Structural studies of the molybdenum center of the pathogenic R160Q mutant of human sulfite oxidase by pulsed EPR spectroscopy and 17O and 33S labeling

Article abstract of DOI:10.1021/ja801406f

Electron paramagnetic resonance (EPR) investigation of the Mo(V) center of the pathogenic R16OQ mutant of human sulfite oxidase (hSO) confirms the presence of three distinct species whose relative abundances depend upon pH. Species 1 is exclusively present at pH ≤ 6, and remains in significant amounts even at pH 8. Variable-frequency electron spin echo envelope modulation (ESEEM) studies of this species prepared with ³³S-labeled sulfite clearly show the presence of coordinated sulfate, as has previously been found for the blocked form of Arabidopsis thaliana at low pH (Astashkin, A. V.; Johnson-Winters, K.; Klein, E. L.; Byrne, R. S.; Hille, R.; Raitsimring, A. M.; Enemark, J. H. J. Am. Chem. Soc. 2007, 129, 14800). The ESEEM spectra of Species 1 prepared in ¹⁷O-enriched water show both strongly and weakly magnetically coupled ¹⁷O atoms that can be assigned to an equatorial sulfate ligand and the axial oxo ligand, respectively. The nuclear quadrupole interaction (nqi) of the axial oxo ligand is substantially stronger than those found for other oxo-Mo(V) centers studied previously. Additionally, pulsed electron-nuclear double resonance (ENDOR) measurements reveal a nearby weakly coupled exchangeable proton. The structure for Species 1 proposed from the pulsed EPR results using isotopic labeling is a six-coordinate Mo(V) center with an equatorial sulfate ligand that is hydrogen bonded to an exchangeable proton. Six-coordination is supported by the ¹⁷O nqi parameters for the axial oxo group of the model compound, (dttd)Mo¹⁷O( ¹⁷Otms), where H₂dttd = 2,3:8,9-dibenzo-1,4,7,10- tetrathiadecane; tms = trimethylsilyl. Reduction of R160Q to Mo(V) with Ti(III) gives primarily Species 2, another low pH form, whereas reduction with sulfite at higher pH values gives a mixture of Species 1 and 2, as well as the primary high pH form of wild-type SO. The occurrence of significant amounts of the sulfate-blocked form of R160Q (Species 1) at physiological pH suggests that this species may be a contributing factor to the lethality of this mutation.

Full text of DOI:10.1021/ja801406f

Published on Web 06/05/2008
Structural Studies of the Molybdenum Center of the
Pathogenic R160Q Mutant of Human Sulfite Oxidase by
Pulsed EPR Spectroscopy and 17O and 33S Labeling
Andrei V. Astashkin,† Kayunta Johnson-Winters,† Eric L. Klein,† Changjian Feng,†,‡
Heather L. Wilson,§ K. V. Rajagopalan,§,* Arnold M. Raitsimring,† and
John H. Enemark†,*
Department of Chemistry, 1306 E UniVersity BlVd, UniVersity of Arizona, Tucson,
Arizona 86721-0041, and Department of Biochemistry, Box 3711, Duke UniVersity Medical
Center, Durham, North Carolina 27710
Received February 29, 2008; E-mail: raj@biochem.duke.edu; jenemark@u.arizona.edu
Abstract: Electron paramagnetic resonance (EPR) investigation of the Mo(V) center of the pathogenic
R160Q mutant of human sulfite oxidase (hSO) confirms the presence of three distinct species whose relative
abundances depend upon pH. Species 1 is exclusively present at pH e 6, and remains in significant amounts
even at pH 8. Variable-frequency electron spin echo envelope modulation (ESEEM) studies of this species
prepared with 33S-labeled sulfite clearly show the presence of coordinated sulfate, as has previously been
found for the “blocked” form of Arabidopsis thaliana at low pH (Astashkin, A. V.; Johnson-Winters, K.;
Klein, E. L.; Byrne, R. S.; Hille, R.; Raitsimring, A. M.; Enemark, J. H. J. Am. Chem. Soc. 2007, 129,
14800). The ESEEM spectra of Species 1 prepared in 17O-enriched water show both strongly and weakly
magnetically coupled 17O atoms that can be assigned to an equatorial sulfate ligand and the axial oxo
ligand, respectively. The nuclear quadrupole interaction (nqi) of the axial oxo ligand is substantially stronger
than those found for other oxo-Mo(V) centers studied previously. Additionally, pulsed electron-nuclear
double resonance (ENDOR) measurements reveal a nearby weakly coupled exchangeable proton. The
structure for Species 1 proposed from the pulsed EPR results using isotopic labeling is a six-coordinate
Mo(V) center with an equatorial sulfate ligand that is hydrogen bonded to an exchangeable proton. Six-
coordination is supported by the 17O nqi parameters for the axial oxo group of the model compound,
(dttd)Mo17O(17Otms), where H2dttd ) 2,3:8,9-dibenzo-1,4,7,10-tetrathiadecane; tms ) trimethylsilyl.
Reduction of R160Q to Mo(V) with Ti(III) gives primarily Species 2, another low pH form, whereas reduction
with sulfite at higher pH values gives a mixture of Species 1 and 2, as well as the “primary” high pH form
of wild-type SO. The occurrence of significant amounts of the “sulfate-blocked” form of R160Q (Species 1)
at physiological pH suggests that this species may be a contributing factor to the lethality of this mutation.
Introduction
(hSO). One of the most extensively studied clinical mutations
that causes isolated sulfite oxidase deficiency is Arg 160 to Gln
Sulfite oxidase (SO) catalyzes the oxidation of sulfite to
sulfate as the terminal step in the metabolism of sulfur amino
acids and is vital for human health. Inherited mutations in SO
can result in severe neurological problems, stunted brain growth,
and early death.1 Investigation of the biochemistry of specific
mutations has been greatly aided by the development of
procedures for producing recombinant human sulfite oxidase
(R160Q),2–4 which results from a single base change (guanine
to adenine) in the gene for SO. No crystal structure is available
for hSO, but the structure of the highly homologous chicken
SO shows that R138 (equivalent to R160 in hSO) is in the
active-site pocket near the Mo center which exhibits ap-
proximately square-pyramidal five-coordinate geometry in the
fully oxidized Mo(VI) state, with axial and equatorial oxo
ligands and three equatorial sulfur ligands (two from the
pyranopterindithiolate cofactor and one from a cysteinyl side
chain).5 In the first step of the proposed catalytic cycle (Scheme
1, AfB) the equatorial oxo ligand reacts with sulfite to form
an enzyme-product complex in which sulfate is coordinated
to Mo(IV). Replacement of the product (sulfate) by water or
hydroxide (BfC) and subsequent one-electron oxidations
(Mo(IV/V/VI)) and concomitant deprotonations (CfDfA)
return the Mo center to the fully oxidized resting state.6–8
† University of Arizona.
§ Duke University Medical Center.
‡ Present address: College of Pharmacy, MSC09 5360, 1 University of
New Mexico, Albuquerque, New Mexico 87131-0001.
(1) Seidahmed, M. Z.; Alyamani, E. A.; Rashed, M. S.; Saadallah, A. A.;
Abdelbasit, O. B.; Shaheed, M. M.; Rasheed, A.; Hamid, F. A.; Sabry,
M. A. Am. J. Med. Genet. Part A 2005, 136A, 205.
(2) Garrett, R. M.; Johnson, J. L.; Graf, T. N.; Feigenbaum, A.;
Rajagopalan, K. V. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 6394.
(3) Johnson, J. L.; Coyne, K. E.; Garrett, R. M.; Zabot, M.-T.; Dorche,
C.; Kisker, C.; Rajagopalan, K. V. Hum. Mutat. 2002, 20, 74.
(4) Lam, C.-W.; Li, C.-K.; Lai, C.-K.; Tong, S.-F.; Chan, K.-Y.; Ng, G. S.-
F.; Yuen, Y.-P.; Cheng, A. W.-F.; Chan, Y.-W. Mol. Genet. Metab.
2002, 75, 91.
(5) Kisker, C.; Schindelin, H.; Pacheco, A.; Wehbi, W. A.; Garrett, R. M.;
Rajagopalan, K. V.; Enemark, J. H.; Rees, D. C. Cell 1997, 91, 973.
9
10.1021/ja801406f CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 8471–8480 8471

A R T I C L E S
Astashkin et al.
Scheme 1. Proposed transformations of the Mo active site for the fatal R160Q mutant of human SO.a
a Species 1, a “blocked” form with bound sulfate, is formed exclusively at pH e 6 and is the subject of this study. Species 2, formed by reduction with
Ti(III) at low pH, is also the major Mo(V) species produced by sulfite reduction at pH ∼7.2.13 Species 3 is formed at high pH and is analogous to the
“primary” high pH form observed in wild type SO.
Previous steady-state kinetic studies of the R160Q hSO
mutant showed a nearly 1000-fold decrease in kcat/Kmsulfite, which
suggested that the positive charge on Arg 160 in hSO makes
an important contribution to the binding of sulfite.2Flash
photolysis studies of intramolecular electron transfer (IET)
between the molybdenum and heme domains showed that IET
was also reduced about 1000-fold in R160Q compared to wild-
type hSO. IET activity was partially restored in the R160K
mutant, which has a positively charged lysine residue rather
than the neutral glutamine of the fatal R160Q mutant.9
The crystal structure of the molybdenum domain of chicken
R138Q shows that this mutation considerably reduces the size
of the binding pocket and changes its overall geometry.10 These
findings suggest that steric factors, in addition to a change in
the electrostatic potential at the active site, may play a role in
attenuation of activity in the R160Q mutant. Similar ideas were
entertained to explain different affinities of the Mo center to
nitrate and sulfate in a related molybdenum enzyme, nitrate
reductase.11,12 To complicate the situation, a recent EXAFS
investigation of the human R160Q mutant of SO suggested that,
in R160Q, the glutamine 160 coordinates to the (sixth) axial
position of the Mo center, trans to the oxo ligand.13 However,
no correlation between the proposed structural modification and
the pathological consequences was explicitly discussed in that
study.
(CW) EPR spectra show significant differences among organ-
isms, with pH, anions in the media and mutation of nearby
amino acid residues, even though all available X-ray crystal
structures reveal that the coordination geometries and structures
of their Mo centers are essentially conserved.5,14,15 The ap-
plication of variable-frequency pulsed EPR techniques allows
these qualitative observations to be rationalized in terms of local
structural variations at the Mo(V) center.16 Importantly, the
differences between the principal g-values of most of the Mo(V)
species formed in SO under various conditions are sufficiently
large that, even in the microwave (mw) X-band, the individual
SO forms present in a mixture can be studied selectively.17 In
addition, pulsed EPR is extremely useful for detecting the
protons in the coordination environment of Mo(V). These
protons are invisible to X-ray crystallography and EXAFS, but
important for understanding the detailed mechanism of SO and
its mutants. Finally, enrichment with magnetic isotopes (e.g.
17O, I ) 5/2; 33S, I ) 3/2) can be used to probe specific features
of the Mo(V) forms of Scheme 1.
Recently we carried out detailed pulsed EPR studies of the
structure of the Mo(V) forms of plant SO from Arabidopsis
thaliana (At-SO).6,18 These studies unambiguously demonstrated
that reduction by sulfite at low pH resulted in a “blocked” form
of the enzyme in which sulfate, the product, remained bound
to Mo(V).6 This “blocked” form is analogous to the phosphate-
and arsenate-bound species of vertebrate SO.19,20 This result
for At-SO raised the question as to whether “blocked” forms
also occur for hSO, particularly for mutants that show greatly
reduced catalytic activity. Here we present a pulsed EPR
investigation of the Mo(V) center of R160Q hSO. Both 17O-
EPR is a sensitive probe of Mo(V) and has long been used
to study sulfite-oxidizing enzymes (SOEs). The continuous wave
(6) Astashkin, A. V.; Johnson-Winters, K.; Klein, E. L.; Byrne, R. S.;
Hille, R.; Raitsimring, A. M.; Enemark, J. H. J. Am. Chem. Soc. 2007,
129, 14800.
(7) Brody, M. S.; Hille, R. Biochim. Biophys. Acta 1995, 1253, 133.
(8) Hille, R. Chem. ReV. 1996, 96, 2757.
(14) Kappler, U.; Bailey, S. J. Biol. Chem. 2005, 280, 24999.
(15) Schrader, N.; Fischer, K.; Theis, K.; Mendel, R. R.; Schwarz, G.;
Kisker, C. Structure 2003, 11, 1251.
(9) Feng, C.; Wilson, H. L.; Hurley, J. K.; Hazzard, J. T.; Tollin, G.;
Rajagopalan, K. V.; Enemark, J. H. Biochemistry 2003, 42, 12235.
(10) Karakas, E.; Wilson, H. L.; Graf, T. N.; Xiang, S.; Jaramillo-Busquets,
S.; Rajagopalan, K. V.; Kisker, C. J. Biol. Chem. 2005, 280, 33506.
(11) Field, S. J.; Thornton, N. P.; Anderson, L. J.; Gates, A. J.; Reilly, A.;
Jepson, B. J. N.; Richardson, D. J.; George, S. J.; Cheesman, M. R.;
Butt, J. N. Dalton Trans. 2005, 3580.
(16) Enemark, J. H.; Astashkin, A. V.; Raitsimring, A. M. Dalton Trans.
2006, 3501.
(17) Raitsimring, A. M.; Astashkin, A. V.; Feng, C. J.; Wilson, H. L.;
Rajagopalan, K. V.; Enemark, J. H. Inorg. Chim. Acta 2008, 361, 941.
(18) Astashkin, A. V.; Hood, B. L.; Feng, C. J.; Hille, R.; Mendel, R. R.;
Raitsimring, A. M.; Enemark, J. H. Biochemistry 2005, 44, 13274.
(19) Pacheco, A.; Basu, P.; Borbat, P.; Raitsimring, A. M.; Enemark, J. H.
Inorg. Chem. 1996, 35, 7001.
(12) Turner, N.; Ballard, A. L.; Bray, R. C.; Ferguson, S. Biochem. J. 1988,
252, 925.
(13) Doonan, C. J.; Wilson, H. L.; Rajagopalan, K. V.; Garrett, R. M.;
Bennett, B.; Prince, R. C.; George, G. N. J. Am. Chem. Soc. 2007,
129, 9421.
(20) George, G. N.; Garrett, R. M.; Graf, T.; Prince, R. C.; Rajagopalan,
K. V. J. Am. Chem. Soc. 1998, 120, 4522.
9
8472 J. AM. CHEM. SOC. VOL. 130, NO. 26, 2008

Molybdenum Center of Mutant Human Sulfite Oxidase
A R T I C L E S
Chart 1. The model complex (dttd)Mo17O(17Otms)
and 33S labeling have been used to gain insight into the
structures that are adopted by R160Q. Our results confirm that
at least three different Mo(V) species of R160Q exist as a
function of pH, as found recently by Doonan et al.13 We show
that a “blocked” form with bound sulfate (Scheme 1, Species
1) is the only species at pH e 6 and remains a significant form
at physiological pH. This form represents a catalytic dead end,
and we propose that it contributes to the lethality of the R160Q
mutation. Additionally, the comparison of the hyperfine (hfi)
and nuclear quadrupole interactions (nqi) of the oxo-17O ligand
with those known for other forms of SO and model oxo-Mo(V)
complexes support the hypothesis that Species 1 of R160Q SO
is six-coordinate.
(2) (dttd)Mo17O2. Mo17O2(acac)2 (25 mg, 0.076 mmol) was
dissolved in 2.0 mL of MeOH. To this was added dropwise a
solution of 2,3:8,9-dibenzo-1,4,7,10-tetrathiadecane (H2dttd) (24 mg,
0.077 mmol) and tert-butyl hydroperoxide (15 µL, 6 M in decane)
in 2.0 mL of CH2Cl2 while rapidly stirring. Stirring was continued
for 1 h at room temperature after the addition was complete. The
reaction mixture was then carefully heated to 40 °C until the volume
was reduced to 0.75 mL. Pentane (3 mL) was added to the resulting
orange-red solution, and the mixture was stirred for 5 h. The desired
product recrystallized from the solution as an orange-red powder
and was filtered off, washed with pentane, a minimal volume of
40% MeOH in Et2O, and then Et2O, and was then dried in Vacuo
(11 mg, 33%). Characterization for natural abundance (dttd)MoO2
prepared by the same procedure: m/z (relative intensity) 438 (100,
Materials and Methods
Enzyme Preparation. Recombinant R160Q hSO was expressed
and purified as previously described.21 The EPR spectra of the low
pH (lpH) (pH 5.8-6.0) forms of R160Q were obtained using sample
buffers of 50 mM Bis-Tris propane. The enzyme was reduced with
a 20-fold excess of sodium sulfite under argon and immediately
frozen in liquid nitrogen. The same buffer system and procedure
were used for reduction with 33S-labeled sulfite, prepared as
previously described.6 The lpH form of R160Q in H217O-enriched
water was prepared by first concentrating a 200 µL solution of the
enzyme in 25 mM Bicine and 25 mM Pipes at pH 6.2 to reduce
the amount of H216O. Next, a solution of 25 mM Bicine and 25
mM Pipes buffer was vacuum centrifuged to evaporate the H216O,
and the pelleted buffer was redissolved in the same volume of
H217O. The concentrated enzyme sample was then incubated in 30
µL of the buffer prepared with H217O for approximately 3 h at 4
°C. Finally, the enzyme was reduced with a 20-fold of excess
sodium sulfite under argon, and the samples were immediately
frozen in liquid nitrogen. Exchange into D2O was accomplished
by concentrating the protein samples to approximately 10 µL and
then diluting them to ∼300 µL with the appropriate buffer in D2O.
The procedure was repeated two times. The value of pD was
calculated as pDtrue ) pDapparent + 0.4.22 For Ti(III) citrate-reduced
lpH samples of R160Q in H2O or D2O, a stock solution of Ti(III)
citrate was diluted to ∼3.76 mM in 100 mM Bis-Tris propane (pH
6.0). Approximately 530 µM of R160Q in 50 mM Bis-Tris propane
was injected into an EPR tube, which was then placed inside a
Schlenk tube. The Schlenk tube was pumped to obtain mild vacuum
and then purged with argon. This procedure was repeated several
times over a period of 40 min to degas the protein. Next, 4 µL of
the diluted Ti(III) citrate solution (250 µM final concentration) in
well-degassed 100 mM Bis-Tris propane buffer (pH 6.0) was
injected into the protein solution under Ar, and the reduced R160Q
enzyme was frozen in liquid nitrogen immediately. The final ratio
of [R160Q SO]/[Ti(III) citrate] was approximately 2:1.
M+). UV-vis (DMF) [λmax, nm (ε, M-1 cm-1)]: 410 (5250). H
1
NMR (CDCl3, 300 MHz): 7.06-7.58 (m, 8H), 3.40 (d, J ) 10.8
Hz, 2H), 3.06 (d, J ) 10.8 Hz, 2H) ppm. Anal. Calcd for
C14H12MoO2S4: C, 38.53; H, 2.77; O, 7.33. Found: C, 38.69; H,
2.86; O, 7.16.
(3) (dttd)Mo17O(17Otms). (dttd)Mo17O2 (10 mg, 0.022 mmol)
was dissolved in 0.5 mL of CH2Cl2. To this was added a solution
of trimethylsilylsulfide ((tms)2S) (18 mg, 0.10 mmol, 21 µL) in
0.5 mL of CH2Cl2 while rapidly stirring. The reaction mixture
was stirred for 12 h during which time its color slowly changed
from orange-red to deep purple. The product was not isolated
from the reaction mixture. Characterization for natural abundance
(dttd)MoO(Otms) prepared by the same procedure: m/z (relative
intensity) 511 (100, M+). UV-vis (toluene) [λmax, nm (relative
intensity)]: 382 (100), 553 (30). CW-EPR (9.340 GHz, 298 K,
2.02 mW, 1.0 G ma): giso ) 1.96, a(95,97Mo) ) 44 G. Chart 1
shows a graphical representation of (dttd)Mo17O(17Otms).
EPR Measurements. The CW EPR experiments were performed
on a Bruker ESP-300 X-band spectrometer at 77 K. The electron spin
echo envelope modulation (ESEEM) and pulsed electron-nuclear
double resonance (ENDOR) experiments were performed on home-
built X/Ku-band (8-18 GHz)24 and Ka-band (26-40 GHz)24,25-pulsed
EPR spectrometers. The exact microwave (mw) frequencies, νmw, for
specific experiments are indicated in the Figure legends. The measure-
ment temperature in the pulsed experiments was about 20 K.
Model System Synthesis. (1) Mo17O2(acac)2. This precursor
was prepared following a slightly modified published procedure.23
Na2MoO4 (84.2 mg, 0.409 mmol) was dissolved in 500 µL of H217O
(80.9 atom%, Isotec). The solution was stirred for 15 h. Freshly
distilled 2,4-pentanedione (148 µL, 1.47 mmol) was added directly
to the rapidly stirring solution, followed immediately by 6.0 M HCl
(∼120 µL, prepared by diluting concentrated reagent grade HCl
by 50% with H217O of the same isotope concentration as above)
slowly and dropwise until the solution pH ) 2, resulting in the
precipitation of the product. The reaction mixture was stirred for
24 h. The precipitate was filtered off, washed generously with water
(natural abundance), pentane, and then Et2O to give the desired
product as a tan powder.
The numerical simulations were done using the SimBud soft-
ware.24 In these simulations the orientations of the hyperfine
interaction (hfi) and nuclear quadrupole interaction (nqi) tensors
with respect to the g-frame (x,y,z) were described by the sets of
Euler angles (ꢀh,θh,Ψh) and (ꢀq,θq,Ψq), respectively. The angles
ꢀh, θh and Ψh describe three consecutive rotations of the hfi
reference frame (1,2,3) from the original state, when it coincided
with the g-frame (1 // x, 2 // y, 3 // z): (1) by an angle ꢀh around
axis 3, (2) by an angle θh around the new axis 2, and (3) by an
(21) Temple, C. A.; Graf, T. N.; Rajagopalan, K. V. Arch. Biochem.
Biophys. 2000, 383, 281.
(22) Glasoe, P. K.; Long, F. A. J. Phys. Chem. 1960, 64, 188.
(23) Dowerah, D.; Spence, J. T.; Singh, R.; Wedd, A. G.; Wilson, G. L.;
Farchione, F.; Enemark, J. H.; Kristofzski, J.; Bruck, M. J. Am. Chem.
Soc. 1987, 109, 5655.
(24) http://quiz2.chem.arizona.edu/epr.
(25) Astashkin, A. V.; Enemark, J. H.; Raitsimring, A. M. Concepts Magn.
Reson., Part B 2006, 29B, 125.
9
J. AM. CHEM. SOC. VOL. 130, NO. 26, 2008 8473

A R T I C L E S
Astashkin et al.
Figure 1. Traces 1-4, CW EPR spectra of sulfite-reduced R160Q SO at
Figure 2. Trace 1, CW EPR spectrum of sulfite-reduced R160Q SO at pH
6.0 (same as trace 1 in Figure 1). Traces 2 and 3, individual spectra
contributing (along with spectrum 1) to the CW EPR spectra of sulfite-
reduced R160Q SO at higher pH (see traces 2-4 in Figure 1). Trace 4,
CW EPR spectrum of Ti(III) citrate-reduced R160Q SO at pH 7.0.
Experimental conditions are the same as those for Figure 1.
pH 6.0, 7.2, 8.1, and 8.9, respectively. Experimental conditions: νmw
)
9.465 GHz, mw power, 0.2 mW; modulation amplitude, 0.4 mT; temper-
ature, 77 K.
angle Ψh around the new axis 3. For the nqi reference frame (X,Y,Z)
similar definitions apply.
Table 1. Principal g-Values of EPR-Active Mo(V) Species
Observed in R160Q SO in Comparison with Those Found in SO
Enzymes from Other Sources
Experimental Results
1. EPR and ESE Field Sweep Spectra. The EPR spectra
observed for R160Q SO recorded at different pHs are shown
in Figure 1. Reduction by sulfite at low pH (pH 6) gives a single
EPR spectrum, as shown by trace 1 in Figure 1. At higher pH
(pH 7-9.5) the EPR spectra showed the presence of several
species. In all cases these spectra could be decomposed into
three distinct spectra shown by traces 1-3 in Figure 2. For
convenience, the forms of the Mo(V) active centers giving rise
to these signals are denoted as Species 1-3, with the numbering
corresponding to that in Figure 2. The principal g-values for
these three Species are summarized in Table 1. For comparison,
Table 1 also gives the principal g-values of the “primary” low-
and high pH forms established previously by Bray et al. for
wild-type chicken SO,26–28 and those of the recently observed
“blocked” (SO42--coordinated) form of At-SO.6,18 The notation
for the principal g-values follows the convention: gz g gy g
gx, as indicated in Figure 2 for Species 1.
From Table 1 it appears that Species 3 is similar to “primary”
hpH SO. However, Species 1 and 2 are clearly different from
all of the other forms. Their most striking difference is the
unusually small value of gx compared to the wild-type (wt)
enzyme. These observations are in agreement with the results
of another recent work on R160Q SO by Doonan et al.,13 who
also found three different EPR spectra that were denoted as low-
pH type 1 and type 2, and high-pH. In contrast, wt hSO exhibits
only two different Mo(V) EPR spectra as a function of pH,
which are the “primary” low-pH and high-pH spectra.29
The CW-EPR spectra of Species 1 and 2 show no splittings
from nearby exchangeable protons and thus cannot provide
Mo(V) species
gz
gy
gx
reference
Species 1
2.006
1.971
1.951
this work
(observed at low pH ∼ 6.0)
2.0061 1.9708 1.9514 13
1.998 1.961 1.939 this work
1.9978 1.9604 1.9391 13
1.988 1.964 1.953 this work
1.9875 1.9636 1.9526 13
2.0037 1.972 1.9658 26, 27
Species 2
(observed at higher pH)
Species 3
(observed at higher pH)
low-pH form of SO (chicken)
high-pH form of SO (chicken) 1.9872 1.9641 1.9531 26, 27
“blocked” form of At-SO 2.005 1.974 1.963 6, 18
information about the nature of the exchangeable equatorial
ligand in these forms. Our previous studies with other types of
SO preparations have shown that two possibilities may be
operative. One possibility is that a coordinated OH group is
present in the equatorial plane of the Mo(V) center, but the O-H
bond is at a large angle with the plane of Mo(V) dxy orbital that
carries the unpaired electron. As a result, the isotropic hfi for
this exchangeable proton becomes close to zero. Under such
conditions, the presence of the equatorial OH ligand is only
evident from a slight broadening at the EPR turning points
resulting mostly from the anisotropic hfi. However, this proton
was readily detectable by ESEEM and pulsed ENDOR, and
detailed spectroscopic information was obtained from these
pulsed EPR measurements. In particular, the anisotropic hfi
constant was found to be statically distributed in rather broad
limits as a result of the presence of various orientations of the
OH ligand relative to the Mo(V) center.31 This situation occurs
(26) Bray, R. C. In Biological Magnetic Resonance; Berliner, L. J., Reuben,
J., Eds.; Plenum Press: New York, 1980; Vol. 2, p 45.
(27) Lamy, M. T.; Gutteridge, S.; Bray, R. C. Biochem. J. 1980, 185, 397.
(28) Bray, R. C.; Gutteridge, S.; Lamy, M. T.; Wilkinson, T. Biochem. J.
1983, 211, 227.
(29) Astashkin, A. V.; Raitsimring, A. M.; Feng, C.; Johnson, J. L.;
Rajagopalan, K. V.; Enemark, J. H. Appl. Magn. Reson. 2002, 22,
421.
(30) Codd, R.; Astashkin, A. V.; Pacheco, A.; Raitsimring, A. M.; Enemark,
J. H. J. Biol. Inorg. Chem. 2002, 7, 338.
9
8474 J. AM. CHEM. SOC. VOL. 130, NO. 26, 2008

Molybdenum Center of Mutant Human Sulfite Oxidase
A R T I C L E S
in wt hpH SO and SDH [see e.g., refs 32–34] and may also be
the case for R160Q SO. Another possibility is that, similar to
sulfite-reduced lpH At-SO,6 Species 1 and/or 2 represent a
“blocked” form, with the equatorial ligand being SO42- instead
of -OH.
One experiment that can be used to address this dilemma is
to reduce the enzyme directly to Mo(V) with Ti(III) citrate,30
a
one-electron reductant, rather than through a pathway involving
two-electron reduction by sulfite followed by subsequent one-
electron oxidation to Mo(V). For At-SO at low pH, reduction
with Ti(III) citrate resulted in formation of the “primary” lpH
form instead of the “blocked” sulfate-coordinated form that was
observed for a sulfite-reduction pathway.18 In the present case
of R160Q SO, reduction by Ti(III) citrate at pH 7.0 resulted in
formation of a pure signal of Species 2 (trace 4 in Figure 2).
Since there was no sulfite in this sample, it is clear that Species
2 does not represent a sulfate-coordinated form and, most likely,
has an OH group as an exchangeable ligand.
If the Ti(III) reduction was performed at still lower pH (pH
e 6), then the EPR signal was a mixture of those of Species 2
and the “primary” (wt-like) signal of lpH SO (see Supporting
Information). The main purpose of this work is to characterize
Species 1. The full characterization of Species 2 will be the
subject of a separate investigation.
Figure 3. Traces 1 and 2 are, respectively, Ka-band and Ku-band two-
pulse ESEEM spectra (cosine FT) of lpH 33S-R160Q SO (Species 1)
obtained at gz. Experimental conditions for trace 1: νmw ) 29.472 GHz; Bo
) 1050.8 mT; mw pulses, 2 × 13 ns. Experimental conditions for trace 2:
νmw ) 17.269 GHz; Bo ) 618 mT; mw pulses, 2 × 10 ns. The temperature
was 20 K for both traces.
The fact that the reduction of Mo(VI) by means of Ti(III)
does not result in formation of Species 1 implies that this species
might well represent a sulfate-coordinated (“blocked”) form of
SO. In order to definitively prove this, high-resolution pulsed
EPR experiments have been performed. These experiments have
simultaneously addressed the nature of the exchangeable ligand
and the general nuclear surroundings of the Mo(V) center.
Specifically, the unusual values of the components of the g
tensor of Species 1 could be a reflection of some identifiable
structural differences between Species 1 and other forms of SO.
2. ESEEM Experiments with 33S-Enriched Samples. The
“blocked” form can be easily identified by ESEEM experiments
on a sample prepared with 33S-enriched sulfite. In this way we
have recently unequivocally shown that the Mo(V) center in
At-SO at low pH is “blocked” with a coordinated sulfate ligand.6
Similar experiments were performed in this work. The samples
of R160Q SO reduced by 33S-enriched sulfite (99% 33S, I )
3/2) will be denoted 33S-R160Q SO, while those prepared using
sulfite with the natural abundance of sulfur isotopes (94.93%
32S, I ) 0) will be denoted 32S-R160Q SO.
the hfi parameters expected for 33S of coordinated sulfate and
have shown that the optimal νmw for ESEEM experiments in
such a system should be between 16 and 24 GHz. The resonance
magnetic fields Bo corresponding to these mw frequencies
provide for the Zeeman/hfi cancelation condition, νS ≈ A/2
(where νS is the Zeeman frequency of 33S and A is the hfi
constant), which results in the largest ESEEM amplitude. Since
our spectrometers do not cover the range of νmw between 18
and 26 GHz, the actual experiments on 33S-At-SO were
performed at νmw ≈ 29 GHz (Ka-band) in order to provide the
so-called weak interaction regime (νS > A/2), which usually
simplifies the qualitative analysis of the ESEEM spectra. Most
of the experiments in this work were also performed at νmw
≈
29 GHz. For comparison, we have also performed some of the
measurements in Ku band (at νmw ≈ 17 GHz) and found that
the ESEEM amplitudes in both mw bands are comparable.
As an example, Figure 3 shows Ka- and Ku-band two-pulse
ESEEM spectra of Species 1 of 33S-R160Q SO obtained at gz.
All of the lines observed in these spectra within the frequency
window shown were missing in the case of 32S-R160Q SO,
which allows one to assign them to the interaction of Mo(V)
with 33S. The feature of interest in these spectra is the peak at
18 MHz that belongs to one of the interdoublet transitions (| (
1/2 T | ( 3/2 ) of 33S within the electron spin manifold where
the Zeeman and hyperfine interactions approximately cancel
each other. The frequency of this transition is:
In our previous work, we have discussed the choice of an
optimal mw operational frequency, νmw, to facilitate the
observation of 33S ESEEM from the coordinated sulfate.6 By
using an analogy with the phosphate- and arsenate-coordinated
Mo(V) centers of SO studied earlier,19,20 we have estimated
(31) Astashkin, A. V.; Mader, M. L.; Pacheco, A.; Enemark, J. H.;
νid ≈ e2Qq ⁄ 2h
(1)
Raitsimring, A. M. J. Am. Chem. Soc. 2000, 122, 5294. After the
publication of this paper, we have often encountered a misinter-
18,31–33
pretation of our results
as indicating a free rotation of the
and we can easily estimate the quadrupole coupling constant
e2Qq/h ≈ 36 MHz, very similar to e2Qq/h ≈ 40 MHz found
for 33S-At-SO.6
Panels a, c, and e in Figure 4 show Ka-band hyperfine sublevel
correlation (HYSCORE) spectra obtained for Species 1 of 33S-
R160Q SO at the EPR turning points. The correlation peaks
seen in these spectra are absent for the sample of 32S-R160Q
SO, which makes their assignment to 33S straightforward. While
these spectra look very simple and similar to those of a system
with I ) 1/2, this similarity is actually superficial because 33S
OH ligand around the Mo-O bond. Therefore, we stress here that
this idea was never entertained by us because at the measurement
temperature of∼20 K dynamic rotation is not plausible. Rather,
our results are based upon a static structural distribution of the
orientations of the OH ligand. The word“rotation”(around the Mo-
O bond) has entered that description to specify the geometrical
transformation relating all the different structural conformations.
(32) Astashkin, A. V.; Raitsimring, A. M. J. Magn. Reson. 2000, 143, 280.
(33) Raitsimring, A. M.; Pacheco, A.; Enemark, J. H. J. Am. Chem. Soc.
1998, 120, 11263.
(34) Raitsimring, A. M.; Kappler, U.; Feng, C.; Astashkin, A. V.; Enemark,
J. H. Inorg. Chem. 2005, 44, 7283.
9
J. AM. CHEM. SOC. VOL. 130, NO. 26, 2008 8475

A R T I C L E S
Astashkin et al.
Table 2. Setsa of hfi Constants A ) (Ax, Ay, Az) Estimated from
Analysis of HYSCORE Spectra at gx, gy, and gz, and the Isotropic
hfi Constants aiso and the Anisotropic hfi Constants T ) (Tx, Ty,
Tz) Estimated from A
set
(Ax, Ay, Az) [MHz]
aiso [MHz]
(Tx, Ty, Tz) [MHz]
I
II
((0, 4.5, 3.4)
((0, 4.5, -3.4)
(2.63
(0.37
((-2.63, 1.87, 0.77)
((-0.37, 4.13, -3.77)
a The sets differ by the choice of relative signs of different
components of A.
was greater than 2, it was reasonable to assume that the direction
of the nqi axis Y was approximately parallel to the axis of gx,
in which case the rhombicity of the nqi tensor could be estimated
as η ≈ 0.2 (see published expressions).6
As a result of the simulations, the following set of parameters
was obtained: aiso ) 2.1 MHz; T ) (-4.1, 2.5, 1.6) MHz;
(ꢀh,θh,Ψh) ) (20°, 10°, 20°), e2Qq/h ) 36 MHz; η ) 0.2;
(ꢀq,θq,Ψq) ) (90°, 90°,0°). The simulated spectra are shown
in panels b, d, and f of Figure 4. These parameters are rather
similar to those obtained earlier for 33S-At-SO, and a nonzero
isotropic hfi constant is within the limits estimated from known
hfi parameters for 31P in coordinated phosphate and 75As in
coordinated arsenate.19,20 The main qualitative conclusion from
the observed significant 33S hfi (both isotropic and anisotropic)
is that Species 1 of R160Q SO represents a “blocked” form,
with sulfate coordinated to the Mo(V) center.
Figure 4. Panels a, c, and e, the (++) quadrants of the HYSCORE spectra
of 33S-R160Q SO (Species 1) obtained at the gz, gy, and gx EPR turning
points, respectively (Bo ) 1060.8, 1078.8, and 1088.6 mT, respectively).
The spectra shown represent sums of the spectra obtained at time intervals
between the first and second mw pulses τ ) 170 and 200 ns. Other
experimental conditions: νmw ) 29.650 GHz; mw pulses, 11, 11, 21, and
11 ns; temperature, 20 K. Panels b, d, and f, simulated HYSCORE spectra
for gz, gy, and gx at the EPR turning points, respectively. Simulation
parameters: aiso ) 2.1 MHz, anisotropic hfi tensor in the principal axes
system, (T11, T22, T33) ) (-4.1, 2.5, 1.6) MHz; e2Qq/h ) 36 MHz; η )
0.2; Euler angles for the orientation of the hfi tensor in the g-frame: ꢀh )
20°, θh ) 10°, Ψh ) 20°; Euler angles for the orientation of the nqi tensor
with respect to the g-frame: ꢀq ) 90°, θq ) 90°, Ψq ) 0°. The simulated
spectra (as experimental ones) represent sums of the spectra calculated for
τ ) 170 and 200 ns.
3. 1H- and 2H-Pulsed ENDOR Experiments. The pulsed
ENDOR experiments on samples of Species 1 of R160Q SO
prepared in H2O and D2O (at pH e 6) have revealed the
presence of an exchangeable proton in close proximity to the
Mo center, but with hfi parameters that are quite different from
those of the “primary” hpH and lpH forms. As an example,
Figure 5 shows the 1H refocused (Re) Mims ENDOR35,36 spectra
obtained as a difference between the normalized (by ESE signal
amplitude) spectra recorded for the samples in H2O and D2O.
For brevity, these spectra will be referred to as the difference
1H ENDOR spectra. The original experimental spectra recorded
for several EPR positions are shown in the Supporting Informa-
tion.
has spin I ) 3/2 and strong nqi (e2Qq/h . νS, A). The theory
for interpreting this kind of spectra was developed in our
previous work.6 According to that theory, the pairs of crosspeaks
seen in HYSCORE spectra belong to |1/2 T |-1/2 transitions
of 33S within the |R and |ꢁ electron spin manifolds. The
frequencies of these transitions, νR and νꢁ, are approximately
equal to:
The largest splitting of ∼5 MHz is observed at gy, which
indicates that the closest distance to an exchangeable proton
from the Mo(V) center is RMoH ≈ 3 Å (as estimated using the
point-dipole approximation, with the spin population on Mo(V)
[
]
νR,ꢁ ) chq · (νS ( A ⁄ 2 )
(2)
where chq is a numerical factor that depends on the orientation
of the magnetic field vector, Bo, relative to the nqi principal
axes frame.
1
of 0.85). Numerical simulations of the difference H ENDOR
spectra were performed under the assumption that the exchange-
able proton of interest is responsible for the exterior features in
all spectra, while the interior features are also contributed by
more distant exchangeable protons. As a result of these
simulations, the isotropic hfi constant, aH, was found to be about
zero, and the anisotropic hfi tensor was found to be rhombic,
with the principal components (T11, T22, T33) ≈ (-3.0, -1.9,
4.9) MHz. The orientation of the hfi tensor with respect to the
g-frame was described by the Euler angles (ꢀh, θh, Ψh) ) (75°,
80°, 0°). The simulations with axial anisotropic hfi tensor did
not produce satisfactory results.
The approximate positions of the correlation peaks in panels
a, c, and e of Figure 4 are respectively (2.9, 8.6), (1.8, 8.1),
and (7.9, 7.9) [MHz], which give the corresponding center
frequencies νc ) (νR+νꢁ)/2 of 5.75, 4.95, and 7.9 MHz. At the
same time, the Zeeman frequency of 33S is νS ≈ 3.5 MHz (varies
from 3.47 MHz at the low-field EPR turning point to 3.56 MHz
at the high-field turning point). From these positions the scaling
factors chq ) νc/νS can be estimated as 1.66, 1.40, and 2.22,
which give the estimates for the hfi constants A ∼ 3.4, 4.5, and
0 MHz. The signs of the effective hfi constants are as yet
undefined, and two possible sets of the hfi constants are shown
in Table 2.
The preliminary simulations of two-pulse ESEEM spectra
similar to those described in our previous work showed that
the experimental ESEEM amplitude can only be reproduced with
the hfi parameters close to those of Set I,6 and therefore this set
was used as a starting approximation in the HYSCORE
simulations. Since the largest chq was obtained for gx, and it
2
Figure 6 shows H Mims ENDOR37 spectra for the sample
prepared in D2O. Deuterons have spin I ) 1 and rather weak
nqi that results in extra splittings in the ENDOR spectra
1
1
(compared to the I ) /2 case of H) which complicates their
(35) Astashkin, A. V.; Kawamori, A.; Kodera, Y.; Kuroiwa, S.; Akabori,
K. J. Chem. Phys. 1995, 102, 5583.
(36) Doan, P. E.; Hoffman, B. M. Chem. Phys. Lett. 1997, 269, 208.
(37) Mims, W. B. Proc. R. Soc. London A 1965, 283, 452.
9
8476 J. AM. CHEM. SOC. VOL. 130, NO. 26, 2008

Molybdenum Center of Mutant Human Sulfite Oxidase
A R T I C L E S
Figure 5. Solid traces, 1H ENDOR spectra at the EPR turning points (traces
1-3) obtained as a difference between the Re-Mims 1H ENDOR spectra
recorded for Species 1 of R160Q in H2O and D2O. Experimental conditions:
Bo ) 1051.3 mT (gz), 1069.4 mT (gy) and 1079.1 mT (gx) for traces 1, 2,
and 3, respectively; νmw ) 29.470 GHz; mw pulses, 3 × 14 ns + 25 ns;
time interval between the first and second mw pulses, τ ) 80 ns; time
interval between the second and third mw pulses, T ) 30 µs; time interval
between the third and fourth (refocusing) mw pulses, τ′ ) 380 ns; RF pulse
length, 20 µs; temperature, 20 K. Dashed lines, numerical simulations for
aiso ) 0; ∆aiso ) 0.5 MHz; T ) (-3.0, -1.9, 4.9) MHz; (ꢀh,θh,Ψh) )
(75°, 80°, 0°).
Figure 7. HYSCORE spectrum of Species 1 in 17O-R160Q SO showing
the lines of oxo-17O in (++) quadrant and equatorial 17O in (+-) quadrant.
Experimental conditions: νmw ) 29.530 GHz; Bo ) 1070 mT (gy); mw
pulses, 10, 10, 20, and 10 ns; time interval between the first and second
mw pulses, τ ) 180 ns; temperature, 20 K.
axis with respect to the g-frame. A reasonable fit of the
experimental spectra was obtained for e2Qq/h ) 0.2 MHz and
(ꢀq, θq, Ψq) ) (30°, -45°, 0°), see Figure 6.
4. ESEEM Experiments with 17O-Enriched Samples. The
ESEEM experiments with samples prepared with water enriched
with 17O (denoted for brevity 17O-R160Q SO) have revealed
the presence of 17O in both equatorial and axial (oxo) positions.
As an example, Figure 7 shows a HYSCORE spectrum of 17O-
R160Q SO (Species 1) obtained at gy. The (++) quadrant of
this spectrum shows two off-diagonal lines centered at the
Zeeman frequency of 17O, νO ≈ 6 MHz. The splitting between
these lines is about 5.5 MHz, which gives an estimate of the
hfi constant similar to that observed earlier for the oxo-17O in
chicken SO38 and a model oxo-molybdenum complex.39
The (+-) quadrant shows a pair of lines due to the equatorial
17O ligand. The splitting between these lines is equal to 2νO ≈
12 MHz, and the center of the doublet is located at the frequency
of about 9 MHz, which gives an estimate of the hfi constant A
≈ 18 MHz. This hfi constant is significantly lower than the
values for the lpH, hpH, and phosphate forms of wt chicken
SO that have a hydroxyl ligand in the equatorial position.40–44
Figure 6. Solid traces, Mims 2H ENDOR spectra for Species 1 of R160Q
in D2O obtained at the EPR turning points. Experimental conditions: Bo )
1056.5 mT (gz), 1074.6 mT (gy) and 1084.3 mT (gx) for traces 1, 2, and 3,
respectively; νmw ) 29.613 GHz; mw pulses, 3 × 14 ns; time interval
between the first and second mw pulses, τ ) 400 ns; time interval between
the second and third mw pulses, T ) 32 µs; RF pulse length, 26 µs;
temperature, 20 K. Dashed traces, simulated numerically with parameters:
aiso ) 0; ∆aiso ) 0.077 MHz; T ) (-0.46, -0.29, 0.75) MHz; (ꢀh, θh,
Ψh) ) (75°, 80°, 0°), e2Qq/h ) 0.2 MHz; η ) 0; (ꢀq, θq, Ψq) ) (30°,
-45°, 0°).
(38) Astashkin, A. V.; Feng, C.; Raitsimring, A. M.; Enemark, J. H. J. Am.
Chem. Soc. 2005, 127, 502.
analysis. The situation is, however, mitigated by the facts that
we have already obtained the hfi estimates from the difference
1H ENDOR spectra, and that the nqi tensor of deuterons is close
(39) Astashkin, A. V.; Neese, F.; Raitsimring, A. M.; Cooney, J. J. A.;
Bultman, E.; Enemark, J. H. J. Am. Chem. Soc. 2005, 127, 16713.
(40) Gutteridge, S.; Lamy, M. T.; Bray, R. C. Biochem. J. 1980, 191, 285.
(41) Bray, R. C.; Gutteridge, S. Biochemistry 1982, 21, 5992.
(42) Gutteridge, S.; Malthouse, J. P. G.; Bray, R. C. J. Inorg. Biochem.
1979, 11, 355.
2
to axial (η ≈ 0). Therefore, in the H ENDOR simulations we
have used the hfi parameters rescaled according to the ratio of
2H and 1H magnetic moments, and varied the quadrupole
(43) Morpeth, F. F.; George, G. N.; Bray, R. C. Biochem. J. 1984, 220,
235.
2
coupling constant, e2Qq/h, and the orientation of the H nqi
9
J. AM. CHEM. SOC. VOL. 130, NO. 26, 2008 8477

A R T I C L E S
Astashkin et al.
Figure 9. 17O sum combination line region of τ-integrated four-pulse
ESEEM spectrum (cosine FT) of (dttd)Mo17O(17Otms) obtained at gz.
Experimental conditions: νmw ) 29.804 GHz; Bo ) 1081.5 mT; mw pulses,
18, 18, 36, and 18 ns; starting time interval between the first and second
mw pulses, τ ) 180 ns; temperature, 20 K.
Figure 8. 17O sum combination line region of τ-integrated four-pulse
ESEEM spectrum (cosine FT) of Species 1 of 17O-R160Q SO obtained at
gz. Experimental conditions: νmw ) 29.530 GHz; Bo ) 1052 mT; mw pulses,
10, 10, 20, and 10 ns; starting time interval between the first and second
mw pulses, τ ) 180 ns; temperature, 20 K.
9 shows the integrated four-pulse ESEEM spectrum obtained
at gz, where the largest quadrupolar splitting of the sum
combination line is observed. Additional experimental data for
this compound can be found in the Supporting Information.
The observation of the equatorial 17O in the ESEEM spectra of
Species 1 of 17O-R160Q implies that both oxo ligands were
exchanged with 17O during incubation of the Mo(VI) form of
the enzyme and before reduction with sulfite.
Discussion
To determine the quadrupole coupling constant of the oxo-
17O, the τ-integrated four-pulse ESEEM has been used.32,45
Previously we have successfully used this technique to obtain
spectra of the sum combination line of the oxo-17O that showed
resolved splittings due to the weak 17O nqi.39 Figure 8 shows
the sum combination line region of the τ-integrated four-pulse
ESEEM spectrum of 17O-R160Q SO obtained at gz, where the
quadrupole splitting of the sum combination line was the largest,
∆νσ ≈ 1.5 MHz. From this splitting, the quadrupole coupling
constant of the oxo-17O can be estimated as e2Qq/h ≈ 5 MHz
(e2Qq/h ≈ (40/12) × ∆νσ, see the relevant theory elsewhere39).
For comparison, the quadrupole coupling constant determined
earlier for [Mo17O(SPh)4]-,39 and hpH chicken SO16 was much
smaller, e2Qq/h ≈ 1.5 MHz.
In the experiments with 33S-R160Q SO we have established
that Species 1 represents a “blocked” form of SO, with sulfate
coordinated to the Mo center. Compared with the “blocked”
form reported for At-SO, however, Species 1 of R160Q SO
has a more rhombic g-factor and an exchangeable proton situated
close to the Mo(V) center. The nqi of the oxo ligand in 17O-
R160Q SO has more than tripled from the values observed for
the model oxomolybdenum complex [Mo17O(SPh)4]-,39 and
hpH SO.16 In contrast, the hfi of this ligand is approximately
the same as in other systems studied so far. Possible structural
reasons for these findings are discussed below.
The anisotropic hfi of the exchangeable proton, while weaker
than that generally observed for the proton of an OH ligand,33,34
has significant rhombicity. This rhombicity can only be ex-
plained by an off-axis interaction with a spin population other
than that located on the Mo center. Such a situation arises, e.g.,
for the proton of the OH ligand that interacts both with the large
spin population on Mo (FMo ≈ 0.85) and with the small spin
population delocalized to the OH ligand oxygen (FO ≈ 0.05).
For R160Q SO such a model would be invalid because the
exchangeable ligand of the Mo center is, as we established, a
sulfate group. However, the rhombicity of the anisotropic hfi
tensor of the proton in question can be explained by considering
the possible hydrogen-bonding interactions. One possibility is
a hydrogen bond to the oxygen atom of sulfate that is
coordinated to the Mo center; a second possibility is a hydrogen
bond to the axial oxo ligand.
The estimation of the anisotropic hfi for the first model
(sp3-hybrid valence orbitals of oxygen, Mo-O distance of
∼2.2 Å and the O · · · H hydrogen bond length of ∼1.7 Å, the
angle between Mo-H and Mo-O ≈ 30°) results in an
anisotropic hfi tensor T ) (-2.9, -2, 4.9) MHz, in qualitative
agreement with the experimental data, T ) (-3.0, -1.9, 4.9).
For the second model (ModO · · · H, FO ≈ -0.05)39, an
anisotropic hfi tensor similar to the experimental one can only
be obtained at a bonding angle of ∼30° from the ModO
direction. Such an angle would indicate a significant change
of the hybridization of the oxo ligand orbitals resulting in
significant s-character of the oxygen orbital that forms a
5. ESEEM Experiments with 17O-Enriched Model Compound
(dttd)Mo17O(17Otms). In addition to R160Q SO, we have also
performed integrated four-pulse ESEEM experiments on the
synthetic model compound (dttd)Mo17O(17Otms) shown in
Structure 1. The primary goal of these experiments was to obtain
hfi and nqi parameters for an axial oxo ligand in a six-coordinate
molecule, as discussed below. The (dttd)Mo17O(17Otms) is well
suited for obtaining 17O parameters because the other coordina-
tion sites of Mo(V) are occupied by sulfur atoms (94.93% 32S,
I ) 0). The HYSCORE and 2D refocused primary (RP) ESEEM
measurements did not reveal the interaction with the oxygen of
the equatorial 17Otms ligand. This may be caused by the
insufficient 17O enrichment (estimated ∼10% from the 17O
modulation amplitude) in combination with the generally greater
difficulty of observing a strongly coupled equatorial oxygen.
The axial oxo-17O was, however, readily observed. The HY-
SCORE and RP ESEEM experiments have shown that the hfi
constant for this oxygen is similar to that in SO and the model
compound studied earlier (∼5 MHz).38,39 The four-pulse experi-
ment has, however, revealed that the quadrupole coupling
constant for this compound, e2Qq/h ≈ 3 MHz, is about twice
that in hpH SO and in [Mo17O(SPh)4]-. As an example, Figure
(44) Cramer, S. P.; Johnson, J. L.; Rajagopalan, K. V.; Sorrell, T. N.
Biochem. Biophys. Res. Commun. 1979, 91, 434.
(45) Van Doorslaer, S.; Schweiger, A. Chem. Phys. Lett. 1997, 281, 297.
9
8478 J. AM. CHEM. SOC. VOL. 130, NO. 26, 2008

Molybdenum Center of Mutant Human Sulfite Oxidase
A R T I C L E S
2
σ-bond with the Mo dz orbital. This essentially nonzero
s-character would lead to a dramatic increase of the isotropic
17O hfi constant. Since the experimental 17O hfi constant is
similar to that found for other systems,38,39 we discard this
model as implausible.
Let us now consider the nqi parameters of the 17O oxo ligand.
A single electron in a p-orbital of 17O produces a nqi tensor
with e2Qq/h ≈ 20 MHz (and two electrons on the same orbital
will result in e2Qq/h ≈ 40 MHz); however, six electrons in three
mutually perpendicular p-orbitals give e2Qq/h ) 0 MHz (the
result of cubic symmetry). If each of these p-orbitals donates
exactly the same electronic population to the bonding orbitals
of the Mo(V), then the oxo-17O nqi will still be zero. The fact
that the experimental nqi is very small (e2Qq/h , 40 MHz)
indicates that the electronic donations from the p-orbitals of
17O are almost perfectly balanced. Indeed, in the detailed study
of the axially symmetric model complex, [MoO(SPh)4]-, it was
found that δpx ) δpy ) δpz ( 0.07.39
Unless a significant change in geometry of the molybdenum
coordination occurs in R160Q SO, the dramatic increase in the
quadrupole coupling constant of its axial oxo ligand relative to
other recently studied systems requires alternative explanations.
One possibility is a modification of the oxo ligand valence
orbitals through asymmetric hydrogen bonding. Such an interac-
tion would destroy the delicate balance between the electric field
gradients produced by the p-orbitals of the oxo group. However,
such a hydrogen-bonding possibility has already been viewed
as unlikely in the above discussion of the weakly coupled
exchangeable proton observed for R160Q SO.
Figure 10. Envisioned structure of the Mo center of Species 1 of R160Q
SO resulting from the pulsed EPR experiments. “X” is a putative axial ligand
coordinated trans to the oxo ligand.
enzyme and model systems. The change in the equatorial ligands
can alter the electronic donation from the oxo ligand to dxz and/
or dyz orbitals of the Mo(V) center, which could result in a
different quadrupole coupling constant. However, the oxo-17O
nqi constants of [Mo(V)17O(SPh)4]- and the Mo(V) center of
hpH SO are very similar even though their equatorial ligands
are different. Likewise, the oxo-17O nqi constants of the lpH
and hpH forms of wt SO are similar even though the orientations
of the OH ligand in the two forms of SO are different, which
should result in different overlaps of the OH oxygen orbitals
with the dxz and dyz orbitals of molybdenum. Therefore, while
we cannot completely rule out the possibility that the change
of the oxo-17O nqi in R160Q SO is simply caused by
the presence of a sulfate ligand, instead of a hydroxyl, in the
coordination sphere of the Mo(V) center, the above examples
make this explanation less plausible than that based on axial
coordination. Further insight into these nqi variations will result
from the detailed quantum chemical calculations on these
systems that are in progress.
Alternatively, the change in the oxo-17O nqi could reflect a
decrease or increase in the electronic donations from some of
the oxygen orbitals without modification of their geometry. This
purely electronic effect could be caused by the weak axial
coordination of a sixth ligand to Mo(V), trans to the oxo ligand,
in the R160Q protein. Coordination of a sixth ligand would
involve the same Mo(V) orbitals used in coordination of the
oxo ligand and may thereby change the electronic donation
pattern from the oxo ligand, which in turn may result in a
different oxo-17O nqi. In the recent investigation of R160Q SO
by XAS and EXAFS the hypothesis of a six-coordinate structure
for its Mo center has been also employed.13
In order to validate the above considerations, we have
performed 17O ESEEM experiments on the model six-coordinate
Mo(V) complex, (dttd)Mo17O(17Otms), which has a thioether
sulfur donor trans to the terminal oxo ligand (Structure 1). The
quadrupole coupling constant of the oxo-17O determined for this
compound is ∼3 MHz, approximately twice that found for the
oxo-17O ligands in [Mo(V)17O(SPh)4]- and the Mo(V) center
of hpH SO. This result demonstrates that axial coordination trans
to the oxo-17O produces an observable increase in the quadru-
pole coupling constant. In addition, the nqi for (dttd)-
Mo17O(17Otms) provides strong, if indirect, support for the
hypothesis that the increase of the quadrupole coupling constant
of the oxo-17O in Species 1 of R160Q SO (compared to the
systems studied earlier) is caused by axial coordination trans
to the oxo ligand, as proposed by Doonan et al.13 Detailed
quantum chemical calculations addressing the 17O hfi and nqi
parameters in (dttd)Mo17O(17Otms) and R160Q SO will be
presented elsewhere.
Conclusion
Three different paramagnetic species of R160Q SO can be
generated, depending upon pH and the mode of reduction
(Scheme 1). Species 1, which is the only one present at pH e
6, is the primary focus of this report. Species 1 contains
coordinated sulfate, as clearly demonstrated by 33S ESEEM
measurements. Moreover, the 33S paramenters for Species 1 are
similar to those previously found for the bound sulfate of the
“blocked” form of At-SO.6 Species 1 also has a nearby
exchangeable proton that is likely to be hydrogen-bonded to an
oxygen of the sulfate ligand. A proposed structure for the Mo(V)
center of Species 1 in R160Q SO based on the experimental
results obtained in this work and the above considerations
regarding the oxo-17O nqi is shown in Figure 10. The proposed
weakly coordinated sixth ligand (X) that is trans to the terminal
oxo ligand is consistent with the larger oxo-17O quadrupole
coupling constant found for the known six-coordinate 17O-
labeled compound, (dttd)Mo17O(17Otms). Two possibilities for
the trans ligand X (O or N) were discussed recently from
EXAFS and DFT studies on R160Q SO.13 From our studies,
only X ) O remains plausible because no 14N modulations,
which are easily detectable by ESEEM, were observed.
The structure of Figure 10 can be viewed as an enzyme-
product complex that is trapped in the Mo(V) oxidation state,
e.g. the hypothesized Mo(IV)-sulfate enzyme-product com-
plex of Scheme 1 has been oxidized to Mo(V) before the
substrate was hydrolyzed. As we already mentioned (see
Figure 1), the fraction of Species 1 present decreases with
increasing pH;13 however, Species 1 remains a significant
Yet another factor, in addition to six-coordination, that could
produce the unusual quadrupole coupling constants found for
the oxo-17O ligands of R160Q SO and (dttd)Mo17O(17Otms)
are the equatorial ligands, which differ from those in previous
9
J. AM. CHEM. SOC. VOL. 130, NO. 26, 2008 8479

A R T I C L E S
Astashkin et al.
Mo(V) form at pH 7.213 and is still observable at pH 8.9.
Thus, it is reasonable to speculate that the “blocked” structure
of R160Q (Figure 10) represents a catalytic dead end that
contributes to the lethality of this mutant under physiological
conditions. The generation of “blocked” forms of SO,
especially at low pH, appears to be an important feature in
the biochemistry of other mutants of hSO, and the catalyti-
cally compromised Y343F mutant of hSO17,47 also exhibits
a “blocked” species with bound sulfate.46
Acknowledgment. We gratefully acknowledge support of this
research by the NIH (GM-37773 to J.H.E., GM 44283 to K.V.R.)
and by Grants from the NSF (DBI-0139459, DBI-9604939, BIR-
9224431) and the NIH (S10RR020959) for development of the
pulsed EPR facility. K.J.-W. thanks the NIH for a Ruth L.
Kirschstein National Service Award.
Supporting Information Available: CW EPR spectra of
1
R160Q SO reduced by Ti(III) citrate at pH 5.8; H Re-Mims
ENDOR spectra of Species 1 of R160Q SO; additional pulsed
EPR data for (dttd)Mo17O(17Otms). This material is available
free of charge via the Internet at http://pubs.acs.org.
(46) Enemark, J. H.; Astashkin, A. V.; Raitsimring, A. M.; Johnson-Winters,
K.; Klein, E. L.; Byrne, R. S.; Hille, R.; Wilson, H. L.; Rajagopalan,
K. V. J. Biol. Inorg. Chem. 2007, 12 (Suppl. 1), S63.
(47) Feng, C.; Wilson, H. L.; Hurley, J. K.; Hazzard, J. T.; Tollin, G.;
Rajagopalan, K. V.; Enemark, J. H. J. Biol. Chem. 2003, 278, 2913.
JA801406F
9
8480 J. AM. CHEM. SOC. VOL. 130, NO. 26, 2008