SUPPLEMENTAL DATA Paradoxical Effects of Serum Amyloid A on the Lipoprotein Oxidation Suggest a New Antioxidant t Function for SAA Shobini Jayaraman, Christian Haupt, and Olga Gursky Figure S1. Lipid peroxidation kinetics by copper in native HDL, nhdl (black), and in SAA- 2 enriched HDL, SAA HDL(total) (gray). Lipoproteins were prepared as described in Figure legend. The characteristic lag followed by a rapid propagation phase (with the maximal rate, V max ) and a saturation phase (with the maximal magnitude, OD max ) are indicated. An extended time scale (0-300 min) for the SAA-enriched HDL shows that the maximal magnitude of OD at saturation, OD max, is much smaller than that in nhdl. 1
Figure S2. Lipid and protein modification ns in HDL oxidized to various stages show dose- to dependent effects of SAA. HDL that contained various amounts of SAA, from 0.5:1 (black) 2:1 SAA:apoA-I molar ratios (red), were oxidized by Cu 2+ to various stages as described in Figure 2A. For the protein ratio of 0:1 or 0.5:1, these stages correspond to the pre-transition (15 min), transition (45 min) and post-transitional regions (100 min) in the OD 234 curve (Fig. 2A) ). (A) Lipid modifications at various oxidation stages were monitored by UV/vis absorption spectra at 25 C reporting on the carotenoid consumption.. The fingerprint absorption peaks of carotenoids at 485, 458 and 4344 nm progressively disappear upon oxidation at 0.5:1 but not at 2:1 ratio. Consequently, at 2:1 ratio SAA effectively delays carotenoid consumption in HDL core. (B) Protein modifications at various oxidation stages were monitored by the intrinsic Trp fluorescence of HDL. The spectra were recorded as described in Methods. At 0.5:1 SAA:apoA-I ratio, progressive decrease in the emission intensity of Trp together with the red shift in the emission maximum indicate protein modifications such as Tpr oxidation mediated by the products of lipid oxidation. At 2:1 SAA:apoA-I ratio, these spectral changes were greatly retarded. Therefore, at 2:1 ratio SAA delays oxidative modifications to HDL proteins. 2
Figure S3. Characterization of the protein released from oxidized SAA HDL. SEC fraction of the protein released from mildly SAA HDL at 2:1 SAA:apoA-I molar ratio (marked free in Figure 3) were collected and analyzed. (A) SDS PAGE of the protein released from HDL shows SAA and apoa-i bands. Quantification of the bands from three independent gels was carriedd out as previously described (30); the results showed approximately 40% apoa-i and 60% SAAA by weight.. (B) MALDI-TOF analysis of the protein released from ox SAA HDL. The spectrum showss two peaks with the masss of 11,656 Da and 28,135 Da corresponding to Met-oxidized forms of SAA and apoa-i, respectively. The spectra of non-oxidized SAA and apoa-i are shown for comparison (bottom panel). 3
Figure S4. Statistical analysis of the effects of SAA on the oxidation of HDL and LDL. HDL incubation mixtures, termed HDL(total), containedd various amounts of SAA as shown in Figure 4A, from 0:1 to 4:1 SAA: :apoa-i mol ratios. LDL were eitherr alone or in the presence of SAA or apoa-i at 1: 1 SAA:apoB or apoa-i: :apob weight ratios (Fig. 4B). These samples were oxidized by Cu 2+ as described in Figure 4 and the OD 234 4 data were recorded (Fig. 4A). The HDL data showed a complete sigmoidal transition at 0:1, 0.5:1 or 1:1 SAA:apoA-I molar ratios (at higher SAA;apoA-I ratios the transition was incompletee and hence, was not analyzed further). The experiments were repeated 3 times. The transitionn midpoint, t ½ (in min) and the maximal rate, V ma ax (in OD units per min) were determined from the sigmoidal fitting of the OD 234 (t) data and its first derivative. The results were subjected too a statistical analysis as described in Methods. LDL oxidation data were analyzed similarly. The results of the statistical analysis for HDL (A) and LDL (B) are shown. The values of t ½ and V max for HDL and LDL are presented as a mean of three independent measurements, with the error bars showing the standard deviation of the mean. P-values are indicated: * (P<0.05); ** (P<0.01); *** (P< <0.001). 4
mox SAA HDL MPO/H 2 O 2 /Cl mox SAA HDL mox SAA HDL ~40 kda Anti apoa-i Anti SAA Anti apoa-ii Figure S5. Identification of protein cross-links formed upon oxidation of SAA-containing HDL by MPO/H 2 O 2 /Cl -. SAA-containing HDL (either the total mixture or isolated particles, with numbers showing SAA:apoA-I mol:mol ratios) were mildly oxidized (mox) and were subjected to SDS PAGE followed by immunoblotting for apoa-i, SAA and apoa-ii. Unmodified native HDL (nhdl) are shown for comparison. Arrow shows the position of the ~40 kda band seen in the blots for apoa-i and SAA but not apoa-ii, which corresponds to the SAA apoa-i heterodimer. This heterodimer was observed only in SAA-enriched (2:1 and 2:1(total)) and oxidized particles (mox SAA HDL). 5
H 2 O 2 + SAA HDL kda St 75 50 37 25 20 15 10 SDS PAGE SAA 3 apoa-i SAA 2 apoa-ii SAA Figure S6. The SAA apoa-i heterodimer does not form in the SAA-enriched and oxidized HDL in the absence of OCl - /Cl -. SAA HDL, which were prepared at indicated SAA:apoA-I molar ratios, have been mildly oxidized using MPO/H 2 O 2 in the absence of Cl - as described in Methods. The oxidative species in this system is the peroxide radical. SDS PAGE shows formation of the SAA homodimer and homotrimer, but no SAA apoa-i heterodimer. This observation, taken together with the data in Figures 5 and 6, indicates that chlorination mediates formation of the SAA apoa-i heterodimer upon HDL oxidation. 6
Figure S7. The SAA apoa-i heterodimer forms only on HDL but not in free proteins in solution. ApoA-I, SAA and their incubation mixture, apoa-i+saa, were subjected to mild oxidation using OCl -, and the oxidative protein cross-linking was analyzed by immunoblotting for apoa-i (A) and for SAA (B). The results showed formation of apoa-i homodimer (A) and SAA homodimer and homotrimer. The position for the SAA apoa-i heterodimer, which did not form in free proteins, is shown by arrow. 7
Figure S8. The impact of SAA enrichment and oxidation on HDL remodeling. (A) Native PAGE of non-oxidized and of minimally oxidized HDL. SAA HDL were prepared at 2:1 molar ratio of SAA to apoa-i. Minimal oxidation promoted protein dissociation from SAAcontaining (mox SAA HDL) but not from SAA-free particles (mox HDL). (B) SDS PAGE of free SAAA that was either unmodified or was oxidized with MPO/H 2 2O 2 in the absence of chloride. No significant SAA cross-linking was detected. (C) SDS PAGE of SAA that was either unmodified or was oxidized by using OCl - or MPO/H2O 2 /Cl -. SAAA homodimer and homotrimer are indicated. 8
Figure S9. Effects of SAA enrichment and mild oxidation by OCl - on the spontaneous biophysical remodeling of HDL at near-physiologic conditions. The lipoproteins were SAA-free (marked 0:1) or SAA-enriched at 2:1 SAA:apoA-I molar ratio (2:1); a subset of these lipoproteins was mildly oxidized by OCl - as described in Methods (mox, panel B). All lipoproteins weree then incubated for 12 h at 37 C in 20 mm MES buffer at ph5.5. SEC analysis of unoxidized (A) and oxidized lipoproteins (B) shows that combined effects off SAA enrichment and oxidation promote dissociation of free protein (gray lines 2:1 in panel B). 9