T-cell immunoglobulin and mucin domain (TIM)-2 is expressed on activated B cells. Here, we provide evidence that murine TIM-2 is a target of ADAM10-mediated ectodomain shedding, resulting in the generation of a soluble form of TIM-2. We identified ADAM10 but not ADAM17 as the major sheddase of TIM-2, as shown by pharmacological ADAM10 inhibition and with ADAM10-deficient and ADAM17-deficient murine embryonic fibroblasts. Ionomycin-induced or 2′(3′)-O-(4-benzoylbenzoyl) ATP triethylammonium salt-induced shedding of TIM-2 was abrogated by deletion of 10 juxtamembrane amino acids from the stalk region but not by deletion of two further N-terminally located blocks of 10 amino acids, indicating a membrane-proximal cleavage site. TIM-2 lacking the intracellular domain was cleaved after ionomycin or 2′ (3′)-O-(4-benzoylbenzoyl) ATP triethylammonium salt treatment, indicating that this domain was not involved in the regulation of ectodomain shedding. Moreover, TIM-2 shedding was negatively controlled by calmodulin. Shed and soluble TIM-2 interacted with H-ferritin. In summary, we describe TIM-2 as a novel target for ADAM10-mediated ectodomain shedding, and reveal the involvement of ADAM proteases in cellular iron homeostasis.
The involvement of the T-cell immunoglobulin and mucin domain (TIM) gene family in immunity and disease was initially detected with a position-cloning approach using a congenic model of asthma. TIM genomic regions were linked to multiple disease states, including allergy, asthma, transplant tolerance, autoimmunity, response to viral infections, and cancer . There are eight TIM genes in mice, but only three in humans. The murine genes encoding TIM-1, TIM-3 and TIM-4 are orthologs of the human genes encoding TIM-1, TIM-3, and TIM-4 . The murine genes encoding TIM-1 and TIM-2 are highly homologous, and may have originated from gene duplication in the rodent lineage, as only the gene encoding TIM-1 is found in primates .
Like all other TIM proteins, TIM-2 is a type I transmembrane glycoprotein with an N-terminal immunoglobulin-like domain (IgV domain), a mucin domain with O-linked and N-linked glycosylations, a single transmembrane domain, and a cytoplasmic region with a tyrosine phosphorylation motif . TIM-2 is expressed on late differentiated Th2 cells  and on activated B cells .
TIM-2 has been shown to be the receptor of H-ferritin and to participate in iron homeostasis in the liver, the kidney [5, 6], and oligodendrocytes . Moreover, it has been demonstrated that TIM-2, when transfected into mouse kidney cells, leads to ferritin uptake into the transfected cells and results in increased levels of intracellular ferritin .
Studies with TIM-2-deficient mice defined TIM-2 as a negative regulator of Th2-type responses [9, 10]. On B cells, TIM-2 signaling influences both proliferation and antibody production of B cells during the early phase of Th1-mediated collagen-induced arthritis . A recombinant soluble TIM–2Fc fusion protein induced T-cell hyperproliferation and enhanced the production of Th2 cytokines in vivo . A subsequent study also showed that treatment with soluble TIM–2Fc exacerbated lung inflammation in the ovalbumin-induced asthma model . These results indicated that a naturally occurring soluble TIM-2 (sTIM-2) ectodomain could also have biological functions.
Soluble protein ectodomains are generated by limited proteolysis of transmembrane proteins, also called ectodomain shedding, resulting in downregulation of protein cell surface expression. In some cases, exemplified by Notch, ectodomain shedding leads to a subsequent cleavage event within the membrane by γ-secretase. The result is the generation of an intracellular fragment, which induces intracellular signaling . Members of the ADAM gene family encode ectodomain shedding proteinases. Among them, ADAM10 and ADAM17 are the major sheddases, with, together, > 100 substrates . There is, however, extensive overlap and compensation between ADAM proteases for several substrates [13, 14]. Different stimuli, including phorbol ester [phorbol-12-myristate-13-acetate (PMA)], ionomycin (IM), ATP [2′ (3′)-O-(4-benzoylbenzoyl) ATP triethylammonium salt (BzATP)], ligands of G-protein-coupled receptors, bacterial toxins, bacterial metalloproteinases, and apoptosis, activate ADAM10-mediated and/or ADAM17-mediated shedding of transmembrane proteins .
Here, we provide evidence that TIM-2 is subject to ectodomain shedding mediated by the metalloprotease ADAM10, leading to rapid downregulation of TIM-2 cell surface expression. sTIM-2 generated by shedding binds H-ferritin, and addition of H-ferritin to B cells increases the proliferation of these cells. Calmodulin binding to TIM-2 negatively regulates shedding of TIM-2. As TIM-2 shedding is observed in primary murine B cells, a role of this process in B-cell biology and iron homeostasis is discussed.
BzATP-induced shedding of endogenous TIM-2 from primary murine B cells
B cells have been shown to express TIM-2 after activation with IgM and mAb against CD40 [4, 6]. We analyzed shedding of TIM-2 from activated primary murine B cells. First, we verified, by flow cytometry, expression of TIM-2 on B cells induced by IgM and mAb against CD40 (Fig. 1A). Cellular expression of TIM-2 was clearly visible after 48 h and 72 h of stimulation with IgM and mAb against CD40. To determine whether primary B cells could be stimulated to shed endogenous, membrane-bound TIM-2 from the cell surface, we treated B cells stimulated with IgM and mAb against CD40 for 30 min with BzATP. As can be seen in Fig. 1B, cell surface expression of TIM-2 was reduced to control levels upon treatment of the cells with BzATP. The loss of cell surface TIM-2 was prevented by cotreatment with the specific ADAM10 inhibitor GI254023X (GI). These results suggested that ADAM10-mediated shedding of TIM-2 can occur in primary B cells.
ADAM10-mediated shedding of murine TIM-2 (TIM-2) after stimulation of cells with BzATP and IM
TIM-2 cDNA was synthesized from mRNA of murine thymus. Here, we generated a TIM-2 expression construct that consists of a signal peptide, an N-terminal FLAG-tag, the TIM-2 protein with extracellular, transmembrane and intracellular domains, and a C-terminally located His-tag (Fig. 1C). HEK293 cells were stably transfected with the expression plasmid coding for TIM-2, and cell surface expression was verified by flow cytometry with a mAb against the FLAG-tag (Fig. 1D). We next investigated whether TIM-2 was cleaved from the cell surface. HEK293 cells expressing TIM-2 were stimulated with PMA, which has been shown to stimulate shedding of numerous cell surface proteins . As can be seen in Fig. 1E, TIM-2 was detected in the cell lysates but not in the supernatant. PMA is thought to activate ADAM17 but not ADAM10 . These results indicated that ADAM17 does not cleave TIM-2.
BzATP and IM are considered to be inducing stimuli for ADAM10 . We stimulated HEK293 cells transiently transfected with the expression plasmid coding for TIM-2 with BzATP or IM in the absence or presence of GI. sTIM-2 released into the supernatant was visualized by western blotting. BzATP and IM treatment of TIM-2-expressing cells led to the release of sTIM-2 into the cell culture supernatant (Fig. 1F,G). GI  completely inhibited the BzATP-induced and IM-induced release of TIM-2 into the cell culture supernatant (Fig. 1F,G). These results indicated that TIM-2 was released from the cell surface by ADAM10. To corroborate this finding, we stimulated HEK293 cells stably transfected with a TIM-2 cDNA with IM in the presence or absence of the recombinant prodomain of ADAM10 (A10pro) . This protein has been shown to specifically block ADAM10 [13, 17]. As expected, A10pro completely inhibited the IM-induced TIM-2 shedding (Fig. 1H). We conclude from these experiments that TIM-2 is cleaved by ADAM10 but not by ADAM17, and that shedding leads to the appearance of a soluble form of TIM-2 in the cell supernatant.
Establishment of a quantitative shedding assay confirmed ADAM10 as an TIM-2 sheddase
To facilitate quantification of TIM-2 shedding, we generated a fusion protein of TIM-2 and alkaline phosphatase (AP), which has been used to detect and quantify shedding reactions . We constructed a eukaryotic expression plasmid coding for a TIM-2 fusion protein with AP, which we named AP–TIM-2. The fusion protein consisted of a signal peptide, AP, a FLAG-tag, TIM-2 with extracellular, transmembrane and intracellular domains, and a C-terminally located His-tag (Fig. 2A). The AP–TIM-2 plasmid was stably transfected into HEK293 cells, and cell surface expression of the AP–TIM-2 fusion protein was verified by flow cytometry (Fig. 2B). As shown for TIM-2, AP-tagged TIM-2 was shed after stimulation of transfected HEK293 cells with BzATP and IM, but not after PMA stimulation, as shown by western blotting and AP assays (Fig. 2C,D), confirming that TIM-2 shedding was mediated by ADAM10 but not ADAM17. This notion was confirmed by the fact that shedding of the AP–TIM-2 fusion protein was inhibited by GI and the combined ADAM10/ADAM17 inhibitor GW280264X (GW) (Fig. 2C,D). As can also be seen in Fig. 2C,D, the results from western blotting and from the measurements of AP activity matched, indicating that the experiments with the AP–TIM-2 fusion protein can be used to quantify shedding of TIM-2.
To independently verify PMA-induced ADAM17 activation, we transiently transfected HEK293 cells with a cDNA coding for AP-tagged transforming growth factor-α (TGF-α), another well-known ADAM17 substrate, and stimulated these cells with PMA. As shown by the AP assay in Fig. 2E, PMA stimulated ADAM17-mediated shedding of TGF-α in HEK293 cells.
HEK293 cells stably transfected with the AP–TIM-2 fusion protein were stimulated with IM in the presence or absence of A10pro [13, 17]. A10pro completely inhibited the IM-induced AP–TIM-2 shedding (Fig. 2F). Finally, we employed murine embryonic fibroblasts (MEFs) deficient for ADAM10, for ADAM17, and for both ADAM10 and ADAM17. These cells were transiently transfected with cDNAs coding for the AP–TIM-2 fusion protein. Without stimulation, constitutive shedding was quantified for 24 h, and the conditioned supernatant was analyzed by AP assay. Here, wild-type and ADAM17-deficient MEFs, but not ADAM10-deficient and ADAM10/ADAM17-deficient MEFs, showed constitutive shedding of TIM-2, suggesting that constitutive shedding of TIM-2 is mainly mediated by ADAM10 (Fig. 2G). As shown in Fig. 2H, shedding of AP–TIM-2 was induced after IM treatment in wild-type and ADAM17-deficient MEFs. In contrast, IM-induced shedding of TIM-2 was almost completely inhibited in ADAM10-deficient and ADAM10/ADAM17-deficient MEFs (Fig. 2H). Taken together, these results clearly indicated that the observed shedding of TIM-2 was a result of ADAM10, but not of ADAM17, activation.
IM-induced shedding of TIM-2 was abrogated by deletion of 10 amino acids from the stalk region of TIM-2
ADAM10 has a relaxed sequence specificity, favoring alanine or threonine at the P1 position, and valine, glutamine or leucine at the P1′ position . ADAM protease cleavage sites are mostly located in close juxtaposition to the plasma membrane . Furthermore, it has been shown that a small 5–10-residue deletion around the cleavage site but not deletions further upstream of the cleavage site lead to abrogation of shedding mediated by ADAM proteases . It was shown that shedding of the type I membrane protein interleukin-6 receptor (IL-6R), which can be cleaved by ADAM10 and ADAM17, is executed by cleavage between Gln357 and Asp358, approximately eight amino acids upstream of the transmembrane region. Deleting 10 amino acids surrounding the cleavage site (Ser353–Gln357) completely inhibited PMA-induced shedding of IL-6R by ADAM17 .
It is not known whether the cleavage site of ADAM10 in IL-6R is identical to the cleavage site used by ADAM17; however, recent data have suggested different cleavage sites . This illustrates why it is difficult to predict the cleavage site of the novel ADAM10 substrate TIM-2 by analysis of the protein sequence. Therefore, we generated three different variants of AP–TIM-2, each lacking 10 amino acids upstream of the transmembrane domain (Δ1, lacking amino acids 221–230; Δ2, lacking amino acids 211–220; and Δ3, lacking amino acids 201–210) (Fig. 3A). As shown in Fig. 3B, all three TIM-2 mutant proteins were expressed on the cell surface of transiently transfected HEK293 cells. IM-induced shedding was completely abrogated for AP–TIM-2_Δ1, reduced for AP–TIM-2_Δ2, and not affected for AP–TIM-2_Δ3, as shown by western blotting and AP assays (Fig. 3C,D). All proteins were expressed to a similar extent, as judged by flow cytometry and western blot analysis (Fig. 3B,C). Importantly, ADAM10-mediated shedding of AP–TIM-2_Δ1, AP–TIM-2_Δ2 and AP–TIM-2_Δ3 was similar for both stimuli used (IM and BzATP), and none of the mutants became a substrate for PMA-induced ADAM17-mediated shedding (Fig. 3E–H). The protease specificities for the different TIM-2 mutants were not different from that for full-length TIM-2, as demonstrated for IM-induced shedding in wild-type, ADAM10-deficient, ADAM17-deficient and ADAM10/ADAM17-deficient MEFs (Fig. 3I). We conclude from our experiments that the cleavage site for IM-induced shedding of TIM-2 mediated by ADAM10 is located close to the transmembrane domain within the region from Pro221 to Lys230 (Fig. 3A).
The intracellular domain (ICD) and the IgV domain of TIM-2 are not important for BzATP and IM-induced shedding
To characterize whether the IgV domain of TIM-2 or the ICD influences ADAM10-mediated shedding of TIM-2, we generated two additional TIM-2 variants, lacking either the IgV domain or the ICD (Fig. 4A). HEK293 cells were transiently transfected with an AP–TIM-2 cDNA lacking the coding sequence of the IgV domain (AP–TIM-2_ΔIgV) or the ICD (AP–TIM-2_ΔICD). As shown in Fig. 4B, both deletion mutants of TIM-2 were expressed on the cell surface of transfected HEK293 cells. It is of note that, in both flow cytometry and western blot analysis, higher expression was seen for the expressed AP–TIM-2_ΔIgV protein than for the AP–TIM-2_ΔICD protein. Like wild-type TIM-2, both mutant proteins were shed efficiently after stimulation with IM and BzATP, but not after stimulation with PMA, as shown by western blotting (Fig. 4C–E) and AP assays (Fig. 4F–H). As shown in Fig. 4I, shedding of both TIM-2 deletion mutants, AP–TIM-2_ΔIgV and AP–TIM-2_ΔICD, was induced by IM in wild-type and ADAM17-deficient MEFs, whereas no shedding was seen in ADAM10-deficient and ADAM10/ADAM17-deficient MEFs. These results show that neither the IgV domain nor the ICD of TIM-2 plays a role in the shedding of TIM-2 mediated by ADAM10.
Binding of H-ferritin to sTIM-2
As TIM-2 is a cell surface receptor for H-ferritin [5-8], we next investigated whether sTIM-2 could bind to recombinant H-ferritin. As shown in Fig. 5A, we expressed sTIM-2 as a C-terminal Fc-fusion protein (sTIM–2Fc) in HEK293 cells. Recombinant PC-tagged H-ferritin (H-ferritin–PC) was expressed in and purified from Escherichia coli. As H-ferritin has been shown to be extremely heat-stable , we removed most bacterial proteins by a temperature shift to 70 °C (Fig. 5B). As shown in Fig. 5D, precipitation of sTIM-2–Fc by protein G–Sepharose in the presence of H-ferritin–PC led to the coprecipitation of H-ferritin–PC as visualized by western blotting. Analysis of the supernatant of the protein G–Sepharose pellet contained no measurable sTIM–2Fc or H-ferritin–PC, indicating that all TIM-2–Fc protein had been precipitated, and that all H-ferritin–PC had bound to sTIM–2Fc. No coprecipitation of H-ferritin–PC was seen when an unrelated Fc-fusion protein, sgp130Fc , was used.
As an alternative approach, we established an ELISA to measure the binding of sTIM–2Fc to H-ferritin. H-ferritin was coated on a plate, and incubated with different dilutions of sTIM–2Fc (expressed in HEK293 cells). To quantify bound sTIM–2Fc, a peroxidase (POD)-coupled antibody against humanFc was used. As shown in Fig. 5E, with lower concentrations of sTIM–2Fc, less sTIM–2Fc was able to bind to plate-bound H-ferritin. When we used sgp130Fc or NaCl/Pi as controls, no binding to H-ferritin was detected.
To demonstrate that full-length TIM-2 and shed TIM-2 (after IM stimulation) were also able to bind to H-ferritin, we developed an AP-ELISA assay (Fig. 5F). An H-ferritin-coated plate was incubated with lysate or supernatant from HEK293 cells transiently transfected with the AP–TIM-2 cDNA. Detection of AP activity was used to measure the bound full-length AP–TIM-2 in the lysate or the bound shed AP–TIM-2 in the supernatant. After IM stimulation, the amount of bound shed AP–TIM-2 was increased in comparison with constitutively shed AP–TIM-2. Treatment with the specific inhibitors GI (blocking ADAM10) and GW (blocking ADAM10 and ADAM17) led to lower H-ferritin binding. Lysate and supernatant from enhanced green fluorescent protein (eGFP)-transfected HEK293 cells stimulated with IM showed no binding to H-ferritin.
A competitive AP-ELISA assay was used to verify the binding of shed AP–TIM-2 to H-ferritin after IM stimulation. As shown in Fig. 5G, with a reduction in the sTIM–2Fc concentration, more shed AP–TIM-2 was able to bind to coated H-ferritin. These results confirm that TIM-2 is a receptor for H-ferritin, and also show that sTIM-2, after IM-induced shedding, can interact with its ligand H-ferritin.
We also analyzed the binding of the AP–TIM-2 deletion variants (∆IgV, ∆ICD, ∆1, ∆2, and ∆3) to H-ferritin. As shown in Fig. 5H, AP–TIM-2, AP–TIM-2_∆ICD, AP–TIM-2_∆1, AP–TIM-2_∆2 and AP–TIM-2_∆3 (HEK293 cell lysate) bound to plate-coated H-ferritin. AP–TIM-2_∆IgV did not bind to H-ferritin, indicating that the immunoglobulin domain of TIM-2 mediated the binding to H-ferritin.
The calmodulin inhibitor N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide (W7) causes TIM-2 shedding
Shedding of the transmembrane proteins L-selectin , amyloid precursor protein (APP)  and semaphorin 4D  has been shown to be regulated by calmodulin. Calmodulin was reported to bind to the transmembrane/intracellular region of these transmembrane proteins, and disruption of the interaction by the calmodulin inhibitor W7 stimulated ADAM-mediated shedding.
We therefore investigated whether the cleavage of TIM-2 was inducible by the calmodulin antagonist W7. HEK293 cells, transiently transfected with TIM-2, were treated with W7 (100 μm) for 1 h in the presence or absence of the ADAM10 inhibitor GI or the ADAM10 and ADAM17 inhibitor GW. The western blot results in Fig. 6A show that W7 treatment induced cleavage of TIM-2, and that shedding was blocked by GI and GW. Identical results were also obtained when HEK293 cells were transiently transfected with the AP–TIM-2 cDNA, as shown by western blotting and AP assays in Fig. 6B,C. Our data indicate that calmodulin might be involved in the regulation of shedding of TIM-2 by ADAM10.
We next investigated whether the ICD of TIM-2 influenced the W7-induced shedding of TIM-2. As shown by western blotting and AP assay (Fig. 6D,F), the TIM-2 deletion mutant lacking the ICD was also shed after W7 stimulation, whereas no shedding was seen when the cells were pretreated with the ADAM10 inhibitor GI or the ADAM10/ADAM17 inhibitor GW.
We also analyzed W7-induced shedding of AP–TIM-2_∆1, which lacked the potential ADAM10 cleavage site. As shown in Fig. 6E,F, there was no detectable W7-induced shedding of AP–TIM-2_∆1.
If calmodulin is involved in the regulation of shedding of TIM-2 by ADAM10, direct binding of calmodulin to TIM-2 should occur. We used calmodulin-coupled Sepharose to precipitate AP–TIM-2, AP–TIM-2_∆1 or AP–TIM-2_∆ICD. As shown in Fig. 6G, TIM-2 was precipitated by calmodulin-coupled Sepharose but not by uncoupled Sepharose, indicating that calmodulin physically interacted with TIM-2 and its mutants. This finding that the ICD of TIM-2 was not needed for calmodulin-prevented shedding was surprising. It has been recently shown that calmodulin binds not only to residues within the ICD, but also to areas within the transmembrane domain . Therefore, calmodulin might also interact with the transmembrane domain of TIM-2. We cannot, however, exclude the possibility that calmodulin interacts with another protein involved in the regulation of ADAM-mediated shedding of TIM-2, which might be part of a larger TIM-2 protein complex.
There are three major findings of this study. We show: (a) that the type I transmembrane protein TIM-2 is cleaved from the cell surface, and that cleavage is exclusively mediated by ADAM10, and not by ADAM17; (b) that shedding of TIM-2 is controlled by the Ca2+-binding protein calmodulin; and (c) that sTIM-2 can interact with H-ferritin, the ligand of transmembrane TIM-2.
Many proteins that are subject to shedding are cleaved by both ADAM10 and ADAM17, depending on the cellular stimulus used. An example of this group of proteins is human IL-6R, which is shed from cells upon stimulation with PMA [28, 29] and IM , whereby PMA induces ADAM17 and IM induces ADAM10 . Shedding of the cadherin proteins, however, is exclusively mediated by ADAM10 [30, 31]. Interestingly, HEK293 cells, which were used in this study, have been shown to shed human IL-6R upon PMA treatment . We therefore conclude that these cells express ADAM17, and that this protease, even upon appropriate stimulation, is not able to cleave TIM-2. This clearly demonstrates the substrate specificity of the ADAM proteases. Interestingly, in an earlier study, it was demonstrated that ADAM10 was responsible for the constitutive cleavage of the membrane-bound chemokine fractalkine, whereas ADAM17 cleaved this membrane protein upon stimulation of the cells with PMA .
Calmodulin binding to the transmembrane domain and part of the cytoplasmic domain of membrane proteins has been implicated in the regulation of the shedding of these proteins . This might explain why the cytoplasmic truncation of TIM-2 did not prevent inducibility of TIM-2 shedding upon treatment with the calmodulin inhibitor W7.
Kahn et al. found that calmodulin binding to L-selectin prevented the cleavage of this protein by ADAM17, and inhibition of calmodulin led to a profound induction of shedding of L-selectin . Calmodulin has also been implicated in the control of semaphorin 4D shedding by ADAM17  and in the processing of APP . In our study, we demonstrated that ADAM10, but not ADAM17, was responsible for cleavage of TIM-2. However, we demonstrated that inhibition of calmodulin led to induction of TIM-2 shedding without further stimulation. These results demonstrate that the observed involvement of calmodulin in the regulation of ADAM proteases is not restricted to ADAM17, but also applies to ADAM10, which otherwise is distinct in its biology.
Little is known about the biological consequences of TIM-2 shedding. This process leads to the loss of cell surface expression of TIM-2, which might have serious consequences. It has been shown previously that blockade of TIM-2 with an antibody exacerbates the development of collagen-induced arthritis in mice . Furthermore, it was shown that an sTIM-2–Fc fusion protein led to hyperproliferation of T cells  and to increased lung inflammation in the ovalbumin-induced asthma model . These results indicate that TIM-2 plays a negative regulatory role in the immune system. Loss of TIM-2 from the surface of immune cells might therefore lead to exacerbated immune processes.
For some proteins, it has been found that, upon shedding, the cleaved protein acquires a novel function. Membrane-bound IL-6R has properties that are completely different from those of the cleaved soluble IL-6R, which renders cells that, by themselves, would be unresponsive to the cytokine responsive to IL-6 . Likewise, it has been reported that membrane-bound tumor necrosis factor-α has anti-inflammatory properties, whereas soluble tumor necrosis factor-α, after cleavage by ADAM17, is one of the strongest proinflammatory cytokines known . In this regard, it is interesting that sTIM-2 binds to H-ferritin and thereby potentially prevents binding of H-ferritin to membrane-bound TIM-2 on other cells. This might lead to changes in iron homeostasis when massive shedding of TIM-2 occurs. Although the major site in the body where iron is absorbed is the intestine, most iron is recycled by monocytes and macrophages . In this regard, an important role of the balance between membrane-bound and sTIM-2 can be hypothesized.
The ORF (excluding the signal peptide-coding sequence) of full-length TIM-2 was PCR-amplified from murine thymus cDNA, with the 5′TIM-2_AflII (5′-GACTCTTAAGCATACAGCAGTGCAGGGCTGG-3′) and 3′TIM-2_NotI (5′-GACTGCGGCCGCGGACTCTTCTTCGGGGTAAGG-3′) primer combination. After digestion with AflII and NotI, the TIM-2 PCR product was cloned into the expression vector pcDNA3.1(+), containing the signal peptide from the murine immunoglobulin κ-chain, an N-terminal FLAG-tag, and a C-terminal His6-tag. For higher expression, the TIM-2 expression cassette was subcloned into the optimized eukaryotic expression plasmid p409 , via a PmeI site. For generation of AP–TIM-2, the vector pCRscript–AP–TGF-α (Agilent Technologies, Waldbronn, Germany) was digested with HindIII, and the AP was subcloned into the pcDNA3.1–TIM-2 plasmid.
For construction of the three TIM-2 deletion mutants (∆1, ∆2, and ∆3), splicing by overlap extension PCR was performed, with the pcDNA3.1–TIM-2 vector as a template. The resulting PCR products were subcloned into pcDNA3.1–AP via AflII and NotI to obtain the plasmids pcDNA3.1–AP–TIM–2_∆1, pcDNA3.1–AP–TIM–2_∆2, and pcDNA3.1–AP–TIM–2_∆3. For the ∆1 deletion from Pro221 to Lys230, the primer pair 5′mut1 (5′-AGTAATCCCTGGCTTCTATGTTGGCATCTC-3′) and 3′mut1 (5′-CATAGAAGCCAGGGATTACTTCAGTGTTAT-3′) was used. For the ∆2 deletion from Pro211 to Pro220, the primer pair 5′mut2 (5′-CTCAGATGACCCACAGAAGCCACAGAAAAACC-3′) and 3′mut2 (5′-GCTTCTGTGGGTCATCTGAGGATGTCACAG-3′) was used. For the ∆3 deletion from Trp201 to Asp210, the primer pair 5′mut3 (5′-CCTGCAGACCCTTGGGATGATAACACTGAAG-3′) and 3′mut3 (5′-CATCCCAAGGGTCTGCAGGAGTAGAGGGTA-3′) was used.
For the deletion of the immunoglobulin domain (ΔIgV) from His22 to Glu129, a PCR was performed with the primer pair 5′deltaIgV_AflII (5′-GATCCTTAAGATTTCCACGAGTCCACCAACAAGG-3′) and 3′TIM-2_NotI(Stop) (5′-GACTGCGGCCGCCTAGGACTCTTCTTCGG-3′), and the plasmid pcDNA3.1–TIM-2 as a template. The TIM-2 cDNA lacking the ICD (ΔICD, from Arg254 to Ser305) was generated by PCR, with the same forward primer as used for the full-length construct, and the reverse primer 3′deltaICD_NotI (5′-GACTGCGGCCGCGGTGATAACCATGGTGCTCAG-3′). The resulting PCR products, ΔIgV and ΔICD, were subcloned into pcDNA3.1–AP–TIM-2 via AflII and NotI to obtain the plasmids pcDNA3.1–AP–TIM-2_ΔIg and pcDNA3.1–AP–TIM-2_ΔICD.
The entire extracellular domain of TIM-2 was expressed as a soluble Fc fusion protein, with a strategy described previously . This construct was generated in a pcDNA3.1(+) vector into which the coding region for His22–Lys230 of TIM-2 and the human IgG1 Fc domain were inserted. For the amplification of sTIM-2–hFc, splicing by overlap extension PCR was performed with the primers 5′sTIM2_HindIII (5′-GATCAAGCTTCATACAGCAGTGCAGGGGCTGGC-3′), i3′sTIM2(EcoRI) (5′-GGTGTGGGTCTTGTCGAATTCCTTATTCAGGTTTTTCTG-3′), i5′hFc(EcoRI) (5′-GAAAAACCTGAATAAGGAATTCGACAAGACCCACACCTG-3′), and 3′hFc_NotI (5′-GATCGCGGCCGCTCACTTGCCAGGAGACAG-3′), with the pcDNA3.1–TIM-2 and the pcDNA_DEST40sgp130-hFc(opt) vectors as templates. The resulting PCR product, sTIM-2–hFc, was subcloned into pcDNA3.1–TIM-2 via HindIII and NotI.
The murine H-ferritin cDNA sequence was amplified by PCR from murine liver cDNA, with the primer pair 5′H-ferritin_NdeI (5′-GATCCATATGACCACCGCGTCTCCCTCGCAAGTGCG-3′) and 3′H-ferritin_BamHI (5′-GATCGGATCCGCTCTCATCACCGTGTCCCAGGGTG-3′). After digestion with NdeI and BamHI, the H-ferritin PCR product was cloned into the vector pet22b(+) (Novagen, Darmstadt, Germany). The coding sequence for a C-terminal PC-tag  was added to H-ferritin using PCR with the primer pair 5′L-PC-tag-BamHI (5′-GATCCAAGCTTGAAGACCAGGTCGATCCACGGCTGATCGATGGCAAGTGAC-3′) and 3′L-PC-tag-XhoI (5′-TCGAGTCACTTGCCATCGATCAGCCGTGGATCGACCTGGTCTTCAAGCTTG-3′). The oligonucleotides were annealed and then subcloned into the BamHI-digested and XhoI-digested plasmid pet22b-H-ferritin to generate the vector pet22b(+)H-ferritin–PC.
HEK293 and COS-7 cells were purchased from the ATTC (Rockville, MD/Manassas, VA, USA). MEFs deficient for ADAM10, for ADAM17 and for ADAM10/ADAM17 have been described previously [13, 39-41]. HEK293 cells, COS-7 cells and MEFs were cultured in DMEM high-glucose culture medium (PAA Laboratories, Cölbe, Germany) containing 10% fetal bovine serum at 37 °C with 5% CO2 in a water-saturated atmosphere.
Transfection of cells
HEK293 cells, COS-7 cells and MEFs were transiently transfected by the use of TurboFect (Fermentas, St Leon-Rot, Germany).
Cells were washed three times with sterile NaCl/Pi (1.4 m NaCl, 27 mm KCl, 100 mm Na2HPO4, 18 mm KH2PO4, pH 7.4). Inhibitors were added 30 min before stimulation with PMA (for 1–2 h), BzATP (for 30 min), IM (for 30 min), or W7 (for 1 h). Stimulation was performed in serum-free culture medium. Cells were then centrifuged 5 min, 10 000 g, and the pellet was lysed in mild lysis buffer (50 mm Tris/HCl, pH 7.5, 150 mm NaCl, 1% Triton X-100), supplemented with complete protease inhibitor cocktail tablets (Roche Diagnostics, Mannheim, Germany). To analyze ADAM10-mediated shedding of TIM-2 via P2X7R after BzATP stimulation, HEK293 cells were transiently transfected with the plasmid pcDNA6-hP2X7R, described previously , and the indicated TIM-2 cDNA. PMA, BzATP and IM were purchased from Sigma-Aldrich (Deisenhofen, Germany). W7 was purchased from Calbiochem, Millipore-Merck (Schwalbach, Germany). The metalloprotease inhibitors GI (ADAM10-selective) and GW (ADAM10-selective and ADAM17-selective) have been described previously . A10pro has been described previously . A10pro was kindly provided by M. L. Moss (BioZyme).
AP shedding assay
To determine the shedding of AP–TIM-2 and its deletion mutants via AP activity, 50 μL of reaction solution (0.1 m glycine, 1 mm MgCl2, 1 mm ZnCl2 containing 1 mg·mL−1 4-nitrophenyl phosphate disodium salt hexahydrate; Sigma-Aldrich) were added to 50 μL of cell culture supernatant or 5 μL of cell lysate plus 45 μL of DMEM. AP activity was measured by colorimetry at an absorbance of 405 nm. The ratio between AP activity in the supernatant and total AP activity in the cell lysate plus supernatant was calculated from the average of three identically treated wells. The ratio reflects the shedding activity of a given protease towards the indicated AP-tagged TIM-2 protein.
The cell culture supernatant from one well of a six-well plate (600 μL) was concentrated to a final volume of ~ 150 μL with a SpeedVac Plus SC110A (ThermoScientific, Dreieich, Germany). For immunochemical detection of TIM-2 proteins, cell lysates or concentrated culture supernatants were separated by SDS/PAGE and transferred to a poly(vinylidene difluoride) membrane (GE Healthcare, Munich, Germany) with a semidry electroblotting procedure. The membrane was blocked with 5% skimmed milk in NaCl/Pi for 1–2 h, and probed with primary antibodies as indicated at 4 °C overnight. After washing with NaCl/Pi plus 0.5% Tween-20, the membranes were incubated with the appropriate secondary antibodies conjugated to horseradish peroxidase (Thermo Scientific/Pierce, Perbio, Bonn, Germany). Protein bands were detected by chemiluminescence with the ECL detection kit (GE Healthcare, Munich, Germany), according to the manufacturer's instructions.
Flow cytometry analysis
For the analysis of cell surface expression of TIM-2 or its mutants, transiently transfected HEK293 cells (0.5 × 106 to 1 × 106) were washed with fluorescence-activated cell sorting (FACS) buffer (NaCl/Pi plus 0.2% BSA) and incubated with 1 μg of FLAG M2 mAb (Sigma-Aldrich, Taufkirchen, Germany) in 100 μL of buffer for 60 min on ice. After repeated washing with FACS buffer, the cells were incubated in a 1 : 100 dilution of allophycocyanin-conjugated AffiniPureF(ab′)2 fragment goat anti-[mouse IgG (H + L)] (Dianova, Hamburg, Germany) for 60 min on ice. Cells were washed twice, suspended in 200 μL of FACS buffer, and analyzed by flow cytometry (BD FACSCanto and diva software; Becton Dickinson, Heidelberg, Germany).
To detect TIM-2 on primary mouse cells, stimulated spleen B-cells (0.5 × 106 to 1 × 106) were first incubated with unlabeled mAb against CD16/32 (BD Biosciences, San Jose, CA, USA) to prevent nonspecific binding of mAbs to Fcγ receptors, and then incubated with phycoerythrin (PE)-conjugated mAb against mouse TIM-2 or PE-conjugated mouse IgG1κ isotype mAb (as control) for 60 min on ice. After being washed with NaCl/Pi twice, the stained cells were suspended in 300 μL of FACS buffer and analyzed by flow cytometry. PE-conjugated mAb against mouse TIM-2, PE-conjugated mouse IgG1κ isotype control, fluorescein isothiocyanate-conjugated mAb against mouse CD45R/B220 and fluorescein isothiocyanate-conjugated rat IgG2aκ isotype mAb (as control) were purchased from BioLegend (San Diego, CA, USA).
Recombinant production of murine H-ferritin
The pet22b(+)H-ferritin–PC vector coding for PC-tagged ferritin was transformed into E. coli BL21. Transformed bacteria were grown at 37 °C in Lysogeny Broth medium supplemented with ampicillin (100 μg·mL−1), and protein expression was induced with 1 mm isopropyl thio-β-d-galactoside (IPTG) at a D600 nm of 0.6–0.8 at 37 °C. After 2 h of culture, the bacteria were harvested by centrifugation (3000 g, 4°C, 10 min), suspended in NaCl/Pi, and lysed by sonication. The resulting supernatant containing murine H-ferritin–PC was heated at 70 °C for 15 min. At this temperature, most of the bacterial proteins were denatured, leading to efficient enrichment of H-ferritin–PC . H-ferritin was separated from the denatured cellular proteins by centrifugation for 10 min at 15 800 g.
Coimmunoprecipitation of sTIM-2 with murine H-ferritin
One milliliter of cell culture supernatant from one 10-cm Petri dish (6 mL) containing sTIM–2Fc or sgp130Fc from transiently transfected HEK293 cells was incubated with 25 μL of protein G–Sepharose (Roche, Mannheim, Germany) overnight at 4 °C. Immune complexes were precipitated, washed with NaCl/Pi twice, and incubated with 75 μL of H-ferritin solution overnight at 4 °C. H-ferritin was preincubated with protein G–Sepharose to prevent nonspecific binding of H-ferritin–PC to the beads. After washing, the coimmunoprecipitates were separated by 12% SDS/PAGE, and subjected to western blot analysis with a POD-coupled antibody against human Fc and a murine antibody against PC (HPC4). When the antibody against PC-tag was applied, buffers were supplemented with 4 mm CaCl2, because the binding of the antibody was calcium-dependent. For western blot detection of the input, 1 mL of sTIM–2Fc or sgp130Fc cell culture supernatant was precipitated with 40% (NH4)2SO4, and the 75-μL H-ferritin solution was concentrated to a final volume of ~ 30 μL with a SpeedVac Plus SC110A.
Isolation and activation of mouse spleen B cells
3Male C57BL/6N and BALB/c mice were obtained from Charles River Laboratories (Koeln, Germany). All experiments were performed according to German guidelines for animal care and protection. B cells were purified from the spleens of 8–12-week-old mice by passage through an iron mesh sieve and by negative selection with auto-MACS columns with the B-cell isolation kit (order no. 130-090-862; MiltenyiBiotec, Bergisch Gladbach, Germany), according to the manufacturer's instructions. Purified B cells were cultured in RPMI-1640 medium with l-glutamine (PAA Laboratories, Cölbe, Germany) supplemented with 10% fetal bovine serum and 50 μm β-mercaptoethanol (Gibco), and stimulated with IgM (5 μg·mL−1) and mAb against CD40 (5 μg·mL−1) for 24–72 h. The mAb against CD40 (HM40-3) was purchased from eBioscience (San Diego, CA, USA), and the IgM (goat anti-[mouse IgMF(ab′)2]) was purchased from Jackson ImmunoResearch Laboratories (West Grove, PA, USA).
To detect the binding of murine H-ferritin to sTIM–2Fc, an ELISA was performed. A 96-well plate was coated with 50 μL of H-ferritin diluted to 10 μg·mL−1 in NaCl/Pi buffer, and incubated overnight at 4 °C. After blocking with 5% BSA/NaCl/Pi for 1–2 h, 100-μL aliquots of different dilutions (1 : 10 to 1 : 1200) of cell culture supernatant from one 10-cm Petri dish (6 mL) containing sTIM–2Fc or sgp130Fc from transiently transfected HEK293 cells were added and incubated for 2 h. Bound sTIM–2Fc (or sgp130Fc as a negative control) was detected by incubation (1 h) with a POD-coupled antibody against human Fc. After each incubation step, the plate was washed twice with NaCl/Pi plus 0.05% Tween-20 and once with NaCl/Pi. To measure the enzymatic POD-coupled reaction, 100 μL of the substrate BM Blue POD (Roche Diagnostics, Mannheim, Germany) was added and incubated in the dark. After 5–10 min, the reaction was stopped by the addition of 50 μL of 1.8 m H2SO4, and the absorbance was read at 450 nm on an SLT Rainbow plate reader (Tecan, Maennedorf, Switzerland).
For the analysis of binding of H-ferritin to AP–TIM-2, sAP-TIM-2 (after IM-induced shedding), or AP–TIM-2 mutants, a 96-well plate was coated with 50 μL of H-ferritin diluted to 10 μg·mL−1 in NaCl/Pi and incubated overnight at 4 °C. After blocking with 5% BSA/NaCl/Pi for 1–2 h, 10 μL of lysate plus 90 μL of NaCl/Pi or 100 μL of culture supernatant from transiently transfected HEK293 cells (unstimulated or stimulated with IM, GI, or GW) was added and incubated for 2 h. After washing with NaCl/Pi (three to five times), AP activity was measured at 405 nm on an SLT Rainbow plate reader to detect the amount of bound AP–TIM-2, sAP–TIM-2, or AP–TIM-2 mutants.
Competitive AP-ELISA assay
A 96-well plate was coated with 50 μL of H-ferritin diluted to 10 μg·mL−1 in NaCl/Pi, and incubated overnight at 4 °C. One hundred microliters of shed AP–TIM-2 culture supernatant from transiently transfected HEK293 cells (stimulated with IM) was incubated with different amounts of sTIM–2Fc or sgp130Fc (as control) for 10 min at room temperature. The different ratios of shed AP–TIM-2/sTIM–2Fc or shed AP–TIM-2/sgp130Fc (ratio from 1 : 3 to 1 : 0.3125) were then added to the H-ferritin-coated wells for 2 h. The plate was washed three times to remove unbound shed AP–TIM-2 or sTIM–2Fc. AP activity was measured at 405 nm on an SLT Rainbow plate reader to detect the amount of bound shed AP–TIM-2.
For the calmodulin-binding studies, HEK293 cells from one 10-cm Petri dish, transiently transfected with AP–TIM-2, AP–TIM-2_∆1, or AP–TIM-2_∆ICD, were lysed in 400 μL of ice-cold lysis buffer (1% Triton X-100, 40 mm Tris, 10 mm EGTA). One hundred and fifty microliters of cell lysate was incubated with 100 μL of calmodulin–Sepharose 4B (GE Healthcare Europe, Freiburg, Germany) in 750 μL of lysis buffer under gentle agitation at 4 °C overnight. As a control, 150 μL of each cell lysate was incubated with 100 μL of Sepharose 4B (GE Healthcare Europe, Freiburg, Germany) under identical conditions. After washing (five times with lysis buffer), the immunoprecipitated proteins were separated by 12% SDS/PAGE, and subjected to western blotting with FLAG M1 primary mAb.
This work was funded by grants from the Deutsche Forschungsgemeinschaft, Bonn, Germany (DFG SCHE 907/2-1; SFB654 projects C5 and C8; SFB877 projects A1 and A4) and by the Cluster of Excellence ‘Inflammation at Interfaces’.