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 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(+) allele was found at a high overall frequency (∼30%). This suggests an adaptive significance for this hypomorphic allele.

The Gimap/IAN proteins are a family of putative GTPases found in vertebrates. They have relatives in higher plants but not in most other organisms studied (1). The importance of the Gimap/IAN GTPase family for T lymphocyte survival became evident from investigations of the rat lymphopenia (lyp) gene. This recessive mutation causes severe peripheral T lymphopenia when homozygous and is an essential susceptibility locus in autoimmune models of both spontaneous diabetes mellitus (the BioBreeding diabetes prone (BB-DP)4 rat) and eosinophilic bowel disease (the PVG-RT1u,lyp/lyp rat) (2, 3, 4). Positional cloning of lymphopenia identified a single base pair deletion in the Gimap5 gene. The predicted frameshift leads to a polypeptide product from this gene that is truncated by about two-thirds compared with wild type (5, 6). The lymphopenic phenotype observed in Gimap5-deficient BB-DP rats is thought to result from the in vivo apoptosis of mature thymocytes and T lymphocytes (7, 8) in the absence of the anti-apoptotic wild-type Gimap5 protein. Consistent with this interpretation, prosurvival properties of human, rat, and mouse Gimap5 have been demonstrated in vitro (9, 10, 11). Gimap5 has a predicted transmembrane domain that appears to direct its expression to the surface of mitochondria (a key site of apoptotic regulation) as well as to some other internal cell membranes (9, 10).

Rats, mice, and humans each express seven or eight Gimap genes clustered tightly on a single autosome. The predicted proteins that they encode share amino-terminal features consisting of a GTPase domain and other sequence motifs leading to their classification in the AIG1 family, a sequence-based category named after a protein involved in responses to bacterial pathogens in plants (12). The carboxyl-terminal features of the Gimap proteins are more diverse, with some carrying obvious transmembrane domains. Given the profound impact of Gimap5-deficiency on T lymphocyte 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 reported physical interactions between Gimap proteins and members of the Bcl-2 protein family (11). In addition, it will be of interest to ascertain whether these proteins act separately or in a coordinated fashion to regulate lymphocyte survival.

We previously demonstrated interesting changes in the expression of the Gimap genes during the development of rat T lymphocytes (13). Among the most dramatic of these were the changes in the expression of Gimap4 (IAN1), previously reported in mice (14), that shows a substantial rise in expression at the thymic CD4+CD8+ double positive to mature single positive transition. Gimap4 lacks an obvious transmembrane domain and appears to lie downstream of the TCR in an as yet ill-defined signaling pathway (14, 15, 16).

We raised Abs against Gimap4 to study its protein expression in rats and mice (Ref. 13 and current report). Using these reagents, we have encountered an unexpected deficiency in Gimap4 protein expression in the inbred Brown Norway (BN) rat, a strain used widely in immunological research as well as in other biomedical disciplines. We also report here functional comparisons of BN rat T lymphocytes with those from mice carrying a targeted deletion in Gimap4, recently described by one of our laboratories (16).

Rats of the strains BN, PVG-RT1n(BN), PVG-RT1u,RT7b, PVG-RT1u,lyp/lyp, BB-disease resistant (BB-DR)/Ed (which is also genetically lyp/lyp; Ref. 17), and DA, as well as some backcross populations (see Results), were bred and maintained in specific pathogen-free conditions at the Babraham Institute. In unpublished analyses we have confirmed that the PVG-RT1u,lyp/lyp strain, which derives its mutant lyp gene from the Edinburgh subline of BB-DP rats (8, 17), carries the frameshift mutation in the Gimap5/Ian5 gene described previously (5, 6). Rats of the strains LEW, LOU/C, WKY, and WAG were purchased from Harlan UK. C57BL/6 mice used for Western blotting experiments were bred at the Babraham Institute. Gimap4 knockout mice on the C57BL/6 background (16) and their heterozygous littermates (used as controls) were bred in specific pathogen-free animal facilities at The Netherlands Cancer Institute. Animals were used at between 8 and 20 wk of age. All animal usage had been previously approved by the respective ethical review committees.

Mouse and rat Gimap4 were PCR amplified from splenic or thymic cDNA and cloned in the GST fusion vector pGEX-4T-1 (Amersham Biosciences). Rats were immunized with purified GST fusion proteins of mouse and rat Gimap4 to produce both a polyclonal antiserum and mAbs reactive against both species. Custom production of a rabbit polyclonal antiserum against rat Gimap4-GST fusion protein was conducted by Harlan Sera-Lab. The rat mAb MAC 417 was derived from a fusion of immune splenocytes with the rat Y3Ag1.2.3 plasmacytoma (18) and is of the IgG2b subclass.

The wells of flat-bottom 96-well plates were incubated overnight at 4°C with recombinant GST, rat Gimap4-GST, or mouse Gimap4-GST at 2.5 μg/ml in 0.1 M sodium bicarbonate (pH 8.3). The wells were washed with water and then blocked with 2% FCS and 0.01% Tween 20 in PBS for 90 min. at 37°C. After washing, the wells were incubated with MAC 417 hybridoma supernatant for 1 h at 37°C. Binding was detected using a biotin-rabbit anti-rat IgG (Dako E0468; mouse absorbed) followed by a streptavidin-HRP complex (catalog no. RPN 1051; Amersham Biosciences), both diluted in blocking buffer. The HRP substrate 3,3′5,5′tetramethylbenzidine (T2885; Sigma-Aldrich) was used for color development. The results were read at 450 nm on a Multiskan EX plate reader (Labsystems).

Single cell suspensions of rat thymocytes or lymph node (LN) cells were 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 with secondary reagents. After final washing, the cells were filtered through a 40-μm filter immediately before cell sorting. The mAbs used were W3/25 (anti-CD4), 341 (anti-CD8β), MARK1 (anti-Igκ L chain), and MARM4 (anti-IgM). These mAbs came either from Serotec or from in-house stocks and were either FITC- or biotin-conjugated. In the latter case, PE-Cy5.5-conjugated streptavidin (eBioscience) was used as a secondary fluorescent reagent. Cell populations were sorted using FACSDiva or FACSAria from BD Biosciences. Purities were checked to be between 90 and 99.9%.

Cell samples at a concentration of 2 × 108/ml were incubated in lysis buffer (2% Nonidet P-40, 20 mM Tris, 150 mM NaCl, and 1 mM MgCl2 (pH 8) supplemented with proteolytic inhibitors) for 30 min at 4°C. Nuclei were removed by centrifugation at 14,000 × g. An equal volume of Laemmli sample buffer (Bio-Rad) was added to each sample. Ten-microliter samples, equivalent to 1 × 106 cells, were run on 15% SDS-polyacrylamide gels and then blotted onto Immobilon P membranes (Millipore). Blots were blocked overnight at 4°C with 4% milk powder in PBS with 0.1% Tween and incubated with primary and secondary Abs for 1 h at room temperature. Blots were developed with a SuperSignal West Pico chemiluminescent substrate (Pierce). After staining for Gimap4, blots were stripped with Restore Western Blot stripping buffer (Pierce) and stained with monoclonal anti-actin (catalog no. A5441; Sigma-Aldrich) as a loading control.

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 Technologies) and cDNA was made using SuperScript III reverse transcriptase (Invitrogen Life Technologies). Methods are essentially as described previously (13, 19), including the use of two control genes, 6-phosphofructokinase C (6PFK) and cirhin. Exon-specific primer pairs were designed to span an intron in genomic DNA such that any products from contaminating genomic DNA could be identified and excluded on the basis of size. Primers for Gimap4 were: 5′-GAGCAGCCATGAGCTTGGAAT-3′ and 5′-TCAACAGGGAACAGCATCCTTG-3′. PCR was performed using a SYBR Green kit in accordance with the manufacturer’s instructions (Applied Biosystems), and samples were amplified using the Chromo4 system (Bio-Rad). A two-way ANOVA was applied to the data to determine the relative expression between BN and wild-type rats for each of the cell types tested. The analysis was performed in triplicate on three biological samples for each cell population.

cDNA sequences encoding the rat Gimap4 proteins predicted to be expressed in PVG-RT1n (BN) or BN rats were cloned downstream of the CMV promoter in plasmid pcDNA3.1/myc-His(−)A in frame with the C-terminal myc-His tags. The resulting plasmid constructs were transfected into the HEK293T cell line using jetPEI transfection reagent (Polyplus Transfection) according to the manufacturer’s instructions (2 μg of plasmid per 35-mm dish of cells). Cells were left overnight and then washed with 2 ml of methionine- and cysteine-free DMEM medium supplemented with 5% (v/v) dialyzed calf serum, 100 U/ml penicillin, and 100 μg/ml streptomycin and then incubated at 37°C for 1 h in the same medium. The medium was replaced with 0.5 ml of the same medium supplemented with a 3.7 MBq/ml mixture of [35S]methionine and [35S]cysteine (Tran35S-label; 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 DMEM containing 10% FCS, 1 mM methionine, and 1 mM cysteine for various times as indicated. At each time point, cells were washed with 1 ml of PBS and lysed into 1 ml of ice-cold CHAPS lysis buffer (10 mM HEPES, 150 mM sodium chloride, and 1% (w/v) CHAPS) containing one tablet per 10 ml of complete Mini EDTA-free protease inhibitor mixture (Roche) for 10 min. The resultant cell lysates were centrifuged at 11,000 × g for 5 min and the supernatant was added to 1 μl of rabbit anti-rat Gimap4-GST fusion protein antiserum (see above) diluted in 100 μl of CHAPS lysis buffer and 50 μl of 1:1 (v/v) suspension of protein A-Sepharose beads, and the suspension was rotated overnight at 4°C. The beads were then washed with 5 × 1 ml CHAPS lysis buffer at 4°C and bound protein was eluted into 100 μl of 2× Laemmli sample buffer by heating to 100°C for 5 min. Proteins were separated on 10% SDS polyacrylamide gels. The gels were then fluorographed, dried, and exposed to x-ray film.

Genomic DNA samples from tissues of inbred rat strains or wild rats were either prepared at the Babraham Institute using the DNeasy tissue kit (Qiagen) or generously supplied by the collaborators indicated in Tables III and IV. Gene-specific primers and PCR were used to amplify Gimap4 from genomic DNA. The resulting PCR products were treated with exonuclease I and shrimp alkaline phosphatase (Amersham Biosciences) before being sent for commercial sequencing (Lark Technologies).

Table III.

Genotyping of a panel of inbred rat strains to show Gimap4 AT status and haplotype as defined by a SNP close to the AT insertion (SNP1054)

Rat strainAT statusHaplotype 1
A990 AT(−) 
AGUS AT(−) 
AS AT(−) 
AUG AT(−) 
BN AT(+) 
COP AT(−) 
DA AT(−) 
DZB/Groa AT(−) 
F344 AT(−) 
LOU AT(−) 
MAXXa AT(+) 
M520 AT(−) 
NEDH AT(−) 
NIGIII AT(−) 
SHR AT(−) 
WAG AT(−) 
WF AT(−) 
WKY AT(−) 
DR-BB/Hb AT(+) C 
Rat strainAT statusHaplotype 1
A990 AT(−) 
AGUS AT(−) 
AS AT(−) 
AUG AT(−) 
BN AT(+) 
COP AT(−) 
DA AT(−) 
DZB/Groa AT(−) 
F344 AT(−) 
LOU AT(−) 
MAXXa AT(+) 
M520 AT(−) 
NEDH AT(−) 
NIGIII AT(−) 
SHR AT(−) 
WAG AT(−) 
WF AT(−) 
WKY AT(−) 
DR-BB/Hb AT(+) C 
Rat StrainAT StatusHaplotype 2
AO AT(−) 
BB-DR/Ed AT(−) 
BDIX AT(−) 
BDE AT(−) 
BDV AT(−) 
BUF AT(−) 
LEJ AT(−) 
LEW AT(−) 
PVG AT(−) 
RP AT(−) 
Rat StrainAT StatusHaplotype 2
AO AT(−) 
BB-DR/Ed AT(−) 
BDIX AT(−) 
BDE AT(−) 
BDV AT(−) 
BUF AT(−) 
LEJ AT(−) 
LEW AT(−) 
PVG AT(−) 
RP AT(−) 
a

DZB/Gro and MAXX DNA were supplied by Emile de Heer (University of Leiden, The Netherlands) and Hein van Lith (University of Utrecht, The Netherlands), respectively.

b

Bold, not sequenced by us; data taken from Ref. 5 .

Table IV.

Genotyping of Gimap4 in wild ratsa

Country/AreaSourcebTotal RatsRat GenotypesAT(+) Alleles/Total
AT(+)/AT(+)AT(−)/AT(−)AT(+)/AT(−)
United Kingdom       
 Cambridgeshire A 15  15  0/30 
 Suffolk B   0/4 
 Berkshire B  2/16 
 Dorset B  2/10 
 Essex A, B   0/12 
 Kent B 3/10 
 Wiltshire B 6/14 
Total United Kingdom  48 38 13/96 (14%) 
       
Germany       
 Hamburg A   0/6 
 Niehoff C 5/10 
 Looz C 3/10 
 Kortenbusch C   0/10 
 Westlinning C   0/6 
Total Germany  21 15 8/42 (19%) 
       
Norway       
 Oslo D 3/14 (21%) 
       
Sweden       
 Lund E 16 22/32 (69%) 
       
Japan       
 Mitake-cho, Gifu F 3*  (3)  0/3c 
 Shitara-cho, Aichi F 1†  (1)  0/1d 
 Matsuyama, Ehime G  6/8 
 Osaka G 13 18/26 
Total Japan  21 10 3 (7) 24/38 (63%) 
       
Grand totals  113 25 64e (68) 20  
Total AT(+) alleles      70/222 (32%) 
       
Japan       
 Tokyo (R. rattusG 3/6 (50%) 
Country/AreaSourcebTotal RatsRat GenotypesAT(+) Alleles/Total
AT(+)/AT(+)AT(−)/AT(−)AT(+)/AT(−)
United Kingdom       
 Cambridgeshire A 15  15  0/30 
 Suffolk B   0/4 
 Berkshire B  2/16 
 Dorset B  2/10 
 Essex A, B   0/12 
 Kent B 3/10 
 Wiltshire B 6/14 
Total United Kingdom  48 38 13/96 (14%) 
       
Germany       
 Hamburg A   0/6 
 Niehoff C 5/10 
 Looz C 3/10 
 Kortenbusch C   0/10 
 Westlinning C   0/6 
Total Germany  21 15 8/42 (19%) 
       
Norway       
 Oslo D 3/14 (21%) 
       
Sweden       
 Lund E 16 22/32 (69%) 
       
Japan       
 Mitake-cho, Gifu F 3*  (3)  0/3c 
 Shitara-cho, Aichi F 1†  (1)  0/1d 
 Matsuyama, Ehime G  6/8 
 Osaka G 13 18/26 
Total Japan  21 10 3 (7) 24/38 (63%) 
       
Grand totals  113 25 64e (68) 20  
Total AT(+) alleles      70/222 (32%) 
       
Japan       
 Tokyo (R. rattusG 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(+), AT(−), 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 für Nematologie und Wirbeltierkunde, Münster, 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.

Total RNA was prepared using TRIzol reagent (Invitrogen Life Technologies) from the LN cells of BN and PVG-RT1u, RT7b rats and from the splenocytes of C57BL/6 mice. Complementary DNA was then synthesized using Superscript III (Invitrogen Life Technologies) and an oligo(dT) anchor primer (5′-GACCACGCGTATCGATGTCGACTTTTTTTTTTTTTTTTV-3′, where V represents A, C, or G). A 1-μl aliquot of the cDNA reaction (final volume 50 μl) was then used for RACE-PCR using a 3′ PCR anchor primer (5′-GACCACGCGTATCGATGTCGAC-3′) and an appropriate 5′ gene-specific primer for rat Gimap4 (5′-GAGAGAGAGAGAGCACGGATAAG-3′) and mouse Gimap4 (5′-AGAGGGAATTCAAGGAGAGG-3′). After the completion of PCR, the products were treated with shrimp alkaline phosphatase and exonuclease 1 to remove unincorporated primers and dNTPs. A 1.5-μl aliquot of reaction product was then used for a second PCR using the 3′ anchor primer (see above) and gene-specific “nested” primers for rat Gimap4 (5′-CAAGATCTGAGAGATGAGCTGG-3′) and mouse Gimap4 (5′-AAGGAAGTTGAGAACACAAGTATG-3′). 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 5′-GCAACTCAAAGCAGGAACCTG-3′ for rat and 5′-AGTTCCAATCTTTGCATGAG-3′ for mouse.

DNA repeats were analyzed using RepeatMasker (http://repeatmasker.genome.washington.edu/). Other bioinformatic tools used were ClustalX (20) and Dotter (21).

For the experiment shown in Fig. 8,A, 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 × 105 cells per milliliter in RPMI 1640 medium supplemented with 20% FCS, 2-ME, and nonessential amino acids. In vitro cell survival 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. 8 B, 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 macrophage (MRC OX42) and anti-rat Igκ (RG7/9.1). Cells bearing bound biotinylated mAbs were retained on magnetic nanoparticles as for mouse T cells. In this manner, cells were 90–96% pure as determined by flow cytometric analysis. T cells (0.2 × 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 αβTCR; clone R73). Apoptotic cells were identified by annexin V and propidium iodide staining according to the apoptosis detection kit from BD Biosciences. Analysis was performed on a FACSCalibur using CellQuest software.

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-RT1n (filled squares, broken line) rats were treated with the apoptotic stimuli dexamethasone (1 μM), etoposide (10 μg/ml), and gamma (γ) 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 animals 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.

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-RT1n (filled squares, broken line) rats were treated with the apoptotic stimuli dexamethasone (1 μM), etoposide (10 μg/ml), and gamma (γ) 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 animals 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.

Close modal

PVG-RT1u,RT7b and BN LN cells were washed twice with PBS and lysed by sonication in 1 ml of ice-cold Break buffer (20 mM HEPES, 1 mM EGTA, 0.5 mM MgCl2, 10 mM NaF, 10 mM β-glycerophosphate, 0.13 M sucrose, 50 mM NaCl, and 1% protease inhibitor mixture (catalog no. P8340; Sigma-Aldrich). An aliquot of the sample was reserved as the “total lysate” and the remainder was centrifuged at 250 × g for 5 min. at 4°C in a swing-out rotor to pellet the nuclei. The supernatant was then ultracentrifuged in a swing-out rotor (Beckman TLS-55) at 195,000 × g for 1 h. at 4°C to obtain the cytosol and membrane fractions. The nuclei and membranes 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 anti-Gimap4 (MAC 417), and anti-Hsp90 (BD Biosciences) and anti-VDAC-1 (Calbiochem) served as fractionation controls.

Rat T cells were isolated by magnetic selection as described above. Cells were treated with fixation buffer (catalog no. 00-8222-49; eBioscience) for 15 min on ice and then incubated with permeabilization buffer (00-8333-56;eBioscience). Intracellular staining was detected by using FITC-conjugated IgGs from normal rabbit antiserum as a negative control and from anti-Gimap4 rabbit antiserum. Staining was detected by flow cytometry using either a FACSCalibur or LSRII analyser from BD Biosciences.

To investigate Gimap4 protein expression in rat lymphocytes, we raised rat antisera and rat mAbs against recombinant (rat plus mouse) Gimap4 and a rabbit antiserum against rat Gimap4 alone. The mAb used in this study is designated MAC 417. It reacts with both rat and mouse Gimap4 as shown by ELISA (Fig. 1,A) and Western blotting (Fig. 1 C).

FIGURE 1.

Gimap4 expression. A, ELISA showing reactivity of mAb MAC 417 with rat (r) and mouse (m) GST-Gimap4 fusion proteins. B, Western blot analysis of Gimap4 in splenocytes from 10 rat strains: Lane 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; lane 9, DA; and lane 10, LOU/C. C, Western blot analysis of BN Gimap4 expression using different reagents. Rat antiserum (panel I), mAb MAC 417 (panel II), and rabbit antiserum (panel III) were used to probe lysates from unseparated LN cells from PVG-RT1n(BN) (lanes 1) and BN rats (lanes 2) and a C57BL/6 mouse (lanes 3). The lysates in this blot were subjected to nuclear spinout, but a duplicate blot with total lysate gave comparable results (data not shown). In blots B and C the top panel shows Gimap4 expression and the lower panel is stained with monoclonal anti-actin as a loading control. Lysate from 1 × 106 cells was loaded in each lane.

FIGURE 1.

Gimap4 expression. A, ELISA showing reactivity of mAb MAC 417 with rat (r) and mouse (m) GST-Gimap4 fusion proteins. B, Western blot analysis of Gimap4 in splenocytes from 10 rat strains: Lane 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; lane 9, DA; and lane 10, LOU/C. C, Western blot analysis of BN Gimap4 expression using different reagents. Rat antiserum (panel I), mAb MAC 417 (panel II), and rabbit antiserum (panel III) were used to probe lysates from unseparated LN cells from PVG-RT1n(BN) (lanes 1) and BN rats (lanes 2) and a C57BL/6 mouse (lanes 3). The lysates in this blot were subjected to nuclear spinout, but a duplicate blot with total lysate gave comparable results (data not shown). In blots B and C the top panel shows Gimap4 expression and the lower panel is stained with monoclonal anti-actin as a loading control. Lysate from 1 × 106 cells was loaded in each lane.

Close modal

MAC 417 was used to test a number of different inbred rat strains for the presence of Gimap4 in spleen cell lysates (Fig. 1,B), with Gimap4 being detected predominantly as a ∼38-kDa band. Some rat strains gave very weak signals. Two of these strains, PVG-RT1u, lyp/lyp and BB-DR/Ed, are inherently T lymphopenic due to a mutation of their Gimap5 (lyp) gene and were previously noted to exhibit poor Gimap4 expression in peripheral lymphocytes (13). The poor expression of Gimap4 by the BN rat spleen, however, was unexpected. To confirm this finding, similar tests were conducted using two polyclonal antisera against Gimap4 raised in rats and rabbits, respectively (Fig. 1,C). These tests also showed weak Gimap4 signals in BN rats compared with a representative “normal” rat (PVG-RT1n). The conclusion that these results were due to a genuine deficiency of the Gimap4 protein in BN rats and not just the absence of a strain-dependent serological epitope was strengthened by the fact that the two polyclonal reagents displayed different serological behavior, i.e., the rabbit anti-rat antiserum cross-reacted much more weakly on mouse Gimap4, indicating that they possessed distinct epitope specificities (Fig. 1 C).

The deficiency in expression of Gimap4 by BN rats was investigated further by analyzing subsets of purified lymphocytes. Flow cytometry was used to generate these cells from the thymuses and LNs of BN and PVG-RT1u, RT7b rats as well as F1 hybrids of these two strains. Western blots of the “wild-type” PVG-RT1u,RT7b cells showed the profile predicted from earlier studies in mice and rats (14, 16, 13), namely the absence of Gimap4 expression in CD4+CD8+ double-positive thymocytes, strong expression in single-positive thymocytes and peripheral T cells, and weaker but significant expression in the double-negative (DN) subset (Fig. 2). Expression by B lymphocytes was low. Similar analysis was performed on BN subsets. Very low levels of Gimap4 were detectable in BN rats although the same subset differences were discernible as for the “wild type” strain, albeit at a lower level of signal. The absence of a Gimap4 signal in DN thymocytes of BN shown in Fig. 2 is presumably due to the overall weakness of signals in this strain and does not indicate a difference in the pattern of expression of this gene in BN rats. Indeed a weak band has been seen for BN DN thymocytes in some experiments using longer exposure. The F1 hybrid rat samples gave strong Gimap4 signals, indicating that the deficiency observed in BN rats is not genetically dominant but probably codominant.

FIGURE 2.

Protein expression of Gimap4 in FACS-sorted populations of thymocytes and lymphocytes in BN vs wild-type (WT) rats. Gimap4 expression was compared in BN, PVG-RT1u,RT7b (WT.), and (BN × PVG-RT1u,RT7b)F1 (WT × BN)F1) rats using MAC 417 in Western blots. Thymocytes were sorted into CD4CD8 DN, double-positive (DP), CD4+ and CD8+ single-positive populations (95–99.8% purity) and LN cells of similar purity were sorted as CD4+ and CD8+ T cells, and B cells (B). The equivalent of 1 × 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.

FIGURE 2.

Protein expression of Gimap4 in FACS-sorted populations of thymocytes and lymphocytes in BN vs wild-type (WT) rats. Gimap4 expression was compared in BN, PVG-RT1u,RT7b (WT.), and (BN × PVG-RT1u,RT7b)F1 (WT × BN)F1) rats using MAC 417 in Western blots. Thymocytes were sorted into CD4CD8 DN, double-positive (DP), CD4+ and CD8+ single-positive populations (95–99.8% purity) and LN cells of similar purity were sorted as CD4+ and CD8+ T cells, and B cells (B). The equivalent of 1 × 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.

Close modal

Bioinformatic searches revealed two forms of rat Gimap4 (Fig. 3). These differ with respect to an AT dinucleotide insertion toward the 3′ end of the transcript but within the open reading frame (ORF). As a result of this frameshift the form with the AT insertion (AT(+)) contains an earlier stop codon, thereby generating a predicted polypeptide that has lost the 21 carboxyl-terminal amino acids of the wild type and that is 18 amino acids shorter than the AT(−) form. The BN genome encodes the AT(+) form; the alternative AT(−) form is similar to the Gimap4 sequences seen in other species, including mouse and human. When Gimap4 was sequenced from three independent rat strains, BN, PVG, and DA, no coding differences other than the AT dinucleotide insertion were encountered although a small number of single nucleotide polymorphisms (SNPs) were recorded. These are annotated in the European Molecular Biology Laboratory (EMBL) Nucleotide Sequence Database (www.ebi.ac.uk/embl/) under accession nos. BC070952, AM285343, and AM285683, respectively.

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.

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.

Close modal

Bioinformatic analysis suggested a second unusual feature of the rat Gimap4 gene. When human and mouse sequences comprising the last exon of Gimap4 were compared, substantial sequence conservation extending through the 3′-UTR to the poly(A) signal was observed (Fig. 4). When rat Gimap4 was entered into the same comparison, however, all sequence conservation ended some 16 nucleotides 3′ of the normal (non-BN) stop codon. A subsequent alignment of sequences from human, chimpanzee, orangoutang, cow, dog, mouse, and rat indicated that the rat has acquired a deletion at this point relative to all the other mammals analyzed. Table I presents the alignment of human, mouse, and rat sequences. To confirm this in silico information, we undertook a 3′ RACE-PCR on Gimap4 from mouse and rat using primers located ∼150 bp 5′ of the normal rat/mouse stop codon. This analysis was done on total RNA from a C57BL/6 mouse and from both BN and PVG-RT1u,RT7b rats. We studied two different rat strains as we wished to discover whether, other than the AT insertion, the BN rat had any additional sequence differences in this region.

FIGURE 4.

Dotter plots showing sequence homology in rat, mouse, and human Gimap4 genes. The rat and mouse forms of Gimap4 are highly conserved within the open reading frame, but in the 3′-UTR there is a ∼1800-bp deletion in the rat as compared with the mouse. The upper panel shows the rat vs mouse comparison, depicting the deletion in the rat gene starting in exon 3 and extending beyond it. The lower panel shows a comparison of human vs mouse.

FIGURE 4.

Dotter plots showing sequence homology in rat, mouse, and human Gimap4 genes. The rat and mouse forms of Gimap4 are highly conserved within the open reading frame, but in the 3′-UTR there is a ∼1800-bp deletion in the rat as compared with the mouse. The upper panel shows the rat vs mouse comparison, depicting the deletion in the rat gene starting in exon 3 and extending beyond it. The lower panel shows a comparison of human vs mouse.

Close modal
Table I.

Multiple sequence alignment of the third exon of Gimap4 (Ian01) from human, mouse, and rata

Multiple sequence alignment of the third exon of Gimap4 (Ian01) from human, mouse, and rata
Multiple sequence alignment of the third exon of Gimap4 (Ian01) from human, mouse, and rata
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).

Our results confirmed as correct the existing database sequences containing the Gimap4 3′-UTRs for both mouse and rat. Hence, we verify that the rat Gimap4 sequence has undergone a ∼1800-bp deletion relative to other mammals immediately downstream of the termination codon, with the consequence that a DNA repeat region containing an MTC (mammalian apparent long-terminal repeat retrotransposon (MaLR)) repeat and a rat MER1B repeat is brought 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 predicted by bioinformatics (Table I) and indicate different message sizes for the two species, these being 1.4 kb in rat and 1.7 kb in mouse, the latter agreeing with previous Northern blot data (14, 16). When we compared the 3′-UTRs of BN and PVG-RT1u,RT7b with each other, we found just three polymorphisms in addition to the AT insertion within the ORF. The three SNPs are located at positions 1052, 1134, and 1219 in the 3′-UTR of the PVG rat (GenBank accession no. AM285343). Because the genomic deletion observed is common to both AT(+) and AT(−) rats, it cannot be directly responsible for the deficiency in Gimap4 at the protein level in BN rats. Searching the rat genome database with the mouse sequence that represents the 1800-bp deletion revealed no significant hits.

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(+)/AT(−) heterozygotes showing high protein expression and 21 AT(+)/AT(+) homozygotes showing low protein expression. Western blots from 18 of the (PVG-RT1n(BN) × BN)F1 × BN rats are shown in Fig. 5,A. The AT(+)/AT(−) 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(+)/AT(+) homozygotes gave either weak bands in LN only or an absence of signal from both the LN and the thymus. The data from a single exceptional rat, an AT(+)/AT(+) rat with high protein expression, are included in Fig. 5 A (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 show that Gimap4 protein expression appears to behave as a Mendelian trait under the control of a gene or genes in, or close to, the Gimap complex.

Table II.

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

BackcrossNo. of RatsGenotypeTissues BlottedPhenotype (Gimap4 Protein)
(PVG-RT1n (BN) × BN)F1 × BN 24 9 × AT(+)/AT(+) LN, spleen, thymus 8 × Low 
  15 × AT(+)/AT(−)  16 × Higha 
     
(PVG-RT1n (BN) × BN)F1 × BN 4 × AT(+)/AT(+) PBLs 4 × Low 
  5 × AT (+)/AT(−)  5 × High 
     
Second backcross of wild-type Gimap4 16 10 × AT(+)/AT(+) PBLs 10 × Low 
 from PVG-RT1u RT7b on to BN  6 × AT(+)/AT(−)  6 × High 
BackcrossNo. of RatsGenotypeTissues BlottedPhenotype (Gimap4 Protein)
(PVG-RT1n (BN) × BN)F1 × BN 24 9 × AT(+)/AT(+) LN, spleen, thymus 8 × Low 
  15 × AT(+)/AT(−)  16 × Higha 
     
(PVG-RT1n (BN) × BN)F1 × BN 4 × AT(+)/AT(+) PBLs 4 × Low 
  5 × AT (+)/AT(−)  5 × High 
     
Second backcross of wild-type Gimap4 16 10 × AT(+)/AT(+) PBLs 10 × Low 
 from PVG-RT1u RT7b on to BN  6 × AT(+)/AT(−)  6 × High 
a

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

FIGURE 5.

Western blots phenotyping BN × PVG (wild type) backcross rats. A, Tissues from 18 (PVG-RT1n(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 (+/+) for homozygous AT(+)/AT(+) and (+/−) 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(+)/AT(+) 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 blots from nine (PVG-RT1n(BN) × BN)F1 × BN backcross rats, the AT(+)/AT(−) heterozygotes exhibited a strong Gimap4 band (lanes b, c, g, h, and i), whereas no band was detected for the AT(+)/AT(+) homozygotes (lanes a, d, e, and f).

FIGURE 5.

Western blots phenotyping BN × PVG (wild type) backcross rats. A, Tissues from 18 (PVG-RT1n(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 (+/+) for homozygous AT(+)/AT(+) and (+/−) 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(+)/AT(+) 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 blots from nine (PVG-RT1n(BN) × BN)F1 × BN backcross rats, the AT(+)/AT(−) heterozygotes exhibited a strong Gimap4 band (lanes b, c, g, h, and i), whereas no band was detected for the AT(+)/AT(+) homozygotes (lanes a, d, e, and f).

Close modal

The observed very low level of Gimap4 in BN T cells could be due either to all BN T cells expressing low levels of this protein or the presence of a small number of T cells expressing normal or high 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-dependent anion channel 1 (VDAC-1) were used as fractionation markers for cytoplasm and membranes, respectively (data not shown). The results in Fig. 6,A show that rat Gimap4 is cytosolic in wild-type cells and, despite the much lower level of expression, the same localization is observed in BN cells. Next, intracellular staining and flow cytometry were used to study Gimap4 expression in purified LN T cells from BN and PVG-RT1n(BN) rats. The results in Fig. 6 B show that all BN T cells express very low levels of Gimap4 (note that the BN profile is slightly shifted relative to the control), whereas all PVG-RT1n(BN) T cells express a high level of Gimap4. This excludes the possibility that there is a subpopulation of high-expressing cells.

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).

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).

Close modal

Having defined large differences in Gimap4 protein expression in immune cells of BN compared with other rats, we wished to determine whether or not these were reflected at the level of mRNA expression. Gimap4 message levels in flow-sorted populations of LN T and B cells and CD4+ and CD8+ single-positive thymocytes from BN and PVG-RT1n(BN) rats were measured using comparative real-time PCR. Transcript expression was measured relative to two control genes, cirhin and 6-phosphofructokinase C (6PFKc), which we had previously selected for their stable expression in the tissues studied (13). In both rat strains the highest Gimap4 expression was seen in LN T cells, whereas thymic CD4 and CD8 cells had lower and relatively similar expression. B cells in both rats had significantly less Gimap4 message than lymphoid T cells. These results were consistent with our previously published data on PVG-RT1u,RT7b and PVG-RT1u,lyp/lyp rats (13). Two-way ANOVA analysis was performed to determine whether there was a difference in mRNA expression between the two strains of rats in four types of cells (Fig. 7 A). No large differences in message levels were observed. BN lymphocytes may express slightly less mRNA, but this difference was of weak significance (p = 0.09). Similarly, no large differences in mRNA levels were observed when spleen RNA isolated from BN and PVG-RT1n(BN) rats was analyzed by Northern blot analysis using a probe for rat Gimap4 mRNA. In addition, no Gimap4 mRNA size difference was seen between the two strains (results not shown).

FIGURE 7.

Gimap4 mRNA levels and protein stability. A, Relative expression of Gimap4 transcript levels in BN and PVG-RT1n (WT) rats in 4 cell types. Real-time PCR results are expressed as a two-way ANOVA relative to two housekeeping genes (HGK). Each value is calculated as the mean of triplicate samples of three separate cell preparations. Error bars show the mean ± 1.0 SE. B, Relative stability of Gimap4 variant proteins from BN and PVG-RT1n (WT). Pulse-chase analysis of Gimap4 protein stability in transfected HEK293T cells was performed as indicated in Materials and Methods. The radiolabeled Gimap4 species are indicated by an arrowhead. The mobilities of coelectrophoresed protein m.w. standards are indicated.

FIGURE 7.

Gimap4 mRNA levels and protein stability. A, Relative expression of Gimap4 transcript levels in BN and PVG-RT1n (WT) rats in 4 cell types. Real-time PCR results are expressed as a two-way ANOVA relative to two housekeeping genes (HGK). Each value is calculated as the mean of triplicate samples of three separate cell preparations. Error bars show the mean ± 1.0 SE. B, Relative stability of Gimap4 variant proteins from BN and PVG-RT1n (WT). Pulse-chase analysis of Gimap4 protein stability in transfected HEK293T cells was performed as indicated in Materials and Methods. The radiolabeled Gimap4 species are indicated by an arrowhead. The mobilities of coelectrophoresed protein m.w. standards are indicated.

Close modal

Having determined that differences in Gimap4 mRNA levels were so small as to be unlikely to explain the observed differences in protein levels between wild-type and BN rats, we were interested to see whether the two proteins made had inherently different stabilities.

To address this, the HEK293T cell line was transfected with plasmids encoding the variant forms of Gimap4 (each with a C-terminal myc-His tag; see Materials and Methods) under the control of a CMV promoter. Newly synthesized proteins were radiolabeled with a short pulse of [35S]methionine and [35S]cysteine. The radiolabel was then chased out of the cells with a large excess of unlabeled amino acids as described in Materials and Methods. At various intervals during the chase, cell lysates were prepared and the radiolabeled Gimap4 was recovered from the lysates by immunoprecipitation. As shown in Fig. 7 B, radiolabeled Gimap4 was still detectable after a 6-h chase period whether the Gimap4 variant expressed was that from BN or PVG-RT1n(BN) rats. In addition, the rate of disappearance of radiolabel from the two variants appears to be grossly similar, suggesting that the half-lives of the two proteins are similar. Similar results were also obtained when the BN variant was tagged at the N terminus or when it was expressed in its native form, suggesting that tagging the protein had no major effect on its turnover rate (results not shown).

Roles in the regulation of cell death have been proposed or demonstrated for members of the Gimap family, most notably in the prosurvival properties of rat Gimap5 (see Introduction); by contrast, recent studies have suggested a prodeath role for Gimap4 in the mouse (11, 16). Initial analysis showed no obvious abnormality in the rate of cell death when FACS-purified BN T cells were cultured for 3 days in tissue culture medium (Fig. 8 A). The numbers of live BN LN T cells declined at approximately the same rate as for wild-type cells (PVG-RT1n) and clearly more slowly than Gimap5-deficient T cells (PVG-RT1u,lyp/lyp) (8).

A more detailed analysis of in vitro apoptotic responses of BN T cells was then performed. This was guided by the recently published analysis of a mouse strain with a targeted deletion in Gimap4 (16). In this study, one of our groups had noticed that T cells from the knockout mice exhibited slower kinetics in the induced “death program” as defined by flow cytometry. Cells appeared to be delayed in proceeding from the annexin V (+), propidium iodide (− (“apoptotic”) stage to the annexin V (+), propidium iodide (+) (“dead”) stage. We therefore performed the same experiment to ask whether T cells from Gimap4-deficient BN rats (AT(+)) would exhibit a similar apoptotic phenotype. As positive controls in these studies we used LN T cells from Gimap4−/− knockout mice (Fig. 8, B and C). As seen previously using mouse splenic T cells (16), in response to various apoptotic stimuli more Gimap4−/− knockout mouse LN cells tended to accumulate at the annexin V (+), propidium iodide (−) (“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 Gimap4−/− knockout and wild-type mouse T cells were somewhat more marked than for Gimap4-deficient and normal rats.

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(+) dinucleotide insertion. Six SNPs within Gimap4 exon 3 defined two major haplotypes. Only the first three SNPs are in the ORF and are silent, whereas the remainder lie in the 3′-UTR (EMBL accession nos. AM285343 and AM285683). Table III organizes the inbred rat strains into the two haplotype groups, with 18 strains in haplotype group 1 and 10 in group 2. BN and MAXX define a subtype of haplotype 1.

Investigations were extended to determine whether the AT(+) form of Gimap4 was present in wild Rattus norvegicus populations. The results are presented in Table IV. Genomic DNA samples were generously donated by colleagues in the United Kingdom, Europe and Japan and were added to a panel we had collected earlier within our own region (see the legend to Table IV for the provenances of the samples tested). Four of the samples from Japan, i.e., those from source F in Table IV, were from rat strains derived by inbreeding from wild-caught animals and thus represented only one haplotype per strain; the remaining 109 DNA samples sequenced were directly from captured or killed wild animals and thus represented two haplotypes per individual. The number of apparent AT(+) homozygotes was 25/109, while the number of heterozygotes was 20/109. Thus, the AT(+) Gimap4 DNA sequence was found in 45 of the 109 samples and the overall frequency of the AT(+) haplotype was ∼32%. The AT(+) haplotype was not detected in samples from some geographic regions/locations, e.g., Cambridgeshire and Essex, U.K., and was, by contrast, common in some others, e.g., Lund, Sweden and Osaka, Japan. We also detected both AT(+) and AT(−) forms of Gimap4 in Rattus rattus DNA samples.

Using newly developed serological reagents, we have discovered a severe deficiency of the Gimap4 GTP binding protein in T cells of the BN rat. This deficiency is linked genetically to the region containing the Gimap gene cluster on rat chromosome 4. Recent evidence has suggested that Gimap4 is a positive regulator of cell death (11, 16). This may be achieved through interaction with the Bcl-2 family members Bax/Bak (11). Gimap4 has an intriguing expression profile across T cell development, with sharp increases associated with intrathymic β-selection and positive selection, followed by high expression in peripheral resting T cells (14, 16, 13, 27). This is clearly confirmed in the rat at the protein level, as seen in Fig. 2. Interestingly, a Gimap4 knockout mouse shows no obvious phenotype in lymphocyte development and selection. Thus, Gimap4 deletion had apparently no effect on thymic positive or negative selection in model (anti-HY) systems (16). Similarly, the successful generation of mature thymocytes and a peripheral T cell repertoire in the BN rat suggests that Gimap4 does not play a critical role in thymic selection events. This conclusion contrasts with suggestions made by Nitta and colleagues on the basis of the overexpression of Gimap4 in mouse fetal thymic organ culture in vitro (11). We suggest that the apoptosis of double-positive thymocytes observed in that study after the forced expression of Gimap4 in DN thymocytes is a nonphysiological phenomenon.

Further analysis of the Gimap4 knockout mice, however, revealed an intriguing phenotype in T lymphocytes in response to in vitro apoptotic stimulation in that knockout cells are delayed in their death program, as defined by the two flow cytometric criteria annexin V and propidium iodide staining for phosphatidylserine exposure and membrane permeabilization, respectively (16). These data suggested that Gimap4 may be acting to accelerate the death process. In the present study. Gimap4-deficient BN LN T cells were put through a similar set of apoptotic tests. Our conclusion is that these cells exhibit a similar phenotype to that of the knockout mouse, although somewhat milder. Thus, with strong apoptotic stimuli clear differences compared with wild-type rat cells were detectable, but the set of analyses conducted in parallel using the knockout mouse (and their controls) showed generally more marked effects. This perhaps reflects the fact that the BN rat appears to be a hypomorph for Gimap4 rather than a complete knockout; we estimate that BN T cells express Gimap4 at ∼5% of normal levels. Analysis of Gimap4 by intracellular staining and flow cytometry indicates that this low level of expression is common to all BN T cells and that there is not a small subset of high-expressing cells that accounts for the blotting signal in this strain. It is possible that this low but finite level of expression allows partial Gimap4 function to persist in BN T cells.

It is probable that the BN Gimap4 phenotype is determined by an AT dinucleotide insertion. In analyses of backcross populations segregating for the two Gimap4 alleles (BN vs wild type), a single animal in 49 showed a lack of agreement between its Gimap4 expression and genotyping for the AT insertion. Although this may indicate control of Gimap4 expression by another closely linked locus, it may, in contrast, reflect the individual-to-individual variations in Gimap4 expression that were apparent among the AT(+)/AT(+) homozygotes in our Western blotting analysis. Protein expression analysis of this backcross litter was performed post mortem, so it was not possible to pedigree test this discordant rat. No similar individual or the obverse, i.e., an AT(+)/AT(−) heterozygote with low protein expression, has yet been encountered. We have now moved over to using blood leukocytes in our Gimap4 phenotyping of rats (Fig. 5,B and Table II) and thereby we will in future be able to follow up unusual individuals by retesting and breeding.

Despite the results with this single backcross rat, it is our preferred hypothesis that the low protein phenotype in BN Gimap4 is due to the AT dinucleotide insertion. This insertion is predicted to bring about premature termination of the BN form of Gimap4 polypeptide, such that the 21 carboxyl-terminal amino acids of the wild-type sequence are lost and replaced by just three different residues. Among the 28 inbred laboratory rat strains screened, this dinucleotide insertion was seen only in BN and one other strain, MAXX, which has BN as an ancestor (28). We found little difference between the mRNA levels of the BN variant and the wild-type alleles of Gimap4 by using quantitative real-time PCR (Fig. 6) or Northern blotting (data not shown), indicating that the effect of the AT(+) variant on protein expression is posttranscriptional. It is possible that the carboxyl-terminal truncation of the Gimap4 polypeptide causes instability at the protein level. We are currently undertaking experiments to try to identify the mechanism involved. To date we have established that the truncation of Gimap4 does not lead to intracellular mislocalization from the cytosol (Fig. 6,A). Also absent is a gross change in polypeptide stability due to the truncation as assessed by a pulse-chase analysis in transfected cells (Fig. 7 B 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 have a pronounced effect on the amount of protein accumulating. The specific kinetic details of Gimap4 synthesis and degradation in lymphocytes may be essential to the phenotype observed.

As noted in Table III (although not included in our panel of DNAs), the Gimap4 AT(+) sequence allele has also been detected in DR-BB/H rats (5). These and some related diabetes-resistant BB sublines, which are Gimap5+/+, are frequently used as controls in studies of autoimmune diabetes in BB-DP rats. That this control strain may also possess a functional defect in a member of the Gimap protein family should be taken into account in interpreting past literature and future investigations. (As mentioned earlier, the BB-DR/Ed subline is different in that it is Gimap5−/− but fails to develop autoimmune diabetes mellitus, presumably on account of resistance alleles at other susceptibility loci.) (17).

The C-terminal region of the Gimap4 polypeptide encoded downstream of the AT insertion site is highly conserved between wild type rat and mouse, with only one conservative difference, 327R↔K, present. This region lies outside of the domains/motifs of Gimap4 so far characterized (Fig. 3 A), namely the GTP binding domain (with G1–G5 motifs), the conserved sites for protein kinase C phosphorylation, and the IQ domain that mediates Ca2+-regulated interaction with calmodulin (16). It will be interesting to determine what function(s) or intermolecular interaction(s) are controlled by this short sequence. Similarly, as and when suitable assays are available it will be interesting to ascertain what functions, if any, the BN Gimap4 protein variant retains in the absence of this peptide sequence.

In addition to the AT dinucleotide insertion found in the BN rat, we have described an additional and larger scale difference between the rat and mouse Gimap4 genes, although this one, which may be unique to the rat, appears to be true of all rat strains. It is an ∼1800-bp deletion beginning at or around the wild-type stop codon. This deletion brings a DNA repeat region into close juxtaposition with the Gimap4 ORF. Some sequence from these repeats is thereby present in the rat Gimap4 3′-UTR. Rat has lost the poly(A) addition signal that is seen in human and mouse (Table I), and all other mammals analyzed (data not shown), and evolution has crafted a different such site elsewhere. This leads to different message sizes for rat and mouse Gimap4 (1.4kb vs 1.7kb). It is possible that this major change to the 3′-region of the rat Gimap4 gene is required to achieve the particular expression of truncated Gimap4 protein observed in the BN rat, i.e., given the different 3′-UTRs we cannot be certain that an identical AT insertion in the mouse would produce the same outcome for the level of protein expression.

As mentioned above, our screen of laboratory rat strains identified only two carrying the Gimap4 AT dinucleotide insertion. At that stage in our investigation it therefore seemed probable that this mutation, like the previously described mutation in Gimap5 of the BB rat (29), had arisen spontaneously in a laboratory rat colony and been fixed and maintained by inbreeding. Nevertheless, we embarked on an analysis of wild rats with unexpected results. Table IV presents a summary of the data from the 113 samples obtained from different parts of the world. These results show not only that we were able to detect the AT(+) allele in the wild but also that this allele was present with a surprisingly high overall frequency of ∼32%. Indeed, in two sets of samples, those from Lund, Sweden and Osaka, Japan, a substantial number of apparent AT(+) homozygotes was present. These findings suggest that the BN-type AT(+) allele of Gimap4 is not necessarily deleterious but that it confers some selective advantage that maintains it in the wild rat population in balance with the more mouse-like AT(−) allele. If this is true, it raises the question of whether Gimap4 variation mediates some developmental or immunomodulatory variation in T cell behavior that is under pathogen-driven selection. It would be very interesting to know whether this is mediated through the apoptosis-related function of Gimap4 or some other, as yet unidentified, function(s) of this molecule.

The BN rat has long been a favored strain in both immunological and toxicological research and was also chosen as the base strain for the Rat Genome Mapping Project (30). It appears to be biased in its immune responses toward Th2-type immunity and it is capable of generating very high IgE responses (31). Similarly, it is highly susceptible to the induction of Th2-mediated autoimmune diseases such as mercuric chloride-induced glomerulonephritis (32) while often being resistant to Th1-mediated diseases, e.g., experimental allergic encephalomyelitis (33). A number of unusual cellular immunological phenotypes have been reported, such as a high CD4:CD8 T cell ratio (34), skewed CD45RC expression (35), and low mitogen responsiveness (36, 37). Although good progress is being made toward the genetic dissection of these phenotypes (38), this work is not complete and it will be interesting to determine whether the BN Gimap4 variant, as a T cell GTPase apparently susceptible to regulation through the TCR (14, 15, 16), plays any part in them. To this end a BN congenic strain carrying the PVG (wild type) Gimap4 is under preparation.

We thank Arthur Davis for flow sorting, Maureen Hamon for assistance with immunochemistry, and Anne Segonds-Pichon for help with statistical analysis.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by Competitive Strategic Grant funding from the Biotechnology and Biological Sciences Research Council (BBSRC), BBSRC GAIN Grant 202/GAN13085 (to G.W.B. and J.R.M.), and The Netherlands Cancer Institute Grant SFN SFR 2.1.29 (to H.J.). S.S. was supported by a travel allowance from Boehringer Ingelheim Fonds.

4

Abbreviations used in this paper: BB-DP, BioBreeding diabetes prone; BN, Brown Norway; DN, double negative; LN, lymph node; ORF, open reading frame; SNP, single nucleotide polymorphism; UTR, untranslated region.

1
Krücken, J., R. M. U. Schroetel, I. U. Müller, N. Saïdani, P. Marinovski, P. M. Benten, O. Stamm, F. Wunderlich.
2004
. Comparative analysis of the human gimap gene cluster encoding a novel GTPase family.
Gene
341
:
291
-304.
2
Guttmann, R. D., E. Colle, F. Michel, T. Seemayer.
1983
. Spontaneous diabetes mellitus syndrome in the rat, II: T lymphopenia and its association with clinical disease and pancreatic lymphocytic infiltration.
J. Immunol.
130
:
1732
-1735.
3
Jackson, R. A., J. B. Buse, R. Rifai, D. Pelletier, E. L. Milford, C. B. Carpenter, G. S. Eisenbarth, R. M. Williams.
1984
. Two genes required for diabetes in BB rats: evidence from cyclical intercrosses and backcrosses.
J. Exp. Med.
159
:
1629
-1636.
4
Cousins, L., M. Graham, R. Tooze, C. Carter, J. R. Miller, F. M. Powrie, G. G. MacPherson, G. W. Butcher.
2006
. Eosinophilic bowel disease controlled by the BioBreeding rat-derived lymphopenia/Gimap5 gene.
Gastroenterology
131
:
1475
-1485.
5
Hornum, L., J. Romer, H. Markholst.
2002
. The diabetes-prone BB rat carries a frameshift mutation in Ian4, a positional candidate of Iddm1.
Diabetes
51
:
1972
-1979.
6
MacMurray, A. J., D. H. Moralejo, A. E. Kwitek, E. A. Rutledge, B. Van Yserloo, P. Gohlke, S. J. Speros, B. Snyder, J. Schaefer, S. Bieg, et al
2002
. Lymphopenia in the BB rat model of type 1 diabetes is due to a mutation in a novel immune-associated nucleotide (Ian)-related gene.
Genome Res.
12
:
1029
-1039.
7
Ramanathan, S., K. Norwich, P. Poussier.
1998
. Antigen activation rescues recent thymic emigrants from programmed cell death in the BB rat.
J. Immunol.
160
:
5757
-5764.
8
Hernández-Hoyos, G., S. Joseph, N. G. Miller, G. W. Butcher.
1999
. The lymphopenia mutation of the BB rat causes inappropriate apoptosis of mature thymocytes.
Eur. J. Immunol.
29
:
1832
-1841.
9
Sandal, T., L. Aumo, L. Hedin, B. T. Gjertsen, S. O. Doskeland.
2003
. Irod/Ian5: an inhibitor of γ-radiation- and okadaic acid-induced apoptosis.
Mol. Biol. Cell
14
:
3292
-3304.
10
Pandarpurkar, M., L. Wilson-Fritch, S. Corvera, H. Markholst, L. Hornum, D. L. Greiner, J. P. Mordes, A. A. Rossini, R. Bortell.
2003
. Ian4 is required for mitochondrial integrity and T cell survival.
Proc. Natl. Acad. Sci. USA
100
:
10382
-10387.
11
Nitta, T., M. Nasreen, T. Seike, A. Goji, I. Ohigashi, T. Miyazaki, T. Ohta, M. Kanno, Y. Takahama.
2006
. IAN family critically regulates survival and development of T lymphocytes.
PLoS Biol.
4
:
E103
12
Reuber, T. L., F. M. Ausubel.
1996
. Isolation of Arabidopsis genes that differentiate between resistance responses mediated by the RPS2 and RPM1 disease resistance genes.
Plant Cell
8
:
241
-249.
13
Dion, C., C. Carter, L. Hepburn, W. J. Coadwell, G. Morgan, M. Graham, N. Pugh, G. Anderson, G. W. Butcher, J. R. Miller.
2005
. Expression of the Ian family of putative GTPases during T cell development and description of an Ian with three sets of GTP/GDP-binding motifs.
Int. Immunol.
17
:
1257
-1268.
14
Poirier, G. M., G. Anderson, A. Huvar, P. C. Wagaman, J. Shuttleworth, E. Jenkinson, M. R. Jackson, P. A. Peterson, M. G. Erlander.
1999
. Immune-associated nucleotide-1 (IAN-1) is a thymic selection marker and defines a novel gene family conserved in plants.
J. Immunol.
163
:
4960
-4969.
15
Cambot, M., S. Aresta, B. Kahn-Perlès, J. de Gunzberg, P.-H. Roméo.
2002
. Human immune associated nucleotide 1: a member of a new guanosine triphosphatase family expressed in resting T and B cells.
Blood
99
:
3293
-3301.
16
Schnell, S., C. Démollière, P. van den Berk, H. Jacobs.
2006
. Gimap4 accelerates T cell death.
Blood
108
:
591
-599.
17
Joseph, S., A. G. Diamond, W. Smith, J. D. Baird, G. W. Butcher.
1993
. BB-DR/Edinburgh: a lymphopenic, non-diabetic subline of BB rats.
Immunology
78
:
318
-328.
18
Galfrè, G., C. Milstein, B. Wright.
1979
. Rat × rat hybrid myelomas and a 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, F. Speleman.
2002
. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes.
Genome Biol.
3
:
RESEARCH0034
20
Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, D. G. Higgins.
1997
. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools.
Nucleic Acids Res.
15
:
4876
-4882.
21
Sonnhammer, E, L. L., and R. Durbin. 1995. A dot-matrix program with dynamic threshold control suited for genomic DNA and protein sequence analysis. Gene 167: GC1–GC10.
22
Rothenburg, S., F. Haag, F. Koch-Nolte, C. Carter, M. Graham, G. W. Butcher.
2005
. Characterization of multiple alleles of the T-cell differentiation marker ART2 (RT6) in inbred and wild rats.
Immunogenetics
57
:
739
-749.
23
Pelz, H.-J., S. Rost, M. Hünerberg, A. Fregin, A.-C. Heiberg, K. Baert, A. D. MacNicoll, C. V. Prescott, A.-S. Walker, J. Oldenburg, C. R. Miller.
2005
. The genetics basis of resistance to anticoagulants in rodents.
Genetics
170
:
1839
-1847.
24
Dissen, E., S. V. Hunt, B. Rolstad, S. Fossum.
1993
. Localization of the rat T-cell receptor β-chain and carboxypeptidase A1 loci to chromosome 4.
Immunogenetics
37
:
153
-156.
25
Olofsson, P., J. Holmberg, J. Tordsson, S. Lu, B. Åkerström, R. Holmdahl.
2003
. Positional identification of Nef1 as a gene that regulates arthritis severity in rats.
Nat. Genet.
33
:
25
-32.
26
Kondo, Y., K. Ohno, S. Oda, T. Serikawa.
1996
. Genetic profiles of three inbred strains derived from wild rats (Rattus norvegicus) trapped in Japan.
Exp. Anim.
45
:
405
-409.
27
Mick, V. E., T. K. Starr, T. M. McCaughtry, L. K. McNeil, K. A. Hogquist.
2004
. The regulated expression of a diverse set of genes during thymocyte positive selection in vivo.
J. Immunol.
173
:
5434
-5444.
28
Festing, M..
1979
. Inbred strains. H. Baker, and J. Lindsey, and S. Weisbroth, eds.
The Laboratory Rat Vol. I: Biology and Diseases
55
-72. Academic Press, New York.
29
Chappel, C., W. Chappel.
1983
. The discovery and development of the BB rat colony: an animal model of spontaneous diabetes mellitus.
Metabolism
32
:
S8
-S10.
30
Rat Genome Sequencing Project Consortium
2004
. Genome sequence of the Brown Norway rat yields insights into mammalian evolution.
Nature
428
:
493
-521.
31
Prouvost-Danon, A., A. Abadie, C. Sapin, H. Bazin, P. Druet.
1981
. Induction of IgE synthesis and potentiation of anti-ovalbumin IgE antibody response by HgCl2 in the rat.
J. Immunol.
126
:
699
-702.
32
Sapin, C., E. Druet, P. Druet.
1977
. Induction of anti-glomerular basement membrane antibodies in the Brown-Norway rat by mercuric chloride.
Clin. Exp. Immunol.
28
:
173
-178.
33
Happ, M., P. Wettstein, B. Dietzschold, E. Heber-Katz.
1988
. Genetic control of the development of experimental allergic encephalomyelitis in rats: separation of MHC and non-MHC effects.
J. Immunol.
141
:
1489
-1494.
34
Damoiseaux, J., B. Cautain, I. Bernard, M. Mas, P. van Breda Vriesman, P. Druet, G. Fournié, A. Saoudi.
1999
. A dominant role for the thymus and MHC genes in determining the peripheral CD4/CD8 T cell ratio in the rat.
J. Immunol.
163
:
2983
-2989.
35
Powrie, F., D. Fowell, A. J. McKnight, D. Mason.
1991
. Lineage relationships and functions of CD4+ T-cell subsets in the rat.
Res. Immunol.
142
:
54
-58.
36
Newlin, C., D. Gasser.
1973
. Genetic control of the in vitro responses of rat peripheral blood lymphocytes to phytohemagglutinin and concanavalin A.
J. Immunol.
110
:
622
-628.
37
Fu, Y., E. Blankenhorn.
1992
. Nitric oxide-induced anti-mitogenic effects in high and low responder rat strains.
J. Immunol.
148
:
2217
-2222.
38
Fournié, G., B. Cautain, E. Xystrakis, J. Damoiseaux, M. Mas, D. Lagrange, I. Bernard, J.-F. Subra, L. Pelletier, P. Druet, A. Saoudi.
2001
. Cellular and genetic factors involved in the difference between Brown Norway and Lewis rats to develop respectively type-2 and type-1 immune-mediated diseases.
Immunol. Rev.
184
:
145
-160.