Isolation and Functional Characterization of cDNA of Serum Amyloid A-Activating Factor That Binds to the Serum Amyloid A Promoter

Abstract

Serum amyloid A (SAA), a plasma protein inducible in response to many inflammatory conditions, is associated with the pathogenesis of several diseases including reactive amyloidosis, rheumatoid arthritis, and atherosclerosis. We have previously reported an element of the SAA promoter, designated SAA-activating sequence (SAS), that is involved in the inflammation-induced SAA expression, and a nuclear factor, SAS-binding factor (SAF), that interacts with the SAS element has been identified previously (A. Ray and B. K. Ray, Mol. Cell. Biol. 16:1584–1594, 1996). To evaluate how SAF is involved in SAA promoter activation, we have investigated structural features and functional characteristics of this transcription factor. Our studies indicate that SAF belongs to a family of transcription factors characterized by the presence of multiple zinc finger motifs of the Cys₂-His₂ type at the carboxyl end. Of the three cloned SAF cDNAs (SAF-1, SAF-5, and SAF-8), SAF-1 isoform showed a high degree of homology to MAZ/ZF87/Pur-1 protein while SAF-5 and SAF-8 isoforms are unique and are related to SAF-1/MAZ/ZF87/Pur-1 at the zinc finger domains but different elsewhere. Although structurally distinct, all members are capable of activating SAS element-mediated expression and display virtually identical sequence specificities. However, varying levels of expression of members of this gene family were observed in different tissues. Functional activity of SAF is regulated by a posttranslational event as SAF DNA-binding and transactivation abilities are increased by a protein phosphatase inhibitor, okadaic acid, and inhibited by a protein kinase inhibitor, H7. Consistent with this observation, increased DNA binding of the cloned SAF and its hyperphosphorylation, in response to okadaic acid treatment of the transfected cells, were observed. Taken together, our results suggest that, in addition to tissue-specific expression, SAFs, a family of zinc finger transcription factors, undergo a modification by a posttranslational event that confers their SAA promoter-binding activity and transactivation potential.

Inflammation induced by infection or injury induces synthesis of a number of proteins which normally are expressed at a very low level. These proteins are also known as acute-phase proteins. Synthesis of serum amyloid A (SAA) protein, a member of the acute-phase proteins, can increase as much as 1,000-fold in response to inflammatory signals (reviewed in reference 21). Aberrant expression of SAA during chronic inflammatory conditions is linked with many diseases (4). In amyloidosis, an N-terminal fragment of SAA, termed amyloid A protein, is deposited in the extracellular areas of various organs including kidney, spleen, heart, and liver. Recent studies have suggested that SAA may be involved in the pathogenesis of atherosclerosis by altering cholesterol metabolism (12, 27, 44, 45). SAA has an ability to associate with high-density lipoprotein (HDL) by displacing apoA1 protein from it. Since apoA1 is the most efficient acceptor of cholesterol and is required as an activator of lecithin-cholesterol acetyltransferase reaction, SAA-rich HDL has a lesser capacity to metabolize cholesterol. Also, SAA-rich HDL particles have a much shorter half-life than HDL particles lacking SAA. Consequently, under chronic inflammatory conditions, persistent high levels of SAA can decrease the clearance rate of cholesterol and increase the risk of cardiovascular disease.

Although SAA biosynthesis primarily occurs in the liver during inflammation, it is also expressed in many other organs including kidney, spleen, and lung cells; adipocytes; vascular cell wall; and monocyte/macrophage cells. Understanding the mechanism of SAA biosynthesis especially in extrahepatic tissues is gaining more attention because many pathological conditions, such as rheumatoid arthritis, amyloidosis, and atherosclerosis, that are linked to abnormal SAA expression are manifested in nonhepatic organs. Increased SAA synthesis during inflammation is attributed primarily to the transcriptional induction of this gene (24). By transient transfection and mutation analysis, we and others have identified multiple transcriptional regulatory elements in the SAA promoter. These regions include a functional NF-κB DNA-binding element (5, 14, 22, 38), two C/EBP DNA-binding elements (5, 14, 16, 33, 34), and a promoter element known as SAA-activating sequence (SAS) (36). C/EBP and NF-κB DNA-binding elements are shown to be important for the transcriptional induction of the SAA gene in the liver. During inflammatory episodes when both NF-κB and the C/EBP family of proteins are increased and activated, they form a heteromeric complex and synergize each other’s function (35). The heteromeric complex of NF-κB and C/EBP can efficiently promote SAA transcription from both NF-κB and C/EBP sites.

Transcriptional induction of SAA in nonhepatic cells is believed to be largely regulated by the SAS element, as mutation of this region results in a severe loss of inducibility in several nonliver cells (36). A protein termed SAS-binding factor (SAF) interacts with the SAS promoter element. To better understand the mechanism by which SAF regulates the expression of the SAA gene, we set out to identify and characterize it. Here we report, on the basis of the ability to interact with the SAS element, identification of several cDNA clones that represent the SAF family of transcription factors. Among the three distinctly different cDNA clones, two isoforms, SAF-5 and SAF-8, are unique and not reported elsewhere. The third cDNA clone, SAF-1, is 85% homologous to MAZ/Pur-1/ZF87 (6, 18, 32), indicating that SAF-1 is a member of this group of transcription factors. All three SAF isoforms have a similar affinity for the SAS element and are able to promote transcription through the SAS element of the SAA promoter. Furthermore, both DNA-binding activity and transactivation potential of SAF isoforms are regulated by posttranslational modification.

MATERIALS AND METHODS

Library screening.

A rabbit brain cDNA expression library (Clontech Corporation) contained in the bacteriophage vector λgt11 was screened by the ligand interaction or Southwestern blot method (43). A ³²P-labeled concatenated SAS DNA containing multiple copies of the SAA sequence from −254 to −226 was used as a probe. Eight positive clones were selected and further analyzed by subcloning the insert cDNAs into pTZ19U vector. DNA sequencing was performed in an automated DNA sequencer.

Plasmids.

pSAS-CAT2 reporter plasmid was constructed by ligating SAA genomic DNA sequences from −280 to −226 into plasmid vector pBLCAT2 (25). A mutant plasmid, pmtSAS-CAT, was constructed by ligating a mutated SAA DNA sequence (5′-CTCAGACAAGACGGTCACTAGACTCCCAATGAGTCGAGACCGTCGACATCCATGG-3′) into pBLCAT2 vector. The underlined bases indicate substitution. Plasmid p3XSAS-CAT was constructed by ligating three tandem copies of the SAA promoter sequences from −254 to −226. The selected clones were analyzed by DNA sequencing to verify their authenticity and orientation. The SAF cDNAs were subcloned into pCMV4 vector to generate pCMVSAF plasmids.

Oligonucleotides.

The oligonucleotides, used as competitors for the binding assays, consisted of the following complementary sequences: wild-type (wt) SAS oligonucleotide, 5′-GGCTTCCTCTCCACCC-3′ and 3′-CCGAAGGAGAGGTGGG-5′; mt1 SAS oligonucleotide, 5′-GGCTTCCTCTGCACCC-3′ and 3′-CCGAAGGAGACGTGGG-5′; mt2 SAS oligonucleotide, 5′-GGCTGCCTCTCGACCC-3′ and 3′-CCGACGGAGAGCTGGG-5′; mt3 SAS oligonucleotide, 5′-GGCTTCCTCTCGGCCC-3′ and 3′-CCGAAGGAGAGCCGGG-5′; mt4 SAS oligonucleotide, 5′-GGCTTCCTCTCGGGCC-3′ and 3′-CCGAAGGAGAGCCCGG-5′; mt5 SAS oligonucleotide, 5′-GGCTAAAGCTCCACCC-3′ and 3′-CCGATTTCGAGGTGGG-5′. In each case, the underlined nucleotides represent the mutated bases. For annealing, equal amounts of complementary strands of oligonucleotides were heated at 95°C for 2 min in 50 mM Tris (pH 7.4)–60 mM NaCl–1 mM EDTA and allowed to cool slowly to room temperature in 2 to 3 h.

Cell culture and transient transfection assays.

HepG2 (human liver), HeLa S3 (human epithelial), HIG82 (rabbit synoviocyte), R9ab (rabbit lung), and THP-1 (human monocyte) cells were obtained from the American Type Culture Collection. All of these cells except THP-1 were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing high glucose (4.5 g/liter) supplemented with 7% fetal calf serum. THP-1 cells were grown in suspension in RPMI 1640 containing 10% fetal calf serum. For induction, cells were stimulated with either lipopolysaccharide (LPS) (10 μg/ml) or interleukin 6 (IL-6) (500 U/ml) and grown for different lengths of time (as indicated). THP-1 cells were transfected by the DEAE-dextran method (42). HeLa S3 cells, rabbit synoviocyte cells (HIG82), and rabbit lung cells (R9ab) were transfected by the calcium phosphate method (15). All transfections were carried out with a mixture of plasmid DNAs containing 2 μg of reporter chloramphenicol acetyltransferase (CAT) plasmid; different amounts of pCMVSAF-1, as indicated in the figure legends; and a carrier plasmid DNA so that the total amount of DNA in each transfection assay remained constant at 10 μg. CAT activity of the transfected cells was determined according to the method described before (33). All transfection experiments were performed at least three times.

Preparation of bacterially expressed SAF proteins.

SAF-1, SAF-5, and SAF-8 cDNAs were ligated in frame with suitable pRSET vectors (Invitrogen). Bacterially expressed SAF proteins were purified by affinity chromatography with a nickel-agarose column according to the manufacturer’s (Invitrogen’s) protocol.

Preparation of antibody to SAF-1.

Polyclonal antibody to SAF-1 was developed in mice by using purified bacterially expressed SAF-1 protein. Specificity of the antibody was evaluated by Western blotting and DNA band shift analyses using preabsorbed antibody. Preabsorption was carried out with purified recombinant SAF-1.

Nuclear extracts and electromobility shift assays (EMSAs).

Nuclear extracts were prepared from uninduced and LPS-induced THP-1 and synoviocyte cells and IL-6-induced HepG2 cells, essentially following a method described previously (39). Protein concentrations were measured as described elsewhere (7). EMSAs were performed following a standard protocol described earlier (35) with ³²P-labeled double-stranded DNA probe. In some binding assays, competitor oligonucleotides were included in the reaction mixture as indicated. For antibody interaction studies, anti-SAF antiserum was added to the reaction mixture during a preincubation period of 30 min on ice.

Northern blot analysis.

Multiple tissue Northern blot was purchased from Clontech, and RNA was isolated as described elsewhere (10). Northern blots were hybridized as indicated with either full-length SAF-1 cDNA, 300-bp SAF-5 unique sequences, or 230-bp SAF-8 unique sequences as the probes. Probe cDNAs were labeled by random priming (42).

In vivo phosphorylation of SAF-1.

Cultured HIG82 cells were transfected with expression plasmid pCMVSAF-1. The cells were metabolically labeled with ³²P_i (0.5 mCi/ml) in phosphate-free DMEM for 6 h either in the presence or in the absence of okadaic acid (OA) (100 nM). Prior to the labeling, the transfected cells were grown in DMEM supplemented with 5% fetal calf serum for 24 h. The ³²P-labeled cells were harvested, washed quickly in phosphate-free DMEM, and lysed by adding hot lysis buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1% Nonidet P-40, 1% sodium dodecyl sulfate [SDS], 3 mM vanadate, 2.5 mM phenylmethylsulfonyl fluoride). After lysis, the lysates were heated at 95°C for 5 min and the SDS concentration was reduced to 0.05% by adding lysis buffer without SDS. The ³²P-labeled SAF was isolated by immunoprecipitation, fractionated in an SDS–12% polyacrylamide gel, and detected by autoradiography.

Western blot analysis.

Nuclear extracts (50 μg of protein) and transfected cell extracts (50 μg of protein) were fractionated in SDS–11% polyacrylamide gels and electroblotted onto nitrocellulose membranes. The membranes were then blocked in phosphate-buffered saline–0.05% Tween 20 supplemented with 5% (wt/vol) nonfat dry milk at room temperature for 1 h. The primary antibody to SAF-1 was diluted 1:1,000 in phosphate-buffered saline–0.05% Tween 20–1% bovine serum albumin and incubated for 1 h at room temperature. Horseradish peroxidase-conjugated goat anti-mouse antibody was used as the secondary antibody. Bands were detected by using a chemiluminescence detection system (Amersham Life Science Ltd.).

Nucleotide sequence accession numbers.

The sequences reported in this paper have been deposited in the GenBank database (accession no. AF076784, AF076785, and AF076786).

RESULTS

Multiple DNA-protein complexes are formed by SAF.

We previously reported that the −280 to −226 region of the SAA promoter designated as SAS is involved in the extrahepatic expression of SAA in response to an inflammatory signal (36). The sequence of the SAA promoter in this region, hereafter referred to as SAS, is highly rich in pyrimidine nucleotides on the sense strand. A transcription factor termed SAF interacts with this promoter element (36). As seen in Fig. 1, the pattern of DNA-protein complexes was somewhat different when different cellular nuclear extracts were used to interact with the SAS element (−254/−226). Several DNA-protein complexes were seen when EMSA was performed with both control and stimulated cell nuclear extracts, indicating induction of multiple SAF-like DNA-binding activities in these cells. In THP-1 monocyte nuclear extracts, we observed three LPS-inducible DNA-protein complexes, designated a, b, and c (lanes 1 and 2). Three major DNA-protein complexes were seen with HepG2 liver cell nuclear extract, two of which, a′ and b′, appeared to be induced by IL-6 (lanes 3 and 4). In rabbit synoviocyte cells, induction of two DNA-protein complexes, a" and b", was seen when the cells were treated with LPS (lanes 5 and 6). A nonspecific complex, NS, was detected in all three cell types. This complex was not inhibited by a competitor oligonucleotide, and anti-SAF antibody did not ablate and/or supershift this complex (data not shown).

FIG. 1.

Band shift of the rabbit SAS element with the nuclear extracts of control and stimulated cells. A radiolabeled SAS element, the sequence from −254 to −226, was incubated with 10 μg of unstimulated (lanes 1, 3, and 5) and stimulated (lanes 2, 4, and 6) cell nuclear extracts. The resulting complexes were resolved in a 6% nondenaturing polyacrylamide gel. Lanes 1 and 2, THP-1 monocyte/macrophage cell nuclear extract; lanes 3 and 4, HepG2 liver cell nuclear extract; lanes 5 and 6, HIG82 synoviocyte cell nuclear extract. THP-1 and HIG82 cells were stimulated with LPS (10 μg/ml) for 24 h, and HepG2 cells were stimulated with 500 U of IL-6 per ml for 24 h. NS, nonspecific.

Cloning of cDNA encoding SAF.

To isolate a cDNA coding for SAF, we screened a rabbit λgt11 cDNA expression library with a concatenated probe containing multiple SASs, nucleotides −254 to −226 in the SAA promoter. This search resulted in the identification of eight positive clones that interacted specifically with the SAS promoter. GenBank and EMBL database searches revealed several homologous sequences. The most closely related, MAZ/ZF87/Pur-1, is 85% identical to SAF-1 at the DNA sequence level (Fig. 2A). This result indicated that SAF-1 may be a rabbit homolog of MAZ/ZF87/Pur-1. The rabbit SAF-1 cDNA is about 200 bp longer at the 3′ end than the human MAZ sequence reported previously (6), and it encodes a protein with an estimated molecular mass of 52.5 kDa, which is very close to the size of 55 kDa estimated earlier from a UV cross-linking assay (36). There were no other proteins that showed a strong homology to SAF, except in the zinc finger region. SAF-1 has six zinc finger domains of the Cys₂-His₂ type, five of which are located at the carboxyl-terminal half of the molecule. It also contains a proline-rich region at the amino-terminal end and three polyalanine tracts. Two other SAF family members, SAF-5 and SAF-8, were also characterized by DNA sequence analysis. The SAF-5 cDNA sequence is highly homologous with SAF-1 at the carboxyl-terminal end, and it contains five of six zinc fingers that are present in SAF-1. The N-terminal end of SAF-5 is different from that of SAF-1 and is rich in serine, arginine, glycine, and proline residues. An in-frame ATG codon was found 39 bp downstream of the 5′ end. Two independent clones of SAF-5 had the same 5′ ends, suggesting that they contain the complete 5′-end sequence. Primer extension analysis (data not shown) also confirmed this possibility. Another clone, SAF-8, contains an insertion of a 230-bp sequence that codes for two additional zinc fingers at the C-terminal end. However, due to this insertion, an apparent frameshift may have created an in-frame TGA codon that precludes downstream common nucleotide sequences coding for a common carboxyl-terminal amino acid in SAF-8. This event resulted in the loss of one polyalanine tract in SAF-8 cDNA, although nucleotide sequences in this region are highly conserved between SAF-1 and SAF-8. Deduced amino acid sequence comparison of the three SAF family members, described for Fig. 2B, shows that homology among all SAF isoforms in the five zinc finger domains is almost 100%, except that SAF-8 has two additional zinc finger sequences (Fig. 2B). The SAF-8 cDNA clone is believed to be partial and lacks a portion of the upstream coding sequences. Zinc fingers present at the 3′ end of all SAF isoforms presumably act as the DNA-binding domains. The proline-rich region in SAF-1 and a serine-arginine-proline-glycine-rich region in SAF-5 may function as transcriptional activation domains. The activation domain of several other members of the Cys₂-His₂ class of zinc finger proteins has been found to contain similar amino acid residues (29). The three polyalanine tracts in SAF-1 and one polyalanine tract in SAF-5 may be involved in the formation of α-helical structure, as they have been found in the Drosophila runt (17), engrailed (31), and evenskipped (26) genes. One notable feature in the amino-terminal region of SAF-5 (Fig. 2B, amino acids 77 to 84) is the potential peptide sequence that resembles SH3 ligand. Proteins carrying the SH3 domain are known to recognize a similar peptide element (1). SAF-5 may, therefore, as an SH3 ligand, be involved in protein-protein interaction with SH3-containing proteins. The amino acid sequence of the three SAF clones revealed several possible sites of phosphorylation. In SAF-1 and SAF-5, a PXTP sequence identified as the consensus phosphorylation site for mitogen-activated protein (MAP) kinases is present (2). MAP kinases (11, 41, 46) are known to be activated in response to a variety of inflammatory signals in various tissues, a condition in which the SAA gene is also induced. Also, possible phosphorylation sites for protein kinase C are identified in SAF isoforms.

FIG. 2.

(A) Nucleic acid and predicted amino acid sequences of cDNA encoding SAF-1. The zinc finger motifs are identified by dashed lines. Polyalanine and proline-rich sequences are identified with solid lines. Possible sites of phosphorylation by MAP kinase and protein kinase C are indicated by open and shaded bars, respectively. The rabbit SAF-1 cDNA sequence is compared with the human MAZ cDNA. Dash indicates absence, and dot indicates common sequence. (B) Amino acid comparison of SAF-1, SAF-5, and SAF-8. Dash indicates common sequence. The zinc finger domains are shaded and polyalanine tracts are underlined. The putative SH3-binding domain in SAF-5 is identified by a jagged underline.

Bacterially expressed SAF proteins can interact with the SAS element.

The identity of cloned SAF-1 cDNA as one coding for an SAF protein was first verified by the DNA-binding ability of the expressed protein. Bacterially expressed SAF-1 protein efficiently interacted with the SAS probe (Fig. 3A). The DNA-protein complex was eliminated by the homologous competitor but not by an unrelated oligonucleotide (lanes 2 and 3). For further characterization of the cloned gene, we used antibody raised against the cloned recombinant SAF-1 in the DNA-binding assay of nuclear proteins (Fig. 3B). All specific complexes were neutralized by the anti-SAF-1 antibody. Only one complex which was previously characterized as a nonspecific (NS) complex (Fig. 1) was unaffected by the antibody. Preabsorbed anti-SAF-1 antibody had no inhibitory effect on the DNA-protein complex formation (data not shown). This indicated that the cloned SAF-1 cDNA expresses SAF protein. Bacterially expressed SAF-5 and SAF-8 proteins can efficiently interact with the SAA promoter (Fig. 3C, lanes 1 to 6). Formation of these complexes is inhibited by anti-SAF-1 antibody and not by preabsorbed antibody (data not shown).

FIG. 3.

SAF protein expressed in bacteria binds specifically to the SAS promoter. (A) A partially purified preparation of bacterially expressed SAF protein (1 μg) was incubated with radiolabeled SAS oligonucleotide (lanes 1 to 3). In addition, a 100-fold molar excess (30 ng) of wt SAS oligonucleotide (lane 2) or a nonspecific oligonucleotide (lane 3) was added as competitor for SAF binding. Lane 4 contains no protein. (B) Radiolabeled SAS oligonucleotide was incubated with LPS-treated THP-1 cell nuclear extract (lanes 1 and 2) or LPS-treated rabbit synoviocyte cell nuclear extract (lanes 3 and 4). In lanes 2 and 4, antibodies prepared against bacterially expressed SAF protein were added. (C) A partially purified preparation of bacterially expressed SAF-5 (lanes 1 to 3) or SAF-8 (lanes 4 to 6) protein (1 μg) was incubated with radiolabeled SAS oligonucleotide. In addition, a 100-fold molar excess (30 ng) of wt SAS oligonucleotide (lanes 2 and 5) or a nonspecific oligonucleotide (lanes 3 and 6) was added as a competitor. (D) Relative affinities of SAF-1 (solid line), SAF-5 (dotted line), and SAF-8 (dashed line) proteins for various SAS oligonucleotides (sequences are described in Materials and Methods). Binding reactions were performed with 1 μg of bacterially expressed purified SAF proteins, ³²P-labeled SAA (−254/−226) probe, and the indicated amounts of unlabeled double-stranded oligonucleotides. The concentration of the unlabeled oligonucleotides is indicated on the abscissa; the amount of the residual complex, expressed as a percentage of that found in the absence of competitor (100%), is indicated on the ordinate. Binding activity was quantitated by densitometry of the DNA-protein complex formed by the interaction of each SAF isoform and SAS probe.

Relative affinities of the three SAF isoforms for the SAS element.

We were interested to know whether the three SAF isoforms differed in terms of their binding affinities for the SAS element. To test this possibility, we constructed a series of oligonucleotides containing slightly altered SAF-binding elements. These altered oligonucleotides were analyzed for their ability to compete with the wt probe for binding to three different bacterially expressed SAF proteins in the DNA-binding assay. As shown in Fig. 3D, introduction of mutant bases allowed these oligonucleotides to compete variably, among which mt5 SAS oligonucleotide did not compete at all. With these different oligonucleotides, however, no significant difference in the recognition pattern among the three SAF isoforms was noticed, which led us to conclude that bacterially expressed SAF-1, SAF-5, and SAF-8 bind indistinguishably to the SAS element of the SAA promoter.

SAF can transactivate an SAA promoter.

To assess the functional activity of SAF-1 in vivo, we cotransfected a reporter gene containing one copy of SAA promoter element (−280/−226) with increasing concentrations of an expression plasmid DNA encoding SAF-1 cDNA. Overexpression of SAF-1 induced the reporter gene activity in a dose-dependent manner (Fig. 4A). The rate of transactivation was much higher when the reporter gene contained three tandem copies of the SAS element (data not shown). Similar results were also obtained when we cotransfected cells with expression plasmids encoding SAF-5 and SAF-8 cDNAs (data not shown). It should be noted that overexpression of SAF was sufficient for transactivation of the reporter gene, and it did not require any further stimulation by LPS or cytokines. Interestingly, SAF-1 transactivated the reporter CAT gene at a much lower level in the HepG2 liver cells than in three other cell types. To determine if such a low rate of transactivation in HepG2 cells was due to any defect in the expression of the transfected SAF-1 plasmid DNA, we performed a Western blot analysis. The same four cell types were transiently transfected with equal amounts of pCMVSAF-1 DNA, and protein extracts prepared from these cells were fractionated in a polyacrylamide gel, transferred onto nitrocellulose membrane, and probed with anti-SAF-1 antibody. As seen in Fig. 4B, transfected SAF-1 was expressed at similar levels in HepG2 and three nonhepatic cell types. Since the protein level of transfected SAF-1 was the same, it can be argued that differential transactivation is not caused by different interactions of SAF-1 with the SAA promoter. Rather, the difference probably results from a function that acts more directly to modify or augment the DNA-binding property or transactivation property of SAF-1.

FIG. 4.

Transactivating ability of cloned SAF. (A) HepG2 liver, HIG82 synoviocyte, HeLaS3 fibroblast, and R9ab lung cells were transfected with increasing amounts (0, 1, 2, 3, and 4 μg) of pCMVSAF-1 DNA, together with 2 μg of either wt or mutant (mt) pSAS-CAT2 reporter plasmid DNA by the calcium phosphate transfection method. The mutated SAS sequence and details of CAT activity determination are described in Materials and Methods. The results represent averages of three separate experiments. (B) Western blot analysis for SAF in transfected cells. HepG2 (lanes 1 and 2), synoviocyte (lanes 3 and 4), HeLa (lanes 5 and 6), and lung (lanes 7 and 8) cells were cotransfected with wt pSAS-CAT2 (2 μg) and 4 μg of either pCMV4 (lanes 1, 3, 5, and 7) or pCMVSAF-1 (lanes 2, 4, 6, and 8). Details of Western immunoblotting are described in Materials and Methods. The migration position of SAF-1 is indicated by an arrow. Numbers at left show molecular masses in kilodaltons.

Differential expression of SAF cDNA isoforms.

The expression pattern of SAF was determined by Northern blot analysis with the open reading frame sequence of SAF-1 cDNA as a hybridization probe. As shown in Fig. 5A, SAF transcripts are expressed widely in the adult tissues but at varying level. Two mRNAs of 2.7 and 4.5 kb in size were detected, of which the 2.7-kb band was more prominent. The hierarchy of SAF mRNA concentrations is brain > testis > liver > lung > heart > skeletal muscle > kidney > spleen. Upon longer exposure, a 5.0-kb mRNA in brain tissue, a 2.0-kb mRNA in lung tissue, and a 1.4-kb mRNA in liver and kidney cells were detected. Detection of multiple transcripts indicated that SAF belongs to a family of proteins that are of variable size, and their expression level is not similar in all tissues. To determine the expression pattern of SAF-5 that has a different amino-terminal end, we used a 300-bp unique sequence as a hybridization probe. We detected a 1.4-kb mRNA predominantly expressed in skeletal muscle tissue (Fig. 5B). Because the SAF-5 clone was isolated from a brain cDNA library, it is anticipated that SAF-5 is also expressed in brain tissue, albeit at a much lower level. SAF-8 mRNA expression was analyzed by using a 230-bp sequence containing two unique zinc finger domains as a probe. The SAF-8 probe detected a 2.7-kb mRNA that is present at almost equal levels in all tissues (Fig. 5C).

FIG. 5.

Northern blot analysis of SAF mRNA. An RNA blot containing 2 μg of mouse poly(A)⁺ RNA per lane from heart (lanes 1), brain (lanes 2), spleen (lanes 3), lung (lanes 4), liver (lanes 5), skeletal muscle (lanes 6), kidney (lanes 7), and testis (lanes 8) cells was obtained from Clontech and hybridized with a probe containing the entire SAF-1 cDNA coding region (A). The blot was subsequently stripped and rehybridized in succession with a 300-bp fragment corresponding to the unique 5′-end region of SAF-5 cDNA (B) or a 230-bp fragment corresponding to the unique region of SAF-8 cDNA (C) or a β-actin probe (D). Autoradiographs shown in panels B and C were exposed for four times longer than the autoradiographs shown in panels A and D. Numbers at the left show sizes in kilobases.

DNA-binding activity and transactivating ability of SAF are regulated by a phosphorylation event.

We next determined whether the increased DNA-binding activity of SAF in THP-1 monocyte and HIG82 synoviocyte cells induced by LPS (Fig. 1) is due to increased gene transcription and/or posttranslational modification. In a Northern blot analysis, total RNA isolated from untreated and 24-h LPS-treated THP-1 monocyte/macrophage and HIG82 synoviocyte cells was fractionated and probed with SAF-1 coding sequence. The relative amount of SAF-1 mRNA remained virtually unchanged upon LPS treatment of the cells, indicating that LPS-mediated increased DNA-binding activity of SAF may be due to some posttranslational modification of this protein (data not shown). A previous study (36) indicated that DNA-binding activity of SAF is reduced upon dephosphorylation of the nuclear extracts in vitro by alkaline phosphatase. Consequently, we first examined whether in vivo inhibition of endogenous phosphatase activity leads to any changes in the DNA-binding activity of SAF. THP-1 cells were treated with OA, a serine/threonine phosphatase inhibitor, for different lengths of time, up to 24 h. DNA-binding assays with the SAA probe, using untreated and various OA-treated nuclear extracts, indicated that SAF DNA-binding activity was gradually increased by OA treatment (Fig. 6A, lanes 1 to 5). When the cells were treated with H7, which inhibits endogenous protein kinases, DNA-binding activity of SAF was severely reduced (Fig. 6B, lane 3). To determine if such treatment resulted in a change of SAF protein level, we performed a Western blot analysis (Fig. 6C), by using nuclear extracts prepared from untreated, OA-treated, and H7-treated THP-1 cells. The anti-SAF antibody detected no change in the level of SAF protein. This result indicated that the change in the DNA-binding activity of SAF during treatments with OA and H7 (Fig. 6B) is mainly due to a posttranslational modification.

FIG. 6.

Effect of OA and H7 on SAF DNA-binding activity. (A) Nuclear extract was prepared from THP-1 cells treated with 10 nM OA for various lengths of time. A radiolabeled SAS element, sequence from −254 to −226, was incubated with 10 μg of protein of the nuclear extracts. (B) EMSA of THP-1 cell nuclear extracts with radiolabeled SAS element. Lane 1 contains nuclear extract from untreated THP-1 cells. Lane 2 contains nuclear extract of THP-1 cells treated with 10 nM OA for 24 h. Lane 3 contains nuclear extract of THP-1 cells treated with 5 μM H7 for 24 h. (C) Western blot analysis of nuclear extracts using anti-SAF antibody. Lane 1 contains nuclear extract from untreated THP-1 cells. Lane 2 contains nuclear extract of THP-1 cells treated with 10 nM OA for 24 h. Lane 3 contains nuclear extract of THP-1 cells treated with 5 μM H7 for 24 h. Numbers at left show molecular mass in kilodaltons.

To further verify that the activity of SAF is modified by a phosphorylation event, we cotransfected pSAS-CAT2 and pmtSAS-CAT2 reporter genes with and without pCMVSAF-1 DNA and incubated the transfected cells in the presence or absence of OA. Inhibition of some endogenous serine/threonine phosphatases stimulated in vivo transactivating ability of SAF by severalfold (Fig. 7A). Sodium orthovanadate, which inhibits tyrosine phosphatase activity, had some stimulatory effect on reporter gene and SAF transactivation potential. Reciprocally, incubation of transfected cells with H7, a serine/threonine protein kinase inhibitor, severely reduced the transactivating ability of SAF. Since expression of the pmtSAA-CAT reporter gene was not affected under these conditions, we conclude that DNA-binding ability as well as transactivating ability of SAF may be regulated by a phosphorylation event. For further examination, we transfected HIG82 cells with cloned pCMVSAF-1 DNA and incubated the cells in the presence or absence of OA. As seen in Fig. 7B, OA considerably stimulated the DNA-binding activity of transfected SAF-1 cDNA. Western blot analysis of the transfected cellular proteins indicated that this increase was not due to any increase of transfected SAF-1 protein content (Fig. 7C) during OA treatment. Taken together, these results suggested that phosphorylation may play a key role in the activation of the SAF-1 gene.

FIG. 7.

(A) The transactivation ability of SAF is controlled by regulators of phosphorylation. HIG82 synoviocyte cells were transfected with 2 μg each of wt and mutant (mt) SAS reporter constructs. In some transfection mixtures, 2 μg of pCMVSAF-1 cDNA plasmid was also included. In addition, as indicated, OA (10 nM), sodium orthovanadate (vanadate; 10 μM), and H7 (5 μM) were separately included in some transfection mixtures. (B) Evidence that treatment of cells with cellular phosphatase inhibitors stimulates DNA-binding activity of SAF-1. Nuclear extracts prepared from SAF-1-transfected HIG82 cells (lane 1) and SAF-1-transfected HIG82 cells grown in the presence of OA (lane 2) were used in the DNA-binding assays with radiolabeled SAS element (−254/−226). (C) Western blot analysis of the nuclear extracts. Thirty micrograms of cellular proteins prepared from SAF-1-transfected HIG82 cells (lane 1) and SAF-1-transfected HIG82 cells grown in the presence of OA (lane 2) was separated by SDS-polyacrylamide gel electrophoresis, transferred onto a nitrocellulose membrane, and probed with anti-SAF-1 antibody. (D) Phosphorylation of SAF-1 in vivo. HIG82 cells were transfected with expression plasmid pCMVSAF-1. The cells were metabolically labeled with ³²P_i (0.5 mCi/ml) in phosphate-free DMEM for 6 h either in the presence (lane 2) or in the absence (lane 1) of OA (100 nM). The ³²P-labeled cellular proteins were immunoprecipitated with the anti-SAF-1 antibody and fractionated in an SDS–12% polyacrylamide gel.

To determine if indeed SAF-1 is hyperphosphorylated in the presence of OA, we labeled pCMVSAF1-transfected HIG82 cells with ³²P_i and immunoprecipitated the ³²P-labeled cellular proteins with anti-SAF-1 antibody. As seen in Fig. 7D, cells transfected with SAF-1 expression plasmid revealed a phosphorylated protein band (lanes 1 and 2), the intensity of which increased when the cells were labeled in the presence of a protein phosphatase inhibitor, OA (lane 2). This increase of phosphorylation was not due to any increase in the expression of the transfected SAF-1 gene, because the level of SAF-1 protein was the same during OA treatment (Fig. 7C).

DISCUSSION

Higher-level expression of SAA in response to chronic inflammatory conditions is linked to a number of pathophysiological conditions, including amyloidosis, arthritis, and atherosclerosis. Previous studies showed that the SAS promoter element of the SAA gene controls its expression and that a novel transcription factor, designated SAF, interacts with this element (36). In this report, we describe the cloning of the cDNA of SAF. The novel findings and conclusions drawn from this study are as follows: (i) SAF belongs to a family of related proteins that have common zinc finger domains at their carboxyl end, (ii) expression of an SAF isoform is tissue specific, (iii) the DNA-binding activity of SAF is increased in response to cytokine or LPS treatment of the cells, and (iv) posttranslational modifications increase DNA-binding and transactivating abilities of SAF.

Among three distinct SAF isoforms, SAF-1 appears to be a rabbit homolog of human MAZ/ZF87 (6, 32) and mouse Pur-1 (18). MAZ/ZF87 was shown to regulate expression of c-myc (6, 32) and serotonin 1A receptor (30) genes, while Pur-1 was identified as a regulator of insulin (18) gene expression. These previous studies (6, 18, 30, 32) suggested that MAZ/ZF87/Pur-1 is a constitutive widely expressed transcription factor that is present in most, if not all, tissues. However, expression of the serotonin 1A receptor gene is highly tissue specific, developmentally regulated, and controlled by hormones. Similarly, the insulin gene is subject to strict transcriptional controls, both at the level of tissue specificity and in its metabolic regulation. Consequently, how constitutively expressed MAZ/ZF87/Pur-1 can regulate expression of these genes was not fully explained. Our study, which identifies tissue-specific, inflammation-responsive, and posttranslationally modified activation of SAF, will be useful in explaining regulated expression of serotonin 1A receptor and insulin genes. Interestingly, the putative SAF-binding element is present in the promoter region of several other inflammation- or acute-phase-responsive genes (Fig. 8). In the human C-reactive protein gene, the potential SAF binding element partially overlaps the STAT3 binding element (51). A previous report (20) identified IL-6-responsive regulatory elements between −852 and −777 bp of the rat α2M gene promoter. This region contains a putative SAF-binding element. The functional importance of these potential SAF-binding elements remains to be experimentally determined.

FIG. 8.

Alignment of promoter sequence for potential SAF binding site comparison. DNA sequences of several acute-phase response genes and those of insulin, islet amyloid polypeptide (IAPP), c-Myc, CD4, and hydroxytryptamine (serotonin) are aligned. For simplicity of comparison, some of the sequences represent the sense strand and others represent the antisense strand. Since SAF interacts with DNA in an orientation-independent manner, either strand is suitable for comparison. This is consistent with the earlier findings where Pur-1 (a homolog of SAF-1) was identified as a transcription factor that binds the purine-rich sequence element (18) while the SAF DNA-binding element in the SAA promoter (39) and some MAZ-binding elements identified in the serotonin 1A receptor promoter (30) contain pyrimidine-rich sequences. Boldface underlined sequences represent the potential SAF binding domain. CRP, C-reactive protein; AGP, α1-acid glycoprotein.

The expression pattern of SAF family members is quite complex. Although transcripts of SAF-1 were identified in most tissues, we believe this ubiquitous expression pattern (Fig. 5A) arises due to the cross-hybridization of probe with different family members. By using unique sequences of SAF-5, we showed that this isoform is predominantly expressed in skeletal muscle tissue (Fig. 5B).

SAF is a strong transcriptional activator in different cell types when tested on the SAA promoter with one copy, as well as multimerized units of SAF-binding elements. SAF-1 contains clusters of proline residues like those present in WT-1 (8) and CTF (28) transcription factors. It has been proposed that clusters of proline amino acids can adopt a structure known as the polyproline II helix (49). RNA polymerase II has such a domain, which is suggested to be involved in interacting with the TFIID transcription initiation complex (19). Multiple proteins may interact via these proline-rich regions to form an optimum preinitiation complex for transcription. We recently reported that SAF can interact with Sp1 to form an SAF-Sp1 heteromeric complex which has a higher level of transactivation potential than SAF or Sp1 alone (39, 40). It will be interesting to know whether the proline-rich domain of SAF is involved in such an interaction.

One other important finding of this study is the phosphorylation-mediated activation of SAF-1 DNA-binding activity. We have shown that OA, a serine/threonine protein phosphatase inhibitor, can increase the DNA-binding ability of SAF as well as its transactivation potential, while the protein kinase inhibitor H7 acts reciprocally (Fig. 6 and 7). Consistent with these results, in SAF sequences, several potential sites for phosphorylation, including those of protein kinase C and MAP kinases, have been located (Fig. 2). It is intriguing that simple inhibition of the endogenous phosphatases can increase the DNA-binding as well as the transactivating ability of SAF. The induction of SAF activity by OA appears to be slow and is detected at 4 h after the addition of OA. This result is unlike those with some other transcription factors, such as STAT3 (48) or NF-κB (3), that are quickly activated, usually within minutes, by posttranslational modification. The slow kinetics of SAF induction in the presence of OA may be explained by assuming the presence of a constitutive but low level of endogenous protein kinase activity that can phosphorylate and increase the DNA-binding activity of SAF. OA, by inhibiting some endogenous protein phosphatase, may allow a gradual increase of the phosphorylated form of SAF. Since the phosphorylated form of SAF can be increased by OA (Fig. 7D), it suggests that protein phosphatases are also involved in the regulation of SAF phosphorylation. Indeed, a similar phenomenon has been observed in the case of cyclic AMP-regulated binding protein (CREB), which is stimulated by cyclic AMP-dependent protein kinase A-mediated phosphorylation. Nuclear protein phosphatase 2A specifically dephosphorylates protein kinase A-phosphorylated CREB (47) and attenuates CREB-mediated transcriptional activation of many eucaryotic genes. Whatever the mechanism, the results presented here suggest that SAF activation by a phosphorylation event may be a dynamic process involving an intricate balance between active phosphorylation and dephosphorylation events in order to establish a steady-state level.