Δ12-Prostaglandin J2 as a product and ligand of human serum albumin: Formation of an unusual covalent adduct at His146

Article abstract of DOI:10.1021/ja908878n

Human serum albumin (HSA), the most abundant protein in plasma, has a very unique function, catalyzing the conversion of prostaglandin J₂ (PGJ₂), a dehydration product of PGD², to yield Δ¹²-PGJ₂. These PGD₂ metabolites are actively transported into cells and accumulated in the nuclei, where they act as potent inducers of cell growth inhibition and cell differentiation, and exhibit their own unique spectrum of biological effects. The facts that (i) arachidonic acid metabolites bind to human serum albumin (HSA) and the metabolism of these molecules is altered as a result of binding, (ii) HSA catalyzes the transformation of PGJ2 into Δ¹²-PGJ₂, and (iii) Δ¹²-PGJ₂ is stable in serum suggest that HSA may bind and stabilize Δ¹²-PGJ₂ in a specific manner. A molecular interaction analysis using surface plasmon resonance (Biacore) indeed suggested the presence of a specific Δ¹²-PGJ ₂-binding site in HSA. To investigate the molecular details of the binding of this PGD₂ metabolite to albumin, we analyzed the cocrystal structure of the HSA-Δ¹²-PGJ₂-myristate complex by X-ray crystallography and found that two Δ¹²-PGJ¹² molecules bind to a primary site in subdomain IB of the protein. The electron density results suggested that one of the two Δ¹²-PGJ ¹² molecules that specifically bind to the site covalently interacted with a histidine residue (His146). Using nano-LC-MS/MS analysis of the HSA-Δ¹²-PGJ₂ complex, the formation of an unusual Δ¹²-PGJ₂-histidine adduct at His146 was confirmed. Thus, our crystallographic and mass spectrometric analyses of the HSA-Δ¹²-PGJ₂ complex provided intriguing new insights into the molecular details of how this electrophilic ligand interacts with its primary producer and transporter.

Full text of DOI:10.1021/ja908878n

Published on Web 12/16/2009
∆12-Prostaglandin J2 as a Product and Ligand of Human
Serum Albumin: Formation of an Unusual Covalent Adduct at
His146
Satoru Yamaguchi,† Giancarlo Aldini,‡ Sohei Ito,§ Nozomi Morishita,†
Takahiro Shibata,† Giulio Vistoli,‡ Marina Carini,‡ and Koji Uchida*,†
Graduate School of Bioagricultural Sciences, Nagoya UniVersity, Nagoya 464-8601, Japan,
Dipartimento di Scienze Farmaceutiche “Pietro Pratesi”, UniVersity of Milan,
I-20131, Milan, Italy, and School of Food and Nutritional Sciences, UniVersity of Shizuoka,
Shizuoka 422-8526, Japan
Received October 17, 2009; E-mail: uchidak@agr.nagoya-u.ac.jp
Abstract: Human serum albumin (HSA), the most abundant protein in plasma, has a very unique function,
catalyzing the conversion of prostaglandin J2 (PGJ2), a dehydration product of PGD2, to yield ∆12-PGJ2.
These PGD2 metabolites are actively transported into cells and accumulated in the nuclei, where they act
as potent inducers of cell growth inhibition and cell differentiation, and exhibit their own unique spectrum
of biological effects. The facts that (i) arachidonic acid metabolites bind to human serum albumin (HSA)
and the metabolism of these molecules is altered as a result of binding, (ii) HSA catalyzes the transformation
of PGJ2 into ∆12-PGJ2, and (iii) ∆12-PGJ2 is stable in serum suggest that HSA may bind and stabilize ∆12-
PGJ2 in a specific manner. A molecular interaction analysis using surface plasmon resonance (Biacore)
indeed suggested the presence of a specific ∆12-PGJ2-binding site in HSA. To investigate the molecular
details of the binding of this PGD2 metabolite to albumin, we analyzed the cocrystal structure of the HSA-∆12-
PGJ2-myristate complex by X-ray crystallography and found that two ∆12-PGJ2 molecules bind to a primary
site in subdomain IB of the protein. The electron density results suggested that one of the two ∆12-PGJ2
molecules that specifically bind to the site covalently interacted with a histidine residue (His146). Using
nano-LC-MS/MS analysis of the HSA-∆12-PGJ2 complex, the formation of an unusual ∆12-PGJ2-histidine
adduct at His146 was confirmed. Thus, our crystallographic and mass spectrometric analyses of the
HSA-∆12-PGJ2 complex provided intriguing new insights into the molecular details of how this electrophilic
ligand interacts with its primary producer and transporter.
cells, mast cells, and megakaryocytes.4 The PGs are physiologi-
cally present in body fluids in picomolar-to-nanomolar concen-
Introduction
trations;5 however, the arachidonate metabolism is significantly
increased under several pathological conditions including hy-
perthermia, infection, and inflammation,6 and local PG concen-
trations in the micromolar range have been detected at sites of
acute inflammation.7 The two major enzymatic pathways
responsible for the catabolism of PGD2 are an 11-ketoreductase
and NADP-linked 15-hydroxy-PG-dehydrogenase.3,8 The former
pathway leads to the production of 11ꢀ-PGF2a. The second
pathway leads to the formation of 13,14-dihydro-15-keto-PGD2,
which does not appear to be biologically active.
The prostaglandins (PGs) are a family of structurally related
molecules that are produced by cells in response to a variety of
extrinsic stimuli and regulate cellular growth, differentiation,
and homeostasis.1,2 Among them, PGD2 is a major cyclooxy-
genase product in a variety of tissues and cells and has marked
effects on a number of biological processes, including platelet
aggregation, relaxation of vascular and nonvascular smooth
muscles, and nerve cell functions.3 PGD synthase catalyzes the
isomerization of the 9,10-endoperoxide group of PGH2, a
precursor of prostanoids, to produce PGD2. Two distinct types
of PGD synthases have been identified: the lipocalin enzyme
(L-PGDS) and hematopoietic enzyme (H-PGDS). Although
L-PGDS is expressed in the central nervous system and male
genital organs of various mammals, H-PGDS is widely ex-
pressed in peripheral tissues, as well as in antigen-presenting
PGD2 is known to be relatively unstable and readily undergoes
dehydration in vivo and in vitro to yield biologically active PGs
(4) Urade, Y.; Eguchi, N. Prostaglandins Other Lipid Mediators 2002,
68-69, 375–382.
(5) Fukushima, M. Eicosanoids 1990, 3, 189–199.
(6) Herschman, H. R.; Reddy, S. T.; Xie, W. AdV. Exp. Med. Biol. 1997,
407, 61–66.
† Nagoya University.
‡ University of Milan.
§ University of Shizuoka.
(7) Offenbacher, S.; Odle, B. M.; Van Dyke, T. E. J. Periodontal Res.
1986, 21, 101–112.
(1) Smith, W. L. Biochem. J. 1989, 259, 315–324.
(2) Smith, W. L. Am. J. Physiol. 1992, 263, F181–191.
(3) Giles, H.; Leff, P. Prostaglandins 1988, 35, 277–300.
(8) Urade, Y.; Watanabe, K.; Eguchi, N.; Fujii, Y.; Hayaishi, O. J. Biol.
Chem. 1990, 265, 12029–12035.
9
824 J. AM. CHEM. SOC. 2010, 132, 824–832
10.1021/ja908878n  2010 American Chemical Society

∆12-PGJ2 as a Product and Ligand of HSA
A R T I C L E S
of the J2 series, such as PGJ2, ∆12-PGJ2, and 15d-PGJ2.9,10 These
cyclopentenone-type PGs (cyPGs) are actively transported into
cells and accumulated in the nuclei, where they act as potent
inducers of cell growth inhibition and cell differentiation, and
exhibit their own unique spectrum of biological effects, includ-
ing the inhibition of cell cycle progression, the suppression of
viral replication, the induction of the heat shock protein
expression, and the stimulation of osteogenesis.11 In addition,
the physiopathological relevance of the enhanced synthesis of
cyPGs in activated macrophages has been speculated to represent
a mechanism in which they function as a feedback regulator of
the inflammatory responses.12-14
The general instability of the PGs in aqueous media has made
it difficult to unravel the many biological roles played by these
highly active signaling molecules. It became apparent early on
in prostaglandin research that proteins in the blood might play
an important role in modulating the biological activities of these
compounds by binding to and stabilizing or destabilizing certain
PGs. A series of binding studies using radiolabeled PGE1, PGE2,
PGA2, and PGF2 found that the only plasma protein that
significantly binds to the above prostaglandins is human serum
albumin (HSA).15 HSA is the most abundant protein in plasma,
where it acts as a transporter of an exceptionally broad spectrum
of compounds that are predominantly fatty acids, but also amino
acids, bile acids, and steroids.16 It is also capable of binding
and transporting a wide range of therapeutic substances. Its
binding abilities have been probed in a number of studies, and
crystal structures are available for HSA in complexes with fatty
acids,17,18 hemin,19 and local anesthetics.20 Although the affinity
of HSA for a variety of biologically active arachidonic acid
metabolites is relatively low,21,22 the high serum HSA concen-
tration makes these interactions physiologically significant. HSA
has been shown to stabilize PGI2,23 an unstable but potent
inhibitor of platelet aggregation derived from PGH2. However,
HSA stabilizes the potent stimulant of irreversible platelet
aggregation TXA2,24 enhancing its activity. In addition, HSA
binds LTA4,25 the unstable precursor of most leukotrienes,
Figure 1. HSA-dependent PGD2 metabolic pathway.
preventing its rapid nonenzymatic degradation to biologically
inactive metabolites in aqueous media.
On the other hand, HSA is known to be an endogenous
catalyst of the PGD2 metabolism. HSA catalyzes the conversion
of PGH2 to PGD2.26 In addition, in aqueous solutions containing
HSA, PGJ2 isomerizes to yield ∆12-PGJ2 (Figure 1).27 Hirata et
al. have previously shown that the injection of monkeys with a
single dose of PGD2 indeed results in a dramatic increase in
the output of urinary ∆12-PGJ2, whereas treatment with a
cyclooxygenase inhibitor results in a significant decrease in
urinary levels of ∆12-PGJ2.28 They have also shown that ∆12-
PGJ2 is present in significant quantities in human urine. These
data provided evidence that ∆12-PGJ2 is a natural PGD2
metabolite that is formed in vivo and excreted in the urine. In
a later study, ∆12-PGJ2 has been identified as substance produced
by murine bone marrow-derived mast cells.29 These findings
and the fact that ∆12-PGJ2 is stable in serum suggest that ∆12-
PGJ2 may be produced and actively carried by HSA in the
bloodstream and transported into cells to exhibit its biological
effects. In this Article, we describe the three-dimensional
structure of HSA complexed with ∆12-PGJ2 and show that,
despite the repeating domain structure of the protein, two
molecules of ∆12-PGJ2 uniquely bind at a fatty acid binding
site. We also show that one of the two ∆12-PGJ2 molecules
covalently interacts with a histidine residue to form an unusual
∆12-PGJ2-histidine adduct. These data suggest an explanation
for observations that ∆12-PGJ2 and related PGs show a unique
spectrum of biological effects.
(9) Fitzpatrick, F. A.; Wynalda, M. A. J. Biol. Chem. 1983, 258, 11713–
11718.
(10) Kikawa, Y.; Narumiya, S.; Fukushima, M.; Wakatsuka, H.; Hayaishi,
O. Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 1317–1321.
(11) Fukushima, M. Prostaglandins, Leukotrienes Essent. Fatty Acids 1992,
47, 1–12.
(12) Gilroy, D. W.; Colville-Nash, P. R.; Willis, D.; Chivers, J.; Paul-
Clark, M. J.; Willoughby, D. A. Nat. Med. 1999, 5, 698–701.
(13) Cuzzocrea, S.; Wayman, N. S.; Mazzon, E.; Dugo, L.; Di Paola, R.;
Serraino, I.; Britti, D.; Chatterjee, P. K.; Caputi, A. P.; Thiemermann,
C. Mol. Pharmacol. 2002, 61, 997–1007.
(14) Itoh, K.; Mochizuki, M.; Ishii, Y.; Ishii, T.; Shibata, T.; Kawamoto,
Y.; Kelly, V.; Sekizawa, K.; Uchida, K.; Yamamoto, M. Mol. Cell.
Biol. 1972, 24, 36–45.
(15) Raz, A. Biochem. J. 1972, 130, 631–636.
(16) Peters T. All About Albumin: Biochemistry, Genetics and Medical
Applications; Academic Press: San Diego, CA, 1995.
(17) Curry, S.; Mandelkow, H.; Brick, P.; Franks, N. Nat. Struct. Biol.
1998, 5, 827–835.
Results
Molecular Interaction between HSA and ∆12-PGJ2. We first
checked the PGJ2-isomerization activity of the commercially
obtained HSA (fatty acid and globulin free, purity 99%) used
(18) Bhattacharya, A. A.; Gru¨ne, T.; Curry, S. J. Mol. Biol. 2000, 303,
721–732.
(19) Wardell, M.; Wang, Z.; Ho, J. X.; Robert, J.; Ruker, F.; Ruble, J.;
Carter, D. C. Biochem. Biophys. Res. Commun. 2002, 291, 813–819.
(20) Zunszain, P. A.; Ghuman, J.; Komatsu, T.; Tsuchida, E.; Curry, S.
BMC Struct. Biol. 2003, 3, 6.
(26) Watanabe, T.; Narumiya, S.; Shimizu, T.; Hayaishi, O. J. Biol. Chem.
1982, 257, 14847–14853.
(21) Unger, W. G. B. J. Pharm. Pharm. Sci. 1972, 24, 470–477.
(22) Gueriguian, J. L. J. Pharmacol. Exp. Ther. 1976, 197, 391–401.
(23) Wynalda, N. A.; Fitzpatrick, F. A. Prostaglandins Other Lipid
Mediators 1980, 20, 853–861.
(27) Shibata, T.; Kondo, M.; Osawa, T.; Shibata, N.; Kobayashi, M.;
Uchida, K. J. Biol. Chem. 2002, 277, 10459–10466.
(28) Hirata, Y.; Hayashi, H.; Ito, S.; Kikawa, Y.; Ishibashi, M.; Sudo, M.;
Miyazaki, H.; Fukushima, M.; Narumiya, S.; Hayaishi, O. J. Biol.
Chem. 1988, 263, 16619–16625.
(24) Folco, G.; Granstrom, E.; Kindahl, H. FEBS Lett. 1977, 82, 321–324.
(25) Fitzpatrick, F. A.; Morton, D. R.; Wynalda, M. A. J. Biol. Chem. 1981,
257, 4680–4683.
(29) Haberl, C.; Hultner, L.; Flugel, A.; Falk, M.; Geuenich, S.; Wilmanns,
W.; Denzlinger, C. Mediators Inflammation 1988, 7, 79–84.
9
J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010 825

A R T I C L E S
Yamaguchi et al.
Figure 3. Surface plasmon resonance (Biacore) detection of the interaction
between HSA and ∆12-PGJ2. (A) Responses from injections at 3.9, 7.8,
15.6, 31.3, 62.5, 125, 250, and 500 µM. Sensorgram are double-referenced
and solvent-corrected. (B) Equilibrium binding responses plotted versus ∆12-
PGJ2 concentration and fitted to a simple binding isotherm to yield an affinity
of 420 µM.
Figure 2. Conversion of PGJ2 by ∆12-PGJ2. (A) HPLC profile of PGJ2
metabolites. PGJ2 (1 mM) was incubated in PBS containing HSA (10 mg/
mL) for 24 h. PGJ2 and its metabolites were separated on a ChiralPak AD-
RH column (0.46 × 15 cm) eluted with a linear gradient of acetonitrile/
water/acetic acid (90/10/0.01, by vol.) (solvent A) - acetonitrile (solvent
B) (time ) 0-5 min, 100% A; 60 min, 0% A), at a flow rate of 0.8 mL/
min. The elution profiles were monitored by UV absorbance at 244 nm.
(B) Dose-dependent conversion of PGJ2 to ∆12-PGJ2 by HSA. PGJ2 was
incubated in PBS containing HSA (0-10 mg/mL) for 24 h. Symbols: 4,
PGJ2; 9, ∆12-PGJ2; b, 15d-PGJ2.
The ∆12-PGJ2-HSA crystal, diffracted up to a 2.5 Å resolution,
showed a positive electron density that protruded from the side-
chain of His146. However, the ∆12-PGJ2 binding configurations
were unclear because of the poor electron densities for the
suspected ∆12-PGJ2 molecules. An improved procedure involv-
ing the addition of ∆12-PGJ2 to the harvesting solution and
cocrystallization with myristate revealed that two ∆12-PGJ2
molecules (∆12-PGJ2-1 and ∆12-PGJ2-2) exist in the subdomain
IB (Figure 4).
in this study. PGJ2 was incubated in PBS containing HSA, and
the products were analyzed by chiral-phase HPLC. It was
observed that PGJ2 was converted to both 15d-PGJ2 and ∆12-
PGJ2 in the presence of HSA (Figure 2A). The isomerization
of PGJ2 to ∆12-PGJ2 was dependent on the incubation time and
the concentration of HSA; when HSA was present in excess,
PGJ2 was almost stoichiometrically converted to ∆12-PGJ2
(Figure 2B).
The omit maps and 2Fo - Fc maps at a 2.2 Å resolution also
showed clear densities for the two ∆12-PGJ2 molecules (Figure
5A and Supporting Information Figure S1). A break in the
electron density around the cyclopentenone ring of ∆12-PGJ2-
1, especially in chain A, suggested a competition against the
myristate. However, the shape of the density, which had a
significant broadening at one end, clearly indicated the position
of the carboxyl moiety and thus defined the orientation of the
bound ∆12-PGJ2 molecules. Figure 5B shows that the cyclo-
pentenone ring and both tails of ∆12-PGJ2-1 are rigidly anchored
by HSA. The ∆12-PGJ2-1 binding site has a large interaction
surface area up to 447 Å2, allowing shape-adaptability with ∆12-
PGJ2 through a specific hydrophobic interaction. It also appears
that ∆12-PGJ2-1 binds at this site in the same orientation with
other saturated medium- and long-chain fatty acids.18 The
calculated interaction surface area of ∆12-PGJ2-2 with HSA was
significantly low (258 Å2), and its carboxyl group and methylene
tail were disordered, especially in chain B. The bond distance
between the imidazole nitrogen Nτ and cyclopentenone ring
(2.25 Å in chain A) and positive electron densities near the
His146 in chain B (Figure 5A and Supporting Information
Figure S1, panel D) strongly suggest that His146 in both
monomers may covalently bind to ∆12-PGJ2-2. No other
significant interaction was observed between ∆12-PGJ2-2 and
HSA, suggesting that the covalent bond may be indispensable
to sustaining the PG molecule (Figure 5A). In addition, no
Despite the fact that HSA is a fatty acid-binding protein, the
protein is not known to bind ∆12-PGJ2. Therefore, we studied
the HSA-∆12-PGJ2 interaction using surface plasmon resonance
analysis. ∆12-PGJ2 was subjected to an exchange reaction to
preload it and tested for an interaction with HSA immobilized
on a Biacore chip. ∆12-PGJ2 showed rapid binding and dis-
sociation kinetics with a dissociation constant of 420 µM (Figure
3). Of interest, at higher concentrations over 125 µM, a fraction
of ∆12-PGJ2 remained bound to the anchored HSA upon
prolonged exposure, suggesting that ∆12-PGJ2 might covalently
interact with HSA. Moreover, HSA recognizes ∆12-PGJ2 with
an Rmax (the total number of accessible binding sites) of 19.1
resonance units, which was almost comparable to the theoretical
Rmax of 18.7 resonance units when the protein binds one ∆12-
PGJ2 molecule. These findings suggest that HSA may have a
specific binding site for ∆12-PGJ2.
Crystallographic Analysis and ∆12-PGJ2 Binding Sites. To
investigate the molecular details of the binding of the PGD2
metabolite to the protein, we prepared the ∆12-PGJ2 complex
of HSA and analyzed it by X-ray crystallography. The ∆12-
PGJ2-HSA complexes were prepared by incubating the protein
with a 5 molar excess of ∆12-PGJ2 with or without myristate.
9
826 J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010

∆12-PGJ2 as a Product and Ligand of HSA
A R T I C L E S
Figure 4. Crystal structure of HSA complexed with ∆12-PGJ2 and myristate.
(A) Electrostatic potential at the molecular surface of both sides of the
complex and location of myristate and ∆12-PGJ2 binding sites. HSA is
composed of three homologous domains (I-III), each of which is divided
into A and B subdomains. Two ∆12-PGJ2 and two of the six molecules of
myristic acid found to bind to HSA are visible in these figures. The
electrostatic potential is represented on a color scale from blue for a positive
potential, white for neutral, to red for a negative potential. Bound ∆12-
PGJ2 and myristic acid (Myr) are shown in a stick representation with atoms
colored by atom-type: carbon, yellow for ∆12-PGJ2-1 and white for ∆12-
PGJ2-2; oxygen, red. (B) Close-up view of ∆12-PGJ2 binding site. Two ∆12-
PGJ2 molecules bound to HSA are shown as a stick model. Cross-sections
of the protein are shown in a gray color, and the R-helix of the subdomain
IIIA at the cross-section is colored cyan. The ∆12-PGJ2 binding site is located
at the positively charged and hydrophobic groove face of the subdomain
IIIA. Near the ∆12-PGJ2 binding site, negatively charged interdomain cleft
was reported to bind to some drugs. (C) Schematic representation of ∆12-
PGJ2 binding site. The subdomains are color-coded as follows: IA, red; IB,
orange; IIA, yellow; IIIA, cyan.
Figure 5. The omit maps and 2 > Fo - Fc electron density maps for ∆12-
PGJ2 binding in subdomain IB of the HSA-∆12-PGJ2-myristate complex.
(A) Close-up views of the omit maps (contoured at 0.9 σ), showing two
∆12-PGJ2 molecules bound to the subdomains IB of both monomers (chains
A and B). The bound ∆12-PGJ2 and HSA are shown in a stick representation
with atoms colored by atom-type: carbon, yellow; oxygen, red; nitrogen,
blue. In the discussion of the fatty acid binding sites, we have adopted the
site-numbering scheme that was established by Curry et al.17,59 The ∆12-
PGJ2-1 binding site is precisely located at the fatty acid binding site 1, and
∆12-PGJ2-1 binds at this site in the same orientation with polyunsaturated
fatty acids.59 (B) Representations of the interaction between ∆12-PGJ2 and
HSA. Left, a representation of the interaction between ∆12-PGJ2-1 and HSA.
Right, a representation of the interaction between ∆12-PGJ2-2 and HSA.
The residues located near the bound ∆12-PGJ2 are also illustrated in italics.
The covalent bond between ∆12-PGJ2-2 and His146 is denoted by a red
line. The chiral conformation around the carbon atom covalently bound to
His146 is (R).
electron density for ∆12-PGJ2 was observed in the other fatty
acid binding sites or coexisted with the myristate molecule
(Supporting Information Figure S2), suggesting that ∆12-PGJ2
can primary replace myristate and bind at the subdomain IB.
The line shape of these sites or the size of the grooves except
for fatty acid binding sites 1 and 6 may prevent the binding of
the D-shaped ∆12-PGJ2 molecule.
On the other hand, HSA has a highly reactive free sulfhydryl
group at position 34 (Cys34), which is a major ligand-binding
site for thiol-containing compounds and R,ꢀ-unsaturated mol-
ecules.30 However, no modification of the cysteine residue was
observed in the crystal structure of the HSA-∆12-PGJ2 and
HSA-∆12-PGJ2-myristate complexes (Supporting Information
Figure S3).
MALDI-TOF MS and Direct Infusion ESI-MS Analyses of
the HSA-∆12-PGJ2 Complex. The crystallographic analyses of
the HSA-∆12-PGJ2 complex suggested the presence of ∆12-
PGJ2-His146 adduct in HSA. To confirm this result, the native
and ∆12-PGJ2-treated HSA were analyzed by MALDI-TOF MS.
The analysis of the native HSA revealed a peak with m/z 67 473,
whereas the protein incubated with 1 mM ∆12-PGJ2 had a peak
with m/z 67 988 (Figure 6A). The peak differs in molecular mass
by approximately 500 Da, which is close to the molecular mass
of ∆12-PGJ2 (334 Da), suggesting the addition of at least one
molecule of the ∆12-PGJ2 per protein. To confirm the formation
of the ∆12-PGJ2-HSA covalent adduct, the native and the ∆12-
PGJ2-treated HSA were analyzed by direct infusion ESI-MS
analysis. Figure 6B represents the positive-ion ESI mass
spectrum of serum HSA (scan range m/z 600-2000), showing
multiple charged peaks ranging from m/z 1055.92 to m/z 1955.17
and corresponding to the addition of 63 to 34 protons. To further
enhance the MS resolution, a reduced scan range m/z 1410-1500
was used to acquire the MS spectrum for 10 min at an m/z 0.5
resolution. Figure 6C shows the ESI-MS spectrum in the
positive-ion mode of the native HSA, characterized by three
abundant ions at m/z 1414.7, 1445.4, and 1477.5, and attributed
(30) Aldini, G.; Gamberoni, L.; Orioli, M.; Beretta, G.; Regazzoni, L.;
Maffei Facino, R.; Carini, M. J. Mass Spectrom. 2006, 41, 1149–
1161.
9
J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010 827

A R T I C L E S
Yamaguchi et al.
40% with respect to mercapto-albumin) (Figure 6D). A decon-
volution analysis indicated a MW 66 778 [+332 Da in respect
to native HSA (66 446 Da)], which can be attributed to the ∆12-
PGJ2-Michael adduct (theoretical mass increment +334.5 Da).
Identification of the ∆12-PGJ2-Binding Site Using Nano-LC-
MS/MS. The ∆12-PGJ2-mediated structural modification(s) of
HSA were further characterized by a nano-LC-MS/MS analysis
of the incubated samples after reduction with NaBH4, an
established procedure for adduct stabilization, followed by
enzymatic digestion. When trypsin was used as the proteolytic
enzyme, peptide mass mapping provided identification of the
peptides, accounting for approximately 80% of the protein
sequence as previously reported.30 The ∆12-PGJ2-adducted
peptide was first identified by using a semiquantitative approach
recently proposed by Aldini et al., which identifies those peptides
undergoing a significant consumption with respect to the
peptides from the native HSA.30 In the HSA:∆12-PGJ2 1:5 molar
ratio samples, only the RHPYFYAPELLFFAK peptide under-
went a consumption over (10% (chosen as threshold level)
(Supporting Information Figure S4). In particular, the peptide
was reduced by almost 50%, in agreement with the relative
abundance of the ∆12-PGJ2-HSA adduct, accounting for almost
50% with respect to the native HSA (calculated by baseline
subtraction). On the basis of the sequence of the quenched
RHPYFYAPELLFFAK peptide, the MW and the corresponding
single and multicharged state ions for the ∆12-PGJ2 Michael
adduct were then predicted. These predicted ion values were
used as filter ions to reconstitute the SIC traces for both the
native (Supporting Information Figure S5, panel A) and the ∆12-
PGJ2-treated HSA (Supporting Information Figure S5, panel B).
This latter molecule only contains a well detectable peak (RT
) 63 min), whose MS spectrum, showing multicharged peaks
at m/z 559.81 (z ) 4) and m/z 746.08 (z ) 3) (Supporting
Information Figure S5, panel C), confirms the MW of the
adducted peptide (MW ) 2234.49). An MS/MS analysis was
used for further confirmation and to identify the site of
adduction. The MS2 spectrum of the [M + 2H]2+ parent ion at
m/z 950.50 relative to the unmodified peptide (native HSA
sample) is characterized by several b and y fragment ions (Figure
7A). The MS2 spectrum of the [M + 3H]3+ parent ion at m/z
746.08 relative to the ∆12-PGJ2 adduct (Figure 7B) shows the
unmodified y-ion series (up to y13), and the b-ion series from
the b2* ion, shifted by 336.5 Da, thus unequivocally identifying
His146 as the adduction site.
Figure 6. Covalent binding of ∆12-PGJ2 to HSA. (A) MALDI-TOF-MS
analysis of HSA treated with and without ∆12-PGJ2. (B-D) ESI-MS spectrum
(linear ion-trap as mass analyzer) of isolated serum HSA using mass ranges
m/z 900-2000 (panel B) and m/z 1410-1500 (panels C and D). On the
right side are depicted the deconvoluted spectra. HSA was incubated for
3 h at 37 °C in the absence (panel C) and presence of ∆12-PGJ2 (panel D)
at the molar ratio of HSA:∆12-PGJ2 ) 1:1. In panel C, MA indicates
mercapto albumin, and Cys and Gly represent the cysteinylated (66 566
Da) and glycated (66 610 Da) forms of HSA, respectively. In panel D, “*”
indicates the signals relative to the monoadduct.
Discussion
Electrophilic ligands, in addition to being noncovalently
bound by the protein, have the potential to undergo nucleophilic
substitution or addition reactions with the protein. Specific amino
acid residues are frequently targeted by electrophilic reactants.
The reasons for this site selectivity are not well understood,
but most small organic molecules appear to be noncovalently
bound in a specific region that contains targeted residues. If
noncovalent binding, which would constrain the rotational and
translational freedom of the electrophile, precedes the nucleo-
philic substitution reaction, it is understandable that only specific
residues would be attacked. It is proposed that docking of an
electrophile in a receptor region of a protein is generally the
first step on the path to covalent adduct formation, and, as a
result, only appropriately situated nucleophiles within the
receptor are available for reaction.31 Members of the J2 series
to the +47, +46, and +45 multicharged ions of the native HSA
(mercapto-albumin; compound MA). As previously reported
working with a triple-stage quadrupole system,30 the MS
spectrum is also characterized by two other ions series relative
to the cysteinylated (compound Cys; +120 Da found; +119
Da theoretical) and glycated forms (compound Gly, +164 Da
found; +162 theoretical), reaching a relative abundance of
30-35% with respect to the native HSA. The deconvoluted
spectra (right panels) show molecular weights of 66 446, 66 566,
and 66 610 Da for compounds MA, Cys, and Gly, respectively.
When HSA was incubated in the presence of ∆12-PGJ2 at a
molar ratio 1:1 for 180 min at 37 °C, an additional ion series
(marked by “*”) was quite detectable (relative abundance ≈
(31) Skipper, P. L. Chem. Res. Toxicol. 1996, 9, 918–923.
9
828 J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010

∆12-PGJ2 as a Product and Ligand of HSA
A R T I C L E S
Figure 7. Identification of the ∆12-PGJ2-binding site using nano-LC-MS/MS. (A) LC-ESI-MS/MS spectra of native and ∆12-PGJ2-adducted RHPYFYAPELL-
FFAK peptide. LC-ESI-MS/MS spectra of native (parent ion m/z 950.50, upper panel) and ∆12-PGJ2-adducted (parent ion m/z 746.08, lower panel)
RHPYFYAPELLFFAK peptide. His146 was identified as the adduction site. The modified fragment ions are labeled with an asterisk and according to the
nomenclature of the peptide fragmentation.60 (B) Main interactions stabilizing the complex between HSA and ∆12-PGJ2, which precede the covalent adduction.
The dashed line underlines the closeness between the imidazole and cyclopentenone ring and confirms that such a configuration is conducive for the Michael’s
addition.
of the PGs, unlike other classes of eicosanoids, characterized
by the presence of an electrophilic R,ꢀ-unsaturated carbonyl
group in the cyclopentenone ring, have their own unique
spectrum of biological effects. The reactive center of the cyPGs
has been proposed to account for some of their receptor-
independent biological actions.32,33 They can covalently react
via the Michael addition reaction with nucleophiles, such as
the free sulfhydryls of glutathione and cysteine residues in
cellular proteins that play an important role in the control of
the redox cell signaling pathways.11,32,33
In the present study, based on the molecular interaction
analysis suggesting the presence of a specific ∆12-PGJ2-binding
site in HSA (Figure 3), we analyzed the crystal structure of the
HSA-∆12-PGJ2 complex and identified the ∆12-PGJ2-binding
site in HSA, which lies in a D-shaped cavity in the center of
the four-helix bundle of subdomain IB (Figure 4). It was
observed that two ∆12-PGJ2 molecules were bound within the
binding pocket. No ∆12-PGJ2 molecule was detected at other
fatty acid binding sites or coexisted with the fatty acid molecule
(myristate), indicating that ∆12-PGJ2 could replace the fatty acid
only in subdomain IB. Of interest, the electron density for the
∆12-PGJ2 was stronger in the HSA-myristate-∆12-PGJ2 complex
than in the HSA-∆12-PGJ2 complex, indicating that ∆12-PGJ2
might be more stably associated with the pockets in the presence
of a fatty acid. It is apparent from these findings that subdomain
IB is the preferred binding site for the PGD2 metabolite. This
site is not observed in the defatted HSA structure34 and may
occur only in the presence of significant amounts of fatty acids.
Selective binding of ∆12-PGJ2 to subdomain IB raises the
question as to why no ∆12-PGJ2 molecule is bound to other fatty
acid-binding sites. It may be simply because, in comparison to
the other major fatty acid binding pockets on the protein,
subdomain IB is relatively open and accessible to solvent.
Indeed, Tyr138 stacks with Tyr161 and occludes the binding
pocket in the absence of fatty acid, whereas, upon fatty acid
binding at this site, the tyrosine side-chains rotate through about
(32) Rossi, A.; Kapahi, P.; Natoli, G.; Takahashi, T.; Chen, Y.; Karin, M.;
Santoro, M. G. Nature 2000, 403, 103–108.
(33) Bui, T.; Straus, D. S. Biochim. Biophys. Acta 1998, 1397, 31–42.
(34) He, X. M.; Carter, D. C. Nature 1992, 358, 209–215.
9
J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010 829

A R T I C L E S
Yamaguchi et al.
90° so that their phenyl rings hold the lipid in a hydrophobic
clamp.35 It has been shown that all of the fatty acids bind at
this site in the same orientation with the carboxyl group
hydrogen-bonded to Arg117 and to a water molecule that is
also coordinated by the side-chain hydroxyl group of Tyr161
and the carbonyl oxygen atom of Leu182. The preferential
binding of ∆12-PGJ2 to subdomain IB may be associated with
the fact that the site represents a catalytic domain where PGJ2
is converted to ∆12-PGJ2.
nonmercaptalbumin, respectively. The redox conversion of
serum albumin has been considered to largely participate in the
maintenance of a constant redox potential in the extracellular
fluids, and the value for the mercaptalbumin fraction on serum
albumin might reflect the redox buffer capacity in the body. In
addition, the oxidized forms of albumin have been found to be
increased under several pathological conditions.40-43 However,
no modification of Cys34 was observed in the crystal structure
of the HSA-∆12-PGJ2 and HSA-∆12-PGJ2-myristate complexes
(Supporting Information Figure S3). The docking searches
around Cys34 have indeed shown that the cavity lined by Cys34,
whose remarkable reactivity toward R,ꢀ-unsaturated aldehydes
was investigated in a previous study,44 is too narrow to
encompass the ∆12-PGJ2 molecule, and therefore the reactive
cyclopentenone moiety of ∆12-PGJ2 is always situated too far
from Cys34 to yield the corresponding Michael adducts (Aldini,
G., unpublished observation). Thus, due to the narrowness of
the cavity, ∆12-PGJ2 can at most insert the aliphatic chains into
the cavity, but the cyclopentenone ring remains in an external
position, which cannot yield covalent adducts with Cys34.
Serum albumin is synthesized in the liver, exported as a
nonglycosylated protein, and is present in the blood at around
40 mg/mL (0.6 mM). It is the major transport protein for
unesterified fatty acids, but is also capable of binding an
extraordinarily diverse range of metabolites, drugs, and organic
compounds.16,39,45 It has also been shown that albumin can
induce a variety of cellular processes that represent a cellular
response program under apoptotic and inflammatory stresses.
Zoellner et al. have shown that serum albumins inhibit apoptosis
of endothelial cells.46 In addition, albumin has been shown to
activate the AKT pathway and promote resistance of B-chronic
lymphocytic leukemia cells to DNA damage-induced apoptosis
in vitro.47 Erkan et al. have reported that albumin overload
induces apoptosis in cultured proximal tubular cells, mediated
at least in part by the Fas-FADD-caspase 8 pathway.48 Albumin
is taken up by the proximal tubule cells via receptor-mediated
endocytosis and can induce the expression of several inflam-
matory molecules. This complex mixture of inflammatory
substances induced by albumin may contribute to a number of
cellular events associated with immune and inflammatory
processes. However, the potential role of albumin, or its multiple
ligands, on the regulation of cellular functions has not been fully
explored.
Crystallographic and mass spectrometric analyses of the
HSA-∆12-PGJ2 complex also showed that ∆12-PGJ2 could form
a covalent adduct at His146. The docking analyses confirmed
that the interdomain cleft between subdomains IB and IIA can
accommodate ∆12-PGJ2 in a pose conducive to the Michael
adduct (Figure 4B and C). The best docking configuration of
∆12-PGJ2 around His146 suggests that the distance between the
cyclopentenone ring and the imidazole nitrogen is reasonably
productive for the Michael adduction (4.7 Å). The identification
of histidine as the covalent binding site of ∆12-PGJ2 was
unexpected because the cyPGs have been considered to be
totally inert to amino acids, except cysteine.36 The ∆12-PGJ2-
histidine Michael adduct may be stabilized toward retro-Michael
reaction, on account of the poorer leaving group ability of
imidazole over other functional groups, such as sulfhydryl and
amine, at neutral conditions. His146 is located at the entrance
of fatty acid binding site 1 in the strong nucleophilic region
(Figure 4) that has been previously reported to be the adduction
site of a lipid peroxidation-derived aldehyde,30 polycyclic
aromatic hydrocarbon epoxides,37 and of the butadiene metabo-
lite, epoxybutanediol.38 As illustrated in Supporting Information
Figure S6, His146 faces a well-known pocket in subdomain IB,
which is characterized by a small number of nucleophile residues
(Lys137, His146, and Lys159) with an exceptional reactivity.
Even if this subdomain has not been regarded as a true drug-
binding site, it is accepted that reactive aromatic electrophiles
have a marked specificity for this region where they can form
covalent adducts with the mentioned nucleophilic residues. Such
specificity can be explained by considering that this region is
particularly rich in aromatic and positively charged residues,
which can favor the noncovalent binding preceding covalent
bond formation. Moreover, the richness of positively charged
residues can also justify the remarkably low pK3 values of the
mentioned basic residues, which favor their neutral form
rendering them reactive nucleophiles to yield covalent adducts.
Thus, it is likely that the presence of a high affinity-binding
site of ∆12-PGJ2 in a specific region might be associated with
the formation of an unusual covalent bond with the histidine
residue.
cyPGs, such as ∆12-PGJ2 and 15d-PGJ2, emerged as a likely
regulator of acute and chronic inflammation. The anti-inflam-
(40) Kawai, K.; Yoh, M.; Hayashi, T.; Imai, H.; Negawa, T.; Tomida, M.;
Sogami, M.; Era, S. Tokai J. Exp. Clin. Med. 2001, 26, 93–99.
(41) Tomida, M.; Ishimaru, J.; Hayashi, T.; Nakamura, K.; Murayama, K.;
Era, S. Jpn. J. Physiol. 2003, 53, 351–355.
One of the most important features of the molecular structure
of serum albumin is the presence of a highly reactive free
sulfhydryl group at Cys34. This residue is a major ligand-
binding site for thiol-containing compounds, such as cysteine
and glutathione, and for various metal ions in a reversible
manner.39,40 Circulating serum albumin contains this Cys34 in
the reduced and oxidized states known as mercaptalbumin and
(42) Mera, K.; Anraku, M.; Kitamura, K.; Nakajou, K.; Maruyama, T.;
Otagiri, M. Biochem. Biophys. Res. Commun. 2005, 334, 1322–1328.
(43) Musante, L.; Bruschi, M.; Candiano, G.; Petretto, A.; Dimasi, N.; Del
Boccio, P.; Urbani, A.; Rialdi, G.; Ghiggeri, G. M. Biochem. Biophys.
Res. Commun. 2006, 349, 668–673.
(44) Aldini, G.; Vistoli, G.; Regazzoni, L.; Gamberoni, L.; Facino, R. M.;
Yamaguchi, S.; Uchida, K.; Carini, M. Chem. Res. Toxicol. 2008, 21,
824–835.
(45) Kragh-Hansen, U. Danish Med. Bull. 1990, 37, 57–84.
(46) Zoellner, H.; Ho¨fler, M.; Beckmann, R.; Hufnagl, P.; Vanyek, E.;
Bielek, E.; Wojta, J.; Fabry, A.; Lockie, S.; Binder, B. R. J. Cell Sci.
1996, 109, 2571–2580.
(35) Curry, S.; Brick, P.; Franks, N. Biochim. Biophys. Acta 1999, 1441,
131–140.
(36) Suzuki, M.; Mori, M.; Niwa, T.; Hirata, R.; Furuta, K.; Ishikawa, T.;
Noyori, R. J. Am. Chem. Soc. 1997, 119, 2376–2385.
(47) Jones, D. T.; Ganeshaguru, K.; Anderson, R. J.; Jackson, T. R.;
Bruckdorfer, K. R.; Low, S. Y.; Palmqvist, L.; Prentice, H. G.;
Hoffbrand, A. V.; Mehta, A. B.; Wickremasinghe, R. G. Blood 2003,
101, 3174–3180.
(37) Brunmark, P.; Harriman, S.; Skipper, P. L.; Wishnok, J. S.; Amin, S.;
Tannenbaum, S. R. Chem. Res. Toxicol. 1997, 10, 880–886.
(38) Lindh, C. H.; Kristiansson, M. H.; Berg-Andersson, U. A.; Cohen,
A. S. Rapid Commun. Mass Spectrom. 2005, 19, 2488–2496.
(39) Carter, D. C.; Ho, J. X. AdV. Protein. Chem. 1994, 45, 152–203.
(48) Erkan, E.; De Leon, M.; Devarajan, P. Am. J. Physiol. Renal. Physiol.
2001, 280, F1107–F1114.
9
830 J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010

∆12-PGJ2 as a Product and Ligand of HSA
A R T I C L E S
for 3 min, and this was followed by a blocking step using
ethanolamine (1 M, pH 8.5) at 10 µL/min for 7 min. Binding studies
were performed by passing ∆12-PGJ2 over the immobilized HSA
at the flow rate of 10 µL/min. The association and dissociation
times were 60 and 60 s, respectively. The running buffer contained
150 mM NaCl, 10 mM HEPES, 3 mM EDTA, 0.05% (v/v)
surfactant P-20, pH 7.4, and 5% DMSO. The complex formation
between ∆12-PGJ2 and the immobilized protein was measured with
3.0-500 µM of ligand in the running buffer. A sample volume of
20-40 µL was injected into the flow cell. Dissociation of the bound
ligand was initiated by injecting running buffer without any ligand.
matory effect was first deduced from an animal model of
carrageenin-induced inflammation in which a two-phase PG
release has been described after the expression of COX-2; the
PGE2 synthesis predominates during the early inflammatory
step, whereas 15d-PGJ2 substitutes for the PGE2 formation at
the end of the process, coincident with the accumulation of
macrophages.12 Cuzzocrea et al. also demonstrated the effects
of 15d-PGJ2 in rodent models of chronic collagen-induced
arthritis inflammation.13 In a later study, Itoh et al. showed that
the accumulation of 15d-PGJ2 activates Nrf2 and regulates the
inflammation process through the induction of a target gene
expression.14 On the basis of these findings, the physiopatho-
logical relevance of the enhanced synthesis of cyPGs in activated
macrophages has been speculated to represent a mechanism, in
which they function as a feedback regulator of the inflammatory
responses. Of interest, our recent study has revealed that the
HSA-∆12-PGJ2 complexes markedly induce the gene expression
of pro-inflammatory molecules, such as COX-2.49 These
observations suggest that the cyPGs, at least ∆12-PGJ2, might
have pro-inflammatory roles and challenges the concept that
they are exclusively involved in the anti-inflammatory responses.
Future studies are required to determine whether and how the
enhanced production of the cyPGs, ∆12-PGJ2 in particular, could
be functionally associated with inflammatory processes and the
net effect in vivo of the interplay between the anti- and pro-
inflammatory products of COX in human diseases.
Protein Purification, Complex Formation, and Crystalliza-
tion. The protocol was based on a previously developed method
for formation of the HSA-myristate complex.17 The HSA fraction
V fatty acid free (Sigma-Aldrich) was further purified by anion-
exchange chromatography on UNO Q-6 (Bio-Rad Laboratories) pre-
equilibrated with buffer A (20 mM phosphate buffer, pH 7.0). The
protein was eluted with a linear gradient of NaCl (0-0.5 M). The
fractions containing HSA were pooled, then applied onto a HiTrap
desalting column (GE Healthcare) pre-equilibrated in buffer A to
remove the NaCl. The purified protein was concentrated to 100
mg/mL using the Vivaspin 6 concentrator (10-kDa molecular-weight
cutoff, Sartorius). Myristate (Wako, Osaka, Japan) dissolved in
buffer A was heated to 50 °C to facilitate the dispersal of myristate,
and then mixed with solution containing HSA to achieve the
myristate molar ratio of 4. To gain the tertiary complex of HSA,
myristate, and ∆12-PGJ2, the ∆12-PGJ2 solution (Cayman Chemical)
was concentrated under a gentle stream of nitrogen, and then a
5-fold molar excess of ∆12-PGJ2 was directly dissolved in the
HSA-myristate solution. Crystals were obtained by the vapor
diffusion hanging drop technique against a reservoir solution
containing 50 mM phosphate buffer, pH 6.5-7.0, 28-32% (w/v)
PEG 3350. Colorless crystals with the dimensions of 0.5 × 0.5 ×
1.0 mm grew within 2 weeks. The crystals were harvested by the
gradual addition of 2 µL of cryoprotectant solution containing ∆12-
PGJ2 and slightly higher PEG 3350 concentrations than were used
for crystallization to the hanging drop. The diffraction data were
collected at 98 K using the beamline BL5A of the Photon Factory
(Tsukuba, Japan).
Experimental Section
Materials. HSA (fatty acid and globulin free, purity 99%) was
obtained from Sigma-Aldrich. ∆12-PGJ2 and free fatty acids were
purchased from the Cayman Chemical Co. (Ann Arbor, MI). The
protein concentration was measured using the BCA protein assay
reagent obtained from Pierce. The sequence grade modified trypsin
was obtained from Promega (Milan, Italy). The LC-MS grade
solvents (Chromasolv) were purchased from Sigma-Aldrich (Milan,
Italy).
In Vitro Modification of Isolated HSA by ∆12-PGJ2. Blood
collected by venopuncture from two healthy male volunteers was
allowed to clot at room temperature for no longer than 30 min and
then centrifuged for 10 min at 1500g. The serum aliquots were
then stored in liquid nitrogen until used. Serum albumin was isolated
using the Affi-gel blue-gel (Biorad, Milan, Italy) and eluted with
1.4 M NaCl in 20 mM phosphate buffer saline (PBS, pH 7.2). The
HSA concentration was measured by absorbance at 279 nm (E1%
1 cm ) 5.31). The samples were then desalted on YM30 Microcon
centrifugal filter devices (Millipore, Milan, Italy) by rinsing with
distilled water, and diluted with 20 mM PBS (pH 7.4) to a final
concentration of 40 µM. ∆12-PGJ2 dissolved in methanol was then
added to obtain the final HSA:∆12-PGJ2 molar ratios of 1:1 and
1:5. After a 180 min incubation at 37 °C, the samples were analyzed
by direct infusion mass spectrometry as detailed below. Additional
samples were then incubated at room temperature for an additional
60 min with NaBH4 (final concentration 5 mM), and then desalted
by Microcon filter devices as already described, lyophylized, and
stored in liquid nitrogen until enzyme digestion.
Data Collection and Structure Determination. Diffraction
images of the data set were indexed, integrated, and scaled with
the HKL2000 program suite.50 The crystals are in the space group
P1, with unit cell dimensions a ) 38.1, b ) 92.1, c ) 94.7, R )
74.8°, ꢀ ) 89.5°, γ ) 80.2°, and contain two molecules (chain
A/B) in the asymmetric unit. The initial phases were obtained by
molecular replacement with the program Phaser in the CCP4
program suite.51 The search model stripped of its myristate was
taken from the PDB code 1BJ5. Model building in the electron
density map and crystallographic refinement were performed using
the programs in the CCP4 program suite and Coot.52 After cycles
of restrained refinement and manual adjustments, the resulting
electron density maps were used to guide positioning of the ligands
and bound water molecules and further refined in Refmac using 3
TLS groups.53 The omit map was also calculated using the program
Sfcheck to reduce the model bias in the electron density.54 The
final model, refined to 2.20, contains two proteins, three ∆12-PGJ2,
12 myristic acids, and the solvent molecules. The surface electro-
static potentials of the PTP1Bs were calculated by the Poisson-
Surface Plasmon Resonance. The binding of ∆12-PGJ2 to HSA
was measured by a Biacore T100 system at 25 °C and a flow rate
of 100 µL/min. HSA (3800 response units) was immobilized on a
CM5 surface via covalent linkage to the N terminus of HSA. A
CM5 chip was activated using 1:4 N-hydroxysuccinimide (NHS)/
1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) at the flow
rate of 10 µL/min for 7 min. HSA (20 µg/mL) in 10 mM acetate
buffer (pH 5.5) was passed over separate flow cells at 10 µL/min
(50) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307–326.
(51) Collaborative Computational Project, Number 4, The CCP4 suite:
programs for protein crystallography. Acta Crystallogr., Sect. D 1994,
50, 760-763.
(52) Emsley, P.; Cowtan, K. Acta Crystallogr., Sect. D 2004, 60, 2126–
2132.
(53) Vagin, A. A.; Steiner, R. A.; Lebedev, A. A.; Potterton, L.; McNi-
cholas, S.; Long, F.; Murshudov, G. N. Acta Crystallogr., Sect. D
2004, 60, 2184–2195.
(49) Morishita, N.; Shibata, T.; Ito, S.; Carini, M.; Aldini, G.; Uchida, K.,
to be submitted.
(54) Vaguine, A. A.; Richelle, J.; Wodak, S. J. Acta Crystallogr., Sect. D
1999, 55, 191–205.
9
J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010 831

A R T I C L E S
Yamaguchi et al.
Boltzmann equation as implemented in the eF-surf server,55 and
all figures were prepared using MolFeat (FiatLux Co.) and Coot.
The coordinates and structure factors have been deposited in the
Protein Data Bank (ID code is 3A73). To evaluate the ligand-protein
interactions within the asymmetric unit, we utilized the Protein
Interfaces, Surfaces and Assemblies (PISA) service.56 The data
collection and refinement statistics are summarized in Supporting
Information Table S1.
Mass Spectrometry. Direct Infusion Electrospray Mass
Spectral Analysis (ESI-MS): Linear Ion Trap Mass Spectrom-
eter. To detect changes in the protein mass and to determine the
stoichiometry of the reaction, undigested native and ∆12-PGJ2-
treated HSA were analyzed by direct infusion using an LXQ Linear
Ion Trap mass spectrometer (Thermo Scientific, Milan, Italy)
equipped with an Electrospray Finnigan Ion Max source. Lyoph-
ilized protein (500 µg) was dissolved in 200 µL of water, mixed
with 200 µL of CH3CN-H2O-HCOOH (60:40:0.4, v/v/v), and
infused into the mass spectrometer at the flow rate of 3 µL/min.
Each sample was analyzed using two different mass ranges, m/z
600-2000 and m/z 1410-1500, under the following instrumental
conditions: positive ion mode, 350 °C capillary temperature; 3.5
kV spray voltage applied to the needle, 46 V capillary voltage,
nebulizer gas (nitrogen) flow rate set at 20 (au), and 5 min
acquisition.
Nano-LC-MS/MS Analysis. Peptide maps were generated by
nano-LC-ESI-MS/MS analysis in data-dependent scan mode and
dynamic exclusion in positive-ion mode. For the sample preparation,
the lyophilized target protein was dissolved in 100 µL of 50 mM
Tris-HCl (pH 7.8) and digested according to the manufacturer’s
procedure. Two microliter aliquots of the digested samples were
diluted 1:100 with 0.1% formic acid and injected into a quaternary
pump HPLC system (Surveyor LC system, ThermoQuest, Milan,
Italy). All of the digested peptide mixtures were separated by online
reversed-phase (RP) nanoscale capillary liquid chromatography
(nanoLC) and analyzed by electrospray tandem mass spectrometry
(ESI-MS/MS). Chromatography was performed using a Surveyor
LC system (ThermoFinnigan Italia, Milan, Italy) on a 180 µm ×
10 cm column packed with 5 µm, Biobasic-18 stationary phase
(Thermo, Superchrom, Milan, Italy). The pump flow rate was split
1:75 for a column flow rate of 2 µL/min. The column effluent was
directly electrosprayed using the silica emitter source without further
splitting. Mobile phases A and B were 0.1% formic acid in water
and in acetonitrile, respectively. The separation of the peptides
obtained by enzymatic digestion was achieved with a gradient of
0-60% B over 60 min. Before the next analysis, both the precolumn
and the column were first washed with 90% solvent B for 10 min
and then equilibrated with 100% solvent A for 20 min. For the
identification of the peptides, an LTQ XL-Orbitrap mass spectrom-
eter was used (Thermo Scientific, Milan, Italy), and the electrospray
interface (dynamic nanospray probe, Thermo Scientific, Milan, Italy)
was set as follows: 1.5 kV spray voltage; 180 °C capillary
temperature, 30 V capillary voltage; 110 V tube lens offset, and
no sheath or auxiliary gas flow. During the analysis, the mass
spectrometer continuously performed scan cycles in which a high-
resolution (resolution r ) 30 000 at m/z 400) full scan (300-2000
m/z) in the profile mode was first made by the Orbitrap, after which
the MS2 spectra were recorded in the centroid mode for the three
most intense ions (isolation width, 3 m/z; normalized collision
energy, 30 CID arbitrary units). Protonated phthlates [dibutylphth-
late (plasticizer), m/z 279.159086; bis(2-ethylhexyl)phthalate, m/z
391.284286] and polydimethylcyclosiloxane ions [(Si(CH3)2O)6 +
H]+; m/z 445.120025] were used for the real time internal mass
calibration. Dynamic exclusion was enabled (repeat count, 3; repeat
duration, 10 s; exclusion list size, 25; exclusion duration, 120 s;
relative exclusion mass width, 5 ppm). Charge state screening and
monoisotopic precursor selection were enabled, and singly and
unassigned charged ions were not fragmented.
Peptide Adduct Identification. Peptide ion responses were
determined by measuring the peak areas in the selected ion
chromatograms (SICs) reconstituted using the most abundant
multicharged ion as the filter ion as previously reported.57 The
peptideresponseswerenormalizedwithrespecttotheLVAASQAAL-
GL peptide, chosen as the reference peptide because it does not
contain nucleophilic residues that can be adducted by ∆12-PGJ2.
The loss of specific peptides in the ∆12-PGJ2-treated HSA with
respect to the native HSA was determined by calculating the relative
peptide consumption using the following equation: peptide con-
sumption (%) ) 100 - [(A12-dPGJ2)/(ANATIVE) × 100], where A12-
dPGJ2 and ANATIVE are the normalized peptide ion responses of the
target peptide from HSA incubated in the presence and in the
absence of ∆12-PGJ2, respectively.
Bioinformatics. The direct infusion ESI-MS spectra were
deconvoluted using the software packages Bioworks 3.1 (Thermo-
Quest, Milan, Italy) and MagTran 1.02.58 The peptide sequences
were identified using the software turboSEQUEST (Bioworks 3.1,
ThermoQuest, Milan, Italy) and using a database containing only
the protein of interest and assuming trypsin digestion. The in silico
tryptic digest of HSA was also used as an aid for peptide
identification. The protein sequence of HSA was obtained from
the SWISS-PROT database (primary accession number P02768).
PEPTIDEMASS was used to calculate the theoretical digested
masses on the basis of the tryptic digest and setting all the Cys as
carbamidomethyl-cysteine. The predicted y and b series ions were
determined using the Peptide Sequence Fragmentation Modeling,
Molecular Weight Calculator software program (ver. 6.37), http://
come.to/alchemistmatt. The Xcalibur software provided the instru-
ment control and data analysis of the Thermo Electron mass
spectrometers.
Acknowledgment. This work was supported by a Grant-in-Aid
for Scientific Research on Innovative Areas (Research in a Proposed
Research Area), MEXT, Japan (K.U.), and by MIUR (Ministero
dell’Istruzione, dell’ Universita` e della Ricerca), Italy (PRIN 2007).
Supporting Information Available: Supplementary figures for
X-ray crystallography and mass spectrometry, and a supple-
mentary table listing data collection and refinement statistics.
This material is available free of charge via the Internet at http://
pubs.acs.org.
JA908878N
(57) Aldini, G.; Gamberoni, L.; Orioli, M.; Beretta, G.; Regazzoni, L.;
Maffei Facino, R.; Carini, M. J. Mass Spectrom. 2006, 41, 1149–
1161.
(58) Zhang, Z.; Marshall, A. G. A. J. Am. Soc. Mass Spectrom. 1998, 9,
225–23.
(59) Petitpas, I.; Grune, T.; Bhattacharya, A. A.; Curry, S. J. Mol. Biol.
2001, 314, 955–960.
(55) Kinoshita, K.; Nakamura, H. Bioinformatics 2004, 20, 1329–1330.
(56) Krissinel, E.; Henrick, K. J. Mol. Biol. 2007, 372, 774–797.
(60) Biemann, K. Biomed. EnViron. Mass Spectrom. 1988, 16, 99–111.
9
832 J. AM. CHEM. SOC. VOL. 132, NO. 2, 2010