S100A2 belongs to a family of at least 16 calmodulin-like proteins displaying substantial sequence divergence (up to 75%) (Schafer et al. 1995). S100 proteins contain two distinct EF-hand motifs flanking a central hinge region. While they share the EF-hand motif with calmodulin, they differ from calmodulin in several respects. They bind calcium only in the millimolar range, as opposed to the micromolar range observed for calmodulin (Heizmann & Cox, 1998). The S100 proteins appear to have arisen relatively late in evolution, being reported thus far only in the animal kingdom, whereas calmodulin is present in all eukaryotes (Kawasaki et al. 1998). Also unlike calmodulin, which is constitutively expressed by most if not all cell types, S100 proteins are expressed in a cell- and tissue-specific fashion (Zimmer et al. 1995). Given their recent origin and extensive divergence, it is attractive to speculate that the S100 proteins may have evolved in vertebrates to perform a variety of cellular tasks; however, the precise nature of these tasks remains largely unknown.
Whereas other members of the S100 family are upregulated in tumors and may participate in the process of metastasis (Weterman et al. 1992;Barraclough, 1998), the identification of S100A2 by subtractive hybridization as a markedly downregulated gene in malignant relative to normal mammary epithelial cell lines suggested that it might function as a tumor suppressor gene (Lee et al. 1992). S100A2 protein was also found to be markedly downregulated in transformed human keratinocyte cell lines, relative to secondary cultures of normal human keratinocytes (NHK) (Celis & Olsen, 1994;Vellucci et al. 1995). In addition, we found marked downregulation of S100A2 mRNA in cell lines derived from several epithelial tumors, relative to cell cultures derived from the corresponding normal tissues (Xia et al. 1997). Utilizing in situ hybridization, we found that S100A2 mRNA was nearly undetectable in 13 of 15 basal cell carcinomas (BCC) (Xia et al. 1997); however, S100A2 mRNA was well-expressed in six of 11 squamous cell carcinomas (SCC) of the skin (three of six tumors) and oral cavity (three of five tumors). Moreover, S100A2 protein has been found in several types of appendageal skin tumors in vivo, including BCC and SCC (Shrestha et al. 1998). While we have confirmed that S100A2 is expressed in BCC, we have observed marked downregulation of S100A2 in BCC tumors relative to adjacent normal skin (T.Z. and J.T.E., manuscript in preparation).
From the foregoing, it is apparent that the patterns of S100A2 expression obtained in actual tumors are not as clear-cut as those originally obtained in cell lines. This variability might be explained if S100A2 expression was controlled by multiple factors whose effects varied in magnitude from tumor to tumor, and from cell line to cell line. Indeed, several lines of evidence identify the epidermal growth factor receptor (EGFR) and P53 as two important signals controlling S100A2 transcription in human keratinocytes and other epithelial cells.
EGFR signaling
We found marked overexpression of S100A2 mRNA in the hyperplastic but nonmalignant epidermis surrounding both SCC and BCC, and in the hyperplastic epidermis of psoriasis lesions (Xia et al. 1997). Expression was predominantly yet not exclusively suprabasal, suggesting a role for S100A2 in regenerative hyperplasia, a recognized mode of ''emergency'' epidermal differentiation (Mansbridge & Knapp, 1987). We and colleagues have advanced the hypothesis that regenerative hyperplasia is driven by activation of keratinocyte EGFR, which are known to be overexpressed in the suprabasal layers of psoriatic and regenerating epidermis (Nanney & King, 1996;Stoll et al. 1997). [Because the EGFR is one of four related ErbB transmembrane tyrosine kinases (Klapper et al. 2000), we will refer to signaling through the EGFR and related receptors as ErbB signaling.] Consistent with our hypothesis, we found that S100A2 transcription, mRNA, and protein levels are all markedly increased in response to ErbB signaling in human keratinocytes (Stoll et al. 1998b). We would speculate that S100A2 was so readily identified as a downregulated gene in tumor cells because of its very strong expression in the corresponding ''normal'' epithelial cells in culture (Lee et al. 1992;Celis & Olsen, 1994;Vellucci et al. 1995), which are all strongly driven by ErbB signaling (Ethier et al. 1991;Barnard et al. 1994;Stoll et al. 1998a).
P53
Recently, P53 has also been shown to regulate S100A2 transcription positively (Tan et al. 1999). Thus, cell lines or tissues carrying deletions or loss-of-function mutations of P53 would be expected to manifest reduced S100A2 expression. Indeed, in each experiment in which downregulation of S100A2 was identified, the tumor cell line utilized was defective in P53 function. Two of the lines, FaDu and 21MT-2, have suffered P53 mutation and/or loss and do not express wild-type P53 (Reiss et al. 1992;Liu et al. 1969–73). The third line, SKV14, was derived by immortalization with SV40 T antigen, which is known to bind to and inactivate P53 (Kim et al. 1998).
This dual mode of S100A2 regulation by ErbB and P53 suggests a dichotomous relationship between S100A2 expression and proliferation in benign versus malignant cells, with increased S100A2 expression in response to EGFR activation in hyperplastic normal cells, and decreased S100A2 expression in response to P53 hypofunction in hyperplastic malignant cells. S100A2 expression is, however, variable but present in oral and skin-derived SCC tumors and cell lines, in which normal P53 is nearly always absent (Ziegler et al. 1994;Shrestha et al. 1998;Xia et al. 1997). It is well known that many epithelial tumors are characterized by high levels of autocrine EGFR signaling (Ethier et al. 1991). It is possible that increased EGFR signaling might compensate for decreased P53 signaling to maintain S100A2 expression in tumor cells.
Because regenerating epidermal cells express high levels of S100A2, whereas tumor cells can apparently ''afford'' to lose S100A2 expression, we suspect that S100A2 has an important function in hyperplastic normal cells that is either selectively neutral or actually deleterious for tumor cells. In a first step towards elucidating this function, we have chosen the time-honored biochemical approach to S100A2 function (Moore, 1965;Calissano et al. 1969) as a starting point. In this report, we characterize several physical properties of S100A2 in human keratinocytes, the cellular target of both regenerative epidermal hyperplasia and cutaneous carcinogenesis. Our findings suggest a possible role for S100A2 as an ''emergency'' electron donor in the context of oxidative stress. This would be consistent with a substantial literature implicating both the EGFR and P53 as important mediators of the cellular response to oxidative stress (Herrlich & Bohmer, 2000;Meplan et al. 2000).
Materials and Methods
Immunochemicals
Mouse monoclonal anti-S100A2 antibodies (IgG1 isotype) have been prepared by others using purified bovine S100L (the bovine homolog of S100A2) (Glenney et al. 1989), porcine stomach extract (Gimona et al. 1997), and human recombinant S100A2 as immunogens. These monoclonal antibodies (MoAb) were obtained from Transduction Laboratories (Nashville, TN), Sigma (St Louis, MO), and DAKO (Carpinteria, CA), respectively. Mouse monoclonal anti-S100A10 was purchased from Transduction Labs. Isotype control IgG1 was from Pharmingen (San Diego, CA). Purified rabbit anti-human S100A2 IgG was obtained from DAKO, and a polyclonal rabbit anti-human S100A2 anti-serum was the generous gift of Prof. Claus Heizmann, Department of Pediatrics, University of Zurich, Switzerland. Protein G-Sepharose was purchased from Zymed (San Francisco, CA). Monoclonal anti-FLAG antibody (M2) was purchased from Eastman Kodak Company (New Haven, CT). Anti-hemagglutinin (anti-HA) was from Boehringer Mannheim (Indianapolis, IN). Goat anti-mouse horseradish peroxidase-conjugated IgG Fc fragment was purchased from Jackson ImmunoResearch Laboratories (Bar Harbor, ME). Fluorescein isothiocyanate (FITC) conjugated goat anti-mouse IgG was from Santa Cruz Biotechnology (Santa Cruz, CA). VectashieldTM Mounting Medium was purchased from Vector Laboratories (Burlingame, CA).
Cell culture
293T human embryonic kidney cells (Pear et al. 1993) were kindly provided by Dr Gabriel Nuñez (Department of Pathology, University of Michigan) and propagated in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (FBS), 100 U penicillin per ml, and 100 
g streptomycin per ml. NHK were propagated from adult human skin in modified (Wille et al. 1984) MCDB 153 medium (M154, Cascade Biologicals, Portland, OR) as previously described (Elder et al. 1991). NHK were used in the first through third passages.
Yeast two-hybrid screening
A commercially available cDNA library derived from HaCaT keratinocytes (Boukamp et al. 1988) was obtained from Clontech Laboratories (Palo Alto, CA). This library contains keratinocyte cDNA fused to the GAL4 activation domain of pACT-2 (Durfee et al. 1993). This library was screened for proteins that interact with S100A2 using the MATCHMAKER kit (Clontech) and an S100A2-GAL4 DNA binding domain construct as ''bait'' (Harper et al. 1993). The bait construct was prepared by excising the cDNA insert from pUC-CaN19, a full-length S100A2 cDNA (Hardas et al. 1996) with EcoRI and PstI and inserting it into the corresponding sites in the polylinker of pAS2-1 (Harper et al. 1993). This construct was transformed into Y190 yeast cells by standard procedures, followed by the pACT2-HaCaT cDNA library in a second round of transformation. Transformed cells were plated on medium lacking tryptophan, leucine, and histidine but containing 50 mM 3-amino-1,2,4-triazole (Sigma) (Durfee et al. 1993). Clones (3 
106) from a single transformation were assayed for 
-galactosidase activity. Positive clones were picked and selectively grown in media lacking leucine and cycloheximide. Yeast plasmid DNA was then prepared and transformed into INVaF'Escherichia coli cells (Invitrogen). Controls for interaction specificity included the empty activation domain vector pACT-2 and pVA3-1, which encodes an irrelevant bait (murine P53). cDNA inserts in the selected plasmids were characterized by restriction enzyme mapping and nucleotide sequence analysis on both strands using an automated DNA sequencer (Applied Biosystems) at the University of Michigan Nucleotide Sequencing Core Facility.
Epitope tagging of S100A2
The HA epitope-tagged S100A2 was constructed from pUC-CaN19 (Hardas et al. 1996) by polymerase chain reaction (PCR) using a 5' primer that incorporated an EcoRI site, a consensus translation initiation site, the 10 amino acid HA epitope and six amino acids from the S100A2 sequence. The sequence for this primer was as follows: 5'-GCC GAA TTC CCC ACC ATG GCT TAC CCA GAT GTT CCA GAT TAC GCT ATG ATG TGC AGT TCT CTG-3'. The 3' primer used corresponded to sequence in the 3'-untranslated region of S100A2 and incorporated a XhoI site. The sequence for the 3'-primer is as follows: 5'-GCC CTC GAG CAA CAG ACA AAA AAA GTT TAT-3'. The FLAG epitope tagged S100A2 was constructed similarly by PCR, using the same 3' primer. The sequence of the 5'-primer for the FLAG construct was as follows: 5'-GCC GAA TTC CCA CCA TGG ACT ACA AAG ACG ATG ACG ATA AAA TGA TGT GCA GTT CTC TG-3'. PCR was performed using pUC-S100A2 as the template DNA through 30 cycles at 95°C for 30 s, 50°C for 1 min, and 72°C for 1 min. The resulting PCR fragments were digested with EcoRI and XhoI and subcloned into the corresponding sites of pcDNA3.1 (Invitrogen). All plasmid constructions were confirmed by DNA sequencing.
Immunoprecipitations
The above constructs were propagated in host INVaF' and plasmid DNAs were purified using QIAprep (Qiagen, Chatsworth, CA). Plasmids were transiently transfected into 293T human embryonic kidney cells as described (Pear et al. 1993). Twenty-hour to 48 h later, cells were lysed in IP lysis buffer [150 mM NaCl, 50 mM Tris–HCl (pH 8), 5 mM ethyleneglycol-bis(-aminoethyl ether)-N, N, N',N'-tetraacetic acid (EGTA), 5 mM ethylenediamine tetraacetic acid (EDTA), 15 mM MgCl2, and 1 mM dithiothreitol (DTT), 0.2% NP-40, supplemented with 1 g leupeptin per ml, 1 g aprotinin per ml, 10 g soybean trypsin inhibitor per ml, 1 g pepstatin per ml, and 0.2 g phenylmethylsulfonyl fluoride per ml]. Lysates were subjected to immunoprecipitation using anti-FLAG and anti-HA antibodies as described (Inohara et al. 1997). Immunoprecipitates were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) followed by western blotting as described below, using anti-FLAG or anti-HA antibodies for detection.
Preparation and western blotting of NHK lysates
NHK were seeded at 104 cells per cm2 in keratinocyte growth medium and grown to approximately 50% confluence. After washing twice with ice-cold phosphate-buffered saline (PBS), cells were subsequently lysed by hypotonic shock, digitonin treatment, or ionic detergent lysis. For hypotonic lysis, NHK were swollen for 2 min in 10 ml of ice-cold reticulocyte standard buffer [RSB, 10 mM Tris–HCl (pH 7.4), 2 mM MgCl2, 10 mM NaCl] (Penman, 1966). The RSB was aspirated, 1 ml of fresh ice-cold RSB was added, and the cells were scraped into a 2 ml Dounce homogenizer (Kontes) and homogenized on ice using 10 strokes of the ''B'' pestle (0.002 inch clearance). For determination of membrane-bound S100A2, the homogenate was then spun at 2800 g for 10 min at 4°C to remove nuclei. The supernatant was then centrifuged again at 100 000 g for 1 h at 4°C. For digitonin lysis, NHK were incubated with rocking for 5 min at 4°C in import buffer (20 mM HEPES, pH 7.35, 110 mM potassium acetate, 2 mM magnesium acetate, 0.1 mM EGTA) containing 50 g digitonin per ml (Adam & Adam, 1994). For ionic detergent lysis, 2.5 ml of ice-cold 1 lithium dodecyl sulfate (LDS) buffer [250 mM Tris–HCl (pH 8.5), 2% LDS, 10% glycerol, with or without 0.019% Serva Blue G and 0.006% Phenol Red] was added per 100 mm dish of NHK on ice. Cells were immediately scraped off the dish using a cell scraper, and pipetted up and down to ensure adequate lysis before transferring into polypropylene tubes. DNA was sheared by multiple passages through a 21G needle. In some experiments RSB containing 2% LDS was used in place of LDS buffer.
Lysates were analyzed using two different methods of gel electrophoresis. In the experiments shown in Figure 1(a, b), 2(a) and 3, pre-cast (Novex, San Diego, CA) Tris–glycine 4–20% polyacrylamide gels were run according to the method ofLaemmli (1970). In these experiments, lysates were mixed prior to loading with one-fourth volume of 250 mM Tris–HCl, pH 6.8, 40% glycerol, 0.01% Bromphenol Blue (Laemmli, 1970), with or without SDS and/or DTT to final concentrations of 2 and 37.5 mM, respectively. In the experiments depicted by Figure 3 (lanes a and b), SDS was also omitted from the running buffer. In the experiments shown in Figures 4, 5, 6, pre-cast Nu-PAGE 4–12% acrylamide gels were used (Novex). These gels utilize a Bis–Tris buffer, rather than the Tris–glycine buffer utilized in the Laemmli system, and run at neutral rather than alkaline pH. Because SDS tends to precipitate as it enters the gel in this system, the manufacturer recommends that LDS be substituted for SDS in the loading buffer. Therefore, in the experiments depicted in Figures 4, 5, 6, hypotonic and digitonin lysates were mixed with one-fourth volume of 1 M Tris–HCl (pH 8.5), 40% glycerol, 0.075% Serva Blue G, 0.025% Phenol Red, with or without LDS, and/or DTT to final concentrations of 2% and 37.5 mM, respectively. The ionic detergent lysates utilized for the experiments depicted in Figure 4 and 5 were loaded directly, without further additions. In order to assess the effects of divalent cations on the mobility of S100A2, EDTA was omitted from the Nu-PAGE loading, running and transfer buffers [1 running buffer = 50 mM 2-(N-morpholino)ethane sulfonic acid, 50 mM Tris base, 0.2% SDS, pH 7.3]. This modification had no effect on the intensity of S100A2 staining (data not shown).
Figure 1.
Subcellular localization of S100A2 in NHK. (A) Lack of S100A2 membrane association in NHK. Hypotonic lysates were subjected to 100 00 g centrifugation followed by SDS–PAGE and western blotting. Note that the 100 000 g pellets lack S100A2 and that addition of nonionic detergent to the lysis buffer did not increase the amount of S100A2 recovered in the supernatant. (B) S100A2 is not associated with the cytoskeleton in NHK. Digitonin lysates were analyzed by SDS–PAGE followed by western blotting. Lane a, digitonin-soluble fraction; lane b, digitonin wash; lane c, digitonin-insoluble fraction. Equal aliquots of a plate were loaded on to each lane; amounts of protein loaded were different in each lane. (C) Digitonin-insoluble S100A2 is localized to nuclei. Immunofluorescence microscopy of S100A2 (upper panels) or S100A10 (lower panels) in NHK fixed with 4% paraformaldehyde after rocking for 5 min at 4°C in phosphate-buffered saline (left panels) or digitonin in import buffer (right panels). Note the presence of nuclear staining and the absence of digitonin-insoluble extranuclear staining for S100A2 (upper panels), as opposed to the lack of nuclear staining and presence of digitonin-insoluble extranuclear staining for S100A10 (lower panels).
Figure 2.
Homodimerization of S100A2. (A) Epitope-tagged S100A2 efficiently homodimerizes in an immunoprecipitation/western blot assay. Western blots of Tris–glycine SDS–PAGE gels are shown. Experimental conditions (described in Materials and Methods) are indicated above and to the left of the autoradiographs. Note that immunoprecipitation with anti-FLAG results in the pulldown of HA-tagged S100A2. (B) S100A2 forms a homodimer in living cells (yeast two-hybrid assay). Various constructs were grown in selective (left panel) or nonselective medium (right panel) as described in Materials and Methods. Note that only the yeast strain transfected with the S100A2 plasmid is capable of growing in selective media.
Full figure and legend (100K)Figure 3.
Presence of disulfide-linked S100A2 dimers in nondenaturing extracts of NHK. Lanes a–d contain digitonin lysates of NHK prepared in the absence of reducing agents, as described in Materials and Methods. All procedures were carried out at 4°C. Tris–glycine PAGE gels were run within 1 h of cell lysis. The gel producing lanes a and b was run without SDS in the loading buffer. Samples were loaded using Laemmli loading buffer without denaturant (SDS) or reducing agent (DTT) (lane a); without SDS, but with DTT (lane b); with SDS, but without DTT (lane c); or with both SDS and DTT (lane d). The molecular weight markers shown to the right of the autoradiograms apply only to lanes c and d.
Full figure and legend (83K)Figure 4.
Lack of disulfide-linked S100A2 dimers in ionic detergent extracts of NHK, and appearance of oxidized S100A2 dimers after H2O2 treatment of intact cells. (A) Comparison of various extraction buffers. Untreated control NHK cultures are compared with cultures treated with 1 m M H2O2 for 90 min prior to extraction. Western blots of Nu-PAGE gels decorated with S100A2 MoAb (Sigma) were prepared as described in Materials and Methods. Extraction conditions were as follows: lane 1, cells were lysed by rocking at 4°C in 1 LDS buffer. These lysates were loaded directly. Lane 2, cells were lysed by rocking at 4°C in hypotonic lysis buffer (RSB) containing 2% LDS. These lysates were adjusted to approximately the composition of 1 LDS buffer prior to loading. Lane 3, cells were allowed to swell for 5 min in hypotonic lysis buffer at 4°C followed by scraping and Dounce homogenization. Lane 4, cells were lysed by rocking at 4°C in import buffer plus 50 g per ml digitonin. In lanes 3 and 4, one-fourth volume of 4 LDS buffer was added to the lysates prior to loading. Results are shown from two of five keratinocyte strains tested. All five strains were derived from different donors; all strains yielded essentially identical results. (B) Time course of H2O2-induced oxidation. Intact NHK were incubated with 1 m M H2O2 for the times shown at the top of the figure. The experiment shown is representative of three independent experiments.
Figure 5.
Exposure of oxidized S100A2 dimers to high calcium concentrations causes formation of cross-linked aggregates of S100A2. (A) Calcium effects on pre-oxidized dimers. Calcium was added to a nondenaturing digitonin lysate of NHK at various concentrations as listed at the top of the figure. Note that the lysates consisted primarily of oxidized S100A2 dimers (open triangles) prior to calcium treatment. Solid triangles indicate mobility of monomeric S100A2. In the left panel, the loading buffer lacked DTT. In the right panel, the loading buffer contained 37.5 mM DTT. Hypotonic lysates prepared using RSB yielded very similar results (data not shown). Also, addition of 1 mM EGTA to nondenaturing extracts of NHK had no detectable effect on the S100A2 band pattern, suggesting that calcium is not required for maintenance of the S100A2 dimer (data not shown). (B) Lack of calcium effect in intact cells. Intracellular calcium levels were artificially elevated by addition of the concentrations of CaCl2 indicated at the top of the figure to NHK together with 3 M of the calcium ionophore A23187. Denaturing extracts were then prepared using 1 LDS buffer as described in Materials and Methods. The lane labeled ''H2O2'' represents a positive control for dimerization. This lane contained a denaturing extract of intact NHK that had been treated with 1 mM H2O2 for 90 min prior to extraction.
Figure 6.
Five different anti-S100A2 antibodies detect similar S100A2 HMW band patterns. Western blots of replicate Nu-PAGE gels decorated with five different anti-S100A2 antibodies are shown. Lanes 1 and 4, digitonin lysates from untreated cells. Lanes 2 and 5, digitonin lysates from cells treated with 1 mM hydrogen peroxide for 1.5 h. Lanes 3 and 6, calcium chloride was added to digitonin lysates from untreated cells, to a final concentration of 1 mM. In lanes 1–3, DTT was omitted from the in loading buffer. In lanes 4–6, 37.5 mM DTT was present in the loading buffer. (A) MoAb from Transduction Labs; (B) MoAb from Sigma; (C) MoAb from DAKO; (D) rabbit polyclonal antibody from DAKO; (E) rabbit polyclonal antibody, gift from DR Claus Heizmann. Open triangles indicate mobility of S100A2 oxidized dimer; closed triangles indicate mobility of monomeric S100A2.
Full figure and legend (99K)In all western blotting experiments, equal aliquots were loaded, usually representing approximately 1.5% of a 100 mm dish. Protein concentrations of replicate dishes did not vary by more than 5% for comparable lysates. Samples were generally not heated prior to electrophoresis; heating had no detectable effect on the results (data not shown). After electrophoresis, the gels were transferred to PVDF membranes (Bio-Rad, Hercules, CA) using 25 mM Bis–Tris, 25 mM bicine, 20% (vol/vol) methanol transfer buffer for 1.5 h at 25 V in constant voltage mode. The membranes were blocked for 30 min with Blotto [5% non-fat dry milk in TBST (25 mM Tris–HCl pH 8.0, 137 mM NaCl, 11 mM KCl, 0.1% Tween-20)], then incubated with 0.5 g anti-S100A2 per ml in Blotto for 1 h at room temperature. The membranes were washed twice with TBST for 20 min each, and the membranes were incubated with goat anti-mouse IgG horseradish peroxidase (Santa Cruz Biotechnology) at 0.35 g per ml in Blotto for 1 h. The membranes were again washed twice with TBST for 20 min each, then chemiluminescent detection was performed using the ECL detection kit according to the manufacturer's instructions (Amersham Pharmacia, Piscataway, NJ).
Immunofluorescence microscopy
Second or third passage adult NHK cultures were plated on Superfrost Plus microscope slides (Fisher Scientific, Pittsburgh, PA) at a density of approximately 5000 cells per cm2 by applying 40 l of a freshly trypsinized NHK suspension (125 000 cells per ml in M154) to the slide. After overnight incubation at 37°C in a humidified 5% CO2 incubator, additional M154 was then added (10 ml per 100 mm dish) and the NHK were incubated until they were approximately 30–40% confluent (1–2 d, depending on plating efficiency and growth rate). Eighteen to 24 h prior to fixation, the medium was replaced with fresh M154. Immediately prior to fixation, some samples were permeabilized by incubating in 50 g digitonin per ml in import buffer at 4°C, as described earlier. All subsequent steps were performed at room temperature unless otherwise stated. Fixation was performed by immersing the slides in freshly prepared 4% paraformaldehyde (Electron Microscopy Sciences, Ft Washington, PA) in PBS for 2 h at room temperature. Slides were then rinsed thrice for 3 min in Tris-buffered saline (TBS) and blocked in 10% FBS in TBS for 1 h, followed by incubation with primary antibody in TBS containing 2% FBS for 2 h (or overnight at 4°C). Anti-S100A2 MoAb (Sigma) was used at a concentration of 1 g per ml. Anti-S100A10 was used at a concentration of 0.5 g per ml. After another three rinses in TBS, slides were incubated with FITC-conjugated goat anti-mouse IgG at 10 g per ml in TBS containing 2% FBS for 45 min. Slides were then rinsed three times in TBS and mounted under glass coverslips in anti-fade mounting medium. Controls for antibody specificity included equivalent molar amounts of the isotype control MoAb MOPC21 and omission of primary antibody. All controls demonstrated minimal or no staining.
Results
Specificity of anti-S100A2 MoAb
S100A2 was tagged at its N-terminus with the FLAG epitope and transfected into 293T cells. This construct was used to characterize the anti-S100A2 MoAb obtained from Transduction Labs. This MoAb recognized a single band of approximately 10 kDa in nonionic detergent lysates prepared from the FLAG-tagged transfectant, but not from vector-transfected cells or cells transfected with a FLAG positive control (data not shown). This MoAb, and the MoAb obtained from Sigma, also yielded identical 10 kDa bands on western blots of NHK extracts and identical immunofluorescence staining patterns on formalin-fixed NHK (data not shown). From these results, we conclude that both MoAb are specific for S100A2, at least among proteins expressed by NHK.
Subcellular fractionation of S100A2
Hypotonic lysis is a well-established method for releasing cytosolic components without solubilizing cell membranes (Penman, 1966). After hypotonic lysis of NHK, essentially all the S100A2 detectable in the postnuclear supernatants failed to sediment at 100 000 g. This indicates that cytoplasmic S100A2 is not associated with cell membranes which are known to pellet at 100 000 g. Addition of calcium, nonionic detergent, or reducing agent (DTT) to the lysis buffer had no effect on this result Figure 1a. To confirm this result, NHK were lysed with 50 g digitonin per ml, which selectively perforates the plasma membrane by solubilizing cholesterol-rich regions (Colbeau et al. 1971) but leaves cellular membranes otherwise intact. Eighty-four per cent of the S100A2 was soluble under these conditions (e.g., present in the pooled lysate plus wash fractions) as determined by densitometry of autoradiograms (Figure 1b, upper panels, sum of lanes a and b). By contrast, only 24% of immunoreactive S100A10, which is known to be associated with the plasma membrane and the cytoskeleton (Gerke & Weber, 1985), was soluble in the same fractions (Figure 1b, lower panels). The digitonin-insoluble S100A2 appears to derive from a nuclear compartment, as immuno-fluorescence microscopy of digitonin-extracted cells revealed no residual cytoplasmic staining but substantial amounts of nuclear staining (Figure 1c, upper panels). In contrast, anti-S100A10 immunostaining of digitonin-extracted NHK revealed substantial amounts of extranuclear (presumably cytoskeletal) fluorescence, but little or no nuclear fluorescence (Figure 1c, lower panels). The cytoskeletal localization of S100A10 is consistent with previous reports (Thiel et al. 1992). From these results, we conclude that at least the cytoplasmic fraction of S100A2 in NHK is not associated with cell membranes, and that a substantial amount of S100A2 resides in the nucleus.
Dimerization of S100A2
293T cells transfected with one or both constructs were lysed in a buffer containing 1 mM DTT and immunoprecipitated using anti-FLAG antibodies, followed by SDS–PAGE and western blotting using anti-FLAG or anti-HA. As shown in Figure 2(a), HA-tagged S100A2 was detected only when FLAG- and HA-tagged constructs were simultaneously transfected into the host cells.
The yeast two-hybrid system exploits the compartmentalized structure of the yeast GAL4 transcriptional activator to identify direct protein–protein interactions by reconstituting an active transcription factor (Chien et al. 1991). A positive result indicates that the two proteins under study can interact without the intervention of other proteins. In our experiments, the GAL4 DNA binding domain fused to S100A2 served as bait, and the GAL4 activation domain fused to a cDNA library prepared from HaCaT keratinocytes served as prey. In a screen of 3 106 yeast transformants, five clones remained positive through a cycle of subcloning in bacteria. As would be expected for a homodimeric protein in this assay, three of these five clones encoded the S100A2 cDNA fused in frame to the GAL4 activation domain Figure 2b. The remaining two cDNA matched mitochondrial sequences that have since been withdrawn from the EST database (data not shown). Based on the yeast two-hybrid results, we conclude that S100A2 is capable of forming homodimers directly; e.g., without the assistance of additional proteins. Moreover, the co-immunoprecipitation results show that N-terminal epitope tagging does not interfere with S100A2 homodimer formation.
Dimerization of S100A2 does not require intermolecular disulfide bonds
Because the IP lysis buffer used to obtain the results shown in Figure 2(a) contained DTT, a sulfhydryl reducing agent, it appeared likely that intermolecular disulfide bonds were not necessary for dimerization of S100A2. To confirm this result in cells that naturally express S100A2, NHK were gently lysed using digitonin in the absence of reducing agent. The lysates were then loaded on to Tris–glycine polyacrylamide gels in the presence or absence of reducing agents and/or SDS, and S100A2 was detected by western blotting. Under these conditions, S100A2 migrated predominantly as a single band when the samples were loaded and run without SDS, whether or not DTT was present (Figure 3, lanes a and b). As shown in lane c, bands of 10 and 20 kDa were present when the samples were loaded and run with SDS but without DTT. Addition of 37.5 mM DTT to the sample buffer (Figure 3, lane d) or to the lysis buffer itself (not shown) eliminated the 20 kDa band, whereas the 10 kDa band increased in intensity. Very similar results were obtained in extracts prepared by hypotonic lysis (data not shown). Based on these results, we conclude that essentially all the S100A2 in NHK is dimeric, as only one band was seen in gels lacking SDS whether or not DTT was present. We also conclude that a substantial portion of S100A2 extracted under non-denaturing conditions forms an oxidized dimer cross-linked by intermolecular disulfide bonds.
Native S100A2 is prone to intermolecular disulfide bond oxidative during extraction under non-denaturing conditions
To determine whether oxidized dimeric S100A2 was actually present in intact cells, we extracted NHK under strongly denaturing conditions (1 LDS loading buffer or RSB containing LDS), and compared these denatured lysates with other lysates prepared using non-denaturing buffers (RSB without detergent or digitonin lysis buffer). The various lysates were analyzed for S100A2 on western blots of Nu-PAGE gels loaded in the presence or absence of DTT. As shown in Figure 4(a), extraction of NHK in a buffer containing 2% LDS effectively eliminated oxidized dimer formation (lane 1). In contrast, oxidized dimer formation was substantial in NHK subjected to hypotonic or digitonin lysis, and several high molecular weight (HMW) bands were also observed (lanes 3 and 4). From these results, we conclude that S100A2 in its native conformation is strongly prone to intermolecular disulfide bond formation during extraction. Presumably, the more complete denaturation afforded by ionic detergent lysis eliminates critical elements of secondary or tertiary structure necessary for oxidative cross-linking of S100A2.
Hydrogen peroxide treatment of intact cells promotes S100A2 intermolecular disulfide bond formation
Treatment of intact NHK with hydrogen peroxide (H2O2) produced a marked change in the mobility of S100A2 bands on non-reducing gels, even when the strong denaturant LDS was present in the lysis buffer (lanes 1 and 2 in Figure 4a). H2O2 treatment caused a reduction in the 10 kDa band (the reduced dimer) and the appearance or increase of a 20 kDa band (the oxidized dimer). The shift was most pronounced in cells extracted under denaturing conditions, due to a lower background of extraction-related oxidative (compare lanes 1 and 2 to lanes 3 and 4 in Figure 4a). The formation of a 20 kDA band corresponding to the oxidized S100A2 dimer was rapid, reaching near-maximal levels by 30 min after H2O2 treatment Figure 4b. Interestingly, the intensity of the 20 kDa band declined as time progressed, returning to near-basal levels after 4 h of treatment. Based on these results, we conclude that hydrogen peroxide treatment rapidly promotes the intermolecular disulfide cross-linking of S100A2 dimers in intact NHK, with subsequent regeneration of the reduced dimer.
Divalent cations promote the formation of HMW bands involving S100A2
As mentioned above, several of HMW S100A2 bands migrating more slowly than the 20 kDa dimer were identified under non-denaturing extraction conditions (Figure 4a, lanes 3 and 4). These bands were also observed after H2O2 treatment of intact NHK followed by ionic detergent extraction (Figure 4a, lanes 1 and 2 and Figure 4b). Clearly, the formation of these HMW bands must involve disulfide cross-linking, as they are destroyed by DTT treatment. The intensity of these HMW bands was variable. Using the strong denaturant LDS to block extraction-related oxidation, HMW bands were much more intense in RSB containing LDS [10 mM Tris–HCl (pH 7.4), 2 mM MgCl2, 10 mM NaCl, 2% LDS], than in 1 LDS buffer [250 mM Tris–HCl (pH 8.5), 10% glycerol, 2% LDS]. Similar HMW bands were observed after digitonin lysis in import buffer (20 mM HEPES, pH 7.35, 110 mM potassium acetate, 2 mM magnesium acetate, 0.1 mM EGTA), and after hypotonic lysis in RBS without LDS (lanes 3 and 4 in Figure 4a). The only common factor that is unique to the buffer systems that display HMW bands is the magnesium ion.
Addition of increasing amounts of calcium chloride to non-denaturing NHK extracts (which already contain mainly oxidized dimeric S100A2) led to the appearance of HMW bands similar to those produced by magnesium (compare Figures 4a and 5a). The half-maximal calcium concentration required for HMW band formation was approximately 0.2 m M Figure 5a. In contrast, treatment of intact NHK with millimolar concentrations of extracellular calcium in the presence of the calcium ionophore A23187 followed by ionic detergent lysis produced no intermolecular disulfide bonds Figure 5b. These HMW bands are unlikely to be explained by divalent cation-dependent antibody cross-reactivity with proteins other than S100A2, as very similar band patterns were detected using two different polyclonal and three different monoclonal anti-S100A2 antibodies Figure 6. From these results, we conclude that concentrations of calcium and magnesium in the 10-4-10-3 M range favor the formation of HMW bands. These bands probably represent further intermolecular disulfide cross-linking of pre-oxidized dimeric S100A2, because HMW bands fail to form after artificially raising the calcium concentration in intact cells (in which nearly all S100A2 is in the reduced dimeric form).
Discussion
Whereas S100 proteins have been studied for over 30 y and numerous intracellular and extracellular functions have been proposed (Zimmer et al. 1995;Schafer & Heizmann, 1996;McNutt, 1998), their mechanism(s) of action remain enigmatic. As the S100 proteins were among the earliest human proteins to be purified from human tissue (Moore, 1965), many groups have approached the question of S100 function by biochemical means. As a result, at least 40 target proteins have been reported to interact with S100 family members, including diverse structural proteins (annexins, tropomyosin, intermediate filament proteins, tubulin, and actin capping proteins); several different enzymes (fructose-1,6-bisphosphate aldolase, guanylate cyclase, casein kinase II, glycogen phosphorylase, adenylate cyclase, and myosin-associated twichin kinase); and the critical cell cycle regulator P53 (Baudier et al. 1992;Fano et al. 1995;Zimmer et al. 1995;Heierhorst et al. 1996;Heizmann & Cox, 1998;Barber et al. 1999); however, the physiologic relevance of these interactions remains to be demonstrated.
The dual regulation of S100A2 expression by ErbB and P53 suggests that it must have an important function in keratinocytes and other epithelial cells. In an initial effort to elucidate this function, we have studied its subcellular localization, dimerization status, and immunolocalization in cultured human keratinocytes. We found that S100A2 partitions into the 100 000 g supernatant after hypotonic lysis Figure 1a, strongly suggesting that it is not associated with cell membranes. Moreover, low concentrations of digitonin released 84% of the S100A2 into the soluble fraction, the remainder being found in the nuclei Figure 1b, c. Nuclear localization of S100A2 has also been observed by others (Mueller et al. 1999). These findings indicate that the bulk of S100A2 is freely soluble in NHK, as digitonin does not lyse intracellular membranes and cholesterol-poor domains of the plasma membrane (Colbeau et al. 1971;Adam et al. 1992). A significant nuclear fraction of S100A2, however, is resistant to repeated extraction in import buffer, which has the approximate ionic concentrations of the cytosol (Adam & Adam, 1994). As dimeric S100A2 has a molecular weight of 20 kDa, it would be expected to equilibrate rapidly between the nuclear and cytoplasmic compartments, based on the known properties of the nuclear pore complex (Gerace, 1992). Therefore, its retention in nuclei even after two cycles of extraction suggests that S100A2 either binds to one or more non-diffusible nuclear component(s), or requires a cytosolic cofactor for nuclear efflux.
As shown by a variety of techniques, most S100 family members form homodimers or heterodimers; however, in some cases S100 proteins are monomeric (Tokumitsu et al. 1991;Skelton et al. 1995). At the outset of this study, it was not known whether S100A2 normally existed as a monomer, homodimer, or heterodimer. The results of co-immunoprecipitation and yeast two-hybrid cloning Figure 2, and cell extraction experiments Figure 3 and 4 indicate that S100A2 must exist as a homodimer in living cells. In the course of these experiments, we observed that a substantial fraction of S100A2 extracted from NHK under non-denaturing conditions formed disulfide cross-linked dimers Figure 3 and 4a. This observation was surprising, given that the cytosol of intact cells constitutes a reducing environment (Hwang et al. 1992). Moreover, the extent of S100A2 oxidative cross-linking was variable in our hands (data not shown). Therefore, we suspected that S100A2 might be undergoing oxidative cross-linking during extraction. This was confirmed by side-by-side comparison of NHK extracts prepared under denaturing versus non-denaturing conditions Figure 4. Consistent with results obtained for other S100 proteins (Mani et al. 1982;Krebs et al. 1995;Kilby et al. 1996;Landar et al. 1997;Gribenko & Makhatadze, 1998;Raftery & Geczy, 1998), we found that these intermolecular disulfide bonds are not necessary for dimer formation.
Recently, others have reported the dimerization of S100A2 (Franz et al. 1998); however, we report the intermolecular disulfide cross-linking of the S100A2 dimer. The biologically active form of S100B is known to be a disulfide-linked dimer (Barger et al. 1992) and disulfide cross-linked homodimers of murine S100A8 are formed in the lung after LPS-induced pulmonary injury (Harrison et al. 1999); however, both of these disulfide-linked S100 dimers are found in the extracellular space. Our identification of oxidized S100A2 dimers in intact cells after oxidative stress Figure 4 is, therefore, novel for the S100 proteins and suggests that S100A2 may play a part in the keratinocyte response to oxidative stress.
We also identified HMW complexes of S100A2 after treatment of non-denatured NHK lysates with 10-4-10-3 M calcium Figure 5a. Moreover, comparison of various extraction buffers indicated that similar HMW complexes could be induced by magnesium ions Figure 4a. HMW complexes were not seen in intact NHK exposed to 10-4-10-3 M calcium in the presence of A23187, suggesting that this effect does not occur unless the S100A2 dimer is pre-oxidized Figure 5b. It is very unlikely that these HMW complexes are due to the exposure of cross-reactive epitope(s) on other cellular proteins, as very similar band patterns of HMW band formation were identified using three monoclonal and two polyclonal antibodies Figure 6. These HMW bands may represent disulfide cross-linked forms of multimeric S100A2. There is precedent for this possibility, as trimers and tetramers of S100A8 and S100A9 have been identified by chemical cross-linking of neutrophilic extracts (Teigelkamp et al. 1991),and S100P forms a mixture of dimers, trimers, and tetramers when analyzed by analytical ultracentrifugation (Gribenko & Makhatadze, 1998). It remains possible, however, that the HMW bands may reflect interaction of S100A2 with one or more additional protein(s); experiments are underway to test this possibility.
These findings are of interest in light of recent structural studies of S100 proteins, which reveal a hydrophobic patch that becomes exposed when both EF hands are occupied by calcium (Heizmann & Cox, 1998;Osterloh et al. 1998;Smith & Shaw, 1998). Similar conformational changes may also be exerted by magnesium (Gribenko & Makhatadze, 1998). S100A2 contains four cysteine residues, located at positions 3, 22, 80, and 85. The presence of cysteine in position 22 is unique among the S100 proteins (Heizmann & Cox, 1998). Based on the emerging three-dimensional structures of S100 family members (Smith & Shaw, 1998), the N- and C-terminal helices containing Cys-3 and Cys-80/-85 would be in close proximity to each other in the proposed hydrophobic patch. (The fourth cysteine, Cys-22, likes within the N-terminal EF hand.) Based on the published Kd for the N-terminal, low-affinity calcium binding site of S100A2 [0.47 m M (Franz et al. 1998)], the concentration of calcium required for S100A2 HMW band formation in intact NHK (
0.2 M, Figure 5a) would be sufficient to occupy both calcium-binding sites on a significant fraction of S100A2 molecules. We would speculate that a calcium-dependent conformational change increases the reactivity of the sulfhydryl groups of S100A2 not already involved in dimer cross-linking. In the physiologic context, it is important to note that the concentration of calcium required for disulfide cross-linking of oxidized dimeric S100A2 is at least 1000 times higher than the intracellular free calcium concentration in resting cells (
0.1 M) and at least 100 times higher than that of maximally stimulated cells ( 5 M) (Alberts et al. 1994). As S100A2 appears to be secreted minimally, if at all, from NHK (data available at http://biobase.dk/cgi-bin/celis), it would probably encounter such high concentrations of calcium only in the context of plasma membrane perforation or frank cell lysis. Whereas the intracellular concentration of free magnesium ion is approximately 0.5 m M (Alberts et al. 1994), only exposure to the extracellular environment would be expected to expose S100A2 to millimolar concentrations of calcium or magnesium and an oxidizing environment at the same time (Hwang et al. 1992), thereby allowing the formation of disulfide cross-linked HMW forms of S100A2.
By analogy with S100B (Barger et al. 1992), it is attractive to speculate that the disulfide-linked form(s) of S100A2 may have biologic activity. Such activity can be envisioned in either an intracellular or an extracellular context. In the intracellular context, we would speculate that S100A2 might participate in oxidative stress responses. S100A2 is positively regulated by P53 in concert with several other genes known to be involved in oxidative stress responses (Tan et al. 1999). P53 contains 10 Cys residues, all located in the central DNA binding region of the P53 molecule where most human mutations are clustered. Eight of the 10 cysteines are evolutionarily conserved, and nine of the 10 Cys residues are mutated in human cancers (Sun & Oberley, 1996). Given the nuclear localization of S100A2 in NHK Figure 1c and its inducibility by P53 (Tan et al. 1999), it is reasonable to hypothesize that nuclear S100A2 undergoes transient oxidative dimerization Figure 4b in order to donate electrons to P53 and perhaps to other oxidation-sensitive transcription factors such as NF-
B (Sun & Oberley, 1996), thereby helping to maintain their integrity in the context of oxidative stress. Given that both P53 and NF-B normally participate in the ''cellular proof-reading'' response of normal epithelial cells in the face of massive oxidative stress (Hill et al. 1999), loss of S100A2 might aid tumor cells in their efforts to escape from this response. Current experiments in our laboratory are designed to test this hypothesis.
In an extracellular context, we would envision that oxidized dimeric and/or HMW form(s) of S100A2 released by oxidatively damaged cells that have succumbed to cellular proof-reading might function as a signal indicative of ongoing epithelial cell lysis in the vicinity. Other S100 proteins known to be secreted into the extracellular space have recently been found to activate NF-B signaling through the receptor for advanced glycation end-products (Hofmann et al. 1999). Moreover, S100A2 has been shown to be a potent leukocyte chemoattractant (Komada et al. 1996). It will be interesting to determine whether oxidized extracellular S100A2 functions to initiate or amplify innate immunologic responses to target cell lysis. If this proves to be the case, loss of S100A2 in tumor cells could contribute to a state of low immunogenicity that has long been suspected to characterize many human cancers (Hewitt, 1979).
References
| 1. | Adam EJ & Adam SA. Identification of cytosolic factors required for nuclear location sequence-mediated binding to the nuclear envelope. J Cell Biol (1994) 125: 547–555. | Article | PubMed | ISI | ChemPort | |
| 2. | Adam SA, Sterne-Marr R & Gerace L. Nuclear protein import using digitonin-permeabilized cells. Methods Enzymol (1992) 219: 97–110. | Article | PubMed | ISI | ChemPort | |
| 3. | Alberts B, Bray D, Lewis J, Raff M, Roberts K & Watson JD. Molecular Biology of the Cell. (1994) New York: Garland. |
| 4. | Barber KR, McClintock KA, Jamieson Ga, Jr, Dimlich RV & Shaw GS. Specificity and Zn2+ enhancement of the S100B binding epitope TRTK-12. J Biol Chem (1999) 274: 1502–1508. | Article | PubMed | ISI | ChemPort | |
| 5. | Barger SW, Wolchok SR & Van Eldik LJ. Disulfide-linked S100 beta dimers and signal transduction. Biochim Biophys Acta (1992) 1160: 105–112. | Article | PubMed | ISI | ChemPort | |
| 6. | Barnard JA, Graves Deal R & Pittelkow MR. et al. Auto- and cross-induction within the mammalian epidermal growth factor-related peptide family. J Biol Chem (1994) 269: 22817–22822. | PubMed | ISI | ChemPort | |
| 7. | Barraclough R. Calcium-binding protein S100A4 in health and disease. Biochim Biophys Acta (1998) 1448: 190–199. | PubMed | ISI | ChemPort | |
| 8. | Baudier J, Delphin C, Grunwald D, Khochbin S & Lawrence JJ. Characterization of the tumor suppressor protein p53 as a protein kinase C substrate and a S100b-binding protein. Proc Natl Acad Sci USA (1992) 89: 11627–11631. | PubMed | ChemPort | |
| 9. | Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Markham A & Fusenig NE. Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line. J Cell Biol (1988) 106: 761–771. | Article | PubMed | ISI | ChemPort | |
| 10. | Calissano P, Moore BW & Friesen A. Effect of calcium ion on S-100, a protein of the nervous system. Biochemistry (1969) 8: 4318–4326. | Article | PubMed | ISI | ChemPort | |
| 11. | Celis JE & Olsen E. A qualitative and quantitative protein database approach identifies individual and groups of functionally related proteins that are differentially regulated in simian virus 40 (SV40) transformed human keratinocytes: an overview of the functional changes associated with the transformed phenotype. Electrophoresis (1994) 15: 309–344. | Article | PubMed | ISI | ChemPort | |
| 12. | Chien CT, Bartel PL, Sternglanz R & Fields S. The two-hybrid system: a method to identify and clone genes for proteins that interact with a protein of interest. Proc Natl Acad Sci USA (1991) 88: 9578–9582. | PubMed | ChemPort | |
| 13. | Colbeau A, Nachbaur J & Vignais PM. Enzymic characterization and lipid composition of rat liver subcellular membranes. Biochim Biophys Acta (1971) 249: 462–492. | PubMed | ISI | ChemPort | |
| 14. | Durfee T, Becherer K & Chen PL. et al. The retinoblastoma protein associates with the protein phosphatase type 1 catalytic subunit. Genes Dev (1993) 7: 555–569. | PubMed | ISI | ChemPort | |
| 15. | Elder JT, Fisher GJ & Zhang QY. et al. Retinoic acid receptor gene expression in human skin. J Invest Dermatol (1991) 96: 425–433. | Article | PubMed | ISI | ChemPort | |
| 16. | Ethier SP, Moorthy R & Dilts CA. Secretion of an epidermal growth factor-like growth factor by epidermal growth factor-independent rat mammary carcinoma cells. Cell Growth Differ (1991) 2: 593–602. | PubMed | ISI | ChemPort | |
| 17. | Fano G, Biocca S, Fulle S, Mariggio MA, Belia S & Calissano P. The S-100: a protein family in search of a function. Prog Neurobiol (1995) 46: 71–82. | Article | PubMed | ISI | ChemPort | |
| 18. | Franz C, Durussel I, Cox JA, Schafer BW & Heizmann CW. Binding of Ca2+ and Zn2+ to human nuclear S100A2 and mutant proteins. J Biol Chem (1998) 273: 18826–18834. | Article | PubMed | ISI | ChemPort | |
| 19. | Gerace L. Molecular trafficking across the nuclear pore complex. Curr Opin Cell Biol (1992) 4: 637–645. | Article | PubMed | ChemPort | |
| 20. | Gerke V & Weber K. The regulatory chain in the p36-kd substrate complex of viral tyrosine-specific protein kinases is related in sequence to the S-100 protein of glial cells. EMBO J (1985) 4: 2917–2920. | PubMed | ISI | ChemPort | |
| 21. | Gimona M, Lando Z, Dolginov Y, Vandekerckhove J, Kobayashi R, Sobieszek A & Helfman DM. Ca2+-dependent interaction of S100A2 with muscle and nonmuscle tropomyosins. J Cell Sci (1997) 110: 611–621. | PubMed | ISI | ChemPort | |
| 22. | Glenney J, Jr, Kindy MS & Zokas L. Isolation of a new member of the S100 protein family: amino acid sequence, tissue, and subcellular distribution. J Cell Biol (1989) 108: 569–578. | Article | PubMed | ISI | ChemPort | |
| 23. | Gribenko AV & Makhatadze GI. Oligomerization and divalent ion binding properties of the S100P protein: a Ca2+/Mg2+-switch model. J Mol Biol (1998) 283: 679–694 10.1006/jmbi.1998.2116. | Article | PubMed | ISI | ChemPort | |
| 24. | Hardas BD, Zhao X, Zhang J, Longqing X, Stoll S & Elder JT. Assignment of psoriasin to human chromosomal band 1q21: Coordinate overexpression of clustered genes in psoriasis. J Invest Dermatol (1996) 106: 753–758. | Article | PubMed | ISI | ChemPort | |
| 25. | Harper J, Adami G, Wei N, Keyomarsi K & Elledge S. The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent kinases. Cell (1993) 75: 805–816. | Article | PubMed | ISI | ChemPort | |
| 26. | Harrison CA, Raftery MJ, Walsh J, Alewood P, Iismaa SE, Thliveris S & Geczy CL. Oxidation regulates the inflammatory properties of the murine S100 protein S100A8. J Biol Chem (1999) 274: 8561–8569. | Article | PubMed | ISI | ChemPort | |
| 27. | Heierhorst J, Kobe B, Feil SC, Parker MW, Benian GM, Weiss KR & Kemp BE. Ca2+/S100 regulation of giant protein kinases. Nature (1996) 380: 636–639. | Article | PubMed | ISI | ChemPort | |
| 28. | Heizmann CW & Cox JA. New perspectives on S100 proteins: a multi-functional Ca (2+) -, Zn (2+) – and Cu (2+) -binding protein family. Biometals (1998) 11: 383–397. | Article | PubMed | ISI | ChemPort | |
| 29. | Herrlich P & Bohmer FD. Redox regulation of signal transduction in mammalian cells. Biochem Pharmacol (2000) 59: 35–41 10.1016/s0006-2952(99)00298-1. | Article | PubMed | ISI | ChemPort | |
| 30. | Hewitt HB. A critical examination of the foundations of immunotherapy for cancer. Clin Radiol (1979) 30: 361–369. | Article | PubMed | ISI | ChemPort | |
| 31. | Hill LL, Ouhtit A, Loughlin SM, Kripke ML, Ananthaswamy HN & Owen-Schaub LB. Fas ligand: a sensor for DNA damage critical in skin cancer etiology. Science (1999) 285: 898–900. | Article | PubMed | ISI | ChemPort | |
| 32. | Hofmann MA, Drury S & Fu C. et al. RAGE mediates a proinflammatory axis. a central cell surface receptor for S100/calgranulin polypeptides. Cell (1999) 97: 889–901. | Article | PubMed | ISI | ChemPort | |
| 33. | Hwang C, Sinskey AJ, Lodish HF & Jones DB. Oxidized redox state of glutathione in the endoplasmic reticulum. Science (1992) 257: 1496–1502. | PubMed | ISI | ChemPort | |
| 34. | Inohara N, Ding L, Chen S & Nunez G. harakiri, a novel regulator of cell death, encodes a protein that activates apoptosis and interacts selectively with survival-promoting proteins Bcl-2 and Bcl-X (L). EMBO J (1997) 16: 1686–1694. | Article | PubMed | ISI | ChemPort | |
| 35. | Kawasaki H, Nakayama S & Kretsinger RH. Classification and evolution of EF-hand proteins. Biometals (1998) 11: 277–295. | Article | PubMed | ISI | ChemPort | |
| 36. | Kilby PM, Van Eldik LJ & Roberts GC. The solution structure of the bovine S100B protein dimer in the calcium-free state. Structure (1996) 4: 1041–1052. | Article | PubMed | ISI | ChemPort | |
| 37. | Kim SH, Banga S, Jha KK & Ozer HL. SV40-mediated transformation and immortalization of human cells. Dev Biol Stand (1998) 94: 297–302. | PubMed | ISI | ChemPort | |
| 38. | Klapper LN, Kirschbaum MH, Sela M & Yarden Y. Biochemical and clinical implications of the ErbB/HER signaling network of growth factor receptors. Adv Cancer Res (2000) 77: 25–79. | PubMed | ISI | ChemPort | |
| 39. | Komada T, Araki R, Nakatani K, Yada I, Naka M & Tanaka T. Novel specific chemotactic receptor for S100L protein on guinea pig eosinophils. Biochem Biophys Res Commun (1996) 220: 871–874 10.1006/bbrc.1996.0496. | Article | PubMed | ISI | ChemPort | |
| 40. | Krebs J, Quadroni M & Van Eldik LJ. Dance of the dimers [news]. Nat Struct Biol (1995) 2: 711–714. | Article | PubMed | ISI | ChemPort | |
| 41. | Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (1970) 227: 680–685. | Article | PubMed | ISI | ChemPort | |
| 42. | Landar A, Hall TL, Cornwall EH, Correia JJ, Drohat AC, Weber DJ & Zimmer DB. The role of cysteine residues in S100B dimerization and regulation of target protein activity. Biochim Biophys Acta (1997) 1343: 117–129. | Article | PubMed | ISI | ChemPort | |
| 43. | Lee SW, Tomasetto C, Swisshelm K, Keyomarsi K & Sager R. Down-regulation of a member of the S100 gene family in mammary carcinoma cells and reexpression by azadeoxycytidine treatment. Proc Natl Acad Sci USA (1992) 89: 2504–2508. | PubMed | ChemPort | |
| 44. | Liu XL, Band H, Gao Q, Wazer DE, Chu Q & Band V. Tumor cell-specific loss of p53 protein in a unique in vitro model of human breast tumor progression. Carcinogenesis (1994) 15: 1969–73. | PubMed | ISI | ChemPort | |
| 45. | Mani RS, Boyes BE & Kay CM. Physicochemical and optical studies on calcium- and potassium-induced conformational changes in bovine brain S-100b protein. Biochemistry (1982) 21: 2607–2712. | Article | PubMed | ISI | ChemPort | |
| 46. | Mansbridge JN & Knapp AM. Changes in keratinocyte maturation during wound healing. J Invest Dermatol (1987) 89: 253–263. | Article | PubMed | ISI | ChemPort | |
| 47. | McNutt NS. The S100 family of multipurpose calcium-binding proteins. J Cutan Pathol (1998) 25: 521–529. | PubMed | ISI | ChemPort | |
| 48. | Meplan C, Richard MJ & Hainaut P. Redox signalling and transition metals in the control of the p53 pathway. Biochem Pharmacol (2000) 59: 25–33 10.1016/s0006-2952(99)00297-x. | Article | PubMed | ISI | ChemPort | |
| 49. | Moore BW A. soluble protein characteristic of the nervous system. Biochem Biophys Res Commun (1965) 19: 739–744. | Article | PubMed | ISI | ChemPort | |
| 50. | Mueller A, Bachi T, Hochli M, Schafer BW & Heizmann CW. Subcellular distribution of S100 proteins in tumor cells and their relocation in response to calcium activation. Histochem Cell Biol (1999) 111: 453–459 10.1007/s004180050381. | Article | PubMed | ISI | ChemPort | |
| 51. | Nanney LB & King LE. Epidermal growth factor and transforming growth factor-alpha. In: Clark RAF, edThe Molecular and Cellular Biology of Wound Repair (1996) New York: Plenum Press. |
| 52. | Osterloh D, Ivanenkov VV & Gerke V. Hydrophobic residues in the C-terminal region of S100A1 are essential for target protein binding but not for dimerization. Cell Calcium (1998) 24: 137–151. | Article | PubMed | ISI | ChemPort | |
| 53. | Pear WS, Nolan GP, Scott ML & Baltimore D. Production of high-titer helper-free retroviruses by transient transfection. Proc Natl Acad Sci USA (1993) 90: 8392–8396. | PubMed | ChemPort | |
| 54. | Penman S. RNA metabolism in the HeLa cell nucleus. J Mol Biol (1966) 17: 117–130. | PubMed | ISI | ChemPort | |
| 55. | Raftery MJ & Geczy CL. Identification of noncovalent dimeric complexes of the recombinant murine S100 protein CP10 by electrospray ionization mass spectrometry and chemical cross-linking. J Am Soc Mass Spectrom (1998) 9: 533–539 10.1016/s1044-0305(98)00020-8. | Article | PubMed | ISI | ChemPort | |
| 56. | Reiss M, Brash DE, Munoz-Antonia T, Simon JA, Ziegler A, Vellucci VF & Zhou ZL. Status of the p53 tumor suppressor gene in human squamous carcinoma cell lines. Oncol Res (1992) 4: 349–357. | PubMed | ISI | ChemPort | |
| 57. | Schafer BW & Heizmann CW. The S100 family of EF-hand calcium-binding proteins: functions and pathology. Trends Biochem Sci (1996) 21: 134–140 10.1016/0968-0004(96)10020-7. | Article | PubMed | ISI | ChemPort | |
| 58. | Schafer BW, Wicki R, Engelkamp D, Mattei MG & Heizmann CW. Isolation of a YAC clone covering a cluster of nine S100 genes on human chromosome 1q21: rationale for a new nomenclature of the S100 calcium-binding protein family. Genomics (1995) 25: 638–743. | Article | PubMed | ISI | ChemPort | |
| 59. | Shrestha P, Murmatus Y & Kudeken W. et al. Localization of Ca (2+) -binding S100 proteins in epithelial tumours of the skin. Virchows Arch (1998) 432: 53–59. | Article | PubMed | ISI | ChemPort | |
| 60. | Skelton NJ, Kordel J & Chazin WJ. Determination of the solution structure of Apo calbindin D9k by NMR spectroscopy. J Mol Biol (1995) 249: 441–462 10.1006/jmbi.1995.0308. | Article | PubMed | ISI | ChemPort | |
| 61. | Smith SP & Shaw GS. A change-in-hand mechanism for S100 signalling. Biochem Cell Biol (1998) 76: 324–333. | Article | PubMed | ISI | ChemPort | |
| 62. | Stoll SW, Benedict M, Mitra R, Hiniker A, Elder JT & Nuñez G. EGF receptor signaling inhibits keratinocyte apoptosis: evidence for mediation by Bcl-XL. Oncogene (1998a) 16: 1493–1499. | Article | ISI | ChemPort | |
| 63. | Stoll S, Garner W & Elder J. Heparin-binding ligands mediated autocrine EGF receptor activation in skin organ culture. J Clin Invest (1997) 100: 1271–1281. | PubMed | ISI | ChemPort | |
| 64. | Stoll SW, Zhao X & Elder JT EGF. stimulates transcription of CaN19 (S100A2) in HaCaT keratinocytes. J Invest Dermatol (1998b) 111: 1092–1097. | Article | PubMed | ISI | ChemPort | |
| 65. | Sun Y & Oberley LW. Redox regulation of transcriptional activators. Free Rad Biol Med (1996) 21: 335–348 10.1016/0891-5849(96)00109-8. | Article | PubMed | ISI | ChemPort | |
| 66. | Tan M, Heizmann CW, Guan K, Schafer BW & Sun Y. Transcriptional activation of the human S100A2 promoter by wild-type p53. FEBS Lett (1999) 445: 265–268 10.1016/s0014-5793(99)00135-0. | Article | PubMed | ISI | ChemPort | |
| 67. | Teigelkamp S, Bhardwaj RS, Roth J, Meinardus-Hager G, Karas M & Sorg C. Calcium-dependent complex assembly of the myeloic differentiation proteins MRP-8 and MRP-14. J Biol Chem (1991) 266: 13462–13467. | PubMed | ISI | ChemPort | |
| 68. | Thiel C, Osborn M & Gerke V. The tight association of the tyrosine kinase substrate annexin II with the submembranous cytoskeleton depends on intact p11- and Ca (2+) -binding sites. J Cell Sci (1992) 103: 733–742. | PubMed | ISI | ChemPort | |
| 69. | Tokumitsu H, Kobayashi R & Hidaka H. A calcium-binding protein from rabbit lung cytosol identified as the product of growth-regulated gene (2A9) and its binding proteins. Arch Biochem Biophys (1991) 291: 401. | PubMed | ISI | ChemPort | |
| 70. | Vellucci VF, Germino FJ & Reiss M. Cloning of putative growth regulatory genes from primary human keratinocytes by subtractive hybridization. Gene (1995) 166: 213–220 10.1016/0378-1119(95)00543-9. | Article | PubMed | ISI | ChemPort | |
| 71. | Weterman MA, Stoopen GM, van Muijen GN, Kuznicki J, Ruiter DJ & Bloemers HP. Expression of calcyclin in human melanoma cell lines correlates with metastatic behavior in nude mice. Cancer Res (1992) 52: 1291–1296. | PubMed | ISI | ChemPort | |
| 72. | Wille J, Jr, Pittelkow MR, Shipley GD & Scott RE. Integrated control of growth and differentiation of normal human prokeratinocytes cultured in serum-free medium: clonal analyses, growth kinetics, and cell cycle studies. J Cell Physiol (1984) 121: 31–44. | Article | PubMed | ISI | ChemPort | |
| 73. | Xia L, Stoll SW & Liebert M. et al. CaN19 expression in benign and malignant hyperplasias of the skin and oral mucosa: evidence for a role in regenerative differentiation. Cancer Res (1997) 57: 3055–3062. | PubMed | ISI | ChemPort | |
| 74. | Ziegler A, Jonason AS & Leffell DJ. et al. Sunburn and p53 in the onset of skin cancer. Nature (1994) 372: 773–776. | Article | PubMed | ISI | ChemPort | |
| 75. | Zimmer DB, Cornwall EH, Landar A & Song W. The S100 protein family: history, function, and expression. Brain Res Bull (1995) 37: 417–429. | Article | PubMed | ISI | ChemPort | |
Acknowledgments
This work was supported by VA Merit Review award no. 321 (JTE, JF, RD, TZ), by an NIH training grant (T32AM07197) in Cell and Molecular Dermatology (TLW), by an award (R01AR42248) from the National Institute for Arthritis, Musculoskeletal, and Skin Diseases, National Institutes of Health (JTE), by the Department of Veterans Affairs Ann Arbor Hospital (JTE), and by the Babcock Memorial Trust. We thank Dr. Claus W. Heizmann of the Department of Pediatrics, University of Zurich, for generous gifts of monoclonal and polyclonal antibodies against S100A2.
