Interleukin-4-Inducing Principle from Schistosoma mansoni Eggs Contains a Functional C-Terminal Nuclear Localization Signal Necessary for Nuclear Translocation in Mammalian Cells but Not for Its Uptake

Abstract

Interleukin-4-inducing principle from schistosome eggs (IPSE/alpha-1) is a protein produced exclusively by the eggs of the trematode Schistosoma mansoni. IPSE/alpha-1 is a secretory glycoprotein which activates human basophils via an IgE-dependent but non-antigen-specific mechanism. Sequence analyses revealed a potential nuclear localization signal (NLS) at the C terminus of IPSE/alpha-1. Here we show that this sequence (125-PKRRRTY-131) is both necessary and sufficient for nuclear localization of IPSE or IPSE-enhanced green fluorescent protein (EGFP) fusions. While transiently expressed EGFP-IPSE/alpha-1 was exclusively nuclear in the Huh7 and U-2 OS cell lines, a mutant lacking amino acids 125 to 134 showed both nuclear and cytoplasmic staining. Moreover, insertion of the IPSE/alpha-1 NLS into a tetra-EGFP construct rendered the protein nuclear. Alanine scanning mutagenesis revealed a requirement for the KRRR residues. Fluorescence microscopy depicted, and Western blotting further confirmed, that recombinant IPSE/alpha-1 protein added exogenously is rapidly internalized by CHO cells and accumulates in nuclei in an NLS-dependent manner. A mutant protein in which the NLS motif was disrupted by triple mutation (RRR to AAA) was able to penetrate CHO cells but did not translocate to the nucleus. Furthermore, the uptake of native glycosylated IPSE/alpha-1 was confirmed in human primary monocyte-derived dendritic cells and was found to be a calcium- and temperature-dependent process. Live-cell imaging showed that IPSE/alpha-1 is not targeted to lysosomes. In contrast, peripheral blood basophils do not take up IPSE/alpha-1 and do not require the presence of an intact NLS for activation. Taken together, our results suggest that IPSE/alpha-1 may have additional nuclear functions in host cells.

Interleukin-4-inducing principle from schistosome eggs (IPSE/alpha-1) is a glycoprotein specifically secreted by the egg stage of Schistosoma mansoni (22, 46), a helminthic parasite which infects more than 200 million people in the tropics and subtropics. The pathology of schistosomiasis is caused mainly by a granulomatous and fibrosing immune reaction in the liver and gut in response to the eggs produced by mature female fertile worms (55). Recently, proteomic analyses identified IPSE/alpha-1 as a highly abundant protein in S. mansoni egg secretions (9, 14, 28, 37). It was previously shown that IPSE/alpha-1 has immunoglobulin-binding properties and activates basophils of immunologically naïve donors, resulting in histamine release and T-helper 2 (Th2)-type cytokine production (45). IPSE/alpha-1 also induces interleukin-4 (IL-4) secretion from murine basophils in vivo (47). However, since there are no known homologs of IPSE/alpha-1 in metazoans outside the Schistosoma genus, little is known regarding its potential functions in host cells or its role in basophil activation.

Our sequence analyses identified a putative monopartite nuclear localization sequence (NLS) at the C terminus of IPSE/alpha-1. Monopartite NLS motifs comprise a small cluster of basic amino acids preceded by a proline residue, with the classic example represented by the PKKKR motif in the simian virus 40 (SV40) T antigen (31). In addition to the NLS, the N terminus of IPSE/alpha-1 contains a classical hydrophobic secretory signal (CSS) sequence (45). CSS motifs are involved in the transport of nascent secretory polypeptide chains into the endoplasmic reticulum (ER), in a process called cotranslational transport, particularly for proteins with lengths of more than 100 amino acids (43). A signal peptidase removes this signal once it has entered the ER lumen. Under such conditions, a protein will normally not be targeted to the nucleus even if an NLS is present. Thus, most known nuclear proteins do not possess CSS motifs. N-terminal sequencing of mature IPSE/alpha-1 confirmed that the CSS motif is removed during export (45). One known example of the rare molecules exhibiting both CSS and NLS motifs on the same polypeptide chain is mouse fibroblast growth factor 3 (FGF-3), which displays both nuclear and extracellular localization (34). Given that IPSE/alpha-1 is secreted in large amounts from the subshell area of eggs (46), which are in close contact with the surrounding host tissues, we hypothesized that the mature IPSE/alpha-1 protein might be targeted to the nuclei of host cells. In this study, we demonstrate that IPSE/alpha-1 contains a functional NLS motif that is necessary and sufficient for translocation to mammalian cell nuclei and that also has a potential role in DNA binding. This suggests that IPSE/alpha-1 could have an important role in modulating the immune response after entering host cell nuclei.

MATERIALS AND METHODS

Subcloning and PCR mutagenesis.

Full-length IPSE/alpha-1 cDNA was amplified by PCR as described previously (45). Since the putative NLS is located close to the C terminus of IPSE/alpha-1, PCR mutagenesis was achieved by introducing the desired mutations in the 3′ primer. Oligonucleotide primer sequences are available in Table S1 in the supplemental material. The primers introduced an alanine substitution for each single amino acid in the putative NLS (125-PKRRRTY-131) (amino acid numbers refer to the full-length protein sequence, including the signal peptide). PCR fragments were amplified using high-fidelity proofreading Pfu Ultra DNA polymerase (Stratagene). The cycling conditions used were as follows: 95°C for 2 min for 1 cycle; 95°C for 30 s, 58°C for 30 s, and 72°C for 1 min 30 s for 30 cycles; and 1 cycle at 72°C for 10 min. The amplified fragment was purified using a QIAquick gel purification kit (Qiagen), digested with BglII and HindIII restriction enzymes, and subcloned into the pEGFP-C1 vector. To investigate whether the nuclear localization of full-length IPSE/alpha-1 is due to the putative monopartite NLS predicted in silico, we used the previously published plasmid pTetra-EGFP, which is an excellent tool for measuring NLS activity (4).

Phosphorylated, annealed oligonucleotide pairs encoding the sequences PKRRRTY (wild type), PARRRTY, and PKAAATY and containing GATC overhangs were subcloned into the BglII site of the pTetra-EGFP vector (4). This inserts the test sequence between the third and fourth copies of a tetra-enhanced green fluorescent protein (tetra-EGFP) fusion protein which is normally cytoplasmic, as it is too large to cross nuclear pores by passive diffusion. All constructs were verified by DNA sequencing.

Cell culture.

The human hepatocellular carcinoma cell line Huh7 D12 (ECACC 01042712) and the human osteosarcoma cell line U-2 OS (ATCC HTB-96) were grown in T25 and T75 cell culture flasks (Nunclon, Denmark) at 37°C in a humidified atmosphere of 95% air and 5% CO₂ in Dulbecco's modified Eagle's medium (DMEM) (GibcoBRL, United Kingdom) supplemented with 10% heat-inactivated fetal calf serum (FCS), 2 mM l-glutamine, 1 U/ml penicillin, and 1 mg/ml streptomycin. Chinese hamster ovary transferrin receptor variant b cells (CHO-TrVb cells; designated CHO-TfR− cells in this study) and CHO-TrVb cells stably transfected with human transferrin receptor 1 (CHO-TrVb1 cells; designated CHO-TfR+ cells herein) were cultured in Ham's F12-HEPES (Autogen Bioclear, United Kingdom) supplemented with 5% heat-inactivated FCS, 2 mM l-glutamine, 1 U/ml penicillin, and 1 mg/ml streptomycin. The medium for CHO-TfR+ cells additionally contained 200 μg/ml G418 (Fisher Bioreagents, United Kingdom).

Transient expression.

Transfections were performed using Transfast (Promega, United Kingdom) as instructed by the manufacturer. All transfections were performed in 6-well plates (Nunc) on autoclaved glass coverslips seeded with cells 24 h prior to transfection (1 × 10⁶ to 3 × 10⁶ cells/well). The experimentally determined optimal conditions (24 h with a 1:2 DNA/Transfast ratio) were used for all transfections.

Fluorescence microscopy.

Recombinant EGFP-IPSE/alpha-1 fusion protein expressed in transfected cells was visualized using fluorescence microscopy. The transfected Huh7 or U-2 OS cell line was washed twice with phosphate-buffered saline (PBS) and fixed in a 4% (wt/vol) paraformaldehyde-PBS solution for 10 min at room temperature. The cells were then washed 4 or 5 times with PBS and incubated with 0.5 μg/ml Hoechst 33258 stain for 10 min at room temperature. The Hoechst stain was removed by 4 or 5 washes with PBS. The coverslips were then mounted on 10 μl of 90% glycerol in PBS, and the edges were sealed with nail polish. Images were taken on a confocal microscope (LSM510 Meta; Zeiss) using the provided software (Zeiss LSM Image Examiner v3.5).

Immunofluorescence.

The antibody solutions used were spun at 14,000 × g for 1 min prior to use, and all steps of incubation were performed at room temperature unless otherwise stated. CHO-TfR− and CHO-TfR+ cells were seeded on coverslips at a density of 0.5 × 10⁵ cells for 48 h. Cells were then incubated with 0.15 nM recombinant IPSE/alpha-1 in serum-free internalization medium (HEPES-buffered Ham's F12 medium containing 10 mM NaHCO₃ and 2 mg/ml bovine serum albumin [BSA; fraction V] [Biomol]) for 30 min at 37°C. Immunolabeling was performed as previously described (49). Briefly, cells were fixed in a 3% (wt/vol) paraformaldehyde-PBS solution for 15 min, washed three times with PBS, incubated in 50 mM NH₄Cl for 10 min, washed three times again, and permeabilized in 0.2% (vol/vol) Triton X-100. The cells were then incubated in blocking buffer (2% [vol/vol] FCS, 2% [wt/vol] BSA) for 30 min prior to addition of anti-IPSE monoclonal antibody culture supernatant at a 1:10 dilution for 30 min. The cells were washed and then incubated with anti-mouse IgG secondary antibody labeled with Alexa Fluor 594 (Invitrogen, United Kingdom) for 30 min at room temperature. The unbound secondary antibody was removed by repeated washing in PBS, and the nucleus was stained by incubating the cells for 10 min in 10 μg/ml Hoechst 33258. After being washed in PBS, the coverslips were mounted on 10 μl of 90% glycerol in PBS, and the edges were sealed with nail polish.

Extraction of cytoplasmic and nuclear fractions.

Recombinant IPSE/alpha-1 was incubated with 5 × 10⁶ CHO-TfR+ and CHO-TfR− cells for 2, 6, 12, and 24 h. After incubation, each cell line was trypsinized, washed twice in ice-cold PBS (Sigma, United Kingdom), and collected by centrifugation at 5,000 rpm (2,655 × g) for 2 min in a microcentrifuge tube (Eppendorf model 5417R centrifuge). To the cell pellets, 200 μl of buffer A (50 mM NaCl, 10 mM HEPES [pH 8.0], 500 mM sucrose, 1 mM EDTA, 0.2% Triton X-100) containing 1× protease inhibitors (Calbiochem, United Kingdom) was added and vortexed briefly. The mixture was spun in a refrigerated microcentrifuge at 5,000 rpm (2,655 × g) for 2 min. The supernatant comprised the cytoplasmic extract and was stored for Western blot detection. The cell pellets containing the nuclei were resuspended and washed in 500 μl buffer B (50 mM NaCl, 10 mM HEPES [pH 8.0], 25% glycerol, 0.1 mM EDTA) containing 1× protease inhibitors (Calbiochem, United Kingdom). The supernatant was discarded, and cell pellets were incubated on ice for 30 min with frequent agitation in 50 μl buffer C (350 mM NaCl, 10 mM HEPES [pH 8.0], 25% glycerol, 0.1 mM EDTA) containing 1× protease inhibitors (Calbiochem, United Kingdom). The mixture was spun at 14,000 rpm (20,817 × g) for 15 min at 4°C. The supernatant was removed carefully and stored as the nuclear extract for Western blot detection. The extracts were run in standard 12% SDS-PAGE gels, blotted on nitrocellulose (Schleicher and Schuell Bioscience, Dassel, Germany [now Whatman/GE Healthcare]), stained with primary monoclonal antibody to IPSE/alpha-1 (1:2,500 for nuclear staining and 1:5,000 for cytosolic staining) and a secondary anti-mouse IgG1 antibody (Invitrogen Western Breeze kit) (1:1,000), and developed by a chemiluminescence detection method (Invitrogen Western Breeze kit), using a Fujifilm luminescence image reader (LAS-4000) as directed by the manufacturer. A primary goat polyclonal antibody to histone H3 (Abcam, United Kingdom) (diluted 1:5,000) and a rabbit anti-goat IgG1-alkaline phosphatase (AP) secondary antibody (1:5,000) for detection of H3, as well as a mouse monoclonal antibody to beta-actin (Sigma, United Kingdom) (1:1,000) and the AP-linked anti-mouse IgG from a Western Breeze kit as a secondary antibody, were used as control antibodies for subcellular extract specificity for both cytosolic and nuclear fractions.

Sequence analysis.

The IPSE/alpha-1 amino acid sequence (GenBank accession no. AAK26170) was used for identification of sequence motifs such as CSS or NLS motifs by using Web-based bioinformatic tools, including Cello (57), ESLPred (6), LOCSVMPSI (56), LocTree (40), MultiLoc (26), NLSpredict (11), Proteome Analyst (51), PLOC (42), PSORTII (41), SignalP 3.0 (5), SherLoc (48), SubLoc (29), TargetP 1.1 (18), and Wolf Psort (27). All programs were used with default settings.

Basophil activation.

Human basophils were obtained from peripheral blood of healthy donors by using a three-step purification protocol as described before (21). Yield and viability were determined by trypan blue exclusion, and basophil purity was determined by staining cytospin preparations with May-Grünwald stain. Purified basophils were incubated overnight in RPMI 1640 medium supplemented with 10% heat-inactivated FCS, 2 mM l-glutamine, 1 U/ml penicillin, and 1 mg/ml streptomycin, with recombinant IPSE/alpha-1, IPSE/alpha-1 ΔNLS, or IPSE/alpha-1 3R at different concentrations, at 37°C in a 5% CO₂ humidified incubator. IL-4 was measured in supernatants by enzyme-linked immunosorbent assay (ELISA) as previously described (45).

Expression of recombinant IPSE/alpha-1.

Recombinant IPSE/alpha-1, IPSE/alpha-1 ΔNLS, and IPSE/alpha-1 3R were expressed in Escherichia coli, purified, and refolded as described earlier (45). Glycosylated recombinant IPSE/alpha-1 used for live-cell imaging of monocyte-derived dendritic cells (MDDCs) and basophils was expressed in human embryonic kidney (HEK) cells transfected with the expression vector pSegTag2-IPSE/alpha-1. The pSegTag2 vector was obtained from Invitrogen. Secreted recombinant HEK-IPSE/alpha-1 was sequentially purified from the culture medium by immobilized metal-affinity chromatography and affinity chromatography with monoclonal anti-IPSE antibodies coupled to an NHS-HiTrap Sepharose column (GE Healthcare).

Labeling of recombinant HEK-IPSE/alpha-1 and live-cell imaging.

Purified recombinant HEK-IPSE/alpha-1 was fluorescently labeled with N-hydroxysuccinimide (NHS)-fluorescein (Pierce; now ThermoScientific) according to the manufacturer's instructions. After protein labeling, nonreacted NHS-fluorescein was removed using Zeba Desalt spin columns (Pierce). Live-cell imaging was performed with a Leica TCS SP5 inverse confocal laser scanning microscope and analyzed with LAS AF software. In detail, 2.5 × 10⁴ MDDCs were added to a channel of an IV0.4 μ-slide (Ibidi, Martinsried, Germany) and incubated for 2 h at 37°C and 6% CO₂ with 1 μl fluorescein-labeled HEK-IPSE/alpha-1 (1 mg/ml). Nuclei and lysosomes of the cells were then counterstained with Hoechst 33342 (1:10,000) and LysoTracker (1:20,000), respectively, for 30 min. A channel containing MDDCs without preincubation of fluorescein-labeled IPSE/alpha-1 was stained with LysoTracker alone as a control.

Purification and labeling of native IPSE/alpha-1.

For flow cytometry binding/uptake studies, IPSE/alpha-1 was purified from SEA via cation-exchange chromatography and affinity chromatography using specific anti-IPSE/alpha-1 monoclonal antibodies coupled to an NHS-HiTrap Sepharose column according to the manufacturer's instructions (GE Healthcare), as described earlier (17, 45). Purified IPSE/alpha-1 was concentrated and dialyzed. IPSE/alpha-1 was fluorescently labeled with PF-647 by use of a Promofluor labeling kit (Promokine, Heidelberg, Germany) according to the manufacturer's recommendations.

IPSE/alpha-1 binding/uptake by human MDDCs.

Monocytes were isolated from venous blood of healthy volunteers according to Institutional Review Board-approved protocols by density centrifugation on Ficoll followed by a Percoll gradient, as described previously (15), and were cultured in RPMI 1640 medium supplemented with 10% FCS, human recombinant granulocyte-macrophage colony-stimulating factor (rGM-CSF) (500 units/ml; a gift from Schering-Plough, Uden, Netherlands), and human rIL-4 (250 units/ml) (R&D Systems). On day 6, 10,000 immature DCs/well were seeded in a 96-well plate. Where indicated, cells were preincubated with 10 mM EGTA. Subsequently, cells were incubated with 500 ng/ml labeled IPSE/alpha-1 at 37°C or 4°C for 1 h and washed in ice-cold PBS or EGTA, where indicated, before analysis using flow cytometry.

RESULTS

Translocation of IPSE/alpha-1 to the nucleus requires a C-terminal NLS.

Sequence motif analysis software such as MultiLoc, PSORTII, and SherLoc predicted the presence of a monopartite NLS (125-PKRRRTY-131) at the C terminus of IPSE/alpha-1 (see Table S2 in the supplemental material). To determine whether this sequence facilitates localization of IPSE/alpha-1 to host cell nuclei, Huh7 or U-2 OS cells were transiently transfected with pEGFP-IPSE/alpha-1 or a series of EGFP-IPSE constructs in which the putative NLS was deleted or mutated. As shown in Fig. 1, EGFP fused to wild-type IPSE/alpha-1 showed a strong and exclusively nuclear staining in both cell lines (Fig. 1A and D), suggesting that IPSE/alpha-1 contains a functional NLS motif. In contrast, EGFP-IPSE/alpha-1 ΔNLS, which lacks amino acids 125 to 134, was detected in both the nucleus and the cytosol (Fig. 1B and E) and showed very similar staining to that of control EGFP alone (Fig. 1C). This result indicates that cytosolic IPSE/alpha-1 can translocate to the nuclei of mammalian host cells and suggests that the PKRRRTY motif at its C terminus may indeed function as an NLS. In the absence of a functional NLS, as seen with the unfused EGFP control, the protein can also enter the nucleus, probably by diffusion, resulting in a mixed cytoplasmic and nuclear localization.

FIG. 1.

Localization of EGFP-IPSE/alpha-1 in transiently transfected mammalian cell lines. (A) Huh7 cells transfected with pEGFP-IPSE/alpha-1, showing complete translocation of EGFP-IPSE/alpha-1 to the nucleus. (B) Huh7 cells transfected with pEGFP-IPSE/alpha-1 ΔNLS, lacking the C-terminal PKRRRTY NLS, showing mixed nuclear and cytoplasmic fluorescence. (C) Control transfection of U-2 OS cells with unmodified EGFP-C1 vector. (D to M) U-2 OS cell line transfected with pEGFP-IPSE/alpha-1 mutants with different alanine-substituted amino acids in the NLS and counterstained with Hoechst stain (insets). (D) pEGFP-IPSE/alpha-1, showing exclusive localization in the nucleus. (E) pEGFP-IPSE/alpha-1 ΔNLS, showing mixed nuclear and cytoplasmic localization. (F, L, and M) pEGFP-IPSE/alpha-1 NLS mutants in which P, T, and Y were replaced, showing no or little effect on nuclear localization. (H, I, and J) Single mutants partially decreasing nuclear localization. Replacement of all three arginines (K) or lysine (G) led to complete disruption of nuclear localization. (N) Summary of the effects of Ala mutations on subcellular localization of the EGFP-IPSE fusion protein. U-2 OS cells were transiently transfected with pEGFP-IPSE/alpha-1 and mutants. One hundred transfected cells were counted for each transfection, and the percentage of cells displaying exclusively nuclear fluorescence, as opposed to mixed nuclear and cytoplasmic fluorescence, was recorded.

To explore this further, alanine scanning mutagenesis was performed on the C terminus of IPSE/alpha-1 to examine the effects of single amino acid substitutions within the putative NLS on subcellular localization in U-2 OS cells. Replacement of P125 (Fig. 1F), T130 (Fig. 1L), or Y131 (Fig. 1M) with alanine had no deleterious effect on the nuclear localization of EGFP-IPSE/alpha-1. However, mutant EGFP-IPSE/alpha-1 proteins harboring alanines at any one of the amino acids in the 126-KRRR-129 region showed significant cytoplasmic localization (Fig. 1G to J). Moreover, replacement of the three arginine residues with alanines (3R) (Fig. 1K) resulted in a severe disruption of the nuclear localization. These results strongly suggest that the 126-KRRR-129 motif is a functional NLS motif.

To perform a more quantitative analysis of the effects of these mutations, 100 transfected cells for each construct were scored for the percentage of cells showing complete nuclear localization of the EGFP fusion protein. As shown in Fig. 1N, while wild-type (100%), P125A (98%), and T130A (95%) constructs were almost entirely nuclear, the K126A mutation (5%) drastically reduced the number of cells showing exclusively nuclear staining. The single-arginine mutations R128A and R129A reduced the percentage of nucleus-only cells by approximately 50%, whereas the R127A mutation had a stronger effect, with only approximately 20% of cells showing strong nuclear localization. For the mutant with all three arginines replaced (IPSE 3R), all cells examined were disrupted for nuclear localization of the fusion protein. In summary, our results show that IPSE/alpha-1 contains an NLS that targets it to the nuclei of mammalian cells, where it may interact with DNA, chromatin, and/or nuclear proteins.

The NLS is sufficient to target IPSE/alpha-1 or EGFP proteins to the nucleus.

Having identified the KRRR residues as necessary for nuclear translocation of IPSE, we next evaluated whether the putative NLS sequence was sufficient on its own to direct targeting of an unrelated protein to the nucleus. Since all EGFP-IPSE NLS mutants showed at least partial nuclear localization, it is possible that the fusion proteins are small enough to undergo passive diffusion into the nucleus. Interactions with DNA or other proteins might permit retention of EGFP-IPSE in the nuclear compartment. To circumvent this problem, IPSE NLS sequences were subcloned into the pTetra-EGFP plasmid construct, encoding a tetra-EFGP fusion protein of approximately 100 kDa that is largely excluded from the nucleus (Fig. 2 A). Insertion of a canonical SV40 NLS between the third and fourth copies of the EGFP sequences on the tetra-EGFP plasmid resulted in complete nuclear localization of the tetra-EGFP protein (Fig. 2B). Similarly, insertion of the IPSE PKRRRTY motif at this position also resulted in complete nuclear localization, showing that this sequence is sufficient to target large polypeptides to the nucleus. As expected, mutation of the three arginines or K126 to alanine resulted in proteins that were exclusively cytoplasmic (Fig. 2D and E), confirming our previous finding that the KRRR motif is a functional NLS.

FIG. 2.

Nuclear translocation of tetra-EGFP. U-2 OS cells were transfected with tetra-EGFP-NLS constructs and counterstained with Hoechst stain (insets). (A) Control tetra-EGFP expressed only in the cytoplasm. (B) Tetra-EGFP fused with canonical SV40 NLS, showing complete nuclear localization. (C) NLS from IPSE/alpha-1 fused with tetra-EGFP, showing exclusively nuclear localization similar to that with SV40 NLS. (D) Tetra-EGFP-NLS constructs in which all the arginines (3R) were replaced with alanine, showing cytoplasmic distribution, which confirms the disruption of nuclear localization. (E) Tetra-EGFP-NLS construct in which lysine (K) was replaced with alanine, also showing complete disruption of nuclear localization.

To confirm that non-EGFP-fused IPSE/alpha-1 can localize to the host cell nucleus, wild-type or mutant mature IPSE/alpha-1 proteins (without EGFP) were transiently expressed in transfected U-2 OS cells. IPSE/alpha-1 proteins were detected using a specific monoclonal antibody raised against the mature protein (45). As shown in Fig. 3, whereas the wild-type IPSE/alpha-1 protein localized entirely to the nucleus (Fig. 3A), the IPSE/alpha-1 ΔNLS (Fig. 3B) and IPSE/alpha-1 3R (Fig. 3C) mutants were entirely cytoplasmic. This provides further evidence that the NLS sequence is essential for translocation to host cell nuclei.

FIG. 3.

Nuclear and cytoplasmic localization of non-EGFP-fused IPSE/alpha-1. (Left) U-2 OS transfection with wild-type IPSE/alpha-1-encoding plasmid, showing exclusively nuclear localization of IPSE/alpha-1. (Middle) U-2 OS transfection with IPSE/alpha-1 ΔNLS-encoding plasmid. (Right) U-2 OS transfection with IPSE/alpha-1 3R-encoding plasmid. Both the middle and right panels show exclusion from the nucleus. After 24 h, the cells were fixed, permeabilized, stained with a monoclonal antibody to IPSE/alpha-1 and an Alexa 594-labeled secondary antibody, and counterstained with Hoechst nuclear stain (insets).

Exogenous IPSE/alpha-1 protein translocates to the nuclei of mammalian cells.

Since IPSE/alpha-1 is a secreted protein with a functional NLS, we investigated whether exogenous IPSE/alpha-1 can be internalized and translocate to mammalian cell nuclei. Our previous studies indicated that IPSE/alpha-1 binds to serum transferrin (Tf) from humans or other mammals, suggesting the Tf/TfR pathway as a possible route of entry into host cells (S. Blindow et al., unpublished data). To test this hypothesis, CHO-TrVb cells, which do not express any endogenous transferrin receptors (designated CHO-TfR− cells), and CHO-TrVb1 cells, which stably express human TfR1 (39) (designated CHO-TfR+ cells), were incubated with 0.15 nM purified recombinant IPSE/alpha-1 expressed in E. coli, and localization of the protein was assessed by immunocytochemistry after 30 min. As shown in Fig. 4 A to D, staining with the IPSE-specific monoclonal antibody (45) revealed that exogenous recombinant IPSE/alpha-1 rapidly entered both CHO-TfR+ and CHO-TfR− cells and showed nuclear staining. These results do not point to a facilitating role of the human TfR1 for cellular uptake of IPSE/alpha-1. Consistent with this, Western blot analysis of cytosolic and nuclear extracts prepared 24 h after addition of exogenous recombinant IPSE/alpha-1 revealed that in CHO-TfR+ and CHO-TfR− extracts, IPSE/alpha-1 protein was also detected in the nuclear fraction (Fig. 4G). Exogenous addition of the 3R mutant, with a nonfunctional NLS, resulted in cellular uptake but not in nuclear translocation (Fig. 4E and F). This suggests that while an intact NLS is needed for nuclear translocation, cellular uptake in this cell type is independent from the NLS.

FIG. 4.

Subcellular localization of exogenously added recombinant IPSE/alpha-1. CHO-TfR+ (A and B) and CHO-TfR− cells (C and D) were treated with recombinant purified IPSE/alpha-1 for 30 min at 37°C and stained with a monoclonal anti-IPSE/alpha-1 antibody and an Alexa 594-labeled secondary antibody. The photographs demonstrate intracellular and nuclear localization of IPSE/alpha-1 in both cell lines. (E and F) Incubation of CHO-TfR+ cells with IPSE/alpha-1 3R mutant, showing internalization but loss of nuclear translocation. Bars, 10 μm (A to D) and 20 μm (E and F). (G) Western blots of nuclear and cytosolic extracts of CHO-TfR+ (lanes 1 and 3) and CHO-TfR− (lanes 2 and 4) cells 24 h after exposure to exogenous recombinant IPSE/alpha-1. Blots were stripped and reprobed with an antibody to beta-actin or histone H3 as a control for extract specificity. IPSE/alpha-1 was found in the cytosolic and nuclear fractions, but the nuclear extracts contained an additional, higher-molecular-weight band, suggesting that IPSE/alpha-1 could undergo modifications during nuclear translocation.

Native glycosylated IPSE/alpha-1 is taken up by human dendritic cells in a Ca²⁺-dependent manner.

To determine whether native IPSE/alpha-1 is also internalized by primary cells, the capacity of human MDDCs to take up IPSE/alpha-1 was assessed. Dendritic cells incubated for 1 h with glycosylated IPSE/alpha-1 purified from S. mansoni eggs and labeled with PF-647 displayed a strong increase in fluorescence as determined by flow cytometry (Fig. 5 A). Since SEA, which also contains IPSE, is internalized by dendritic cells in a C-type lectin- and calcium-dependent manner (52), we tested the calcium dependency of these cells for recognition and internalization of IPSE/alpha-1. Indeed, pretreatment with EGTA, a calcium chelator, almost totally abolished the ability of the cells to bind and take up IPSE/alpha-1 (Fig. 5A). Finally, to ascertain that IPSE/alpha-1 is truly internalized by MDDCs, the incubated cells were washed in EGTA to remove all surface C-type lectin-bound IPSE/alpha-1 (Fig. 5B). Importantly, the latter treatment did not reduce the fluorescence intensity of the cells, while the same treatment on cells incubated with IPSE/alpha-1 at 4°C, which prevents receptor-mediated uptake, did lower the fluorescence of the cells back to background levels (Fig. 5B). This shows that native IPSE/alpha-1 is efficiently internalized by human dendritic cells, in a Ca²⁺- and temperature-dependent manner. Interestingly, uptake by the dendritic cells was dependent on glycosylation, as recombinant IPSE/alpha-1 expressed in bacteria, in contrast to that taken up in CHO cells, was not internalized by the primary cells (Fig. 5C). Together, the calcium and glycosylation dependencies of uptake by monocyte-derived dendritic cells point to a C-type lectin-mediated mechanism.

FIG. 5.

(A) Uptake/binding of IPSE/alpha-1 by human dendritic cells. Immature MDDCs were preincubated with EGTA, where indicated, followed by a 1-h incubation with PF-647-labeled IPSE/alpha-1 at 37°C. Uptake of antigens by MDDCs was evaluated by fluorescence-activated cell sorter (FACS) analysis. (B) To determine that IPSE/alpha-1 is truly internalized by dendritic cells, incubated cells were washed in EGTA to remove all surface C-type lectin-bound IPSE/alpha-1. Cells incubated with IPSE/alpha-1 at 4°C, preventing receptor-mediated uptake, were used as a control. Fluorescence intensity was set to 100% for conditions without EGTA treatment. (C) Uptake/binding of native purified glycosylated IPSE/alpha-1 (nIPSE) in comparison with the nonglycosylated bacterial recombinant. (A to C) Data are shown as means plus standard deviations (SD) for duplicates. Data for one representative experiment out of two are shown. P < 0.001 for significant differences compared to the control (one-sided t test). (D to G) Live-cell imaging of MDDCs incubated with fluorescein-labeled HEK-IPSE/alpha-1 for 2 h at 37°C before staining of the nucleus and the lysosomes with Hoechst 33342 and LysoTracker, respectively. (D) Overlay image of panels E to G, showing that IPSE/alpha-1 does not colocalize significantly with lysosomes. (E) Hoechst 33342 staining (blue). (F) Fluorescein-HEK-IPSE/alpha-1 (green). (G) LysoTracker staining (red). Bar, 10 μm.

IPSE/alpha-1 is not targeted to lysosomes.

To rule out the possibility that IPSE/alpha-1 undergoes lysosomal degradation after uptake via endocytic mechanisms, live-cell imaging was performed with human monocyte-derived dendritic cells incubated for 2.5 h with fluorescein-HEK-IPSE/alpha-1 (green) and LysoTracker (red) and counterstained with Hoechst 33342 (blue). The results shown in Fig. 5D to G suggest that internalized IPSE/alpha-1 found in the cytosol of the cell does not colocalize with lysosomes. The lack of nuclear localization in these experiments was probably due to the interference of fluorescein, which is conjugated via free amino groups in the protein, including mainly the KRRR residues in the NLS, and thus very likely to affect its functionality, as suggested by the results shown in Fig. 1. Since the uptake of this protein is mediated via carbohydrates, fluorescein labeling did not affect the protein's entry into MDDCs.

Neither a functional NLS nor cellular uptake is required for IPSE/alpha-1-mediated activation of human basophils.

IPSE/alpha-1 is known to trigger IL-4 release from naïve human basophils (45), although the exact underlying mechanism, which is known to involve immunoglobulin E binding (22, 45), is not fully understood. For example, and in light of our findings, it is not known whether this phenomenon requires nuclear localization of the parasite protein after internalization. To investigate this, we compared basophil activation by recombinant wild-type protein and the IPSE/alpha-1 ΔNLS or IPSE/alpha-1 3R mutant. As shown in Fig. 6, wild-type IPSE showed a dose-dependent induction of IL-4 release from peripheral blood basophils, as shown previously (45). In contrast, IPSE/alpha-1 ΔNLS showed little or no ability to induce IL-4 production under the same conditions. However, basophil activation was not affected in the NLS-defective IPSE/alpha-1 3R mutant, suggesting that the NLS activity per se is not required for this function. The apparent reduction in IL-4 production with IPSE/alpha-1 at concentrations above 100 μg/ml was presumably due to stimulation in the supraoptimal concentration range, which can lead to Src homology 2 domain-containing inositol 5′-phosphatase (SHIP)-mediated downregulation of mediator release (20). This was not observed with the IPSE/alpha-1 3R mutant, as indicated by the dose-response curve (Fig. 6A). The reason for the missing suppression of IL-4 production at high concentrations of the IPSE/alpha-1 3R mutant resulting in IL-4 levels beyond those reached upon stimulation of basophils via anti-IgE at optimal concentrations (data not shown) is presently under study.

FIG. 6.

(A) Dose-response curve for interleukin-4 secretion by naïve purified human peripheral blood basophils. Basophils were stimulated with different concentrations of recombinant IPSE/alpha-1, IPSE/alpha-1 ΔNLS, and IPSE/alpha-1 3R in the presence of IL-3 (2.5 ng/ml) (n = 3 independent experiments with cells from different donors). Values represent the mean percentages of maximum IL-4 release ± standard errors of the means (SEM). (B) Nonreducing 12% SDS-PAGE of unglycosylated IPSE/alpha-1, IPSE/alpha-1 3R, and IPSE/alpha-1 ΔNLS. While the first two recombinant proteins appear as a double or single band with an apparent molecular mass of 32 to 36 kDa, the truncated mutant, in which the seventh and terminal Cys residue has been removed, appears as monomeric bands of approximately 16 to 18 kDa. (C) Live-cell imaging depicting a human peripheral blood basophil (lobulated nucleus; stained with Hoechst 33342 [red]) incubated with 1 μl fluorescein-labeled HEK-IPSE/alpha-1 (1 mg/ml) for 60 min at 37°C and 6% CO₂. Serial sections (1 to 10) along the z axis show that fluorescein-labeled IPSE/alpha-1 (green) is not internalized but is seen in clusters on the plasma membrane in sections not containing the nucleus. Bar, 10 μm.

The above results suggest that nuclear localization of IPSE/alpha-1 is not necessary for IgE binding and subsequent IL-4 induction in human basophils. However, the results also indicate that deletion of the C terminus, including the NLS, abrogates this function. As an explanation for this, our recent studies have shown that dimerization of native IPSE/alpha-1 requires a cysteine residue (C132) at the C terminus of IPSE, which forms an interchain disulfide bond (54). This cysteine is removed in IPSE/alpha-1 ΔNLS but is present in the IPSE/alpha-1 3R mutant. Since IPSE/alpha-1 ΔNLS therefore occurs as a monomer (Fig. 6B), we concluded that dimerization of IPSE/alpha-1, but not nuclear localization, is necessary for its ability to induce IL-4 release by human basophils. Finally, live-cell imaging of basophils incubated with fluorescein isothiocyanate (FITC)-labeled HEK-IPSE/alpha-1 (Fig. 6C, showing successive sections along the z axis) depicts that while the parasitic protein aggregates on the surface, probably clustering the high-affinity IgE receptor via binding to IgE, it is not internalized.

Altogether, these results show that while the NLS in IPSE/alpha-1 is functional, i.e., is able to translocate large proteins from the cytosol to the nucleus, it is not needed for cellular uptake or for activation of peripheral blood basophils. This study demonstrates that a protein secreted by S. mansoni contains an NLS motif that is functional in host cells.

DISCUSSION

Several pathogens use host cell surface receptors to gain entry into host cells. For example, the mannose receptor (MR) (which is probably responsible, at least in part, for the uptake of IPSE/alpha-1 by MDDCs) is used by Trypanosoma cruzi to gain entry into cardiomyocytes (50), and cellular uptake has also been shown for the opportunistic pathogenic yeast Candida albicans (8). Dendritic cell-specific intercellular adhesion molecule-3-grabbing nonintegrin (DC-SIGN) and MR on immature dendritic cells are also used by viruses such as human immunodeficiency virus (HIV), via gp120 (35), as well as by larger pathogens such as Leishmania (12), for internalization. Because of their relative sizes, with a few exceptions, such as Trichinella spiralis and the whipworm Trichuris, most helminth parasites cannot enter host cells. However, it is possible that they use similar strategies involving hijacking host cell receptors for internalization and subsequent nuclear translocation of switch factors which may, e.g., affect host cell transcriptional patterns.

This study has demonstrated for the first time that a protein secreted by the human parasite S. mansoni contains an NLS motif that is functional in host cells. IPSE/alpha-1 is produced in the subshell area of the egg and is not detectable in the miracidium, the parasitic larval stage present in mature eggs, as either protein or mRNA (46). After secretion, IPSE/alpha-1 comes into close contact with inflammatory cells recruited to the vicinity of the egg surface (46). There is no evidence to date for an intracellular or nuclear localization of IPSE/alpha-1 in schistosome tissues. Altogether, the data suggest that IPSE/alpha-1 may have functional interactions with host cells rather than in the parasite. In order to reach the nucleus, a secreted protein has to be internalized back to the cytoplasm or, in the case of IPSE/alpha-1, into the cytoplasm of host cells. This may involve endocytosis-dependent processes, as reported for angiogenin (23) and parathyroid hormone-related protein (PtHrP) (3).

Using standard approaches, we have demonstrated that IPSE/alpha-1 contains a monopartite functional NLS motif (PKRRRTY) that is necessary and sufficient for its translocation to the nucleus. We have also demonstrated that this sequence is functional in the context of other proteins, such as EGFP. The PKRRRTY sequence is in agreement with the basic core consensus sequence K(K/R)X(K/R) of monopartite NLS motifs (24, 32). Our results showed that different alanine mutations had different effects on disruption of NLS functionality (which we have defined as exclusively nuclear localization), depending on the position. Mutations of 126K and 127R appeared to have the strongest effects. This is in good agreement with the work of Hodel et al., who determined the energy profiles for the binding of the SV40 NLS (PKKKRKV) to importin alpha and, based on these, defined the relative importance of each residue in the NLS (25).

Native IPSE/alpha-1 released by schistosome eggs is a dimer as a consequence of an interchain disulfide bond involving the C-terminal cysteine residue C132 (54). Despite the relatively low molecular mass (33 to 35 kDa) of IPSE/alpha-1 dimers, the mutants defective for NLS activity were excluded from the nucleus (Fig. 3 and 4E and F). This may indicate retention of IPSE/alpha-1 by cytosolic components or the presence of a nuclear export signal (NES), although we were unable to identify a consensus NES motif (36) within the IPSE/alpha-1 sequence.

As stated earlier, examples of proteins that contain both CSS (secretory) and NLS motifs are relatively rare, with an exception being the mouse FGF-3 protein (34). This raises the question of how proteins containing both motifs are secreted rather than targeted to the nucleus in the cell of origin. In FGF-3, the CSS cleavage site is adjacent to a bipartite NLS which is separated by only six amino acids. Increasing the distance between the signal sequence cleavage site and the NLS in FGF-3, by introducing a 15-amino-acid neutral linker sequence, results in exclusively secretory pathway localization, favoring recognition of the secretory signal over the NLS (34). In IPSE/alpha-1, the distance between the two motifs is 104 amino acids. Thus, this separation of the CSS motif from the NLS may be important in directing IPSE/alpha-1 for secretion from the producing structures in schistosome eggs.

We have shown that introduction of E. coli-expressed IPSE protein to the culture medium of growing CHO cells results in its internalization and subsequent targeting to the nucleus. Moreover, deletion of the C terminus of IPSE, including the NLS sequence, disrupted nuclear uptake (Fig. 3 and 4) but not internalization (Fig. 4E and F). These findings raise questions regarding which mechanisms IPSE/alpha-1 uses to gain access to the cytosolic and nuclear compartments of host cells. We have previously found that IPSE/alpha-1 binds mammalian Tfs of different species (human, mouse, rat, and goat) (Blindow et al., unpublished data), suggesting the Tf/TfR pathway as a possible route of entry. Our data suggest that the presence of TfR1 is not a prerequisite for IPSE/alpha-1 entry into cells (Fig. 4). This is in good agreement with the findings of Cervi and coauthors showing that labeled SEA (a complex antigenic mixture also containing IPSE/alpha-1) does not colocalize with TfR in murine dendritic cells (10). Entry via this route would pose an additional problem, as this receptor is internalized mainly through clathrin-coated pits and the ligand and receptor are recycled back to the plasma membrane (13). In order to gain access to the host cell cytoplasm, IPSE/alpha-1 would have to dissociate from the Tf-TfR complex and escape the endosomal recycling compartment.

Incubation of primary human MDDCs with native as well as HEK cell-expressed recombinant IPSE/alpha-1 showed that uptake of IPSE/alpha-1 into these cells is mediated by a calcium-dependent mechanism. C-type lectins are carbohydrate-binding receptors expressed on both immune and structural cells and are known to depend on calcium for binding to and internalization of glycosylated molecules. Given that native IPSE/alpha-1 is glycosylated (54), it is likely that in dendritic cells and possibly other cell types, such as macrophages, IPSE/alpha-1 can gain access to the cytosolic and nuclear compartments through C-type lectin-dependent internalization. Previous work demonstrated uptake of SEA (containing IPSE/alpha-1) by human MDDCs via the MR, macrophage galactose-type lectin (MGL), and DC-SIGN (52). However, that study failed to identify any uptake into the nucleus. In our view, this was due to the coupling technique used, which is likely to lead to a covalent modification of the NLS leading to a loss of function. The lack of uptake by peripheral blood basophils (Fig. 6C) is in agreement with the lack of C-type lectin expression on basophils. To the best of our knowledge, neither MR (CD206), DC-SIGN (CD209), nor MGL (CD301) has been described as present on human basophils. The lack of uptake at 4°C (Fig. 5B) points to an endocytic process and is compatible with a C-type lectin-mediated process, as is the complete lack of binding and uptake of unglycosylated recombinant IPSE/alpha-1 by MDDCs (Fig. 5C).

In the case of endocytic uptake, a cytosolic phase has to be assumed for subsequent translocation to the nucleus. Interestingly, the MR in particular has been shown to be involved in antigen cross-presentation (7), a process by which exogenous antigens are presented in association with major histocompatibility complex class I (MHC I) (33). While different mechanisms have recently been suggested to account for cross-presentation in dendritic cells (1, 16), most include a cytosolic transit phase. Such a cytosolic phase would provide an opportunity for an interaction of IPSE/alpha-1 with the nuclear import machinery and would result in nuclear translocation. Furthermore, while the work of van Liempt et al. (52) suggested targeting of SEA to the MHC II lysosomes of MDDCs, our results (Fig. 5D to G) clearly indicate that, at least for IPSE/alpha-1, these compartments are distinct. In addition to C-type lectin-dependent uptake, other mechanisms of uptake can be envisaged. Dendritic cells can internalize antigens via Fc receptors (44), and IPSE/alpha-1 is an immunoglobulin binding factor (45).

It remains to be established whether IPSE/alpha-1 uptake by CHO cells may rely on such a process as well. Since the CHO experiments were carried out with unglycosylated recombinant protein and MDDCs did not take up the unglycosylated form, underlying pathways appear to differ between cell types and may reflect alternative surface receptor expression patterns. Alternatively, IPSE/alpha-1 entry into CHO cells in these experiments might be mediated by a receptor-independent pathway. Indeed, some proteins are able to enter cells due to the presence of short, positively charged Arg/Lys-rich stretches of amino acids, termed protein transduction domains or cell-penetrating peptides (CPPs), described, e.g., for a positively charged domain in HIV-1 Tat protein (GRKKRRQRRR) (53). It is now thought that CPPs can be internalized via several different endocytic mechanisms or via direct translocation through the plasma membrane (2, 30). Thus, the PKRRRTY NLS in IPSE/alpha-1 might fulfill a dual role as a CPP and an NLS, transporting this secretory protein directly from the extracellular environment into the nuclei of host cells surrounding the parasite's eggs. However, the finding that the recombinant IPSE/alpha-1 3R mutant, with a strongly reduced positive charge, was still able to enter mammalian cells (without translocating to the nucleus, due to the disruption of the NLS motif) and, to some extent, the finding that IPSE/alpha-1 is not taken up by basophils argue against such a CPP-like mechanism. Nevertheless, the CPP-like sequence might still be necessary for endosomal escape and access to the cytosol, as it has been shown that CPPs such as that of HIV-1 Tat use a mechanism requiring endosomal acidification in order to escape to the cytosol before translocating to the nucleus (19).

Regarding the potential roles of IPSE/alpha-1 in host cell nuclei, we have found that IPSE/alpha-1 has DNA binding activity and is also associated with DNA in SEA (G. Schramm et al., unpublished data). It is interesting to speculate that binding of IPSE/alpha-1 to DNA or chromatin may have a role in altering gene expression in its target host cells. DNA binding experiments (data not shown) indicated an overlap of the NLS and DNA binding activities, as deletion of the NLS also fully ablated DNA binding. This is in line with the work of Cokol et al. (11), who found an overlap between the NLS and DNA binding regions for 90% of the proteins for which both the NLS and DNA binding regions were known. Our preliminary experiments using whole-genome DNA arrays point to dramatic changes in transcriptional patterns in MDDCs treated with IPSE/alpha-1 (data not shown).

IPSE/alpha-1 has no clear sequence homology with any other known protein, and its recently elucidated three-dimensional structure has yet to reveal more about its potential functions (38). This study has demonstrated that the C terminus of IPSE appears to have multiple functions, as an NLS and in basophil activation, in addition to other potential roles described above. Future work will focus on further characterizing the receptors enabling IPSE/alpha-1 to enter mammalian (primary) cells, the potential mechanisms of endosomal escape, and the consequences of nuclear targeting of IPSE/alpha-1 in these cells, e.g., via transcriptional profiling. The properties of IPSE/alpha-1 described here also make it an interesting potential vehicle for intracellular and nuclear delivery.