Chromophoric spin-labeled beta-lactam antibiotics for ENDOR structural characterization of reaction intermediates of class A and class C beta-lactamases.

Article abstract of DOI:10.1016/j.saa.2003.10.024

The chromophoric spin-label substrate 6-N-[3-(2,2,5,5-tetramethyl-1-oxypyrrolin-3-yl)-propen-2-oyl]penicillanic acid (SLPPEN) was synthesized by acylation of 6-aminopenicillanic acid with the acid chloride of 3-(2,2,5,5-tetramethyl-1-oxypyrrolinyl)-2-propenoic acid and characterized by physical methods. By application of angle-selected electron nuclear double resonance (ENDOR), we have determined the molecular structure of SLPPEN in solution. SLPPEN exhibited UV absorption properties that allowed accurate monitoring of the kinetics of its enzyme-catalyzed hydrolysis. The maximum value of the (substrate-product) difference extinction coefficient was 2824 M(-1) cm(-1) at 275 nm compared to 670 M(-1) cm(-1) at 232 nm for SLPEN [J. Am. Chem. Soc. 117 (1995) 6739]. For SLPPEN, the steady-state kinetic parameters kcat and kcat/KM, determined under initial velocity conditions, were 637 +/- 36 s(-1) and 13.8 +/- 1.4 x 10(6) M(-1) s(-1), respectively, for hydrolysis catalyzed by TEM-1 beta-lactamase of E. coli, and 0.5 +/- 0.04 s(-1) and 3.9 +/- 0.4 x 10(4) M(-1) s(-1) for hydrolysis catalyzed by the beta-lactamase of Enterobacter cloacae P99. We have also observed burst kinetics for the hydrolysis of SLPPEN with P99 beta-lactamase, indicative of formation of an acylenzyme reaction intermediate. In DMSO:H2O (30:70, v:v) cryosolvent mixtures buffered to pH* 7.0, the half-life of the acylenzyme intermediate formed with the P99 enzyme at -5 degrees C was > or = 3 min, suitable for optical characterization. The observation of burst kinetics in the hydrolysis of SLPPEN catalyzed by P99 beta-lactamase suggests that this chromophoric spin-labeled substrate is differentially sensitive to active site interactions underlying the cephalosporinase and penicillinase reactivity of this class C enzyme.

Full text of DOI:10.1016/j.saa.2003.10.024

Spectrochimica Acta Part A 60 (2004) 1279–1289
Chromophoric spin-labeled ␤-lactam antibiotics for ENDOR
structural characterization of reaction intermediates of
class A and class C ␤-lactamases
Devkumar Mustafia,∗, Jennifer E. Hofera, Wanzhi Huangb,
Timothy Palzkillb, Marvin W. Makinena
a
Department of Biochemistry and Molecular Biology, Cummings Life Science Center, The University of Chicago,
920 East 58th Street, Chicago, IL 60637, USA
Department of Molecular Virology and Microbiology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA
b
Received 16 June 2003; received in revised form 13 July 2003; accepted 16 October 2003
Abstract
The chromophoric spin-label substrate 6-N-[3-(2,2,5,5-tetramethyl-1-oxypyrrolin-3-yl)-propen-2-oyl]penicillanic acid (SLPPEN) was syn-
thesized by acylation of 6-aminopenicillanic acid with the acid chloride of 3-(2,2,5,5-tetramethyl-1-oxypyrrolinyl)-2-propenoic acid and
characterized by physical methods. By application of angle-selected electron nuclear double resonance (ENDOR), we have determined the
molecular structure of SLPPEN in solution. SLPPEN exhibited UV absorption properties that allowed accurate monitoring of the kinetics of its
enzyme-catalyzed hydrolysis. The maximum value of the (substrate−product) difference extinction coefficient was 2824 M−1 cm−1 at 275 nm
compared to 670 M−1 cm−1 at 232 nm for SLPEN [J. Am. Chem. Soc. 117 (1995) 6739]. For SLPPEN, the steady-state kinetic parameters
kcat and kcat/KM, determined under initial velocity conditions, were 637 36 s−1 and 13.8 1.4 × 106 M−1 s−1, respectively, for hydrolysis
catalyzed by TEM-1 ␤-lactamase of E. coli, and 0.5 0.04 s−1 and 3.9 0.4 × 104 M−1 s−1 for hydrolysis catalyzed by the ␤-lactamase
of Enterobacter cloacae P99. We have also observed “burst kinetics” for the hydrolysis of SLPPEN with P99 ␤-lactamase, indicative of
formation of an acylenzyme reaction intermediate. In DMSO:H2O (30:70, v:v) cryosolvent mixtures buffered to pH∗ 7.0, the half-life of the
acylenzyme intermediate formed with the P99 enzyme at −5 ◦C was ≥3 min, suitable for optical characterization. The observation of burst
kinetics in the hydrolysis of SLPPEN catalyzed by P99 ␤-lactamase suggests that this chromophoric spin-labeled substrate is differentially
sensitive to active site interactions underlying the cephalosporinase and penicillinase reactivity of this class C enzyme.
© 2003 Elsevier B.V. All rights reserved.
Keywords: ␤-Lactamases; Electron nuclear double resonance; Electron paramagnetic resonance; Enzyme mechanism; Cryoenzymology; Nitroxyl spin-labels
1. Introduction
␤-Lactam resistant bacteria command important attention
because they constitute the second-most frequent source of
nosocomial infections [1,2]. Since the primary mechanism
of antibiotic resistance in these bacteria are hydrolytic en-
zymes known as ␤-lactamases, study of their catalytic mech-
anisms is of importance for antibacterial drug design. To
this end, it is important to understand not only the struc-
tural determinants of antibiotic substrate recognition among
the various classes of ␤-lactamases but also the structural
differences that underlie their mechanism of action. There
are four classes of ␤-lactamases, three of which function
via an active site serine residue (classes A, C, and D) while
the fourth (class B) is comprised by Zn2+-metalloenzymes
Abbreviations: DMSO, dimethylsulfoxide; ENDOR, electron nu-
clear double resonance; EPR, electron paramagnetic resonance; HEPES,
N-(2-hydroxyethyl)piperazine-Nꢀ-2-ethanesulfonic acid; hf, hyperfine;
hfc, hyperfine coupling; MeOH, methanol; SLCEP, 7-N-(2,2,5,5-
tetramethyl-1-oxypyrrolinyl-3-carboxyl)-cephalosporanic acid or spin-
labeled cephalosporin; SLPA, 3-(2,2,5,5-tetramethyl-1-oxypyrrolinyl)-
2-propenoic acid or spin-labeled propenoic acid; SLPEN, 6-N-
(2,2,5,5-tetramethyl-1-oxypyrrolinyl-3-carboxyl)-penicillanic acid or spin-
labeled penicillin; SLPPEN, 6-N-[3-(2,2,5,5-tetramethyl-1-oxypyrrolin-3-
yl)-propen-2-oyl]penicillanic acid or spin-labeled propenoyl penicillin;
U∗V, ultraviolet
Corresponding author. Tel.: +1-773-702-1667; fax: +1-773-702-0439.
E-mail address: dmustafi@midway.uchicago.edu (D. Mustafi).
1386-1425/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.saa.2003.10.024

1280
D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
[3,4]. Although structures of class A and class C enzymes
have been solved to high resolution [5–8], there is still
relatively little agreement about the chemical roles of ac-
tive site residues except for that of the nucleophilic serine
residue.
out the suggestion that the general base in acylation is the
active site Lys73 residue [16].
In earlier studies we employed spin-labeled penicillin and
cephalosporin derivatives in which the amino nitrogen was
derivatized with the 2,2,5,5-tetramethyl-1-oxypyrrolinyl-
3-formyl group [17,18]. While these spectroscopic probe
substrates are highly suitable for ENDOR structural charac-
terization, their ultraviolet absorption properties are not read-
ily employed for monitoring kinetics of enzyme-catalyzed
reactions. Here we report characterization of a spin-labeled
penicillin substrate in which the amino nitrogen is acylated
with 3-(2,2,5,5-tetramethyl-1-oxypyrrolinyl)-2-propenoic
acid [19]. By application of ENDOR to determine dipolar,
through-space electron–nucleus distances, we establish the
structure and conformation of 6-N-[3-(2,2,5,5-tetramethyl-1-
oxypyrrolin-3-yl)-propen-2-oyl]penicillanic acid (SLPPEN)
in solution. The conjugated, unsaturated double bond sys-
tem in this spin-labeled penicillin results in enhanced ul-
traviolet absorption compared to that of other spin-labeled
antibiotics employed earlier [17,18]. The ultraviolet absorp-
tion properties of SLPPEN not only allow facile monitoring
of the steady-state kinetics of its enzyme-catalyzed hy-
drolysis but also reveal “burst” kinetics in the hydrolytic
reaction that is catalyzed by the class C P99 ␤-lactamase
but not observed with the class A TEM-1 ␤-lactamase.
These different kinetic properties are likely due to active
site structural features distinguishing class A and class
C enzymes as penicillinases and cephalosporinases. Fur-
thermore, observation of “burst” kinetics, due to a much
slower rate of deacylation in the class C enzyme, allows
optical characterization of an acylenzyme reaction interme-
diate near 0 ◦C. ENDOR characterization of this reaction
intermediate, therefore, may reveal structural differences in
class A and class C ␤-lactamases that distinguish them as
penicillinases and cephalosporinases.
Virtually the same types of active site residues are found
in class A and class C ␤-lactamases with similar spatial
relationships to each other [5,6,9]. Nonetheless, consider-
able variability is found in the effectiveness with which
they hydrolyze ␤-lactam antibiotics. In general, steady-state
kinetic parameters show that class A enzymes are good
penicillinases while in contrast class C enzymes are good
cephalosporinases [5]. Since class A enzymes hydrolyze
penicillin substrates very efficiently, it is not clear why
cephalosporins are associated with lower kinetic speci-
ficity. For class C enzymes the reversed relationship is
observed. Because active site residues are responsible for
catalytic activity, differences in catalytic properties must
be due to subtle variations in their spatial relationships.
Since the catalytic properties of class A and class C en-
zymes differ markedly against the same substrates, it is
difficult to propose a common mechanism operative in
both. For these reasons, it is believed that the catalytic
machinery of the class A and class C enzymes is distinct
and that they function according to different mechanisms
[5,8–10].
We have shown through a series of studies published from
this laboratory that application of electron nuclear double
resonance (ENDOR) spectroscopy for structure analysis of
true, catalytically competent intermediates of enzymes cat-
alyzed reactions provides an incisive means to identify cat-
alytically important enzyme–substrate interactions that are
unique and underlie catalysis [11–14]. We have applied this
approach to characterize the acylenzyme reaction interme-
diate of TEM-1 ␤-lactamase, using a kinetically specific,
nitroxyl spin-labeled penicillin as a paramagnetic probe of
active site structure [14]. In addition to assigning the cat-
alytically competent conformation of the acyl moiety of the
substrate in the acylenzyme reaction intermediate, we have
extended this approach through site-directed cysteine mu-
tagenesis to assign structural relationships of nearby active
site residues [14,15]. By comparative ENDOR spectroscopy
of the reaction intermediate of the wild-type enzyme and
the deacylation impaired Glu166Asn mutant, we have iden-
tified a sequestered water molecule in the active site of the
wild-type enzyme within hydrogen-bonding distance to the
carboxylate group of Glu166, positioned for nucleophilic at-
tack on the scissile ester bond in the acylenzyme [14]. This
water molecule is not detected in the acylenzyme formed
with the Glu166Asn mutant [16], and it has not been possible
to demonstrate the position of the hydrolytic water molecule
by X-ray crystallography. Not only do the ENDOR results
provide direct support of Glu166 in its functional role as the
general base catalyst for hydrolytic cleavage of the acylen-
zyme reaction intermediate in wild-type TEM-1 ␤-lactamase
[14], but also, through these studies, we were able to rule
2. Materials and methods
2.1. General
The parent spin-label 2,2,5,5-tetramethyl-1-oxypyrrolinyl-
3-carboxylic acid was obtained by hydrolysis of the cor-
responding carboxamide (Aldrich Chemical Co., Inc.,
Milwaukee, WI, USA) according to the method of Rozantsev
[20]. Specifically deuterated solvents and 6-aminopenici-
llanic acid were also obtained from Aldrich. Benzylpeni-
cillin, cephalotin, and HEPES buffer were obtained from
Sigma Chemical Co. (St. Louis, MO, USA). The synthesis
procedures for 3-(2,2,5,5-tetramethyl-1-oxypyrrolinyl)-2-
propenoic acid (SLPA), 6-N-(2,2,5,5-tetramethyl-1-oxypyr-
rolinyl-3-carboxyl)-penicillanic acid (SLPEN), and 7-N-
(2,2,5,5-tetramethyl-1-oxypyrrolinyl-3-carboxyl)-cephalos-
poranic acid (SLCEP) have been described previously
[17–19]. All other reagents were of analytical grade quality.
Deionized distilled water was used throughout.

D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
1281
2.2. 6-N-[3-(2,2,5,5-tetramethyl-1-oxypyrrolin-3-yl)-
propen-2-oyl]penicillanic acid (SLPPEN)
The resulting plasmid was introduced into E. coli BL21
(DE3) cells for periplasmic expression. The enzyme was pu-
rified by anion exchange (DE-52) and gel filtration (G-25)
chromatography, similarly to TEM-1 ␤-lactamase as previ-
ously published [21].
To a suspension of SLPA (0.87 g, 4.14 mmol) in 20 ml
of benzene maintained at 6–8 ◦C, 5 ml of benzene were
added dropwise containing 0.36 ml (4.94 mmol) of multiply
distilled, colorless SOCl2. To this mixture 5 ml of benzene
containing 0.33 ml (4.10 mmol) dry pyridine were added
dropwise over a period of 5 min with vigorous stirring. The
mixture was then brought to room temperature, and the
pyridinium chloride precipitate was removed by suction fil-
tration. Benzene (100 ml) was added to the orange filtrate,
and the solution was reduced to ∼10 ml in volume. Resid-
ual pyridinium chloride precipitate was removed by gravity
filtration.
The filtrate was evaporated in vacuo to obtain a dark
orange-red oil, 15 ml acetone was added, and the mix-
ture was cooled to 0 ◦C. 6-Aminopenicillanic acid (1.08 g,
5 mmol) suspended in a mixture of 30 ml acetone and 32 ml
4% (w/v) KHCO3 was added dropwise over 30 min. The
mixture was brought to room temperature and stirred for an-
other 30 min. Then 50 ml benzene were added, the organic
phase was separated and discarded. Similarly, 30 ml ether
were added and the ether layer was discarded. Then 40 ml
ether were added to the aqueous layer, the mixture was
cooled on an ice bath, the pH was slowly lowered to 2.5 with
10% phosphoric acid, and the ether layer was separated.
The aqueous layer was extracted with two successive 20 ml
aliquots of cold ether, and the organic extracts were com-
bined. The ether layer was dried over MgSO4 and filtered.
The ether was removed by evaporation in vacuo. The crude
product was recrystallized from 50:50 (v/v) ether/pentane
at −20 ◦C, and 0.88 g (58% yield) of crystalline product
was recovered with mp 140 ◦C (dec.). Thin layer chro-
matography in n-butanol:ethanol:water::4:1:5 (v/v) showed
a single spot, and the infrared spectrum (KBr disc) showed
an intense band at 1792 cm−1 characteristic of the in-
tact ␤-lactam bond. Elemental analysis for C19H26N3O5S
showed: calculated (found) C, 55.87 (54.72); H, 6.42 (6.31);
N, 10.29 (9.85); S, 7.85 (7.34). Fast atom bombardment
mass spectrometric analysis showed the highest molecular
ion peak to correspond to the expected m/e ratio of 409.
2.4. Enzyme kinetics
Initial velocity data were collected with a Cary 15 spec-
trophotometer modified by On-Line Instrument Systems,
Inc. (Jefferson, GA, USA) for microprocessor-controlled
data acquisition. The change in absorbance was evaluated
at 232 nm for benzylpenicillin (ꢀε = 1015 M−1 cm−1),
at 230 nm for SLPEN (ꢀε = 670 M−1 cm−1), at 275 nm
for SLPPEN (ꢀε = 2824 M−1 cm−1), at 265 nm for SL-
CEP (ꢀε = 3589 M−1 cm−1), and at 264 nm for cephalotin
(ꢀε = 7464 M−1 cm−1) to estimate the steady-state kinetic
parameters kcat and kcat/KM [14,17,18]. Initial velocity data
were analyzed with use of the program ENZKIN, as previ-
ously described [24]. For kinetic studies of SLPPEN hydrol-
ysis near and below 0 ◦C, DMSO:H2O cryosolvent mixtures
were prepared as described by Douzou et al. [25,26].
2.5. EPR and ENDOR spectroscopy
EPR and ENDOR spectra were recorded with use of a
Bruker ESP300E spectrometer operating at 9.5 GHz and
equipped with an Oxford Instruments ESR910 liquid he-
lium cryostat and Bruker ENDOR accessory, as previously
described [27]. Typical experimental conditions for EPR
measurements were: sample temperature, 20 K; microwave
frequency, 9.45 GHz; incident microwave power, 64 ␮W;
modulation frequency, 12.5 kHz; modulation amplitude,
0.8 G. Typical experimental conditions for ENDOR were:
microwave power, 6.4 mW; rf power, 50 W; rf modulation
frequency, 12.5 kHz; rf modulation depth, 10 kHz. The static
laboratory magnetic field was not modulated for ENDOR.
Spin-labeled samples were dissolved to a concentration of
4.0 × 10−3 M for EPR and ENDOR studies.
2.6. Molecular modeling
Atomic coordinates of SLPPEN were derived on the basis
of X-ray defined molecular fragments [28,29]. Superpo-
sitioning of 3-(2,2,5,5-tetramethyl-1-oxypyrrolinyl)-2-pro-
penoic acid determined in earlier studies [19] onto the
X-ray defined acyl moiety of amoxycillin [29] was carried
out following previously described graphics procedures
[17–19]. Hydrogen atoms were calculated according to ide-
alized bond lengths and valence angles [30]. The chemical
bonding structure and the atomic numbering scheme of
SLPPEN are shown in Fig. 1.
2.3. TEM-1 and P99 β-lactamases
The procedures for overexpression and purification of re-
combinant TEM-1 ␤-lactamase have been published previ-
ously [21]. For expression of class C ␤-lactamase, the P99
gene was first fused to the ompA leader sequence. For these
experiments, the DNA segment encoding the mature portion
of the P99 enzyme was amplified from the pYY12 plasmid
[22] and a segment encoding the ompA leader sequence was
amplified from the TEM-1-ompA pET24a plasmid described
previously [21]. The resulting PCR fragments were fused by
overlap extension PCR [23] and ligated into the pET24a vec-
tor as previously constructed for TEM-1 ␤-lactamase [21].
Molecular modeling was carried out with use of IN-
SIGHT95 (Accelrys, San Diego, CA, USA) running on an
SGI R4400 Indigo2 Workstation with Solid Impact Graph-
ics. With the program package SYBYL (Tripos, St. Louis,

1282
D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
shows that the propenoyl derivatized substrate is associated
with equivalent specificity to that of classical penicillin
substrates. It is interesting to note here that the values of
steady-state kinetic parameters kcat and kcat/KM in Table 1
clearly illustrate that TEM-1 is an efficient penicillinase but
a poor cephalosporinase. For P99 the opposite relationship
is observed.
3.2. Principles of angle-selected ENDOR of nitroxyl
spin-labels
Fig. 1. Illustration of the chemical bonding structure and the atomic
numbering scheme of SLPPEN.
ENDOR spectroscopy provides the most precise method
of measuring the strength of electron–nucleus hf interac-
tions. Consequently, its application results in very high
resolving power for study of molecular structure. Since an
ENDOR spectrum is a nuclear resonance spectrum, one
must determine the total field at the nucleus to calculate the
transition frequency. The first-order transition frequencies
MO, USA), a systematic conformational search was carried
out with SEARCH, which checks van der Waals contacts
among nonbonded atoms by scanning all possible torsion
angles around rotatable bonds [31,32]. It then identifies
within van der Waals allowed conformational space those
conformations that are compatible with ENDOR-determined
electron–nucleus distances as added constraints, together
with their respective uncertainties. Torsion angle search
calculations were performed in 2◦ increments. The effective
position of the unpaired spin density of the nitroxyl group
as a point dipole was applied as the reference position for
distance constraints, as described earlier [33].
␯
for a nucleus are given in Eq. (1), where the li are the
direction cosines of H0 in the molecular axis system, Aji

1/2





2

ꢁ
ꢂ
3
3
ꢀ
ꢀ
ms
ge





ν
=
gjljAji − liνn
(1)
i=1
j=1
is the orientation dependent value of the hf coupling, νn
the nuclear Larmor frequency, ms the electron spin quantum
number, and ge represents the effective g value defined by
the relation ge = [l12gx2x+l2gy2y +l32gz2z]1/2 [34]. In systems of
low g-anisotropy, as for s2pin-labels, we assume the axis of
quantization of the unpaired electron spin to be the direction
of the applied field H0.
If the applied field is oriented parallel to the principal axis
j of the hf tensor A, Eq. (1) then simplifies to that of Eq. (2)
where the separation of ν about νn is called the ENDOR
3. Results and discussion
3.1. Steady-state kinetic studies
As earlier noted, the chromophoric properties of SLP-
PEN allow convenient monitoring of the kinetics of
enzyme-catalyzed reactions. For SLPPEN, the value of ꢀε
at 275 nm was 2824 M−1 cm−1, compared to ꢀε at 230 nm
of 670 M−1 cm−1 for SLPEN [17]. Kinetic parameters gov-
erning the hydrolysis of SLPPEN catalyzed by TEM-1 and
P99 ␤-lactamase are compared in Table 1. Values of kinetic
parameters for hydrolysis of benzylpenicillin, cephalotin,
SLPEN, and SLCEP obtained in earlier studies [14,17,18]
are provided for comparison. As expected, SLPPEN exhibits
greater reactivity towards TEM-1 ␤-lactamase. Compari-
son of the kinetic parameters for hydrolysis of SLPPEN to
those of benzylpenicillin catalyzed by TEM-1 ␤-lactamase
1
ν
= νn 2 |Ajj|
(2)
shift. For symmetric separations, the hf coupling is, thus,
twice the value of this frequency spacing. Eq. (2) applies
under the condition νn > |A|/2, as for hydrogen nuclei.
For some nuclei the condition νn < |A|/2 may apply, in
which case Eq. (2) becomes ν = |A|/2 νn. Within the
strong-field approximation, the observed hf coupling A is
Table 1
Steady-state kinetic parametersa for the hydrolysis of benzylpenicillin, SLPEN, SLPPEN, cephalotin, and SLCEP, catalyzed by class A TEM-1 and class
C P99 ␤-lactamases
Substrate
Class A TEM-1 ␤-lactamase
Class C P99 ␤-lactamase
kcat (s−1
)
kcat/KM (106 M−1 s−1
)
kcat (s−1
)
kcat/KM (106 M−1 s−1
)
Benzylpenicillin
SLPEN
SLPPEN
Cephalotin
SLCEP
1840
1273
637
14.0
2.0
20
29
36
1.2
1.0
19.9
17.5
13.8
0.15
0.04
0.5
1.0
1.4
0.02
0.01
14.0
4.0
0.50
72
2.0
0.2
0.02
0.04
7.64
3.6
0.8
1.1
0.04
4
0.01
0.01
0.81
0.6
184
17
a
At 24 ◦C in aqueous solution containing 0.1 M NaCl buffered with 0.02 M HEPES to pH 7.

D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
1283
Fig. 2. Magnetic interactions governing angle-selected ENDOR of a nitroxyl spin-label. At the top, the relationships of the molecular axes to the principal
axes of the ge tensor of the nitroxyl group and the A tensors of nearby protons are shown. Relative positions of protons in the x, y plane or on the
z-axis are also indicated. In the lower left-hand portion, a frozen solution EPR spectrum of a spin-label is shown. Setting A is defined by H0 set on
the low-field spectral feature while setting B is defined by H0 set to the central EPR spectral feature. Two ENDOR spectra are shown which would
correspond to spectra of a spin-labeled penicillin totally deuterated at all positions except at H(6). The abscissa measures the observed ENDOR shift,
ꢀν = νENDOR − νH, where νENDOR represents the observed frequency and νH represents the proton Larmor frequency.
given in Eq. (3) as a function of r and α, where h is the
strong-field approximations are less than 3% for r ≥ 5 Å in
Planck’s constant, r is the modulus of the electron–nucleus
ge|βe|gn|βn|
(3)
spin-labeled compounds [36].
Detailed descriptions of the physical basis of angle-selec-
ted ENDOR of spin-labels are provided in earlier publica-
tions [14,15,17–19,35,36]. We briefly review here salient
points with reference to Fig. 2 to facilitate presentation of
ENDOR spectra and discussion. For a spin-label immobi-
lized in a polycrystalline state or in frozen glassy solutions,
as in this investigation, Fig. 2 illustrates the basic geomet-
rical relationships underlying the orientation dependence of
hf couplings of the unpaired electron of the nitroxyl group
with nearby protons as a function of the static (laboratory)
magnetic field H0. The hf interactions are governed by the ge
tensor, which defines the principal axes of the interactions of
the unpaired electron, and the A tensor, which correspond-
ingly describes the principal hyperfine axes of a proton. In
spin-labels, the axes of the ge and A tensors are essentially
A =
(3 cos2α − 1) + Aiso
hr3
position vector r, and α is the angle between H0 and r,
the g’s and β’s having classical definitions [34,35]. For ni-
troxyl spin-labels, the observed principal hfc components
A and A correspond to the maximum and minimum EN-
ꢁ
⊥
DOR shifts in the spectrum, respectively [35]. The principal
hfc components due to dipole–dipole interaction ADꢁ and A⊥D
correspond to the first term of Eq. (3) for values of α = 0◦
and 90◦, respectively. The dipolar Eq. (3) includes implicit
assumption that the electron and nuclear spins are quantized
along the applied field direction [34]. We have evaluated the
quantitative limit of Eq. (3) to nitroxyl spin-labels and have
shown that the errors associated with the point dipole and

1284
D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
coincident and g-anisotropy is very small. The gz compo-
nent of ge is coincident with the molecular z-axis and, there-
fore, perpendicular to the molecular x, y plane. In Fig. 2 the
low-field absorption feature of the EPR spectrum arises from
those molecules for which the gz component is coincident
with the static magnetic field H0. Thus, the five-membered
ring is oriented perpendicularly to H0. Microwave power
saturation that drives the transitions of the unpaired electron
then selects perpendicularly oriented molecules for moni-
toring nuclear resonance transitions with resultant ‘single
crystal-like spectra’ [34–38]. We have termed this condition
setting A. On the other hand, microwave saturation of the in-
tense central feature of the EPR spectrum selects molecules
of all orientations. This is termed setting B.
directly to estimate electron–nucleus distances. The valid-
ity of this approximation has been demonstrated in numer-
ous ENDOR studies of spin-labeled compounds, including
spin-labeled penicillin and cephalosporin [17,18].
3.3. Assignment of ENDOR features and estimation of
electron–proton distances
Fig. 3 illustrates the proton ENDOR spectra of SLPPEN
with H0 at the low-field turning point of the EPR spectrum.
Only the perpendicular hfc components for protons in the
propenoyl penicillin moiety are observed. The top spectrum
(a) was obtained for the sample that was first dissolved in
[2H6]DMSO and then rapidly mixed with [2H4]methanol
and frozen in liquid nitrogen, and the spectrum (b) was
obtained after thawing the sample to exchange the amide
proton H(14). By comparing the two spectra in Fig. 3, the
As illustrated in Fig. 2 for setting A, a proton in the plane
of the spin-label gives rise to the perpendicular hf interac-
tion A while a proton located on or near the molecular
⊥
z-axis gives rise to the parallel hf interaction A . On the
ENDOR features labeled a are attributed to the perpendic-
ꢁ
⊥
other hand, for setting B of the laboratory magnetic field,
an axially located proton gives rise to a perpendicular inter-
ular hfc component of H(14), because in the bottom spec-
trum the peak-to-peak amplitudes of the ENDOR features
are only diminished. With respect to the hf splittings and
action A while a proton in the molecular plane gives rise
⊥
to both parallel (A ) and perpendicular (A ) interactions.
ꢁ
⊥
This differential appearance of parallel and perpendicular
hf couplings dependent on orientation of the spin-label with
respect to the (static) laboratory magnetic field H0 provides
a diagnostic signature for assigning the relative position
of the detected magnetic nucleus according to the axes of
the ge tensor [35,39]. In addition, of critical diagnostic im-
portance is that the ENDOR splitting for the perpendicular
hf interaction A is to first-order independent of the set-
⊥
ting of the magnetic field H0 while the ENDOR splitting
for the parallel interaction A varies and passes through a
ꢁ
maximum value when H0 reaches the canonical orientation
corresponding to the principal hfc component [34]. In the
description of ENDOR spectra of the spin-labeled com-
pounds, assignment of resonance features is based on these
physical relationships and confirmed through experiment.
By measuring the energy of the electron–nucleus hf interac-
tion A, the separation between the unpaired electron of the
nitroxyl group [33] and each ENDOR-active nucleus can be
estimated according to Eq. (3). The position of the nucleus
relative to the molecular plane of the oxypyrrolinyl ring is as-
signed according to the dependence of the resonance pattern
on H0.
The effective dipolar position of the unpaired electron spin
associated with the nitroxyl group lies approximately at the
mid-point of the N–O bond and is defined to within 0.04 Å
[33]. By application of ENDOR we have shown that dipolar
electron–nucleus distances can be determined over a 4–11 Å
range with errors not greater than 5% based on ENDOR line
widths [14,15,17–19,33,35,36,39]. This precision of struc-
tural analysis is greater than that achieved by other physical
methods for molecules in solution and in fact is exceeded
only by that of single crystal X-ray diffraction. For nuclei
for which r ≥ 5 Å, the isotropic contributions are negligi-
bly small. We can then use the observed hfc components
Fig. 3. Proton ENDOR spectra of SLPPEN with H0 at setting A of the
EPR spectrum. (a) The spectrum was obtained for the sample that was first
dissolved in [2H6]DMSO and then rapidly mixed with [2H4]methanol and
frozen in liquid nitrogen. (b) The spectrum was obtained after thawing the
sample to exchange the amide proton H(14). Five line pairs are identified
in the stick diagram and are assigned to the perpendicular hfc components
of specific protons (see text).

D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
1285
its relationship in the spectrum to a , the pair of resonance
⊥
features labeled b is assigned to that of H(6), similar to
⊥
H␣ of spin-labeled amino acids [35,40]. The most intense
pair of resonance features is labeled as the perpendicular
hfc components of H(␣) and H(␤) of the propenoyl moiety
of SLPPEN. We have previously assigned these resonance
features by employing site-specifically deuterated com-
pounds [19]. Since perdeuterated solvents were employed,
the other resonance features, labeled c , must come from
⊥
covalent hydrogens of the penicilloate moiety. In addition to
H(6), there are three classes of hydrogens in the fused ring:
H(2), H(5), and methyl protons. However, the assignment
of c to a specific proton in the penicilloyl moiety cannot
⊥
be made unambiguously.
Fig. 4 illustrates the proton ENDOR spectra of SLPPEN
under identical solvent conditions as in Fig. 3 but with H0
setting at B. At this magnetic field setting, both parallel and
perpendicular hfc components for each proton are observed.
As in Fig. 3, the parallel features for H(14) are absent in the
bottom spectrum of Fig. 4 due to solvent exchange. The de-
crease in ENDOR intensities labeled a and b in Fig. 4 is
ꢁ
ꢁ
more pronounced compared to their perpendicular features
in Fig. 3. The principle hfc components for H(␣), H(␤), H(6),
and H(14) are identified in the stick diagram. The parallel hfc
component c is not observed because it overlaps with the
ꢁ
intense perpendicular features of H(6) and H(14). In Table 2
we have summarized the principal hfc components of spe-
cific protons of SLPPEN identified in the ENDOR spectra
of Figs. 3 and 4. In Table 2 we have also listed the values of
isotropic and dipolar hfc components for protons together
with corresponding electron–proton separations r calculated
on the basis of Eq. (3). As seen in Table 2, the isotropic
hfc contributions of H(14) and H(6) are vanishingly small,
as expected for distant protons. Accordingly, we have cal-
culated the corresponding values of r for other penicilloate
protons labeled H(P) on the basis of the observed perpen-
Fig. 4. Proton ENDOR spectra of SLPPEN with H0 at setting B of the
EPR spectrum. Experimental conditions for obtaining spectra of (a) and
(b) were described in Fig. 3. At this magnetic field setting, both parallel
and perpendicular hfc components are observed, and are identified in the
stick diagram. Detailed assignments are described in the text.
within van der Waals allowed conformational space. The
general methodology of application of torsion angle search
calculations to define molecular conformation constrained
by ENDOR-determined electron–nucleus distances has
been described in earlier publications [17–19,35,36,39,40].
With reference to the atomic numbering scheme in Fig. 1,
the search calculations were applied for simultaneous ro-
tation around the C(3ꢀ)–C(␣), C(␣)–C(␤), C(␤)–C(15),
C(15)–N(14), and N(14)–C(6) bonds. The torsion angle
search calculations were carried out within van der Waals
hard-sphere limits and constrained by electron–proton
dicular hfc component c alone.
⊥
3.4. ENDOR-determined conformation of SLPPEN
To determine the conformation of SLPPEN, we have
carried out torsion angle search analyses to identify con-
formers that accommodate the ENDOR-based values of r
Table 2
Summary of hfc components (A, MHz) and estimated electron–proton distances (r, Å) in SLPPEN
Protona
A
A
Aiso
AꢁD
A⊥D
rb
ꢁ
⊥
H(␣)
H(␤)
H(14)
H(6)
H(P)
1.195
1.416
0.485
0.477
–
0.807
0.817
0.247
0.243
0.107
−0.140
−0.072
−0.003
−0.003
–
1.335
1.488
0.488
0.480
–
−0.667
−0.745
−0.244
−0.240
−0.107
4.92
4.73
6.87
6.91
9.05
0.03
0.03
0.05
0.05
0.19
a
Line pairs are assigned in Figs. 3 and 4: ␣ , ␣ for H(␣); ␤ , ␤ for H(␤); a , a for H(14); b , b for H(6); c for H(P) (representing H(2),
ꢁ
⊥
ꢁ
⊥
ꢁ
⊥
ꢁ
⊥
⊥
H(5), or the geometrically averaged methyl proton; see Fig. 1 for atom designations).
Under the conditions |Aiso| ꢂ |A |, |A | and AD > 0 > A⊥D, the dipolar hfc components AꢁD and A⊥D have been calculated using the constraint
b
ꢁ
⊥
ꢁ
(A + 2A ) = 3Aiso. Uncertainty in frequency of 10–20 kHz due to each ENDOR absorption is included in the calculation of electron–proton distances.
ꢁ
⊥

1286
D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
Fig. 5. Stereo diagram of the ENDOR-determined conformation of spin-labeled propenoyl penicillin in solution. The conformation was obtained according
to the ENDOR-determined distance constraints to H(␣), H(␤), H(14), and H(6). The dihedral angles are listed in Table 3. The ENDOR significant
hydrogen positions with the corresponding electron–proton distances are shown in addition to the nonhydrogen atoms.
distances to H(␣), H(␤), H(14), and H(6) and their uncer-
tainties as listed in Table 2. Since we determined an addi-
tional electron–nucleus distance of 9.05 0.19 Å to either
H(2), H(5), or the geometrically averaged methyl proton in
the penicilloate moiety, labeled H(P), the search calculation
was also constrained by the requirement that the distances
to these covalent hydrogens must be ≥9.05 0.19 Å. In the
torsion angle search calculation, the ENDOR distance con-
straints to these five classes of hydrogens resulted in a single
family of conformers. In Fig. 5 we have illustrated the molec-
ular structure of the ENDOR-compatible conformation of
SLPPEN. Table 3 summarizes the ENDOR-determined
values of torsion angles defining the conformation of the
acylamido group with respect to the penicilloate moiety and
compares them to corresponding results similarly obtained
for SLPA [19] and SLPEN [17] in earlier studies. The dif-
ferences in the values of the first four dihedral angles listed
in Table 3 reflect the influence of the chemical groupings
that comprise the acyl moiety. Of critical importance is the
value of the [C(15)–N(14)–C(6)–C(7)] dihedral angle that
defines the relative conformation of acylamido group with
respect to the ␤-lactam ring of penicillin. The value of this
Table 3
Comparison of ENDOR-determined values of dihedral angles (◦) involving the acylamido group in SLPPEN to corresponding dihedral angles in SLPA
[19], SLPEN [17], and benzylpenicillin [16,29]
Dihedral angle
SLPPEN
SLPA
SLPEN
Benzylpenicillin
C(4ꢀ)=C(3ꢀ)–C(␣)=C(␤)
C(3ꢀ)–C(␣)=C(␤)–C(15)
C(␣)=C(␤)–C(15)=O(16)
C(␤)–C(15)–N(14)–C(6)a
C(15)–N(14)–C(6)–C(7)
140
−148
−120
−102
−120
28
32
24
12
14
180
168
0
–
–
12
18
22
–
–
–
–
–
–
−168
3
14
−172
−118
−129
For SLPEN and benzylpenicillin, the dihedral angles were measured for C(3ꢀ)–C(15)–N(14)–C(6) and C(17)–C(15)–N(14)–C(6), respectively
a
[16,17,29].

D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
1287
dihedral angle in SLPPEN is essentially identical to that of
benzylpenicillin.
by methanol, and the P99 enzyme was stable in DMSO:H2O
mixtures over the same wide temperature range as found for
TEM-1 ␤-lactamase in MeOH:H2O mixtures [14].
Fig. 6 compares the progress curves for hydrolysis of
SLPPEN under different reaction conditions. In contrast to
the rapid hydrolysis of SLPPEN catalyzed by the TEM-1
enzyme, expected in view of the high kcat value in Table 1,
P99 ␤-lactamase exhibits a pattern of “burst kinetics” at
25 ◦C with substrate in excess, i.e., [S0] ꢃ [E0]. This pat-
tern occurs under conditions of rapid conversion of the
Michaelis complex into a reaction intermediate, followed
by slow breakdown of the intermediate into reaction prod-
ucts. This situation is represented in Eq. (4), where E is the
enzyme, S the substrate, ESꢀ the reaction intermediate, the
3.5. Kinetic isolation of a reaction intermediate of P99
β-lactamase
We have carried out preliminary experiments to identify
conditions under which reaction intermediates of TEM-1 or
P99 ␤-lactamase formed with SLPPEN can be kinetically
isolated for ENDOR structural characterization, as in pre-
vious studies [14]. Since we have already characterized re-
action intermediates of TEM-1 ␤-lactamase, we first sought
to identify conditions for detection of reaction intermediates
of the P99 enzyme. In contrast to the TEM-1 enzyme, for
which cryosolvent mixtures of aqueous methanol proved op-
timal, we found that aqueous mixtures of DMSO were more
suitable as a cryosolvent for the P99 enzyme. DMSO was
found to act as a noncompetitive inhibitor of the P99 en-
zyme up to concentrations of 50:50 DMSO:H2O (v:v), sim-
ilar to the noncompetitive inhibition of the TEM-1 enzyme
k1
k2
k3
E + SꢀES−→ESꢀ −→E + P
(4)
k
−1
breakdown of which is rate-limiting, i.e., k2 ꢃ k3, and P
the product of the reaction. However, in contrast to the clas-
sical description of “burst kinetics” [41] in which formation
Fig. 6. Time dependent change in optical density (OD) at 275 nm due to hydrolysis of SLPPEN catalyzed by TEM-1 and P99 ␤-lactamases at ambient
and subzero temperatures. The concentrations of enzyme and substrate were 16 × 10−6 and 12 × 10−6 M, respectively, in 0.1 M NaCl buffered to pH 7
with 0.02 M HEPES. Evidence for formation of a reaction intermediate with P99 ␤-lactamase according to Eq. (4) under subzero temperature conditions
with SLPPEN is shown by the difference spectra in the inset.

1288
D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
of the acylenzyme reaction intermediate ESꢀ is associated
with release of the first product of the reaction P1, acylation
of the nucleophilic serine residue in ␤-lactamases occurs
through cleavage of the ␤-lactam bond with the antibiotic
remaining covalently attached to the enzyme. Furthermore,
as shown in the inset of Fig. 6, under conditions of enzyme
in excess at −5 ◦C, a long-lived intermediate can be spec-
trally characterized because of the enhanced UV absorption
properties of the spin-labeled propenoyl chromophore. For
the kinetic trace of substrate hydrolysis at −5 ◦C, only after
2.5 h did the absorbance begin to show further decline to the
limiting value indicative of complete hydrolysis. Since the
half-life of ESꢀ at −5 ◦C is much greater than the mixing
time, kinetic isolation of the reaction intermediate under
these conditions results in essentially no contamination of
the reaction mixture by unbound substrate or hydrolyzed
product. Under these conditions, the ENDOR spectrum
would arise only from the enzyme-bound, spin-labeled
antibiotic.
Acknowledgements
This work was supported by grants from the National
Science Foundation (MCB-0092524) awarded to M.W.M.
and the National Institutes of Health (AI32956) awarded to
T. P.
References
[1] H.C. Neu, Science 257 (1992) 1064.
[2] G.A. Jacoby, Eur. J. Clin. Microbiol. Infect. Dis. 13 (1994) 2.
[3] B. Joris, J.M. Ghuysen, G. Dive, A. Renard, O. Dideberg, P. Char-
lier, J.M. Frère, J. Kelly, J.C. Boyington, P.C. Moews, J.R. Knox,
Biochem. J. 250 (1988) 313.
[4] K. Bush, G.A. Jacoby, A.A. Medeiros, Antimocrob. Agents
Chemother. 39 (1995) 1211.
[5] J.R. Knox, P.C. Moews, J.M. Frère, Chem. Biol. 3 (1996) 937.
[6] J.R. Knox, Antimicrob. Agents Chemother. 39 (1995) 2593.
[7] M. Paetzel, F. Danel, L. De Castro, S.C. Mosimann, M.G.P. Page,
N.C.J. Strynadka, Nature Struct. Biol. 7 (2000) 918.
[8] L. Maveyraud, D. Golemi, L.P. Kotra, S. Tranier, S. Vakulenko, S.
Mobashery, J.P. Samama, Structure 8 (2000) 1289.
[9] I. Massova, S. Mobashery, Antimicrob. Agents Chemother. 42 (1998)
1.
4. General conclusions
[10] G.S. Weston, J. Blazquez, F. Baquero, S.K. Shoichet, J. Med. Chem.
41 (1998) 4577.
[11] G.B. Wells, D. Mustafi, M.W. Makinen, J. Biol. Chem. 269 (1994)
4577.
[12] D. Mustafi, M.W. Makinen, J. Biol. Chem. 269 (1994) 4587.
[13] M.W. Makinen, Spectrochim. Acta A 54 (1998) 2269.
[14] D. Mustafi, A. Sosa-Peinado, M.W. Makinen, Biochemistry 40 (2001)
2397.
[15] D. Mustafi, A. Sosa-Peinado, V. Gupta, D.J. Gordon, M.W. Makinen,
Biochemistry 41 (2002) 797.
[16] N.C.J. Strynadka, H. Adachi, S.F. Jensen, K. Johns, A. Sielecki, C.
Betzel, K. Sutoh, M.N.J. James, Nature 359 (1992) 700.
[17] D. Mustafi, M.W. Makinen, J. Am. Chem. Soc. 117 (1995) 6739.
[18] D. Mustafi, M.M. Knock, R.W. Shaw, M.W. Makinen, J. Am. Chem.
Soc. 119 (1997) 12619.
[19] D. Mustafi, W.E. Boisvert, M.W. Makinen, J. Am. Chem. Soc. 115
(1993) 3674.
[20] E.G. Rozantsev, Free Nitroxyl Radicals, Plenum Press, New York,
1970, Chapter 9, pp. 203–246.
Our modeling studies [14,15,42] and X-ray crystallo-
graphic investigations [5,8–10,16] have shown that the
active site of ␤-lactamases provides relatively little steric
restriction to the acyl moiety of substrates. On this basis, we
expected that the spin-labeled propenoyl acylamido group
would be accommodated as a convenient chromophoric re-
porter group in the active site of the TEM-1 or P99 enzymes;
however, we did not anticipate that SLPPEN would prove
to be unusual in its kinetic behavior since the reactivity of
the corresponding chromophoric spin-labeled cephalospo-
ranic acid with TEM-1 ␤-lactamase was not associated with
“burst kinetics” (D. Mustafi, M.W. Makinen, unpublished
results). Thus, the kinetic properties of SLPPEN illustrated
in Fig. 6 suggest that SLPPEN, as a probe of the class C
P99 ␤-lactamase, will help to differentiate active site inter-
actions underlying cephalosporinase reactivity from those
responsible for its behavior as a poor penicillinase. Because
of burst kinetics, the optically detected reaction interme-
diate identified in Fig. 6 will be structurally distinct from
the acylenzyme intermediate formed with, for instance, the
corresponding chromophoric spin-labeled cephalosporanic
acid. It is possible that the optically detected species in
Fig. 6 represents a long-lived intermediate that decays to
a structurally “relaxed” species prior to formation of the
enzyme-product complex. ENDOR structural characteriza-
tion of such a species may be feasible through application
of a rapid heating–cooling cycle to a reaction mixture in
which the ESꢀ has been kinetically isolated. This technique
has been applied to other cryoenzymologically isolated
reaction intermediates [43,44], and we have demonstrated
its application to convert the noncovalent Michaelis com-
plex of TEM-1 ␤-lactamase with SLPEN into the covalent
acylenzyme reaction intermediate [14].
[21] A. Sosa-Peinado, D. Mustafi, M.W. Makinen, Prot. Expres. Purific.
19 (2000) 235.
[22] Z. Zhang, Y. Yu, J.M. Musser, T. Palzkill, J. Biol. Chem. 276 (2001)
46568.
[23] S.N. Ho, H.D. Hunt, R.M. Horton, J.K. Pullen, L.R. Pease, Gene 77
(1989) 51.
[24] W. Maret, M.W. Makinen, J. Biol. Chem. 266 (1991) 20636.
[25] P. Douzou, Cryobiochemistry, Academic Press, New York, 1977.
[26] P. Douzou, G. Hui Bon Hoa, P. Maurel, F. Travers, in: G.D. Fasman
(Ed.), Handbook of Biochemistry and Molecular Biology, 3rd ed.,
Chemical Rubber Publishing Co., Cleveland, OH, 1976, pp. 520–540.
[27] D. Mustafi, Y. Nakagawa, Proc. Natl. Acad. Sci. USA 91 (1994)
11323.
[28] J.W. Turley, F.P.B. Boer, Acta Crystallogr. B 28 (1972) 1641.
[29] M.O. Boles, R.J. Girven, P.A. Gane, Acta Crystallogr. B 34 (1978)
461.
[30] O. Kennard, in: C.H. MacGillavry, G.D. Rieck (Eds.), International
Tables for X-ray Crystallography, vol. 3, Clarendon Press, Oxford,
UK, 1968, pp. 275–276.
[31] S. Naruto, J. Motoc, G.R. Marshall, S.B. Daniels, M.J. Sofia, J.A.
Katzenellenbogen, J. Am. Chem. Soc. 107 (1985) 5262.

D. Mustafi et al. / Spectrochimica Acta Part A 60 (2004) 1279–1289
1289
[32] H. Iijima, J.B. Dunbar Jr., G.R. Marshall, Proteins Struct. Funct.
Genet. 2 (1987) 330.
[38] B.M. Hoffman, J. Martinsen, R.A. Venters, J. Mag. Res. 59 (1984)
110.
[33] D. Mustafi, H. Joela, M.W. Makinen, J. Mag. Res. 91 (1991) 497.
[34] G.C. Hurst, T.A. Henderson, R.W. Kreilick, J. Am. Chem. Soc. 107
(1985) 7294.
[39] G.B. Wells, M.W. Makinen, J. Am. Chem. Soc. 110 (1988) 6343.
[40] G.B. Wells, D. Mustafi, M.W. Makinen, J. Am. Chem. Soc. 112
(1990) 2566.
[35] D. Mustafi, J.R. Sachleben, G.B. Wells, M.W. Makinen, J. Am.
Chem. Soc. 112 (1990) 2558.
[36] M.W. Makinen, D. Mustafi, S. Kasa, in: L.J. Berliner (Ed.), Bio-
logical Magnetic Resonance, vol. 14, Spin-Labeling IV: The Next
Millennium, Plenum Press, New York, 1998, pp. 181–249.
[37] G.H. Rist, J.S. Hyde, J. Chem. Phys. 52 (1970) 4633.
[41] B.S. Hartley, B.A. Kilbey, Biochem. J. 56 (1954) 288.
[42] B.P. Atanasov, D. Mustafi, M.W. Makinen, Proc. Natl. Acad. Sci.
USA 97 (2000) 3160.
[43] P. Douzou, R. Sireix, F. Travers, Proc. Natl. Acad. Sci. USA 66
(1970) 787.
[44] M.W. Makinen, A.L. Fink, Ann. Rev. Biophys. Bioeng. 6 (1977) 301.