Design, discovery, modelling, synthesis, and biological evaluation of novel and small, low toxicity s-triazine derivatives as HIV-1 non-nucleoside reverse transcriptase inhibitors

Article abstract of DOI:10.1016/j.bmc.2016.04.018

A set of top-ranked compounds from a multi-objective in silico screen was experimentally tested for toxicity and the ability to inhibit the activity of HIV-1 reverse transcriptase (RT) in cell-free assay and in cell-based assay using HIV-1 based virus-like particles. Detailed analysis of a commercial sample that indicated specific inhibition of HIV-1 reverse transcription revealed that a minor component that was structurally similar to that of the main compound was responsible for the strongest inhibition. As a result, novel s-triazine derivatives were proposed, modelled, discovered, and synthesised, and their antiviral activity and cellular toxicity were tested. Compounds 18a and 18b were found to be efficient HIV-1 RT inhibitors, with an IC₅₀ of 5.6 ± 1.1 μM and 0.16 ± 0.05 μM in a cell-based assay using infectious HIV-1, respectively. Compound 18b also had no detectable toxicity for different human cell lines. Their binding mode and interactions with the RT suggest that there was strong and adaptable binding in a tight (NNRTI) hydrophobic pocket. In summary, this iterative study produced structural clues and led to a group of non-toxic, novel compounds to inhibit HIV-RT with up to nanomolar potency.

Full text of DOI:10.1016/j.bmc.2016.04.018

Accepted Manuscript
Design, discovery, modelling, synthesis, and biological evaluation of novel and
small, low toxicity s-triazine derivatives as HIV-1 non-nucleoside reverse tran-
scriptase inhibitors
Birgit Viira, Anastasia Selyutina, Alfonso T. García-Sosa, Maarit Karonen, Jari
Sinkkonen, Andres Merits, Uko Maran
PII:
S0968-0896(16)30252-8
DOI:
http://dx.doi.org/10.1016/j.bmc.2016.04.018
BMC 12933
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
28 December 2015
10 March 2016
8 April 2016
Please cite this article as: Viira, B., Selyutina, A., García-Sosa, A.T., Karonen, M., Sinkkonen, J., Merits, A., Maran,
U., Design, discovery, modelling, synthesis, and biological evaluation of novel and small, low toxicity s-triazine
derivatives as HIV-1 non-nucleoside reverse transcriptase inhibitors, Bioorganic & Medicinal Chemistry (2016),
doi: http://dx.doi.org/10.1016/j.bmc.2016.04.018
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.

Design, discovery, modelling, synthesis, and biological evaluation of novel and small, low
toxicity s-triazine derivatives as HIV-1 non-nucleoside reverse transcriptase inhibitors.
Birgit Viiraa, Anastasia Selyutinab, Alfonso T. García-Sosaa, Maarit Karonenc, Jari Sinkkonenc,
Andres Meritsb*, Uko Marana*
a Institute of Chemistry, University of Tartu, Tartu 50411, Estonia
b Institute of Technology, University of Tartu, Tartu 50411, Estonia
c Department of Chemistry, University of Turku, FI-20014 Turku, Finland
Author contributions:
BV, AS and ATGS have contributed equally.
Corresponding author:
* Phone, +3727375254; fax, +3727375264; e-mail, uko.maran@ut.ee (UM)
* Phone, +3727375007; e-mail, andres.merits@ut.ee (AM)
Abstract
A set of top-ranked compounds from a multi-objective in silico screen was experimentally tested
for toxicity and the ability to inhibit the activity of HIV-1 reverse transcriptase (RT) in cell-free
assay and in cell-based assay using HIV-1 based virus-like particles. Detailed analysis of a
commercial sample that indicated specific inhibition of HIV-1 reverse transcription revealed that
a minor component that was structurally similar to that of the main compound was responsible
for the strongest inhibition. As a result, novel s-triazine derivatives were proposed, modelled,
discovered, and synthesised, and their antiviral activity and cellular toxicity were tested.
Compounds 18a and 18b were found to be efficient HIV-1 RT inhibitors, with an IC50 of 5.6 ±
1.1 µM and 0.16 ± 0.05 µM in a cell-based assay using infectious HIV-1, respectively.
Compound 18b also had no detectable toxicity for different human cell lines. Their binding
mode and interactions with the RT suggest that there was strong and adaptable binding in a tight
(NNRTI) hydrophobic pocket. In summary, this iterative study produced structural clues and led
to a group of non-toxic, novel compounds to inhibit HIV-RT with up to nanomolar potency.
Keywords: HIV, non-nucleoside reverse transcriptase inhibitors, synthesis, antiviral activity,
molecular docking
1

Content
(Only for reviewing purposes)
1
2
Introduction.........................................................................................................................3
Results and Discussion ........................................................................................................6
2.1 Biological activity of in silico screened compounds and their samples ..........................7
2.2 Chemical and biological analysis of active sample ........................................................8
2.3 Proposing novel s-triazine derivatives ...........................................................................9
2.4 Analysis of protein-ligand binding interactions of novel s-triazine derivatives.............13
2.5 Cytotoxicity and antiviral activity of novel s-triazine derivatives.................................17
Conclusions.......................................................................................................................20
Materials and Methods ......................................................................................................22
4.1 In silico screened and selected compounds..................................................................22
4.2 Computational methods...............................................................................................22
4.3 Chemical synthesis......................................................................................................23
4.4 Materials and instruments. ..........................................................................................27
4.5 Cell-culture based assays used for cytotoxicity and anti-HIV-1 activity measurements29
Acknowledgments.............................................................................................................34
Supplementary data...........................................................................................................35
References.........................................................................................................................36
3
4
5
6
7
2

1
Introduction
Human immunodeficiency virus type 1 (HIV-1) infection is a serious health threat worldwide: in
2014, a total of 37 million people were living with HIV-1 [1]. The progression of the infection
leads to the development of acquired immunodeficiency syndrome (AIDS), which, if untreated,
eventually results in patients’ deaths, mostly due to opportunistic infections or malignant tumors.
To suppress HIV-1 replication, antiretroviral therapy (ART) is used. Currently, ART is a
combination of three or more drugs that have different mechanisms of antiviral action. One
important group of antiretroviral drugs is Non-Nucleoside Reverse Transcriptase Inhibitors
(NNRTIs). Currently, five NNRTIs have been approved by the Food and Drug Administration
agency: nevirapine, efavirenz, delavirdine, etravirine, and rilpivirine. They have very different
chemical structures but share the same mode of action: their binding to the NNRTI binding
pocket of HIV-1 reverse transcriptase (RT) leads to conformational changes in the catalytic
subunit of the enzyme and inhibition of its DNA polymerase activity [2]. Because HIV-1 has a
high mutation rate during its replication, resistant forms of the virus that are not susceptible to
NNRTI treatment arise very quickly. Therefore, new compounds with higher genetic barriers of
resistance that are nontoxic for humans are required.
Among the first discovered NNRTI-s were HEPT [3] and diarylpyrimidine (DAPYs) [4]
compounds. In silico screening has been notably successful in finding new inhibitors for HIV-1
RT. Barreca et al. discovered compounds with new molecular scaffolds, including sulfonamides
and an oxadiazole that inhibits HIV-1 RT using a structure-based pharmacophore method that
includes docking [5]. Jorgensen and co-workers used a pool of inactive compounds from a
molecular mechanics-General Born/Surface Area (MM-GB/SA) screen and recovered active
compounds by using a substituent scan of core structures placed in the active site [6]. Herschhorn
3

et al. used virtual screening with docking to two protein crystal structures to find thiourea
inhibitors against HIV-1 RT [7]. Nichols et al. reported HIV-1 RT inhibitors using virtual
screening against wild-type (wt) and NNRTI resistant mutant (Y181C) versions of the protein [8].
Among their reported inhibitors, several had structures that included sulphur atoms that were
close to conjugated nitrogens or that had an oxazole or cyanouracil group. Physicochemical
filters and the rigid docking of conformers led to active compounds against the enzyme [9].
Distinto et al. [10] described active sulfonohydrazone and hydrazone HIV-1 RT inhibitors using
a method that was based on shape-based filters, 2D-fingerprints, and 3D-pharmacophore virtual
screening that inhibits both DNA polymerase and ribonuclease H (RNase H) activities of HIV-1
RT. Another study showed that docking to X-ray crystal and MD-simulation structures led to
two inhibitors [11]. Recent sur-flex docking screening lead to the discovery of new
dihydroalkoxybenzyloxopyrimidines (DABOs) with modification in the C6 position of the
4(3H)-pyrimidinone scaffold [12]. Our own multi-objective in silico screening of chemical
compound libraries against several protein X-ray crystal structures of wt and NNRTI resistant
Y181C HIV-1 RT has resulted in a ranked list of potentially interesting compounds [13]. In that
study, protein structures included several conformations of the critical tyrosine 181 residue as
well as its substitution for cysteine in addition to the presence or absence of an explicit water
molecule in the NNRTI binding pocket. A battery of anti-target proteins was also included in the
workflow to profile inhibitor candidates and find compounds that have a moderate-to-low
toxicity. The anti-targets included important metabolism proteins, such as the pregnane X
receptor, sulfotransferase 1A3, cytochrome P450 2a6, cytochrome P450 2c9, and cytochrome
P450 3a4. The resultant compounds had interesting structures, chemistry, and proposed binding
modes. The compounds among the top scorers were also profiled and ranked according to the
4

number of hydrogen bonds they made to backbone atoms of the protein in order to reduce the
loss of activity due to a mutation. In addition, compounds were profiled according to their ligand
efficiencies [14, 15, 16, 17, 18, 19, 20], which allowed selecting those that were best suited for
further development and promising from a physicochemical properties point of view.
The present study experimentally tests the result of the above-described multi-objective
in silico screening [13] and includes the discovery and synthesis of several new chemical entities
from a family of s-triazines that were not reported before. The toxicity and antiretroviral activity
in the cell-free and cell culture assays of the top-ranked compounds and their derivatives were
measured experimentally. The described research shows how the iteration of in silico screening
and analysis in exploring the NNRTI binding site with chemical compounds can approach
systematically closer to the desired result of small, novel, low toxicity NNRT and whole-cell
viral HIV inhibitors.
5

2
Results and Discussion
In total, twenty compounds were experimentally tested. First, the 16 commercially available
samples (Figure 1) were tested for toxicity and antiviral activity. Then, the detected active
sample was studied further, and new compounds were proposed, synthesised, and tested for
toxicity and antiviral activity. All of the new compounds were supported by computational
verification.
Figure 1. Structures of compounds selected from in silico screening for toxicity and activity
testing.
6

Table 1. Commercial compounds 1-16, their toxicities and abilities to inhibit HIV-1 activities in
different assays.
Compound
MW
(g/mol)
Toxicity assay,
CC50, µM
Cell culture assay,
IC50, µM
Cell free assay,
IC50, µM
373.4
387.4
405.4
341.8
376.3
378.4
323.4
390.5
375.4
389.5
406.5
347.4
323.4
319.3
319.3
320.4
266.9
> 5
> 50
> 50
> 50
> 50
> 50
> 50
> 50
> 50
> 50
> 50
> 50
> 50
> 5
> 5
> 50
> 50
> 50
> 50
> 200
> 200
> 200
> 200
> 200
> 200
> 200
105.7 ± 30
> 200
> 200
> 200
> 50
> 200
> 200
> 200
> 200
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
> 50
> 50
0.33 ± 0.07
> 50
> 50
> 5
> 50
> 50
> 5
> 5
> 50
0.42±0.07
> 5
> 50
not toxic
16
0.92 ± 0.15
nevirapine
2.1
Biological activity of in silico screened compounds and their samples
We determined the toxicities of 16 commercial samples of compounds at a 50-µM concentration
and at a 5-µM concentration for samples that were toxic at 50 µM. Subsequently in cell-culture
experiments, the samples were studied only at their corresponding non-toxic or moderate-toxic
concentrations (50 µM and/or 5 µM) (Table 1 and Table S1, in supplementary material). The
first screen for antiretroviral activities was performed using a virus-like particle (VLP)-based
assay because of its increased sensitivity and safety. In addition, the direct effect on the reverse
transcription reaction was assessed in a cell-free assay; 200 µM and 50 µM concentrations were
used (Table 1 and Table S1, in supplementary material). The highest antiviral activity was
observed for 8 (Figure 1), and it was the only active sample in both tests (Table 1). Subsequent
experiments with different concentrations of 8 revealed characteristic sigmoidal inhibition curves
(Figure S1, in supplementary material). The inhibitory concentration 50 (IC50) of 8 in the VLP-
based assay was 0.33 ± 0.07 µM, which was only 10 times higher than the IC50 for the potent
known NNRTI nevirapine, which was determined in the same assay. The IC50 of 8 in a cell-free
7

assay was 105.7 ± 30 µM (Table 2). A much higher IC50 in the cell-free reaction, compared to
the cell culture assay, has also been shown for NNRTIs elsewhere [21, 22, 23, 24]. The results
from these two assays allow to conclude that sample 8 contains an NNRTI HIV-1 inhibitor.
Table 2. Anti HIV-1 activities of compound 8 and its derivatives.a
IC50, µM
CC50, µM
Cell free assay
HIV-1
VLP (wt)
TZM-bl (HIV-1) U2OS (VLP)
8(n.p.)
8
8(i1)
105.7 ± 30
NA
NA
ND
NA
NA
0.33 ± 0.07
16.0 ± 3.5
18.2 ± 5.2
>50
>50
>50
>50
>50
>50
8(i2)
NA
ND
NA
ND
32 ± 12
NA
>50
>50
>50
>50
17
116 ± 12
55.1 ± 7.5
5.6 ± 1.1
0.16 ± 0.05
2.5 ± 0.3
0.067 ± 0.009
57 ± 22
>12.5
42.2 ± 5.6
>12.5
18a
18b
ND
ND
NA
>50
>50
19
nevirapine
0.92 ± 0.15
0.026 ± 0.005
0.043 ± 0.007
>100
>100
a IC50: the concentration of the compound that is required to inhibit the signal generated by HIV-
1 or HIV-1 VLP infection or the activity of purified reverse transcriptase of HIV-1 by 50%;
CC50: the concentration of the compound that reduces the cell viability (cell index) by 50%; NA
– no activity, ND – not determined.
2.2
Chemical and biological analysis of active sample
Compound 8 has not been described previously in HIV-1 research in any publication or patent. It
contains a diaryltriazine core substituted by three aminoalkyl groups, with two N-p-tolyl groups
and one N-methyltetrahydrofuranyl group.
The purity of the commercial sample 8 (n.p. – non-purified) was verified by HPLC.
Three additional distinctive components (Figure S2), denoted 17, 18, and 19, were detected,
which together represented approximately 11% of the mass of the sample. These components
were purified to homogeneity with HPLC and collected. Because 17 and 18 failed to be
separated well under the used conditions/equipment, a mixture of 17/18 was collected. Because
the structure of 8 possessed a chiral centre, its two isomers were also separated with a chiral
HPLC column.
8

To verify the antiretroviral activity of compound 8, all of the obtained components of the
sample (purified compound 8(p.), 19, and mixture 17/18) and the isomers of 8 were tested for
their cytotoxicity and antiviral activity. Testing revealed that purified 8 and its separated isomers
as well as 19 did not show prominent antiretroviral activity, whereas the mixture of 17/18 was
found to be the most active (Figure S3). The mixture of 17/18 also lacked any cytotoxic effect,
even at the highest concentration (Figure S4). These results allow to conclude that instead of the
original belief about compound 8, components 17 and/or 18 must represent a potent HIV-1 RT
inhibitor with a good therapeutic index. This finding also provided the grounds to speculate that
the components discovered in the sample could be “side” products from synthesis or
intermediates and have structures similar to that of compound 8. Experimental and analytical
testing and verification of the original components 17/18 and 19 were complicated due to the low
amount (approximately 0.2 mg) in the commercial sample. Therefore, to identify the structures
of 17, 18, and 19, the synthesis methods for compound 8 and its analogs were investigated.
2.3
Proposing novel s-triazine derivatives
Considering the symmetric nature of the compound 8 core and the molar masses from the
mass spectra of the sample components 17/18 and 19 and the synthesis pathways, possible
structures (Table 3: 17, 18a, 18b, and 19) were proposed. The structures of these new
compounds had the same core (1,3,5-triazine) and differed only in the combination of functional
groups in substitution positions on the core ring of compound 8. The molar mass for component
18 was 307.4 g/mol, but considering the possible substitutions, it could match two structures that
+
differ in one substituent (18a: –NH3 or 18b: –OH). This conclusion led to the use of high-
resolution LC-ESI-QTOF-MS to separate the original mixture 17/18 and determine the structure
for 18. The spectrum of 18 exhibited an ion at m/z 308.1530 that corresponded to [M+H]+
9

(C17H18N5O, calculated 308.1506, error 7.8 ppm), and the molecular formula for the structure 18
was found to be C17H17N5O. The NMR spectrum for the mixture 17/18 was also measured.
Comparing the 1H NMR spectrum with compound 17 showed that the structure of 18 must be
rather similar to 17, but it appeared to be lacking the furan-containing aminoalkyl group, while
the tolyl group was easily identified. Based on this arrangement, it was concluded that
component 18 must have a structure that contained two amino-tolyl and one hydroxyl group, i.e.,
compound 18b.
Table 3. Structures for compounds related to 8, their molecular mass and binding score.
Structure
MW
(g/mol)
384.5
Score
(kcal/mol)
–12.42
17
307.4
307.4
396.5
–14.22
–14.36
–12.8
18a
18b
19
The proposed structures (Table 3: 17, 18a, 18b, and 19) of the sample components were
not available from any commercial provider. Consequently, methods for the synthesis of a family
of chemical structures with the scaffold (1,3,5-triazines or s-triazines) were adopted from the
literature and modified. Even though compound 8 and component 19 did not show prominent
10

antiretroviral activity (see section 2.2), they were synthesised and tested to verify previous
findings with known structure in the case of 19.
The 1,3,5-triazine derivatives have received substantial attention in the literature because
of their potent biological activity as anticancer drugs [25], estrogen receptor modulators [26],
antibacterials [27, 28], antimicrobials [28, 29, 30, 31, 32, 33, 34] and also tumor growth
inhibition activities [35]. The 1,3,5-triazine derivatives also have several other applications as
chiral stationary phases, for the preparation of luminescent complexes and as metal complexes,
liquid crystals, calixarenes, dendrimers, and polymers, and as optical brighteners for household
washing powders [36]. Considering the above, effort has been devoted for the synthesis of s-
triazine derivatives in recent years [36, 37, 38, 39, 40].
After careful consideration of the possible methods for the synthesis of substituted 1,3,5-
triazines, the functionalization of the less expensive reagent cyanuric chloride was selected as a
promising approach. It has been previously described that the nucleophilic substitution of each
chloride is successfully controlled by taking advantage of the decrease in reactivity with the
increasing number of substituents [36]. The 2,4,6-substituted 1,3,5-triazines were obtained from
cyanuric chloride by sequential substitution of the chloride atom using nitrogen and oxygen
centered nucleophiles (Scheme 1). Sequential nucleophilic substitution of the C-Cl bond of
1,3,5-triazine by the C-O and C-N bonds has been described elsewhere [36]. In this work,
different solvents and substituents were used to achieve the desired products.
11

Scheme 1. General synthetic pathways used for the synthesis of 2,4,6-trisubstituted-1,3,5-
triazines. i) cyanyric chloride, p-toluene (2 eq), diisopropylamine (2.85 eq), DCM, 0 °C to rt; ii)
cyanuric chloride, tetrahydrofurane-4-amine (2 eq), diisopropylamine (2.85 eq), DCM, 0 °C to rt;
iii) cyanyric choride, NH4OH (1 eq), DCM, rt; iv) 2,4-dichloro-1,3,5-triazine, tetrahydrofurane-
4-amine (2.5 eq), 1,4-dioxane, reflux; v) 2-chloro-1,3,5-triazine, p-toluene (1.5 eq), 1,4-dioxane,
reflux; vi) 2-chloro-1,3,5-triazine, tetrahydrofurane-4-amine (1.5 eq), 1,4-dioxane, reflux; and
vii) 2-chloro-1,3,5-triazine, H2O (0.5 eq), dest.H2O, TFA, reflux.
This synthesis scheme allowed obtaining large (in gram scale) quantities of 8, 17, 18a,
18b, and 19 (Table 3). Mono – and di-substituted derivatives of 8 were prepared by the reaction
of cyanuric chloride with the corresponding aminoalkyl group in DCM using di-isopropyl amine
as a base. Di-substituted derivatives were achievable by performing the reaction at ambient
temperature (methods i and ii, Scheme 1). To obtain the monosubstituted compound, the reaction
had to be carried out at 0°C or below. In case of substitution of the first chlorine with an amino
group, the solution was kept at room temperature, and the reaction finished rapidly (method iii,
Scheme 1). The substitution of the remaining chloride atom(s) usually required considerably
more vigorous reaction conditions (such as higher temperatures and longer reaction time)
(methods iv-vi, Scheme 1).
12

2.4
Analysis of protein-ligand binding interactions of novel s-triazine derivatives.
The docking of compounds 17, 18a, 18b and 19 into the NNRTI binding site of protein
structure 2be2 [41], similar to that conducted during virtual screening [13], resulted in a ranking
(18b > 18a > 19 > 17), which is presented in Table 3. The triazine chemical motif that is shared
between the compounds was indeed present in some of the compounds that were identified in the
previous in silico screen.
Figure 2. Docked binding mode of compound 18a with the HIV-1 RT crystal structure.
Hydrogen atoms are in white, oxygen atoms are in red, and nitrogen atoms are in blue. Ligand
carbon atoms are in cyan, and protein carbon atoms are in pink. All of the hydrogen atoms are
shown for the ligand, and the polar hydrogens are shown for the protein. Compound 18a donates
a hydrogen bond and accepts a hydrogen bond (yellow dashes) from Lys101, as well as forming
an ionic interaction with the sidechain of Glu138. Hydrophobic contacts are provided by the
sidechains of Phe 227, Tyr181, Tyr 188, Val 179, and Trp229.
Further detailed computational analysis of the compound 18a binding mode to the protein
(see Figure 2) shows that 18a binds deep into the NNRTI pocket, with its p-tolyl groups
engaging in - interactions with residues Tyr 181, Tyr 188, Trp 229, and Phe 227. It also
possesses hydrophobic contacts with residues Leu 234 and Val 179, the former interacting with
the central ligand heterocycle. It has hydrogen bonds with the backbone C=O and NH groups of
13

Lys 101 as well as a strong ionic interaction between its charged ammonium group with the
charged carboxylic group of Glu 138, at the mouth of the binding site.
Compound 18a has a low molecular weight of 307.4 g/mol, which enables further
modification of the ligand, and it has a calculated logP of 4.11, three hydrogen bond donors,
three hydrogen bond acceptors, four rotatable bonds, and three rings. Other diaryltriazines have
been reported to have promising anti-HIV activity [42], but the main advantage that compound
18a has over other compounds with its butterfly structure is that its charged ammonium group is
ideally placed to interact strongly with the Glu 138 carboxylic group, and its other ring
substitutions, two p-toluene groups, are perfectly symmetric. This combination of features is
absent from the other compounds.
The binding mode resembles that of the known and approved NNRTI drugs, and it
especially resembles those of a different class of compounds, the binding mode of
diarylpyrimidine [43] and the approved drugs etravirine [44] and rilpivirine [45]. Moreover, it
has been proposed that this particular binding mode has advantages for allowing a strong fit in
the tight, closed hydrophobic binding pocket, but that at the same time, the flexibility in the
“wings” of compound 18a could provide it with a degree of adaptability to conformational and
mutational change in the protein [42]. The molecular symmetry of 18a also allows it to bind
tightly in a 180° flip, which would increase the number of favorable conformers and binding
modes of the protein.
The computational analysis shows that compared to 18a, compound 18b has an even
stronger interaction with the protein (Table 3). Compound 18b represents a difference of one
+
isosteric substituent with compound 18a, the replacement of –NH3 by -OH, and they possess the
same molar mass and similar polarity. Similarly to 18a, 18b also has interactions with protein
14

such as donating and receiving hydrogen bonds to the backbone NH and CO of Lys 101 using an
exocyclic and an endocyclic nitrogen, with 18b using in addition a hydroxyl group, as shown in
Figure 3. This change loses the ionic interaction of 18a, but it gains a slightly deeper binding
mode inside the protein binding-site. Such symmetrical compounds have the advantage of being
capable of binding to the protein in the same way after a 180° flip.
Figure 3. Docked binding mode of compound 18b with the HIV-1 RT crystal structure.
Hydrogen atoms are in white, oxygen atoms are in red, and nitrogen atoms are in blue. Ligand
carbon atoms are in grey, and protein carbon atoms are in pink. All of the hydrogen atoms are
shown for the ligand, and the polar hydrogens are shown for the protein. Compound 18b donates
and accepts three conjugated (close) hydrogen bonds (yellow dashes) from Lys101 backbone
atoms, as well as forming deep and enclosed hydrophobic contacts with the sidechains of Phe
227, Tyr181, Tyr 188, Val 179, Trp229, and Leu 100.
Compound 18b has the same low molecular weight of 307.4 g/mol as 18a, and three
hydrogen bond donors, three hydrogen bond acceptors, four rotatable bonds, three rings, and a
moderate, calculated logP of 5.16. Compound 18b, on the other hand, has a slightly deeper
binding than compound 18a; its hydroxyl group can make stronger hydrogen bonds than the
amine group in 18a and has a large number of resonance structures that can stabilize tautomers
and protonation states, which could provide its stronger activity. These equivalent states could
enhance the binding to the protein. The nitrogen groups on the heterocycle as well as on the side
15

chains, together with the less polar toluene functional groups, allow a balance in the compound’s
polarity and solubility properties. Compound 18b has not been reported in HIV-1 or in any other
biological or material research.
The loss of the salt bridge inside the protein can incur a loss of around 4 kcal/mol in the
free energy of binding, yet the charged amino group can also have a high desolvation energy, as
well as difficulties in membrane transport. The better activity of 18b compared to 18a can thus
be related to a tighter (deeper) fit in the binding site, the availability of a higher number of
resonance structures for 18b, or stronger hydrogen bonding of OH groups relative to amino
groups, as has been reported elsewhere [46]. In addition, the side chain of Glu138 is the
hydrogen bond partner of 18a, whereas 18b makes all three hydrogen bonds to Lys101 backbone
atoms. Since side chains are more dynamic (especially so the flexible side chain of glutamate), it
would be expected that the hydrogen bonds to the backbone atoms of Lys101 would be more
conserved and easier to maintain than a hydrogen bond to the mobile side chain of Glu138. The
actual score breakdowns show a higher Rewards term for 18b of -38.5 kcal/mol compared to -
35.2 kcal/mol for 18a. The Rewards term in Glide accounts for rewards and penalties to the
scoring function for various features, such as buried polar groups, a hydrophobic enclosure,
correlated hydrogen bonds, amide twists, among others [47]. The deeper, tighter binding of 18b
compared to 18a would favor a better burial of polar groups, and better enclosure of the
hydrophobic groups from the bulk solvent, and possibly stronger correlated hydrogen bonds (that
is, hydrogen bonds at a small distance separation – one or two bonds - from each other).
16

2.5
Cytotoxicity and antiviral activity of novel s-triazine derivatives
Toxicity and antiretroviral activity of the synthesised compounds 8 (isomers si1, si2), 17,
18a, 18b, and 19 were analysed. The obtained results were consistent with previously obtained
experimental data for sample components (see chapter 2.2) and with computational predictions
(see chapter 2.4). It was confirmed that synthesised 8 and its stereoisomers were weak inhibitors
of HIV-1 infection. They showed weak activity in experiments with HIV-1-based VLPs and no
activity against purified HIV-1 RT in cell-free assay or against infectious HIV-1 in cell culture
assay (Table 2, see also: Figure S3: 8(p.) and Figure S5). Consistent with the results obtained
for component 19 (Figure S3), the synthesised compound 19 also lacked anti-HIV-1 activity
(Table 2). The same was found to be the case for synthesised compound 17 (Table 2 and data
not shown). This data indicates and confirms that 17 and 19 were not responsible for the anti-
HIV-1 activity of the original sample of 8.
In accordance with in silico predictions (see section 2.4), synthesised 18a was found to be
an efficient HIV-1 RT inhibitor. In the assay based on the use of HIV-1 VLPs, the IC50 of 18a
was found to be 2.5 ± 0.3 µM (Table 2; Figure S6: A). The cytotoxic concentration (CC50, the
concentration of the compound that reduces the cell viability by 50%) for U2OS cells was 42.2 ±
5.6 µM, and the SI (selectivity index, the ratio CC50/IC50) was approximately 17. Importantly,
comparable values were obtained in the assay using infectious HIV-1 (IC50 = 5.6 ± 1.1 µM, CC50
= 57 ± 22 µM (when measured for TZM-bl cells), and SI was approximately 10 (Table 2;
Figure S6: B). Expectedly, 18a also showed a dose-dependent inhibition of purified HIV RT in
the cell free assay (IC50 = 116 ± 12 µM) (Table 2, Figure S6: C). Finally, the anti-HIV-1 effect
of 18a was clearly specific to the wt HIV-1 RT because the compound was not active against
17

HIV-1 VLPs that contained RT with mutations that are associated with resistance against
NNRTIs (Figure S7).
Figure 4. Compound 18b reduces the infectivity of HIV-1 VLPs (A) and HIV-1 virions (B) in
cell-based assays and the activity of HIV-1 RT in the cell-free test-tube reaction (C). Compound
18b also reduces the infectivity of HIV-1 VLPs that contain RT that harbors the K103N and
Y181C mutations (D). Each experiment was performed in triplicate; DMSO was used as a
vehicle control, and the corresponding values were set to 100%. The error bars represent the
standard deviation. Data from one reproducible experiment are shown.
Compound 18b revealed no cytotoxicity at concentrations up to 12.5 µM (data not
shown); higher concentrations could not be assayed due to the limited solubility of 18b. At the
same time, 18b was the most potent HIV-1 inhibitor among all of the tested compounds: in the
assay based on the use of HIV-1 VLPs, the IC50 of 18b was 0.067 ± 0.009 µM, which is
approximately 5-fold lower than that of the original 8 (n.p.) from stock (Figure 4: A, Table 2).
18

Importantly, similar results (Figure 4: B, Table 2) were obtained in the assay in which infectious
HIV-1 was used instead of VLPs (IC50 = 0.16 ± 0.05 µM). It was also shown (Figure 4: C,
Table 2) that this compound is active against the purified RT of HIV (IC50 = 55.1 ± 7.5 µM).
The effect of 18b was specific for wt HIV-1 RT because the compound had no effect on the
HIV-1 VLPs that contained RT with a single K103N or single Y181C mutation (data not shown).
However, the compound demonstrated moderate activity (Figure 4: D) against HIV-1 VLPs that
contained RT with both indicated mutations, although in this case, the calculated IC50 (8.0 ± 4.0
µM) was more than 100-fold higher than in the case in which the VLPs contained wt HIV-1 RT.
Thus, although compound 18a is less efficient than compound 18b, they represent novel
HIV-1 NNRTI-s. Both compounds have not been reported in HIV or any other biological or
material research before. The discovery of 18b and 18a allows us to conclude that the in silico
screen was successful because their chemical motif is also shared with 8 and for both compounds
+
low toxicity was found. The isosteric replacement of –NH3 by –OH was found to be beneficial
to HIV inhibitory activity, which can be rationalised based on its molecular properties and
interactions. The relative ranking of the compounds as calculated computationally reflects well
the experimentally observed trend.
Other 1,3,5-triazines with different substituents have been reported to bind to the ATP
binding site of DNA topoisomerase II, a target protein in cancer [48, 49], and therefore, further
tests can be conducted on these compounds to assess and develop them as regards to their
inhibition profile and safety, though they appear to have low toxicity. Compound 18b is non-
toxic at up to 12.5 M, which is its estimated solubility in the current experimental setting,
though 18b is active in the nanomolar range in cells which provides a window of safety for the
possible use of this compound.
19

3
Conclusions
On the basis of the earlier multi-objective in silico screening and the current work, novel s-
triazine derivatives are shown to be HIV-1 NNRTIs. For this, the biological properties of the 16
top-ranked compounds (their commercial samples) were investigated. The commercial non-
purified sample 8 was shown to be a potent HIV-1 inhibitor, although it was discovered that
closely related compounds that share the s-triazine and some of the functional groups were
present in the initial commercial preparation and had a strong anti-HIV-1 effect at low
concentrations. Iterative and systematic work was performed in the virological, synthetical,
analytical, and computational analysis of the data. Altogether, five compounds were synthesised
and analysed for their antiviral activity: 8 (its 2 stereoisomers), 17, 18a, 18b, and 19. The latter
four were not available commercially and showed different biological activity. All of them
possess a s-triazine core and similar functional group substituents. The difference between the
+
strongest inhibitors, 18a and 18b, was one conserved functional group: –NH3 vs. –OH,
respectively. The most potent, 18b, exists in several tautomeric forms that can be stabilised
through resonance. The compounds proved to be non-toxic and effective in both cell-based and
enzyme experiments; the in silico methods and systematic follow-up were able to identify a
novel group of non-toxic compounds and a new chemotype (1,3,5-triazinone/s-triazinone/1,3,5-
triazinol/s-triazinol) that has the ability to bind to HIV-RT in the NNRTI site and inhibit HIV
infectivity. Their binding mode and interactions with the protein suggest strong binding in a tight
(NNRTI) hydrophobic pocket, using hydrogen bonds, hydrophobic contacts and - interactions.
Their molecular symmetry affords them with at least two potent binding poses that are equally
strong, and their wing-type structure provides adaptability to protein conformational movement.
Resonance structures can stabilise the binding modes to the protein. Their small molecular size,
20

potent ligand efficiencies, and measured low toxicities permit further exploration of their
derivatives and the NNRTI binding site. It is also interesting to emphasise that compounds with
different degrees of activity can be among the components of commercial samples and that
computations and a systematic approach can provide valuable clues toward them.
21

4
Materials and Methods
4.1
In silico screened and selected compounds
Samples for 16 compounds of the top thirty compounds from the multi-objective screening and
ranking described previously [13] were available commercially. They formed a library for the
experimental testing. The compounds were diverse (structures shown in Figure 1) as measured
by Tanimoto similarity coefficients (mean = 0.24, median = 0.21) that compare the amount of
structural commonalities between two compounds. They also had good ligand efficiencies
(G/PSA deeper than -0.07 kcal/(mol•Å2) [13], G/MW deeper than -0.02 (kcal•g)/mol2 [13],
and G/NHA deeper than -0.24 kcal/(mol•NHA) [50]). The ligand efficiencies normalise the
binding energy of a compound by a factor that can be related to the size or other
physicochemical features of the compound. This initial set included quinolin-2-ones, chromen-2-
ones, purines, and related rings that are present among the top-ranked compounds. Nitrogen
compounds also featured prominently. Compounds 14 and 15 were different protonation states in
the in silico screen, although they were tested experimentally with the same compound.
4.2
Computational methods
Chemical compounds were drawn and energy minimised with MacroModel version 9.9 [51].
Protein X-ray crystal structures were assigned polar hydrogen atoms and prepared (including
energy minimisation) using the Protein Preparation Wizard version 1.0 [52]. Docking was
conducted using Glide version 5.9 [47] with rigid protein and fully flexible ligand structures,
scoring with XP GlideScore. LogP was calculated with the program xlogP3.0 [53].
22

4.3
Chemical synthesis
4.3.1 Chemicals.
Samples of 1, 7, 11, 13, and 14 were purchased from Vitas-M Laboratory, Ltd.; samples of 2, 4,
5, 6, and 16 from Enamine Ltd.; samples of 3 and 12 from InterBioScreen Ltd.; a sample of 9
from Asinex Ltd.; and a sample of 10 from Otava Ltd. Samples of 8 were purchased from both
InterBioScreen Ltd. and ChemBridge Corporation, and these samples were equivalent in
composition and behaviour.
Other solvents and reagents were obtained commercially and used without further
purification. The 2,4,6-trisubstitued-pyrimidine derivatives (8, 17, 18a, 18b and 19) were
synthesised from 2,4,6-trichloro-1,3,5-triazine, by sequential substitution of the chloride atom
using oxygen and nitrogen centered nucleophiles. In general, the first chlorine was displaced at a
temperature of 0 °C or below, and the second chlorine was displaced at room temperature
(between 22 °C and 26 °C); the third chlorine was displaced with a temperature between 80 °C
and 90 °C. All of the substituted derivatives were purified by column chromatography using
different eluents (specified in the methods section). The reaction sequence is given in Scheme 1.
All of the synthesised compounds were fully characterised by spectroscopic data obtained by
HPLC-DAD, MS (high-resolution ESI-QTOF and ESI-FT-ICR-MS) and NMR.
4.3.2 2-N,4-N-bis(4-methylphenyl)-6-N-(oxolan-2-ylmethyl)-1,3,5-triazine-2,4,6-triamine (8)
A solution of cyanuric chloride (5.05 g, 27.33 mmol) in anhydrous dichloromethane (50 mL)
was added via cannula (dropwise, 6 min.) to a stirred solution of p-toluidine (2 eq) and
diisopropylamine (2.85 eq) in anhydrous dichloromethane (60 mL) at 0 °C under argon. Then,
the reaction mixture was allowed to reach room temperature. After 14 h, the mixture was
partitioned between CH2Cl2 (40 mL) and the saturated aqueous sodium chloride (60 mL), and the
23

combined organic layers were dried with MgSO4, filtered, evaporated to dryness and purified by
flash chromatography (dichloromethane/diethyl ether 4:1) to afford 8(I) a 97 % yield.
Compound 8(I) (3.72 g, 11.42 mmol) and tetrahydrofurane-4-amine (1.5 eq) were
dissolved in dioxane (20 mL) under argon, and the reaction mixture was refluxed for 21 h. The
solvent was evaporated in vacou, and the residue was purified with flash chromatography
(dichloromethane/ethyl acetate 1:1) to afford compound 8 a 93 % yield. The parameters that
were relevant are the following: 1H NMR (CDCl3, +50°C, 700 MHz) δ: 7.44 (AA’ part of AA’
BB’ multiplet, 4H, 4×CHar); 7.10 (BB’ part of AA’BB’ multiplet, 4H, 4×CHar); 6.95 (br. s, 2H,
2×PhNH); 5.76 (br. s, 1H, CH2NH); 4.08 (dddd, J=6.9, 6.5, 6.3, 3.9 Hz, 1H, CH-O); 3.90 (ddd,
J=8.4, 7.2, 6.2 Hz, 1H, CH2a-O); 3.77 (ddd, J=8.4, 7.4, 6.3 Hz, 1H, CH2b-O); 3.67 (ddd, J=13.8,
6.3, 3.9 Hz, 1H, CH2a-N); 3.47 (ddd, J=13.8, 6.5, 5.7 Hz, 1H, CH2b-N); 2.33 (m, 6H, 2×CH3);
1.99 (ddddd, J=12.2, 8.5, 6.9, 5.3, 0.4 Hz, 1H CH2a-C); 1.91 (m, 2H, 2×CH2b-C); 1.65 (ddddd,
J= 12.2, 8.8, 7.4, 7.2, 0.5 Hz, 1H, CH2a-C); 13C NMR (CDCl3, +50 °C, 176 MHz) δ: 166.3;
164.4; 136.4; 132.6; 129.2; 120.8; 78.0; 68.1; 44.6; 28.7; 25.9; 20.7; HRMS (ESI)
+
m/z 391.22393 [M+H]+ (C22H27ON6 requires 391.22409).
4.3.3 2-N-(4-methylphenyl)-4-N,6-N-bis(oxolan-2-ylmethyl)-1,3,5-triazine-2,4,6-triamine (17)
A solution of cyanuric chloride (2.14 g, 10.93 mmol) in anhydrous dichloromethane (20 mL)
was added via cannula (6 min) to a stirred solution of tetrahydrofurane-4-amine (2 eq) and
diisopropylamine (2.85 eq) in anhydrous dichloromethane (40 mL) at 0 °C under argon, and the
reaction mixture was allowed to warm to room temperature. After 14 h, the mixture was
partitioned between CH2Cl2 (30 mL) and the saturated aqueous sodium chloride (60 mL). The
combined organic layers were dried with MgSO4, filtered, evaporated to dryness and purified by
flash chromatography (ethyl acetate/hexane 9:1) to afford compound 17(I) a 61 % yield.
24

A mixture of compound 17(I) (2.09 g, 6.68 mmol) and p-toluene (1.5 eq) in dioxane (26
mL) under argon was refluxed for 21 h. The solvent was evaporated in vacuo, and the residue
was purified with flash chromatography (acetonitrile/methanol 1:0.1) to afford compound 17 an
89 % yield. The parameters that were relevant are the following: 1H NMR (CDCl3, +50 °C, 700
MHz) δ: 7.44 (AA’ part of AA’ BB’ multiplet4, 4H, 4×CHar); 7.08 (BB’ part of AA’BB’
multiplet, 4H, 4×CHar); 6.9 (br. s, 1H, PhNH); 5.47 (br. s, 2H, 2×CH2NH); 4.05 (dddd, J=7.0,
6.8, 6.7, 4.1 Hz, 2H, CH-O); 3.89 (ddd, J=8.4, 7.2, 6.1 Hz, 2H, CH2a-O); 3.75 (ddd, J=8.4, 7.6,
6.3 HZ, 2H, CH2b-O); 3.61 (m, 2H, CH2a-N); 3.44 (m, 2H, CH2b-N); 2.31 (m, 6H, 2×CH3); 1.93
(m, 4H, 2×CH3, 2×CH2a-C); 1.63 (m, 4H, 2×CH2b-C, 2×CH2a-C); 13C NMR (CDCl3, +50 °C, 176
MHz) δ: 166.0; 164.1; 136.7; 132.2; 129.1; 120.3; 77.9; 77.2; 77.0; 76.8; 68.1; 44.5; 28.7; 25.9;
+
20.7; HRMS (ESI) m/z 385.23414 [M+H]+ (C20H29O2N6 requires 385.23465).
4.3.4 2-N-hydroxy-4-N,6-N-bis(4-methylphenyl)-1,3,5-triazine-2,4,6-triamine (18a)
To a stirred solution of cyanuric chloride (2.16 g, 11.69 mmol) in anhydrous dichloromethane
(10 mL) ammonium hydroxide (1 eq) was added dropwise. The residue was purified with flash
chromatography (ethyl acetate/dichloromethane 8:2) to afford compound 18(I) a 99% yield.
P-toluene (2.5 eq) was added under argon to a stirred solution of compound 18(I) (4 g,
24.25 mmol) in dioxane (80 mL). The reaction mixture was stirred for 20 h. The residue was
purified with flash chromatography (dichloromethane/methanol 1:0.1) to afford compound 18a a
71 % yield. The parameters that were relevant are the following: 1H NMR (DMSO-d6, +50 °C,
700 MHz) δ: 9.72 (br. s, 2H, 2×NH); 7.54 (AA’ part of AA’ BB’ multiplet, 4H, 4×CHar); 7.45
(br. s, 2H, NH2); 7.11 (BB’ part of AA’BB’ multiplet, 4H, 4× CHar); 2.28 (s, 6H, 2×CH3); 13C
NMR (DMSO-d6, +20 °C, 176 MHz) δ: 160.9; 135.9; 132.5; 129.0; 121.3; 40.0; 39.9; 39.7; 39.6;
25

+
39.5; 39.4; 39.3; 39.2; 20.5; HRMS (ESI) m/z 307.16661 [M+H]+ (C17H19N6 requires
307.16657).
4.3.5 2-N,4-N,6-N-tris(4-methylphenyl)-1,3,5-triazine-2,4,6-triamine (19)
A stirred solution of compound 8(I) (2.01 g, 10.93 mmol) and p-toluidine (1.5 eq) in dioxane (25
mL) under argon was refluxed for 21 h. The solvent was evaporated in vacuo, and the residue
was purified with flash chromatography (dichloromethane/n-hexane/methanol 5:5:01) to afford
compound 19 a 98% yield. The parameters that were relevant are the following: 1H NMR
(DMSO-d6, +50 °C, 700 MHz) δ: 7.35 (AA’ part of AA’ BB’ multiplet, 4H, 4×CHar); 7.03 (BB’
part of AA’BB’ multiplet, 4H, 4× CHar); 6.86 (br. s, 3H, 3×NH); 2.25 (m, 9H, 3×CH3); 13C NMR
(CDCl3, +50 °C, 176 MHz) δ: 164.6; 136.2; 133.2; 129.5; 121.2; 77.3; 77.2; 76.9; 20.9; HRMS
+
(ESI) m/z 397.21353 [M+H]+ (C24H25N6 requires 397.21352).
4.3.6 Bis[(4-methylphenyl)amino]-1,3,5-triazine-2-ol a.k.a. 4,6-bis[(4-methylphenyl)amino]-
1H-1,3,5-triazin-2-one a.k.a. 4,6-bis[(4-methylphenyl)amino]-5H-1,3,5-triazin-2-one (18b)
The mixture of distillated H2O (42 mL) and TFA (0.5 eq) was added to compound 8(I) (350 mg,
1,07 mmol) in the reaction vessel. The reaction mixture was refluxed for 120 h, evaporated to
dryness and purified by recrystallisation from ethanol to afford compound 18b with 60 % yield.
The following data was measured for the synthesised comp 18b: 1H NMR (DMSO-d6, +50 °C,
700 MHz) δ: 770 (AA’ part of AA’ BB’ multiplet, 4H, 4×CHar); 7.05 (BB’ part of AA’BB’
multiplet, 4H, 4× CHar); 3.23 (br. s, 3H, 3×NH); 2.61 (br. d, J=9.6 Hz, 18H, 6×CH3
phosphazene); 2.26 (s, 6H, 2×Ph-CH3); 1.18 (br. s, 9H, C-CH3)3 phosphazene); 13C NMR
(DMSO-d6, +50 °C, 176 MHz) δ: 165.91; 163.86; 138.7; 129.0; 120.3; 49.2; 40.5; 40.4; 40.2;
40.1; 40.0; 39.89, 39.77; 39.90; 36.90; 36.88, 29.8; 20.8; HRMS (ESI) m/z 308.15057 [M+H]+
+
(C17H18ON5 requires 308.15059).
26

4.4
Materials and instruments.
All of the final compounds were purified to >95% by column chromatography, using Silica gel
40 MN (Fluka -60735, particle size 0.063-0.2 mm, 70-230 mesh ASTM) and aluminium backed
Silica gel 60 F254 from Merck was used for analytical TLC. Mass spectra (MS) were obtained
from the Mass Spectrometry Voyager DE Pro (Applied Biosystems, USA).
1H and 13C NMR measurements were conducted using either Bruker AVANCE III 700
(700.1 MHz for 1H and 176.1 MHz for 13C) or Bruker AVANCE 500 (500.1 MHz for 1H and
125.8 MHz for 13C) instruments at +20 or + 50 °C. The higher temperature was used to make the
rate of the tautomeric equilibrium faster and therefore to obtain sharper signals in the spectra. 1H
chemical shifts were directly referenced to internal TMS (0.00 ppm) or indirectly to TMS via the
solvent signal (CHCl3, 7.26 ppm, DMSO-d5, 2.50 ppm), and 13C chemical shifts were indirectly
referenced to TMS via the solvent signal (CDCl3, 77.00 ppm, DMSO-d6, 39.52 ppm).
Sample analysis was performed with the Shimadzu LC Solution (Prominence) system by
using an injector and a diode array detector (SPD 20A) with a mass detector (LCMS 2020).
Separation was achieved with a Gemini C18 5μm column (250×4.6 mm i.d., Phenomenex)
protected by a 5μm Gemini C18 guard column (4×2.0 mm i.d., Phenomenex). Mobile phase A:
0.1 % TFA, mobile phase B: 0.1 % TFA in ACN and a flow rate of 1 mL/min were employed.
The linear gradient speed was 3 % B/min, and elution was started at 3 min (the injection time
was also at 3 min) with 30 % B. Sample peaks were confirmed by the retention time and mass
data. Electrospray ionization was achieved at a positive mode for [M+H]+. The desolvation line
temperature was 250 °C, the heat block temperature 200 °C, the ionization voltage 4.5 kV, and
the qarray RF voltage 60 V. The drying gas flow was set at 15 L/min and the drying gas
temperature at 250 °C; the scan speed was 15,000 μ/sec.
27

The mixture 17/18 was analysed by Agilent’s 1200 series HPLC combined with a diode
array detector and microTOFQ mass spectrometer (Bruker Daltonics), i.e., by HPLC-ESI-QTOF-
MS. Separation was achieved by a BEHC18 column (Acquity UPLC, Waters) with a gradient
elution. Two mobile phases, A: ACN and B: 0.1 % HCOOH, were used as follows: 0-5 min 0-30
% A in B, 5-6 min 30-70 % A in B, and 6-8 min 70 % A in B. The flow rate was 0.45 mL/min,
and the injection volume was 5 μL. Electrospray ionization was used in a positive ion mode. The
capillary voltage was set at –4500 V, with the end plate offset at –500 V. The nebulizer gas (N2)
was set at 1.6 bar, and the dry gas (N2) flow was set at 8 L/min and the temperature at 200 °C. A
mass range of m/z 50–3000 was used. The data were handled by Bruker Compass DataAnalysis
(version 4.0).
High-resolution ESI-ICR spectra were obtained on a hybrid Varian 910-FT-ICR-MS
system, which was coupled to Varian-J320 triple-quadrupole MS. The ionization parameters
were as follows: ESI-needle voltage, 5000 V; spray chamber temperature, 40°C; nebulizing gas
(N2) pressure, 30 psi at 300 °C; API-drying gas (N2), 18 psi at 300 °C; shield voltage, 600 V; and
capillary voltage, 60 V; and infusion rate, 10 l/min. The ion guide and FT-ICR parameters were
optimised for the mass range m/z 100–800 as described in ref. [54]. The ion collection time was
500 ms, after which ions were passed to the FT-ICR analyser cell for excitation and detection
(detection mode: direct (broadband); ADC rate: 4 MHz; transient length: 2048 K; 524.288 ms).
Samples were dissolved in 1 % HCOOH MeOH solution, so that final concentration of analyte
was about 10 g/ml. For steady calibration of the m/z axis, samples were spiked with the in-
house prepared internal calibration solution, which contained ions with the following exact m/z
+
values: 1. C16H36N+ (m/z = 242.28423); C19H29N4PF3 (m/z = 401.20765) [55]; 3. C26H45N7P2Cl+
28

+
(m/z = 552.28947) [55]; 4. C26H64N13P4 (m/z = 682.43526) [56]. The concentration of the
calibrants in the infused solutions remained within range 0.5–1.0 mM.
4.5
Cell-culture based assays used for cytotoxicity and anti-HIV-1 activity measurements
4.5.1 Materials used.
Dimethyl sulfoxide (DMSO), polybrene, Phorbol 12-myristate 13-acetate (PMA), nevirapine and
trichloracetic acid (TCA) were purchased from Sigma Aldrich (USA). Na4P2O7 was purchased
from Applichem (Germany). Recombinant purified HIV-1 reverse transcriptase was purchased
from Calbiochem (USA).
4.5.2 Cells and growth media.
ACH-2 and TZM-bl cell lines were obtained through the NIH AIDS Reagent Program, Division
of AIDS, NIAID, NIH: ACH-2 [57,58] from Dr. Thomas Folks and TZM-bl [59,60,61,62,63]
from Dr. John C. Kappes, Dr. Xiaoyun Wu and Tranzyme Inc.
U2OS human osteosarcoma cells were obtained from ATCC and grown in Iscove's
Modified Dulbecco's Medium (IMDM) supplemented with 10 % fetal bovine serum (FBS), 100
U Penicillin and 100 µg/mL Streptomycin (Pen/Strep). ACH-2 cells were grown in Roswell Park
Memorial Institute medium 1640 (RPMI 1640) supplemented with 25 mM HEPES, 0.3 g/L L-
Glutamine, 10 % heat-inactivated FBS and Pen/Strep. TZM-bl cells were grown in Dulbecco's
Modified Eagle's medium (DMEM) supplemented with 10 % heat-inactivated FBS and
Pen/Strep. All of the cells were grown at 37 °C in the presence of 5 % CO2. All of the reagents
and media used for cell cultivation were purchased from Naxo OÜ (Estonia).
4.5.3 Cytotoxicity assay.
The cytotoxicity of compounds 1-19 was measured using the xCELLigence RTCA DP
Instrument (ACEA Biosciences, Inc) as follows. 3.5×103 U2OS cells or 8.5×103 TZM-bl cells
29