Total synthesis of the ribosomally synthesized linear azole-containing peptide plantazolicin a from bacillus amyloliquefaciens

Article abstract of DOI:10.1002/anie.201302266

A cyclodehydration reaction has been used to synthesize the linear azole-containing peptide plantazolicin A. The target compound was synthesized from two heterocyclic fragments derived from dipeptide building blocks. The acid-labile oxazolines and thiazoles necessitate the use of a Teoc/TMSE-based protection-group strategy, which allows structure modification of plantazolicin. Copyright

Full text of DOI:10.1002/anie.201302266

Angewandte
Chemie
Communications
Total Synthesis
S. Banala, P. Ensle,
R. D. Sꢀssmuth*
&&&& — &&&&
Total Synthesis of the Ribosomally
Synthesized Linear Azole-Containing
Peptide Plantazolicin A from Bacillus
amyloliquefaciens
A cyclodehydration reaction has been
used to synthesize the linear azole-con-
taining peptide plantazolicin A. The
target compound was synthesized from
two heterocyclic fragments derived from
dipeptide building blocks. The acid-labile
oxazolines and thiazoles necessitate the
use of a Teoc/TMSE-based protection-
group strategy which allows structure
modification of plantazolicin.
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Figure 1. Structure of plantazolicin A (1a) and suggested course of the
biosynthesis from a ribosomally synthesized prepeptide: multiple
cyclodehydration (PznB/C/D), removal of the leader (PznE), and N,N-
bismethylation (PznL).[12,13]
2
www.angewandte.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
Figure 2. Structure of plantazolicin A (1a) and its retrosynthetic analysis by dissection into two main peptide fragments and breakdown into
single heterocyclic amino acid building blocks.
DOI: 10.1002/anie.201302266
Total Synthesis
Total Synthesis of the Ribosomally Synthesized Linear Azole-
Containing Peptide Plantazolicin A from Bacillus
amyloliquefaciens**
Srinivas Banala, Paul Ensle, and Roderich D. Sꢀssmuth*
Dedicated to the Bayer company on the occasion of its 150th anniversary
thiazoles, heterocyclic amino acids derived from Ser/Thr and
The lack of new lead structures is a major obstacle in
developing drugs for the treatment of bacterial infections. A
promising yet under-investigated source for such compounds
is ribosomally synthesized and posttranslationally modified
peptides (RiPPs), a growing group of peptides, mostly
produced by microorganisms.[1] Besides the cyclic azole
peptides of the cyanobactin type (patellamides, ulithiacycla-
mide)[2–4] and thiopeptides (thiostrepton, noshiheptide),[5–8]
the linear azole-containing peptides (LAPs), which include
prominent natural products such as microcin B17 and goad-
sporin,[9,10] are an important subgroup of the RiPPs. The
distinguishing structural features of LAPs are oxazoles and
Cys[11] by enzymatic cyclodehydration and dehydrogenation
reactions.[12] Recently, we extended the LAP family with
plantazolicin A (1a) and B (1b; Figures 1 and 2). These are
produced from the Gram-positive soil bacterium Bacillus
amyloliquefaciens FZB42,[13,14] which is also used in the
industrial production of a-amylase and serine proteases.[15]
Interestingly, studies reported that plantazolicin A showed
a selective growth inhibition against nine Bacillus species,[13,16]
but not against Staphylococcus or Enterococcus species. Thus,
plantazolicin can be considered as a new lead compound to
fight Bacillus anthracis (anthrax) infections.
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
The structure of plantazolicin A was elucidated
by 15N-labeling studies facilitated by B. amylolique-
faciens cultures growing in 15N-enriched media. This
approach provided crucial signals from 15N nuclei in
2D NMR spectra (1H-15N-HSQC, 1H-15N-HMBC) to
solve the structure. From the structure elucidation of
1a, information concerning the functions of enzymes
of the corresponding biosynthesis gene cluster (pzn
cluster) was deduced (Figure 1). Hence, the struc-
tural gene PznA codes for a ribosomally synthesized
linear 41-mer precursor peptide with the 14-mer core
peptide RCTCTTIISSSSTF. The trimeric protein
complex PznBCD (cyclodehydratase, dehydrogen-
ase, and docking/scaffolding protein) likely encodes
posttranslational modifications, that is ten cyclode-
hydrations followed by nine dehydrogenations.[13]
After protease PznE cleaves off the leader peptide
to yield desmethylplantazolicin 1b, a final N,N-
bismethylation by methyltransferase PznL gives 1a.
Structurally, plantazolicin comprises of two
unprecedented extended stretches of contiguous
heterocycles. The configuration of the only non-
oxidized methyloxazoline (5-MeOxH13) stereocen-
ters were assigned as 4S,5R from the biosynthesis
logic of it being derived from l-Thr.[12] Fermentation
only yielded minute amounts of 1a, thus hampering
further studies on the molecule and its mode of
action. Furthermore, the complex structure of 1a also
represents an attractive synthetic target for total
synthesis. In this study, we present the first total
synthesis of plantazolicin A (1a). Our synthetic route
is designed in such a way that structural analogues for
structure–activity relationship (SAR) studies can be
readily prepared by iterative couplings of precursor
fragment peptides.
Scheme 1. Synthesis of heterocyclic amino acids (7, 8, 9, and 10): Reagents and
conditions: a) DAST, CH2Cl2, À788C, 2 h, 95%. b) DBU, CBrCl3, CH2Cl2, À108C,
1 h to 258C, 2 h, 79%. c) TFA, Et3SiH, CH2Cl2, 258C, 40 h, 99%. d) Dess–Martin
periodinane, CH2Cl2, 258C, 0.5 h, 95%. e) PMe3/I2/NEt3, CH2Cl2, À408C, 2 h,
68%. f) Piperidine, EtOAc (1:5, v/v), 258C, 0.5 h, 99%. g) ClCO2Et, NEt3, THF,
08C, 0.5 h, then 30% aq NH3, 258C, 3 h, 99%. h) Lawesson’s reagent, CH2Cl2,
408C, 3 h, 72%. i) BrCH2COCO2Me, KHCO3, DME, À408C, 0.5 h to À178C, 14 h;
then TFAA, 2,6-lutidine, 08C, 14 h, 86%. j) LiOH, THF, MeOH, H2O (1:1:1),
258C, 3 h, 99%. k) SiMe3Cl, NEt3, CH2Cl2, 08C, 0.5 h, 408C, 0.5 h, then (CH3)3Si-
C2H4-OCO-Suc (Teoc-OSu), NEt3, 258C, 3 days, 99%. l) 4m HCl in dioxane,
258C, 0.5 h. m) 37% HCHO (in H2O), THF, H2O (2:1), NaOAc 08C, 10 min,
then NaCNBH3, 08C, 1 h, 99%. DBU=1,8-diazabicyclo[5.4.0]undec-7-ene, TFA=
trifluoroacetic acid, TFAA=trifluoroacetic anhydride.
Our retrosynthesis strat-
egy (Figure 2) was inspired
by enzymatic cyclodehydra-
tions, which represent key
steps in the biosynthesis of 1.
To begin with, we identified
the two extended stretches of
contiguous
heterocycles
including one acid-sensitive
oxazoline as well as the selec-
tive N,N-bismethylation of the
arginyl residue as major syn-
thetic challenges. For the syn-
thesis of the heterocycles we
selected dimethylaminosulfur
trifluoride (DAST) as the
cyclodehydration agent to
synthesize oxazolines from
the respective aminoacyl-Ser
or aminoacyl-Thr precursor
peptides.[17] As an alternative,
cyclodehydrations with cata-
lytic amounts of molybdenum
oxide (MoO2) under azeo-
Scheme 2. Synthesis of N-terminal fragment 2: Reagents and conditions: a) HATU, DIPEA, CH2Cl2, 08C to
258C, 14 h, then MeOH, 90% (60% of 22, 30% of 18). b) 4m HCl in dioxane, 258C, 6 h, 99%. c) HATU,
DIPEA, DMF 258C, 14 h, then MeOH, 56% (45% of 24, 11% of 20). d) DAST, CH2Cl2, À788C, 14 h, 86%.
e) DBU, CBrCl3, CH2Cl2, 08C to 258C, 3 days, 60% (45% of 26, 15% of 26-H2). f) 37% aq HCHO, THF/
H2O (2:1), NaOAc, 08C, 10 min, then NaCNBH3, 08C, 1 h. g) Me3SnOH, 1,2-DCE, 858C, 18 h, 97%
(3 steps). HATU=O-(7-azabenzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexafluorophosphate,
DIPEA=N,N-diisopropylethylamine, DCE=1,2-dichloroethane.
4
www.angewandte.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
tropic removal of water were considered.[18]
in quantitative yield, but the dehydrogenation to 16 was not
The assembly of acid-sensitive 5-MeOxH13 (methyloxazo-
line) was planned using DAST in the penultimate step to
removal of protecting groups from the 14-mer heterocyclic
precursor of 1. For protection of the C terminus and of the
guanidine side chain (Arg1), we chose concomitantly remov-
able, fluoride-labile 2-(trimethylsilyl)ethyl ester (TMSE),[19]
and 2-(trimethylsilyl)ethyl carbamate (Teoc),[20] respectively,
both being orthogonal to acid- and base-labile protecting
groups. Protection as an TMSE ester was considered advanta-
geous to suppress diketopiperazine formation during coupling
reactions of dipeptides (e.g. 6 with 5).[21] The N,N-bismethy-
lation was planned to be established on a late-stage fragment
because the purification of Boc-protected peptides is easier.
In the retrosynthetic analysis, the structure of 1 was
dissected at the amide bond between Ile7 and 5-MeOxz6 to
obtain N-terminal fragment 2 and C-terminal fragment 3, as
activation of 5-MeOxz6 does not lead to racemization. The
fragments 2 and 3 were further dissected at (methyl)oxazole
rings 5-MeOxz3, 5-MeOxz5, Oxz9, and Oxz11 to obtain the
corresponding heterocyclic Thr/Ser-containing precursor pep-
tides 4 and 5. Subsequent amide-bond disconnections led to
dipeptide building blocks Thr-Phe 6, Ile-Ile 11, and hetero-
cyclic amino acids 7, 8, 9, and 10. The total synthesis
commenced with the synthesis of dipeptide building blocks
from protected Ser/Thr by using TBTU/DIPEA (see the
Supporting Information).[22] The subsequent cyclodehydra-
tion was achieved with DAST, as this reagent is known to
mediate cyclodehydrations of various Ser- and Thr-containing
dipeptides with inversion of the configuration at b-C.[17]
Indeed the reaction of Boc-Ser(tBu)-Ser-OMe (12) with
DAST at À788C in CH2Cl2 gave the corresponding oxazoline
in > 95% yield. Subsequent dehydrogenation with CBrCl3/
DBU provided the protected oxazole 13 in an overall yield of
70% (Scheme 1).[23] Analogously Boc-Thr(tBu)-Thr-OMe
(14) was converted into its corresponding 5-methyloxazoline
successful with the above-mentioned reagents (CBrCl3/DBU)
or under various other conditions (CCl4/DBU/pyridine,[24]
MnO2,[25] CuBr/Cu(OAc)2/tert-butyl peroxybenzoate,[26] and
DDQ). This prompted us to apply Wipfꢀs modified Gabriel–
Robinson oxazole synthesis.[27] Accordingly, the oxidation of
14 to ketone 15 with Dess–Martin periodinane and subse-
quent treatment with PMe3/I2/NEt3 provided 5-methyloxazole
16 in 68%. Thiazoles 7 and 8 were prepared by using
a modified Hantzsch thiazole procedure.[28,29] Thionation of
the primary carboxamides of Boc-Thr(OtBu) or 19 with
Lawessonꢀs reagent followed by the KHCO3-mediated addi-
tion of methyl bromopyruvate in 1,2-DME and further
treatment with TFAA/2,6-lutidine provided thiazoles 18 and
20, both in an overall yield of 60%. Initially, the N,N-
bismethylation of 7 by reductive amination seemed to be
trivial; however, considerable optimization was required.
Finally, selective cleavage of the Boc group in 7 with HCl
followed by imine formation with 37% aq HCHO in the
presence of NaOAc at 08C and reduction with NaCNBH3 at
08C gave 21 in quantitative yield.[30]
Having the heterocyclic amino acids in hand, their
iterative coupling was performed (Scheme 2). For the syn-
thesis of N-terminal fragment 4, a HATU/DIPEA-mediated
coupling of 8 and 9 yielded 22 in 60%, which was followed by
cleavage of the Boc and tBu groups with HCl to give the
bisheterocyclic compound 23. Further coupling of dipeptide 7
in the presence of HATU/DIPEA furnished heterocyclic
peptide 24 in 45% yield.
DAST-mediated cyclodehydration of peptide 24 gave 25
in 86% yield. However, the dehydrogenation of 5-methylox-
azolines with CBrCl3/DBU proved difficult. Complete dehy-
drogenation of one of the newly formed methyloxazolines
was observed after 6 h at 08C (26-H2, LC-MS analysis), but
only trace amounts of the twofold dehydrogenated compound
26 were observed. Finally we found that an excess of reagents
Scheme 3. Synthesis of C-terminal fragment 3: Reagents and conditions: a) HATU, DIPEA, DMF, À308C to 258C, 20 h, 64%. b) DAST, CH2Cl2,
À788C to À208C, 4 h. c) MnO2, toluene, 708C, 3 days, 26%. d) LiOH, THF, H2O, 258C, 2 h. e) DBU, CBrCl3, CH2Cl2, À208C to 258C, 3 days,
75%. f) 4m HCl, dioxane, 258C, 5 h.
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Scheme 4. Final stages of the total synthesis of plantazolicin A (1a): coupling of N-terminal and C-terminal peptides, oxazoline formation, and
deprotection. Reagents and conditions: a) HATU, DIPEA, DMF, 258C, 14 h, 30%. b) DAST, CH2Cl2, À788C, 12–16 h, 81%. c) TBAF, DMF, 258C,
1.5 h. d) (CF3)2CHOH, 258C, 48 h, 91% (2 steps). TBAF=tetra-n-butylammonium fluoride.
in a reaction over three days was needed to effect dehydro-
genation of the second oxazoline, although it did not drive the
reaction to completion. Fortunately, the mono- and bisoxi-
dized products could be separated by flash chromatography
(yields: 15% 26-H2 and 50% 26). N,N-Bismethylation was
carried out on fragment 26 by reductive amination under the
conditions optimized for 21. Final hydrolysis of the methyl
ester with Me3SnOH provided the N-terminal fragment 2 in
97% yield.
For the construction of the C-terminal fragment
(Scheme 3), Boc-Ile-Ile-OH (11) was coupled with oxazole
10 in the presence of HATU/DIPEA to obtain peptide 27 in
64% yield. Cyclodehydration with DAST proceeded quanti-
tatively; however, dehydrogenation with MnO2 in hot toluene
gave 28 in 26% yield only. Here, dehydrogenation with
CBrCl3/DBU was not successful, even after prolonged
reaction times, presumably for steric reasons. After hydrolysis
of the methyl ester with LiOH, a second coupling of 10 in the
presence of HATU/DIPEA gave 30 in 57% yield. Saponifi-
cation of the methyl ester with LiOH set the stage for the
third coupling of dipeptide 6. To our delight, HATU-
mediated coupling proceeded smoothly to give the 8-mer
fragment 32 in 81% yield. The DAST-mediated oxazoline
formation and CBrCl3/DBU dehydrogenation readily gave
the contiguous tetraoxazole product 33 in 75% yield. Con-
current removal of the Boc and tBu protecting groups with
HCl gave the C-terminal fragment 3 ready for final coupling
with N-terminal fragment 2. This optimized fragment cou-
pling route enables the synthesis of plantazolicin analogues,
i.e. the introduction of labels via fragment 6.
ring formation and dehydrogenation reactions. In our initial
approaches, a “zipper”-like multiple cyclodehydration was
pursued on protected linear peptides, synthesized by Fmoc-
based solid-phase synthesis. Reaction of N-terminal fragment
Boc-Arg(Z)2-Cys-Thr(Bn)-Cys-Thr(Bn)-Thr(Bn)-OMe with
MoO2 gave two thiazolines along with multiple eliminations
of benzyl groups to yield dehydroalanine-type products.[31]
Likewise the use of Tf2O/OPPh3 gave an intractable mixture
of by-products.[32] Similarly, the reaction of C-terminal frag-
ment Boc-Ile-Ile-Ser-Ser-Ser-Ser-Thr(tBu)-Phe-OMe with
DAST resulted in four cyclodehydrations, but the oxidation
was incomplete when either CBrCl3/DBU or MnO2 in hot
toluene was used, as well as a variety of other methods
mentioned above during the synthesis of 16. During the
stepwise synthesis of the contiguous tetraoxazole peptide
Boc-Ile-Ile-Oxz9-Oxz10-Oxz11-Oxz12-OMe, the cyclodehydra-
tion of Boc-Ile-Ile-Oxz9-Oxz10-Ser11-Oxz12-OMe with DAST
followed by CBrCl3/DBU gave the desired tetraoxazole
product, but it turned out that its solubility in a range of
organic solvents was too low to continue along this route. The
solubility of the azoles also caused difficulties in later stages
of the total synthesis, especially during chromatographic
purifications, which could be attributed to low yields in later
steps.
Having heterocylic fragments 2 and 3 in hand, fragment
condensation with HATU/DIPEA provided the 14-mer
heterocyclic peptide 34 in 30% yield (Scheme 4). As planned,
DAST-mediated final cyclodehydration went smoothly to
afford 5-methyloxazoline 35 in 81% yield. Deprotection with
TBAF trihydrate in DMF did remove the TMSE, but only one
of the guanidine Teoc protecting groups. Excess reagent and
prolonged reaction times did not yield the desired product.
The use of TFA, TFA/NEt3, or NEt3·3HF cleaved all the
At this point we consider it necessary to discuss other
attempts and problems we encountered during the synthesis
of plantazolicin, which were mostly related to heterocyclic
6
www.angewandte.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!