ϩ)/AT(ϩ) LN, spleen, thymus 8 ϫ Low 15 ϫ AT(ϩ)/AT(Ϫ)16ϫ Higha

(PVG-RT1n (BN) ϫ BN)F1 ϫ BN 94ϫ AT(ϩ)/AT(ϩ) PBLs 4 ϫ Low 5 ϫ AT(ϩ)/AT(Ϫ)5ϫ High

Second backcross of wild-type Gimap4 16 10 ϫ AT(ϩ)/AT(ϩ) PBLs 10 ϫ Low from PVG-RT1u RT7b on to BN 6 ϫ AT(ϩ)/AT(Ϫ)6ϫ High

a Denotes the backcross in which the genotype and phenotype were different for one rat.

T cells. These results were consistent with our previously pub- labeled with a short pulse of [35S]methionine and [35S]cysteine. lished data on PVG-RT1u,RT7b and PVG-RT1u,lyp/lyp rats (13). The radiolabel was then chased out of the cells with a large excess Two-way ANOVA analysis was performed to determine whether of unlabeled amino acids as described in Materials and Methods. there was a difference in mRNA expression between the two At various intervals during the chase, cell lysates were prepared strains of rats in four types of cells (Fig. 7A). No large differences and the radiolabeled Gimap4 was recovered from the lysates by Downloaded from in message levels were observed. BN lymphocytes may express immunoprecipitation. As shown in Fig. 7B, radiolabeled Gimap4 slightly less mRNA, but this difference was of weak signiﬁcance was still detectable after a 6-h chase period whether the Gimap4 ( p ϭ 0.09). Similarly, no large differences in mRNA levels were variant expressed was that from BN or PVG-RT1n(BN) rats. In observed when spleen RNA isolated from BN and PVG-RT1n(BN) addition, the rate of disappearance of radiolabel from the two vari- rats was analyzed by Northern blot analysis using a probe for rat ants appears to be grossly similar, suggesting that the half-lives of

Gimap4 mRNA. In addition, no Gimap4 mRNA size difference the two proteins are similar. Similar results were also obtained http://www.jimmunol.org/ was seen between the two strains (results not shown). when the BN variant was tagged at the N terminus or when it was expressed in its native form, suggesting that tagging the protein Stability of the variant Gimap4 proteins expressed by BN and had no major effect on its turnover rate (results not shown). PVG-RT1n(BN) rats Having determined that differences in Gimap4 mRNA levels were T lineage cells from BN rats display an unusual apoptotic so small as to be unlikely to explain the observed differences in pro- phenotype in vitro tein levels between wild-type and BN rats, we were interested to see Roles in the regulation of cell death have been proposed or dem- whether the two proteins made had inherently different stabilities. onstrated for members of the Gimap family, most notably in the To address this, the HEK293T cell line was transfected with prosurvival properties of rat Gimap5 (see Introduction); by con- by guest on September 26, 2021 plasmids encoding the variant forms of Gimap4 (each with a C- trast, recent studies have suggested a prodeath role for Gimap4 in terminal myc-His tag; see Materials and Methods) under the con- the mouse (11, 16). Initial analysis showed no obvious abnormality trol of a CMV promoter. Newly synthesized proteins were radio- in the rate of cell death when FACS-puriﬁed BN T cells were

FIGURE 5. Western blots phenotyping BN ϫ PVG (wild type) backcross rats. A, Tissues from 18 (PVG- n ϫ ϫ RT1 (BN) BN)F1 BN backcross rats were analyzed by Western blotting using mAb MAC 417 to stain for rat Gimap4 and anti-actin as a loading control. One million unseparated LN cells and thymocytes were blotted for each rat. AT status is indicated as (ϩ/ϩ) for homozygous AT(ϩ)/AT(ϩ) and (ϩ/Ϫ) for heterozygotes. With the exception of rat no. 10 (see Results), heterozygous individuals exhibited a heavy band for LN and a weaker band for thymocytes. AT(ϩ)/AT(ϩ) homozygotes were typiﬁed by either a weak band or no band in LN tissue only. B, Demonstration that rats may also be phenotyped using 106 PBLs per lane. In Western n ϫ ϫ blots from nine (PVG-RT1 (BN) BN)F1 BN backcross rats, the AT(ϩ)/AT(Ϫ) heterozygotes exhibited a strong Gimap4 band (lanes b, c, g, h, and i), whereas no band was detected for the AT(ϩ)/AT(ϩ) homozygotes (lanes a, d, e, and f). The Journal of Immunology 1791

FIGURE 6. A, Subcellular distribution of Gimap4. PVG-RT1u,RT7b and BN LN cells were fractionated into nuclei (N), cytoplasm (C), membrane (M), and total lysate (TL). To enable us to see the BN bands clearly without overloading the wild-type samples, 3-fold less wild-type material was used. Gimap4 was detected by Western blotting. B, Intracellular staining of LN T cells. Puriﬁed T cells from PVG-RT1n and BN rats were stained with FITC-normal rabbit IgG (solid line) and FITC-anti-Gimap4 rabbit IgG (dotted line, shaded area). Downloaded from cultured for 3 days in tissue culture medium (Fig. 8A). The num- results are presented in Table IV. Genomic DNA samples were gen- bers of live BN LN T cells declined at approximately the same rate erously donated by colleagues in the United Kingdom, Europe and as for wild-type cells (PVG-RT1n) and clearly more slowly than Japan and were added to a panel we had collected earlier within our Gimap5-deﬁcient T cells (PVG-RT1u,lyp/lyp) (8). own region (see the legend to Table IV for the provenances of the A more detailed analysis of in vitro apoptotic responses of BN samples tested). Four of the samples from Japan, i.e., those from

T cells was then performed. This was guided by the recently pub- source F in Table IV, were from rat strains derived by inbreeding http://www.jimmunol.org/ lished analysis of a mouse strain with a targeted deletion in from wild-caught animals and thus represented only one haplotype Gimap4 (16). In this study, one of our groups had noticed that T per strain; the remaining 109 DNA samples sequenced were directly cells from the knockout mice exhibited slower kinetics in the in- from captured or killed wild animals and thus represented two hap- duced “death program” as defined by flow cytometry. Cells ap- lotypes per individual. The number of apparent AT(ϩ) homozygotes peared to be delayed in proceeding from the annexin V (ϩ), pro- was 25/109, while the number of heterozygotes was 20/109. Thus, the pidium iodide (Ϫ (“apoptotic”) stage to the annexin V (ϩ), AT(ϩ) Gimap4 DNA sequence was found in 45 of the 109 samples propidium iodide (ϩ) (“dead”) stage. We therefore performed the and the overall frequency of the AT(ϩ) haplotype was ϳ32%. The same experiment to ask whether T cells from Gimap4-deficient BN rats (AT(ϩ)) would exhibit a similar apoptotic phenotype. As pos- by guest on September 26, 2021 itive controls in these studies we used LN T cells from Gimap4Ϫ/Ϫ knockout mice (Fig. 8, B and C). As seen previously using mouse splenic T cells (16), in response to various apoptotic stimuli more Gimap4Ϫ/Ϫ knockout mouse LN cells tended to accumulate at the annexin V (ϩ), propidium iodide (Ϫ) (“apoptotic”) stage than did wild-type cells when assayed by flow cytometry for up to 2 days. The Gimap4-deficient BN rat LN T cells showed similar excesses of cells at the “apoptotic” stage as compared with “normal” rats when using strong apoptotic stimuli, i.e., dexamethasone, etoposide, and gamma irradiation (Fig. 8, B and C). By contrast, no significant differences were seen between BN and normal T cells in response to a weaker apoptotic stimulus, i.e., serum-free culture (data not shown). Overall, the differences seen for Gimap4Ϫ/Ϫ knockout and wild-type mouse T cells were somewhat more marked than for Gimap4-deficient and normal rats. Gimap4 variants in inbred rat strains Gimap4 polymorphism was analyzed by DNA sequencing of a panel of 28 inbred laboratory rat strains (Table III). Only two strains, BN and MAXX, were found to carry the AT(ϩ) dinucleotide insertion. Six SNPs within Gimap4 exon 3 defined two major FIGURE 7. Gimap4 mRNA levels and protein stability. A, Relative ex- n haplotypes. Only the first three SNPs are in the ORF and are silent, pression of Gimap4 transcript levels in BN and PVG-RT1 (WT) rats in 4 whereas the remainder lie in the 3Ј-UTR (EMBL accession nos. cell types. Real-time PCR results are expressed as a two-way ANOVA AM285343 and AM285683). Table III organizes the inbred rat strains relative to two housekeeping genes (HGK). Each value is calculated as the mean of triplicate samples of three separate cell preparations. Error bars into the two haplotype groups, with 18 strains in haplotype group 1 show the mean Ϯ 1.0 SE. B, Relative stability of Gimap4 variant proteins and 10 in group 2. BN and MAXX define a subtype of haplotype 1. from BN and PVG-RT1n (WT). Pulse-chase analysis of Gimap4 protein Gimap4 variants in wild rats stability in transfected HEK293T cells was performed as indicated in Ma- terials and Methods. The radiolabeled Gimap4 species are indicated by an Investigations were extended to determine whether the AT(ϩ) form arrowhead. The mobilities of coelectrophoresed protein m.w. standards are of Gimap4 was present in wild Rattus norvegicus populations. The indicated. 1792 CHARACTERIZATION OF A HYPOMORPHIC Gimap4 ALLELE

FIGURE 8. Survival of lymphocytes and thymocytes in vitro. A, In vitro survival curves for BN, PVG-RT1n(BN), and PVG.RT1u,lyp/lyp LN T cells. B, Kinetics of stress-induced T cell death from BN rats. Puriﬁed LN T cells from

BN (open squares, continuous line) and PVG- Downloaded from RT1n (ﬁlled squares, broken line) rats were treated with the apoptotic stimuli dexamethasone (1 ␮M), etoposide (10 ␮g/ml), and gamma (␥) irradiation (2 Gray). Live, apoptotic, or dead cells were distinguished by ﬂuorochrome-annexin V and propidium iodide labeling. The data represent the means of three independent ani- http://www.jimmunol.org/ mals per strain. As a positive control, pooled LN T cells from three Gimap4 knockout mice (open circles, continuous line) and their heterozygous littermates (solid circles, broken line) were used (16). C, Flow cytometric analysis of T cells incubated with dexamethasone as described above. As examples, this panel shows the results for T cells from one animal per strain incubated with dexamethasone. by guest on September 26, 2021

AT(ϩ) haplotype was not detected in samples from some geographic associated with intrathymic ␤-selection and positive selection, fol- regions/locations, e.g., Cambridgeshire and Essex, U.K., and was, by lowed by high expression in peripheral resting T cells (14, 16, 13, contrast, common in some others, e.g., Lund, Sweden and Osaka, 27). This is clearly confirmed in the rat at the protein level, as seen Japan. We also detected both AT(ϩ) and AT(Ϫ) forms of Gimap4 in in Fig. 2. Interestingly, a Gimap4 knockout mouse shows no ob- Rattus rattus DNA samples. vious phenotype in lymphocyte development and selection. Thus, Gimap4 deletion had apparently no effect on thymic positive or Discussion negative selection in model (anti-HY) systems (16). Similarly, the Using newly developed serological reagents, we have discovered a successful generation of mature thymocytes and a peripheral T cell severe deficiency of the Gimap4 GTP binding protein in T cells of repertoire in the BN rat suggests that Gimap4 does not play a the BN rat. This deficiency is linked genetically to the region con- critical role in thymic selection events. This conclusion contrasts taining the Gimap gene cluster on rat chromosome 4. Recent ev- with suggestions made by Nitta and colleagues on the basis of the idence has suggested that Gimap4 is a positive regulator of cell overexpression of Gimap4 in mouse fetal thymic organ culture in death (11, 16). This may be achieved through interaction with the vitro (11). We suggest that the apoptosis of double-positive thy- Bcl-2 family members Bax/Bak (11). Gimap4 has an intriguing mocytes observed in that study after the forced expression of expression profile across T cell development, with sharp increases Gimap4 in DN thymocytes is a nonphysiological phenomenon. The Journal of Immunology 1793

Table III. Genotyping of a panel of inbred rat strains to show Gimap4 expression and genotyping for the AT insertion. Although this may AT status and haplotype as deﬁned by a SNP close to the AT insertion indicate control of Gimap4 expression by another closely linked (SNP1054) locus, it may, in contrast, reﬂect the individual-to-individual vari- ations in Gimap4 expression that were apparent among the AT(ϩ)/ Rat strain AT status Haplotype 1 AT(ϩ) homozygotes in our Western blotting analysis. Protein ex- A990 AT(Ϫ)Cpression analysis of this backcross litter was performed post AGUS AT(Ϫ)Cmortem, so it was not possible to pedigree test this discordant rat. Ϫ AS AT( )CNo similar individual or the obverse, i.e., an AT(ϩ)/AT(Ϫ) AUG AT(Ϫ)C BN AT(ϩ)Cheterozygote with low protein expression, has yet been encoun- COP AT(Ϫ)Ctered. We have now moved over to using blood leukocytes in DA AT(Ϫ)Cour Gimap4 phenotyping of rats (Fig. 5B and Table II) and a Ϫ DZB/Gro AT( )Cthereby we will in future be able to follow up unusual individ- F344 AT(Ϫ)C LOU AT(Ϫ)Cuals by retesting and breeding. MAXXa AT(ϩ)CDespite the results with this single backcross rat, it is our pre- M520 AT(Ϫ)Cferred hypothesis that the low protein phenotype in BN Gimap4 is Ϫ NEDH AT( )Cdue to the AT dinucleotide insertion. This insertion is predicted to NIGIII AT(Ϫ)C SHR AT(Ϫ)Cbring about premature termination of the BN form of Gimap4 WAG AT(Ϫ)Cpolypeptide, such that the 21 carboxyl-terminal amino acids of the WF AT(Ϫ)Cwild-type sequence are lost and replaced by just three different Downloaded from WKY AT(Ϫ)Cresidues. Among the 28 inbred laboratory rat strains screened, this DR-BB/Hb AT(؉)C dinucleotide insertion was seen only in BN and one other strain, Rat Strain AT Status Haplotype 2 MAXX, which has BN as an ancestor (28). We found little difference between the mRNA levels of the BN variant and the wild- AO AT(Ϫ)T BB-DR/Ed AT(Ϫ)Ttype alleles of Gimap4 by using quantitative real-time PCR (Fig. 6) BDIX AT(Ϫ)Tor Northern blotting (data not shown), indicating that the effect of http://www.jimmunol.org/ BDE AT(Ϫ)Tthe AT(ϩ) variant on protein expression is posttranscriptional. It is Ϫ BDV AT( )Tpossible that the carboxyl-terminal truncation of the Gimap4 BUF AT(Ϫ)T LEJ AT(Ϫ)Tpolypeptide causes instability at the protein level. We are currently LEW AT(Ϫ)Tundertaking experiments to try to identify the mechanism in- PVG AT(Ϫ)Tvolved. To date we have established that the truncation of Gimap4 RP AT(Ϫ)Tdoes not lead to intracellular mislocalization from the cytosol (Fig. a DZB/Gro and MAXX DNA were supplied by Emile de Heer (University of 6A). Also absent is a gross change in polypeptide stability due to Leiden, The Netherlands) and Hein van Lith (University of Utrecht, The Netherlands),

the truncation as assessed by a pulse-chase analysis in transfected by guest on September 26, 2021 respectively. b Bold, not sequenced by us; data taken from Ref. 5. cells (Fig. 7B and data not shown). Of course, a caveat in interpreting such an analysis is that this technique would not be sensitive enough to measure small differences that, over a long period of time, could Further analysis of the Gimap4 knockout mice, however, re- have a pronounced effect on the amount of protein accumulating. The vealed an intriguing phenotype in T lymphocytes in response to in specific kinetic details of Gimap4 synthesis and degradation in lym- vitro apoptotic stimulation in that knockout cells are delayed in phocytes may be essential to the phenotype observed. their death program, as defined by the two flow cytometric criteria As noted in Table III (although not included in our panel of ϩ annexin V and propidium iodide staining for phosphatidylserine DNAs), the Gimap4 AT( ) sequence allele has also been detected exposure and membrane permeabilization, respectively (16). in DR-BB/H rats (5). These and some related diabetes-resistant BB ϩ/ϩ These data suggested that Gimap4 may be acting to accelerate the sublines, which are Gimap5 , are frequently used as controls in death process. In the present study. Gimap4-deficient BN LN T studies of autoimmune diabetes in BB-DP rats. That this control cells were put through a similar set of apoptotic tests. Our con- strain may also possess a functional defect in a member of the clusion is that these cells exhibit a similar phenotype to that of the Gimap protein family should be taken into account in interpreting knockout mouse, although somewhat milder. Thus, with strong past literature and future investigations. (As mentioned earlier, the Ϫ/Ϫ apoptotic stimuli clear differences compared with wild-type rat BB-DR/Ed subline is different in that it is Gimap5 but fails to cells were detectable, but the set of analyses conducted in parallel develop autoimmune diabetes mellitus, presumably on account of using the knockout mouse (and their controls) showed generally resistance alleles at other susceptibility loci.) (17). more marked effects. This perhaps reflects the fact that the BN rat The C-terminal region of the Gimap4 polypeptide encoded appears to be a hypomorph for Gimap4 rather than a complete downstream of the AT insertion site is highly conserved between knockout; we estimate that BN T cells express Gimap4 at ϳ5% of wild type rat and mouse, with only one conservative difference, normal levels. Analysis of Gimap4 by intracellular staining and 327R7K, present. This region lies outside of the domains/motifs flow cytometry indicates that this low level of expression is com- of Gimap4 so far characterized (Fig. 3A), namely the GTP binding mon to all BN T cells and that there is not a small subset of domain (with G1–G5 motifs), the conserved sites for protein ki- ϩ high-expressing cells that accounts for the blotting signal in this nase C phosphorylation, and the IQ domain that mediates Ca2 - strain. It is possible that this low but finite level of expression regulated interaction with calmodulin (16). It will be interesting to allows partial Gimap4 function to persist in BN T cells. determine what function(s) or intermolecular interaction(s) are It is probable that the BN Gimap4 phenotype is determined by controlled by this short sequence. Similarly, as and when suitable an AT dinucleotide insertion. In analyses of backcross populations assays are available it will be interesting to ascertain what func- segregating for the two Gimap4 alleles (BN vs wild type), a single tions, if any, the BN Gimap4 protein variant retains in the absence animal in 49 showed a lack of agreement between its Gimap4 of this peptide sequence. 1794 CHARACTERIZATION OF A HYPOMORPHIC Gimap4 ALLELE

Table IV. Genotyping of Gimap4 in wild ratsa

Rat Genotypes

Country/Area Sourceb Total Rats AT(ϩ)/AT(ϩ) AT(Ϫ)/AT(Ϫ) AT(ϩ)/AT(Ϫ) AT(ϩ) Alleles/Total

United Kingdom Cambridgeshire A 15 15 0/30 Suffolk B 2 2 0/4 Berkshire B 8 6 2 2/16 Dorset B 5 3 2 2/10 Essex A, B 6 6 0/12 Kent B 5 1 3 1 3/10 Wiltshire B 7 2 3 2 6/14 Total United Kingdom 48 3 38 7 13/96 (14%)

Germany Hamburg A 3 3 0/6 Niehoff C 5 1 1 3 5/10 Looz C 5 1 3 1 3/10 Kortenbusch C 5 5 0/10 Westlinning C 3 3 0/6 Total Germany 21 2 15 4 8/42 (19%) Downloaded from

Norway Oslo D 7 1 5 1 3/14 (21%)

Sweden Lund E 16 9 3 4 22/32 (69%) http://www.jimmunol.org/ Japan Mitake-cho, Gifu F 3* (3) 0/3c Shitara-cho, Aichi F 1† (1) 0/1d Matsuyama, Ehime G 4 2 2 6/8 Osaka G 13 8 3 2 18/26 Total Japan 21 10 3 (7) 4 24/38 (63%)

Grand totals 113 25 64e (68) 20 Total AT(ϩ) alleles 70/222 (32%)

Japan by guest on September 26, 2021 Tokyo (R. rattus) G 3 1 1 1 3/6 (50%)

a DNA samples were obtained from 113 wild rats in five countries. A 293-bp fragment was amplified and sequenced to show AT status. The table segregates the rats from different areas and gives numbers for those with the homozygous AT(ϩ), AT(Ϫ), or heterozygous genotype. b Source notes for wild rats: A, See Ref. 22; B, Samples provided by Dr. A. MacNicoll (Central Science Laboratory, Defra, York, U.K.) and H. Zain (University of Leicester, Leicester, U.K.); C, Samples provided by Hans-Joachim Pelz (Institut fu¨r Nematologie und Wirbeltierkunde, Mu¨nster, Germany (23); D, Samples provided by E. Dissen (University of Oslo, Oslo, Norway) (24); E, Samples provided by M. Hultqvist and R. Holmdahl (University of Lund, Lund, Sweden) (25); F, Samples provided by T. Serikawa (Kyoto University, Kyoto, Japan) and Sen-ichi Oda (Nagoya City Public Health Research Institute, Nagoya, Japan) (26); and G, Samples provided by Makoto Kawahara (Nagoya City Public Health Research Institute, Nagoya, Japan). c Partially inbred wild-derived strains MITB, MITC, and MITE defining only one haplotype per strain. d Partially inbred strain DOB defining only one wild-derived haplotype. e Excludes samples from source F because they were not derived from segregating populations.

In addition to the AT dinucleotide insertion found in the BN rat, that stage in our investigation it therefore seemed probable that this we have described an additional and larger scale difference be- mutation, like the previously described mutation in Gimap5 of the tween the rat and mouse Gimap4 genes, although this one, which BB rat (29), had arisen spontaneously in a laboratory rat colony may be unique to the rat, appears to be true of all rat strains. It is and been fixed and maintained by inbreeding. Nevertheless, we an ϳ1800-bp deletion beginning at or around the wild-type stop embarked on an analysis of wild rats with unexpected results. Ta- codon. This deletion brings a DNA repeat region into close jux- ble IV presents a summary of the data from the 113 samples ob- taposition with the Gimap4 ORF. Some sequence from these retained from different parts of the world. These results show not Ј peats is thereby present in the rat Gimap4 3 -UTR. Rat has lost the only that we were able to detect the AT(ϩ) allele in the wild but poly(A) addition signal that is seen in human and mouse (Table I), also that this allele was present with a surprisingly high overall and all other mammals analyzed (data not shown), and evolution frequency of ϳ32%. Indeed, in two sets of samples, those from has crafted a different such site elsewhere. This leads to different Lund, Sweden and Osaka, Japan, a substantial number of apparent message sizes for rat and mouse Gimap4 (1.4kb vs 1.7kb). It is ϩ possible that this major change to the 3Ј-region of the rat Gimap4 AT( ) homozygotes was present. These findings suggest that the ϩ gene is required to achieve the particular expression of truncated BN-type AT( ) allele of Gimap4 is not necessarily deleterious but Gimap4 protein observed in the BN rat, i.e., given the different that it confers some selective advantage that maintains it in the 3Ј-UTRs we cannot be certain that an identical AT insertion in the wild rat population in balance with the more mouse-like AT(Ϫ) mouse would produce the same outcome for the level of protein allele. If this is true, it raises the question of whether Gimap4 expression. variation mediates some developmental or immunomodulatory As mentioned above, our screen of laboratory rat strains iden- variation in T cell behavior that is under pathogen-driven selection. tified only two carrying the Gimap4 AT dinucleotide insertion. At It would be very interesting to know whether this is mediated The Journal of Immunology 1795 through the apoptosis-related function of Gimap4 or some other, as 13. Dion, C., C. Carter, L. Hepburn, W. J. Coadwell, G. Morgan, M. Graham, N. Pugh, G. Anderson, G. W. Butcher, and J. R. Miller. 2005. Expression of the yet unidentified, function(s) of this molecule. Ian family of putative GTPases during T cell development and description of an The BN rat has long been a favored strain in both immunolog- Ian with three sets of GTP/GDP-binding motifs. Int. Immunol. 17: 1257–1268. ical and toxicological research and was also chosen as the base 14. Poirier, G. M., G. Anderson, A. Huvar, P. C. Wagaman, J. Shuttleworth, E. Jenkinson, M. R. Jackson, P. A. Peterson, and M. G. Erlander. 1999. Immune- strain for the Rat Genome Mapping Project (30). It appears to be associated nucleotide-1 (IAN-1) is a thymic selection marker and defines a novel biased in its immune responses toward Th2-type immunity and it gene family conserved in plants. J. Immunol. 163: 4960–4969. is capable of generating very high IgE responses (31). Similarly, it 15. Cambot, M., S. Aresta, B. Kahn-Perle`s, J. de Gunzberg, and P.-H. Romeó. 2002. Human immune associated nucleotide 1: a member of a new guanosine triphos- is highly susceptible to the induction of Th2-mediated autoimmune phatase family expressed in resting T and B cells. Blood 99: 3293–3301. diseases such as mercuric chloride-induced glomerulonephritis 16. Schnell, S., C. De´mollie`re, P. van den Berk, and H. Jacobs. 2006. Gimap4 ac- (32) while often being resistant to Th1-mediated diseases, e.g., celerates T cell death. Blood 108: 591–599. 17. Joseph, S., A. G. Diamond, W. Smith, J. D. Baird, and G. W. Butcher. 1993. experimental allergic encephalomyelitis (33). A number of unusual BB-DR/Edinburgh: a lymphopenic, non-diabetic subline of BB rats. Immunology cellular immunological phenotypes have been reported, such as a 78: 318–328. high CD4:CD8 T cell ratio (34), skewed CD45RC expression (35), 18. Galfre`, G., C. Milstein, and B. Wright. 1979. Rat ϫ rat hybrid myelomas and a and low mitogen responsiveness (36, 37). Although good progress monoclonal anti-Fd portion of mouse IgG. Nature 277: 131–133. 19. Vandesompele, J., K. De Preter, F. Pattyn, B. Poppe, N. Van Roy, A. De Paepe, is being made toward the genetic dissection of these phenotypes and F. Speleman. 2002. Accurate normalization of real-time quantitative RT-PCR (38), this work is not complete and it will be interesting to deter- data by geometric averaging of multiple internal control genes. Genome Biol. 3: RESEARCH0034. mine whether the BN Gimap4 variant, as a T cell GTPase appar- 20. Thompson. J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins. 1997. ently susceptible to regulation through the TCR (14–16), plays any The CLUSTAL_X windows interface: flexible strategies for multiple sequence align- part in them. To this end a BN congenic strain carrying the PVG ment aided by quality analysis tools. Nucleic Acids Res. 15: 4876–4882. Downloaded from 21. Sonnhammer, E, L. L., and R. Durbin. 1995. A dot-matrix program with dynamic (wild type) Gimap4 is under preparation. threshold control suited for genomic DNA and protein sequence analysis. Gene 167: GC1–GC10. Acknowledgments 22. Rothenburg, S., F. Haag, F. Koch-Nolte, C. Carter, M. Graham, and G. W. Butcher. We thank Arthur Davis for flow sorting, Maureen Hamon for assistance 2005. Characterization of multiple alleles of the T-cell differentiation marker ART2 (RT6) in inbred and wild rats. Immunogenetics 57: 739–749. with immunochemistry, and Anne Segonds-Pichon for help with statistical 23. Pelz, H.-J., S. Rost, M. Hu¨nerberg, A. Fregin, A.-C. Heiberg, K. Baert, A. D. analysis. MacNicoll, C. V. Prescott, A.-S. Walker, J. Oldenburg, and C. R. Miller. 2005. The http://www.jimmunol.org/ genetics basis of resistance to anticoagulants in rodents. Genetics 170: 1839–1847. Disclosures 24. Dissen, E., S. V. Hunt, B. Rolstad, and S. Fossum. 1993. Localization of the rat T-cell receptor ␤-chain and carboxypeptidase A1 loci to chromosome 4. Immu- The authors have no financial conflict of interest. nogenetics 37: 153–156. 25. Olofsson, P., J. Holmberg, J. Tordsson, S. Lu, B. Åkerstro¨m, and R. Holmdahl. References 2003. Positional identification of Nef1 as a gene that regulates arthritis severity 1. Kru¨cken, J., R. M. U. Schroetel, I. U. Mu¨ller, N. Saı¨dani, P. Marinovski, in rats. Nat. Genet. 33: 25–32. P. M. Benten, O. Stamm, and F. Wunderlich. 2004. Comparative analysis of the 26. Kondo, Y., K. Ohno, S. Oda, and T. Serikawa. 1996. Genetic profiles of three human gimap gene cluster encoding a novel GTPase family. Gene 341: 291–304. inbred strains derived from wild rats (Rattus norvegicus) trapped in Japan. Exp. 2. Guttmann, R. D., E. Colle, F. Michel, and T. Seemayer. 1983. Spontaneous Anim. 45: 405–409. diabetes mellitus syndrome in the rat, II: T lymphopenia and its association with 27. Mick, V. E., T. K. Starr, T. M. McCaughtry, L. K. McNeil, and K. A. Hogquist. clinical disease and pancreatic lymphocytic infiltration. J. Immunol. 130: 2004. The regulated expression of a diverse set of genes during thymocyte pos- by guest on September 26, 2021 1732–1735. itive selection in vivo. J. Immunol. 173: 5434–5444. 3. Jackson, R. A., J. B. Buse, R. Rifai, D. Pelletier, E. L. Milford, C. B. Carpenter, 28. Festing, M. 1979. Inbred strains. In Baker, H., J. Lindsey, and S. Weisbroth, eds., G. S. Eisenbarth, and R. M. Williams. 1984. Two genes required for diabetes in The Laboratory Rat Vol. I: Biology and Diseases, Academic Press, New York, BB rats: evidence from cyclical intercrosses and backcrosses. J. Exp. Med. 159: pp. 55–72. 1629–1636. 29. Chappel, C., and W. Chappel. 1983. The discovery and development of the BB 4. Cousins, L., M. Graham, R. Tooze, C. Carter, J. R. Miller, F. M. Powrie, rat colony: an animal model of spontaneous diabetes mellitus. Metabolism 32: G. G. MacPherson, and G. W. Butcher. 2006. Eosinophilic bowel disease con- S8–S10. trolled by the BioBreeding rat-derived lymphopenia/Gimap5 gene. Gastroenter- 30. Rat Genome Sequencing Project Consortium 2004. Genome sequence of the Brown ology 131: 1475–1485. Norway rat yields insights into mammalian evolution. Nature 428: 493–521. 5. Hornum, L., J. Romer, and H. Markholst. 2002. The diabetes-prone BB rat carries 31. Prouvost-Danon, A., A. Abadie, C. Sapin, H. Bazin, and P. Druet. 1981. Induc- a frameshift mutation in Ian4, a positional candidate of Iddm1. Diabetes 51: tion of IgE synthesis and potentiation of anti-ovalbumin IgE antibody response by 1972–1979. HgCl2 in the rat. J. Immunol. 126: 699–702. 6. MacMurray, A. J., D. H. Moralejo, A. E. Kwitek, E. A. Rutledge, B. Van Yserloo, 32. Sapin, C., E. Druet, and P. Druet. 1977. Induction of anti-glomerular basement P. Gohlke, S. J. Speros, B. Snyder, J. Schaefer, S. Bieg, et al. 2002. Lymphopenia membrane antibodies in the Brown-Norway rat by mercuric chloride. Clin. Exp. in the BB rat model of type 1 diabetes is due to a mutation in a novel immune- Immunol. 28: 173–178. associated nucleotide (Ian)-related gene. Genome Res. 12: 1029–1039. 33. Happ, M., P. Wettstein, B. Dietzschold, and E. Heber-Katz. 1988. Genetic control 7. Ramanathan, S., K. Norwich, and P. Poussier. 1998. Antigen activation rescues of the development of experimental allergic encephalomyelitis in rats: separation recent thymic emigrants from programmed cell death in the BB rat. J. Immunol. of MHC and non-MHC effects. J. Immunol. 141: 1489–1494. 160: 5757–5764. 8. Hernańdez-Hoyos, G., S. Joseph, N. G. Miller, and G. W. Butcher. 1999. The 34. Damoiseaux, J., B. Cautain, I. Bernard, M. Mas, P. van Breda Vriesman, lymphopenia mutation of the BB rat causes inappropriate apoptosis of mature P. Druet, G. Fournie´, and A. Saoudi. 1999. A dominant role for the thymus and thymocytes. Eur. J. Immunol. 29: 1832–1841. MHC genes in determining the peripheral CD4/CD8 T cell ratio in the rat. J. Im- 9. Sandal, T., L. Aumo, L. Hedin, B. T. Gjertsen, and S. O. Doskeland. 2003. munol. 163: 2983–2989. ␥ 35. Powrie, F., D. Fowell, A. J. McKnight and D. Mason. 1991. Lineage relationships Irod/Ian5: an inhibitor of -radiation- and okadaic acid-induced apoptosis. Mol. ϩ Biol. Cell 14: 3292–3304. and functions of CD4 T-cell subsets in the rat. Res. Immunol. 142: 54–58. 10. Pandarpurkar, M., L. Wilson-Fritch, S. Corvera, H. Markholst, L. Hornum, 36. Newlin, C., and D. Gasser. 1973. Genetic control of the in vitro responses of rat D. L. Greiner, J. P. Mordes, A. A. Rossini, and R. Bortell. 2003. Ian4 is required for peripheral blood lymphocytes to phytohemagglutinin and concanavalin A. J. Im- mitochondrial integrity and T cell survival. Proc. Natl. Acad. Sci. USA 100: munol. 110: 622–628. 10382–10387. 37. Fu, Y., and E. Blankenhorn. 1992. Nitric oxide-induced anti-mitogenic effects in 11. Nitta, T., M. Nasreen, T. Seike, A. Goji, I. Ohigashi, T. Miyazaki, T. Ohta, high and low responder rat strains. J. Immunol. 148: 2217–2222. M. Kanno, and Y. Takahama. 2006. IAN family critically regulates survival and 38. Fournie´, G., B. Cautain, E. Xystrakis, J. Damoiseaux, M. Mas, D. Lagrange, development of T lymphocytes. PLoS Biol. 4: E103. I. Bernard, J.-F. Subra, L. Pelletier, P. Druet, and A. Saoudi. 2001. Cellular and 12. Reuber, T. L., and F. M. Ausubel. 1996. Isolation of Arabidopsis genes that genetic factors involved in the difference between Brown Norway and Lewis rats differentiate between resistance responses mediated by the RPS2 and RPM1 dis- to develop respectively type-2 and type-1 immune-mediated diseases. Immunol. ease resistance genes. Plant Cell 8: 241–249. Rev. 184: 145–160.