Enantioselective sulfoxidations catalyzed by horseradish peroxidase, manganese peroxidase, and myeloperoxidase

Article abstract of DOI:10.1007/s007060070097

Horseradish peroxidase (HRP), myeloperoxidase (MPO), and manganese peroxidase (MnP) have been shown to catalyze the asymmetric sulfoxidation of thioanisole. When H₂O₂ was added stepwise to MPO, a maximal yield of 78% was obtained at

Full text of DOI:10.1007/s007060070097

Monatshefte fuÈr Chemie 131, 687±695 (2000)
Enantioselective Sulfoxidations Catalyzed
by Horseradish Peroxidase, Manganese
Peroxidase, and Myeloperoxidase
Antonin Tuynman1, Hans E. Schoemaker2, and Ron Wever1;Ã
1
E.C. Slater Institute, BioCentrum, University of Amsterdam, NL-1018 TV Amsterdam, The
Netherlands
2
DSM Research, Bio-Organic Chemistry, NL-6160 MD Geleen, The Netherlands
Summary. Horseradish peroxidase (HRP), myeloperoxidase (MPO), and manganese peroxidase
(MnP) have been shown to catalyze the asymmetric sulfoxidation of thioanisole. When H2O2 was
added stepwise to MPO, a maximal yield of 78% was obtained at pH 5 (ee 23%), whereas an
optimum in the enantiomeric excess (32%, (R)-sulfoxide) was found at pH 6 (60% yield). For MnP a
yield of 18% and a high enantiomeric excess of 91% of the (S)-sulfoxide were obtained at pH 5 and a
yield of 36% and an ee of 87% at pH 7.0. Optimization of the conversion catalyzed by horseradish
peroxidase at pH 7.0 by controlled continuous addition of hydrogen peroxide during turnover and
monitoring the presence of native enzyme as well as of intermediates I, II, and III led to the formation
of the sulfoxide in high yield (100%) and moderate enantioselectivity (60%, (S)-sulfoxide).
Keywords. Enantioselective sulfoxidation; Horseradish peroxidase; Myeloperoxidase; Manganese
peroxidase.
Introduction
General
The use of enzymes in oxidative conversions is a potentially attractive method for the
synthesis of optically active compounds. In this way oxidative processes employing
stoichiometric amounts of heavy metal salts can be avoided, whereas the high regio-
and/or enantioselectivity of enzymatic processes is a promising feature which may be
exploited in various ways. Other oxidative bioconversions involve the use of whole
cells, making the whole process rather cumbersome and being therefore often
restricted to conversions leading to compounds with a high added value like
pharmaceuticals [1]. Peroxidases, however, might be potentially attractive biocata-
lysts for ®ne chemicals production since cosubstrate recycling procedures are not
necessary and hydrogen peroxide is used as a clean and cheap oxidant.
Various peroxidases have been shown to catalyze the formation of a number of
alkyl arylsulfoxides or dialkylsulfoxides in an enantioselective manner [2±5]. The
Ã
Corresponding author

688
A. Tuynman et al.
compounds are useful intermediates in organic synthesis, both as chiral auxiliaries in
catalytic processes and as stoichiometric intermediates. Also epoxidations [6, 7],
benzylic [8] and propargylic [9] hydroxylation, and indole oxidations [10] are
catalyzed by peroxidases. The potential application in organic synthesis is still
hampered by the moderate stability of the peroxidases under turnover conditions, the
limited solubility of organic reactants in water, and the oxidative inactivation of the
heme. The catalytic performance of peroxidases may be improved by continuous
controlled addition of hydrogen peroxide [11, 12], by variation of pH [4], and by the
addition of some organic solvents [11]. We recently have shown that under carefully
choosen conditions also lactoperoxidase (a commercially available enzyme) is
capable of enantioselective sulfoxidation of thioanisole, affording the (R) isomer in a
yield of 85% with an ee of 80% [5]. We have also shown that the opposite antipode,
the (S)-sulfoxide, could be obtained in a yield of 84% with an ee of 73% using the
commercially available fungal peroxidase from Coprinus cinereus. In addition, we
have described a spectroscopic method to monitore the active enzyme intermediates
during turnover. In this way we were able to optimize the conditions for maximum
yield and optimum e.e. [5].
The aim of the present research was twofold. First we investigated if we also
could use our spectroscopic method to further optimize the sulfoxidation of
thioanisole to form the (S)-sulfoxide in high yield and high ee using horseradish
peroxidase. Second, we investigated two other peroxidases (myeloperoxidase and
manganese peroxidase) to see if these enzymes could also be used in enantioselective
sulfoxidation reactions. Myeloperoxidase was expected to give the (R)-sulfoxide due
to its structural ressemblance with lactoperoxidase [13]. Moreover, in analogy to the
heme CPO it is able to oxidize chloride to hypochlorous acid [14]. The manganese
peroxidase is also a fungal enzyme like CiP, and it was expected that with this
enzyme ± the mnp-gene has recently been cloned and expressed [15] ± we would
obtain the (S)-sulfoxide.
Mechanistic considerations
The general mechanism of peroxidases starts by the addition of H2O2 to the
enzyme in its resting state, the native Fe3‡ state [16]. Upon release of a molecule of
‡Á
water, an oxoferryl species (Fe(IV) ˆ O P ) called compound I is created
containing two oxidative equivalents, one of which is located in the porphyrin or on
a protein residue as a radical-cation. The substrate is oxidized to a substrate radical
in this step. Addition of an equivalent of a traditional peroxidase substrate, a one-
electron donor, will reduce compound I to a Fe(IV) ˆ O species that is one
oxidative equivalent above the resting state, the so-called compound II, which does
not contain an electron hole an either porphyrin or protein anymore. A second
equivalent of substrate will reduce this species back to the native enzyme upon the
release of a water molecule and a second substrate radical. The enzyme can also
return directly from compound I back to the native enzyme upon reaction with a
second molecule of H2O2, releasing a molecule of oxygen and a molecule of water.
This is called the catalase activity of the enzyme. Another way for the direct
reconversion of compound I to the native enzyme proceeds via oxygen-transfer to a
suitable substrate such as a sul®de or an alkene (Scheme 1). This is called the oxo-

Peroxidase-Catalyzed Enantioselective Sulfoxidations
689
Scheme 1. Oxene and oxygen-rebound mechanisms for the oxygen transfer to sul®des; the formation
of compound III under a large excess of H2O2 is indicated as well as the direct racemic reaction
between H2O2 and the sul®de
ferryl or oxene mechanism [17]. Alternatively, oxygen-transfer may take place via
a two-step mechanism: in this so-called oxygen-rebound mechanism [17±22] a
molecule of substrate is ®rst oxidized by compound I to a substrate radical-cation
that forms a complex with compound II (compound II Á S‡Á Scheme 1). Subsequently
the oxygen of compound II is transferred to this substrate radical-cation, a
molecule of oxygenated product is released, and the enzyme returns to its native
state. In the case of an oxygen-rebound mechanism the rate of transition from
compound I to compound II should depend linearly on the sul®de concentration,
whereas in an oxene mechanism the formation of compound II should not be
observed. As we have shown, lactoperoxidase and myeloperoxidase operate via an
oxygen-rebound mechanism [5].
At high H2O2 concentrations a third intermediate of the enzyme, compound III,
can be formed (Scheme 1). In most reactions this intermediate is believed to be a
catalytic inactive, mostly a dead-end species and leading to inactivated enzyme for
most peroxidases [23]. However, compound III is not necessarily irreversible
formed and may return to the native state [23]. Under operational conditions,

690
A. Tuynman et al.
involving discontinuous addition of hydrogen peroxide [3], compound III may well
be transiently formed. It is of importance to avoid the presence of compound III,
since it is indicative of a high H2O2 concentration [5]. This will lead to an
enhanced racemic direct reaction between the sul®de and H2O2 and therefore lower
the ee of the sulfoxides formed [5].
Results and Discussion
Several studies have been published on the catalysis of asymmetric sulfoxidation
by horseradish peroxidase [3, 4, 24±28]. Yet there is no agreement on the best
method to perform these incubations, and therefore the yield and the enantiomeric
excess of the sulfoxide formed differ from report to report. This may also be due to
the experimental conditions used in these studies. Differences include amount of
enzyme, concentration of the sul®de, open reaction vessels vs. closed ones, size
headspace to prevent evaporation of the substrate, pH values [3, 4, 24±28], and the
way H2O2 is added (in one step [24] or stepwise [3, 4, 25±28]).
For horseradish peroxidase, we have attempted to optimize the sulfoxidation
with respect to the yield whilst maintaining a high enantioselectivity. Therefore, the
method previously described in detail for CiP and LPO [5] of continuously adding
H2O2 and simultaneously monitoring the enzyme intermediates, in particular
compound III, was applied. This method guarantees that no excessive accumulation
of H2O2 will occur. The rate of addition of H2O2 is to be lowered when a small
amount of compound III is formed, thus minimizing the direct racemic reaction
between H2O2 and the sul®de. The reactions were carried out at pH ˆ 7.0, since at
that pH value the highest ee for this conversion had been reported. Initially it was
found that by incubating 19.2 mM HRP with 1.7 mM thioanisole at pH 7.0 at a H2O2
in¯ux of 1 mmol/h, after 5 hours a small amount of compound III could be detected
in the mixture of compound I and II. Therefore, the rate of H2O2 addition was
lowered to 0.4 mmol/h, and consequently only compound I and II were present. The
presence of a mixture of compounds I and II is indicative of an oxygen-rebound
mechanism which could be con®rmed by stopped-¯ow studies [21, 22]. None-
theless, during the incubations the absorbance of the Soret band decreased. About
20 mM enzyme were necessary to fully convert the substrate before the enzyme was
inactivated.
After 48 hours all of the thioanisole was converted into the sulfoxide with an
enantioselectivity of 60% of the (S)-enantiomer. Variation of the substrate or the
enzyme concentration did not lead to an altered ee provided that the reaction was
quenched as soon as compound III started to be formed.
It is intriguing that the enantioselectivity does not change when the substrate
concentration is changed. Kobayashi et al. have shown that 10% of the oxygen
incorporated in the sul®de derives from water [17]. It is believed that the
incorporation of an oxygen atom from water occurs when a sul®de radical-cation
recombines with a second molecule of radical-cation to give a molecule of sul®de
and a sul®de dication; the latter will react with water. The fact that the ee is
independent on the sul®de concentration implies that these processes take place in
the active site. Indeed, the stopped-¯ow experiments of Dunford et al. are indicative
of the presence of two molecules of substrate in or near the active site [22].

Peroxidase-Catalyzed Enantioselective Sulfoxidations
691
From the decrease in the absorption at 280 nm we determined an initial
1
turnover frequency of 0.3 min . This is signi®cantly lower than the turnover
1
frequency of 3 min determined by Savenkova et al. [29, 30]. The turnover
number obtained with our optimization technique is 52, which is signi®cantly
higher than those previously reported [3, 4]. It must be mentioned that we did not
use sul®de concentrations exceeding the solubility limits and that the reaction
vessels were completely ®lled and sealed to prevent evaporation of the sul®de.
Using less favourable conditions resulted in lower yields and/or lower ees.
When the reactions were carried out by stepwise addition of H2O2 and were
quenched after one hour [25±28], the yields did not exceed 11% (results not
shown). After quenching, the enzyme is still active, and higher yields could have
been obtained if the procedure would have been carried out for a longer time. In
Refs. [25±28] higher enantioselectivities were claimed, but the amont of sulfoxides
formed in a parallel reaction without enzyme was subtracted. This is not correct
since the H2O2 concentration in the enzyme catalyzed reaction is much lower than
in the absence of enzyme due to its catalase activity. In a later publications of the
same group the subtraction of the racemic parallel reaction has been omitted,
leading to enantioselectivities of 58% [29, 30].
When the reactions were carried out as described by Morishima et al. [24], i.e.
by a single addition of 1.0 mM H2O2 at the start of the reaction, a yield of only 4%
of the (S)-sulphoxide was obtained in one hour with an ee of only 52%. The low ee
is probably due to a signi®cant contribution of the direct reaction. Figure 1 shows
that using this method the conversion of thioanisole only takes place during the ®rst
Fig. 1. Spectra of the enzyme intermediate of HRP in the sulfoxidation of thioanisole after 10, 130,
250, and 490 s (ÐÐ); hereafter, the enzyme starts to return to its native state, the spectrum of which
is represented after one hour (- - - -); in the insert the conversion of thioanisole is shown by the
decrease in absorption at 280 nm; see experimental procedures for details

692
A. Tuynman et al.
Fig. 2. pH dependence of the formation of (R)-methyl phenyl sulfoxide and its ee by MPO (O: ee,
&: yield); see experimental procedures for details
500 seconds as judged from the decrease in absorption at 280 nm (insert). The
following slower increase in absorption is probably due to conversion of compound
I or II into native enzyme: the enzyme is mostly present as compound II and starts
to return to its native state after the ®rst 500 seconds.
In conclusion, ee and yield of the sulfoxidation of thioanisole by HRP depend
strongly on the conditions and the design of the experiment. It was possible to
convert the thioanisole completely into the sulfoxide using a slow continuous
in¯ux of H2O2 (on average 0.46 mmol/h) for 48 h. This in¯ux is considerably
slower than that for the same reaction catalyzed by LPO and CiP [5]. Continuous
addition of H2O2 only is not suf®cient to obtain a high ee it is also imperative to
follow the enzyme intermediates spectroscopically to prevent the accumulation of
H2O2 and the formation of catalytically inactive species as e.g. compound III.
The sulfoxidation of thioanisole by MPO was carried out via the traditional
method of adding H2O2 stepwise. The pH dependence of the enantioselectivity and
the yield are presented in Fig. 2. Although the yield is maximal at pH 5 (78%), the
best enantioselectivity is obtained at pH 6.0 (32%). At higher pH values the
enantiomeric excess remains fairly constant. The yield decreases showly at
increasing pH values. These enantioselectivities are much higher than previously
reported (4±8%) for the sulfoxidation of p-methylthioanisole by MPO [31]. In
these experiments, much lower enzyme and substrate concentrations have been
used. In agreement with Ref. [31] the rate for this reaction is maximal at pH 5.0.
For MPO we have not optimized the yield via our method of controlled addition of
H2O2 and monitoring at the same time the enzyme intermediates present. The high
yields obtained employing the traditional stepwise addition of H2O2 are in contrast
to our ®ndings with other heme peroxidases.
Here we also report for the ®rst time the enantioselective sulfoxidation by MnP.
Duetoalimitedamountofpuri®edenzymeavailablewewereonlyabletocarryoutthe

Peroxidase-Catalyzed Enantioselective Sulfoxidations
693
experiment at pH 5.0 and 7.0. Best results were obtained at pH 5.0. Although the yield
was very low (18%), the enantioselectivity was very high (91% ee of (S)-sulfoxide).
This is the highest value reported for a heme peroxidase producing the (S)-sulfoxide
and signi®cantly higher than reported for structurally related enzymes [32] (CiP: 73%
ee HRP: 60% ee). At pH 7.0 the yield was considerably higher (36%), but the ee
dropped to 87%. CPO also produces the sulfoxidewith a high ee of even 99% [11, 12],
but in this case the (R)-enantiomer is formed.
Unfortunately, the very limited amount of MnP available did not permit us to
optimize this reaction with respect to pH value and controlled continuous addition
of H2O2. Small genetic modi®cations for MnP (such as the F41L mutation for HRP
[25]) might enhance the ee and make this enzyme one of its mutants a promising
catalyst for the production of (S)-sulfoxides.
Experimental
MPO was puri®ed from human leukocytes as described (A428nm/280nm ˆ 0.8) [33]. Enzyme
1
1
concentrations were determined using a molar extinction coef®cient of 102 mM Á cm
at
1
1
1
1
403 nm for HRP [34], 89 mM Á cm at 428 nm for MPO [33], and 127 mM Á cm at 406 nm for
MnP [35]. Absorption spectra in the UV/Vis range were recorded on a Hewlett Packard 8452 A
spectrophotometer. Figure 3 shows the spectra of native MPO, MPO-I, MPO-II, and MPO-III. The
spectra of native HRP, HRP-1, HRP-II, and HRP-III are shown in Fig. 4.
To determine the pH optimum for the sulfoxidation of thioanisole by MPO, reactions were
carried out in 100 mM buffer at 25ꢀC. Buffers used were sodium acetate ( pH ˆ 4.0 and 5.0),
potassium phosphate ( pH ˆ 6.0±8.0), and sodium carbonate ( pH ˆ 9.0).
The reaction mixture contained 10 mM MPO and 1 mM thioanisole. During one hour, 20Â10 mm3
of 10 mM H2O2 were added. The same procedure was applied to the reaction with MnP as a catalyst,
however only at pH ˆ 5.0 and 7.0 due to the very limited amount of enzyme available. Kinetic
experiments following the sulfoxidation of thioanisole by HRP were conducted using a HP 8452 A
spectrophotometer at 280 nm. The difference of the molar extinction coef®cients of methyl phenyl
1
1
sul®de and the methyl phenyl sulfoxide was determined to be 0.84 mM Á cm at 280 nm. The
amounts of compound I, II, or III and of native enzyme were followed simultaneously. Hydrogen
Fig. 3. Optical absorption spectra of MPO, native enzyme, compound II, and compound III; ÐÐ:
2 mM MPO in 100 mM potassium phosphate ( pH 7.0), - - - -: MPO-II generated by addition of 40 mM
hydrogen peroxide to 2 mM MPO, Á Á Á Á Á Á: MPO-III generated by addition of 4 mM hydrogen
peroxide to 2 mM MPO

694
A. Tuynman et al.
Fig. 4. Optical absorption spectra of native HRP, HRP-I, HRP-II, and HRP-III; ÐÐ: 10 mM HRP in
100 mM potassium phosphate ( pH 6.5), - Á - Á - Á - Á - Á : HRP-I generated by addition of 10 mM
hydrogen peroxide, - - - - - -: HRP-II generated by addition of 10 mM ferrocyanide to HRP-I, Á Á Á Á Á:
HRP-III generated by addition of a large excess hydrogen peroxide (2.5 mM)
peroxide was added continuously via a syringe pump (Cole Parmer 74900-10) with a 250 mm3
Hamilton syringe with a te¯on luer lock connected to the reaction cuvet via a PEEK tubing
1/1600OD/0.2000ID that went through a capillary in the te¯on cap that sealed the cuvet.
For HRP, the average rate of hydrogen peroxide addition was 0.46 mmol/h. Typically, reactions
were carried out in a 1.67 cm3 quartz cuvet sealed with a te¯on cap with two capillaries: one to add
the hydrogen peroxide solution and one to dispose of the over¯ow. The cuvets were completely ®lled
with the reaction mixture in order to prevent partitioning of the methyl phenyl sul®de into a gas-
phase headspace. The contents of the cuvet were stirred continuously. General conditions were
10 mM enzyme, 100 mM potassium phosphate buffer, and 0.15±1.7 mM methyl phenyl sul®de.
The reactions were quenched with sodium sul®te [3] to consume the excess of H2O2 after 1 h for
MPO and MnP and for HRP when the sulfoxidation process had ended as judged from a constant
absorbance at 280 nm.
1 mmol of acetophenone was added as internal standard, and the reaction mixture was extracted
twice with 3.4 cm3 CH2Cl2. The organic layer was concentrated under a stream of nitrogen to a
volume of about 20 mm3 and diluted with 1 cm3 of 80% hexane/20% isopropyl alcohol. A 20 mm3
sample was loaded onto a chiracel OD HPLC column (Daicel Chemical Industries, 0.46 cmÂ25 cm)
equipped with a Pharmacia LKB-HPLC pump 2248 and a LKB Bromma 2140 rapid spectral detector
connected to a PC. The BorwinTM program was used to evaluate peak areas. The (R)- and (S)-methyl
phenyl sulfoxides were eluted isocratically with the same solvent mixture at a ¯ow rate of 0.5 ml/min
and detected at 254 nm. Retention times of methyl phenyl sul®de, acetophenone, and the (R)- and
(S)-methyl phenyl sulfoxides were 8.4, 9.4, 15.0, and 17.8 min, respectively [5].
H2O2 solutionswere freshly prepared bydilutionof a 30%stocksolution(Merck). The concentration
1
1
was determined spectrophotometrically using an absorption coef®cient of 43.6 M Á cm at 240 nm
[36]. All other chemicals were of the highest purity. Methyl phenyl sul®de was purchased from Fluka,
methyl phenyl sulfoxide and acetophenone from Aldrich.
Acknowledgements
We wish to thank Prof. M. H. Gold from the Oregon Graduate Institute of Science and Technology
for a gift of manganese peroxidase. This work was supported by the BioCentrum and the Innovation
Oriented Research Programs Catalysis (IOP catalysis). We also received ®nancial support from the
Council for Chemical Sciences of the Netherlands Organisation for Scienti®c Research (NWO) and
the Netherlands Technology Foundation (STW).

Peroxidase-Catalyzed Enantioselective Sulfoxidations
695
References
[1] Holland HL (1992) In: Holland HL (ed) Organic synthesis with oxidative enzymes. VCH, New
York
[2] Colonna S, Gaggero N, Manfredi A, Casella L, Gulotti M, Carrea G, Pasta P (1990) Biochemistry
29: 10465
[3] Colonna S, Gaggero N, Carrea G, Pasta P (1992) J Chem Soc Chem Commun 357
[4] Colonna S, Gaggero N, Richelmi C, Carrea G, Pasta P (1995) Gaz Chim Ital 125: 479
[5] Tuynman A, Vink MKS, Dekker HL, Schoemaker HE, Wever R (1998) Eur J Biochem 258: 906
[6] Allain AJ, Hager LP, Deng L, Jacobsen EN (1996) J Am Chem Soc 117: 4415
[7] Tuynman A, Lutje Spelberg J, Kooter IM, Schoemaker HE, Wever R (2000) J Biol Chem 275:
3025
[8] Miller VP, Tschirret-Guth, RA, Ortiz de Montellano PR (1995) Arch Biochem Biophys 319: 333
[9] Hu S, Hager LP (1999) J Am Chem Soc 121: 872
[10] Corbett MD, Chipko BR (1979) Biochem J 183: 269
[11] Van Deurzen MPJ, Remkes IJ, van Rantwijk F, Sheldon RA (1997) J Mol Cat A: Chemical 117:
329
[12] Van Deurzen MPJ, Seelbach K, van Rantwijik F, Kragl U, Sheldon RA (1997) Biocat Biotrans
15: 1
[13] De Gioia L, Ghibaudi EM, Laurenti E, Salmona M, Ferrari RP (1996) J Bioinorg Chem 1: 476
[14] Klebanoff SJ (1968) J Bacteriol 95: 2131
[15] May®eld MB, Kishi K, Alic M, Gold MH (1994) Appl Environ Microbiol 60: 4303
[16] Ortiz de Montellano PR (1992) Annu Rev Pharmacol Toxicol 32: 89
[17] Kobayashi S, Nakano M, Goto T, Kimura T, Schaap AP (1986) Biochem Biophys Res Commun
135: 166
[18] Kobayashi S, Nakano M, Kimura T, Schaap PA (1987) Biochemistry 26: 5019
[19] Casella L, Gullotti M, Ghezzi R, Poli S, Beringhelli T, Colonna S, Carrea G (1992) Biochemistry
31: 9451
[20] Baciocchi E, Lanzalunga O, Malandrucco S (1996) J Am Chem Soc 118: 8973
[21] Perez U, Dunford HB (1990) Biochem Biophys Acta 1038: 98
[22] Perez U, Dunford, HB (1990) Biochemistry 29: 2757
[23] Huwiler M, Jenzer H, Kohler H (1986) Eur J Biochem 158: 609
[24] Tanaka M, Ishimori K, Mukai, M, Kitagawa T, Morishima I (1997) Biochemistry 36: 9889
[25] Ozaki S, Ortiz de Montellano PRO (1994) J Am Chem Soc 116: 4487
[26] Harris RZ, Newmyer SL, Ortiz de Montellano PR (1993) J Biol Chem 268: 1637
[27] Ozaki S, Ortiz de Montellano PR (1995) J Am Chem Soc 117: 7056
[28] Newmyer SL, Ortiz de Montellano PR (1995) J Biol Chem 270: 19430
[29] Savenkova MI, Ortiz de Montellano PR (1998) Arch Biochem Biophys 351: 286
[30] Savenkova MI, Ortiz de Montellano PR (1998) Biochemistry 37: 10828
Á
[31] Capeillere-Blandin C, Martin C, Gaggero N, Pasta P, Carrea G, Colonna S (1998) Biochem J
335: 27
[32] Baunsgaard L, Dalbùge H, Houen G, Rasmussen EM, Welinder KG (1993) Eur J Biochem 213:
605
[33] Bakkenist ARJ, Wever R, Vulsma T, Plat H, van Gelder BF (1978) Biochim Biophys Acta 524: 45
[34] Schonbaum GR, Lo S (1972) J Biol Chem 247: 3353
[35] Millis CD, Cai D, Stankovich MT, Tien M (1994) Biochemistry 28: 8484
[36] Beers RF Jr and Sizer IW (1952) J Biol Chem 195: 133
Received November 18, 1999. Accepted January 21, 2000