Catalytic Mechanism of S-Ribosylhomocysteinase (LuxS): Direct Observation of Ketone Intermediates by 13C NMR Spectroscopy

Article abstract of DOI:10.1021/ja0369663

S-Ribosylhomocysteinase (LuxS) catalyzes the cleavage of the thioether linkage of S-ribosylhomocysteine (SRH) to produce l-homocysteine and 4,5-dihydroxy-2,3-pentanedione (DHPD). This is a key step in the biosynthetic pathway of the type II autoinducer (AI-2) in both Gram-positive and Gram-negative bacteria. Previous studies demonstrated that LuxS contains a catalytically essential Fe²⁺ ion. The catalytic mechanism of LuxS was investigated using 2- and 3-¹³C-labeled SRH as substrate and ¹³C NMR spectroscopy. These studies revealed the presence of 2- and 3-keto intermediates in the catalytic pathway. The 2-keto intermediate was chemically synthesized, and its chemical and kinetic competence was demonstrated. The results support a catalytic mechanism in which the metal ion catalyzes an internal redox reaction, shifting the carbonyl group from the C-1 position to the C-3 position. Subsequent β-elimination at the C-4 and C-5 positions releases homocysteine as a free thiol. The results also suggest that Cys-84 and Glu-57 are the possible general acids/bases for proton transfer during catalysis and that the keto intermediates are released from the enzyme active site before rebinding and completion of the reaction. Copyright

Full text of DOI:10.1021/ja0369663

Published on Web 10/11/2003
Catalytic Mechanism of S-Ribosylhomocysteinase (LuxS): Direct Observation
of Ketone Intermediates by 13C NMR Spectroscopy
Jinge Zhu, Xubo Hu, Eric Dizin, and Dehua Pei*
Department of Chemistry and Ohio State Biochemistry Program, The Ohio State UniVersity, 100 West 18th AVenue,
Columbus, Ohio 43210
Received June 29, 2003; E-mail: pei.3@osu.edu
Bacterial cell-cell communication, i.e., quorum sensing, is
mediated through the production, release, and detection of small
signaling molecules called autoinducers (AIs).1 There are two types
of AIs. The type 1 AIs (AI-1s) are species-specific and are
acylhomoserine lactone and oligopeptide derivatives for Gram-
negative and Gram-positive bacteria, respectively. The type 2 AI
(AI-2) is a universal signal shared by all bacteria, presumably for
interspecies communication.2 AI-2 is biosynthesized from S-
adenosylhomocysteine (SAH), which is hydrolyzed to adenine and
S-ribosylhomocysteine (SRH) by the nucleosidase Pfs (Figure 1).
Subsequently, SRH is converted into homocysteine and 4,5-
dihydroxy-2,3-pentanedione (DHPD) by S-ribosylhomocysteinase
Figure 1. Biosynthetic pathway of AI-2.
(LuxS).3-5 DHPD spontaneously cyclizes and complexes with
borate to form a furanosyl borate diester as the AI-2.6
δ 213.3 peak remained a singlet, while the other peaks were split
into doublets. When the LuxS reaction was quenched with methanol
and incubated with NH2OH, the δ 213.3 peak disappeared, while
two new peaks appeared at δ 160.9 and 161.2. These results indicate
that the peak at δ 213.3 is the 13C-labeled 2-keto intermediate 1.
The two peaks at δ ∼161 upon treatment with NH2OH correspond
to the cis and trans isomers of the oxime derivative. When the above
experiment was repeated with a catalytically impaired C84D mutant,
the slow accumulation of the δ 213.3 peak was observed (Figure
3b). Most importantly, the appearance of ketone 1 preceded the
products, suggesting that the 2-keto species is a bona fide
intermediate of the LuxS reaction. A control experiment conducted
with an inactive mutant (C84A) resulted in no signal for either the
intermediate or product.
The involvement of 3-keto intermediate 2 was investigated using
[3-13C]SRH as substrate. The free substrate has two peaks at δ 74.8
and 74.2 (Figure 3c). Upon incubation with E57A LuxS, a weak
singlet appeared at δ 214.5 at 2 h, reached maximal intensity at
∼6 h, and decayed after 14 h. Following the appearance of the δ
214.5 peak, two new peaks emerged at δ 100.5 and 100.1 at 4-6
h and remained even after the δ 214.5 signal decayed. We assign
the δ 214.5 peak as the 3-keto intermediate and the signals at δ
100.5 and 100.1 as the hydrated cyclic furanone products. Note
that an intense peak appeared at δ 78.3 immediately after the reac-
tion was initiated and remained throughout the experimental period
(0-14 h). The kinetics of its formation and its chemical shift suggest
that it is the 3-13C-labeled ketone 1. A very weak peak at δ 214.0
was also observed with wild-type LuxS when a spectrum was
acquired over the entire 2.5-h reaction period. The 3-keto inter-
mediate was not detected for C84D LuxS. In fact, the 3-keto
intermediate was also observed as a singlet at δ 77.0 in the above
experiments with [2-13C]SRH as substrate and two of the less active
mutants (E57D and E57A), but not with wild-type or C84D LuxS.
We noted that the peak widths for the keto intermediates (4-6
Hz) were unusually narrow for enzyme-bound species, especially
when the paramagnetic Co2+ ion is expected to further broaden
the bound ketone signals. We performed further experiments to
test whether the keto intermediates were released from the LuxS
The reaction catalyzed by LuxS is mechanistically intriguing; it
is formally the nonredox cleavage of a stable thioether bond. LuxS
contains a tetrahedrally coordinated, catalytically essential Fe2+ ion,
which can be substituted by Co2+ without significant change in
catalytic properties.7 Deuterium was incorporated at the C1, C4,
and C5 positions of DHPD when the LuxS reaction was carried
out in D2O.7 On the basis of these data, we have proposed a metal-
catalyzed carbonyl migration mechanism for the LuxS reaction
(Figure 2).7 SRH exists as an equilibrium between the aldehyde
and hemiacetal forms. Coordination of the aldehyde carbonyl with
the metal ion increases the acidity of the C2 proton, which is
abstracted by a general base (possibly Cys-84 in Bacillus subtilis
LuxS) to form an enolate. Isomerization of the enolate and proton-
ation at C1 forms a 2-keto intermediate (1). Repetition of the above
sequence shifts the keto group to the C3 position to give a 3-keto
intermediate (2). Subsequent â-elimination releases homocysteine
and DHPD. Here we provide direct evidence for the involvement
of the ketone intermediates by taking advantage of the slow turnover
rate of B. subtilis LuxS (kcat ) 0.03 s-1 at 23 °C) and monitoring
reactions in real time by 13C NMR spectroscopy. In addition, the
2-keto intermediate was chemically synthesized and demonstrated
as a true intermediate on the catalytic pathway of LuxS.
To observe the 2-keto intermediate, Co2+-substituted B. subtilis
LuxS was rapidly mixed with 1 equiv of [2-13C]SRH at 4 °C, and
broadband-decoupled 13C NMR spectra were collected every 4 min.
The free substrate (no LuxS) showed two peaks at δ 76.2 and 71.6,
reflecting the presence of two hemiacetal anomers (Figure 3a). Upon
the addition of LuxS, three new peaks appeared immediately (∼4
min) at δ 213.3, 74.9, and 74.0. The intensity of the δ 74.9 and
74.0 peaks increased as the substrate was depleted and reached
maximum when all other peaks disappeared. These two peaks were
assigned as the two furanone products. The peak at δ 213.3 reached
maximal intensity at 4 min, gradually decreased with time, and
finally disappeared at ∼4 h when most of the substrate was depleted.
Similar spectra were obtained with two LuxS mutants, E57D and
E57A, although their reactions were much slower. When a 13C
NMR spectrum was recorded under off-resonance conditions, the
9
10.1021/ja0369663 CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 13379-13381
13379

C O M M U N I C A T I O N S
Figure 2. Proposed mechanism for LuxS-catalyzed reaction. RSH, L-homocysteine.
active site (see Supporting Information for details). First, NMR
experiments with different LuxS/SRH ratios (1:1-8 mol/mol) all
gave the same relative intensity for ketone 1 and SRH signals
(ketone 1/SRH ∼1:6), indicating that the 2-keto intermediate and
SRH are in rapid equilibrium (vide infra). Next, a pulse-chase
experiment showed that addition of excess unlabeled SRH into the
reaction of LuxS and [2-13C]SRH can dramatically slow the decay
of the 13C-labeled 2-keto intermediate. Third, severe substrate
inhibition was observed, especially for the less active mutants.
Taken together, our data strongly argue that the keto intermediates
are at least partially released into the solution during catalysis.
The intermediacy of ketone 1 was further assessed through its
chemical synthesis and kinetic characterization. Ketone 1 is a
2-ketopentose linked to L-homocysteine through a thioether bond.
We decided to prepare the corresponding alditol, couple it to homo-
cysteine, and then oxidize the alditol to a ketose (Scheme 1). Syn-
thesis of the alditol started with commercially available D-lyxose,
which was converted into allyl D-lyxoside, followed by protection
of the 2-, 3-, and 5-hydroxyl groups with the p-methoxybenzyl
(PMB) group. The resulting glycoside 3 (as a mixture of R and â
anomers) was treated with diisobutylaluminum hydride (DIBAL)
in the presence of NiCl2(dppp)8 to give alditol 4.9 Note that oxida-
tion of the 4-OH of alditol 4 gives a ketopentose of the same stereo-
chemical configuration as that in intermediate 1. Coupling of diol
4 and N-Boc-homocysteine tert-butyl ester (5) proved quite
problematic. After numerous other attempts failed, we converted
the diol 4 into cyclic sulfate diester 6 by reacting it with thionyl
chloride followed by oxidation of the resulting sulfite with period-
ate.10 Cyclization should make it more accessible and reactive to
nucleophilic attack. Indeed, sulfate ester 6 readily reacted with
homocysteine 5 to give the desired alcohol 7 in good yield and a
small amount of alcohol 8, due to nucleophilic attack at the second-
ary C4 position. Oxidation of the mixture of alcohols 7 and 8 by
the method of Dess and Martin11 gave ketone 9 and aldehyde 10,
respectively, which were readily separated on a silica gel column.
Treatment of ketone 9 with trifluoroacetic acid removed all of the
protecting groups to afford ketone 1 in 25% overall yield (10 steps).
Ketone 1 is stable for weeks in neutral aqueous solution at room
temperature. The addition of LuxS resulted in rapid release of
homocysteine as detected by a DTNB assay.7 When the LuxS
reaction was carried out in the presence of 1,2-phenylenediamine,
a single quinoxaline derivative (compound 11 in Scheme 2) was
Figure 3. 13C NMR spectra (150 MHz) of LuxS-catalyzed reactions. (A)
isolated and was found to be identical to that from LuxS catalyzed
cleavage of SRH, indicating that DHPD is the second reaction
product. Thus, intermediate 1 is chemically competent.
To demonstrate the kinetic competence of intermediate 1, the
catalytic activities of various LuxS forms toward ketone 1 were
[2-13C]SRH (2.0 mM) and wild-type B. subtilis LuxS (2.0 mM) were mixed
at 4 °C, and spectra were recorded every 4.1 min. (B) [2-13C]SRH (2.8
mM) and C84D LuxS (2.8 mM) were mixed at 4 °C, and spectra were
recorded every 2 h. (C) [3-13C]SRH (1.8 mM) and E57A LuxS (1.4 mM)
were mixed at 4 °C, and spectra were recorded every 2 h.
9
13380 J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003

C O M M U N I C A T I O N S
Table 1. Catalytic Properties of Co(II)-Substituted LuxS Variants toward SRH and 2-Ketone Intermediate 1
SRH
2-ketone intermediate 1
LuxS varianta
k
cat (s-1
)
KM (µM)
kcat/KM (M-1 s-1
)
kcat (s-1
)
KM (µM)
kcat/KM (M-1 s-1
)
WT (Bs)
0.030 ( 0.005
0.48 ( 0.01
ND
1.4 ( 0.1
39 ( 1
ND
2.1 × 104
0.031 ( 0.004
0.48 ( 0.02
0.010 ( 0.001
0.007 ( 0.001
0.030 ( 0.001
1.5 ( 0.3
38 ( 1
29 ( 2
50 ( 4
58 ( 3
2.1 × 104
1.3 × 104
3.4 × 102
1.4 × 102
5.2 × 102
WT (Vh)
1.2 × 104
E57D (Vh)
C83D (Vh)
C83S (Vh)
148
5
18
ND
ND
ND
ND
a Bs, B. subtilis LuxS; Vh, V. harVeyi LuxS; WT, wild-type; ND, not determined.
Scheme 1. Synthesis of 2-Keto Intermediate 1a
faster than its formation. Thus, the rate-limiting step is the shift of
the carbonyl group from C2 to C3 (1 to 2). The C84D mutation
slows down the formation of ketone 1 and causes the disappearance
of ketone 2. The C84A mutant is catalytically inactive and does
not form the 2-keto intermediate. These results suggest that Cys-
84 (Cys-83 in V. harVeyi LuxS) is involved in the early steps of
the reaction, likely responsible for the proton transfer at C1-C3
positions (Figure 2). Increased accumulation of the 3-keto inter-
mediate upon the E57A mutation suggests that Glu-57 plays a
critical role in its decay, presumably acting as the general base for
the final â-elimination reaction. Mutation of Glu-57 also slowed
the formation of the 2-keto intermediate (Table 1), as evidenced
by its peak intensity reaching maximum at ∼4 min for WT LuxS
(Figure 3a) but at 4-6 h for the E57A mutant (Figure 3c). We
propose that Glu-57 is also involved in the early steps of the
mechanism, by acting as a proton shuttle to facilitate the isomer-
ization of the enolate intermediates (Figure 2). Structural studies
show that both Glu-57 and Cys-84 are properly positioned in the
active site to perform the proposed functions.12
In summary, we have obtained direct evidence for the involve-
ment of 2- and 3-ketone intermediates in the LuxS-catalyzed
reaction, validating the catalytic mechanism previously proposed.
The keto intermediates are at least partially released from the
enzyme active site during catalysis, providing a rare example of
one active site catalyzing three distinct chemical reactions. The data
also suggest Cys-84 and Glu-57 as the possible general acids/bases
in the proposed catalytic mechanism.
a Reagents and conditions: (a) allyl alcohol, cat. H2SO4; (b) PMBCl,
Bu4NBr, 50% aq KOH/THF (1/1), 90 °C, 76% for 2 steps; (c) DIBAL, cat
NiCl2(dppp), THF; (d) NaBH4, EtOH, 73% for 2 steps; (e) SOCl2, TEA,
CH2Cl2, -78 °C; (f) NaIO4, cat. RuCl3, CCl4/CH3CN/H2O (4/4/6), 0 °C,
86% for 2 steps; (g) BocNH-CH(CH2CH2SH)CO2-t-Bu (5), BuLi, DMF,
0 °C; (h) H2SO4/H2O/THF, rt, 74% for 2 steps, 7:8 ) 4:1; (i) Dess-Martin,
CH2Cl2, rt, 76%; (j) TFA/anisole (5:1), rt, 93%.
Acknowledgment. We thank Dr. Charles Cottrell for assistance
in the NMR experiments. This work was supported by National
Institutes of Health (AI40575 and GM62820).
Scheme 2
Supporting Information Available: Experimental details and
additional NMR spectra (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Miller, M. B.; Bassler, B. L. Annu. ReV. Microbiol. 2001, 55, 165-199.
(2) Bassler, B. L.; Wright, M.; Showalter, R. E.; Silverman, M. R. Mol.
Microbiol. 1993, 9, 773-786.
determined and compared with those of SRH (Table 1). Wild-type
LuxS from either B. subtilis or Vibrio harVeyi exhibited similar
activity toward the keto intermediate and SRH. This is consistent
with the notion that formation of intermediate 1 is a rapid step in
WT LuxS (vide infra). However, three catalytically compromised
V. harVeyi LuxS mutants (E57D, C83D, and C83S) all had much
higher activities (2-29-fold) toward the keto intermediate than SRH
(Table 1). For example, the C83D LuxS has a kcat value of 0.007
s-1 and a kcat/KM value of 140 M-1 s-1 toward the keto intermediate,
whereas its activity toward SRH was barely detectable by the DTNB
assay (kcat/KM ) 5 M-1 s-1). These results indicate that the 2-keto
intermediate is kinetically competent. They also suggest that
formation of the 2-keto intermediate is at least partially rate limiting
in the mutants.
(3) Miller, C. H.; Duerre, J. A. J. Biol. Chem. 1968, 243, 92-97.
(4) Surette, M. G.; Miller, M. B.; Bassler, B. L. Proc. Natl. Acad. Sci. U.S.A.
1999, 96, 1639-1644.
(5) Schauder, S.; Shokat, K.; Surette, M. G.; Bassler, B. L. Mol. Microbiol.
2001, 41, 463-476.
(6) Chen, X.; Schauder, S.; Potier, N.; Van Dorsselaer, A.; Pelczer, I.; Bassler,
B. L.; Hughson, F. M. Nature 2002, 415, 545-549.
(7) Zhu, J.; Dizin, E.; Hu, X.; Wavreille, A.; Park, J.; Pei, D. Biochemistry
2003, 42, 4717-4726.
(8) Taniguchi, T.; Ogasawara, K. Angew. Chem., Int. Ed. 1998, 37, 1136-
1137.
(9) Reduction by DIBAL gave alditol 3 plus a small amount of hemiacetal,
which was converted into 3 by reduction with NaBH4.
(10) Byun, H.-S.; He, L.; Bittman, R. Tetrahedron 2000, 56, 7051-7091.
(11) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155-4158.
(12) Lewis, H. A.; Furlong, E. B.; Laubert, B.; Eroshkina, G. A.; Batiyenko,
Y.; Adams, J. M.; Bergseid, M. G.; Marsh, C. D.; Peat, T. S.; Sanderson,
W. E.; Sauder, J. M.; Buchanan, S. G. Structure 2001, 9, 527-537. (b)
Hilgers, M. T.; Ludwig, M. L. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
11169-11174. (c) Ruzheinikov, S. N.; Das, S. K.; Sedelnikova, S. E.;
Hartley, A.; Foster, S. J.; Horsburgh, M. J.; Cox, A. G.; McCleod, C. W.;
Mekhalfia, A.; Balckburn, G. M.; Rice, D. W.; Baker, P. J. J. Mol. Biol.
2001, 313, 111-122.
The above results provide additional insight into the LuxS
mechanism. Accumulation of intermediate 1 indicates that its
formation is fast relative to its decay. The lack of significant
accumulation of intermediate 2 for WT LuxS suggests that it decays
JA0369663
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13381