Design, Synthesis, and Biochemical Evaluation of Alpha-Amanitin Derivatives Containing Analogs of the trans-Hydroxyproline Residue for Potential Use in Antibody-Drug Conjugates

Article abstract of DOI:10.1002/chem.202101373

Alpha-amanitin, an extremely toxic bicyclic octapeptide extracted from the death-cap mushroom, Amanita phalloides, is a highly selective allosteric inhibitor of RNA polymerase II. Following on growing interest in using this toxin as a payload in antibody-drug conjugates, herein we report the synthesis and biochemical evaluation of several new derivatives of this toxin to probe the role of the trans-hydroxyproline (Hyp), which is known to be critical for toxicity. This structure activity relationship (SAR) study represents the first of its kind to use various Hyp-analogs to alter the conformational and H-bonding properties of Hyp in amanitin.

Full text of DOI:10.1002/chem.202101373

Accepted Article
Accepted Article
Title: Design, synthesis, and biochemical evaluation of alpha-amanitin
derivatives containing analogs of the trans-hydroxyproline
residue for potential use in antibody-drug conjugates
Authors: Kaveh Matinkhoo, Antonio AWL Wong, Chido M. Hambira,
Brandon Kato, Charlie Wei, Torsten Hechler, Christoph
Muller, Alexandra Braun, Francesca Gallo, Andreas Pahl, and
David Perrin
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202101373
Link to VoR: https://doi.org/10.1002/chem.202101373
01/2020

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
Design, Synthesis, and Biochemical Evaluation of Alpha-amanitin
Derivatives Containing Analogs of the Trans-hydroxyproline
Residue for Potential Use in Antibody-drug Conjugates
Kaveh Matinkhooa, Antonio A. W. L. Wonga, Chido M. Hambiraa, Brandon Katoa, Charlie Weia,
Christoph Müllerb, Torsten Hechlerb, Alexandra Braunb, Francesca Gallob, Andreas Pahl*b, David M.
Perrin*a
[a]
Dr. Kaveh Matinkhoo, Mr. Antonio A. W. L. Wong, Dr. Chido M. Hambira, Mr. Brandon Kato, Mr. Charlie Wei, Prof. David M. Perrin*
Chemistry Department
University of British Columbia
2036 Main Mall
Vancouver, British Columbia
V6T-1Z1, CANADA
Email: dperrin@chem.ubc.ca
@perrinlab, #amanitin, @UBCCHEM
[b]
Dr. Christoph Müller, Dr. Torsten Hechler, Dr. Alexandra Braun, Dr. Francesca Gallo, Dr. Andreas Pahl*
Heidelberg Pharma
Gregor-Mendel-Straße 22
68526 Ladenburg
Germany
E-mail: andreas.pahl@hdpharma.com
Supporting information for this article is given via a link at the end of the document.
Abstract: Alpha-amanitin, an extremely toxic bicyclic octapeptide
extracted from the death-cap mushroom, Amanita phalloides, is a
highly selective allosteric inhibitor of RNA polymerase II. Following on
growing interest in using this toxin as a payload in antibody-drug
conjugates, herein we report the synthesis and biochemical
evaluation of several new derivatives of this toxin to probe the role of
the trans-hydroxyproline (Hyp), which is known to be critical for toxicity.
This structure activity relationship (SAR) study represents the first of
its kind to use various Hyp-analogs to alter the conformational and H-
bonding properties of Hyp in amanitin.
HDP-101, an amanitin-based ADC for the treatment of multiple
myeloma and the first amanitin-based ADC that is advancing
towards clinical trials.[8] Others have explored targeted amanitin
conjugates underscoring interest in amanitin and its chemistry.[9]
With this interest in using α-amanitin for therapeutic applications,
new knowledge as to the molecular basis of toxicity will be
essential for designing new toxicophores for therapeutic
applications. XRD studies on α-amanitin-Pol II co-crystal obtained
by Kornberg and coworkers have suggested multiple interactions
between this toxin and the bridge helix of Pol II from S. cerevisiae
(Figure 1, A)[10] including numerous interactions with backbone
amides, π-stacking/π-cation interactions with the hydroxy-
tryptathionine, and putative H-bonds to the hydroxyl groups of
Hyp and DHIle. These results were corroborated by cryo-EM
structures obtained by Cramer and coworkers (Figure 1, B).[11]
Introduction
Alpha-amanitin, isolated over 70 years ago, is the principal toxin
in Amanita phalloides - the notorious “death-cap” mushroom. With
a rich scientific history,[1] it is one of the deadliest toxins found in
nature (LD50 = 50-100 µg/kg in humans). Its unique bicyclic
structure comprises
a
6'-hydroxytryptathionine-(R)-sulfoxide
staple along with two oxidized amino acids that are considered
critical for cytotoxicity: trans-4-hydroxyproline (Hyp) and
(2S,3R,4R)-4,5-dihydroxyisoleucine (DHIle).[2] Featured in most
modern biochemistry textbooks, α-amanitin is a highly selective
allosteric inhibitor of eukaryotic RNA polymerase II (Pol II) with a
Ki value of 1-10 nM,[3] the inhibition of which leads to rapid
proteolytic degradation of Pol II and cell death.[4]
Because Pol II-catalyzed transcription is essential for cellular
function, α-amanitin kills dividing and quiescent cells by a unique
mechanism of action. It is precisely this generalized cytotoxicity
that makes α-amanitin an attractive payload for antibody drug
conjugates (ADCs).[5] Following an early report on an amanitin-
antibody conjugate against Thy 1.2 antigen towards T lymphoma
S49.1 cells,[6] Moldenhauer et al. demonstrated the extraordinary
promise of α-amanitin as a payload for ADC development; an anti-
EpCAM antibody-amanitin conjugate remarkably cured 60% (3 of
5) mice with pancreatic tumor xenografts,[7] paving the way for
Figure 1. (A) Co-crystal structure of α-amanitin bound to yeast (S. cerevisiae)
Pol II at 2.8 Å resolution (reproduced from Bushnell et al. ref. 10 with permission
from Proc. Natl. Acad. Sci.). The H-bond between Glu822 of Pol II and the
hydroxyl group of Hyp is shown. (B) Observed interactions between different
residues of α-amanitin 1 and the backbone of Pol II (Brueckner et al.).[11]
Hydrogen bonds are shown in dashed green lines and other interactions are
marked with dashed red lines. Note the H-bonding interactions between the
hydroxyl group of Hyp and Glu822 and His1085 of the Rpb1 subunit of Pol II.
1
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
Recent cryo-EM structures on porcine (S. scrofa) RNA Pol II
show that Glu845 plays the analogous role of Glu 822. Yet an H-
bond acceptor, analogous to His1085, is not evident this addition
interaction is not clearly supported in the known structures of
mammalian RNA Pol II.[12]
the increased interest in α-amanitin, here we begin to tackle SAR
profiling of the Hyp with choice analogs introduced into new toxin
analogs based on structural and conformational considerations.
These fascinating structural studies cohere in part with earlier
structure activity relationships (SARs) reported in the 1980s,
based on naturally-sourced amatoxins. For example,
deoxygenation of the (R)-sulfoxide to the thioether or oxidation to
the sulfone resulted in no loss of activity in the case of the 6'-O-
methyl ether of α-amanitin.[13] Naturally occurring amanullin and
proamanullin, which lack the hydroxyl groups on DHIle or on both
Results and Discussion
We based our analysis on the synthetically more accessible
“dideoxy-amanitin” (6'-deoxy-S-deoxo-α-amanitin) (2) lacking
both the 6'-OH and the (R)-sulfoxide, where the thioether is known
to be equipotent in cytotoxicity assays,[13, 19] confirmed by us and
17, 20]
others.[9d,
Typically, toxins constructed with a simple
tryptathionine staple show near-native potencies.[2a, 14a, 21] Hence,
we prepared the dideoxy-amanitin scaffold so as to make this
comparison internally consistent across all proline analogs
studied here.
DHIle and Hyp respectively, show Ki values of 10-40 nM (near-
[2a, 14]
native) and 5-20 µM (greatly attenuated) respectively.
The
synthetically accessible “tetradeoxy-amanitin” (lacking the
sulfoxide, the 6'-hydroxy group on the tryptathionine, and where
Ile replaces DHIle) was reported to have a Ki value of 80 nM in an
in vitro transcription assay with calf-thymus polymerase,[14a] but a
much elevated Ki of 1 µM in a different report on drosophila RNA
Pol II [2b] (despite considerable interspecies homology cf. B. taurus,
S. scrofa, C. capitata). Yet these values had been obtained using
transcription assays that often differ in terms of the species of Pol
II used in the assay as well as how the assays were conducted.
Nevertheless, taken together, these studies, along with
In choosing 4-substituted prolines, we considered studies on
collagen that employed conformationally biased (4R)-substituted
prolines that have questioned the contribution of the H-bonding
by the Hyp to embedded water molecules or to carbonyls on
opposing collagen strands.[22] Indeed, (4R)-substituted prolines
that are incapable of H-bonding, e.g. 4-fluoro-, 4-chloro-, 4-azido-
proline form highly stable collagen helices[23] as reviewed,[24]
suggesting that ring-pucker contributions are more important than
H-bonding for determining the structural integrity of collagen. This
bias challenges analog development in amanitin because of the
need to also address defined H-bonds to Glu845 with a suitable
functional group, while also ensuring a Cγ-exo conformation that
is seen in the crystal structures of α- and β-amanitin.[14a, 15c, 19]
To maintain the conformational bias of Hyp while affording H-
bonding functionalities, we looked at comprehensive work by
Zondlo et al.[25] and synthesized five trans-4-substituted prolines
for insertion into amanitin. Hence, we considered: (i) potential for
H-bond donation or acceptance and (ii) the ring pucker induced
by the presence of a trans-4-substitutent on the proline residue.
Given limitations in throughput, the cyano-, amino-, mercapto-,
and guanidino- analogs appeared compelling due to their
potential to form at least one H-bond with Glu845. It was further
anticipated that the guanidine would form bifurcated H-bonds or
alternatively a salt-link with Glu845.[26] The use of the amino-
proline was rationalized similarly in terms of favorable
electrostatic interactions and/or H-bonding interactions.
Additionally, we investigated two protected intermediates, Acm-
protected trans-4-mercapto-proline and Boc-protected trans-4-
guanidino-proline along with a trans-4-methylcarbamoly-proline,
that adventitiously formed upon azide reduction (vide infra), all of
which provide H-bonding functionalities (Table 1).
crystal/cryo-EM studies suggest that the Hyp is
contributor to toxicity.
a major
In Lipscomb’s structure,[15] the trans-(4R)-hydroxyproline exists
in the trans-amide bond and adopts the classic Cγ-exo ring pucker
conformation commonly seen for Hyp in collagen. The Cγ-exo ring
pucker is favored by trans-oriented electron-withdrawing groups
(EWGs) at the 4-position, that intensify the gauche interaction
between the amide bond and the EWG (Figure 2).[16]
Figure 2. (A) Exo and endo ring puckers observed in the proline residue of a
trans-4-substituted proline-containing peptide; (B) hyperconjugation of the σ(C-
Hδ) orbital and the electron-deficient σ*(C-R) where (R = EWG).
As such, the hydroxyl group may contribute to amanitin toxicity,
not only via H-bonds but via a Cγ-exo conformation that may
reinforce subtle conformational differences throughout the entire
macrocycle. Nevertheless, the precise role of this hydroxyl group,
either in terms of H-bonding or ring-puckering remains entirely
unaddressed in the vast literature on amanitin derivatives.
Furthermore, a systematic SAR investigation based on rationally
designed Hyp analogs represents a worthy challenge that has yet
to be undertaken, likely owing to synthetic inaccessibility of DHIle.
In 2018, we completed the first total synthesis of α-amanitin
providing an enantioselective synthesis of (2S,3R,4R)-4,5-
dihydroxyisoleucine on moderate scale.[17] In 2020, two additional
total syntheses highlighted the need for synthetic access.[18] With
2
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
Table 1. Proposed analogs of Hyp for incorporation into amanitin. Compound 1
represents the natural product, α-amanitin, while 2 shows dideoxy-α-amanitin,
which is the basis for our synthetic analogs. Compound 32 is an adventitious
by-product that was not initially proposed as an analog.
Figure 3. Synthesis of Boc-protected methyl esters of azido-, cyano- and
mercapto-proline analogs. Reagents and conditions: a) PPh3 (1.2 eq.), DIAD
(1.2 eq.), THF, 0°C to RT, 20 h, 85%. b) NaN3 (2.0 eq.), MeOH, 40°C, 16 h,
75%. c) MsCl (1.6 eq.), Et3N (1.4 eq.), DCM, 0°C, 16 h, 95%. d) NaN3 (2.0 eq.),
DMSO, 80°C, 4 h, 79%. e) KCN (1.5 eq.), DMSO, 80°C, 4 h, 30%. f) KSAc (1.3
eq.), DMF, 70°C, 4 h, 52%.
Cmpd.
X
OH
OH
N3
CN
SH
NH2
Gdn
Y
Z
(R)-SO
1
2
3
4
5
6
7
8
9
32
OH
H
H
H
H
H
H
H
H
H
S
S
S
S
S
S
S
S
S
For the synthesis of azido- and cyano-proline analogs, their
corresponding methyl esters were saponified to the free acid, and
the Boc protecting group was swapped with Fmoc to yield SPPS
compatible monomers 21 and 22 (Figure 4, A). In the case of
mercapto-proline, the thioacetate group and the methyl ester of
16 were concomitantly saponified, the Boc protecting group was
removed, then an Acm (acetamidomethyl) protecting group was
introduced on the free thiol. Finally, the free amine was protected
with Fmoc to yield the fully protected monomer 26 (Figure 4, B).
SAcm
Boc-Gdn
MeOCONH
In terms of ring-puckering, whereas the mercapto-proline
prefers a Cγ-endo conformation in approximately 5:1 ratio over the
Cγ-exo,
[27]
both the amino- and guanidino-prolines (extant at
physiological pH as ammonium and guanidinium cations) are
known to strongly favor a Cγ-exo conformation.[28] Trans-4-amido
prolines as well as carbamoylated ones (N-Boc-modified) are also
known to adopt a Cγ-exo conformation,[29] leading us to rationalize
testing a methylcarbamate 32 that formed serendipitously upon
azide reduction (vide infra). The cyano-proline, which is strongly
electron withdrawing, is expected to favor the Cγ-exo
conformation, and has the potential to accept an H-bond, albeit
weakly.[30] Finally, the azido-proline, also known to favor the Cγ-
exo conformation,[31] provides a control for conformation in the
absence of H-bonding. The logical application of 4-fluoroproline
was not carried forward in this analysis as it did not provide a
cytotoxic analog of dideoxy-amanitin, and in our hands proved
difficult to synthesize (data not shown).
The synthesis of the Hyp analogs began with preparation of cis-
Nα-Boc-4-hydroxyproline methyl ester (cis-Boc-Hyp-OMe, 12)
from the commercially available trans-Nα-Boc-4-hydroxyproline
(Boc-Hyp, 10); Boc-Hyp was subjected to an intramolecular
Mitsunobu reaction using PPh3 and DIAD to afford the lactone of
cis-Hyp 11.[25] An azide-assisted saponification of this lactone with
methanol afforded the methyl ester of cis-Boc-Hyp 12.[32]
Subsequently, the hydroxyl group was mesylated and the
resulting compound was subjected to SN2 conditions to yield the
desired trans-isomer of various analogs (Figure 3).
Figure 4. Synthesis of fully protected, SPPS compatible monomers of: (A)
azido-, cyano-proline and (B) mercapto-proline. Reagents and conditions: a)
LiOH (20 eq.), THF/H2O (1:1), 0°C, 16 h. b) TFA/DCM (1:2), RT, 30 min. c)
Fmoc-OSu (1.1 eq.), NaHCO3 (2.0 eq.), 1,4-dioxane/H2O (2.3:1), RT, 3 h. d) (N-
hydroxymethyl)acetamide (1.2 eq.), 12 M aq. HCl (1.0 eq.), H2O, 0°C to RT, 48
h, 25% over 4 steps.
Following a solid phase strategy similar to our reported total
synthesis, we incorporated the three aforementioned monomers
into the corresponding dideoxy-amanitins (Figure 5). To
summarize, Hyp along with the synthetic Hyp analogs were
separately loaded on 2-chlorotrityl chloride (CTC) resin, followed
by the coupling of Fmoc-Asn(Trt)-OH, Fmoc-Cys(Trt)-OH, Fmoc-
Gly-OH, Fmoc-Ile-OH, Fmoc-Gly-OH and Boc-Fpi-OH (30),
where Fpi = 3a-fluoro-hexahydropyrrolo-[2,3-b]indoline,[33] to
afford the linear heptapeptide. Treating the resin with TFA/DCM
(1:1) resulted in the global deprotection of the acid-labile
protecting groups and tryptathionine formation via the Savige-
3
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
Fontana reaction[34] to yield monocyclic heptapeptides of various
amanitin analogs (27c, 28c, 29c). Next, the DHIle 31 was grafted
onto the N-terminus, followed by in situ removal of Fmoc and TBS
protecting groups. The resulting monocyclic octapeptides (27d,
28d, 29d) were macrolactamized to yield the final bicyclic
octapeptides containing different Hyp analogs (3, 4 and 8).
Figure 5. Synthesis of amanitin derivatives containing three Hyp analogs:
azido-, cyano- and SAcm-prolines. A solid-phase peptide synthesis involving an
Fmoc strategy was utilized (see Supporting Information for reagents and
reaction conditions).
Figure 6. (A) Removal of Acm to afford the mercapto-proline analog 5. (B)
Reduction of the azide under Staudinger conditions (PPh3, DMSO) to yield the
speculated methyl carbamate analog 32, and using DTT to obtain the amino-
proline analog 6. (C) Guanidinylation of amino-proline-amanitin 6 to afford the
Boc-protected (9) and unprotected guanidino-proline 7 amanitins. Reagents and
conditions: a) PdCl2 (20 eq.), 6 M aq. Gdn.HCl, 37°C, 30 min. b) PPh3 (12 eq.),
DMSO, RT, 16 h. c) DTT (5.0 eq.), DMSO, RT, 3 h. d) i. N-Boc-N'-TFA-pyrazole-
1-carboxamidine (6.0 eq.), DIPEA (2.5 eq.), THF/DMSO (2:1), RT, 16 h. ii.
K2CO3 (13 eq.), MeOH/H2O (2.5:1), RT, 3 h. e) TFA/DCM (1:2), RT, 1 h.
To synthesize analog 5 (mercapto-proline), the Acm-protected
intermediate 8 was treated with a large excess of PdCl2 in 6M aq.
guanidinium hydrochloride (Gdn·HCl) followed by quenching with
DTT to remove the Pd (Figure 6, A).[35] To prepare the amino-
proline analog from the azido-prolyl amanitin, Staudinger
reduction conditions were initially employed using PPh3 in
aqueous DMSO.[36] To our surprise, instead of obtaining the
To prepare the guanidino analog 7, amanitin 6 was reacted with
N-Boc-N'-TFA-pyrazole-1-carboxamidine in THF/DMSO in the
presence of DIPEA.[39] The resulting doubly protected guanidine
was treated with K2CO3 in MeOH and H2O to remove the TFA
protecting group, affording intermediate 9. Finally, the Boc
protecting group was removed using TFA in DCM (1:1) to yield an
amanitin analog 7, containing a guanidino-proline residue (Figure
6, C). Lastly, to prepare the dideoxy-amanitin (2), we simply
followed our previous reports on the synthesis of amanitin starting
with Hyp(OtBu) (see Supporting Information).
Eight analogs were purified and compared to both α-amanitin
(1) and the more chemically similar dideoxy-amanitin (2). To
compare the overall conformation of these amanitins, circular
dichroism (CD) spectra of all the analogs were obtained in two
solvents: 1) MeOH, which is the standard solvent used to evaluate
amatoxins,[2] and 2) MeOH/0.1% aqueous formic acid (10:1) at pH
4.5. Although 2 showed a slightly different spectrum from the
others (in MeOH), all analogs exhibited similar CD spectra
(Figure 7, A), with a negative Cotton effect generally observed
between 215 and 250 nm, and a positive Cotton effect above 250
nm.
desired amino analog, we obtained
a byproduct that we
characterized to be the methyl carbamate of the amino-proline
residue (32), based on the mass spectrometry and preliminary 1H-
NMR studies showing a characteristic singlet (see Supporting
Information). There is literature precedent for the reaction of
organic azides with triphenylphosphine and CO2 to give
isocyanates that may in turn react with nucleophilic solvents (i.e.
MeOH in this case, see Supporting Information for
mechanism).[37] Without prejudice, we incorporated this derivative
into our panel of analogs for biochemical characterization.
Attempting alternative conditions for the reduction, the azide
group on 3 was successfully reduced to an amine 6 using DTT in
DMSO (Figure 6, B).[38]
Curiously, in the acidic environment of MeOH/H2O, pH 4.5, four
analogs (N3-, CN-, SH- and Acm-S-proline amanitins: 3, 4, 5, and
8 respectively) demonstrated similar CD spectra, while the others
(amino-, Gdn-, Boc-Gdn-, and methylcarbamoyl-prolines: 6, 7, 9,
and 32 respectively) showed highly disparate spectral signatures
(Figure 7, B). Interestingly, unlike the other analogs, 2 seemed to
vary minimally upon this solvent change.
4
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
conclusive IC50 values could be obtained for these analogs
(Figures ESI.7 and ESI.8 in Supporting Information).
Figure 8. MTT colorimetric assay for evaluation of the toxicity of the synthetic
analogs against CHO cells. (A) azido, Acm-S- and cyano-proline analogs, (B)
mercapto, MeOCONH, amino, Boc-Gdn and Gdn-proline analogs; (C) controls
– dideoxy-amanitin and α-amanitin.
Surprised by these findings, we questioned whether poor
internalization of certain derivatives, in particular 6 (amino-
proline) and 7 (guanidino-proline), resulted in an apparent lack of
toxicity. To address this hypothesis, all synthetic analogs (with the
exception of the putative MeOCONH-proline analog 32 due to
limited amounts) were then subjected to an in vitro transcription
assay using HeLaScribe® nuclear extract containing Pol II and all
other transcription factors using a DNA template containing a
strong cytomegalovirus (CMV) promoter (see Supporting
Information for detailed procedures).[41]
Figure 7. Circular dichroism (CD) spectra of α-amanitin and the synthetic
amanitins containing analogs of Hyp. (A) CD in MeOH. (B) CD in MeOH/(0.1%
FA in H2O), 10:1 at pH 4.5.
To test Pol II inhibition, increasing concentrations of α-amanitin
(1, 1 to 100 nM) were added to the transcription reaction.
Following autoradiography, an inhibition curve for α-amanitin was
obtained, with an IC50 value of 7.9 ± 0.9 nM that matched the
reported values in the literature.[42] The in vitro activities of all
compounds were measured at various concentrations (generally
ranging from 10 nM to 3 μM) and inhibition curves for most of the
tested analogs were obtained (Figure 9).
We then investigated cytotoxicity using CHO (Chinese hamster
ovary) cells as they are readily killed by α-amanitin (1, Ki ca. 0.5
µM) even though they do not overexpress the organic anion-
transporting protein (OATP) implicated in active toxin transport.[40]
To assess the cytotoxicity of these synthetic analogs, CHO cells
were treated with various concentrations ranging from 0.078 to 20
μM, and the percentages of viable cells were measured using an
MTT colorimetric assay (see Supporting Information). To our
disappointment, none of the synthetic analogs exhibited any
appreciable level of toxicity towards CHO cells, even at
concentrations as high as 20 μM (Figure 8, A and B). As controls,
IC50 values of 0.3 μM that were obtained for both α-amanitin and
dideoxy-amanitin in this assay were in line with the values
previously reported by us[9d, 17] and others[2a, 14] (Figure 8, C).
In a separate experiment, the unprotected analogs (3-7 and 32)
were tested against human embryonic kidney 293 (HEK293) cells,
with and without overexpression of the OATP1B3 protein, to
investigate the effect of this anion-transporting protein on the
cytotoxicity of the synthetic analogs. While some analogs showed
slightly elevated toxicity against the HEK293 cells, which
overexpress OATP1B3, this increase was marginal, and no
Figure 9. IC50 curves obtained for amanitin analogs CN, SH, NH2, Gdn-AMA,
dideoxy-α-AMA and α-amanitin in an in vitro transcription assay using the HeLa
nuclear extract (n=3).
5
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
A full IC50 curve was not acquired for the analogs that did not
exhibit any appreciable inhibitory effects up to 10 μM. The results
for the transcription assay of amanitin analogs containing various
proline residues are summarized in Table 2.
into HER2-targeting ADCs: CN-Pro (4) and NH2-Pro (6). In both
cases,
a
maleimide-based Val-Ala linker with
a
para-
aminobenzyl-based conjugation site was employed (Figure 10).
Two different linkers were adopted to address the versatility of
linking strategies; in the case of 6, we exploited the masking of
the amino proline with an immolative p-aminobenzylcarbamate
while in the case of 4, we applied a p-aminobenzyldioxolane
formed from the DHIle; upon enzymatic cleavage, the p-
aminobenzyldioxolane is known to immolate to give the diol as
previously reported.[43]
Table 2. Results of the in vitro transcription assay for α-amanitin and synthetic
analogs using the HeLa nuclear extract (n=3). Note that the R2 value for CN-
Pro-AMA 4 shows an extremely poor fit for the IC50 curve.
Cmpd.
X
OH
OH
N3
CN
SH
NH2
Gdn
Y
Z
IC50 (nM)
7.9 ± 0.90
74.2 ± 5.4
> 1000
480 ± 380
910 ± 230
550 ± 90
410 ± 80
> 1000
R2
(R)-SO
1
2
3
4
5
6
7
8
9
32
OH
H
H
H
H
H
H
H
H
H
0.991
0.980
-
-5.760
0.860
0.983
0.918
-
S
S
S
S
S
S
S
S
S
SAcm
Boc-Gdn
MeOCONH
> 600
N/A
-
N/A
Of all synthetic analogs, dideoxy-amanitin 2 was the most active
inhibitor with an IC50 value of 74 ± 5 nM; the loss of activity
compared to the natural product is most likely understood in terms
of the loss of the H-bonds formed with the 6'-OH and/or a slightly
weaker π-cation interaction owing to an indole ring with lower
electron density. An explanation for its near-native cytotoxicity
may yet be due to fact that RNA Pol II is highly expressed and
may attain concentrations as high at 500 nM, a concentration that
is higher than the Kd of 2.
Cpd 7 (guanidine) shows an IC50 value of 410 ± 80 nM, which is
approximately 50-fold higher than that of the natural product and
yet only 6-fold less active than 2. Other potentially active analogs
include amino-proline 6 (IC50 = 550 ± 90 nM) and mercapto-
proline 5 (IC50 = 910 ± 230 nM). Analogs that did not exhibit any
noticeable inhibitory effects up to 1 μM concentrations include N3-,
SAcm- and Boc-guanidino-Pro amanitins. All compounds were
investigated three times (n=3) except for 4 (n=4), which proved
puzzling in terms of its inhibitory effects on Pol II leading to an
unexplained but reproducible inability to fit an IC50 curve to the
data points making curve-fitting especially difficult and leading us
to question the IC50-value that we obtained for this analog (Figure
9). Based on the transcription results for the mercapto-proline
analog, IC50 value of 910 ± 230 nM (R2 = 0.860), we cannot
completely exclude the possibility of disulfide formation to give
dimers of 5. Nevertheless, we note that 5 was resynthesized and
immediately subjected to the transcription assay in buffer
supplemented with DTT, which likely would have reduced any
disulfide bonds that might have formed.
Figure 10. Chemical structure of synthetic ADCs of: (A) NH2-Pro-AMA analog
(34), and (B) CN-Pro-AMA analog (37). A maleimide Val-Ala-based linker has
been used in both cases (see Supporting Information for reagents and reaction
conditions).
In the case of 6, the amine group on NH2-Pro was conveniently
used to attach the linker as a carbamate (Figure 10, A). However,
to install said linker on CN-Pro-amanitin, the hydroxyl groups of
the DHIle residue were protected by cyclic acetal formation with
the para-aminobenaldehyde terminus (Figure 10, B). Both linker-
toxin intermediates were site-specifically conjugated to an anti-
HER2 antibody (T-D265C thiomab, generated by Heidelberg
Pharma Research GmbH) to furnish the corresponding ADCs
(see Supporting Information). The resulting ADCs were subjected
to in vitro cell-based assays against three HER2+ cell lines (SK-
BR-3, SK-OV-3 and JIMT-1) and a HER2- cell line (MDA-MB-231)
(Figure 11). Interestingly, while the antibody conjugates against
these two analogs did not show detectable cytotoxicity against
SK-OV-3 and JIMT-1 or MDA-MB-231, there was significant
toxicity against SK-BR-3. As the SK-BR-3 have the highest
expression of HER-2, they are likely to be more sensitive to
otherwise less toxic payloads.
Of note, this result demonstrates the potential for designing
amanitin analogs that may have specific toxicity against certain
cell lines and not others.
To investigate the possibility of using synthetic amanitin analogs
as ADC payloads, two analogs were selected for incorporation
6
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
While CD is used extensively to characterize amanitins and
many other peptides, we caution that CD is qualitative at best for
inferring meaningful structural differences that would inform
bioactivity; case in point: amanitin-sulfoxide analogs with very
disparate CD spectra are found to be very structurally similar.[43]
Indeed, of several amanitins that have been crystallized, these
show nearly superimposable structures despite exhibiting
significantly different CD spectra and equally disparate
cytotoxicity IC50 values.[15c]
We next considered the putative H-bond interactions with Pol II.
To date, cryo-EM and XRD structures have not directly observed
the toxin at sufficient resolution to unequivocally reveal specific H-
bonds and instead have been refined in accord with the ground-
state structure determined by Lipscomb and coworkers for β-
amanitin.[15a, b] As shown by Kornberg[10, 44] and Cramer[11-12, 45] in
several reports, the most notable interaction of α-amanitin with
Pol II is an H-bond between the hydroxyl group of Hyp and the
glutamic acid residue (A845 Glu) of Pol II (Figures 1 and 12).
Figure 12. (A) Observed H-bonding interactions (red lines) between the
hydroxyl group of Hyp in α-amanitin and Pol II from S. scrofa. The oxygen of the
OH can accept a proton from Glu845 and the proton of the OH could act as an
H-bond donor although the acceptor is currently unknown. (B) Expected
interactions between the ammonium group of NH2-Pro-AMA analog and Pol II.
Figure 11. Results of the in vitro cell-based assay of anti-HER2 ADCs of CN-
Pro- and NH2-Pro-amanitins against (A) JIMT-1 (HER2+), (B) SK-BR-3 (HER2+),
(C) SK-OV-3 (HER2+), and (D) MDA-MB-231 (HER2-) cells.
+
A proton from the NH3 group can form a salt bridge with Glu845, sill still
providing an H-bond donor. Similarly, a bifurcated set of H-bonds could be
envisioned for the Gdn-Pro-AMA analog (not shown).
Among the Hyp analogs incorporated into the dideoxy-amanitin
core, surprisingly none was as active as dideoxy-amanitin (2), as
assayed in cytotoxicity assays and by run-off in vitro transcription
assays. This disparity between the hydroxyproline and known
analogs is challenging to rationalize and merits further discussion.
Although MMFF-based molecular modeling (Spartan, data not
shown) suggested varying degrees of success with all of these
derivatives, we recognized the challenge of replacing a (4R)-
Nevertheless, the net energetic contribution of these H-bonding
interactions is likely low since the toxin must be desolvated from
bulk solvent wherein the same H-bonds will exist. Notwithstanding
the thermoneutrality of H-bonding, we reasoned that the (4R)-
amino-, (4R)-guanidino-, (4R)-methylcarbamoyl-, and (4R)-
cyano-proline could be recognized by Glu845, either through H-
bonding, charge-charge complementarity, or both. Similarly, the
methylcarbamoyl-proline proved inactive despite the strong
potential for H-bonding and an expected Cγ-exo conformation. As
a control, we used the (4R)-azido-proline that is incapable of H-
bonding. The fact that 6 (ammonium) and 7 (guanidinium) were
the strongest inhibitors of Pol II in this series may point to the
potential for accessing the H-bonding interactions seen in the
crystal and cryo-EM structures while assuming the correct Cγ-exo
conformation.
hydroxyl
group
on
proline
with
a
hydrogen-bond
donating/accepting group that would enforce the Cγ-exo
conformation.
CD spectra suggest that the overall 3D structures of the
synthetic analogs differ slightly from dideoxy amanitin, but
generally resemble each other (Figure 7, A). Alpha-amanitin (1)
and all synthetic analogs show a positive Cotton effect past 250
nm, yet behave differently at wavelengths below 250 nm with the
synthetic analogs exhibiting a local minimum at 230-234 nm, while
α-amanitin shows a local minimum at 220 nm. The difference
between the synthetic toxins and α-amanitin may be explained by
the fact that α-amanitin (1) has a different chromophore owing to
the 6'-OH. Greater differences were observed when the CD-
spectra were acquired at pH 4.5. Although the relevance of
working at pH 4.5 could be questioned, we chose this to
complement our study in an attempt to find a medium where
differences could be revealed. However, it is not readily
understood why changing the pH would cause such dramatic
deviation since none of the prolines has an ionizable group that
would be perturbed upon decreasing the pH from neutrality to 4.5.
We expected lower IC50 values for NH2- and Gdn-Pro amanitins
as we anticipated these to form H-bonds or salt bridges with
+
Glu845 of Pol II. Whereas as cations (NH3 and GdnH+,
respectively), these would not serve as H-bond acceptors, yet
they could form strong salt bridges with Glu845 while still donating
an H-bond to a putative acceptor. Protected versions including the
N-Boc-protected guandidine-proline and S-Acm-protected
mercapto-proline were also evaluated as these present additional
functionalities capable of further H-bonding interactions that may
7
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
find supplementary interactions within the binding site. Whereas
we did not expect success with these protected analogs given
their extra steric bulk, we felt their evaluation would complement
this study.
HEK cells – assuming unspecific and thus comparable uptake of
all variants into these cells lacking the OATP1B3 transporter.
The reduced yet detectable binding affinity of selected Hyp
variants on Pol II, along with the considerably reduced cytotoxic
effect on OATP1B3-overexpressing HEK293 cells, prompted us
to develop ADCs based on specific variants for targeted delivery
to cancer cells and controlled release of the payload through
cleavage of the inter-positioned protease-sensitive linker. We
selected NH2-Pro-amanitin, which showed a strong inhibitory
potential on Pol II, for attachment of a valine-alanine linker via
carbamate linkage. Additionally, we modified CN-Pro-amanitin
with this Val-Ala linker through a cyclic acetal (see Supporting
Information). For PoC in vitro studies, these amanitin-linker
variants derived from NH2-Pro-amanitin and CN-Pro-amanitin
were conjugated to Trastuzumab for targeted delivery to HER2+
cells using site-specific conjugation chemistry.
In vitro characterization was performed on three HER2+ cell
lines (SK-BR-3, SK-OV-3 and JIMT-1) and on one HER2-
(HER2low) cell line (MDA-MB-231; Figure 11). The cytotoxic
potency of the two ADCs with Hyp analogs was compared to an
ADC with an α-amanitin variant as the cytotoxic payload using the
same Val-Ala linker (T-D265C-OH-Pro-amanitin). All HER2+ cell
lines showed high sensitivity towards the ADC based on α-
amanitin with IC50 values of 0.12 ± 0.01 nM for JIMT-1, 7.84 ± 0.14
pM for SK-BR-3 and 5.54 ± 0.71 pM for SK-OV-3 cells,
respectively. In contrast, JIMT-1 and SK-OV-3 cells did not show
any response towards the Hyp-amanitin variants (Figure 11, C).
However, on SK-BR-3 cells (highest HER2-expression), the Hyp
variants were able to induce a mild cytotoxic effect by reducing
cell viability to approx. 50% with an IC50 value of 1.16 ± 0.89 nM
for T-D265C-NH2-Pro-amanitin, and an IC50 value of 0.55 ± 0.10
nM for T-D265C-CN-Pro-amanitin, respectively (Figure 11, C).
The HER2- cell line MDA-MB-231 did not respond to ADC
treatment indicating that no unspecific effects are responsible for
the cytotoxicity on HER2+ cell lines (Figure 11, D).
With respect to the mercapto-prolyl-amanitin (5), we recognized
that the thiol may serve as a possible analog of a hydroxyl group
albeit being a poor H-bond donor and incapable of accepting an
H-bond.[46] A preference for Cγ-endo ring puckering in this case
may further explain the generally poor inhibitory activity of this
analog but not the others that prefer the Cγ-exo conformation.
Nevertheless, the energetic difference between the exo and endo
forms is on the order of ca.-1 kcal/mol at 25°C, which translates
to an exo/endo ratio of only ca. 6.[16c, 47] This conformational bias
alone is not enough to explain why this analog is so much poorer
an inhibitor compared to 2. However the added differences in H-
bonding capacity may further erode inhibitory activity. More
challenging however is to rationalize why prolines that are
properly biased in favor of the Cγ-exo conformation show such
poor inhibitory activity and correlated cytotoxicity.
The in vitro toxicity of the synthetic analogs was further
assessed on HEK293 (Figure ESI.7 in Supporting Information)
and in HEK293-OATP1B3 cells that overexpress the organic
anion-transporting polypeptide 1B3 (OATP1B3) (Figure ESI.8 in
Supporting Information) to assess whether OATP1B3 mediates
active transport of α-amanitin and the Hyp analogs of amanitin
into the cell. Human embryonic kidney HEK293 cells lack the
OATP1B3 transporter (wt-HEK293) and hence showed only
micromolar susceptibility to α-amanitin (IC50 = 0.13± 0.08 µM) and
β-amanitin (IC50 = 0.12 ± 0.09 µM), while the amanitin derivatives
showed no cytotoxic effect. On the contrary, transfected HEK293
cells constitutively expressing OATP1B3 (HEK293-OATP1B3)
showed clearly enhanced sensitivity to α-amanitin, with an IC50
value of 43 ± 7.4 nM. The 3-fold increase in toxicity of α-amanitin
on HEK293-OATP1B3 cells relative to HEK293 cells indicates
that this toxin is subject to OATP1B3-mediated transport (note β-
amanitin – used as a positive control, owing to an anionic
aspartate in lieu of the asparagine is an even better substrate for
the OATP1B3 transporter as indicated by a 15-fold reduction of
the IC50 value to 8.5 ± 1.1 nM).
These analogs showed no cytotoxicity on HEK293 cells and only
a very mild cytotoxic effect only at high concentrations in the µM
range on HEK293-OATP1B3 cells. For example, at the highest
used concentration of 1 µM, cell viability of NH2-Pro-amanitin
showed the lowest value of 33%, in comparison to N3-Pro-
amanitin (68%), SH-Pro-amanitin (50%), CN-Pro-amanitin (43%),
Gdn-Pro-amanitin (86%), MeOCONH-Pro-amanitin (61%), α-
amanitin (0%) and β-amanitin (0%). This slightly lower cell viability
of OATP1B3-overexpressing cells compared to HEK293 cells
after exposure to Hyp analogs at identical concentrations could
be an indication that these variants are only partially transported
into the cell via the organic anion-transporting polypeptide 1B3
(OATP1B3). However, in direct comparison to α-amanitin, these
variants exhibit substantially lower cytotoxicity on these
overexpressing cells as the natural compound. This reduced
cytotoxic potential could be attributed to (i) a reduced binding
affinity to RNA Pol II and/or (ii) limited uptake into the cell via the
OATP1B3 transporter. The correlation between the inhibitory
effect on Pol II and the cytotoxic potency on HEK-OATP1B3 cells
of the different Hyp analogs supports the assumption that Pol II
binding affinity plays a decisive role in mediating cytotoxicity. This
finding is in line with the absence of any cytotoxic effect on wt-
The reduced inhibitory activity of some of the modified amanitin
derivatives nevertheless show the promising potential of these
analogs for therapeutic applications if used as payload for ADCs.
The combination of being a poor substrate for OATP1B3
transporters whilst retaining inhibitory activity of amanitin to a
certain degree might help to develop ADCs with reduced payload-
mediated toxicity and an improved target-specific effect. ADCs
based on low potency payloads like SN38 (e.g. ENHERTU®)
make use of high drug-antibody ratios to overcome the limitation
of the payload. Notwithstanding that toxicity was greatly
diminished, the aim of synthesis must be to address SARs. This
work will inform the design of provide toxins with attenuated
toxicity such that these modified derivatives may still serve as
candidate payloads for the development of an ADC.
Conclusion
At this juncture, we can conclude that Hyp residue is essential
for the cytotoxicity and in vitro inhibitory activity of α-amanitin and
replacement with analogs that afford the same conformational Cγ-
exo pucker of Hyp significantly erode toxicity. Surprisingly, all
analogs were far less cytotoxic than α-amanitin and dideoxy-
amanitin however when tested in vitro, two analogs (6 & 7) were
1
only 5-fold less active than 2. While an H-15N NMR study could
provide more insight into conformational effects, intra-annular
8
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
[1] a) H. Wieland, R. Hallermayer, Justus Liebigs Ann. Chem. 1941, 548, 1-18;
b) T. Wieland, H. Faulstich, CRC Critical Reviews in Biochemistry 1978,
5, 185-260; c) T. Wieland, Int. J. Pept. Protein Res. 1983, 22, 257-276.
[2] a) G. Zanotti, C. Mohringer, T. Wieland, Int. J. Pept. Protein Res. 1987, 30,
450-459; b) G. Zanotti, C. Birr, T. Wieland, Int. J. Pept. Protein Res. 1981,
18, 162-168.
NOE distances are unlikely to differ by more than an angstrom
and such would be difficult to accurately quantify. Furthermore,
such would only describe an envelope of ground state structures
and would not fully interrogate the bound form.
X-ray or cryo-EM structures of Pol-II/amanitin would be more
informative, yet it is unlikely that these will provide the resolution
needed. As the acquisition of such structural data goes far beyond
the scope of this study, we cautiously assert that the effects of the
proline substitution must induce subtle structural effects and/or
mismatched interactions within the polymerase that conspire to
effect significantly lower potency. Future studies to obtain
structural information are anticipated to complement this work in
defining the critical nature of the Hyp on cytotoxicity.
[3] a) M. Cochetme, P. Chambon, Biochim. Biophys. Acta 1974, 353, 160-184;
b) C. Kedinger, M. Gniazdow, J. L. Mandel, F.P. Gissinge, P. Chambon,
Biochem. Biophys. Res. Commun. 1970, 38, 165-171.
[4] V. T. Nguyen, F. Giannoni, M. F. Dubois, S. J. Seo, M. Vigneron, C. Kedinger,
O. Bensaude, Nucleic Acids Res. 1996, 24, 2924-2929.
[5] K. C. Nicolaou, S. Rigol, Angew. Chem.-Int. Edit. 2019, 58, 11206-11241.
[6] M. T. B. Davis, J. F. Preston, Science 1981, 213, 1385-1388.
[7] G. Moldenhauer, A. V. Salnikov, S. Luttgau, I. Herr, J. Anderl, H. Faulstich,
J. Natl. Cancer Inst. 2012, 104, 622-634.
In conclusion, this is the first report to address the role of the
Hyp in amanitin by testing analogs thereof, four of which were
chosen for their well-known Cγ-exo conformation characteristic of
Hyp while providing a test of H-bonds that engage the target RNA
Pol II. These analogs failed to show significant toxicity highlighting
the mystery that continues to shroud this venerated toxin.
Replacing the hydroxyl group on Hyp with either an amine,
guanidine, or thiol may yet provide useful chemical handles for
the design of bio-reducible and self-immolating conjugates. If
potency can be augmented by the use of a targeting agent such
as an antibody, lower inherent cytotoxicity might actually prove
advantageous for therapeutic applications, particularly if the less-
potent composition provides a chemical handle for further
mediating intracellular activation or reducing overall systemic
toxicity. Hence these analogs may still find use in the design of
new amanitin-bioconjugates for ADCs.
[8] P. Ramkumar, A. B. Abarientos, R. L. Tian, M. Seyler, J. T. Leong, M. Chen,
P. Choudhry, T. Hechler, N. Shah, S. W. Wong, T. G. Martin, J. L. Wolf,
K. T. Roybal, A. Pahl, J. Taunton, A. P. Wiita, M. Kampmann, Blood Adv.
2020, 4, 2899-2911.
[9] a) M. A. Krzyscik, L. OpalinskiJ. Otlewski, Mol. Pharm. 2019, 16, 3588-3599;
b) Y. J. Li, Y. H. Liu, H. C. Xu, G. L. Jiang, K. Van der Jeught, Y. Z. Fang,
Z. L. Zhou, L. Zhang, M. Frieden, L. F. Wang, Z. H. Luo, M. Radovich, B.
P. Schneider, Y. B. Deng, Y. L. Liu, K. Huang, B. He, J. Wang, X. M. He,
X. N. Zhang, G. Ji, X. B. Lu, Nat Commun. 2018, 9, 15; c) L. Bodero, P.
L. Rivas, B. Korsak, T. Hechler, A. Pahl, C. Muller, D. Arosio, L. Pignataro,
C. Gennari, U. Piarulli, Beilstein J. Org. Chem. 2018, 14, 407-415; d) L.
Zhao, J. P. May, A. Blanc, D. J. Dietrich, A. Loonchanta, K. Matinkhoo,
A. Pryyma, D. M. Perrin, ChemBioChem 2015, 16, 1420-1425; e) A.
Moshnikova, V. Moshnikova, O. A. Andreev, Y. K. Reshetnyak,
Biochemistry 2013, 52, 1171-1178.
[10] D. A. Bushnell, P. Cramer, R. D. Kornberg, Proc. Natl. Acad. Sci. USA 2002,
99, 1218-1222.
[11] F. Brueckner, P. Cramer, Nat. Struct. Mol. Biol. 2008, 15, 811-818.
[12] X. Y. Liu, L. Farnung, C. Wigge, P. Cramer, J. Biol. Chem. 2018, 293, 7189-
7194.
[13] A. Buku, R. Altmann, T. Wieland, Annalen Der Chemie-Justus Liebig 1974,
1580-1586.
Conflict of Interest
[14] a) G. Shoham, D. C. Rees, W. N. Lipscomb, G. Zanotti, T. Wieland, J. Am.
Chem. Soc. 1984, 106, 4606-4615; b) T. Wieland, C. Gotzendorfer, G.
Zanotti, A. C. Vaisius, Eur. J. Biochem. 1981, 117, 161-164.
[15] a) E. C. Kostansek, W. N. Lipscomb, R. R. Yocum, W. E. Thiessen, J. Am.
Chem. Soc. 1977, 99, 1273-1274; b) E. C. Kostansek, W. N. Lipscomb,
R. R. Yocum, W. E. Thiessen, Biochemistry 1978, 17, 3790-3795; c) G.
Shoham, W. N. Lipscomb, T. Wieland, J. Am. Chem. Soc. 1989, 111,
4791-4809.
Compositions of matter described herein are the subject of a PCT
filing. Heidelberg Pharma GmBH holds patents on derivatives of
amanitin and conjugates thereof.
Experimental Section
[16] a) N. Panasik, E. S. Eberhardt, A. S. Edison, D. R. Powell, R. T. Raines,
Int. J. Pept. Protein Res. 1994, 44, 262-269; b) S. K. Holmgren, L. E.
Bretscher, K. M. Taylor, R. T. Raines, Chem. Biol. 1999, 6, 63-70; c) M.
L. DeRider, S. J. Wilkens, M. J. Waddell, L. E. Bretscher, F. Weinhold,
R. T. Raines, J. L. Markley, J. Am. Chem. Soc. 2002, 124, 2497-2505.
[17] K. Matinkhoo, A. Pryyma, M. Todorovic, B. O. Patrick, D. M. Perrin, J. Am.
Chem. Soc. 2018, 140, 6513-6517.
Experimental Details are given in the electronic supporting
information.
Acknowledgements
[18] a) M. A. J. Siegert, C. H. Knittel, R. D. Sussmuth, Angew. Chem.-Int. Edit.
59 5500-5504; b) C. Lutz, W. Simon, S. Werner-Simon, A. Pahl, C. Muller,
Angew. Chem.-Int. Edit. 2020, 59, 11390-11393.
We acknowledge funding from the Canadian Institute of Health
Research #220656 and the Canadian Cancer Society Research
Initiative Grant #703374, and from Heidelberg Pharma Research
GmbH for providing the necessary funds to conduct this work. Dr.
C. M. Hambira was supported in part by a competitively awarded
postdoctoral fellowship of the Michael Smith Foundation for
Health Research. The authors thank Dr. Maria Ezhova for
assisting with NMR spectra acquisition, and Dr. Elena Polishchuk
and Ms. Jessie Chen for help with cytotoxicity assays.
[19] T. Wieland, C. Gotzendorfer, J. Dabrowski, W. N. Lipscomb, G. Shoham,
Biochemistry 1983, 22, 1264-1271.
[20] A. Pryyma, K. Matinkhoo, A. A. Wong, D. M. Perrin, Chem. Sci. 2020.
[21] G. Zanotti, G. Petersen, T. Wieland, Int. J. Pept. Protein Res. 1992, 40,
551-558.
[22] S. K. Holmgren, K. M. Taylor, L. E. Bretscher, R. T. Raines, Nature 1998,
392, 666-667.
[23] a) F. W. Kotch, I. A. Guzei, R. T. Raines, J. Am. Chem. Soc. 2008, 130,
2952-2953; b) M. D. Shoulders, I. A. Guzei, R. T. Raines, Biopolymers
2008, 89, 443-454.
Keywords: amanitin • amatoxin • hydroxyproline • peptide •
[24] M. D. Shoulders, R. T. Raines, Annu. Rev. Biochem 2009, 78, 929-958.
[25] A. K. Pandey, D. Naduthambi, K. M. Thomas, N. J. Zondlo, J. Am. Chem.
Soc. 2013, 135, 4333-4363.
antibody-drug conjugate
9
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
[26] D. S. Eggleston, D. J. Hodgson, Int. J. Pept. Protein Res. 1985, 25, 242-
253.
[27] S. A. Cadamuro, R. Reichold, U. Kusebauch, H. J. Musiol, C. Renner, P.
Tavan, L. Moroder, Angew. Chem.-Int. Edit. 2008, 47, 2143-2146.
[28] a) I. R. Babu, K. N. Ganesh, J. Am. Chem. Soc. 2001, 123, 2079-2080; b)
M. Umashankara, I. R. Babu, K. N. Ganesh, Chem. Commun. 2003,
2606-2607; c) M. Umashankara, M. Nanda, M. SonarK. N. Ganesh,
Chimia 2012, 66, 936-940; d) M. Nanda, K. N. Ganesh, J. Org. Chem.
2012, 77, 4131-4135; e) M. Umashankara, M. V. Sonar, N. D. Bansode,
K. N. Ganesh, J. Org. Chem. 2015, 80, 8552-8560; f) J. Egli, C. Siebler,
B. Maryasin, R. S. Erdmann, C. Bergande, C. Ochsenfeld, H.
Wennemers, Chem.-Eur. J. 2017, 23, 7938-7944.
[29] a) M. Kuemin, Y. A. Nagel, S. Schweizer, F. W. Monnard, C. Ochsenfeld,
H. Wennemers, Angew. Chem.-Int. Edit. 2010, 49, 6324-6327; b) R. S.
Erdmann, H. Wennemers, J. Am. Chem. Soc. 2012, 134, 17117-17124.
[30] a) C. Laurence, M. Berthelot, Perspect. Drug Discovery Des. 2000, 18, 39-
60; b) N. Ziao, J. Graton, C. Laurence, J. Y. Le Questel, Acta Crystallogr.
Sect. B-Struct. Sci.Cryst. Eng. Mat. 2001, 57, 850-858; c) D. L. Chen, N.
Oezguen, P. Urvil, C. Ferguson, S. M. Dann, T. C. Savidge, Sci. Adv.
2016, 2, 16.
[31] a) L. S. Sonntag, S. Schweizer, C. Ochsenfeld, H. Wennemers, J. Am.
Chem. Soc. 2006, 128, 14697-14703; b) M. Kumin, L. S. Sonntag, H.
Wennemers, J. Am. Chem. Soc. 2007, 129, 466-467; c) R. S. Erdmann,
M. Kumin, H. Wennemers, Chimia 2009, 63, 197-200.
[32] J. A. Gomez-Vidal, M. T. Forrester, R. B. Silverman, Org. Lett. 2001, 3,
2477-2479.
[33] A. Pryyma, Y. J. Bu, Y. Wai, B. O. Patrick, D. M. Perrin, Org. Lett. 2019, 21,
8234-8238.
[34] W. E. Savige, A. Fontana, Int. J. Pept. Protein Res. 1980, 15, 102-112.
[35] M. Jbara, S. Laps, M. Morgan, G. Kamnesky, G. Mann, C. Wolberger, A.
Brik, Nat. Commun. 2018, 9, 11.
[36] A. L. Cimecioglu, D. H. Ball, D. L. KaplanS. H. Huang, Macromolecules
1994, 27, 2917-2922.
[37] a) A. Yagodkin, K. Loschcke, J. Weisell, A. Azhayev, Tetrahedron 2010, 66,
2210-2221; b) D. Carnaroglio, K. Martina, G. Palmisano, A. Penoni, C.
Domini, G. Cravotto, Beilstein J. Org. Chem. 2013, 9, 2378-2386.
[38] C. Z. Gao, Z. B. Fisher, K. J. Edgar, Cellulose 2019, 26, 445-462.
[39] M. Q. Tian, M. Yan, P. S. Baran, J. Am. Chem. Soc. 2016, 138, 14234-
14237.
[40] K. Letschert, H. Faulstich, D. Keller, D. Keppler, Toxicol. Sci. 2006, 91, 140-
149.
[41] J. D. Dignam, R. M. Lebovitz, R. G. Roeder, Nucleic Acids Res. 1983, 11,
1475-1489.
[42] a) F. Novello, L. Fiume, F. Stirpe, Biochem. J 1970, 116, 177-&; b) T. J.
Lindell, F. Weinberg, P. W. Morris, R. G. Roeder, W. J. Rutter, Science
1970, 170, 447-449.
[43] W. Schmitt, G. Zanotti, T. Wieland, H. Kessler, J. Am. Chem. Soc. 1996,
118, 4380-4387.
[44] a) D. Wang, D. A. Bushnell, X. H. Huang, K. D. Westover, M. Levitt, R. D.
Kornberg, Science 2009, 324, 1203-1206; b) C. D. Kaplan, K. M. Larsson,
R. D. Kornberg, Mol. Cell 2008, 30, 547-556.
[45] A. C. M. Cheung, P. Cramer, Cell 2012, 149, 1431-1437.
[46] H. S. Biswal, P. R. Shirhatti, S. Wategaonkar, J. Phys. Chem. A 2010, 114,
6944-6955.
[47] R. W. Newberry, R. T. Raines, Topics in heterocyclic chemistry 2017, 48,
1-25.
10
This article is protected by copyright. All rights reserved.

10.1002/chem.202101373
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents
Getting at the Hyp of Amanitin: A formal study of structure-function relationships for trans-hydroxyproline – a key functionality the
specifies the toxicity of amanitin is undertaken to probe aspects of H-bonding and other interactions with analogs of hydroxyl-proline.
The toxin is surprisingly refractory to replacing the hydroxyproline.
11
This article is protected by copyright. All rights reserved.