3-Hydroxylysine, a potential marker for studying radical-induced protein oxidation

Article abstract of DOI:10.1021/tx980118h

γ-Irradiation of several amino acids (Val, Leu, Ile, Lys, Pro, and Glu) in the presence of O₂ generates hydroperoxides. We have previously isolated and characterized valine and leucine hydroperoxides, and hydroxides, and have detected these products in both isolated systems [e.g., bovine serum albumin (BSA) and human low-density lipoprotein (LDL)] and diseased human tissues (atherosclerotic plaques and lens cataractous proteins). This work was aimed at investigating oxidized lysine as a sensitive marker for protein oxidation, as such residues are present on protein surfaces, and are therefore likely to be particularly susceptible to oxidation by radicals in bulk solution. HO°attack on lysine in the presence of oxygen, followed by NaBH₄ reduction, is shown to give rise to (2S)-3-hydroxylysine [(2S)-2,6-diamino-3- hydroxyhexanoic acid], (2S)-4-hydroxylysine [(2S)-2,6-diamino-4- hydroxyhexanoic acid], (2S,5R)5-hydroxylysine [(2S,5R)-2,6-diamino-5- hydroxyhexanoic acid], and (2S,5S)-5-hydroxylysine [(2S,5S)-2,6-diamino-5- hydroxyhexanoic acid]. 5-Hydroxylysines are natural products formed by lysyl oxidase and are therefore not good markers of radical-mediated oxidation. The other hydroxylysines are however useful markers, with HPLC analysis of 9- fluorenylmethyl chloroformate (FMOC) derivatives providing a sensitive and accurate method for quantitative measurement. Hydroxylysines have been detected in the hydrolysates of peptides (Gly-Lys-Gly and Lys-Val-Ile-Leu- Phe) and proteins (BSA and histone H1) exposed to HO°/O₂, and subsequently treated with NaBH₄. Quantification of the hydroxylysines yields, and comparison with hydroxyvalines and hydroxyleucines, supports the hypothesis that surface residues give higher yields of oxidized products than the hydrophobic leucines and valines, at least with globular proteins such as BSA. Hydroxylysines, and particularly 3-hydroxylysine, may therefore be sensitive and useful markers of radical-mediated protein oxidation in biological systems.

Full text of DOI:10.1021/tx980118h

Chem. Res. Toxicol. 1998, 11, 1265-1273
1265
3-Hyd r oxylysin e, a P oten tia l Ma r k er for Stu d yin g
Ra d ica l-In d u ced P r otein Oxid a tion †
Be´ne´dicte Morin,‡ William A. Bubb,§ Michael J . Davies,‡ Roger T. Dean,*,‡ and
Shanlin Fu‡
The Cell Biology and EPR Groups, The Heart Research Institute, 145 Missenden Road,
Camperdown, Sydney NSW 2050, Australia, and Department of Biochemistry, University of Sydney,
Sydney NSW 2006, Australia
Received May 26, 1998
γ-Irradiation of several amino acids (Val, Leu, Ile, Lys, Pro, and Glu) in the presence of O2
generates hydroperoxides. We have previously isolated and characterized valine and leucine
hydroperoxides, and hydroxides, and have detected these products in both isolated systems
[e.g., bovine serum albumin (BSA) and human low-density lipoprotein (LDL)] and diseased
human tissues (atherosclerotic plaques and lens cataractous proteins). This work was aimed
at investigating oxidized lysine as a sensitive marker for protein oxidation, as such residues
are present on protein surfaces, and are therefore likely to be particularly susceptible to
oxidation by radicals in bulk solution. HO• attack on lysine in the presence of oxygen, followed
by NaBH4 reduction, is shown to give rise to (2S)-3-hydroxylysine [(2S)-2,6-diamino-3-
hydroxyhexanoic acid], (2S)-4-hydroxylysine [(2S)-2,6-diamino-4-hydroxyhexanoic acid], (2S,5R)-
5-hydroxylysine [(2S,5R)-2,6-diamino-5-hydroxyhexanoic acid], and (2S,5S)-5-hydroxylysine
[(2S,5S)-2,6-diamino-5-hydroxyhexanoic acid]. 5-Hydroxylysines are natural products formed
by lysyl oxidase and are therefore not good markers of radical-mediated oxidation. The other
hydroxylysines are however useful markers, with HPLC analysis of 9-fluorenylmethyl
chloroformate (FMOC) derivatives providing a sensitive and accurate method for quantitative
measurement. Hydroxylysines have been detected in the hydrolysates of peptides (Gly-Lys-
Gly and Lys-Val-Ile-Leu-Phe) and proteins (BSA and histone H1) exposed to ΗÕ/O2, and
subsequently treated with NaBH4. Quantification of the hydroxylysines yields, and comparison
with hydroxyvalines and hydroxyleucines, supports the hypothesis that surface residues give
higher yields of oxidized products than the hydrophobic leucines and valines, at least with
globular proteins such as BSA. Hydroxylysines, and particularly 3-hydroxylysine, may
therefore be sensitive and useful markers of radical-mediated protein oxidation in biological
systems.
Glu, and Lys) give significantly higher yields of hydro-
peroxides than the others (1, 2). Such selectivity is in
agreement with the hypothesis that most of the hydro-
peroxide groups occur on the side chains of free aliphatic
amino acids.
In tr od u ction
Free radicals are known to be generated either by
normal metabolic processes during electron transport
chains and redox reactions of enzymes or as a result of
exposure to exogenous factors such as UV light, radiation,
and various chemicals. Until recently, it had been
thought that radical damage on proteins was essentially
a chain-terminating process and that the damage pro-
duced (cross-linking, fragmentation, and chemical modi-
fication of the amino acids) was mainly the result of
inactive products. However, it has been shown recently
that two products of protein oxidation, i.e., protein-bound
reducing species, believed to be mainly 3,4-dihydroxy-
phenylalanine and protein hydroperoxides, are able to
initiate further chemical reaction. Studies on free amino
acids have shown that six amino acids (Val, Leu, Ile, Pro,
Carbonyl formation on proteins has been used exten-
sively to determine the oxidative state of proteins (3, 4);
however, this method is not specific for carbonyl groups
on proteins associated with oxygen radical attack, and
it can be confounded by carbonyls resulting from other
processes such as glycoxidation and lipid peroxidation.
As a result, it is often difficult to use in biological
samples. Therefore, more specific methods for analyzing
the oxidative state of proteins are needed. In the course
of searching for a suitable marker for studying protein
oxidation, we have previously focused on the oxidation
products of valine and leucine. We have extended earlier
work of Garrison and Kopoldova (reviewed in ref 5) by
characterizing the structures of three hydroxides of valine
and five oxidation products of leucine produced upon
γ-radiolysis in the presence of oxygen. These oxidation
products have been detected with both isolated proteins
and a number of biological samples, including human
† This study was supported in part by the Australian Research
Council and the award of an ARC QE2 fellowship to M.J .D.
* Corresponding author: The Heart Research Institute, 145 Mis-
senden Rd., Camperdown, Sydney NSW 2050, Australia. Telephone:
+61 2 9550 3560. Fax: +61 2 9550 3302. E-mail: r.dean@hri.org.au.
‡ The Heart Research Institute.
§ University of Sydney.
10.1021/tx980118h CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/03/1998

1266 Chem. Res. Toxicol., Vol. 11, No. 11, 1998
Morin et al.
from EMScience (Gibbstown, NY). Tetrahydrofuran was from
BDH (Poole, England).
atherosclerotic plaques, normal human plasma and low-
density lipoprotein (LDL1), cataract crystallins, and
glycated rat-tail collagen (6-11) (S.-L. Fu and R. T. Dean,
submitted for publication). Hydroxyleucines, like hy-
droxyvalines, have been proposed as useful markers to
assess the extent of damage occurring on proteins during
various pathological processes. Another novel hydroxyl-
ated product (5-hydroxy-2-aminovaleric acid) has been
identified in radical-damaged proteins (12). This product
is thought to arise from the oxidation of proline and
arginine residues and has been postulated to be a specific
marker of radical attack on proteins.
γ-Ir r a d ia tion of Lysin e, P ep tid es, a n d P r otein s. Ir-
radiations of lysine (4 mM), tripeptide GKG (2 mM), histone
H1 (4 mg/mL), and BSA (5 mg/mL) solutions were carried out
using a 60Co source (dose rate of ca. 21 Gy/min), and all samples
were gassed throughout the irradiation with oxygen. After
irradiation, the solutions were stored at -20 °C until they were
used. The absolute quantity of hydroperoxides generated from
γ-radiolysis was determined using an iodometric assay after
adding a small volume of catalase (5 µg/mL) to the samples to
remove radiation-generated hydrogen peroxide (6, 16).
Ga s-P h a se Am in o Acid Hyd r olysis. The gas-phase hy-
drolysis method developed to obtain the best recovery for DOPA
from protein hydrolysis was adopted (6). With this method, we
utilized 1 mL of 6 M HCl containing 1% (w/v) phenol and 50 µL
of mercaptoacetic acid as reductants. The hydrolysate was
subsequently analyzed by HPLC.
In this study, we focused on hydroxyl radical attack
on lysine residues in proteins. Although valine and
leucine as free amino acids are among the principal
targets of hydroxyl radicals as assessed by the high yield
of hydroperoxides, lysine is of potential interest because,
as revealed by structural databases, the majority of these
residues are preferentially located at the protein surface.
EPR studies have shown that charged amino acids such
as lysine and glutamate are important sites of formation
of hydroperoxides (13, 14). Furthermore, protein surfaces
play an important role in cellular mechanisms through
their interactions with other molecules. 5-Hydroxylysine
has been identified previously with both free lysine and
lysine-containing peptides, as a product of metal ion-
catalyzed oxidation of lysine (15). We report here the
identification of four oxidation products of lysine, (2S,5R)-
5-hydroxylysine, (2S,5S)-5-hydroxylysine, 4-hydroxy-
lysine, and 3-hydroxylysine, from γ-radiolysed lysine in
the presence of oxygen. Among these, 3-hydroxylysine
has been isolated by high-performance liquid chroma-
tography with 9-fluorenylmethyl chloroformate (FMOC)
precolumn derivatization. The characterization of this
product was carried out on the basis of extensive spec-
troscopic measurements, including electrospray mass
F MOC Der iva tiza tion of Am in o Acid s. To 10 µL of
sample were added 90 µL of borate buffer (0.1 M, pH 9.4) and
100 µL of FMOC (15 mM in acetone). The mixture was vortex
mixed and after 1 min was extracted with 1 mL of pentane. The
extraction was repeated twice to remove FMOC-OH, produced
by reaction of FMOC-Cl with water. The aqueous solution was
then diluted 10-fold for HPLC analysis.
Sep a r a tion of F MOC-Der iva tized Lysin e Hyd r oxid es
for Str u ctu r a l An a lysis. To obtain enough material for
structural work, 500 mL of a 2 mM lysine solution was
γ-irradiated (1000 Gy). After reduction with NaBH4, the
solution was concentrated 100-fold under reduced pressure and
derivatized with FMOC by addition to 45 mL of sodium borate
buffer (0.1 M, pH 9.4) and 50 mL of FMOC in acetone (50 mM).
The solution was vortexed for 5 min, and the lysine derivatives
were extracted twice with 50 mL of pentane. The mixture was
separated by HPLC on a semipreparative Supelco C18 column
(25 cm × 10 mm, 5 µm particle size) with UV detection at 300
nm, isocratic elution with 55% solvent A [acetonitrile/tetrahy-
drofuran/20 mM sodium acetate at pH 4.2 (19.5:0.5:80 v/v)] and
45% solvent B [acetonitrile/20 mM sodium acetate at pH 4.2
(80:20 v/v)] with a flow rate of 4 mL/min. After concentration,
the four fractions collected from 30 injections were repurified
on the same column. An extra HPLC purification step was
performed using 55% solvent A [acetonitrile/20 mM ammonium
acetate at pH 4.2 (20:80 v/v)] and 45% solvent B [acetonitrile/
20 mM ammonium acetate at pH 4.2 (80:20 v/v)]. The samples
were then lyophilized a minimum of three times to free the
samples of ammonium acetate for MS and NMR studies.
1
spectrometry (ES-MS) and H and 13C NMR. 3-Hydroxy-
lysine has been detected on oxidized peptides and pro-
teins, and a comparison of 3-hydroxylysine formation to
that of other markers of protein oxidation (valine and
leucine hydroxides) has been made in various peptides
and proteins.
Ma ter ia ls a n d Meth od s
HP LC An a lysis of F MOC a n d OP A Der iva tives of
Am in o Acid s. The HPLC system consisted of a SIL-10A auto
injector (Shimadzu, Kyoto, J apan), two LC-10AT pumps (Shi-
madzu), a F-1080 fluorescence detector (Hitachi, Tokyo, J apan),
a SPD-10A UV detector (Shimadzu), and a column oven (30 °C,
Waters, Millipore, Milford, MA). Data were digitized using a
CBM-10A interface (Shimadzu) and processed on an IBM PC
123 computer.
Ma ter ia ls. o-Phthalaldehyde (OPA) crystals and OPA dil-
uent [3% KOH and 3% boric acid (pH 10.4)] were from Pickering
Laboratories (Mountain View, CA). 2-Mercaptoethanol, fatty
acid-free BSA, histone H1 (calf thymus), 5-hydroxylysine (mixed
DL and DL-allo), 2,4-dinitrophenylhydrazine, and 9-fluorenyl-
methyl chloroformate were from Sigma (St. Louis, MO). L-
Lysine and sodium borohydride were provided by Aldrich
(Milwaukee, WI). DL-trans-4,5-Dehydrolysine and the tripeptide
Gly-Lys-Gly (GKG) were obtained from Bachem (Budendorf,
Switzerland). (2S,5R)-5-Hydroxylysine [5(R)OHLys] was pur-
chased from Fluka (Bushs, Switzerland). Sodium borate and
sodium acetate (analytical grade) were from Merck (Darmstadt,
Germany). Water was purified by passage though a four-stage
Milli-Q system (Millipore-Waters) equipped with a 0.2 µm pore
size final filter. HPLC-grade methanol and acetonitrile were
(1) F MOC Der iva tiza tion of Lysin e. The FMOC deriva-
tives of amino acids were separated by HPLC using a Supelco
C18 column (4.6 mm × 25 cm, 5µm particle size) with
a
Pelliguard column (2 cm, Supelco) at a flow rate of 1 mL/min,
eluted with a gradient of solvent A [acetonitrile/tetrahydrofuran/
20 mM sodium acetate at pH 4.05 (20:5:75 v/v)] and solvent B
[same solvents (80:5:15 v/v)]. The gradient was generated as
follows: isocratic elution for 30 min at 32% B, then a gradient
to 70% B in 10 min, isocratic elution for 8 min, and then re-
equilibration at 32% B for 7 min for the next analysis. Oxidized
lysine products were monitored by fluorescence with an excita-
tion wavelength of 265 nm and an emission wavelength of 310
nm. A calibration curve was carried out with authentic pure
5(R)OHLys. The parent amino acids were monitored by UV
detection at 260 nm. The amount of protein analyzed was
calculated from the lysine peak area after appropriate calibra-
1
Abbreviations: BSA, bovine serum albumin; DNPH, 2,4-dinitro-
phenylhydrazine; DQF-COSY, double-quantum filtered correlated
spectroscopy; ES-MS, electrospray mass spectrometry; FMOC, 9-fluo-
renylmethyl chloroformate; HMQC, heteronuclear multiple-quantum
coherence; LDL, human low-density lipoprotein; 3OHLys, (2S)-3-
hydroxylysine; 4OHLys, (2S)-4-hydroxylysine; 5(R)OHLys, (5R,2S)-5-
hydroxylysine; 5(S)OHLys, (5S,2S)-5-hydroxylysine; OPA, o-phthal-
aldehyde.

Hydroxylysines as Markers of Protein Oxidation
Chem. Res. Toxicol., Vol. 11, No. 11, 1998 1267
tion. All the data are expressed as the amount of oxidized
product per lysine parent.
(2) OP A Der iva tiza tion of Va lin e a n d Leu cin e. The
method, which has been previously described (9), involves a
HPLC purification step on a LC-NH2 column (4.6 mm × 25 cm,
5µm particle size, Supelco) prior to the OPA-HPLC assay. The
eluting window (13.7-15.3 min) of both 3-hydroxyvaline (13.8
min) and 4-hydroxyleucines (14.4 and 15.0 min) was collected
using an isocratic elution with 83% acetonitrile in 10 mM
phosphate buffer (pH 4.3), eluting at 1.5 mL/min. The eluent
was monitored at 210 nm. After evaporation to dryness, this
fraction was derivatized with freshly made OPA reagent in the
presence of 2-mercaptoethanol (0.5% v/v) using an SIL-10A
autosampler (Shimadzu, South Rydalmere, Australia) as de-
scribed previously (6). 3-Hydroxyvaline and 4-hydroxyleucines
were analyzed using HPLC on an ODS column (9) and moni-
tored by fluorescence (λex ) 340 nm and λem ) 440 nm). The
second HPLC step was also used to measure the lysine parent
in the hydrolysate after appropriate dilution (usually ca. 1000-
fold).
Sp ectr oscop ic Mea su r em en ts. Electrospray (ES) MS was
performed on a VG Platform mass spectrometer (Fisons, Home-
bush, Australia). The samples were dissolved in 50% aqueous
acetonitrile. Formic acid (1%) was added to the solutions to
assist with protonation of the sample. The solvent was delivered
by a Phoenix (Fisons) syringe pump at a flow rate of 5 µL/min;
10 µL of each solution was injected. A dry nitrogen bath gas at
atmospheric pressure was employed to assist evaporation of the
electrospray droplets. The electrospray probe tip potential was
3.5 kV with 0.5 kV on the chicane counter electrode. A sample
cone potential of 15 V was used. The 1H spectra, double-
quantum filtered COSY (DQF-COSY) spectra, and hetero-
nuclear multiple-quantum coherence (HMQC) correlated NMR
spectra were recorded in DMSO-d6 (99.9%, Cambridge Isotopes,
Andover, MA) in the Fourier transform mode on a Bruker AMX-
600 NMR spectrometer. The chemical shifts are expressed in
parts per million with respect to the DMSO-d6 signal at δ 2.50
for the 1H spectra and δ 39.50 for the 13C spectra. Two-
dimensional spectra were acquired, using standard Bruker
software, over 2K data points and typically 512 t1 increments
with phase discrimination achieved using time-proportional
phase incrementation (TPPI). The spectra were routinely
processed with zero-filling in F1 and a π/2 sine-bell function in
each dimension.
Syn th esis of 4-Hyd r oxylysin e a n d 5-Hyd r oxylysin e by
Hyd r a tion of 4,5-Deh yd r olysin e. 4,5-Dehydrolysine (7 mg,
10 µΜ) was dissolved in 3.8 mL of water. Perchloric acid (70%,
185 µL, 0.55 M final concentration) was added and the mixture
stirred at reflux for 24 h. An aliquot was then removed,
neutralized with NaOH, and derivatized by the FMOC method
described above for HPLC analysis. 5OHLys was identified by
comparison with pure standard 5(R)OHLys and the mixture of
DL- and DL-allo-5-hydroxylysine.
Der iva tiza tion of Ald eh yd es by 2,4-Din itr op h en ylh y-
d r a zin e (DNP H). The purpose of this experiment was to
determine whether carbonyls were a significant source of
hydroxides upon borohydride reduction; this was addressed by
derivatizing carbonyls prior to borohydride reduction. Thus,
γ-irradiated lysine was mixed with a solution of 1 mM DNPH
in 1 M HCl (1:1) or with 1 M HCl alone (1:1) as a control, to
derivatize carbonyls. After incubation for 1 h at ambient
temperature, NaBH4 was added to the sample. Lysine hy-
droxides were then derivatized using FMOC and analyzed by
HPLC.
F igu r e 1. HPLC detection of FMOC derivatives of γ-irradiated
lysine: (a) a γ-irradiated lysine solution and (b) a γ-irradiated
and reduced lysine solution. For further details, see Materials
and Methods.
high hydrophilicity and the charged nature of lysine.
Derivatization with 9-fluorenylmethyl chloroformate
(FMOC) was therefore employed as this gives highly
fluorescent species which are stable (17, 18) in contrast
to those with o-phthalaldehyde (19). This derivatization
method was optimized by examining the effects of dif-
ferent extraction and FMOC solvents, and different
reaction times, and the method tested for linearity and
reproducibility.
(1) E ffect s of E xt r a ct ion Solven t a n d F MOC
Solven t. Two pH values (7.7 and 9.4) for the borate
buffer (0.1 M) and two FMOC solvents (acetone and
acetonitrile) were used to determine the best reaction
conditions. FMOC in acetone and borate buffer at pH
9.4 provided the best recovery for lysine and 5(R)OHLys
as well as a minimal contamination with hydrolyzed
FMOC after appropriate extraction (data not shown).
Two extraction solvents, ethyl acetate and pentane, were
evaluated; pentane was employed as ethyl acetate,
though efficient at removing hydrolyzed FMOC, caused
loss of the lysine derivatives as previously described (17).
(2) Effects of Rea ction Tim e. A reaction time of 30
s to 1 min gave the maximum fluorescence yield; longer
times gave lower values.
(3) Lin ea r ity a n d Rep r od u cibility. The linearity
of the method was tested with lysine standards (1, 2.5,
5, 7.5, and 10 mM) with UV detection (260 nm) as well
as with (2S,5R)-5-hydroxylysine (10, 25, and 50 µΜ) with
fluorescence detection (λex ) 265 nm and λem ) 310 nm).
Five separate derivatizations were performed at each
concentration level. Correlation coefficients for lysine
and 5(R)OHLys were 0.997 and 0.998, respectively.
Isolation an d Ch ar acter ization of th e Main FMOC
Der iva tive P r od u cts of γ-Ir r a d ia ted Lysin e Solu -
tion s. The oxidation products of lysine, derivatized by
FMOC, were analyzed by HPLC before and after reduc-
tion by NaBH4. The HPLC chromatograms showed
similar traces for the reduced and nonreduced samples
(Figure 1) with the exception of the two latest eluting
peaks (5 and 6) in the nonreduced samples. At least four
different products (1-4) can be detected from the γ-ra-
diolysed and reduced lysine. These products have been
identified as FMOC derivatives of (2S,5R)-5-hydroxy-
Resu lts
F MOC Der iva tiza tion of th e Sa m p les. Initial at-
tempts to detect and separate the oxidized products of
lysine on an analytical LC-NH2 column with an aceto-
nitrile/aqueous 10 mM sodium phosphate solvent system
gave broad overlapping peaks; this is probably due to the

1268 Chem. Res. Toxicol., Vol. 11, No. 11, 1998
Morin et al.
F igu r e 2. Structures of lysine and its oxidation products together with some of the corresponding valine and leucine materials. For
those marked with a superscript a, the absolute configuration was not determined.
lysine [(2S,5R)-2,6-diamino-5-hydroxyhexanoic acid] (1),
its diastereoisomer (2S,5S)-5-hydroxylysine [(2S,5S)-2,6-
diamino-5-hydroxyhexanoic acid] (2), (2S)-3-hydroxy-
lysine [(2S)-2,6-diamino-3-hydroxyhexanoic acid] (3), and
(2S)-4-hydroxylysine [(2S)-2,6-diamino-4-hydroxyhex-
anoic acid] (4), on the basis of spectroscopic measure-
ments and chemical evidence (vide infra) (Figure 2).
None of these products were present in the native lysine
(Figure 3).
ES mass spectra exhibit a pseudomolecular ion at m/z
607.8 [M + H+], which differs by 16 mass units from the
lysine parent at m/z 591.7. This is indicative of the
addition of a hydroxy group to the molecule and the fact
that they differ from each other only by the position of
the hydroxy group in their structures. The ion at m/z
369.5 in L-lysine corresponds to a molecular ion minus
one FMOC group (mass of 223) presumably arising from
the cleavage of one FMOC group and a concomitant
transfer of two hydrogens to the residual fragment
containing one FMOC group. This was not observed in
the oxidized products.
ES-MS. The molecular weight of the four oxidized
lysine derivatives 1-4 purified by HPLC was determined
to be 607 by mass spectrometry (data not shown). The

Hydroxylysines as Markers of Protein Oxidation
Chem. Res. Toxicol., Vol. 11, No. 11, 1998 1269
age), and 4-hydroxylysine is poorly recovered from pro-
tein hydrolysis, the structure of the FMOC derivative of
compound 3 was investigated as a potential marker for
oxidative damage.
Ch a r a cter iza tion of 3-Hyd r oxylysin e (3). Further
support for the assignment of peak 3 to 3-hydroxylysine
was provided by the detailed NMR data of the di-FMOC
derivative (Tables 1 and 2). 1H assignments were
1
determined from a H-1H COSY spectrum (Figure 4) by
sequential correlation with the NH resonances at each
end of the molecule. The corresponding 13C resonance
frequencies were established from an HMQC experi-
ment.
First, we observed the loss of two protons in the upfield
1
region of the H spectrum with respect to the six protons
corresponding to the three methylenic groups at C(3),
C(4), and C(5) in lysine. This is in agreement with the
addition of an OH group to one of the methylenic groups
of the lysine side chain. The methinic proton in the
resultant CHOH group resonates at low field (δ 3.58
ppm). The 1H chemical shifts (Table 1) are evidently
strongly influenced by conformational effects and, except
for the proton H(3) directly attached to the hydroxylated
carbon, cannot be reliably used for confirmation of the
structure (21). We can nevertheless note the magnetic
nonequivalence of the two protons at C(4) due to the
formation of a new chiral center at C(3).
F igu r e 3. HPLC detection of FMOC derivatives of lysine
hydroxides: (a) γ-irradiated and reduced lysine solution after
protein hydrolysis treatment (see Materials and Methods), (b)
γ-irradiated and reduced lysine solution, and (c) unirradiated
lysine.
Ch a r a cter iza tion of (2S,5S)-5-Hyd r oxylysin e (1)
a n d (2S,5R)-5-Hyd r oxylysin e (2). Formation of the
two diastereoisomers of 5-hydroxylysine in γ-irradiated
lysine solutions was confirmed by injection in the HPLC
system of a pure (2S,5R)-5-hydroxylysine standard and
a mixture of (2S,5R)- and (2S,5S)-5-hydroxylysine stand-
ards after derivatization by FMOC.
Ch a r a ct er iza t ion of (2S)-4-H yd r oxylysin e (4).
Chemical evidence supporting the identification of 4OHLys
came from the synthesis of a mixture of 4-hydroxylysine
and 5-hydroxylysine by hydration of 4,5-dehydrolysine
using perchloric acid. The mixture was derivatized by
FMOC and analyzed by HPLC. The HPLC trace was
compared to those of γ-irradiated lysine solutions and a
mixture of 5R and 5S diastereoisomers of the 5OHLys
standard. Four products were formed during this syn-
thesis corresponding to the two diastereoisomers of
5OHLys and presumably the two diastereoisomers of
4OHLys. It is interesting to note that only one dia-
stereoisomer of 4OHLys was detected with irradiated
lysine as no HPLC peak was observed to elute after the
4OHLys diastereoisomer formed in irradiated lysine
(Figure 1b).
Sta bility of Lysin e Hyd r oxid es u n d er Acid Hy-
d r olysis Con d ition s. Previous studies have shown that
acid-catalyzed protein hydrolysis can result in loss of
amino acids which contain a hydroxy group in their
structures (9, 20). Therefore, we studied the stability of
the various lysine hydroxides under acidic conditions. The
recoveries of the hydroxylysines were ca. 40 and 80% for
5(S)OHLys (1) and 5(R)OHLys (2), respectively, 80% for
compound 3, and <10% for 4OHLys (4) (Figure 3a). This
is in agreement with previous studies where recoveries
of 80 and 20% were reported for 3-hydroxyvaline and
4-hydroxyvaline, respectively (20). Similarly, protein
hydrolysis of a γ-radiolysed leucine solution gave 70% and
<5% recoveries for 5-hydroxyleucines and 4-hydroxy-
leucine, respectively (9). The low recoveries observed for
4-hydroxyamino acids are probably due to the formation
of γ-lactones under acidic conditions. Lysine was almost
fully recovered under similar conditions.
The 13C data (Table 2) are entirely consistent with
incorporation of a hydroxy group at C(3), with chemical
shifts to higher frequency for the directly attached and
adjacent C(2) and C(4), and to lower frequency for C(5),
compared with the parent lysine derivative (22). Similar
effects are observed for 5(R)OHLys.
The assignments for the 3OHLys FMOC derivative are
in accordance with previous literature data for 3OHLys
where the 13C resonances were obtained in D2O (23).
F or m a tion of Lysin e Hyd r oxid es, Hyd r op er ox-
id es, a n d Ald eh yd es in γ-Ir r a d ia ted Lysin e Solu -
tion s. 60Co radiolysis (1000 Gy) of L-lysine (4 mM) in
the presence of O2 gave ca. 80 µΜ lysine hydroperoxides
(i.e., ca. 2000 hydroperoxides/105 lysines, mean of three
determinations) as determined by iodometric assay (16,
24). No hydroperoxides were detected after reduction by
NaBH4. The lysine hydroperoxides present in the γ-ra-
diolysed lysine solution were not detected with HPLC
either after FMOC derivatization and reversed-phase
HPLC or by injection of the nonderivatized solution onto
an amino-phase HPLC column coupled with chemilumi-
nescence detection. The hydroperoxides are presumably
degraded during FMOC derivatization or HPLC analysis.
(1) P r od u ction of Lysin e Hyd r oxid es by Hyd r oxyl
Ra d ica ls. FMOC measurement of the γ-radiolysed
lysine solution gave a total yield of 2000 lysine hydroxides/
105 lysines, and the γ-radiolysed and reduced lysine
solution gave a yield of 4700 lysine hydroxides/105 lysines
[5(R)OHLys, 2545/105 Lys; 5(S)OHLys, 774/105 Lys;
4OHLys, 1149/105 Lys; and 3OHLys, 336/105 Lys]. This
means that 2700 lysine hydroxides/105 lysines were
formed by NaBH4 reduction. Almost half of the lysine
hydroxides were therefore generated during γ-radiolysis
as discussed below.
(2) P r od u ction of Lysin e Ald eh yd es d u r in g γ-Ra -
d iolysis. The high yield of lysine hydroxides formed by
NaBH4 reduction (2700/105 lysines) which is greater than
the total content of lysine hydroperoxides (2000/105
lysines, see above) may indicate that other oxidation
Since 5-hydroxylysine is a natural amino acid (and
hence probably not a useful marker for oxidative dam-

1270 Chem. Res. Toxicol., Vol. 11, No. 11, 1998
Ta ble 1. 1H Ch em ica l Sh ifts for F MOC-Lysin e Der iva tivesa
Morin et al.
L-lysine
5(R)OHLys
3OHLys
lysine moiety
H(2) (R)
H(3a) (â)
H(3b) (â)
H(4a) (γ)
H(4b) (γ)
H(5a) (δ)
H(5b) (δ)
2H(6) (ꢀ)
NH(R)
3.84
1.59
1.70
1.30
1.30
1.39
1.39
2.96
7.37
7.26
-
4.20
3.90
1.55
1.91
1.25
1.51
3.44
-
2.96
7.59
7.22
4.68
4.20
3.60
3.58
-
1.01
1.31
1.32
1.60
2.95
6.23
NH(ꢀ)
7.25
OH
CH2
ndb
FMOC moiety
4.20
CH
4.27
4.26
4.25
aromatic ring
7.31, 7.40,
7.67, 7.88
7.32, 7.40,
7.70, 7.88
7.32, 7.39,
7.66, 7.87
a
Reference CD3SOCD2H of δ 2.50. The NH chemical shifts were obtained from one-dimensional spectra. The 1H data represent the
b
midpoints of multiplets rather than true chemical shifts. Not determined.
Ta ble 2. 13C Ch em ica l Sh ifts for F MOC-Lysin e Der iva tivesa
L-lysine
5(R)OHLys
3OHLys
lysine moiety
FMOC moiety
C(2) (R)
C(3) (â)
C(4) (γ)
C(5) (δ)
C(6) (ꢀ)
CH2
53.9
30.6
22.5
28.8
39.8
65.0
53.9
27.1
30.9
68.7
46.5
65.2
57.1
69.8
29.0
26.5
40.6
65.1
CH
46.4
46.4
46.5
aromatic ring
119.8, 124.9,
126.7, 127.3
119.8, 124.9,
126.7, 127.3
119.8, 124.9,
126.9, 127.3
a
Reference CD3SOCD3 of δ 39.5. The 13C chemical shifts were obtained from HMQC spectra.
correspond to 200/105 lysines. This value does not
account for all the non-hydroperoxide precursors of the
hydroxides; other sources have not been investigated
further.
Detection of 3-Hyd r oxylysin e on Oxid ized P ep -
tid es a n d P r otein s. 60Co radiolysis of Gly-Lys-Gly (2
mM), Lys-Val-Ile-Leu-Phe (1 mM), BSA (5 mg/mL), and
calf thymus histone H1 (4 mg/mL) was carried out with
various doses in the presence of O2. Significant amounts
of peptide-bound hydroperoxides and protein-bound hy-
droperoxides were measured by iodometric assay. For
instance, BSA or histone H1 exposed to 2000 Gy of
radiation gave 158 or 335 µΜ protein-bound hydroper-
oxides, respectively. After reaction with NaBH4, which
removed all the hydroperoxides, the samples were acid-
hydrolyzed and analyzed using both the OPA- and
FMOC-HPLC method.
The yields of 3-hydroxylysine in γ-irradiated (1500 Gy)
and NaBH4-reduced Gly-Lys-Gly and Lys-Val-Ile-Leu-
Phe were determined to be, respectively, 2170/105 Lys
and 230/105 Lys after taking into account their recovery.
The extent of formation of 3-hydroxyvaline and 4-hy-
droxyleucines was also determined in γ-radiolysed and
reduced Lys-Val-Ile-Leu-Phe (Table 3).
3-Hydroxylysine formation during γ-radiolysis of his-
tone H1 was readily detected (Figure 5). An irradiation
dose of 2000 Gy yielded a production of 269 3OHLys/105
Lys (Table 3). Species co-migrating with the other hy-
droxylysines were also formed (Figure 5), but were not
investigated further for reasons already given.
Generation of 3OHLys during irradiation of BSA
(Figure 6) showed a dose-dependent response increase
from 500 to 2000 Gy (the yield at the lowest dose was
below the detection limit). About 0.1% of the total lysine
present in 5 mg/mL BSA was converted to 3OHLys. The
formation of 3OHLys in γ-radiolysed BSA and histone
F igu r e 4. Double-quantum filtered COSY spectrum of the
3-hydroxylysine derivative. The spectrum was processed with
zero-filling in F1 and a π/2 shifted sine-bell function in each
dimension, and ridges of t1 noise due to the solvent and water
resonances diminished by subtraction of a partial projection.
Negative contours are shown with lighter shading; correlations
from each of the NH protons are indicated by dashed lines.
Artifacts due to quadrature images of the solvent resonances
are enclosed in triangles. For clarification of chemical shifts not
readily distinguished in the COSY spectrum, the relevant
section of a HMQC spectrum is shown in the inset.
products are also produced during γ-radiolysis. Carbonyl
formation during γ-radiolysis has been detected previ-
ously (4, 20, 25, 26). To investigate this, we examined
the reactivity of these non-hydroperoxide precursors with
2,4-dinitrophenylhydrazine, a compound which deriva-
tizes carbonyl groups such as aldehydes (27). When a
γ-radiolysed lysine solution was reacted with DNPH
before reduction by NaBH4, the level of hydroxylysines
was slightly lower than that from direct NaBH4 reduc-
tion. By this analysis, the amount of lysine aldehydes

Hydroxylysines as Markers of Protein Oxidation
Chem. Res. Toxicol., Vol. 11, No. 11, 1998 1271
Ta ble 3. F or m a tion of Oxid ized P r od u cts in γ-Ir r a d ia ted P ep tid e, H1, a n d BSAa
dose (Gy)
3OHLys/105 Lys
ValOH1b/105 Val
LeuOH2b/105 Leu
LeuOH3b/105 Leu
Lys-Val-Iso-Leu-Phe
H1
BSA
1500
2000
500
1000
1500
2000
232 ( 44
269 ( 19
ndc
37 ( 1
80 ( 5
90 ( 5
220 ( 52
112 ( 15
15 ( 1
1421 ( 486
499 ( 103
34 ( 7
1086 ( 47
326 ( 86
31 ( 8
23 ( 6
64 ( 4
46 ( 1
38 ( 4
52 ( 3
150 ( 25
223 ( 11
101 ( 19
154 ( 29
a
b
Data represent mean ( SD of three separate experiments conducted as described in Materials and Methods. Structures of ValOH1,
LeuOH2, and LeuOH3 are given in Figure 2. c Value below the HPLC detection limit.
Ta ble 4. Ra tio of 3OHLys to Hyd r oxyla ted Va lin e
(Va lOH1a ) a n d Leu cin e (Leu OH2a or Leu OH3a )
Der iva tives in γ-Ir r a d ia ted P ep tid e, H1, a n d BSA
Lys-Val-Iso-Leu-Phe
at 1500 Gy
H1
BSA
at 2000 Gy at 1500 Gy
3OHLys/ValOH1
3OHLys/LeuOH2
3OHLys/LeuOH3
ValOH1/LeuOH2
ValOH1/LeuOH3
1.05
0.16
0.21
0.15
0.20
2.42
0.54
0.82
0.22
0.34
2.09
0.53
0.79
0.21
0.37
F igu r e 5. HPLC detection of the FMOC derivative of 3-hy-
droxylysine in oxidized lysine and histone H1 solutions. (a)
Native and (b) γ-irradiated histone H1 (4 mg/mL) were reduced
by NaBH4, hydrolyzed, and derivatized using FMOC (see
Materials and Methods). (c) A γ-irradiated lysine solution (4
mM) was treated with NaBH4 and subsequently derivatized
with FMOC (cf. Figure 2).
a
Structures of ValOH1, LeuOH2, and LeuOH3 are given in
Figure 2.
for the generation of lysine hydroperoxides similar to that
proposed previously for formation of valine and leucine
hydroperoxides is shown in Figure 7. The initial step
involves hydrogen abstraction by hydroxyl radicals to
generate a carbon-centered radical and subsequent for-
mation of a peroxyl radical species via O2 addition. The
formation of lysine hydroxides during γ-radiolysis can be
explained by several pathways which have been de-
scribed in some detail (7, 8). The first route is the
reduction of the preformed lysine hydroperoxides by
electrons generated during water radiolysis (8). Another
possibility is via a transient tetroxide formed by dimer-
ization of two peroxyl radicals. This species can decom-
pose to form oxygen and alkoxyl radicals (which can be
expected to undergo hydrogen atom abstraction reaction
with suitable R-H bonds) or decompose directly to form
oxygen, an aldehyde, and a hydroxide (the Russell
mechanism) (28, 29).
The predominant formation of 5OHLys (3300/105 Lys)
compared to those of 4OHLys (1100/105 Lys) and 3OHLys
(330/105Lys) is consistent with EPR spin trapping studies
which have demonstrated that hydrogen abstraction
occurs mainly at C(4) and C(5) on the side chain when
lysine is treated with HO• (30). As 5OHLys is a naturally
occuring amino acid (present in collagen as a result of
lysine oxidation by the enzyme lysyl oxidase) and 4OHLys
is not fully recovered after protein hydrolysis treatment,
these species are probably not useful markers for protein
oxidation. However, 3OHLys which is stable during our
procedure and apparently a specific product of oxidative
conditions may be suitable as a marker for studying
protein oxidation. Hydroxylation at C(2) and C(6) is not
observed as protonated R-amino groups in free amino
acids deactivate the carbon nearby with respect to
hydrogen atom abstraction by HO• (29, 31, 32).
F igu r e 6. Detection of 3-hydroxylysine, 3-hydroxyvaline, and
5-hydroxyleucines on γ-radiolysed BSA. BSA (5 mg/mL in water)
was exposed to the stated doses of radiation in the presence of
O2. After treatment with NaBH4, the samples were hydrolyzed,
derivatized by FMOC (for 3-hydroxylysine) or OPA (for 3-hy-
droxyvaline and 5-hydroxyleucines), and analyzed by HPLC as
described in Materials and Methods. Results are means ( SD
from two different experiments.
H1 was confirmed by spiking experiments using a
mixture of the four lysine hydroxides.
Discu ssion
On the basis of chemical and spectroscopic evidence,
we have identified the formation of four lysine hydroxides
generated during γ-radiolysis of lysine in the presence
of O2. 3-Hydroxylysine derivatized by FMOC has been
fully characterized by detailed structural studies using
1
electrospray MS and H and 13C NMR.
The γ-radiolysed lysine solution, before reduction with
NaBH4, contains a mixture of lysine hydroperoxides (as
inferred from iodometric assay), lysine hydroxides (as
determined by HPLC), and some carbonyl species (as
shown by DNPH derivatization). Carbonyl groups have
been previously observed in irradiated valine and leucine
solutions, although in the latter case the carbonyls
undergo further rearrangement to generate cyclic prod-
ucts via Schiff base formation (9). A possible mechanism
This study was undertaken in the light of the hypoth-
esis that lysine residues on proteins might be especially
susceptible to HO• attack because of their presence on
the surface of globular proteins. We have therefore
compared the formation of lysine hydroxides to that of
valine and leucine hydroxides in γ-radiolysed peptides
and proteins BSA and histone H1 (Table 3); the latter in
particular has a very high lysine content. The extent of

1272 Chem. Res. Toxicol., Vol. 11, No. 11, 1998
Morin et al.
F igu r e 7. Proposed reaction mechanisms for the generation of lysine hydroperoxides and hydroxides.
hydroperoxide formation on free lysine (70 µΜ in a 1 mM
lysine solution) was lower compared to that of valine (112
µΜ in a 1 mM valine solution) and leucine (88 µΜ in a 1
mM leucine solution) (6). However, we have demon-
strated the formation of high yields of lysine hydroxides
during the irradiation. In contrast, γ-radiolysis of valine
and leucine did not directly produce hydroxide com-
pounds.
Table 4 shows an analysis of our data in terms of ratios
between 3OHLys and the leucine and valine hydroxides,
and between the leucine and valine hydoxides them-
selves. The indices for the number of 3OHLys per leucine
or valine derivative are approximately 2-4-fold higher
in the two proteins (histone H1 and BSA) than in the
peptide Lys-Val-Ile-Leu-Phe. In contrast, the indices for
ValOH1/LeuOH2 and ValOH1/LeuOH3 are similar in all
three cases. Thus, it seems that protein-bound lysine is
preferred as target for radical attack compared to peptide
lysine, whereas this is not true for Val (or Leu). This
can be most readily explained as a consequence of the
preferential location of lysine residues at protein surfaces,
and the contrasting more internal disposition of the
valine and leucine residues.
We conclude that 3-hydroxylysine may be a useful
marker of radical-mediated protein oxidation and, in
particular, that it may help in the further analysis of
radical reactions localized on the surfaces of proteins.
Refer en ces
(1) Gebicki, S., and Gebicki, J . M. (1993) Formation of peroxides in
amino acids and proteins exposed to oxygen free radicals. Bio-
chem. J . 289, 743-749.

Hydroxylysines as Markers of Protein Oxidation
Chem. Res. Toxicol., Vol. 11, No. 11, 1998 1273
(2) Simpson, J . A., Narita, S., Gieseg, S., Gebicki, S., Gebicki, J . M.,
and Dean, R. T. (1992) Long-lived reactive species on free-radical-
damaged proteins. Biochem. J . 282, 621-624.
(3) Stadtman, E. R. (1992) Protein oxidation and aging. Science 257,
1220-1224.
(18) Lewis, J . R. E., Morley, J . S., and Venn, R. F. (1993) Analysis of
human â-endorphin 28-31 (melanotropin potentiating factor) and
analogues by high-performance liquid chromatography of their
9-fluorenylmethoxy-carbonyl derivatives. J . Chromatogr. 615, 37-
45.
(4) Oliver, C. N., Ahn, B. W., Moerman, E. J ., Goldstein, S., and
Stadtman, E. R. (1987) Age-related changes in oxidized proteins.
J . Biol. Chem. 262, 5488-5491.
(5) Garrison, W. M. (1987) Reaction mechanisms in the radiolysis of
peptides, polypeptides, and proteins. Chem. Rev. 87, 381-398.
(6) Fu, S., Hick, L. A., Sheil, M. M., and Dean, R. T. (1995) Structural
identification of valine hydroperoxides and hydroxides on radical-
damaged amino acid, peptide, and protein molecules. Free Radical
Biol. Med. 19, 281-292.
(19) Chen, R. F., Scott, C., and Trepman, E. (1979) Fluorescence
properties of o-phthaldialdehyde derivatives of amino acids.
Biochim. Biophys. Acta 576, 440-455.
(20) Fu, S., Gebicki, S., J essup, W., Gebicki, J . M., and Dean, R. T.
(1995) Biological fate of amino acid, peptide and protein hydro-
peroxides. Biochem. J . 311, 821-827.
(21) J ackman, L. M., and Sternhell, S. (1969) Applications of Nuclear
Magnetic Resonance Spectroscopy in Organic Chemistry, pp
1-163, Pergamon Press, Oxford, U.K.
(7) Davies, M. J ., and Dean, R. T. (1997) Radical-mediated protein
oxidation: from chemistry to medicine, pp 1-443, Oxford Uni-
versity Press, Oxford, U.K.
(8) Dean, R. T., Fu, S., Stocker, R., and Davies, M. J . (1997) The
biochemistry and pathology of radical-mediated protein oxidation.
Biochem. J . 324, 1-18.
(9) Fu, S., and Dean, R. T. (1997) Structural characterisation of the
products of hydroxyl radical damage to leucine and their detection
on proteins. Biochem. J . 324, 41-48.
(10) Fu, S., Fu, M. X., Baynes, J . W., Thorpe, S. R., and Dean, R. T.
(1998) Presence of DOPA and amino acid hydroperoxides in
proteins modified with advanced glycation end products (AGEs):
amino acid oxidation products as a possible source of oxidative
stress induced by AGE proteins. Biochem. J . 330, 233-239.
(11) Fu, S., Davies, M. J ., Stocker, R., and Dean, R. T. (1998) Evidence
for roles of radicals in protein oxidation in advanced human
atherosclerotic plaque. Biochem. J . 333, 519-525.
(22) Kalinowski, H. O., Berger, S., and Braun, S. (1994) Carbon-13
NMR Spectroscopy, pp 1-179, Wiley, Chichester, U.K.
(23) Hughes, P. F., Smith, S. H., and Olson, J . T. (1994) Two chiral
syntheses of threo-3-Hydroxylysine. J . Org. Chem. 59, 5799-5802.
(24) Gebicki, J . M., and Guille, J . (1989) Spectrophotometric and high
performance chromatographic assays of hydroperoxides by iodo-
metric technique. Anal. Biochem. 176, 360-364.
(25) Stadtman, E. R. (1990) Covalent modification reactions are
marking steps in protein turnover. Biochemistry 29, 6323-6331.
(26) Stadtman, E. R. (1993) Oxidation of free amino acids and amino
acid residues in proteins by radiolysis and by metal-catalyzed
reactions. Annu. Rev. Biochem. 62, 797-821.
(27) Levine, R. L. (1984) Mixed-function oxidation of histidine residues.
Methods Enzymol. 107, 370-376.
(28) Russell, G. A. (1957) Deuterium-isotope effects in the autoxidation
of aralkyl hydrocarbons. Mechanism of the interaction of peroxy
radicals. J . Am. Chem. Soc. 79, 3871-3877.
(12) Ayala, A., and Cutler, R. G. (1996) The utilization of 5-hydroxyl-
2-amino valeric acid as a specific marker of oxidised arginine and
proline residues in proteins. Free Radical Biol. Med. 21, 65-80.
(13) Davies, M. J ., Fu, S., and Dean, R. T. (1995) Protein hydroper-
oxides can give rise to reactive free radicals. Biochem. J . 305,
643-649.
(14) Davies, M. J . (1996) Protein and peptide alkoxyl radicals can give
rise to C-terminal decarboxylation and backbone cleavage. Arch.
Biochem. Biophys. 336, 163-172.
(15) Trelstad, R., Lawley, L. R., and Holmes, L. B. (1981) Nonenzy-
matic hydroxylations of proline and lysine by reduced oxygen
derivatives. Nature 289, 310-312.
(16) J essup, W., Dean, R. T., and Gebicki, J . M. (1994) Iodometric
determination of hydroperoxides in lipids and proteins. Methods
Enzymol. 233, 289-303.
(29) von Sonntag, C. (1987) The chemical basis of radiation biology,
pp 1-515, Taylor and Francis, London.
(30) Hawkins, C. L., and Davies, M. J . (1997) Oxidative damage to
collagen and related substrates by metal ion/hydrogen peroxide
systems: random attack or site-specific damage? Biochim. Bio-
phys. Acta 1360, 84-96.
(31) Simic, M., Neta, P., and Hayon, E. (1970) Selectivity in the
reactions of e-aq and OH radicals with simple peptides in aqueous
solution. Optical absorption spectra of intermediates. J . Am.
Chem. Soc. 92, 4763-4768.
(32) Rao, P. S., and Hayon, E. (1975) Reactions of hydroxyl radicals
with oligopeptides in aqueous solutions. J . Phys. Chem. 79, 109-
115.
(33) Buxton, G. V., Greenstock, C. L., Helman, W. P., and Ross, A. B.
(1988) Critical review of rate constants for reactions of hydrated
electrons, hydrogen atoms and hydroxyl radicals (.OH/.O-) in
aqueous solution. J . Phys. Chem. Ref. Data 17, 513-886.
(17) Einarsson, S., J osefsson, B., and Lagerkvist, S. (1983) Determi-
nation of amino acids with 9-fluorenylmethylchloroformate and
reverse-phase high-performance liquid chromatography. J . Chro-
matogr. 282, 609-618.
TX980118H