Triphospharylated 2'-5'-linked triadenilates (ppp2'-5'A₃) activate ribonuclease L (RNase L) and trigger interferon-induced antiviral protection of the cell [1]. Interaction of ppp2'-5'A₃ with the enzyme is the essential condition for the induction of ribonuclease activity [2]. RNase L binding to the 5'-end of 2'-5'-oligonucleotide requires the presence of at least one phosphate group, while the activation of enzyme requires at least two phosphate groups [3]. Three active domains of the RNase L have been discovered so far. The first one is responsible for direct binding to ppp2'-5'A₃, the second one has nuclease activity, and the third one possesses protein kinase (PK) activity [4]. In the absence of ppp2'-5'A₃, the inhibitory sequence within the third domain of RNase L inhibits RNase domain. Interaction of ppp2' 5'A₃ with RNase L induces structural alterations within protein’s structure, causing the formation of a dimeric form. As a result, the ribonuclease domain acquires the capacity of viral RNA disintegration. Dephosphorylated 2'-5'-tryadenilates (2'-5'A₃) do not bind to RNAse L. Their role in cell metabolism is not completely clear. PKs are the key mediators of various cellular reactions (Table 1). In many diseases, including cancer pathology and hypertension, down regulation of PK activity is observed. That is why many protein kinase inhibitors exhibit antitumor activity. Previously we have shown that the 2'-5'A₃ and their analogues have specific biological effects, affecting the voltage-dependent calcium channels, activity of which is associated with the protein kinase A [5]. It was shown that 2'-5'A₃ and their analogues affect the contraction of smooth muscle of aorta and femoral artery in rats. The action of the 2'-5'A₃ and their analogues can be eliminated by using the inhibitors of calcium-dependent potassium channels of high conductivity or PK A inhibitors [6-13]. We’ve demonstrated the capability of 2'-5'A₃ binding to several biologically relevant proteins [14], as well as its capability of altering protein’s secondary structure [15]. We have suggested a possible concept of 2'-5'A₃ involvement in various biological processes, which includes its unspecific binding to cell proteins, including protein kinases, followed by conformational changes within target proteins, which, in the end, leads to activity value shifts, both negative and positive. In this paper we demonstrate the direct evidence for the capability of 2'-5' A₃ and its epoxy-modified analogue to modulate the activity of protein kinase Aurora. Our previous data demonstrate that 2'-5'A₃ are capable of affecting the value of L-currents of Ca2+ channels [16]. Considering the fact that this process is regulated via the specific protein kinase’s activity, current study is focused on the investigation of possible direct 2'-5'A₃ influence on protein kinase activity.

Table 1: Biological activities of different protein kinases.

We used protein kinases produced by Millipore (USA), CK2 and peptide substrate RRREEETEEE (CK-tide) by New England Biolabs, (USA).

Synthesis of oligoadenylates

Natural 2'-5' triadenilate and its epoxy modified analogue, 2'- 5' epoxytriadenilate (2'-5'A₃-epo), were synthesized as described previously (Figure 1 for oligoadenylates’ chemical formulae) [5].

Figure 1: Structures of synthesized 2'-5'-oligoadenilates.

Assaying protein kinase activity in the presence of oligoadenylates

The study of the protein kinase activity in the presence of 2'-5'A₃ was conducted in test system by directly determining the product of PK activity using [32P]-ATP in vitro [17]. The radioactivity was determined with the use of scintillation counter TriSarb-7000 (Perkin- Elmer, USA). Protein kinase activity level was determined by the ratio of the radioactivity of reaction products in the presence and absence of 2'-5'A₃. Protein kinase was titrated with 2'-5'A₃ and 2'-5'A₃-epo at concentrations from 5 μM to 160 μM in the presence of 25 and 100 μM ATP. To determine the IC₅₀ value, oligoadenylates were used at the concentration range from 50 nM to 50 μM.

Statistical analysis

Each protein kinase activity experiments were conducted at least 3 times. The data were probed with the use of one-way analysis of variance (Anova), followed by the Student t-test. Results are expressed as mean ± SD (standard deviation). Obtained data has been shown to be statistically significant (n=3, p<0.05).

Computer modeling

Docking experiments were carried out using the Gold Suite software. Considering the absence of any information within literature sources, binding site was set by single atom selection (c-gamma atom of Leucin 194) deep in the protein’s cavity and spread to 15 angstrom diameter sphere. Visualization of docking results was performed with the use of PyMol software.

We have investigated the influence of 2'-5'A₃ and 2'-5'A₃-epo (Figure 1) on the activity of several PKs: Aurora, FGFR, TIE2, ROCK1, JNK1, CK2 and ASK. For clarity sake, we have only provided the titration curves for protein kinase Aurora in the presence of 2'-5'A₃ and 2'-5'A₃- epo within this study, since we believe that this particular kinase is the most relevant to suggested biological function of oligoadenylates. As it follows from the obtained results, these PKs behave differently upon binding to oligoadenylates. In most cases, the bot 2'-5'A₃ and 2'-5'A₃- epo stimulate the activity of protein kinases. It turned out that 2'-5'A₃ mainly activates studied protein kinases and particularly strongly (84%) increases the activity of protein kinase CK2. JNK1 and ROCK1 kinases are up regulated moderately (31% and 40% respectively) by 2'- 5'A₃. The protein kinases TIE2, ASK, and FGFR activity is increased by 18-25% upon binding of 2'-5'A₃. On the contrary, in case of Aurora kinase in particular, 2'-5'A₃ demonstrates opposite effect – 35% inhibition. A similar to 2'-5'A₃, yet more pronounced effect was demonstrated by the 2'-5'A₃-epo: its stimulating activity was higher than 2'-5'A₃. On the contrary to 2'-5'A₃, which slightly stimulates the activity of TIE2, 2’-5’epoA₃ inhibits this protein kinase by 24%, and increases the inhibition of the Aurora kinase by 5% (40% inhibition vs. 35%). It also significantly inhibits the ASK (21%) and CK2 (12%) kinases. It is worth mentioning, that 2'-5'A₃ and especially its epoxy modified analogue had pronounced stimulatory effect on protein kinases, specifically ASK. However, 2'-5'A₃ had little stimulatory effect on the protein kinase TIE2, in contrast to the inhibitory action of 2'-5' A₃-epo. At the next stage of our research, we determined the dependence of Aurora activity on the concentration of oligoadenylates. The enzymatic reaction was performed in the presence of 2'-5'A₃ and 2'-5'A₃-epo at different concentrations – from 5 to 160 μM), at two ATP concentrations – 25 μM and 100 μM. Titration curves were constructed by plotting the activity of protein kinase versus the logarithm of concentration of particular oligoadenylate in μM (lg C). It was found that at both ATP concentrations obtained curves are not linear. At the ATP concentration of 100 μM, the curves for both 2'-5'A₃ and 2'-5'A₃- epo are V-shaped (Figures 2 and 3). Titration of Aurora with 2'-5'A₃ resulted in slight rise of its enzymatic activity within the lgC 0.7-1.0 concentration range. Further increase in lgC to 1.3 lead to sharp activity drop, while lgC rise to 1.9 led to the restoration of Aurora enzymatic capability. Similar effect has been demonstrated by the 2'-5'A₃-epo (Figure 3). Titration curve of the enzyme by 2'-5'A₃ is V-shaped with falling activity in the region, where lgC ranges from 0.7 to 1.3, then there is a non-linear growth with a local maximum at lgC 1.6, and after lgC 1.9, further 2'-5'A₃ concentration increase ends with a sharp drop in inhibitory effect and almost complete restoration of the basic activity of the enzyme at lgC 2.2. Inhibition of Aurora by 2'-5'epoA₃ in fact is sharply different from the one, demonstrated by 2'-5'A₃: it is deeper and remains at a high level in a wider range of concentrations. From the obtained titration curves, IC₅₀ values of 2'-5'As were determined. It was found that the IC₅₀ value in the area of low concentration for 2'-5'A₃ was 19.3 μM, and for 2'-5'epoA₃ 11.6 μM in the presence of 100 mM ATP. Thus, 2'-5'epoA₃ had significantly greater inhibitory effect at lower concentrations, than the 2'-5'A₃. After conducting molecular docking study, the best pose with the most negative Gibbs free energy value for each ligand, 2'-5'A₃ and 2'-5'A₃-epo, was chosen after (Figure 4). Estimated association energies equaled -5.827 kJ and -4.512 kJ for 2'-5'A₃ and 2'-5'A₃-epo respectively. Detailed docking results analysis showed higher affinity of 2'-5'A₃ to Aurora then 2'-5'A₃-epo. The association energies are quite comparable though – only 1.3 kJ difference. This can be explained considering ≈ 96% structural identity of both ligands. Nevertheless, considerable differences in interaction patterns of those 2 with protein were postulated. While 2'-5'A₃ did not directly interact with aromatic aminoacid residues, 2'-5'A₃-epo was shown to create H-bonds with water molecules, which were associated via hydrogen bonds with several aromatic aminoacid residues – Trp152 and Tyr87. Water molecules were shown to additionally stabilize the complex via H-bonding closely-spaced atoms.

Figure 2: Dependence of the Aurora activity on the concentration of 2'-5'A₃ and 2'-5' A₃-epo. ATP concentration was 100 μM.

Figure 3: Dependence of the Aurora activity on the concentration of 2'-5'A₃ and 2'-5'A₃-epo. ATP concentration was 25 μM.

Figure 4: Visualization of docking results for binding of Aurora to 2'-5'A₃ (A) and 2'-5'A₃-epo (B). Oligoadenylate molecules are represented as white sticks, lying in the cavity of the protein, represented as a surface.

2'-5'A₃ and its analogues under the described conditions showed different, both inhibiting and stimulating the effects on PKs. Specifically, we would like to emphasize almost two-fold increase of CK2 activity under the influence of 2'-5'A₃ and 2'-5'A₃-epo. In most cases, 2'-5'A₃-epo provided stronger stimulatory effect compared with the unmodified 2'-5'A₃. Thus, the data shows that the 2'-5'As significantly affect the activity of protein kinases, and can act as inhibitors and activators of PK, depending on the enzyme and the nature of the oligoadenylate. Therefore, their effect on the activity of PK depends on the concentration of oligoadenylate and ATP concentration. Hence, we believe that dephosphorylated 2'-5' oligoadenylates might modulate the activity of regulatory proteins, such as protein kinases, and thus affecting various cellular processes.

Let us examine some possible mechanisms of the outlined above effects. There are many protein kinases, however, all of them possess the ATP binding site, target protein binding site, and a site for transferring of phosphate from PK to protein, which is phosphorylated. If the ligand binds effectively to any of these sites, the inhibition of phosphorylation occurs. It is known that adenine and adenosine interact with the ATP binding site and may be the potential inhibitors of protein kinases. It can be assumed that 2'-5'oligoadenylates also operate in a similar manner, which is demonstrated by the obtained molecular docking simulations data. Low-molecular compounds that compete with ATP for the ATP-binding site prevent the phosphorylation of cellular proteins thus blocking the transmission of cellular signals. Known inhibitors of PKs often interact with the same ATP binding site.

Apart from the direct binding of ligands to the main active sites, there is another mechanism, which suggests somewhat similar to allosteric regulation of activity, interaction outside the active functional PK sites [17]. This causes conformational alterations within protein’s globule, which in turn leads to changes in the enzyme activity. The results on the influence of oligoadenylates on the activity of PK we obtained prove the correctness of this assumption: not only inhibition but also the increased activity of enzymes has been witnessed. In case of oligoadenylates binding to the active site of PK (for example ATPbinding), the inhibition of enzymatic activity should be observed. Therefore it is likely, that the 2'-5'A₃ and its analogues modulate the activity of PK by influencing their conformation. 2'-5'A₃ impact the activity of protein kinases, but, contrary to classical inhibitors, they could act as mediators, which are capable of demonstrating both stimulating and inhibiting effects.