Synthesis and biological effects of small molecule enhancers for improved recombinant protein production in plant cell cultures

Article abstract of DOI:10.1016/j.bioorg.2019.103452

Histone deacetylases are involved in chromatin remodelling and thus play a vital role in the epigenetic regulation of gene expression. HDAC inhibitors alter the acetylation status of histone and non-histone proteins to regulate various cellular events suc

Full text of DOI:10.1016/j.bioorg.2019.103452

Journal Pre-proofs
Synthesis and biological effects of small molecule enhancers for improved re-
combinant protein production in plant cell cultures
Bárbara A. Rebelo, Rita B. Santos, Osvaldo S. Ascenso, Ana Cláudia
Nogueira, Diana Lousa, Rita Abranches, M. Rita Ventura
PII:
S0045-2068(19)31574-3
DOI:
https://doi.org/10.1016/j.bioorg.2019.103452
YBIOO 103452
Reference:
To appear in:
Bioorganic Chemistry
Received Date:
Revised Date:
Accepted Date:
20 September 2019
4 November 2019
15 November 2019
Please cite this article as: B.A. Rebelo, R.B. Santos, O.S. Ascenso, A. Cláudia Nogueira, D. Lousa, R. Abranches,
M. Rita Ventura, Synthesis and biological effects of small molecule enhancers for improved recombinant protein
production in plant cell cultures, Bioorganic Chemistry (2019), doi: https://doi.org/10.1016/j.bioorg.2019.103452
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. 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.
© 2019 Published by Elsevier Inc.

Synthesis and biological effects of small molecule enhancers for improved
recombinant protein production in plant cell cultures
Bárbara A. Rebelo1,2, Rita B. Santos1, Osvaldo S. Ascenso2, Ana Cláudia Nogueira1,
Diana Lousa3, Rita Abranches1#, M. Rita Ventura2#
1Plant Cell Biology Laboratory, Instituto de Tecnologia Química e Biológica António Xavier
(ITQB NOVA), Universidade Nova de Lisboa, 2780-157 Oeiras, Portugal.
2Bioorganic Chemistry Laboratory, Instituto de Tecnologia Química e Biológica António
Xavier (ITQB NOVA), Universidade Nova de Lisboa, 2780-157 Oeiras, Portugal.
3Protein Modelling Laboratory, Instituto de Tecnologia Química e Biológica António Xavier
(ITQB NOVA), Universidade Nova de Lisboa, 2780-157 Oeiras, Portugal.
# Corresponding authors:
Rita Abranches (ritaa@itqb.unl.pt)
M. Rita Ventura (rventura@itqb.unl.pt)
1. GMP (good manufacturing practice
2. Nicotiana tabacum L. cv. Bright Yellow 2 (BY-2)
3. Lipocalin-type prostaglandin D2 synthase (L-PGDS)
4. Small molecule enhancers (SMEs)
5. Histone deacetylases inhibitors (HDACi)
6. Histone deacetylases (HDAC)
7. Histone acetyltransferase (HAT)
8. Suberoylanilide hydroxamic acid (SAHA)
9. Sodium butyrate (SB)
10. Dimethylsulfoxide (DMSO)
11. Trichostatin A (TSA)
12. Binding free energy (BFE)
13. Total soluble protein (TSP)
14. Basic Local Alignment Search Tool (BLASTp)

Abstract
Histone deacetylases are involved in chromatin remodelling and thus play a vital role in the
epigenetic regulation of gene expression. HDAC inhibitors alter the acetylation status of
histone and non-histone proteins to regulate various cellular events such as transcription.
Novel HDAC inhibitors were designed and synthesized to promote higher levels of
recombinant protein production in tobacco cell cultures. The effect of these chemical
enhancers on the epigenetic profiles in plant cells has been evaluated by molecular docking,
in vitro and in vivo studies. The addition of these novel enhancers led to an increase in
histone H3 acetylation levels that promoted an increase in the accumulation levels of the
recombinant protein in cell culture. These results can pave the way for the application of
these enhancers to improve the production of high value products in plant cell based
systems.
Key words
Histone acetylation; HDAC inhibitors; Recombinant protein; Lipocalin-type prostaglandin D2
synthase; Nicotiana tabacum BY-2 cells; Organic synthesis; Hydroxamic acids; Molecular
modelling.
2

1. Introduction
For the past few decades, the use of plant-based systems for the production of recombinant
proteins has become an increasingly interesting alternative. Roughly 250
biopharmaceuticals have received licenses in the USA and Europe, and several more are
currently in clinical trials [1]. These products, such as traditional vaccines, proteins and
biomolecules are usually extracted from a biological source and represent an important
category of pharmaceutical products [1].
Plant cell cultures offer significant advantages when compared with traditional systems, such
as bacterial or animal cells, namely in terms of cost reduction and increase in safety [2].
They are able to secrete products into the culture medium efficiently [3,4], they can grow in a
controlled and confined environment and the risk of contamination with pathogens or
endotoxins is reduced [5,6]. The downstream processing is simple and low-cost and this
expression system is GMP compliant (good manufacturing practice; [1,2,4,6]).
Nicotiana tabacum L. cv. Bright Yellow 2 (BY-2) cell cultures are the most widely used
system for the production of recombinant proteins due to well-established growth
parameters, efficient gene transfer and expression, as well as generation of high biomass
[7,8]. In this work, we used a cell line of this species expressing a human recombinant
lipocalin-type prostaglandin D2 synthase (L-PGDS) that had been previously established in
the lab. This protein is one of the main compounds of the cerebrospinal fluid and catalyses
the isomerization of PGH2, a common precursor of various prostanoids, to produce PGD2. It
has many functions, both physiological and pathological [9]. Given the high number of
potential applications for this enzyme, the recombinant production of L-PGDS can be a new
and important research path for the putative use of this protein in therapeutic and diagnostic
applications [10].
One of the major bottlenecks of plant-based systems, which needs to be overcome, is the
low yields obtained so far. In mammalian cell cultures, there has been investment in
optimizing culture conditions for maximum production through the use of small molecule
enhancers (SMEs). These enhancers are generally divided into three broad classes:
acetamides, carboxylic and hydroxamic acids [11]. The most commonly used SMEs are
described in Figure 1.
3

Acetamides
Carboxylic acids
Hydroxamic acids
O
O
O
S
NHOH
N
O
DMSO
TSA
ONa
O
SB
H
O
OH
N
NHOH
O
VPA
SAHA
Figure 1. Chemical structure of sodium butyrate (SB), dimethylsulfoxide (DMSO), valproic
acid (VPA), trichostatin A (TSA) and suberoylanilide hydroxamic acid (SAHA).
These molecules, also known as histone deacetylase inhibitors (HDACi), enhance
transcription by blocking histone deacetylases (HDAC) activity. HDAC are responsible for
the removal of acetyl groups from histones and their activity is regulated in the cell by the
availability of acetyl groups and by the HDAC/HAT (histone acetyltransferase) ratio. The
blocking of the HDAC enzyme activity results in the hyperacetylation of histones. This
potentially neutralizes the positively charged histone tails. In turn, acetylated histones
associate less tightly with the negatively charged DNA, resulting in the relaxation of
chromatin, facilitating access of the transcriptional machinery and potentially leading to
higher protein expression [12,13]. Histone acetylation and deacetylation play an important
role in the regulation of gene expression. Decreasing HDAC activity facilitates gene
expression, potentially leads to higher mRNA synthesis and, consequently, protein
production [14]. We have applied this strategy for the first time to plant cell cultures, showing
that a commercial HDACi SAHA (suberoylanilide hydroxamic acid) is able to promote
histone acetylation and increase transgene expression [15].
HDAC proteins are grouped into four classes by their size, number of catalytically active
sites, homology and cellular localization. Class III operates in a NAD+-dependent manner,
while other classes present a metal ion-dependent mechanism. Class I, with homology to
yeast protein RPD3, features an active site with a Zn2+ ion and some adjacent conserved
residues, important for the removal of an acetyl group from the histone tails [16,17].
Commercial HDACi are designed to structurally mimic the HDAC aliphatic acetyl-lysine
substrate and can be divided in three essential domains: capping group, linker and metal
binding moiety. The optimization of these domains is crucial for the design of compounds
4

that better fit into the catalytic centre and, therefore, have higher effectiveness as HDACi
[14,17].
In the present work, we designed and synthesised novel compounds which could act more
efficiently as HDACi in plant cell cultures, and promote higher levels of recombinant protein
production. We evaluated the biological effects of HDACi addition in BY-2 cell cultures. We
analysed whether these compounds promoted cell toxicity and, if not, whether they showed
a significant capacity of inhibiting BY-2 HDAC. We measured the synthesised HDACi
effectiveness by an in vitro test and analysed the levels of recombinant L-PGDS
accumulated in the culture medium. Additionally, docking studies were performed as a tool
to understand the ligand-receptor interactions and to select the compounds to be tested in
the following step. Finally, we investigated the acetylation levels of BY-2 histones after the
addition of two selected HDACi. With this work, we aim to show that designing and using
HDACi as a strategy for transcription enhancement has multiple applications and can be
applied to different fields of research.
5

2. Results and Discussion
2.1.
Design and synthesis of new HDACi
A library of compounds (see Table 1 and Table 2) for enhancing recombinant protein
production in plant cell cultures was designed and synthesised. Our starting point was based
on previously published syntheses of HDACi, like TSA and SAHA (Figure 1), and their
structure can be divided in tree domains, as represented in Figure 2.
Capping group
Metal binding moiety
O
H
N
NHOH
O
Linker
Figure 2. Structure of SAHA representing the capping group, linker and metal binding
moiety.
The capping group is solvent-exposed and interacts with amino acid residues at the
entrance region of the active site. Thus, the aromatic ring was maintained and we introduced
functional groups at different positions (ortho, meta and para). It was relevant to introduce
fluorine atoms since the carbon-fluorine bond is commonly found in pharmaceutical and
agrochemical products. The fluorine atom is stable and acts as a bioisostere of the hydrogen
atom [18,19]. The metal binding moiety resides in the protein interior and coordinates with
the zinc atom within the HDAC active site. In the present work we compared the activity of
two good metal chelators, hydroxamic and carboxylic acid groups. The linker positions the
domains described above for interaction in the active site. Our compounds possessed an
eight carbon saturated chain (Table 2) or a trans 2-butene linker (Table 1). This last linker is
constrained with the presence of the double bound, which has been described in other
HDACi [14,17,20]. We were interested in evaluating the influence on the biological activity
prompted by the presence of functional groups and the size and rigidity of the carbon chain.
2.1.1. Synthesis of the new compounds
From (E)-4-ethoxy-4-oxobut-2-enoic acid 28, compounds 3-7 were synthesised in two steps,
briefly described in Scheme 1. The coupling reaction, which led to the formation of amides,
was accomplished with 63-83% yield. The resulting product may follow two synthetic routes.
Reaction with hydroxylamine hydrochloride and NaOH in MeOH afforded the corresponding
6

hydroxamic acids 3-5 in 17-46% yield. Treatment of 29 with NaOH in THF gave the
corresponding carboxylic acids 6 and 7, in 90% and 87% yield respectively (Table 1).
R
O
O
NH2
H
N
HO
OEt
OEt
(a)
O
O
29
28
R
O
H
N
(b)
(c)
NHOH
29
O
O
R
3 5
-
O
H
N
OH
29
R
6 7
-
Scheme 1. Synthesis of compounds 3-7. Reactions conditions: (a) EDC, DMF, rt; (b)
NH2OH·HCl, NaOH, MeOH, 0˚C; (c) NaOH, THF, rt.
R
O
O
H
NH2
N
HO
OMe
OMe
(a)
O
O
31
30
R
O
H
N
(b)
(c)
31
31
NHOH
O
O
8 15
-
R
R
O
H
N
OH
18 26
-
Scheme 1. Synthesis of compounds 8-15 and 18-26. Reactions conditions: (a) EDC, DMF,
rt; (b) NH2OH·HCl, NaOH, MeOH, 0˚C; (c) NaOH, THF, rt.
From 8-methoxy-8-oxooctanoic acid 30, the same synthetic sequence was followed
(Scheme 2). The aromatic amides were obtained in 30-95% yield, the hydroxamic acids 8-15
in 30-60% yield and the carboxylic acids 18-26 in 49-95% yield (Table 2).
7

Scheme 3 represents the synthetic route for compounds 16, 17 and 27. In order to increase
the ratio of the hydroxamic acid product over carboxylic acid impurities, a nucleophilic base
(Scheme 3) was used to obtain compound 16 in 27% yield. Phosphate-derived compounds
were synthesised in order to assess the effect of a bulkier and charged capping group.
Compound 32 was phosphorylated with dibenzylphoshite (Scheme 3 – 93% yield), followed
by the synthesis of hydroxamic (Scheme 3 – 35% yield) and carboxylic acids (Scheme 3 –
83% yield). After that, it was necessary to deprotect the hydroxyl groups (Scheme 3) to
obtain the final compounds 17 and 27, in 91% and 93% yield, respectively.
O
H
(d)
N
NHOH
O
O
HO
16
17
O
O
O
H
N
H
(e), (b), (f)
(e), (c), (f)
N
NHOH
OMe
O
O
HO P O
OH
HO
32
H
N
OH
O
O
HO P O
OH
27
Scheme 2. Synthesis of compounds 16, 17 and 27. Reactions conditions: (b) NH2OH·HCl,
NaOH, MeOH, 0˚C; (c) NaOH, THF, rt; (d) NH2OH, DBU, MeOH, rt; (e) CCl4, DIPEA, DMAP,
dibenzylphosphite, ACN, -10˚C; (f) Pd/C, H2, EtOH.
The final compounds, represented in Table 1 and Table 2, were obtained in the 60 mg to
460 mg range. They were obtained in moderate to excellent yields, allowing a quick and
efficient synthesis of a large number of structurally diverse compounds to be evaluated. All
compounds to be tested, including the commercial compound SAHA, were dissolved in
DMSO at a concentration of 100 mM and stored at -20˚C.
8

Table 1. Yields for the reactions depicted in Scheme 1.
Yield (%)
a
Yield (%)
b or c
Final compound
O
H
N
83a
17b
NHOH
O
3
F
O
H
N
63a
34b
NHOH
O
4
F
Br
O
H
N
NHOH
79a
46b
O
5
O
O
H
N
63a
79a
90c
87c
OH
O
F
Br
6
7
O
H
N
OH
O
O
Reaction conditions: (a) EDC, DMF, rt; (b) NH2OH·HCl, NaOH, MeOH, 0˚C; (c) NaOH, THF, rt.
9

Table 2. Yields for the reactions depicted in Scheme 2 and 3.
Yield (%)
a or e
Yield (%)
b, c, d or f
Final compound
O
O
H
N
54a
45b
60b
38b
30b
NHOH
NHOH
O
8
F
H
N
30a
87a
46a
O
9
I
O
H
N
NHOH
NHOH
O
O
F
F
10
O
H
N
Br
11
12
O
H
N
NHOH
87a
81a
81a
33a
95a
30b
38b
42b
39b
27d
O
O
O
H
N
NHOH
O
I
13
14
15
16
O
H
N
NHOH
O
Br
O
H
N
NHOH
O
O
Br
H
N
NHOH
O
HO
10

O
H
N
NHOH
93e
35b, 91f
O
O
HO P O
OH
17
18
O
O
H
N
54a
30a
87a
46a
59c
83c
57c
49c
OH
OH
O
O
F
H
N
I
19
20
21
O
H
N
OH
OH
O
F
F
O
H
N
O
Br
O
H
N
OH
87a
82c
O
22
O
O
O
H
N
81a
81a
33a
95a
95c
92c
93c
93c
OH
OH
O
I
23
24
H
N
O
Br
O
H
N
OH
O
25
26
Br
O
H
N
OH
O
HO
11

O
H
N
OH
93e
83c, 99f
O
27
O
HO P O
OH
Reaction conditions: (a) EDC, DMF, rt; (b) NH2OH·HCl, NaOH, MeOH, 0˚C; (c) NaOH, THF, rt; (d)
NH2OH, DBU, MeOH, rt; (e) CCl4, DIPEA, DMAP, dibenzyl phosphite, ACN, -10˚C; (f) Pd/C, H2, 50
psi, EtOH.
2.2.
Docking studies
The interaction of all synthesised compounds with HDAC was tested by performing docking
calculations. Since the structure of Nicotiana tabacum histone deacetylase 6 is unknown,
Blastp was used to select the HDAC with known structure that showed the highest sequence
identity with our target. The proteins retrieved by Blastp with highest identity to Nicotiana
tabacum belong to the species Arabidopsis thaliana, Homo sapiens and Saccharomyces
cerevisiae. The results indicated that there is a high degree of homology between the HDAC
of these organisms (identity values between 59 and 78%). The human HDAC2 in complex
with SAHA structure [Protein Data Bank (PDB) code: 4LXZ_A] was selected for the
molecular docking studies since it displayed 78% of query cover and 67% of identity. All the
residues present in the active center are conserved between the human HDAC2 and
Nicotiana tabacum histone deacetylase 6. Moreover, this structure was obtained in a
complex with the SAHA ligand and thus its active site is in the appropriate conformation to
accommodate ligands of the same type, which is an advantage for the docking studies
performed with the synthesised compounds.
The main aim of this computational study was to restrict the number of the synthesised
compounds to be tested experimentally. The molecular docking studies provided information
about the preferential binding mode and estimated the binding free energy (BFE). The
lowest binding free energy value corresponds to the best predicted binding mode for each
ligand. In order to rank the compounds regarding their efficiency in inhibiting the enzyme
activity, the preferred binding mode of each ligand was selected and the respective BFE
were compared (see Figure 3). When analysing the bar plot, compounds with a smaller
linker showed lower BFE values when compared with compounds that have an 8-carbon
linker. However, the difference between the BFE values is small and, therefore, we could not
establish which compound would perform better as HDACi, based only on this parameter.
For hydroxamic acids, compounds 17 and 13, with phosphate and iodine in the aromatic
ring, respectively, showed the best BFE value. In the case of carboxylic acids, the best result
was obtained for compound 25 with the lowest BFE value (-16.5 kcal/mol). This compound
12

has a capping group consisting of an aromatic ring with a bromine substituent. Another
carboxylic acid compound, 16, which has an iodine substituent in the aromatic ring, had a
BFE value of -16.4 kcal/mol. After analysis of these BFE values, we can conclude that there
were no relevant differences in the BFE values between different capping groups, only
between different linkers.
Figure 3. Clustering histogram of all docking solutions, as a function of BFE, for all the
synthesised hydroxamic and carboxylic acid compounds.
One of the goals of the docking calculations was to understand the influence of the
substituents in the capping group and the size of the linker. As shown in Figure 4, the 8-
carbon linker from SAHA allows a hydrogen bond to occur between the secondary amine
and the ASP104 residue of HDAC2, with a theoretical distance of 1.8 Å between the amine
hydrogen and the ASP oxygen. By analysing the predicted binding of compound 4 to
HDAC2, we conclude that the smaller linker is too spaced apart from the residues (more
than 3 Å) to form that hydrogen bond. The hydrogen from the secondary amine is 5.4 Å from
the residue ASP104 and 5.0 Å from the residue GLY154 and, thus, cannot form hydrogen
bonds with these residues. Therefore, the linker length seems to be important for the
interaction with the enzyme.
13

Figure 4. Predicted binding of SAHA (a) and compound 4 (b) to HDAC2. The elements of
the ligand are represented by red – oxygen, blue – nitrogen, white – hydrogen, yellow –
carbon, light blue – fluorine, brown – bromine, grey solid sphere – zinc. The elements of the
protein are represented by red – oxygen, blue – nitrogen, white –hydrogen, green – carbon,
light brown – sulphur.
Regarding the capping group, the results displayed in Figure 5 allowed the evaluation of the
difference caused by substituents in the aromatic ring. SAHA, which does not have
substituents on the aromatic ring, was compared with compounds with different substituents:
compound 10, which has a fluorine in the para position, and compound 4, which has a
fluorine and a bromine in the para and ortho positions, respectively. These ligands adopted
different binding modes in the HDAC active site when compared to SAHA. These halogens
are bulkier, taking up more space at the entrance of the active site, which can confer them
an improvement as HDACi. Also, the ligands were exposed to the solvent and the
hydrophilic substituents may help to stabilise the interaction with water.
14

Figure 5. Molecular docking results of HDAC2 crystal structure with docked SAHA (a and
b), compound 4 (c and d), compound 10 (e and f) into catalytic site. The images a, c and e
represent the best conformation of these compounds in the catalytic site. The images b, d
and f are a visual representation of how the conformation of these compounds occupies the
15

entrance of the catalytic site. The elements of the ligand and protein are represented with
the same colour scheme as described in Figure 4.
In a previous work, we synthesised and analysed other SAHA analogues. Those with
bromine or iodine at the para position of the aromatic ring were tested in vivo and performed
poorly (results not shown). Therefore, the compounds with these substituents were not
chosen for the following experiments.
All the other compounds with good BFE values and with different capping groups were
selected, the hydroxamic acid compounds 8-10 and 15-17, and the carboxylic acid
compounds 18-20 and 25-27. Furthermore, the compounds with a smaller linker and that
presented a methoxide group in the meta position, compounds 5 and 7, or bromine and
fluorine in the ortho and para positions of the aromatic ring, compounds 4 and 6, were
chosen. This selection allowed the understanding of how the linker size of the compounds
influenced their activity.
2.3.
Biological evaluation
2.3.1. In vitro HDAC inhibition assay
The ability of the synthesised compounds to inhibit HDAC activity was analysed using an in
vitro HDAC inhibition assay (Epigentek, USA). This assay aimed at selecting the best
performing compounds to be used in the in vivo assays with BY-2 cell cultures. Compounds
to be tested in the in vitro assays were chosen from all the synthesised compounds based
on the following criteria: 1) compounds with shorter unsaturated linker and with an eight
carbon linker; 2) compounds with different aromatic group substituents; 3) hydroxamic and
carboxylic acid compounds with the same capping group. This selection was also performed
taking into account the molecular docking studies, previously described. Based on these
studies, the hydroxamic acid compounds 3, 11-14 and the carboxylic acid compounds 21-24
were excluded. The commercial HDACi SAHA was used as a reference and the results
obtained are presented in Figure 6.
16

Figure 6. Inhibition of HDAC enzymatic activity by synthesised compounds and SAHA.
Control: sample without inhibitor addition. Error bars indicate S.D. for n=3. Horizontal dashed
line represents average % of HDAC activity in the presence of SAHA. ns: non-significant; **:
p value < 0.01; ***: p value < 0.001.
Figure 6 represents the % of HDAC activity in the presence of synthesised compounds,
SAHA and control sample, i.e., BY-2 nuclear extracts without the addition of an HDACi.
Overall, most compounds tested caused a decrease in HDAC activity, except for carboxylic
acids 6, 25 and 27. In fact, carboxylic acids performed worse than the hydroxamic acids, in
general terms. In contrast, all the hydroxamic acid compounds, with the exception of
compound 17, had a stronger effect in decreasing HDAC activity when compared to SAHA.
This analysis allowed for a more informed choice of which compounds should be tested in
the in vivo assays. For the in vivo experiments, the pairs of hydroxamic and carboxylic acids
were tested and then selected as follows:
i)
4 and 6: compounds with a 4 carbon linker and F and Br as substituents in ortho
and para positions, respectively; 4 performed better than SAHA.
17

ii)
8 and 18: compounds with an 8 carbon linker and F as substituent in the ortho
position; 8 performed better than SAHA.
iii)
iv)
10 and 20: compounds with an 8 carbon linker and F as substituent in the para
position; 10 performed better than SAHA.
15 and 25: compounds with an 8 carbon linker and Br as a substituent in the
ortho position; 15 performance was similar to SAHA and this pair allows for
analysis of the role of Br in the inhibition of HDAC.
2.3.2. In vivo assays and analysis of recombinant protein production
To assess the effect of the compounds in vivo, we used a BY-2 cell culture producing a
human recombinant protein L-PGDS that had been previously generated in the lab. Each
compound was added to the cell culture medium on day 2 of growth, to a final concentration
of 10 µM. Cells and culture medium were recovered on days 0, 2, 4, 7 and 10 of growth. A
cell culture with no added inhibitor was monitored as a control. The dry weight of the cells
was measured and used to build a growth curve. Culture medium was used to quantify total
soluble protein (TSP) by Bradford assay and to perform western blots to detect the
recombinant protein.
From all the compounds tested, the hydroxamic acid 8 and the carboxylic acid 18 were the
best performers. In general terms, hydroxamic acids perform better as HDACi than
carboxylic acids due to their higher affinity to the zinc ion present in the catalytic pocket of
the enzyme. However, in this particular case, the carboxylic acid 18 had a similar
performance of that of hydroxamic acids. It should be noted that carboxylic acids are also a
good chelating group [21]. In Figure 7 and 8, the growth curves, TSP, western blots and
relative quantification of recombinant L-PGDS are presented for both compounds 8 and 18.
18

Figure 7. Analysis of BY-2 cultures expressing L-PGDS with and without the addition of 10
µM of compound 8. a) chemical structure of compound 8; b) growth curve (solid line) and
total soluble protein (dashed line); c) relative expression of L-PGDS with and without
addition of 10 µM of compound 8 at day 2 obtained from d) western blot analysis of
biological triplicates. Error bars represent the standard deviation of three biological
replicates. **: p value < 0.01.
In Figure 7b, BY-2 cell growth and TSP along the 10 days of culture with compound 8 is
represented. The addition of compound 8 at day 2 affected negatively cell growth on day 7.
However, on day 10, both control culture and the culture with 10 µM of compound 8
achieved the same dry weight value. Although cell growth decreases when this compound
was added, the TSP was not altered in relation to control, and protein values were similar
along the 10 days of growth.
19

The production of L-PGDS throughout growth was monitored by western blot (see Figure 7c
and 7d). The addition of compound 8 resulted in a significant increase in accumulation of the
protein. The higher accumulation of L-PGDS protein is observed 5 days after the addition of
the compound (day 7), with a 1.5-fold increase when compared to the control culture. This
result is in line with the performance of this compound on the in vitro assay. The % HDAC
activity decreased in the presence of this compound (see Figure 6) and its average value
was below of that of SAHA. Even though the decrease observed on day 4 is not statistically
significant, it is possibly due to retention of the recombinant protein in the cell, as this has
been previously observed in [15].
Figure 8. Analysis of BY-2 cultures expressing L-PGDS with and without the addition of 10
µM of compound 18. a) chemical structure of compound 18; b) growth curve (solid line) and
total soluble protein (dashed line); c) relative expression of L-PGDS with and without
20

addition of 10 µM of compound 18 at day 2 obtained from d) western blot analysis of
biological triplicates. Error bars represent the standard deviation of three biological
replicates. *: p value < 0.05.
The effect of compound 18 on the cell growth and TSP of BY-2 cell cultures is shown in
Figure 8b. A slight decrease in cell growth can be observed on day 7 in the culture when
compound 18 was added at day 2. This decline in growth affected TSP values on day 10 of
growth, when compared to the control culture. The production of L-PGDS was also
monitored by western blot (Figure 8c and 8d). The addition of this compound resulted in an
increase in the accumulation of L-PGDS in the culture medium. In the same way as for
compound 8, the increase in recombinant protein was observed on day 7, five days after the
addition of compound 18. A 1.4-fold increase was observed when comparing to the control
culture. On day 10, there was a similar decrease on recombinant L-PGDS protein
accumulation in the medium, as seen in Figure 7 (compound 8), which is in accordance with
the presence of degradation bands in the western blot on day 10. Although this compound
showed less inhibitory effect of HDAC activity in the in vitro assay when compared to SAHA
(see Figure 6), it increased the level of recombinant protein in BY-2 cell medium.
The in vivo assays with compounds 4, 6, 10, 20, 15, 25 and SAHA are shown in the
supplementary results (Figure S1 to Figure S7). The addition of these compounds did not
result in an increase of recombinant protein in the cell culture. SAHA, which has been widely
described has a HDACi [11,22], did not promote an increase in protein accumulation in BY-2
cell medium. In fact, it was possible to observe a significant decrease on protein
accumulation of 44% and 88% on day 4 and 7, respectively. Compounds 4, 6, 10 and 20
promoted no significant differences or led to a decrease on TSP. Compound 4 strikingly
decreased the accumulation of recombinant protein in the cell culture medium, while
compound 6 had no significant effect. Compounds 10 and 20 promoted a decrease on
recombinant L-PGDS protein accumulation in the medium, on day 10 and 7, respectively. In
contrast, compounds 15 and 25 promoted a 2-fold increase of recombinant protein in the cell
medium on day 10. However, western blots showed the presence of degradation bands.
This might be due to the action of proteases which may compromise the final yield and
quality of the protein [6,23].
2.3.3. Quantification of L-PGDS in the cell medium
21

The concentration of L-PGDS in the cell medium of control and assay cultures was
quantified by FIJI software. A curve with BSA standards (5, 10, 20, 30 and 40 μg/mL) was
built and L-PGDS was quantified relatively to this standard curve. Only day 7 of the cell
growth was analysed and biological triplicates of control and assay samples were pooled.
SDS-PAGE gel and final quantification values are present in Figure 9.
Figure 9. Quantification of L-PGDS in the cell medium of control assays and assays with 10
μM of compounds 8 and 18. a) SDS-PAGE gel of culture medium samples of day 7 (C:
control; 8: compound 8; 18: compound 18) along with BSA standards (5, 10, 20, 30 and 40
μg/mL). The orange rectangles highlight the bands that were quantified by FIJI software.
Quantification results are presented in b).
As already described in the previous section, the addition of compounds 8 and 18 at day 2 of
cell growth resulted in a higher accumulation of L-PGDS in the cell culture medium. The
results shown in Figures 7c and 8c compared the amount of L-PGDS protein between
control and assay cell medium, resulting in a relative quantification. With this analysis (see
Figure 9b) it was possible to determine the L-PGDS concentration of the pooled samples,
with and without the presence of compound 8/compound 18. The concentration values of L-
PGDS confirmed that the relative quantification performed in Figure 7 and 8 correctly
predicted the increase of protein in the cell medium of assays with compounds 8 and 18.
This increase was of 1.44-fold for compound 8 and 1.6-fold for compound 18, which is in line
with the values obtained in the in vivo assays.
2.3.4. Histone acetylation levels
The acetylation levels of histone H3 of cell extracts from control culture and cultures with the
addition of compound 8 and 18 were analysed by western blot (see Figure 10). This specific
22

histone was chosen to be analysed since it has been described as the major target of
acetylation in plants [12]. In western blot analysis, for the calculation of fold changes, the
acetylation levels of histone H3 were normalized using total histone H3.
Figure 10. Histone H3 relative acetylation levels. a) Relative acetylation level of histone H3
with and without compound 8 and 18. Relative levels were obtained from b) western blot
analysis of nuclei extracts samples with anti acetyl-histone H3 and anti histone H3
antibodies. Error bars indicate S.D for n=3. C: control lanes; arrow heads indicate acetyl-
histone H3 and histone H3 bands.
The addition of compounds 8 and 18 led to an increase in the acetylation levels of histone
H3. Compound 8 promoted an increase of 1.26-fold in histone acetylation, while compound
18 led to a 1.56-fold increase. This surge on acetylation was observed two days after the
addition of the compounds, on day 4. On day 7 and 10 of growth, a decrease in histone H3
acetylation was observed. This has already been described by Waterborg and colleagues in
alfalfa cell cultures. In their work, authors reported that transient hypoacetylation of histone
H3 was observed after the addition of TSA to alfalfa cultures [24].
This increase in acetylation confirms the results observed in the in vitro assays, where these
compounds were shown to reduce the activity of BY-2 HDAC. Moreover, an increase in
histone acetylation on day 4 led to an increase in protein accumulation in the cell medium on
day 7. These successive events have also been described for Medicago truncatula cell
cultures upon the addition of SAHA, where an increase in acetylation and protein
accumulation was observed 2 days after SAHA addition [15].
23

24

3. Conclusion
The use of chemically synthesised compounds has gained attention in the past few years,
since they are able to generate desirable outcomes in biological systems. There has been
some progress in the production of SMEs for the treatment of medical dysfunctions or
pathologies in mammalian cell cultures or to improve recombinant protein production in
biological systems [11]. To our knowledge, this is the first report where new analogues of
commercial HDACi were tested in plant cell cultures, with the aim of improving protein
production yields, in order to overcome low yields observed in this type of expression
system. In this work, a library of 25 compounds was synthesised as potential HDACi. These
compounds were grouped in hydroxamic or carboxylic acids, by using different reaction
conditions from a common ester intermediate, with a quick and efficient synthesis. Our main
aim was to evaluate the influence of the aromatic substitution in the amide group and the
length of the carbon linker. Based on the preliminary screenings in BY-2 cultures, the part of
the molecule that is more likely to have an impact on the activity is the carbon chain linker.
The best results were obtained with the 8 carbon linker (the same linker of SAHA). However,
the nature and the position of the halogen substituent at the aromatic ring of the amide was
also important. Compounds 8 and 18 showed a higher activity in promoting the production of
L-PGDS than SAHA. Compound 8 only differs from SAHA by the presence of a fluorine
atom at the ortho position in the aromatic ring. This apparently small structural difference
had an impact in the activity of the molecule in promoting the L-PGDS production by BY-2
cell cultures. This might be due to electronic and stereochemical factors. A substituent at the
aromatic ring’s ortho position is more sterically hindered than that at the para position. Even
though no rotamers were observed for the ortho amides in NMR spectra, this steric
difference might have an influence at the HDAC binding site. Compound 18 presents a
carboxylic acid instead of a hydroxamic acid and performed better than compound 8. This
shows that this particular carboxylic acid was either more efficient as a metal chelator than
the hydroxamic acid in this set of compounds or its chemical structure probably conferred
other properties that enhanced its activity. A carboxylic acid is considerably easier to
synthesise than the corresponding hydroxamic acid, so this is a further advantage shown in
this study.
All the research accomplished derived from the application of tailor-made compounds that
were effective and non-toxic for BY-2 cells. These results were an extensive process that
went through several steps: the synthesis of each compound, docking calculations, in vitro
testing and, finally, in vivo experiments in BY-2 cell suspension cultures (Figure 11). These
findings pave the way for the rational design and synthesis of new improved HDACi, by
25

further acting at the level of chromatin-mediated regulation of gene expression to obtain
higher levels of recombinant product. This work is a substantial contribution to the
development of plant cells cultures as commercially viable platforms for the cost-effective
production of high value products.
Figure 11. Representative screening of the selection of synthesised compounds for in vivo
studies.
26

4. Experimental section
4.1. Chemistry
4.1.1. Chemical synthesis – General procedures:
1
Chemical shifts for H-NMR, acquired at a Bruker 400 MHz spectrometer, are given in ppm
13
downfield from tetramethylsilane. For C NMR, the spectra were acquired at 101 MHz and
calibrated according to the respective solvent. (77.16 ppm for CDCl3, 39.52 ppm for DMSO-
d6). Assignments are supported by 2D correlation NMR studies. Purification of each product
was done either by recrystallisation, silica flash-column chromatography on Silica gel Merck
60 or preparative thin layer chromatography. Analytical TLC: Aluminium-backed Silica Gel
Merck 60 F254. Reagents and solvents were purified and dried by conventional techniques
according to reference [25]. All the reactions were carried out under an inert atmosphere
(argon).
4.1.2. General reaction experimental procedures:
Ethyl (2E)‐3‐[(2‐bromo‐4‐fluorophenyl)carbamoyl]prop‐2‐enoate 29. To a solution of 28
(6.94 mmol, 1g) in DMF (5 mL), 2-bromo-4-fluoroaniline (10.35 mmol, 1.18 mL) and EDC
(10.35 mmol, 1.994 g) were added. The resulting solution was stirred overnight. Water was
added; the resulting mixture was extracted with AcOEt, dried with MgSO4 anhydrous,
concentrated under vacuum and recrystallised in AcOEt/Hex to afford the amide ethyl
(2E)‐3‐[(2‐bromo‐4‐fluorophenyl)carbamoyl]prop‐2‐enoate as a white powder (63% yield,
1.378 g). mp = 196 - 200˚C. 1H NMR (400 MHz, CDCl3): δ 8.42 (dd, J = 8.9, 5.6 Hz, 1H, Ar),
7.77 (s, 1H, NH), 7.33 (dd, J = 7.7, 2.8 Hz, 1H, Ar), 7.14 – 7.04 (m, 2H, Ar), 7.08 – 7.04 (m,
1H, CH=CH), 6.96 (d, J = 15.3 Hz, 1H, CH=CH), 4.30 (q, J = 7.1 Hz, 2H, OCH2CH3), 1.35 (t,
J = 7.1 Hz, 3H, OCH2CH3). 13C NMR (101 MHz, CDCl3): δ 165.1 (C=O, quaternary), 161.4
(C=O, quaternary), 159.9 (J = 48 Hz, C-F, quaternary), 135.9 (CH=CH), 132.2 (CH=CH),
131.5 (C-NH, quaternary), 123.2 (J = 8 Hz, Ar), 119.5 (J = 26 Hz, Ar), 115.4 (J = 21 Hz, Ar),
113.8 (Br-C, Ar, quaternary), 61.4 (OCH2CH3), 14.1 (OCH2CH3). FT-IR (ATR): (vmax/cm-1)
3266 (NH), 1713 (C=O), 1643 (C=O, amide).
Methyl 7‐[(2‐fluorophenyl)carbamoyl]heptanoate 31. 2-Fluoroaniline (7.26 mmol, 0.53
mL) was mixed with 30 (4.84 mmol, 0.87 mL) in DMF (4.600 mL), with further addition of
EDC (7.26 mmol, 1.380 g) and stirred overnight. Water was added, the resulting mixture was
extracted with AcOEt, dried with MgSO4 anhydrous, concentrated under vacuum and
recrystallised
in
AcOEt/Hex
to
afford
the
amide
methyl
7‐[(2‐fluorophenyl)carbamoyl]heptanoate as a white powder (54%, 0.740 g). mp = 52 -
1
60˚C. H NMR (400 MHz, CDCl3) δ 8.33 (t, J = 7.8 Hz, 1H, Ar), 7.33 (s, 1H, NH), 7.15 – 6.99
27

(m, 3H, Ar), 3.67 (s, 3H, OCH3), 2.40 (t, J = 7.5 Hz, 2H, CH2), 2.32 (t, J = 7.5 Hz, 2H, CH2),
1.70 (m, 4H, CH2-CH2), 1.39 (dd, J = 7.7, 4.3 Hz, 4H, CH2-CH2). 13C NMR (101 MHz, CDCl3)
δ 174.1 (C=O, quaternary), 171.2 (C=O, quaternary), 152.3 (J = 241 Hz, C-F, quaternary),
126.4 (C-NH, quaternary), 124.6 (J = 4 Hz, Ar), 124.1 (J = 8 Hz, Ar), 121.7 (Ar), 114.7 (J =
19 Hz, Ar), 51.4 (OCH3), 37.6 (CH2), 33.9 (CH2), 28.7 (CH2-CH2), 25.2 (CH2), 24.7 (CH2). FT-
IR (ATR): (vmax/cm-1) 3305 (NH), 2923 (OCH3), 1726 (C=O), 1667 (C=O, amide).
(2E)-N'‐(2‐bromo‐4‐fluorophenyl)‐N‐hydroxybut‐2‐enediamide 4. Compound ethyl
(2E)‐3‐[(2‐bromo‐4‐fluorophenyl)carbamoyl]prop‐2‐enoate 29 (1.21 mmol, 0.384 g) and
NH2OH·HCl (1.21 mmol, 0.088 g) were dissolved in MeOH (2.70 mL) and cooled to 0˚C.
Next, a solution of NaOH (4 M, 9.68 mmol, 2.40 mL) was added and the reaction mixture
was stirred until all starting material has been consumed. The aqueous phase was washed
with Et2O, acidified with HCl 10% until pH 3.5. The resulting mixture was extracted with
AcOEt, dried with anhydrous MgSO4 and concentrated under vacuum. Purification gave a
1
yellowish powder (34%, 0.124 g). mp = 196 - 200˚C. H NMR (400 MHz, DMSO-d6) δ 11.07
(s, 1H, NH), 10.04 (s, 1H, OH), 9.51 (s, 1H, NH), 7.71 – 7.59 (m, 2H, Ar), 7.30 (ddd, J = 8.6,
8.6, 2.9 Hz, 1H, Ar), 7.19 (d, J = 15.2 Hz, 1H, CH=CH), 6.83 (d, J = 15.2 Hz, 1H, CH=CH).
13C NMR (101 MHz, DMSO-d6) δ 163.1 (C=O, quaternary), 160.9 (C=O, quaternary), 158.6
(C-F, quaternary), 133.0 (C-NH, quaternary), 132.3 (CH=CH), 131.8 (CH=CH), 129.3 (J = 9
Hz, Ar), 120.1 (J = 25 Hz, Ar), 119.4 (C-Br, quaternary), 115.5 (J = 22 Hz, Ar). FT-IR (ATR):
(vmax/cm-1) 3299 (NH), 2943 (OH), 1738 (C=O), 1661 (C=O, amide). HR-MS: calcd. for
C10H9BrFN2O3 [M+H]+: 302.9775; found: 302.9775.
(2E)‐3‐[(2‐bromo‐4‐fluorophenyl)carbamoyl]prop‐2‐enoic acid 6. Compound ethyl
(2E)‐3‐[(2‐bromo‐4‐fluorophenyl)carbamoyl]prop‐2‐enoate 29 (0.53 mmol, 0.166 g) was
dissolved in THF (1 mL), and then a solution of NaOH (1 M, 0.53 mmol, 0.53 mL) was
added. The reaction mixture was stirred until all starting material has been consumed and
the aqueous phase was washed with Et2O and acidified with HCl 10% until pH 3.5. The
resulting mixture was extracted with AcOEt, dried with anhydrous MgSO4 and concentrated
under vacuum. Purification gave a white powder (90%, 0.136 g). mp = 231 - 232˚C. 1H NMR
(400 MHz, DMSO-d6) δ 13.04 (s, 1H, NH), 10.13 (s, 1H, OH), 7.72 – 7.62 (m, 2H, Ar), 7.35 –
7.28 (m, 1H, Ar), 7.26 (d, J = 15.5 Hz, 1H, CH=CH), 6.68 (d, J = 15.5 Hz, 1H, CH=CH). 13C
NMR (101 MHz, DMSO-d6) δ 166.7 (C=O, quaternary), 162.7 (C=O, quaternary), 159.9 (J =
246 Hz, C-F, quaternary), 158.7 (C-Br, quaternary), 136.6 (CH=CH), 132.7 (C-NH,
quaternary), 131.9 (CH=CH), 129.3 (J = 9 Hz, Ar), 120.1 (J = 26 Hz, Ar), 115.6 (J = 22 Hz,
Ar). FT-IR (ATR): (vmax/cm-1) 3261 (NH), 3050 (OH), 1698 (C=O), 1660 (C=O, amide). HR-
MS: calcd. for C10H8BrFNO3 [M+H]+: 287.9666; found: 287.9664.
28

N‐hydroxy‐N'‐(4‐hydroxyphenyl)octanediamide 16. DBU (2.36 mmol, 0.35 mL) was
added to a solution of 32 (0.788 mmol, 0.220 g) in anhydrous MeOH (0.69 mL). The mixture
was stirred at rt and an aqueous solution 50% of NH2OH (7.88 mmol, 0.24 mL) was added.
The reaction was stopped with water and the workup, generally similar to the one described
on experiment (b), afforded 0.060 mg (27%) of 16 as a light brown solid [26]. mp = 88 -
90˚C. 1H NMR (400 MHz, DMSO-d6) δ 10.33 (s, 1H, NH), 9.57 (s, 1H, OH), 9.12 (s, 1H, OH),
8.65 (s, 1H, NH), 7.35 (d, J = 8.8 Hz, 2H, Ar), 6.67 (d, J = 8.8 Hz, 2H, Ar), 2.23 (t, J = 7.4 Hz,
2H, CH2), 1.94 (t, J = 7.4 Hz, 2H, CH2), 1.61 – 1.42 (m, 4H, CH2-CH2), 1.34 – 1.21 (m, 4H,
CH2-CH2). 13C NMR (101 MHz, DMSO-d6) δ 170.9 (C=O, quaternary), 169.5 (C=O,
quaternary), 153.5 (C-OH, quaternary), 131.5 (C-NH, quaternary), 121.2 (Ar), 115.4 (Ar),
36.6 (CH2), 32.7 (CH2), 28.8 (CH2-CH2), 25.6 (CH2), 25.5 (CH2). FT-IR (ATR): (vmax/cm-1)
3301 (NH), 2938 (OH), 1654 (C=O), 1608 (C=O, amide).
{4‐[7‐(hydroxycarbamoyl)heptanamido]phenoxy}phosphonic acid 17. Compound 32
(2.15 mmol, 0.600 g) was dissolved in anhydrous acetonitrile (12 mL). A bath of ice and
NaCl cooled the solution to -10˚C; then CCl4 (1.06 mL) was added. The mixture was treated
with DIPEA (4.52 mmol, 0.79 mL) followed by DMAP (0.22 mmol, 0.026 g), and dropwise
addition of dibenzyl phosphite (3.12 mmol, 0.69 mL), keeping the mixture at or below -10˚C
for a further 70 min, at which point TLC indicated the consumption of the starting materials.
The mixture was quenched by the addition of a solution of KH2PO4 (0.5 M, 3.90 mL), allowed
to warm to rt and then extracted with AcOEt three times. The organic phases were washed
successively with water and a saturated solution of NaCl, dried and concentrated. The
residue was purified by silica flash-chromatography, which gave the compound Methyl
7‐[(4‐{[bis(benzyloxy)phosphoryl]oxy}phenyl)carbamoyl]heptanoate as an yellowish oil (93%,
1.083 g). 1H NMR (400 MHz, DMSO-d6) δ 9.92 (s, 1H, NH), 7.57 (d, J = 8.5 Hz, 2H, Ar), 7.37
(s, 10H, O-CH2-Ph), 7.11 (d, J = 8.0 Hz, 2H, Ar), 5.15 (d, J = 8.3 Hz, 4H, O-CH2-Ph), 3.58 (s,
3H, O-CH3), 2.36 – 2.21 (m, 4H, CH2-CH2), 1.64 – 1.47 (m, 4H, CH2-CH2), 1.29 (s, 4H, CH2-
13
CH2). C NMR (101 MHz, DMSO-d6) δ 173.3 (C=O, quaternary), 171.1 (C=O, quaternary),
145.3 (C-O-P, quaternary), 136.5 (C-NH, quaternary), 128.5 (O-CH2-Ph, Ar, quaternary),
127.9 (O-CH2-Ph, Ar, quaternary), 120.2 (Ar), 69.3 (O-CH2-Ph), 51.1 (O-CH3), 36.2 (CH2),
33.2 (CH2), 28.2 (CH2-CH2), 24.9 (CH2), 24.3 (CH2). Compound Methyl
7‐[(4‐{[bis(benzyloxy)phosphoryl]oxy}phenyl)carbamoyl]heptanoate (0.11 mmol; 0.060 g)
was mixed with NH2OH·HCl (0.11 mmol; 0.008 g) in MeOH (0.40 mL), cooled to 0 °C and
the NaOH solution (4 M; 0.88 mmol; 0.22 mL) was then added. The product dibenzyl
4‐[7‐(hydroxycarbamoyl)heptanamido]phenyl phosphate was a greenish oil (35%, 0.021 g).
1H NMR (400 MHz, DMSO-d6) δ 10.33 (s, 1H, NH), 9.92 (s, 1H, OH), 8.65 (d, J = 1.6 Hz, 1H,
NH), 7.57 (d, J = 9.0 Hz, 2H, Ar), 7.45 – 7.29 (m, 10H, O-CH2-Ph), 7.11 (d, J = 8.0 Hz, 2H,
29