Steric control of oxygenation regiochemistry in soybean lipoxygenase-1

Full text of DOI:10.1021/ja003855k

J. Am. Chem. Soc. 2001, 123, 2931-2932
Scheme 1. Simplified Chemical Mechanism of SLOa
2931
Steric Control of Oxygenation Regiochemistry in
Soybean Lipoxygenase-1
Michael J. Knapp, Florian P. Seebeck,‡ and
Judith P. Klinman*
Department of Chemistry, UniVersity of California
Berkeley, California 94720
ReceiVed NoVember 2, 2000
Mammalian lipoxygenases catalyze the formation of important
biological secondary messengers and have been implicated in
carcinogenesis1 and inflammation.2 While the catalytic reaction
of all known lipoxygenases leads to the formation of a substrate-
derived radical intermediate, most lipoxygenase products are
highly regio- and stereopure.3 Soybean lipoxygenase-1 (SLO),
the prototypical model lipoxygenase, catalyzes the oxidation of
linoleic acid (LA) by O2, producing 13-(S)-hydroperoxyoctadeca-
dienoic acid (HPOD). The first chemical step in this reaction4 is
abstraction of the pro-S hydrogen atom from C-11 by the active-
site Fe3+-OH to generate a putatively delocalized carbon based
radical (L•) and Fe2+-OH2. This intermediate reacts with O2 to
regenerate Fe3+-OH and the product 13-(S)-HPOD (Scheme 1).
Reactions in solution between carbon radicals and triplet O2
typically lead to many isomers and are generally diffusion
limited.5 While SLO reacts with O2 at a rapid rate (kcat/KM(O2) )
4.4 × 106 M-1 s-1), this reaction remains highly selective for O2
attack at C-13. This has led to the notion that the active-site Fe2+
influences the site of O2 insertion via a purple (Fe-OOL)2+
intermediate.18 However, this idea is difficult to reconcile with
the fact that H-abstraction by Fe3+-OH is antarafacial to O2
insertion.6 Inspection of the crystal structure for SLO7 reveals a
a WT-SLO produces predominantly 13-(S)-HPOD at high pH.
‡ Current address: Laboratory of Organic Chemistry, Swiss Federal Institute
of Technology (ETH), CH-8092, Zu¨rich, Switzerland.
(1) Rioux, N.; Castonguay, A. Carcinogenesis 1998, 19, 1393-1400.
(2) Samuelsson, B.; Dahlen, S.-E.; Lindgren, J.; Rouzer, C. A.; Serhan, C.
N. Science 1987, 237, 1171-1176.
(3) Kuhn, H.; Thiele, B. J. FEBS Lett. 1999, 449, 7-11.
(4) Glickman, M. H.; Klinman, J. P. Biochemistry 1996, 35, 12882-12892.
(5) Walling, C. Autoxidation. In ActiVe Oxygen in Chemistry, 1st ed.; Foote,
C. S., Valentine, J. S., Greenberg, A., Liebman, J. F., Eds.; Chapman and
Hall: London, 1995; Vol. 2, pp 24-65.
Figure 1. Substrate cavity and O2 access channel of SLO shown as a
white surface. The substrate LA (C, green; H, white; O, red) has been
modeled with C-1 toward the top of the figure, and with the C-9-C-13
moiety of substrate between Ile538 and Gln495 (orange). The postulated
O2 channel, on top of Ile553 (red), intersects the substrate cavity close to
the C-13 position of LA. Leu546 and Leu754 (blue) constrict the LA
channel, preventing O2 access to C-9 of LA. Fe3+-OH is presented as
CPK spheres (Fe, magenta; O, red; H, white).16
(6) Gardner, H. W. Biochim. Biophys. Acta 1989, 1001, 274-281.
(7) Minor, W.; Steczko, J.; Stec, B.; Otwinowski, Z.; Bolin, J. T.; Walter,
R.; Axelrod, B. Biochemistry 1996, 35, 10687-10701.
(8) Schwarz, K.; Borngraber, S.; Anton, M.; Kuhn, H. Biochemistry 1998,
37, 15327-15335.
(9) Jisaka, M.; Kim, R. B.; Boeglin, W. E.; Brash, A. R. J. Biol. Chem.
2000, 275, 1287-1293.
(10) Gan, Q. F.; Browner, M. F.; Sloane, D. L.; Sigal, E. J. Biol. Chem.
1996, 271, 25412-25418.
(11) Hornung, E.; Walther, M.; Kuhn, H.; Feussner, I. Proc. Natl. Acad.
Sci. U.S.A. 1999, 96, 4192-4197.
side channel intersecting the substrate pocket near the reactive
C-11, proposed7 to permit O2 access. The current report details
our investigation into the role played by bulky hydrophobic
residues in controlling the regiospecificity of SLO. Steady-state
kinetics and product distribution data from single-point mutants
of SLO show that Leu546 and Leu754 grant selectivity for 13-(S)-
HPOD by blocking O2 access to C-9 of LA, and that O2 enters
the active site via the postulated side channel. These results
indicate that SLO controls O2 access to the intermediate L•, and
suggest that Fe2+ plays a secondary role in oxygenation.
(12) Nelson, M. J.; Seitz, S. P. Curr. Opin. Struct. Biol. 1994, 4, 878-
884.
(13) Gillmor, S. A.; Villasenor, A.; Fletterick, R.; Sigal, E.; Browner, M.
F. Nat. Struct. Biol. 1997, 4, 1003-1009.
(14) Holman, T. R.; Zhou, J.; Solomon, E. I. J. Am. Chem. Soc. 1998,
120, 12564-12572.
(15) Rickert, K. W.; Klinman, J. P. Biochemistry 1999, 38, 12218-12228.
(16) All modeling was performed using InsightII/Discover software (Bio-
sym/MSI, San Diego, CA). The CVFF force field was used, and the pro-S
hydrogen of LA was constrained to lie within 3.0 Å of the oxygen atom of
Fe(III)-OH. Only residues within the substrate cavity (channel IIa7) were
minimized, with the exception that Fe and its ligands were held fixed. The
internal Connolly surface was constructed by use of a 0.8 Å solvent sphere.
Using a larger solvent sphere (1.4 Å) resulted in a surface that was interrupted
by a few amino acid side chains (Ile553, Leu564) which could easily reorient to
form a continuous surface, as noted before.7
The X-ray crystal structure of SLO revealed the LA pocket,
cavity IIa,7 as a twisted channel, constricted at the Fe3+-OH by
the side chains of Leu546 and Leu754 (Figure 1). A sharp bend
(17) An O2 electrode was used to determine initial rates at constant [LA]
as a function of [O2], at pH 9.0, 20 °C. Kinetic parameters were determined
by nonlinear least-squares fitting to the Michaelis-Menten equation. Rates
were normalized for Fe content of the enzymes (ICP-AES).
(18) Nelson, M. J.; Cowling, R. A.; Seitz, S. P. Biochemistry 1994, 33,
4966-4973.
near Fe3+-OH is introduced by the side chains of Gln495 and Ile538
.
On the basis of extensive experiments with many lipoxygenases8-13
models have been published6,10,11,13 describing substrate binding
in which the reactive C-11 lies near the constriction formed by
10.1021/ja003855k CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/06/2001

2932 J. Am. Chem. Soc., Vol. 123, No. 12, 2001
Communications to the Editor
Table 1. Kinetic Constantsa (pH 9.0, 20 °C) and Product
Distributions (pH 10.0, 23 °C) for WT-SLO and Mutants
near C-9. The Gln495fAla and Ile538fAla mutations should each
reduce the amount of strain that might favor a localized ene-
allyl radical. However, these mutations had only modest effects
on the rate of oxygenation, suggesting that such an ene-allyl
radical is not formed during turnover, or else that there is rapid
interconversion between the ∆12-[9,10,11] and ∆9-[11,12,13]
radicals. Additionally, these mutants produce the same product
distribution as WT-SLO (vide infra).
% HPODc
kcat
kcat/KM(O2)
enzyme
(s-1
)
(µM-1s-1
)
13(S) 13(R) 9(S) 9(R)
WT-SLO
210
209
4.4
6.1
2.4
95
97
95
3
1
2
3
2
12
2
<1
<1
1
2
<1
2
Ile553fAla
Gln495fAla 120
Ile538fAla
Leu546fAla
Leu754fAla
Ile553fPhe
110
3.0
3.2 (0.2) 93
0.4 (0.3) 85
2
3
3
Table 1 lists the stereo- and regiochemical product distributions
for each enzyme, as determined by chiral HPLC analysis.19,20 It
is immediately apparent that the dominant product for WT-SLO
is 13-(S)-HPOD, with trace amounts of 13-(R), 9-(S), and 9-(R)
isomers. This product distribution is observed for all other
mutants, except for Leu546fAla and Leu754fAla. Both of these
mutants replace a bulky Leu side chain with that of Ala at the
constriction within the substrate cavity. It is noteworthy that
Leu546fAla yields 13-(S)- and 9-(R)-HPOD, the stereoisomers
formed by addition of O2 on the same face of L•; inverse (head-
to-tail) substrate binding would yield the 9-(S) isomer.6 Leu754fAla
yields an increase not only in the 9-(R) isomer, but also a similar
increase in the 13-(R) and 9-(S) isomers, with ca. 3% of the all-
trans-HPOD isomers formed as well (supplementary material).
The reasons for the formation of these other isomers are unclear,
but Leu754 does sit roughly adjacent to the Fe3+-OH, and removal
of this bulky side chain could easily permit the LA substrate to
adopt alternative binding modes. These data indicate that the
constriction in cavity IIa formed by Leu546 and Leu754 sterically
prevents O2 access to the C-9 position within the putative
pentadienyl radical, thereby imparting regioselectivity to this
reaction. This obviates any requirement for the (Fe-OOL)2+
intermediate during turnover, suggesting that purple lipoxygenase
is not catalytically relevant.
These results present a departure from typical views of
substrate-activating oxygenases, in that O2 has seldom been
thought to require a discrete and separate access channel. The
present analysis shows how simple sterics can control the reaction
of an enzyme-bound radical with O2, and it provides a structural
rationale for product distributions within the lipoxygenase family.
Other substrate-activating oxygenases,21,22 such as the cofactor-
generating activity of TOPA-containing amine oxidase,23 may also
sterically direct O2 attack to a specific site of the substrate, thereby
controlling product identity.
10
16
2
0.18 >0.2b
105 0.72
62
95 (2)
10
1
a Corrected for Fe content; see Supporting Information. Standard
error is less than 15% of the parameter, unless specified in parentheses.
b KM(O2) was too low to be measured (e1 µM). c Average of two
separate experiments. The estimated error ((run 1 - run 2)/2) is less
than 1, unless noted in parentheses.
Leu546 and Leu754. Such binding predicts that Gln495 and Ile538
will interact with C-14 and C-8 of L•, respectively, which would
disfavor a planar pentadienyl moiety.7 The side channel bordered
by Ile553 intersects cavity IIa between Leu546 and Gln495, a position
that would be occupied by C-13 of bound L•. Additionally, this
channel encounters the active site on the face of cavity IIa opposite
that of Fe2+. It seems feasible that O2 enters through this channel,
reacting with substrate before encountering Fe2+. To test this
hypothesis, the bulkiness of critical side chains near this intersec-
tion was changed by site-directed mutagenesis,14,15 and the
characteristics of the O2 reaction with L• were compared to those
of WT-SLO.
Table 1 contains a summary of kinetic data17 for these mutants
and WT-SLO. The overall rate of catalysis (kcat) is limited by H•
abstraction, whereas the rate of O2 chemistry (kcat/KM(O2)) reflects
only those steps from O2 encounter to product release. The
mutants exhibited three types of behavior, depending on how the
rates of overall catalysis and of O2 chemistry were affected by
the mutation. Ile553fAla has little effect on the rates of overall
catalysis (kcat) and oxygenation chemistry (kcat/KM(O2)), while
Gln495fAla and Ile538fAla reduce these parameters by ca. 2-fold.
Leu546fAla and Leu754fAla greatly reduce overall catalysis (kcat
decreases by ca. 100-1000-fold), while having a lesser effect
on oxygenation chemistry. Finally, Ile553fPhe is interesting in
that it greatly reduces the rate of oxygenation chemistry relative
to WT-SLO and Ile553fAla, while leaving kcat relatively un-
changed. This suggests that Ile553fPhe impedes O2 access to the
carbon radical intermediate (L•), and supports the view that O2
enters the active site through the channel that passes residue 553.
The radical intermediate L• has generally been depicted as a
pentadienyl radical, in which radical character is spread over C-9-
C-13, generating greater stability than for an allyl radical. This
would make C-9 and C-13 equally reactive to O2 attack, barring
accessibility differences. However, the side chains of Ile538 and
Gln495 provide sufficient bulk to disfavor the planar pentadienyl
intermediate,7 raising the possibility that L• is an ene-allyl radical.
If the unpaired electron density of L• were localized near C-13,
in the form of a ∆9-[11,12,13] allyl radical, then O2 would be
expected to selectively react at C-13. However, an EPR investiga-
tion into the identity of L• showed predominantly a ∆12-[9,10,11]
ene-allyl radical,18 suggesting that radical character is localized
Acknowledgment. We thank Professor T. Holman (UCSC) for
providing the Q495A mutant plasmid. We thank the Swiss National
Foundation and the NIH (GM25765, F32-GM19843) for funding.
Supporting Information Available: Chromatograms showing peak
assignments and a table listing Fe content of each enzyme (PDF). This
material is available free of charge via the Internet at http://pubs.acs.org.
JA003855K
(19) Martini, D.; Iacazio, G. J. Chromatogr. A 1997, 790, 235-241.
(20) Lopez-Nicolas, J. M.; Perez-Gilabert, M.; Garcia-Carmona, F. J.
Chromatogr. A 1999, 859, 107-111.
(21) Orville, A. M.; Lipscomb, J. D.; Ohlendorf, D. H. Biochemistry 1997,
36, 10052-10066.
(22) Stubbe, J.; vanderDonk, W. A. Chem. ReV. 1998, 98, 705-762.
(23) Dove, J. E.; Schwartz, B.; Williams, N. K.; Klinman, J. P. Biochemistry
2000, 39, 3690-3698.