Induction of Serum Amyloid A (SAA) Gene by SAA-activating Sequence-binding Factor (SAF) in Monocyte/Macrophage Cells

Serum amyloid A (SAA) is a plasma protein that is associated with many inflammatory diseases including amyloidosis, arthritis, and atherosclerosis. SAA level is significantly increased during inflammatory condition, and such abnormal expression of this protein is linked to the pathogenesis of the above-mentioned diseases. A promoter element, designated as SAA-activating sequence (SAS), located between −280 and −226 has been implicated in the induction mechanism and a nuclear factor, SAS-binding factor (SAF), has been shown to bind to this region. In this report, using a cloned SAF gene in transient transfection assay, we provide evidence that SAF potentiates SAA gene expression through SAS element. Furthermore, we show that during lipopolysaccharide-mediated induction of SAF, heteromeric complex with transcription factor Sp1 is formed. Transfection assays using both transcription factor genes have demonstrated that SAF-Sp1 heteromer is a highly potent transactivator of SAA expression.

Serum amyloid A (SAA), 1 a member of the acute-phase group of proteins, is synthesized in the body in response to inflammatory signals. Normally, SAA is expressed at an undetectable level, but infection, tissue injury, or even a prolonged high fat diet can highly induce the synthesis of this protein (1)(2)(3). SAA synthesized in various cell types (liver being the major source), circulates as constituents of high density lipoprotein (HDL). Whereas HDL of a normal healthy individual may contain negligible amount of SAA, under inflammatory conditions, concentration of SAA in HDL can rise significantly and at the peak of inflammation may constitute up to 80% of total HDL protein.
Recent studies have suggested that SAA may alter lipid metabolism because: (i) SAA has the ability to displace apoA1 protein from HDL (4 -6), (ii) SAA can itself bind to cholesterol (7), and (iii) SAA can modulate lecithin-cholesterol acyltransferase activity associated with HDL (8).
SAA is a well conserved protein among various species. In mammals, SAA comprises multiple isoforms and is coded by multiple genes. Increase of SAA biosynthesis is due primarily to its increased transcription (9). In addition to transcriptional induction, several reports have indicated that mRNA stability also contributes to the enhanced expression of SAA in mouse (10) and human (11,12). Studies have shown that C/EBP and NF-B are involved in regulating SAA gene expression in human (13), mouse (14), rat (15), and rabbit (16 -18). Recently, a new promoter called SAA-activating sequence (SAS) is identified to control expression of a rabbit SAA gene in some nonhepatic cells (19). By promoter function analysis, we previously showed that SAS promoter element can regulate LPS-mediated SAA gene induction in monocyte/macrophage cells (20). This study, however, did not reveal whether the SAS-binding factor (SAF) is directly involved in potentiating SAA transcription. In the present study, we have used a cloned cDNA of SAF transcription factor 2 and have shown that transient overexpression of SAF can activate a rabbit SAA2 gene promoter in monocyte/ macrophage cells. We also present evidence for heteromeric complex formation between SAF and Sp1. Furthermore, we demonstrate that cooperative interaction between SAF and Sp1 synergistically activates SAA gene expression through the SAS element.

MATERIALS AND METHODS
Plasmid Constructs-Plasmid construct pSAS-CAT was prepared by ligating SAA genomic DNA sequences from Ϫ280 to Ϫ226 into plasmid vector pBLCAT2 (21). Plasmid pmtSAS-CAT, a mutant derivative of pSAS-CAT, contained the mutated DNA sequence (5Ј-CAAGACGGT-CACTAGACTCCCAATGAGTCGAGACCGTCGACATCC-3Ј) ligated into pBLCAT2 vector. The underlined bases indicate substitution. Plasmid p3XSAS-CAT was constructed by ligating three tandem copies of the SAA promoter sequences from Ϫ254 to Ϫ226. The selected clones were analyzed by DNA sequencing to verify their authenticity and orientation. CMV (cytomegalovirus)-SAF contained cDNA encoding rabbit SAF gene subcloned into pCMV4 vector. pSVSp1-F and pSVSp1-FX expression plasmids contained wild-type and mutated Sp1 cDNAs, respectively (22).
Cell Culture and Transfection-Monocyte cell THP-1 was obtained from the American Type Culture Collection and maintained in suspension in RPMI 1640 containing 10% fetal calf serum. For induction, cells were stimulated with 10 g of LPS/ml and were grown for different lengths of time (as indicated). For transient transfection of THP-1 cells, 2 ϫ 10 7 cells were transfected with 10 g of plasmid DNA by the DEAE-dextran method (23). Rabbit synoviocyte cells HIG-82, obtained from ATCC, were transfected by calcium-phosphate method (24). Chloramphenicol acetyltransferase (CAT) activity of the transfected cells was performed according to the method described before (16). All transfection experiments were performed at least three times.
Preparation of Bacterial Extracts Containing SAF-The preparation of bacterial extracts containing SAF was carried out as described (25). A gt11 clone carrying SAF cDNA was isolated from a gt11-cDNA library of rabbit brain using ligand interaction or Southwestern blot method as described (26). The selected clone 2 was lysogenized in host Escherichia coli that was grown in LB medium, induced by isopropyl-␤-D-thiogalactopyranoside, and harvested by centrifugation. The cells were resuspended in 10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.5 mM PMSF, 0.5 mg/ml benzamidine, 2.5 M urea and lysed by repeated freeze-thaw cycles. The cell extracts were partially purified by passing through a DE-52 column. Flowthrough fractions that contain SAF DNA binding activity were dialyzed against 20 mM HEPES, pH 7.9, 100 mM KCl, 0.2 mM EDTA, 0.5 mM dithiothreitol, 20% glycerol, 0.5 mM phenylmethylsulfonyl fluoride, 0.5 mg/ml benzamidine.
Nuclear Extracts and Electromobility Shift Assays-Nuclear extracts were prepared from uninduced and LPS-induced THP-1 cells essentially following a method described previously (20). Protein concentrations were measured as described (27). DNA binding assays were performed following a standard protocol described earlier (17) with 32 Plabeled double-stranded DNA probe. In some binding assays, competitor oligonucleotides were included in the reaction mixture. For antibody interaction studies, anti-Sp1 (Santa Cruz Biotechnology) and anti-SAF antiserum (20) were added to the reaction mixture during a preincubation period of 30 min on ice. Purified Sp1 protein was obtained from Promega Corporation.

Transactivation of SAA Promoter by SAF in Monocyte/
Macrophage Cells-SAA gene transcription is induced in monocyte/macrophage cells by LPS (20,28). We previously showed that nonhepatic transcriptional induction of SAA is primarily regulated by SAS promoter located between Ϫ280 and Ϫ226 nucleotide sequences (19). In keeping with the promoter activity, using THP-1 monocytic cell nuclear extract, specific DNAprotein complex binding to this promoter was detected. The DNA-binding factor was characterized as SAF (20). These studies, although demonstrating involvement of SAF in DNA binding to SAA promoter, did not determine the transactivation potential of this factor in SAA gene expression.
To assess the in vivo role of SAF in SAA gene transcription, we used a transient expression system in which a cloned cDNA of SAF transcription factor 2 was overexpressed in THP-1 monocytic/macrophage cells. A reporter gene, pSAS-CAT, containing one copy of the promoter element of the SAA gene from Ϫ280 to Ϫ226, was transfected alone or with an expression plasmid containing SAF cDNA under the control of CMV promoter. As shown in Fig. 1, overexpression of SAF increased the reporter gene expression in a dose-dependent manner. The level of reporter gene activation was much higher on a reporter gene containing multiple copies of the SAS sequence. The transactivating effect of pCMV-SAF was negligible on pmtSAS-CAT reporter gene containing mutated sequence of SAS element. These results indicate that SAF is involved in the expression of SAA gene in monocyte/macrophage cells.
Multiple DNA Binding Activities of SAF Are Induced in THP-1 Monocyte/Macrophage Cells by LPS-To test how LPS treatment of monocyte cells affects DNA binding activity of SAF, gel mobility shift assays were performed with untreated and LPS-treated THP-1 monocytic/macrophage cell nuclear extracts. As the DNA probe, we used a minimal binding region, sequences from Ϫ254 to Ϫ226 base pairs, containing SAFbinding elements that produced a better resolution of the DNAprotein complexes than that obtained with the larger probe (Ϫ280/Ϫ226) used earlier (20). Interestingly, the shorter DNAbinding element (Ϫ254/Ϫ226) gave rise to multiple DNA-protein complexes designated as a, c, d, and e by the control untreated THP-1 nuclear extract (Fig. 2, lane 1). It is possible that nuclear proteins interacted less efficiently with the larger probe (Ϫ280/Ϫ226) due to the steric hindrances posed by extra 5Ј-flanking sequences. Occasionally, we have seen a slightly higher level of complex d in control untreated THP-1 cell extract than that seen in lane 1. This pattern was changed when the probe was incubated with same protein amount of LPStreated (48 h) THP-1 cell nuclear extract (lane 2). LPS treatment increased the intensities of complexes a and d severalfold and caused the appearance of a new complex designated as b that was not formed by untreated cell nuclear extract (lane 1).
The intensity of minor complex e, seen in lane 1, was reduced in LPS-treated nuclear extract. This complex often showed variability in its intensity; its level decreased variably during LPS treatment of the cells. Intensity of complex c remained virtually unchanged during LPS treatment. Competition with a molar excess of a nonradioactive homologous probe inhibited to Ϫ226 base pairs. Plasmid p3XSAS-CAT contains three copies of the SAS element (Ϫ254 to Ϫ226). These SAA sequences were ligated in front of the tk promoter in the pBLCAT2 vector. Identical experiments were performed with a reporter plasmid, pmtSAS-CAT, which contains mutated SAS element, and pBLCAT2 parent vector plasmids. The reporter plasmids (10 g) were cotransfected in THP-1 cells with increasing amounts of the expression plasmid pCMV-SAF.

FIG. 2.
Analysis of the DNA-protein complexes using electrophoretic mobility shift assay. Nuclear extracts (10 g of protein) prepared from THP-1 cells, both uninduced (lane 1) and LPS-induced for 48 h (lanes 2-9), were incubated with the SAS DNA probe (Ϫ254 to Ϫ226) containing the following sequence Ϫ254 CCCTTCCTCTCCAC-CCACAGCCCCCATGG Ϫ226 . The resulting complexes were resolved in a 6% nondenaturing polyacrylamide gel. Five DNA-protein complexes (a-e) were seen with LPS-induced nuclear extract. In some binding reactions (lanes 4 -9), increasing levels of unlabeled homologous probe were used for competition analysis of the DNA-protein complexes formed by the LPS-48 h THP-1 cell nuclear extract. Lanes 1-3 contain no competitor DNA. complexes a, b, and d but not complex c (Fig. 2, lanes 4 -9). At a very high level of the homologous probe, some reduction of the intensity of complex c was seen (data not shown). This suggested that complex c was formed by a very low affinity DNAprotein interaction between a nuclear factor and the SAA promoter. In some experiments, we have noticed the appearance of some faster migrating DNA-protein complexes that are sequence-specific, as judged from efficient competition by excess unlabeled homologous probe.
Kinetics of LPS-mediated Increase in the DNA Binding Activity of SAF-Because the above experiment indicated LPSmediated increase in the levels of multiple DNA-protein complexes, namely a, b, and d, we sought to determine the kinetics of activation of these DNA-binding proteins. Nuclear extracts were prepared from THP-1 cells that were incubated in the presence of LPS for various lengths of time (Fig. 3, lanes 1-4).  3 and 4). The level of complex c remained virtually unchanged during LPS exposure of the cells. Some additional minor bands composed of faster migrating DNA-protein complexes that are sequence-specific, as judged from efficient competition by excess unlabeled homologous probe (described earlier in Fig. 2), were also detected. Preincubation of LPS-24 h nuclear extract with anti-SAF antibody completely inhibited complexes b and d and partially blocked complex a formation (Fig. 3, lanes 6 and 7). In the SAF DNA-binding element, a Sp1 transcription factor DNA-binding sequence, a 5Ј-CCACCC-3Ј sequence element, is embedded (see legend of Fig. 2). To determine if any of these complexes are formed by Sp1, we used anti-Sp1 antibody (lane 8). Sp1 antibody partially supershifted the complex a and had no effect on complexes b and d. The addition of nonspecific preimmune serum also had no effect on any of these complexes (lane 9). In a separate experiment when both anti-SAF and anti-Sp1 antibodies were included in a single reaction, complete inhibition of complex a as well as complexes b and d were noted (data not shown). These results taken together indicated that complex a is formed by the combined interaction of Sp1 and SAF, whereas complexes b and d are formed by SAF only. It is noteworthy to mention that complex a has a higher affinity for the SAS element because it was most efficiently competed with the homologous probe (Fig. 2, lanes 3-9).
Interaction of SAF and Sp1 Transcription Factors with SAS Promoter Element-In light of the facts that Sp1 can interact with the SAS promoter and Sp1 is present ubiquitously in almost all cell types, we addressed whether Sp1 and SAF can simultaneously interact with SAS element and, if so, what is the functional consequence of such an interaction. In vitro interaction of these transcription factors was studied using a fractionated system where both of these factors could be provided in different combinations. DNA binding assays were performed with a SAF protein preparation obtained from a gt11 recombinant phage containing SAF cDNA and affinity column purified Sp1 available from a commercial source (Promega). As the probe we used the SAS element, sequences from Ϫ254 to Ϫ226. The addition of bacterially produced SAF fusion protein formed a single DNA-protein complex (Fig. 4, lane 1). The anti-SAF antibody considerably inhibited this complex formation (lane 4), which indicated the specific nature of this complex. It should be noted that the cloned SAF-␤gal fusion protein produced only one complex with SAS element as opposed to several DNA-protein complexes (a, b, d, and e) formed by the crude nuclear extract of THP-1 cells (Figs. 2 and 3). This difference could arise due to the presence of either several different isoforms of SAF family in THP-1 cells or due to other factors present in the THP-1 cells that may have some influence in SAF binding or both. In comparison to SAF, purified Sp1 protein interacted very weakly, almost at an undetectable level with SAS DNA (lane 2). This SAS-Sp1 complex was detectable only after prolonged exposure of the autoradiogram (data not shown). When we added both SAF and Sp1 proteins  in the DNA binding assay (lane 3), two DNA-protein complexes, one migrating with SAF-specific complex and the other migrating quite slowly, were detected. Although SAF antibody considerably reduced the levels of both of these two complexes (lane 6), the Sp1 antibody supershifted the slower migrating complex (lane 7). These results indicated that the top complex in lane 3 is a heteromer of SAF and Sp1 and the bottom complex is composed of only SAF. It was interesting that although Sp1 alone could not form a very visible complex with SAS promoter DNA, in the presence of SAF it formed a heteromeric SAF-Sp1 complex that was readily detectable.
Because protein-protein interactions between transcription factors often lead to both high level and cell-specific gene activation, we evaluated the rate of heteromeric SAF-Sp1 complex formation with SAS DNA. The rate of SAF-Sp1 heteromeric complex formation was favored (Fig. 5, lanes 1-4) when a constant amount of SAF was used in conjunction with increasing amounts of Sp1. Addition of increasing levels of Sp1 recruited more SAF to form the SAF-Sp1 heteromer, and because a constant amount of SAF was provided, this somewhat lowered the level of SAF-specific complex (lanes 2-4). In a reciprocal experiment, an increasing dose of SAF in the presence of a constant amount of Sp1 again favored the formation of SAF-Sp1 heteromer rather than the SAF-specific complex (lanes 5-8).
Synergistic Transactivation of SAS Promoter Element by SAF and Sp1-To evaluate the in vivo effect of SAF and Sp1 interaction in the transcriptional activation of SAA gene, we performed cotransfection experiments using pSAS-CAT reporter plasmid and expression vectors encoding SAF and Sp1. Transfection assays were conducted in HIG-82 cells because these cells have lower levels of endogenous Sp1 and SAF activity than that present in THP-1 cells (data not shown). Lower levels of endogenous SAF and Sp1 made it easier to observe transactivation in these transfection experiments. Transfection of pSAS-CAT reporter gene with SAF expression plasmid yielded about 6-fold activation, whereas the same amount of Sp1 expression plasmid transactivated the reporter gene about 2.5-fold over the control (Fig. 6). However, combination of equal amounts of both SAF and Sp1 activated the pSAS-CAT reporter gene at a level much higher than that expected from their additive value (Fig. 6). The expression plasmid, pSVSp1-FX (containing a frameshift mutation in the Sp1 coding region), had no positive transactivating effect.
The transactivation potential of the SAF-Sp1 heteromer was further analyzed by cotransfecting cells with SAS-promoter containing reporter plasmid plus a combination of a constant amount of SAF and increasing amounts of Sp1 expression plasmids. The results, described in Fig. 7A, revealed a synergistic dose-dependent transactivation of the reporter gene. Western blot analysis (data not shown) was performed, which verified that the increase of reporter gene expression was proportional to increasing Sp1 protein expression in the transfected cells, whereas the expression level of SAF protein remained same. In a reciprocal experiment, similar synergistic transactivation was seen with a fixed amount of Sp1 and an increasing dose of SAF expression plasmids (Figs. 7B). Western blot analysis verified that the increase in reporter gene expression is proportional to increased SAF production with similar levels of Sp1 production (data not shown). In both experiments, the observed induction level of the reporter gene was always higher than that achieved by a simple additive effect of the two transcription factors. These results suggested that Sp1 may cooperate with SAF and synergize its transactivating ability. DISCUSSION SAA, a member of the acute phase protein group is gaining more attention in the pathogenesis of atherosclerosis because recent studies have suggested that SAA, by displacing apoA1 from HDL, is capable of affecting lipoprotein metabolism (4). Also, SAA is found to be a constituent of the atherosclerotic lesions (29), which is composed of macrophage-derived foam cells. In an effort to control the SAA production as a therapeutic measure against pathophysiological inflammatory conditions, we undertook this investigation. The mechanism of SAA gene induction was studied, particularly in monocyte/macrophage cells, which play a key role in atherogenesis. In this report we have shown that SAF, a zinc finger transcription factor is involved in the regulation of SAA induction in THP-1 monocyte/macrophage cells. Also, we provide evidence for physical and functional interactions between the transcription factors SAF and Sp1 that have a synergistic transactivating ability on SAS promoter. SAS promoter containing Ϫ280 to Ϫ226 base pairs sequences of SAA gene is induced in response to LPS treatment of THP-1 cells. SAF has been found to interact with this element, and this protein was detected in both hepatic and several nonhepatic cells (19,20). In an effort to understand whether such interaction of SAF with SAS element results in transcriptional induction, we have studied the transactivation potential of a cloned SAF gene on a reporter gene carrying a SAS element. Increasing reporter CAT activity as more of the SAF cDNA is included in the cotransfection assay ( Fig. 1) confirmed that SAA gene induction through SAS promoter is indeed regulated by SAF. Consistent with the ability of SAF to induce expression via the SAS element, an increase of the DNA binding ability of SAF in response to LPS-treatment was seen (Figs. 2 and 3). Although control THP-1 cells contain a low level of SAF, our results indicate that such a low level of SAF may not be sufficient to promote SAA gene expression.
We have also assessed the functional consequence of SAF and Sp1 interaction on SAA gene induction. SAF is a zinc finger transcription factor containing six zinc finger motifs. Although SAF and Sp1 both contain zinc finger domains, SAF is structurally different from Sp1 and from three of its family members, Sp2, Sp3, and Sp4. Sp1 alone does not interact very efficiently with the SAS promoter but forms a stable heteromeric SAF-Sp1 complex with SAS DNA. In fact, in vitro DNAbinding studies suggested that SAF-Sp1 heteromeric complex is more favorably formed than the SAF-or Sp1-specific complex (Figs. 4 and 5). We also noticed that SAF-Sp1 heteromer has a higher affinity for the SAS element than the SAF alone (Fig. 2). The fact that complex a, which is composed of SAF-Sp1 heteromer, is readily competed by the SAS oligonucleotide, suggests that this complex has high affinity for SAS promoter sequence. In the competition assay, preincubation of nuclear extract with unlabeled SAS oligonucleotide allows virtually irreversible binding of only high affinity factors to the SAS oligonucleotide, thus making these factors unavailable for binding to radiolabeled SAS oligonucleotide, which is subsequently added in the assay mixture. Rapid sequestration of complex a by the competitor SAS oligonucleotide thus indicates high affinity binding for the involved factors. Because the level of SAA gene activation with combined SAF and Sp1 is much higher than that seen by these factors alone, this suggested that SAF-Sp1 heteromer is a superior transactivator of SAA than SAF or Sp1 alone. Accumulative evidence indicates that the cell-specific high level expression of a gene is achieved through the combined actions of a group of transcription factors. In fact, previous studies showed that under some inflammatory conditions NF-B and C/EBP synergistically transactivate SAA expression (18). Sp1 was recently shown to potentiate the transactivation role of GATA-1, a major erythroid transcription factor (30). The mechanism by which Sp1 potentiates SAF activity is not yet known. It is tempting to speculate that the SAF-Sp1 heteromeric complex establishes a stable contact with basal transcription initiation complex with the help of Sp1 when interacting with the SAS element via SAF. Indeed, the glutamine-rich activation domains of Sp1 were found to bind to a glutamine-rich domain of TAFII 110, which is a major protein component of basal transcription initiation complex (31,32), and such interaction helps to mediate transcription activation possibly via conformational or covalent modification of the basal transcription complex (reviewed in Ref. 33). SAF-Sp1 interaction may play a critical role in this process where SAF via interacting with the SAS promoter of SAA gene recruits Sp1 to the vicinity of basal transcription complex formed at the TATA box and facilitates Sp1 interaction to TAFII 110.
In summary, our data show that SAF activates SAA gene induction in THP-1 monocyte/macrophage cells. The in vivo biological consequence of the fact that SAF can functionally cooperate with Sp1, a ubiquitously present transcription factor, and synergistically transactivate SAA promoter remains to be determined.