Mechanistic investigation of the radical S -adenosyl- L -methionine enzyme DesII using fluorinated analogues

Article abstract of DOI:10.1021/jacs.5b02545

DesII is a radical S-adenosyl-l-methionine (SAM) enzyme that can act as a deaminase or a dehydrogenase depending on the nature of its TDP-sugar substrate. Previous work has implicated a substrate-derived, C3-centered α-hydroxyalkyl radical as a key interm

Full text of DOI:10.1021/jacs.5b02545

Communication
pubs.acs.org/JACS
Mechanistic Investigation of the Radical S‑Adenosyl‑L‑methionine
Enzyme DesII Using Fluorinated Analogues
†
†
‡
,†,‡
Geng-Min Lin, Sei-Hyun Choi, Mark W. Ruszczycky, and Hung-wen Liu*
†
Department of Chemistry, University of Texas at Austin, Austin, Texas 78712, United States
‡
Division of Medicinal Chemistry, College of Pharmacy, University of Texas at Austin, Austin, Texas 78712, United States
S Supporting Information
*
8
catalyzes the elimination of trimethylamine from choline. In
contrast, however, the DesII reaction is initiated by the reductive
ABSTRACT: DesII is a radical S-adenosyl-L-methionine
SAM) enzyme that can act as a deaminase or a
(
homolysis of SAM via electron transfer from an active-site [4Fe−
+
6
dehydrogenase depending on the nature of its TDP-
sugar substrate. Previous work has implicated a substrate-
derived, C3-centered α-hydroxyalkyl radical as a key
intermediate during catalysis. Although deprotonation of
the α-hydroxyalkyl radical has been shown to be important
for dehydrogenation, much less is known regarding the
course of the deamination reaction. To investigate the role
played by the C3 hydroxyl during deamination, 3-deutero-
4
S] cluster. This produces a 5′-deoxyadenosyl radical (5′-
•
dAdo-CH ) that subsequently abstracts the C3-hydrogen atom
from 1 (see Schemes 2 and 3). The resulting substrate radical 7
2
6
is thus activated for the elimination of ammonia; however, the
mechanism by which this takes place is unclear.
Of particular interest is the role played by the C3-hydroxyl in
the deamination of 1. One hypothesis is that the C3-hydroxyl is
not directly involved in the subsequent deamination. In this case,
the key step is proposed to be a radical-induced 1,2-migration of
the protonated C4-amino functionality to generate a C3-
carbinolamine-C4-radical (7 → 8), as shown in Scheme 2.
3-fluoro analogues of both substrates were prepared and
characterized with DesII. In neither case was deamination
or oxidation observed; however, in both cases deuterium
was efficiently exchanged between the substrate analogues
and SAM. These results imply that the C3 hydroxyl plays a
key role in both reactionsthereby arguing against a 1,2-
migration mechanism of deaminationand that homolysis
of SAM concomitant with H atom abstraction from the
substrate is readily reversible when forward partitioning is
inhibited.
Scheme 2. Deamination via 1,2-Migration
1
esII is a radical S-adenosyl-L-methionine (SAM) enzyme
D
that is responsible for the key step in the biosynthesis of
2
,3
TDP-D-desosamine, which is required for the glycosylation of
various macrolide antibiotics such as erythromycin and
3
−5
pikromycin.
Specifically, DesII acts as a lyase to catalyze the
radical-mediated deamination of TDP-4-amino-4,6-dideoxy-D-
glucose (1) to yield TDP-4,6-dideoxy-3-keto-D-glucose (2) (see
3
,6
Scheme 1). This chemistry is similar to the radical deamination
reactions catalyzed by B -dependent ethanolamine ammonia
1
2
7
lyase (EAL) and the glycyl-radical enzyme CutC, which
Intermediate 8 then either undergoes reduction to a carbinol-
amine followed by elimination of ammonia (8 → 10 → 2) or
eliminates ammonia prior to the reduction (8 → 9 → 2). Such a
Scheme 1. Reactions Catalyzed by DesII
1
,2-migration mechanism is analogous to the current models of
B -dependent catalysis for EAL and diol dehydratase and is
12
supported by electron paramagnetic resonance and isotope-
7,9,10
labeling studies.
The 1,2-migration mechanism is also
consistent with gas-phase computations on the EAL system,
where the corresponding transition state was found to be
energetically reasonable and, in most of the cases considered,
Received: March 10, 2015
©
XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b02545
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
more favorable than mechanisms that do not involve direct
migration.
cannot be easily oxidized. To this end, we designed and
synthesized C3-fluorinated 5 as a probe to test the importance of
the C3-hydroxyl during DesII catalysis. Because of the increased
1
1
Despite evidence for the 1,2-migration mechanism, experi-
ments involving the alternative DesII substrate TDP-D-
quinovose (3) suggest that the C3-hydroxyl may actually play a
more direct role in the deamination reaction. TDP-D-quinovose
is the C4-hydroxy analogue of 1 and is readily accepted by DesII
17
electronegativity of a fluoro substituent, its presence should
impede the redistribution of electron density required for direct
elimination (Scheme 3). In this case, deamination of 5 would not
be observed. In contrast, the C3-fluoro group should have less
impact on a radical-induced 1,2-migration of the amino group
(Scheme 2). The result would be a 3-amino-3-fluoro species that
subsequently decomposes to give 2. We also prepared the
analogue 6 to determine the effect of C3-fluoro substitution on
the DesII-catalyzed oxidation reaction. Deuterium was intro-
duced at C3 of both analogues in order to monitor C3-radical
generation even in the absence of net turnover. Our data indicate
that the C3-hydroxyl is indeed required for both the deamination
and oxidation reactions. Reported herein are our observations
with these compounds in the DesII system along with a
discussion of the mechanistic implications.
1
2,13
as a substrate.
As is the case with 1, TDP-D-quinovose also
undergoes C3-hydrogen atom abstraction within the DesII active
site, which affords an α-hydroxyalkyl radical intermediate
1
4
analogous to 7; however, the reaction does not result in the
6
,12
corresponding dehydration but rather C3-dehydrogenation.
This is in stark contrast to expectation because water is rapidly
eliminated from 1,2-dihydroxyalkyl radicals generated via pulse
15
radiolysis in bulk media. Furthermore, the B -dependent diol
1
2
dehydratases are well-known to catalyze analogous radical-
mediated dehydration reactions rather than dehydrogen-
9
,10
ations.
indicated that deprotonation of the C3 α-hydroxyalkyl radical of
to generate a ketyl radical intermediate is necessary to facilitate
Interestingly, kinetic isotope effect measurements
The deuterated 4-amino-3-fluoro substrate analogue 5D was
prepared according to Scheme 4. The protected azide 19 was
3
2
+
electron transfer to the [4Fe−4S] cluster and thus complete
13
the oxidation of 3 to 4. In light of these results, there is good
reason to believe that the DesII active site contains a residue that
can deprotonate the C3 α-hydroxyalkyl radical of whichever
substrate is bound.
Scheme 4. Synthesis of 4-Amino-3-fluoro Analogue 5D
These observations suggest the alternative deamination
mechanism shown in Scheme 3, in which the amino group is
Scheme 3. Deamination via Direct Elimination
eliminated from C4 directly following the deprotonation of 7 to
form a ketyl radical intermediate (i.e., 11). Despite parallels with
the dehydrogenation reaction (i.e., 3 → 4), such an E1cB-type
obtained from methyl α-D-mannopyranoside (14) in a manner
18
similar to a previously reported synthesis. Intermediate 17 was
sequentially oxidized and reduced to invert the stereochemistry
at C4 (17 → 18) prior to installation of an azide moiety using a
Mitsunobu-type azidation reaction (18 → 19). Acidic hydrolysis
of isopropylidene 19 and selective benzyl protection of the C2-
elimination of NH (11 → 13) conflicts with the hypothesis that
3
the formation of a strongly reducing ketyl radical promotes
2
+
16
electron transfer back to the [4Fe−4S] cluster. However, a
process wherein the NH leaving group accepts a proton from
3
19
the C3 α-hydroxyalkyl radical in concert with its elimination (i.e.,
hydroxyl (20 → 21) permitted the equatorial introduction of
7
→ 12) not only would resolve this conflict but also is consistent
deuterium at C3 via Swern oxidation followed by NaBD
4
20
with computations that imply a similar activation energy
reduction (21 → 22). Epimerization at C2 occurred during
this reaction. Fluorination of 22 using diethylaminosulfur
trifluoride (DAST) gave 23 with inversion of stereochemistry
11
compared with migration. In this case, the putative active-site
base may operate to facilitate the proton transfer as shown in
Scheme 3, though a general acid may alternatively serve as a
proton donor counterpart to the base.
One way to evaluate these mechanistic alternatives (i.e., 1,2-
migration vs direct elimination) is to consider the effect of
replacing the C3 hydroxyl with a less nucleophilic moiety that
20
at C3. This was followed by conversion to dibenzyl phosphate
25 via a glycosyl phosphite intermediate with the α-anomer as
the major product. Hydrogenolysis of 25 led to benzyl
deprotection and reduction of the C4-azido moiety to give 26.
Coupling of 26 with TMP-morpholidate provided 5D.
B
DOI: 10.1021/jacs.5b02545
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Compound 6D was prepared in a similar manner (see the
Supporting Information (SI)).
reaction. This was further substantiated by the observation of low
levels of 5′-deoxyadenosine at retention times earlier than 5 min.
One possible explanation for the lack of substrate turnover
despite 5′-deoxyadenosyl radical generation is poor formation of
a productive Michaelis complex between DesII and the 3-fluoro
analogues. Furthermore, an α-fluoro substituent is known to be
less effective at stabilizing an alkyl radical in comparison to an α-
When 5D or 6D (250 μM) was incubated at room
temperature for up to 8 h in the presence of 300 μM SAM, 1
mM Na S O , 25 mM 3-[4-(2-hydroxyethyl)piperazin-1-yl]-
2
2
4
propane-1-sulfonate (EPPS) buffer (50 μL, pH 8.0) and DesII
3.7 to 30 μM), no consumption of the corresponding fluoro
(
1
7,22
analogue was observed by HPLC (see Figure 1 and Figure 1S in
hydroxyl group.
The combined effect of poor substrate
binding and inefficient H atom transfer could thus preclude C3-
radical generation, making the C3-fluoro analogues ineffective
mechanistic probes.
To investigate this possibility, the DesII reactions with 5D and
6D were monitored for deuterium exchange between the fluoro
analogues and SAM. Reaction mixtures containing 600 μM 5D
or 6D, 300 μM SAM, 500 μM Na S O , and 20 μM DesII in 400
2
2
4
μL of 25 mM EPPS buffer (pH 8.0) were prepared. The reactions
were carried out under anaerobic conditions at 30 °C. At 2 h
intervals, additional fresh Na S O was added and a 100 μL
2
2
4
aliquot was removed. HPLC was then used to isolate the residual
TDP-sugar substrates and SAM, which were subsequently
analyzed by ESI-MS. Deuterium enrichment of each species
was determined from the relative MS peak intensities after
13
correction for natural-abundance C.
Analysis of the changes in isotopic enrichment as a function of
reaction time demonstrated that both 5D and 6D underwent
significant depletion in deuterium content while at the same time
SAM was enriched in both the mono- and dideuterated
isotopologues (see Figure 3S). The exchange reactions with 20
μM DesII began to plateau within the first 2 h of incubation.
Furthermore, at the 8 h time point, the fractional concentrations
of monodeuterated 5 and 6 were 50% and 60%, respectively.
Given the initial 2:1 ratio of the deuterated substrate
isotopologue (one exchangeable hydrogen) and SAM (two
exchangeable hydrogens), these values approach the expected
equilibrium value of 50%, assuming that there is little difference
Figure 1. HPLC traces for incubations involving 5D, SAM, and DesII in
the presence of Na S O . Reactions were run anaerobically at room
2
2
4
24
temperature in 25 mM EPPS buffer (pH 8.0). (A) Constant-time
incubations of 2 h with variable DesII concentration (3.7−30 μM,
indicated on each trace) and 250 μM 5D, 300 μM SAM, and 1 mM
in the isotopic fractionation factor associated with the C3
position of the substrate and the 5′ position of SAM. Likewise,
the expected enrichments of mono- and dideuterated SAM are
50% and 25%, respectively, which are similar to the values of ca.
Na S O . (B) Variable-time incubations of 2−8 h (indicated on each
2
2
4
trace) with 3.7 μM DesII, 250 μM 5D, 300 μM SAM, and 1 mM
5
0% and 20% observed at 8 h in the reactions with both C3-
Na S O . Methylthioadenosine also contributes to the peaks at early
2
2
4
fluoro analogues. Finally, analysis of the 5′-deoxyadenosine
produced demonstrated greater than 95% monodeuteration at
early time points in the reaction.
23
retention times (peak a, <5 min), and the minor peaks b correspond to
thymidine monophosphate (TMP), impurities in the SAM preparation,
and 3-epi-5.
The observation of efficient hydrogen exchange between the
C3-fluoro analogues and SAM in the presence of DesII indicates
that both substrate analogues undergo C3-hydrogen atom
abstraction by the 5′-deoxyadenosyl radical generated in the
active site. Careful inspection of HPLC traces from the reaction
with 5D also revealed a minor peak (<5%) consistent with 3-epi-
5 by ESI-MS (see Figure 2S). This minor product was not
isotopically enriched beyond natural abundance and can account
for the observed low levels of monodeuterated 5′-deoxyadeno-
sine in these experiments. The epimerization likely results from
reduction of the C3-radical intermediate 7 by a solvent-derived H
atom equivalent, as has been previously reported in the case of
the SI). Changes in the analogue concentration (0.25 to 2 mM)
had no significant impact on the appearance of the HPLC traces
(
data not shown). Furthermore, there was no visible develop-
ment of new HPLC peaks in the 30−40 min region, where the
deamination and dehydration products (i.e., 2 and 4) are
6
,12
expected to elute. However, slow consumption of SAM was
noted in both cases. The disappearance of SAM was dependent
on both the enzyme concentration (Figure 1A) and the
incubation time (Figure 1B) and was accompanied by the
appearance of a new HPLC peak with a retention time of 27.3
min. Collection and electrospray ionization mass spectrometry
21
TDP-D-fucose.
(
ESI-MS) analysis of this new peak demonstrated a signal at m/z
The inability of 5D and 6D to undergo elimination or
oxidation is thus not due to a failure to generate the substrate
radical but is a consequence of the nonreactivity of the C3-
fluoroalkyl radicals within the DesII active site. These results are
mechanistically significant for three reasons. First, the fact that
formation of a C3-radical adduct of 5 does not prompt
deamination disfavors a mechanism involving a radical-induced
3
16.1 (positive ion), which is consistent with formation of 5′-
+
deoxyadenosylsulfinate (5′-dAdo-SO H, predicted [M + H] m/
2
−
z 316.07). The generation of 5′-dAdo-SO during prolonged
2
incubations of SAM and Na S O with elevated concentrations
2
2
4
21
of DesII has been reported previously and implies that
formation of the 5′-deoxyadenosyl radical indeed occurs in the
C
DOI: 10.1021/jacs.5b02545
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
,2-migration of the amino group. Second, the results further
Communication
1
REFERENCES
■
substantiate the hypothesis that an increase in oxidation potential
of the substrate radical is required for oxidation, since the
ionization energy of a fluoromethyl radical is approximately 1 eV
(
1) (a) Frey, P. A.; Hegeman, A. D.; Ruzicka, F. J. Crit. Rev. Biochem.
Mol. Biol. 2008, 43, 63−88. (b) Broderick, J. B.; Duffus, B. R.; Duschene,
K. S.; Shepard, E. M. Chem. Rev. 2014, 114, 4229−4317.
(2) Xue, Y.; Zhao, L.; Liu, H.-w.; Sherman, D. H. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 12111−12116.
3) Szu, P.-h.; He, X.; Zhao, L.; Liu, H.-w. Angew. Chem., Int. Ed. 2005,
4, 6742−6746.
4) Zhao, L.; Borisova, S.; Yeung, S.-M.; Liu, H.-w. J. Am. Chem. Soc.
2001, 123, 7909−7910.
5) Ruszczycky, M. W.; Ogasawara, Y.; Liu, H.-w. Biochim. Biophys.
Acta 2012, 1824, 1231−1244.
6) Szu, P.-H.; Ruszczycky, M. W.; Choi, S.-h.; Yan, F.; Liu, H.-w. J. Am.
1
7,25
greater than that of a hydroxymethyl radical.
Third, the
apparent equilibration of deuterium between the 3-fluoro
analogues and SAM indicates that reductive homolysis of SAM
and H atom abstraction from the substrate are readily reversible.
Previous reports have provided evidence for this possibility
(
4
(
6
,26,27
under turnover conditions,
and enzymes such as lysine
(
2
8
29 30
amino mutase, spore photoproduct lyase, and QueE are
well-known to regenerate SAM as part of their catalytic cycles.
However, the present results suggest that the reverse partitioning
of the substrate radical, methionine, and 5′-deoxyadenosine to
regenerate SAM and the substrate (i.e., 7 → 1) can become the
dominant course of the reaction when forward partitioning (i.e.,
(
Chem. Soc. 2009, 131, 14030−14042.
(7) Bandarian, V.; Reed, G. H. In Chemistry and Biochemistry of B ;
1
2
Banerjee, R., Ed.; John Wiley & Sons: New York, 1999; pp 811−833.
8) Craciun, S.; Marks, J. A.; Balskus, E. P. ACS Chem. Biol. 2014, 9,
(
1
(
(
(
8
(
1
408−1413.
7
→ 8, 11, or 12) is impeded.
Taken together, these results indicate that the mechanism of
9) Toraya, T. Chem. Rev. 2003, 103, 2095−2127.
10) Toraya, T. Arch. Biochem. Biophys. 2014, 544, 40−57.
11) Semialjac, M.; Schwarz, H. J. Am. Chem. Soc. 2002, 124, 8974−
the DesII-catalyzed deamination reaction appears to be most
consistent with that shown in Scheme 3. Upon formation of the
ternary Michaelis complex between DesII, SAM, and 1, SAM is
reduced concomitant with C3-hydrogen atom abstraction from
the substrate to produce the initial substrate radical 7. This
process is reversible, and the 5′-deoxyadenosyl or substrate
radical may also be quenched unproductively to result in
uncoupling, sulfinate formation, or C3-epimerization. However,
these latter pathways are relatively minor in the case of DesII
unless the substrate exhibits significant structural perturba-
983.
12) Ruszczycky, M. W.; Choi, S.-h.; Liu, H.-w. J. Am. Chem. Soc. 2010,
32, 2359−2369.
(13) Ruszczycky, M. W.; Choi, S.-h.; Liu, H.-w. Proc. Natl. Acad. Sci.
U.S.A. 2013, 110, 2088−2093.
(14) Ruszczycky, M. W.; Choi, S.-h.; Mansoorabadi, S. O.; Liu, H.-w. J.
Am. Chem. Soc. 2011, 133, 7292−7295.
15) (a) Steenken, S.; Davies, M. J.; Gilbert, B. C. J. Chem. Soc., Perkin
Trans. 2 1986, 1003−1010. (b) Steenken, S. J. Phys. Chem. 1979, 83,
595−599. (c) Bansal, K. M.; Gratzel, M.; Henglein, A.; Janata, E. J. Phys.
(
2
1
̈
tions.
Chem. 1973, 77, 16−19.
The subsequent reaction of 7 depends on the presence of an
ionizable hydroxyl group at C3, since radical-induced deami-
nation of 5 does not take place. Instead, deamination likely
proceeds via a mechanism of direct elimination wherein the
proton from the C3 α-hydroxyalkyl radical is transferred to the
amine leaving group at C4 (i.e., 7 → 12). This may be mediated
by the putative active-site base and likely proceeds in a concerted
manner to avoid possible dehydrogenation due to the formation
(
(
(
16) Ruszczycky, M. W.; Liu, H.-w. Isr. J. Chem. 2015, 55, 1−11.
17) Dolbier, W. R., Jr. Chem. Rev. 1996, 96, 1557−1584.
18) Bundle, D. R.; Gerken, M.; Peters, T. Carbohydr. Res. 1988, 174,
239−251.
(19) Peters, T.; Bundle, D. R. Can. J. Chem. 1989, 67, 497−502.
́ ̌
(20) Poirot, E.; Chang, A. H. C.; Horton, D.; Kovac, P. Carbohydr. Res.
2001, 334, 195−205.
21) Ko, Y.; Ruszczycky, M. W.; Choi, S.-H.; Liu, H.-w. Angew. Chem.,
(
13
Int. Ed. 2015, 54, 860−863.
of a ketyl radical (i.e., 11) as a discrete intermediate. The
(
(
(
(
9
22) Bordwell, F. G.; Zhang, X.-M. Acc. Chem. Res. 1993, 26, 510−517.
23) (a) Iwig, D. F.; Booker, S. J. Biochemistry 2004, 43, 13496−13509.
b) Hoffman, J. L. Biochemistry 1986, 25, 4444−4449.
24) (a) Hartshorn, S. R.; Shiner, V. J., Jr. J. Am. Chem. Soc. 1972, 94,
002−9012. (b) Venkatasubban, K. S.; Schowen, R. L. CRC Crit. Rev.
Biochem. 1984, 17, 1−44.
25) Dyke, J. M.; Ellis, A. R.; Jonathan, N.; Keddar, N.; Morris, A.
Chem. Phys. Lett. 1984, 111, 207−210.
26) Yokoyama, K.; Numakura, M.; Kudo, F.; Ohmori, D.; Eguchi, T. J.
catalytic cycle is then completed upon reduction of enol radical
1
2 by an external electron donor in a process that has yet to be
clarified. Overall, the use of C3-fluoro analogues to investigate
the DesII reaction has shed significant light on its mechanism of
catalysis and radical SAM enzymes in general.
(
ASSOCIATED CONTENT
■
(
*
S
Supporting Information
Am. Chem. Soc. 2007, 129, 15147−15155.
Synthetic procedures, enzyme assays, and spectroscopic
characterization of all compounds discussed. This material is
available free of charge via the Internet at http://pubs.acs.org.
(27) Duffus, B. R.; Ghose, S.; Peters, J. W.; Broderick, J. B. J. Am. Chem.
Soc. 2014, 136, 13086−13089.
28) (a) Chirpich, T. P.; Zappia, V.; Costilow, R. N.; Barker, H. A. J.
(
Biol. Chem. 1970, 245, 1778−1789. (b) Moss, M.; Frey, P. A. J. Biol.
Chem. 1987, 262, 14859−14862. (c) Baraniak, J.; Moss, M. L.; Frey, P.
A. J. Biol. Chem. 1989, 264, 1357−1360.
AUTHOR INFORMATION
■
Corresponding Author
(29) (a) Cheek, J.; Broderick, J. B. J. Am. Chem. Soc. 2002, 124, 2860−
2
2
(
861. (b) Buis, J. M.; Cheek, J.; Kalliri, E.; Broderick, J. B. J. Biol. Chem.
006, 281, 25994−26003.
*h.w.liu@mail.utexas.edu
Notes
30) McCarty, R. M.; Krebs, C.; Bandarian, V. Biochemistry 2013, 52,
The authors declare no competing financial interest.
188−198.
ACKNOWLEDGMENTS
■
We dedicate this paper to Prof. Koji Nakanishi on the occasion of
his 90th birthday. This work was supported by grants from the
National Institutes of Health (GM035906) and the Welch
Foundation (F-1511).
D
DOI: 10.1021/jacs.5b02545
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX