Perfluorinated hydroxamic acids are potent and selective inhibitors of HDAC-like enzymes from Pseudomonas aeruginosa

Article abstract of DOI:10.1016/j.bmcl.2017.02.050

A series of perfluorinated SAHA (PFSAHA) was prepared and profiled against a panel of human and bacterial members of the Histone deacetylase (HDAC) family. Some of the active substances show nanomolar inhibitory activity and several hundred fold selectivity for the HDAC like enzyme PA3774 from P. aeruginosa. The extraordinary selectivity against human HDACs results from the distinct oligomeric state of PA3774 which consists of two head-to-head dimers. The binding pocket is defined by the surface of both opposite monomers confining the access of ligands to the active site. In addition, the aromatic cap group of PFSAHA undergoes an edge-to-face aromatic interaction with phenylalanine from the opposite monomer.

Full text of DOI:10.1016/j.bmcl.2017.02.050

Accepted Manuscript
Perfluorinated hydroxamic acids are potent and selective inhibitors of HDAC
like enzymes from Pseudomonas aeruginosa
Christian Meyners, Benjamin Wolff, Alexander Kleinschek, Andreas Krämer,
Franz-Josef Meyer-Almes
PII:
S0960-894X(17)30182-8
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.02.050
BMCL 24722
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
23 January 2017
17 February 2017
18 February 2017
Please cite this article as: Meyners, C., Wolff, B., Kleinschek, A., Krämer, A., Meyer-Almes, F-J., Perfluorinated
hydroxamic acids are potent and selective inhibitors of HDAC like enzymes from Pseudomonas aeruginosa,
Bioorganic & Medicinal Chemistry Letters (2017), doi: http://dx.doi.org/10.1016/j.bmcl.2017.02.050
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Perfluorinated hydroxamic acids are potent and
selective inhibitors of HDAC like enzymes from
Pseudomonas aeruginosa
Christian Meyners1, Benjamin Wolff1, Alexander Kleinschek1, Andreas Krämer1 , Franz-Josef
Meyer-Almes1,*
1
Department of Chemical Engineering and Biotechnology, University of Applied Sciences,
Haardtring 100, 64295 Darmstadt, Germany.
* Corresponding author:
Franz-Josef Meyer-Almes
Email : franz-josef.meyer-almes@h-da.de
Phone: +49 6151168406
Fax : +49 6151168404
1

ABSTRACT
A series of perfluorinated SAHA (PFSAHA) was prepared and profiled against a panel of human
and bacterial members of the Histone deacetylase (HDAC) family. Some of the active substances
show nanomolar inhibitory activity and several hundred fold selectivity for the HDAC like
enzyme PA3774 from P. aeruginosa. The extraordinary selectivity against human HDACs
results from the distinct oligomeric state of PA3774 which consists of two head-to-head dimers.
The binding pocket is defined by the surface of both opposite monomers confining the access of
ligands to the active site. In addition, the aromatic cap group of PFSAHA undergoes an edge-to-
face aromatic interaction with phenylalanine from the opposite monomer.
2

Human zinc-dependent histone deacetylases (HDACs) are established cancer targets with four
inhibitors (vorinostat, romidepsin, belinostat and panabinostat) being approved by the FDA for
the treatment of cutaneous T-cell lymphoma or multiple myeoloma (Fig. 1).1-4 HDACs have also
suggested as targets in other indication areas, such as neurodegenerative, heart or inflammatory
diseases.5-9 HDAC like amidohydrolases are also found in protozoa or bacteria. Several agents
against malaria and Schistosomiasis have been developed with a strong focus on selectivity
against human HDACs in order to avoid side effects.10-12 While the development of anti-parasitic
HDAC inhibitors has been pushed for almost a decade, the potential of HDAC inhibition as
antibacterial strategy has only recently been explored. In the light of ongoing increases of
antibiotic resistance in bacteria, there is an urgent need for effective alternatives in antibacterial
therapeutics.13 Interestingly, P. aeruginosa, one of the most threatening hospital pathogen,
contains three HDAC like enzymes (PA3774, PA0321 and PA1409), of which two are capable to
deacetylate acetylated polyamines, and one of them being an acetyl-lysine deacetylase with
hitherto unknown substrate.14 Although the function of these enzymes is not fully understood, we
could show an inhibitory effect on the proliferation of P. aeruginosa upon inhibition of the
HDAC-like amidohydrolases PA0321 and PA1409 under glucose starvation.14 We also
described, for the first time, novel photoswitchable inhibitors of the bacterial HDAC homologs
as potential agents in photopharmacological antimicrobial chemotherapy.15 In earlier work we
synthesized analogs of PFSAHA as weak inhibitors of human HDACs with some preference for
class II HDACs.16
3

Figure 1: HDAC inhibitors approved by the FDA for the treatment of different types of cancer.
Notably, these compounds showed also activity against a bacterial HDAC homolog from
16
Bordetella/Alcaligenes which was tested as reference enzyme.
The selectivity of the
perfluorinated hydroxamic acids for this bacterial enzyme with respect to human HDACs (up to
50-fold for 1) was even greater than among human HDAC classes (< 3-fold for 1). This
observation led us to the hypothesis that these compounds could also inhibit homologous
HDACs from P. aeruginosa.
In this study, we synthesized a series of new perfluorinated hydroxamic acids and explored the
structure-activity and structure-selectivity relationship of this class of inhibitors against a panel
of bacterial and human HDAC homologs. The new molecules described here were essentially
prepared following the previously published synthetic route.16 In short, dodecafluorooctanedioic
acid was transformed to its dichloride and further reacted with 1.5 mol equivalent benzylalcohol
to yield the corresponding mono ester. The final compounds were obtained after reaction with
different
4

Table 1: Selectivity profile of perfluorated hydroxamic acids against bacterial and human members of the HDAC protein family.
KI (µM)
#
R1
R2
PA0321
PA1409
PA3774
HDAH
HDAC1
HDAC6
HDAC7
HDAC8
-OH
0.7±0.2
2.6±0.1
0.022±0.001
0.54±0.02a
29±5a
11±2a
11±1a
17±1a
1
-OH
-OH
2.3±0.1
6.2±0.4
11±1
0.19±0.02
0.27±0.02a
0.45±0.08
14±1a
>50
>50a
>50
>50a
>50
34±5a
>50
2
3
0.27±0.03
0.092±0.004
-OH
84±0.1
2.1±0.1
>50
0.16±0.01
0.62±0.08
>50
>50
>50
>50
4
-OH
-OH
-OH
5.7±0.4
11±1
>50
0.17±0.01
0.046±0.004
0.16±0.01
0.086±0.008
0.17±0.02
0.24±0.08
12±3a
>50
23±5a
>50
19±6a
>50
35±3a
>50
5
6
7
0.073±0.0
06
0.8±0.1
>50
>50
>50
>50
5

-OH
-OH
-OH
0.49±0.05
1.1±0.1
>50
0.46±0.04
0.046±0.004
0.32±0.04
0.33±0.08
0.39±0.03a
0.23±0.01a
>50
>50
4.2±1a
>50a
>50
>50
8
9
4.1±0.4
7.1±0.4
14±9a
18±1a
4.1±0.4a
32±4a
36±5a
35±4a
2.8±0.4
10
-OH
3.6±0.2
7.6±0.4
0.32±0.04
1.3±0.2a
18±3a
9.2±2a
13±5a
36±5a
11
-OH
-OH
0.37±0.02
>50
>50
>50
1.2±0.1
2.7±0.2
0.18±0.03
>50
>50
>50
>50
>50
>50
>50
>50
>50
12
13
-OH
1.6±0.1
5.8±0.4
0.10±0.01
0.11±0.01
7.0±0.8
20±10
2.5±0.4
8.4±0.6
14
-OH
-OH
2.1±0.1
7±1
6.7±0.4
21±3
0.24±0.02
0.35±0.04
0.19±0.02a
0.37±0.03a
16±4a
29±3a
1.3±0.2a
32±8a
3.0±0.4a
20±4a
13±1a
19±6a
15
16
6

-OH
-OH
1.2±0.2
1.5±0.2
3.8±0.2
3.1±0.4
0.53±0.1
0.16±0.02a
3.8±1.1a
12±1a
6.8±1.8a
49±36a
11±1a
14±1a
49±12a
>50a
17
18
0.046±0.007
31±13a
-OH
1.5±0.1
4.4±0.4
0.10±0.01
0.45±0.05a
30±6a
15±5a
14±1a
>50a
19
-OH
-OH
-OH
12±2
24±7
9±3
17±3
>50
1.1±0.2
2.0±0.2
1.1±0.2
0.86±0.05a
9.1±0.8a
49±6a
>50a
9.2±0.9a
>50a
14±3a
38±8a
13±2a
5.3±0.5a
>50a
20
21
22
11±1
0.74±0.06a
14±1a
5.9±0.8a
1.4±0.3a
-OH
6±1
11±1
0.25±0.03
0.29±0.04a
18±2a
10±2a
11±1a
5.3±0.8a
23
-OH
-OH
4.1±0.5
>50
16±3
>50
0.23±0.02
2.6±0.2
0.6±0.2a
7.0±1a
22±6a
>50a
17±45a
8.2±2a
>50a
47±11a
>50a
24
25
30±7a
7

-OH
-OH
22±5
15±1
3.1±0.3
4.5±0.5a
2.2±0.2a
>50a
47±9a
12±2a
>50a
26
27
8.5±
>50
>50
>50
12±1
10±1
2.6±0.2a
5.4±1.3a
1.4±0.2a
1.5±0.3a
27±3a
>50a
>50a
>50a
>50a
28
0.32±0.02 0.12±0.01 0.014±0.001
0.30±0.02a
0.04±0.01a
0.3±0.07a
36±16 a
2.9±0.2a
SAHA
a
16
Data from Ref.
, HDAH is the HDAC-like amidohydrolase from Bordetella/Alcaligenes. All KI values were determined using a fluorogenic enzyme activity assay.17 For
HDAC1/6 Boc-Lys(Ac)-AMC and for HDAC4/5/7/8 Boc-Lys(trifluoroacetyl)-AMC were used as substrates. After an incubation time of 30 min the deacetylated substrate was
converted into a fluorescent dye upon addition of trypsin. KI values were calculated from IC50 values according to Cheng and Prusoff.18 IC50 values were calculated from dose
response curve containing 10 different ligand concentrations. The KI values for HDAC4 and HDAC5 were > 50 µM for all compounds except for 1 (HDAC5: 13 ± 2µM) and
SAHA (HDAC4: 14 ± 2µM, HDAC5: 12 ± 2µM).
8

aniline analogs and subsequent transformation into hydroxamic acids by hydroxylamine. Details
of organic synthesis and chemical analysis are provided in the Supplemental Data.
The biochemical activity data are summarized in table 1 and demonstrate the remarkable
selectivity of the perfluorinated compounds for the HDAC like enzymes from P. aeruginosa.
Compound 1 for example is at least 500-fold more active against PA3774 than for any human
HDAC isoform (HDAC1: 1300-fold; HDAC4, 2300-fold; HDAC5: 590-fold; HDAC6: 500-fold;
HDAC7: 500-fold and HDAC8: 770-fold).19 With the exception of 12 and 27, all PFSAHA
analogs inhibit PA3774 stronger than PA0321 and PA1409. Obviously, the binding pocket of
PA3774 is better suited to accommodate the perfluorinated alkyl chain than any other human
HDAC or HDAC like enzyme from P. aeruginosa. At the time of this study no structural data of
the HDAC homologs from P. aeruginosa was available. Therefore, we performed a systematic
variation of substituents at the phenyl ring of 1 starting with methyl- and chlorine-scanning
(compounds 2-7) to dissect the influence of the cap group on potency and selectivity. Methyl- as
well as chlorine- substitution leads to a dramatic drop of potency against PA3774 that was not so
pronounced for the other enzymes of P. aeruginosa. Substitution at meta-positions leads to the
highest activity against PA3774 in both, methyl and chlorine, scans indicating favorable
hydrophobic interactions which is confirmed by a comparable activity of a derivative with
phenyl group in meta-orientation (19). In case of PA0321 the meta-substituted analogs with
methyl and chlorine were even more potent than the chemical starting point 1. The potency of
PFSAHA analogs increases from F- to Cl- and I-substitution in para-position (8 < 5 < 9). In
addition, a phenyl group at this position (18) leads to similar activity than iodine pointing
towards hydrophobic interactions which were not so obvious from the methyl- and chlorine-
scans. This finding is in agreement with unfavorable effects of polar groups (10, 11, 12, 13) in
para-position.
9

Fig. 2: Binding pose of 1 within the active site of PA3774 (PDB: 5G11). The ligand chelates the catalytic zinc ion at the bottom
of the binding pocket in a bidentate fashion and is mainly surrounded by hydrophobic amino acid residues. The binding pocket is
formed by two protein monomers (colored in blue and pink) resulting in a particularly hydrophobic upper rim (L25, P26, P339,
L340, F343). A characteristic feature is the edge-to-face aromatic interaction between the cap group of 1 and F343 of the
opposite monomer.
During the phase of preparation of this report the crystal structure of a complex between the
Y313F mutant of PA3774 and 1 (PDB: 5G11) was determined by our group revealing molecular
details of this interaction.19 Although the phenylalanine adopts an “outward” conformation in the
mutant protein and the tyrosine in the unliganded wildtype PA3774 (PDB: 5G0Y) and its
complex with SATFMK (PDB: 5G10) shows an “inward” conformation, all other amino acids of
the binding pocket perfectly superimpose on one another. In addition, both PA3774 complexes,
with PFSAHA and SATFMK, show the same characteristic trigonal bipyramidal zinc
complexation. We demonstrated that the PA3774 forms a tetramer which is more correctly
classified as a dimer of two head-to-head dimers. Furthermore, the head-to-head dimer interface
confines the entrance area of the binding pocket suggesting a significant impact on the molecular
10

recognition of ligands.19 The complex structure of PA3774 and 1 not only confirms the bidentate
chelation of the catalytic zinc ion at the bottom of the active site binding pocket as observed in
numerous crystal structures of homologeous HDAC enzymes (e.g. HDAC4 (PDB: 4LXZ) or
HDAC8 (PDB: 2V5X), but also a striking edge-to-face aromatic interaction between the cap
group of 1 and F343 of the opposite monomer (Fig. 2; PDB: 5G11). Moreover, the upper rim of
the binding pocket is particularly hydrophobic due to amino acids L25, P26, P339, L340 and
F343 from both monomers. This region has considerable potential for forming hydrophobic
interactions with unpolar substituents in para-position of the phenyl ring in PFSAHA analogs.
However, none of the PFSAHA analogs with small group substitutions at the aromatic ring
shows higher activity than 1 on PA3774 suggesting that potential hydrophobic interactions with
substituents in para- and meta-position of the cap group are counteracted by steric hindrance
within the binding pocket of the enzyme. This is supported by the observation that bulkier cap
groups like naphthalene, anthracene, phenanthrene or benzpyrene also decrease the activity
against PA3774. Interestingly, the unfavorable effects of polar groups in para-position of the cap
group are more pronounced than the adverse steric effects of the large benzpyrene group. The
sterically demanding phenyl-triazol substituents in para- and meta-position (25, 26) are, probably
on account of steric hindrance, also detrimental to potency. If the phenyl ring in 1 is replaced by
a thiazole ring, the activity against PA3774 is particularly strongly affected suggesting a vital
role of the phenyl group in the molecular recognition of PA3774 with was verified as edge-to-
face aromatic interaction in the crystal structure of a PA3774 / 1 complex (Fig. 2; PDB: 5G11).
These qualitative insights into the modulation of biological activity by variations at the cap group
of PFSAHA were expanded by a quantitative structure-activity relationship (QSAR) model on
the basis of IC50 values for all perfluorinated compounds listed in table 1 to summarize the
relationship between chemical structure and biological activity against the target protein PA3774
11

using 15 2D physico-chemical descriptors. The model shows good correlation between
experimental and predicted log10(IC50) values with a coefficient of determination of 0.85 (Fig. 3).
A direct comparison of SAHA and its perfluorinated analog 1 reveals similar inhibitory activity
against PA3774, but dramatically decreased activity of 1 against human HDAC1 and HDAC6
and to a lesser extent the other tested human HDAC isoforms (Table 1). The high selectivity of 1
for PA3774 is supposed to be induced by the only difference between SAHA and 1 with regard
to the fluorination of the alkyl spacer. Apart from the L1-loop that is much longer in PA3774
than in human HDACs the crystal structures of the binding site pocket between the entrance and
the catalytic zinc ion are highly conserved and superimposing in PA3774 and human HDACs.
Moreover, the fundamental importance of flexible loops lining the binding pocket of HDACs,
particularly HDAC8, for the molecular recognition of substrates and inhibitors has been
20-23
recognized by several groups.
Therefore, we hypothesize that the bulkier nature of 1 along
with different flexibilities of the target enzymes may have a considerable effect on selectivity in
ligand recognition.
In summary, we prepared and profiled a variety of PFSAHA derivatives on a panel of bacterial
and human HDAC enzymes. Compound 1 is a potent inhibitor of PA3774 from P. aeruginosa in
the nano-molar range and simultaneously extraordinarily selective against human HDACs. Two
major structural features were identified which are responsible for both, potency and selectivity,
of the PFSAHA analogs against PA3774. Firstly, the perfluorinated alkylchain of the active
substances fits better into the binding pocket of PA3774 than in human HDACs and, secondly,
the active site binding pocket of PA3774 is formed by the surfaces of two interacting monomers
in head-to-head dimers and the phenyl ring of 1 undergoes edge-to-face pi-stacking with a
phenylalanine of the opposite monomer. The structure-activity relationship for PA3774 suggests
that potentially favorable hydrophobic interactions of meta- or para- substituents at the phenyl
12

ring of PFSAHA analogs with the upper hydrophobic rim of the dimeric binding pocket are
overcompensated by adverse steric effects.
Fig. 3: Correlation of experimental and predicted pIC50 values against HDAC like PA3774 from P. aeruginosa. Each data point
represents one perfluorinated hydroxamic acid analog listed in table 1. The predicted pIC50 values were calculated using a QSAR
model based on 15 physico-chemical descriptors.
ACKNOWLEDGEMENTS
This work was supported by the DFG (grant GZ: ME 3122/2-1).
Supporting Information
13

Detailed synthesis procedures, analytical data, description of biochemical assays and QSAR
model.
14

References
1.
Marks PA, Breslow R. Dimethyl sulfoxide to vorinostat: development of this histone
deacetylase inhibitor as an anticancer drug. Nat Biotechnol. 2007;25(1): 84-90.
2.
Reddy SA. Romidepsin for the treatment of relapsed/refractory cutaneous T-cell
lymphoma (mycosis fungoides/Sezary syndrome): Use in a community setting. Crit Rev Oncol
Hematol. 2016;106: 99-107.
3.
cell lymphoma. Future oncology (London, England). 2015;11(11): 1659-1664.
4. Laubach JP, Moreau P, San-Miguel JF, Richardson PG. Panobinostat for the Treatment
of Multiple Myeloma. Clin Cancer Res. 2015;21(21): 4767-4773.
5. Mielcarek M, Zielonka D, Carnemolla A, Marcinkowski JT, Guidez F. HDAC4 as a
Rashidi A, Cashen AF. Belinostat for the treatment of relapsed or refractory peripheral T-
potential therapeutic target in neurodegenerative diseases: a summary of recent achievements.
Front Cell Neurosci. 2015;9: 42.
6.
disease. Pharmacol Ther. 2015;147: 55-62.
7. Fan Z, Yang J, Yang J, Yang C, Ping Z. HDAC inhibition: A novel therapeutic approach
for atherosclerosis. International journal of cardiology. 2016;202: 722-723.
8. Felice C, Lewis A, Armuzzi A, Lindsay JO, Silver A. Review article: selective histone
Abend A, Kehat I. Histone deacetylases as therapeutic targets--from cancer to cardiac
deacetylase isoforms as potential therapeutic targets in inflammatory bowel diseases. Alimentary
pharmacology & therapeutics. 2015;41(1): 26-38.
9.
Das Gupta K, Shakespear MR, Iyer A, Fairlie DP, Sweet MJ. Histone deacetylases in
monocyte/macrophage development, activation and metabolism: refining HDAC targets for
inflammatory and infectious diseases. Clinical & translational immunology. 2016;5(1): e62.
10.
inhibitor analogues. Antimicrob Agents Chemother. 2008;52(4): 1454-1461.
11. Giannini G, Battistuzzi G, Vignola D. Hydroxamic acid based histone deacetylase
Andrews KT, Tran TN, Lucke A, et al. Potent antimalarial activity of histone deacetylase
inhibitors with confirmed activity against the malaria parasite. Bioorg Med Chem Lett.
2015;25(3): 459-461.
12.
Heimburg T, Chakrabarti A, Lancelot J, et al. Structure-Based Design and Synthesis of
Novel Inhibitors Targeting HDAC8 from Schistosoma mansoni for the Treatment of
Schistosomiasis. J Med Chem. 2016.
13.
Antimicrob Agents Chemother. 2013;57(10): 4605-4607.
14. Kramer A, Herzer J, Overhage J, Meyer-Almes FJ. Substrate specificity and function of
acetylpolyamine amidohydrolases from Pseudomonas aeruginosa. BMC Biochem. 2016;17(1): 4.
15. Weston CE, Kraemer A, Colin F, et al. Towards photopharmacological antimicrobial
chemotherapy using photoswitchable amidohydrolase inhibitors. ACS Infectious Diseases. 2016.
16. Henkes LM, Haus P, Jager F, Ludwig J, Meyer-Almes FJ. Synthesis and biochemical
Shlaes DM, Sahm D, Opiela C, Spellberg B. The FDA reboot of antibiotic development.
analysis of 2,2,3,3,4,4,5,5,6,6,7,7-dodecafluoro-N-hydroxy-octanediamides as inhibitors of
human histone deacetylases. Bioorg Med Chem. 2012;20(2): 985-995.
17.
Wegener D, Hildmann C, Riester D, Schwienhorst A. Improved fluorogenic histone
deacetylase assay for high-throughput-screening applications. Anal Biochem. 2003;321(2): 202-
208.
15

18.
Cheng Y-C, Prusoff WH. Relationship between the inhibition constant (K1) and the
concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction.
Biochem Pharmacol. 1973;22(23): 3099-3108.
19.
Kraemer A, Wagner T, Yildiz Ö, Meyer-Almes FJ. Crystal structure of a histone
deacetylase homolog from Pseudomonas aeruginosa. Biochemistry. 2016: DOI
10.1021/acs.biochem.1026b00613.
20.
Deschamps N, Simões-Pires CA, Carrupt P-A, Nurisso A. How the flexibility of human
histone deacetylases influences ligand binding: an overview. Drug Discov Today. 2015;20(6):
736-742.
21.
Decroos C, Clausen DJ, Haines BE, Wiest O, Williams RM, Christianson DW. Variable
active site loop conformations accommodate the binding of macrocyclic largazole analogues to
HDAC8. Biochemistry. 2015;54(12): 2126-2135.
22.
insights into the class I histone deacetylases. Structure. 2004;12(7): 1325-1334.
23. Dowling DP, Gantt SL, Gattis SG, Fierke CA, Christianson DW. Structural Studies of
Somoza JR, Skene RJ, Katz BA, et al. Structural snapshots of human HDAC8 provide
Human Histone Deacetylase 8 and Its Site-Specific Variants Complexed with Substrate and
Inhibitors†,‡. Biochemistry. 2008;47(51): 13554‐13563.
16

Graphical abstract
17