52 kd Ro/SS-A LOCALIZES TO PUNCTATE STRUCTURES IN THE CYTOPLASM OF EPITHELIAL CELLS Łodź, Poland

Transcription

1 CELLULAR & MOLECULAR BIOLOGY LETTERS Volume 8, (2003) pp Received 4 November 2002 Accepted 31 January kd Ro/SS-A LOCALIZES TO PUNCTATE STRUCTURES IN THE CYTOPLASM OF EPITHELIAL CELLS DANIEL P. McCAULIFFE 1 and ANNA WOŹNIACKA 2 * 1 Department of Dermatology, University of North Carolina at Chapel Hill, USA 2 Department of Dermatology, Medical University of Łódź, Krzemieniecka 5, Łodź, Poland Abstract: Autoantibodies directed against 52 kd and 60 kd Ro/SS-A are frequently found in the sera of patients with lupus erythematosus and Sjögren s syndrome-related disorders. Their location in the cell is subject to continuous debate in literature. It has been postulated that 52 kd Ro (52 Ro) co-localizes with the 60 kd Ro autoantigen in the nucleus, while others demonstrated that 52 Ro is primarily cytoplasmic. In order to resolve this controversy, 52 Ro protein was tagged with green fluorescence protein, overexpressed in A431 keratynocytes, and its location determined using fluorescence confocal microscopy. The intracellular location of the fusion protein was revealed via GFP autofluorescene and indirect immunofluorescence microscopy, using purified anti-52 Ro antibodies. The cellular locations of native 52 Ro in normal human keratinocytes, and in human A431 keratinocyte and HepG2 hepatocyte cell lines were similarly determined by utilizing 2 human anti-52 Ro antibodies purified from two different non-overlapping fragments of recombinant 52 Ro. In addition, colocalization of 52 Ro with mitochondria, lysosomes and endosomes was evaluated. It was found that both the 52 Ro-GFP fusion protein and the native 52 Ro localize in discrete cytoplasmic punctate structures separately from the mitochondria, lysosomes and endosomes. Furthermore, human autoantibodies that are reactive with denaturation-sensitive epitopes on 52 Ro recognize these cytoplasmic punctate structures, whereas antibodies directed against denaturation-resistant 52 Ro epitopes do not. This explains why the previously used antibody against denaturation-resistant 52 Ro epitopes failed to detect the protein in such punctate structures. Key Words: Autoantibodies, Green Fluorescence Protein, Immunofluorescence, Ro, SS-A * Corresponding author, Tel/Fax: (+48 42) , wozniacka@bmp.net.pl Abbreviations used: GFP - green fluorescent protein; ATCC - American Type Culture Collection

2 134 CELL. MOL. BIOL. LETT. Vol. 8. No INTRODUCTION Autoantibodies directed against the 52 kd and 60 kd Ro/SS-A autoantigens are frequently found in the sera of patients with lupus erythematosus and Sjögren s syndrome-related disorders [1]. The cellular functions of these autoantigens remain unknown and the precise intracellular location of the 52 kd Ro/SS-A autoantigen (Ro) remains a matter of debate. Initially, 52 kd Ro (52 Ro) was described as co-localizing with the 60 kd Ro autoantigen, primarily in the nucleus [2]. Others reported similar findings [3-6]. However, more recent findings have challenged this notion by demonstrating that 52 Ro is primarily cytoplasmic [7-9]. We now demonstrate that 52 Ro localizes to discrete punctate structures in the cytoplasm, as well as to a faint, diffuse cytoplasmic and nuclear distribution. MATERIALS AND METHODS Cell culture Normal human keratinocytes were harvested and cultured as previously described [8]. HepG2 cells (a human hepatocellular carcinoma cell line, American Type Culture Collection (ATCC) HB-8065, Rockville, MD) were cultured in Eagle s minimal essential media with 10% fetal bovine serum according to the ATCC recommended guidelines. A431 cells (a human keratinocyte cell line, ATCC CRL-1555) were cultured in Dulbecco s modified Eagle s medium with 4.5 g/l glucose, and 10% fetal bovine serum according to the ATCC recommended guidelines. GFP transfection construct A plasmid construct was made that encodes a GFP fused to the carboxy-terminal end of the 52 Ro protein (pegfp, Clontech Laboratories, Palo Alto, CA). This construct was transfected into A431 using Effectene transfection reagent according to the manufacturer s recommendations (Qiagen, Valencia, CA). Clones of stable tranfectants were established by culturing the transfected A431 cells in the presence of the neomycin/g18 selection marker (Gibco BRL Life Technologies, Grand Island, NY). Immunofluorescence technique Cells were cultured in LabTech chamber slides (Fisher Scientific, Atlanta, GA), fixed and examined via epiluminescence microscopy as previously described [8]. The cell-permeant compartmental markers used (Molecular Probes, Eugene, OR) included LysoTracker Red to localize lysosomes and tetramethylrhodamine methylester to localize mitochondria. Anti-52 Ro antibodies were purified from human autoimmune sera as previously described [8]. B3/25 anti-transferrin receptor antibodies (Boehringer-Mannheim, Indianapolis, IN) were used to locate endosomes. Secondary antibodies included 488 nm (green) and 594 nm

3 CELLULAR & MOLECULAR BIOLOGY LETTERS 135 (red) Alexa-conjugated antibodies (Molecular Probes, Eugene, Oregon). Some transfected A431 cells over-expressing 52 Ro-GFP fusion protein were examined live, without fixation. Antibody purification Recombinant bacterially produced 52 Ro protein and protein fragments were purified as preciously described [10]. Full length 52 Ro recombinant protein was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS- PAGE), transferred to polyvinylidine difluoride paper and blocked with milk protein according to the established methods [8]. A monospecific anti-52 Ro autoimmune serum was incubated with this paper and the antibodies stripped off for use in immunofluorescence studies as previously described [8]. Antibodies were similarly harvested from the 52 Ro zinc finger fragment, with the exception that the protein was directly coated on the paper without being first subjected to SDS-PAGE. RESULTS In previous immunofluorescent localization studies with anti-52 Ro antibodies purified from denatured recombinant 52 Ro protein, we demonstrated a predominantly diffuse weak cytoplasmic fluorescence, and weaker nuclear fluorescence [8]. To confirm these earlier findings, and to better visualize the cytoplasmic location of 52 Ro, a construct was made that over-expresses 52 Ro as a green fluorescent protein (GFP) fusion protein in A431 cells, a human keratinocyte cell line. The transfected cells showed intense punctate cytoplasmic fluorescence that differed from the diffuse cellular pattern generated by GFP alone (Fig. 1). Flourescence microscopy of both fixed and live transfected cells produced the same cytoplasmic punctate pattern, indicating that it did not represent a fixation artifact (data not shown). To verify that the punctate cytoplasmic fluorescence was emanating from the 52 Ro-GFP fusion protein, immunofluorescence localization was performed with human anti-52 Ro antibodies that were purified from denatured recombinant 52 Ro (Fig. 2.). Fig. 1. Localization of 52 Ro in cultured human A431 keratinocytes. Fluorescence signal from GFP (A) and 52 Ro-GFP fusion protein (B) in transiently transfected live (non-fixed) A431 cells.

4 136 CELL. MOL. BIOL. LETT. Vol. 8. No Fig. 2. Co-localization of the autofluorescence in transfected A431 cells from the 52 Ro-GFP protein (A) and indirect immunofluorescence generated with human anti-52 Ro antibodies purified from denatured 52 Ro, and Alexa Red conjugated anti-human antibodies (B). Panels A and B represent the same microscopic field that was examined with different colored filters to separately isolate the GFP-produced green fluorescence from antibody-produced red fluorescence. Fig. 3. Indirect immunofluorescent localization of 52 Ro in cultured normal human keratinocytes (A and D), HepG2 cells (B and E) and non-transfected A431 cells (C and F). Localization with antibodies purified from the non-denatured 52 Ro zinc finger fragment (A, B, C), and antibodies from denatured full-length recombinant 52 Ro (D, E and F). The cytoplasmic punctate structures are revealed in all three epithelial cell types with the zinc finger fragment antibodies, but not with the antibodies purified from denatured 52 Ro. Fig. 4. Co-localization of 52 Ro-GFP and 52 Ro zinc finger fragment antibodies. Transfected A431 cells expressing 52 Ro-GFP fusion protein were fixed and then subjected to immunofluorescence microscopy utilizing purified human anti-52 Ro zinc finger fragment antibodies and Alexa Red conjugated anti-human antibodies. Panels A and B represent the same microscopic field that was examined with different colored filters to separately isolate the GFPproduced green autofluorescence (A) from the zinc finger fragment antibody-produced red fluorescence (B). The photographs were converted to black and white, so the color distinction is no longer apparent.

5 CELLULAR & MOLECULAR BIOLOGY LETTERS 137 Fig. 5. Double immunofluorescence localization of 52 Ro in non-transfected A431 cells with antizinc finger fragment antibodies (A and C), and markers for mitochondria (B) and lysosomes (D). Panels A and B represent the same field, as do panels C and D. The same microscopic field was examined with different colored filters to separately isolate green and red colored fluorescence signals. The photographs were converted to black and white, so the color distinction is no longer apparent. The fluorescent patterns produced by the mitochondrial and lysosomal markers are clearly different to the patterns produced by the 52 Ro zing finger fragment antibodies, indicating that the 52 Ro containing cytoplasmic punctate structures do not represent mitochondria or lysosomes. Fig. 6. Double immunofluorescence localization of endosomes and 52 Ro in non-transfected A431 cells. Endosomes were localized with murine anti-transferrin receptor antibodies and Alexa Red conjugated anti-murine antibodies (red). The 52 Ro molecules were localized with anti-zinc finger fragment antibodies and Alexa Green conjugated anti-human antibodies (green). The green and red colors are separate indicating that 52 Ro localizes to non-endosomal cytoplasmic punctate structures. The white horizontal bar represents ~10 μm. Our previous localization work, which utilized these same anti-52 Ro antibodies, had revealed weak fluorescence signals throughout the cytoplasm and the nuclei of cells, but failed to reveal the cytoplasmic punctate structures [8]. It is possible that the punctate structures are specious, related to over-expression of 52 Ro, and/or related to abnormal localization due to the GFP component, although GFP alone localized diffusely throughout the cell in both the nuclear and cytoplasmic compartments (Fig. 1A). It is also possible that earlier work failed to show these cytoplasmic structures due to masking of 52 Ro antibody binding sites in these locations. Perhaps these antibody binding sites become more accessible through over-expression of the protein, or by conformational changes

6 138 CELL. MOL. BIOL. LETT. Vol. 8. No induced by the GFP. To explore this later speculation, human anti-52 Ro antibodies were purified from a non-denatured 52 Ro zinc finger containing fragment that we had previously shown to contain a different 52 Ro epitope [10]. Via immunofluorecence microscopy, these non-denatured zinc finger epitope antibodies revealed a punctate cytoplasmic pattern similar to that of the 52 Ro- GFP fusion protein (Figs. 3 A, B & C), but unlike the diffuse cytoplasmic pattern generated with antibodies purified from denatured recombinant 52 Ro as seen in Figs. 3 D, E & F, and in previous work [8]. The zinc finger fragment antibodies bound to the same punctate structures revealed by GFP s autofluorescence in the 52 Ro-GFP transfected A431 cells (Fig. 4.). Double immunofluorescence localization studies revealed that the 52 Rocontaining punctate structures in non-transfected cells were not lysosomes, mitochondria or endosomes (Figs. 5 and 6). The 52 Ro-GFP fusion protein in transfected A431 cells also did not co-localize with these organelles (data not shown). The over-expressed 52 Ro-GFP fusion protein aggregated in larger punctate structures that at times appeared rod-like (Figs. 1, 2 and 4), compared to the smaller and rounder punctate structures seen in the non-transfected cells (Figs. 3 A-C, 5 A &C, and 6). DISCUSSION In previous studies, it has been demonstrated that anti-52 Ro autoimmune sera contain antibodies mainly directed at a denaturation-resistant epitope [11] near the leucine zipper motif in the mid portion of the molecule [12]. A minor native epitope near the amino terminal zinc finger region has also been described [10, 13]. We reported earlier that 52 Ro antibodies purified from denatured recombinant 52 Ro produce a weak diffuse cytoplasmic pattern by immunofluorescence microscopy [8]. Keech et al. reported a similar diffuse cytoplasmic pattern in transfected cells over-expressing 52 kd as determined by immunofluorescence microscopy using a monospecific anti-52 Ro autoimmune serum [7]. Contrary to these earlier findings of a diffuse cytoplasmic pattern, we have now found an intense punctate pattern of 52 Ro expression in transfected cells that over-express this protein with GFP fused to its carboxy-terminus. The GFP fusion to the carboxy-terminus in our construct is unlikely to have affected the localization of 52 Ro via interference with this end of the 52 Ro molecule as Pourmand et al. have recently demonstrated similar cytoplasmic punctate structures with GFP fused to the amino terminus of 52 Ro in HeLa cells [9]. Until now, however, it was uncertain whether this intense punctate pattern was an aberration from over-expression of 52 Ro in these cells or whether it might represent the normal distribution of 52 Ro that might go unrecognized by conventional immunofluorescence microscopic localization due to epitope masking or fixation effects. We have now demonstrated that there is an intense cytoplasmic punctate pattern in both transfected and non-transfected cells when

7 CELLULAR & MOLECULAR BIOLOGY LETTERS 139 utilizing human antibodies purified from a non-denatured 52 Ro zinc finger fragment. The punctate pattern is not discernable in non-transfected cells when using antibodies purified from denatured 52 Ro, suggesting that there may be epitope masking of these denatured Ro antibody binding sites in these punctate bodies. It is of interest to note that Pourmand et al. have demonstrated that GFP fused to the full length 52 Ro and to a fragment of the mid portion of 52 Ro both localize to cytoplasmic punctate structures, while the amino- and carboxy-terminal fragments of 52 Ro fused to GFP do not. This would suggest that the mid portion of 52 Ro is required for 52 Ro to localize to these structures. Previous studies have revealed a major denaturation resistant epitope near the leucine zipper motif in the mid portion of 52 Ro. It is possible that this epitope becomes masked when this portion of native 52 Ro interacts with other molecules which may serve to anchor it in these punctate structures. However, these antibodies do bind to the 52 Ro-GFP protein in the punctate structures (Fig. 2), possibly from epitope unmasking due to the larger amount of the 52 Ro expressed and/or structural changes related to the GFP portion of the fusion protein. Several proteins with a ring finger motif have been shown to have ubiquitin-related function. Di Donato et al. have found possible associations of 52 kd Ro with ubiquitin [14]. We conducted IF studies with anti-ubiquitin antibodies but did not detect similar punctate structures (personal unpublished observation). It is uncertain what these cytoplasmic punctate structures represent. The colocalization studies presented here show that these structures are separate from endosomes, lysosomes and mitochondria. Pourmand et al. have similarly failed to show that these structures are related to the Golgi apparatus or an intermediate compartment at the endoplasmic reticulum-golgi interface utilizing antimannosidase II and anti-58 monoclonal antibodies, respectively [9]. Preliminary immunoelectron microscopic localization studies have failed to reveal evidence that these structures are membrane bound (personal unpublished observation). Although we now know more about where 52 Ro resides in cells, we still know very little about what roles it plays in cellular processes. Further exploration into the nature of these cytoplasmic punctate structures should provide valuable insight into 52 Ro s cellular functions and the putative role that its autoantibodies play in the pathogenesis of disease [1]. Acknowledgements. This work was supported by the National Institutes of Health, grant AR (Dr. Woźniacka was supported by a Fulbright Scholarship) and grant from Medical University of Łódź No REFERENCES 1. McCauliffe, D.P. Cutaneous disease in adults associated with anti-ro/ss-a autoantibody production. Lupus 6 (1997)

8 140 CELL. MOL. BIOL. LETT. Vol. 8. No Ben-Chetrit, E., Chan, E.K.L., Sullivan, K.F. and Tan, E.M.A. 52-kD protein is a novel component of the SS-A/Ro antigenic particle. J. Exp. Med. 167 (1988) Slobbe, R.L., Pruijn, G. J.M., Damen, W.G.M., Van Der Kemp, M.J. and Van Venrooj, W.J. Detection and occurrence of the 60- and 52-kD Ro (SS- A) antigens and of autoantibodies against these proteins. Clin. Exp. Immunol. 86 (1991) Casciola-Rosen, L.A., Anhalt, G. and Rosen, A. Autoantigens targeted in systemic lupus erythematosus are clustered in two populations of surface structures on apoptotic keratinocytes. J. Exp. Med. 179 (1994) Peek, R., Van Venrooij, W.J., Simons, F. and Pruijn, G. The SS-A/SS-B autoantigenic complex: localization and assembly. Clin. Exp. Rheumatol. 12 (1994) S15- S Kelekar, A., Saitta, M.R. and Keene, J.D. Molecular composition of Ro small ribonucleoprotein complexes in human cells - Intracellular localization of the 60- and 52-kD proteins. J. Clin. Invest. 93 (1994) Keech, C.L., Gordon, T.P. and McCluskey, J. Cytoplasmic accumulation of the 52 kda Ro/SS-A nuclear autoantigen in transfected cell lines. J. Autoimmunity 8 (1995) Yell, J., Wang, A.L., Yin, H. and McCauliffe, D.P. Disparate locations of the 52 and 60 kd Ro/SS-A antigens in cultured human keratinocytes. J. Invest. Dermatol. 107 (1996) Pourmand, N., Blange, I., Ringertz, N. and Pettersson, I. Intracellular localisation of the Ro 52kD auto-antigen in He-La cells visualized with green fluorescent protein chimeras. Autoimmunity 28 (1998) McCauliffe, D.P., Yin, H., Wang, L. and Lucas, L. Autoimmune sera react with multiple epitopes on recombinant 52 and 60 kda Ro(SSA) proteins. J. Rheumatol. 21(1994) Itoh, Y. and Reichlin, M. Autoantibodies to the Ro/SSA antigen are conformation dependent. I.: Anti-60kD antibodies are mainly directed to the native protein; anti-52 kd antibodies are mainly directed to the denatured protein. Autoimmunity 14 (1992) K. 12. Dorner, T., Feist, E., Wagenmann, A.,. Kato, T., Yamamoto, K. Nishioka, G., Burmester, R. and Hiepe, F. Anti-52 kda Ro(SSA) autoantibodies in different autoimmune diseases preferentially recognize epitopes on the central region of the antigen. J. Rheumatol. 23 (1996) Miyagawa, S., Okada, N., Inagaki, Y., Kitano, Y., Ueki, H., Sakamoto, K. and Steinberg. M.L. SSA/Ro antigen expression in simian virus 40- transformed human keratinocytes. J. Invest. Dermatol. 90 (1988) Di Donato, F., Chan, E.K., Askanase, A..D., Miranda-Carus, M. and Buyon, J.P. Interaction between 52kDa Ssa/Ro and deubiquitinating enzyme UnpEL: a clue to function. Int. Biochem. Cell. Biol. 33 (2001)