Reevaluation of the final steps in the biosynthesis of blasticidin S by Streptomyces griseochromogenes and identification of a novel self-resistance mechanism

Article abstract of DOI:10.1016/S0040-4020(99)01060-1

The final steps in the biosynthesis of the antifungal peptidyl- nucleoside blasticidin S (3) have been revised to include a novel self- resistance mechanism wherein the previously proposed final precursor, demethylblasticidin S (7), is modified with a leucine residue yielding leucyldemethylblasticidin S (10) which exhibits reduced antibiotic activity. Methylation of 10 yields leucylblasticidin S (9) which can be exported from the bacterium and hydrolyzed to 3. Also disclosed is the finding of a blasticidin S N-acetyltransferase activity that may function to detoxify 3 and 7 inadvertently produced prior to export or which gain reentry to the cell. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4020(99)01060-1

TETRAHEDRON
Pergamon
Tetrahedron 56 (2000) 693–701
Reevaluation of the Final Steps in the Biosynthesis of Blasticidin S
by Streptomyces griseochromogenes and Identification of a Novel
Self-Resistance Mechanism
Qibo Zhang,a Martha C. Cone,a Steven J. Goulda,b and T. Mark Zabriskiec,
*
aDepartment of Chemistry, Oregon State University, Corvallis, OR 97331, USA
bMerck Research Laboratories—Basic Research, Rahway, NJ 07065, USA
cCollege of Pharmacy, Oregon State University, Corvallis, OR 97331, USA
Received 14 October 1999; revised 28 November 1999; accepted 7 December 1999
Abstract—The final steps in the biosynthesis of the antifungal peptidyl-nucleoside blasticidin S (3) have been revised to include a novel self-
resistance mechanism wherein the previously proposed final precursor, demethylblasticidin S (7), is modified with a leucine residue yielding
leucyldemethylblasticidin S (10) which exhibits reduced antibiotic activity. Methylation of 10 yields leucylblasticidin S (9) which can be
exported from the bacterium and hydrolyzed to 3. Also disclosed is the finding of a blasticidin S N-acetyltransferase activity that may
function to detoxify 3 and 7 inadvertently produced prior to export or which gain reentry to the cell. ᭧ 2000 Elsevier Science Ltd. All rights
reserved.
Introduction
Members of the peptidyl-nucleoside family of antibiotics
are characterized by a modified nucleoside possessing an
amino sugar coupled via an amide linkage to an amino
acid or short peptide.1,2 Three well-studied examples
with potent biological activities are puromycin (1), the
nikkomycins (e.g. nikkomycin Z (2)) and blasticidin S (3).
Puromycin and blasticidin S block protein biosynthesis in
eukaryotic and prokaryotic cells3–5 while nikkomycins are
inhibitors of fungal chitin synthase.6 Studies on the bio-
syntheses of 1–3 have advanced to the molecular level and
gene clusters encoding structural and resistance genes for
these three antibiotics have been cloned.7–9 Investigations
of peptidyl-nucleoside biosynthesis in this laboratory have
focused primarily on blasticidin S (BS, 3), a Streptomyces
griseochromogenes metabolite once used on a large scale in
Asia as a fungicide for the prevention of rice blast.10
Incorporation experiments with isotopically labeled
compounds identified the primary precursors to BS as cyto-
sine, d-glucose, l-arginine, and l-methionine.11 More
advanced intermediates identified in the biosynthesis of 3
include b-arginine (4),12 cytosylglucuronic acid (CGA,
5),13,14 cytosinine (6),15 and demethylblasticidin S (DBS,
7)16 and led to the proposed overall pathway to 3 depicted
in Scheme 1.
One of the unanswered questions in BS formation is how
S. griseochromogenes is protected from the action of BS.
Antibiotic producing organisms usually require a self-
resistance mechanism to prevent being subject to the lethal
Keywords: amino acids and derivatives; antifungals; nucleosides; biosynth-
esis.
* Corresponding author; e-mail: mark.zabriskie@orst.edu
0040–4020/00/$ - see front matter ᭧ 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(99)01060-1

694
Q. Zhang et al. / Tetrahedron 56 (2000) 693–701
Scheme 1. Proposed pathway for the biosynthesis of blasticidin S (3) by S. griseochromogenes.
effect of the substances they make. Several self-resistance
mechanisms have been described in antibiotic-producing
Streptomyces including antibiotic inactivation, target modi-
fication, efflux pumps, and combinations of these mecha-
nisms.17 Antibiotic inactivation is generally achieved by
N-acetylation or O-phosphorylation. In some instances, inter-
mediates generated during the biosynthesis are toxic and
must be converted to an inactive form. Examples of such
cases are found in puromycin8,18–22 and bialaphos23–25 forma-
tion. The fully assembled antibiotics are then deprotected
when they no longer present a threat to the producer. In the
example of puromycin, three methylation steps proceed
after N-acetylation of an active intermediate.8 The last
step of the biosynthesis involves active transport of N-acetyl-
puromycin from the cell and hydrolysis of the acetamide on
the exterior to release the active substance into the surround-
ings.26 Here we report results of recent experiments to deter-
mine the mechanism of self-resistance towards blasticidin S
in the producing bacterium S. griseochromogenes and the
findings that led us to further examine and redefine the final
steps in the biosynthetic pathway.
not been previously identified in S. griseochromogenes but
a blasticidin S acetyltransferase has been detected, isolated
and characterized from the BS-producing actinomycete
Streptoverticillium sp. JCM 4673.29,30 The enzyme has an
Mr of 15 kDa and exhibits Kms of 2 and 3 mM for BS and
acetyl CoA, respectively. The inhibition of protein synthesis
by BS could be relieved in a cell-free system when the
enzyme was added along with acetyl CoA. Other nucleoside
antibiotics including puromycin and streptothricin were not
substrates for the purified BS N-acetyltransferase.30 A gene
encoding the acetyltransferase, bls, has been cloned and
expressed in Streptomyces lividans.31 The bls gene is now
a commercially available resistance marker used in generat-
ing stable mammalian cell lines (Invitrogen Corp., Carlsbad,
CA).
The only other documented mechanism of resistance
towards BS is deamination of the cytosine moiety to
produce the corresponding inactive uracil analog. Blasticidin
S deaminase has been characterized in the non-producing
bacterium Bacillus cereus.32 The most extensively studied
deaminase is from the fungus Aspergillus terreus. The
fungal bsd gene has been cloned and expressed, and
crystals of the enzyme suitable for X-ray diffraction
studies have been produced.33 S. lividans 66 was shown to
inactivate BS by a means other than acetylation but
the product of the degradation/derivatization was not
characterized.34
Results and Discussion
Mechanisms of blasticidin S self-resistance in
S. griseochromogenes
Antibiotic modification. Some peptidyl-nucleoside anti-
biotics such as puromycin and streptothricin are N-acetyl-
ated by the producing organism to prevent binding to its own
ribosomes.20,27 In the case of puromycin, the biologically
inactive N-acetylpuromycin is exported via a proton-
dependent electrochemical gradient-driven efflux pump
and deacetylated by an amidohydrolase to generate the
active antibiotic.8,26,28 N-Acetylblasticidin S (AcBS) has
An unanswered issue pertaining to BS acetylation is which
site on the molecule becomes modified. In the earlier
studies on AcBS formation the evidence for acetylation
was based on the consumption of acetyl CoA or the loss
of antibacterial activity.30,31 The product of the acetyl-
transferase reaction was not characterized and the actual
function that AcBS plays in the cell was not addressed.

Q. Zhang et al. / Tetrahedron 56 (2000) 693–701
695
Scheme 2. Preparation of N-acetylblasticidin S (8).
Synthesis of N-acetylblasticidin S and detection in fermen-
tation broth. Blasticidin S (3) was treated with acetic
anhydride in acetic acid to give N,N0-diacetylblasticidin
S. Selective basic hydrolysis of the cytosine acetamide in
dilute ammonium hydroxide gave N-acetylblasticidin S
specifically acetylated at the b-arginine amine (AcBS, 8,
Scheme 2). Cation-exchange HPLC analysis of the synthetic
material revealed a peak with a retention time and UV
chromophore closely matching a peak observed in a
chromatogram of wild-type S. griseochromogenes complex
medium fermentation broth. Confirmation that the fermen-
tation broth contained 8 came from HPLC coinjection
analysis on both ion exchange and C18 reverse phase
columns and FAB mass spectrometry. This finding is
consistent with the report that small amounts of N-acetyl-
puromycin are found in the culture medium of the producing
organism.26 AcBS was not detected when S. griseochromogenes
was grown in a chemically defined medium.
Antibacterial activity of DBS and AcDBS. Studies on puro-
mycin illustrated that the first biosynthetic intermediate in a
pathway with potentially lethal activity is the logical target
for self-resistance through derivatization. Demethylblasti-
cidin S (DBS, 7) was reported to exhibit antibiotic activity
slightly less than that of 3 and it is reasonable to assume that
acetylation might be required to block this action.35 DBS
was isolated from a complex medium fermentation of
S. griseochromogenes supplemented with methionine
hydroxamate, a methionine adenosyltransferase inhibitor.36
A standard of N-acetyldemethylblasticidin S (AcDBS) was
synthesized from 7 using the same procedure described for
the preparation of AcBS (Scheme 2). At 3.8 mg/disk DBS
produced an 18 mm clear zone of inhibition in the disk
diffusion assay versus B. circulans. No antibiotic activity
was observed with AcDBS at levels up to 190 mg/disk.
Acetylation of demethylblasticin S in a cell-free extract. If
AcBS is a true pathway intermediate, the previous result
suggests that DBS is the more likely target for inactivation
by acetylation. To test this, DBS and acetyl CoA were
incubated with an aliquot of S. griseochromogenes CFE
and analyzed for the presence of AcDBS. HPLC analysis
revealed a peak identical in retention time and UV profile to
that of authentic AcDBS. Coinjection confirmed this find-
ing. FABMS analysis showed a molecular ion at m/z 451,
consistent with the expected [MϩH]ϩ ion for AcDBS.
Control assays using either boiled CFE or lacking added
acetyl CoA contained no detectable AcDBS.
Antibacterial activity of AcBS. The ability of 8 to inhibit the
growth of Bacillus circulans, a BS-sensitive bacterium, was
evaluated by disk-diffusion bioassay on solid medium. No
activity was observed with levels of 8 as high as 410 mg/
disk. Control assays using 3 at 3.8 mg/disk resulted in a
24 mm zone of inhibition. Thus, the antibiotic activity of
3 is effectively eliminated by acetylation at the b-arginine
amine and implicates this as a possible self-protection
process in the producing organism.
Detection of blasticidin S N-acetyltransferase activity. To
determine if S. griseochromogenes possesses detectable BS
acetyltransferase activity, BS and acetyl CoA were
incubated with aliquots of a S. griseochromogenes cell-
free extract (CFE) and analyzed for AcBS production.
Cation-exchange HPLC analysis confirmed the formation
of 8 and this finding was further supported by coinjection.
AcBS was not detected in control experiments containing
boiled CFE or in which acetyl CoA was omitted. Similar
activity was also detected in the supernatant of a CFE
sample centrifuged at 100,000×g for 30 min. These results
indicate that S. griseochromogenes possesses a soluble
N-acetyltransferase similar to that found in Streptoverticillium
sp. JCM 4673.
Table 1. Apparent kinetic parameters for the acetylation of BS (3) and DBS
(7)
app
max
Substrate
Kampp (mM)
V
(nmol minϪ1 mgϪ1
)
Vmax/Km
BS (3)
DBS (7)
4.6^0.2
14^2
0.60^0.01
1.0^0.1
0.13
0.07
Identification of the preferred substrate for the N-acetyl-
transferase. To determine if either BS or DBS is preferen-
tially modified by the N-acetyltransferase(s) in S.
griseochromogenes, the kinetic parameters Km and Vmax
were measured for both substrates at a constant acetyl
CoA concentration (Table 1). Using Vmax/Km as a measure
of specificity, these data suggest BS should be the preferred
substrate at low concentrations in vivo. A direct competition
experiment was also conducted in which equimolar quanti-
ties of BS and DBS were incubated with the CFE and acetyl
CoA and the rate of production of each product was moni-
tored (Fig. 1). The competition study revealed that AcDBS
accumulated more rapidly and in greater quantity than
AcBS. The amount of the AcDBS was 20% greater than
Role of N-acetylblasticidin S. The actual function of
AcBS in the blasticidin biosynthetic pathway has never
been fully addressed. If AcBS is a true intermediate in the
pathway, Scheme 1 must be amended to accommodate
points where the acetyl group is added and removed. We
therefore proceeded to determine at which stage acetylation
occurs and identify a AcBS amidohydrolase activity.

696
Q. Zhang et al. / Tetrahedron 56 (2000) 693–701
that of AcBS after 1 h and 30% greater after 3 h. The results
of these specificity experiments suggest neither compound
is the overwhelmingly preferred substrate for the acetyl-
transferase. It should be noted that a crude preparation
from Streptoverticillium sp. JCM 4673 was found to
acetylate BS as well as puromycin and streptothricin
while the purified enzyme exhibited specificity for BS.30
The possibility that more than one enzyme is acting on
these compounds prevents a conclusion to be made about
substrate specificity in this case. It is plausible that the
results are consistent with the action of a general detoxify-
ing enzyme.
levels were periodically monitored by HPLC over 48 h. The
amount of AcBS remained constant during the fermentation
while a substantial quantity of BS was produced. The
amount of BS produced was negligibly different than a
control fermentation that did not contain added AcBS and
AcBS was not detected in this control. These findings
indicate that AcBS is unlikely to serve as an intermediate
of BS biosynthesis. The absence of an enzyme to deacetyl-
ate AcBS suggests the N-acetyltransferase may act as a
detoxifying enzyme used by S. griseochromogenes when
free DBS or BS is inadvertently formed or produced in
excess under artificial culture conditions.
Intermediacy and role of leucylblasticidin S in the BS
pathway. In 1968, Seto et al. isolated two compounds that
were viewed as potential precursors to BS. One was
demethylblasticidin S (DBS 7) discussed above, the
second was leucylblasticidin
S
(LBS, 9).35,37 Both
compounds require a single transformation for conversion
to 3. LBS was originally identified as a minor metabolite of
S. griseochromogenes when culture conditions were kept
below pH 4. LBS was later isolated from Streptomyces sp.
SCC 1785 along with Sch36605, a LBS analog possessing a
5-hydroxymethylcytosine.38 Similar to AcBS, these
compounds are acylated at the b-arginine amine. LBS was
suggested to be a direct precursor to BS when it was
observed that incubating LBS with washed S.
griseochromogenes cells resulted in the formation of BS.37
To determine if the leucyl modification serves as a self-
resistance mechanism we conducted similar experiments
to those used to evaluate the role of AcBS.
Figure 1. Progress curves for the formation of AcBS and AcDBS.
Methylation of acetyldemethylblasticidin S. For AcDBS to
be a true biosynthetic intermediate to BS it must first
be methylated at the guanidine followed by acetamide
hydrolysis. Previous isotope trapping experiments using
14C-labeled S-adenosylmethionine (AdoMet) demonstrated
that DBS is a substrate for a methyltransferase present in a
CFE.16 However, only 0.85% of the added DBS was
converted to BS after 20 h. When we incubated AcDBS
and AdoMet with S. griseochromogenes CFE, HPLC analy-
sis revealed a 21% conversion of AcDBS to AcBS in 1 h.
The identity of AcBS was confirmed by mass spectrometry.
In a control experiment using boiled CFE, AcBS was not
detected. When DBS and AdoMet were incubated with CFE
the amount of BS detected was insignificantly different from
the low background level of BS present in the boiled
control. This establishes AcDBS as the superior substrate
for the guanidine methyltransferase.
Antibacterial activity of leucylblasticidin S and leucylde-
methylblasticidin S. LBS was isolated from the fermentation
broth of S. griseochromogenes grown in a chemically
defined medium using ion exchange chromatography
followed by reverse phase HPLC. In our studies LBS
production did not require the fermentation be kept below
pH 4.37 The production of LBS by Streptomyces sp. SCC
1785 was also reported under normal fermentation condi-
tions.38 If LBS is the last intermediate in the biosynthetic
pathway to 3, the earlier finding that AcDBS is a better
methyltransferase substrate than DBS suggests that LBS
should be formed by methylation of leucyldemethylblacti-
cidin S (LDBS, 10). LDBS was synthesized by coupling 7
with N-t-BOC-l-leucine-N-hydroxysuccinimide followed
directly by removal of the BOC group in trifluoroacetic
acid without purification (Scheme 3). LDBS was purified
by reverse phase HPLC and characterized spectroscopically.
Deacetylation of N-acetylblasticidin S. With strong
evidence suggesting S. griseochromogenes acetylates DBS
as a self-resistance mechanism, it remained to demonstrate
the final step—removal of the acetyl group from AcBS.
There are no reports on detection of an amidohydrolase
activity that can convert AcBS to BS and all our attempts
to directly detect such an activity in either fermentation
broth or CFE have failed. In another experiment to see if
AcBS is converted to BS, AcBS was added to a fermenta-
tion in chemically defined medium at the beginning of anti-
biotic production (48 h after inoculation) and AcBS and BS
The antibiotic activity of LDBS and LBS was evaluated
against B. circulans. At levels of 75 mg/disk, 15 and
20 mm zones of inhibition were observed, respectively,
for 9 and 10. For comparison, 3.8 mg of BS produced a
24 mm zone of inhibition on the same agar plate. Formation

Q. Zhang et al. / Tetrahedron 56 (2000) 693–701
697
Scheme 3. Preparation of leucyldemethylblasticidin S (10).
of the leucyl derivative does not abolish activity towards the
test bacterium as does acetylation, but activity is decreased
by 20-fold and this supports modification of the toxic DBS
with a leucine as a plausible mechanism of self protection.
Formation of leucylblasticidin S in cell-free extracts. Seto et
al. also described efforts to detect LBS formation from BS
and l-leucine.37 This was addressed because of concerns
that low pH fermentation conditions could promote the
reversal of the hydrolysis reaction. The authors provide
evidence that, under the conditions used, radiolabeled BS
is not a precursor in vivo to LBS. Because BS might not
readily be transported into S. griseochromogenes cells we
attempted to detect synthesis of LDBS from DBS and
l-leucine in a cell-free system. ATP was included because
the leucine carboxyl is most likely activated as an acyl-
adenylate or a phosphate mixed anhydride. We observed
no detectable LDBS formation using HPLC analysis. One
obvious explanation for the failure to detect LDBS is the
facile hydrolysis to DBS. It is also conceivable that a
b-arginyl–leucine dipeptide is coupled to cytosinine (6) to
yield LDBS. The investigation of this possibility will have
to await preparation of a properly labeled precursor.
Conversion of leucyldemethylblasticidin S to leucylblasticidin
S. To determine if LDBS is a substrate for the previously
detected guanidine methyltransferase, LDBS was incubated
with S. griseochromogenes CFE and AdoMet for 3 h at 30ЊC.
HPLC analysis showed that all the LDBS had disappeared but
only a small amount of LBS was produced. The major peak
formed was DBS, which should arise from hydrolysis of
LDBS. The BS peak increased three-fold compared to the
small amount always present in the boiled controls. Increasing
the incubation time to 8 h resulted in complete disappearance
of the LBS peak. In control assays using boiled CFE, amounts
of LDBS remained unchanged and no DBS or LBS was
detected. In a second control experiment that did not contain
AdoMet, LBS was not detected and the BS peak was at the
background level, whereas all the LDBS was converted to
DBS. The identities of peaks in the chromatograms were
confirmed by coinjecting standard samples.
The above findings provide evidence for a revision of the
final steps in blasticidin S formation as outlined in Scheme
4. The results indicate that modification of BS and DBS with
an acetyl group may serve as a detoxification process while
the actual self-resistance mechanism is the previously
unreported formation of a leucyl derivative. To our
knowledge, this is the first example of a bioactive secondary
metabolite being derivatized with an amino acid as a means
of self-resistance.
HPLC analysis of the assay mixtures indicated 19% of the
added LDBS was methylated and subsequently converted to
BS. If the competing hydrolysis of LDBS to DBS is con-
sidered (see below), LDBS appears to be a very good
substrate for the methyltransferase. Because DBS was
shown to be a poor substrate, the sequence where LDBS
is transformed to DBS followed by methylation to yield
BS should not contribute significantly to the production of
BS. To verify this, a parallel experiment was carried out
where DBS was included in the assay mixture. The total
amount of BS present increased by only 2%, confirming
the minor contribution of this route.
Efflux mechanism. As previously mentioned, active
removal of an antibiotic or a precursor from a producing
bacterium via an efflux pump is a common mechanism of
self-resistance. We recently reported the cloning of two
DNA fragments (2.6 and 4.8 kb) from S. griseochromogenes
which confer increased resistance to BS upon S. lividans.9
Further sequence analysis of the two fragments has revealed
the presence of genes characteristic of ATP-binding cassette
transporters (ABC-transporters) associated with antibiotic
efflux pumps in some Streptomycetes.39 The 2.6 kb
resistance fragment was found to carry a gene encoding a
protein with two ATP-binding motifs while a gene on the
4.8 kb fragment of DNA appears to encode a trans-
membrane protein.40 Thus, a type I ABC-transporter,
where the hydrophilic and hydrophobic components are
encoded by different genes, appears to function in the
mechanism of resistance to BS in S. griseochromogenes.
Conversion of leucylblasticidin S to blasticidin S. Seto et al.
reported that incubating LBS with washed S. griseochromogenes
cells resulted in the formation of BS.37 We also identified
this leucylpeptidase activity in the CFE. Incubating LBS
overnight at 30ЊC with S. griseochromogenes CFE resulted
in complete conversion to BS. The level of LBS was
unchanged in the boiled control. Centrifuging the CFE
and assaying the pellet and supernatant revealed nearly
equal hydrolysis activity in both fractions. This may indi-
cate the enzyme is only loosely associated with membrane
but more extensive solubility studies are required. There
was no detectable activity in the fermentation broth.
Attempts to block LBS hydrolysis with protease inhibitors,
such as benzamidine, pepstatin A, leupeptin, and phenyl-
methylsulfonyl fluoride were unsuccessful.
The resistance-encoding DNA fragments were used to probe
a S. griseochromogenes cosmid library to identify bio-
synthetic genes clustered with the resistance determinants.9
Several cosmids hybridizing with the larger resistance

698
Q. Zhang et al. / Tetrahedron 56 (2000) 693–701
Scheme 4. Revised final steps for the biosynthesis of blasticidin S.
fragment were identified which, when expressed in
S. lividans, resulted in production of CGA (5) and other
BS-related metabolites including DBS (7), but not 3 itself
(cf. Scheme 1). There was no evidence for the presence of
AcBS and BS N-acetyltransferase activity was not detec-
table in S. lividans harboring cosmids carrying the BS
gene cluster or which were transformed with plasmids
carrying the 4.8 kb resistance fragment. Since the initial
report, we have determined that next to CGA (5), the most
abundant blasticidin-related metabolite produced by these
transformants is LBS (9) (Fig. 2). These findings further
support the revised pathway in Scheme 4 and help exclude
AcBS as a biosynthetic intermediate to BS.
Conclusion
The overall process of self-protection by S. griseochromogenes
towards BS shares many similarities with the mechanism
used by S. alboniger to escape the effects of puromycin.
Both organisms appear to acylate early intermediates and
export the final products in protected form out of the cell
where liberation of the active species occurs. The detection
of LDBS synthetase activity remains to be demonstrated but
the steps elaborating LDBS to BS have all been established.
Our findings require the revision of the earlier proposal that
DBS is the last precursor to BS and do not support acetyl-
ation as the self-resistance mechanism used under normal
conditions. Currently, efforts are underway in our laboratory
to complete the sequencing of the BS biosynthetic gene
cluster. This information will allow us to rationally target
specific genes for disruption and subcloning, permitting
the full characterization of each enzyme and chemical
transformation in the blasticidin S biosynthetic pathway.
Experimental
General experimental procedures
Many of the general procedures have been previously
described.9,41 NMR spectra were obtained on either a Bruker
AC 300 or AM 400 spectrometer. For NMR spectra
recorded in D2O, t-BuOH was added as an internal refer-
ence; d 1.27 ppm and 31.2 ppm for 1H and 13C, respectively.
HPLC analysis
Figure 2. HPLC chromatograms of S. lividans fermentation broth extract
containing blasticidins. The broad line corresponds to the extract from a
transformant expressing the blasticidin gene cluster. The thinner line repre-
sents the same extract to which authentic LBS was added.
The conditions for cation-exchange analytical HPLC were
as follows: polysulfoethylaspartamide cation exchange

Q. Zhang et al. / Tetrahedron 56 (2000) 693–701
699
column (The Nest Group, Southboro, MA); 4.6×200 mm;
mobile phase, (A) 5 mM potassium phosphate in 25%
aqueous MeCN, pH 3.0, (B) 5 mM potassium phosphate
in 25% aqueous MeCN containing 0.25 M KCl, pH 3.0,
30 min linear gradient from 0 to 100% B followed by
10 min at 100% B; flow rate 1.0 mL/min. Conditions for
reverse phase HPLC were: C18 column (Microsorb-MV,
Varian) 4.6×250 mm; mobile phase, 5% aqueous MeCN,
0.1% TFA, isocratic, flow rate 1.0 mL/min. The UV region
from 200 to 300 nm was scanned with the photodiode array
detector.
reaction. After centrifugation the resulting supernatant
(50 mL) was analyzed by ion-exchange HPLC.
N-Acetyltransferase competition experiment
A 1.0 mL assay mixture containing 0.7 mL CFE, 5 mM BS,
5 mM DBS and 8 mM acetyl CoA was incubated in a
30ЊC water bath and aliquots (100 mL) were removed at 0,
15, 30, 45, 60, 90, 120, 150 and 180 min. Each sample was
immediately mixed with an equal volume of cold EtOH,
centrifuged, and the resulting supernatant (50 mL) analyzed
by HPLC. A control experiment lacking acetyl CoA was
carried out under identical conditions. The rate of pro-
duction of AcBS and AcDBS was calculated from the
differences of the two incubations.
Preparation of cell-free extracts
Standard culture maintenance and fermentation conditions
have been described previously.12 S. griseochromogenes
grown in a chemically defined medium was generally
harvested at 72 h after inoculation. Mycelia were pelleted
by centrifugation at 10,000×g for 10 min at 4ЊC and washed
twice by resuspending in five volumes of 50 mM potassium
phosphate buffer, pH 7.0, containing 1 mM dithiothreitol
and centrifuging. The pellet was either used directly or
stored at Ϫ80ЊC. In general, pellets were suspended in
two volumes of buffer, cooled in an ice/ethanol bath, and
disrupted by sonication (Heat Systems Ultrasonic, Inc.
model W-225, power level 8, 40% duty cycle). Sonication
was for 3×1 min with 1 min cooling intervals. Boiled CFE
was obtained by immersion in boiling water for 10 min.
Methylation of acetyldemethylblasticidin S
In a total volume of 0.5 mL, 0.35 mL CFE, 1 mM AcDBS
and 23 mM AdoMet were combined and incubated at 30ЊC
water for 10 h. The reaction was stopped with the addition
of 0.5 mL cold EtOH and the mixture was centrifuged. The
supernatant was analyzed by ion-exchange HPLC. Control
experiments contained boiled CFE or lacked AdoMet.
Diacetylblasticidin S. Blasticidin
S
hydrochloride
(800 mg) was dissolved in water (50 mL) and the pH
adjusted to 11.0 with Amberlite 410 resin (OHϪ form).
Removal of the resin and lyophilization of the filtrate gave
630 mg of BS as free base. The free base (500 mg) was
dissolved in AcOH (50 mL) and Ac2O (50 mL) and the
reaction mixture was stirred overnight at room temperature.
The resulting mixture was evaporated to a small volume
(15 mL) and stored at 4ЊC overnight. The resulting crystals
were collected and recrystallized from H2O (5 mL) to give
Bioassay
Sterile peptone agar solution (50ЊC) was mixed with
0.2% v/v of a stock spore suspension of B. circulans and
aliquots (10 mL) were dispensed into Petri dishes and
allowed to solidify. Sterile paper bioassay disks were treated
with the solution to be assayed and distributed evenly on the
agar plates. The diameter of the inhibition zone was
measured after incubation at 37ЊC for 16 h.
1
260 mg (43%) of diacetylblasticidin S. H NMR (D2O) d
1.84 (m, 1H), 1.96 (m, 1H), 2.03 (s, 3H), 2,27 (s, 3H), 2.50
(d, Jˆ6.6 Hz, 2H), 3.03 (s, 3H), 3.37 (m, 1H), 3.47 (m, 1H),
4.15 (m, 1H), 4.23 (d, Jˆ8.5 Hz, 1H), 4.78 (1H, under
HOD), 5.94 (d, Jˆ10.3 Hz, 1H), 6.19 (d, Jˆ10.2 Hz, 1H),
6.58 (s, 1H), 7.35 (d, Jˆ7.5 Hz, 1H), 8.07 (d, Jˆ7.5 Hz,
1H). 13C NMR (D2O) d 23.6, 25.7, 31.8, 37.1, 42.8, 46.3,
46.9, 48.6, 79.2, 81.8, 100.1, 126.7, 134.7, 148.8, 158.2,
158.9, 164.8, 173.9, 175.2, 175.6, 176.2. HRFABMS
calcd for C21H31N8O7 [MϩH]ϩ 507.2316, found m/z
507.2317.
Detection of N-acetyltransferase activity in S.
griseochromogenes CFE
For typical assays, 3 mM BS or DBS and 6 mM acetyl CoA
were incubated with 0.4 mL CFE in a total volume of
0.5 mL in a 30ЊC water bath. After 8–12 h, 0.5 mL cold
EtOH was added and the mixture was centrifuged. The
supernatant was analyzed directly by ion-exchange HPLC
or evaporated under reduced pressure and the residue redis-
solved in water for reverse phase HPLC analysis. Control
experiments contained boiled CFE or lacked acetyl CoA.
N-Acetylblasticidin S (8). Diacetyl BS (100 mg) was stir-
red in 1 M aqueous ammonium hydroxide at room tempera-
ture for 5 h. The solution was evaporated to dryness and the
residue was crystallized from water to give 82 mg (89%) of
1
8. H NMR (D2O) d 1.83 (m, 1H), 1.98 (m, 1H), 2.03 (s,
Determining kinetic parameters for the acetylation of
blasticidin S (3) and demethylblasticidin S (7)
3H), 2.49 (d, Jˆ6.7 Hz, 2H), 3.03 (s, 3H), 3.37 (m, 1H),
3.46 (m, 1H), 4.16 (m, 1H), 4.19 (d, Jˆ8.9 Hz, 1H), 4.70
(ddd, Jˆ8.8, 3.0, 2.4 Hz, 1H), 5.88 (dt, Jˆ10.3, 1.8 Hz, 1H),
6.06 (d, Jˆ7.5 Hz, 1H), 6.13 (dt, Jˆ10.3, 2.0 Hz, 1H), 6.50
(s, br, 1H), 7.63 (d, Jˆ7.5 Hz, 1H). 13C NMR (D2O) d 23.6,
31.8, 37.1, 42.8, 46.3, 47.1, 48.6, 79.3, 81.2, 98.2, 127.4,
134.5, 144.5, 158.2, 159.3, 168.0, 174.0, 175.2, 176.4.
HRFABMS calcd for C19H29N8O6 [MϩH]ϩ 465.2210,
found m/z 465.2209.
For BS: 0.5 mL incubation mixtures contained 0.4 mL CFE,
3 mM acetyl CoA and one of the following concentrations
of 3: 0.25, 0.31, 0.40, 0.56, 1.07 or 6.0 mM. For DBS:
0.25 mL incubation mixtures contained 0.2 mL CFE,
3 mM acetyl CoA and one of the following concentrations
of 7: 5.0, 6.1, 7.8, 10.9, 17.8, or 50 mM. The assay mixtures
were incubated at 30ЊC for 30 min and an equal volume of
cold EtOH was added to each incubation to terminate the
N-Acetyldemethylblasticidin S. DBS free base (7)

700
Q. Zhang et al. / Tetrahedron 56 (2000) 693–701
(110 mg) in AcOH (6 mL) and Ac2O (6 mL) was stirred for
2 h at room temperature and then evaporated to dryness.
Spectroscopic analysis showed the residue contained a
mixture of mono- and di-acetylated 7. The residue was
then dissolved in 5 mL of 1 M aqueous ammonium
hydroxide and stirred for 2 h at room temperature. After
lyophilization and recrystallization from MeOH/H2O,
56 mg of AcDBS was obtained from two crops of crystals
(46%). 1H NMR (D2O) d 1.79 (m, 1H), 1.89 (m, 1H), 2.02
(s, 3H), 2.50 (d, Jˆ6.8 Hz, 2H), 3.25 (m, 2H), 4.20 (d,
Jˆ8.9 Hz, 1H), 4.24 (m, 1H), 4.72 (ddd, Jˆ8.8, 3.4,
2.1 Hz, 1H), 5.88 (d, Jˆ10.3 Hz, 1H), 6.07 (d, Jˆ7.5 Hz,
1H), 6.13 (d, Jˆ10.2 Hz, 1H), 6.51 (s, 1H), 7.64 (d,
Jˆ7.5 Hz, 1H). 13C NMR (D2O) d 23.6, 33.7, 39.5, 42.7,
46.2, 47.1, 79.4, 81.2, 98.1, 127.3, 134.6, 144.6, 158.4,
158.8, 167.6, 174.0, 175.3, 176.4. HRFABMS calcd for
C18H27N8O6 [MϩH]ϩ 451.2054, found m/z 451.2054.
Acknowledgements
Prof. Haruo Seto is thanked for helpful advice on the
preparation of AcBS. Dr Yukio Miyazaki, Kaken Pharma-
ceutical Co., Ltd, is thanked for providing a culture of
S. griseochromogenes and samples of blasticidin S. This
work was supported by grant GM 32110 from the National
Institutes of Health.
References
1. Isono, K. J. Antibiot. 1988, 41, 1711.
2. Gould, S. J. Blasticidin S and Related Peptidyl Nucleoside
Antibiotics. Biotechnology of Antibiotics; Strohl, W. R. Ed.;
Marcel Dekker: New York, 1997, p 703.
3. Nathans, D.; Lipmann, F. Proc. Natl. Acad. Sci. U.S.A. 1961,
47, 497.
4. Yamaguchi, H.; Yamamoto, C.; Tanaka, N. J. Biochem. 1965,
57, 667.
5. Yamaguchi, H.; Tanaka, N. J. Biochem. 1965, 60, 632.
6. Gaughran, J. P.; Lai, M. H.; Kirsch, D. R.; Silverman, S. J.
J. Bacteriol. 1994, 176, 5857.
7. Bormann, C.; Mohrle, V.; Bruntner, C. J. Bacteriol. 1996, 178,
1216.
Leucyldemethylblasticidin S (10). To a solution of DBS
hydrochloride (7, 50 mg) and NaHCO3 (15 mg) in H2O
(2 mL), was added a solution of BOC–l-leucine–N-
hydroxysuccinimide (72 mg) in EtOH (3 mL). The reaction
mixture was stirred at room temperature for 2 h and then
acidified to pH 3 with 1N HCl. The solution was evaporated
to dryness under vacuum, the residue dissolved in a 1:1
mixture of TFA/CH2Cl2 (10 mL), stirred for 2 h at room
temperature and evaporated to dryness. The residue was
dissolved in H2O and purified by reverse phase HPLC
(Varian C18 Microsorb-MV, 4.6×250 mm, mobile phase:
7% aqueous CH3CN containing 0.1% TFA, flow rate
1.0 mL/min). Fractions containing 10 were pooled and
´
8. Tercero, J. A.; Espinosa, J. C.; Lacalle, R. A.; Jimenez, A.
J. Biol. Chem. 1996, 271, 1579.
9. Cone, M. C.; Petrich, A. K.; Gould, S. J.; Zabriskie, T. M.
J. Antibiot. 1998, 51, 409.
10. Yonehara, H. Blasticidin S: Properties, Biosynthesis and
Fermentation. Biotechnology of Industrial Antibiotics; Vandamme,
E. Ed.; Marcel Dekker: New York, 1984, p 651.
11. Seto, H.; Yamaguchi, I.; Otake, N.; Yonehara, H. Agric. Biol.
Chem. 1968, 32, 1292.
1
evaporated to yield 54 mg (92%). H NMR (D2O) d 0.98
(d, Jˆ6.0 Hz, 3H), 0.99 (d, Jˆ5.9 Hz, 3H), 1.70 (m, 3H),
1.87 (m, 2H), 2.52 (dd, Jˆ14.8, 7.9 Hz, 1H), 2.63 (dd,
Jˆ14.9, 5.6 Hz, 1H), 3.26 (m, 2H), 4.00 (t, Jˆ7.3 Hz,
1H), 4.33 (m, 1H), 4.49 (d, Jˆ7.6 Hz, 1H), 4.8 (1H under
HOD), 5.98 (d, Jˆ10.3 Hz, 1H), 6.24 (d, Jˆ7.8 Hz, 1H),
6.27 (dt, Jˆ10.3, 2.3 Hz, 1H), 6.57 (d, Jˆ2.0 Hz, 1H), 7.82
(d, Jˆ7.8 Hz, 1H). 13C NMR (D2O) d 22.5, 23.3, 25.4, 34.1,
39.5, 41.7, 42.0, 45.9, 46.6, 53.5, 76.5, 80.6, 96.9, 125.8,
134.8, 147.1, 150.2, 158.4, 161.0, 171.6, 173.5, 173.9.
HRFABMS calcd for C22H36O6N9 [MϩH]ϩ 522.2789,
found m/z 522.2782.
12. Prabhakaran, P. C.; Woo, N. T.; Yorgey, P. S.; Gould, S. J.
J. Am. Chem. Soc. 1988, 110, 5785.
13. Guo, J.; Gould, S. J. J. Am. Chem. Soc. 1991, 113, 5898.
14. Gould, S. J.; Guo, J. J. Bacteriol. 1994, 176, 1282.
15. Gould, S. J.; Zhang, Q. J. Antibiot. 1995, 48, 652.
16. Gould, S. J.; Guo, J. Bioorg. Med. Chem. Lett. 1991, 1, 497.
17. Cundliffe, E. Annu. Rev. Microbiol. 1989, 43, 207.
18. Perez-Gonzalez, J. A.; Vara, J.; Jimenez, A. Biochem.
Biophys. Res. Commun. 1983, 113, 772.
19. Sugiyama, M.; Paik, S. Y.; Nomi, R. J. Gen. Microbiol. 1985,
131, 1999.
Conversion of leucylblasticidin S (9) to blasticidin S (3)
´
20. Vara, J.; Malpartida, F.; Hopwood, D. A.; Jimenez, A. Gene
1985, 33, 197.
S. griseochromogenes CFE (0.2 mL) was incubated with
1 mM 9 in a total volume of 0.25 mL at 30ЊC bath overnight.
Cold ethanol (0.25 mL) was added and the mixture was
centrifuged. The supernatant was analyzed by ion-exchange
HPLC. Control experiments were prepared the same way
using boiled CFE.
21. Vara, J.; Perez-Gonzalez, J. A.; Jimenez, A. Biochemistry
1985, 24, 8074.
22. Vara, J. A.; Pulido, D.; Lacalle, R. A.; Jimenez, A. Gene 1988,
69, 135.
23. Murakami, T.; Anzai, H.; Imai, S.; Satoh, A.; Nagaoka, K.;
Thompson, C. J. Mol. Gen. Genet. 1986, 205, 42.
24. Kumada, Y.; Anzai, H.; Takano, E.; Murakami, T.; Hara, O.;
Itoh, R.; Imai, S.; Satoh, A.; Nagaoka, K. J. Antibiot. 1988, 41,
1838.
25. Raibaud, A.; Zalacain, M.; Holt, T. G.; Tizard, R.; Thompson,
C. J. J. Bacteriol. 1991, 173, 4454.
26. Lacalle, R. A.; Tercero, J. A.; Vara, J.; Jimenez, A.
J. Bacteriol. 1993, 175, 7474.
27. Horinouchi, S.; Furaya, K.; Nishiyama, M.; Suzuki, H.;
Beppu, T. J. Bacteriol. 1987, 169, 1929.
Methylation of leucyldemethylblasticidin S (10)
Assay mixtures (0.25 mL) containing 0.20 mL CFE, 10 mM
AdoMet and 1 mM LDBS (10) were incubated in a 30ЊC
water bath overnight. After 0.25 mL cold ethanol was
added, the mixture was centrifuged and the supernatant
analyzed by ion-exchange HPLC. Control experiments
contained boiled CFE, lacked AdoMet or contained 1 mM
DBS (7).

Q. Zhang et al. / Tetrahedron 56 (2000) 693–701
701
28. Tercero, J. A.; Lacalle, R. A.; Jimenez, A. Eur. J. Biochem.
1993, 218, 963.
34. Nomura, H.; Paik, S. Y.; Izuhara, M.; Sugiyama, M. Korean
Biochem. J. 1991, 24, 327.
35. Seto, H.; Yonehara, H. J. Antibiot. 1977, 30, 1022.
36. Gould, S. J.; Guo, J.; Dejesus, K.; Geitmann, A. Can. J. Chem.
1994, 72, 6.
37. Seto, H.; Otake, N.; Yonehara, H. Agric. Biol. Chem. 1968, 32,
1299.
29. Sugiyama, M.; Takeda, A.; Paik, S.-Y.; Nimi, O.; Nomi, R.
J. Antibiot. 1986, 39, 827.
30. Sugiyama, M.; Takeda, A.; Paik, S.-Y.; Nimi, O. J. Antibiot.
1989, 42, 135.
31. Perez-Gonzalez, J.-A.; Ruiz, D.; Esteban, J. A.; Jimenez, A.
Gene 1990, 86, 129.
32. Endo, T.; Furuta, K.; Kaneko, A.; Katsuki, T.; Kobayashi, K.;
Azuma, A.; Watanabe, A.; Shimazu, A. J. Antibiot. 1987, 40, 1791.
33. Nakasako, M.; Kimura, M.; Yamaguchi, I. Acta Crystallogr.,
Sect. D: Biol. Crystallogr. 1999, 55, 547.
38. Cooper, R.; Conover, M.; Patel, M. J. Antibiot. 1988, 41, 123.
´
39. Mendez, C.; Salas, J. A. FEMS Microbiol. Lett. 1998, 158, 1.
40. Cone, M. C.; Petrich, A. K.; Gould, S. J.; Zabriskie, T. M.
Unpublished data.
41. Zhang, Q.; Gould, S. J.; Zabriskie, T. M. J. Nat. Prod. 1998,
61, 648.