Investigating the initial steps in the biosynthesis of cyanobacterial sunscreen scytonemin

Article abstract of DOI:10.1021/ja807192u

The cyanobacterial natural product scytonemin (1) functions as a sunscreen, absorbing harmful UV-A radiation. Using information from a recently identified gene cluster, we propose a biosynthetic route to this pigment. We also report the characterization of two enzymes, NpR1275 and NpR1276, which are involved in the initial stages of this pathway. A regioselective acyloin reaction between indole-3-pyruvic acid (4) and p-hydroxyphenylpyruvic acid (5) is a key step in assembling the carbon framework of a proposed monomeric scytonemin precursor (2). Copyright

Full text of DOI:10.1021/ja807192u

Published on Web 10/28/2008
Investigating the Initial Steps in the Biosynthesis of Cyanobacterial
Sunscreen Scytonemin
Emily P. Balskus and Christopher T. Walsh*
Department of Biological Chemistry and Molecular Pharmacology,
HarVard Medical School, Boston, Massachusetts 02115
Received September 10, 2008; E-mail: christopher_walsh@hms.harvard.edu
Scheme 1
Photosynthetic cyanobacteria have evolved a variety of strategies
for coping with exposure to damaging solar UV radiation,1 including
DNA repair processes,2 UV avoidance behavior,3 and the synthesis
of radiation-absorbing pigments.4 Scytonemin (1) is the most
widespread and extensively characterized cyanobacterial sunscreen.5
This lipid soluble alkaloid accumulates in the extracellular sheaths
of cyanobacteria upon exposure to UV-A light, where it absorbs
further incident radiation (λmax ) 384 nm).6 The natural product
also possesses interesting biological activity; as a micromolar
inhibitor of polo-like kinase 1, it exhibits antiproliferative and anti-
inflammatory properties.7
A gene cluster responsible for the production of scytonemin was
recently identified in Nostoc punctiforme ATCC 29133; however,
no functions were assigned to specific gene products.8 In this
communication, we outline a possible biosynthetic route to the
natural product (Scheme 1a). We also report the characterization
of two enzymes involved in the initial steps of this pathway and
identify a remarkably selective acyloin reaction as a key step in
constructing the carbon framework of the pigment.
The scytonemin gene cluster consists of 18 open reading frames
(ORFs), including a total of eight genes involved in the biosynthesis
of probable precursors tryptophan and tyrosine and a number of
ORFs bearing no significant homology to characterized proteins
(for the complete annotation, see Supporting Information). Among
the remaining genes, we identified two potential candidates for
enzymes utilized in the early stages of scytonemin assembly:
NpR1275, which resembles leucine dehydrogenase,9 and NpR1276,
homologous to the thiamin diphosphate (ThDP)-dependent enzyme
acetolactate synthase.10 We also noted the possible involvement
of a putative tyrosinase, NpR1263, in later oxidation steps, pot-
entially paralleling the well-established role of these enzymes in
melanin biosynthesis.11
We hypothesized that NpR1275 could oxidize tryptophan and/
or tyrosine to the corresponding pyruvic acid derivative (4 and 5).
A subsequent acyloin reaction involving these two substrates,
catalyzed by NpR1276, would assemble the precursor (3) to one-
half of the scytonemin skeleton. Decarboxylation, cyclization, and
oxidation steps, either enzyme-catalyzed or spontaneous, could
provide monomer 2. Notably, the structure of this proposed in-
termediate is closely related to a known cyanobacterial metabolite,
nostodione A.12 A final oxidative dimerization, either spontaneous
or promoted by tyrosinase NpR1263, would afford scytonemin.
To explore the validity of this hypothesis, NpR1275 and
NpR1276 were amplified from N. punctiforme ATCC 29133
genomic DNA and their products overexpressed in E. coli as C-His-
tagged fusions. The tryptophan dehydrogenase activity of NpR1275
was confirmed using a spectrophotometic assay measuring the
formation of NADH.13 Although this enzyme readily catalyzed the
oxidative deamination of tryptophan, the analogous reaction with
tyrosine was extremely slow. It is possible that p-hydroxyphe-
nylpyruvic acid (5) is generated directly by putative prephenate
dehydrogenase NpR1269 and consumed without prior conversion
to the amino acid, as the scytonemin cluster lacks the transaminase
used to convert 5 to tyrosine in the final step of the tyrosine bio-
synthetic pathway.
With the role of NpR1275 in generating putative substrates for
NpR1276 established, we explored the reactivity of this ThDP-
dependent enzyme. Treating a mixture of indole-3-pyruvic (4) and
p-hydroxyphenylpyruvic acids (5) (500 µM each) with NpR1276
(400 nM) in the presence of MgCl2 (5 mM) and ThDP (1 mM) in
pH 7.5 Tris-HCl buffer resulted in the formation of a single new
peak by HPLC analysis (Figure 1a). This reactivity was dependent
on the presence of NpR1276, as well as both Mg2+ and ThDP
cofactors.
Isolation and characterization of the new product peak from a
large-scale enzymatic preparation afforded isomeric acyloins 6 and
7 in a 1.4:1 ratio (Scheme 1b). However, the HPLC retention time
of this material did not match that of the original product as
confirmed by coinjection of a synthetic standard of 7 with the
NpR1276 reaction mixture. Further experiments demonstrated that
the acyloin products 6 and 7 formed during incubation of quenched
reaction mixtures at room temperature (Figure 1b).
Suspecting that 6 and 7 arose from an initial ꢀ-ketoacid adduct
(3), a reductive quench was attempted in hopes of preventing the
presumed decarboxylation event. Treatment of the NpR1276
reaction mixture with sodium borohydride (100 mM) afforded two
9
15260 J. AM. CHEM. SOC. 2008, 130, 15260–15261
10.1021/ja807192u CCC: $40.75
2008 American Chemical Society

C O M M U N I C A T I O N S
Figure 1. (a) HPLC time course for NpR1276-catalyzed formation of ꢀ-ketoacid 3 (280 nm, MeOH quench). (b) HPLC time course of ꢀ-ketoacid
decarboxylation during rt incubation of acidified reaction mixtures (220 nm). (c) HPLC traces of NaBH4 reduction products (280 nm).
major products (8 and 9) that were not cofactor-derived (Figure
1c). NMR characterization (1H, 13C, HMBC) revealed these
compounds to be diastereomeric diols generated from the reduction
of a single ꢀ-ketoacid regioisomer (3) (Scheme 1b).14 A mixture
of 8 and 9 was also obtained by treating 3 with sodium borohydride
immediately upon its collection from an analytical HPLC run.
Approximately 10% of the material underwent decarboxylation, as
indicated by the presence of additional diol products.
Acknowledgment. This work is supported by the NIH (GM-
20011). E.P.B. is the recipient of an NIH postdoctoral fellowship.
Dr. Elizabeth Nolan is acknowledged for helpful discussions.
Supporting Information Available: Experimental details and
characterization data for new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
References
The structure of ꢀ-ketoacid 3 provides important information
about the timing of key bond-forming events in the NpR1276-
catalyzed transformation. Isolation of a single regioisomer is
indicative of a highly selective reaction of the ThDP cofactor with
p-hydroxyphenylpyruvic acid (5), followed by nucleophilic attack
of cofactor-bound 5 onto indole-3-pyruvic acid (4) (for mechanism,
see Scheme S4, Supporting Information). The lack of products
arising from coupling between two identical pyruvic acid derivatives
(Figure 1b) indicates an exquisite level of enzymatic control over
the binding and activation of both substrates.15 Although ThDP-
dependent enzymes have been engineered to perform selective CsC
bond-forming reactions,16 this degree of specificity is largely
unprecedented in natural systems.17,18 Elucidation of the factors
responsible for this selectivity may provide valuable information
for future engineering efforts.
(1) For a review, see: Castenholz, R. W.; Garcia-Pichel, F. In The Ecology of
Cyanobacteria; Whitton, B. A.; Potts, M., Eds.; Kluwer Academic
Publishers: Dordrecht, The Netherlands, 2000; pp 591-611.
(2) Levine, E.; Thiel, T. J. Bacteriol. 1987, 169, 3988–3993.
(3) Bebout, B. M.; Garcia-Pichel, F. Appl. EnViron. Microbiol. 1995, 61, 4215–
4222.
(4) For an overview, see: Cockell, C. S.; Knowland, J. Biol. ReV. 1999, 74,
311–345.
(5) Scytonemin structure determination: Proteau, P. J.; Gerwick, W. H.; Garcia-
Pichel, F.; Castenholz, R. Experientia 1993, 49, 825–829.
(6) Garcia-Pichel, F.; Castenholz, R. W. J. Phycol. 1991, 27, 395–409.
(7) (a) Stevenson, C. S.; Capper, E. A.; Roshak, A. K.; Marquez, B.; Grace,
K.; Gerwick, W. H.; Marshall, L. A. Inflamm. Res. 2002, 51, 112–114. (b)
Stevenson, C. S.; Capper, E. A.; Roshak, A. K.; Marquez, B.; Eichman,
C.; Jackson, J. R.; Mattern, M.; Gerwick, W. H.; Jacobs, R. S.; Marshall,
L. A. J. Pharmacol. Exp. Ther. 2002, 303, 858–866.
(8) Soule, T.; Stout, V.; Swingley, W. D.; Meeks, J. C.; Garcia-Pichel, F. J.
Bacteriol. 2007, 189, 4465–4472.
(9) Pang, S. S.; Duggleby, R. G.; Schowen, R. L.; Guddat, L. W. J. Biol. Chem.
2004, 279, 2242–2253, and references therein.
(10) Ohshima, T.; Nishida, N.; Bakthavatsalam, S.; Kataoka, K.; Takada, H.;
Yoshimura, T.; Esaki, N.; Soda, K. Eur. J. Biochem. 1994, 222, 305–312.
(11) For a review, see: Land, E. J.; Ramsden, C. A.; Riley, P. A. Methods
Enzymol. 2004, 378, 88–109.
(12) Kobayashi, A.; Kajiyama, S.; Inawaka, K.; Kanzaki, H.; Kawazu, K. Z.
Naturforsch., C: Biosci. 1994, 49c, 464–470.
(13) Oshima, T.; Nagata, S.; Soda, K. Arch. Microbiol. 1985, 141, 407–411.
(14) The absolute and relative stereochemistry of 8 and 9 have not yet been
assigned.
(15) When treated with 4 or 5 individually, NpR1276 will catalyze homocoupling
at a greatly reduced rate (see Supporting Information).
(16) For a review, see: Phol, M.; Lingen, B.; Mu¨ller, M. Chem. Eur. J. 2002,
8, 5289–5295.
(17) Regioselective coupling of nonidentical but similarly reactive substrates
by ThDP-dependent enzymes is limited to certain isozymes of acetohy-
droxyacid synthase: McCourt, J. A.; Duggleby, R. G. TRENDS Biochem.
Sci. 2005, 30, 222–225.
(18) The instability and poor aqueous solubility of pyruvic acids 4 and 5 have
prevented quantitative kinetic analysis and comparison of NpR1276 with
related ThDP-dependent enzymes.
(19) For the structures and isolation references of these related natural products,
see Supporting Information.
Finally, a number of biologically active acyloin- and diol-
containing natural products from various sources appear to be
assembled using related biosynthetic logic.19 Each of these
molecules could conceivably arise from the action of a ThDP-
dependent enzyme on pyruvic acid containing substrates diverted
from primary metabolism. Such enzymes may represent a rich,
untapped source of biocatalysts capable of selective CsC bond
formation.
In summary, we have proposed a biosynthesis for the cyano-
bacterial pigment scytonemin and characterized two enzymes
utilized in the early stages of this pathway. Further study of the
NpR1276-catalyzed reaction, as well as the other enzymatic
transformations involved in the assembly of scytonemin, will be
the focus of future investigations.
JA807192U
9
J. AM. CHEM. SOC. VOL. 130, NO. 46, 2008 15261