High-Affinity Copolymers Inhibit Digestive Enzymes by Surface Recognition

Article abstract of DOI:10.1021/acs.biomac.7b00162

This account presents a general method for the construction of polymeric surface binders for digestion enzymes. Two prominent parts, namely, the modification of the copolymer composition and the screening assay for the most powerful inhibitors are both am

Full text of DOI:10.1021/acs.biomac.7b00162

Subscriber access provided by HACETTEPE UNIVERSITESI KUTUPHANESI
Article
High-Affinity Copolymers Inhibit Digestive Enzymes by Surface Recognition
Patrick Gilles, Kirstin Wenck, Inga Stratmann, Michael Kirsch, Daniel A.
Smolin, Torsten Schaller, Herbert de Groot, Arno Kraft, and Thomas Schrader
Biomacromolecules, Just Accepted Manuscript • DOI: 10.1021/acs.biomac.7b00162 • Publication Date (Web): 18 Apr 2017
Downloaded from http://pubs.acs.org on April 24, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Biomacromolecules is published by the American Chemical Society. 1155 Sixteenth
Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 26
Biomacromolecules
1
2
3
4
5
6
7
8
9
HighꢀAffinity Copolymers Inhibit Digestive
Enzymes by Surface Recognition
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Patrick Gilles,‡ Kirstin Wenck,‡ Inga Stratmann,‡ Michael Kirsch,† Daniel A. Smolin,‡ Torsten
Schaller, ‡ Herbert de Groot, † Arno Kraft,§ Thomas Schrader*‡
‡
†University of DuisburgꢀEssen, Faculty of Medicine; University of DuisburgꢀEssen, Faculty of
§
Chemistry, Universitaetsstr. 7, 45117 Essen, Germany; HeriotꢀWatt University, Edinburgh,
Scotland, UK EH14 4AS
KEYWORDS. Proteases, Hydrolases, Enzyme Inhibition, Supramolecular Chemistry, Radical
Copolymerization
ABSTRACT.
This account presents a general method for the construction of polymeric surface binders for
digestion enzymes. Two prominent parts, namely the modification of the copolymer composition
and the screening assay for the most powerful inhibitors are both amenable to parallelization.
The concept hinges on the appropriate selection of aminoꢀacidꢀselective comonomers, their free
radical copolymerization, and subsequent screening of the resulting copolymer library for
efficient enzyme inhibition. A microscale synthetic procedure for the copolymerization process
was developed, which produces waterꢀsoluble affinity polymers that can be stored for years at
room temperature. Initial parallel screening was conducted in standard enzyme assays to identify
polymeric inhibitors, which were subsequently subjected to determination of IC50 values for their
target enzyme. For all digestion enzymes except elastase a number of polymer inhibitors were
found, some of which were selective towards one or two protein targets. Since the key monomers
of the best inhibitors bind to amino acid residues in the direct vicinity of the active site, we
conclude that efficient coverage of the immediate environment by the copolymers is critical.
Strong interference with enzymatic activity is brought about by blocking the substrate access and
product exit to and from the active site.
Introduction.
Many proteins operate by the same common mechanism and hence share very similar active
sites. Thus, serine proteases all contain the catalytic triad AspꢀHisꢀSer and an oxyanion hole, and
achieve specificity mainly through additional binding pockets for their substrates. The classical
approach for the development of potent enzyme inhibitors involves rigid substrate mimetics
ACS Paragon Plus Environment
1

Biomacromolecules
Page 2 of 26
1
2
3
4
5
6
7
8
9
which carry tailored substituents for the optimal filling of specificity pockets. A very elegant
example is the development of a thrombine inhibitor in the Diederich group, which occupies the
S1ꢀ, Dꢀ and Pꢀpocket and thus achieves very high affinity and thrombin/trypsin selectivity.1 The
preparation of this lowꢀmolecularꢀweight drug, however, requires a multistep enantioꢀselective
synthesis. Moreover, if an enzyme does not offer large pockets, which are characteristic in shape
and charge distribution, the development of enzymeꢀspecific active site inhibitors becomes
problematic.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In those cases it would be advantageous to overcome the limitation of active site targeting and
include the protein surface for molecular recognition and inhibition. Quite often the size,
topology and amino acid distribution of similar serine proteases differ greatly and create
opportunities for their selective inhibition. In this respect the development of linear, branched or
crossꢀlinked affinity copolymers is an attractive approach, because these can provide large
surface areas and offer a number of different binding sites within one molecule, exploiting the
concept of multivalency and heteroꢀavidity.2 If surface recognition is thus combined with steric
hindrance to the entrance of the active site, a selected enzyme will stop to process its substrate.
Such a supramolecular inhibition mode will be reversible in the presence of a competitor.
However, it requires multipoint binding of the copolymer on the enzyme surface.
Various new concepts have evolved in recent years which focus on tailored dendrimers and
polymers for protein surface recognition: Haag developed dendrimeric polyglycerol sulfates
(dPGS) for multivalent Selectin binding on leukocytes for the control of viral inꢀfections.3 Chiral
ruthenium(II) trisbipyridine complexes with up to six Lꢀaspartyl side chains in the ligand have
been found to bind to the active site of proteases, making it possible to inhibit chymotrypsin with
high efficiency. (Wilson, Ohkanda).4ꢀ5 Detection and identification of proteins has been
accomplished with nanoparticle–fluorescent polymer ‘chemical nose’ sensors (Rotello), whereas
selective sensing of metalloꢀproteins could be achieved from nonselective binding using a
fluorogenic amphiphilic polymer (Thayumanavan).6ꢀ7
Lightly crosslinked nanoparticles recognize protein surfaces and exploit multivalency of
unspecific noncovalent attraction (Shea).8 Molecularly imprinted hydrogels add the templateꢀ
effect in order to achieve improved affinity and selectivity (Haupt et al.).9 Inspired by this
pioneering work, coreꢀshell particles were developed, which combine immobilization of an
anchor monomer and selfꢀassembly of the protein template (Zhang).10 Various new methods
have been recently developed, exploiting the concept of molecular imprinting by peptidic and
protein templates.11ꢀ15 However, all these approaches have hitherto been limited to conventional
monovalent binding monomers and lack the element of specific molecular recognition.
Our group adds the following supramolecular component: over the past years, we have designed
a large variety of amino acidꢀselective binding monomers. A simple conventional radical
copolymerization of selected monomers gives access to new waterꢀsoluble materials which
ACS Paragon Plus Environment
2

Page 3 of 26
Biomacromolecules
1
2
3
4
5
6
7
8
9
display high affinities through multivalency16 and surprising protein specificity by cooperative
binding of the appropriate functional monomers.17 Molecular recognition of characteristic amino
acid residues on the surface of lysozyme then leads to efficient enzyme inhibition, which can be
switched on and off.18 Very recently we could show that the correct monomer choice also allows
epitopeꢀselective binding, e.g. of the Fc part in IgGs.19
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Serine proteases are digestion enzymes which operate in the intestines and provide a challenging
medicinally relevant target. Their upregulation inside the pancreas after shock, during cardiac
surgery or organ transplants can lead to ischemia and reperfusion lesions.20 These in turn may
cause acute pancreatitis, a lifeꢀthreatening emergency.21 Trypsin holds a key role because it
activates other hydrolases whose uncontrolled action may ultimately damage multiple organs.22
Treatment of such a condition with antiꢀsecretory hormones23 or protease inhibitors is
unsatisfactory, most likely because most of them act unspecifically on proteases of similar
structure.24
Experimental Section.
Materials. All reagents and proteins were used as received. Trizma Base, BAPNA,
dodecylamine were purchased from Fluka (Buchs, Switzerland), dry solvents from Acros
Organics (geel, Belgium), HꢀLys(Z)ꢀOtBu HCl from Bachem (Bubendorf, Switzerland) and
deuterated solvents were purchased from Deutero (Kastellaun, Germany). All other chemicals
were purchased from Sigma Aldrich (Steinheim, Germany).
Methods. Nuclear Magnetic Resonance (NMR). 1H and 31P spectra were recorded with a Bruker
1
13
DMX 300, further H and C spectra were recorded with a Bruker DRX 500 using deuterated
solvents. Residual solvent peak was used as an internal standard for peak calibration.
Size Exclusion Chromatography (SEC). Polymer molecular weights were determined by size
exclusion chromatography with a JASCO PUꢀ980 pump and refractive index detector RIꢀ930.
For cationic polymers two consecutive PSS NOVEMA Max 300 x 80 mm columns (PSS, Mainz,
Germany) were used. Anionic polymers were characterized over two consecutive PSS HEMA
Bio linear 300 x 80 mm columns (PSS, Mainz, Germany). As standards pullulanes with different
molecular weights were used (PSS, Mainz, Germany).
Mass spectrometry. Mass spectra were recorded with a Bruker maXis 4G and with a Bruker
BioTOF III time of flight mass detector with ESI ion source.
Isothermal Titration Calorimetry (ITC). ITC measurements were recorded with a VPꢀITC
calorimeter from MicroCal.
Lyophylization. Synthesized polymers were lyophilized using a device from Christ, Model Alpha
2ꢀ4 LSC.
ACS Paragon Plus Environment
3

Biomacromolecules
Page 4 of 26
1
2
3
4
5
6
UV/Vis and fluorescence spectroscopy. The absorption spectra were recorded with a Tecan
Infinite M200 Microplates reader. The used NonꢀTissue CultureꢀTreated Plate, 96ꢀwell, flat
bottom with low evaporation lid (polystyrene) purchased from BD Falcon.
7
8
9
Parallel copolymerization. For the parallel polymerization a HLCꢀHeating MHL 23 thermomixer
from Ditabis were used.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General procedure for the parallel free radical copolymerization. Stock solutions of all
monomers were degassed according to the „freezeꢀandꢀexhaust” method. Under argon these
solutions were transferred into a 2 mL polymerization vessel and 5 mol % AIBN were added.
The resulting total volume never exceeded 1300 µL. Subsequently the polymerization was
conducted in a thermomixer at 70 °C und 600 rpm. After 24 ꢀ 30 h the polymerization solutions
were transferred to a small conical shaped flask and subjected to lyophilization.
The resulting residue was dissolved in little water and transferred into a small centrifugal filter
unit (Amicon Ultra 2mL 3K Ultracel ꢀ 3K Membrane, Millipore). Ultrafiltration was carried out
three times at 4000 rpm for 60 min, then a fourth ultracentrifugation followed at 4000 rpm for a
total of 100 min using a Eppedorf Centrifuge 5706. The resulting oligomerꢀ and monomerꢀfree
solution was subsequently transferred into a 2mL Eppendorf tube and subjected again to
lyophilization. The voluminous colorless lyophylizate was further examined: NMR spectra
indicated the degree of conversion and the stoichiometric ratio of comonomers inside the final
copolymer. SEC was used for the molecular weight determination of the polymers.
Enzyme assays. General remarks. All substrate and enzyme solutions were freshly prepared and
stored at 0°C during the measurements. Stock solutions of copolymers had a mass concentration
of 20 mg/mL in doubly distilled water. Measurements were conducted in 96ꢀwell microplates.
Before each assay, absorption spectra of substrate, enzyme and chromophore were obtained, in
order to determine the absorption maximum of the chromophore and to secure that absorption of
enzyme and substrate was negligible at the selected wavelength.
In order to validate each method, a known reference inhibitor was tested, its IC50 value was
determined and compared to the literature value. Subsequently, a screening was conducted in
triplicate with all polymers. Each NTAꢀcontaining polymer was measured in the absence and in
the presence of 50 molꢀ% Ni2+ ions with respect to the NTA amount within the polymer. If an
inhibitory effect was observed for a polymer within the error limits of the screening, its IC50
value was subsequently determined. To this end, the enzyme velocity was measured for 10
different polymer concentrations; each experiment was carried out in triplicate (see more details
in the Supporting Information).
ACS Paragon Plus Environment
4

Page 5 of 26
Biomacromolecules
1
2
3
4
Results and Discussion.
5
6
7
8
We asked ourselves whether or not it would be possible to design a synthetic polymer which
would be capable of distinguishing, e. g., between trypsin and chymotrypsin and lead to efficient
inhibition simply by binding to the protein surface? Thus we embarked on a program aimed at
the discovery of highly specific polymeric inhibitors for diseaseꢀrelevant proteases (Fig. 1). For
economic reasons, this required the development of a general method for the parallel synthesis of
affinity copolymers in small amounts. In addition, we needed to probe their interaction with
digestion enzymes. This was realized in three stages: synthesis of new binding monomers,
development of a small scale combinatorial polymerization, and an assay for the parallel
screening of target enzymes.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Schematic representation of linear affinity copolymers recognizing the surface of serine
proteases. A A benzamidine comonomer binds to aspartates and glutamates on the surface and within the
active site. B A bisphosphonate comonomer selectively recognizes arginines on the protein surface – it is
supported by an anchor monomer with a terminal alanine unit filling the specificity pocket of elastase.
For this study we chose a representative series of seven digestive enzymes involved in
pancreatitis and intestinal injuries. Inspection of all protein surfaces revealed very different sizes,
topologies and amino acid compositions ꢀ ideal to test the polymers’ potential to distinguish
between protein surfaces within the same family of enzymes. Table 1 displays the protein sizes
and charges together with their main cleavage sites and typical inhibitors.
Amino
Acid
Residues
MW
[kD]
Cleavage
Site
Protein
Mechanism
pI
pH Optimum Inhibitor
ACS Paragon Plus Environment
5

Biomacromolecules
Page 6 of 26
1
2
3
4
5
6
7
8
9
Serine
protease
Serine
protease
Serine
protease
α1ꢀAntitrypsin
Aprotinin
Ala, Gly
Phe, Tyr, Trp
Arg, Lys
Phe, Trp, Leu
Arg
Elastase
240
241
233
307
619
308
346
26
25
24
35
86
37
48
8.8
8.7
10.1
6.0
4.5
7.2
6.0
8.5
αꢀChymotrypsin
Trypsin
7.8
pꢀAminoꢀ
7.5ꢀ8.5
7.5
benzamidine
Benzomercapto
ꢀpropanoic acid
Zinc protease
Carboxypeptidase A
Kallikrein
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Serine
protease
Serine
protease
Aspartate
protease
Aprotinin
7.0ꢀ8.0
7.5ꢀ8.5
3.5ꢀ5.0
Argatroban
Pepstatin A
Arg
Thrombin
PheꢀPhe
Cathepsin D
Table 1. Overview of seven selected digestion enzymes for pancreatitis and intestinal injury, illustrating
their diverse sizes, charges and pI values.
Binding monomers. For each major class of amino acid residues we developed a series of
methacrylamides which carry a specific binding site ꢀ the current status of this library is
presented in Scheme 1: Lysine and arginine are recognized by a methacrylamideꢀbased
bisphosphonate dianion (BP, C5ꢀBP, C6ꢀBP); serine, threonine as well as diols in glycosylated
proteins can form cyclic esters with aminomethylphenylboronic acid units (Bor), whereas nonꢀ
polar amino acids can interact with cyclohexyl or dodecyl methacrylamides (Dod, C6ꢀCyH).
Further inclusion of a glucosamineꢀbased monomer (Glu) is beneficial because it greatly
enhances water solubility, and a dansyl moiety (Dan) provides a fluorescence label for
convenient detection in titration or microscopy experiments.
In order to complete the series, several new binding monomers were synthesized for acidic
amino acids, aromatic amino acids and histidine. A guanidiniocarbonylpyrrole, previously
introduced by Schmuck et al. as a powerful aspartate and glutamate receptor25 was converted into
a polymerizable methacrylamide CBS. For improved steric accessibility, the bisphosphonate
monomers C5ꢀSp and C6ꢀSp were generated with a C5 and C6 spacer. Finally, Ni2+ complexation
of the nitrilotriacetic acid headgroup converted NTA and NTAꢀGlu into monomers for histidine
recognition.26 In an attempt to further improve specificity to the proteinꢀselective polymers, we
also synthesized methacrylamides of amino acids which can occupy the S1ꢀpocket, but avoided
Cꢀterminal functionalization of the monomers as carboxyamides in order to prevent their
proteolytic cleavage. A benzamidine methacrylamide was used as an anchor monomer for
Trypsin as mentioned by Haupt et al. Similar to the BP unit, several other key monomers for
unpolar and charged residues (BP, BA, CyH, ED) were also prepared with a C5ꢀ or C6ꢀ spacer
between receptor unit and methacrylamide, in order to facilitate their interaction with remote
amino acid residues on the protein surface.
ACS Paragon Plus Environment
6

Page 7 of 26
Biomacromolecules
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Methacrylamideꢀbased comonomer pool for the main classes of amino acids and special
monomers. BP = Bisphosphonate; CBS = Guanidiniocarbonylpyrrole; BA = Benzamidine; NTA =
Nitrilotriacetic acid; Bor = Boronic acid; Glu = Glucosamide; Iso = Amidoisopropanol; EA =
Ethyleneamine; Sp = Spacer; Dod = Dodecyl; Ani = C6ꢀAnilide; CyH = Cyclohexyl; Dan = Dansyl.
The synthetic pathways towards the main new compounds are depicted in Scheme 2. Briefly,
guanidiniopyrrole carboxamide as a powerful carboxylate binding site (CBS) was generated
from an NꢀBoc benzyl ester precursor, hydrogenated and coupled with ethylenediamine monoꢀ
methacrylamide (EA); mild Boc removal was effected with HCl in dioxane.
A bisphosphonate nitroarene precursor was reduced to the respective aniline, and subjected to
amide coupling with C5ꢀSP or C6ꢀSP. Subsequent mild dealkylation with LiBr furnished the
extended spacer monomers C5ꢀBP and C6ꢀBP. A modified procedure was used to prepare the
NTA building block: 2 equivs. of O-t-butylꢀbromoacetate attached to the primary αꢀamine in the
ACS Paragon Plus Environment
7

Biomacromolecules
Page 8 of 26
1
2
3
4
5
mixed tBu/Bn glutamate diester, debenzylated and coupled with excess EA. Mild removal of all
Boc groups yielded NTAꢀGlu.27
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Synthetic routes to three new comonomers for charged amino acid residues: A
guanidiniocarbonylpyrrole carboxamide monomer CBS, B extended bisphosphonate monomer C5ꢀBP and
C nitrilotriacetic acid comonomer NTAꢀGlu.
Parallel copolymerization. In order to avoid polymerꢀanalogous reaction steps which often
suffer from incomplete conversion, free monomers were used without protecting groups. All
monomers were soluble in water, DMF or mixtures thereof.
ACS Paragon Plus Environment
8

Page 9 of 26
Biomacromolecules
1
2
3
4
5
6
7
8
9
Copolymerization of ternary mixtures of comonomers of different size and polarity strongly
depends on monomer reactivity ratios.28 To confirm that all key binding monomers are readily
incorporated into a copolymer, we carried out a simple copolymerization experiment in an NMR
tube. A 1:1:1 mixture of a charged monomer (BP), a polar neutral monomer (Glu) and a nonꢀ
polar monomer (C6ꢀCyH) was dissolved in D2O together with a radical initiator (AIBN),
1
degassed and heated to 70 °C for 96 hours. H NMR spectra were recorded before and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
throughout the polymerization. The decrease in the integrals of the alkene signals (relative to the
aromatic signals) was followed over time. The dianionic bisphosphonate was more slowly
incorporated than the other two monomers (Fig. 2), most likely due to electrostatic repulsion.
Obviously, the BP likes to be surrounded by less polar monomer units, which in turn prefer
formation of homopolymer spacer fragments. The resulting isolation of binding monomer units
supports their specific interaction with amino acid binding partners on the protein surface.
Figure 2. Determination of relative copolymerization kinetics from 1H NMR spectroscopic monitoring of
the monomer consumption starting with a ternary 1:1:1 mixture BP/Glu/C6ꢀCyH (300 MHz in D2O).
Further optimization revealed optimal copolymerization conditions for 24 h at 70 °C with good
reproducability. Each monomer was kept in a DMF/water stock solution at a concentration of ~
100 mg/mL; prior to the copolymerization, the monomer solutions were freshly degassed and
transferred with an Argonꢀfilled syringe into a small dry 2 mL vial, which also contained a
known amount of AIBN (usually 5 mol%). Maximum solvent loading was 2 mL per vial,
maximum monomer loading was 100 mg. All copolymerizations were carried out in clear
solutions of aqueous DMF. After sealing with glass stopcocks, 24 vials were placed on a
thermomixer and were gently shaken for 24 h at 70 °C. During this time, copolymerization was
complete. The solvent was evaporated to dryness and the residue was lyophilized to give an
almost colorless amorphous solid.
ACS Paragon Plus Environment
9

Biomacromolecules
Page 10 of 26
1
2
3
4
5
6
7
8
NMR spectra of the crude products revealed the presence of small amounts of DMF and
monomers, which were subsequently removed by ultrafiltration through centrifugal filters.29
Final lyophilization afforded colorless voluminous solids of pure material in overall yields
around 50%.
9
Almost all new polymeric inhibitors proved to be waterꢀsoluble; copolymers containing the CBS
unit with its guanidiniopyrrole cation required 2ꢀ3 equivs. of ethyleneamine (EA) monomer or 10
equivs. of glucosamine (Glu) monomer to be sufficiently soluble in water. They were examined
by size exclusion chromatography (SEC) in neutral aqueous solution at a concentration of 1
mg/mL. Poly(ethylene glycol) standards were used for calibration and produced Numberꢀaverage
molecular weights between 35 kD and 200 kD. Since the PEG skeleton bears little structural
similarity to methacrylamides, calibration with polyacrylates was also attempted, but produced
higher molecular weights with irregular large deviations, even for polymers with comparable
composition. Polydispersities (PDI) or Mw/Mn ratios were between 2 and 3 as expected from
conventional radical copolymerization.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
Finally, as a purity check, each filtered polymer was examined by H NMR spectroscopy in
order to determine the content of rest monomer; evaluation of the integrals for the sharp
methacrylamide peaks usually remained below 1 molꢀ%. Integration of broad polymer peaks was
sometimes problematic; it was, however, essential to determine copolymer composition. Usually,
only 5% less of charged comonomers were built into the copolymers, especially if the mixture
contained large amounts of the glucose monomer. 31P NMR spectra of all phosphonateꢀ
containing copolymers just displayed a broad singlet at 23.7 ppm.
Table 2 gives a survey of over 50 purified copolymers, 49 of which were water soluble.
Structurally related copolymers are shown as small series which could be used to study the
contributions of specific comonomer binding sites to the overall recognition event. The number
of glucosamide equivalents (1.0 – 5.0) necessary for water solubility directly reflects the content
of nonpolar monomer units. In some dodecylꢀcontaining polymers (P05, P13 and P41) the
1
aliphatic protons were invisible in the H NMR spectrum, and aqueous solutions appeared
blurred, pointing to selfꢀassembly into micellar aggregates. Attempted formation of zwitterionic
polymers was successful for P30, which combined BP and CBS monomers and produced clear
solutions in aqueous buffer, which showed the characteristic signals for both comonomers. In
general, copolymerization results were reproducible, as exemplified in three cases by comparison
of NMR spectra as well as SEC parameters (Mn and PDI).
Table 2. List of all affinity copolymers with their exact compositions and molecular weights.
ACS Paragon Plus Environment
10

Page 11 of 26
Biomacromolecules
1
2
3
Monomer Building Blocks (in eq.)
4
5
6
7
8
9
3
2
1.5
1.5
1.5
1.5
3
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
3
0.5 135
139
P01
P02
P03
P04
P05
P06
P07
P08
P09
P10
P11
P12
P13
P14
P15
P16
P17
P18
P19
P20
P21
P22
P23
P24
P25
P26
P27
P28
P29
P30
P31
P32
P33
P34
P35
P36
P37
P38
P39
P40
P41
P42
P43
P44
P45
P46
P47
P48
P49
P50
P51
P52
0.5
122
133
150
158
132
122
142
0.5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0.5
0.5
1
1
1
1
1
1
0.5
0.5
0.3 145
133
132
0.3 124
133
0.5
0.5
1
1
1
1
0.3
0.3
1
0.5
0.5
0.5
1
1
3
3
3
1
1
1
1
123
81
78
116
71
118
158
196
0.5
1
4
0.5
0.5
0.5
1
1
1
3
3
3
1
3
4
4
5
3
3
1
1
1
0.5
0.5
1
1
1
1
1
1
1
202
107
142
180
147
167
148
123
87
1
1
1
1
1
1
1
1
1
1
1
1
4
1
1
4
0.5
134
ꢀ
1
1
4
120
222
103
155
37
1
1
1
1
1
1
1
1
1
0.3
3
3
5
1
1
0.5 100
94
0.4 103
112
1
3
2
1
1
1
1
1
1
1
2
3
3
1.5
1.5
1
1
1
1
1
1
1
1
83
92
104
1
119
104
118
95
1.5
3
1.5
3
113
0.3 67
73
1
2
5
1
1
ACS Paragon Plus Environment
11

Biomacromolecules
Page 12 of 26
1
2
3
4
5
6
7
8
9
It should be noted, that no evidence of aggregate formation was obtained for any of the
copolymers, such as, e. g., excessive broadening of NMR signals or turbidity of solutions. Very
recent DLS as well as DOSY measurements indicated hydrodynamic radii corresponding to
nonaggregated species. Moreover, to avoid aggregation, the composition of all copolymers had
at least 65% of highly waterꢀsoluble functional groups.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In Table 2 the different series of copolymers with a common major binding monomer are
presented in groups. From top to bottom: P01ꢀP13 all contain the bisphosphonate monomer BP
and are therefore negatively charged. Within this large group, three subgroups represent
combinations with nonpolar monomers Ani, C6ꢀCyH and Dod: P03ꢀ05, P07ꢀ10 and P11ꢀ13. The
next major series comprises P14ꢀ17 which share the cationic ethyleneamine unit EA. Strongly
overlapping is the series P15ꢀ27 which all carry the boronic acid moiety Bor. P19ꢀ21 contain the
negatively charged NTA unit. With P22ꢀ26, a series of cationic benzamidineꢀcontaining
polymers (BA) is introduced, followed by P26ꢀ31, presenting the positively charged CBS unit.
P39ꢀP36, P38 and P42 are all homopolymers or polymers with only one amino acid recognition
unit, namely with NTA, BP, BA or Glu monomers. Finally, P42ꢀP50 are all NTAꢀbased
copolymers, the first of them (P43ꢀP46) rich in BP, the next subgroup (P47ꢀP50) rich in BP with
spacers.
Parallel enzyme assays. These polymer materials were subsequently used directly to investigate
their interaction with proteolytic enzymes. Since determination of protein affinities by titration
methods is very timeꢀconsuming, and does not necessarily coincide with inhibitory power, a
parallel screening method was developed which measured the attenuation of enzymatic activity.
To this end, a 10ꢀfold molar excess of each polymer was incubated with the respective enzyme
and treated with excess substrate. From the relative decrease in enzyme velocity the best
inhibitors were selected for a subsequent determination of their IC50 values (Figure 3). All
titrations were done at physiological pH and ionic strength.
Standard enzyme assays were established which used typical peptidic enzyme substrates whose
proteolytic cleavage produced or consumed a colored product with a strong absorption in the
visible range of the electromagnetic spectrum.30 For each enzyme assay UVꢀvis spectra
confirmed the absorption maximum of the released chromophore and ascertained that enzyme as
well as substrate absorption at the selected wavelength were negligible. Almost all enzyme
assays were crossꢀchecked by determination of the IC50 value of a known reference inhibitor,
whose literature value was reproduced. Buffers (Trizma, Hepes) were adjusted to optimal pH for
each enzyme (5.0 and 7.4 ꢀ 8.0), and contained physiological salt concentrations.
Elastase31 (PDB code 1H1B, 2Z7F) is a serine protease, produced in zymogenic form in the
exocrinic part of the pancreas, and activated by trypsin cleavage in the duodenum. It shares the
same catalytic triad with the structurally related endopeptidases trypsin and αꢀchymotrypsin, but
ACS Paragon Plus Environment
12

Page 13 of 26
Biomacromolecules
1
2
3
4
5
6
7
8
9
due to its small binding pocket cuts its substrates, mainly casein and elastin, only after small
uncharged residues like alanine. Its high pI value of 8.7 and pH optimum at 8.0 suggests strong
binding to anionic polymers; it was therefore very frustrating that even at pH 8.0 none of the 50
polymers was an effective inhibitor. Inspection of elastase’s crystal structure32 gives two hints
for a plausible explanation: no basic or acidic residue is found in the direct vicinity of the active
site; and many arginines and lysines are located close to aspartates and glutamates, facilitating
internal ion pair formation which hinders complexation by an ionic external binder (Scheme
3A).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
This is different with αꢀchymotrypsin (PDB code 1AFQ),33 which is also proteolytically
activated by trypsin in the duodenum. The enzyme also possesses a high pI value of 8.7 and a pH
optimum of 7.8, but it cuts selectively after aromatic amino acids due to its spacious S1 pocket.
In our polymer screen only very few polymers were efficient inhibitors. Contrary to our
expectation these were cationic and carried benzamidine or guanidiniocarbonylpyrrole binding
sites, which were combined with extended aromatic or aliphatic nonpolar arms. Intriguingly,
there is a patch of two phenylalanines with two adjacent acidic residues directly next to αꢀ
chymotrypsin’s active site, whereas most basic resides are located on the back side of the protein
(Scheme 3B).34
ACS Paragon Plus Environment
13

Biomacromolecules
Page 14 of 26
1
2
3
4
5
6
Figure 3. Polymer screening assay for determining IC50 values of the active compounds, illustrated for
trypsin inhibition. A Only P18 is active; B Doseꢀdependent trypsin inhibition by P18; C Determination of
the IC50 value for P18 (0.5 µM); D LineweaverꢀBurk plot indicating competitive inhibition.
7
8
9
From a medicinal point of view, trypsin35 (PDB code 4I8G–4I8L, 1S0R), is the most interesting
serine protease, since it activates numerous other proteolytic enzymes and its overexpression and
premature activation may ultimately be responsible for multiple organ damage. Trypsin is
formed in the exocrinic pancreas part, and activated by autocatalysis in the duodenum. The
strongly basic enzyme (pI 10.0) is especially rich in lysines and operates at a pH optimum
between 7.5 and 8.5. Contrary to αꢀchymotrypsin, trypsin cuts after arginine and lysine. Various
polymers turned out to be good to excellent trypsin inhibitors. They all contained the
bisphosphonate arginine binder. However, the simultaneous presence of a small amount of
boronic acid binders again lowered the IC50 value 6ꢀfold. Indeed, trypsin contains 34 serines,
which represent the most abundant residues (Scheme 3C). Serines in close proximity will readily
form covalent cyclic boronate esters in a reversible manner.36
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A similar effect was achieved by adding small amounts of NTA head groups. Surprisingly, the
NTA groups increased affinity in the absence of Ni2+ ions, whereas especially the NTA
homopolymer P42 (but also the related P43) completely lost all their inhibitory power after
equimolar Ni2+ addition. These results strongly indicate that histidine complexation is not the
driving force behind the NTA/protein interaction. It seems as transition metal complexation even
masks the NTA trianion – leaving a direct electrostatic attraction of lysine cations as the most
probable interaction. The same phenomenon was already observed with efficient IgG
complexation by NTAꢀrich polymers, and led to the discovery of a direct interaction between
NTA head groups and lysine pairs on the surface of basic proteins.12 It could be corroborated by
ITC measurements (Fig. 4):37 in a titration experiment trypsin was bound by the NTA
homopolymer at a 2:1 stoichiometry (protein/polymer) in an enthalpically driven process (ꢁH = ꢀ
7.3 kcal/mol) with a small entropic cost (TꢁS = ꢀ1.7 kcal/mol). The resulting polymer protein
affinity is relatively low (ꢁG = ꢀ5.5 kcal/mol ~ 100 µM Kd). In the presence of 50 molꢀ% of Ni2+
ions, no binding isotherm could be obtained. Control experiments proved that Ni2+ ions as well
as NTA monomers did not display any protein affinity, ruling out any unspecific metal/protein
interactions and confirming a direct electrostatic NTA protein attraction. The fact that lysines are
only complexed in the absence of Ni2+ ions, however strongly suggests, that the metal complex is
thermodynamically more stable. If this is true, Ni2+ ions should be able to compete with a
preformed enzyme NTA homopolymer complex. This was indeed observed when the saturated
polymer enzyme complex was titrated against an equimolar solution of Ni2+ ions – in a
competitive exothermic process the Ni2+ ions displaced the NTA homo polymer from the protein
and formed the more stable Ni2+/NTA complex.
If the affinity polymers bind to critical amino acids around the active site, they may inhibit
enzyme activity by blocking the substrate access. In an exemplary case study, enzyme kinetics
ACS Paragon Plus Environment
14

Page 15 of 26
Biomacromolecules
1
2
3
4
5
6
7
8
9
for trypsin inhibition by P52 were investigated and evaluated by LineweaverꢀBurk plots (Figure
3D). Here the kinetic graphs for the enzyme in the absence and presence of increasing amounts
of inhibitor polymer indeed had a common intersection with the ordinate, clearly indicating a
competitive mechanism with reference to the substrate. In this case, the polymer competes with
the substrate to be bound on the protein surface. Most likely, this is achieved, when the polymer
sterically blocks the entrance to the active site.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. A Exothermic ITC titration of trypsin with increasing amounts of NTA homopolymer P42. B
Lysine complexation by the NTA head in the absence of Ni2+ ions. C Back titration of the
protein/polymer complex with Ni2+ ions is also enthalpyꢀdriven because the stable NTA/Ni2+ complex is
formed and the protein is released.
A good example for an acidic protease is carboxypeptidase A (PDB code 1F57, 5CPA).38 It is a
metalloprotease with a catalytically active Zn2+ ion and, like the serine proteases, it is activated
by trypsin cleavage in the duodenum. Its pI is 6.0 with a pH optimum at 7.5. A large
hydrophobic pocket accommodates both aromatic or aliphatic amino acid side chains. The Zn2+
ion together with an arginine activates the amide bond by Brønsted as well as Lewis acid
catalysis and together with a glutamate helps to deprotonate a water nucleophile. Again, very few
polymers were able to inhibit this metalloprotease. They all contain prominent NTA groups, but
do not all require Ni2+ ions for shutting down the enzyme. The crystal structure shows several
histidines, two pairs in or close to the active site (Scheme 3D).39 Thus, direct interaction with the
ACS Paragon Plus Environment
15

Biomacromolecules
Page 16 of 26
1
2
3
4
5
6
7
8
9
catalytically active Zn2+ꢀ histidine diad by NTA complexation at Zn2+ or even transmetallation
by the polymer NTAꢀNi2+ moieties may be the key interaction to interrupt enzyme activity. The
presence of several bisphosphonate dianions will help to attach the polymer around the active
site which is surrounded by multiple lysines, whereas most acidic residues are located on the
back side of carboxypeptidase A. This may explain why no cationic polymers inhibit the acidic
protease.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Kallikrein40 (PDB code 5TJX, 1SPJ) is a very interesting case, because this serine protease is
important in the early stage of the blood coagulation cascade. It is formed in the hepatozytes of
the liver and is again activated by trypsin, but also by Factor XII (Hagemann factor). Kallikrein
also operates in tissues and releases kinins, tissue hormones with general relevance in
inflammatory processes. It is a strongly acidic enzyme with a pI of 4.0 ꢀ 4.5 and a pH optimum at
7ꢀ8. Contrary to expectations, again no cationic polymer was able to interfere with peptide
cleavage by kallikrein. A relatively large number of BP and NTAꢀcontaining polymers however,
turned out to be potent inhibitors, with a substantial overlap in the group of previously identified
powerful trypsin inhibitors. In the crystal structure41 two histidines are located on opposite ends
at the entrance of the active site, whose complexation with a single NTAꢀrich polymer strand
would entirely block the substrate approach (Scheme 4A). The active site is also surrounded by
four serines, explaining the potency of two boronic acidꢀbased polymers. Interesting enough,
most glutamates and aspartates are located on the back side of kallikrein, leaving a lysineꢀrich
patch underneath the active site. Most likely, this explains the inhibitory power of BP and NTAꢀ
rich polymers, even in the absence of Ni2+ ions.
ACS Paragon Plus Environment
16

Page 17 of 26
Biomacromolecules
1
2
3
4
5
6
7
8
9
Scheme 3. Crystal structure of A pancreas elastase, B αꢀchymotrypsin, C trypsin and D carboxypeptidase
A. Red spots represent carboxylate groups of acidic amino acids (Asp, Glu). Light blue spots indicate
ammonium groups from lysines and dark blue spots arginines’ guanidinium cations. The violet molecule
is an inhibitor occupying the active site. The catalytic triad is depicted in green. Black areas are surfaceꢀ
exposed phenylalanines. Orange areas are surfaceꢀexposed serines. Green areas show histidines –
turquoise are histidine pairs.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The most interesting feature of kallikrein inhibition by our affinity polymers, however, is their
drastically subꢀstoichiometric nature: 0.5 ꢂM enzyme solutions were effectively inhibited at 3ꢀ12
nM polymer concentrations, corresponding to enzyme:polymer ratios of up to 190:1. For steric
reasons, even polymers with a molecular weight of 100 kD are not able to complex such a large
number of enzyme molecules at the same time. Digestion of the polymers to small oligomers or
monomers is very unlikely because of their allꢀcarbon backbone and can further be ruled out,
since those small entities had no inhibitory effect on kallikrein. The underlying cause for this
interesting subꢀstoichiometric enzyme inhibition will be investigated in the near future.42
In order to avoid unwanted interference with blood coagulation, an orthogonal inhibition mode
towards Thrombin43ꢀ44 (PDB code 1A2C,1A3B) is especially important. This hydrolase is
formed in the liver and activated during the blood coagulation cascade by factors Xa and Va. It
exerts a fourfold arginineꢀglycine cleavage and thus activates fibrinogen to fibrin, which
ultimately forms the thrombus. Thrombin is inhibited by antithrombin III, a process of
fundamental importance for fibrinolysis (thrombus dissolution). The enzyme has a pI of 7.2 and
a pH optimum of 7.5ꢀ8.5.
Only a few BPꢀrich copolymers were able to inhibit thrombin; and these required usually high
polymer/protein excesses of up to 50 equivalents. Moreover, inhibition was limited to around
70%. There are some arginine and lysine clusters on thrombin’s protein surface which will
represent anchor points for the bisphosphonate units (Scheme 4B).45 Contrary to many other
hydrolases, NTAꢀrich polymers have no effect on thrombin activity. However, the presence of
nonpolar binding sites such as dansyl or dodecyl within a polymer rich in bisphosphonates
significantly lowers its IC50 value. A large hydrophobic patch stretches along one flank of the
active site – well suited for hydrophobic and dispersive interactions.
Our final example, Cathepsin D,46ꢀ47 represents a lysosomal aspartate endopeptidase (PDB code
1LYA, 4OBZ). Preprocathepsin D is formed in the rough endoplasmatic reticulum (ER) and
converted by proteolytic cleavage and glycosylation to procathepsin D, which is then
autocatalytically activated. A catalytic diad is formed by two aspartate residues, only one of
which is protonated. Together, they deprotonate a nucleophilic water molecule, cut between two
phenylalanines and protonate the released amide anion. Cathepsin is a markedly acidic protease,
and operates at a pH range of 3.5ꢀ5 (Scheme 4C). In the assay, a selfꢀquenched coumarin labeled
decapeptide releases the fluorophore in a turn-on process and provides a highly sensitive
detection mode.
ACS Paragon Plus Environment
17

Biomacromolecules
Page 18 of 26
1
2
3
4
5
6
7
8
9
It has to be taken into consideration that at a pH below 4 most carboxylates on cathepsine’s
surface48 will be protonated, and only weakly bound by cationic polymers. However, diꢀ and
trianionic BP and NTA head groups are able to form a network of ionic hydrogen bonds towards
these ꢀCO2H groups. It is hence not surprising, that most BPꢀ and NTAꢀcontaining polymers
were inhibitors of Cathepsin D with IC50 values in the low nanomolar range. Importantly, the
single BP and NTA monomers had no inhibitory effect in the screening, ruling out that
hydrolytic polymer cleavage might have produced small monomeric inhibitors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Affinities and Selectivities. With the full account of all polymer/protein interactions at hand, a
synoptic comparison of all polymers with respect to their protein selectivities becomes possible.
Our results are summarized in Table 3. Each blank represents a polymer whose inhibitory power
is at least 100 times weaker than that of all other inhibitors for a given enzyme. It should be
emphasized that the overall enzyme concentrations in all assays varied between 2 nM and 2 µM.
Thus the efficiency/effectivity of a polymer should not be measured by absolute IC50 values, but
rather by comparison of the respective stoichiometric polymer/protein ratio. These values are
always presented in the accompanying columns.
Scheme 4. Crystal structure of A kallikrein, B thrombin and C cathepsin D. Red spots represent
carboxylate groups of acidic amino acids (Asp, Glu). Light blue spots indicate ammonium groups from
lysines and dark blue spots arginines’ guanidinium cations. The violet molecule is an inhibitor occupying
the active site. The catalytic triad is green. Green areas show histidines – orange areas indicate serines.
ACS Paragon Plus Environment
18

Page 19 of 26
Biomacromolecules
1
2
3
4
5
6
7
8
9
In general, efficient inhibitors were found for all enzymes except for elastase. Quite often, the pI
value of the protein is not a good guide for the choice of appropriately charged comonomers,
because a majority of the complementary amino acid residues are located on the back of the
protein, far away from the active site. On the contrary, for strong inhibition the direct
environment around the active site seems to be critical. If a copolymer contains a significant
amount of monomer units which bind directly next to the active site, it will efficiently prevent
substrate access and product exit. This seems to be a very important factor that renders a given
polymer a good enzyme inhibitor. In addition, a general affinity increase can be brought about by
clusters of equally charged amino acids whose simultaneous interaction with an oppositely
charged affinity polymer will lead to unspecific, but powerful attraction, which reaches far into
the aqueous environment around the protein.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. Summary of all active polymers with their corresponding IC50 values and enzyme/polymer
stoichiometries. Bold fields indicate the most proteinꢀselective copolymer.
3.8µM
2.8µM
1.6µM
3.8µM
3.6µM
3.9nM
7.1nM
7.6nM
9.7nM
136:1
75:1
70:1
55:1
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
P01
+Ni
P02
+Ni
P08
P18
P24
P32
P34
P36
P39
+Ni
P40
P41
P42
+Ni
P43
+Ni
P44
P45
P46
P47
P48
P49
P50
P03
P04
P05
P06
0.6µM
1:0.3
7.6nM
70:1
1:0.9
1:12
1:53
1:3.3
1:7.8
0.44µM
110nM
470nM
30nM
1.3µM
2.3µM
3.2µM
1.0µM
1.6µM
1.4µM
1:0.7
1:4.6
1:1.6
1:0.5
1:0.8
1:0.7
3.7nM
2.8nM
4.4nM
4.6nM
9.2nM
6.2nM
7.7nM
3.3nM
6.1nM
8.9nM
11.5nM
6.5nM
6.0nM
143:1
189:1
120:1
115:1
58:1
31nM
1:12
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
<0.1µM
70nM
85:1
69:1
17nM
1:1.9
0.5µM
0.9µM
1:1.0
1:1.8
160:1
87:1
1.3µM
60:1
46:1
0.9µM
1.4µM
1.8µM
1.7µM
1.5µM
1.5µM
1.5µM
1:1.8
1:2.9
1:3.6
1:3.4
1:3.0
1:3.0
1:3.0
82:1
88:1
ACS Paragon Plus Environment
19

Biomacromolecules
Page 20 of 26
1
2
3
0.5µM
2.0 / 0.5µM
n.d.
0.53µM
9nM
2.5nM
c(E)
4
5
6
7
8
9
Most enzyme/polymer combinations did not lead to any change in enzyme activity – a
prerequisite for selectivity. In light of all the above results, it can be inferred that only few
favorable comonomer combinations and ratios lead to multipoint binding of our linear
copolymers around the active site. Except P24 all polymers interact with cathepsin D. Some
polymers shut down four of all seven highly diverse proteases, most of them with a dominating
BP content (P32, P34 and P36).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
There is a strong overlap between copolymers which are active against the related serine
proteases trypsin and kallikrein, and all carboxypeptidase inhibitors likewise stop kallikrein.
However, some affinity polymers are unique: thus P03ꢀP06 selectively block cathepsin D. P24
only attenuates chymotrypsin and thrombin, and P08 selectively shuts down carboxypeptidase A
(and cathepsin D). Similarly, P46ꢀP50 block only trypsin (and cathepsin D), whereas P42 is
active against carboxypeptidase A and kallikrein (as well as cathepsin) in the presence of Ni2+
ions. Finally, P41 targets kallikrein, thrombin (and cathepsin D).
In this protein series, monomers with mediumꢀsized spacers (C5/C6) between their binding site
and the methacrylamide moieties did not perform differently to their short counterparts without
the spacer. We conclude that enough flexibility is maintained in the copolymers due to the
presence of glucose “dummy” monomers. Even if the synthetic binding site (e.g., BP) is attached
close to the polymer backbone, it is able to reach its complementary amino acid functionality on
the protein surface (e.g., Lys/Arg). It remains to be examined whether or not the intense induced
fit process of a linear copolymer on the complex protein surface, which ultimately leads to
optimized multivalent and heteroavidic recognition, requires time and lowers entropy
significantly.
Figure 5. Visualization of a representative polymer fragment taken directly from polymer P18 (ball and
sticks) on bovine trypsin (Connolly surface). The picture shows several bisphosphonate units engaged in
ACS Paragon Plus Environment
20

Page 21 of 26
Biomacromolecules
1
2
3
4
5
chelate complexes with lysines and boronic acid moieties which formed covalent ester links to serines.
(MacroModel 9.9, OPLSꢀ2005, water/GBA solvation, MM calculation 10000 steps).
6
7
8
9
Several polymers with nonpolar comonomers were among the most efficient protease inhibitors.
We tentatively explain this with the fast homopolymerization kinetics of the nonpolar reference
comonomer, pointing to a potential formation of nonpolar polymer blocks which target extended
nonpolar patches on a protein surface and thus contribute to high affinity binding.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Taken together, the synopsis in Table 3 reveals strengths and weaknesses of the underlying
concept: various enzymes with similar active sites can be effectively distinguished by polymeric
surface binders; and relative small variations in monomer composition greatly alter the inhibition
potential and selectivity of these polymers. However, only a limited number of monomer binding
sites contributes to strong binding; in particular, less polar monomers must be more efficiently
isolated within the copolymer strand so that they can help to discriminate the immediate
environment around a given active site.
For a visualization of the polymer binding mode on the protein surface we calculated the
structure between a polymer fragment taken directly from P18 and bovine trypsin (Fig. 5). After
placing the polymer fragment closely above the trypsin surface next to the active site, a forceꢀ
field calculation reached a minimum which shows several polymer binding sites engaged in
specific interactions with the complementary amino acid residues; other polymer binding sites
point away from the protein into the aqueous solvent. It should be taken into consideration, that
this is only one possible snapshot out of many conformations due to the highly dynamic nature of
the protein/polymer interaction.
Conclusion and Outlook. This account presents a general method for the construction of
polymeric surface binders for digestion enzymes. Two prominent parts, namely the modification
of the copolymer composition and the screening assay for the most powerful inhibitors are both
amenable to parallelization. The concept hinges on the appropriate selection of aminoꢀacidꢀ
selective comonomers, their free radical copolymerization, and subsequent screening of the
resulting copolymer library for efficient enzyme inhibition. A microscale synthetic procedure for
the copolymerization process was developed, which produces waterꢀsoluble affinity polymers
that can be stored for years at room temperature. Initial parallel screening was conducted in
standard enzyme assays to identify polymeric inhibitors, which were subsequently subjected to
determination of IC50 values for their target enzyme. For all digestion enzymes except elastase a
number of polymer inhibitors were found, some of which were selective towards one or two
protein targets. Since the key monomers of the best inhibitors bind to amino acid residues in the
direct vicinity of the active site, we conclude that efficient coverage of its immediate
environment by the copolymers is critical. Strong interference with enzymatic activity is brought
about by blocking the substrate access and product exit to and from the active site.
ACS Paragon Plus Environment
21

Biomacromolecules
Page 22 of 26
1
2
3
4
5
6
7
8
9
In the future we will further study the postulated inhibition mechanism in more detail, in order to
find out how the extended linear affinity polymer explores the protein surface and leads to
efficient enzyme inhibition. A special emphasis will be placed on the unexpected
substoichiometric action of polymer inhibitors towards kallikrein. The most efficient and
selective affinity polymers of this study will be subjected to toxicity and activity tests in isolated
segments of rat intestines, and finally in animal models by collaboration partners at the
university clinics. Since no membranes have to be passed to reach the intestinal walls, we
propose oral administration of potential drugs in the final phase. The concept presented here may
also be applied to the blood coagulation cascade or other essential multiꢀprotein
processes/pathways.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ASSOCIATED CONTENT.
Electronic supplementary information (ESI) available free of charge: Synthetic Procedures
to monomers and polymers, NMR spectra, SEC spectra, calibration and reference curves for the
enzyme assays, kinetics screening, doseꢀresponse plots for IC50 determination, Copolymerization
kinetics (in NMR spectra), ITC titrations.
AUTHOR INFORMATION.
Corresponding Author
* Eꢀmail: thomas.schrader@uniꢀdue.de
Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval
to the final version of the manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENT.
Financial support from the DFGꢀfunded Collaborative Research Centre 1093 is gratefully
acknowledged.
REFERENCES.
1. Schweizer, E.; HoffmannꢀRöder, A.; Olsen, J. A.; Seiler, P.; ObstꢀSander, U.; Wagner, B.;
Kansy, M.; Banner, D. W.; Diederich, F., Org. Biomol. Chem. 2006, 4, 2364ꢀ2375.
2. Schmitt, L.; Tampé, R., Prot. Eng. Oxford Univ. Press UK: 1995.
3. Dernedde, J.; Rausch, A.; Weinhart, M.; Enders, S.; Tauber, R.; Licha, K.; Schirner, M.;
Zügel, U.; von Bonin, A.; Haag, R., PNAS 2011, 107, 19679ꢀ19684.
ACS Paragon Plus Environment
22

Page 23 of 26
Biomacromolecules
1
2
3
4
5
4. Filby, M. H.; Muldoon, J.; Dabb, S.; Fletcher, N. C.; Ashcroft, A. E.; Wilson, A. J., Chem.
Commun. 2011, 47, 559ꢀ561.
6
7
8
9
5. Yamaguchi, Y.; Kato, N.; Azuma, H.; Nagasaki, T.; Ohkanda, J., Bioorg. Med. Chem. Lett.
2012, 22, 2354ꢀ2358.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6. You, C.ꢀC.; Miranda, O. R.; Gider, B.; Ghosh, P. S.; Kim, I.ꢀB.; Erdogan, B.; Krovi, S. A.;
Bunz, U. H. F.; Rotello, V. M., Nat. Nanotechn. 2007, 2, 318ꢀ323.
7. Sandanaraj, B. S.; Demont, R.; Aathimanikandan, S. V.; Savariar, E. N.; Thayumanavan, S.,
J. Am. Chem. Soc. 2006, 128, 10686ꢀ10687.
8. Lee, S.ꢀH.; Hoshino, Y.; Randall, A.; Zeng, Z.; Baldi, P.; Doong, R.; Shea, K. J., J. Am.
Chem. Soc. 2012, 134, 15765ꢀ15772.
9. Cutivet, A.; Schembri, C.; Kovensky, J.; Haupt, K., J. Am. Chem. Soc. 2009, 131, 14699ꢀ
14702.
10. Zhang, H.; Jiang, J.; Zhang, H.; Zhang, Y.; Sun, P., Macro Lett. 2013, 2, 566ꢀ570.
11. Nishino, H.; Huang, C.ꢀS.; Shea, K. J., Angew. Chem. Int. Ed. 2006, 45, 2392ꢀ2396.
12. Nematollahzedeh, A.; Sun, W.; Aureliano, C. S. A.; Lütkemeyer, D.; Stute, J.; Abdekhodaie,
M. J.; Shojaei, A.; Sellergren, B., Angew. Chem. Int. Ed. 2001, 50, 495ꢀ498.
13. Canfarotta, F.; Poma, A.; Guerreiro, A.; Piletsky, S., Nat. Prot. 2016, 11, 443ꢀ455.
14. Turner, N. W.; Jeans, C. W.; Brain, K. R.; Allender, C. J.; Hlady, V.; Britt, D. W.,
Biotechnol. Prog. 2006, 22, 1474ꢀ1489.
15. Verheyen, E.; Schillemans, J. P.; van Wijk, M.; Demeniex, M.ꢀA.; Hennink, W. E.; van
Nostrum, C. F., Biomaterials 2011, 32, 3008ꢀ3020.
16. Renner, C.; Piehler, J.; Schrader, T., J. Am. Chem. Soc. 2006, 128, 620ꢀ628.
17. Koch, S.; Renner, C.; Xie, X.; Schrader, T., Angew. Chem. Int. Ed. 2006, 45, 6352ꢀ6355.
18. Wenck, K.; Koch, S.; Renner, C.; Sun, W.; Schrader, T., J. Am. Chem. Soc. 2007, 129,
16015ꢀ16019.
19. Latza, P.; Gilles, P.; Schaller, T.; Schrader, T., Chem. Eur. J. 2014, 20, 11479ꢀ11487.
20. Oldenburg, W. A.; Lau, L. L.; Rodenberg, T. J.; Edmonds, H. J.; Burger, C. D., Arch. Intern.
Med. 2004, 164, 1054ꢀ1062.
ACS Paragon Plus Environment
23

Biomacromolecules
Page 24 of 26
1
2
3
4
21. Whitcomb, D. C., New Engl. J. Med. 2006, 354, 2142ꢀ2150.
5
6
22. Frossard, J. L.; Steer, M. L.; Pastor, C. M., Lancet 2008, 371, 143ꢀ152.
7
8
9
23. Uhl, W.; Buchler, M. W.; Malfertheiner, P.; Beger, H. G.; Adler, G.; Gaus, W., Gut 1999,
45, 97ꢀ104.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
24. Buchler, M. W.; Malfertheiner, P.; Uhl, W.; Scholmerich, J.; Stockmann, F.; Adler, G.;
Gaus, W.; Rolle, K.; Beger, H. G., Gastroenterology 1993, 104, 1165ꢀ1170.
25. Schmuck, C., Chem. Eur. J. 2000, 6, 709ꢀ718.
26. Hart, B. R.; Shea, K. J., Macromolecules 2002, 35, 6192ꢀ6201.
27. Lata, S.; Reichel, A.; Brock, R.; Tampé, R.; Piehler, J., J. Am. Chem. Soc. 2005, 127, 10205ꢀ
10215.
28. Aguilar, M. R.; Gallardo, A.; del Mar Fernández, M.; San Román, J., Macromolecules 2002,
35, 2036ꢀ2041.
29. 2 mL Amicon Ultra 3K Membrane Filters, M. M.
30. Bergmeyer, H. U., 2nd ed.; Academic Press: New York, 1974; Vol. I.
31. Macdonald, S. J. F.; Dowle, M. D.; Harrison, L. A.; Clarke, G. D. E.; Inglis, G. G. A.;
Johnson, M. R.; Shah, P.; Smith, R. A.; Amour, A.; Fleetwood, G.; Humphreys, D. C.;
Molloy, C. R.; Dixon, M.; Godward, R. E.; Wonacott, A. J.; Singh, O. M. P.; Hodgson, S.
T.; Hardy, G. W., J. Med. Chem. 2002, 45, 3878ꢀ3890.
32. Tamada, T.; Kinoshita, T.; Kurihara, K.; Adachi, M.; Ohhara, T.; Imai, K.; Kuroki, R.; T.,
T., J. Am. Chem. Soc. 2009, 131, 11033ꢀ11040.
33. Tsukada, H.; Blow, D. M., J. Mol. Biol. 1985, 184, 703ꢀ711.
34. Kashima, A.; Inoue, Y.; Sugio, S.; Maeda, I.; Nose, T.; Shimohigashi, Y., Eur. J. Biochem.
1998, 255, 12ꢀ23.
35. Liebschner, D.; Dauter, M.; Brzuszkiewicz, A.; Dauter, Z., Acta Cryst. 2013, 69, 1447ꢀ1462.
36. Transue, T. R.; Gabel, S. A.; London, R. E., Bioconj. Chem. 2006, 17, 300ꢀ308.
37. O'Brien, R.; Ladbury, J. E.; Chowdry, B. Z., Isothermal titration calorimetry of
biomolecules. In Protein-Ligand interactions: hydrodynamics and calorimetry, Harding, S.
E.; Chowdry, B. Z., Eds. Oxford Univ. Press: UK, 2000.
ACS Paragon Plus Environment
24

Page 25 of 26
Biomacromolecules
1
2
3
4
5
6
38. van Aalten, D. M. F.; Chong, C. R.; JoshuaꢀTor, L., Biochemistry 2000, 39, 10082ꢀ10089.
39. Jensen, A. F.; Bukrinsy, J. T.; Bjerrum, M. J.; Larsen, S., Inorg. Chem. 2002, 7, 490ꢀ499.
7
8
9
40. Li, Z.; Partridge, J. R.; Silva, A.; Rademacher, P.; Betz, A.; Xu, Q.; Sham, H.; Hu, Y.; Shan,
Y.; Liu, B.; Zhang, Y.; Shi, H.; Ma, X.; Zhang, L., ACS Med. Chem. Lett. 2017, 8, 185–190.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
41. Laxmikanthan, G.; Blaber, S. I.; Bernett, M. J.; Scarisbrick, I. A.; Juliano, M. A.; Blaber,
M., Proteins 2005, 58, 802ꢀ814.
42. Evaluation of Enzyme Inhibitors in Drug Discovery: A Guide for Medicinal Chemists and
Pharmacologists, Copeland, R. A., John Wiley & Sons: Hoboken, New Jersey, 2005.
43. RiosꢀSteiner, J. L. R.; Murakami, M.; Tulinsky, A., J. Am. Chem. Soc. 1998, 120, 597ꢀ598.
44. Skordalakes, E.; Elgendy, S.; Goodwin, C. A.; Green, D.; Scully, M. F.; Kakkar, V. V.;
Freyssinet, J. M.; Dodson, G.; Deadman, J. J., Biochemistry 1998, 37, 14420ꢀ14427.
45. Biela, A.; Khayat, M.; Tan, H.; Kong, J.; Heine, A.; Hangauer, D.; Klebe, G., J. Mol. Biol.
2012, 418, 350ꢀ366.
46. Metcalf, P.; Fusek, M., EMBO J. 1993, 12, 1293ꢀ1302.
47. Gradler, U.; Czodrowski, P.; Tsaklakidis, C.; Klein, M.; Werkmann, D.; Lindemann, S.;
Maskos, K.; Leuthner, B., Bioorg. Med. Chem. Lett. 2014, 24, 4141ꢀ4150.
48. Baldwin, E. T.; Bhat, T. N.; Gulnik, S.; Hosur, M. V.; Sowder II, R. C.; Cachau, R. E.;
Collins, J.; Silva, A. M.; Erickson, J. W., PNAS 1993, 90, 6796ꢀ6800.
For Tables of Contents Use only
HighꢀAffinity Copolymers Inhibit Digestive
Enzymes by Surface Recognition
Patrick Gilles,‡ Kirstin Wenck,‡ Inga Stratmann,‡ Michael Kirsch,† Daniel A. Smolin,‡ Torsten
Schaller, ‡ Herbert de Groot, † Arno Kraft,§ Thomas Schrader*‡
ACS Paragon Plus Environment
25

Biomacromolecules
Page 26 of 26
1
2
3
4
5
6
TOC graphics:
7
8
9
New affinity polymers recognize the
surface of digestion enzymes, block
the access to their active sites and
thus stop enzyme activity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
26