Cross-talk between Cys 34 and Lysine Residues in Human Serum Albumin Revealed by N -Homocysteinylation*

Protein N -homocysteinylation involves a post-transla-tional modification by homocysteine (Hcy)-thiolactone. In humans, about 70% of circulating Hcy is N -linked to blood proteins, mostly to hemoglobin and albumin. It was unclear what protein site(s) were prone to Hcy at-tachment and how N -linked Hcy affected protein func-tion. Here we show that Lys 525 is a predominant site of N -homocysteinylation in human serum albumin in vitro and in vivo . We also show that the reactivity of albumin lysine residues, including Lys 525 , is affected by the status of Cys 34 . The disulfide forms of circulating albumin, albumin-Cys 34 -S-S-Cys and albumin-Cys 34 -S-S-Hcy, are N -homocysteinylated faster than albumin-Cys 34 -SH. Although N -homocysteinylations of albumin-Cys 34 -SH and albumin-Cys 34 -S-S-Cys yield different primary products, subsequent thiol-disulfide exchange reactions result in the formation of a single product, N -(Hcy-S-S-Cys)-albu-min-Cys 34 -SH. We also show that N -homocysteinylation affects the susceptibility of albumin to oxidation and proteolysis. The data suggest that

Since the 1960s, it has been known that elevated levels of homocysteine (Hcy), 1 resulting from mutations in genes encoding Hcy-metabolizing enzymes, are harmful to humans (1,2). During the past decade it has been established that even a mild increase in Hcy level is a risk factor for cardiovascular disease and stroke in humans (3,4) and predicts mortality independently of traditional risk factors in patients with coronary artery disease (5). Plasma Hcy is also a risk factor for neurodegenerative disorders, such as dementia and Alzheimer's disease (6). In tissue cultures, Hcy does not support growth and induces apoptotic death in human endothelial cells (7). Animal and cell culture studies have shown that Hcy induces cell death and potentiates amyloid ␤-peptide toxicity in neurons (8).
In humans, Hcy, formed from dietary methionine as a byproduct of cellular methylation reactions, is detoxified by folic acid-and vitamin B 12 -dependent re-methylation to methionine (2) or vitamin B 6 -dependent trans-sulfuration to cysteine (1). Whereas Hcy is formed in all human organs, most of its detoxification occurs in the liver and kidneys. Detoxification of Hcy in human vascular tissues and skin occurs only by re-methylation; enzymes of the trans-sulfuration pathway are not expressed in these tissues (9).
Hcy is perhaps the most reactive amino acid in biological systems (1,2). In addition to re-methylation to methionine or trans-sulfuration to cysteine (via cystathionine), Hcy is also metabolically converted to Hcy-thiolactone, S-nitroso-Hcy, AdoHcy, Hcy-containing disulfides, or homocysteic acid, each of which has been implicated in the pathology of hyperhomocystinemia (1,10).
Because of its similarity to the protein amino acid methionine, Hcy can exert its biological effects by interfering with protein biosynthesis (10 -18). For example, Hcy is metabolized to Hcy-thiolactone by methionyl-tRNA synthetase in a two-step reaction (19,20). In the first step (Reaction 1), methionyl-tRNA synthetase catalyzes activation of Hcy with ATP, which yields methionyl-tRNA synthetase (MetRS)-bound homocysteinyl adenylate.
MetRS ϩ Hcy ϩ ATP 7 MetRS⅐Hcy ϳ AMP ϩ PP i REACTION 1 The second step (Reaction 2), in which the side chain thiolate of Hcy reacts with the activated carboxyl group of Hcy, yields Hcy-thiolactone.
The energy of the anhydride bond in HcyϳAMP is conserved in the thioester bond in Hcy-thiolactone. Because of this, Hcythiolactone reacts with proteins, forming Hcy-containing adducts, in which the carboxyl group of Hcy is linked by an amide bond with ⑀-amino group of a protein lysine residue (21)(22)(23).
Molecular mechanism and functional consequences of albumin N-homocysteinylation were not known. In this study we show that Lys 525 is a predominant site of N-homocysteinylation in human serum albumin. We also show that the status of Cys 34 affects the reactivity of albumin lysine residues, including Lys 525 , and that N-homocysteinylation affects the susceptibility of albumin to proteolysis and oxidation. Preparation of Albumin-Cys 34 -SH and Albumin-Cys 34 -S-S-Cys-Human serum albumin (50 mg/ml) was converted to albumin-Cys 34 -SH by treatment with 2 mM DTT in 0.1 M potassium phosphate buffer, pH 7.4, 0.2 mM EDTA for 5 min at room temperature, diluted 10-fold with 0.01 M potassium phosphate buffer, pH 5.8, and purified by anion exchange HPLC. Fractions containing albumin-Cys 34 -SH were concentrated to 10 mg/ml in 0.1 M potassium phosphate buffer, pH 7.4, 0.2 mM EDTA, and treated with 2-fold molar excess of Cys overnight at 37°C. Excess Cys was removed from albumin-Cys 34 -S-S-Cys by ultrafiltration through an Ultrafree-0.5 10-kDa cut-off membrane (Millipore) at 4°C. The conversion of albumin-Cys 34 -SH to albumin-Cys 34 -S-S-Cys was monitored by anion exchange HPLC.

Preparation of [ 35 S]Hcy-thiolactone-Carrier-free
Preparation of Albumin-Cys 34 -S-IAA-Albumin-Cys 34 -SH (10 mg/ml) was incubated with 4-fold molar excess IAA in 0.1 M potassium phosphate buffer, pH 7.4, 0.2 mM EDTA for 1 h at 37°C in the dark. Excess IAA was removed by ultrafiltration. The conversion of albumin-Cys 34 -SH to albumin-Cys 34 -S-IAA was complete as determined by anion exchange HPLC.
Preparation of N-Hcy-Proteins-Individual proteins (obtained from Sigma) were dissolved at 50 mg/ml in 0.1 M potassium phosphate, pH 7.4, 0.2 mM EDTA, and incubated for 16 h at 37°C with 1-2 molar excess of unlabeled L-Hcy-thiolactone. Under these conditions the extent of modification was 0.6 -0.75 mol of Hcy/mol of protein. The extent of modification has been determined by the formation of N-[ 35 S]Hcyprotein, assayed either by precipitation with trichloroacetic acid (22), in parallel reactions with identical concentrations of [ 35 S]Hcy-thiolactone or by monitoring the increase in protein thiol groups with Ellman's reagent (Sigma) (34). Both methods gave similar results.
Susceptibility of Hcy-thiolactone-modified Albumin to Oxidation-Unmodified albumin or N-Hcy-albumin (at 10 mg/ml) were incubated with or without 5-fold molar excess of H 2 O 2 in 0.1 M potassium phosphate buffer, pH 7.4, for 15 min at 37°C. The formation of albumin aggregates after oxidation was analyzed by anion exchange HPLC.
Anion Exchange HPLC-Albumin species (5 mg/ml) were analyzed by anion exchange HPLC on a Vydac 301VHP575 DEAE column (7.5 ϫ 50 mm, 5 m, 900 Å) equilibrated with 10 mM potassium phosphate, pH 5.8 (buffer A), and eluted with a linear gradient 0 -250 mM NaCl (40 min, 1 ml/min) in buffer A. The elution conditions were somewhat modified for each experiment to achieve optimal separation of analyzed albumin species. In some experiments, a nonporous Tosohaas TSK-gel DNA-NPR column (4.6 ϫ 75 mm, 2.5 m) was used with similar results. The effluent was monitored at 280 nm using a diode array detector.
Purification of Native Albumin from Human Serum-Human serum (0.3 ml) was diluted with 1 volume of 0.1 M potassium phosphate, pH 6.8, and loaded onto a 1-ml High-Trap Blue column (Amersham Biosciences). After washing off unbound protein, albumin was eluted with 2 M NaCl and desalted by ultrafiltration.
Enrichment of N-Hcy-albumin with Thiopropyl-Sepharose-Crystal structure of albumin-Cys 34 -SH shows that Cys 34 is located in a crevice about 10 Å deep (30), and because of this, access to Cys 34 is somewhat limited. In contrast, N-linked Hcy in N-(Hcy-SH)-albumin-Cys 34 -SH is expected to be freely accessible. Indeed, in control experiments we found that N-(Hcy-SH)-albumin-Cys 34 -SH binds well, whereas albumin-Cys 34 -SH binds very poorly to thiopropyl-Sepharose. We utilized this property to enrich native albumin in N-(Hcy-SH)-albumin. Native albumin (50 mg/ml) was first converted into mercaptoalbumin by treatment with 2 mM DTT in 0.1 M potassium phosphate, pH 7.4, 0.2 mM EDTA. After removal of low molecular weight thiols by ultrafiltration on Ultrafree-05 30-kDa cut-off membrane (Millipore), mercaptoalbumin (50 mg/ml) was allowed to bind to washed thiopropyl-Sepharose (120 mg) in 0.4 ml of 0.1 M potassium phosphate, pH 5.0, for 20 min at 21°C. The resin was washed with 0.1 M potassium phosphate, pH 5.0, 0.1 M NaCl to remove unbound protein. Bound albumin was eluted with 0.5 ml of 25 mM DTT in 0.1 M potassium phosphate, pH 7.4, desalted, and concentrated to 50 l in 0.1 M potassium phosphate, pH 5.0. Resulting albumin preparation was subjected to Thiopropyl-Sepharose enrichment two more times. About 7% of native albumin was bound to and recovered from thiopropyl-Sepharose after three cycles of enrichment.
Reversed Phase C18 HPLC-HPLC analyses of tryptic peptides were carried out using a reversed phase C18 peptide-protein column (4.6 ϫ 100 mm, 5 , 300 Å) from Vydac. The detection was by monitoring at 215 nm using a diode array detector. After sample application, the C18 HPLC column was eluted with a linear gradient from 0 to 40% acetonitrile (60 min) in 0.1% trifluoroacetic acid at a flow rate of 1 ml/min. Cation Exchange HPLC-N-Hcy-peptides from a C18 column were further purified by cation exchange on a PolySULFOETHYL Aspartamide TM column (2.1 ϫ 200 mm, 5 m, 300 Å, PolyLC, Inc.). Solution A (50 mM potassium phosphate-HCl, pH 3.4) and solution B (500 mM NaCl in solution A) were used as solvents, and the flow rate was 0.6 ml/min. After sample application, the column was eluted with a linear gradient from 0 to 60% solution B for 20 min.
HPLC Instrumentation-HPLC instrumentation was Beckman-Coulter System Gold Noveau as described before (24,26). It consisted of the following modules: advanced gradient solvent delivery module 126 and a high resolution diode array 168 detector module. Manual injector (7725i Rheodyne) with a 0.1-ml loop was used. Chromatograms were analyzed using Gold Noveau chromatography work station software for Windows.
Mass Spectroscopy-Peptide mass analyses were carried out by Dr. Hong Li on a MALDI-TOF Voyager-DE TM PRO Biospectrometry TM work station (PerSeptive Biosystems) at the New Jersey Medical School Center for Advanced Proteomics Research facility (njms.umdnj.edu/ biochemistry/proteomics/index.htm).

Hcy-thiolactone Reacts Faster with Albumin-Cys 34 -S-S-Cys and Albumin-Cys 34 -S-S-Hcy than with Albumin-Cys 34 -SH-To
determine whether the status of Cys 34 in human serum albumin affects the susceptibility of that protein to N-homocysteinylation, albumin-Cys 34 -S-S-Cys, albumin-Cys 34 -S-S-Hcy, and albumin-Cys 34 -SH ( Fig. 1), major forms of circulating albumin (30, 31), were prepared as described under "Materials and Methods" and incubated with [ 35 S]Hcy-thiolactone. The forma- REACTION 2 tion of N-[ 35 S]Hcy-albumin was monitored by precipitation with trichloroacetic acid. As shown in Fig. 2, albumin-Cys 34 -S-S-Cys and albumin-Cys 34 -S-S-Hcy were N-homocysteinylated at similar rates. However, both disulfide forms of albumin were N-homocysteinylated 47 Ϯ 5% faster than mercaptoalbumin, albumin-Cys 34 -SH. Similar results were obtained when corresponding forms of albumin were prepared in situ by treatments with 5-fold molar excess Cys, Hcy, or DTT. These results suggest that the reactivity of albumin lysine residues depends on the status of Cys 34 .
Thiol-Disulfide Exchange in N-Homocysteinylated Albumin-Cys 34 -S-S-Cys Leads to Formation of Free Sulfhydryl at Cys 34 -Albumin-Cys 34 -S-S-Cys or albumin-Cys 34 -SH was modified with equimolar amounts of Hcy-thiolactone and analyzed by anion exchange HPLC. Unmodified albumin-Cys 34 -S-S-Cys was eluted from the HPLC column at 13.2 min (Fig. 3A, broken trace). After 4 h of modification, about 50% of albumin-Cys 34 -S-S-Cys was N-homocysteinylated and eluted from the column as a new peak at 12.6 min, before the peak of unmodified form (Fig. 3A, thin trace). After 20 h of modification, all albumin-Cys 34 -S-S-Cys disappeared, and only a single peak of N-homocysteinylated albumin-Cys 34 -S-S-Cys at 12.6 min was observed (Fig. 3A, thick trace). Both native and N-homocysteinylated albumin-Cys 34 -SH were eluted from the anion exchange HPLC column at 12.6 min (not shown).
Altered chromatographic behavior of albumin-Cys 34 -S-S-Cys after modification with Hcy-thiolactone could be due to two alternative mechanisms. First, blocking an ⑀-amino group of lysine residue(s) by N-homocysteinylation may have changed the charge of albumin, thus weakening its binding to the anion exchange column. However, this seems unlikely because similar modification of albumin-Cys 34 -SH did not affect its elution position from the column. Instead, N-homocysteinylation of albumin-Cys 34 -S-S-Cys may have caused regeneration of a free thiol at Cys 34 as a result of thiol-disulfide exchange between a free thiol of newly incorporated Hcy and the disulfide at Cys 34 (Reaction 3). This interpretation is supported by an observation that regeneration of a disulfide at Cys 34 by incubation of N-(Hcy-S-S-Cys)-albumin-Cys 34 -SH with 2-fold molar excess of cysteine also restored its stronger binding to the anion exchange column (Fig. 3B).
To determine whether a thiol of Cys 34 in N-(Hcy-SH)-albumin-Cys 34 -SH is required for the thiol-disulfide exchange, albumin was first treated with IAA to block the Cys 34  N-Homocysteinylation Affects Susceptibility of Albumin to Oxidation-Because N-homocysteinylation results in addition of thiol group(s) to albumin, modified albumin should become more prone to formation of intermolecular aggregates after oxidation. Indeed, treatment of N-homocysteinylated albumin with H 2 O 2 resulted in formation of albumin aggregates (eluting as a broad peak at 18 min), which bound more strongly than monomeric forms to the DEAE column (Fig. 7A). Treatment of N-homocysteinylated and oxidized sample with DTT resulted in the disappearance of aggregates, suggesting that they were formed as a result of intermolecular disulfide bond formation. Native albumin did not form appreciable aggregates upon treatment with H 2 O 2 (Fig. 7B), which suggests that thiols introduced by N-homocysteinylation are required for aggregation.
N-Homocysteinylation Affects Susceptibility of Albumin to Proteolysis-Structural changes caused by N-homocysteinylation may affect proteolytic turnover of the modified protein.
To determine this, we examined the susceptibility of unmodified and N-homocysteinylated forms of albumin to trypsin, chymotrypsin, elastase, and cathepsin D. As shown in Fig. 8, two major forms of circulating albumin, albumin-Cys 34 -S-S-Cys and albumin-Cys 34 -SH differed in their susceptibility to diges- After DTT treatment the peak at 18 min becomes smaller, the peak of albumin-Cys 34 -S-S-Cys disappears (2nd peak), and the 1st peak, corresponding to albumin-Cys 34 -SH, becomes larger (broken trace). B, minor peak is visible at 18 min on HPLC profile of H 2 O 2 -oxidized native albumin (solid trace). After DTT treatment the minor peak at 18 min is still visible (labeled Alb(ox)), but the peak of albumin-Cys 34 -S-S-Cys (2nd peak) disappears, and the peak of albumin-Cys 34 -SH (1st peak) becomes larger (broken trace). Identity of the 3rd peak is not known. tion by trypsin, with the former being degraded faster than the latter. The nature of a thiol bound by a disulfide linkage to Cys 34 of albumin does not seem to affect its sensitivity to proteolytic digestion by trypsin; albumin-Cys 34 -S-S-Hcy was degraded by trypsin as efficiently as was albumin-Cys 34 -S-S-Cys (not shown). After N-homocysteinylation, albumin-Cys 34 -SH became more resistant, but the sensitivity of N-(Hcy-S-S-Cys)-albumin-Cys 34 -SH to trypsin was not much different from that of unmodified form (Fig. 8). N-(Hcy-S-S-Cys)-albumin-Cys 34 -SH and N-(Hcy-S-S-Cys)-albumin-Cys 34 -S-S-Cys exhibited similar sensitivities to trypsin (not shown). Similar pattern of susceptibilities to proteolysis were also observed with chymotrypsin (Fig. 8). There were only minor differences in susceptibilities of different forms of albumin to elastase or cathepsin D (Fig. 8).
Lys 525 in Albumin Undergoes Preferential N-Homocysteinylation in Vitro-To identify site(s) susceptible to N-homocysteinylation, albumin was modified with Hcy-thiolactone at a molar ratio 1:1, reduced with DTT, acetamidated with IAA to block thiols, and digested with trypsin. Tryptic peptides were separated by reversed phase HPLC (Fig. 9A). A predominant N-Hcy-peptide (eluting in 32 min fraction) was further purified by cation exchange HPLC (Fig. 9C). N-Homocysteinylated peptides were detected by using tryptic digests of N-[ 35 S]Hcyalbumin in parallel experiments (Fig. 9, B and D). Purified N-Hcy-peptide was subjected to MALDI-TOF mass spectrometric analysis. A single signal at m/z of 1,302.83 (Fig. 10A) was observed on mass spectra of the N-Hcy-peptide. This mass corresponds to an acetamidated and N-homocysteinylated peptide containing Lys 525 , 525 KQTALVELVK 534 (calculated mass of the derivatized peptide is 1,302.8; the site of N-homocysteinylation is in boldface). This peptide was also observed on mass spectra of unpurified tryptic digests of N-Hcy-albumin (not shown).
Lys 525 in albumin-Cys 34 -S-S-Cys was N-homocysteinylated significantly faster than the same lysine residue in N-Hcyalbumin-Cys 34 -SH. By analyzing tryptic digests on a reversed phase C18 HPLC column (Fig. 9B), we found that 9.22 Lys 525 in Native Albumin Is N-Homocysteinylated-Although in vitro data indicated that Lys 525 is a predominant site susceptible to N-homocysteinylation, it is unclear whether this site is also N-homocysteinylated in vivo. To determine this, native albumin was isolated from two human subjects (having elevated plasma total Hcy levels of 40 -80 M) and analyzed for the Lys 525 modification. To facilitate detection of N-linked Hcy, native human serum albumin was enriched in N-Hcycontaining species by using thiopropyl-Sepharose. Such enriched preparations were reduced with DTT, modified with IAA, and digested with trypsin. Putative peptide containing N-homocysteinylated Lys 525 was purified by HPLC and subjected to mass spectrometric analysis. A peptide with a mass of 1,302.92, corresponding to the mass of Lys 525 -containing peptide, was present in a MALDI-TOF spectrum (Fig. 10B). The signal at m/z 1,302.92 was about 2-fold smaller for albumin  9. Purification of a predominant N-homocysteinylated peptide from N-Hcy- from a subject who had 40 M tHcy compared with a subject who had 80 M tHcy. The signal at m/z 1,302.92 was also observed in analysis of commercial native albumin but was much less intense (not shown). These data strongly suggest that Lys 525 is a site of N-homocysteinylation in native albumin in vivo.

DISCUSSION
Human serum albumin is the major plasma protein (30) which is also a major target for N-homocysteinylation by Hcythiolactone in vitro (22) and in vivo (26). The present work identifies Lys 525 as a preferential site of N-homocysteinylation in human serum albumin in vitro and in vivo, and provides evidence for specific structural alterations caused by N-homocysteinylation. The findings of this work also suggest that a disulfide at Cys 34 , a conserved residue in albumins from various organisms, facilitates conversion of N-homocysteinylated mercaptoalbumin to a proteolytically sensitive form.
The present data support the following mechanisms of albumin N-homocysteinylation. Hcy-thiolactone reacts with lysine residues of both major forms of circulating albumin. However, the status of Cys 34 affects the rate of N-homocysteinylation; the reaction with albumin-Cys 34 -S-S-Cys is faster than with albumin-Cys 34  Thus, N-homocysteinylation of a mixture of albumin-Cys 34 -SH and albumin-Cys 34 -S-S-Cys, which is present in circulation (31), leads to a single N-homocysteinylated product, N-(Hcy-S-S-Cys)-albumin-Cys 34 -SH. The equilibrium is strongly shifted toward N-(Hcy-S-S-Cys)-albumin-Cys 34 -SH because the Cys 34 thiolate anion has an unusually low pK a of ϳ5 (30) and thus is more thermodynamically stable than Hcy thiolate anion. The low pK a of the Cys 34 thiolate also makes the thiol-disulfide exchange of N-(Hcy-SH)-albumin-Cys 34 -SH with albumin-Cys 34 -S-S-Cys thermodynamically more favored than with cystine. For example, under conditions where the thiol-disulfide exchange between N-(Hcy-SH)-albumin-Cys 34 -SH and albumin-Cys 34 -S-S-Cys went to completion (Fig. 3), the exchange with cysteine was Ͻ20% complete (not shown). Thiol-disulfide exchange reactions between albumin-Cys 34 -SH and cystine, homocystine, or Cys-S-S-Hcy disulfide are known to be slow, being about 20% or less complete in 4 h (32,36).
It has been hypothesized that metabolic conversion of Hcy to Hcy-thiolactone, the reactivity of Hcy-thiolactone toward proteins, and resulting protein damage contribute to pathologies associated with elevated Hcy levels in human beings (21)(22)(23). If this hypothesis is correct, it is likely that protective mechanism(s) against Hcy-thiolactone have evolved. One possible protective mechanism can be provided by Hcy-thiolactonase/ paraoxonase, a component of high density lipoprotein, which detoxifies Hcy-thiolactone by hydrolyzing it to Hcy (28,29).
Human serum albumin, because of its abundance and the ability to avidly react with Hcy and Hcy-thiolactone, is likely to serve an important dual protective role. Consistent with this suggestion are the observations that S-Hcy-albumin and N-Hcy-albumin account for most (Ͼ90%) of S-linked and N-linked Hcy, respectively, present in human plasma protein and for Ͼ90% of total plasma Hcy (22,26,28). In addition to S-Hcyprotein (26,31,32) and N-Hcy-protein (12, 13, 26 -28), Hcy exists in circulation as free reduced Hcy (31), Hcy-thiolactone (24,25), and disulfide forms with low molecular weight thiols (31). Of these, Hcy-thiolactone (10, 17, 21-23, 26, 28, 29) and free reduced Hcy (37), which comprise 1.4 (24,25) and 2% (31) of plasma total Hcy, respectively, have been suggested to be the deleterious forms of Hcy. For example, free reduced Hcy is associated with endothelial dysfunction, whereas disulfide forms of Hcy, including albumin-Cys 34 -S-S-Hcy, are not (37). Hcy-thiolactone induces caspase-independent vascular endothelial cell death with apoptotic features (38). Thus, the conversion of Hcy into albumin-Cys 34 -S-S-Hcy would prevent cellular uptake of Hcy and therefore minimize the conversion to Hcy-thiolactone. Indeed, albumin-Cys 34 -S-S-Hcy comprises 82%, whereas reduced Hcy accounts for only 2% of "total" Hcy in human serum (31). In addition, albumin detoxifies a significant fraction of Hcy-thiolactone by virtue of N-homocysteinylation. Plasma pool of N-linked Hcy comprises up to 25% of total plasma Hcy in humans, with N-Hcy-albumin accounting for 90% of plasma N-Hcy-protein (22,26). Both N-linked and S-linked protein Hcy are most likely detoxified in the liver where transmethylation and trans-sulfuration pathways of Hcy metabolism are the most active (1,2). Consistent with this suggestion are our findings that N-(Hcy-S-S-Cys)-albumin-Cys 34 -SH, albumin-Cys 34 -S-S-Cys, and albumin-Cys 34 -S-S-Hcy are more susceptible to proteolysis than albumin-Cys 34 -SH. Although N-(Hcy-SH)-albumin-Cys 34 -SH is resistant to proteolytic digestion, this form is unlikely to exist in circulation because it undergoes the thiol-disulfide exchange with albumin-Cys 34 -S-S-Cys, which converts it into proteolysis-prone form, N-(Hcy-S-S-Cys)-albumin-Cys 34 -SH (Reaction 5). These findings suggest that a disulfide at residue Cys 34 in albumin may have an important role in facilitating proteolytic turnover of N-homocysteinylated albumin.
Different proteolytic susceptibilities of albumin-Cys 34 -SH and albumin-Cys 34 -S-S-Cys suggest that albumin adopts a different structure depending on the state of Cys 34 . Indeed, a structural transition in albumin dependent on the state of Cys 34 has been detected by NMR spectroscopy (39). Our data suggest that N-homocysteinylation interferes with this structural transition.
Because it is a downstream metabolite that, most likely, reflects damage caused by Hcy, N-linked Hcy could be a new marker of cardiovascular risk, possibly more predictive than total Hcy. To determine this, it would be important to monitor protein N-homocysteinylation in human beings. Present methods of monitoring protein N-homocysteinylation are relatively complex and thus not very useful in a clinical setting. However, identification of Lys 525 as a predominant N-homocysteinylation site in human serum albumin opens up a way of designing new diagnostic tools for monitoring cardiovascular risk associated with elevated levels of plasma Hcy. For example, specific antibodies can be raised against N-homocysteinylated peptides of albumin containing Lys 525 and used to monitor the status of albumin N-homocysteinylation in human beings.