A New Fluorogenic Small-Molecule Labeling Tool for Surface Diffusion Analysis and Advanced Fluorescence Imaging of β-Site Amyloid Precursor Protein-Cleaving Enzyme 1 Based on Silicone Rhodamine: SiR-BACE1

Article abstract of DOI:10.1021/acs.jmedchem.8b00387

β-site APP-cleaving enzyme 1 (BACE1) is a major player in the pathogenesis of Alzheimer's disease. Structural and functional fluorescence microscopy offers a powerful approach to learn about the physiology and pathophysiology of this protease. Up to now, however, common labeling techniques require genetic manipulation, use large antibodies, or are not compatible with live cell imaging. Fluorescent small molecules that specifically bind to the protein of interest can overcome these limitations. Herein, we introduce SiR-BACE1, a conjugate of the BACE1 inhibitor S-39 and SiR647, as a novel fluorogenic, tag-free, and antibody-free label for BACE1. We present its chemical development, characterize its photophysical and pharmacologic properties, and evaluate its behavior in solution, in overexpression systems, and in native brain tissue. We demonstrate its applicability in confocal, stimulated emission depletion and dynamic single-molecule microscopy. The first functional studies with SiR-BACE1 on the surface mobility of BACE1 revealed a markedly confined diffusion pattern.

Full text of DOI:10.1021/acs.jmedchem.8b00387

Subscriber access provided by - Access paid by the | UCSB Libraries
Article
A new fluorogenic small molecule labeling tool for surface diffusion
analysis and advanced fluorescence imaging of #-site amyloid precursor
protein (APP)-cleaving enzyme 1 based on silicone rhodamine: SiR-BACE1
Sandra Karch, Johannes Broichhagen, Julia Schneider, Daniel Böning, Stephanie Hartmann,
Benjamin Schmid, Philipp Tripal, Ralf Palmisano, Christian Alzheimer, Kai Johnsson, and Tobias Huth
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00387 • Publication Date (Web): 25 Jun 2018
Downloaded from http://pubs.acs.org on June 25, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A new fluorogenic small molecule labeling tool for surface diffusion
analysis and advanced fluorescence imaging of βꢀsite amyloid precursor
protein (APP)ꢀcleaving enzyme 1 based on silicone rhodamine: SiRꢀ
BACE1
§,1
§,2,3
§,1
4
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Sandra Karch , Johannes Broichhagen
, Julia Schneider , Daniel Böning , Stephanie
1
5
5
5
1
Hartmann , Benjamin Schmid , Philipp Tripal , Ralf Palmisano , Christian Alzheimer ,
§§,2,3
§§,1
Kai Johnsson
, Tobias Huth*
1
Institute of Physiology and Pathophysiology, FriedrichꢀAlexanderꢀUniversität Erlangenꢀ
Nürnberg, Universitaetsstraße 17, 91054 Erlangen, Germany.
2
Department of Chemical Biology, Max Planck Institute for Medical Research, Jahnstraße
29, 69120 Heidelberg, Germany.
3
Laboratory of Protein Engineering, Institut des sciences et ingénierie chimiques, Sciences de
base, École Polytechnique Fédérale Lausanne, 1015 Lausanne, Switzerland.
4
Max Planck Institute for the Science of Light, Staudtstraße 2, 91058 Erlangen, Germany.
5
Optical Imaging Centre, FriedrichꢀAlexanderꢀUniversität ErlangenꢀNürnberg,
Hartmannstraße 14, 91052 Erlangen, Germany.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
βꢀsite APPꢀcleaving enzyme 1 (BACE1) is a major player in the pathogenesis of Alzheimer’s
disease. Structural and functional fluorescence microscopy offers a powerful approach to
learn about the physiology and pathophysiology of this protease. Up to now, however,
common labeling techniques either require genetic manipulation, use large antibodies, or are
not compatible with live cell imaging. Fluorescent small molecules that specifically bind to
the protein of interest can overcome these limitations. Herein, we introduce SiR-BACE1, a
conjugate of the BACE1 inhibitor Sꢀ39 and SiR647, as a novel fluorogenic, tagꢀfree, and
antibodyꢀfree label for BACE1. We present its chemical development, characterize its photoꢀ
physical and pharmacologic properties, and evaluate its behavior in solution, in overꢀ
expression systems, and in native brain tissue. We demonstrate its applicability in confocal,
stimulated emission depletion (STED), and dynamic single molecule microscopy. First
functional studies with SiR-BACE1 on the surface mobility of BACE1 revealed a markedly
confined diffusion pattern.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
ACS Paragon Plus Environment

Page 3 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
First described by Alois Alzheimer over a century ago, Alzheimer’s disease (AD) is the most
common form of dementia with progressive cognitive impairment including memory and
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
speech deficits, spatial disorientation, apraxia, and psychiatric symptoms. Most cases can be
2
,3
classified as sporadic AD, for which age is a major risk factor. Due to the aging population,
4
AD is expected to become a growing socioꢀeconomic challenge in the upcoming decades.
The brains of AD patients display two histopathological characteristics: i) amyloid plaques,
which mainly consist of accumulated Aβ peptides, and ii) neurofibrillary tangles formed by
5
aggregated hyperꢀphosphorylated tau protein. Despite a still incompletely understood
etiology, vast evidence supports the amyloid cascade hypothesis, which suggests a
6
–9
deregulated Aβ homeostasis as an early step in the pathogenesis. Aβ is generated by twoꢀ
step proteolysis from amyloid precursor protein (APP). The first and rateꢀlimiting step is
10–14
catalyzed by βꢀsite APPꢀcleaving enzyme 1 (BACE1).
BACE1 is an aspartic protease,
mainly located in the transꢀGolgi network and the endosomal pathway, which undergoes
pronounced surface trafficking and recycling to early endosomes, where it coꢀlocalizes with
1
5–17
and processes APP.
Predisposing or protecting hereditary APP mutations that affect
1
8–21
processing rates at the βꢀsite underscore the pathogenic impact of BACE1.
Current knowledge on physiological and pathological, proteolytic and nonꢀproteolytic
functions of BACE1 is based on decades of research including animal, histological,
5
,17,22,23
biochemical, electrophysiological, and optical studies.
Light microscopy has
contributed to some of the key findings: Immunofluorescence demonstrated that BACE1 is
expressed in mossy fiber synapses of the hippocampal formation, which belongs to the brain
regions that are affected by the earliest pathologic changes and are central for memory
2
4,25
formation and retrieval.
Antibody staining also showed that BACE1 accumulates near
amyloid plaques in dystrophic neurites, being associated with impaired lysosomal degradation
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
6,27
and possibly fostering Aβ secretion.
BACE1 and APP surface trafficking is independent
from each other, as illustrated by total internal reflection fluorescence (TIRF) microscopy
2
8
using fusions of the proteins to pHꢀsensitive fluorescent indicators. Fluorescence resonance
energy transfer (FRET) between antibodies directed against tagged proteins revealed close
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
15
encounter between BACE1 and APP in early endosomes. A fluorescence complementation
approach with BACE1 and APP fusion proteins recently visualized their interaction in neurite
microdomains and axonal BACE1ꢀAPP coꢀtrafficking, translating previous work to the
29
neuronal environment.
Interrogation of BACE1 in overꢀexpression systems and in nonꢀneuronal cell lines may not
necessarily give a full and unbiased account of its many actions in the healthy and diseased
brain. This holds in particular for all approaches using fluorescence imaging, where bulky
labels impose steric hindrances, thereby possibly affecting BACE1 trafficking and regulation.
Given the constraints of conventional imaging tools, alternative specific, bright, and easyꢀtoꢀ
apply labeling solutions are eagerly awaited to visualize BACE1 molecules and track their
subcellular pathways and interaction sites at the highest spatiotemporal resolution.
Because of the central role of BACE1 in AD pathogenesis, great efforts were made to develop
30,31
small molecule BACE1 inhibitors.
Despite some setbacks, inhibition of BACE1, most
promising as part of a multitarget combination therapy, has remained a rational strategy to
32–35
prevent and treat AD.
Aminohydantoins have been introduced as a class of inhibitors that
contact the catalytic aspartates, inhibit BACE1 at subꢀmicromolar concentrations, and possess
3
6
good brain permeability. The derivative Sꢀ39 exhibits a low nanomolar inhibitory potency
3
7,38
and shows excellent selectivity over related proteases in screenings in vitro.
To meet the urgent needs for innovative BACE1 visualization strategies, we made use of the
preꢀclinical inhibitor Sꢀ39 as a starting point to build a novel BACE1 labeling tool. Here, we
introduce an organic conjugate that exploits the high affinity and specificity of Sꢀ39 in
3
ACS Paragon Plus Environment

Page 5 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
combination with the farꢀred siliconꢀrhodamine derivative SiR647 and demonstrate the
viability and functionality of this new approach.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Results
A novel fluorogenic probe for BACE1 visualization
We aimed to develop a probe for BACE1 labeling independent of tags and with a good signalꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
39
toꢀnoise ratio based on fluorogenicity (Figure 1A). Therefore, we started to design
conjugates of the fluorogenic siliconeꢀrhodamine derivative SiR647 (Figure 1B). The small
3
7
molecule BACE1 inhibitor Sꢀ39 was linked to the dye via methylene linkers of different
lengths that drag the conjugate into the active site pocket of BACE1. Binding to the protein
should bring the SiR647 component into the polar environment of the protein surface and
cause a conformational switch from its nonꢀfluorescent spirolactone form to its fluorescent
zwitterꢀion, as shown in Figure 1B. The linkers ranged from 2 to 8 methylene spacers
(conjugates C2 to C8) and were synthetically introduced lateꢀstage, a fact that allows flexible
alteration, optimization and synthetic economy (Figure 1C). In order to do so, a multistep
synthetic sequence was performed, starting from acetonedicarboxylic acid (1) and propionic
anhydride (2) to obtain bromo pyridine 3 according to a previously reported procedure from
40
Li et al. Silicon rhodamine (SiR) 4 was functionalized after activation with TSTU and
coupling to diamines with several linker lengths (denoted by n), before a guanidine was
installed on the primary amine to obtain 5a-f (see Supplementary Tables 1 and 2). In parallel,
the main pharmacophore was built from commercially available boronic acid 6 and 3ꢀ
bromoiodobenzene (7) by a iodoꢀselective Suzuki reaction that gave access to bromo aryl 8.
Next steps were a Sonogashira reaction with trimethylsilyl acetylene (to obtain 9),
deprotection (to obtain 10) and subsequent Sonogashira reaction with previously obtained 3 to
obtain internal alkyne 11. Triple bond oxidation of 11 with permanganate yielded 1,2ꢀ
bisketone 12, that underwent benzilic rearrangement with SiRꢀlinked guanidines. This
straightforward access to SiRꢀlinked probes 13a-f gave access to several possible fluorogenic
BACE1 inhibitors (see Supplementary Table 3).
5
ACS Paragon Plus Environment

Page 7 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
We next investigated the photoꢀphysical properties of our conjugates in solution. Figure 1D
and E show the excitation and emission spectra for the C4 conjugate, termed SiR-BACE1.
Absorption was maximal at 651 nm and emission was observed at around 660ꢀ680 nm with
its maximum at 674 nm (Figure 1D). To control whether SiR647 had preserved its
fluorogenicity as a component of the conjugate, we recorded the absorption in PBS with or
without the addition of 0.2% SDS (Figure 1E). In the absence of detergent, the probe lost its
ability to absorb light in the range of 650 nm, hence our conjugates were fluorogenic. In this
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
39
experiment, fluorogenicity was promoted by the nonꢀpolar environment of SDS micelles. To
determine whether the fluorogenic onꢀswitch is also mediated by specific binding to BACE1,
we recorded SiR-BACE1 emission in the absence or presence of recombinant BACE1 protein
(Figure 1F). The target protein increased fluorescence in the expected spectral range, which
confirmed BACE1ꢀmediated fluorogenic onꢀswitch.
BACE1 binding, inhibition, and visualization by a C4-linker S-39/SiR647
conjugate
Effective BACE1 labeling by the compound happens in two steps: binding of the probe and
switching to its fluorescent form. Thus, we aimed to identify conditions that promote both. In
a first experiment, we overꢀexpressed hBACE1ꢀEGFP in HEK293T cells and stained with all
conjugates of different linker length. hBACE1ꢀEGFP located mainly to the plasma membrane
and intracellular compartments near the nucleus (Figure 2A, middle panel). At medium linker
lengths (C4 to C6), our conjugates showed the strongest coꢀstaining (Figure 2A, bottom
panel), indicating efficient binding to BACE1 and switching to the onꢀstate, with C4
displaying the highest intensities.
From the data in Figure 2A, we generated colorꢀcoded 2D histograms of EGFP and SiR647
intensities to further evaluate the coꢀstaining (Figure 2B). For all linker lengths, we observed
6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
two areas of coꢀlocalized pixels in the whole frame coꢀlocalization plots, which could be
assigned to pixels at the plasma membrane (membrane ROI, middle panel) or pixels of
perinuclear clusters (cluster ROI, bottom panel). At the plasma membrane, the C4 and C5
conjugates showed the most intense staining, indicated by the steepest correlation and
numerous pixels with high SiR647 intensities (Figure 2B, middle panel). In perinuclear
cluster regions, C4 was clearly superior to the other substances, giving the highest SiR647
signals (Figure 2B, bottom panel). Given these results, we chose C4 as our BACE1 labeling
probe, which we termed SiR-BACE1.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Next, we reasoned that binding of the conjugate to the active site pocket of BACE1 should
inhibit its proteolytic activity, like the parent drug Sꢀ39. To determine and quantify inhibition
of BACE1 by increasing concentrations of the different conjugates, we performed a FRETꢀ
based enzymatic assay (Figure 2C) using a synthetic BACE1 substrate coupled to a donor
fluorophore, which is quenched until substrate cleavage by BACE1. All conjugates indeed
inhibited BACE1 in a doseꢀdependent manner, with BACE1 being more efficiently
suppressed by conjugates with longer linkers (Figure 2D). Plotting the IC50 values of all linker
lengths and performing a cluster analysis, the IC50 of conjugates with 2 or 3 methylene
spacers (C2, C3) clustered around 1700 nM, while IC50 of conjugates with 4 or more
methylene spacer groups (C4ꢀC6, C8) clustered around 1100 nM. We conclude that
conjugates were slightly hindered in binding by a linker length ≤3 carbon atoms but perform
equally well above that threshold. Staining differences as observed in Figure 2A must hence
be due to varying fluorogenic conversion.
Acidic conditions promote fluorogenicity of SiR-BACE1
41
Both binding of small molecule inhibitors to the active site pocket of BACE1 and SiR647
protonation for the fluorescent onꢀswitch are pHꢀdependent, preferring acidic conditions.
Consequently, we evaluated the effect of acidic pH on SiR-BACE1 binding to BACE1 and its
7
ACS Paragon Plus Environment

Page 9 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
fluorogenicity. In order to have an experimental setup with controlled environment, we chose
a cellꢀfree approach using fluorescence correlation spectroscopy (FCS) as a readꢀout,
analyzing SiR-BACE1 diffusion and its brightness. First, we characterized free SiR-BACE1
diffusion in HBS pH 5 in the absence of BACE1 (Figure 3A). The autoꢀcorrelation curves
from 3 independent experiments over 30x 10 s were fitted to a 1ꢀcomponent 3D translational
model, determining the diffusion time, triplet fraction and relaxation time, number of
molecules, and the structural parameter of the point spread function. The structural parameter
was freely fitted in all FCS experiments and typically ranged from 5 to 7. The diffusion time
of 100 nM SiR-BACE1 was 72 ± 3 µs (mean ± SE), serving as a reference in following
experiments. We next explored how SiR-BACE1 diffusion changes upon adding recombinant
BACE1 extracellular domain (ECD). Onꢀstage, 2 µM BACE1 ECD were added to SiR-
BACE1, instantly followed by timeꢀlapse FCS recording over 500x 10 s. The autoꢀcorrelation
curves of each 10 sꢀinterval were fitted to a 2ꢀcomponent 3D translational model with one
diffusion time being fixed to 72 µs for freely diffusing SiR-BACE1 (see above). The
diffusion time of the second fraction was determined by the fit and was always higher than the
fixed first fraction, indicating a slowly diffusing second population. The fraction of SiR-
BACE1 that belonged to this slow fraction increased over time (Figure 3B) with a t0.5 of
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
13 ± 1 min (mean ± SE). This prompted us to incubate SiR-BACE1 with BACE1 ECD for
30 min before performing endꢀpoint FCS recordings in the following set of experiments to
allow SiR-BACE1 to bind and convert into the fluorescent zwitterꢀionic conformation. The
diffusion time of the slow fraction in Figure 3B equaled 549 ± 3 µs (mean ± SE, N = 3
recordings) and did not change systematically over time (Figure 3C), serving as a reference
for diffusion of BACE1ꢀbound SiR-BACE1.
To evaluate the effect of pH, endꢀpoint FCS was recorded over 30x 10 s from SiR-BACE1 in
the presence of different concentrations of BACE1 ECD in pH 5 or 7.4. Upon binding to
BACE1 ECD, we expected SiR-BACE1 diffusion to decelerate (indicated by longer diffusion
8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
times, see Figure 3D) and SiR-BACE1 fluorescence to increase (indicated by rising count
rates). Autoꢀcorrelation curves (Figure 3D) were parameterized by fitting to the 2ꢀcomponent
3D translational model with the diffusion times being fixed to 72 and 550 µs (see above).
With increasing concentrations of BACE1 ECD, a growing SiR-BACE1 fraction diffused
slowly, i.e. BACE1ꢀdependent (Figure 3D), and the fluorescence signal became more intense.
We first compared the influence of pH on the fraction of slowlyꢀdiffusing SiR-BACE1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Figure 3E). At pH 5, the halfꢀmaximal effect was reached at 27 ± 4 nM BACE1 ECD (mean
± SEM), while at pH 7.4, the apparent EC50 was 87 ± 33 nM. Note that the analysis is biased
towards the slowly diffusing BACE1ꢀbound fluorophores because of the fluorogenic onꢀ
switch. When we focused on photon counts, acidic pH enabled bright fluorescence at lower
BACE1 ECD concentrations than slightly alkaline pH (Figure 3F), with EC50 values of
45 ± 6 nM and 620 ± 130 nM at pH 5 and 7.4, respectively. These results showed that acidic
pH likely supports both binding and fluorogenicity. While the effect on binding was not
pronounced enough to be statistically significant, acidic conditions clearly support the
fluorescent onꢀswitch.
We next evaluated whether a positive effect of acidic pH holds true for staining of BACE1ꢀ
EGFP in HEK293T cells. In a first experiment, we stained living cells with SiR-BACE1 in
pH 5 or pH 7.4 (Figure 3G). The intracellular compartments, presumably transꢀGolgi network
(pH ~6), were stained brightly in both conditions. The plasma membrane, however, where
BACE1 molecules are exposed to our staining buffer’s pH, was labelled by SiR-BACE1 at
pH 5, but not at pH 7.4. To determine whether lacking membrane staining was due to the
absence of binding or to a nonꢀfluorescent conformation of the fluorophore, we performed a
twoꢀstep experiment with fixed HEK293T cells overꢀexpressing hBACE1ꢀEGFP. Cells were
stained with SiR-BACE1 in pH 7.4, washed, and observed in timeꢀlapse imaging for 60 min
in pH 7.4 before the imaging buffer’s pH was adjusted to pH 5, followed by another 60ꢀmin
observation (Figure 3H). The change of buffer pH was followed by a decrease of the EGFP
9
ACS Paragon Plus Environment

Page 11 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
fluorescence signal, while in the SiR-BACE1 channel, fluorescence increased and plasma
membranes of the transfected cells became visible. This observation indicates that SiR-
BACE1 can be bound at pH 7.4 in its fluorescent offꢀstate, and can switch to its fluorescent
conformation upon pH change.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
To further discriminate between pHꢀdependent binding and pHꢀdependent fluorescent
switching, we performed an additional cellꢀfree experiment. Binding of SiR-BACE1 to
BACE1 was examined with the FRETꢀbased BACE1 activity assay (see Figure 2C),
performed in HBS of pH 5, 7.4, or 8 instead of the supplied assay buffer. Processing of the
BACE1 substrate was controlled by a kinetic assay run, as suggested by the manufacturer. In
an endpoint assay, activity of BACE1, indicated by the intensity of unquenched cleavage
product, was significantly lower in pH 7.4 and 8 when compared to pH 5, as expected (Figure
42,43
3I, upper panel).
For all three conditions, we determined the IC for BACE1 inhibition by
50
SiR-BACE1 as illustrated in Figure 2C and found that inhibition was not significantly
changed by pH (Figure 3I, lower panel).
Cellular affinity and specificity
Applying acidic conditions for staining, we quantified the affinity of SiR-BACE1 for BACE1
in a cellular environment. HEK293T cells overꢀexpressing hBACE1ꢀEGFP were stained with
different concentrations of SiR-BACE1, ranging from 1 to 5000 nM, in pH 5. Flow cytometry
allowed us to correlate intensities of hBACE1ꢀEGFP and SiR647 signals from individual cells
(Figure 4A). The slope of the correlation curve indicates how efficiently SiR-BACE1 has
stained the hBACE1ꢀEGFP molecules. Figure 4A depicts the distribution of cells, which were
stained with 1 (grey dots), 100 (green dots), or 1000 nM SiR-BACE1 (red dots). At low
EGFP channel intensities (i.e. in nonꢀtransfected cells), no correlation with SiR-BACE1
signal was observed. For higher hBACE1ꢀEGFP intensities (i.e. transfected cells), the SiR-
BACE1 signal positively correlated with the expression level of hBACE1ꢀEGFP. A bilinear
10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
custom fit with a constant first and ascending proportional second part appeared suitable to
parameterize this distribution. A small fraction of cells with a very high intensity in the EGFP
channel deviated from the correlation towards higher SiR-BACE1 signals. Since we had
observed that such “hyperꢀexpressing cells” lose their membrane integrity, this population
was not considered.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Increasing SiR-BACE1 concentration lead to a systematic change of the cells’ fluorescence
distribution, represented by the bilinear fit (Figure 4A): The level of the first component stays
low for staining with 100 nM, but rises for staining with 1000 nM of SiR-BACE1. We
interpreted an increased level as an unspecific accumulation of SiR-BACE1 in all cells,
irrespective of the presence of hBACE1ꢀEGFP. The slope of the second, proportional
component of the fit increased with increasing SiR-BACE1 concentrations (cf. 1 to 100 nM),
indicating enhanced probability for a hBACE1ꢀEGFP molecule to be stained by SiR-BACE1.
Upon further increase of SiR-BACE1 concentration (cf. 100 to 1000 nM), the slope hardly
increased further, indicating saturation of hBACE1ꢀEGFP. The proportionality factor was
plotted as a function of SiR-BACE1 concentration to illustrate specific staining (Figure 4B).
The EC50 of SiR-BACE1 for BACE1 staining was determined to be 78 ± 18 nM. In this
concentration range of SiR-BACE1 (see green dataset in Figure 4A), the constant part of the
fit did not indicate relevant unspecific accumulation of SiR-BACE1 in nonꢀtransfected cells,
demonstrating that specific staining is feasible in HEK293T cells.
In this experimental system, specific BACE1 staining could be unambiguously confirmed by
a competition experiment against the benchmark small molecule BACE1 inhibitor IV
44
(
Merck), also known as C3 (IC = 15 nM). HEK293T cells expressing hBACE1ꢀEGFP
50
were stained with 70 nM SiR-BACE1 in the presence of different concentrations of inhibitor
IV. Increasing concentrations of inhibitor IV displaced SiR-BACE1 (Figure 4C) with an IC50
of 55 ± 8 nM, which was determined by a fit according to Equation 1. The successful
11
ACS Paragon Plus Environment

Page 13 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
competition between the two substances argues in favor of the same orthosteric binding site
for inhibitor IV and SiR-BACE1.
In addition, we performed control experiments to confirm that SiR-BACE1 specifically binds
to the BACE1 active site in HEK293T cells (Figure 4D). The proteins of interest were overꢀ
expressed as fusion proteins with EGFP. We stained cells overꢀexpressing the proteolytically
inactive mutant D289N of BACE1, which was expected not to bind SiR-BACE1 due to the
mutated catalytic site. Additionally, we tested whether SiR-BACE1 stains the BACE1
homolog BACE2 or the BACE1ꢀrelated aspartyl protease cathepsinꢀD. Indeed, none of these
three proteins evoked significant staining by SiR-BACE1.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
SiR-BACE1 is STEDable
With this new tool in our hands, we explored several feasible applications that may profit
from SiR-BACE1’s unique feature of being a fluorogenic, specific, and tagꢀindependent label
for BACE1. As first application, we used SiR-BACE1 for superꢀresolution imaging of overꢀ
expressed BACE1 in the neuronal cell line N1Eꢀ115. The probe reproduced the pattern of
BACE1 indicated by its EGFP tag (Figure 5AꢀB). We scanned neurite details in confocal and
stimulated emission depletion (STED) mode (Figure 5C). BACE1, visualized by SiR-
BACE1, was present in the soma and in neurites. Unlike the confocal recording, the STED
image resolved the vesicular distribution of the protein. For comparison, we plotted
representative profiles from a confocal versus a STED image for a cross profile (Figure 5C,
green box and graph) and neighboring vesicular structures (red box and graph), which
demonstrate the gain in resolution by using SiR-BACE1 for STED microscopy.
12
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Visualization of endogenous BACE1
Having applied SiR-BACE1 successfully on samples overꢀexpressing EGFPꢀtagged BACE1,
we next used it to visualize endogenous BACE1. As a sample, we chose native hippocampal
brain slices from wildꢀtype and BACE1 knockout (KO) mice. SiR-BACE1 stained the hilus
of the dentate gyrus and the mossy fiber terminals in the CA3 region (Figure 6A and B, left
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
24,45,46
panels), consistent with the BACE1 expression pattern described before.
In slices from
knockout animals, no equivalent BACE1 staining was observed (Figure 6A, right panel),
confirming specificity of SiR-BACE1 for inꢀsituꢀdetection of endogenous BACE1. To further
support the finding that SiR-BACE1 does indeed specifically detect BACE1, we performed a
competition experiment with BACE1 inhibitor IV (see above). At 20ꢀfold excess, inhibitor IV
displaced most of the SiR-BACE1 from the active site and suppressed BACE1 staining
(Figure 6B, right panel).
We also tested SiR-BACE1 for staining of living hippocampal neuronal cultures, in which the
probe provoked intense staining of somatic and neuritic structures, both in wildꢀtype and
knockout preparations (Figure 6C). To elucidate the origin of these accumulations, we coꢀ
stained mouse embryonic fibroblasts (MEF) from BACE1 knockout mice with SiR-BACE1
and LysoTracker. Coꢀlocalization implied SiR-BACE1 accumulations in acidic lysosomes
(Figure 6D, upper panel). We asked whether SiR-BACE1 accumulated because of an offꢀ
target present in lysosomes or due to the acidic pH of the compartment. With bafilomycin A1
(BafA1), we disrupted lysosomal acidification. Disappearance of the LysoTracker signal
confirmed neutralization. With BafA1 treatment, lysosomal SiR-BACE1 staining was
abolished (Figure 6D, lower panel), arguing for a pHꢀdependent accumulation. A preꢀ
treatment of the primary neuronal cultures with BafA1 to establish a specific staining was not
considered because control experiments with BafA1 in MEF cells had shown that BACE1ꢀ
EGFP, as well as additional transmembrane and membraneꢀassociated proteins, lost its typical
expression pattern. We concluded that SiR-BACE1 is not suited for staining of primary
13
ACS Paragon Plus Environment

Page 15 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hippocampal neurons under the tested circumstances, but well applicable for BACE1
detection in frozen sections, which do no longer contain acidified vesicles (Figure 6A,B).
Single particles tracking reveals confined membrane diffusion of BACE1
Finally, we wanted to investigate the behavior of single BACE1 molecules at the cell surface.
The diffusion and trafficking dynamics of BACE1 have been recently implicated in the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5,47,48
pathogenesis of Alzheimer´s disease.
Therefore, BACE1 surface movement is getting
increasing attention, but dynamic data are not available yet. The development of SiR-BACE1
offers now the unprecedented opportunity to track single BACE1 molecules without the
hindrance of tags or huge antibody labels.
The experiments with neuronal cultures and MEF cells (Figure 6) taught us that
accumulations of SiR-BACE1 in acidic intracellular compartments may impair BACE1
visualization when it comes to detection of low endogenous levels – or single molecules. We
4
9
therefore worked with plasma membrane lawns (PML) of CHOꢀK1 cells, with intracellular
organelles being removed to achieve the best possible signalꢀtoꢀbackground ratio.
Wildꢀtype human BACE1 was transiently expressed at low levels and stained with SiR-
BACE1, resulting in spotꢀlike, diffractionꢀlimited signals of single BACE1 molecules (Figure
7A, right panel). Unstained transfected PML and nonꢀtransfected stained PML did not show
relevant autofluorescence or unspecific staining, respectively (Figure 7A, left and middle
panel). Tracks of BACE1 diffusion (Figure 7B) were generated from timeꢀlapse recordings of
6
PML regions from 4 cells. Track fluorescence intensity was averaged over time and plotted
for all tracks (Figure 7C). Intensities followed a biꢀmodal Gaussian distribution. 50% of the
tracks showed a relative intensity of 6.6 ± 2.4 AU (mean ± SE). The intensity of the
remaining tracks was 12 ± 5 AU. The velocity of the molecules did not correlate with their
intensity (Figure 7D). Tracks were typically shorter than 1ꢀ2 s (Figure 7E). Mobility of the
tracked molecules was independent from track lifetime (Figure 7F). Therefore, the track
14
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
population was assumed to be homogenous regarding diffusion behavior. Plotting the mean
square displacement (MSD) over time revealed a confined diffusion of the BACE1 molecules,
indicated by the negative deviation of the data from a linear diagonal (Figure 7G, black dots).
50
As described previously, the mobility distribution (Figure 7H) did not fit to a γꢀdistribution,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
thereby excluding free diffusion. 15% of the tracks, i.e. 220 tracks, were immobile (defined
5
0
by a mobility < 0.02 µm²/s) . When we excluded these from the analysis, the mobile BACE1
fraction still showed restricted movement (Figure 7G, green dots).
To validate that the observed traces originated from single SiR-BACE1 molecules, PMLs
were fixed to immobilize the proteins and then stained with SiR-BACE1. The presence of a
PML was confirmed by an area showing smooth fluorescence of membraneꢀtargeted EGFP
(Figure 7I). Expression of SNAPꢀtagged hBACE1 was confirmed by staining with the SNAP
substrate SNAPꢀsurface549. SiR-BACE1 staining resulted in distinct BACE1 signals
comparable to the SiR-BACE1 pattern in live PML (Figure 7I, cf. A, right panel). The signals
had been efficiently immobilized by the fixation (cf. 0 s vs. 10 s in Figure 7J) and typically
disappeared from one frame to the other in oneꢀstep bleaching, confirming the presence of
single fluorophores (Figure 7J, arrows).
15
ACS Paragon Plus Environment

Page 17 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discussion and Conclusions
The aim of this study was to design and evaluate a small molecule BACE1 labeling tool that
fulfils the following criteria: It marks the protease for fluorescence imaging without genetic
manipulation or antibodies, it is easy to apply, and it is compatible with a broad range of
imaging systems. We succeeded in developing the fluorogenic probe SiR-BACE1 as a
conjugate of the small molecule BACE1 inhibitor Sꢀ39 for selective targeting and the farꢀred
dye SiR647 (Figure 1). Following a oneꢀstep staining procedure, SiR-BACE1 can visualize
recombinant BACE1 in solution (Figure 2CꢀD, Figure 3AꢀF), overꢀexpressed BACE1 (Figure
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
AꢀB, Figure 3GꢀH, Figure 4, Figure 5, Figure 7), and endogenous BACE1 (Figure 6). SiR-
BACE1 staining rendered BACE1 accessible for confocal laser scanning microscopy (Figure
A, Figure 6), superꢀresolution imaging by Stimulated Emission Depletion (STED, Figure 5),
2
and single molecule timeꢀlapse analysis (Figure 7).
Choice of a linker
A key aspect in the synthetic strategy was the lateꢀstage coupling of the pharmacophore to the
fluorophore, which allows easy modulation of the derivatives’ linker. We selected a linker
length of 4 methylene spacers (C4), which proved to be critical for sensitive labeling.
Conjugates with shorter linker had a slightly lower affinity for the BACE1 active site (Figure
2D). Longer linkers, like in the C8 compound, did allow undisturbed binding to and inhibition
of BACE1, but resulted in low fluorescence, which likely was a consequence of insufficient
fluorescent conversion of bound conjugate from its nonꢀfluorescent spirolactone form to the
fluorescent zwitterꢀion. Overall, the C4 compound, which we termed SiR-BACE1, allowed
binding of the inhibitor with low steric hindrance by SiR647 and brought the fluorophore
close enough to the protein surface to promote its fluorogenic onꢀswitch.
16
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
pH dependency
5
1
In the endosomal pathway, pH decreases from a slightly alkaline extracellular pH to pH 5ꢀ6.
39
Based on the chemistry of SiR647, the acidic pH should support labeling by enforcing the
fluorescent onꢀswitch of the dye. Indeed, that effect was clearly observed for BACE1 in
solution (Figure 3F) as well as for overꢀexpressed BACE1 in cells (Figure 3H).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In addition, BACE1 active site residues have been shown to change their hydration state
4
2,43
depending on pH, which results in the enzymatic pH optimum of 4.5.
Like substrate
cleavage, binding of small molecule inhibitors into the active site pocket has been shown to
41,52
depend on the charges of the respective side chains.
Previous work has pointed out that
the pH profile of small molecule binding to BACE1 is dependent on how the small molecule
5
3
interacts with BACE1. We checked for pH dependency by cellꢀfree FCS and found a
tendency for enhanced binding at pH 5 (Figure 3E). Superior staining of overꢀexpressed
BACE1 in pH 5, as demonstrated in Figure 3G, is mainly attributable to a promoted
fluorogenic onꢀswitch.
Acidic accumulation
Inhibitory potency of SiR-BACE1 was 100ꢀfold lower than that of the parent drug Sꢀ39 in a
3
7
cellꢀfree enzymatic assay (IC Sꢀ39: 10 nM). Only a factor 2 can be attributed to the fact
5
0
that SiR-BACE1 was obtained as a racemate. We conclude that the conjugation of the drug to
SiR647 diminishes its potency of binding to the BACE1 active site pocket, an effect that has
5
4
been observed before. Arguing for steric hindrance is supported by the finding that short
linker lengths, which bring SiR647 close to the pharmacophore, slightly decrease BACE1
inhibition.
When compared to the inhibitory potency in the cellꢀfree FRET assay, the effective
concentration for staining BACE1 in the cellular environment was 14ꢀfold lower with an EC50
17
ACS Paragon Plus Environment

Page 19 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of 78 ± 18 nM. Such an observation is not unprecedented: A study that evaluated a small
55
molecule labeling tool for cathepsinꢀD, another aspartyl protease, revealed a similar effect.
The compound in that study accumulated in acidic compartments and thereby reached an
increased affinity for cathepsinꢀD in cells versus inꢀvitroꢀassays by one to two orders of
magnitude. Considering such a mechanism for SiR-BACE1, it is important to note that
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
9
SiR647 changes its conformation in acidic pH. Acidic trapping of SiR-BACE1 therefore
seemed possible. Support for this hypothesis came from the staining of primary hippocampal
neurons by SiR-BACE1: When we analyzed those for endogenous BACE1 staining,
prominent unspecific SiR-BACE1 accumulations occurred in an irregular pattern and
impeded BACE1 visualization in this sample (Figure 6C). Further characterization in BACE1
knockout MEF cells confirmed that SiR-BACE1 is entrapped into acidic lysosomes. This
local accumulation can fully explain the higher apparent affinity of SiR-BACE1 in the
cellular system when compared to the cellꢀfree FRET assay.
Given the highꢀaffinity EC50 in cellular staining, it may seem surprising that only a minor
fraction of BACE1 in plasma membrane lawns (PMLs) was stained by 100 nM SiR-BACE1
(Figure 7I). However, the staining situation rather resembles the cellꢀfree assay than the
cellular staining with locally increased SiR-BACE1 concentrations. In the homogenous
staining solution above PMLs, SiR-BACE1 does not accumulate and may therefore not be
able to saturate all present BACE1 molecules in the preparation. An alternative explanation
for nonꢀsaturated staining in PMLs could be that SiR-BACE1 may prefer binding to a certain
multimeric form of BACE1 that may occur predominantly intracellularly. However, that
seems unlikely since recent work has identified BACE1 multimers in an equal distribution
56
between subcellular compartments.
18
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
SiR-BACE1 for confocal imaging and nanoscopy
When a protein of interest is labelled using primary and dyeꢀcoupled secondary antibodies,
one target molecule is estimated to be stained by approximately 25 fluorophores yielding a
bright signal. In superꢀresolution imaging, however, this apparent advantage due to “antibody
piling” has the adverse effect of separating the fluorophores by approximately 20 nm from the
target protein. Such spacing is unacceptable for techniques that offer spatial resolution in the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
7
nanometer or even Angstrom scale, like the novel extension of STORM imaging at cryo
5
8
temperatures (Cryogenic Optical Localization in 3D, COLD) . A superꢀresolution label
hence is a tradeꢀoff between photon yield and label size. In addition, ratiometric analysis and
subꢀnanometer structural investigations require a 1ꢀtoꢀ1 labeling. SiR-BACE1 meets these
diverging criteria by i) depositing one fluorophore per BACE1 molecule, ii) dragging the dye
into the active site pocket, resulting in the minimal possible distance between BACE1 and the
label, and iii) containing SiR647, which offers a high photon yield and excellent photoꢀ
3
9
stability . The latter compensates partially for the low number of deposited fluorophores.
Owing to these features, SiR-BACE1 allowed STED imaging, which required long exposure
times because of low expression level and fluorophore concentration. In addition,
visualization of endogenous BACE1 in native hippocampal slices, where fluorophore
concentration is naturally limited, (Figure 6A,B) similarly profited from SiR-BACE1’s
features. Both measurements would have been impossible with a lowꢀyield or fastꢀbleaching
fluorophore. Judging from these applications, we can advocate SiR-BACE1 for superꢀ
resolution, lowꢀlight, and single molecule applications.
Single molecule BACE1 diffusion analysis
Label size is not just a factor when discussing the distance to the target protein, but also
matters because of its sheer mass and diameter: Analysis of single molecules in the plasma
19
ACS Paragon Plus Environment

Page 21 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
0,59,60
membrane has mostly been done with fluorescent protein fusion constructs,
antibodies
61
62,63
labelled with quantum dots, or selfꢀlabeling enzymes like the SNAP tag.
Since these
labels can reach the same size as the analytes or even exceed their dimensions, they may
confound optical recordings in several aspects. For instance, SNAP tag has been shown to
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
63
slow the movement of a protein of interest and antibodies are commonly used to even
64
immobilize proteins at the cell surface . Additionally, tags interfere with proteinꢀprotein
interactions. Incorporation of fluorescent unnatural amino acids has minimized tag size but
6
5
requires genetic manipulation of the target protein. Interferometric detection of scattering
(iSCAT) allows observation of molecules without any label but is not yet available to the
66,67
broad life science community and is currently limited to the detection of secreted proteins.
Thus, single molecule tracking of membrane proteins still requires a label, which should be as
6
3,65,68
small as possible, such as fluorophore conjugates of receptor agonists or antagonists.
These small molecules are fabricated against specific targets and, with SiR-BACE1, we
introduce the first such label for BACE1.
Employing this imaging tool, we are now able to probe the motility of BACE1 in cellular
membranes. Our data argue for a confined diffusion behavior of the secretase, which was not
dependent on diffusion coefficient, track length, or intensity. This finding strongly suggests
that BACE1 molecules form a homogenous population in the plasma membrane that does not
diffuse freely. We are aware that our characterization of BACE1 diffusion may be cell typeꢀ
5
0
specific, as reported for other transmembrane proteins. The observed confined diffusion may
be a result of proteinꢀprotein interactions (as discussed below for ion channels), anchoring to
the cytoskeleton, lipid microdomains, or spatial partitioning of the plasma membrane, as
69–73
recently demonstrated for the neuronal surface.
We have previously revealed the interaction of BACE1 with the voltageꢀgated sodium
channel Na
V V
1.2, voltageꢀgated potassium channel K 3.4, and potassium channels of the
KCNQ family (KCNQ1ꢀ3), for which BACE1 acts as a nonꢀproteolytic accessory
20
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
2,74ꢀ77
subunit.
Both Na 1.2 and KCNQ2/3 are anchored to the actin cytoskeleton at the axon
V
78
initial segment by interaction with ankyrinꢀG. A similar anchoring to actin has been
50,79
described for KCNQ1 via the protein kinase A anchoring protein (AKAP) yotiao.
It is
therefore conceivable that the restricted motility of BACE1 results, at least in part, from its
physical association with anchored membrane proteins. In future studies, tracking BACE1
diffusion with SiR-BACE1 might offer a means to identify putative new interaction partners
and regulators of the secretase.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Membrane composition and protein surface diffusion has garnered increasing attention in AD
8
0
research. For instance, increased cholesterol levels support encounters of BACE1 with APP
8
1,82
in membrane subdomains, facilitating joint endocytosis and APP processing.
Vice versa,
membrane partitioning can preclude proteaseꢀsubstrate interaction, as demonstrated for
8
3
ADAM10 and its chemokine substrate CX3CL1. This raises the question of whether the
efficacy of BACE1 inhibitors depends on membrane composition and subꢀcellular BACE1
localization. In a successful attempt to optimize inhibition, the pharmacophore was directed
towards endosomes, where BACE1 is located, active, and in a protonation state favoring drug
8
4,85
binding.
Just as well, membrane partitioning and protein confinement could determine an
inhibitor’s potency already at the cell surface. Remarkably, clustering of target molecules has
been discussed to allow enzymatic reaction bursts despite the presence of high doses of
86
inhibitors.
In conclusion, we have introduced SiR-BACE1 as new tool for live cell imaging of BACE1.
The probe should become valuable for monitoring the distribution and dynamics of BACE1 in
cellular membranes and the development of AD therapeutics that aim to either modulate the
membrane to prevent BACE1ꢀAPP interaction or target inhibitors to the protease.
21
ACS Paragon Plus Environment

Page 23 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Experimental Section
Chemistry
All compounds used in the results section (13a-f) were determined to be >95% pure by
analytical HPLC. All chemical reagents and anhydrous solvents for synthesis were purchased
from commercial suppliers (SigmaꢀAldrich, Fluka, Acros, Fluorochem, TCI) and were used
without further purification or distillation. If necessary, solvents were degassed either by
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
freezeꢀpumpꢀthaw or by bubbling N
minutes.
2
through the vigorously stirred solution for several
NMR spectra were recorded in deuterated solvents on a BRUKER DPX 400 or on a
1
13
BRUKER DPX 400 instruments and calibrated to residual solvent peaks ( H/ C in ppm):
CDCl (7.26/77.00), DMSOꢀd (2.50/39.52), acetoneꢀd (2.05). Multiplicities are abbreviated
3
6
6
as follows: s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, br = broad, m =
multiplet. Coupling constants J are reported in Hz. Spectra are reported based on appearance,
not on theoretical multiplicities derived from structural information.
Highꢀresolution mass spectra (HRMS) were measured on a Micromass QꢀTOF Ultima
spectrometer with electrospray ionization (ESI).
LCꢀMS was performed on a Shimadzu MS2020 connected to a Nexerra UHPLC system
equipped with a Waters ACQUITY UPLC BEH C18 1.7 ꢁm 2.1x50 mm column. Buffer A:
2
0.05% HCOOH in H O Buffer B: 0.05% HCOOH in acetonitrile. Analytical gradient was
from 10% to 90% B within 5.0 min with 0.5 or 1.0 mL/min flow.
Preparative RPꢀHPLC was performed on a Dionex system equipped with an UVD 170U UVꢀ
Vis detector for product visualization on a Waters SunFireTM Prep C18 OBDTM 5 ꢁm
2
10×150 mm column. Buffer A: 0.1% TFA in H O Buffer B: acetonitrile. Typical gradient was
from 10% to 90% B within 32 min with 4 mL/min flow. After lyophilization of HPLCꢀ
purified compounds, the solid residue was generally dissolved in dry DMSO and the
22
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
concentration of the SiRꢀderivatives was measured by UVꢀVis spectroscopy in PBS
ꢀ1
ꢀ1
containing 0.1% SDS, using the SiR molar extinction coefficient of 100,000 M cm at 650
nm.
Flash column chromatography was performed on a Teledyne ISCO CombiFlash® Rf+ with
preꢀpacked silica columns or manually on silica gel (SilicaDlash® P60, 0.040–0.063 mm,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
230ꢀ400 mesh, Silicycle). Reactions and chromatography fractions were monitored by thin
layer chromatography (TLC) on Merck silica gel 60 F254 glass plates. The spots were
visualized either under UV light at 254 nm and/or 366 nm.
4-Bromo-2,6-diethylpyridine (Figure 1C, 3)
40
4ꢀBromoꢀ2,6ꢀdiethylpyridine was prepared as previously described
from 1,3ꢀ
acetonedicarboxylic acid (1) and propionic anhydride (2) over five steps and obtained as a
yellow to colorless light oil in 0.8% yield.
1
3
H NMR (400 MHz, CDCl ): δ [ppm] = 7.15 (s, 2H), 2.76 (q, J = 7.6 Hz, 4H), 1.28 (t, J = 7.6
Hz, 6H).
13
C NMR (101 MHz, CDCl
3
): δ [ppm] = 164.4, 133.3, 122.3, 31.3, 13.9.
+
HRMS (ESI): calc. for C H BrN [M+H] : 214.0226 and 216.0205, found: 214.0230 and
9
13
216.0211.
R
f
(TLC; hexanes/DCM = 1/1) = 0.42.
3,7-Bis(dimethylamino)-N-(3-guanidinoalkyl)-5,5-dimethyl-3'-oxo-3'H,5H-
spiro[dibenzo[b,e]siline-10,1'-isobenzofuran]-6'-carboxamide (5a-f)
General procedure: A 2 mL vial was charged with SiRꢀ6ꢀCOOH 4 (1.0 equiv.) in DMSO (1
mL / 10 mg SiRꢀ6ꢀCOOH) and DIPEA (4.0 equiv.) and TSTU (1.2 equiv.) was added in one
portion. Upon addition, the mixture turned from blue to colorless (DIPEA) to yellow (TSTU).
After stirring at rt for 5 min, 1,nꢀdiaminoalkane (10 equiv.) was added (liquid diamines were
23
ACS Paragon Plus Environment

Page 25 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pipetted directly into the solution while solids were dissolved in a minimal amount of DMSO
and then added in one portion) and the reaction mixture was vigorously stirred for 1 h (turning
colorless) it was quenched with HOAc (25 equiv.) and 0.1% aqueous TFA (200 µL / 10 mg
SiRꢀ6ꢀCOOH) and the crude subjected to RPꢀHPLC (MeCN/H
2
O/TFA = 10/90/0.1 ꢀ
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
90/10/0.1 over 32 minutes). The desired monoamine was collected (λ = 650 nm), freezeꢀdried
and the obtained blue powder was dissolved in DMF (1 mL / 10 mg SiRꢀ6ꢀCOOH) and NEt3
(15 equiv.) before 1Hꢀpyrazoleꢀ1ꢀcarboximidamide hydrochloride (10 equiv.) was added in
one portion. The reaction mixture was vigorously stirred for 2 h before it was quenched with
HOAc (20 equiv.) and the crude subjected to RPꢀHPLC (MeCN/H
2
O/TFA = 10/90/0.1 ꢀ
90/10/0.1 over 32 minutes). The obtained guanidine was freezeꢀdried to obtain the desired
product as a blue powder and briefly characterized by LCꢀMS analysis (MeCN/H O/formic
2
acid = 10/90/0.1 ꢀ 100/10/0.1 over 6 min).
3,7ꢀBis(dimethylamino)ꢀNꢀ(3ꢀguanidinobutyl)ꢀ5,5ꢀdimethylꢀ3'ꢀoxoꢀ3'H,5Hꢀ
spiro[dibenzo[b,e]silineꢀ10,1'ꢀisobenzofuran]ꢀ6'ꢀcarboxamide (n = 4, 5c) was obtained in 85%
1
yield over 2 steps and subjected to further H NMR and HRMS characterization:
1
H NMR (400 MHz, acetoneꢀd ): δ [ppm] = 8.47 (s, 1H), 8.24 (s, 1H), 8.15 (dd, J = 10.3, 7.6
6
Hz, 1H), 8.06–7.91 (m, 1H), 7.82–7.74 (m, 1H), 7.49 (br s, 3H), 7.20–7.11 (m, 2H), 6.83–
6.74 (m, 2H), 6.71–6.65 (m, 2H), 3.43 (t, J = 5.3 Hz, 2H), 3.40–3.28 (m, 2H), 2.99 (s, 12H),
1.70–1.66 (m, 4H), 0.68 (s, 3H), 0.55 (s, 3H).
+
41 6 3
HRMS (ESI): calc. for C32H N O Si [M+H] : 585.3004, found: 585.3004.
3
-(3-Bromophenyl)-2-fluoropyridine (8)
A Schlenk flask was charged with (2ꢀfluoropyridinꢀ3ꢀyl)boronic acid (6) (249 mg, 1.77 mmol,
.0 equiv.), 1ꢀbromoꢀ3ꢀiodobenzene (7) (500 mg, 1.77 mmol, 1.0 equiv.) and Cs CO (577
mg, 1.77 mmol, 1.0 equiv.) suspended in DME / H O (8 mL / 2 mL). The suspension was
degassed by bubbling with nitrogen for 10 min under stirring before Pd(PPh ) (204 mg,
1
2
3
2
3
4
24
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
.177 mmol, 0.1 equiv.) was added in one portion. The flask was sealed and the reaction
mixture was heated to 85 °C to become homogenous and stirred for 18 h before it was
allowed to cool to rt and then quenched by the addition of water (100 mL) and extracted with
EtOAc (2 x 100 mL). The combined organic layers were washed with water (100 mL) and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
brine (100 mL), dried over MgSO and the volatiles were removed in vacuo. The crude was
taken up in DCM and hexanes was added to precipitate the majority of PPh that was filtered
3
off before the solution was subjected to FCC (80 g silica: 3 CV 100% hexanes ꢀ 10 CV
gradient to 100% DCM ꢀ 1 CV 100% DCM; the product eluted ~90% DCM) to obtain 88.0
mg (0.350 mmol) of the desired product as a clear oil in 20% yield.
1
3
H NMR (400 MHz, CDCl ): δ [ppm] = 8.20 (dt, J = 4.9, 1.6 Hz, 1H), 7.83 (ddd, J = 9.6, 7.4,
2
.0 Hz, 1H), 7.69 (d, J = 1.7 Hz, 1H), 7.52 (ddd, J = 8.0, 2.0, 1.0 Hz, 1H), 7.50–7.46 (m, 1H),
.32 (t, J = 7.9 Hz, 1H), 7.27 (ddd, J = 7.5, 4.9, 1.8 Hz, 1H).
7
1
3
3
C NMR (101 MHz, CDCl ): δ [ppm] = 160.1 (d, J = 240.5 Hz), 146.9 (d, J = 14.7 Hz),
140.5 (d, J = 4.1 Hz), 135.8 (d, J = 5.1 Hz), 131.6 (d, J = 3.1 Hz), 131.4, 130.1, 127.4 (d, J =
3.3 Hz), 122.6, 122.3 (d, J = 28.3 Hz), 121.9 (d, J = 4.4 Hz).
1
9
F NMR (376 MHz, CDCl ): δ [ppm] = –70.73 (d, J = 9.7 Hz).
3
+
HRMS (ESI): calc. for C H BrFN [M+H] : 251.9819 and 253.9798, found: 251.9823 and
11
8
2
53.9804.
UV/Vis (LCMS): λmax1 = 226 nm; λmax2 = 268 nm.
(LCMS; MeCN/H O/formic acid = 10/90/0.1 ꢀ 90/10/0.1 over 6 min) = 3.900 min.
R (TLC; hexanes/DCM = 1/1) = 0.41.
t
R
2
f
2
-Fluoro-3-(3-((trimethylsilyl)ethynyl)phenyl)pyridine (9)
A Schlenk flask was charged with 3ꢀ(3ꢀbromophenyl)ꢀ2ꢀfluoropyridine (8) (302 mg, 1.20
mmol, 1.0 equiv.), TMSꢀacetylene (141 mg, 1.44 mmol, 205 µL, 1.2 equiv.), CuI (2.2 mg, 12
2 3 2
µmol, 0.01 equiv.) and PdCl (PPh ) (25 mg, 36 µmol, 0.03 equiv.) dissolved in degassed
25
ACS Paragon Plus Environment

Page 27 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
DMF / NEt (5 mL / 1 mL) under a nitrogen atmosphere. The reaction mixture was heated to
3
65 °C o.n. before it was allowed to cool to rt and quenched by the addition of water (50 mL)
and extracted with EtOAc (50 mL). The organic layer was washed with water (2 x 100 mL)
and brine (100 mL), dried over MgSO
was subjected to FCC (40 g silica: 2 CV 100% hexanes ꢀ 10 CV gradient to 100% DCM ꢀ
CV 100% DCM; the product eluted ~85% DCM) to obtain 251 mg (0.93 mmol) of the
4
and the solvents were removed in vacuo. The crude
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
desired product as a yellow oil in 78% yield.
1
3
H NMR (400 MHz, CDCl ): δ [ppm] = 8.18 (ddd, J = 4.9, 1.9, 1.1 Hz, 1H), 7.83 (ddd, J =
9.6, 7.4, 2.0 Hz, 1H), 7.64 (q, J = 1.6 Hz, 1H), 7.51–7.47 (m, 2H), 7.42–7.35 (m, 1H), 7.29–
.21 (m, 1H), 0.25 (s, 9H).
7
13
C NMR (101 MHz, CDCl ): δ [ppm] = 160.2 (d, J = 240.3 Hz), 146.6 (d, J = 14.7 Hz),
3
140.6 (d, J = 4.2 Hz), 133.9 (d, J = 5.0 Hz), 132.1 (d, J = 2.6 Hz), 131.7, 128.8 (d, J = 3.3
Hz), 128.6, 122.9 (d, J = 28.3 Hz), 121.8 (d, J = 4.5 Hz), 104.3, 94.9, –0.2.
1
9
F NMR (376 MHz, CDCl
3
): δ [ppm] = –70.9 (d, J = 9.8 Hz).
+
HRMS (ESI): calc. for C16
H
17FNSi [M+H] : 270.1109, found: 270.1111.
UV/Vis (LCMS): λmax1 = 244 nm; λmax2 = 257 nm.
t (LCMS; MeCN/H O/formic acid = 10/90/0.1 ꢀ 90/10/0.1 over 6 min) = 4.897 min.
R
2
R
f
(TLC; hexanes/DCM = 1/1) = 0.45.
3-(3-Ethynylphenyl)-2-fluoropyridine (10)
A round bottom flask was charged with 2ꢀfluoroꢀ3ꢀ(3ꢀ((trimethylsilyl)ethynyl)phenyl)pyridine
(
9) (251 mg, 0.93 mmol, 1.0 equiv.) and Cs CO (363 mg, 1.12 mmol, 1.2 equiv.) dissolved in
2 3
EtOH / DCM (10 mL / 10 mL) and stirred at rt for 30 min. The volatiles were removed in
vacuo and the crude was filtered through a plug of silica with DCM to obtain 180 mg (0.91
mmol) of the desired product as an orangeꢀbrown oil in 98% yield.
26
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
H NMR (400 MHz, CDCl ): δ [ppm] = 8.18 (dt, J = 4.9, 1.6 Hz, 1H), 7.83 (ddd, J = 9.6, 7.4,
3
2.0 Hz, 1H), 7.66 (q, J = 1.6 Hz, 1H), 7.54–7.49 (m, 2H), 7.40 (t, J = 7.8 Hz, 1H), 7.31–7.20
m, 1H), 3.11 (s, 1H).
(
13
C NMR (101 MHz, CDCl
40.5 (d, J = 4.3 Hz), 134.0 (d, J = 5.0 Hz), 132.2 (d, J = 2.9 Hz), 131.9, 129.1 (d, J = 3.3
Hz), 128.7, 122.9, 122.6, 121.8 (d, J = 4.5 Hz), 82.9, 77.8.
3
): δ [ppm] = 160.2 (d, J = 240.4 Hz), 146.6 (d, J = 14.7 Hz),
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
9
F NMR (376 MHz, CDCl ): δ [ppm] = –70.9 (d, J = 9.7 Hz).
3
+
HRMS (ESI): calc. for C13
H
9
FN [M+H] : 198.0714, found: 198.0717.
UV/Vis (LCMS): λmax1 = 229 nm; λmax2 = 268 nm.
(LCMS; MeCN/H O/formic acid = 10/90/0.1 ꢀ 90/10/0.1 over 6 min) = 3.618 min.
t
R
2
R (TLC; hexanes/DCM = 1/1) = 0.39.
f
2,6-Diethyl-4-((3-(2-fluoropyridin-3-yl)phenyl)ethynyl)pyridine (11)
A Schlenk flask was charged with 3ꢀ(3ꢀethynylphenyl)ꢀ2ꢀfluoropyridine (10) (180 mg, 0.91
mmol, 1.0 equiv.), 4ꢀbromoꢀ2,6ꢀdiethylpyridine (196 mg, 0.91 mmol, 1.0 equiv.), CuI (1.7
2 3 2
mg, 9.1 µmol, 0.01 equiv.) and PdCl (PPh ) (19.2 mg, 27.3 µmol, 0.03 equiv.) dissolved in
degassed DMF / NEt (5 mL / 1 mL) and flushed with nitrogen. The reaction mixture was
3
heated to 65 °C o.n. before it was allowed to cool to rt and quenched by the addition of water
(50 mL) and extracted with EtOAc (50 mL). The organic layer was washed with water (2 x
4
100 mL) and brine (100 mL), dried over MgSO and the solvents were removed in vacuo. The
crude was subjected to FCC (40 g silica: 20 CV 20% EtOAc/hexanes) to obtain 199 mg (0.60
mmol) of the desired product as a lightꢀyellowish oil in 66% yield.
1
3
H NMR (400 MHz, CDCl ): δ [ppm] = 8.27–8.15 (m, 1H), 7.88 (ddd, J = 9.6, 7.4, 1.9 Hz,
1
H), 7.74 (q, J = 1.6 Hz, 1H), 7.58–7.55 (m, 2H), 7.46 (t, J = 7.7 Hz, 1H), 7.31–7.26 (m, 1H),
.10 (s, 2H), 2.80 (q, J = 7.6 Hz, 4H), 1.30 (t, J = 7.6 Hz, 6H).
7
27
ACS Paragon Plus Environment

Page 29 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
3
C NMR (101 MHz, CDCl ): δ [ppm] = 163.2, 160.3 (d, J = 240.4 Hz), 146.8 (d, J = 14.6
3
Hz), 140.6 (d, J = 4.2 Hz), 134.3 (d, J = 5.0 Hz), 132.1 (d, J = 2.9 Hz), 131.7, 131.5, 129.3 (d,
J = 3.2 Hz), 128.9, 123.1, 122.9 (d, J = 28.2 Hz), 121.9 (d, J = 4.5 Hz), 120.9, 91.8, 88.2,
3
1.4, 14.0.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
19
F NMR (376 MHz, CDCl
3
): δ [ppm] = –70.8 (d, J = 9.8 Hz).
+
HRMS (ESI): calc. for C H FN [M+H] : 331.1605, found: 331.1632.
22
20
2
UV/Vis (LCMS): λmax1 = 225 nm; λmax2 = 277 nm, λmax3 = 302 nm, λmax4 = 321 nm.
t
R
2
(LCMS; MeCN/H O/formic acid = 10/90/0.1 ꢀ 90/10/0.1 over 6 min) = 3.178 min.
R
f
(TLC; hexanes/EtOAc = 8/2) = 0.26.
1
-(2,6-Diethylpyridin-4-yl)-2-(3-(2-fluoropyridin-3-yl)phenyl)ethane-1,2-dione (12)
round bottom flask was charged with 2,6ꢀdiethylꢀ4ꢀ((3ꢀ(2ꢀfluoropyridinꢀ3ꢀ
yl)phenyl)ethynyl)pyridine (11) (199 mg, 0.60 mmol, 10 equiv.) dissolved in acetone (20
mL). NaHCO (30 mg, 0.36 mmol, 0.6 equiv.) and MgSO (108 mg, 0.90 mmol, 1.5 equiv.)
was added to the solution to obtain a slurry, to which KMnO (209 mg, 1.32 mmol, 2.2
A
3
4
4
equiv.) was added spatulaꢀwise at rt. The reaction mixture was warmed to and stirred at 40 °C
for 1 h before it was diluted with water (100 mL) and extracted with Et O (2 x 100 mL). The
2
combined organic layer was washed with water (100 mL) and brine (100 mL), dried over
4
MgSO and the solvents were removed in vacuo. The crude was subjected to FCC (40 g silica:
2 CV 10% EtOAc/hexanes ꢀ 5 CV gradient to 50% EtOAc/hexanes ꢀ 5 CV 50%
EtOAc/hexanes) to obtain 73 mg (0.20 mmol) of the desired product as a yellow wax in 34%
yield.
1
3
H NMR (400 MHz, CDCl ): δ [ppm] = 8.23 (ddd, J = 4.9, 1.9, 1.2 Hz, 1H), 8.17–8.15 (m,
1
H), 7.98 (dt, J = 7.8, 1.4 Hz, 1H), 7.92–7.87 (m, 2H), 7.63 (t, J = 7.8 Hz, 1H), 7.44 (s, 2H),
.30 (ddd, J = 7.4, 4.9, 1.7 Hz, 1H), 2.86 (q, J = 7.6 Hz, 4H), 1.29 (t, J = 7.6 Hz, 6H).
7
28
ACS Paragon Plus Environment