Communications to the Editor
J. Am. Chem. Soc., Vol. 121, No. 35, 1999 8125
Scheme 2
Upon reduction of 4, whose instability is highlighted by its
ready loss of TDP, the more stable product 5 is formed, which is
sufficiently stable to enable purification and characterization. It
is tempting to speculate whether some form of protein-protein
interaction leading to substrate channeling occurs in vivo to
maximize the utilization of the unstable intermediate 4 by TylC1.
However, attempts to detect interactions between TylX3 and
2
0
TylC1 both in vivo using the yeast two-hybrid system and in
vitro using the gel filtration chromatography were not fruitful.
Not only does TylX3 have interesting catalytic properties, but
of further significance is that its deduced sequence shows a high
degree of homology to those of eryBVI in the erythromycin path-
weight of 36 kDa.13 The UV-visible spectra of both proteins
were transparent above 300 nm. However, inductively coupled
plasma (ICP) analysis showed the presence of 0.8 equiv of Zn2
bound per TylX3 monomer. Prolonged dialysis of TylX3 against
9
21
way, lanS from the landomycin pathway, orf23 in the vanco-
+
22
23
mycin pathway, snoH in the nogalamycin pathway, dnmT from
the doxorubicin pathway,24 and orf27 in the granaticin pathway.
Thus, it is plausible that these enzymes constitute a small family
25
5
mM 1,10-phenanthroline led to a total loss of activity, further
2
+
14
2+
implicating a role for Zn in catalysis. ICP analysis of the 1,-
of Zn -dependent dehydratases involved in C-2 deoxygenation
2
+
10-phenanthroline-treated enzyme showed the absence of Zn ,
for the generation of the 2,6-dideoxyhexose moieties in their
respective systems. Although TylC1 and Gra Orf26 both are
typical NADPH-dependent keto-reductases, the lack of any
significant sequence similarity between the corresponding genes,
and the fact that the product of TylC1 catalysis is stereochemically
which correlates well with the loss of enzyme activity. The
specific activity of the native enzyme was also found to increase
about 8-fold when it was dialyzed against 5 mM ZnSO .
4
When TylX3 was incubated with the substrate 3, which was
prepared from TDP-D-glucose using the 4,6-dehydratase RfbB15
from the rhamnose pathway, consumption of substrate and the
concomitant formation of TDP and a new product were detected
by HPLC (see Scheme 2).16 This new compound was isolated
from an ethyl acetate extract and was identified as maltol (6) by
11
distinct from that of Gra Orf 26, seem to indicate that these
two enzymes have different evolutionary backgrounds.
In conclusion, these results confirm the mode of C-2 deoxy-
genation in the biosynthesis of 2,6-dideoxyhexoses. Our bio-
chemical characterization of TylX3 and TylC1 clearly revealed
that nature has opted for a metal-assisted dehydration/reduction
sequence to remove a â-OH (at C-2) from a ketohexose (4-keto)
precursor. A similar â-elimination mechanism has also been found
for C-6 deoxygenation in the biosynthesis of 6-deoxyhexoses, in
which the expulsion of the 6-OH (â-OH) from a 4-ketohexose
1
comparing its H NMR spectrum with that of an authentic sample
(from Aldrich). A similar observation was also noted by Draeger
et al. in which formation of maltol by Gra Orf27 had been
suggested as a result of degradation of the expected dehydration
product 4.11 Interestingly, when 3 was incubated with TylX3
together with TylC1 and NADPH, a new product was detected
intermediate generates a 4-keto-6-deoxyhexose product (such as
by HPLC.1
7,18
This product was purified by gel filtration chro-
1,2,26
3).
Clearly, a more elaborate mechanism has been evolved
matography (Bio-Rad P2, extra fine, 1 × 150 cm, elution with
to remove an R-OH from a ketohexose precursor, as in the case
of C-3 deoxygenation. Taken together, these differences are a
testament to the evolutionary diversity of biological C-O bond
cleavage events, and as more sugar biosynthetic systems become
characterized, it will be interesting to see what other novel
deoxygenations may be discovered.
10 mM KCl at 8 mL/h) and identified as TDP-2,6-dideoxy-D-
1
13
19
glycero-D-glycero-4-hexulose (5) by H and C NMR. Again,
a reductase (Gra Orf26) was also required to afford a stable sugar
product, as was observed by Draeger et al.11
When the incubation was carried out in buffer prepared with
2
D O, deuterium incorporation was found at the C-2 equatorial
position in 5. These results, while clearly indicating that
displacement of the 2-OH with a solvent hydrogen proceeds with
retention of configuration, also imply that the hydride is trans-
ferred from NADPH to the C-3 position of 4. It should be noted
that the C-3 hydroxyl group in 5 is axial, as revealed by the
coupling constants for H-3 (JH2a-H3 ) 3.5, JH2e-H3 ) 3.0 Hz),
which is also consistent with the lack of nuclear Overhauser effect
1
9
Acknowledgment. We express our great appreciation to Dr. Eugene
Seno and the Lilly Research Laboratories for their generous gifts of the
plasmids pSET552, pHJL309, pHJL311, and pOJ190. This work is
supported by the National Institutes of Health Grants GM35906 and
GM54346. H.-w.L. also thanks the National Institute of General Medical
Sciences for a MERIT Award.
JA991713O
(
NOE) between H-3 and H-5. This assignment is in contrast to
1
(
19) H NMR (D
2
O, hydrated form): δ 1.05 (3H, d, J ) 6.5 Hz, 5-Me),
11
the Gra Orf26 product isolated by Draeger et al.
1.80 (3H, s, 5′′-Me), 1.97 (1H, ddd, J ) 15.0, 3.0, 1.5, 2-Heq), 2.09 (1H, dq,
J ) 15.0, 3.5, 2-Hax), 2.18-2.27 (2H, m, 2′-H), 3.69 (1H, dd, J ) 3.0, 3.5,
3-H), 3.96-4.04 (3H, m, 4′-H, 5′-H), 4.18 (1H, q, J ) 6.6, 5-H), 4.47-4.48
(1H, m, 3′-H), 5.43 (1H, m, 1-H), 6.16 (1H, t, J ) 6.6, 1′-H), 7.58 (1H, s,
6
(12) After thrombin digestion to remove the His tag, the N-terminal amino
acid sequencing confirmed that the first 10 residues of TylX3 (AHSSAT-
AGPQ) and TylC1 (SGMYVQLGR) are identical to the respective translated
tylX3 and tylC1 sequences, except for the deletion of the first methionine
residue in both cases.
1
3
6′′-H). C NMR (D O, hydrated form): δ 11.5, 11.6, 33.7 (d, J ) 8.4 Hz),
2
38.0, 64.7 (d, J ) 5.2 Hz), 65.6, 68.9, 70.2, 84.6, 85.0 (d, J ) 8.6 Hz), 91.8,
3
1
93.5 (d, J ) 6.0 Hz), 111.6 (d, J ) 2.9 Hz), 137.3, 151.6, 166.5. P NMR
(D O): δ -10.7 (d, J ) 21.9 Hz), -12.9 (d, J ) 21.9 Hz).
(
13) The calculated molecular mass for TylX3 is 55 795 Da, and that for
TylC1 is 36 920 Da.
14) The actual role of Zn in TylX3 catalysis must await further
investigation.
15) Romana, L. K.; Santiago, F. S.; Reeves, P. R. Biochem. Biophys. Res.
Commun. 1991, 174, 846-852.
16) The incubation mixture was loaded on an Adsorbosphere SAX column
5 µm, 4.6 × 250 mm), and a linear gradient from 140 to 320 mM potassium
phosphate buffer (pH 3.6) over 20 min was used to elute the reaction products
monitored at 278 nm). Under these conditions, the retention times were 4.8
min for maltol (6), 13.0 min for substrate 3, and 17.6 min for TDP.
17) A preparative incubation contained the RfbB product (3, 14.0 mg,
3.7 µmol) and NADPH (21.7 mg, 26.1 µmol) in 1.5 mL of 100 mM potassium
phosphate buffer (pH 7.5). The reaction was initiated by the addition of TylX3
6.5 nmol) and TylC1 (10.5 nmol). Due to the instability of the TylX3 product,
2
(20) (a) Fields, S.; Song, O.-k. Nature 1989, 340, 245-246. (b) Vojtek,
2
+
(
A. B.; Hollenberg, S. M.; Cooper, J. A. Cell 1993, 74, 205-214.
(21) Westrich, L.; Domann, S.; Faust, B.; Bedford, D.; Hopwood, D. A.;
Bechthold, A. FEMS Microbiol. Lett. 1999, 170, 381-387.
(
(22) van Wageningen, A.; Kirkpatrick, P.; Williams, D.; Harris, B.;
Kershaw, J.; Lennard, N.; Jones, M.; Jones, S.; Solenberg, P. Chem. Biol.
1998, 3, 155-162.
(
(
(23) Torkkell, S.; Ylihonko, K.; Hakala, J.; Skurnik, M.; Mantsala, P. Mol.
Gen. Genet. 1997, 256, 203-209.
(
(24) Scotti, C.; Hutchinson, C. R. J. Bacteriol. 1996, 178, 7316-7321.
(25) Ichinose, K.; Bedford, D. J.; Tornus, D.; Bechthold, A.; Bibb, M. J.;
Revill, W. P.; Floss, H. G.; Hopwood, D. A. Chem. Biol. 1998, 5, 647-659.
(26) Glaser, L.; Zarkowsky, R. In The Enzymes; Boyer, P., Ed.; Academic
Press: New York, 1971; Vol. 5, pp 465-480. (b) Gabriel, O. In Carbohydrates
in Solution; Gould, R., Ed.; Advances in Chemistry Series 117, American
Chemical Society: Washington, DC, 1973; pp 387-410. (c) Snipes, C. E.;
Brillinger, G.-U.; Sellers, L.; Mascaro, L.; Floss, H. G. J. Biol. Chem. 1977,
252, 8113-8117. (d) Yu, Y.; Russell, R. N.; Thorson, J. S.; Liu, L.-d.; Liu,
H.-w. J. Biol. Chem. 1992, 267, 5868-5875.
(
2
(
an excess of TylC1 was used. The reaction was incubated at room temperature
and was followed by monitoring the consumption of NADPH at 340 nm. The
reaction was usually complete within 2 h.
(18) The same HPLC conditions as described in ref 16 were used. The
retention time for product 5 was 13.8 min.