In Vitro and in Vivo Interactions of Homocysteine with Human Plasma Transthyretin*

Hyperhomocysteinemia is an independent risk factor for cardiovascular disease and an emerging risk factor for cognitive dysfunction and Alzheimer's disease. Greater than 70% of the homocysteine in plasma is disulfide-bonded to protein cysteine residues. The identity and functional consequences of protein homocysteinylation are just now emerging. The amyloidogenic protein transthyretin (prealbumin), as we now report, undergoes homocysteinylation at its single cysteine residue (Cys10) both in vitro and in vivo. Thus, when human plasma or highly purified transthyretin was incubated with 35S-l-homocysteine followed by SDS-PAGE and PhosphorImaging, two bands corresponding to transthyretin dimer and tetramer were observed. Treatment of the labeled samples with β-mercaptoethanol prior to SDS-PAGE removed the disulfide-bound homocysteine. Transthyretin-Cys10–S–S–homocysteine was then identified in vivo in plasma from normal donors, patients with end-stage renal disease, and homocystinurics by immunoprecipitation and high performance liquid chromatography/electrospray mass spectrometry. The ratios of transthyretin-Cys10–S–S–homocysteine and transthyretin-Cys10–S–S–sulfonate to that of unmodified transthyretin increased with increasing homocysteine plasma concentrations, whereas the ratio of transthyretin-Cys10–S–S–cysteine to that of unmodified transthyretin decreased. The hyperhomocysteinemic burden is thus reflected in the plasma levels of transthyretin-Cys10–S–S–homocysteine, which in turn may contribute to the pathological consequences of amyloid disease.

Individuals with elevated plasma total homocysteine (tHcy) 1 (hyperhomocysteinemia) are at greater risk for cardiovascular disease (1), and the prognosis for patients with cardiovascular disease and other diseases in combination with the highest levels of tHcy is poor (2)(3)(4). Recent studies (5)(6)(7)(8) suggest that hyperhomocysteinemia is also a risk factor for Alzheimer's disease and other disorders of cognitive dysfunction. Most of the homocysteine in circulation (Ͼ70% of tHcy) is disulfidelinked to albumin and other plasma proteins (9 -11). The remaining free homocysteine is found as low molecular weight disulfide forms such as homocystine and homocysteine-cysteine mixed disulfide (12). Less than 1% of tHcy is found as free, reduced (i.e. -SH) form (13). Protein-homocystamide (homocysteine-N-protein), the reaction product formed between a protein lysine residue and homocysteine thiolactone, is also found in circulation (14).
The functional consequences of protein homocysteinylation are beginning to emerge. For example, in vitro studies have shown that homocysteinylation of the Cys 9 residue of annexin II, the endothelial cell surface docking protein for tissue plasminogen activator, inhibits the binding (15,16). Homocysteinylation of factor Va in vitro makes it resistant to inactivation by activated protein C (17). Homocysteinylation appears to activate latent elastolytic metalloproteinase pro-MMP-2 by disulfide bond formation with the "cysteine switch" on the propeptide (18). Recently, we reported that homocysteine binds to the fibrin-binding domain of plasma fibronectin in vitro and inhibits its ability to bind fibrin (19). Taken together, these in vitro studies suggest that post-translational modification of proteins by homocysteine may have important functional consequences.
Transthyretin (prealbumin) is a 13.8-kDa protein that is synthesized predominantly in the liver and secreted into plasma. As a homotetramer transthyretin binds and transports the hormone thyroxine and the retinol-binding protein-retinal complex (20). Transthyretin has been implicated in the formation of amyloid deposits in familial transthyretin amyloidosis and senile systemic amyloidosis (21,22). Familial transthyretin amyloidosis is an autosomal dominant disorder involving the deposition of transthyretin as amyloid fibrils in tissues and * This work was supported by National Institutes of Health Grants P41 RR 10888 (to C. E. C.) and R01 HL 52234 (to D. W. J.). This work was initially presented at the 12th International Symposium on Pteridines and Folates, June 17-22, 2001 TTR-Cys 10 -S-S-Cys, transthretin-Cys 10 -S-S-cysteine; TTR-Cys 10 -S-S-CysGly, transthyretin-Cys 10 -S-S-cysteinylglycine; TTR-Cys 10 -S-SG, transthyretin-Cys 10 -S-S-glutathione; TTR-Cys 10 -S-S-Hcy, transthyretin-Cys 10 -S-S-homocysteine; TTR-Cys 10 -S-SO 3 H, transthyretin-Cys 10 -S-sulfonate. organs. The deposits may contain wild type transthyretin, transthyretin variants, and/or their fragments. Variant tetrameric transthyretin is destabilized by certain amino acid substitutions, which form self-associating amyloidogenic intermediates and amyloid fibrils that deposit in tissues and organs (23)(24)(25). Senile systemic amyloidosis is a nonhereditary disorder that affects about 25% of individuals over 80 years old (26). Here the amyloid fibrils and deposits are usually made up of wild type transthyretin and/or its fragments, which are found mainly in the heart.
Monomeric transthyretin has a single cysteine residue at position 10. In the normally folded tetrameric protein, the Cys 10 residues are in exposed sites at the start of the helical regions. Because the Cys 10 residue of transthyretin can conjugate with cysteine and other sulfur-containing ligands (27,28), we hypothesized that it could be homocysteinylated and, with increasing homocysteine burden, stable transthyretin-Cys 10 -S-S-homocysteine might become the predominant form of circulating transthyretin. In vitro as well as in vivo evidence in support of this hypothesis is now reported.
␤ 2 -Microglobulin is found in amyloid deposits of patients with end-stage renal disease who are on dialysis. It too is a potential candidate for homocysteinylation because this 11.8-kDa protein has a ␤-sandwich structure stabilized by a singe cross-sheet disulfide bond formed by Cys 25 and Cys 80 (29). It theoretically could be homocysteinylated in a thiol-disulfide exchange reaction by homocysteine thiolate anion. Moreover, nearly all dialysis patients with end-stage renal disease have elevated levels of ␤ 2 -microglobulin (30) and hyperhomocysteinemia (31). It would be of interest to determine whether ␤ 2 -microglobulin might be another carrier of homocysteine in human plasma. 35 S-L-Homocysteine thiolactone was synthesized from L-[ 35 S]methionine by a slight modification of the method of Mudd et al. (32) and purified as described previously (11). 35 S-L-Homocysteine (500 M final concentration; specific activity 50 Ci/mol) was prepared from 35 S-L-homocysteine thiolactone (33). The thiol content of L-homocysteine was determined using Ellman's reagent (34). All experiments were conducted with fresh preparations of L-homocysteine and 35 S-Lhomocysteine. Purified human transthyretin was obtained from Lee Scientific (St. Louis, MO), and purified ␤ 2 -microglobulin was obtained from Sigma. Human plasma was obtained from healthy donors, subjects with chronic renal failure, and subjects with homocystinuria using protocols approved by the Institutional Review Boards of the Cleveland Clinic Foundation. All other reagents and solvents of analytical grade or better were obtained from Sigma.

Materials-
In Vitro Binding of 35 S-L-Homocysteine to Human Plasma Proteins, Purified Transthyretin, and Purified ␤ 2 -Microglobulin- 35 S-L-Homocysteine (500 M final concentration) was incubated with 50% human plasma in 0.050 M TES buffer at pH 7.4, or with purified human transthyretin (1 mg/ml in 0.050 M TES buffer, pH 7.4), or with purified human ␤ 2 -microglobulin (1 mg/ml in 0.050 M TES buffer, pH 7.4) for 5 h at 37°C. Plasma proteins, transthyretin, and ␤ 2 -microglobulin were then precipitated with 1.5 M perchloric acid, washed three times with 1.5 M perchloric acid, and dissolved in nonreducing SDS-PAGE sample buffer (0.062 M Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 0.2% bromphenol blue). To one-half of each sample, 3 l of ␤-mercaptoethanol was added, and the sample was heated at 100°C for 5 min to reduce disulfide bonds. Aliquots of the ␤-mercaptoethanol-treated and untreated samples were applied to a 10% SDS-polyacrylamide gel, and the electrophoresis was carried out according to standard protocol (35). The gels were dried and analyzed using a PhosphorImager to identify 35 Slabeled proteins.
Binding of L-Homocysteine to Transthyretin-L-Homocysteine (500 M final concentration) was added to purified transthyretin (180 M final concentration) in 0.050 M TES buffer, pH 7.4, and the reaction mixture was incubated at 37°C for 2 h in a shaking water bath. Aliquots were withdrawn after 30 min and added directly to tubes containing 0.1 ml of 1.5 M perchloric acid to precipitate the transthyretin. The tubes were vortexed, incubated on ice for 10 min, and centri-fuged at 12,000 rpm for 10 min. The protein pellet was washed 3 times with perchloric acid and was solubilized in Tris-HCl (0.50 M, pH 8.5).
The concentrations of S-cysteinylated and S-homocysteinylated transthyretin were determined as described previously (11). Isolation and Purification of Transthyretin from Human Serum-Transthyretin was immunoprecipitated from human serum as described previously (36). Briefly, 100 l of serum was treated with 80 l of goat anti-human transthyretin antiserum (Diasorin, Stillwater, MN) for 12 h at 37°C. The mixture was centrifuged at 14,000 rpm for 20 min at room temperature. The supernatant was removed, and the pellet was washed three times with 100 l of water, dried in a Speedvac concentrator, and then stored at Ϫ80°C. The transthyretin-antibody immunoprecipitate pellets were thawed, dissolved in 80:10:10 (v/v/v) water/ acetonitrile/acetic acid, and passed through a Millipore Microcon YM-100 centrifugal filter (100,000 molecular weight cut-off) to remove the antibody. The filtrate was then applied to an analytical Vydac C-4 HPLC column (25 ϫ 0.46 cm, 5-m particle size) and eluted at 0.75 ml/min over 30 min using a gradient of 40 -85% acetonitrile. Transthyretin and its Cys 10 conjugates eluted between 52 and 54% acetonitrile. The solvent mixture was removed from the protein using a Speedvac concentrator.
Mass Spectrometry of Transthyretin and Transthyretin-Cys 10 Conjugates-The masses of the intact transthyretin-related components were determined using electrospray-ionization mass spectrometry (ESI MS). Briefly, each HPLC-purified transthyretin sample was dissolved in 60 l of 50:50:0.01 methanol/water/formic acid (ESI buffer). A 5-l aliquot was taken and diluted in the ESI buffer; 3 l of this transthyretin solution was nanosprayed (37) into an Applied Biosystems/MDS-SCIEX QSTAR Pulsar i ESI quadrupole/orthogonal acceleration time-of-flight mass spectrometer. The capillary potential was increased slowly from 0 up to 1.2 kV until a stable ion current was observed. The declustering potential was held at 35 V. The instrument was calibrated using the [M ϩ 2H] 2ϩ ion (m/z 879.9704) and [M ϩ 4H] 4ϩ ion (m/z 440.4892) of porcine renin substrate tetradecapeptide. After calibration, this instrument was capable of achieving ϳ10 ppm mass accuracy with a minimum resolution of 1:9000 (full-width half-maximal). The relative abundance of the transthyretin-related components was determined using the abundance at the apex of each peak shown in the deconvoluted ESI mass spectra.
Determination of tHcy-tHcy was determined as described previously (38). Briefly, the plasma samples were reduced with sodium borohydride to liberate free reduced homocysteine, cysteine, cysteinylglycine, and glutathione. These reduced products were then derivatized with monobromobimane. The protein was removed by perchloric acid precipitation and centrifugation. The thiol-bimane adducts were separated by reversed-phase HPLC and detected fluorometrically (38).

RESULTS
The objectives of this study were to determine whether homocysteine forms disulfide conjugates with the amyloid proteins transthyretin and ␤ 2 -microglobulin under in vitro conditions and, more importantly, in vivo. Human plasma from healthy donors and purified transthyretin were incubated with 35 S-Lhomocysteine for 5 h at 37°C. The samples, before and after  2) and after (lanes 3 and 4) treatment with ␤-mercaptoethanol (BME), were subjected to SDS-PAGE and Phospho-rImaging analysis as described under "Experimental Procedures." treatment with ␤-mercaptoethanol, were subjected to SDS-PAGE and analyzed by PhosphorImaging. As shown in Fig. 1, lane 1, incubation of normal human plasma with 35 S-L-homocysteine produced faint bands corresponding to the transthyretin dimer (27.5 kDa) and tetramer (55 kDa). The heavily labeled band in lane 1 is albumin, which forms a disulfide bond with homocysteine (11). When purified human transthyretin was incubated with 35 S-L-homocysteine (Fig. 1, lane 2), labeling was found primarily on the dimeric form of transthyretin. Treatment of the plasma or purified transthyretin sample with ␤-mercaptoethanol prior to SDS-PAGE analysis resulted in the removal of homocysteine from transthyretin (Fig. 1, lanes 3 and  4), suggesting that the homocysteine was indeed disulfidelinked to the protein. These in vitro experiments provide strong evidence that L-homocysteine readily reacts with Cys 10 , the only cysteine residue found in transthyretin. In contrast, when human plasma from healthy donors and purified ␤ 2 -microglobulin were incubated with 35 S-L-homocysteine for 5 h at 37°C, subjected to SDS-PAGE, and then analyzed by Phospho-rImaging, no bands corresponding to homocysteinylated ␤ 2microglobulin were visualized, suggesting that homocysteine does not react with the single disulfide bond of the molecule (data not shown). Our attention was then focused on transthyretin alone.
Human plasma from a healthy donor was treated with increasing concentrations of L-homocysteine (0 -500 M) for 5 h at 37°C. Plasma transthyretin was then immunoprecipitated and purified by reversed-phase HPLC. The masses of the transthyretin-related components were determined using ESI MS. The deconvoluted ESI mass spectra of immunoprecipitated transthyretin and purified by HPLC from the plasma of a healthy individual without and with the addition of 500 M homocysteine is shown in Fig. 2, A and B, respectively. The ESI mass spectrum of the transthyretin sample in the absence of homocysteine ( Fig. 2A) showed peaks corresponding to the unmodified transthyretin molecule (mass ϭ 13,761 Da) and the Cys 10 adducts for S-sulfonate (TTR-Cys 10 -S-SO 3 H, mass ϭ 13,841 Da), S-cysteine (TTR-Cys 10 -S-S-Cys, mass ϭ 13,880 Da) Scysteinylglycine (TTR-Cys 10 -S-S-CysGly, mass ϭ 13,937 Da), and S-glutathione (TTR-Cys 10 -S-SG, mass ϭ 14,067 Da). S-Homocysteine of transthyretin (TTR-Cys 10 -S-S-Hcy, mass ϭ 13,894) was detected as a minor component. In contrast, the ESI mass spectrum of the transthyretin sample in the presence of 500 M homocysteine (Fig. 2B) displayed a major peak corresponding to TTR-Cys 10 -S-S-Hcy, in addition to peaks corresponding to the transthyretin-related components seen in Fig. 2A.
These ESI mass spectra were analyzed to determine the relative abundance of the TTR-Cys 10 -S-S-Cys and TTR-Cys 10 -S-S-Hcy adducts to that of the unmodified transthyretin molecule. These ratios were plotted as a function of Lhomocysteine concentration in the in vitro dose-response study (Fig. 3). In Fig. 3A, the ratio of the relative abundance of TTR-Cys 10 -S-S-Cys to that of unmodified transthyretin decreased from 1.71 in normal plasma (0 M exogenous homocysteine) to 1.09 in the presence of 50 M exogenous homocysteine. The ratio then increased up to 1.21 when 250 M homocysteine was added. The ratio then substantially increased when the transthyretin was incubated with 500 M homocysteine (Fig.  3A). This substantial increase is attributed to the increase in the concentration of free cysteine in plasma due to the interaction of homocysteine with albumin-Cys 34 -S-S-Cys (S-cys- teine albumin), wherein albumin-Cys 34 -S-S-Hcy is formed releasing free cysteine in a two-step process (11). In contrast, the ratio of TTR-Cys 10 -S-S-Hcy to that of the unmodified transthyretin remained essentially unaltered up to 50 M added homocysteine (Fig. 3B). The ratio then increased, albeit not substantially, up to 250 M homocysteine. Like TTR-Cys 10 -S-S-Cys (Fig. 3A), the ratio of the TTR-Cys 10 -S-S-Hcy to that of the unmodified transthyretin substantially increased at 500 M homocysteine (Fig. 3B). However, the ratios of the TTR-Cys 10 -S-SO 3 H, TTR-Cys 10 -S-S-CysGly, and TTR-Cys 10 -S-S-G to that of the unmodified TTR remained relatively constant over the entire concentration range of L-homocysteine (data not shown).
We had reported earlier (11) that during the reaction of homocysteine with human serum albumin, homocysteine strips the cysteine attached to Cys 34 of albumin (albumin-Cys 34 -S-S-Cys), resulting in the formation of albumin thiolate anion and homocysteine-cysteine mixed disulfide in the first step of the reaction. To determine whether a similar reaction is involved here, we incubated 500 M homocysteine with 180 M purified transthyretin and found that after 30 min, homocysteine had stripped almost all of the cysteine from TTR-Cys 10 -S-S-Cys (Fig. 4). During the same time period, only 10 M of TTR-Cys 10 -S-S-Hcy was formed, indicating that homocysteine attacks the S-cysteine sulfur on TTR-Cys 10 -S-S-Cys, resulting in the formation of TTR-Cys 10 -S Ϫ thiolate anion and homocysteine-cysteine mixed disulfide. These observations are similar to those reported by us previously (11) for the reaction of homocysteine with albumin.
All the results clearly suggest that homocysteine readily forms a disulfide conjugate with transthyretin in vitro. Next, we wanted to determine whether transthyretin is a carrier of homocysteine in vivo. Transthyretin from plasma of healthy individuals and patients with hyperhomocysteinemia due to chronic renal failure or homocystinuria was immunoprecipitated, HPLC-purified, and subjected to the same analysis as those used in the in vitro studies. Representative ESI deconvoluted mass spectra of the transthyretin isolated from the plasma of a patient with end-stage renal disease (tHcy ϭ 20.7 M) and a homocystinuric patient (tHcy ϭ 434 M) are shown in Fig. 5, A and B, respectively. In the plasma from the patient with end-stage renal disease, the transthyretin existed predominantly as in S-cysteine form (Fig. 5A). Only a small amount of the S-homocysteine transthyretin was detected (Fig.  5A). In contrast, the majority of the transthyretin isolated from the plasma of the homocystinuric patient was S-sulfonate (Fig.  5B). Additionally, the S-homocysteine transthyretin was now a major adduct, whereas the amount of the S-cysteine transthyretin was greatly reduced (Fig. 5B).
Like the results obtained for the in vitro studies, the ratios of the relative abundance of the S-sulfonate, S-cysteine, and Shomocysteine adducts of the transthyretin to that of the unmodified transthyretin molecule were plotted as function of L-homocysteine concentration (Fig. 6). The ratio of the relative abundance of the S-sulfonate transthyretin to that of the unmodified transthyretin increased with increasing concentrations of plasma homocysteine and plateaued at about 200 M homocysteine (Fig. 6A). In contrast, the ratio of the S-cysteine transthyretin to that of the unmodified transthyretin from the hyperhomocysteinemic patients decreased up to ϳ200 M plasma homocysteine and then increased (Fig. 6B). However, the ratio of the S-homocysteine transthyretin to that of the unmod- After 30 min, aliquots were withdrawn; the transthyretin was precipitated using perchloric acid, and the concentrations of TTR-Cys 10 -S-S-Cys and TTR-Cys 10 -S-S-Hcy were determined using HPLC with fluorescence detection as described previously (38). ified transthyretin increased initially to a small extent and then there was a substantial increase after the 200 M plasma homocysteine concentration (Fig. 6C). Overall, the ratio of Cys 10conjugated transthyretin to Cys 10 -free transthyretin increased linearly as the concentration of tHcy increased (Fig. 6D). DISCUSSION These studies show that L-homocysteine reacts with transthyretin in human plasma to form a stable covalent adduct both in vitro and in vivo. Transthyretin is the third plasma protein, after albumin (9 -11) and fibronectin (19), to be identified as a carrier of homocysteine in vivo. Because transthyretin contains only one cysteine residue (Cys 10 ), the homocysteine adduct must be TTR-Cys 10 -S-S-Hcy. Earlier reports (27,28,36,39) have identified other transthyretin-Cys 10 adducts as S-sulfonate, S-cysteine, S-cysteinylglycine, and S-glutathione. However, to our knowledge, this is the first report demonstrating the presence of S-homocysteine transthyretin in normal human serum. The relatively low abundance of S-homocysteine transthyretin in normal human serum probably explains why it was not detected in earlier studies. It should be noted that while this work was under review, Sass et al. (40) also identified TTR-Cys 10 -S-S-Hcy in the plasma and serum from hyperhomocysteinemic individuals.
Our in vitro studies show that homocysteine displaces cysteine from TTR-Cys 10 -S-S-cysteine (Fig. 4) to form homocysteine-cysteine mixed disulfide (Hcy-S-S-Cys) and transthyretin thiolate anion (TTR-Cys 10 -S Ϫ ) (Reaction 1), Hcy™S Ϫ ϩ TTR-Cys 10 ™S™S™Cys 3 TTR™Cys 10 ™S Ϫ ϩ Hcy™S™S™Cys REACTION 1 which is consistent with our studies on the interaction of homocysteine with albumin-Cys 34 -S-S-cysteine (11,41). (About one-third of the albumin molecules in normal plasma are cysteinylated at Cys 34 (42).) When normal human plasma was treated with increasing concentrations of L-homocysteine, the ratio of the relative abundance of S-homocysteine transthyretin to that of the unmodified transthyretin remained relatively constant up to 50 M added L-homocysteine. This was followed by a small increase in the ratio up to about 200 M added L-homocysteine. However, at higher concentrations of added homocysteine (Ͼ250 M), the formation of S-homocysteine transthyretin increased substantially (Fig. 3B).
This phenomenon can be explained if we consider that, in plasma, albumin is the most abundant protein, accounting for ϳ50 -60% of total plasma proteins. We propose that transthyretin thiolate anion and the albumin thiolate anion competitively attack the low molecular weight disulfides homocysteinecysteine mixed disulfide or homocystine. Because the concentration of albumin is about 100 times greater than that of transthyretin in plasma, homocysteinylation of albumin would be the predominant reaction. However, the in vivo binding capacity of plasma albumin for homocysteine is ϳ150 -200 M (43,44). We determined that the in vitro binding capacity of plasma albumin for homocysteine was similar (10,11). Therefore, when the binding capacity of albumin for homocysteine is reached, homocysteine will then react with transthyretin. This explains the substantial increase in the formation of S-homocysteine transthyretin when exogenously added homocysteine exceeds 200 M.
The in vivo study results are similar. The in vivo ratio of the relative abundance of S-homocysteine transthyretin to that of unmodified transthyretin increased to a small extent with increasing concentrations of tHcy (Fig. 6C). However, at homocysteine concentrations Ͼ200 M, the ratio increased dramatically. In contrast, the ratio of the relative abundance of the S-cysteine transthyretin to that of the unmodified transthyretin decreased initially and then increased at higher homocysteine concentrations (Fig. 6B). Based on these results, we propose that homocysteine, upon entering circulation, preferentially strips cysteine from both albumin-Cys 34 -S-S-Cys and transthyretin-Cys 10 -S-S-Cys forming the respective protein thiolate anions and homocysteine-cysteine mixed disulfide. The protein thiolate anions then react with the mixed disulfide to form the respective S-homocysteine protein adducts and cysteine thiolate anion (Reaction 2). Protein™S Ϫ ϩ Hcy™S™S™Cys 3 Protein™S™S™Hcy ϩ Cys™S Ϫ

REACTION 2
Because the pK a of cysteine (ϳ8.3) is at least an order of magnitude lower than the pK a of Hcy (ϳ9.5) (45), the cysteine thiolate anion would be much more stable and the preferred leaving group at neutral pH as we found for the albumin reaction (11).
The in vivo ratio of the relative abundance of the S-sulfonate transthyretin (TTR-Cys 10 -S-SO 3 H) to that of the unmodified transthyretin also increased with increasing homocysteine concentrations until it plateaued at about 200 M added homocysteine (Fig. 6A). The in vitro ratio of the relative abundance of the S-sulfonate transthyretin to that of the unmodified transthyretin remained essentially unchanged on addition of exogenous homocysteine (data not shown). Why the concentration of the S-sulfonate transthyretin increases in hyperhomocysteinemic patients is unknown. It is possible that the flux of cysteine through the catabolic pathway leading to cysteine sulfinic acid and then sulfite may increase in homocystinurics. Enhanced cysteine flux may be due to the reaction between homocysteine and S-cysteine albumin followed by the reaction of albumin thiolate anion with homocysteine-cysteine mixed disulfide (Reaction 2). The sulfite formed as a result of enhanced cysteine catabolism in homocystinurics could then react with transthyretin and/or its S-conjugated forms to produce Ssulfonate transthyretin. The formation of S-sulfonate albumin and S-sulfonate fibronectin has been reported previously (46) in rabbit plasma. Irrespective of the underlying mechanism, the formation of S-sulfonate transthyretin may have pathophysiological consequences.
Transthyretin forms amyloid fibrils under weakly acidic conditions (pH 4.0 to 4.5) (23). Kishikawa et al. (39) reported that S-sulfonation enhanced the amyloidogenicity of transthyretin. The formation of transthyretin fibrils was studied with three different preparations: unmodified transthyretin (with thiol compounds attached to Cys 10 and containing ϳ20% S-sulfonated transthyretin), dithiothreitol-treated transthyretin (with a free sulfhydryl group at Cys 10 ), and transthyretin conjugated with sulfite (S-sulfonate transthyretin). At pH 4.0 there was a 3-fold enhancement of fibril formation with S-sulfonate transthyretin compared with unmodified transthyretin, whereas reduced transthyretin had very low capacity to form fibrils. These results show that sulfonation of Cys 10 of transthyretin might increase the fibril forming capacity of transthyretin, which could lead to a more rapid progression of familial transthyretin amyloidosis or senile systemic amyloidosis. Interestingly, a higher percentage of S-conjugated transthyretin to the unmodified transthyretin has been reported in patients with symptomatic amyloid disease (47).
Our studies show that the amyloid protein transthyretin can undergo homocysteinylation in human plasma. In contrast, the amyloid protein ␤ 2 -microglobulin, although it has a single disulfide bond that could be targeted by homocysteine thiolate anion, is not homocysteinylated under the same in vitro reaction conditions used to homocysteinylate transthyretin. The reason that ␤ 2 -microglobulin is resistant to homocysteinylation is probably due to the buried nature of the disulfide bond in the native protein (48). The ratio of the relative abundance of the S-homocysteine transthyretin to that of the unmodified transthyretin increases with increasing tHcy concentrations. Thus, homocysteinylated transthyretin is a novel indicator of plasma homocysteine burden in hyperhomocysteinemia and homocystinuria. It remains to be determined whether post-translational modification of transthyretin by homocysteine plays a role in its pathogenicity in amyloidosis and other diseases.