DNase X Is a Glycosylphosphatidylinositol-anchored Membrane Enzyme That Provides a Barrier to Endocytosis-mediated Transfer of a Foreign Gene*

DNase X is the first mammalian DNase to be isolated that is homologous to DNase I. In this study, we have examined its function using a novel monoclonal antibody and showed it to be expressed on the cell surface as a glycosylphosphatidylinositolanchored membrane protein. High level expression was observed in human muscular tissues and in myotubes obtained in vitro from RD rhabdomyosarcoma cells. We observed that RD myotubes incorporated a foreign gene, lacZ, by endocytosis but that expression of the encoded coding product, β-galactosidase, was strongly inhibited. Overexpression of DNase X inhibited endocytosis-mediated gene transfer, whereas knockdown of DNase X with small interfering RNA had the opposite effect. These results reveal that DNase X provides a cell surface barrier to endocytosis-mediated gene transfer.

DNase I is a secretory endonuclease that hydrolyzes DNA to 3Ј-hydroxyl oligonucleotides in the presence of divalent metal ions, such as Ca 2ϩ , Mg 2ϩ , and Mn 2ϩ (1)(2)(3). Its physiological significance in body fluids has long been unclear; however, a recent study demonstrates that targeted disruption of murine dnase1 increases the generation of anti-DNA antibodies and the development of a systemic lupus erythematosus-like syndrome (4). These findings indicate that DNase I plays an important role in eliminating extracellular DNA that can cause autoimmune diseases in animals.
In early investigations, most divalent cation-dependent DNase activities were regarded as being carried out by DNase I. However, more recent studies revealed the existence of several DNase I-like DNases, DNase X/Xib, DNase ␥/DNAS1L3, and DNAS1L2 (5). DNASEX is located at q28 of the human X chromosome and was the first gene to be found that encoded a protein homologous to DNase I (6 -8).
DNase X has an extra hydrophobic stretch at its C terminus, which is regarded as its most outstanding structural feature (5). This hydrophobic domain is conserved among mammalian DNase X proteins, suggesting that it has functional importance (9) and that DNase X might play a unique physiological role.
In the current study, we have generated a monoclonal antibody (mAb) 3 specific for human DNase X and used it to characterize the molecule. We have presented evidence that DNase X is a glycosylphosphatidylinositol (GPI)-anchored membrane DNase located on the cell surface and on early endocytic vesicles. Furthermore, we have demonstrated that DNase X hydrolyzes endocytosed extracellular DNA, thereby protecting cells from invasion by foreign genes.

EXPERIMENTAL PROCEDURES
Cell Culture-HeLa S3, COS-7, and CHO-K1 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 100 units/ml penicillin, and 0.1 mg/ml streptomycin. Human embryonic rhabdomyosarcoma RD cells, obtained from the Health Science Research Resources Bank of Japan, were maintained in growth medium (GM) (Dulbecco's modified Eagle's medium supplemented with 20% fetal calf serum, 100 units/ml penicillin, and mg/ml streptomycin). To induce myogenic differentiation, subconfluent cultures grown on collagen type I-coated dishes (Iwaki Scitech) were shifted to differentiation medium (DM) (Dulbecco's modified Eagle's medium supplemented with 2% horse serum, 100 units/ml penicillin, and 0.1 mg/ml streptomycin) and cultured for the indicated number of days with daily medium replacement.
The efficiency of endocytosis-mediated gene transfer was evaluated by measuring the ␤-galactosidase activity expressed from placZ. Cells in 60-mm culture dishes were loaded with 10 g of placZ precomplexed with Transome IV. Control experiments were also performed using an empty vector to subtract the basal ␤-galactosidase activity. After 24 h, cells were harvested and ␤-galactosidase activity was assayed using the Beta-Gro assay system (Promega) according to the manufacturer's recommendations. Total protein in cell lysates was determined with protein assay solution (Bio-Rad). Endocytotic activities were determined by monitoring the uptake of Alexa 594-dextran (D22913, Molecular Probes) or FITC-lacZ-Transome IV complex. Cells maintained on coverslips in 60-mm culture dishes were cultured for 60 min in the presence of 1 mg/ml Alexa 594-dextran or 10 g of FITC-lacZ precomplexed with Transome IV. Cells were washed twice with Dulbecco's modified phosphate-buffered saline (DPBS), fixed with 3.7% formaldehyde at room temperature for 10 min, washed three times in DPBS, and observed with a fluorescence microscope after counterstaining the nuclei with 100 nM Hoechst 33342.
Antibodies-A synthetic peptide (RSLLHTAAAFDFPTS-FQL) corresponding to residues 237-254 of human DNase X was conjugated to thyroglobulin through its N-terminal Cys and used as an antigen. BDF1 mice were immunized subcutaneously with 50 g of peptide in complete Freund's adjuvant and then injected with 50 g of peptide in incomplete Freund's adjuvant after 2, 3, and 4 weeks. A final 50-g peptide booster was delivered by intravenous injection. Four days later, spleen cells were isolated from the mice and fused to X63-Ag8 -653 mouse myeloma cells (10) using polyethylene glycol. Growing hybridomas were selected in hypoxanthine/aminopterin/thymidine medium, and culture supernatants were screened using 96-well enzyme-linked immunosorbent assay plates coated with the peptide. The resulting hybridomas were cloned twice by limited dilution, and a hybridoma producing monoclonal antibody was established and designated clone 1B1.
We used the same method to obtain the hybridoma clone 10A2-producing monoclonal antibody against DNase I, which recognizes the synthetic peptide GAVVPDSALPFNFQAAYG corresponding to residues 245-262 of human DNase I. The subclass of the monoclonal antibodies was determined using mouse monoclonal antibody-isotyping reagents (IBL), and both 1B1 and 10A2 were found to be IgG1 . Mouse anti-SERCA-1 mAb (VE12 1 G9) was purchased from Lab Vision, anti-GAPDH (MAB374) from Chemicon, and anti-Myc tag (R950-25) from Invitrogen. The goat polyclonal antibody against GM130 (P-20) was obtained from Santa Cruz Biotechnology. The rabbit polyclonal antibody against mouse Fas (341289) was purchased from Calbiochem, anti-EEA1 (PA1-063) was from Affinity BioReagents, anti-LAMP-1 (H-228) from Santa Cruz Biotechnology, and anti-calreticulin (C4606) and anti-pan-cadherin (C3678) were from Sigma.
Immunohistochemistry-Sections prepared from formalinfixed, paraffin-embedded normal human tissues were obtained from BioChain. Immunohistochemistry was performed using a cell and tissue staining kit (HRP-DAB, R & D Systems). In brief, the tissue slides, pretreated with an antigen retrieval dewax solution (BioChain), were incubated with anti-DNase X mAb at a concentration of 3 g/ml overnight at 4°C. The sections were counterstained with Contrast BLUE solution (KPL), and images were captured using a CK40 light microscope system (Olympus). The specificity of each immunostaining procedure was confirmed by a blocking experiment performed in the presence of an ϳ10-fold excess (200 ng/ml) of the immunizing DNase X peptide.
Indirect Immunofluorescence-Immunofluorescence analyses were performed as described previously with some modifications (9). In brief, cells grown on a coverslip were fixed with 3.5% formaldehyde and permeabilized by incubating in a Cytonin solution (R & D Systems) at room temperature for 30 min. When detecting cell surface targets, the incubation with Cytonin was omitted. The first antibody reactions were performed in a humid sealed chamber overnight at 4°C. The working concentrations or dilution factors of the antibodies were for DNase X, 3 g/ml; DNase I, 1 g/ml; calreticulin, 1:2000; SERCA-1, 1:50; GM130, 1:50; Fas, 1:100; pan-cadherin, 1:400; EEA1, 1:1000; and LAMP-1, 1:100. The cells were then washed in DPBS and incubated with the following species-specific secondary antibodies (1:2000 dilution) for 30 min at room temper-ature: Alexa 488 chicken anti-mouse IgG, Alexa 594 donkey anti-goat IgG, and Alexa 594 goat anti-rabbit IgG (Molecular Probes). Double staining with mAbs for DNase X and SERCA-1 was performed using a Zenon Tricolor Mouse IgG1 labeling kit number 2 (Molecular Probes) according to the manufacturer's protocol; anti-DNase X and anti-SERCA-1 mAbs were prelabeled with Alexa 488 and 594, respectively, and the antibody reaction was performed for 90 min at room temperature. The cells were washed in DPBS, counterstained with 100 nM Hoechst 33342, and observed with a fluorescence microscope (BX60, Olympus) or confocal laser scanning microscope (TCS SP2, Leica).
PI-PLC Treatment-Cells grown on coverslips were rinsed twice with DPBS and incubated in Leibovidz culture medium containing 5 units/ml Bacillus thuringiensis phosphatidylinositol-specific phospholipase C (PI-PLC) (Sigma) at 37°C for 30 min. The cells were rinsed twice with DPBS and fixed with 3.7% formaldehyde. Cell surface DNase X, mFas⌬C, and cadherin were detected by indirect immunofluorescence as described above. The attached cells were rinsed with DPBS, collected in a 1.5-ml tube, and incubated in 100 l of 5 units/ml PI-PLC at 37°C for 20 min. They were then centrifuged at 2000 ϫ g for 10 min, and the supernatant was collected in a separate tube (Sup fraction). The cell pellet was washed twice with DPBS by repeated suspension and centrifugation and resuspended in 500 l of Leibovidz medium (cell fraction). Identical aliquots of the two fractions were subjected to Western blotting.
TUNEL Assay-Cells grown on coverslips were cultured for 30 min in the presence of placZ-Transome IV complex as described above. They were washed twice with DPBS, fixed with 3.7% formaldehyde in DPBS at room temperature for 10 min, and the 3Ј-OH ends produced in the lacZ plasmids were detected by TUNEL using a FlowTACS in situ TUNEL detection Kit (R & D Systems). The overall assay was performed according to the manufacturer's protocol with one exception, FITC-conjugated streptavidin was replaced by Alexa 488-conjugated streptavidin (Molecular Probes). The cells were washed three times in DPBS, the endocytosed DNA counterstained with 10 M Hoechst 33342 in DPBS for 10 min at room temperature, and then observed with a fluorescence microscope (BX60, Olympus).
Capture of DNA Images in Fluorescence Microscopy-Nuclear and/or endocytosed DNAs were counterstained with Hoechst 33342 (a cell-permeable fluorescent dye) as described in each section. Observations of the DNA images were performed as follows unless otherwise noted. Images of nuclear DNA were taken through two neutral density filters (U-ND6 and U-ND25, Olympus) because of their strong fluorescence, whereas no neutral density filters were used for observing endocytosed DNA.
RNA Interference-Stealth RNA interference for human DNase X and a negative control RNA (ncRNA) duplex were obtained from Invitrogen. The small interfering RNA (siRNA) was directed against the target sequence 5Ј-CCTGCTTCGAG-AACTCAATCGATTT-3Ј. Subconfluent cultures of RD cells, seeded in 60-mm dishes, were shifted to DM and transfected with 120 pmol of siRNA using Lipofectamine 2000 transfection reagent (Invitrogen). The resulting culture was maintained in DM for 4 days with daily medium replacement. We observed no induction of 2Ј,5Ј-oligoadenylate synthetase 2 or signal transducer and activator of transcription 1b mRNA in siRNA-transfected cells, indicating the absence of detectable interferon responses (data not shown).

DNase X Is Expressed during Myogenic Differentiation of RD
Cells-To examine DNase X expression at the protein level, we performed an immunohistochemical analysis of several human tissues using the newly developed anti-DNase X mAb 1B1 (supplemental Fig. S1). As shown in Fig. 1A, both cardiac muscle cells and skeletal muscle fibers were highly immunoreactive for DNase X, whereas there was little positive staining in the brain. These results are consistent with the gene expression profile determined previously (5-9) and confirm the expression of the DNase X protein in muscle cells.
The human rhabdomyosarcoma cell line RD forms myotubes when it is grown under low mitogen conditions (Fig. 1B). To monitor the changes in DNase X expression associated with myogenic differentiation, we performed Western blots and showed that DNase X expression began to increase on day 3 after induction of differentiation, reaching a maximum on day 5 (Fig. 1C, top panel). The induction of a muscle marker protein, sarcoplasmic reticulum Ca 2ϩ -ATPase 1 (SERCA-1), confirmed the occurrence of myogenic differentiation (Fig. 1C, middle  panel).
DNase X Is Targeted to the Secretory Pathway and Delivered to the Cell Surface-We analyzed the intracellular distribution of DNase X by confocal laser scanning microscopy (Fig. 1D). Colocalization of DNase X with calreticulin and SERCA-1 revealed its association with the endoplasmic reticulum and sarcoplasmic reticulum, respectively. DNase X immunofluorescence also paralleled that of GM130, a marker of the Golgi apparatus. These results indicate that newly synthesized DNase X is directed to the secretory pathway of the cell. To determine its final destination, we immunostained non-permeabilized cells and showed that the surface of multinucleated RD myotubes was strongly positive (Fig. 1E), whereas the surface of undifferentiated cells was only weakly stained (Fig. 1E, white arrowheads). The membrane expression of DNase X was also confirmed by confocal microscopy (Fig. 1F). These results demonstrate that newly synthesized DNase X is targeted to the secretory pathway and delivered to the cell surface in RD myotubes.
The C-terminal Hydrophobic Domain Is Essential for Cell Surface Localization of DNase X-To determine whether membrane association is a specific hallmark of DNase X, COS-7 cells were transfected with expression vectors for DNase X and DNase I, and the membrane distribution of each DNase was analyzed by immunostaining. To locate DNase I, we generated an anti-DNase I mAb, 10A2. A mutant form of the murine Fas antigen (mFas⌬C), which is expressed on the cell surface as a transmembrane protein, was co-expressed with the DNases, and its membrane expression was monitored as a marker of transfection (Fig. 2, A and B, right panels).
Indirect immunofluorescence revealed cell surface localization of the ectopic DNase X (Fig. 2A, upper panels). In contrast, there was no sign of DNase I membrane expression ( Fig. 2A,  lower panels).
DNase X is unique among family members in having a conserved hydrophobic domain at its C terminus; this suggested to us that the hydrophobic domain might play a role in the membrane attachment of DNase X. To evaluate this possibility, we constructed two mutant DNase derivatives and observed their cell surface localization, DNase X with a deleted hydrophobic domain (DNase X⌬C) and DNase I C-terminally fused to the hydrophobic domain of DNase X (DNase I-HD). As shown in Fig. 2B, DNase X⌬C failed to attach to the plasma membrane, whereas DNase I-HD acquired a membrane location. These results show that cell surface localization is a specific feature of DNase X and that the C-terminal hydrophobic domain is necessary and sufficient for this localization.
Membrane Binding of DNase X Is Mediated by a GPI Anchor-Eukaryotic proteins become membrane-anchored by several mechanisms, through being transmembrane proteins, through protein-protein interaction, or through binding to a GPI anchor (11,12). Significantly, attachment to a GPI anchor requires a hydrophobic signal sequence located at the C termini of proteins (12). Given that DNase X has a C-terminal hydrophobic domain, we hypothesized that GPI anchoring might account for its membrane association. To evaluate this possibility, we tested the effect of phosphoinositide-specific phospholipase C (PI-PLC), an enzyme that releases GPI-anchored proteins, on DNase X membrane association. Indirect immunofluorescence revealed that PI-PLC treatment greatly reduced the signal intensity of an exogenous DNase X without affecting that of the co-transfected transmembrane protein mFas⌬C (Fig. 3A).
The PI-PLC-mediated release of DNase X was further analyzed by Western blotting. As shown in Fig.  3B, DNase X (but not mFas⌬C) was released into the supernatant by PI-PLC treatment. We obtained essentially the same results with the endogenous DNase X of RD myotubes (Fig. 3C); PI-PLC effected the release of DNase X without solubilizing cadherin, a membrane-integrated protein responsible for stable contact between cells (13). It is of note that an anti-Myc antibody failed to detect the ectopic DNase X released by PI-PLC treatment (data not shown), suggesting that the C-terminal hydrophobic domain and the Myc tag had been replaced by the GPI anchor. Based on these results, we conclude that membrane binding of DNase X is mediated by a GPI anchor and that the C-terminal hydrophobic domain serves as a signal sequence for GPI attachment.
Endocytosis-mediated Gene Transfer Is Impaired in RD Myotubes-The endocytic activity of eukaryotic cells internalizes various cell surface and extracellular molecules, including foreign genes. Endocytosed genes are usually delivered to lysosomes and degraded by lysosomal acid DNases; however, certain cationic peptides are known to promote the escape of incorporated genes from endosomal degradation, thereby permitting transfer of the incorporated genetic information into the cell (14 -16). As an initial approach to unveiling the function of DNase X, we observed endocytosis-mediated gene transfer in RD myotubes that express DNase X on the cell surface at high levels. We employed an artificial ␣-helical cationic peptide to evaluate the efficiency of gene transfer by endocytosis (17,18). RD myotubes incorporated a fluorophore-conju- The resulting cells were harvested, and DNase X and SERCA-1 expression was analyzed by Western blotting. GAPDH signals are shown as an internal control. D, subcellular localization of DNase X. Differentiated RD cells, cultured in DM for 6 days, were fixed, permeabilized, and subjected to indirect immunofluorescence for DNase X (green) and the indicated marker proteins (red). Fluorescent images were observed by confocal laser scanning microscopy. Merged images are shown in the right panels. The yellow and blue areas show co-localization of the two proteins and nuclei, respectively. Scale bar represents 50 m. E and F, cell surface localization of DNase X. DNase X membrane expression in 6-day-differentiated RD cells (non-permeabilized) was analyzed by indirect immunofluorescence. Images were observed by fluorescent microscopy (E) or confocal laser scanning microscopy (F). White arrowheads indicate undifferentiated cells. Scale bars represent 50 m. gated transgene, FITC-lacZ, more effectively than undifferentiated cells (Fig. 4A). However, expression of the encoded product, ␤-galactosidase, in the differentiated cells was Ͻ10% of that observed in the undifferentiated cells (Fig. 4B).
To explain this finding, we hypothesized that the high level of membrane DNase X expression was responsible for the poor transgene expression in differentiated cells. To test this idea, we examined the association of DNase X with early and late endocytic vesicles in RD myotubes. As shown in Fig. 4C, DNase X immunofluorescence co-localized with that of the early endosomal marker EEA1, whereas there was no detectable overlap with that of LAMP-1, a marker of mature lysosomes. We next performed terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) to detect 3Ј-OH nick ends on the endocytosed plasmids. This method was originally developed to detect apoptotic DNA fragmentation in situ (19). As expected, we found that the transgenes incorporated by RD myotubes were positive for TUNEL (Fig. 4D). In view of the fact that DNase X catalyzes a 3Ј-OH/5Ј-P type of DNA hydrolysis (5), these results suggest that it introduces multiple strand-  breaks into the DNA, probably both at the cell surface and within early endocytic vesicles.
Overexpression of the Membrane-bound Form of DNase X Suppresses Endocytosis-mediated Gene Transfer-To further address the role of DNase X, we introduced several mutant forms of DNase X into cells possessing a low level of endogenous DNase X activity and observed their effects on endocytosis-mediated gene transfer. Indirect immunofluorescence revealed ectopic DNase X expression on the surface of both HeLa S3 and undifferentiated RD cells (Fig. 5A, DNase X), whereas a deletion mutant lacking the C-terminal hydrophobic domain displayed no specific membrane binding (Fig. 5A, DNase X⌬C). These cultures were then loaded with the placZcationic peptide complex, and expression of the reporter gene was measured. As shown in Fig. 5B, overexpression of DNase X inhibited ␤-galactosidase expression by Ͼ70% in both HeLa S3 and undifferentiated RD cells, and inhibition was greatly diminished by the C-terminal deletion (Fig. 5B, ⌬C).
To establish the importance of DNase X activity, we expressed an inactive mutant in the cells and measured endocytosis-mediated gene transfer levels. DNase X H130A, in which one of the two catalytic His residues is replaced by Ala, was expressed normally on the cell surface (Fig. 5A, DNase X H130A); however, it did not inhibit transgene expression (Fig.  5B). The DNA strand breaks produced by the ectopic forms of DNase X were visualized using the TUNEL technique. As shown in Fig. 5C, wild type DNase X lacZ plasmids incorporated by HeLa S3 cells were TUNEL-positive, whereas lacZ plasmids ectopically expressing inactive DNase X were TUNEL-negative.
SiRNA-mediated Knockdown of DNase X Reverses Inhibition of Endocytosis-mediated Gene Transfer in RD Myotubes-To establish whether endogenous DNase X acts as an inhibitor of gene transfer, we asked whether down-regulation of endogenous DNase X would reverse the impaired endocytosis-mediated gene transfer in RD myotubes. Fig. 6A shows that the intensity of the DNase X band was greatly diminished in cells treated with an siRNA specific for human DNase X, whereas no such down-regulation was observed with an ncRNA duplex. The suppression of DNase X did not affect SERCA-1 expression (Fig. 6A) or myotube formation (Fig. 6C), indicating that DNase X per se is not essential for myogenic differentiation of RD cells.
We next cultured siRNA-treated myotubes in the presence of the placZ-peptide complex and examined the effect of DNase X suppression on endocytosis-mediated gene transfer (Fig. 6B). The siRNA-mediated knockdown of DNase X stimu-  's t test). C, localization of DNase X in endocytic vesicles. RD cells, cultured in DM for 6 days, were fixed, permeabilized, and double-stained with antibodies for DNase X (green) and the indicated marker proteins (red). Fluorescent images were obtained by confocal laser scanning microscopy. The yellow area in the merged images, also marked by white arrowheads, shows co-localization of the two proteins. Scale bar represents 10 m. D, generation of 3Ј-OH nick ends in placZ incorporated by RD myotubes. placZ-peptide complex (ϩplacZ) was added to 4-day-differentiated RD cells, and they were incubated for 1 h, washed, fixed, and subjected to TUNEL assay. After counterstaining with 100 nM Hoechst dye, 3Ј-OH ends of the placZ plasmid (green) and the nuclei (red) were observed by fluorescence microscopy. A control experiment performed using the cationic peptide alone (ϪlacZ) shows background fluorescence. Scale bar represents 50 m. lated transgene expression Ͼ4-fold in differentiated RD cells. Moreover, TUNEL-positive strand breaks, which were apparent in the placZ endocytosed by control cells, were scarcely visible in the plasmids incorporated by siRNAtreated myotubes (Fig. 6C). These results indicate that GPIanchored membrane DNase X catalyzes the degradation of incorporated foreign genes, thereby acting as a suppressor of endocytosis-mediated gene transfer.

Identification of DNase X as a GPI-anchored Membrane
Protein-Our results demonstrate that DNase X is expressed on the cell surface and on early endocytic vesicles as a GPIanchored membrane protein. Surprisingly, the surface of muscle cells is also covered with this GPI-anchored endonuclease.
A diverse set of eukaryotic cell surface proteins is anchored to membranes by covalent linkage to GPI (11,12). However, to the best of our knowledge, this is the first demonstration of a eukaryotic GPI-anchored ecto-DNase. Posttranslational modification with a GPI anchor occurs at the C terminus (-site) of proteins after proteolytic cleavage of their C-terminal hydrophobic signal sequences (12,20,21). We show here that the C-terminal hydrophobic domain of DNase X is essential for its membrane binding and that membrane DNase X is solubilized by PI-PLC.
In a previous study, we observed that ectopic DNase X was found exclusively in the endoplasmic reticulum, as determined by immunofluorescence of its C-terminal Myc tag (9). This contradictory observation is explained by the fact that the GPI anchor is attached within the endoplasmic reticulum lumen, and at the same time, the protein to be anchored is cleaved of its C-terminal propeptide. Thus, the membrane-bound form of DNase X loses its C-terminal tag.
Endocytosis-mediated Gene Transfer and DNase X-We observed that endocytosis-mediated gene transfer is strongly suppressed in cells expressing DNase X at high levels and that DNase X activity is responsible for TUNEL-positive strand breaks in the endocytosed DNA. Furthermore, siRNA-mediated knockdown of DNase X reverses the impairment of gene transfer in the RD myotubes. Our results thus demonstrate that DNase X protects cells from endocytosis-mediated gene transfer by inactivating the incorporated foreign genes.
GPI-anchored proteins are associated with lipid rafts, which are detergent-insoluble membrane microdomains enriched in cholesterol and glycosphingolipids (11,12). The biological significance of lipid rafts is not yet fully understood, but they have been implicated in diverse cellular activities, including signal transduction, cholesterol trafficking and endocytosis (11,12). To date, at least three types of endocytic pathway have been identified in eukaryotic cells, the classical clathrin-mediated pathway, caveolae/raft-dependent endocytosis, and constitutive fluid phase endocytosis, also known as pinocytosis (22, 23). The endocytic pathways by which cells take up foreign genes remain unclear. However, the caveolae/raft-dependent and the fluid phase pathways are the most likely candidates, because both mediate nonspecific internalization of micro-and macrosubstances from the extracellular environment (24,25). Importantly, GPI-anchored proteins are known to be internalized  t test). C, generation of 3Ј-OH nick ends in endocytosed lacZ plasmids. HeLa S3 cells, transfected with DNase X or DNase X H130A, were cultured in the presence of placZ peptide complex for 1 h, fixed, and subjected to TUNEL assay. After counterstaining with 10 M Hoechst dye, the 3Ј-OH ends of placZ (upper panels) and placZ plus nuclei (lower panels) were observed by fluorescence microscopy. The DNA images were taken through a neutral density filter (U-ND25). White arrowheads indicate placZ incorporated by the cells. Scale bar represents 50 m.
into recycling and early endosomes via the caveolae/raft-dependent and the fluid phase endocytic pathways (26 -28). These observations suggest a possible reason for DNase X to be a GPI-anchored protein; the GPI anchor directs it to those pathways that support nonspecific endocytosis, thereby making it possible to intercept genetic invaders not only at the cell surface but also within endocytic vesicles. This idea is supported by the observation that the DNA-LL-37 complex is internalized into cells by a raft-dependent endocytic activity (16).
Physiological Significance of the DNase X Barrier-It is a known fact that animal cells have the ability to take up foreign genes and express their encoded products, and this has been exploited by the development of transfection procedures suitable for experimental and clinical applications. However, the occurrence of spontaneous gene transfer is apparently harmful to animals, because the frequent integration and unregulated expression of foreign genes disturbs cellular activity and increases the risk of malignant transformation. Thus, suppression of foreign gene transfer by DNase X is probably important in maintaining healthy cell communities in vivo.
A basal level of expression of DNase X is seen in several mammalian tissues (5,9). We have performed immunohistochemistry on several non-muscular tissues and observed that epithelial cells are generally positive for DNase X (data not shown). This shows that the DNase X membrane barrier is not restricted to muscle cells.
A Possible Clinical Application of Breaking through the DNase X Barrier-Skeletal muscle is an attractive target for therapeutic gene transfer for several reasons; it is the most abundant tissue in the human body, it is accessible to many systems of gene delivery, and turnover of its cells is relatively slow so that the effects of transgene expression are long lasting (29). In addition, skeletal muscle is rare among tissues in incorporating naked DNA (30).
Naked DNA transfer is a method of therapeutic gene delivery in which plasmid vectors are simply injected into a target tissue (31,32). Because of its safety, simplicity, and lack of size limitation, naked DNA transfer is now receiving attention as an attractive gene delivery method for future gene therapies. Naked gene transfer has been tested in mice and shown to deliver genes into skeletal muscle (30); however, it is still of little practical use in humans because of its low efficiency (29). Although the factors limiting naked DNA transfer in vivo are not fully understood, our data strongly suggest that DNase X is a major barrier. Based on the results of this study, we propose that artificial control of DNase X activity using a specific inhibitor may provide a way to improve the results of naked gene transfer in future gene therapies.