A direct spectropolarimetric assay of arabinose 5-phosphate isomerase

Article abstract of DOI:10.1016/j.ab.2021.114116

Arabinose 5-phosphate isomerase (API) catalyzes the reversible isomerization of Ribulose 5-phosphate (Ru5P) to Arabinose 5-Phosphate (Ar5P) for the production of 3-deoxy-2-octulosonic acid 8-phosphate (KDO), a component of bacterial lipopolysaccharide (LPS) of gram-negative bacteria. API is an attractive target for therapeutic development against gram-negative bacterial pathogens. The current assay method of API activity utilizes a general reaction for keto sugar determination in a secondary, 3-h color development reaction with 25 N sulfuric acid which poses hazard to both personnel and instrumentation. We therefore aimed to develop a more user friendly assay of the enzyme. Since Ru5P absorbs in the UV region and contains at least 2 chiral centers, it can be expected to display circular dichroism (CD). A wavelength scan revealed indeed Ru5P displays a pronounced negative ellipticity of 30,560 mDeg M^?1cm^?1 at 279 nm in Tris buffer pH 9.1 but Ar5P does not have any CD. API enzymatic reactions were monitored directly and continuously in real time by following the disappearance of CD from the Ru5P substrate, or by the appearance of CD from Ar5P substrate. The CD signal at this wavelength was not affected by absorption of the enzyme protein or of small molecules, or turbidity of the solution. Common additives in protein and enzyme reaction mixtures such as detergents, metals, and 5% dimethylsulfoxide did not interfere with the CD signal. Assay reactions of 1–3 min consistently yielded reproducible results. Introduction of accessories in a spectropolarimeter will easily adapt this assay to high throughput format for screening thousands of small molecules as inhibitor candidates of API.

Full text of DOI:10.1016/j.ab.2021.114116

Analytical Biochemistry 622 (2021) 114116
Contents lists available at ScienceDirect
Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio
A direct spectropolarimetric assay of arabinose 5-phosphate isomerase
a
,
1
b
a
c
Todd M. Kijek , Joel A. Bozue , Rekha G. Panchal , Vladislav A. Litosh ,
d
a,*
Ronald W. Woodard , S. Ashraf Ahmed
a
Systems and Structural Biology Division, United States Army Medical Research Institute of Infectious Diseases, Fort Detrick, MD, 21702, USA
Bacteriology Division, United States Army Medical Research Institute of Infectious Diseases, Fort Detrick, MD, 21702, USA
Medical Countermeasures Division, United States Army Medical Research Institute of Infectious Diseases, Fort Detrick, MD, 21702, USA
Department of Chemistry, University of Michigan, Ann Arbor, MI, 48109-1065, USA
b
c
d
A R T I C L E I N F O
A B S T R A C T
Keywords:
Arabinose 5-phosphate isomerase (API) catalyzes the reversible isomerization of Ribulose 5-phosphate (Ru5P) to
Arabinose 5-Phosphate (Ar5P) for the production of 3-deoxy-2-octulosonic acid 8-phosphate (KDO), a compo-
nent of bacterial lipopolysaccharide (LPS) of gram-negative bacteria. API is an attractive target for therapeutic
development against gram-negative bacterial pathogens. The current assay method of API activity utilizes a
general reaction for keto sugar determination in a secondary, 3-h color development reaction with 25 N sulfuric
acid which poses hazard to both personnel and instrumentation. We therefore aimed to develop a more user
friendly assay of the enzyme. Since Ru5P absorbs in the UV region and contains at least 2 chiral centers, it can be
expected to display circular dichroism (CD). A wavelength scan revealed indeed Ru5P displays a pronounced
Arabinose 5-phosphate isomerase
Francisella tularensis
KdsD
KDO
Assay method
Circular dichroism
Spectropolarimeter
ꢀ 1
ꢀ 1
negative ellipticity of 30,560 mDeg M cm at 279 nm in Tris buffer pH 9.1 but Ar5P does not have any CD. API
enzymatic reactions were monitored directly and continuously in real time by following the disappearance of CD
from the Ru5P substrate, or by the appearance of CD from Ar5P substrate. The CD signal at this wavelength was
not affected by absorption of the enzyme protein or of small molecules, or turbidity of the solution. Common
additives in protein and enzyme reaction mixtures such as detergents, metals, and 5% dimethylsulfoxide did not
interfere with the CD signal. Assay reactions of 1–3 min consistently yielded reproducible results. Introduction of
accessories in a spectropolarimeter will easily adapt this assay to high throughput format for screening thousands
of small molecules as inhibitor candidates of API.
1
. Introduction
Arabinose 5-phosphate isomerase (API) catalyzes the reversible
isomerization of Ribulose 5-phosphate (Ru5P) to Arabinose 5-Phosphate
(
(
Ar5P). Ar5P is the precursor of 3-deoxy-2-octulosonic acid 8-phosphate
KDO), a component of bacterial lipopolysaccharide (LPS) of gram-
negative bacteria [1]. API and other enzymes of the KDO pathway are
sufficiently conserved among gram-negative bacteria and are essential
for growth of the bacteria. Recent investigations in our laboratory have
established that API is essential for Francisella tularensis virulence [2].
Isomerization of Ar5P by API
F. tularensis is a gram-negative facultative intracellular bacterial
pathogen and is the causative agent of tularemia that infects many
mammals including humans. Because of its ability to cause a lethal
infection, low infectious dose, and aerosolizable nature, F. tularensis
subspecies tularensis is considered a potential biowarfare agent [3].
There is a critical need for additional therapeutic development against
this organism. API is an attractive target for such therapeutic
*
Corresponding author.
E-mail address: syed.a.ahmed.civ@mail.mil (S.A. Ahmed).
Present Address: 2769 PGA Blvd, Navarre FL, 32566, USA.
1
https://doi.org/10.1016/j.ab.2021.114116
Received 13 October 2020; Received in revised form 20 December 2020; Accepted 18 January 2021
Available online 12 March 2021
0
003-2697/© 2021 Elsevier Inc. All rights reserved.

T.M. Kijek et al.
Analytical Biochemistry 622 (2021) 114116
development [2,4–7]. The current approach to therapeutic development
against most pathogens heavily relies on high throughput screening
assay of the enzymes and other targets in the presence of thousands of
small organic molecules.
placed into a 2 mm NMR tube. 10
μ
l of 10 mM Tris buffer was added, and
1
3
13
1
13
31
C NMR spectrum with no C– H or C– P was recorded on the 500
MHz Bruker Avance III NMR spectrometer (NCI branch at Fort Detrick,
MD). After recording the initial spectrum, 5 l of 81 M solution of the
API enzyme was added. The 13C NMR spectra were recorded at 5 min
interval for 30 min, followed by 10 min interval for 100 min, at 420 min
and 1260 min since commencement of the reaction.
μ
μ
The main hurdle to this approach is to devise a convenient assay
method for API enzymatic activity. The current assay method relies on
the sulfuric acid conversion of the ketosugar Ru5P into furfural followed
by its prolonged incubation with cysteine and carbazole in the presence
of 25 N sulfuric acid to produce a color adduct [7]. With an extinction
3. Results
ꢀ 1
ꢀ 1
coeffient Ɛ549 of 250,200, M cm the method is sufficiently sensitive
8]. However, the method is not specific for Ru5P but is a general
[
3.1. CD properties of substrates & API
method for measuring almost all ketosugars. In addition, use of such
high concentration of sulfuric acid is health, workplace, and environ-
mental hazard. Moreover, it takes as much as 3 h for the color devel-
opment alone [7]. A direct UV-spectrophotometric assay of Ru5P should
eliminate these drawbacks. Although Ru5P absorbs at 280 nm, it has a
We collected CD data of both Ar5P and Ru5P over a wavelength
◦
range of 180–900 nm at pH 6.83 and 20 C. While Ar5P did not display
any significant CD, Ru5P showed only one intense CD signal centered at
279 nm (Fig. 1A). This signal was not affected by changing the tem-
ꢀ 1
ꢀ 1
◦
◦
very low extinction coefficient Ɛ280 of 50–60 M cm [4,9] which
makes it extremely insensitive, and thus less useful. Besides, many small
organic molecules strongly absorb at 280 nm to obscure the absorption
changes due to enzymatic activity determined by absorbance. Therefore,
we explored for an alternative approach to API activity assay.
perature from 5 C to 50 C or by adding 5%DMSO (see later, Fig. 5), or
0.05% Triton X-100 or 0.05% Tween-20, or turbidity of the solution
(data not shown). The signal was not changed by the presence of 1 mM
metal ions, EDTA or DTT, the common reagents used in enzyme assays.
These results suggested that an enzyme assay by following CD at 279 nm
should not be affected due the changes in the Ru5P substrate properties
by these common reagents. Fig. 1B shows the far-UV CD spectra of 5.6
Assay of many amino acid racemases takes the advantage of a chiral
center in either of its D- or L-substrate which display specific but
opposite optical rotation at one or more wavelengths in a polarimeter
μ
M API indicating a typical helical secondary structure [17], but having
[
10–16]. Like most sugars, Ru5P or Ar5P can also be expected to display
no signal at 260 nm, and above using 1 mm path length cuvettes (not
shown). Therefore, an enzymatic assay following CD changes at 279 nm
will not be affected by the enzyme itself.
optical rotation of polarized light. As such, Ru5P having more than one
chiral center is also expected to display circular dichroism (CD) [17]. We
conducted a CD wavelength scan of Ru5P and found it displays a pro-
nounced negative ellipticity centered at 279 nm. In contrast, Ar5P did
not display any CD at any wavelength. We found that the API enzymatic
reactions can be monitored directly and continuously in real time for as
little as 1–3 min by following the disappearance of CD from the Ru5P
substrate, or the appearance of CD from the conversion of Ar5P substrate
into RuP product. The CD signal was not affected by absorbance or
fluorescence of small molecules, or turbidity of the reaction mixture.
These advantages coupled with a very short assay time could be
conveniently applied to screen large batch of small molecule inhibitors
of API activity.
The CD of Ru5P is not adversely affected by its concentration
(Fig. 2A). In addition, data generated from 0.2 mM to 10 mM Ru5P
perfectly follows a straight line at pH < 1 (0.5 N HCl), 6.8 (water) and
9.1 (50 mM Tris buffer) (Fig. 2B). Linearity of the straight line was
maintained up to 30 mM Ru5P (not shown), the maximum concentra-
tion used in these assays. These results show that depending on the
purpose of the assay and availability, very low (0.2 mM) to very high
(10 mM) substrate concentration can be employed without any
compromise to the CD signal (Fig. 2B). From the slope of the straight
lines in Fig. 2B, it is evident that the intensity of the CD signal decreases
with pH. However, retention of more than 70% of the maximum CD
signal at the low pH of 1 or below will be ideal to follow enzyme re-
actions that are stopped by acidifying the reaction mixtures (see later).
2
. Materials and methods
ꢀ 1
ꢀ 1
We calculated a CD of 30,560 M cm for a molar solution of Ru5P at
pH 9.1 in 50 mM Tris-HCl buffer. This value is over 60 times more
sensitive than the molar absorbance of Ru5P at 280 nm (Ɛ280 of 50–60
Francisella tularensis API was purified from E. coli harboring a
multicopy recombinant plasmid as previously described [7]. The sub-
strates Ar5P, Ru5P, and most common reagents used were purchased
from Sigma-Aldrich. Enediol compounds were custom synthesized and
purified by New England Discovery Partners (23 Business Park Drive,
Branford Drive, CT 06404, USA).
ꢀ
1
ꢀ 1
cm ) [4,9] but almost an eighth of the cysteine-carbazole color
M
ꢀ
1
ꢀ 1
adduct (Ɛ549 of 250,200 M cm ) [8]. The nearly perfect straight lines
in Fig. 2B are also advantageous because absorbance of all chromo-
phoric solutions decrease with the increase in concentration of the
chromophore generating a curved line according to the Bear-Lambert’s
Law. As such, a measured absorbance above 2.0 usually is not accepted
as valid. CD measurement is thus free from the substrate (or product)
concentration effect.
◦
CD spectra were recorded in 1 mm light path quartz cuvettes at 20 C
or as indicated with a Jasco 718 spectropolarimeter equipped with a
Jasco Peltier-temperature controller. An average of 5 scans at 20 nm/
min scan speed with a response time of 8 s was recorded to increase
signal-to-noise ratio. In all spectral measurements, a buffer blank was
recorded separately and subtracted from sample recordings. Time
course of enzyme reactions were measured by following the ellipticity
At pH 0.1–1.0 (0.5 N HCl) intensity of the negative peak at 279 nm is
reduced by 28% (Fig. 2B), but it does not change further over time (data
not shown). Should a reaction be terminated by acidifying the reaction
mixture, a standard curve must be constructed at that pH. We calculated
◦
changes at 30 C or as indicated.
ꢀ 1
ꢀ 1
Unless and otherwise mentioned, all enzyme reaction mixtures
contained 5 mM Ru5P and 1 mM EDTA in 50 mM Tris-HCl pH 9.1 and
a molar ellipticity of Ru5P in 0.5 N HCL as 22,200 M cm . At pH 14
(1 N NaOH) the CD signal of Ru5P gradually disappeared (not shown),
most likely due to its non-enzymatic conversion to Ar5P or another sugar
(Lobry De van Bryn-Ekestein Reaction) that display no ellipticity
(Fig. 1A) [19]. However, by leaving the Ru5P solution at pH 9.5 for
several hours did not decrease the CD signal (data not shown). But the
molar ellipticity of Ru5P is highest (Fig. 2B).
variable amounts of API in a total volume of 200
μ
l, that was incubated
◦
at 30 C. Data collection was started within 10 s of addition of 1–5
to start the reaction. Unless mentioned otherwise, all results are average
of at least two assays.
μl API
NMR identification of product of D-arabinose 5-phosphate isomeri-
zation. 2-[13C]-D-Arabinose-5-monophosphate was obtained by enzyme-
driven selected phosphorylation as described [18] of commercially
3.2. Isomerization of Ru5P & Ar5P by API
1
3
available 2-[ C]-D-arabinose (Omicron Biochemicals, Inc.). A 500
μ
l
solution 50 mM 2-[13C]-D-Arabinose-5-monophosphate in D
O was
Incubation of Ru5P with API resulted in gradual decrease in
2
2

T.M. Kijek et al.
Analytical Biochemistry 622 (2021) 114116
Fig. 1. Circular dichroism spectra of the substrate Ru5P (20 mM) and Ar5P (50 mM) (A), and 5.6
μ
M API from Francisella tularensis (B) in 50 mM NaPB pH 6.83 at
◦
2
0 C. Spectra were collected from 900 nm to 180 nm in 1 mm path length quartz cuvettes. Only the portions of the wavelength range displaying noticeable ellipticity
are shown.
◦
Fig. 2. CD spectra of various concentrations of Ru5P in water at 20 C (A), and standard curve of Ru5P concentrations in water, HCl, and buffer versus CD at 279 nm
B). The data points in B were taken from figures similar to A. Each spectrum in A is an average of 5 scans.
(
ellipticity (Fig. 3A and B), presumably due to its conversion to Ar5P
which does not show any ellipticity. Although Ar5P does not show CD,
incubation of this substrate with API resulted in appearance of a nega-
tive CD at 279 nm, again most likely due to its conversion into Ru5P
maximum Ru5P ellipticity (Fig. 3B). For an isomerase reaction with an
equilibrium constant of 1, the CD values from both substrates should
meet at midpoint. However starting with 10 mM of either substrate, CD
values from both incubations became almost superimposable that was in
equilibrium after overnight incubation (Fig. 3B). With an ellipticity of
(
Fig. 3A).
In order to obtain proof that the products identified in our newly
established CD assay for API are indeed Ru5P and Ar5P, we enzymati-
◦
ꢀ 7.5 at equilibrium (Fig. 3A), we calculated approximately 76% of the
sugar exist as Ar5P, and the remaining as Ru5P. This yields an equilib-
rium constant ([Ru5P]/Ar5P]) of 0.24 favoring Ar5P. This compares
with the E. coli API equilibrium constant of 0.3, greatly favoring Ar5P
[20]. Fig. 3C shows a typical linearity of the progress curve of the re-
action of 5 mM Ru5P with API at the initial stage of the reaction. Thus
the rate of the API reaction can be measured by following as little as 1–2
min. This compares with a total of more than 3 h needed for the
cysteine-carbazole reaction for assaying API [7]. Fig. 3D shows linearity
of the reaction rate of 1 mM Ru5P with varying concentration of the API
enzyme. The range of concentration of the enzyme used in these assays
(40–240 nM) are well within 200 nM employed in the cysteine-carbazole
assay of API [7].
1
3
cally synthesized C-labelled Ar5P from D-arabinose and ATP and pu-
rified by chromatography. We verified its authenticity by MS and 13
NMR in D O. By CD we have verified that this labelled 10 mM substrate
in D O is converted into Ru5P by API in a time dependent fashion. We
C
2
2
1
3
13
then recorded the C NMR spectra of 35 mM [2– C]Ar5P before and
after 2–1250 min of reaction with 1.5
μ
M API in Tris.HCl buffer pH 9.0 at
◦
2
0
C in a 500 MHz Bruker Avance III NMR spectrometer (NCI, Fort
Detrick, MD). A signal at 81 ppm corresponding to Ar5P gradually
decreased with the concomitant increase in a new signal at 208 ppm that
was unambiguously identified as belonging to Ru5P (data not shown).
This result confirmed that the API enzymatic product that we are
measuring by the continuous CD assay method is indeed Ru5P.
Ellipticity of Ru5P incubated with API never reached zero, and in-
cubation of an equimolar Ar5P with API never equilibrated to 50% of
3

T.M. Kijek et al.
Analytical Biochemistry 622 (2021) 114116
Fig. 3. (A) and (B): (A), CD spectra of 10 mM Ru5P
and 10 mM Ar5P alone both in 50 mM Tris.HCl pH
◦
8
.1, 20 C is compared with those of the substrates at
equilibrium in the presence of 0.81 M API in 1 mm
μ
path length quartz cuvettes. Ellipticity values at 279
nm were continuously collected upon addition of
API. After initial incubation of the substrates with
◦
0
.81
μ
M API at 30 C for 30 min (as shown in
Fig. 3B), the reaction mixtures were incubated for
◦
additional 16.5 h at 20 C. Almost superimposable
spectra of the two reaction mixtures after this long
incubation indicate that they are at equilibrium. (B),
Time course of CD changes at 279 nm for enzymatic
reactions of 10 mM Ar5P or 10 mM Ru5P in 50 mM
◦
Tris.HCl pH 8.1, 20 C. C and D: C, Linearity of API
reaction rate with time. 5 mM Ru5P was incubated
with 2 M API in 50 mM Tris.HCl pH 8.1. D, Line-
μ
arity of the reaction rate with increasing concentra-
tion of API. 1 mM Ru5P was incubated in triplicate
with the indicated concentrations of API in 50 mM
Tris.HCl containing 1 mM EDTA (see later). Initial
rates of reactions were calculated from plots similar
to those shown in C at each API concentration.
3
.2. Effects common additives on the activity of API
We investigated the effect of several common metal salts and re-
agents on the reactions of Ru5P with API as measured by monitoring
ellipticity at 279 nm (Fig. 4). Some of these reagents, such as detergents,
dithiothreitol (DTT) and ethylenediamine tetraacetate (EDTA) are
frequently used with enzymes and enzyme reactions [21]. None of them
had any effect on either the intensity or the shape of the CD spectra (data
Fig. 5. Effect of increasing concentrations of DMSO on the API enzymatic ac-
tivity. Activities were calculated from initial rates of reactions containing 10
mM Ru5P, 1 mM EDTA and 2 M API in 50 mM Tris.HCl pH 9.0.
μ
not shown). Some of these however, had profound effect on the API
activity (Fig. 4). Divalent metal salts of mercury and zinc almost
completely abolished the activity, while those of magnesium, calcium,
manganese, cobalt, nickel, and copper increasingly inhibited the activity
in this order. That the counterion sulfate has very little effect is
demonstrated by the fact that magnesium chloride and magnesium
sulfate has very similar effect on the enzyme activity (Fig. 4).
Fig. 4. Effect of divalent metal salts and other common reagents on the cata-
In a simulated 3-D model, two histidine residues at the active site of
the API have been implicated to participate in catalysis [22], as was
found in the Pseudomonas aeruginosa enzyme ([23] by Mark Olson
lytic activity of API followed by monitoring ellipticity at 279 nm. The reaction
mixtures contained 10 mM Ru5P as a substrate and 2
μ
M API and 50 mM Tris.
HCl pH 9.1. Final concentration of the salts and EDTA was 1 mM, and of Tween-
(
USAMRIID, personal communication). The divalent metal induced in-
2
3
0 or triton X-100 was 0.05%. Activities were calculated from initial rates of
hibition observed in Fig. 4 may be due to the metal-histidine interactions
–4 reactions in each case. Solid bars represent standard deviations.
4

T.M. Kijek et al.
Analytical Biochemistry 622 (2021) 114116
at the active site. The metal chelator EDTA increased the activity
twofold. Inhibition by divalent metal salts and activation by the metal
chelator EDTA strongly suggest that our API preparation contained
metal ions. In fact, the enzyme was purified from Ni-NTA affinity col-
umn [7]. The protein might have picked up nickel from the column
during the purification step, and our results described here show that
concentration, reaction time, and enzyme concentration, are the foun-
dations of a dependable enzyme assay [24]. We have demonstrated that
the ellipticity of the API substrate (or product) Ru5P linearly decreases
with its increasing concentration (Fig. 2B), ellipticity of the API sub-
strate Ru5P increases linearly in the presence of API with the passage of
time (Fig. 3B and D), and the initial rates of the API enzyme reaction are
linear with the API concentration (Fig. 3D). Thus, the CD based assay of
API fulfills the three essential criteria of a dependable enzyme assay.
Enzyme assays often require component reagents in addition to the
substrate(s) and enzyme. We have tested several such reagents including
DTT, EDTA, detergents, and metal salts to determine if these compounds
might have any adverse effect on the CD property of the substrate(s) or
on the enzyme itself. None of these reagents had any adverse effect on
the CD of the substrate Ru5P (Fig. 4). Some Metals, such as zinc and
nickel, had strong inhibitory effect on the API activity. A similar effect
was also observed with the enzyme from E. coli [7]. The inhibitory effect
of zinc can be utilized as an internal control in screening compound
databases as inhibitor or drug candidates. Small organic molecules are
often insoluble in aqueous solvents and require DMSO to solubilize. In
most cases, DMSO is inhibitory to enzyme activity. Our demonstration in
Fig. 5 that low concentrations of DMSO had no effect on Ru5P CD and
API enzyme activity makes this spectropolarimetric assay an ideal tool
to screen for small molecule inhibitors.
4
NiSO strongly inhibited the API. Therefore, all further enzyme activity
reaction mixtures included 1 mM EDTA. The divalent iron of ferrous
sulfate on the other hand stimulated the activity remarkably. A struc-
tural role of iron in API would be an important aspect of further
investigation.
3
.3. Low concentrations of dimethyl sulfoxide has no effect on CD of
Ru5P and on API activity
API is an attractive target for therapeutic development against
tuleramia and other gram-negative infections [2]. Since drug-like
organic enzyme inhibitors are often insoluble in aqueous solvents,
dimethyl sulfoxide (DMSO) is frequently used to solubilize those [24].
Thus in many enzyme reactions, low concentrations of DMSO inadver-
tently get introduced that may adversely affect the enzyme activity.
Determination of the threshold of DMSO concentration is a prerequisite
for small molecule inhibitor screening [24]. We therefore looked at the
effects of increasing concentrations of DMSO on API activity (Fig. 5).
DMSO at 5% or lower concentration had an insignificant effect on the
enzyme activity. Above 5%, the solvent progressively inhibited the
enzyme activity. At 10% DMSO, the enzyme retained more than 80%
activity. Therefore we used 5–10% DMSO in the assay mixture as a
control (with no inhibitor) to keep the inhibitors in solution.
We have also probed the dependability of the CD method by
comparing results of our assay with those of the literature determined by
a more cumbersome yet conventional method of cysteine-carbazole
method [7,8]. Although the latter method is eightfold more sensitive
in detecting the substrate Ru5P than our CD method (Fig. 2B), we were
able to use 0.4–1 mM substrate consistently (an example is shown in
Fig. 3D), a concentration close to 0.5 mM used in the cysteine-carbazole
method [7]. The 1 mM concentration of the substrate used is also above
4
. Discussions
the K
Ar5P (0.57 mM) and Ru5P (0.3 mM) [20]. Validity of our CD method is
further demonstrated by the similarity of Ru5P K
m
as reported for F. tularensis Ar5P (0.3 mM) [7], and for E. coli
Rotation of polarized light is a unique property of molecules con-
taining one or more asymmetric carbon atoms. Rotation of plane
polarized light by such molecules has been routinely used for some
racemase enzyme assays [10–16]. On the other hand, extensive appli-
cation of circularly polarized light, such as CD, is used in protein sec-
ondary structure determination [17]. CD of a molecule is the difference
in the absorbance of left circularly polarized light and right circularly
polarized light. In a limited extent, it has also been used for poly-
saccharide secondary structure determination [25]. However, compared
to UV–Vis absorption spectrophotometry, the utility of CD for routine
enzyme activity assays has remained largely unexplored. Only a handful
of examples are available. CD-based assays of ornithine decarboxylase
m
of 0.7 ± 0.1 mM
determined by CD with the above mentioned reported values in litera-
ture that were determined by the cysteine-carbazole method. We
determined this value by measuring the reaction rates of 0.4–2 mM
Ru5P with a fixed 20 nM API concentration. Most importantly, the CD
method described here eliminates the use of highly corrosive 25 N sul-
furic acid in the cysteine carbazole method.
The CD method described here is thus fully compatible with the
conventional cysteine-carbazole method. In addition, unlike the latter,
our method is direct, avoids the toxic sulfuric acid, and at least 90 times
faster. Therefore the CD method should be a routine method of choice
for API enzyme assay in screening for inhibitors from large compound
libraries.
[
26], mandelate racemase [14], triose phosphate isomerase [27], and
amino acid racemase [28] have been used in addition to other more
conventional methods. These CD based assays are direct, without the
need for additional coupling enzyme or chemical reaction, and can
easily be adapted to initial velocity measurements. The basic principle of
this assay is that either the reactant or the principal product, but not
both, of the enzymatic reaction has a CD ellipticity at a particular
wavelength. Progress of the enzyme reaction is accompanied by an in-
crease or decrease of the CD ellipticity. Because no other reactant(s) is
necessary, the technique is a direct measure of the enzyme activity and
should be a preferred method.
Although the API assay can be accomplished in 1 min, the method
described here manually handles one assay at a time. However, CD
spectroplolarimeter manufactures have come up with capabilities to
handle 96 samples at a time. For example, both Jasco and Bio-Logics and
Hinds Instruments market CD machines that has an accessory to screen
96-well plates in 2 min. Adaptation of such instruments in the API ac-
tivity assay would accelerate the discovery of small drug-like molecules
against targeting API for the discovery of new antibiotics for gram
negative bacterial infections.
The reaction of API described here uses one substrate Ru5 and pro-
duces one product Ar5P. Because the reaction does not require an
extraneous energy source, such as ATP, the reaction is also reversible. In
other words the product can also serve as a substrate. In this paper, we
have shown that Ru5P has a strong negative ellipticity at 279 nm of
In conclusion, the API enzymatic assay described here is a direct,
rapid, and ‘real time’ assay that is free from health and environmental
hazard, and should be in routine use for screening of compound libraries
for drug discovery against Francisella tularensis.
In conclusion, the API enzymatic assay described here is a direct,
rapid, and ‘real time’ assay that is free of sulfuric acid use and should be
in routine use for screening of compound libraries for targets allowing
CD measurements.
ꢀ 1 ꢀ 1
3
0,560 M cm , while Ar5P has no ellipticity (Fig. 1A). The lack of
ellipticity of the latter is due to the fact that it does not have any
noticeable absorbance at 279–280 nm. The difference in CD properties
of these two compounds formed the basis of development of our API
assay.
Linearity of a measurement parameter with substrate (or product)
5

T.M. Kijek et al.
Analytical Biochemistry 622 (2021) 114116
Author contributions
[11] S.A. Ahmed, N. Esaki, H. Tanaka, K. Soda, Mechanism of alpha-amino-epsilon-
caprolactam racemase reaction, Biochemistry 25 (2) (1986) 385–388. Retrieved
from, https://www.ncbi.nlm.nih.gov/pubmed/3955003.
Todd M. Kijek – Data Analysis.
Joel A. Bozue – Paper writing & Project supervision.
Rekha G. Panchal – Data Analysis.
[
12] L.M. Fisher, J.G. Belasco, T.W. Bruice, W.J. Albery, J.R. Knowles, Energetics of
proline racemase: transition-state fractionation factors for the two protons
involved in the catalytic steps, Biochemistry 25 (9) (1986) 2543–2551. Retrieved
from, https://www.ncbi.nlm.nih.gov/pubmed/3521738.
Vladislav A. Litosh – NMR experiments and results interpretation.
Ronald W. Woodard – Enzyme purification.
S. Ashraf Ahmed – Conceive method, conduct CD experiments, and
write manuscript.
[
[
[
13] D.L. Schonfeld, U.T. Bornscheuer, Polarimetric assay for the medium-throughput
determination of alpha-amino acid racemase activity, Anal. Chem. 76 (4) (2004)
1
184–1188, https://doi.org/10.1021/ac035012s.
14] T.R. Sharp, G.D. Hegeman, G.L. Kenyon, A direct kinetic assay for mandelate
racemase using circular dichroic measurements, Anal. Biochem. 94 (2) (1979)
3
29–334. Retrieved from, https://www.ncbi.nlm.nih.gov/pubmed/464301.
15] T. Yamauchi, S.Y. Choi, H. Okada, M. Yohda, H. Kumagai, N. Esaki, K. Soda,
Properties of aspartate racemase, a pyridoxal 5’-phosphate-independent amino
acid racemase, J. Biol. Chem. 267 (26) (1992) 18361–18364. Retrieved from, http
s://www.ncbi.nlm.nih.gov/pubmed/1526977.
Acknowledgement
This work was supported by a grant from JSTO-CBD CB10645
[
16] T. Yorifuji, H. Misono, K. Soda, Arginine racemase of Pseudomonas graveolens. II.
Racemization and transamination of ornithine catalyzed by arginine racemase,
J. Biol. Chem. 246 (16) (1971) 5093–5101. Retrieved from, https://www.ncbi.
nlm.nih.gov/pubmed/5570440.
(
USAMRIID Project number 92138743). Opinions, interpretations,
conclusions, and recommendations are those of the authors and are not
necessarily endorsed by the U.S. Army.
[
[
17] N. Berova, K. Nakanishi, R.W. Woody, in: R.W. Woody, N. Berova, K. Nakanishi
(
2
Eds.), Circular Dichroism: Principles and Applications, second ed., Wiley-VCH,
000 second ed.
References
18] M. Ahn, F.C. Cochrane, M.L. Patchett, E.J. Parker, Arabinose 5-phosphate
analogues as mechanistic probes for Neisseria meningitidis 3-deoxy-D-manno-
octulosonate 8-phosphate synthase, Bioorg. Med. Chem. 16 (22) (2008)
[
1] J. Lodowska, D. Wolny, L. Weglarz, The sugar 3-deoxy-d-manno-oct-2-ulosonic
acid (Kdo) as a characteristic component of bacterial endotoxin – a review of its
biosynthesis, function, and placement in the lipopolysaccharide core, Can. J.
Microbiol. 59 (10) (2013) 645–655, https://doi.org/10.1139/cjm-2013-0490.
2] T. Chance, J. Chua, R.G. Toothman, J.T. Ladner, J.E. Nuss, J.L. Raymond, J.
A. Bozue, A spontaneous mutation in kdsD, a biosynthesis gene for 3 Deoxy-D-
manno-Octulosonic Acid, occurred in a ciprofloxacin resistant strain of Francisella
tularensis and caused a high level of attenuation in murine models of tularemia,
PloS One 12 (3) (2017), e0174106, https://doi.org/10.1371/journal.
pone.0174106.
9
830–9836, https://doi.org/10.1016/j.bmc.2008.09.056.
[
[
19] E.R. Garrett, J.F. Young, Alkaline transformations among glucose, fructose, and
mannose, J. Org. Chem. 35 (10) (1970) 3502.
[
20] T.C. Meredith, R.W. Woodard, Characterization of Escherichia coli D-arabinose 5-
phosphate isomerase encoded by kpsF: implications for group 2 capsule
biosynthesis, Biochem. J. 395 (2) (2006) 427–432, https://doi.org/10.1042/
BJ20051828.
[
[
[
[
21] B. Rowe, J.J. Schmidt, L.A. Smith, S.A. Ahmed, Rapid product analysis and
increased sensitivity for quantitative determinations of botulinum neurotoxin
proteolytic activity, Anal. Biochem. 396 (2) (2010) 188–193, https://doi.org/
[
[
3] D.T. Dennis, T.V. Inglesby, D.A. Henderson, J.G. Bartlett, M.S. Ascher, E. Eitzen,
B. Working Group on Civilian, Tularemia as a biological weapon: medical and
public health management, J. Am. Med. Assoc. 285 (21) (2001) 2763–2773,
https://doi.org/10.1001/jama.285.21.2763.
4] L.J. Gourlay, S. Sommaruga, M. Nardini, P. Sperandeo, G. Deho, A. Polissi,
M. Bolognesi, Probing the active site of the sugar isomerase domain from E. coli
arabinose-5-phosphate isomerase via X-ray crystallography, Protein Sci. 19 (12)
1
0.1016/j.ab.2009.09.034.
22] C. Airoldi, S. Sommaruga, S. Merlo, P. Sperandeo, L. Cipolla, A. Polissi, F. Nicotra,
Targeting bacterial membranes: NMR spectroscopy characterization of substrate
recognition and binding requirements of D-arabinose-5-phosphate isomerase,
Chemistry 16 (6) (2010) 1897–1902, https://doi.org/10.1002/chem.200902619.
23] L. Bzdzion, H. Krezel, K. Wrzeszcz, I. Grzegorek, K. Nowinska, G. Chodaczek,
W. Swietnicki, Design of small molecule inhibitors of type III secretion system
ATPase EscN from enteropathogenic Escherichia coli, Acta Biochim. Pol. 64 (1)
(
2010) 2430–2439, https://doi.org/10.1002/pro.525.
[
[
5] T.C. Meredith, R.W. Woodard, Identification of GutQ from Escherichia coli as a D-
arabinose 5-phosphate isomerase, J. Bacteriol. 187 (20) (2005) 6936–6942,
https://doi.org/10.1128/JB.187.20.6936-6942.2005.
6] S. Sommaruga, L.D. Gioia, P. Tortora, A. Polissi, Structure prediction and
functional analysis of KdsD, an enzyme involved in lipopolysaccharide
biosynthesis, Biochem. Biophys. Res. Commun. 388 (2) (2009) 222–227, https://
doi.org/10.1016/j.bbrc.2009.07.154.
(
2017) 49–63, https://doi.org/10.18388/abp.2016_1265.
24] R.A. Copeland, in: Evaluation of Enzyme Inhibitors in Drug Discovery: A Guide for
Medicinal Chemists and Pharmacologists, second ed., John Wiley & Sons, Inc, USA,
2
013.
[
[
25] D.J. Lewis, W.C. Johnson Jr., Optical properties of sugars: circular dichroism of
amylose and glucose oligomers, Biopolymers 17 (1978) 1439–1449.
[
7] A. Yep, R.J. Sorenson, M.R. Wilson, H.D. Showalter, S.D. Larsen, P.R. Keller, R.
W. Woodard, Enediol mimics as inhibitors of the D-arabinose 5-phosphate
isomerase (KdsD) from Francisella tularensis, Bioorg. Med. Chem. Lett 21 (9)
26] H.B. Brooks, M.A. Phillips, Circular dichroism assay for decarboxylation of
optically pure amino acids: application to ornithine decarboxylase, Anal. Biochem.
(
2011) 2679–2682, https://doi.org/10.1016/j.bmcl.2010.12.066.
2
38 (2) (1996) 191–194, https://doi.org/10.1006/abio.1996.0274.
[
[
8] Z. Dische, E. Borenfreund, A new spectrophotometric method for the detection and
determination of keto sugars and trioses, J. Biol. Chem. 192 (2) (1951) 583–587.
Retrieved from, https://www.ncbi.nlm.nih.gov/pubmed/14907652.
[
[
27] R.C. Fahey, E.F. Fischer, A nonenzyme-coupled assay for triosephosphate
isomerase based upon circular dichroism of glyceraldehyde-3-phosphate, Anal.
Biochem. 57 (2) (1974) 547–554. Retrieved from, https://www.ncbi.nlm.nih.
gov/pubmed/4819742.
9] F.C. Knowels, M.K. Pon, M.G. Pon, Direct spectrophotometric measurement of the
ribosephosphate isomerase catalyzed reaction, Anal. Biochem. 29 (1969) 40–47.
28] M. Noda, Y. Matoba, T. Kumagai, M. Sugiyama, A novel assay method for an amino
acid racemase reaction based on circular dichroism, Biochem. J. 389 (Pt 2) (2005)
[
10] S.A. Ahmed, N. Esaki, K. Soda, Purification and properties of alpha-amino-epsilon-
caprolactam racemase from Achromobacter obae, FEBS Lett. 150 (2) (1982)
4
91–496, https://doi.org/10.1042/BJ20041649.
3
70–374. Retrieved from, https://www.ncbi.nlm.nih.gov/pubmed/7160482.
6