PROTEINS: Structure, Function, and Genetics 30:381–387 (1998) A Fusion Protein Between Serum Amyloid A and Staphylococcal Nuclease—Synthesis, Purification, and Structural Studies Alan K. Meeker1 and George H. Sack, Jr.1,2* of Biological Chemistry, The Johns Hopkins University School of Medicine, Baltimore, Maryland 2Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, Maryland 1Department ABSTRACT We developed a recombinant DNA system to overexpress a fusion protein between the small, minimally soluble acute phase serum protein, serum amyloid A (SAA), and the bacterial enzyme staphylococcal nuclease (SN). This fusion protein is very soluble and is immunoreactive to polyclonal anti-SAA antibodies. Tryptophan fluorescence shows smooth denaturation curves for the fusion protein in guanidinium HCl or potassium thiocyanate. Fluorescence also indicates that only a single tryptophan residue (of the four present) is accessible to iodide quenching and, presumably, is exposed on the surface of the fusion protein. Circular dichroism (CD) shows a significant signal indicating a-helix, which can be attributed to the SAA portion of the molecule; these are the first CD spectral data available for SAA. pH titration shows persistence of helix domains for the fusion protein at pH 3.0, in contrast to the denaturation of SN under the same conditions. (The entire fusion protein shows a random coil pattern below pH 3.0.) By exploiting the structural and solubility properties of SN, this fusion protein has provided the first structural data about SAA—the precursor of the amyloid deposits in secondary amyloidosis. This fusion protein should be useful for further physical and physiologic studies of SAA. Proteins 30:381–387, 1998. r 1998 Wiley-Liss, Inc. Key words: serum amyloid A; fluorescence; circular dichroism; acute phase; denaturation; nuclease; amyloidosis INTRODUCTION Serum amyloid A (SAA) is the most prominent protein of the acute phase response in humans and many other animals.1 Human SAA comprises 104 amino acids and the protein has been highly conserved in mammalian evolution. Proteolytic cleavage of the native monomer yields an N-terminal fragment of , 64 aa which forms the highly ordered, protease-resistant polymeric fibrils of ‘‘secondary’’ or ‘‘reactive’’ amyloidosis.2,3 SAA-derived amyloid fibrils are prototypes for understanding the biology of r 1998 WILEY-LISS, INC. amyloid disease, which is an important, generally lethal complication of neoplastic, genetic, and chronic inflammatory disorders throughout the world. Because of the widespread pathologic involvement of SAA proteins we isolated and characterized members of the human SAA gene family,4,5 a cluster of similar loci on chromosome 11.6 At least one member of this gene/protein family may function as an autocrine stimulator of collagenase secretion.7,8 This protein has multiple N-terminal aa differences from the precursor of amyloid fibrils, but they are otherwise quite similar.9 Despite the important information available regarding the SAA genes, the basic problem in understanding amyloid disease is related to the structure and biology of the proteins themselves. Thus, we sought to investigate the structure of SAA proteins in order to develop models for them and the fibrils to which they contribute. Unfortunately, the low solubility of SAA (in most reports a maximum of about 50 µg/ml) has prevented detailed structural studies in the past. Visual inspection of the SAA aa sequence does not indicate why it should be so insoluble, especially since 40% of its residues are charged. Because of this, we hypothesized that much of the insolubility may be related to the folded structure of SAA monomers. Isolating SAA from serum is complicated by its low concentration and limited solubility. Thus, we sought to use recombinant systems to produce SAA proteins suitable for structural and physiologic studies. Producing native SAA monomers is complicated by their apparent toxicity in bacterial systems (unpublished observations) and their intrinsic insolubility, even if such systems were successful. Thus, we studied fusions between staphylococcal nuclease (SN)—an extremely soluble and well-characterized bacterial enzyme10,11—and SAA. We showed earlier Abbreviations: SAA: serum amyloid A; SN: staphylococcal nuclease; CD: circular dichroism; Gdn-HCl: guanidinium chloride; KSCN: potassium thiocyanate; PCR: polymerase chain reaction; PIPES: piperazine-N,N’-bis(2-ethanesulfonic acid); IPTG: isopropyl β-D-thiogalactoside. *Correspondence to: George H. Sack, Jr., Blalock 1008, The Johns Hopkins Hospital, Baltimore, MD 21287. Received 1 July 1997; Accepted 2 September 1997 382 A.K. MEEKER AND G.H. SACK, JR. that a fusion protein between a truncated form of SN and SAA was recognized by anti-SAA antibodies but still had limited solubility.12 Thus, we investigated the properties of a fusion protein combining fulllength SN and SAA that displays excellent solubility. We present spectral data indicating that this very soluble fusion protein has significant, stable structure making it an appropriate reagent for subsequent analysis. METHODS Preparation of Recombinant Expression Vector Plasmid pGS112 (containing the complete cDNA sequence for human SAA1 and derived from pA113) was linearized by BamH1 digestion and pFOG506 (containing the complete SN coding region14) was linearized by AlwN1 digestion (Fig. 1A). 25 ng of each linearized DNA was mixed with 1.3 µg primer 1 (5’ GCAACTAGTACTAAAAAATTACATAAAGAACCT 3’), 1.3 µg primer 2 (5’ CGGAGGATCCTTAGTATTTCTCAGACAGGCC 3’), 5 ng primer 3 (5’ TGGAGCGAAGACAACCGAAGCTTCTTTTCGTTCCTTGGC 3’), and 5 ng primer 4 (5’ GCCAAGGAACGAAAAGAAGCTTCGGTTGTCTTCGCTCCA 3’) in 10 mM KCl, 20 mM Tris-HCl, pH 8.8, 10 mM (NH4)2SO4, 2 mM MgSO4, 0.1% Triton X-100, 0.2 mM deoxynucleoside triphosphates, and 0.5 µl VenttPolymerase (New England Biolabs, Beverly, MA) in 100 µl. The mix was overlaid with oil and run for 30 cycles of (1 min @ 95°, 1 min @ 55°, 1 min @ 72°). Following the polymerase chain reaction (PCR), 8 U of Klenow DNA polymerase was added and incubation continued for 20 min at 37°. The reaction then was precipitated in ethanol and the product was resuspended at 50 ng/µl in 2.5 mM Tris, pH 8.0, 0.25 mM EDTA. The fragment was digested with BamH1 and SpeI and ligated into the pET12A vector (Novagen, Madison, WI) with matching ends. The product was used to transform Mg-T7 Escherichia coli; ampicillin-resistant colonies were selected. Ampicillin-resistant colonies were screened for expression of a protein of appropriate size (visualized by Coomassie-staining of SDS-PAGE gels) following induction by isopropyl β-D-thiogalactoside (IPTG). Induction and Purification of Fusion Protein The culture was inoculated from a single transformed colony and grown for 16 hr at 31°C in medium containing 10 gm tryptone, 5 gm yeast extract, 5 gm NaCl, and 200 mg ampicillin per liter; 1.5 ml was then used to inoculate 200 ml cultures of the same medium. Cultures were grown 4 hr at 36°C, IPTG was then added to 1.0 mM and growth continued for 2.5 hr. Cells were harvested by centrifugation at 3.5 K rpm, resuspended in 1/10th volume 6M urea, 25 mM Tris-HCl, pH 8.1, 5 mM EDTA at 4°C and placed on ice for 20 min. The suspension was centrifuged at 8K rpm for 15 min in a GSA rotor and Fig. 1. (A) Protocol for assembling DNA construct used to produce the fusion protein. Numbered oligonucleotides correspond to sequences in Methods. Complete pET 12A vector not shown. (B) (Left) SDS-polyacrylamide gel electrophoresis of purified fusion protein stained with Coomassie blue. Left lane 5 nuclease, right lane 5 fusion protein. (Right) Reactivity of a Western blot of the same electrophoresis gel with polyclonal anti-SAA antibodies. (C) Sequence of the fusion protein (253 amino acids). Underlined amino acids correspond to SN; those in bold represent SAA. the supernatant was diluted 1:8 in the original buffer, applied to a fast flow sulfonyl Sepharoset (Pharmacia, Piscataway, NJ) column and washed with 5 bed volumes of 6 M urea, 2.5 mM EDTA, 20 mM glycine, pH 9.5. The fusion protein was eluted with the same buffer containing an additional 100 383 SAA/NUCLEASE FUSION PROTEIN mM NaCl. The eluate was dialyzed against 5 mM NaCl, 5 mM PIPES, pH 6.5 at 4°C and a fine precipitate formed which was discarded; the remaining soluble protein was concentrated for further study. Electrophoresis and Antigenicity The protein migrated as a single band of the predicted molecular weight using Coomassie staining on 15% SDS-polyacrylamide gel electrophoresis and appeared .95% pure (Fig. 1B). It reacted with commercial polyclonal anti-SAA antibodies (Calbiochem, La Jolla, CA, #566705) on Western blots; no reaction with native SN was observed in control lanes. Spectroscopic Studies For denaturation studies, fluorescence was recorded on a SPEX Fluorolog II (SPEX Industries, Edison, NJ) at 20°C. The 2 ml cuvette contained protein at 50 µg/ml in 0.1 M NaCl, 25 mM NaPO4, pH 7.0. 6 M guanidinium HCl (Gdn-HCl) in the same buffer was titrated into the mix with excitation at 295 nm and emission monitored at 325 nm.15 Stability (DG) of the fusion protein was calculated from fluorescence denaturation curves as previously described.15 Fluorescence quenching was measured at 20°C on a Perkin-Elmer LS50B Luminescence spectrometer in a 1 ml cuvette containing 20 µg/ml protein in the same buffer. 5 M NaI in the same buffer was titrated into the mix with excitation at 295 nm. Emission was monitored from 310 to 420 nm and calculations were based on peak heights measured above the background. Studies were performed in triplicate. Circular dichroism (CD) spectra were collected on an Aviv CD-60 spectropolarimeter. Standard spectra were obtained using 0.5 mg/ml protein in 5 mM Tris-HCl, 5 mM NaCl, pH 8.0, with a Suprasil quartz cuvette with 0.2 mm path length at 19.6°C. pH studies used 5 mM NaPO4, 5mM NaCl. Below pH 4.0, 5 mM glycine, 5 mM NaCl was used with a protein concentration of 0.8 mg/ml. Spectra represent the average of 5 scans made at 0.5 nm intervals with 1 sec signal integration time and 1.5 nm bandwidth. They were corrected relative to a background scan (buffer alone) and smoothed using a third-order polynomial with 62-point averaging algorithm. Concentrations were determined for each sample by measuring absorbance at 280 nm and using extinction coefficients of 0.93 for SN16 and 1.14 for the fusion protein (calculated on the basis of the sequence shown in Fig. 1C); these values were used to calculate protein concentrations used to calculate mean residue weight ellipticity. Estimates of helix content were made using the deconvolution program PROSEC17. For measurements of ellipticity at 220 nm as a function of Gdn-HCl concentration, a 1 cm X 1 cm pathlength suprasil quartz cuvette with a small magnetic stir bar in the bottom was used. RESULTS Figure 1B shows the single stained band on SDSpolyacrylamide gel electrophoresis of the purified fusion protein following chromatography and dialysis. This band reacts strongly and uniquely with anti-SAA antibodies on Western blots (Fig. 1B). The primary aa sequence of the fusion protein is shown in Figure 1C. The calculated molecular weight is 32,899 daltons. This fusion protein exhibits high (a minimum of 10 mg/ml), prolonged solubility at 4°C for at least 8 months in several buffers of low ionic strength. We investigated structural aspects of the fusion protein using fluorescence and CD spectroscopy. Figure 2A shows individual fluorescence spectra for the fusion protein and SN as well as the difference spectrum. The SN spectrum primarily originates from a single Trp residue buried within the hydrophobic core of the protein. By contrast, the fusion contains three additional Trps and therefore has greater fluorescence intensity. Note also the significant shift of the fusion and the difference spectra to longer wavelengths (red shift), indicating possible exposure of one or more Trps in the fusion protein to the solvent. In order to investigate the extent of solvent exposure for Trp residues, the effect of NaI quenching on fluorescence intensity was measured. Figure 2B shows reduced fluorescence intensity with increasing NaI concentration, with no further changes beyond 0.2 M NaI. Figure 2C shows a modified Stern-Volmer plot18 for these data, indicating that a single Trp fluorophore is susceptible to iodide quenching under these nondenaturing conditions and, therefore, is likely to be on the surface of the parent protein. A blue shift of the fluorescence peak occurs during the titration from 343 nm (top curve, no NaI) to 339 nm (bottom curve, 200 mM NaI), supporting the notion that the Trp being quenched is associated with SAA. Figure 3A presents the denaturation profiles of the fusion protein and SN by monitoring Trp fluorescence as a function of Gdn-HCl concentration. The fusion protein shows a single transition (midpoint 5 0.72 M Gdn-HCl) which is less steep than that of SN (midpoint 5 0.82 M Gdn-HCl). The calculated free energy values indicate that the fusion protein is 2 kcal/mole less stable (5.6 kcal/mole vs. 3.6 kcal/mole) than that of the wild-type SN. The residual Trp fluorescence of the fusion protein persisted through 3.4 M Gdn-HCl (data not shown). Similar measurements were made using potassium thiocyanate (KSCN) as the denaturant (Fig. 3B). All denaturation profiles were fully reversible. Under both sets of conditions, the residual fluorescence levels were consistent with the total number of Trp residues, all in a fully denatured environment. Denaturation also was assessed by CD based on the prominent signal at 220 nm for the fusion 384 A.K. MEEKER AND G.H. SACK, JR. Fig. 3. (A) Fluorescence measurement during Gdn-HCl denaturation of fusion protein (open squares), and SN alone (closed squares). Note single transition for fusion protein. (B) Fluorescence measurement during KSCN denaturation of fusion protein (open squares) and SN alone (closed squares). The fusion protein also shows a single transition in this system. (C) CD denaturation measurements at u5220 nm during Gdn-HCl addition. Three sets of symbols correspond to three repeats of the same experiment and are superposable. Fig. 2. (A) Fluorescence spectra of fusion protein (a), SN (b), and the difference between them (c 5 a-b). (B) Quenching of fluorescence in fusion protein by NaI. NaI concentrations (top to bottom) are: 0, 20, 50, 70, 120, 200 mM. There is minimal fluorescence change at concentrations above 200 mM. (C) Modified Stern-Volmer plot based on peak height data from Figure 2B. The intercept corresponds to having one exposed Trp residue on the fusion protein. protein, presumably representing a-helical domains (see below). As shown in Figure 3C, mean specific ellipticity/residue showed a single transition during Gdn-HCl titration with a midpoint similar to that seen in Figure 3A. Thus, two complementary techniques—Trp fluorescence and molar ellipticity at 220 nm—show a single transition during denaturation at the same level of Gdn-HCl. There is no evidence for any persistent structure to the fusion protein under these denaturing conditions. CD spectra were obtained for both the fusion and SN. Figure 4A shows the CD spectra at pH 8.0. The difference spectrum (Fig. 4B) is compatible with a-helical domains in the SAA portion of the molecule. However, shallow negative signals at 208 and 222 nm are seen in the difference spectrum, possibly reflecting cancellations due to increases in aromatic components in the far UV region.19 The helix content of the difference spectrum at pH 8.0 was estimated at 33% (PROSEC). CD spectra also were obtained for both SN and the fusion protein at different pH values. The spectra for both proteins show relatively few changes from pH 8 through pH 4. Figure 4C shows the spectra for pH SAA/NUCLEASE FUSION PROTEIN 385 Fig. 4. (A) CD spectra for fusion protein and SN at pH 8.0 showing values at consecutive nm wavelengths. Closed squares 5 fusion protein; open squares 5 SN. (B) Difference CD spectrum at pH 8.0. (C) CD spectra for fusion protein and SN at pH 3.0. Note prominent changes in SN spectrum. Closed squares 5 fusion protein; open squares 5 SN. (D) Difference CD spectrum at pH 3.0. Note similarities to curve in 3B and considerable differences from SN, which is denatured under these conditions. 3.0. The SN spectrum displays a ‘‘random coil’’ pattern resulting from the loss of secondary structure following denaturation,20 but the fusion protein spectrum shows persistent structure, compatible with significant a-helical content. Figure 4D shows the difference spectrum (fusion 2 SN) at pH 3.0. Comparing the difference-derived SAA spectra at pH 8.0 and pH 3.0 shows shifting of the maximum and minima to shorter wavelengths at lower pH. The positive peak is shifted to a slightly longer wavelength in the acidic environment. The helix content at pH 3.0 was estimated at 44%. Below pH 3.0 there was little evidence for any structure of the SAA portion of the molecule and the SN remained denatured (data not shown). Trp in SN. Equilibrium denaturation experiments using either Gdn-HCl or KSCN show that Trp residues in the fusion protein undergo significant quenching with a single inflection compatible with global denaturation of the protein with increasing denaturant concentrations in both environments. This conclusion is supported by monitoring Gdn-HCl denaturation behavior with CD measurements. The shape of the difference CD spectrum at pH 8.0, including a prominent positive peak at 191 nm, suggests helical domains in the SAA region of the fusion protein. A similar peak persists in the difference spectrum obtained at pH 3.0 (see Fig. 4). SN exhibits near random coil CD spectra at pH #3.0 due to extensive denaturation20 supporting, but not proving, the notion that the residual secondary structure resides within SAA domain(s). The helix content of the fusion protein has been estimated at 33–44% for these conditions based on PROSEC deconvolutions but falls quickly below pH 3.0. These observations are important for several reasons. First, few structural data have been available for SAA because of its poor solubility. Second, the antigenic reactivity of the fusion protein with polyclonal anti-SAA antibodies indicates that at least some SAA epitopes are intact in the fusion protein. Third, the earlier proposal of Turnell et al.,21 used predictive algorithms to suggest that SAA likely contained two a-helices; this is compatible with our DISCUSSION We purified and characterized a soluble recombinant derivative of the human acute phase protein SAA expressed in E. coli. Our protein combines the full lengths of both SN and SAA, can be readily overproduced and purified, and is stably soluble to at least 10 mg/ml. It retains SAA-antigenicity as measured with polyclonal anti-SAA antiserum. Structural information was obtained by spectroscopic techniques. The fusion protein has four Trp residues, three derived from SAA. Quenching of the fluorescence with iodide indicates that at least one Trp is exposed to solvent, unlike the single, buried 386 A.K. MEEKER AND G.H. SACK, JR. CD data consistent with increased helical content for the fusion protein. Fourth, the evidence for helical domains in the SAA monomer contrasts with earlier proposals that SAA-derived amyloid fibrils are largely comprised of β-sheets.22 However, it is also possible that important structural transition(s) of the monomers may accompany their polymerization into fibrils. The CD spectra indicating substantial a-helical content in SAA through pH 3.0 are consistent with the existence of relatively stable and, possibly, independent structure for the region of this fusion protein derived from SAA. The persistence of SAA structure at pH 3.0 strengthens this notion because it is in contrast to the known denaturation of SN at low pH, which we also have reproduced in these data (Fig. 4). By contrast, fluorescence studies show smooth denaturation curves in response to increasing concentrations of either Gdn-HCl or KSCN with single inflection points, indicating that the relative independence of SN and SAA domains as suggested by the CD data at lower pH values is not maintained for all denaturants. Nevertheless, our data show that we have achieved our goal of maintaining SAA-specific epitopes in the fusion protein as measured by antigenicity. Our fusion protein was designed intentionally to minimize the distance between the SN and SAA domains in order to improve its likelihood of forming crystals in future studies. We did not plan to be able to separate the SN and SAA domains because this would regenerate insoluble SAA monomers that would have the recognized insolubility problems. We thus presume that the SAA component of the fusion protein has a structure similar to that of SAA alone, although the only direct evidence of this is the reactivity of the fusion molecule with polyclonal anti-SAA antibodies (Fig. 1B). The fusion protein also reacts with selected anti-SAA epitope-specific monoclonal antibodies (preliminary observations). The impressive solubility (a minimum of 10 mg/ ml) and ease of preparation of this fusion protein make it an ideal reagent for further structural and physiologic studies, most of which have been prevented by past problems of limited availability and poor solubility of SAA. Further structural studies should permit testing of the SAA monomer model proposed by Turnell et al.21,23 They also will improve interpretation of data about SAA fibril structure because polymerization of SAA into fibrils requires removing the C-terminal third of the protein. Understanding fibril formation is basic to the design of rational pharmacologic intervention for secondary (i.e., SAA-derived) amyloid disease and also will be essential in clarifying the role of other amyloid fibrils in conditions, including Alzheimer’s disease and diabetes, where fibrils are also prominent.24,25 Amyloid fibrils form by the aggregation of monomers in an apparently unassisted fashion, as shown by the early observation that dissociated monomers can reform fibrils in solution.22 Clearly, the structures of the monomer precursors are the basic determinants of this reaction. The studies reported here establish this fusion protein as an important reagent for investigating the basis of reactive amyloid fibril formation. SAA has amphipathic character26 and is generally isolated from the high-density lipoprotein fraction of serum. It has been proposed that SAA is an authentic apolipoprotein27,28 but this has not been established unequivocally. More recently, SAA mRNA expression has been found in atherosclerotic vascular lesions.29 Whatever its actual function in serum is, however, the lipophilicity of SAA has been established. Our fusion protein presents an opportunity to obtain structural information about such a lipophilic molecule in a novel and rigorous manner. Our observations with this derivative of SAA show that fusing a small, relatively insoluble protein to SN is a useful approach to obtaining important structural data about a protein which is otherwise difficult to study. This general approach should be useful in analyzing other small proteins which have hitherto resisted detailed analysis because of insolubility, instability, or scarcity. SN has ideal features as a fusion partner for protein structure studies because of its small size, remarkable solubility, ease of crystallization, complete crystal structure availability, and extensive study for its folding properties.10,11,30–32 The effects of deletions and mutations on SN also have been studied in detail.33–36 Multiple genetic constructs with excellent expression levels are available for SN and recombinant DNA methods can rapidly modify the basic SN sequence and/or the structure of the appended protein, permitting a remarkable range of study possibilities. In addition to its useful physical features, the enzymatic function of SN also has been useful in fusion proteins. For example, a fusion protein between SN and a retroviral Gag protein reduced infectious viral titers 20- to 60-fold.37,38 Thus, this small bacterial enzyme has been useful in different types of recombinant systems. This fusion protein should be the basis for further study of the structure of the SAA monomer. Such information is essential for understanding amyloid fibril formation. Recent studies have shown the growth of microcrystals of our fusion protein in the presence of polyethylene glycol, indicating that this recombinant may provide the substrate for crystal studies of SAA. ACKNOWLEDGMENTS We are particularly grateful to Dr. David Shortle for continuous advice and discussion, to Dr. Joel Gillespie for assistance with CD measurements and thoughtful comments, to Dr. Peter Pedersen for the use of his spectrometer, to Dr. Wesley Stites for SAA/NUCLEASE FUSION PROTEIN corroborative fluorescence measurements, to Ronald Yap for help with protein purification, to Tina Saxowsky for denaturation CD measurements, and to Mrs. Mary A. Mix for secretarial assistance. Plasmid pA1 was the kind gift of Dr. Jean Sipe. The work was supported by generous gifts from Mr. and Mrs. William M. Griffin and Mr. Daniel M. Kelly. REFERENCES 1. Kushner, I. The phenomenon of the acute phase response. Ann. N.Y. Acad. Sci. 111:39–48, 1982. 2. Benditt, E.P., Eriksen, M., Hermodsen, M.A., Ericsson, L.H. The major proteins of human and monkey smyloid substance: Common properties including unusual N-terminal amino acid sequences. FEBS Lett. 19:169–173, 1981. 3. Parmelee, D.C., Titani, K., Ericsson, L.H., Eriksen, N., Benditt, E.P., Walsh, K.A. Amino acid sequence of amyloidrelated apoprotein (apoSAA1) form human high-density lipoprotein. Biochemistry 21:3298–3303, 1982. 4. Sack, G.H. Jr. Molecular cloning of human genes for serum amyloid A. Gene 21:19–24, 1983. 5. Sack, G.H. Jr., Talbot, C.C. Jr. The human serum amyloid A gene SAA4 is a pseudogene. Biochem. Biophys. Res. Commun. 183:362–366, 1992. 6. Sack, G.H. Jr., Talbot, C.C. Jr, Seuanez, H., O’Brien, S.J. Molecular analysis of the human serum amyloid A (SAA) gene family. Scand. J. Immunol. 29:113–119, 1989. 7. Brinckerhoff, C.E., Mitchell, T.I., Karmilowicz, M.J., KluveBeckerman, B., Benson, M.D. Autocrine induction of collagenase by serum amyloid A-like and β2-microglobulin-like proteins. Science 243:655–657, 1989. 8. Sack, G.H. Jr., Zink, M.C. Serum amyloid A (SAA) gene expression in synovial cells in retroviral arthritis. Am. J. Pathol. 141:525–529, 1992. 9. Sack, G.H. Jr., Talbot, C.C. Jr. The human serum amyloid A (SAA)-encoding gene GSAA1: Nucleotide sequence and possible autocrine-collagenase-inducer function. 84:509– 515, 1989. 10. Tucker, P.W., Hazen, E.E. Jr., Cotton, F.A. Staphylococcal nuclease reviewed: A prototypic study in contemporary enzymology. I. Isolation, physical and enzymatic properties. Mol. Cell. Biochem. 22:66–77, 1978. 11. Loll, P.J., Lattman, E.E. The crystal structure of the ternary complex of staphylococcal nuclease, Ca21, and the inhibitor pdTp, refined at 1.65 angstroms. Proteins 5:183– 201, 1989. 12. Sack, G.H. Jr., Aquilina, P., Shortle, D., Meeker, A. A recombinant system for serum amyloid A (SAA) gene and protein expression. In: ‘‘Amyloid and Amyloidosis.’’ Kisilevsky, R. et al. (eds.). New York: Parthenon Press, 1994: 146–148. 13. Sipe, J.D., Colten, H.R., Goldberger, G., Edge, M.D., Tack, B.F., Cohen, A.S., Whitehead, A.S. Human serum amyloid A (SAA): Biosynthesis and postsynthetic processing of preSAA and structural variants defined by complementary DNA. Biochemistry 24:2931–2936, 1985. 14. Shortle, D. A genetic system for analysis of staphylococcal nuclease. Gene 22:181–189, 1983. 15. Shortle, D. Guanidine hydrochloride denaturation studies of mutant forms of staphylococcal nuclease. J. Cell. Biochem. 30:281–289, 1986. 16. Fuchs, S., Cuatrecasas, P., Anfinsen, C.B. An improved method for the purification of staphylococcal nuclease J. Biol. Chem. 242:4768–4770, 1967. 17. Chang, C.T., Wu, C.-S.C., Yang, Y.T. Circular dichroism analysis of protein conformation: Inclusion of the β turns. Anal. Biochem. 91:13–31, 1978. 387 18. Lakowicz, J.R. ‘‘Principles of Fluorescence Spectroscopy’’ New York: Plenum Press, 1983:260–273. 19. Johnson, W.C. Jr. Secondary structure of proteins through circular dichroism spectroscopy. Ann. Rev. Biophys. Biophys. Chem. 17:145–166, 1988. 20. Schechter, A.N., Chen, R.F., Anfinsen, C.B. Kinetics of folding of staphylococcal nuclease. Science 167:886–887, 1970. 21. Turnell, W., Sarra, R., Glover, I.D., Baum, J.O., Caspi, D., Baltz, M.L., Pepys, M.B. Secondary structure prediction of human SAA1: Presumptive identification of calcium and lipid binding sites. Mol. Biol. Med. 3:387–407, 1986. 22. Glenner, G.G., Ein, D., Eanes, E.D., Bladen, H.A., Terry, W. Page, D. Creation of ‘‘amyloid’’ fibrils from Bence Jones proteins in vitro. Science 174:712–714, 1971. 23. Turnell, W., Sarra, R., Baum, J.O., Caspi, D., Baltz, M.L., Pepys, M.B. X-ray scattering and diffraction by wet gels of AA amyloid proteins. Mol. Biol. Med. 3:409–424, 1986. 24. Selkoe, D. Aging, amyloid, and Alzheimer’s disease. New Engl. J. Med. 320:1484–1487, 1989. 25. Lorenzo, A., Razzaboni, B., Weir, G.C., Yanker, B.A. Pancreatic islet cell toxicity of amylin associated with type 2 diabetes mellitus. Nature 368:756–760, 1994. 26. Segrist, J.P., Pownall, H.J., Jackson, R.L., Glenner, G.G., Pollock, P.S. Amyloid A: Amphipathic helices and lipid binding. Biochemistry 15:3187–3191, 1976. 27. Skogen, B., Borresen, A.L., Natvig, J.B., Berg, K., Michaelsen, T.E. High-density lipoprotein as a carrier for amyloid-related protein SAA in rabbit serum. Scand. J. Immunol. 10:39–45, 1979. 28. Benditt, E.P., Eriksen, N., Hansen, R.H. Amyloid protein SAA is an apoprotein of mouse plasma high density lipoprotein. Proc. Natl. Acad. Sci. USA 76:4092–4096, 1979. 29. Meek, R.L., Urieli-Shoval, S., Benditt, E.P. Expression of apolipoprotein serum amyloid A mRNA in human atherosclerotic lesions and cultured vascular cells: Implications for serum amyloid A function. Proc. Natl. Acad. Sci. USA 91:3186–3190, 1994. 30. Tucker, P.W., Hazen, E.E. Jr., Cotton, F.A. Staphylococcal nuclease reviewed. II. Solution studies of the nucleotide binding site and the effects of nucleotide binding. Mol. Cell. Biochem. 23:3–16, 1979. 31. Tucker, P.W., Hazen, E.E. Jr., Cotton, F.A. Staphylococcal nuclease reviewed. III. Correlation of the three-dimensional structure with the mechanisms of enzyme action. Mol. Cell. Biochem. 23:67–86, 1979. 32. Tucker, P.W., Hazen, E.E. Jr., Cotton, F.A. Staphylococcal nuclease reviewed. IV. The nuclease as a model for protein folding. Mol. Cell. Biochem. 23:131–142, 1979. 33. Shortle, D., Meeker, A.K. Mutant forms of staphylococcal nuclease with altered patterns of guanidine hydrochloride and urea denaturation. Proteins 1:81–89, 1986. 34. Shortle, D., Meeker, A.K. Residual structure in large fragments of staphylococcal nuclease: Effects of amino acid substitutions. Biochemistry 28:936–944, 1989. 35. Shortle, D., Stites, W.E., Meeker, A.K. Contributions of the large hydrophobic amino acids to the stability of staphylococcal nuclease. Biochemistry 29:8033–8041, 1990. 36. Green, S.M., Meeker, A.K., Shortle, D. Contributions of the polar, uncharged amino acids to stability of staphyloccal nuclease: Evidence for mutational effects on the free energy of the denatured state. Biochemistry 31:5717–5728, 1992. 37. Natsoulis, G., Boeke, J.D. New antiviral strategy using capsid-nuclease fusion proteins. Nature 352:632–635, 1991. 38. Schumann, G., Qin, L., Rein, A., Natsoulis, G., Boeke, J.D. Therapeutic effect of Gag-nuclease transport on retrovirusinfected cell cultures. J. Virol. 70:4329–4337, 1996.