Conjugation of biologically relevant molecules
using SuFEx reagents
Master’s thesis in Material’s Chemistry

MALIN FUHRMAN

DEPARTMENT OF CHEMISTRY AND CHEMICAL ENGINEERING
CHALMERS UNIVERSITY OF TECHNOLOGY
Gothenburg, Sweden 2025
www.chalmers.se

www.chalmers.se


Master’s thesis 2025

Conjugation of biologically relevant molecules
using SuFEx reagents

Malin Fuhrman

Department of Chemistry and Chemical Engineering
Division of Chemistry and Biochemistry

Chalmers University of Technology
Gothenburg, Sweden 2025



Conjugation of biologically relevant molecules using SuFEx reagents:
Malin Fuhrman

© Malin Fuhrman, 2025.

Supervisor: Dr. Maximilian Häfner, AstraZeneca
Examiner: Prof. Jerker Mårtensson, Department of Chemistry and Chemical Engi-
neering

Master’s Thesis 2025
Department of Chemistry and Chemical Engineering
Division of Chemistry and Biochemistry

Chalmers University of Technology
SE-412 96 Gothenburg
Telephone: +46 31 772 1000

Cover: Linking together molecule A and molecule B using SuFEx linkers in a high
throughput format.

Typeset in LATEX
Printed by Chalmers Digital Printing
Gothenburg, Sweden 2025



Conjugation of biologically relevant molecules using SuFEx reagents
Malin Fuhrman
Department of Chemistry and Chemical Engineering
Chalmers University of Technology

Abstract
Sulfur fluoride exchange (SuFEx) chemistry is a type of click chemistry utilizing
the reactivity of the S(VI)-F bond, where nucleophiles can selectively attack the
sulfur center and exchange with fluoride upon activation with e.g., Lewis acids.
SuFEx chemistry has the potential to be a simple and easy strategy to create new
molecular connections by forming covalent S-N, S-O and S-C bonds. In this project,
bi-functional SuFEx linkers have been investigated with the purpose of creating novel
conjugates of biologically relevant molecules.

Six novel bi-functional SuFEx linkers were synthesized in the project. All contain
a SuFEx group for reaction with amines, and another functionality for reaction
with thiols. The SuFEx linkers were screened in a high throughput format where a
two-step, one-pot reaction was performed, using the linker to join together a model
thiol and model amines. Two self-synthesized and two commercially available SuFEx
linkers were compared in different reaction conditions. The results showed proof of
concept, where all screened linkers successfully formed the thiol intermediate and
the SuFEx product, however, the SuFEx reaction is highly dependent on the choice
of reactions conditions and the properties of the amine.

SuFEx linkers were finally used to form conjugates between biologically relevant
molecules containing either a thiol or an amine. All but one linker successfully
yielded the thiol intermediate, however, all failed to produce the SuFEx product.
Further research is required to elucidate the reasons behind this and to find and
improve better reaction conditions.

Keywords: Click Chemistry, SuFEx, HTE, Conjugation

iii



Acknowledgements
First of all, I would like to thank my supervisor Dr. Maximilian Häfner for all your
help and support throughout my thesis work. Your guidance and belief in me during
this spring has allowed me to develop as a chemist and I am very grateful to have
had you as my supervisor. Thank you for designing a project that allowed me to
work with many different techniques and areas in organic chemistry. It has truly
been a great learning experience and a highlight of my studies!

I would also like to thank the New Modalities Team and others in early CVRM
at AstraZeneca for your support and for welcoming me into your team this spring.
A special thanks to my team lead Dr. Ranganath Gopalakrishnan, Dr. Magnus
Johansson and Dr. Malin Lemurell for giving me the opportunity to do my thesis
work at your department.

Thank you Dr. Daniele Antermite for taking the time and energy to help me with
high throughput experimentation. The introductions to the instruments, setting up
the experiments and analyzing the reaction plate and as well as the second step of
the linker screen took many hours from your schedule. The project would not have
been the same without your help.

Finally, I would like to thank Prof. Jerker Mårtensson for being my examiner at
Chalmers University of Technology, and my partner Eric Nilsson for your support
and feedback on the report.

Malin Fuhrman, Gothenburg, June 2025

iv



Table of contents

List of Abbreviations 2

1 Introduction 3
1.1 Aim and Importance . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2 Scope and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2 Theory 9
2.1 SuFEx Linkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.1.1 The SuFEx group . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.2 The thiol accepting group . . . . . . . . . . . . . . . . . . . . 10

2.2 High Throughput Experimentation . . . . . . . . . . . . . . . . . . . 11
2.2.1 Reactions and Mechanisms . . . . . . . . . . . . . . . . . . . . 12

2.3 Purification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.1 Column chromatography . . . . . . . . . . . . . . . . . . . . . 12
2.3.2 Aqueous work-up . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.4 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.1 TLC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.2 LC-MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.3 NMR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3 Results and Discussion 17
3.1 Synthesis of SuFEx linkers . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 High Throughput Experimentation . . . . . . . . . . . . . . . . . . . 21

3.2.1 Determination of conditions for thiol addition . . . . . . . . . 23
3.2.2 Determination of conditions for the SuFEx reaction . . . . . . 24
3.2.3 Heat Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3.2.3.1 Thiol Addition . . . . . . . . . . . . . . . . . . . . . 25
3.2.3.2 SuFEx reaction . . . . . . . . . . . . . . . . . . . . . 26

3.3 Linker Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3.1 Thiol addition . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.3.2 SuFEx reaction . . . . . . . . . . . . . . . . . . . . . . . . . . 34

3.4 Conjugation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

4 Conclusion and Outlook 51

5 Experimental 54
5.1 Synthesis of SuFEx linkers . . . . . . . . . . . . . . . . . . . . . . . . 54
5.2 High Throughput Experimentation . . . . . . . . . . . . . . . . . . . 59
5.3 Linker Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

v



Table of contents Table of contents

5.4 Synthesis of starting materials for conjugation . . . . . . . . . . . . . 63
5.5 Synthesis of conjugates . . . . . . . . . . . . . . . . . . . . . . . . . . 68

References 71

A First appendix A-1
A.1 SuFEx Linkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
A.2 Conjugation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-17

vi



Table of contents Table of contents

1



List of Abbreviations

DABCO 1,4-diazabicyclo[2.2.2]octane

DBU 1,8-Diazabicyclo[5.4.0]undec-7-ene

DCM Dichloromethane

DIPEA N,N -Diisopropylethylamine

DMAP 4-Dimethylaminopyridine

DMF Dimethylformamide

DMSO Dimethylsulfoxide

EtOAc Ethyl Acetate

iPrOH Isopropanol

MeCN Acetonitrile

Me-THF 2-Methyltetrahydrofuran

t-AmylOH Tert-Amyl alcohol

TCEP Tris(2-carboxyethyl)phosphine

TEA Triethylamine

TFA Trifluoroacetic acid

2



1. Introduction

Click chemistry is a synthetic strategy enabling easy formation of carbon-heteroatom
and heteroatom-heteroatom linkages [1], and was first coined by Sharpless in 2001
as a response to the synthetic challenges posed by many drug candidates in the
pharmaceutical industry. The idea was to explore different types of linking chemistry
that could be used to build molecules in a simple way, in order to decrease time and
costs during the drug discovery process [2].

In order for a reaction to be classified as click chemistry, it must fulfill a list of
criteria. The most important requirements are that the reaction must be simple to
perform and generate high yields of the product in a reliable manner. Purification
steps should be limited and the reaction should have applications in many areas and
use only commercially available starting materials [2]. There are four main categories
of click reactions, namely cycloaddition reactions (e.g., Diels-Alder cycloadditions),
nucleophilic substitution reactions, (e.g., ring-opening of strained rings such as
epoxides and aziridines), carbonyl condensation reactions and addition reactions
(e.g., Michael addition) [3]. The first reaction to be described as click chemistry in
2001 was the cycloaddition between an alkyne and an azide, catalyzed by Cu(I) (the
‘CuAAC’ reaction) [2]. This reaction is regarded as the click reaction [1].

Figure 1.1: Cu(1) catalyzed cycloaddition of an alkyne and an azide.

Being biocompatible, click chemistry is an excellent strategy for use in drug discovery
and is also heavily used in medicinal chemistry [4]. For example, click chemistry in
combination with high throughput strategies can accelerate the drug discovery process
by forming libraries of drug candidates to be screened for e.g., targets or cellular
assays [5]. The CuAAC reaction has been used to form peptide-peptide conjugates [5]
and conjugates of oligodeoxyribonucleotides (ODNs) and small fluorescent molecules
for the purpose of labeling [6] (figure 1.2). It has also been used in library syntheses
of small molecule enzyme inhibitors such as protein tyrosine phosphatasesenzyme
(PTP) warheads and protein kinases inhibitors [5]. Furthermore, the Diels-Alder
click reaction has been used to synthesize many drugs such as Bisantrene (to treat
leukemia), Benzoctamine (against anxiety) and Oselmivir (influenza medication)
but has also been used in bioconjugation and in the synthesis of biologially active
molecules [4].

3



1. Introduction 1. Introduction

Figure 1.2: Modifying ODNs with fluorescent small molecules using CuAAC
click chemistry [6].

Sulfur fluoride exchange (SuFEx) chemistry was first described by Sharpless et al.∼
in 2014 [7] and is a type of click chemistry utilizing the stability and reactivity of the
S(VI)-F bond. The S(VI)-F bond is thermodynamically stable and resistant to acidic
and basic conditions and compared to its S-Cl sibling, it is also resistant to reduction
and to nucleophilic attack. The reactivity is however unlocked when fluoride, a poor
leaving group and strongly bonded to sulfur, is activated by coordination to H+ or
R3Si+ [7] or by addition of a Lewis acid or base [8]. This allows for substitution
to occur upon collision with a nucleophile, and due to the high electronegativity
of fluorine, nucleophilic attack occurs selectively at the sulfur center resulting in
exchange of fluoride [7].

Figure 1.3: Reactivity of sulfur fluoride exchange (SuFEx).

SuFEx molecules contain an S(VI)-F bond that can undergo SuFEx reactions. Dif-
ferent variations exist, including sulfuryl fluoride (SO2F2), sulfonyl fluoride (RSO2F)
and sulfamoyl fluoride (R1R2NSO2F).

(a) (b) (c)

Figure 1.4: Examples of common SuFEx functional groups: (a) sulfuryl
fluoride, (b) sulfonyl fluoride and (c) sulfamoy fluoride.

Sulfuryl fluoride is a harmful gas normally used as fumigant [9], and displays high
reactivity towards phenols and secondary amines [10]. Sulfonyl fluorides react
with oxygen, nitrogen or carbon-based nucleophiles, whilst sulfamoyl fluorides are
less explored and have only been reported to react with nitrogen-based nucleophiles
[8].

SuFEx connectors can thus be used to create molecular connections by forming
covalent S-O, S-N and S-C bonds between molecules. There is an increasing number

4



1. Introduction 1. Introduction

of applications in bio-conjugation [11], medicinal chemistry [12], polymer chemistry
[13], organic synthesis and drug discovery [9], to name a few. Sulfonyl fluoride groups
on drug candidates can inhibit enzyme function by covalently link to amino acids on
the active sites [1], and the use of sulfuryl fluoride for late stage functionalization of
phenols showed improved anticancer activities for certain compounds [14]. The fact
that the SuFEx reaction proceeds under metal–free conditions is especially important
for applications in biology and drug discovery [1].

Thionyl fluoride (SOF4) derived reagents can react with primary amines on the
protein bovine serum albumin as a way to modify proteins via bio-conjugation [11],
and the sulfonyl fluoride group has been shown to act as covalent warhead for the
inhibition of the Glutathione S-transferase (GST) Pi enzyme, which is present to a
great extent on various types of cancer cells [15].

Figure 1.5: Sulfonyl fluoride as a covalent warhead binding to GST Pi
enzyme [15].

Furthermore, sulfonyl fluoride groups on stapled alpha-helical peptides are used to
target His252 on hMcl-1, another protein that is abundantly expressed on cancer
cells, by covalently bind to the imidazole on histidine via a SuFEx reaction [16].

Figure 1.6: Stapled alpha-helical peptides used to target His252 on hMcl-1
via a SuFEx reaction [16].

SuFEx molecules containing another type of functionality can serve as bi-functional
linkers. An example is ethene sulfonyl fluoride (ESF), which apart from its SuFEx
moiety contains an activated double bond to which nucleophiles can add.

Figure 1.7: Ethene sulfonyl fluoride (ESF).

5



1. Introduction 1.1. Aim and Importance

The advantage compared to other SuFEx molecules is that the bi-functional linkers
contain different types of reactivity that can be used to form molecular connections
[1]. In this project, different bi-functional SuFEx linkers have been investigated.
All contain a SuFEx functionality (sulfonyl fluoride or sulfamoyl fluoride) to which
amines can exchange with fluoride, and one functionality with which thiols can react,
either via conjugate addition or via substitution.

1.1 Aim and Importance
The aim of the project was to compare different bi-functional SuFEx linkers for
the purpose of linking together biologically relevant molecules. The project was
divided into three parts, where the first part consisted of synthesizing bi-functional
SuFEx linkers containing a SuFEx group and a functionality to which thiols can
add (the thiol accepting group). In the second part of the project, synthesized and
commercially available SuFEx linkers were screened for amine and thiol reactivity in
a high throughput format. The SuFEx linkers were compared whilst also exploring
different reaction conditions. Product yield was used to determine the optimal
conditions for an extended comparison between additional linkers. The final part of
the project consisted of conjugating biologically relevant molecules using the results
from the screening processes. The amine functionality is abundant in drug molecules
and is therefore a good target for conjugation [17], and primary and secondary amines
and anilines on commercial small molecule drugs were used to react with the SuFEx
group on the linker. Thiols on various drug modalities could add to the linker via
conjugate addition or via substitution.

SuFEx linkers provide a simple route to create molecular connections [9] and to the
best of my knowledge the linkers investigated in this project are under–explored in the
scientific community. The demand for fast, efficient and physiologically compatible
methods to form chemical conjugations increases, and versatile approaches to building
complex molecules are therefore essential in order to meet these demands [11].
Investigating whether the SuFEx linkers in this project could be used to conjugate
biologically relevant molecules is thus important as this can expand the current linker
chemistry available. The drug modalities included in this project were a peptide,
a dye, lipids, a biotin derivative and an E3 ligase ligand. Each of these classes of
molecules play important roles in science and drug discovery.

Peptide therapy has become increasingly important in medicine as peptides usually
exhibit good binding to molecular targets with high specificity and potency. By
conjugating a peptide to a small molecule drug, the aqueous solubility, metabolism
and cell permeability of the drug can improve whilst also allowing targets to be
reached inside the cell, something that previously has been a limitation to peptide-
based therapeutics. Additionally, the peptide can provide targeted delivery to the
right cells, which can increase the local concentration of the drug, and also decrease
accumulation at healthy cells and thus limit toxicity. Some examples of peptide-based
therapeutics include glucagon-like peptide-1 agonists (GLP-1) and somatostatins
(growth hormone-inhibiting hormone). Somastostatin conjugated with small molecule

6



1. Introduction 1.1. Aim and Importance

radioisotope 177Lu displayed higher efficacy and improved safety for cancer treatment
compared to the individual constituents alone [18].

Figure 1.8: A peptide conjugated to a small molecule drug via a linker.

Proteolysis-targeting chimeras, or PROTACs, are bi-functional small molecules
intended to degrade protein and consist of two small molecule ligands connected by
a linker. One ligand binds to the protein target whilst the other (the E3 ligase
ligand) binds to an E3 ubiquitin ligase. The small molecule PROTAC enables
protein degradation by the E3 ligase due to the close proximity of the two entities
[19]. PROTACs are an interesting group of new modality recently reaching clinical
trial in several occasions, and improving the linker part of the molecule is an important
part of the extensive optimization process required due to the increased complexity
of the larger molecule [20].

Figure 1.9: The mechanism of protein degradation via E3 ligase ligand
binding.

Small molecules conjugated to fluorescent dyes enables imaging of drug candidates
to elucidate transport and distribution pathways on cellular levels in vitro and
in vivo, thus aiding in the time-consuming process of drug optimization [21]. To
modify drug properties, small molecule drugs can be conjugated to lipids in order
to make them more lipophilic. This can have positive effects such as improved
bioavailability via increased cell membrane interaction, and increased blood–brain
barrier penetration for compounds that have to be lipophilic in order to pass.

Figure 1.10: A fluores-
cent dye elucidating the
pathway of a drug.

It can also alter drug release profiles to allow release
for an extended period of time with fewer dose admin-
istrations as a consequence [22]. Biotinylation is the
process of attaching biotin (vitamin B7) to biologically
active molecules, and it can be used for small molecules
drugs to elucidate interactions with the protein target
[23]. Biotinylation has also showed increased uptake of
small molecule drugs, especially in cancer cells [24] as
they have an increased number of biotin uptake systems
[25].

7



1. Introduction 1.2. Scope and Limitations

1.2 Scope and Limitations
The overall goal of the project was to produce final conjugates using SuFEx linkers
synthesized in the project. Sub-goals included successfully synthesizing SuFEx linkers
and testing them in a high throughput format. To ensure that this would be possible
in the 20 weeks that the project progressed, the time allowed for linker synthesis
was limited to 10 weeks. The synthesized SuFEx reagents were limited to sulfamoyl
fluoride linkers as this SuFEx moiety can be attached using a solid imidazolium
salt and an amine, a procedure already described in literature. High throughput
experiments (HTE) were limited to 96 reactions in order to fit in one reaction plate,
and the number of HTE screens performed was limited to one. The number of
conjugates was limited to a maximum of five, including five small molecule drugs,
five different drug modalities and five different linkers. Each molecule was part of
only one conjugate.

8



2. Theory

This chapter covers the theory behind the synthesis and reactions of the different
SuFEx linkers used in the project, as well as of the analysis and purification methods
used.

2.1 SuFEx Linkers
The first part of the project consisted of synthesizing the SuFEx linkers that were
to be used to conjugate biologically available molecules. The SuFEx linkers in this
project contained a SuFEx group and an thiol accepting group (for thiol addition)
connected by a linker backbone. Figure 2.1 displays a representative example of such
a linker, where the SuFEx group is the sulfamoyl fluoride and the thiol accepting
moiety is the activated double bond of the acrylate.

Figure 2.1: A representative example of a SuFEx linker synthesized in the
project.

2.1.1 The SuFEx group
The SuFEx group on the linkers used in the project were of two types: sulfamoyl
fluorides and sulfonyl fluorides.

Sulfamoyl fluorides
The SuFEx group on the synthesized linkers were in most cases of the sulfamoyl
fluoride type (R1R2NSO2F). A common method to synthesize sulfamoyl fluorides
is to use sulfuryl fluoride (SO2F2) but because this reagent is both harmful and
a gas, it comes with some challenges [8]. Laboratories require a special setup to
handle harmful gas, and there might be regulatory limitations in the use of sulfuryl
fluoride. Additionally, the reaction between sulfuryl fluoride and amines to yield
sulfamoyl fluorides are not as straightforward as with phenols due to slow reaction
times, complicated reaction conditions and undesired elimination reactions. To avoid
these problems, a solid imidazolium salt was developed by Sharpless and co-workers
in 2018 [26], which possesses a greater reactivity towards nitrogen nucleophiles and
give good yields of sulfamoyl fluorides without the addition of base or activating
agents, something that is required when using sulfuryl fluoride. Other advantages

9



2. Theory 2.1. SuFEx Linkers

include shorter reaction times and the fact that the imidazolium salts are solid, which
makes the synthesis simpler [26]. For these reasons, this is the method used in this
project. The overall reaction is displayed in figure 2.2 below.

Figure 2.2: The overall reaction of making sulfamoyl flourides using imida-
zolium salts.

Sulfonyl fluorides
The SuFEx group on the commercially available linkers used in this project were
of the sulfonyl fluoride type (RSO2F). Sulfonyl fluorides can be synthesized using
gaseous sulfuryl fluoride or sulfur dioxide, however solid reagents are desired due to
practical reasons. For example, DABSO forms sulfonyl fluorides by reaction with
Grignard reagents and sulfonyl chlorides can be converted to sulfonyl fluorides by
reaction with KHF2, [8]. Below is an example of a SuFEx linker with a sulfonyl
fluoride group.

Figure 2.3: Ethene sulfonyl fluoride (ESF) bearing a sulfonyl fluoride group.

2.1.2 The thiol accepting group
The thiol-accepting group on the SuFEx linkers were Michael acceptors for conjugate
addition of the thiol, and alkyl bromides for substitution reactions. The different
thiol accepting groups installed on the molecules were chosen based on what is
commonly used in bio-conjugation of thiols [27], [28], [29]. They were installed on
the compounds via amide formation, using different reagents for different linkers. A
representative reaction is shown below, where amide formation is successful using
acryloyl chloride.

Figure 2.4: A representative example of installing the thiol accepting group
via amide formation.

10



2. Theory 2.2. High Throughput Experimentation

2.2 High Throughput Experimentation
The second part of the project consisted of screening successfully synthesized SuFEx
linkers for amine and thiol reactivity in a high throughput format. High Throughput
Experimentation (HTE) is a method used to run a large number of reactions simulta-
neously to allow for fast experimentation [30]. HTE was originally a technique used
in biology but is currently used in chemistry for e.g., research on novel synthetic
strategies, optimization of reaction conditions such as solvents or catalysts [31] or to
quickly synthesize many compounds for use in drug discovery [32]. By systematically
planning and performing reactions in a high throughput format, scientists can gain a
better understanding of e.g., specific reaction steps, reaction mechanisms or reagent
scope by running reactions in an extensive array of conditions. The small scale
results in minimal use of reagents, and the lack of purification steps and work-up
generates results quickly [31].

Figure 2.5: High Throughput Experimentation.

The purpose of using the SuFEx linkers in HTE was to connect an amine with a
thiol using the two different types of reactivities of the linker. The general reaction
is displayed below using ethene sulfonyl fluoride (ESF) as SuFEx linker and a model
thiol and amine.

Figure 2.6: The general reaction performed in HTE.

The reaction is a two-step reaction, where the thiol is added before the amine. As
the amine can react both with the SuFEx group and the alkyl bromide/Michael
acceptor, it must be added after the thiol has had time to completely add to the thiol
accepting moiety, leaving only the SuFEx group available for attack by the amine.

11



2. Theory 2.3. Purification

2.2.1 Reactions and Mechanisms
Thiols can add to the SuFEx reagents via a substitution mechanism, where the thiol
substitutes a bromide.

Thiols can add to the SuFEx reagents via conjugate addition to an activated alkene
or alkyne.

Adding an amine to the SuFEx group goes via a substitution mechanism, and is the
same for primary and secondary amines and anilines.

2.3 Purification
After synthesis, compounds need to be purified to obtain the pure product. The
different methods for purification that were used during this project are described
below.

2.3.1 Column chromatography
Column chromatography is a method used to separate compounds from a chemical
mixture in order to isolate the desired product. The sample is injected into a column
filled with a stationary phase, and a mobile phase is passed through the column. In
this project, liquid chromatography was used, where the mobile phase is a liquid.

12



2. Theory 2.3. Purification

There are two options: 1) polar stationary phase and nonpolar mobile phase (normal
phase chromatography) and 2) nonpolar stationary phase and polar mobile phase
(reversed phase chromatography). As the stationary phase and the mobile phase
differ in polarity, compounds will interact with varying strength to either phase
depending on the polarity of the compound itself. Due to differences in polarity,
compounds in the chemical mixture retain on the column to varying extents and will
elute at different times, which makes separation possible [33].

Figure 2.7: A schematic overview of separation via column chromatography.

In this project, Biotage® flash column chromatography was used. For normal phase
chromatography, the stationary phase then consists of silica spheres and for reversed
phase C18 or C8 carbon chains [34]. For normal phase chromatography, the mobile
phase can consist of e.g., ethyl acetate in heptane and for reversed phase of acetonitrile
in water [35]. A gradient system is used, where the composition of the mobile phase
is adjusted during the separation to allow for compounds with different polarity to
elute in a reasonable time frame [33].

2.3.2 Aqueous work-up
Work-up is the process of isolating the product from the reaction mixture and
include for example filtration, distillation and extraction. When the reaction mixture
contains compounds that are soluble in aqueous and organic media, an aqueous
work-up can be performed to separate aqueous compounds from organic products
using their differences in solubility. To the reaction mixture is added an aqueous
solvent (such as water) and the mixture is transferred to a separating funnel and
extracted with an organic solvent (such as ethyl acetate). Organic compounds will
prefer the organic layer, and aqueous compounds will stay in the aqueous layer.
Addition of acid and/or base can make molecules such as organic amines or acids
more water soluble and thereby exclude them from the organic layer. The extraction
can be done in steps to exclude one or more compounds at a time until only the

13



2. Theory 2.4. Analysis

product dissolves in the organic layer. A drying agent can finally be used to remove
residual water from the organic layer [36].

2.4 Analysis
To analyze reactions mixtures and products throughout the project, the following
methods were used.

2.4.1 TLC
Thin Layer Chromatography, TLC, is a type of liquid chromatography that uses a
plate with a thin (∼ 0.25 mm) layer of absorbing material as the stationary phase
[37]. The mobile phase is a liquid in which the plate is placed. TLC is used to
separate compounds in a chemical mixture by the difference in polarity between
them. In this project, the stationary phase was polar silica gel and the mobile phase
was usually a mixture of ethyl acetate and heptane. The sample is placed on the
TLC plate, which is slightly submerged in the mobile phase allowing the molecules
in the mixture to interact with both phases [38], as they move up the plate due to
capillary forces [37]. Polar molecules will interact to a greater extent with the polar
silica gel than with the less polar mobile phase and will therefore retain on the plate
close to the starting point. Less polar compounds will on the other hand display
a stronger interaction with the mobile phase than the stationary phase, and will
therefore migrate farther up the plate together with the mobile phase [38]. After
the plate is taken out of the mobile phase and the solvent has evaporated, spots
of different components from the mixture can be detected by e.g., UV detection or
by staining. In this project, potassium permanganate (KMnO4) was used as stain
the majority of the time. Permanganate oxidizes functional groups such as alcohols,
alkenes, ethers, amines, amides, esters etc. to form MnO2 which shows up as yellow
spots on the TLC plate upon heating [39].

2.4.2 LC-MS
LC-MS stands for Liquid Chromatography with Mass Spectrometry and is an ana-
lytical method that separates compounds from a chemical mixture and subsequently
identifies them based on their mass. In this project, the separation is provided by
reversed phase High Performance Liquid Chromatography (HPLC). The stationary
phase is nonpolar and consists of a column covered with C8 or C18 carbon chains
bound to porous silica particles. The particles are small (3-10 µm) and tightly packed
to increase the surface area of the stationary phase. Smaller particles create smaller
interstitial spaces for the mobile phase to occupy, and thus a high pressure is required
to push the mobile phase through the column. In reversed phase HPLC, the mobile
phase is polar and can for example be a mixture of water and acetonitrile. Less polar
compounds will remain on the stationary phase to a greater extent and will elute
at a later time than polar molecules that prefer to interact with the mobile phase
rather than with the stationary phase. The composition of the mobile phase can be
adjusted during the separation to form a gradient, going from more polar to less

14



2. Theory 2.4. Analysis

polar during the separation. The varying polarity will ensure that compounds with
different polarity still elute in a reasonable time frame [33].

The mass spectrometer ionizes the compounds eluted from the column, and can also
cause fragmentation of the molecules into smaller parts. The ionization makes it
possible to separate them based on their mass-to-charge ratio using a magnetic or
electrical field, and a detector counts the number of ions to obtain both quantitative
and qualitative information [33]. There are different types of ion sources and mass
analyzers. In Electron Impact Ionization, the sample is evaporated and passed
through an electron beam forming positive ions. The ions are unstable and fragment,
and the method is classified as a hard ionization method [40]. Electrospray Ionization
converts the sample into a charged spray of droplets that will ionize the compounds
as solvent molecules are evaporated. This method does not impact the molecules
with much energy and they are therefore not very fragmented [41]. The Time
of Flight (TOF) mass analyzers measure the distance that ions travel after being
accelerated to reach the same kinetic energy. Their mass-to-charge ratio gives them
different velocities which makes it possible to separate them by mass [42]. Quadrupole
analyzers on the other hand use a set of four parallel metal rods through which
molecules with a certain mass-to-charge ratio can travel as the voltage is varied [41].

2.4.3 NMR
Nuclear Magnetic Resonance (NMR) utilizes the quantum property of spin angular
momentum of protons and neutrons in atomic nuclei. The spins of the elementary
particles couple to create a net spin, which exist for all atomic nuclei with an odd
atomic number. The nuclear spins give rise to a magnetic field [43] and because spin
is a quantum property, the spin states of atomic nuclei are quantized with discrete
energy levels. The most common nuclei for NMR are 1H and 13C, and 1H NMR will
be discussed for conceptual purposes. The hydrogen nucleus has two spin states, and
the lack of an external magnetic field will result in two degenerate states with the
same energy [44]. However, once an external magnetic field is applied to the nucleus,
the energy levels split and become separated by an energy difference. Nuclei can
align with the magnetic field and occupy a lower energy level or align against the
external field and occupy a higher energy level [43]. During an NMR experiment, an
oscillating magnetic field supplies the energy required to excite nuclei in the lower
energy level to the upper [44], and the frequency at which the transition occurs is
called the resonance frequency [43]. The excited nuclei will eventually relax back to
the lower energy level and in doing so they emit energy corresponding to the energy
difference between the two spin states. The NMR machine records the frequency at
which the relaxation occurs and measures the intensity to produce a signal [43].

The reason NMR can be used to characterize molecules is that different nuclei
resonance at different frequencies depending on their chemical environment. The
chemical environment of a nucleus is determined by the electron cloud surrounding
it, which in turn is affected by electron densities on neighboring atoms [44]. Electron
densities generate their own magnetic fields that work opposite the applied magnetic

15



2. Theory 2.4. Analysis

field and thus cause the nuclei to experience a smaller net magnetic field. This effect
of subtracting from the applied magnetic field is called shielding and the extent
at which it will occur depends on the electronic environment around each nucleus,
giving rise to different resonance frequencies. However, for nuclei close to electron
withdrawing substituents, the opposite occurs. These substituents will pull electron
density from the nucleus and decrease the shielding effect, and they are said to be
de-shielding. The protons will therefore experience a larger net magnetic field than
shielded protons and the resonance frequencies become higher [43]. The frequency
at which resonance occurs is called the chemical shift, and is reported relative to a
reference [44].

For 1H NMR, an additional feature is present. For a CH2 proton, there are four
combinations of spin states for the two nuclei. Both nuclei can occupy the upper
energy level, the lower energy level state or occupy different energy levels. Because
there are two alternatives of the two protons occupying different energy levels, the
signal intensity of them doing so will be twice the intensity of both of them occupying
different levels, and the intensity ratio is thus 1:2:1. A proton on an adjacent carbon
will therefore interact with the protons on the CH2 in three different ways, as there
are three spin combinations to interact with. Its signal will thus be split into three
according to the 1:2:1 ratio. The result is called signal splitting, and affects protons
that are able to interact with protons on adjacent carbons. The integrated peak
area intensities for each signal correspond to the number of protons giving rise to
the signal, and can be used to deduce which nuclei correspond to which peak in the
NMR spectrum [44].

16



3. Results and Discussion

This chapter covers the results and discussion on the different parts of the project:
linker synthesis, high throughput experimentation, linkers screen and conjugation.

3.1 Synthesis of SuFEx linkers
Six SuFEx linkers were successfully synthesized in the project, and are presented
in table 4.1. L1, L2, L4 and L6 were synthesized via a three step synthesis where
the SuFEx group was installed first using a solid imidazolium salt according to a
procedure described in litearture [26]. The starting materials were Boc-protected
amines and the resulting SuFEx groups were thus sulfamoyl fluorides. Two Boc-
protected amines were used, which allowed the SuFEx linkers to have different
degree of flexibility. After the SuFEx group had been installed, the Boc-group was
deprotected and the thiol accepting group was added to the compound via amide
formation. L3 and L5 were each synthesized in one step; L3 via amide formation
to create a sulfonyl fluoride linker, and L5 via installation of the sulfamoyl fluoride
group using the imidazolium salt. The different thiol accepting groups installed on
the molecules were chosen based on what is commonly used in bio-conjugation of
thiols [27], [28], but also based on what was commercially available.

Table 3.1: SuFEx linkers synthesized in the project.

Name Structure Molecular Weight Label
4-(but-2-
ynoyl)piperazine-1-
sulfonyl fluoride

234.245 g/mol L1

(3-(2-
bromoacetamido)
propyl) sulfamoyl
fluoride

277.108 g/mol L2

3-((2-(2,5-dioxo-2,5-
dihydro-1H-pyrrol-1-
yl)ethyl)
carbamoyl)benzene
sulfonyl fluoride

326.298 g/mol L3

4-acryloylpiperazine-
1-sulfonyl fluoride

222.234 g/mol L4

(2-(2,5-dioxo-2,5-
dihydro-1H-pyrrol-1-
yl)ethyl)sulfamoyl
fluoride

222.190 g/mol L5

4-(2-bromoacetyl)
piperazine-1-sulfonyl
fluoride

289.119 g/mol L6

17



3. Results and Discussion 3.1. Synthesis of SuFEx linkers

Adding the sulfamoyl fluoride group to the Boc-protected amine 2 using 1 gave the
SuFEx product 3 in good yield.

Subsequently adding the thiol accepting group was first attempted with the NHS-
esters 5 and 6, but in both cases it formed the undesired addition and substitution
products 7 and 8, respectively.

The reason is assumed to be the bulkiness of the secondary amine, making it sterically
more accessible to attack the activated double bond or the alkyl bromide rather
than the carbonyl of the NHS-esters 5 and 6. L6 was also attempted by reacting
piperazine 4 with bromoacetyl bromide 9, but despite the relief in steric hindrance,
the reaction was unsuccessful.

However, using the less sterically hindered reagent acryloyl chloride 10 to install the
acrylate on piperazine 4 gave L4 in good yield upon scaling up the reaction to 1 g
of piperazine 4. At smaller scale (∼ 100 mg), the yield from the same reaction was
about 30 %.

Butynoic acid 11 was reacted with 4 using EDC as an attempt to install the alkyne,
however this yielded product in low yield (∼ 10%). As the reason for this could have
been the choice of coupling agent, T3P was explored as an alternative in the reaction.
T3P proved to be very efficient for the amide coupling and furthermore provided
pure material after aqueous work-up, making further purification unnecessary.

18



3. Results and Discussion 3.1. Synthesis of SuFEx linkers

The success of amide formation using T3P motivated the attempt to synthesize
product L6 again, using T3P and bromoacetic acid 12, and this reaction formed
the product in good yield. However, bromoacetyl chloride, a reagent ordered as
bromoacetyl bromide 9 did not work, also proved to be a good reagent for the
formation of L6. The increased reactivity in using the acetyl chloride instead of the
acetyl bromide might be needed due to the bulkiness of the amine.

Attaching the SuFEx group onto amine 13 using 1 was also a successful reaction.

However, installing the thiol accepting groups on Boc-deprotected amine 15 came
with some challenges. For example, linker 16 was obtained in extremely low yield of
about 5% upon reacting 15 with acryloyl chloride 10. This is surprising as acryloyl
chloride 10 formed product of significant yield when reacted with piperazine 4. The
reason could be an intramolecular cyclization of the amine, or a polymerization
reaction. The same product was also attempted by reacting 15 with the NHS-ester of
the acrylate 5, but this also yielded product only in low yield (6%). A final attempt
was performed using T3P and acrylic acid, but despite the success of T3P in previous
reactions, this reaction yielded no product.

As for piperazine 4, reaction of amine 15 with butynoic acid 11 using EDC as
coupling reagent was unsuccessful, and for amine 10 the reaction yielded no product
at all. Perhaps EDC is not a suitable coupling reagent for these types of compounds.
Amine 15 was only successful in forming a SuFEx linker when reacted with the
NHS-ester of bromoacetate 6, yielding 23% of product. The fact that this reaction
worked but not with the acrylate equivalent 5 is surprising considering both reagents

19



3. Results and Discussion 3.1. Synthesis of SuFEx linkers

are sterically comparable. The difference might be due to the conjugated system of
the acrylate, which increases electron density at the carbonyl and thus decreases the
electrophilicity.

As attaching the thiol accepting group had proven to be problematic, the synthesis
was attempted backwards with Boc-protected amine 15 and acryloyl chloride 10,
first attaching the thiol accepting group followed by Boc-deprotection and subsequent
installation of the SuFEx group. The first step proved to be a successful reaction
and gave product in 40% yield. However, upon attaching sulfamoyl fluoride on 17
using 1, no product was obtained after purification, despite the fact that 16 was
UV-active in LC-MS.

The maleimide-based linker L5 was synthesized in one step forming an amide between
amine 18 and carboxylic acid 19, using T3P as coupling reagent. As for most other
reactions using T3P, L3 was produced in high yield requiring only aqueous work-up
to form the pure product.

However, attaching the sulfamoyl fluoride on amine 18 using the imidazolium salt
1 proved to be less successful, which is surprising considering the high yields in
reactions of 1 with amines 2 and 13. This reaction was attempted four times, with
three different equivalents of base, however the highest yield recorded was only 22%.

The SuFEx linkers in this project are small and some of them are not UV-active,
which make them difficult to detect in LC-MS analysis and after purification by
column chromatography. For example, compound 20 was attempted twice with
imidazolium salt 1 to install the sulfamoyl fluoride, but failed both times. The
compound ionizes well and was clearly visible in the LC-MS analysis of the reaction
mixture, however, it is not very UV-active and could not be detected after column
chromatography. Another attempted linker is the polyethylene glycol (PEG) based
linker 21 shown below, however due to time limitations, installation of the thiol

20



3. Results and Discussion 3.2. High Throughput Experimentation

accepting group was never attempted. Addition of the SuFEx group was however
successful and different PEG-linkers could thus be explored in future studies.

The sulfonyl fluorides presented in table 3.2 are commercially available and were
used as SuFEx linkers in the project.

Table 3.2: Commercially available SuFEx linkers used in the project.

Name Structure Molecular Weight Label
4-(bromomethyl) benzenesulfonyl
fluoride

253.085 g/mol L7

2-phenylvinylsulfonyl fluoride 186.20 g/mol L8

Ethylenesulfonyl fluoride 110.102 g/mol L9

3.2 High Throughput Experimentation
Reactions in high throughput format were performed in a 96 well plate containing
1 mL glass vials. Two subsequent reactions were performed: thiol addition (Michael
addition/alkylation) and SuFEx reaction. The same 96 well plate was used for both
reactions and the solvent was evaporated between steps. Due to time limitations the
thiol addition reaction was performed over the weekend, while the SuFEx reaction
was performed overnight. Both steps were performed at room temperature.

Two synthesized (L3 and L4) and two commercial (L7 and L8) linkers were tested in
a high throughput format. The linkers were chosen based on two things: 1) to have
four different thiol accepting groups: a maleimide, an acrylate, an alkyl bromide and
a vinyl sulfonyl fluoride, 2) based on the amount of linker available. Each of the four
linkers used in HTE were reacted with one model thiol in the first step, and with
three model amines in the second step (the SuFEx reaction). The requirements on
the model compounds were that they should exhibit UV activity but not have their
nucleophilic group affected by other functional groups, as to be able to examine only
thiol and amine reactivity. Furthermore, three amines were chosen to allow SuFEx
reactions with 1) a primary amine, 2) a secondary amine and 3) an aniline. The
structures of the model compounds are presented in table 3.3 below.

21



3. Results and Discussion 3.2. High Throughput Experimentation

Table 3.3: Model thiol and amines used in HTE.

Name Structure Molecular Weight Label
2-phenylethane-1-thiol 138.228 g/mol S1

phenylmethanamine 121.18 g/mol A1

N-Methyl-N-phenethylamine 135.21 g/mol A2

4-aminobenzenesulfonyl fluoride 171.22 g/mol A3

The plate layout for the first step, the thiol addition, is presented in figure 3.1 below.
Using four linkers in one 96 well plate allows for 24 reactions per linker. The same
conditions were used for all reactions using L4, L3 and L8 as these are all Michael
acceptors. Other conditions were used for the first step using L7 as this is an alkyl
bromide. Determination of conditions was done according to the procedure described
in section 3.2.1 below.

Figure 3.1: Plate Layout of the thiol addition reaction. The colors represent
the four different linkers.

The plate layout for the second step, the SuFEx reaction, is presented in figure
3.2 below. The four linkers produce four thiol intermediates from the first step,
which are all reacted with the three amines. In a 96 well plate, this allows for 8
different SuFEx reactions per linker–amine combination to take place, and thus 8
different conditions were tried. Determination of conditions was done according to

22



3. Results and Discussion 3.2. High Throughput Experimentation

the procedure described in section 3.2.2 below.

Figure 3.2: Plate Layout of the SuFEx reaction. The colors represent the
four different thiol intermediates.

All solid materials were dispensed automatically into the reaction plate using
Chronect®XPR Robotic Dispensing. Stock solutions of all liquid reagent com-
binations were prepared manually and were then automatically dispensed into the
reaction plate. The liquid dispensing was performed by Andrew+ Alliance Pipetting
Robot. Removal of solvent after the thiol addition was performed using Genevac
EZ-2 Elite Evaporator.

Analysis was performed using plate LC-MS, where area integration was done by the
software with manual editing when necessary. Product conversion was determined by
analyzing product peak area in relation to total peak area of product and remaining
starting material as

Conversion = Product Peak Area
Product Peak Area + Starting Material Peak Area . (3.1)

3.2.1 Determination of conditions for thiol addition
To determine which conditions would be suitable for thiol addition in a high through-
put format, test reactions were performed. Conditions to test in these reactions were
based on experiments reported in literature.

Alkylation of alkyl bromide L7 (1 eq) was performed in three solvents; acetonitrile,

23



3. Results and Discussion 3.2. High Throughput Experimentation

DMSO and butanone, all with the addition of thiol S1 (1.5 eq), KI (1 eq) and a
base (2 eq; cesium carbonate for acetonitrile and DMSO and potassium carbonate
for butanone). Conversion to the alkylated products was difficult to determine in
the reactions with acetonitrile and DMSO due to poor ionization in LC-MS. The
primary amine A1 (1.5 eq) was therefore added to all reaction mixtures, together
with DIPEA (2 eq) and Ca(NTf2)2 (1 eq), and the conversion to the SuFEx product
was determined by LC-MS. As there was no significant difference in yield, acetonitrile
was chosen as solvent as it is accessible and easy to evaporate. Potassium carbonate
was chosen as base as this was already available in the dispensing robot.

The conjugate addition of the thiol to Michael acceptor L8 (1 eq) was performed
with thiol S1 (1 eq) in acetonitrile using three different activating agents; DBU (1
eq), TCEP (0.05 eq) and DMAP (0.05 eq). Conversion to the addition product could
not be determined as the product did not ionize in LC-MS. The primary amine A1
(1.5 eq) and Ca(NTf2)2 (1 eq) were therefore added to the reaction mixtures with
TCEP and DMAP and conversion to the SuFEx product could be determined by
LC-MS. A crude NMR of the reaction with DBU was measured which showed the
addition product, but the reaction was considerably slower than the others. Because
the LC-MS of the reaction with DMAP was cleaner, it was decided to use it as
activating agent in HTE.

3.2.2 Determination of conditions for the SuFEx reaction
To determine which conditions would be suitable for the SuFEx reaction in a high
throughput format, a literature review was done. Using DABCO as base and
Ca(NTf2)2 as lewis acid is reported to give excellent yields of SuFEx products in
various solvents [45], [46], [10]. Other sets of conditions such as MgO in MeCN:H2O [7]
and TEA in MeCN [26], as well as with aqueous buffer [11] are also reported. Solvents,
bases and lewis acids were combined to form both new and existing combinations to
be used in HTE. One objective in determining these combinations was to be able to
test the reactions in greener conditions.

3.2.3 Heat Maps
Below are heat maps of the 96 well plate in which the thiol addition and SuFEx
reaction were performed. Values come from LC-MS analyses. Peak integration was
done by the LC-MS software and additional editing was done manually when needed.
Among the 96 samples, green indicates the most desired results and red indicates
the most undesired results, where the color intensity represents the degree to which
a value is positive or negative in relation to the rest. For the heat maps showing
product peak area and conversion, green represents high values as a large amount of
product and high conversion are desired. For the heat maps showing starting material
peak area and impurity peak area, green represent low values, as the presence of
remaining starting material or impurities is undesired.

24



3. Results and Discussion 3.2. High Throughput Experimentation

3.2.3.1 Thiol Addition

(a)

(b)

(c)

(d)

Figure 3.3: Heat maps from the thiol addition reaction showing (a) product
peak area (%), (b) starting material peak area (%), (c) impurity peak area
(%) and (d) conversion (%).

According to figure 3.3, all linkers except L4 performed well in the thiol addition
reaction. L4 did convert to product but also formed an impurity which is believed to
be the DMAP-intermediate, where DMAP has added to the double bond to form an

25



3. Results and Discussion 3.2. High Throughput Experimentation

intermediate that does not react further with the thiol. The assumption is based
on the detected mass in LC-MS analysis. Additionally, there was starting material
left in the reaction mixture. With only one experiment, it is difficult to draw any
clear conclusions about the usefulness of L4. The reactivity of the thiol might be
too low in combination with this linker, and other activating agents could perhaps
have resulted in more of the thiol intermediate. Additionally, further exploration of
conditions could result in a more successful reaction for this linker.

For the other linkers, no starting material remained and no linker-specific impurities
had formed. However, a general impurity was present in all reaction mixtures,
both from the first and second step, which in some reaction mixtures constituted
a significant amount. This impurity is the main reason that Product Peak Area
(hereon PPA) is not 100% for the linkers that converted to product. The impurity
did not ionize and therefore does not show a mass in the LC-MS spectra, but because
it is present in all reactions it is assumed to be some kind of thiol-impurity as the
thiol is the only common denominator in all reactions, perhaps the disulfide of the
thiol. However, to ascertain what compound it is it would be necessary to purify the
reaction mixtures and isolate the different components in order to measure NMR.
This is something that was not manageable in the time frame of the project.

Even though all 24 reactions per linker were performed in the same conditions for
the first step, there are some differences in PPA. For example, A4, A5 and A6
yielded a higher PPA value than, say, E4, E5 and E6. Such random differences could
be attributed to mistakes in the solid or liquid dispensing, either in preparing the
reaction or the analysis plate. Reagents could be heterogeneously distributed in the
vials and for example get stuck on the walls due to movement or from stirring which
would yield an uneven result.

3.2.3.2 SuFEx reaction

Looking at the heat maps from the SuFEx reaction presented in figure 3.4, there are
significantly more variations in the appearance compared to the first step. To begin
with, there is a major difference in PPA between the aniline (A3, columns 3, 6, 9
and 12) and the primary and secondary amines. PPA is especially low for A3 in L4,
and this is believed to be partly because the linker did not completely convert to the
thiol intermediate in the first step. However, the reactivity of the aniline probably
also suffers from the electron withdrawing nature of the sulfonyl substituent in the
para position, pulling electron density from the nitrogen and making it less reactive.

26



3. Results and Discussion 3.2. High Throughput Experimentation

(a)

(b)

(c)

(d)

Figure 3.4: Heat maps from the SuFEx reaction showing (a) product peak
area (%), (b) starting material peak area (%), (c) impurity peak area (%) and
(d) conversion (%).

27



3. Results and Discussion 3.2. High Throughput Experimentation

The reason that this specific aniline was chosen is mainly because it is a solid, which
the naked aniline is not, and because it was commercially available. As it turned out
that stock solutions of all liquid reagents had to be made before HTE could proceed,
it did not matter whether reagents are solids or liquids. Future high throughput
experiments could incorporate the naked aniline instead as this might be more
reactive.

As can be seen in figure 3.4 (a), the aniline performed especially poor in conditions
F, G and H across all linkers, which are the conditions not using Ca(NTf2)2 and
DABCO, a successful combination of reagents reported in literature [45], [46], [10],
[47], [48]. However, the results probably have more to do with the reactivity of the
aniline rather than the conditions themselves, as both F and G yielded the SuFEx
product for the primary and secondary amine in similar amounts as in conditions A
through E, which all used Ca(NTf2)2 and DABCO.

Condition H turned out to be a poor condition for any amine and linker, however
this is not too surprising considering it is reported in literature to yield the SuFEx
product when stirred at reflux [7]. Using only one reaction plate, it was not possible
to attempt the reaction at different temperatures. However, looking at the heat
map of starting material and impurities for condition H it can be seen that, at least
for L2-L4, there is no starting material left but a lot of impurities have formed.
The condition is thus giving rise to some kind of reaction that consumes the thiol-
intermediate. The kind of impurities formed cannot be deduced from the LC-MS
data, and purification would be necessary to ascertain what has occurred.

Conditions A, D and E seemed to work best across all linkers and amines, with
D and E yielding product in all reactions, although with varying success. These
conditions also gave the highest value of PPA obtained out of the 96 reactions. The
two conditions are similar as the solvents used are both alcohols. Condition D using
t-AmylOH is reported in literature [47], [48], [46], however, condition E using iProH
is not. These results are especially exciting as iProH is considered a green solvent.
The PPA appearance from condition A is less varying and generally works better for
the aniline, giving the highest PPA values among all conditions for that amine.

It is important to note that the general conditions are not necessarily the best
conditions for every linker-amine combination. For example, L3 performed the best
with A1 in condition F, which is interesting considering only base is added in this
reaction. Condition G, with only DMSO and DIPEA, and C, using MeCN:H2O,
Ca(NTf2)2 and DABCO also yielded a good amount of product. Both are important;
DMSO is a common solvent for storage of compounds and MeCN:H2O is a common
solvent for e.g., reactions with peptides.

28



3. Results and Discussion 3.3. Linker Screen

Regarding the linkers, it is clear, as for the first step, that L4 performed poorly
compared to the others. Conditions F, G and H were especially poor, and this
might be due to the fact that these conditions are the ones not using Ca(NTf2)2 and
DABCO. However, conditions B and C were also unsuccessful for L4 and the same
fact does not apply here. As only conditions A, D and E yielded product in any
significant amount, the success of the SuFEx reaction using L4 is therefore assumed
to be heavily dependent on the choice of conditions. However, the fact that the linker
did not convert completely to the thiol intermediate in the first step probably affects
the results from the SuFEx reaction and makes it difficult to draw any conclusions.

L3 and L7 were the most successful linkers in the SuFEx reaction and yielded product
in a significant amount across most conditions, implying that they are not as strongly
dependent on the choice of conditions to produce the SuFEx product. However
for L7, conditions E gave a significant increase in product formation, and seems to
be to optimal conditions for this linker. L8 on the other hand yielded product in
variable success. The problem for this linker was not the SuFEx reaction but the
fact that the SuFEx product did retro-Michael reaction and eliminated the thiol to
form an impurity consisting of the linker–amine molecule (figure 3.5), owing to the
high values of Impurity Peak Area.

Figure 3.5: Retro-Michael reaction of linker L8.

3.3 Linker Screen
After high throughput experimentation, a linker screen was performed using the
synthesized and commercially available linkers that were not part of the 96 well
plate reactions. Four synthesized linkers (L1, L2, L5 and L6) and one commercial
linker (L9) were tested in a thiol addition and SuFEx reaction, using thiol S1 for
thiol addition and amine A1 for the SuFEx reaction. The amount of linker used
in thiol addition was decided based on a theoretical amount of product yield after
the SuFEx reaction of about 30 mg in case it would be necessary to isolate the
product. In total, 10 reactions were performed. The conditions used in the first step
were the same as during HTE and two sets of conditions were used for the SuFEx
reactions. Condition E from HTE, with iPrOH, was used because of the successful
results from HTE and because iPrOH is a green solvent. The other condition for the
SuFEx reaction, MeCN:H2O, DABCO and Ca(NTf2)2 in 40 °C, was chosen because
it resembles physiological conditions, which is interesting for bio-conjugation. The
reactions were performed in fume hoods using one 1.5 mL vial per reaction. All solid
and liquid dispensing were done manually, and the solvent was removed between
steps under a stream of N2. Figure 3.6 gives an overview of the reactions performed
in the linker screen.

29



3. Results and Discussion 3.3. Linker Screen

Figure 3.6: Reactions performed in the linker screen.

3.3.1 Thiol addition
The peak appearances from LC-MS spectra from the thiol addition reactions 1–10
are presented below. The numbers by each structure represents the molecular weight
(g/mol) of that compound.

30



3. Results and Discussion 3.3. Linker Screen

Reactions 1 and 6

Figure 3.7 shows the LC-MS spectra from reactions 1 and 6, respectively. None of
the peaks contain a mass that corresponds to either thiol intermediate mass or linker
mass, however, this is assumed to be because these compounds do not ionize and
therefore cannot be identified. It is thus unclear whether the thiol addition worked
or not.

(a) (b)

Figure 3.7: LC-MS peak appearance from thiol addition reactions (a) 1 and
(b) 6.

Reactions 2 and 7

Figure 3.8 and figure 3.9 show the LC-MS spectra from reactions 2 and 7, respectively.

Figure 3.8: LC-MS peak appearance from thiol addition 2.

31



3. Results and Discussion 3.3. Linker Screen

Figure 3.9: LC-MS peak appearance from thiol addition 7.

How much linker remains in the reaction mixtures of reactions 2 and 7 is difficult to
estimate. The linker itself is not very UV-active according to the LC-MS spectra, and
it could be assumed that there is barely any linker remaining in the mixture. However,
from the looks of the ionization spectra it is possible that there is a significant amount
left. Purification would be necessary to obtain a quantitative result.

Reactions 3 and 8

Figure 3.10 and 3.11 show the LC-MS spectra from reactions 3 and 8, respectively.

Figure 3.10: LC-MS peak appearance from thiol addition 3.

Figure 3.11: LC-MS peak appearance from thiol addition 8.

32



3. Results and Discussion 3.3. Linker Screen

As for reactions 2 and 7, it is difficult to estimate how much linker is unreacted.
Having the phenyl ring on the thiol intermediate should make it more UV-active
than the linker and the UV-signal might therefore be misleading. The ionization
spectra of the two compounds might display a more accurate result, a give a lower
value percentage of linker remaining.

The mass corresponding to the mass of the thiol intermediate from reactions 2 and 7
is also present in the LC-MS spectra from reactions 3 and 8 (at 1.19, 361.1 in positive
mode; 359.1 in negative mode). There is a possibility of contamination between
the vials, but due to the amount detected it is probably more likely that this mass
does not correspond to the mass of the thiol intermediate from reaction 2 and 7 but
to another compound with the same mass, or some adduct or impurity from the
LC-column. This assumption is backed up by the fact that the masses 357.1 and
361.1 are also present in negative mode in both LC-MS spectra from reactions 2 and
7, something that the thiol intermediate would not give rise to.

Reactions 4 and 9

Figure 3.12 shows the LC-MS spectra from reaction 4 and 9, respectively. The peak
at 0.65 is assumed to be thiol intermediate peak due to the mass of 333.1 being
detected in negative mode. However, the mass of 331.1 is more prevalent and should
not correspond to the thiol intermediate, and might correspond to something else in
the reaction mixture.

(a) (b)

Figure 3.12: LC-MS peak appearance from thiol addition (a) 4 and (b) 9.

Reactions 5 and 10

Figure 3.13 shows the LC-MS spectra from reaction 5 and 10, respectively.

33



3. Results and Discussion 3.3. Linker Screen

(a) (b)

Figure 3.13: LC-MS peak appearance from thiol addition (a) 5 and (b) 10.

Only in reactions 4, 9, 5, and 10 is there thiol intermediate without remaining linker,
however for reactions 4 and 9 there are several other masses detected as well. More
importantly, the thiol intermediate peak contains another mass, which might mean
that something else is giving rise to the signal. The spectra from reactions 5 and
10 on the other hand show almost only thiol intermediate in UV and represent the
cleanest results out of all reactions. Both linkers should ionize and be visible in the
ionization spectra, and with no traces of remaining linker, L2 and L6 are considered
the best in the first step of the linker screen. The trend in the results reveals that
the alkyl bromides performed better than the Michael acceptors. However, as the
conditions used in the first step have not been optimized, the reason might not be
the reactivity of the linkers but the fact that there are other conditions that are
more suitable for reaction with Michael acceptors.

3.3.2 SuFEx reaction
The peak appearances from LC-MS spectra from the SuFEx reactions are presented
below. The LC-MS analysis for the second step of the linker screen was performed
using the same instrument as for the analysis of the HTE plate, with the aim to better
compare the data. Reactions 1–5 were performed in iPrOH at room temperature,
and reactions 6–10 were performed in MeCN:H2O 1:1 at 40 °C.

Reactions 1 and 6

Figure 3.14 shows the LC-MS spectra from reactions 1 and 6, respectively. No
product is detected, however this is assumed to be because the product does not
ionize very well, as was the problem for the first step. Only an impurity at 1.77
is detected, which is the same as the impurity detected in all reactions in HTE.
Again, this compound does not ionize and therefore does not show a mass in LC-MS.
Purification and NMR analysis would be necessary to elucidate which structure is
giving rise to the signal.

34



3. Results and Discussion 3.3. Linker Screen

(a)

(b)

Figure 3.14: LC-MS peak appearance from SuFEx reaction (a) 1, in iPrOH
and (b) 6, in MeCN:H2O 1:1.

Reactions 2 and 7

Figure 3.15 shows the LC-MS spectra from reactions 2 and 7, respectively. The
results show the impurity peak at 1.77, but also several other peaks, as opposed
to the LC-MS data from reactions 1 and 6. However, none of the peaks contain a
mass corresponding to the mass of the SuFEx product. With promising LC-MS data
from the first step, this either means that the thiol intermediate is consumed in an
undesirable partway or that the mass detected from the first step did not arise as a
result of thiol intermediate formation. The fact that other masses were detected in
the thiol intermediate peak from the first step and that the same mass was detected
in other reactions mixtures support the latter theory.

35



3. Results and Discussion 3.3. Linker Screen

(a)

(b)

Figure 3.15: LC-MS peak appearance from SuFEx reaction (a) 2, in iPrOH
and (b) 7, in MeCN:H2O 1:1.

Reactions 3 and 8

Figure 3.16 shows the LC-MS spectra from reactions 3 and 8, respectively. Due to
L1 being an alkyne, there are two isomers of the thiol intermediate and two isomers
of the SuFEx product. They all have different retention times (in the area 1.36–1.45)
but the retention time of one of the isomers of the thiol product is similar enough
to the retention time of one isomer of the SuFEx product to make these peaks
overlap. This makes it difficult to determine a conversion, however, with the peak
at 1.36 corresponding to thiol intermediate and 1.46 corresponding to product, it
can be concluded that there is more thiol intermediate remaining than the amount
of product formed. A 4 minute method was run for this sample as an attempt
to separate the peaks, however this did not help. A comparison between the two
conditions reveals that the reaction run in MeCN:H2O resulted in more product
formation.

36



3. Results and Discussion 3.3. Linker Screen

(a)

(b)

Figure 3.16: LC-MS peak appearance from SuFEx reaction (a) 3, in iPrOH
and (b) 8, in MeCN:H2O 1:1.

Reactions 4 and 9

Figure 3.17 shows the LC-MS spectra from reactions 4 and 9, respectively. As
for reactions 1 and 6, the results show nothing but the impurity at 1.77. This
is surprising considering the thiol intermediate was visible in the LC-MS spectra
from the first step. The change of instruments should not be the reason for this
discrepancy, and the SuFEx product should exhibit a significant UV-activity due to
its two aromatic groups. Perhaps the masses detected from the first step did not
arise as a result of thiol intermediate formation but correspond to the mass of an
impurity in the reaction mixture or on the LC-column. The fact that the mass 331.1
was most prevalent in the thiol intermediate peak from the first step supports this
theory. If this is the case, it is impossible to deduce whether the problem is the
first or the second step, and the reactions would have to be remade, followed by
purification and NMR analysis in order to elucidate this.

37



3. Results and Discussion 3.3. Linker Screen

(a)

(b)

Figure 3.17: LC-MS peak appearance from SuFEx reaction (a) 4, in iPrOH
and (b) 9, in MeCN:H2O 1:1.

Reactions 5 and 10

Figure 3.18 shows the LC-MS spectra from reactions 5 and 10, respectively. The
spectra show a clear product (1.31) and thiol intermediate (1.36) peak, representing
the cleanest results out of all reaction even if there is more thiol intermediate than
product in the reaction mixtures. A rough calculation of conversion reveals that the
reaction performed in MeCN:H2O 1:1 yielded more product (47% conversion) than
the reaction performed in iPrOH (27% conversion).

(a)

(b)

Figure 3.18: LC-MS peak appearance from SuFEx reaction (a) 5, in iPrOH
and (b) 10, in MeCN:H2O 1:1.

38



3. Results and Discussion 3.4. Conjugation

In summary, the results from the SuFEx reaction showed that only reactions 3, 5, 8
and 10 (L1 and L6) gave the SuFEx product. As opposed to the linker success from
the first step, the best linkers in the SuFEx reaction are both Michael acceptors.
With significant variability in the results using different types of linkers, a more
extensive screen of linkers, conditions and reagents is needed to elucidate if there
is an actual difference between Michael acceptor linkers and alkyl bromide linkers.
For the four reactions in which product formed, there is a difference in conditions
favoring MeCN:H2O 1:1 at 40 °C.

There is a common impurity in all reactions with the same retention time (1.77) as
the impurity found in all reactions in HTE. Again, this compound does not ionize
and can therefore not be identified via LC-MS analysis, however it is assumed to
be some kind of thiol-adduct as the thiol is the only common denominator in all
reactions; perhaps the disulfide of the thiol. Purification and NMR analysis would
be necessary to deduce which compound is giving rise to the signal, however this
was not possible in the time frame of the project. For some reactions, it constituted
a significant amount and future work to elucidate the structure would be helpful in
further understanding the reactions.

There were some problems handling the small vials and small amounts used in the
linker screen. For example, when removing the solvent from the first step under a
stream of N2, some of the reaction mixtures were blown out by the pressure from
the nitrogen stream. For reaction 10, at least half the reaction mixture was blown
out, and for reaction 7, a smaller amount was removed along with the solvent. This
did not have any apparent effect on the results from the SuFEx reaction, however,
influences are impossible to ascertain as the reactions were performed only once.

When adding the amounts of reagents to the vials in the linker screen (in both
steps), all solids were weighted out separately. In hindsight, it would have been
more accurate to prepare stock solutions of reagents that were added in amounts
less than ∼ 5 mg, such as DMAP. This solution would also have made the reaction
preparation less laborious.

3.4 Conjugation
Using the results from the high throughput experimentation and the linker screen,
the aim was to form conjugates of biologically relevant molecules using SuFEx
linkers from the project. The biologically relevant molecules used in conjugation
are presented in table 3.5 and the attempted conjugates are presented in table 3.6.
Some of the starting materials for the conjugations had to be synthesized before
conjugations could proceed. Reagents used in the synthesis of these starting materials
are presented in table 3.4 below.

The synthesis of conjugates proceeded in a one-pot, two-step reaction as for HTE
and the linker screen. The thiol addition was performed in the first step, followed
by SuFEx reaction with the corresponding amines. For all conjugates except C2

39



3. Results and Discussion 3.4. Conjugation

and C3, using a peptide, condition E from HTE was used, i.e., iPrOH, Ca(NTf2)2
and DABCO. The conjugates with the peptide used MeCN:H2O 1:1, Ca(NTf2)2 and
DABCO for solubility reasons.

Table 3.4: Reagents used in the synthesis of starting materials for conjugation.

Name Structure Molecular Weight Label
4-((2-aminoethyl)
amino)-2-(2,6-
dioxopiperidin-3-
yl)isoindoline-1,3-dione
hydrochloride

316.317 g/mol R6

2,5-dioxopyrrolidin-1-
yl
3-(pyridin-2-
yldisulfaneyl)
propanoate

312.358 g/mol R7

3’,6’-dihydroxy-3-oxo-
3H-spiro[isobenzo
furan-1,9’-xanthene]-5-
carboxylic acid

376.320 g/mol R8

2-(tritylthio)ethan-1-
amine

319.466 g/mol R9

2,5-dioxopyrrolidin-1-
yl
5-((3aS,4S,6aR)-2-
oxohexahydro-1H-
thieno[3,4-d]imidazol-
4-yl)pentanoate

341.382 g/mol R10

tert-butyl
(2-(2-(2-aminoethoxy)
ethoxy)
ethyl)carbamate

248.323 g/mol R11

40



3. Results and Discussion 3.4. Conjugation

Table 3.5: Biologically relevant molecules used in conjugation.

Molecule Structure Molecular Weight Type Label
Thiocholesterol 402.725 g/mol Thiol T1

Peptide 1 1554.853 g/mol Thiol T2

E3 ligase ligand 404.441 g/mol Thiol T3

Dye 435.450 g/mol Thiol T4

Biotin derivative 462.624 g/mol Thiol T5

Niraparib 320.396 g/mol Amine N1

Sulfaphenazole 314.363 g/mol Amine N2

Hexadecanamine 241.463 g/mol Amine N3

Flecainide 414.348 g/mol Amine N4

Mirabegron 396.509 g/mol Amine N5

Norfloxacin 319.336 g/mol Amine N6

1 Sequence: Ac-R-F-Pen-T-G-H-F-G-Sar-NMeLeu-Y-P-C-NH2

41



3. Results and Discussion 3.4. Conjugation

T
ab

le
3.

6:
C

on
ju

ga
te

s
at

te
m

pt
ed

in
th

e
pr

oj
ec

t.

La
be

l
T

hi
ol

Li
nk

er
A

m
in

e
M

ol
ec

ul
ar

W
ei

gh
t

T
hi

ol
P

ro
du

ct

M
ol

ec
ul

ar
W

ei
gh

t
Su

F
E

x
P

ro
du

ct
C

1
57

4.
89

8
g/

m
ol

87
5.

28
8

g/
m

ol

C
2

18
79

.7
64

g/
m

ol
21

73
.8

42
g/

m
ol

C
3

18
79

.7
64

g/
m

ol
21

01
.0

35
g/

m
ol

C
4

62
6.

16
3

g/
m

ol
10

20
.2

94
g/

m
ol

C
5

66
9.

69
5

g/
m

ol
10

46
.1

98
g/

m
ol

C
6

65
8.

82
0

g/
m

ol
95

8.
15

0
g/

m
ol

42



3. Results and Discussion 3.4. Conjugation

Below are the peak appearances from selected LC-MS spectra obtained from analyzing
the reaction mixtures from the conjugates.

Conjugate 1

Figure 3.19 shows the LC-MS spectrum from the thiol addition reaction of conjugate
C1. None of the peaks contain a mass corresponding to the mass of either starting
material or the thiol intermediate. The thiol intermediate is assumed not to ionize,
and would therefore not be visible in the LC-MS spectrum even if it was present
in the reaction mixture. It is thus impossible to say whether the first step worked
or not. In hindsight, it would probably have been more productive to choose an
ionizable linker in order to increase the chances of detecting the thiol intermediate.

Figure 3.19: LC-MS peak appearance of the thiol addition reaction of
conjugate C1, run in an acidic method.

Figure 3.20 shows the LC-MS spectrum from the SuFEx reaction of conjugate C1.
Because the amine, Niraparib, is both UV-active and ionizes well according to the LC-
MS spectra, the SuFEx product would too if it had formed. No mass corresponding
to the mass of the SuFEx product is found and it is thus assumed that it is not
present in the reaction mixture.

The linker (L7) was successful in yielding both the thiol intermediate and the SuFEx
product in HTE and should not be the major problem in forming the SuFEx conjugate.
The fact that Niraparib did not dissolve in the reaction mixture is probably a major
reason that the SuFEx reaction did not work, as it is unclear how much of the amine
was actually available for reaction with the thiol intermediate. However, as the
results from the first step does not show any thiol intermediate at all, there might
also be little or no thiol intermediate in the reaction mixture to begin with.

43



3. Results and Discussion 3.4. Conjugation

Figure 3.20: LC-MS peak appearance of the SuFEx reaction of conjugate
C1, run in a basic method.

Conjugate 2 and 3

Figure 3.21 shows the LC-MS spectrum from the thiol addition reaction of conjugate
C2 and C3, which have identical first steps. The thiol intermediate formation for
C2 and C3 is the most successful among all conjugates, with 82% purity and no
remaining starting materials.

Figure 3.21: LC-MS peak appearance of the thiol addition reaction of
conjugate C2 and C3, run in an acidic method.

44



3. Results and Discussion 3.4. Conjugation

Figure 3.22 shows the LC-MS spectrum from the SuFEx reaction of conjugate C2 and
C3; these were performed in the same vial. No peaks contain a mass corresponding
to the mass of either SuFEx products. C2 was attempted first and when it did not
yield the SuFEx product, C3 was attempted in the same reaction vial by adding
the corresponding amine. The peak of Sulfaphenazole was the most prominent peak
throughout the reaction. Eventually, a small peak contained a mass that corresponds
to the mass of the cyclized peptide product, meaning that an intramolecular SuFEx
reaction had occurred.

Increased complexity, such as the number of functionalities of the larger molecule,
make side reactions more probable than for small model compounds used in HTE.
However, the major reason why the desired SuFEx reactions were unsuccessful is
believed to be connected to the amines. As for C1, none of the amines dissolved
in the reaction mixture and the amount available for reaction is thus difficult to
estimate. The linker and peptide should not be fully responsible for the failed reaction
considering the success of the linker (L3) in previous experiments and the fact that
another, although undesired, SuFEx reaction occurred between linker and peptide.

C2

C3

45



3. Results and Discussion 3.4. Conjugation

Figure 3.22: LC-MS peak appearance of the SuFEx reaction of conjugate
C2 and C3, run in an acidic method.

Conjugate 4

Figure 3.23 show the LC-MS spectra from the thiol addition reaction of conjugate C4
before and after addition of base. By the look of the UV-spectra, there is significantly
more starting material than thiol intermediate in figure 3.23 (a), however the opposite
is true after addition of base. The thiol intermediate should be as UV-active as the
E3 ligase ligand and the UV-spectrum can thus be assumed to give an accurate
picture of what has occurred in the vial. It is therefore clear that base is needed for
the reaction to proceed.

(a) (b)

Figure 3.23: LC-MS peak appearance of the thiol addition reaction of
conjugate C4 (a) before addition of base and (b) after addition of base, both
run in an acidic method.

Figure 3.24 shows the LC-MS spectrum from the SuFEx reaction of conjugate C4.
No peak contains a mass that corresponds to the mass of the product. The fact
that Fleacinide did not dissolve in the reaction mixture and that the linker (L4) was

46



3. Results and Discussion 3.4. Conjugation

slightly problematic in HTE (especially in the first step), can be reasons that the
SuFEx reaction was unsuccessful for C4.

Figure 3.24: LC-MS peak appearance of the SuFEx reaction of conjugate
C4, run in an acidic method.

Conjugate 5

Figure 3.25 shows the LC-MS spectrum from the thiol addition reaction of conjugate
C5, before and after addition of base. Both starting material and thiol intermediate
are assumed to be equally UV-active and the UV-spectra shown in figure 3.25 are
assumed to accurately represent the reaction status. As for the first step of C4, it is
clear that this reaction requires base in order to proceed.

It is interesting to note that even though the linkers used for C2 and C3 (L3), C4
(L4) and C5 (L1) are all Michael acceptors and should display similar trends, the
first step of C3 proceeded without addition of base, and gave the purest result out of
all conjugates. Previous experiments also show a superior behavior of L3 compared
to L4 and L1. This might be attributed to the reactivity of the linker, with L3 being
a maleimide and L4 and L1 being activated double- and triple bonds. L4 and L1
might have been more successful in previous experiments if base had been added to
the reaction mixtures.

47



3. Results and Discussion 3.4. Conjugation

(a) (b)

Figure 3.25: LC-MS peak appearance of the thiol addition reaction of
conjugate C5 (a) before addition of base and (b) after addition of base, both
run in a basic method.

Figure 3.26 shows the LC-MS spectrum from the SuFEx reaction of conjugate C5
after having stirred at 60 °C over the weekend. The second step of C5 also did not
yield the SuFEx product, however, the linker (L1) should not be responsible for this.
Even though it was a bit problematic during the first step of the linker screen, it
was one of two linkers that yielded the SuFEx product at all. Also in this reaction
mixture, the amine did not dissolve, which makes it difficult to assess whether there
is a problem with reactivity or simply a need for optimizing reaction conditions.
Some changes to the reaction conditions were attempted for C1, C3 and C5 during
the reactions, such as increased temperature and addition of other solvents, however
without positive results.

48



3. Results and Discussion 3.4. Conjugation

Figure 3.26: LC-MS peak appearance of the SuFEx reaction of conjugate
C5, run in a basic method.

Conjugate 6

Figure 3.27 shows the LC-MS spectrum from the thiol addition reaction of conjugate
C6. Due to time limitations, the first step of C6 could only run for approximately
5 hours, and to be able to perform the SuFEx reaction at all, it had to be started
regardless of whether the results from the thiol addition showed the thiol intermediate
in LC-MS analysis or not. Luckily, the thiol intermediate had formed already after 1
hour, however, the biotin derivative still remained in the reactions mixture.

Figure 3.27: LC-MS peak appearance of the thiol addition reaction of
conjugate C6, run in an acidic method.

49



3. Results and Discussion 3.4. Conjugation

Figure 3.28 shows the LC-MS spectrum from the SuFEx reaction of conjugate C6.
As for the other conjugates, the SuFEx reaction yielded no product. The thiol
intermediate is however no longer visible in LC-MS after the seconds step, which
could mean that it was consumed in an undesired pathway. The linker used for the
final conjugate has not produced a SuFEx product in a previous experiment as the
results from the linker screen showed nothing but the impurity at 1.77 minutes. It
was however successful in the corresponding first step, and it is therefore improbable
that it would be the major problem in the SuFEx reaction.

Figure 3.28: LC-MS peak appearance of the SuFEx reaction of conjugate
C6, run in a basic method.

In summary, LC-MS spectra of the first step showed the thiol intermediate for all
conjugates except C1, but did not show the desired SuFEx product for any conjugates.
One reason that the SuFEx reactions were unsuccessful could be solubility issues.
The solvents did not dissolve all components in any reaction during the second step,
and the amines were especially difficult, together with Ca(NTf2)2. Small amounts of
reagents and low solubility increase the uncertainty of the actual amounts available
for reaction with the thiol intermediate. Additionally, the increasing complexity
that comes with larger molecules, such as more functionalities, makes side reactions
and intramolecular reactions more probable, as with the SuFEx cyclization of the
peptide. Furthermore, the reactions were only attempted once at small scale, and
other solvents, bases or temperatures might have been more successful. Some changes
to the conditions were attempted, e.g., by addition of other solvents or by increasing
the temperature, however this had no positive results. This is something that could
be further explored in future studies.

50



4. Conclusion and Outlook

Six sulfamoyl fluoride linkers with different thiol accepting groups were successfully
synthesized in this project using a solid imidazolium salt for installation of the
sulfamoyl fluoride group. The linker backbones were both rigid and flexible and
the thiol accepting groups were of two kinds: Michael acceptors and alkyl bromides,
where the Michael acceptors included an acrylate, maleimides, an alkyne and an
alkene. The linkers are presented in table 4.1.

Table 4.1: SuFEx linkers synthesized in the project.

Label Structure Label Structure
L1 L4

L2 L5

L3 L6

Synthesized SuFEx linkers were compared with commercially available linkers in
a high throughput format which provided proof of concept: the SuFEx linkers
successfully yield a thiol intermediate and a SuFEx product with a model thiol
and three model amines. The general reaction performed in HTE is displayed in
chapter 4. However, the success of the reactions depended on linker, amine and the
conditions used. The commercially available linkers used in HTe are presented in
chapter 4. Linkers L3 and L7 were the best linkers across all amines and conditions.
L4 did not convert to thiol intermediate completely during the first step, and L8 did
retro-Michael reaction of the thiol.

Figure 4.1: The general reaction performed in HTE.

51



4. Conclusion and Outlook 4. Conclusion and Outlook

(a) (b)

Figure 4.2: Commercially available SuFEx linkers used in HTE; (a) L7 and
(b) L8.

A linker screen was performed, which revealed that the synthesized linker L6 was the
only linker yielding the thiol intermediate without remaining linker and simultaneously
yielding the SuFEx product. L1 was the only other linker yielding the SuFEx product,
however this linker did not completely convert to the thiol intermediate during the
first step. No linkers converted to SuFEx product completely.

SuFEx linkers successfully yielded the thiol intermediate for four different conjugates
of biologically relevant molecules. However, no SuFEx reactions yielded the desired
SuFEx conjugates between the thiol intermediates and corresponding amines. This
is assumed to be due to solubility issues and sub-optimal reaction conditions.

Further research needs to be done to elucidate why the SuFEx reaction was unsuc-
cessful for all conjugates. Investigating the reactivity of the different components
in simpler environments could elucidate if there is a problem with certain com-
pounds. An HTE screen of only the SuFEx reaction using a simple SuFEx molecule,
phenylsulfonyl fluoride, with some of the amines used in conjugation is planned
(figure 4.3). Additionally, purifying and analyzing the different components in the
reaction mixtures from the HTE and linker screens would give valuable insight in
whether or not side reactions occurred, and would also allow for a quantitative result
on how much product formation there was.

Figure 4.3: Exploring only the SuFEx step in a planned HTE screen.

It is possible that optimization of reaction conditions could provide a solution,
considering the solubility issues during the experiment. Conditions using nucleophilic
catalysis such as HOBT in DMSO have been reported and could potentially be tested
in the future [49], [50].

Figure 4.4: Possible conditions to try in future SuEFx reactions [49], [50].

52



4. Conclusion and Outlook 4. Conclusion and Outlook

A recent paper show how similar SuFEx linkers were used in covalent targeting of
proteins via a SuFEx reaction, indicating that modification of conditions and linkers
can lead to a positive conclusion [51].

53



5. Experimental

5.1 Synthesis of SuFEx linkers
Below are the procedures of the synthesis of all SuFEx linkers. NMR summaries are
attached, but complete NMR data with spectra can be found in appendix B.

tert-butyl 4-(fluorosulfonyl)piperazine-1-carboxylate I1
In a 100 mL round bottom flask tert-butyl piperazine-1-carboxylate AM1 (1.26 g,
6.78 mmol) was dissolved in acetonitrile (25 mL). 1-(fluorosulfonyl)-2,3-dimethyl-1H-
imidazol-3-ium trifluoromethanesulfonate SM1 (2.67 g, 8.14 mmol) was added and
the solution was stirred overnight at room temperature. Water (100 mL) was added
and the mixture was extracted with EtOAc (3x150 mL). The combined organic
phase was washed with water (100 mL) with some added brine and with brine (100
mL), dried over Na2SO4 and filtered. The solvent was removed din vacuo to yield
tert-butyl 4-(fluorosulfonyl)piperazine-1-carboxylate I1 (1.72 g, 95 %) as a colorless
solid.

LCMS: product did not ionize.

1H NMR (500 MHz, CDCl3, 25°C) δ 1.47 (s, 9H), 3.39–3.45 (m, 4H), 3.54–3.6 (m,
4H).
13C NMR (126 MHz, CDCl3) δ 154.2, 81.1, 47.0, 42.7, 28.5.
19F NMR (471 MHz, CDCl3) δ 39.7.

piperazine-1-sulfonyl fluoride, I2a
In a 10 mL round bottom flask tert-butyl 4-(fluorosulfonyl)piperazine-1-carboxylate
(100 mg, 0.37 mmol) I1 was dissolved in Hydrogen chloride solution 4.0 M in Dioxane
(2 mL, 57.60 mmol) and the solution was stirred at room temperature for 1 hour.
The solvent was removed in vacuo and the residue was used in the next step without
further treatment.

LCMS (m/z, ES+) mass found = 169.0 [M+H]+, expected = 169,04 [M+H]+ for
[C4H10FN2O2S+].

54



5. Experimental 5.1. Synthesis of SuFEx linkers

piperazine-1-sulfonyl fluoride, I2b
In a 100 mL round bottom flask (3-acrylamidopropyl)sulfamoyl fluoride (1.00 g, 3.73
mmol) was dissolved in DCM (15 mL) and 2,2,2-trifluoroacetic acid (15 mL, 194
mmol) was added. The solution was stirred at room temperature for 3 hours. The
solvent was removed in vacuo and the residue was used in the next step without
further treatment.

LCMS (m/z, ES+) mass found = 169.1 [M+H]+, expected = 169,04 [M+H]+ for
[C4H10FN2O2S+].

tert-butyl (3-((fluorosulfonyl)amino)propyl)carbamate, I3
In a 50 mL round bottom flask tert-butyl (3-aminopropyl)carbamate AM2 (1.5 mL,
8.61 mmol) was dissolved in acetonitrile (25 mL) and the solution was cooled down
to 0° C. 1-(fluorosulfonyl)-2,3-dimethyl-1H-imidazol-3-ium trifluoromethanesulfonate
SM1 (3.39 g, 10.33 mmol) was added and the solution was stirred overnight while
slowly warming up to room temperature. Water (100 mL) was added and the
mixture was extracted with EtOAc (3x150 mL). The combined organic phase was
washed with water (100 mL) with some added brine and with brine (100 mL), dried
over Na2SO4 and filtered. The solvent was removed in vacuo to yield tert-butyl
(3-((fluorosulfonyl)amino)propyl)carbamate I3 (2.19 g, 99 %) as a colorless oil.

LCMS (m/z, ES-) mass found = 255.1 [M-H]-, expected = 255,08 [M-H]- for
[C8H16FN2O4S−].

1H NMR (500 MHz, CDCl3, 25°C) δ 1.44 (s, 9H), 1.72 (p, J = 5.9 Hz, 2H), 3.30
(ddt, J = 14.7, 12.1, 6.3 Hz, 4H), 4.78 (s, 1H), 7.09 (s, 1H).
13C NMR (126 MHz, CDCl3) δ 157.9, 80.7, 41.0, 36.5, 30.2, 28.4.
19F NMR (471 MHz, CDCl3) δ 52.1.

(3-aminopropyl)sulfamoyl fluoride, I4
In a 100 mL round bottom flask tert-butyl (3-((fluorosulfonyl)amino)propyl)carbamate
I2 (1.00 g, 3.90 mmol) was dissolved in DCM (15 mL) and 2,2,2-trifluoroacetic acid
(15 ml, 194 mmol) was added. The solution was stirred at room temperature for 3
hours. The solvent was removed in vacuo and the residue was used in the next step
without further treatment.

55



5. Experimental 5.1. Synthesis of SuFEx linkers

4-(but-2-ynoyl)piperazine-1-sulfonyl fluoride, L1
In a 10 mL round bottom flask piperazine-1-sulfonyl fluoride I2 (62.7 mg, 0.37 mmol)
and but-2-ynoic acid R2 (34.5 mg, 0.41 mmol) were dissolved in DCM (3 mL) and
the mixture was cooled down to 0 °C. N-ethyl-N-isopropylpropan-2-amine (0.33 mL,
1.86 mmol) was added, followed by a dropwise addition of 2,4,6-tripropyl-1,3,5,2,4,6-
trioxatriphosphinane 2,4,6-trioxide (0.28 mL, 0.93 mmol). The solution was stirred
at 0 °C for 10 minutes and then at room temperature for 2 hours. Sat. NaHCO3
(20 mL) was added and the mixture was extracted with EtOAc (3x20 mL). The
combined organic phase was washed with 1M HCl solution (20 mL) and with brine
(20 mL), dried over Na2SO4 and filtered. The solvent was removed in vacuo to yield
4-(but-2-ynoyl)piperazine-1-sulfonyl fluoride L1 (72.5 mg, 83 %) as a colorless solid.

LCMS (m/z, ES+) mass found = 235.2 [M+H]+, expected = 235,05 [M+H]+ for
[C8H12FN2O3S+].

1H NMR (500 MHz, CDCl3, 25°C) δ 2.04 (d, J = 0.6 Hz, 3H), 3.42–3.52 (m, 4H),
3.76–3.92 (m, 4H).
13C NMR (126 MHz, CDCl3) δ 153.2, 91.3, 72.3, 47.3, 46.7, 45.7, 40.2, 4.3.
19F NMR (471 MHz, CDCl3) δ 40.5.

(3-(2-bromoacetamido)propyl)sulfamoyl fluoride, L2
In a 100 mL round bottom flask (3-aminopropyl)sulfamoyl fluoride I4 (609 mg, 3.90
mmol) was dissolved in acetonitrile (30 mL). 2,5-dioxopyrrolidin-1-yl 2-bromoacetate
R1 (1.84 g, 7.80 mmol) was added, followed by a slow addition of N-ethyl-N-
isopropylpropan-2-amine (1.36 mL, 7.80 mmol). The solution was stirred at room
temperature overnight. The solvent was removed in vacuo and the residue was
purified on flash column chromatography using a gradient of 12-100% ethyl acetate
in heptane. Fractions containing the product were combined and the solvent was
removed in vacuo. (3-(2-bromoacetamido)propyl)sulfamoyl fluoride L2 (247 mg,
22.85 %) was obtained as a colorless oil.

LCMS (m/z, ES-) mass found = 274.9 [M-H]-, expected = 274,95 [M-H]- for
[C5H9BrFN2O3S−].

1H NMR (500 MHz, DMSO, 25°C) δ 1.67 (p, J = 7.1 Hz, 2H), 3.09–3.21 (m, 4H),
3.83 (s, 2H), 8.31 (d, J = 5.9 Hz, 1H), 9.35 (q, J = 5.5 Hz, 1H).
13C NMR (126 MHz, CDCl3) δ 167.9, 41.1, 36.7, 29.4, 28.8.
19F NMR (471 MHz, DMSO) δ 50.9.

56



5. Experimental 5.1. Synthesis of SuFEx linkers

3-((2-(2,5-dioxo-2,5-dihydro-1H-pyrrol-1-yl)ethyl)carbamoyl)benzene sul-
fonyl fluoride, L3
In a 25 mL round bottom flask 1-(2-aminoethyl)-1H-pyrrole-2,5-dione hydrochloride
AM3 (100 mg, 0.57 mmol) and 3-(fluorosulfonyl)benzoic acid R4 (127 mg, 0.62
mmol) were dissolved in DCM (5 mL) and the mixture was cooled down to 0 °C.
N-ethyl-N-isopropylpropan-2-amine (0.49 mL, 2.83 mmol) was added, followed by a
dropwise addition of 2,4,6-tripropyl-1,3,5,2,4,6-trioxatriphosphinane 2,4,6-trioxide
(0.42 mL, 1.42 mmol). The solution was stirred at 0 °C for 10 minutes and then
at room temperature for 1 hour. Sat. NaHCO3 (40 mL) was added and the mix-
ture was extracted with EtOAc (3x30mL). The combined organic layer was washed
with 1M HCl solution (30 mL) and brine (30 mL), dried over Na2SO4 and filtered.
The solvent was removed in vacuo to yield 3-((2-(2,5-dioxo-2,5-dihydro-1H-pyrrol-
1-yl)ethyl)carbamoyl)benzenesulfonyl fluoride L3 (170 mg, 92 %) as a colorless
solid.

LCMS (m/z, ES+) mass found = 327.2 [M+H]+, expected = 327,04 [M+H]+ for
[C13H12FN2O5S+].

1H NMR (500 MHz, DMSO, 25°C) δ 3.4–3.47 (m, 2H), 3.61 (dd, J = 6.5, 4.9 Hz,
2H), 7.02 (s, 2H), 7.90 (t, J = 7.9 Hz, 1H), 8.25–8.33 (m, 2H), 8.41 (t, J = 1.8 Hz,
1H), 9.01 (t, J = 5.9 Hz, 1H).
13C NMR (126 MHz, DMSO) δ 171.1, 164.1, 136.1, 135.1, 134.6, 132.0, 131.8, 130.9,
130.8, 126.6, 37.9, 36.9.
19F NMR (471 MHz, DMSO) δ 66.4.

4-acryloylpiperazine-1-sulfonyl fluoride, L4
In a 100 mL round bottom flask piperazine-1-sulfonyl fluoride I2b (627 mg, 3.73
mmol) was dissolved in DCM (30 mL) and the solution was cooled down to 0 °C.
acryloyl chloride R3 (0.30 mL, 3.73 mmol) was added, followed by a slow addition of
N-ethyl-N-isopropylpropan-2-amine (1.30 mL, 7.46 mmol). The solution was stirred
at 0 °C for 30 minutes and was then allowed to slowly warm up to room temperature.
The solution was stirred until completion (1.5 hours). The solvent was removed in
vacuo and the residue was purified on flash column chromatography using a gradient
of 12-100% ethyl acetate in heptane. Fractions containing the product were combined
and the solvent was removed in vacuo. 4-acryloylpiperazine-1-sulfonyl fluoride L4
(534 mg, 64.5 %) was obtained as a colorless solid.

LCMS (m/z, ES+) mass found = 223.1 [M+H]+, expected = 223,05 [M+H]+ for
[C7H12FN2O3S+].

57



5. Experimental 5.1. Synthesis of SuFEx linkers

1H NMR (500 MHz, DMSO, 25°C) δ 3.44–3.48 (m, 4H), 3.65–3.75 (m, 4H), 5.74 (dd,
J = 10.5, 2.3 Hz, 1H), 6.14 (dd, J = 16.7, 2.3 Hz, 1H), 6.79 (dd, J = 16.7, 10.5 Hz,
1H).
13C NMR (126 MHz, DMSO) δ 164.5, 128.1, 127.9, 46.9, 46.5, 43.9.
19F NMR (471 MHz, DMSO) δ 41.2.

(2-(2,5-dioxo-2,5-dihydro-1H-pyrrol-1-yl)ethyl)sulfamoyl fluoride, L5
In a 25 mL round bottom flask1-(2-aminoethyl)-1H-pyrrole-2,5-dione hydrochlo-
ride AM1 (177 mg, 1 mmol) was dissolved in dichloromethane (5 mL) and 1-
(fluorosulfonyl)-2,3-dimethyl-1H-imidazol-3-ium trifluoromethanesulfonate SM1 (492
mg, 1.50 mmol) was added in one portion. triethylamine (0.14 mL, 1.00 mmol)
was added drop-wise and the mixture was stirred for 1 hour at room temperature.
The solvent was evaporated and the residue was dissolved in EtOAc (20 mL) and
the mixture was washed with 1 M HCl (3x20 mL) and brine (1x20 mL), dried over
Na2SO4, and filtered. The solvent was evaporated and the residue was dissolved in
DCM and purified by flash column chromatography using a gradient of 12-100%
EtOAc in heptane. Fractions containing the product were combined and the solvent
was evaporated in vacuo. (2-(2,5-dioxo-2,5-dihydro-1H-pyrrol-1-yl)ethyl)sulfamoyl
fluoride L5 (48.9 mg, 22.01 %) was obtained as a colorless solid.

LCMS (m/z, ES-) mass found = 221.0 [M-H]-, expected = 221,00 [M+H]+ for
[C6H6FN2O4S−].

1H NMR (500 MHz, CDCl3, 25°C) δ 3.5–3.57 (m, 2H), 3.77–3.84 (m, 2H), 5.61 (s,
1H), 6.79 (s, 2H).
13C NMR (126 MHz, CDCl3) δ 170.9, 134.6, 43.9, 37.3.
19F NMR (471 MHz, CD3CN) δ 50.0.

4-(2-bromoacetyl)piperazine-1-sulfonyl fluoride, L6
In a 25 mL round bottom flask piperazine-1-sulfonyl fluoride I2a (125 mg, 0.74
mmol) and 2-bromoacetic acid R5 (0.059 mL, 0.82 mmol) were dissolved in DCM (5
mL) and the mixture was cooled down to 0 °C. N-ethyl-N-isopropylpropan-2-amine
(0.65 mL, 3.72 mmol) was added, followed by a dropwise addition of 2,4,6-tripropyl-
1,3,5,2,4,6-trioxatriphosphinane 2,4,6-trioxide (0.55 mL, 1.86 mmol). The solution
was stirred at 0 °C for 10 minutes and then at room temperature for 1.5 hours. Sat.
NaHCO3 (40 mL) was added and the mixture was extracted with EtOAc (3x40 mL).
The combined organic phase was washed with 1M HCl solution (40 mL) and brine
(40 mL), dried over Na2SO4 and filtered. The solvent was removed in vacuo to yield
4-(2-bromoacetyl)piperazine-1-sulfonyl fluoride L6 (141 mg, 65.8 %) as a colorless
solid.

58



5. Experimental 5.2. High Throughput Experimentation

LCMS (m/z, ES+) mass found = 289.1 [M+H]+, expected = 288,97 [M+H]+ for
[C6H11BrFN2O3S+].

1H NMR (500 MHz, CDCl3, 27°C) δ 3.45–3.59 (m, 4H), 3.66–3.82 (m, 4H), 4.08 (s,
2H).
13C NMR (126 MHz, CDCl3) δ 165.4, 47.0, 46.8, 45.4, 41.0, 40.5.
19F NMR (471 MHz, CDCl3) δ 40.5.

5.2 Hig