Long-chain N-acyl amino acid antibiotics isolated from heterologously expressed environmental DNA [20]

Full text of DOI:10.1021/ja002990u

J. Am. Chem. Soc. 2000, 122, 12903-12904
12903
Table 1. HRMS Data for CSL12-A through CSL12-M (1-13)
Long-Chain N-Acyl Amino Acid Antibiotics Isolated
from Heterologously Expressed Environmental DNA
(m/z) [M + H]+
molecular
formula
acyl side
chain
CSL12-
calcd
found
Sean F. Brady and Jon Clardy*
A (1)
B (2)
C (3)
D (4)a
E (5)
C17H25N1O4
C18H27N1O4
C19H29N1O4
C20H31N1O4
C21H33N1O4
C22H35N1O4
C23H37N1O4
C24H39N1O4
C25H41N1O4
C21H31N1O4
C23H35N1O4
C25H39N1O4
C27H43N1O4
308.1862
322.2018
336.2175
350.2331
364.2488
378.2644
392.2801
406.2957
420.3114
362.2331
390.2644
418.2957
446.3270
308.1864
322.2011
336.2176
350.2336
364.2471
378.2638
392.2797
406.2955
420.3131
362.2329
390.2652
418.2967
446.3260
C8H15O
C9H17O
C10H19O
C11H21O
C12H23O
C13H25O
C14H27O
C15H29O
C16H31O
C12H21O
C14H25O
C16H29O
C18H33O
Department of Chemistry and Chemical Biology
Cornell UniVersity, Ithaca, New York 14853-1301
ReceiVed August 11, 2000
F (6)
G (7)
H (8)
I (9)
The tiny minority of soil microorganisms that can easily be
cultured with standard techniques, roughly 0.1 to 1.0%,1 produce
a spectacular array of biologically active natural products. The
uncultured majority likely produces natural products with chemical
diversity and biological activity similar to that of cultured
microorganisms. To access the natural products produced by
uncultured microorganisms,2 a cosmid library of DNA extracted
directly from soil samples (environmental DNA, eDNA) was
constructed and screened for the production of biologically active
small molecules.3 One of the active cosmid clones, CSL12,
produces a series of long-chain N-acyl-L-tyrosine antibiotics.
Long-chain N-acyl amino acids are a growing family of bacterial
natural products for which no biosynthesis genes have yet been
identified. Analysis of the eDNA cloned in CSL12 indicated that
a single open reading frame (ORF) was responsible for the
production of these antibiotics and thus led to the identification
of what we believe to be the first long-chain N-acyl amino acid
biosynthesis gene. In this communication we report the charac-
terization of these new natural product antibiotics and the sequence
for a long-chain N-acyl amino acid synthase.
J (10)a
K (11)
L (12)
M (13)
a 4 and 10 were isolated by RPHPLC as a mixture.
cosmid from CSL12 was retransformed into E. coli, it continued
to confer antibacterial activity indicating that the cloned eDNA
was responsible for the observed activity.
A sequential two antibiotic selection scheme was used to
identify and then recover cosmid clones that produce antibacterial
activities. The cosmid library of eDNA was originally selected
on LB plates containing kanamycin and allowed to incubate at
30 °C for 2-4 days. The mature colonies were then overlayed
with top agar containing kanamycin resistant Bacillus subtilis.
After an additional 24 h of incubation at 30 °C, colonies that
produced zones of growth inhibition in the B. subtilis lawn,
indicating the production of antibacterial activity, were recovered
by streaking bacteria picked from the active colonies onto LB
plates containing ampicillin (50 µg/mL). The second selection,
on ampicillin, removes the B. subtilis assay strain and allows for
the recovery of antibacterial hits directly from the assay plates.
Of the approximately 700 000 clones screened, 65 antibacterial
active colonies were found. Since we were most interested in small
molecule antibiotics, ethyl acetate extracts from small-scale
cultures of the active clones were assayed for antibacterial activity.
One of the clones that produced a very active organic extract,
CSL12, was chosen for further characterization. When the purified
The active constituents in the ethyl acetate extract from cultures
of CSL12 were isolated using a bioassay-guided fractionation
against B. subtilis. Crude ethyl acetate extracts were obtained from
neutralized cultures grown in LB (30 µg/mL of kanamycin) at
30 °C for 60 h. The ethyl acetate extract was partitioned by normal
phase flash chromatography (CHCl3:MeOH step gradient with
0.1% HOAc) and the active material that eluted from the silica
column was then further partitioned by reversed-phase LC-CN
flash chromatography (CH3CN:H2O step gradient with 0.1%
triethylamine). Thirteen related compounds trivially named
CSL12-A through CSL12-M were isolated by reversed-phase
HPLC from the antibacterial active material that eluted from the
second flash column.4 The active material from the ethyl acetate
extract of a 3.5 kb BamH I subclone of CSL12, CSL12.1,
produced an identical reversed-phase HPLC trace, and this
subclone was used for the remainder of our characterization
studies.
(1) Hugenholtz, P.; Goebel B. M.; Pace, N. R. J. Bacteriol. 1998, 180,
4765. Bintrim, S. B.; Donohue, T. J.; Handelsman, J.; Roberts, G. P.;
Goodman, R. M. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 277. Stackebrandt,
E.; Liesack, W.; Goebel, B. M. FASEB J. 1993, 7, 232. Liesack, W.;
Stackebrandt, E. J. Bacteriol. 1992, 174, 5072. Torsvik, V.; Goksoyr, J.; Daae,
F. L. Appl. EnViron. Microbiol. 1990, 56, 782. Ward, D. M.; Weller, R.;
Bateson, M. M. Nature 1990, 345, 63.
(2) Handelsman, J.; Rondon, M. J.; Brady, S. F.; Clardy, J.; Goodman, R.
M. Chem. Biol. 1998, 5, R245.
Mass spectral analysis of CSL12-A through CSL12-M sug-
gested that they were a family of long-chain saturated and
unsaturated acyl derivatives of tyrosine, 1-13. Tyrosine was
apparent from the fragments observed in the LRESI MS spectra
(m/z 182 and 136) and the acyl side chains were deduced from
the high-resolution mass spectra (Table 1). Compounds 1-13
were all negative in a ninhydrin assay suggesting the acyl
substituents were attached through an amide bond to the nitrogen
(3) Construction of the cosmid library: A 1:1 (w:v) mixture of soil and
lysis buffer (Zhou, J.; Bruns, M. A.; Tiedge, J. M. Appl. EnViron. Microbiol.
1996, 62, 316-322.) was heated at 70 °C for 2 h and extracted once with an
equal volume of chloroform, and the eDNA was then precipitated with
2-propanol (0.6 vol) from the centrifuge clarified aqueous phase. Crude eDNA
was gel purified on preparative 0.7% agarose gels (100 V for 1 h then 20 V
for 24 h) and the high molecular weight (HMW) DNA collected by
electroelution (100 V, 3 × 30 min). Purified eDNA was partially digested
with BamH I and dephosphorylated with calf intestinal alkaline phosphatase
prior to ligating into the SuperCos I vector (Sratagene). DNA ligated into the
BamH I site in the SuperCos I vector was subsequently packaged into Gigapack
III Gold packaging extracts (Stratagene) and transformed into E. coli XL1
blue MR cells.
(4) HPLC conditions: ODP50, 10 mm × 25 cm, 4 mL/min, gradient CH3-
CN:H2O 10:90 to 40:60 over 30 min with 0.1% triethylamine.
10.1021/ja002990u CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/27/2000

12904 J. Am. Chem. Soc., Vol. 122, No. 51, 2000
Communications to the Editor
cluster, are present in CSL12.1. A BLAST7 search did not identify
any deposited sequences with sequence similarity to these ORFs.
A BLAST search against deposited sequences found one se-
quence, a hypothetical protein (MJ1207) from Methanococcus
jannaschii,8 similar to ORF1. MJ1207 is a member of a small
superfamily of hypothetical proteins that are predicted N-acyl
transferases. Although ORF1 is only distantly related to this group
of hypothetical N-acyl transferases by sequence homology, its
functional similarity, the production of N-acyl products, argues
that the sequence homology could be significant.
Long-chain N-acyl aminoacylases have been characterized from
bacteria and appear to be widespread among microorganisms.9
However, no enzymes that catalyze the formation of these
compounds have been identified and the role of this growing
family of natural products remains unclear.10 With the observation
that these long-chain N-acyl amino acids possess antibacterial
activities, it is likely that long-chain N-acyl amino acids represent
a general class of microbial antibiotics. The identification of ORF1
as an N-acyl amino acid biosynthesis gene should aid in
determining the true role of this family of natural products as
well as help in the assignment of function to other hypothetical
N-acyl transferases from bacterial and archaeal sequencing
projects.
The heterologous expression of eDNA accesses the chemical
diversity of uncultured microorganisms. CSL12-A through
CSL12-M (1-13) are the first new biologically active natural
products that have been characterized from the heterologous
expression of environmental DNA.11 This heterologous expression
approach automatically couples biologically active natural prod-
ucts with their biosynthetic genes. In this case, the characterization
of the eDNA cloned in CLS12 identified the first biosynthesis
gene for long-chain N-acyl amino acids. The straightforward
strategy for constructing and screening large eDNA libraries that
we have described should provide ready access to many of the
natural products produced by previously inaccessible microorgan-
isms. This method can also be used for heterologous expression
in alternative host organisms using cosmid shuttle vectors and
such studies are currently underway.
Figure 1. Genome priming systems (GPS) transposon mutagenesis of
CSL12.1. Transposon insertions that knock out, reduce, or do not affect
the antibacterial activity produced by CSL12.1 are marked with red,
yellow, or green flags, respectively: (a) 3.5 kb insert from CSL12.1, (b)
expanded region containing the predicted promoter for ORF1, and (c)
proposed eDNA ribosome binding site (RBS), -35 and -10 sequences
aligned with the corresponding consensus sequences from E. coli.
on tyrosine. The N-acyl-tyrosine structure and the configuration
of the tyrosine in this series of compounds was ultimately
confirmed by the total synthesis of two representative examples:
CSL12-C (3), the most abundant compound isolated from
CSL12.1, and CSL12-G (7), one of the most active compounds
isolated from CSL12.1.5 Both 3 and 7 are spectroscopically
identical to their synthetic counterparts N-decanoyl-L-tyrosine and
N-myristoyl-L-tyrosine, respectively.
CSL12-A to CSL12-M (1-13) show varying degrees of
antibacterial activity, with the length of the aliphatic substitution
on tyrosine markedly affecting the antibacterial activity of these
natural products. In assays against B. subtilis, C13 to C16 saturated
and unsaturated N-acyl substituted L-tyrosines were the most
active. Tyrosine derivatives with side chains one and two carbons
longer or shorter were markedly less active and derivatives with
even shorter side chains were essentially inactive.
The genome priming system transposon mutagenesis (New
England BioLabs, MA) of the CSL12.1 insert indicated that a
single open reading frame (ORF1) produced the observed
antibacterial activity and permitted the complete sequencing of
the CSL12.1 insert.6 With two exceptions, all of the transposon
insertions that knock out the production of antibacterial activity
were found in ORF1 (Figure 1a). The remaining two knock-out
insertions were found in what is predicted to be the endogenous
promoter in the eDNA for ORF1 (Figure 1b,c). In addition to
transposons which completely knocked out the antibacterial
activity, we also found a group of insertions both upstream and
downstream of the proposed promoter but not in ORF1 that
reduced the antibacterial activity observed in agar plate assays
(Figure 1b). The transposon insertion studies suggest that ORF1
is responsible for the antibacterial activity and that the eDNA
contains its own promoter that is successfully used by E. coli. In
vitro biochemical studies are currently underway to confirm that
ORF1 is an N-acyl transferase and not functioning indirectly to
cause the production of N-acyl-tyrosine compounds in the E. coli
host.
Acknowledgment. We thank Jo Handelsman and Robert Goodman
for helpful discussions. Mass spectra were obtained in the Mass
Spectrometry Facility, Department of Chemistry and Chemical Biology,
Cornell University. This work was supported by the NIH (CA24487)
and the David and Lucile Packard Foundation.
Supporting Information Available: NMR spectral data for 3 and
characterization of the monounsaturated sides in 11, 12, and 13 (PDF).
This material is available free of charge via the Internet at http://pubs.acs.org.
JA002990U
(7) WU-BLAST (Washington University) 2.0.
(8) Bult, C. J.; et al. Science 1996, 272, 1058.
(9) Fukada, H.; Iwade, S.; Kimura, A. J. Biochem. 1982, 91, 1734. Shintani,
Y.; Fukuda, H.; Okamoto, N.; Murata, K.; Kumura, A. J. Biochem. 1984, 96,
6343. Matsumoto, J.; Nagai, S. J. Biochem. 1972, 72, 269.
Three additional hypothetical ORFs (>150 amino acids), which
do not appear to be organized as a secondary metabolite gene
(10) Pohnert, G.; Jung, V.; Haukioja, E.; Lempa, K.; Boland, W.
Tetrahedron 1999, 55, 11275. Alborn, H. T.; Turlings, T. C. J.; Jones, T. H.;
Stenhagen, G.; Loughrin, J. H.; Tumlinson, J. H. Science 1997, 276, 945.
Yagi, H.; Corzo, G.; Nakahara, T. Biochim. Biophys. Acta 1997, 1336, 28
and references therein. Yagi, H.; Corzo, G.; Yochochi, T.; Kamisaka, Y.;
Yamaoka, M.; Nakahara, J. Jpn. Oil Chem. Soc. 1995, 44, 635. Asselineau,
J. Prog. Chem. Org. Nat. Prod. 1991, 56, 1.
(5) A slight excess of the appropriate acid chloride was stirred with 20 mg
of L-tyrosine in 1 mL of DMF at room temperature. After 3 h the reaction
was diluted with 10 mL of 1 N HCl, extracted 3× with ethyl acetate, and the
synthetic N-acyl-tyrosines were purified by reversed-phase HPLC.
(6) The nucleotide sequence of CSL12.1 has been deposited in GenBank
under Accession Number AF324335.
(11) Wang, G.; Graziani, E.; Waters, B.; Pan, W.; Li, X.; McDermott, J.;
Meurer, G.; Saxena, G.; Anderson, R.; Davies, J. Org. Lett. 2000, 2, 2401.