Biosynthesis of riboflavin. The reaction catalyzed by 6,7-dimethyl-8-ribityllumazine synthase can proceed without enzymatic catalysis under physiological conditions

Article abstract of DOI:10.1021/jo001120s

6,7-Dimethyl-8-ribityllumazine is the biosynthetic precursor of the vitamin, riboflavin. The biosynthetic formation of the lumazine by condensation of 5-amino-6-ribitylamino-2,4(1H,3H)-pyrimidinedione and 3,4-dihydroxy-2-butanone 4-phosphate is catalyzed by the enzyme, lumazine synthase. We show that the condensation reaction can proceed without enzyme catalysis in dilute aqueous solution at room temperature and neutral pH. The reaction rate is proportional to e ^pH. The activation energy of the uncatalyzed reaction is E_a = 46.3 kJ mol^-1. The regioselectivity of the uncatalyzed reaction increases with pH and temperature (70% at 65 °C and pH 7.75). The data suggest partitioning of the uncatalyzed reaction via two different reaction pathways. The value of k_cat/k_uncat may be indicative for an entropy driven process for the enzyme-catalyzed reaction.

Full text of DOI:10.1021/jo001120s

J . Org. Chem. 2001, 66, 2555-2559
2555
Biosyn th esis of Ribofla vin . Th e Rea ction Ca ta lyzed by
6
,7-Dim eth yl-8-r ibityllu m a zin e Syn th a se Ca n P r oceed w ith ou t
†
En zym a tic Ca ta lysis u n d er P h ysiologica l Con d ition s
Klaus Kis, Karl Kugelbrey, and Adelbert Bacher*
Institut f u¨ r Organische Chemie und Biochemie, Technische Universit a¨ t M u¨ nchen,
Lichtenbergstr. 4, D-85747 Garching, Germany
adelbert.bacher@ch.tum.de
Received J uly 24, 2000
6
,7-Dimethyl-8-ribityllumazine is the biosynthetic precursor of the vitamin, riboflavin. The
biosynthetic formation of the lumazine by condensation of 5-amino-6-ribitylamino-2,4(1H,3H)-
pyrimidinedione and 3,4-dihydroxy-2-butanone 4-phosphate is catalyzed by the enzyme, lumazine
synthase. We show that the condensation reaction can proceed without enzyme catalysis in dilute
pH
aqueous solution at room temperature and neutral pH. The reaction rate is proportional to e
.
-
1
The activation energy of the uncatalyzed reaction is E ) 46.3 kJ mol . The regioselectivity of the
a
uncatalyzed reaction increases with pH and temperature (70% at 65 °C and pH 7.75). The data
suggest partitioning of the uncatalyzed reaction via two different reaction pathways. The value of
k
cat/ kuncat may be indicative for an entropy driven process for the enzyme-catalyzed reaction.
In tr od u ction
intermediate 5-amino-6-ribitylamino-2,4(1H,3H)-pyrim-
idinedione (1). The pyrimidine is converted into 6,7-
dimethyl-8-ribityllumazine (3) under the catalytic action
of lumazine synthase by condensation with 3,4-dihy-
droxy-2-butanone 4-phosphate (2) which is formed from
ribulose 5-phosphate. An unusual dismutation of the
lumazine 3 yields riboflavin. The second product of the
dismutation reaction, 5-amino-6-ribitylamino-2,4(1H,3H)-
pyrimidinedione, is returned into the biosynthetic cycle.
Paterson and Wood found that riboflavin can be
obtained by boiling a solution of its biosynthetic precur-
It is generally assumed that constituent molecules of
early organisms were generated by prebiotic chemical
reactions occurring in aqueous solution. More than a
century ago, L o¨ w and Emil Fischer showed that carbo-
1
,2
hydrates could be obtained from formaldehyde. These
reactions have been studied in more detail by Eschen-
3
moser and co-workers. Miller, Urey, and their co-workers
obtained evidence for the formation of amino acids by
electric discharge in a gas phase which was considered
to be similar to the early atmosphere, but more recent
geophysical evidence suggests that the primitive atmo-
sphere was different from that which was assumed by
these authors.4
sor, 6,7-dimethyl-8-ribityllumazine, in phosphate buffer,
pH 7.3, without a catalyst.1
3-15
The spontaneous and the
enzyme-catalyzed reaction were shown to proceed with
the same regiospecificity. Plaut and Beach subsequently
showed that the nonenzymatic formation of riboflavin
-6
More recently, W a¨ chtersh a¨ user proposed that the
formation of the first biomolecules could have occurred
in hot submarine environments providing a supply of
hydrogen sulfide, carbon monoxide, elemental hydrogen,
from 3 can proceed under neutral as well as acidic
16,17
conditions.
More recently, Eschenmoser and Strupp
obtained the lumazine 3 in 9.4% yield by boiling a neutral
7
and transition metal ions. Recently, Huber and W a¨ cht-
solution containing ribulose 1,5-bisphosphate and the
ersh a¨ user showed that amino acids and peptides can be
obtained under these reaction conditions in relatively
high yield.8
Flavins are a group of redox catalysts which are
indispensable in all cellular organisms. Their biosynthe-
sis has been studied in considerable detail (Figure 1; for
review, see refs 9-12). Briefly, the heterocyclic part of
the riboflavin molecule (4) is derived from GTP via the
3,18
pyrimidine 1.
We have found that the lumazine 3 can be formed from
the biosynthetic precursors 1 and 2 in dilute, neutral
aqueous solution at room temperature without enzyme
catalysis, albeit with a low regioselectivity. This low
regioselectivity may be a consequence of two reaction
(10) Plaut, G. W. E.; Smith, C. M.; Alworth, W. Annu. Rev. Biochem.
1
974, 43, 895-922.
*
Corresponding author. Tel.: +49-89-289-13360, Fax: +49-89-289-
(11) Bacher, A. In Chemistry and Biochemistry of Flavoproteins;
1
7
3363.
M u¨ ller, F., Ed.; CRC Press: Boca Raton, Florida, 1991; pp 215-259.
(12) Bacher, A.; Eberhardt, S.; Richter, G. In Escherichia coli and
Salmonella; Neidhardt, F. C., Ed.; American Society for Microbiology:
Washington, DC, 1996; pp 657-664.
(13) Paterson, T.; Wood, H. C. S. J . Chem. Soc., Chem. Commun.
1969, 290-291.
(14) Rowan, T.; Wood, H. C. S. Proc. Chem. Soc. 1963, 21-22.
(15) Rowan, T.; Wood, H. C. S. J . Chem. Soc. (C) 1968, 452-458.
(16) Beach, R. L.; Plaut, G. W. E. Biochemistry 1970, 9, 760-770.
(17) Plaut, G. W. E.; Beach, R. L. In Flavins and Flavoproteins;
Singer, T. P., Ed.; Elsevier Scientific Publishing Co.: Amsterdam, 1976;
pp 737-746.
†
Dedicated to Professor Albert Eschenmoser on the occasion of his
5th birthday.
(
(
(
(
(
(
(
(
(
1) L o¨ w, A. J . Prakt. Chem. 1886, 33, 321.
2) Fischer, E. Ber. Dtsch. Chem. Ges. 1890, 23, 370-394.
3) Eschenmooser, A.; Loewenthal, E. Chem. Soc. Rev. 1992, 1-16.
4) Miller, S. Science 1953, 117, 528.
5) Mattioli, G. S.; Wood, B. J . Nature 1986, 322, 626.
6) Kasting, J . F. Origins Life 1990, 20, 199.
7) Huber, C.; W a¨ chtersh a¨ user, G. Science 1997, 276, 245-247.
8) Huber, C.; W a¨ chtersh a¨ user, G. Science 1998, 281, 670-672.
9) Plaut, G. W. E.; Beach, R. L.; Aogaichi, T. Biochemistry 1970, 9,
7
71-785.
(18) Strupp, C. J . Thesis, ETH Z u¨ rich, 1992.
1
0.1021/jo001120s CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/08/2001

2
556 J . Org. Chem., Vol. 66, No. 8, 2001
Kis et al.
F igu r e 1. Biosynthesis of riboflavin. A, 3,4-dihydroxy-2-butanone 4-phosphate synthase; B, lumazine synthase; C, riboflavin
synthase.
pathways with different regioselectivities. These observa-
tions support the hypothesis that flavin type molecules
could have been formed naturally in prebiotic environ-
ments without enzyme catalysis.
The temperature dependence of the uncatalyzed reac-
tion in the range from 20 °C to 80 °C is shown in Figure
2. The Arrhenius-plot of log(v) versus 1/ T gave a straight
line indicating an overall activation energy of 46.3 kJ
-
1
-1
mol with a preexponential term P of 180 M min
.
pH
Using the overall rate law v ) k × [1] × [2] × e ) A ×
Resu lts
(
-Ea/ RT)
pH
e
× [1] × [2] × e
the frequency factor A can be
In the course of kinetic studies on lumazine synthase,
we observed that 6,7-dimethyl-8-ribityllumazine (3) was
formed from 5-amino-6-ribitylamino-2,4(1H,3H)-pyrim-
idinedione (1) and 3,4-dihydroxy-2-butanone 4-phosphate
calculated from the preexponential factor P by the
pH
equation P ) A × [1] × [2] × e . At pH 7, A had a
3
-1
-1
calculated value of 122 × 10 M min , and thus the
-
1
-1
value of k at pH 7 and 37 °C was 0.0020 M min
.
(
2) in blank samples devoid of lumazine synthase. This
The enzyme-catalyzed formation of the lumazine 3 has
observation prompted a more detailed study of the
uncatalyzed condensation reaction.
19,20
been shown to be regiospecific.
This suggested that
the initial step in the enzyme-catalyzed reaction is a
nucleophilic attack of the position 5 amino group of 1 on
the carbonyl group of 2 yielding the Schiff’s base 5
(Figure 3).
Kinetic analysis of the spontaneous condensation reac-
tion was performed in aqueous solution containing the
biosynthetic intermediates 1 and 2 in the mM concentra-
tion range. At pH 7.0 and 37 °C, the reaction was found
to be pseudo first order for the substrate 1. A logarithmic
plot of the reaction velocity versus pH gave a straight
To study the regiospecificity of the uncatalyzed reac-
13
1
tion, we prepared [1- C ]-2 by a modification of the
19
procedure reported earlier. Reaction with the pyrimi-
-
1
line indicating a pH dependence of v ) 0.0185 µM min
pH
pH
×
e
. The rate law is v ) k × [1] × [2] × e . At a
(
19) Kis, K.; Volk, R.; Bacher, A. Biochemistry 1995, 34, 2883-2892.
temperature of 37 °C, the rate constant k has a value of
(20) Nielsen, P.; Neuberger, G.; Fujii, I.; Bown, D. H.; Keller, P. J .;
-
1
-1
0
.0022 M min
.
Floss, H. G.; Bacher, A. J . Biol. Chem. 1986, 161, 3661-3669.

Biosynthesis of Riboflavin
J . Org. Chem., Vol. 66, No. 8, 2001 2557
which is similar to the pathway of the enzyme-catalyzed
reaction (pathway A, Figure 3). The alternative pathway
(B) could involve the formation of diacetyl from 2 by
â-elimination of phosphate. The subsequent condensation
of diacetyl with the pyrimidine 1 would be characterized
by the absence of regiospecificity due to its symmetric
structure.
The velocity of the uncatalyzed lumazine formation
increases with increasing pH. At high pH, the rate of the
â-elimination of 2 should be increased because the
removal of the acidic proton at C-3 of 2 is facilitated. On
the other hand, the 5-amino group of 1 obviously is
deprotonated in basic solution, and therefore the nucleo-
philic attack of this amino group to the carbonyl group
of 2 proceeds faster. The enhanced regioselectivity of the
lumazine formation shows that pathway A is favored at
higher pH and that the reactivity of the deprotonated
5
-amino group of 1 exceeds the â-elimination of 2. An
F igu r e 2. Arrhenius plot of uncatalyzed formation of 3 from
additional point of evidence may arise directly from
Figure 5. The point of inflection of the pH dependent
regioselectivity at pH 6.8 may reflect the pK of a certain
a
1
1
and 2. The fit of the data afforded the relationship ln(v) )
2
8.99 - 5567.3 × 1/ T with R ) 0.998 corresponding to an
-
1
activation energy of E ) 46.3 kJ mol ; 1.0 mM 1, 1.2 mM 2,
pH 7.0.
a
species in the proposed reaction sequence.
Studies on the velocity of the spontaneous lumazine
formation showed an exponential dependence on tem-
perature. The data gave rise to a linear Arrhenius plot,
dine 1 afforded 3 which was analyzed by NMR spectros-
copy.
-
1
indicating an activation energy of about 46 kJ mol for
the overall reaction. As shown recently, the abstraction
of the proton at C-3 of 2 irreversibly leads to the
elimination of the phosphate group under formation of
diacetyl.19 This reaction should proceed faster at elevated
temperature. However, the enhanced regioselectivity at
elevated temperature shows that pathway A is preferred.
This could imply that the elimination of phosphate is
facilitated by Schiff’s base formation which is supported
1
The H NMR signal of the position-6 methyl group of
1
3
3
1
obtained from [1- C ]-2 shows a central line for the
1
2
13
3 3
CH isotopomer and two satellite lines for the CH
isotopomer. The isotopomer composition can be obtained
by integration of these different signal components. It is
1
3
relevant for the calculation that the C enrichment of
1
3
[
1- C
formed with [1- C
synthesis or by enzymatic conversion from [2- C
1
]-2 was about 99%. Data from experiments per-
1
3
1
]-2 samples obtained by chemical
1
3
1
]-D-
2
1
by experiments of Bender and Williams and Hine et
glucose gave closely similar results. It follows that the
presence of small amounts of the 3,4-dihydroxy-2-bu-
tanone 3-phosphate (an impurity from the chemical
synthesis) did not influence the observed regioselectivity
of the condensation reaction.
2
2
al. who studied the deuterium exchange of the R-protons
of carbonyl compounds in the presence of primary amines.
They observed a dramatic increase in the exchange
velocity and explained their observation by the formation
of a Schiff’s base.
The regioselectivity of lumazine formation as a function
of temperature and pH is shown in Figures 4 and 5. The
regioselectivity of the reaction was found to increase with
temperature in the range of 27 °C to 65 °C. It was also
found to increase with pH in the range of pH 5.25 to 8.25.
The highest value of regioselectivity observed in this
sequence of experiments was 70% at 65 °C and pH 7.75.
The biomimetic formation of lumazine seems to be
attributed to a combination of the two pathways A and
B. In basic solution and at elevated temperature the
contribution of pathway A increases due to an enhanced
reactivity of 1 (at high pH) and a faster Schiff’s base
formation. The Schiff’s base may be easily protonated,
and the enhancement of the velocity is a consequence of
the basicity of the Schiff’s base. A high pH is necessary
for the formation of a Schiff’s base from the free amino
group of 1. However, if the pH was too high, a protonation
of the Schiff’s base would be no longer possible. Therefore,
the inflection point of the regioselectivity at pH 6.8 may
Discu ssion
The formation of 6,7-dimethyl-8-ribityllumazine (3)
from 1 and 2 catalyzed by lumazine synthase is strictly
1
9,20
regiospecific (at least 98%).
The mechanism of the
a a
reflect the pK of the Schiff’s base 5. pK Values for
enzyme-catalyzed reaction has been studied in some
detail, and it has been proposed that the initial step of
23,24
Schiff’s bases typically are in the range of 5.8 to 7.5.
The rate constant kuncat of the spontaneous lumazine
formation was calculated from the pH-dependence and
temperature-dependence experiments and had a value
this reaction is the nucleophilic attack of the carbonyl
group of 2 by the 5-amino group of 1 (Figure 3).1
9,20
It
was further proposed that the Schiff’s base intermediate
could eliminate phosphate with subsequent formation
-
1
-1
of about 0.002 M min . The corresponding rate con-
5
-
1 19
stant kcat of the enzyme-catalyzed reaction is 0.074 s
.
of the pyrazine ring of the lumazine chromophore.
The uncatalyzed reaction has only a relatively low
regioselectivity (at pH 7.5, 37 °C, 60%). However, the
velocity of the uncatalyzed reaction and its degree of
regioselectivity increase with pH and with temperature.
Our regioselectivity data suggest that the uncatalyzed
reaction can proceed via two different pathways, one of
(
21) Bender, H. C.; Williams, A. J . Am. Chem. Soc. 1966, 88, 2502-
508.
(22) Hine, J .; Menon, B. C.; J ensen, J . H.; Mulders, J . J . Am. Chem.
2
Soc. 1966, 88, 3367-3373.
(
(
23) Culbertson, J . B. J . Am. Chem. Soc. 1951, 73, 4818-4823.
24) Cordes, E. H.; J encks, W. P. J . Am. Chem. Soc. 1963, 85, 2843-
2848.

2
558 J . Org. Chem., Vol. 66, No. 8, 2001
Kis et al.
F igu r e 3. Proposed mechanisms for the uncatalyzed formation of 3. 13C labels are denoted by asterisks.
F igu r e 4. Regioselectivity of the biomimetic lumazine forma-
F igu r e 5. Regioselectivity of the biomimetic lumazine forma-
1
3
13
tion at different temperatures. 40 mM 1, 1.8 mM [1- C
1
]-2;
tion at different pH. 40 mM 1, 1.8 mM [1- C
were fitted to the function r ) (low + high × 10 )/(10
+ 1) according to Leatherbarrow; values for low, high,
1
]-2, 37 °C. Data
(
pH - w)
(pH
O, pH 6.70; b, pH 7.75 (pH was not corrected for tempera-
tures).
-
w)
31
and w were solved by the Newton method implemented in
Microsoft Excel 8.0 Solver; low ) 2.2, high ) 70.1, w ) 6.8,
sum of square errors ) 2.14; the point of inflection at pH 6.8
corresponds to a regioselectivity of 39.5%
The value of kcat/ kuncat is therefore 2220 M. It is difficult
to interpret this value. However, a possible explanation
may be that the nonenzymatic reaction could be as fast
as the enzymatic reaction when the substrate concentra-
3
from ribulose 1,5-bisphosphate require a temperature
3
tions are in the order of 10 M. In light of the concentra-
around the boiling point of water. The uncatalyzed
reaction discussed in this paper proceeds under even
milder conditions and has an appreciable velocity at room
temperature. Although biomimetic reaction conditions for
the formation of the riboflavin precursor 1 have not been
found so far, these observations address the question
whether riboflavin or a related isoalloxazine could have
been formed prior to the evolution of enzyme catalysis.
3
tion of pure water (56 M) a concentration of 10 M is not
possible. The large value of kcat/ kuncat may show that the
enzymatic reaction is an entropy driven process.
The terminal two steps in the pathway of riboflavin
biosynthesis can proceed under relatively mild conditions
in the absence of catalyst. The spontaneous dismutation
of the lumazine1
3,15,16
and the formation of the lumazine

Biosynthesis of Riboflavin
Exp er im en ta l Section
J . Org. Chem., Vol. 66, No. 8, 2001 2559
monitored fluorimetrically (excitation, 408 nm; emission, 487
nm). An eluent containing 40% methanol and 0.1 M am-
monium formate was used for the determination of riboflavin
1
3
]Methyl iodide (99% 13C) was purchased
1
Ma ter ia ls. [1- C
from Campro, Emmerich, Germany. [2- C
1
3
1
]-D-glucose (99%
C) was purchased from Omicron Biochemicals Inc., South
Bend, IN. 5-Nitro-6-ribitylamino-2,4(1H,3H)-pyrimidinedi-
one, 6,7-dimethyl-8-ribityllumazine, and (2S)-2,3-O-isopro-
pylideneglyceraldehyde were prepared by published proce-
dures. Solvents were distilled before use.
(
4); the effluent was monitored fluorimetrically (excitation, 445
1
3
nm; emission, 516 nm). The molar amount of 2 consumed by
lumazine synthase corresponds to the amount of lumazine 3
plus two times the amount of riboflavin (4).
2
5
10
1
9
In flu en ce of th e Con cen tr a tion of 5-Am in o-6-r ibity-
la m in o-2,4(1H,3H)-p yr im id in ed ion e on th e Non en zy-
m a tic F or m a tion of Lu m a zin e. Assay mixtures (100 µL)
contained 22.5 mM potassium phosphate (pH 7.0), 5 mM 2,
and 1 (from 11 to 75 mM) at 37 °C. Aliquots (15 µL) were
retrieved at intervals, quenched with 50 µL of TCA (15%), and
analyzed by HPLC as described above.
P r ep a r a tion of 5-Am in o-6-r ibityla m in o-2,4(1H,3H)-p y-
r im id in ed ion e (1). 5-Nitro-6-ribitylamino-2,4(1H,3H)-pyri-
midinedione (36.7 mg, 0.1 mmol) was suspended in 10 mL of
water. The suspension was hydrogenated over Pd/charcoal at
room temperature and atmospheric pressure for 2 days. The
solution was then quickly passed through a 0.2 µm membrane
filter under an atmosphere of inert gas. A 10 µL aliquot was
removed and diluted with 1 mL of hydrochloric acid (0.1 M)
p H Dep en d en ce of Non en zym a tic F or m a tion of 6,7-
Dim eth yl-8-r ibityllu m a zin e. Assay mixtures containing 2
mM EDTA, 22.5 mM 1, and 90 mM potassium phosphate
buffer at pH values ranging from 5.8 to 7.3 were preincubated
for 3 min at 37 °C. The reaction was started by the addition
of 2 to a final concentration of 3.3 mM in a total assay volume
of 75 µL. Aliquots (10 µL) were retrieved and acidified with
50 µL of TCA (15%). 6,7-Dimethyl-8-ribityllumazine was
analyzed by HPLC as described above.
for the determination of the product concentration (ꢀ268
)
-
1
-1 25
2
4500 M cm ). The main solution was stabilized by the
addition of dithiothreitol (123 mg, 0.8 mmol) and stored at -70
°
C.
P r ep a r a tion of [4- C
1
3
1
]-(2S,3RS)-1,2-O-Isop r op ylid en e-
1
,2,3-bu ta n etr iol. [1- C ]Methylmagnesium iodide was pre-
1
3
1
1
3
pared from magnesium (35 mmol, 0.85 g) and [1- C
iodide (35 mmol, 5.0 g) in absolute ether. A solution of (2S)-
,3-O-isopropylideneglyceraldehyde in absolute ether (pre-
pared from 17.5 mmol (5S)-5,6-O-isopropylidene-L-gulono-1,4-
1
]methyl
Tem p er a tu r e Dep en d en ce of Non en zym a tic F or m a -
tion of 6,7-Dim eth yl-8-r ibityllu m a zin e. A solution contain-
ing 112 mM potassium phosphate, pH 7.0, 2.2 mM EDTA, and
1.1 mM 1 in a total volume of 420 µL was preincubated at the
indicated temperatures for 3 min. The reaction was started
by the addition of 5 µL of 12.1 mM 2. Aliquots (10 µL) were
retrieved and acidified with 50 µL of TCA (15%). 6,7-Dimethyl-
2
1
9
lactone ) was added. After 17 h, the reaction mixture was
poured on ice and treated with a saturated solution of
ammonium chloride to solubilize the precipitate. The phases
were separated, and the water phase was extracted with ether.
The combined organic layers were washed with 10 mL of water
8
-ribityllumazine was analyzed by HPLC as described above.
and dried over MgSO
reduced pressure.
4
. The solvent was removed under
R egioch em ist r y of t h e Non en zym a t ic F or m a t ion of
,7-Dim eth yl-8-r ibityllu m a zin e. To avoid oxidative decom-
6
1
3
P r ep a r a tion of [1- C
1
]-(3S)-3,4-Dih yd r oxy-2-bu ta n on e
]-2). The reaction steps for the prepara-
1
]-(3S)-3,4-dihydroxy-2-butanone 4-phosphate
position of 5-amino-6-ribitylamino-2,4(1H,3H)-pyrimidinedione
1), these experiments were carried out under an atmosphere
of N /H (9:1).
Solutions containing 6.2 mM [1- C
1
3
4
-P h osp h a te ([1- C
1
(
1
3
tion of [1- C
(
with [4- C
2
2
1
3
19
1
[1- C ]-2) were carried out according to Kis et al. starting
1
3
1
3
1
]-2 or 85.3 mM 1 or 2
1
]-(2S,3RS)-1,2-O-isopropylidene-1,2,3-butanetriol.
1
3
M potassium phosphate were preincubated at the desired
temperature. Aliquots of these solutions (buffer, 120 µL; 2, 220
µL; and 1, 300 µL) were combined and incubated at the
respective temperature. pH was controlled at the end of the
En zym a tic Syn th esis of [1- C
bu tan on e 4-P h osph ate ([1- C
1
]-(3S)-3,4-Dih yd r oxy-2-
1
3
13
1
]-2). Small amounts of [1- C
1
]-
1
3
2
were prepared enzymatically from [2- C ]-D-glucose as
1
2
6
described elsewhere.
1
3
1
incubation period, when all of [1- C ]-2 was consumed.
P r otein s. The lumazine synthase/riboflavin synthase com-
plex (formerly designated heavy riboflavin synthase) was
purified from cell extracts of the derepressed mutant H94 of
Hydrochloric acid (1 M, 400 µL) was added, and precipitate
was removed by centrifugation. 6,7-Dimethyl-8-ribityllumazine
(3) was purified by HPLC using a column of Nucleosil 10C18
(20 × 250 mm). The eluent contained 10% methanol and 50
mM formic acid. The flow rate was 40 mL/min. The effluent
was monitored fluorimetrically (excitation, 408 nm; emission,
2
7,28
Bacillus subtilis by published procedures.
NMR Sp ectr oscop y. H and 13C NMR spectra were re-
1
corded at room temperature with AM 360, AC 250, and AC
2
00 spectrometers from Bruker Instruments, Karlsruhe, Ger-
many.
Assa y of (3S)-3,4-Dih yd r oxy-2-bu ta n on e 4-P h osp h a te.
4
87 nm). Fractions were pooled and freeze-dried. The residue
was dissolved in 540 µL of 50 mM formic acid and 60 µL of
O were added. 13C abundance at the methyl groups of the
D
2
The concentration of (3S)-3,4-dihydroxy-2-butanone 4-phos-
phate (2) was determined enzymatically using the lumazine/
riboflavin synthase complex (heavy riboflavin synthase).29
Assay mixtures contained 100 mM potassium phosphate, pH
3
0
lumazine 3 was analyzed by NMR spectroscopy.
Defin ition of Regioselectivity. The regioselectivity of the
lumazine formation is defined by ((L6 - L6R)(a + i ))/(L/
/
+
R
6R
/
7
.0, 2 mM EDTA, 0.7 mM 1, 2 µg of heavy riboflavin synthase,
L
6R)(a - i )) where L6Rand L6R denote the amount of lumazine
and 2 in a total volume of 100 µL. The concentration of 2
should be in the range of 20-50 µM for optimum accuracy.
Assay mixtures were incubated at 37 °C for 1 h. Protein was
precipitated by the addition of 15% trichloroacetic acid (TCA,
13
with or without the C label in the 6R-methyl position,
13
respectively; a is the C enrichment of the starting material
99%), and i is the natural abundance of 13C (1.1%).
(
1
00 µL). The precipitate was removed by centrifugation (14000
Ack n ow led gm en t. This work was supported by
rpm, 5 min). The samples were analyzed by reversed phase
HPLC using a column of Nucleosil 10C18 (4 × 250 mm)
grants from the Deutsche Forschungsgemeinschaft and
the Fonds der Chemischen Industrie. We thank Prof.
Eschenmoser, Prof. Arigoni, Prof. Cushman, and Prof.
W a¨ chtersh a¨ user for helpful discussions. We also thank
Dr. Klaus Kis for creation of the cover art.
(
Macherey & Nagel, D u¨ ren, Germany). An eluent containing
1
0% methanol and 30 mM formic acid was used for determi-
nation of 6,7-dimethyl-8-ribityllumazine (3); the effluent was
(
25) Cresswell, R. M.; Wood, H. C. S. J . Chem. Soc. 1960, 4768-
775.
26) Richter, G.; Krieger, C.; Volk, R.; Kis, K.; Ritz, H.; G o¨ tze, E.;
4
J O001120S
(
Bacher, A. Methods Enzymol. 1997, 280, 374-382.
(
27) Bacher, A. Methods Enzymol. 1986, 122, 192-199.
(30) Bown, D. H.; Keller, P. J .; Floss, H. G.; Sedlmaier, H.; Bacher,
A. J . Org. Chem. 1986, 51, 2461-2467.
(31) Leatherbarrow, R. J . Trends Pharmacol. Sci. 1990, 15, 455-
458.
(
28) Schott, K.; Ladenstein, R.; K o¨ nig, A.; Bacher, A. J . Biol. Chem.
1
990, 265, 12686-12689.
(29) Volk, R.; Bacher, A. J . Biol. Chem. 1990, 265, 19479-19485.