Preparation of thermocleavable conjugates based on ansamitocin and superparamagnetic nanostructured particles by a chemobiosynthetic approach

Article abstract of DOI:10.1002/chem.201404502

A combination of mutasynthesis, precursor-directed biosynthesis and semisynthesis provides access to new ansamitocin derivatives including new nanostructured particle-drug conjugates. These conjugates are based on the toxin ansamitocin and superparamagnetic iron oxide-silica core shell particles. New ansamitocin derivatives that are functionalized either with alkynylor azido groups in the ester side chain at C-3 are attached to nanostructured iron oxide core-silica shell particles. Upon exposure to an oscillating electromagnetic field these conjugates heat up and the ansamitocin derivatives are released by a retro-Diels-Alder reaction. For example, one ansamitocin derivative exerts strong antiproliferative activity against various cancer cell lines in the lower nanomolar range while the corresponding nanostructured particle-drug conjugate is not toxic. Therefore, these new conjugates can serve as dormant toxins that can be employed simultaneously in hyperthermia and chemotherapy when external inductive heating is applied.

Full text of DOI:10.1002/chem.201404502

DOI: 10.1002/chem.201404502
Full Paper
&
Medicinal Chemistry
Preparation of Thermocleavable Conjugates Based on
Ansamitocin and Superparamagnetic Nanostructured Particles by
a Chemobiosynthetic Approach
Lena Mancuso,[a] Tobias Knobloch,[a] Jessica Buchholz,[a] Jan Hartwig,[a] Lena Mçller,[a]
Katja Seidel,[a] Wera Collisi,[b] Florenz Sasse,[b] and Andreas Kirschning*[a]
Abstract: A combination of mutasynthesis, precursor-direct-
ed biosynthesis and semisynthesis provides access to new
ansamitocin derivatives including new nanostructured
particle–drug conjugates. These conjugates are based on
the toxin ansamitocin and superparamagnetic iron oxide–
silica core shell particles. New ansamitocin derivatives that
are functionalized either with alkynyl- or azido groups in the
ester side chain at C-3 are attached to nanostructured iron
oxide core–silica shell particles. Upon exposure to an oscillat-
ing electromagnetic field these conjugates heat up and the
ansamitocin derivatives are released by a retro-Diels–Alder
reaction. For example, one ansamitocin derivative exerts
strong antiproliferative activity against various cancer cell
lines in the lower nanomolar range while the corresponding
nanostructured particle-drug conjugate is not toxic. There-
fore, these new conjugates can serve as dormant toxins that
can be employed simultaneously in hyperthermia and
chemotherapy when external inductive heating is applied.
Introduction
mutasynthesis can ideally be combined with chemical synthe-
sis to generate even greater structural diversity.[4,5] Precursor-
directed biosynthesis (PDB) is related to mutasynthesis in that
biosynthetic precursor-analogues are supplemented to the fer-
mentation cultures, but here producer strains are used that do
not depend on external supplementation of advanced key pre-
cursors. In contrast to mutasynthesis, yields of the desired ana-
logues are often low due to the internal competition between
natural and unnatural precursors and separation from the con-
comitantly formed natural derivatives poses extra difficulties.[2]
Among other natural products,[6] the ansamitocins (3–5),[8]
being representatives of the group maytansinoids[7] that
belong to the class of ansamycin antibiotics, have become
a showcase for combining chemical synthesis with mutasyn-
thesis. In this context, particularly mutants of the ansamitocin
producer Actinosynnema pretiosum, blocked in the biosynthesis
of the starting building block 3-amino-5-hydroxybenzoic acid
(AHBA, 1)[9,10] or late stage-blocked mutants have been
employed.[11]
Besides total and semisynthesis the combination of chemical
synthesis with methods from the field of modern biotech-
nology and metabolic engineering have emerged as new strat-
egies to carry out natural product synthesis with the aim to
generate compound libraries of complex secondary metabo-
lites.[1] In this context mutasynthesis,[2] a technique originating
from precursor-directed biosynthesis,[3] and related concepts
based on interference with biosynthesis are highly useful to
create structural diversity, otherwise difficult to achieve purely
by total or semisyntheses. Mutasynthesis is based on the gen-
eration of mutants of a producer organism blocked in the for-
mation of a biosynthetic pathway intermediate. Mutasynthons,
which are, chemically prepared, modified biosynthetic inter-
mediates are administered to the blocked mutant thereby
potentially restoring the biosynthetic flux to finally provide
new metabolites. Our group and others have shown that
The ansamitocins are medically highly important natural
products that inhibit growth of cancer cell lines as well as
human solid tumors at subnanomolar concentrations.[12] The
mode of action for this activity is the prevention of tubulin
polymerization that subsequently leads to apoptosis.[13] The
structural features essential for bioactivity have been identified
as the presence of an ester side-chain at C-3, the a-orientation
of the hydroxyl group at C-9 and the presence of the cyclic
structure across C-7 to C-9.[11] Currently, the maytansinoids are
one of the most intensely studied group of natural products.
In early 2013 the antibody–maytansine conjugate Trastuzumab
emtansine (T-DM1; tradename: Kadcyla) was approved by the
[a] Dr. L. Mancuso, Dr. T. Knobloch,+ J. Buchholz,+ J. Hartwig, Dr. L. Mçller,
K. Seidel, Prof. Dr. A. Kirschning
Institut fꢀr Organische Chemie und Biomolekulares Wirkstoffzentrum
Leibniz Universitꢁt Hannover, Schneiderberg 1B
30167 Hannover (Germany)
E-mail: andreas.kirschning@oci.uni-hannover.de
[b] W. Collisi, Dr. F. Sasse
Abteilung Chemische Biologie
Helmholtz Zentrum fꢀr Infektionsforschung (HZI)
Inhoffenstraße 7, 38124 Braunschweig (Germany)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201404502.
Chem. Eur. J. 2014, 20, 1 – 12
1
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FDA as chemotherapeutic agent against Her2 positive breast
cancer.[14]
Here, we report, for the first time, on the introduction of
ester side chains functionalized for further semisynthetic modi-
fications during fermentation by precursor-directed biosynthe-
sis (PDB) and show that this approach can be combined with
mutasynthesis to achieve two structural changes in the ansa-
mitocins during one bioprocess. Besides standard chemical
manipulations and oligomerizations of ansamitocins, we also
conjugate ansamitocin via the ester side chain to superpara-
magnetic nanostructured particles (size: 10–40 nm) called
MAGSILICA employing a thermolabile linker (Figure 1).
The maytansinoids are biosynthesized by a type I polyketide
synthase (PKS) into which a nonribosomal peptide synthetase-
like loading module introduces AHBA as starter building
block.[16] Chain extension by seven consecutive PKS modules,
followed by macrolactamization catalyzed by the amide syn-
thase Asm9 provides proansamitocin (2; Scheme 1).[17,18] Six
post-PKS tailoring enzymes furnish the ansamitocins 3–5.[15] In
vivo, the acyltransferase Asm19 transfers carboxylic acids such
as isobutyrate 8 in form of its activated CoA-thioester onto the
hydroxyl group at C-3. Ansamitocin P3 (4) is the main
metabolite formed by A. pretiosum under most fermentation
conditions. Isobutyrate (8) can be derived from l-valine (7) by
oxidative deamination (by the branched-chain amino acid
transaminase) and oxidative decarboxylation (by the branched-
chain a-keto acid dehydrogenase enzyme complex).[19] Wheth-
er the apparent preference for the isobutyrate side chain can
be attributed to the substrate preference of the enzyme or
simply results from the abundant internal supply of this partic-
ular activated carboxylic acid, is unclear. Nevertheless, utiliza-
tion of 3 gLꢀ1 l-valine in the fermentation medium is known
to stimulate ansamitocin production and to lead to almost ex-
clusive formation of the P-3 type (4), greatly improving yields
and simplifying chromatographic clean-up.[12f]
Figure 1. Concept for the combination of hyperthermia with chemotherapy
using drug–nanostructured particle conjugates 9.
Modification of ansamitocins at C-3 has proven attractive,
especially for conjugating tumor-specific ligands to an ansami-
tocin warhead.[20] This can be achieved by a semisynthetic
approach starting from a mixture of ansamitocins produced by
fermentation. After LiAlH(OMe)3-promoted removal of the
different ester side chains, the reagent system DCC/ZnCl2
mediates re-esterification with the desired acyl side chain in
about 30% yield over two steps.[21]
The nanostructured MAGSILICA particles are composed of
a ferritic core (maghemite g-Fe2O3 and magnetite Fe3O4)
encased in a silica shell and fulfil the requirements for being
applied in the design of nanostructured–drug particle conju-
gates. The size is suitable for passing through blood vessels
while the silica shell prevents agglomeration, ensures chemical
resistance and provides options for chemical modification of
its surface. A distinguishing fea-
ture of the particles is their su-
perparamagnetism.
When
brought into close proximity of
solid tumors, these new conju-
gates could induce apoptosis in
two ways: 1) through hyperther-
mia[22] mediated by superpara-
magnetic nanostructured parti-
cles when an external oscillating
electromagnetic field is applied,
and 2) by the chemotoxic effect
ensuing from thermal cleavage
of the linker and liberation of
the highly toxic ansamitocin de-
rivative. While we prepared this
report, Cheon et al. reported on
the development of drug–nano-
structured particle conjugate
based on the toxin geldanamy-
cin and a thermolabile linker
that relies on a radical cleavage
protocol.[23]
Scheme 1. Biosyntheses of ansamitocins P-2 (3), P-3 (4) and P-4 (5) via proansamitocin (2) (postketide transforma-
tions from 2 to 3–5 are listed along with gene annotations according to the predominant route of metabolic flux)
and biosynthesis of isobutyrate (8) from l-valine (7).[15]
&
&
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
The present report covers two main topics: 1) the side chain
modification and functionalization of ansamitocin by advanced
PDB and mutasynthesis, and 2) the fusion with superparamag-
netic nanostructured particles that can release the attached
ansamitocin derivatives through inductive heating.
acids still affected production strongly, resulting in yields esti-
mated by analytical reversed phase HPLC to drop to about
10%. With a view on the potential symbiosis of maytansinoid-
producing strains with certain plants,[7] fresh coconut water
was investigated as an additive. Fresh coconut water is known
as a stimulating additive in plant cell culture.[24] In our experi-
ments it proved to be the additional key additive to apparently
counter the toxic effects of the unnatural carboxylic acids
when added to the fermentation broth in 3.3% (v/v). Further-
more, our usual mutasynthetic approach of using a drop-wise
addition of precursors via syringe pump-driven feeding capilla-
ries over 3 days proved essential for these experiments (see
the Supporting Information).
Results and Discussions
Side-chain modifications at C3 by PDB and mutasynthesis
In view of the fact that AP3 (4) is the major metabolite formed
by A. pretiosum under standard fermentation conditions with
3 gLꢀ1 l-valine (7) added as production stimulating and side
chain-directing additive (for mechanism see Scheme 1), with
yields usually ranging from 80–110 mgLꢀ1, we investigated the
introduction of functionalized ester side chains at C-3 under al-
tered fermentation conditions by supplementing the function-
alized carboxylic acids 10 or, in case of the more complex
amides, the corresponding SNAC ester 11, a SCoA mimetic
(Scheme 2).
Under the optimized fermentation conditions three new
ester side chains could be introduced at C-3 and provided an-
samitocin derivatives 12a, 13a and 14a in excellent, syntheti-
cally useful amounts. To a small degree also derivatives not N-
methylated by Asm10 (12b, 13b and 14b) were formed. The
two shortest alkyne-functionalized carboxylic acids 10a and
10b proved pronouncedly toxic and inhibited cell growth,
likely by disruption of fatty acid b-oxidation (vide infra). On the
other hand, 6-heptynoic acid 10e, which was not incorporated
as such, first underwent b-oxidation to the SCoA-activated de-
rivative of 10c, so that ansamitocin 12a was formed instead.
Likewise, efforts to incorporate the maytansine side chain
derivatives 10g/h and 11 were not successful.[8a] In case of the
N-demethylated precursor 10g, first amide hydrolysis must
have occurred because only formation of metabolite 12a was
detected. Neither acids 10 f, 10h, 10j, 10k, nor SNAC ester 11
resulted in corresponding ansamitocin derivatives.
These results are remarkable since the xenobiotic carboxylic
acids can exert inhibitory effects on fatty acid b-oxidation once
being activated as CoA-thioesters, disrupting bacterial metabo-
lism and growth. Particularly w-alkynyl-functionalized carboxy-
lates such as 4-pentynylate show broad toxicity.[25] To the best
of our knowledge, the literature only provides one example for
the successful incorporation of an alkynylate, but analysis was
merely based on mass spectrometry.[26] In contrast to alkyne
moieties, the azido group is not found in nature. Some reports
note that organic azides may have mutagenic properties.[27]
We also achieved the first example of a combined mutasyn-
thetic and precursor-directed biosynthetic (PDB) approach em-
ploying the AHBA(ꢀ) mutant strain of A. pretiosum supple-
mented with 5-hexynoic acid (10c) and aminobenzoate 15
(Scheme 3). Expectedly, isolated yields of new secondary
metabolites 16a and 16b were lower than for fermentations
with mutasynthon 15 under standard fermentation conditions
(~17%).[5]
Scheme 2. Side chain modification of ansamitocins by PDB. [a] Conditions:
K-medium, AHBA 1 (1.25 mmolLꢀ1), 10 (12.5 mmolLꢀ1) both in H2O/DMSO
(1:1), 7–10 days of fermentation (see also the Supporting Information);
[b] yields in mgLꢀ1 refer to yields achieved under optimum conditions for
isolated products.
In addition to the toxic effects of alkynoic and azido carbox-
ylic acids, neutralized prior to addition, the lack of the stimulat-
ing additive l-valine resulted in very low productivity in initial
experiments. With a view on the studies carried out by Hatano
et al.,[12e] different amino acid additives were investigated. A
combination of l-valine and l-threonine added to a final con-
centration of 1 and 3 gLꢀ1, respectively, resulted in excellent
stimulation of productivity, but did not favor the isobutyrate
side chain unduly. However, addition of the toxic carboxylic
Under the optimized fermentation conditions three new
ester side chains could be introduced at C-3 and provided an-
samitocin derivatives 12a, 13a and 14a in useful amounts. To
a lesser extent also the N-demethylated metabolites (12b, 13b
and 14b) were formed. The two shortest alkyne-functionalized
carboxylic acids 10a and 10b inhibited cell growth that we as-
cribe to inhibition of fatty acid b-oxidation (vide supra). On the
other hand, 6-heptynoic acid (10e) first underwent b-oxidation
before being incorporated as activated SCoA-ester 10c, so that
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
3
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Scheme 4. 1,3-Dipolar cycloadditions of alkynes 12a and 13a, respectively,
with benzyl azide (17).
Scheme 3. Access to ansamitocin derivatives 16a,b by combined muta-
synthesis and PDB (see also Scheme 2 and the Supporting Information).
We extended this semisynthetic concept to the preparation
of dimeric ansamitocins 22 and 23 as well as trimers 24 and
25, which were formed by cycloaddition of alkynes 12a and
13a, respectively, and the bis- as well as trisazide 20 and 21
(Scheme 5). The concept of creating multivalency by dimeriz-
ing complex natural products has been pursued for several
drugs[31] including neomycin.[32,33]
metabolite 12a was formed instead. Likewise, efforts to incor-
porate the maytansine side chain were not successful.[8a] In
case of the N-demethylated precursor 10g first amide hydroly-
sis must have occurred because only formation of metabolite
12a was detected. Neither acids 10 f, 10h, 10j, 10k nor SNAC
ester 11 were incorporated during fermentation.
On each stage of our studies towards conjugates composed
of ansamitocin and nanostructured particles, we evaluated the
antiproliferative profiles using cultured mouse fibroblast cells
and different human tumor cell lines (Table 1). All new deriva-
tives 12a/b, 13a, 14a and 16a/b with modified ester side
chains show strong antiproliferative activity similar to AP-3 (4)
towards mouse and human cell lines, the latter being a prereq-
uisite for the development of tumor-targeting conjugates.
Table 1. Antiproliferative activities IC50 [nmolLꢀ1] of 12a,b, 13a, 14a and
16a,b and comparison with AP-3 (4) towards different mammalian cell
lines.[a]
Cell line[b]
12a
12b
13a
14a
16a
16b
4
L-929
1.3
–
5.7
0.42
–
1.1
–
1.0
1.1
0.2
KB-3-1
U-937
A-431
SK-OV-3
PC-3
0.11
0.11
0.2
0.35
0.25
0.35
0.25
0.04
0.04
0.1
0.09
0.07
0.08
0.1
0.05
0.05
0.1
0.17
0.13
0.15
0.2
0.17
0.01
0.08
0.05
0.06
–
0.03
0.16
0.1
0.19
0.13
0.1
0.02
0.08
0.05
0.35
0.06
0.05
0.08
0.12
0.07
0.19
0.16
0.1
MCF-7
HUVEC
0.02
[a] Values shown are means of two determinations in parallel. [b] L-929
(mouse fibroblasts), KB-3-1 (cervix carcinoma), U-937 (histiocytic lympho-
ma), PC-3 (prostate adenocarcinoma), SK-OV-3 (ovary adenocarcinoma),
A-431 (epidermoid carcinoma), MCF-7 (breast adenocarcinoma), HUVEC
(umbilical vein endothelial cells).[28]
Scheme 5. 1,3-Dipolar cycloadditions with bis- and trisazides 20 and 21,
respectively.
Antiproliferative properties of triazoles 18, 19 and 22–25
were also determined and except for the cell line U-937 (histio-
cytic lymphoma, see Table 2), which still showed overall high
sensitivity, biological activities were lower by a factor of about
10 (for the dimeric ansamitocins 22 and 23), 100 (for benzyl-
triazoles 18 and 19) and 1000 (for trimers 24 and 25) com-
pared to AP-3 (4). Still, all values are in the lower nanomolar
range, a sufficiently low value for anticancer drugs. But it
needs to be noted, that in the present case the concept of
oligomerizing drugs and creating multivalency did not lead to
improved biological activities.
1,3-Dipolar cycloadditions with modified side chains
En route to new ansamitocin conjugates composed of side
chain modified ansamitocins and superparamagnetic, nano-
structured particles we first optimized the linker design which
included investigations on 1,3-dipolar cycloadditions by
making use of the additional chemical functionality introduced
by PDB. Both ansamitocin derivatives 12a and 13a reacted
smoothly with benzyl azide 17 in the presence of a stoichio-
metric amount of the copper(I) source and Hꢁnig’s base to
yield triazoles 18 and 19, respectively (Scheme 4).[29,30]
&
&
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
4
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Table 2. Antiproliferative activities IC50 [nmolLꢀ1] of 18, 19 and 22–25.[a]
Cell line
18
19
22
23
24
6.4
8.7
12.4
4.1
50
15.1
3.8
25
9.5
9.9
21
4.1
72
L-929
33
0.55
10.2
0.19
1.6
0.64
1.2
0.23
0.66
0.95
0.74
0.55
0.58
1.8
0.4
6.6
U-937
A-431
SK-OV-3
PC-3
MCF-7
HUVEC
2.7
3.1
5.3
0.81
2.3
1.0
1.1
1.6
1.5
1.1
1.8
1.1
17.1
2.8
[a] For details on cell lines see Table 1.
Development of nanostructured particle–ansamitocin
conjugates
Having established a new biotechnological route for introduc-
ing functionalized ester side chains into ansamitocins and
having demonstrated that these side chains can further be
modified by 1,3-dipolar cycloadditions, we consequently
pursued the conjugation of these ansamitocin derivatives with
superparamagnetic nanostructured particles[34] while imple-
menting a thermolabile linker at the same time. Approval of
the antibody–drug conjugate Kadcyla by the FDA made it
a frontrunner example for a general and strong trend in cur-
rent anticancer drug development, the conjugation of
extremely toxic molecules with tumor-specific ligands such as
monoclonal antibodies,[35] vitamins[5,36] and epidermal growth
factors.[37] These conjugates can guide the toxin to solid
tumors with high selectivity. Commonly, they are internalized
by endocytosis followed by liberation of the toxin inside
the tumor cell.
Alternatively, the linker system can be designed in a way
that allows specific cleavage of the drug outside the cancer
cell by an external trigger. This can be the lower pH value in
close vicinity of solid tumors,[38] enzyme-mediated hydroly-
sis[38–40] or heat.[23] This field of research is governed by: 1) the
Scheme 6. Side-chain functionalization of alkyne 13a and azide 14a, respec-
tively, by Huisgen-type 1,3-dipolar cycloadditions.
choice of the ideal toxin, 2) the development of a smart linker nanostructured particle was equipped with maleimide. For that
system that is cleavable under bioorthogonal conditions, and purpose we coupled modified azide 26 with alkyne 13a by
3) the quest for a target specific delivery systems.[41]
a
copper-mediated cycloaddition to furnish ansamitocin
In order to combine hyperthermia with chemotherapy derivative 27 (Scheme 6).
based on toxins such as ansamitocin derivatives, a thermocleav-
Likewise, azido-functionalized ansamitocin derivative 14a
able linker is required. The modified drug should be attached was modified with alkynes 29 and 30 to provide new ansami-
to the nanoparticle surface in a defined way and it must be ef- tocins 31 and 32, respectively. As the reaction between azide
ficiently released during exposure to an external oscillating 14a and alkyne 30 proceeded in low yield, we restricted the
electromagnetic field. Loading and unloading should proceed Diels–Alder cycloaddition as well as retro Diels–Alder reaction
in bioorthogonal manner. Additionally, the drug must be ther- studies to furan derivatives 27 and 31. N-Methylmaleimide
mally stable, a prerequisite fulfilled by the ansamitocins.[42]
served as a model dienophile and yielded adducts 28 (from
A simple and reliable thermocleavable linker can be adapted 27) and 33 (from 31) in good isolated yield under low temper-
from by a concept known from the field of self-healing poly- ature conditions (acetonitrile, 658C, 3 days). The retro-Diels–
mers.[43] The concept relies on a Diels–Alder system composed Alder reaction proceeded at elevated temperature (1108C) in
of a maleimide and furan. The coupling occurs slowly at tem- the same polar solvent within a few hours and furnished
peratures between 50–708C, while cleavage rapidly takes place ansamitocin derivatives 27 and 31 in moderate to very good
between 110–1208C. During the preparation of this paper, the isolated yields (Scheme 6).
combination of superparamagnetic nanostructured particles
Next, the medical relevance of new ansamitocin derivatives
with this linker system able to undergo retro-Diels–Alder reac- 27, 28 and 31–33 was investigated by determining their anti-
tion when inductively heated was reported.[44] We chose to proliferative properties. In addition, the cytotoxic properties of
locate the furan ring in the toxin while the superparamagnetic functionalized nanostructured particles 38, 41 and 42 were
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
tested (Table 3). Although ansamitocin derivatives 27, 31 and
32 substantially differ from AP3 (4) in the ester side chain at C-
3, all derivatives that result from 1,3-dipolar cycloadditions
showed strong or very strong antiproliferative activity against
various cancer cell lines. Another important outcome of these
antiproliferation assays is that the Diels–Alder cycloaddition
test products 28 and 33 with extended side chain at C-3 as
well as the modified nanostructured particles 38, 41 and 42
are not cytotoxic. These latter results are another important
prerequisite when utilizing these conjugates in a combined
Table 3. Antiproliferative activity IC50 [nmolLꢀ1] of 27, 28 and 31–33 and
comparison to AP3 (4).[a]
Scheme 7. Functionalization of core shell nanostructured particles (MAGSILI-
CA 37); two modes of attachment (via one or two siloxy-linkages) are
depicted.
Cell line
27
28
31
32
7.4
3.2
6.9
10.5
6.2
33
L-929
4.9
>40000
250
4.9
14
35
12
2700
190
330
310
290
KB-3-1
A-431
PC-3
0.41
0.77
2.2
–
–
–
–
ternal standard; see the Supporting Information). Then, sur-
face-modified particles 39 (20–80 mmolgꢀ1) were heated at
1208C for 2 h and the amount of coumarin derivative 40 re-
leased into solution was measured again. From these studies,
the loading was determined to be in the range of 2 to
7 mmolgꢀ1 for different samples. When these samples were
heated up a second time at 1208C for 2 h no detectable
SK-OV-3
1.6
[a] Values shown are means of two determinations in parallel. Com-
pounds 38, 41 and 42 were tested with L-929 but no inhibitory effect
was measured up to 10 mgmLꢀ1
.
therapy based on hyperthermia and chemotherapy.
Having established the principal chemical and biological pa-
rameters, we turned our attention to the surface modification
of MAGSILICA 37. The preparation of the thermolabile linker to
be attached to the nanostructured particle commenced with
the coupling of maleic anhydride (34) and 3-aminopropyltrie-
thoxysilane (APTES, 35) to yield maleimide (36; Scheme 7). In
order to prevent autopolymerization through the ethoxysilane
moieties, the preparation of maleimide and functionalization
of MAGSILICA 37 was carried out in a one-pot protocol. Forma-
tion of the open-chain form of maleimide was monitored by
the disappearance of maleic anhydride. Subsequent ring clo-
sure to 36 proceeded in the presence of the Lewis acid zinc
chloride and HMDS at 1208C within 2 h.[45] Attachment to
MAGSILICA 37 was completed after stirring at 658C for about
four days. Noteworthy, the preparation of the Diels–Alder
system in its detached form paves the way to rapidly create
various surface-modified nanostructured particles and testing
them independently with any other ansamitocin derivative.
XPS (X-ray photoelectron spectroscopy) measurements pro-
vided information on the amount of nitrogen present either as
amine or as maleimide on the surface of modified MAGSILICA.
We found that 88% of nitrogen on the silica surface of 38 was
part of maleimide moieties.
Scheme 8. Functionalization of core shell nanostructured particles 38 with
coumarin derivative 40.
amount of coumarin derivative 40 was found in the solution.
Next, the surface-modified nanostructured particles 38 un-
derwent Diels–Alder cycloadditions in the presence of the
furan-modified ansamitocin derivatives 27 and 31, respectively,
under conditions established for N-methylmaleimide
(Scheme 6). The cycloadditions were monitored (by TLC and by
HPLC) as the soluble ansamitocin derivatives slowly disap-
peared and functionalized nanostructured particles 41 and 42
formed (Scheme 9).
These qualitative results were complemented with quantifi-
cation of loading using fluorescent coumarin derivative 40 that
was attached onto functionalized nanostructured particles 38
by Diels–Alder cycloaddition. Likewise the fluorescence label
can be released by retro Diels–Alder reaction (Scheme 8). Thus,
functionalized particles 38 were treated with furan 40 at 658C
for three days and the solution was analyzed for remaining
coumarin 40 by LC-MS (using AHBA hydrochloride (1) as an in-
Finally, we studied the release of ansamitocin derivatives 27
and 31 from the modified nanostructured particles by: 1) heat-
ing with a conventional oil bath, or 2) in the presence of an ex-
ternal oscillating electromagnetic field (Table 4). In an oil bath
at 1108C complete release of ansamitocin derivatives 27 and
31, respectively, was achieved within 4 h as judged by HPLC-
MS analysis.[28] The reaction mixture was decanted from the
magnetic nanostructured particles under the influence of
&
&
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
perature of individual nanoparti-
cles is a principal problem. Com-
monly, the temperature of the
suspension is determined, which
is much lower than the surface
temperature of a suspended and
inductively heated superpara-
magnetic nanoparticle. This fact
explains why release of toxins
from nanostructured particles
like 41 proceeds more efficiently
with inductive heating than in
the oil bath and why high fre-
quency conditions are preferred
to medium frequencies (Table 3).
About 11–14% of the theoretical
amount of ansamitocin deriva-
tive 27 was released from conju-
gate 41 by inductive heating at
850 kHz. This corresponds to
about 25–30 nmol per mg nano-
particle. If one takes the mea-
sured cytotoxicity of ansamitocin
derivative 27 into account
(Table 3) only 0.16–0.19 mgmLꢀ1
of functionalized nanoparticle
41 should be sufficient to
Scheme 9. Functionalization of nanostructured particles 38 with ansamitocin derivatives 27 and 31 and thermal
cleavage by retro-Diels–Alder reaction (see also Table 3).
a magnet, the particles were washed with acetonitrile, taken
up again in the same solvent and heated in the oil bath. Sur-
prisingly, no traces of ansamitocin derivatives were detected in
the supernatant. Conventional heating proved ineffi-
cient for releasing the toxins 27 or 31 under more
physiological conditions such as in water or in the
exert complete cell destruction under the conditions of assay-
ing.
Table 4. Release of ansamitocin derivatives 27 and 31, respectively, from functional-
ized nanostructured particles 41 and 42 using conventional and inductive heating.[a]
Ringer solution.[46]
Alternatively, superparamagnetic nanostructured
particles can directly be heated under medium (mf=
15–25 kHz) or high frequency (hf=up to 780–
850 kHz) conditions. The choice of frequency can
have a dramatic effect on the efficiency of heating
nanoparticles. A stronger magnetic field (H) and
a higher frequency (f) induce more heat inside most
conductive materials as well as superparamagnetic
nanoparticles. Medium frequencies are preferentially
employed when larger objects need to be heated,
whereas high frequencies are ideally suited for small-
er objects including superparamagnetic nanomateri-
als.[34,47]
NP
retro-Diels–Alder reaction conditions (oil
bath; solvent, temperature, reaction time)
T
Yield [%]
[8C][b]
27/31[c]
41
42
MeCN, 1108C, 4 h
MeCN, 1108C, 4 h
110
110
>99
>99
NP
retro-Diels–Alder reaction conditions (mf;
solvent, power in%, sample, reaction time)
T
[8C]
Yield [%]
27/31[c]
41
42
MeCN, 100% mf, 50 mgmLꢀ1, 4 h
MeCN, 100% mf, 60 mgmLꢀ1, 4 h
40
40
2
25
NP
retro Diels–Alder reaction conditions (hf;
solvent, power in%, sample, reaction time)[b]
T
Yield [%]
27[c]
[8C]
41
41
41
41
41
41
MeCN/R, 40% hf, 10 mgmLꢀ1, 1 h
H2O, 40% hf, 10 mgmLꢀ1, 1 h
MeCN/H2O, 70% hf, 2.3 mgmLꢀ1, 1 h
H2O, 70% hf, 2.3 mgmLꢀ1, 1 h
H2O, 90% hf, 2.4 mgmLꢀ1, 1 h
H2O, 100% hf, 2.2 mgmLꢀ1, 1 h
49 (IR)
50 (IR)
31 (IR)
32 (IR)
35 (IR)
36–39 (IR)
57[e]
n.d.[d]
traces
10
traces
traces
11–14
The first experiments were carried out in acetoni-
trile at 15 kHz, which did not lead to cleavage of the
toxin from modified MAGSILICA 41. When increasing
the oscillating field to 25 kHz a small amount of ansa-
mitocin derivative 27 (2%) was released as judged by
HPLC-MS. In contrast, 25% of ansamitocin derivative
31 was liberated from the nanostructured particle
under the same conditions.
[a] NP=functionalized nanostructured particle; mf=medium frequency (15–25 kHz);
hf=high frequency (750–850 kHz); R=Ringer solution.[46] [b] Temperature of the solu-
tion was measured conventionally using a thermometer; [c] determined with a calibra-
tion curve (LC-MS-UV);[28] [d] quantification was not possible because of overlapping
signals in the UV but release of toxin was detected; [e] temperature was measured in
test tube.
It needs to be stressed, that unlike conventional
oil-bath heating, determination of the surface tem-
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
7
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
amino acid solution [l-valine 1.5% (w/v), l-threonine 4.5% (w/v)]+
1.5 mL young coconut water [sterile filtered]+3 mL preculture+
1 drop of SAG 471 anti-foam {GE Bayer Silicones}[48]) using non-baf-
fled 250 mL Erlenmeyer flasks (final volume: 45 mL per flask, with
additional steel spring). AHBA (1) and carboxylic acid derivative
10d were dissolved in DMSO/water and neutralized with NaHCO3-
(aq.) [1m] [DMSO/water 1:1; volume of feeding solution should not
exceed 10% (v/v) with respect to the recipient culture] and steri-
lized by filtration. Cultures were shaken for two days at 298C
before continuous (drop-wise) addition of 1 and carboxylic acid de-
rivative 10d over the time-course of 3 days was started using auto-
clavable, syringe pump-driven feeding capillaries (Braintree Scien-
tific BS-9000-8 syringe pump with Upchurch Scientific high-purity
Teflon PFA tubing {1/16“ OD, 0.1” ID} and Tefzel connectors). Shak-
ing was continued to a total cultivation time of 7–10 days. For de-
tection of novel products from test cultures, samples of the culture
broth (200 mL) were mixed with ethanol (200 mL), centrifuged
(20800 g, 3 min, 48C) and the clear supernatant was subjected to
UPLC-ESI-MS analysis.
Conclusions
In summary, we have reported on the development of new an-
samitocin derivatives that are conjugated via a thermolabile
linker to superparamagentic core-shell nanostructured particles
based on a ferritic core (maghemite g-Fe2O3 and magnetite
Fe3O4) and a silica shell. Synthetically, a combination of muta-
synthesis, precursor-directed biosynthesis and semisynthesis
was pursued to access new ansamitocin derivatives that are
either functionalized with an alkyno- or azido-function in the
ester side chain at C-3. To the best of our knowledge, this is
the first description of successful supplementation experiments
with such toxic carboxylic acids, resulting in isolated and syn-
thetically useful amounts of new ansamitocin derivatives. The
essential synthetic feature of the linkage between the toxin
and the nanostructured particle is a Diels–Alder cycloaddition
between a furan and a maleimide group which undergoes
a retro-Diels–Alder reaction upon heating. As the conjugates
of ansamitocin and the nanostructured particle have superpar-
amagnetic properties these can be heated in an oscillating
electromagnetic field and the cytotoxic ansamitocin derivatives
are released into solution. Most importantly, the ansamitocin
derivatives liberated after heating show strong anti-
For isolation of product 13a from large-scale fermentation, the
combined fermentation broth was extracted with ethyl acetate,
and the crude extract was subjected to a sequence of chromato-
graphic purifications as described in Table 5. Derivative 13b was
detected by HRMS and MS-MS from fermentation supernatant.
proliferative activity against several mammalian
cancer cell lines, while the same cell lines were found
Table 5. Chromatographic purification of ansamitocin derivative 13a.
to be insensitive towards nanostructured particle
drug conjugates. As cleavage can also be induced in
aqueous solutions, these conjugates can serve as
dormant toxins that can be triggered through exter-
nal inductive heating to act simultaneously in hyper-
thermia and in chemotherapy. Further studies in
animal models are currently being pursued in our
laboratories.
Sample
Phase Conditions of elution
SiO2 ethyl acetate
Fractions
crude
extract
F-1
F-1 (Rf (ethyl
acetate) 0.5–0.2)
C18-P[B] H2O [+0.1% formic acid]/MeOH [+0.1% F-2 (tR =70.0–72.0 min)
formic acid] [A/B], flow rate=15 mLminꢀ1
gradient (t [min]/B [%]):
(0/20) (5/20) (55/70) (99/85)
F-2
(five
CN-SP H2O/MeCN [A/B], flow rate=2.5 mLminꢀ1 13a (tR =85.0 min)
gradient (t [min]/B [%]):
portions)
(0/5) (5/5) (105/45)
Experimental Section
The details provided below cover synthetic key protocols for pre-
paring conjugate 27 composed of an ansamitocin derivative and
a nanostructured particle. Additional synthetic procedures towards
all other derivatives, relevant background information on fermenta-
tion experiments, testing of antiproliferative activities and copies
of NMR spectra are found in the Supporting Information.
13a: 1H NMR (400 MHz, [D4]MeOH, CHD2OD=3.31 ppm): d=7.15
(d, J=2.0 Hz, 1H, 21-H), 6.96 (d, J=2.0 Hz, 1H, 17-H), 6.63 (dd, J=
15.4, 11.0 Hz, 1H, 12-H), 6.32 (d, J=11.0 Hz, 1H, 13-H), 5.62 (dd, J=
15.4, 9.0 Hz, 1H, 11-H), 4.8 (dd, J=12.0, 2.7 Hz, 1H, 3-H), 4.19 (ddd,
J=11.0, 10.9, 3.0 Hz, 1H, 7-H), 3.98 (s, 3H, 20-OMe), 3.58 (d, J=
9.0 Hz, 1H, 10-H), 3.56 (d, J=13.0 Hz, 1H, 15-Ha), 3.37 (s, 3H, 10-
OMe), 3.32 (d, J=13.0 Hz, 1H, 15-Hb), 3.16 (s, 3H, N-Me), 2.76 (d,
J=9.6 Hz, 1H, 5-H), 2.69 (t, J=7.3 Hz, 2H, 2’-H), 2.59 (dd, J=13.9,
12.0 Hz, 1H, 2-Ha), 2.39 (t, J=2.7 Hz, 1H, 6’-H), 2.37–2.21 (m, 2H, 4’-
H), 2.14 (dd, J=13.9, 2.7 Hz, 1H, 2-Hb), 1.91–1.83 (m, 2H, 3’-H), 1.73
(s, 3H, 14-Me), 1.61–1.48 (m, 1H, 6-H), 1.58 (dd, J=13.8, 3.0 Hz, 1H,
8-Ha), 1.52 (dd, J=13.8, 11.0 Hz, 1H, 8-Hb), 1.23 (d, J=6.5 Hz, 3H, 6-
Me), 0.88 ppm (s, 3H, 4-Me); 13C NMR (100 MHz, [D4]MeOH,
[D4]MeOH=49.0 ppm): d=173.4 (s, C-1’), 171.4 (s, C-1), 157.5 (s, C-
20), 155.3 (s, 7-OCONH), 143.1 (s, C-18), 142.7 (s, C-16), 141.1 (s, C-
14), 133.8 (d, C-12), 130.0 (d, C-11), 126.0 (d, C-13), 123.3 (d, C-17),
119.8 (s, C-19), 114.9 (d, C-21), 89.7 (d, C-10), 84.7 (s, C-5’), 81.9 (s,
C-9), 78.1 (d, C-3), 75.9 (d, C-7), 70.9 (d, C-6’), 67.9 (d, C-5), 61.8 (s,
C-4), 57.2 (q, 20-OMe), 57.0 (q, 10-OMe), 47.5 (t, C-15), 39.2 (d, C-6),
37.5 (t, C-8), 36.2 (q, N-Me), 33.8 (t, C-2), 33.8 (t,C-2’), 24.9 (t, C-3’),
18.4 (t, C-4’), 15.8 (q, 14-Me), 14.7 (q, 6-Me), 12.4 ppm (q, 4-Me);
UPLC-MS [MeOH] tR =2.42 min, [MeCN] tR =2.10 min; HRMS [ESI]
Preparation of ansamitocin derivatives 13a/b using an
AHBA(ꢀ) mutant strain of A. pretiosum
Actinosynnema pretiosum HGF073[17a] was stored as spore suspen-
sion in 40% (v/v) glycerol/water at ꢀ808C, and used for the inocu-
lation of YMG agar plates. Following incubation of the plates for
4 days at 308C, 5–8 well-sporulated colonies were transferred to
a 1.5 mL tube charged with 1 mL of sterile distilled water and filled
to approximately 50% height with sterile glass beads (Ø=2 mm,
washed with dilute hydrochloric acid). After vortex-mixing, the re-
sulting suspension was used for the inoculation of precultures in
bottom-baffled 250 mL Erlenmeyer flasks charged with YMG
medium (50 mL per flask, with additional steel spring). Precultures
were shaken for 2 days at 298C before inoculation of main produc-
tion cultures (1:15 dilution). Cultivations were performed in K medi-
um with additives (37.5 mL K medium, basal composition+3 mL
&
&
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
8
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
m/z for C34H43ClN2O9Na [M+Na]+: calcd 681.2555, found 681.2563;
MS-MS {659.28 [M+H]+}: 547.22 [MꢀHOC(O)(CH2)3CꢁCH+H]+.
13b: UPLC-MS [MeCN] tR =1.92 min; HRMS [ESI] m/z for
C33H42ClN2O9 [M+H]+: calcd 645.2579, found 645.2580; MS-MS
{645.25 [M+H]+}: 533.20 [MꢀHOC(O)(CH2)3CꢁCH+H]+.
sealed and stirred at 658C for 100 h. After completion the reaction
mixture was decanted by fixing the nanostructured particles with
an external magnet and the solid was washed 12 times with 50 mL
methanol. The functionalized MAGSILICA 38 was dried under high
vacuum.
Ansamitocin derivative 27
Diels–Alder cycloaddition of modified MAGSILICA 38 with
ansamitocin derivative 27 and formation of conjugate 41
CuIBr (8.5 mg, 59 mmol, 26.0 equiv) was dissolved in degassed
methanol (3 mL) under argon atmosphere and DIPEA (30 mL,
0.16 mmol, 76.0 equiv) was added. The resulting yellowish suspen-
sion was degassed for 30 min under a stream of argon and with
exclusion of light. Ansamitocin derivative 13a (1.5 mg, 2.3 mmol,
1.0 equiv) was dissolved in degassed methanol (0.5 mL) under
argon atmosphere and treated with 240 mL of the freshly prepared
suspension of CuIBr-DIPEA and 2-(3-azidopropyl)furan 26 (0.8 mg,
5.3 mmol, 2.3 equiv). The reaction mixture was stirred at room tem-
perature under argon atmosphere and with exclusion of light for
20 h. EDTA [pH 8.0, (0.5m), 0.5 mL] was added and after being
stirred for 5 min the reaction mixture was directly purified by HPLC
(Trentec Reprosil 100 C18-ISIS 5 mm, 250 mmꢂ8 mm, with guard
column, 40 mm; gradient H2O/MeOH 95:5!90:10 over 10 min,
then 90:10!40:60 over 50 min, 40:60!20:80 over 20 min and
Ansamitocin derivative 27 (1 mg, 1.24 mmol, 1.0 equiv) was dis-
solved in acetonitrile (1 mL) and modified MAGSILICA 38 (100 mg)
was added. The mixture was heated at 658C and was stirred at this
temperature for 4 days. After cooling to room temperature the su-
pernatant was decanted using a pipette by fixation of modified
MAGSILICA 41 with an external magnet. The particles were sus-
pended in methanol several times until no ansamitocin derivative
27 could be detected as judged by LC/MS. The starting material
was recovered and purified by HPLC (C-18-ISIS phase; gradient
H2O/MeOH 95:5!90:10 over 10 min, then 90:10!40:60 over
50 min, 40:60!20:80 over 20 min and 20:80!100% MeOH over
10 min; flow 2.25 mLminꢀ1; tR =68.5 min). Ansamitocin derivative
27 (0.4 mg, 0.49 mmol, 40%) was isolated as a colorless foam.
Functionalized MAGSILICA 41 was dried in vacuo.
20:80!100% MeOH over 10 min; flow 2.25 mLminꢀ1
, tR =
76.5 min). Furfurylansamitocin derivative 27 (1.6 mg, 1.97 mmol,
87%) was isolated as a colorless foam.
Retro Diels–Alder reaction of functionalized MAGSILICA 41
1H NMR (500 MHz, [D4]MeOH CD2HCOD=3.31 ppm): d=7.83 (s,
1H, 6’-H), 7.34 (d, J=1.6 Hz, 1H, 13’-H), 7.12 (d, J=1.6 Hz, 1H, 21-
H), 6.80 (d, J=1.6 Hz, 1H, 17-H), 6.57 (dd, J=11.1, 15.4 Hz, 1H, 12-
H), 6.32 (dd, J=1.6, 3.1 Hz, 1H, 12’-H), 6.08 (d, J=11.1 Hz, 1H, 13-
H), 6.04 (d, J=3.1 Hz, 1H, 11’-H), 5.37 (dd, J=15.4, 8.9 Hz, 1H, 11-
H), 4.76 (dd, J=11.9, 2.8 Hz, 1H, 3-H), 4.39 (t, J=7.2 Hz, 2H, 7’-H),
4.18 (td, J=2.6, 10.9 Hz, 1H, 7-H), 3.97 (s, 3H, 20-OCH3), 3.56 (d, J=
8.9 Hz, 1H, 10-H), 3.54 (m, 1H, 15-Ha), 3.35 (d, J=2.2 Hz, 3H, 10-
OCH3), 3.26 (m, 1H, 15-Hb), 3.11 (s, 3H, N-CH3), 2.82 (t, J=7.3 Hz,
2H, 4’-H), 2.76 (d, J=9.8 Hz, 1H, 5-H), 2.63 (t, J=7.4 Hz, 2H, 9’-H),
2.57 (m, 3H, 3’-H, 2-Ha), 2.20 (q, J=7.2 Hz, 2H, 8’-H), 2.12 (dd, J=
12.9 Hz, 1H, 2-Hb), 2.03 (m, 2H, 2’-H), 1.69 (s, 3H, 14-CH3), 1.56–1.47
(m, 3H, 8-Ha, 8-Hb, 6-H), 1.22 (d, J=6.4 Hz, 3H, 6-CH3), 0.85 ppm (s,
3H, 4-CH3); 13C NMR (125 MHz, [D4]MeOH, [D4]MeOH=49.0 ppm):
d=173.5 (p, C-1), 171.4 (p, C-1‘), 157.4 (p, C-20), 155.5 (p, OCONH),
155.3 (p, C-10’), 148.4 (p, C-5’), 142.9 (p, C-18), 142.7 (p, C-16),
142.5 (p, C-10’), 141.1 (p, C-14), 133.8 (s, C-12), 129.7 (s, C-11), 125.9
(s, C-13), 123.6 (s, C-6’), 123.2 (s, C-17), 119.7 (p, C-19), 114.9 (s, C-
21), 111.2 (s, C-12‘), 106.6 (s, C-11‘), 89.7 (s, C-10), 81.8 (p, C-9), 78.1
(s, C-3), 75.9 (s, C-7), 67.9 (s, C-5), 61.8 (p, C-4), 57.2 (q, 20-OCH3),
57.1 (q, 10-OCH3), 50.6 (t, C-7‘), 47.4 (t, C-15), 39.2 (t, C-8), 37.5 (d,
C-6), 36.2 (q, N-CH3), 34.3 (t, C-3’), 33.8 (t, C-2), 29.9 (t, C-8‘), 25.8 (t,
C-2‘), 25.7 (t, C-4‘), 25.5 (t, C-9‘), 15.8 (q, 14-CH3), 14.7 (q, 6-CH3),
12.4 ppm (q, 4-CH3); UPLC-MS tR =2.03 min; HRMS [ESI] m/z for
C41H53N5O10Cl [M+H]+: calcd 810.3481, found 810.3450.
Functionalized MAGSILICA 41 was suspended in acetonitrile (5 mL)
and heated at 1108C for 4 h. The suspension was cooled to room
temperature in an ice bath and after fixation of nanostructured
particles with an external magnet the supernatant was decanted
using a pipette. The particles were suspended with methanol sev-
eral times. The particles were suspended in methanol several times
until no ansamitocin derivative 27 could be detected as judged by
LC/MS. The combined organic extracts were removed under re-
duced pressure and the residue was purified by HPLC (C-18-ISIS
phase; gradient H2O/MeOH 95:5!90:10 over 10 min, then
90:10!40:60 over 50 min, 40:60!20:80 over 20 min and 20:80!
100% MeOH over 10 min; flow 2.25 mLminꢀ1; tR =68.5 min). Ansa-
mitocin derivative 27 (0.6 mg, 0.74 mmol) was isolated as a colorless
foam.
The analytical and spectroscopic data are listed above.
Acknowledgements
The work was funded by the Deutsche Forschungsgemein-
schaft (Cluster of Excellence REBIRTH; “From Regenerative Biol-
ogy to Reconstructive Therapy” EXC 62), the Ministry of Sci-
ence and Culture of Lower Saxony (M.W.K.; graduate school
HSN) and by the Fonds der Chemischen Industrie (Ph.D. schol-
arship for J.H.). We thank L. Kupracs (LU Hannover), H. Herzog
and Prof. S. Katusic (EVONIK Industries AG, Essen, Germany) for
technical advice.
Preparation of surface-modified MAGSILICA 38
MAGSILICA 300 (37; 6.0 g) was dried at 1208C for 6 h under high
vacuum. Maleic anhydride (34; 1.72 g, 18 mmol, 3.0 mmolgꢀ1),
120 mL dry toluene and 3-aminopropyltriethoxy silane (35; 2.8 mL,
12 mmol, 2.0 mmolgꢀ1) were added consecutively and the suspen-
sion was gently stirred for 30 min at room temperature under an
argon atmosphere. To this suspension ZnCl2 (2.46 g, 18 mmol,
3.0 mmolgꢀ1) was added and gentle stirring was continued for
30 min at room temperature. Then, hexamethyldisilazane (5.0 mL,
24 mmol, 4.0 mmolgꢀ1) was added and the reaction mixture was
heated under refluxing conditions for 2 h. The reaction vessel was
Keywords: antitumor agents · hyperthermia · maytansinoids ·
mutasynthesis · nanoparticles
[1] A. Kirschning, F. Hahn, Angew. Chem. 2012, 124, 4086–4096; Angew.
Chem. Int. Ed. 2012, 51, 4012–4022.
[2] a) S. Weist, R. D. Sꢁssmuth, Appl. Microbiol. Biotechnol. 2005, 68, 141–
150; b) A. Kirschning, F. Taft, T. Knobloch, Org. Biomol. Chem. 2007, 5,
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
9
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
3245–3295; c) J. Kennedy, Nat. Prod. Rep. 2008, 25, 25–34; d) K. Weiss-
man, Trends Biotechnol. 2007, 25, 139–142.
[20] a) A. Kawai, H. Akimoto, Y. Kozai, K. Ootsu, S. Tanida, N. Hashimoto, H.
Nomura, Chem. Pharm. Bull. 1984, 32, 3441–3451; b) W. C. Widdison,
[3] R. Thiericke, J. Rohr, Nat. Prod. Rep. 1993, 10, 265–289.
[4] a) F. Taft, M. Brꢁnjes, H. G. Floss, N. Czempinski, S. Grond, F. Sasse, A.
Kirschning, ChemBioChem 2008, 9, 1057–1060; b) A. Deb Roy, S. Grꢁ-
schow, N. Cairns, R. J. M. Goss, J. Am. Chem. Soc. 2010, 134, 1224–
12245; c) K. Harmrolfs, L. Mancuso, B. Drung, F. Sasse, A. Kirschning, Beil-
stein J. Org. Chem. 2014, 10, 535–543.
R. V. J.
No. US20070112188.
[21] W. C. Widdison, S. D. Wilhelm, E. E. Cavanagh, K. R. Whiteman, B. A.
Leece, Y. Kovtun, V. S. Goldmacher, H. Xie, R. M. Stevens, R. J. Lutz, R.
Zhao, L. Wang, W. A. Blꢃttler, R. V. J. Chari, J. Med. Chem. 2006, 49,
4392–3408.
Chari
(Immunogen
Inc.,
Cambridge,
MA),
Patent
[5] F. Taft, K. Harmrolfs, I. Nickeleit, A. Heutling, M. Kiene, N. Malek, F. Sasse,
A. Kirschning, Chem. Eur. J. 2012, 18, 880–886.
[22] a) D. Ortega, Q. A. Pankhurst, Nanoscience 2013, 1, 60–88; b) J. van der
Zee, Ann. Oncol. 2002, 13, 1173–1184; c) M. H. Falk, R. D. Issels, Int. J.
Hyperthermia 2001, 17, 1–18.
[23] D. Yoo, H. Jeong, S.-H. Noh, J.-H. Lee, J. Cheon, Angew. Chem. 2013, 125,
13285–13289; Angew. Chem. Int. Ed. 2013, 52, 13047–13051.
[24] a) J. van Overbeek, M. E. Conklin, A. F. Blakeslee, Science 1941, 94, 350–
351; b) J. van Overbeek, Science 1966, 152, 721–731.
[25] a) C. Thorpe, Trends Biochem. Sci. 1989, 14, 148–151. 4-Pentynoic acid
as inhibitor of b-oxidation: b) Y. Yoshizawa, Z. Li, P. B. Reese, J. C. Vede-
ras, J. Am. Chem. Soc. 1990, 112, 3212–3213; c) D. E. Cane, G. Luo, C.
Khosla, C. M. Kao, L. Katz, J. Antibiot. 1995, 48, 647–651.
[26] 2-Methyl-pent-4-ynoic acid: C. J. Dutton, S. P. Gibson, A. C. Goudie, K. S.
Holdom, M. S. Pacey, J. C. Ruddock, J. D. Bu’Lock, M. K. Richards, J. Anti-
biot. 1991, 44, 357–365.
[6] a) B. Kusebauch, N. Brendel, H. Kirchner, H.-M. Dahse, C. Hertweck,
ChemBioChem 2011, 12, 2284–2288; b) S. Weist, C. Kittel, D. Bischoff, B.
Bister, V. Pfeifer, G. J. Nicholson, W. Wohlleben, R. D. Sꢁssmuth, J. Am.
Chem. Soc. 2004, 126, 5942–5943; c) R. D. Sꢁssmuth, W. Wohlleben,
Appl. Microbiol. Biotechnol. 2004, 63, 344–350; d) S. Weist, B. Bister, O.
Puk, D. Bischoff, S. Pelzer, G. J. Nicholson, W. Wohlleben, G. Jung, R. D.
Sꢁssmuth, Angew. Chem. 2002, 114, 3531–3534; Angew. Chem. Int. Ed.
2002, 41, 3383–3385.
[7] The term maytansinoids relates to the macrolactam antibiotic Maytan-
sine (6) which was first isolated from the plant Maytenus serrata: S. M.
Kupchan, Y. Komoda, W. A. Court, G. J. Thomas, R. M. Smith, A. Karim,
C. J. Gilmore, R. C. Haltiwanger, R. F. Bryan, J. Am. Chem. Soc. 1972, 94,
1354–1356.
[27] a) Y. Matsumura, T. Shiozawa, H. Matsushita, Y. Terao, Biol. Pharm. Bull.
1995, 18, 1805–1807; b) M. A. A. Sarhan, J. Appl. Sci. Res. 2007, 3, 886–
889.
[8] a) J. M. Cassady, K. K. Chan, H. G. Floss, E. Leistner, Chem. Pharm. Bull.
2004, 52, 1–26; b) A. Kirschning, K. Harmrolfs, T. Knobloch, C. R. Chim.
2008, 11, 1523–1543.
[28] Details are given in the Supporting Information.
[9] a) T. Kubota, M. Brꢁnjes, T. Frenzel, J. Xu, A. Kirschning, H. G. Floss,
ChemBioChem 2006, 7, 1221–1225; b) A. Meyer, M. Brꢁnjes, F. Taft, T.
Frenzel, F. Sasse, A. Kirschning, Org. Lett. 2007, 9, 1489–1492; c) F. Taft,
M. Brꢁnjes, T. Knobloch, H. G. Floss, A. Kirschning, J. Am. Chem. Soc.
2009, 131, 3812–3813; d) K. Harmrolfs, M. Brꢁnjes, G. Drꢃger, H. G.
Floss, F. Sasse, F. Taft, A. Kirschning, ChemBioChem 2010, 11, 2517–
2520; e) T. Knobloch, H. G. Floss, K. Harmrolfs, F. Sasse, F. Taft, B. Tho-
maszewski, A. Kirschning, ChemBioChem 2011, 12, 540–547; f) T. Kno-
bloch, G. Drꢃger, W. Collisi, F. Sasse, A. Kirschning, Beilstein J. Org. Chem.
2012, 8, 861–869; g) L. Mancuso, G. Jꢁrjens, J. Hermane, K. Harmrolfs,
S. Eichner, J. Fohrer, W. Collisi, F. Sasse, A. Kirschning, Org. Lett. 2013, 15,
4442–4445.
[29] a) R. Huisgen, in 1,3-Dipolar Cycloaddition Chemistry, Wiley, New York,
1984, 1–176; b) V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless,
Angew. Chem. 2002, 114, 2708–2711; Angew. Chem. Int. Ed. 2002, 41,
2596–2599; c) C. W. Tornøe, C. Christensen, M. Meldal, J. Org. Chem.
2002, 67, 3057–3062; d) H. C. Kolb, M. G. Finn, K. B. Sharpless, Angew.
Chem. 2001, 113, 2056–2075; Angew. Chem. Int. Ed. 2001, 40, 2004–
2021.
[30] a) J.-F. Lutz, Z. Zarafshani, Adv. Drug Delivery Rev. 2008, 60, 958–970;
b) E. M. Sletten, C. R. Bertozzi, Angew. Chem. 2009, 121, 7108–7133;
Angew. Chem. Int. Ed. 2009, 48, 6974–6998.
[31] S.-K. Choi, Synthetic Multivalent Molecules: Concepts and Biomedical Ap-
plications, Wiley, Hoboken, 2004.
[10] C. T. Walsh, S. W. Haynes, B. D. Ames, Nat. Prod. Rep. 2012, 29, 37–59.
[11] S. Eichner, T. Knobloch, H. G. Floss, G. J. Fohrer, K. Harmrolfs, J. Hermane,
A. Schulz, F. Sasse, P. Spiteller, F. Taft, A. Kirschning, Angew. Chem. 2012,
124, 776–781; Angew. Chem. Int. Ed. 2012, 51, 752–757.
[12] a) P. Workman, Curr. Cancer Drug Targets 2003, 3, 297–300; b) L. Neck-
ers, K. Neckers, Expert Opin. Emerging Drugs 2005, 10, 137–149; c) L.
Whitesell, S. L. Lindquist, Nat. Rev. Cancer 2005, 5, 761–772; d) C. Pro-
dromou, S. M. Roe, R. O’Brien, J. E. Ladbury, P. W. Piper, L. H. Pearl, Cell
1997, 90, 65–75; e) K. Hatano, E. Higashide, S. Akiyama, M. Yoneda,
Agric. Biol. Chem. 1984, 48, 1721–1729; f) S. Bandi, Y. J. Kim, S. O. Sa, Y.-
K. Chang, J. Microbiol. Biotechnol. 2005, 15, 930–937.
[32] S. Kumar, D. P. Arya, J. Am. Chem. Soc. 2011, 133, 7361–7375.
[33] This concept was shown to work well, if the biological target operates
as a homodimer, as was recently shown for the inhibition of heat shock
protein Hsp90: H. Wahyudi, Y. Wang, S. R. McAlpine, Org. Biomol. Chem.
2014, 12, 765–773.
[34] Reviews on superparamagnetic nanoparticles: a) A. Kirschning, L. Ku-
pracz, J. Hartwig, Chem. Lett. 2012, 41, 562–570; b) A.-H. Lu, E. L. Sala-
bas, F. Schꢁth, Angew. Chem. 2007, 119, 1242–1266; Angew. Chem. Int.
Ed. 2007, 46, 1222–1244; c) Y.-W. Jun, J.-S. Choi, J. Cheon, Chem.
Commun. 2007, 1203–1214; d) X. K. Zhang, Y. F. Li, J. Q. Xiao, E. D.
Wetzel, J. Appl. Phys. 2003, 93, 7124–7126.
[13] M. A. Biamonte, R. van de Water, J. W. Arndt, R. H. Scannevin, R. H.
Perret, D. Perret, W.-C. Lee, J. Med. Chem. 2010, 53, 3–17.
[14] FDA approval see: http://www.accessdata.fda.gov/drugsatfda_docs/
label/2013/125427lbl.pdf.
[15] P. Spiteller, L. Bai, G. Shang, B. J. Carroll, T.-W. Yu, H. G. Floss, J. Am.
Chem. Soc. 2003, 125, 14236–14237.
[16] a) C. G. Kim, T. W. Yu, C. B. Fryhle, S. Handa, H. G. Floss, J. Biol. Chem.
1998, 273, 6030–6040; b) K. Arakawa, R. Mꢁller, T. Mahmud, T. W. Yu,
H. G. Floss, J. Am. Chem. Soc. 2002, 124, 10644–10645; c) T.-W. Yu, R.
Mꢁller, M. Mꢁller, X. Zhang, G. Drꢃger, C.-G. Kim, E. Leistner, H. G. Floss,
J. Biol. Chem. 2001, 276, 12546–12555.
[17] T.-W. Yu, L. Bai, D. Clade, D. Hoffmann, S. Toelzer, K. Q. Trinh, J. Xu, S. J.
Moss, E. Leistner, H. G. Floss, Proc. Natl. Acad. Sci. USA 2002, 99, 7968–
7973.
[35] a) A. M. Wu, P. D. Senter, Nat. Biotechnol. 2005, 23, 1137–1146; b) R. V.
Chari, Acc. Chem. Res. 2008, 41, 98–107; c) S. C. Alley, N. M. Okeley, P. D.
Senter, Curr. Opin. Chem. Biol. 2010, 14, 529–537.
[36] P. S. Low, W. A. Henne, D. D. Doorneweerd, Acc. Chem. Res. 2008, 41,
120–129.
[37] a) W. Arap, R. Pasqualini, E. Ruoslahti, Science 1998, 279, 377–380;
b) M. L. Janssen, W. J. Oyen, I. Dijkgraaf, L. F. Massuger, C. Frielink, D. S.
Edwards, M. Rajopadhye, H. Boonstra, F. H. Corstens, O. C. Boerman,
Cancer Res. 2002, 62, 6146–6151.
[38] Review: L. F. Tietze, K. Schmuck, Curr. Pharm. Des. 2011, 17, 3527–3547.
[39] L. F. Tietze, J. M. von Hof, M. Mꢁller, B. Krewer, I. Schuberth, Angew.
Chem. 2010, 122, 7494–7497.
[40] L. F. Tietze, R. Hannemann, W. Buhr, M. Lçgers, P. Menningen, M. Lieb,
D. Starck, T. Grote, A. Dçring, I. Schuberth, Angew. Chem. 1996, 108,
2840–2842; Angew. Chem. Int. Ed. 1996, 35, 2674–2677.
[41] Target specific drug conjugates were recently reviewed by A. Mullard,
Nat. Rev. Drug Discovery 2013, 12, 329–333.
[18] S. C. Wenzel, R. M. Williamson, C. Grꢁnanger, J. Xu, K. Gerth, R. A. Marti-
nez, S. J. Moss, B. J. Carroll, S. Grond, C. J. Unkefer, R. Mꢁller, H. G. Floss,
J. Am. Chem. Soc. 2006, 128, 14325–14336.
[19] a) E. W. Hafner, B. W. Holley, K. S. Holdom, S. E. Lee, R. G. Wax, D. Beck,
H. A. I. McArthur, W. C. Wernau, J. Antibiot. 1991, 44, 349–356; b) H. B.
Bode, P. Meiser, T. Klefisch, N. S. d. J. Cortina, D. Krug, A. Gçhring, G.
Schwꢃr, T. Mahmud, Y. A. Elnakady, R. Mꢁller, ChemBioChem 2007, 8,
2139–2144.
[42] Ansamitocin and its derivatives are stable in refluxing toluene for 22 h;
see ref. [9].
[43] S. D. Bergman, F. Wudl, J. Mat. Chem. 2008, 18, 41–62.
[44] T. T. N’Guyen, H. T. T. Duong, J. Basuki, V. Montembault, S. Pascual, C.
Guibert, J. Fresnai, C. Boyer, M. R. Whittaker, T. P. Davis, L. Fontaine,
&
&
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
10
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Angew. Chem. 2013, 125, 14402–14406; Angew. Chem. Int. Ed. 2013, 52,
14152–14156.
[45] H. I. Lim, P. M. Oliver, J. Marzillier, D. V. Vezenov, Anal. Bioanal. Chem.
2010, 397, 1861–1872.
[48] Use of SAG 471 anti-foam for fermentation of A. pretiosum was suggest-
ed in Patent No. US6790954 (J. Chung, G. S. Byng; Immunogen Inc.,
Cambridge, MA; Genentech Inc., South San Francisco, CA).
[46] Ringer’s solution is an aqueous solution of several salts. It resembles an
isotonic solution.
Received: July 21, 2014
[47] A. Gagnoud, IEEE Trans. Magn. 2004, 40, 29–36.
Published online on && &&, 0000
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
11
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FULL PAPER
A dual strategy to fight cancer: New
ansamitocin derivatives were prepared
by a combination of precursor-directed
biosynthesis and semisynthesis, leading
to conjugates with a thermolabile linker
attached to superparamagnetic core-
shell nanostructured particles. These
nontoxic conjugates can be heated with
release of the toxin in an external
&
Medicinal Chemistry
L. Mancuso, T. Knobloch, J. Buchholz,
J. Hartwig, L. Mçller, K. Seidel, W. Collisi,
F. Sasse, A. Kirschning*
&& – &&
Preparation of Thermocleavable
Conjugates Based on Ansamitocin and
Superparamagnetic Nanostructured
Particles by a Chemobiosynthetic
Approach
oscillating electromagnetic field. The
conjugates can act simultaneously in
hyperthermia and in chemotherapy.
&
&
Chem. Eur. J. 2014, 20, 1 – 12
www.chemeurj.org
12
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!