Characterization of the enzyme BtrD from Bacillus circulans and revision of its functional assignment in the biosynthesis of butirosin

Article abstract of DOI:10.1002/anie.200604194

(Chemical Equation Presented) A functional reassignment: BtrD, a protein encoded in the butirosin gene cluster, functions as a deacetylase rather than a nucleotidyltransferase as previously reported. BtrD was found to selectively catalyze the conversion o

Full text of DOI:10.1002/anie.200604194

Communications
DOI: 10.1002/anie.200604194
Biosynthetic Enzymes
Characterization of the Enzyme BtrD from Bacillus circulans and
Revision of Its Functional Assignment in the Biosynthesis of
Butirosin**
Andrew W. Truman, Fanglu Huang, Nicholas M. Llewellyn, and Jonathan B. Spencer*
The
2-deoxystreptamine(2-DOS)-
containing aminoglycosides are a clin-
ically important class of antibiotics,
whose members include neomycin,
gentamycin, and the semisynthetic
amikacin.A detailed understanding
of the biosynthetic enzymes that con-
struct these natural products is
required for the rational engineering
of biosynthetic pathways to produce
novel aminoglycosides.Butirosin ( 5)
is a 2-DOS aminoglycoside that con-
tains a novel (S)-4-amino-2-hydroxy-
butyrate side chain.In 2005, Kudo
et al.[1] proposed that BtrD, a protein
encoded by the butirosin gene clus-
ter,[2] functions as a nucleotidyltrans-
ferase that specifically synthesizes
thymidine 5’-diphosphoglucosamine
(TDP-GlcN) from thymidine 5’-tri-
phosphate (TTP) and glucosamine-1-
phosphate, thus generating the sugar
donor for a putative glycosyltransfer-
ase, BtrM (Scheme 1a).It was also
proposed that homologous enzymes
Scheme 1. Proposed biosynthesis of butirosin B (5).
in other gene clusters also function as hexose-1-phosphate
case, glucosamine originates from uridine diphospho-N-
acetyl-d-glucosamine (UDP-GlcNAc), an abundant primary
metabolite essential for cell wall biosynthesis.A UDP- N-
acetyl-glucosaminyl transferase glycosylates an aglycone,
then the aminosugar is subsequently deacetylated.A similar
pathway also exists in the biosynthesis of lipid A using the
phylogenetically unrelated deacetylase LpxC.[6] This biosyn-
thetic strategy ensures that UDP-GlcNAc is not unnecessarily
hydrolyzed (a transformation that could be detrimental to cell
wall biosynthesis), and the rate of UDP-GlcNAc sequestra-
tion is controlled by glycosyltransferase activity and the rate
of aglycone production.
nucleotidyltransferases.[1]
Here, we report that this characterization is incorrect and
that BtrD actually functions as a deacetylase (Scheme 1b),
which is consistent with its phylogenetic similarity to charac-
terized deacetylases PIG-L (glycosylphosphatidylinositol bio-
synthesis[3]), MshB (mycothiol biosynthesis[4]) and Orf2*
(teicoplanin biosynthesis[5]).These all have a conserved
domain called COG2120 and function in an analogous fashion
to incorporate glucosamine into natural products.In each
[*] A. W. Truman, Dr. F. Huang, N. M. Llewellyn, Dr. J. B. Spencer
University of Cambridge
The recent discovery[5] that Orf2*, a protein encoded in
the teicoplanin gene cluster, functions as a deacetylase
prompted a reinvestigation into the function of BtrD.The
gene btrD was cloned from cosmid DNA and BtrD was
expressed in E. coli BL21(DE3) as an N-terminally His6-
tagged protein.Unexpectedly, incubation of the enzyme with
TTP or uridine triphosphate with glucosamine-1-phosphate
or glucose-1-phosphate gave no reaction.[7] The utility of BtrD
as a deacetylase was therefore explored.
University Chemical Laboratory
Lensfield Road, Cambridge CB21EW (UK)
Fax: (+44)122-333-6362
E-mail: jbs20@cam.ac.uk
[**] This research was supported financially by the BBSRC (support for
J.B.S.), the BBSRC and St. John’s College, Cambridge (studentship
for A.W.T.), and the Marshall Scholarships and the National Science
Foundation (support for N.M.L.). Many thanks go to Professor
Anthony Clarke for the donation of AAC(2’)-Ia.
As has been previously observed,[1] BtrD does not
deacetylate UDP-GlcNAc, so a pathway of glucosamine
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
1462
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1462 –1464

Angewandte
Chemie
incorporation analogous to the biosyntheses of mycothiol,
glycosylphosphatidylinositol, and teicoplanin was investi-
gated.To accomplish this, the synthesis of acetylated biosyn-
thetic intermediates was required.N-acetylation is a common
resistance mechanism towards aminoglycosides,[8] and a
number of coenzyme A dependent aminoglycoside 2’-N-
acetyltransferases (AAC(2’)) have been reported.[9] The
well-characterized[10] AAC(2’)-Ia from Providencia stuartii
was expressed as an N-terminally His6-tagged protein in
E. coli and its activity exploited to selectively 2’-N-acetylate
paromamine (2), neamine (3), ribostamycin (4), and butirosin
(5).
from d = 1.92 ppm for 1 to d = 1.79 ppm for free acetate. Since
BtrD showed no activity with the other acetylated com-
pounds, it strongly suggests that the timing of deacetylation in
butirosin biosynthesis is directly after the formation of 2’-N-
acetylparomamine (1; Scheme 1b).This is consistent with the
observation that paromamine restored butirosin production
in a btrD gene disrupted mutant of Bacillus circulans.[1]
The reassignment of BtrD as a deacetylase raises the
possibility that all proteins containing the COG2120 domain
are deacetylases.This domain has a conserved sequence motif
near the N terminus ((A/P)HXDD) and another towards the
[4]
middle of the protein (HXDH).Crystallographic studies
These compounds represent all major intermediates
beyond 2-DOS in butirosin biosynthesis (Scheme 1), so
activity with at least one of these substrates would be
expected if BtrD functions as a deacetylase.Incubation of
BtrD with each of these substrates and analysis by LC/ESI-
MS demonstrated that BtrD is active only with 2’-N-acetyl-
paromamine (1); this was evident from the disappearance of
the signal at m/z 366.1 at t = 6.0 min and the appearance of the
signal at m/z 324.2 at t = 10.0 min attributed to paromamine
(2; Figure 1a).The transformation was also observed in situ
by 1H NMR spectroscopy (Figure 1b).This clearly shows the
movement of the singlet for the 2’-N-acetyl methyl group
imply that these conserved regions participate in metal
binding, substrate binding, and base catalysis for MshB.
Figure 2 clearly demonstrates the presence of these conserved
Figure 2. Partial amino acid sequence alignment of BtrD and amino-
glycoside gene cluster homologues with known deacetylases. Residues
proposed (from MshB crystallization studies) to participate in metal
~
*
binding ( ), substrate binding ( ) and base catalysis (N) are
indicated. Fully conserved residues have white text on a red back-
ground and similar (at least 70%) amino acids are framed in blue (the
similar residues are colored red).
regions in a wide array of homologues encoded in amino-
glycoside gene clusters.In addition to these highly conserved
regions, other more isolated residues proposed to bind to the
glucosaminyl substrate (Arg68 and Asp95 in MshB[4]) are also
well conserved throughout these proteins (see the Supporting
Information).
The homologue in the apramycin gene cluster, AprN,
suggests a common biosynthetic pathway to paromamine
prior to formation of apramycinꢀs novel fused ring system
(Scheme 2).It is worth noting that BtrD exhibits relatively
poor overall identity to the aminoglycoside gene cluster
homologues, probably because butirosinꢀs producer (Bacillus
circulans) is the only non-actinomycete aminoglycoside
producer.
A glucosaminyl group (sometimes heavily derivatized) is
present in almost every natural product whose gene cluster
encodes a protein with the COG2120 domain, thus suggesting
a common function.Notable exceptions are chloroeremomy-
cin[11] and balhimycin,[12] glycopeptide antibiotics whose gene
clusters contain putative proteins with high identity to Orf2*
and lincomycin.[13] Although all three of these molecules
contain aminosugars, it is not clear whether they are derived
from glucosamine.Further experimental work will be needed
Figure 1. Spectra demonstrating the conversion of 2’-N-acetylparom-
amine to paromamine. a) Chromatogram and mass spectra from LC/
ESI-MS scanning for m/z 324.2 and 366.2; top: 2’-N-acetyl-parom-
amine prior to BtrD addition; bottom: after incubation of 1 with BtrD.
b) Section of the 1H NMR spectrum showing the conversion of the 2’-
N-acetyl group to acetate.
Angew. Chem. Int. Ed. 2007, 46, 1462 –1464
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1463

Communications
Keywords: aminoglycosides ·
.
biosynthesis · butirosin ·
deacetylation
[1] F.Kudo, K.Kawabe, H.
Kuriki, T.Eguchi, K.Kaki-
numa, J. Am. Chem. Soc.
2005, 127, 1711 – 1718.
[2] Y.Ota, H.Tamegai, F.Kudo,
H.Kuriki, A.Koike-Take-
shita, T.Eguchi, K.Kaki-
numa, J. Antibiot. 2000, 53,
1158 – 1167.
[3] R.Watanabe, K.Ohishi, Y.
Maeda, N.Nakamura, T.
Kinoshita, Biochem. J. 1999,
339, 185 – 192.
[4] a) G.L.Newton, Y.Av-Gay,
R.C. Fahey, J. Bacteriol.
2000,
182,
6958 – 6963;
b) JT..Maynes, C.Garen,
M.M. Cherney, G. Newton,
D.Arad, Y.Av-Gay, RC. .
Fahey, M.N. James, J. Biol.
Chem. 2003, 278, 47166 –
47170; c) A.A. McCarthy,
N.A. Peterson, R. Knijff,
E.N. Baker, J. Mol. Biol.
2004, 335, 1131 – 1141.
[5] a) A.W. Truman, L. Robin-
son, J.B.Spencer, ChemBio-
Chem 2006, 7, 1670 – 1675;
b) T.-L. Li, F. Huang, S. F.
Haydock, T.Mironenko, P.F.
Leadlay, J.B.Spencer, Chem.
Biol. 2004, 11, 107 – 119.
[6] J.E. Jackman,
C.R.H.
Raetz, C.A. Fierke, Bio-
chemistry 1999, 38, 1902 –
1911.
Scheme 2. A proposed general biosynthetic pathway for glucosamine incorporation into aminoglycosides.
Blue rings indicate a second sugar that is likely to be incorporated in an analogous fashion. The BtrD
homologue in each aminoglycoside cluster is listed in parentheses.
[7] A mixture of nucleotidyl-
transferase activity similar to
that reported by Kudo et al.
and nucleotide 5’-triphosphate phosphatase activity was
observed in the cell-free supernatant (possibly a glucose-1-
phosphate thymidyltransferase) of both BL21 containing the
insert and those that did not.This activity was lost when BtrD
was purified by nickel-affinity chromatography.
to ascertain the true biosynthetic pathway of these com-
pounds and the roles of the BtrD homologues.
Another detail to be confirmed is the biosynthetic path-
way of aminoglycosides that incorporate more than one
glucosamine group, such as neomycin.The neomycin gene
cluster[14] contains only one BtrD homologue (Neo16),
suggesting it may act twice in the formation of neomycin
(Scheme 2).
In conclusion we have shown that BtrD functions as a
deacetylase in the formation of butirosin, therefore strongly
suggesting that UDP-GlcNAc is the true substrate for the
glycosyltransferase BtrM.Since there are many homologues
of BtrD encoded in the gene clusters of other natural
products, it implies a common strategy for incorporating a
glucosamine moiety into secondary metabolites.
[8] K.J.Shaw, P.N.Rather, R.S.Hare, G.Miller,
Microbiol. Rev.
1993, 57, 138 – 163.
[9] D.R.Macinga, P.N.Rather, Front. Biosci. 1999, 4, 132 – 140.
[10] a) M.Chevereau, PJL. .Daniels, J.Davies, F.LeGoffic,
Biochemistry 1974, 13, 598 – 603; b) K.Franklin, A.J.Clarke,
Antimicrob. Agents Chemother. 2001, 45, 2238 – 2244.
[11] A.M.A. van Wageningen, P.N. Kirkpatrick, D.H. Williams,
B.R. Harris, J.K. Kershaw, N.J. Lennard, M. Jones, S.J.M.
Jones, P.J.Solenberg, Chem. Biol. 1998, 5, 155 – 162.
[12] S.Pelzer, R.Süssmuth, D.Heckmann, J.Recktenwald, P.Huber,
G.Jung W.Wohlleben, Antimicrob. Agents Chemother. 1999, 43,
1565 – 1573.
[13] a) U.Peschke, H.Schmidt, H-.Z.Zhang, W.Piepersberg,
Mol.
Microbiol. 1995, 16, 1137 – 1156; b) N.M.Brahme, J.E.Gonza-
lez, S.Mizsak, J.R.Rolls, E.J.Hessler, L.H.Hurley,
Chem. Soc. 1984, 106, 7878 – 7883.
J. Am.
[14] F.Huang, S.F.Haydock, T.Mironenko, D.Spiteller, Y.Li, J.B.
Spencer, Org. Biomol. Chem. 2005, 3, 1410 – 1418.
Received: October 13, 2006
Published online: January 17, 2007
1464
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1462 –1464