ADP-ribosylation of membrane proteins on mouse T cells by ecto-ADP-ribosyltransferase(s) (ARTs) can down-regulate proliferation and function. The lack of mAbs against mouse ARTs has heretofore prevented analysis of ART expression on T cell subsets. Using gene gun technology, we immunized a Wistar rat with an Art2b expression vector and produced a novel mAb, Nika102, specific for ART2.2, the Art2b gene product. We show that ART2.2 is expressed as a GPI-anchored protein on the surface of mature T cells. Inbred strain-dependent differences in ART2.2 expression levels were observed. C57BL/6J and C57BLKS/J express the Ag at high level, with up to 70% of CD4+ and up to 95% of CD8+ peripheral T cells expressing ART2.2. CBA/J and DBA/2J represent strains with lowest expression levels. T cell-deficient mice and NZW/LacJ mice with a defective structural gene for this enzyme were ART2.2 negative. In the thymus, ART2.2 expression is restricted to subpopulations of mature cells. During postnatal ontogeny, increasing percentages of T cells express ART2.2, reaching a peak at 6–8 wk of age. Interestingly, ART2.2 and CD25 are reciprocally expressed: activation-induced up-regulation of CD25 is accompanied by loss of ART2.2 from the cell surface. Nika102 thus defines a new differentiation/activation marker of thymic and postthymic T cells in the mouse and should be useful for further elucidating the function of the ART2.2 cell surface enzyme.
ADP-ribosylation represents an important posttranslational protein modification affecting key biological functions in pro- and eucaryotic organisms (1, 2). In the reaction, the ADP-ribose moiety of NAD+ is transferred onto a specific amino acid residue in a target protein, while the nicotinamide moiety is released. The ADP-ribosyltransferase family includes some of the most potent bacterial toxins, such as diphtheria and cholera toxins. These toxins interfere with cellular functions by catalyzing mono-ADP-ribosylation of key cellular target proteins in their human hosts, such as elongation factor EF2 and the α subunit of heterotrimeric G proteins (3, 4, 5). Ample biochemical evidence has shown that endogenous mono-ADP-ribosylation reactions also occur in animal tissues. Recent findings suggest that this posttranslational protein modification may be used to control important physiological functions such as the induction of long-term potentiation in the brain, terminal muscle cell differentiation, and the cytotoxic activity of T cells (6, 7, 8). In the mouse, ADP-ribosylation of cell surface proteins by a GPI-anchored ADP-ribosyltransferase (ART)3 activity has been shown to inhibit important T cell functions such as cell proliferation, target cell binding, and cytolytic activity (9). LFA-1 and CD8 have been identified as key target proteins of ADP-ribosylation on the T cell surface (10, 11, 12).
Multiple ecto-ADP-ribosyltransferase (ART)-encoding genes have been identified in mice (13, 14, 15, 16, 17). A duplicated pair of genes on mouse chromosome 7, designated Art2a and Art2b, are homologues of the single RT6 locus in the rat (13, 18, 19, 20). There are two alloantigenic forms of RT6 expressed in rats as 25–35-kDa GPI-anchored maturation markers on the surface of T cells (21, 22). RT6-expressing T cells exert a regulatory role in rat models for autoimmune insulin-dependent diabetes mellitus (23, 24). Although the two alloantigenic forms of RT6 degrade NAD+, it is unclear whether they are capable of ADP-ribosylating T cell surface proteins other than themselves (25, 26, 27). The recombinant proteins expressed from mouse Art2a and Art2b genes (ART2.1 and ART2.2, respectively) ADP-ribosylate a variety of synthetic substrates (28, 29).
As mentioned above, ADP-ribosylation of integral T cell membrane proteins is associated with CTL down-regulation (9). The lack of mAbs against the mouse ART gene products has heretofore made it difficult to establish whether the T cell activities were indeed encoded by the RT6 homologues or by another member of the ART gene family (30, 31). In this study, we describe the production and staining characteristics of a rat mAb, Nika102, that reacts with the mouse ART2.2 ortholog of rat RT6. In this first report of a mAb against mouse T cell surface-expressed ART, we describe the use of this Ab for monitoring the Ag expression during postnatal T cell development and during T cell activation. We further demonstrate considerable variation among inbred strains in the percentages of T cells expressing this Ag, and we show that activation of T cells induces rapid Ag release from the cell surface, while other GPI-anchored proteins such as Thy-1 (CD90) and Ly-6A/6E are not affected.
Materials and Methods
Wistar rats were from the animal facility of the University Hospital (Hamburg, Germany). ALS/Lt, ALR/Lt, NOD/Lt, NOR/Lt, NOcCB-1/Lt, NZO/Lt, and C3H/HeJBir mice were obtained from a research colony (E.H.L.); C57BLKS/J, C57BL/6J, NZB/BINJ, BXSB/MpJ, BALB/cByJ, Cast/Ei, CBA/J, C57BL/6.Tcrbtm1Mom, 129/SvJ, DBA/2J, (NZB × NZW)F1/J, C57BL/6.scid, and NZW/LacJ (f/8) mice were obtained from the Animal Resources Unit of The Jackson Laboratory (Bar Harbor, ME); CTS/Shi mice were purchased from TGC (Tokyo, Japan) with the sale stipulation that they could not be distributed by the purchaser. For dexamethasone-mediated depletion of thymocytes, mice received a single injection i.p. of 0.5 mg dexamethasone (Sigma, St. Louis, MO) in 0.5 ml PBS or 0.5 ml PBS alone.
mAbs used in this study for immunofluorescence staining and activation assays include anti-CD3ε (145-2C11), anti-CD4 (GK1.5), anti-CD5 (53-7.3), anti-CD8 (53-6.72), anti-CD24 (J11d), anti-CD25 (7D4), anti-CD62L (Mel-14), anti-CD69 (H1.2F3), anti-CD90 (HO13.4), anti-γδTCR (GL3), anti Ly-6A/E (Sca-1, E13-161), and anti-Ly-6C (AL-21). Biotin, PE, and FITC conjugates were purchased from PharMingen (San Diego, CA). Streptavidin-Alexa350 was purchased from Molecular Probes (Eugene, OR). Polyclonal rabbit antiserum K48 was raised against a peptide derived from the deduced amino acid sequence of ART2.2 (residues 48–59), as described previously (13). GK1.5 (anti-CD4) was labeled with Cy3 using a commercially available kit (FluroLink Cy3 Reactive Dye; Amersham, Piscataway, NJ).
Cloning and expression of rART2.2
Vectors for expressing Art2b in Escherichia coli and mammalian cells were cloned by standard techniques (32), replacing sequences for the N-terminal leader and/or C-terminal GPI-signal sequence. In brief, PCR products generated from Art2b cDNA with fusiogenic primers derived from the published Art2b sequence (13) were restriction digested, gel purified, and cloned into pASK (Biometra, Göttingen, Germany), pCDM8.AP (kindly provided by Sandy Zurawski, DNAX Research Institute, Palo Alto, CA), and pME.CD8LF (kindly provided by Rob Kastelein, DNAX Research Institute). The pASK.ART2.2-HF construct encodes the OMP leader (E. coli outer membrane protein) in place of the ART2.2 leader and a chimeric His(6x)-FLAG tag (HF) (33) in place of the ART2.2 GPI-signal sequence. The pCDM8.ART2.2-AP construct retains the ART2.2 leader, but encodes AP in place of the GPI-signal sequence. The pME.CD8LF-ART2.2 construct encodes the CD8 leader, followed by a FLAG tag in place of the ART2.2 signal sequence, but retains the ART2.2 GPI-signal sequence.
C-terminally His6x/FLAG-tagged ART2.2-HF was expressed in E. coli periplasm, as described previously (34). In brief, 1 L cultures of pASK.ART2.2-HF-transformed E. coli NM522 cells were propagated at 24°C for 20 h, and cells were harvested by centrifugation and resuspended in 10 ml lysis buffer (20 mM Tris (pH 8) and 500 mM glucose) for 30 min at 4°C. Lysates were cleared by high speed centrifugation for 15 min at 4°C, and supernatants were stored at 4°C. For purification of rART2.2, 200 μl Talon-Matrix (Clontech, Palo Alto, CA) was added to 10 ml lysate. After overnight incubation at 4°C, matrix-bound protein was pelleted by low speed centrifugation and washed four times with PBS. For preparation of soluble ART2.2, the matrix was pelleted again and resuspended in 100 mM EDTA, pH 7, for 10 min at room temperature. The matrix was pelleted by centrifugation and the supernatant was dialyzed against PBS.
For the production of ART2.2-AP fusion protein, human kidney 293T cells were lipofectamine (Life Technologies, Grand Island, NY) transfected with pCDM8.ART2.2-AP (20 μg/107 cells). After 5 days of culture, supernatants were harvested, cleared by high speed centrifugation, and stored at 4°C. N-terminally FLAG-tagged ART2.2 was expressed in mouse EL4 lymphoma cells after electroporation with pME.CD8LF-ART2.2 (20 μg/107 cells) and selection of stable transfectants with G418.
Immunization and generation of mAbs
pME.CD8LF-ART2.2 was coated onto 1-μm gold particles (Bio-Rad, Hercules, CA). DNA-coated gold particles (1 μg DNA/mg gold) were injected ballistically into the abdominal skin of Wistar rats with the Helios gene gun (Bio-Rad, Munich, Germany) at a pressure setting of 400 psi (6 shots per animal at each immunization). Gene gun immunization was repeated three times every 3–5 wk. Five days before sacrificing, animals received 50 μg purified rART2.2-HF in 500 μl PBS i.v. and 50 μg rART2.2-HF attached to 50 μl Talon-agarose beads in 500 μl PBS i.p. Splenocytes were fused with the Sp2/0 myeloma cell line using standard polyethylene glycol fusion techniques with PEG 6000 (Boehringer Mannheim, Mannheim, Germany). A total of 100 μl of supernatant was collected from wells containing growing cells after 14 to 20 days in selective medium and was tested by ELISA for reactivity with rART2.2-AP fusion protein. Supernatants from positive wells were further tested for reactivity with ART2.2-transfected EL4 cells and for lack of reactivity with parental EL4 cells. Hybridomas from positive wells were cloned by limiting dilution in the presence of irradiated rat thymocyte feeder layers.
ELISA was performed on 96-well plates that had been coated overnight with goat anti-rat Ig (Pierce, Rockford, IL) at 10 μg/ml in PBS. After blocking of the anti-rat Ig-coated plates with 10% goat serum, hybridoma supernatants were added for 30 min at room temperature. Plates were washed and then incubated with 293T cell supernatants containing ART2.2-AP fusion protein (diluted 1/4 in PBS). After 30 min at room temperature, plates were washed again and incubated with p-nitrophenylphosphate (Pierce) for 30 min at 37°C.
For purification of mAbs, Ab was precipitated from hybridoma supernatants with an equal volume of saturated ammonium sulfate (35). Precipitates were washed, and resuspended in and dialyzed against PBS. Ab was further purified by passage over a gel filtration column (molecular mass cutoff, 1 × 106 kDa; Bio-Rad, Hercules, CA). Purified Ab was conjugated to FITC (Sigma) following standard protocols (35). mAb isotypes were determined by Ouchterlony assay using a commercially available rat Ig isotyping kit (The Binding Site, Birmingham, U.K.).
Preparation of cells and immunofluorescence analyses
For preparation of PBLs, animals were bled retroorbitally using heparinized capillary tubes and blood was diluted 10-fold with HBSS containing 5 mM EDTA. In case of thymus, extraneous blood and parathymic lymph nodes were carefully removed before tissue disruption. Single cell suspensions from thymus, spleen, and lymph nodes were prepared in HBSS/2% FCS by passage through Nitex membrane (110 μm mesh; Tetko, Kansas City, MO). Erythrocytes in blood and spleen cell suspensions were lysed by incubation in two washes of Gey’s buffer for 5 min at 4°C, followed by a wash in FACS buffer. Cells (1 × 106) were incubated with saturating amounts of Abs for 30 min at 4°C and were then washed once in 2 ml HBSS. After washing, cells were resuspended in 250 μl FACS buffer (PBS, 1% BSA, 0.05% NaN3) and 10 μl propidium iodide (20 μg/ml). Cells were analyzed by flow cytometry on a FACStar or FACScan (Becton Dickinson, Mountain View, CA) using viable cell gating through exclusion of propidium dye.
Stimulation of cells
PMA (100 ng/ml) and ionophore A23187 (200 ng/ml) (Sigma) were added to cells (106/ml) suspended in HBSS. Cells were incubated for 2 h at 37°C and subjected to immunofluorescence analyses, as described above, with the exception that NaN3 was omitted from FACS buffer. For kinetic analyses, cells were prestained with appropriately labeled Abs before addition of PMA and A23187. For analysis of coexpression of CD25 and ART2.2, splenocytes (2 × 106 cells/ml) from 12-wk-old C57BL/6J males were cultured in RPMI medium supplemented with 10% FCS, l-glutamine, HEPES, and antibiotics for 20 h at 37°C with or without 2 μg/ml anti-CD3 mAb. Then the cells were collected, washed in PBS/1% BSA/0.05% sodium azide, and stained with biotin-conjugated A102 mAb and anti-CD25 FITC, followed by incubation with PE-labeled streptavidin, and analyzed by flow cytometry, as described above.
Treatment of cells with PI-PLC
Cell suspensions were washed with PBS, resuspended at 108 cells/ml in RPMI medium with or without 1 U PI-PLC (Molecular Probes), and incubated on a roller for 1 h at 37°C. Cells were pelleted by centrifugation, and supernatants were harvested and stored at −20o. Cells were washed once with PBS and subjected to immunofluorescence analyses, as described above.
Immunoprecipitation and Western blotting
E. coli lysates and supernatants of PI-PLC-treated cells were cleared by high speed centrifugation. After addition of Abs, samples were rotated for 20 h at 4°C. Protein G-Sepharose was then added to bind Abs and immune complexes. After rotation for 60 min at 4°C, the matrix was pelleted by centrifugation in an Eppendorf centrifuge, washed four times with TBS, 1% Triton X-100, and resuspended in SDS-PAGE sample buffer. All samples were heated for 5 min at 95°C and cleared by centrifugation before loading onto precast 10–12% polyacrylamide gels (Novex, Frankfurt, Germany). Gels were run in Tricine buffer under reducing conditions. Fractionated proteins were blotted onto nitrocellulose or polyvinylidene difluoride membranes and membranes were blocked with 10% goat serum in TBS.
Blots were incubated for 2–16 h at 4°C with anti-ART2.2 peptide rabbit antiserum K48 (13) diluted 1/2000 in TBS, 0.5% Tween 20 (TBST), and 10% goat serum, and washed extensively in TBST. Secondary reagent for detection of bound K48 Abs was peroxidase-conjugated goat anti-rabbit Ig (1:5000) (Amersham). After extensive washing in TBST, bound Ab was detected with the enhanced chemiluminescence (ECL) system, according to the manufacturer’s instructions and by exposure to ECL film (Amersham).
Expression of FLAG-tagged ART2.2 in stably transfected EL4 cells
Because no Abs suitable for detecting ART2.2 by fluorocytometry were available at the time, we engineered a FLAG tag onto the N terminus of ART2.2 to make the protein detectable by the anti-FLAG tag mAb, M2 (33). To this end, we constructed the expression vector pME.CD8LF-ART2.2, in which the presumptive N-terminal ART2.2 leader was replaced by the CD8 leader, followed by a FLAG tag, and in which the presumptive C-terminal GPI-anchor signal peptide of ART2.2 was retained. Fig. 1 shows FACS analyses of EL4 cells stably transfected with this expression vector. ART2.2-transfected cells are stained by M2 (Fig. 1 a), while parental cells are negative (not shown).
Rat RT6 is attached to the cell membrane via a GPI anchor that can be cleaved by bacterial PI-PLC (36). Staining of ART2.2-transfected cells by M2 also is sensitive to treatment of cells with PI-PLC (Fig. 1 a). These results are consistent with the interpretation that M2 recognizes FLAG-tagged, GPI-anchored ART2.2 on these transfectants.
Production of mAb Nika102 to mouse ART2.2
Having shown that pME-CD8LF-ART2.2 directs cell surface expression of ART2.2, we reasoned that this vector might be useful also for raising ART2.2-specific antisera by DNA immunization. To this end, Wistar rats were gene gun immunized with pME.CD8LF-ART2.2. Sera were tested for reactivity with ART2.2-transfected EL4 cells. Spleen cells from animal R8, whose immune serum contained ART2.2-specific Abs (Fig. 1,b), were fused with the mouse myeloma cell line Sp2/0. Supernatants of growing hybridoma cultures were screened by ELISA for reactivity with ART2.2-AP fusion protein produced by appropriately transfected 293T cells (not shown). Supernatants from positive wells were further tested for reactivity with pME.CD8LF-ART2.2-transfected cells and for lack of reactivity with mock-transfected cells. One hybridoma that showed particularly bright staining with ART2.2-transfected cells (Fig. 1 c) was selected for further analyses. The mAb produced by this hybridoma, Nika102, was determined to be of the IgG2a isotype by Ouchterlony assay (data not shown).
Nika102 and an anti-ART2.2 peptide antiserum react with the same molecule
To further corroborate the specificity of Nika102 for ART2.2, we next attempted to immunoprecipitate rART2.2 with Nika102. As shown in Fig. 2, mAb Nika102 specifically immunoprecipitates proteins with apparent molecular mass of 30 kDa (lane 4) and 45 kDa (lane 10) from lysates of ART2.2-transformed E. coli and ART2.2-transfected EL4 cells, respectively, but not from mock-transfected cells (lane 6). The same proteins are detected by anti-ART2.2 peptide antiserum K48, which reacts with denatured ART2.2 in Western blot analyses (13). These results demonstrate that Nika102 recognizes ART2.2.
ART2.2 is expressed by the majority of peripheral T cells and by a small fraction of thymocytes
Rat RT6 is expressed by the majority of peripheral helper and cytotoxic T cells, but not by B cells or thymic lymphocytes, and is down-regulated on activated T cells (22, 37). To assess the distribution of ART2.2 on mouse T cells, we performed flow-cytometric analyses on C57BL/6J mouse lymphocytes using FITC-conjugated Nika102 in combination with Abs against the T cell markers CD3, CD4, CD8, and CD25 (IL-2R). As shown in Fig. 3 and Table I, the majority of peripheral CD3+ cells (Fig. 3, b and c), but only a small fraction of CD3+ thymocytes (Fig. 3,a) express ART2.2. In this mouse strain, ART2.2 is expressed by 93% of peripheral CD8+ and by 66% of peripheral CD4+ T cells (Fig. 3,e). Note that use of Cy3-conjugated anti-CD4 together with PE-conjugated anti-CD8 vs FITC-conjugated Nika102 allows direct comparison of ART2.2 expression levels on CD4+ and CD8+ peripheral T cells in the same plots (Fig. 3, d and e). Note also that ART2.2 is expressed by few if any CD3− cells (Fig. 3, a–c), and that expression of ART2.2 and CD25 is reciprocal (Table I).
|Activation .||None .||Anti-CD3 .|
|% ART 2.2+ CD25−||25.5||2.1|
|% ART 2.2− CD25+||5.2||80.4|
|% ART 2.2− CD25−||68.0||15.3|
|% ART 2.2+ CD25+||0.6||2.3|
|Activation .||None .||Anti-CD3 .|
|% ART 2.2+ CD25−||25.5||2.1|
|% ART 2.2− CD25+||5.2||80.4|
|% ART 2.2− CD25−||68.0||15.3|
|% ART 2.2+ CD25+||0.6||2.3|
C57BL/6J splenocytes were cultured with or without 2 μg/ml anti-CD3 Ab for 20 h at 37°C and then stained for expression of ART2.2 and CD25. Numbers indicate the percentage of totals cells expressing ART2.2 and CD25 as indicated.
ART2.2 is expressed by mature thymocytes
Considering that rat thymocytes are negative for RT6 and, furthermore, that peripheral T cells express RT6 soon after emigrating from the thymus (22, 38), we suspected that the small fraction of ART2.2+ thymocytes might represent mature cells and set up a series of experiments to test this. First, we performed triple fluorescent staining analyses of C57BL/6J mouse thymocytes (Fig. 4 a). The results reveal that neither CD4−/CD8− double-negative nor the immature CD4+/CD8+ double-positive cells contain any sizable fractions of ART2.2-expressing cells (gates 1 and 2). In contrast, subpopulations of mature CD4+ and CD8+ single-positive cells in the thymus do express ART2.2, with much larger percentages of CD8 single-positive than CD4 single-positive thymocytes expressing ART2.2 (gate 4 vs gate 3).
Second, we tested the sensitivity of ART2.2+ and ART2.2− thymocytes to cortisone treatment, which is known to deplete immature thymocytes. Indeed, dexamethasone treatment almost completely depletes the population of ART2.2−/CD4+/CD8+ cells (Fig. 4 b, gate 2), while the ART2.2+ cells in the subpopulations of mature CD4+ and CD8+ single-positive thymocytes (gates 3 and 4) are largely dexamethasone resistant.
Third, we examined ART2.2 expression on thymocytes of the CTS/Shi mouse, in which a genetic defect prevents the emigration of mature thymocytes (39), so that mature cells accumulate in the thymus. The results show a very high percentage of ART2.2+ cells among the CD4+ and CD8+ single-positive but not in the CD4/CD8 double-negative and double-positive thymocyte subpopulations (Fig. 4 c), consistent with an accumulation of mature (ART2.2+) thymocytes in this mouse strain.
Finally, we compared expression of ART2.2 with expression of other cell surface markers known to be either up (CD3, CD5)- or down-regulated (CD24, CD69, Ly-6C) during the final stages of thymocyte maturation. Fig. 5 shows double stainings of C57BL/6J thymocytes for ART2.2 and a series of other cell surface markers. The results show that ART2.2+ cells are significantly enriched in the following subpopulations: CD3high (Fig. 5,a), CD5high (Fig. 5,b), CD69low (Fig. c), CD24low (Fig. d), and Ly-6Clow (Fig. 5,e). Note, in particular, that almost all of the CD24low cells, which include the most mature thymocytes (40), are ART2.2+ (Fig. 5 d). Cumulatively, these results support the interpretation that ART2.2 is a marker for the most mature population of thymocytes.
ART2.2 is expressed by the majority of γδT cells, but only by few if any NK cells
Next, we set out to determine whether ART2.2, like RT6 in the rat (41), is expressed also on γδT cells and NK cells. To this end, we performed dual staining analyses of ART2.2 vs γδT cell and NK cell markers (the γδTCR and NK1.1, respectively) on splenocytes of C57BL/6.Tcrbtm1Mom (denoted Tcrb−/−), C57BL/6J-Prkdcscid (denoted C57BL/6.scid), and C57BLKS/J mice (Fig. 6). Tcrb−/− mice lack TCR-α/β T cells, but still contain TCR-γ/δ T cells and B cells (42). As shown in Fig. 6,e, about one-half of the TCR-γ/δ splenocytes in these animals express ART2.2. C57BL/6.scid mice lack mature T and B cells and are known to contain a high proportion of NK1.1+ splenocytes (43). These NK1.1+ splenocytes evidently do not express ART2.2 (Fig. 6,i). However, a small fraction of NK1.1+ cells in the Tcrb−/− mice appears to express ART2.2 (Fig. 6 h).
The number of ART2.2-expressing cells increases during postnatal ontogeny
To assess whether ART2.2, like RT6 in the rat (22, 44), is developmentally regulated during postnatal ontogeny, expression of ART2.2 by thymic and splenic lymphocytes obtained from C57BL/6J mice of different ages was assessed by flow cytometry. As shown in Fig. 7, the proportion of ART2.2-expressing CD4+ and CD8+ cells increases steadily during the early postnatal ontogeny. ART2.2 expression levels reach a maximum at 6–8 wk of age and then fall off again.
Inbred strains of mice show marked differences in ART2.2 expression levels
To determine whether ART2.2, like RT6 in the rat (22, 44), is expressed in different extents by different inbred strains, we compared different strains of mice for ART2.2 expression by flow cytometry (Fig. 8). Strains having high proportions of ART2.2-expressing splenocytes include C57BL/6J, C57BLKS/J, NOR/Lt, and NOD/Lt. In contrast, DBA2/J, C3H/HejBir, CBA/J, NZO/Lt, and (NZB × W)F1/J mice have low proportions of ART2.2-expressing splenocytes. In general, the mean fluorescent intensity of ART2.2-expressing cells was relatively low in mice with low percentages of ART2.2-expressing splenocytes. ART2.2 is not detectable in NZW mice in which the ART2.2 gene is deleted (45). ART2.2 is also not detectable in T cell-deficient C57BL/6.scid mice (see also Fig. 6).
Expression of ART2.2 is sensitive to treatment of cells with PI-PLC
The deduced amino acid sequence of ART2.2 contains a hydrophobic C-terminal signal sequence characteristic of GPI-anchored proteins (13). To assess whether the predicted GPI anchor of native ART2.2 on mouse T cells, like that of rART2.2 (Fig. 1), is accessible to PI-PLC, spleen cells were incubated with PI-PLC for 45 min at 37°C before flow cytometry. As shown in Fig. 9, staining for ART2.2 is abolished by treatment of cells with PI-PLC (Fig. 9,b). In contrast, staining for the type I membrane proteins CD4 and CD8 is resistant to treatment with PI-PLC (Fig. 9 b).
ART2.2 expression is down-modulated upon activation of T cells
Considering that RT6 in the rat is down-modulated upon T cell activation (37) and that cell surface expression of ART2.2 and the IL-2R (CD25) is inversely correlated on unstimulated splenocytes (see above, Table I), we analyzed the expression of ART2.2 upon activation of T cells with anti-CD3 Abs or with direct activation of protein kinase C with PMA. As shown in Table I, CD25 expression on splenic T cells is up-regulated following 20 h of anti-CD3 stimulation, whereas this activation elicits a reciprocal down-regulation in ART2.2 expression.
It has previously been shown that an ADP-ribosyltransferase activity is released from mouse T cells upon treatment with the protein kinase C activator PMA (46). To assess whether ART2.2 is released from the cell surface by T cell activation, splenocytes were first stained with fluorescently labeled Abs and then activated by the addition of phorbol ester (100 ng/ml PMA) and Ca2+-ionophore (200 ng/ml A23187) (Fig. 9, c and d). Staining for ART2.2 is reduced more than 10-fold within 60 min after PMA activation (Fig. 9,d). In contrast, staining for the type I membrane proteins CD4 and CD8 is resistant to treatment with PMA under these conditions (Fig. 9,d). In a second experiment, splenocytes were first incubated for 90 min in the presence or absence of PMA and A23187 and then stained for surface expression of ART2.2 and other cell surface proteins (Fig. 9, e–l). Again, ART2.2 expression is reduced more than 10-fold in PMA treated vs control cells (Fig. 9, f vs e). In contrast to GPI-anchored ART2.2, neither the GPI-anchored membrane proteins CD90 (Thy-1) (Fig. 9, g and h) and Ly-6A/E (Fig. 9, i and j) nor the type I membrane proteins CD3 (data not shown) and CD8 (Fig. 9, e–l) are affected by treatment of cells with PMA/A23187. Expression of the type I membrane protein CD62L, which is known to be cleaved by a metalloprotease under these conditions (47, 48), is reduced more than 10-fold (Fig. 9, l vs k). Staining for CD4, however, which is known to be endocytosed under these conditions (49), is decreased ∼2-fold after 2 h of PMA/A23187-induced cell activation (Fig. 9, f, h, j, and l vs e, g, i, and k). Note that when cells are prestained with Ab before T cell activation (panels c and d), CD4 down-modulation is no longer visible, whereas down-modulation of ART2.2 (Fig. 9 d) and CD62L (data not shown) proceeds unabated.
We have used gene gun technology to immunize a Wistar rat with an expression vector encoding mouse ART2.2, the Art2b gene product. We then used hybridoma fusion technology to raise an ART2.2-specific mAb, designated Nika 102, from splenocytes of this rat. We initially determined the specificity of Nika102 by its reactivity with rART2.2-AP fusion protein. Flow-cytometric analyses demonstrated that Nika102 also reacts with murine EL4 lymphoma cells transfected with Art2b cDNA, but not with untransfected EL4 cells (Fig. 1,c). Moreover, Nika102 specifically immunoprecipitates a 45-kDa protein from ART2.2-transfected EL4 cells that is not present in mock-transfected cells (Fig. 2, lane 10 vs lane 6). Nika102 also specifically immunoprecipitates a 30-kDa protein from periplasma lysates of ART2.2-transformed E. coli cells, but not from control E. coli cells (Fig. 2, lane 4). The proteins precipitated by Nika102 are recognized in Western immunoblot analyses by K48, a rabbit antiserum produced against a peptide deduced from the sequence of mouse Art2b cDNA (13). These results demonstrate that we have identified a mAb specific for the product of the mouse Art2b gene.
Based on the Art2b cDNA sequence and the ART2.2 amino acid sequence deduced from it, we had previously predicted that ART2.2, like its rat RT6 homologue (36), is expressed as a cell surface protein anchored within the cell membrane via a GPI anchor (13). We now corroborate this prediction by visualizing ART2.2 on the T cell surface by FACS analyses with Nika102 and by demonstrating that ART2.2 disappears from the cell surface after treatment of cells with PI-PLC (Figs. 1 a and 9b).
The apparent m.w. of ART2.2 from E. coli (Fig. 2, lane 4) approximates the m.w. of native polypeptide deduced from the cDNA sequence (29.3 kDa). The higher apparent m.w. of ART2.2 from EL4 cells (Fig. 2, lane 10) most likely reflects posttranslational modification of ART2.2 in these cells. It is possible, for example, that EL4 cells glycosylate ART2.2 at one or both of the potential N-linked glycosylation sites deduced from the Art2b cDNA sequence (13). The fact that Nika102 immunoprecipitates both unglycosylated ART2.2 from E. coli and posttranslationally modified ART2.2 from EL4 cells, indicates that ART2.2 recognizes an epitope on the peptide backbone of ART2.2 and, furthermore, that this epitope is not blocked by posttranslational modification(s) in EL4 cells.
The distribution of ART2.2 on mouse T cells shows many similarities to that of its structural orthologue in the rat. Like rat RT6 (22, 44), mouse ART2.2 is expressed by peripheral CD4+ and CD8+ T cells, but not by CD3− cells in the adult mouse (Fig. 3). Similar to RT6 in the rat, expression of ART2.2 by T cells is low in the newborn mouse and increases during postnatal ontogeny to reach a peak at 6–8 wk of age (Fig. 7). Like immature rat thymocytes that are RT6− (22, 44, 50, 51), immature mouse CD4+/CD8+ double-positive thymocytes are ART2.2− (Fig. 4, a–c, gate 2). One notable difference between rat and mouse regards expression of ART2.2 by mature thymocytes (Fig. 4, a–c, gates 3 and 4, and Fig. 5). Mature rat thymocytes also are RT6− (22, 44, 50, 51). In contrast, the mouse thymus contains a small but distinctly ART2.2+ population of CD3highCD4+/CD8− and CD3highCD4−/CD8+ cells (Figs. 4 and 5).
The expression of ART2.2 in the thymus is of particular interest because it appears to identify primarily the most mature thymocytes: i.e., Nika102 reacts with <1% of CD4+/CD8+ double-positive cells, but stains 15% of CD4+ and 58% of CD8+ single-positive cells (Fig. 4,a). ART2.2+ cells expressed high levels of CD3 (Figs. 3,a and 5a) and CD5 (Fig. 5,b), but low levels of CD69 (Fig. 5,c), which is transiently expressed during positive selection (52), and low levels of CD24 (Fig. 5d), which is down-regulated during the final stages of thymocyte maturation (40). Moreover, ART2.2+ thymocytes are resistant to dexamethasone-induced apoptosis (Fig. 4,b). Cumulatively, these data are consistent with ART2.2 being expressed late in thymic development after the most intense phase of positive selection. This interpretation is also supported by the finding that ART2.2+ cells are enriched in the thymi of CTS/Shi mice (Fig. 4 c) in which a genetic defect prevents the emigration of mature cells from the thymus (39). Thus, ART2.2 appears to be a useful marker for the most mature thymocytes.
RT6-expressing T cells exert a regulatory role in the BB rat model for autoimmune insulin-dependent diabetes mellitus (23, 24). Moreover, it has been shown that T cells from diabetic BB rats can transfer disease to nondiabetic rats only if the cells are stimulated before the transfer with phorbol ester or mitogen (53, 54). In this context, it may be of interest to point out that ART2.2 and the IL-2R (CD25) are reciprocally expressed (Table I). The results indicate that ART2.2 is down-modulated from the cell surface and/or that ART2.2-negative cells preferentially expand upon anti-CD3-mediated T cell activation. It will be of interest to analyze the potential immunoregulatory role of ART2.2-expressing cells and of ART2.2 itself in murine models of autoimmune disease. The results presented in this study indicate that Nika102 holds promise as a new experimental tool to address these questions.
Using CTLs induced in a 10-day mixed lymphocyte reaction, Dennert and coworkers have previously observed that CTL functions such as proliferation, target cell binding, and cytotoxicity can be down-regulated by ecto-NAD+ (10, 11, 12), and further, that this is due to the action of a GPI-anchored cell surface ADP-ribosyltransferase (8, 9). These authors also noted that ADP-ribosyltransferase activity is released from the cell surface upon stimulation of CTLs with PMA (46). In accordance with these findings is our observation that mouse ART2.2 disappears from the surface of lymph node and spleen cells upon stimulation of these cells with PMA and ionophore (Fig. 9, c–f) or anti-CD3 (Table I).
Disappearance of an Ag from the cell surface can, in principle, be caused by shedding or endocytosis of the Ag. Precedences for both types of mechanisms have been reported for other GPI-anchored cell surface proteins (55, 56). Our results are more compatible with shedding rather than with endocytosis of ART2.2. PMA-induced disappearance of ART2.2 from the cell surface more closely resembles that of CD62L than that of CD4 in both magnitude and kinetics. CD62L is known to be shed from the cell surface, whereas CD4 is known to be endocytosed under these conditions (47, 49). CD4 staining levels remain unaltered when cells are prestained with Abs before stimulation with PMA, as expected for an Ag that is endocytosed together with its bound fluorescently labeled Ab (Fig. 9, c and d). In contrast, like CD62L (not shown), disappearance of ART2.2 from the cell surface can still be observed under these conditions. Whatever the mechanism of PMA-induced loss of ART2.2 from the cell surface may be, it is of interest to note that expression levels of other GPI-anchored cell surface proteins such as CD90 and Ly-6A/E are not similarly affected (Fig. 9, g–h). The finding that ART2.2 disappears from the cell surface of PMA-stimulated T cells raises interesting questions to be addressed in future investigations, such as: 1) how is ART2.2 released from the cell surface, e.g., by a phospholipase acting on the GPI anchor or by a protease acting on the polypeptide backbone; 2) how is T cell function affected by shedding of ART2.2; and 3) is ART2.2 released in an enzymically active form?
Interestingly, we have observed significant differences in the percentage of peripheral T cells expressing ART2.2 as a function of inbred strain (Fig. 8). T cells from NZW/BinJ mice, unable to express ART2.2 because the Art2b gene is deleted in this strain (45), did not react with Nika102, nor did splenocytes from T cell-deficient C57BL/6J.scid mice (Figs. 5 and 8). High ART2.2-expressing strains include C57BL/6J, C57BLKS/J, NOR/Lt, and NOD/Lt; low expressors include DBA/2J, C3H/HeJBir, CBA/JLt, and NZO/Lt. In general, ART2.2 cell surface expression levels as detected by Nika102 correlate well with results of previous studies showing marked differences in ART2.2-specific mRNA, as detected by RT-PCR in different strains of mice (13, 57, 58). The strain-specific differences in ART2.2 expression may also be of relevance for graft rejection across the H1 minor histocompatibility barrier. The ART2.2 gene maps near the H1 minor histoincompatibility locus and ART2 haplotypes correlate with H1 allotypes (59). We note that H1a mice (DBA/2J, C3H/HeJBir, CBA/JLt) all are low ART2.2 expressors, whereas H1c mice (C57BL/6J, C57BLKS/J) are high ART2.2 expressors. It is conceivable that differences in ART2.2 expression levels contribute to graft rejection across the H1 barrier. It remains to be determined whether these strain-specific differences in ART2.2 expression on T cells are compensated by expression of other ecto-ADP-ribosyltransferases on these cells, e.g., ART1 (15, 16, 60) or ART2.1 (13, 28, 29). In case of C57BL/6J and BXSB/MpJ mice, in which the closely linked ART2.1 gene is inactivated by a premature stop codon (57, 58), it is possible that ART2.2 is up-regulated to high expression levels in compensation for loss of ART2.1.
Abs to the other ecto-ARTs are necessary to complete the picture. The strategy described in this study for raising mAb Nika102 against ART2.2 by ballistic DNA immunization may be an attractive alternative to traditional protein immunization strategies for raising mAbs to other ARTs. Indeed, reports from other laboratories (61, 62, 63) lend support to the idea that DNA immunization is generally applicable for raising mAbs against molecularly cloned cell surface proteins.
In conclusion, we have demonstrated that gene gun technology is an efficient means to raise mAbs against a T cell surface protein. We have used one such mAb, Nika102, to show that mouse T cells express ecto-ART ART2.2 as a GPI-anchored cell surface molecule. In the thymus, ART2.2 is not expressed by immature CD4+/CD8+ double-positive cells, but is expressed by mature CD4+ and CD8+ single-positive cells. In peripheral lymphoid organs, ART2.2 is expressed at high levels on CD8+ cells, and at slightly lower levels on CD4+ cells. ART2.2 expression levels are low in the newborn mouse, but increase during the early postnatal ontogeny. ART2.2 is rapidly down-modulated from the cell surface after PMA-induced T cell activation. Thus, the ART2.2-specific mAb Nika102 described in this study should be of use as new differentiation marker of thymic and postthymic T cells in the mouse and as a tool to further elucidate the function of this intriguing cell surface enzyme in the mouse.
We gratefully acknowledge the assistance provided by Bruce Regimbal (The Jackson Laboratory) for animal care, and by Drs. Jens Dimigen and Franz Iglauer (Central Animal Facility, University Hospital Hamburg) for gene gun immunizations. We further thank Christiane Beig, Roman Girisch, Inga Heinsohn, Nicolle Schröder, Vivienne Welge, Hamburg; Steve Langley (The Jackson Laboratory); and Jason Dietrich (University of California, San Francisco), for expert technical assistance. F.K.-N. and M.N. thank the staff of E.H.L.’s lab for the hospitality during their stay at The Jackson Laboratory as visiting investigators. Parts of the work described in this study represent the partial fulfillment of the requirements for the graduate thesis of S.K. at the Faculty of Medicine, Hamburg. F.K.-N., N.K., F.H., and E.H.L. supervised this study. F.K.-N. and M.N. did the fusion and raised Nika102. S.K. performed the experiments described in Figs. 1 and 2; T.D., M.N., and F.K.-N. those described in Figs. 3, 4, and 6–9; N.K. that in Fig. 5; and V.A. that in Table I. We thank Drs. David Serreze, Robert Graser, and Heinz-Günter Thiele for critical reading of the manuscript.
This work was supported by Deutsche Forschungsgemeinschaft Grants 310 (to F.K.-N.) and SFB 545/B9 to (F.K.-N. and F.H.), and by National Institutes of Health Grants 27722 and 36175 to (E.H.L.). Institutional Shared Services at The Jackson Laboratory were supported by National Cancer Institute Center Support Grant CA-34196.
Abbreviations used in this paper: ART, ADP-ribosyltransferase; AP, alkaline phosphatase; HF, chimeric His(6x)-FLAG tag; PI-PLC, phosphatidylinositol-specific phospholipase C; NZW, New Zealand White.