(Received for publication, September 4, 1996, and in revised form, October 11, 1996)
From the Department of Microbiology & Molecular Genetics, University of Medicine & Dentistry of New Jersey-New Jersey Medical School, Newark, New Jersey 07103
Editing of the non-protein amino acid
homocysteine, a frequent type of error-correcting process in amino acid
selection for protein synthesis by an aminoacyl-tRNA synthetase,
results in formation of a cyclic thioester, homocysteine thiolactone.
Here it is shown that human cells in which homocysteine metabolism is
deregulated by a mutation in the cystathionine -synthase gene and/or
by an antifolate drug, aminopterin (which prevents remethylation of
homocysteine to methionine by methionine synthase), produce more
homocysteine thiolactone, in addition to homocysteine, than unaffected
cells. The thiolactone is incorporated into cellular and extracellular
proteins, in addition to being secreted and hydrolyzed to homocysteine.
Experiments with model proteins and amino acids suggest that the
mechanism of incorporation involves acylation of side chain amino
groups of lysine residues by the activated carboxyl group of the
thiolactone. The metabolic conversion of homocysteine to homocysteine
thiolactone and the reactivity of the thiolactone toward proteins may
explain pathological consequences of elevated levels of homocysteine
such as observed in vascular disease.
Elevated levels of homocysteine (Hcy)1 are an independent risk factor for cardiovascular disease in humans (reviewed in Ref. 1, see also Refs. 2-4). However, it is not known what aspect of Hcy metabolism is harmful to human cells. We have been studying for some time an error editing reaction in protein synthesis in which Hcy is first misactivated by methionyl-tRNA synthetase to form an enzyme-bound Hcy-AMP. The Hcy-AMP is subsequently destroyed by the enzyme with the formation of a cyclic thioester, Hcy thiolactone (5). This reaction is apparently universal and prevents misincorporation of Hcy into protein in all living cells tested thus far (Refs. 6-10; this work). Although only methionyl-tRNA synthetase is involved in editing of both endogenous (formed in the methionine biosynthetic pathway) and exogenous (supplied into culture media) Hcy, two other synthetases, isoleucyl- and leucyl-tRNA synthetase, also edit exogenous Hcy, at least in Escherichia coli (10).
Oncogenic transformation of human and rodent cells is associated with
enhanced synthesis of Hcy thiolactone (9). Experiments with wild type
and temperature-sensitive aminoacyl-tRNA synthetase mutants of Chinese
hamster ovary cells showed that Hcy is edited by methionyl-tRNA
synthetase in these mammalian cells (9). As might be expected, the
synthesis of Hcy thiolactone is minimized in normal mammalian cells.
Mammalian, including human, cancer cells synthesize excessive amounts
of Hcy thiolactone probably because Hcy metabolism in these cells is
deregulated. Thus, one can predict that other human cells in which Hcy
metabolism is deregulated should also synthesize the thiolactone. Hcy
thiolactone, a thioester, is expected to acylate proteins under
physiological conditions, which would lead to cell damage. With this in
mind, I have examined effects of the antifolate drug aminopterin (which prevents re-methylation of Hcy to methionine by methionine synthase) on
synthesis of Hcy thiolactone in tissue cultures of human fibroblasts from homocystinuria patients (heterozygous and homozygous for cystathionine -synthase (CBS)-deficiency) and human breast tumor cells. The fate of Hcy thiolactone in cultures of human cells and its
reactivity toward proteins and amino acids under physiological conditions were studied. The data suggest a mechanism by which Hcy,
through its metabolic conversion to the thiolactone which in turn
acylates proteins, can lead to cell damage resulting in pathological
consequences such as vascular disease.
Human fibroblasts from homocystinuria patients homozygous for CBS deficiency, GM01128 (vitamin B6 nonresponsive) and GM01374 (vitamin B6 responsive), as well as corresponding fibroblasts from clinically unaffected parents heterozygous for CBS deficiency, GM01126 and GM01375, respectively, were obtained from the NIGMS Human Genetic Mutant Cell Repository (Camden, NJ). Breast tumor cell line HTB-132 was from American Type Culture Collection (Rockville, MD).
Cell Culture and [35S]Methionine-labeling ConditionsCells were maintained as monolayer cultures on 10-cm dishes at 37 °C in media supplemented with 10% fetal bovine serum (Life Technologies, Inc.). The culture medium for CBS-deficient and unaffected human fibroblasts was MEM (Sigma) in an atmosphere of 5% CO2, 95% air. For breast cancer cells HTB-132, L15 medium (Life Technologies, Inc.) in an atmosphere of 100% air (no extra CO2) was used. For [35S]methionine labeling, cells were plated in 3.5-cm dishes and grown to confluence. The medium was then replaced with 0.5 ml of methionine-free MEM supplemented with 10% dialyzed fetal bovine serum and 5 µM [35S]methionine (Amersham Corp.), and the cultures were maintained at 37 °C in an atmosphere of 5% CO2. Hypoxanthine/aminopterin/thymidine supplement (Life Technologies, Inc.) was used as a source of aminopterin (0.4 µM). Hypoxanthine (100 µM, sodium salt) and thymidine (16 µM) were included in the labeling medium in the absence of aminopterin.
Determination of Hcy ThiolactoneHcy thiolactone was separated from other sulfur-containing compounds by two-dimensional TLC chromatography using previously published procedures (7-9) with a slight modification (omitting ammonia from the second dimension solvent). At specified time intervals, aliquots of cell-free media from the [35S]methionine-labeled cultures were applied onto the origin line of 10 × 6.7 cm TLC cellulose plates (Kodak). Authentic Hcy thiolactone (Sigma), which was visualized on TLC plates under UV, was included as a standard. The plates were developed with butanol/acetic acid/water (4:1:1, v/v) in the first dimension and with 2-propanol/ethyl acetate/water (5:5:1, v/v) in the second dimension. Radioactivity co-migrating with the thiolactone standard on TLC plates was visualized by authoradiography using Kodak BioMax MR film and quantitated by scintillation counting. The radioactivity associated with Hcy thiolactone spots was sensitive to NaOH treatment, as expected.
Determination of HcySamples from [35S]methionine-labeled cultures were reduced with 10 mM dithiothreitol (DTT) and treated with 25 mM iodoacetate (pH 7.5) to convert Hcy into S-carboxymethyl-Hcy. Control experiments established that the conversion was quantitative. The S-carboxymethyl-Hcy was separated from other sulfur-containing compounds by one-dimensional TLC on cellulose plates developed with butanol/acetic acid/water (4:1:1, v/v) at a distance of 15 cm. Authentic radiolabeled S-carboxymethyl-Hcy was run in parallel as a standard. The spots of radioactive S-carboxymethyl-Hcy were visualized by authoradiography and quantitated by scintillation counting. Commercial preparations of [35S]methionine were found to contain <0.01% [35S]Hcy thiolactone and 0.4% [35S]Hcy. This level of [35S]Hcy contamination in preparations of [35S]methionine did not interfere with the interpretation of the data.
Determination of Hcy Incorporation into ProteinConfluent cultures in 3.5-cm dishes were labeled for 24-48 h with 5 µM [35S]methionine (0.1 mCi/ml) in 0.5 ml of methionine-free MEM medium supplemented with 10% dialyzed fetal bovine serum in the absence and presence of 0.4 µM aminopterin. The medium was removed, the cells were overlaid with 2 × 0.125 ml of ice-cold 1 M formic acid, and scraped off the dish. Protein was precipitated with 10% trichloroacetic acid. The resulting pellets were washed once with 5% trichloroacetic acid and once with acetone, all on ice. To remove Hcy bound to protein through disulfide bonds, as well as any free Hcy, the resulting pellets were taken up in 0.1 M ammonium bicarbonate, reduced with 50 mM DTT for 15 min at 37 °C, and precipitated with 10% trichloroacetic acid. The cycle of washing, treatment with DTT, and precipitation with trichloroacetic acid was repeated four times. In the last cycle 10 mM DL-Hcy (Sigma) (to dilute out any remaining unincorporated [35S]Hcy) was used together with DTT. After the last cycle, the trichloroacetic acid-precipitated protein was washed with acetone, dried under vacuum, and resuspended in 0.25 ml of 6 N HCl, 0.5 M 3-mercaptopropionic acid. To control for the completeness of removal of free Hcy from protein preparations, one-half of each sample was incubated for 15 min at 70 °C. During this treatment, all free Hcy, if present, is converted into the thiolactone, but protein is not hydrolyzed. The other half of each sample was sealed in an ampoule under vacuum and hydrolyzed at 110 °C for 20 h. This treatment would release any protein-bound Hcy (through amide bonds) in the form of the thiolactone. The recovery of Hcy thiolactone and methionine from known standards was 25 and 100%, respectively. Control and hydrolyzed samples were lyophilized, taken up in 10 µl of water, and subjected to two-dimensional TLC. Hcy thiolactone was present only in hydrolyzed samples. The radioactivity associated with the thiolactone and methionine spots on the TLC plates was determined by scintillation counting.
Preparation of [35S]Hcy ThiolactoneThe
method is based on observations that methionine synthase-deficient
mutants of E. coli (metE) (7) and yeast
(met6) (8), unable to transmethylate Hcy to methionine,
quantitatively convert Hcy (and its precursors such as cysteine and
sulfate) to Hcy thiolactone when starved for methionine (7, 8). A
CBS-deficient yeast mutant (cys2cys4), unable to metabolize
Hcy to cystathionine, converts Hcy (and sulfate) to the thiolactone (8)
when starved for cysteine. The thiolactone is secreted from cells and
accumulates in the culture media (7, 8). Either
[35S]sulfate in sulfate-free minimal media,
[35S]cysteine, or [35S]methionine in
minimal media can be used as a precursor of [35S]Hcy
thiolactone. Bacterial cells from a fresh overnight culture of
metE grown in M9 medium supplemented with 0.15 mM methionine were collected by centrifugation, washed once
with M9, and resuspended in 1 ml of M9 at a cell density of 4 × 108 cells/ml. Carrier-free [35S]cysteine (1 mCi, Amersham Corp.) and unlabeled cysteine (0.1 ml of 0.1 mM solution) were added to the culture (to obtain a specific activity of 105 Ci/mol). After a 5-h incubation at
37 °C with aeration all cysteine was consumed. The labeled culture
was centrifuged and the supernatant, containing >95% of the
thiolactone present in the culture, was extracted with 4 ml of
chloroform/methanol (2:1, v/v). The thiolactone was re-extracted from
the organic layer with 0.5 ml of 0.1 M HCl and lyophilized.
The residue was taken up in 20 µl of water, and the thiolactone was
further purified by two-dimensional TLC on a 20 × 20-cm Kodak
cellulose plate. The TLC plate was developed in the first dimension
with butanol/acetic acid/water (4:1:1, v/v) for 2 h and in the
second dimension with 2-propanol/ethyl acetate/water (5:5:1, v/v) for
another 2 h. The [35S]thiolactone spot, localized by
autoradiography using Kodak XAR-5 x-ray film (exposed for 20 min), was
eluted with 1 mM HCl at 4 °C. The first 0.25 ml of the
eluate contained 97% of the [35S]thiolactone. Out of 1 mCi (10 nmol) [35S]cysteine used, 0.24 mCi (2.4 nmol) was
recovered in [35S]Hcy thiolactone, a 24% yield.
Analytical two-dimensional TLC showed that the radiolabeled thiolactone
was >96% pure. A 10% yield of [35S]Hcy thiolactone was
obtained when [35S]methionine was used as a precursor.
The preparations were stored frozen at 20 °C. For experiments
requiring [35S]Hcy, the radiolabeled thiolactone was
hydrolyzed to completion with 10 mM NaOH at room
temperature for 15 min and used immediately.
In human cells, the transsulfuration
pathway (11) converts methionine, an essential amino acid, into
cysteine with Hcy as an intermediate (see Fig. 5). The pathway starts
with the formation of S-adenosylmethionine (AdoMet), which
yields S-adenosylhomocysteine (AdoHcy) in subsequent
transmethylation reactions. Hydrolysis of AdoHcy yields Hcy. Hcy is
then re-methylated back to methionine by a folate- and vitamin
B12-dependent enzyme, methionine synthase. Hcy is also
converted into cysteine via cystathionine with a vitamin B6-dependent enzyme, CBS, which is the first enzyme in the
pathway. In an additional pathway, originally observed in tumor cells, methionyl-tRNA synthetase converts Hcy into Hcy thiolactone (9). In
this study, effects of CBS deficiency and of the antifolate drug
aminopterin on Hcy thiolactone synthesis by human cells were determined.
Several different human cell lines growing in monolayer cultures were
labeled with [35S]methionine for up to 48 h. Most
(~90%) of [35S]methionine was incorporated into
protein within 6-12 h as shown by trichloroacetic acid precipitation
of cell extracts. The ~10% of [35S]methionine that
remained in the media even after 48 h (as demonstrated by
two-dimensional TLC) is due, most likely, to protein turnover. Two-dimensional TLC analysis of 35S-compounds from the
media and cell extracts demonstrated that all cell lines synthesized
Hcy thiolactone but with different efficiencies. Most (95%) of the
thiolactone was secreted from cells, as found previously with other
cell types (6-9). Fig. 1 shows examples of kinetics of
the thiolactone formation (open squares) determined in
CBS-deficient (GM01374; Fig. 1A), unaffected (GM01375; Fig.
1C), and breast tumor (HTB-132; Fig. 1E) cell
cultures. The most efficient synthesis of Hcy thiolactone was observed
in cultures of human breast tumor cells HTB-132, with a level of 4 nM reached just after a 2-h labeling with
[35S]methionine. Slightly lower levels of the thiolactone
(3 nM) were reached by CBS-deficient cells after 12 h.
The least efficient synthesis of Hcy thiolactone was observed in
cultures of unaffected human fibroblasts GM01375 in which the level of
the thiolactone was increasing steadily up to 2 nM at
48 h (Fig. 1C).
The differences between the three cell lines in their ability to synthesize Hcy thiolactone were paralleled by similar differences in their ability to synthesize Hcy (Fig. 1, B, D, and F). The CBS-deficient cells GM01374 produced more Hcy than unaffected cells GM01375 (e.g. 77 nM versus 38 nM at 6 h; open squares in Fig. 1, B and D). In the absence of aminopterin (open squares), the highest levels of Hcy (170 nM at 4 h; Fig. 1F) were observed in cultures breast tumor cells HTB-132.
The addition of the antifolate drug aminopterin, which prevents re-methylation of Hcy to methionine by methionine synthase, to tissue cultures resulted in enhanced synthesis of Hcy thiolactone (solid circles in Fig. 1, A, C, and E) by all cell lines. In the presence of aminopterin, the levels of Hcy thiolactone in cultures of CBS-deficient (GM01374; Fig. 1A) and unaffected cells (GM01375; Fig. 1C) were increasing with similar kinetics and reached a concentration of about 16 nM in each culture at 48 h. There were also no major differences between GM01374 and GM01375 cells in their ability to produce Hcy in the presence of aminopterin; production of Hcy followed similar kinetics and a concentration of 155 nM Hcy (representing ~3.5% of the metabolized [35S]methionine) was reached in both cultures at 48 h (compare B and D in Fig. 1). Apparently, in the presence of aminopterin, the contribution of CBS deficiency to the synthesis of the thiolactone and Hcy by GM01374 cells is minor.
The most efficient synthesis of Hcy thiolactone (Fig. 1E) and Hcy (Fig. 1F) occurred in cultures of breast cancer cells HTB-132 in the presence of aminopterin (solid circles in E and F, respectively). The levels of the thiolactone measured in the presence of aminopterin were higher (7 nM) than in its absence (4 nM) already after a 2-h labeling with [35S]methionine (Fig. 1E). In contrast to subsequent leveling off observed in Hcy thiolactone levels in the absence of aminopterin, the thiolactone levels in the presence of aminopterin were increasing continuously up to a concentration of 23 nM at 20 h. The levels of Hcy measured in the presence of aminopterin were 1.5-3-fold higher than in its absence throughout the course of the experiment. The highest concentration of Hcy in the cultures of HTB-132, 330 nM, observed after 8 h (solid circles in Fig. 1F) represents ~7% of the metabolized [35S]methionine.
Hcy thiolactone levels were also measured in cultures of vitamin B6 nonresponsive CBS-deficient cell line GM01128 and a corresponding unaffected cell line GM01126. As shown in Table I, GM01128 produced more thiolactone (0.7 nM at 20 h) than GM01126 (0.2 nM at 20 h) in the absence of aminopterin. As with other cell lines, the addition of aminopterin resulted in enhancement of the production of Hcy thiolactone (from 0.7 to 10.2 nM in the culture of GM01128 and from 0.2 to 2.7 nM in the culture of GM01126). However, in the presence of aminopterin, GM01128 produced about 4-fold more thiolactone than GM01126 (10.2 versus 2.7 nM; Table I), in contrast to GM01374 and GM01375 which produced similar amounts of Hcy thiolactone (7.5 versus 7.0 nM; see the "+" aminopterin lines in Table I). Thus, in the presence of aminopterin, CBS deficiency contributes significantly to the synthesis of Hcy thiolactone in GM01128 but not in GM01374.
|
The quantitative differences between GM01128 and another CBS-deficient cell line GM01374 in the ability to produce Hcy thiolactone are most likely due to different natures of lesions in their CBS proteins. CBS deficiency in GM01374 is vitamin B6-responsive whereas that of GM01128 is vitamin B6 nonresponsive. Because cell culture media contain vitamin B6, CBS deficiency of GM01128 will be more severe than that of GM01374 which is indeed manifested by greater production of Hcy by GM01128 than by GM01374 (Table I).
Effects of the age of the cell cultures on their ability to produce Hcy thiolactone was also determined. For this purpose, Hcy thiolactone levels in confluent cultures of GM01374 and GM01375 were compared with the thiolactone levels in the cultures that were maintained in growth media for an additional 7 days after reaching confluence. In the absence of aminopterin, the aged cultures of GM01374 and GM01375 produced 7-14-fold more Hcy thiolactone than the corresponding younger cultures (Table I). In the presence of aminopterin, the aged cultures produced 28% less thiolactone that the younger cultures. Hcy levels were affected less than the thiolactone levels in the aged cultures. For example, in the absence of aminopterin, the aged cultures of GM01375 and GM01374 produced 1.3- and 2.8-fold, respectively, more Hcy than the younger cultures (Table I). Thus, aging of a culture is associated with enhanced ability to produce Hcy thiolactone, most likely due to less efficient metabolism of the thiolactone in the older cultures. These findings most likely explain our inability to detect the thiolactone in GM01375 in our previous studies (9) in which, in contrast to the present study, subconfluent cultures were used.
Metabolism of Exogenous Hcy in Tissue CulturesThe formation of both Hcy and the thiolactone in tissue cultures raises the question of a precursor-product relationship (a "chicken and egg" question) between the two. To answer this question, cell cultures were incubated with either [35S]Hcy or [35S]thiolactone, and the fate of the 35S-compounds was determined.
Hcy thiolactone was detected in CBS-deficient fibroblast cultures (GM01128), but not in unaffected fibroblasts (GM01126), even in the presence of aminopterin, labeled with [35S]Hcy (Table II). Breast tumor cells produced more thiolactone than GM01128. Addition of aminopterin to these cultures resulted in up to 3-fold increase in the levels of Hcy thiolactone (Table II). In these experiments, up to 0.1% of [35S]Hcy was metabolized to [35S]thiolactone and up to 5% was converted into protein [35S]methionine (Table II). These levels of conversion indicate that the metabolism of exogenous Hcy is less efficient than the metabolism of exogenous methionine; up to 0.5 and 95% of [35S]methionine added to culture media is converted into [35S]thiolactone and into protein [35S]methionine, respectively. An apparent inhibition by aminopterin of the conversion of Hcy into protein methionine in HTB-132 is most likely due to greater sensitivity of cancer cells to inhibition by antifolates, relative to nontransformed cells (17).
|
It is remarkable that HTB-132 cells are more efficient than GM01126 and
GM011128 in converting exogenous Hcy into the thiolactone. In the
absence of aminopterin, the conversion of exogenous Hcy by HTB-132,
GM01126, and GM01128 into the thiolactone represents 5.4, <0.4, and
0.2%, respectively, of the conversion into protein methionine (see the
thiolactone/protein ratios in the "" aminopterin lanes in Table
II). Aminopterin increased these relative extents of Hcy conversion
into the thiolactone even more (see the "+" aminopterin lines in
Table II). Because Hcy thiolactone is metabolized very efficiently in
tissue cultures, in particular by breast cancer cells HTB-132
(half-life of the thiolactone in cultures of HTB-132 is 1 h; see
the following section below), these levels of Hcy editing represent
minimal values. Thus, more Hcy is diverted from the re-methylation
pathway into the thiolactone pathway (see Fig. 5) in breast cancer
cells HTB-132 than in nontransformed cells GM01126 and GM01128. These
observations further support our original suggestion (9) that the
inability of some tumor cells to grow on Hcy could be due to enhanced
editing of Hcy, which generates an ATP-consuming futile cycle, in these
cells relative to nontransformed cells (see Fig. 5).
Cell
cultures of GM01374, GM01375, and HTB-132 were incubated with
[35S]Hcy thiolactone. At time intervals, the remaining
thiolactone was determined by two-dimensional TLC. The kinetics of the
disappearance of [35S]Hcy thiolactone are shown in Fig.
2. The fastest metabolism of the thiolactone was
observed in cultures of HTB-132; the half-life of the thiolactone in
these cultures was 1 h. In cultures of GM01374 and GM01375, the
thiolactone was metabolized with a half-life of about 7 h. In
controls with culture media without cells, the thiolactone disappeared
with a half-life of 12 h, about twice as fast as in a buffer of
the same pH as the medium (see Table IV below). This indicates that
serum alone contributes to the decomposition of the thiolactone (see
below) in tissue cultures. The addition of 0.1 mM
methionine to the GM01375 culture had only a minor effect on the
metabolism of the thiolactone (half-life of 9 h), suggesting that
most of the metabolism of [35S]thiolactone does not
involve incorporation into protein methionine (which would be abolished
by excess unlabeled methionine). In fact, most of the thiolactone was
hydrolyzed to Hcy which was present on the two-dimensional TLC
chromatograms of the media (not shown). About 20% of the
[35S]Hcy thiolactone was converted into protein
[35S]methionine which was demonstrated by two-dimensional
TLC analysis of acid hydrolysates of the labeled protein (not shown).
Thus, efficient turnover of the exogenous thiolactone and inefficient metabolism of exogenous Hcy suggest that the secreted thiolactone is a
precursor of the extracellular Hcy in the cultures maintained on
methionine.
|
In experiments with exogenous [35S]Hcy thiolactone, only [35S]methionine was recovered from acid hydrolysates of cellular proteins. No detectable [35S]Hcy was present in the hydrolysates of proteins from cultures labeled with [35S]Hcy thiolactone. However, because exogenous [35S]Hcy thiolactone is rapidly turned over in tissue cultures and only limited quantities of [35S]Hcy thiolactone were available, the sensitivity of the analyses would allow detection of only relatively high levels (>1%) of [35S]Hcy incorporation into protein in these experiments. These limitations can be at least partially circumvented in tissue cultures labeled with [35S]methionine in which [35S]Hcy thiolactone is produced continuously.
35S-Labeled proteins were prepared from cell cultures labeled with [35S]methionine. Free [35S]Hcy, [35S]thiolactone, as well as any [35S]Hcy bound to protein through the disulfide bonds were removed from protein preparations by four rounds of precipitations, washes, and reductions with DTT and unlabeled Hcy. The final preparation of the labeled protein was hydrolyzed with 6 N HCl using standard procedures. 3-Mercaptopropionic acid was present during acid hydrolysis to prevent oxidation of any Hcy that might be present in protein. Because Hcy easily cyclizes to the thiolactone under acidic conditions (12, 13), any protein Hcy, if present, would be converted to Hcy thiolactone during acid hydrolysis of protein. In a control experiment, complete conversion of Hcy to the thiolactone occurred within 5 min in the presence of 6 N HCl at 100 °C. To determine levels of Hcy thiolactone formed in acid hydrolysates of protein, the hydrolysates were lyophilized, taken up in 10 µl of water, and subjected to two-dimensional TLC. Quantitation of the levels of Hcy incorporated into proteins in HTB-132, as well as quantitation of similar analyses carried out with proteins from CBS-deficient GM01128, GM01374, and unaffected fibroblasts GM01126, GM01375, are presented in Table III.
|
Hcy was present in proteins from each analyzed cell culture. The level
of Hcy incorporation, normalized to methionine incorporation, was from
0.0004 to 0.0024 (Table III). Similar levels of Hcy incorporation into
protein from CBS-deficient (normalized Hcy incorporation was 0.0005 for
GM01374 and 0.0007 for GM01128) and unaffected fibroblasts (the values
were 0.0006 for GM01375 and 0.0004 for GM01126) were observed. However,
consistently higher values (by up to 67%) were obtained for Hcy
incorporation into protein from aminopterin-treated cells than from
untreated cells of CBS-deficient and unaffected fibroblasts (compare
the "" and "+" aminopterin lines for each cell line in Table
III). Incorporation of Hcy into protein from breast tumor cells HTB-132
was enhanced 2.4-fold by aminopterin (from a value of 0.0010 to
0.0024). The experiments described in the following section suggest
that incorporation of Hcy into protein occurs most likely
post-translationally as a result of acylation of side chain lysine
groups by Hcy thiolactone (Fig. 4).
Reactions of Hcy Thiolactone with Proteins and Amino Acids
Hcy thiolactone, a thioester, is expected to acylate any nucleophilic group, in particular free amino groups in proteins (14, 15). However, it is not known how efficient and extensive these reactions might be under physiological conditions. To answer this question, the reactions of the thiolactone with amino acids and proteins under physiological conditions (pH 7.4, 37 °C) were followed by UV spectrometry, and the products were also analyzed by TLC.
As shown in Table IV, Hcy thiolactone is relatively stable at physiological values of pH and temperature and hydrolyzes with a half-life of 25 h, consistent with previous measurements (15). In the presence of some amino acids or glutathione, but not DTT or 2-mercaptoethanol, the thiolactone disappeared much faster. Each amino acid accelerated decomposition of the thiolactone to a different degree. The fastest decomposition of the thiolactone, with a half-life of 3 h, was observed in the presence of lysine or cysteine. Less effective were histidine > arginine, serine, glutathione, homocysteine > alanine.
Fig. 3 shows results of one-dimensional TLC analyses of
products of reactions of [35S]Hcy thiolactone with some
amino acids, proteins, and human serum. Fig. 3A shows direct
TLC analysis (without DTT treatment; some products are therefore
present in disulfide forms that either stay at the origin or migrate
just above the origin) of the products after a 4-h reaction, which does
not lead to extensive hydrolysis of the thiolactone in controls (see
also Table IV). Fig. 3B shows TLC analyses of samples (after
22-h reactions) that have been treated with DTT to convert disulfide
forms of products into free sulfhydryl forms. With 4 × diluted
human serum, most of the thiolactone has reacted within 4 h, and a
thiolactone-serum protein adduct, that stays at the origin of the TLC
plate, was a major product (Fig. 3A, lane 8).
About half of the protein-bound Hcy from the adduct was released after
DTT treatment (Fig. 3B, lane 8) suggesting that
the other portion of Hcy in the adduct is bound via non-disulfide bonds. These data are consistent with previous analyses of reactions of
Hcy thiolactone with nondiluted human serum (15). Similar non-disulfide
bond-bound Hcy was also a major product in reactions of the thiolactone
with bovine serum albumin (Fig. 3, A and B, lane 3) and histone (Fig. 3, A and B, lane
2). More thiolactone was incorporated to histone (a protein of
high lysine content) than to albumin (a protein of average lysine
content), which suggests that a major site for homocysteinylation are
lysine residues in the proteins. This conclusion is supported by the
observation that lysine itself reacts with Hcy thiolactone to give an
adduct that contains free thiol groups (see the major spot just below Hcy in lane 6 of Fig. 3B). A disulfide form of
the Hcy-Lys adduct stays at the origin of the TLC plate (see lane
6 in Fig. 3A). Hcy thiolactone also reacts with
cysteine (lanes 4 in Fig. 3, A and B)
to give a Hcy-Cys adduct. Products of the reactions of the thiolactone
in the presence of serine (lanes 5) and alanine (lanes
7) are not much different from the control which yields mostly
homocystine disulfide (which migrates close to the origin in lane
1 of Fig. 3A) or Hcy after DTT treatment (lane
1 in Fig. 3B).
The data presented in this paper demonstrate that 1) Hcy thiolactone is an important component of the metabolism of thio-amino acids in human cells, in particular under pathological conditions in which Hcy metabolism is deregulated; and 2) Hcy is incorporated into proteins in human cells, most likely by post-translational acylation of lysine residues by the thiolactone (Fig. 4). These data suggest a mechanism by which excessive levels of Hcy can initiate cellular damage and lead to pathological consequences such as vascular disease. The data also suggest that a futile cycle of synthesis and hydrolysis of Hcy thiolactone, which wastes ATP, would be generated under conditions of deregulated Hcy metabolism, which could account for the inability of tumor cells to grow on Hcy.
The metabolism of Hcy thiolactone and its relationship to the metabolism of methionine and Hcy in human cell cultures is depicted in Fig. 5. Most of the exogenous methionine (~90%) that is taken up by cells is incorporated into tRNA by methionyl-tRNA synthetase (MRS in Fig. 5) to provide Met-tRNA, the immediate substrate for protein synthesis. About 7% of methionine is converted into Hcy (in HTB-132 cells, this work) via AdoMet and AdoHcy, the immediate precursor of Hcy. In other studies, 12% of methionine was converted into Hcy in low folate cultures of human fibroblasts (16) and in methotrexate-treated cultures of transformed mouse fibroblasts (17). Hcy is then re-methylated back to methionine by a folate- and vitamin B12-dependent enzyme, methionine synthase (MS in Fig. 5). Hcy is also converted into cysteine via cystathionine with CBS, a vitamin B6-dependent enzyme, which is the first enzyme in the pathway (11). In another pathway, originally observed in tumor cells, methionyl-tRNA synthetase converts Hcy into Hcy thiolactone (9). The present work demonstrates production of Hcy thiolactone in CBS-deficient and unaffected human fibroblasts, presumably also by methionyl-tRNA synthetase. As also demonstrated in the present work, limiting or preventing the flow of Hcy through the CBS pathway (as in CBS-deficient fibroblasts) and/or the methionine synthase pathway (by using the antifolate drug aminopterin) leads to an increased flow of Hcy into the thiolactone pathway in human cells. Because of its mostly neutral character under physiological conditions (pK of the amino group of the thiolactone is 7.1, Ref. 18), Hcy thiolactone is secreted from cells and accumulates in culture media (Fig. 5).
As shown in the present work, Hcy thiolactone is efficiently metabolized in tissue cultures, mostly by hydrolysis to Hcy which also accumulates in the culture media. This provides a mechanistic explanation for "Hcy export" from cultured human cells observed originally in other studies (16, 19, 20). A mechanism involving intracellular conversion of Hcy to Hcy thiolactone, excretion of the thiolactone from tissues, followed by its extracellular hydrolysis (Fig. 5) may also explain the origin of Hcy in human plasma. A fraction of Hcy that enters cells is either re-methylated and incorporated into protein as methionine (21), converted into the thiolactone (this work, Table II), or trans-sulfurylated via the CBS pathway to cysteine which in turn can be incorporated into protein (21, 22). High concentrations of (nonradiolabeled) cysteine in media used in the present work prevent detection of radiolabeled protein cysteine. Diminishing or preventing the flow of Hcy through the methionine synthase and CBS pathways (Fig. 5), which frequently occurs in tumor cell lines (23-26), would lead to a futile cycle that involves ATP-consuming conversion of Hcy to the thiolactone by methionyl-tRNA synthetase (9), followed by hydrolysis of the thiolactone back to Hcy (Fig. 5). It is well established that at least half of the known tumor cells are, in contrast to normal cells, methionine-dependent and do not grow on media in which Hcy has replaced methionine (24-28), a property that has been exploited for design of cancer chemotherapy (28-31). The futile cycling between Hcy and the thiolactone may explain the inability of some cancer cells to grow on Hcy.
Hcy is also incorporated into cellular proteins in all cell lines tested. Although at present it has not been possible to directly determine the mechanism (translational versus post-translational) and the site of Hcy incorporation in human cell proteins, studies in model systems suggest a plausible post-translational mechanism. For example, proteins in human serum (Ref. 15 and this work), as well as pure proteins (this work), are easily acylated by Hcy thiolactone under physiological conditions of pH and temperature. A protein with a high lysine content, such as histone, is homocysteinylated to a greater extent than a protein with an average lysine content, such as bovine serum albumin. Of several amino acids tested in this work, lysine (but not alanine) was shown to easily react with the thiolactone forming a Hcy-Lys adduct containing a free sulfhydryl group. Thus, acylation of the side chain amino group of a lysine residue in protein by the activated carboxyl group of the thiolactone is a likely mechanism for the observed incorporation of Hcy into protein in human cells (Fig. 4).
There was a weak correlation between the degree of protein
homocysteinylation and the levels of either Hcy or the thiolactone present in tissue culture. For example, although the "+"
aminopterin cultures of GM01374 and GM01375 produced 5-7-fold more
thiolactone than the "" aminopterin cultures (Table I), there was
only 40-67% increase in protein homocysteinylation in the "+"
aminopterin cultures (Table III). This suggests that the fibroblasts
may have a mechanism that prevents or corrects excessive
homocysteinylation of cellular proteins. This mechanism may be somewhat
impaired in breast tumor cells HTB-132 in which a 2.4-fold increase in homocysteinylation of proteins in the "+" aminopterin cultures, relative to the "
" aminopterin cultures, better correlates with a
6.5-fold increase in Hcy thiolactone levels observed in the "+"
versus "
" aminopterin cultures.
As shown in this work, pure proteins or proteins present in human serum can be homocysteinylated by the thiolactone under physiological conditions of pH and temperature. However, because of rapid metabolism of the exogenous thiolactone, there was no detectable homocysteinylation of cellular proteins in tissue cultures incubated with the thiolactone. These observations, as well as a previous report of enzymatic activities that hydrolyze Hcy thiolactone in human plasma and in cultured arterial endothelial cells (15), also suggest that there is a mechanism that prevents excessive protein homocysteinylation in the cell.
Although one can expect that Hcy thiolactone is produced in whole
organisms as it is in cultured cells, subsequent turnover of the
thiolactone in body tissues and fluids might be faster than that
observed in tissue culture experiments. The levels of Hcy thiolactone
in tissue cultures, determined in this study, are no higher than
one-tenth of the levels of Hcy (Fig. 1 and Table I). If the thiolactone
represents a similar fraction of Hcy in the whole body, one can expect
to find 1 µM thiolactone, for example, in human serum.
To detect these levels of the thiolactone, more sensitive methods than
those used so far are required. Hcy thiolactone has been reported to be
absent in human serum and plasma (15, 32, 33) and in rabbit plasma and
urine (34), but the methods used were relatively insensitive with
detection limits of 10-50 µM. A claim that Hcy
thiolactone is present in pig liver is not supported by the presented
evidence (35).
Numerous epidemiological studies have established that elevated levels of Hcy are an independent risk factor for premature atherosclerotic vascular disease (1-4). However, it is not known what aspect of Hcy metabolism is harmful to human cells and induces atherosclerosis. Several hypotheses have been proposed to explain the mechanisms involved in the pathology caused by Hcy. Hcy has been shown to affect the function of the blood clotting system thereby promoting thrombotic tendency (36-38). Hcy has also been shown to induce oxidative damage of endothelial cells (39, 40), although this has been questioned recently (41), and to promote vascular smooth muscle growth and inhibit regeneration of endothelial cell (42) which would result in atherosclerosis. All these effects are unlikely to have physiological relevance because they were observed at much higher than physiological concentrations of Hcy and, where reported, other thiols, such as cysteine and 2-mercaptoethanol, had similar effects. Thus, these hypotheses are based on experiments whose major weaknesses are the lack of demonstrable specificity with respect to Hcy (the effects were caused by thiols in general) and the use of extremely high concentrations of Hcy (thiols).
The data described in this paper provide the basis for an alternative hypothesis that explains the mechanisms involved in the pathology of Hcy but does not have the weaknesses of the other hypotheses: post-translational modification of cellular proteins by homocysteinylation can be a mechanism that initiates cellular damage resulting in pathological consequences such as cardiovascular disease. For example, deficient collagen cross-linking observed in patients with homocystinuria (43) can be due to modification of collagen lysine residues by homocysteinylation. Purified human low density lipoprotein, a major factor in atherosclerosis, can also be homocysteinylated by the thiolactone (44). Homocysteinylation of proteins can modify their function and/or be autoimmunogenic which would lead to cell damage. Homocysteinylation of cellular or extracellular proteins is absolutely specific for Hcy because Hcy thiolactone, the actual agent creating the modification, can only form from Hcy in any cell type. Moreover, this mechanism does not require nonphysiological concentrations of Hcy; it is easily observed in tissue cultures that contain as little as 0.2 nM thiolactone in addition to 60 nM Hcy and is enhanced at higher levels of the thiolactone and Hcy. Initial small amounts of protein modification can accumulate to significant levels over extended periods that are usually required for development of atherosclerotic disease, in analogy to the accumulation of protein glycation products in the brain of Alzheimer's patients which occurs over their life spans (reviewed in Ref. 45).