Substrate Substitution in Kanosamine Biosynthesis Using Phosphonates and Phosphite Rescue

Article abstract of DOI:10.1021/acs.biochem.1c00283

Kanosamine is an antibiotic and antifungal compound synthesized from glucose 6-phosphate (G6P) inBacillus subtilisby the action of three enzymes: NtdC, which catalyzes NAD-dependent oxidation of the C3-hydroxyl; NtdA, a PLP-dependent aminotransferase; and

Full text of DOI:10.1021/acs.biochem.1c00283

pubs.acs.org/biochemistry
Article
Substrate Substitution in Kanosamine Biosynthesis Using
Phosphonates and Phosphite Rescue
Natasha D. Vetter and David R. J. Palmer*
Cite This: Biochemistry 2021, 60, 1926−1932
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Kanosamine is an antibiotic and antifungal compound
synthesized from glucose 6-phosphate (G6P) in Bacillus subtilis by the action
of three enzymes: NtdC, which catalyzes NAD-dependent oxidation of the
C3-hydroxyl; NtdA, a PLP-dependent aminotransferase; and NtdB, a
phosphatase. We previously demonstrated that NtdC can also oxidize
substrates such as glucose and xylose, though at much lower rates, suggesting
that the phosphoryloxymethylene moiety of the substrate is critical for
effective catalysis. To probe this, we synthesized two phosphonate analogues
of G6P in which the bridging oxygen is replaced by methylene and
difluoromethylene groups. These analogues are substrates for NtdC, with
second-order rate constants an order of magnitude lower than those for G6P. NtdA converts the resulting 3-keto products to the
corresponding kanosamine 6-phosphonate analogues. We compared the rates to the rate of NtdC oxidation of glucose and xylose
and showed that the low reactivity of xylose could be rescued 4-fold by the presence of phosphite, mimicking G6P in two pieces.
These results allow the evaluation of the individual energetic contributions to catalysis of the bridging oxygen, the bridging C6
methylene, the phosphodianion, and the entropic gain of one substrate versus two substrate pieces. Phosphite also rescued the
reversible formation 3-amino-3-deoxy-D-xylose by NtdA, demonstrating that truncated and nonhydrolyzable analogues of
kanosamine 6-phosphate can be generated enzymatically.
anosamine, 3-amino-3-deoxy-D-glucose, is an amino sugar
Kknown to inhibit the growth of a variety of pathogenic
bacterial species, including Staphylococcus aureus, as well as
certain fungi such as Candida spp. Inhibition of these
organisms by kanosamine is a result of intracellular
phosphorylation to kanosamine 6-phosphate (K6P), which
then inhibits glucosamine-6-phosphate synthase during cell
wall synthesis.1,2 Kanosamine is also a precursor to more
complex natural products and antibiotics, such as kanamycin,
rifamycins, and ansamycins.1,3,4
A variety of Bacillus species have been shown to produce
kanosamine3−6 and, as a result, have found utility in the
treatment of plant disease.1,7 Kanosamine was first isolated
from cultures of Bacillus pumilis in 19685 and later from
Bacillus cereus.1,6 The proposed biosynthetic pathway of
kanosamine from these species involves first the C3 oxidation
of UDP-glucose by a dehydrogenase, then transamination
using a PLP-dependent aminotransferase, and hydrolysis of the
UDP at C1. We have more recently demonstrated that in
Bacillus subtilis, this pathway starts not with UDP-glucose but
rather with glucose 6-phosphate (G6P) (Figure 1A). The final
Figure 1. (A) Kanosamine biosynthesis in B. subtilis. (B) Enolate
formation from 3oG6P. (C) Structures of G6P, phosphonate
analogues of G6P, and xylose and inorganic phosphite mimicking
G6P in two pieces.
step catalyzed by NtdB, a phosphatase, irreversibly cleaves the
Received: April 23, 2021
C6 phosphate group. This was the first time the functions of
Revised: June 1, 2021
these enzymes and this pathway were demonstrated
Published: June 7, 2021
unambiguously.8,9 Prior to that work, these three enzymes
were mistakenly believed to make 3,3′-neotrehalosadiamine,
resulting in the ntd designation. We have since examined
https://doi.org/10.1021/acs.biochem.1c00283
© 2021 American Chemical Society
Biochemistry 2021, 60, 1926−1932
1926

Biochemistry
pubs.acs.org/biochemistry
Article
similar enzymes from B. cereus UW85 (KabABC) and have
shown that they have the same enzymatic functions as
NtdABC.10
Scheme 1. Synthesis of Phosphonates 1 and 2
The first step in this pathway catalyzed by NtdC, a G6P 3-
dehydrogenase, oxidizes the C3 hydroxyl of G6P to generate a
unique 3-keto reducing sugar. This product, 3-oxo-G6P
(3oG6P), is only one of very few examples of known
biologically relevant 3-keto sugars.11−14 In solution, G6P and
3oG6P rapidly undergo mutarotation, and therefore as a result,
3oG6P transiently exists in the open-chain form as a 1,3-
dicarbonyl, which is then deprotonated at C2 under alkaline
conditions to give the enolate form shown in Figure 1B,
producing a strong, characteristic ultraviolet (UV) absorbance
at 310 nm.8,15 We recently demonstrated a carbocyclic
analogue of the α-anomer of G6P was a substrate for NtdC,
and the resulting product, which is incapable of the ring
opening and enolization shown in Figure 1B, can act as a
substrate for NtdA.16 This confirmed all previous observations
and established that these two enzymes had co-evolved so that
both process the α-anomer, avoiding the observed side
reactions associated with ring opening that would be required
for mutarotation.
Having observed that NtdC and NtdA can accept a non-
natural sugar as a substrate, we chose to investigate whether
other substrate analogues could be accepted into the active
sites of these enzymes. Our previous work with the Ntd
enzymes showed that the C6 phosphate group of the substrate
is crucial for the proper activity of the enzyme, with both
glucose and xylose displaying only minimal activity with the
enzyme. John Richard’s laboratory has shown that the low
activity of several enzymes with truncated substrates can be
rescued with the addition of various dianions, mimicking the
substrate in pieces.17−20 Here we describe the ability of NtdC
to catalyze the oxidation of G6P analogues in which the C6-
phosphate group has been replaced by a nonhydrolyzable
methylenephosphonate or difluoromethylenephosphonate
group, or removed and rescued by the presence of phosphite
(Figure 1C), and the energetic contributions of the substituted
groups to catalysis. We also show that these non-natural
substrates can be converted to the corresponding 3-amino-3-
deoxy analogues by NtdA.
product. Several alternate deprotection attempts did not
improve yields. For example, TMSI is reported to cleave the
alkyl protecting groups of phosphonates much more rapidly
than TMSBr,22−24 and while this was the case, we also
observed partial deprotection of the benzyl groups and
substitution at the anomeric position. Regardless of the excess
of TMSI used, a complex mixture was always obtained, which
could not be resolved. Final products were purified by flash
chromatography followed by ion exchange chromatography
and converted to the monosodium salt by careful titration with
acidic DOWEX, as our experience has shown that the
monosodium salt is easier to handle and less hygroscopic
than the free acid.
Activity of the NtdC-Catalyzed Reaction with
Phosphonates 1 and 2. Initial assays with NtdC and NAD
revealed that both methylene- and difluoromethylenephopho-
nate derivatives 1 and 2 were good substrates for the enzyme.
Observation of the complete UV spectra evolving over time
revealed two peaks corresponding to formation of both NADH
(340 nm) and enolate (310 nm) (Figure 2A−C), similar to
what is observed with G6P.8,15 When the reaction mixture was
supplemented with excess L-glutamate and the next enzyme in
the pathway, NtdA, only NADH production was observed,
while the 310 nm peak was no longer observed (Figure 2D−
F). This indicates that NtdA could consume the 3-keto
phosphonate products, converting them to the corresponding
kanosamine phosphonates. Additionally, the apparent rate of
the NtdC reaction with 1 and 2 was enhanced in the presence
of NtdA, as was previously observed for the reaction with
G6P,8,15 but not to the same extent, suggesting NtdA is not as
tolerant of the phosphonate analogues of 3oG6P.
RESULTS AND DISCUSSION
■
Synthesis of Phosphonates 1 and 2. We were able to
synthesize the desired phosphonate analogues of G6P
according to a published procedure with minor changes
(Scheme 1).21 Methyl tri-O-benzyl glucopyranoside 5 was
synthesized easily in two steps from commercially available 4,6-
O-benzylidene 3 as previously reported.16 The resulting C6
hydroxyl was converted to triflate 6 in quantitative yield,
followed by displacement with either diethyl methylphospho-
nate or diethyl difluoromethyl phosphonate to give the desired
protected glucose phosphonate 7 or 8, respectively. It is
important to note that excess methyl or difluoromethyl
phosphonate was difficult to remove by flash column
chromatography, as it co-elutes with the product and could
be removed by rotary evaporation only at higher temperatures
under high vacuum for extended periods of time.
The kinetics of 1 and 2 with NtdC was examined in detail by
varying the concentrations of both substrates. Because of the
overlapping peaks that evolve in the UV spectra over time,
spectra were deconvoluted using a procedure developed
previously to obtain kinetic information about both the
NtdC reaction and enolate formation.15 A complete kinetic
profile of the enzyme with these substrates was obtained by
varying the substrate concentration at different fixed
concentrations of the second substrate, NAD (Figure 3).
Kinetic parameters are summarized in Table 1 and compared
to those of G6P.15
Final deprotection was achieved in three steps giving
compounds 1 and 2 in 25% and 16% yields, respectively.
The loss of yield after deprotection is likely due to
decomposition of the product during the acidic cleavage of
the methyl glycoside, as well as difficulties in purifying the final
Both 1 and 2 were shown to be good substrates for NtdC,
with kcat values 4- and 12-fold lower than for G6P, respectively.
The Km parameter for 2 was on the same order of magnitude
compared to that of G6P, while that of 1 was 3-fold higher.
1927
https://doi.org/10.1021/acs.biochem.1c00283
Biochemistry 2021, 60, 1926−1932

Biochemistry
pubs.acs.org/biochemistry
Article
difference in pKa should have a minimal effect on binding of
the substrates. Carbon−fluorine bonds have a van der Waals
radius of 1.47 Å, significantly larger than carbon−hydrogen
bonds (1.2 Å); CF2 cannot be considered a closely isosteric
replacement for oxygen. The increased size of the bridging CF2
compared to oxygen suggests that the active site has either
adequate space or flexibility to accommodate changes in the
substrate phosphate region. No change in the Km of NAD was
observed in the presence of either phosphonate compared to
that of G6P. These results are similar to those of previous
studies of yeast glucose-6-phosphate dehydrogenase (G6PDH)
and 6-phosphogluoconate dehydrogenase (6PGDH) with
phosphonate analogues of their respective substrates. Both
G6PDH and 6PGDH showed approximately 4-fold increases
in the Km of the phosphonate derivatives compared to those of
their natural substrates, and no effect on the Km of the cofactor.
Additionally, G6PDH showed a 2-fold decrease in Vmax while
that of 6PGDH decreased 10-fold.26,27 We have previously
observed that with the natural substrate, G6P, NtdC follows a
random bisubstrate mechanism, and kinetic isotope effect
experiments revealed that the hydride transfer step is at least
partially rate-limiting.15 The observed decreases in the kcat
values suggest that hydride transfer is fully rate-limiting for
these substrates.
Kinetics of Enolate Formation. The reactions of NtdC
with 1 and 2 both showed the formation of a second peak in
the UV spectra at 310 nm, characteristic of the non-enzymatic
conversion of the 3-keto products to the corresponding
enolates. We were able to measure both the enzymatic
reaction and non-enzymatic enolate formation simultaneously,
as we have previously done for 3oG6P formation,15 by
deconvolution of the wavelength traces evolving over the
course of the reaction. From the deconvoluted peaks, plots of
[NADH] versus A310 for the reaction with either phosphonate
revealed extinction coefficients for the absorbing enolate
species similar to that of 3oG6P (Table 1 and Figure
S1A,B). Using the derived extinction coefficients, we were
able to quantify both the absorbing enolate species and the
non-absorbing 3-keto species throughout each reaction. Plots
of the rate of enolate formation versus the steady state
concentration of the non-absorbing 3-keto species revealed
apparent rate constants of enolate formation, ke, only
marginally slower than that observed for 3oG6P (Table 1
and Figure S1C,D).
Figure 2. Wavelength traces of the reaction of NtdC with (A) G6P,
(B) phosphonate 1, and (C) phosphonate 2. (D−F) Same reactions
as in panels A−C, respectively, coupled to excess NtdA. All reactions
were carried out at 25 °C in 100 mM AMP-HCl (pH 9.5) with 1 mM
substrate, 1 mM NAD, and 10 mM L-glutamate.
NtdC Activity with Smaller Monosaccharides. NtdC
has been shown to be active in the presence of a small subset of
monosaccharides, with the best substrates being G6P, glucose,
and xylose.8,28 While the enzyme has the greatest activity with
G6P, glucose and xylose are <1% as active in comparison,
which emphasizes the importance of the C6 phosphate on
substrate binding and reactivity. We had therefore postulated
that addition of either free phosphate or phosphite might
rescue the low activity of glucose and xylose with NtdC. The
addition of inorganic phosphate had no effect on the reaction
with glucose or xylose; at high concentrations, slight inhibition
was seen with glucose. In the presence of sodium phosphite, no
effect was observed on the reaction with glucose; however, we
were pleased to observe an increase in the rate of oxidation of
xylose by NtdC. The apparent Km of phosphite with xylose was
found to be 27 5 mM, which was calculated from a series of
experiments in which the concentration of phosphite was
varied, with fixed NAD and xylose concentrations [10 and 100
mM, respectively (Figure 4)].
Figure 3. Saturation curves of the reaction of NtdC with (A) 1 and
(B) 2. All reactions were carried out at 25 °C in 100 mM AMP-HCl
■
(pH 9.5), 28 nM NtdC, and fixed NAD concentrations of 5.00 ( ),
◆
●
○
▲
2.50 ( ), 1.00 ( ), 0.50 ( ), 0.25 (×), and 0.10 mM ( ). All
reactions were performed in duplicate.
This could indicate that hydrogen bonding of the phosphate-
bridging group is important for properly binding and orienting
the phosphate in the active site. Replacement of the oxygen
atom with methylene and difluoromethylene groups will
impact the pKa of the substrate analogues. While G6P has a
second pKa value of 6.4, 1 has a value of 7.6, while 2 has a value
of 5.6.25 At the pH of the assay, all three substrates should exist
predominantly as the dianion in solution, meaning the
1928
https://doi.org/10.1021/acs.biochem.1c00283
Biochemistry 2021, 60, 1926−1932

Biochemistry
pubs.acs.org/biochemistry
Article
Table 1. Kinetic Parameters of NtdC with 1 and 2
a
G6P
CH2, 1
CF2, 2
kcat (s−1
)
4.1 0.1
1.08 0.01
0.14 0.01
0.05 0.01
3.7 0.4
(7.7 0.6) × 103
(20 3) × 103
(2.2 0.1) × 104
(3.3 0.1) × 10−3
0.34 0.01
0.065 0.007
0.03 0.01
sugar
Km
Km
(mM)
(mM)
0.043 0.004
0.040 0.004
0.9 0.1
(9.6 0.8) × 104
(10 1) × 104
(2.34 0.01) × 104
(5.2 0.1) × 10−3
NAD
KiNAD (mM)
6.0 0.8
b
kcat/Km
(M−1 s−1
)
(5.3 0.5) × 103
(10 4) × 103
(1.9 0.1) × 104
sugar
b
NAD
kcat/Km
(M−1 s−1
)
c
ε310 (M−1 cm−1
)
d
ke (s−1
)
(3.6 0.1) × 10−3
a
b
c
From ref 15. Calculated from a global fitting of the kinetic data to eq S1 rearranged to treat Vmax/Ka or Vmax/Kb as a single parameter. Calculated
d
from eq S2. Calculated from the pseudo-first-order rate equation (eq S3).
normalized to the initial velocities in the absence of phosphite,
v0 (Figure S2), fit to the equation describing a two-substrate
mechanism (eq S4) to give the kinetic parameters listed in
Table 2, and compared to those obtained in the absence of
Table 2. Kinetic Parameters of NtdC with Xylose and
Phosphite
0 mM Na2HPO3
10−100 mM Na2HPO3
kcat (s−1
kXyl (M)
cat/kXyl (M−1 s−1
kHPO (mM)
)
1.2 0.1
0.52 0.09
2.3 0.2
−
2.2 0.1
0.24 0.03
k
)
9
1
46
5
3
k
cat/kXylkHPO (M−2 s−1
)
−
190 30
Figure 4. Estimation of the Km of phosphite. All reactions were
carried out in duplicate at 25 °C in 100 mM AMP-HCl (pH 9.5) with
100 mM xylose, 10 mM NAD, and 28 nM NtdC. Data were fit to the
Michaelis−Menten equation ().
3
phosphite. Examination of the kinetics of NtdC with xylose
showed that in the presence of ≤100 mM phosphite, the Km of
xylose decreases by half, and the kcat doubles, resulting in a
nearly 4-fold increase in the second-order rate constant (Table
2 and Figure 5B). This indicates a 2-fold effect of phosphite,
whereby both catalysis and substrate binding are enhanced.
We examined the coupled NtdC−NtdA reaction with xylose,
in the same manner as with 1 and 2; however, there was no
observable effect of NtdA on the NtdC reaction. This is likely a
result of the very slow turnover of the reaction of NtdC with
xylose, such that any effect is too subtle to detect. The same
was observed when we compared the rates of the uncoupled
and coupled reactions in the presence of phosphite. Addition-
ally, because very little enolate is visible in the reaction with
xylose, we could not observe any activity by NtdA in the
coupled reaction, so it was unclear if NtdA could accept
oxidized xylose as a substrate.
We next examined the kinetics of NtdC in more detail by
varying both xylose and phosphite concentrations, keeping the
NAD concentration fixed under saturating conditions. We had
observed that the maximum activity of the enzyme with xylose
was obtained at 10 mM NAD, with some substrate inhibition
at higher concentrations. Therefore, the apparent kinetic
constants of NtdC with xylose were obtained at saturating
NAD concentrations; the same experiments were then carried
out in the presence of 10−100 mM phosphite, and the curves
are shown in Figure 5A. The data for each curve were fit to the
Michaelis−Menten equation to give apparent kcat and Km
values, with the effect of phosphite concentration on these
parameters shown in Figure 5B. The same data were also
While phosphite was shown to activate the reaction with
xylose, we observed inhibition of the reaction with G6P. This is
unsurprising because binding of phosphite in the active site
would likely interfere with the binding of G6P. By varying both
substrate concentrations at fixed phosphite concentrations, we
were able to show that phosphite is a competitive inhibitor of
both NAD and G6P [Ki = 20 3 mM (Figure S3)], suggesting
that it is capable of binding to both the phosphate binding
region of the G6P binding site and the diphosphate binding
region of the NAD binding site.
Our observations suggest that the size of the phosphate
binding pocket is ideally spaced for the CH2OPO3 of the
sugar substrate, necessary for orienting the substrate properly
in the active site for catalysis. The lack of any significant effect
of phosphate with either glucose or xylose as the substrate
suggests that inorganic phosphate is too sterically bulky to
occupy this binding pocket while still allowing the sugar to
Figure 5. (A) Saturation curves of NtdC with xylose. Each experiment
was carried out at 25 °C in 100 mM Amp-HCl (pH 9.5), 10 mM
NAD, and 28 nM NtdC. The solid curves are the fits of the data to
2−
◆
●
▲
the Michaelis−Menten equation at 0 ( ), 10 (×), 25 ( ), 50 ( ),
■
and 100 mM ( ) phosphite. (B) Effect of phosphite on the apparent
●
kinetic parameters of NtdC with xylose from panel A: kcat ( ), Km
(
■
), and kcat/KXyl ( ).
▲
1929
https://doi.org/10.1021/acs.biochem.1c00283
Biochemistry 2021, 60, 1926−1932

Biochemistry
pubs.acs.org/biochemistry
Article
Figure 6. (A) Reaction of NtdC with 3oX and NADH and simultaneous 3oX enolate formation. (B) Wavelength traces of the reaction of NtdC
with 3oX and NADH. (C) Disappearance of NADH obtained by deconvolution of the wavelength traces in panel B. (C) Appearance of 3oX
enolate obtained by deconvolution of the wavelength traces in panel B. Arrows in panels B−D represent the direction of peak formation over time.
bind. The greater activity of glucose compared to that of xylose
suggests that the CH2OH is large enough to partially occupy
the binding pocket of the phosphate and can satisfy some
hydrogen bonding interactions. Xylose, however, has no
moiety to occupy this site when bound in the reactive α-
pyranose form, allowing phosphite to bind and satisfy many
bonding interactions in the site, which consequently improves
the catalytic efficiency and Km.
Having shown that enzyme activity with xylose could be
rescued with phosphite, we were interested to see if the same
was true of substrates in the reverse direction with NtdC and
NtdA. This prompted us to synthesize both 3-oxo-D-xylose
[3oX (Figure 6A)] and 3-amino-3-deoxy-D-xylose [3aX
(Figure 7A)]. We hypothesized that the lack of the C6
phosphate would cause 3oX to be much more stable in
solution than 3oG6P and allow us to assay both NtdC in the
reverse direction and NtdA in the forward direction of the
pathway.
Activity of NtdA and NtdC with Xylose Analogues.
Synthesis of both 3oX and 3aX followed similar synthetic
routes that were seen for 3-oxo-D-glucose and kanosamine
(Scheme S1),3,8 and they were obtained in 37% (over four
steps) and 23% (over six steps) yields, respectively. Aqueous,
unbuffered solutions of 3oX showed very little enolate
formation at 310 nm, while in an alkaline solution, a prominent
310 nm peak was visible in the UV spectrum. This is similar to
our previous observations with 3-oxoglucose, consistent with
the non-enzymatic formation of the enolate species.8 Our
initial assays of 3oX with NtdC in the presence of NADH
showed the simultaneous disappearance of the 340 nm peak of
NADH and the appearance of the 310 nm enolate peak
(Figure 6). In the presence of 100 mM phosphite, the rate of
NADH consumption was approximately 1.5 times faster,
indicating that phosphite binding activates the reverse catalytic
activity of NtdC. Additionally, the rate of enolate formation
also nearly doubled. Quantification of the species in solution
over the course of the reaction showed that approximately 70%
of the reacted 3oX is consumed by the NtdC reaction, while
nearly 30% is being diverted to the unproductive enolate.
The NtdA−NtdC coupled reaction starting from 3aX,
NADH, and 2-ketoglutarate (2KG) resulted in the expected
disappearance of NADH at 340 nm (Figure 7B), with very
little enolate being formed, indicating that NtdA recognizes
this substrate, and as expected, NtdC reduces the product,
3oX. When the reaction mixture also included 100 mM
phosphite, we observed an initial increase in absorbance at 310
nm and a 2-fold increase in the rate of NADH disappearance
(Figure 7C), indicating that phosphite binding activates the
transamination of 3aX by NtdA also.
Effect of the C6 Bridging Group on Catalysis. The
ability of phosphite to rescue low enzyme activity with small,
truncated substrates has been observed with other enzymes,
notably phosphoglucomutase (PGM),29,30 G6P isomerase
(PGI), G6P dehydrogenase (G6PDH), and 6-phosphogluco-
nate dehydrogenase (6PGDH).19,20 With PGM, a 1000-fold
increase in the reaction rate is observed for smaller, truncated
substrates, in the presence of phosphite. Work from Richard et
al. has similarly postulated that phosphite binding interactions
allow for the stabilization and activation of the enzyme
compared to the unliganded catalyst that would otherwise exist
Figure 7. (A) Coupled NtdA−NtdC reaction with 3aX, 2KG, and
NADH. Wavelength traces of the reverse NtdA−NtdC reaction with
3aX (B) without added phosphite and (C) with 100 mM phosphite.
Reactions were carried out at 25 °C in 100 mM Tris-HCl (pH 8.0)
with 0.25 M 3aX, 5 mM 2KG, 0.2 mM NADH, 36 nM NtdC, and
0.18 μM NtdA over the course of 1 h. Arrows in panels B and C
represent the direction of peak formation over time.
1930
https://doi.org/10.1021/acs.biochem.1c00283
Biochemistry 2021, 60, 1926−1932

Biochemistry
pubs.acs.org/biochemistry
Article
in an open, flexible, and inactive form. Their studies compare
second-order rate constants of whole (phosphorylated) and
truncated substrates to determine intrinsic binding energies.
They have shown that the phosphate groups of the whole,
natural substrates for PGI, G6PDH and 6PGDH, have intrinsic
binding energies of 11−13 kcal/mol, 50% of which can be
recovered by addition of phosphite for smaller, truncated
substrates.20
G6P, and the activity with xylose can be partially rescued by
the addition of inorganic phosphite, improving the catalytic
efficiency 4-fold. Taken together, we have been able to evaluate
2−
the role of the individual pieces of the CH2OPO3 on the
binding of G6P. Our results show that the electrostatic
interactions of the phosphate, the bridging oxygen, and the
substrate as a whole each contribute to the binding of the
substrate in the active site. The resulting 3-keto products of
these reactions are all substrates for the aminotransferase
NtdA, and phosphite rescues NtdA-catalyzed 3-amino-3-
deoxy-D-xylose transamination, reinforcing our observation
that these two enzymes have co-evolved to process the
relatively unstable 3-keto intermediate and minimize un-
productive enolate formation. The ability of NtdC and NtdA
to accept non-natural substrates such as carbocyclic and
phosphonate analogues suggests that these enzymes could be
used to produce novel and nonhydrolyzable kanosamine
derivatives, which could in turn be incorporated into more
complex natural product analogues with improved therapeutic
properties.
In the case presented here, the second-order rate constants
for NtdC-catalyzed oxidation can be used to approximate the
relative activation barriers in the case of G6P (ΔG⧧G6P), xylose
(ΔG⧧ ), and xylose in the presence of phosphite (ΔG⧧XP).
X
Our results indicate that the phosphate of G6P accounts for
approximately 6 kcal/mol of intrinsic binding energy (ΔG⧧
−
X
ΔG⧧G6P) in the active site of NtdC, and with xylose, 2.6 kcal/
mol (41%) is recovered in the presence of phosphite (ΔG⧧
−
X
ΔG⧧XP) as shown in Figure 8. Much of the remaining binding
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.biochem.1c00283.
Full experimental procedures, including synthetic
methods, protein purification, and kinetic methods,
and supplemental figures and H and 13C NMR spectra
of compounds synthesized in this study (PDF)
1
Figure 8. Contributions of substrate pieces to the intrinsic binding
energy of G6P. Energies were calculated using second-order rate
Ä
É
Å
Å
Ñ
Ñ
kcat
Km κbT
h
constants according to the equation ΔG⧧ = −RT ln
. For
Å
Å
Å
Å
Ñ
Ñ
( )Ñ
Accession Codes
NtdC, O07564; NtdA, O07566.
Ñ
Å
Ñ
Å
Ç
Ñ
Ö
Xyl
+
PO3, the third-order rate constant was used:
Ä
É
Å
Å
Ñ
Ñ
kcat
h
Å
Å
Å
Å
Å
Ñ
Ñ
ΔG⧧ = −RT ln
.
( )Ñ
Ñ
K
mKHPO κbT
3
AUTHOR INFORMATION
Ñ
Å
Ç
Ñ
Ö
■
Corresponding Author
energy may be identified as entropic contributions for a whole
substrate compared to a substrate in pieces, and to lost
interactions with the bridging CH2OH. This is more evident
upon comparison of the intrinsic binding energies of
David R. J. Palmer − Department of Chemistry, University of
Saskatchewan, Saskatoon, SK, Canada S7N 5C9;
orcid.org/0000-0001-9300-7460; Email: dave.palmer@
usask.ca
phosphonates 1 (ΔG⧧ − ΔG⧧G6P) and 2 (ΔG⧧
−
1
2
ΔG⧧G6P), where 1.5−1.7 kcal/mol can be attributed to
interactions, either electrostatic or steric, with the bridging
oxygen alone, and is consistent with previous observations on
G6PDH and 6PGDH.26,27 This also suggests an approximately
2 kcal/mol advantage is gained from the binding of the whole
Author
Natasha D. Vetter − Department of Chemistry, University of
Saskatchewan, Saskatoon, SK, Canada S7N 5C9;
orcid.org/0000-0002-0963-8858
substrate compared to the two substrate fragments (ΔGXP
−
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.biochem.1c00283
ΔG1/2).31 The estimated intrinsic binding energy of the
phosphate is smaller than those observed for other enzymes,
but this may be because the chemical step of the whole, natural
substrate is not fully rate-limiting for NtdC.
Funding
This work was funded by an NSERC Discovery Grant to
D.R.J.P. and an NSERC PGS award and University of
Saskatchewan Gerhard Herzberg Memorial Fellowship to
N.D.V.
CONCLUSION
■
We showed previously that NtdC, the enzyme that catalyzes
the first step in kanosamine biosynthesis, can accept non-
natural substrate analogues and that the phosphoryl group of
G6P is crucial for efficient catalysis. Here we show that
phosphonate analogues 1 and 2 can be oxidized by NtdC with
∼10% of the efficiency of the natural substrate as measured by
second-order rate constants, comparable to the effect of
replacement of the C5 oxygen atom with a methylene group.
In addition, we have shown that NtdC-catalyzed oxidation of
D-xylose proceeds at a rate that is 40000-fold lower than that of
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank K. Brown, P. Zhu, K. Fransishyn, K. Thoms,
and other staff of the Saskatchewan Structural Sciences Centre
for their expertise and D. Sanders for helpful discussions and
shared equipment.
1931
https://doi.org/10.1021/acs.biochem.1c00283
Biochemistry 2021, 60, 1926−1932

Biochemistry
pubs.acs.org/biochemistry
Article
by triosephosphate isomerase: enzyme activation by phosphite
dianion. J. Phys. Org. Chem. 27, 269−276.
(19) Reyes, A. C., Zhai, X., Morgan, K. T., Reinhardt, C. J., Amyes,
T. L., and Richard, J. P. (2015) The activating oxydianion binding
domain for enzyme-catalyzed proton transfer, hydride transfer, and
decarboxylation: Specificity and enzyme architecture. J. Am. Chem.
Soc. 137, 1372−1382.
ABBREVIATIONS
■
3aX, 3-amino-3-deoxy-D-xylose; G6P, D-glucose 6-phosphate;
G6PDH, G6P dehydrogenase; K6P, kanosamine 6-phosphate;
2KG, 2-ketoglutarate; 3oG6P, 3-oxo-D-glucose 6-phosphate;
3oX, 3-oxo-D-xylose; 6PGDH, 6-phosphogluconate dehydro-
genase; PGI, G6P isomerase; PGM, phosphoglucomutase.
(20) Fernandez, P. L., Nagorski, R. W., Cristobal, J. R., Amyes, T. L.,
and Richard, J. P. (2021) Phosphodianion Activation of Enzymes for
Catalysis of Central Metabolic Reactions. J. Am. Chem. Soc. 143,
2694−2698.
(21) Berkowitz, D. B., Bose, M., Pfannenstiel, T. J., and Doukov, T.
(2000) α-Fluorinated phosphonates as substrate mimics for glucose 6-
phosphate dehydrogenase: the CHF stereochemistry matters. J. Org.
Chem. 65, 4498−4508.
(22) Blackburn, G. M., and Ingleson, D. (1978) Specific dealkylation
of phosphonate esters using iodotrimethylsilane. J. Chem. Soc., Chem.
Commun., 870−871.
(23) Blackburn, G. M., and Ingleson, D. (1980) The dealkylation of
phosphate and phosphonate esters by iodotrimethylsilane: a mild and
selective procedure. J. Chem. Soc., Perkin Trans. 1, 1150−1153.
(24) Zygmunt, J., Kafarski, P., and Mastalerz, P. (1978) Preparation
of Oxoalkanephosphonic Acids. Synthesis 1978, 609−612.
(25) Berkowitz, D. B., and Bose, M. (2001) (α-Monofluoroalkyl)-
phosphonates: a class of isoacidic and “tunable” mimics of biological
phosphates. J. Fluorine Chem. 112, 13−33.
(26) Adams, P. R., Harrison, R., Inch, T. D., and Rich, P. (1976)
Dehydrogenation of the phosphonate analogue of glucose 6-
phosphate by glucose 6-phosphate dehydrogenase. Biochem. J. 155,
1−4.
REFERENCES
■
(1) Milner, J. L., Silo-Suh, L., Lee, J. C., He, H., Clardy, J., and
Handelsman, J. (1996) Production of kanosamine by Bacillus cereus
UW85. Appl. Environ. Microbiol. 62, 3061−3065.
(2) Janiak, A. M., and Milewski, S. (2001) Mechanism of antifungal
action of kanosamine. Med. Mycol. 39, 401−408.
(3) Guo, J., and Frost, J. W. (2002) Kanosamine biosynthesis: A
likely source of the aminoshikimate pathway’s nitrogen atom. J. Am.
Chem. Soc. 124, 10642−10643.
(4) Floss, H. G., Yu, T.-W., and Arakawa, K. (2011) The
biosynthesis of 3-amino-5-hydroxybenzoic acid (AHBA), the
precursor of mC7N units in ansamycin and mitomycin antibiotics:
a review. J. Antibiot. 64, 35−44.
(5) Umezawa, S., Shibahara, S., Omoto, S., Takeuchi, T., and
Umezawa, H. (1968) Studies on the biosynthesis of 3-amino-3-deoxy-
D-glucose. J. Antibiot. 21, 485−491.
(6) Kevany, B. M., Rasko, D. A., and Thomas, M. G. (2009)
Characterization of the complete zwittermicin a biosynthesis gene
cluster from Bacillus cereus. Appl. Environ. Microbiol. 75, 1144−1155.
(7) Shafi, J., Tian, H., and Ji, M. (2017) Bacillus species as versatile
weapons for plant pathogens: a review. Biotechnol. Biotechnol. Equip.
31, 446−459.
(8) Vetter, N. D., Langill, D. M., Anjum, S., Boisvert-Martel, J.,
Jagdhane, R. C., Omene, E., Zheng, H., van Straaten, K. E., Asiamah,
I., Krol, E. S., Sanders, D. A. R., and Palmer, D. R. J. (2013) A
previously unrecognized kanosamine biosynthesis pathway in Bacillus
subtilis. J. Am. Chem. Soc. 135, 5970−5973.
(9) Boisvert-Martel, J. (2013) NtdB: A kanosamine-6-phosphate
phosphatase. Ph.D. Thesis, University of Saskatchewan, Saskatoon,
SK.
(27) Roach, D. J. W., and Harrison, R. (1981) Dehydrogenation by
gluconate 6-phosphate dehydrogenase of an isosteric phosphonate
substrate analogue. Biochem. J. 197, 731−735.
(28) Langill, D. M. (2010) Synthesis of neotrehalose; kinetics and
mutagenesis of NtdC. Ph.D. Thesis, University of Saskatchewan,
Saskatoon, SK.
(29) Ray, W. J., Long, J. W., and Owens, J. D. (1976) An analysis of
the substrate-induced rate effect in the phosphoglucomutase system.
Biochemistry 15, 4006−4017.
(10) Prasertanan, T., and Palmer, D. R. J. (2019) The kanosamine
biosynthetic pathway in Bacillus cereus UW85: Functional and kinetic
characterization of KabA, KabB, and KabC. Arch. Biochem. Biophys.
676, 108139.
(30) Ray, W. J., and Long, J. W. (1976) Thermodynamics and
mechanism of the PO3 transfer process in the phosphoglucomutase
reaction. Biochemistry 15, 3993−4006.
(31) Jencks, W. P., and Meister, A. (2006) Binding Energy,
Specificity, and Enzymatic Catalysis: The Circe Effect. Adv. Enzymol.,
219−410.
(11) Fukui, S., and Hayano, K. (1969) Micro methods for
determination of 3-ketosucrose and 3-ketoglucose. Agric. Biol. Chem.
33, 1013−1017.
(12) Hayano, K., Tsubouchi, Y., and Fukui, S. (1973) 3-Ketoglucose
reductase of Agrobacterium tumefaciens. J. Bacteriol. 113, 652−657.
̌
(13) Volc, J., Sedmera, P., Halada, P., Prikrylova, V., and Daniel, G.
(1998) C-2 and C-3 oxidation of D-Glc, and C-2 oxidation of D-Gal
by pyranose dehydrogenase from Agaricus bisporus. Carbohydr. Res.
310, 151−156.
(14) Kojima, K., Tsugawa, W., and Sode, K. (2001) Cloning and
Expression of Glucose 3-Dehydrogenase from Halomonas sp. α-15 in
Escherichia coli. Biochem. Biophys. Res. Commun. 282, 21−27.
(15) Vetter, N. D., and Palmer, D. R. J. (2017) Simultaneous
Measurement of Glucose-6-phosphate 3-Dehydrogenase (NtdC)
Catalysis and the Nonenzymatic Reaction of Its Product: Kinetics
and Isotope Effects on the First Step in Kanosamine Biosynthesis.
Biochemistry 56, 2001−2009.
(16) Vetter, N. D., Jagdhane, R. C., Richter, B. J., and Palmer, D. R.
J. (2020) Carbocyclic Substrate Analogues Reveal Kanosamine
Biosynthesis Begins with the α-Anomer of Glucose 6-Phosphate.
ACS Chem. Biol. 15, 2205−2211.
(17) Tsang, W.-Y., Amyes, T. L., and Richard, J. P. (2008) A
Substrate in Pieces: Allosteric Activation of Glycerol 3-Phosphate
Dehydrogenase (NAD+) by Phosphite Dianion. Biochemistry 47,
4575−4582.
(18) Zhai, X., Malabanan, M. M., Amyes, T. L., and Richard, J. P.
(2014) Mechanistic imperatives for deprotonation of carbon catalyzed
1932
https://doi.org/10.1021/acs.biochem.1c00283
Biochemistry 2021, 60, 1926−1932