ISSN 1068-1620, Russian Journal of Bioorganic Chemistry, 2008, Vol. 34, No. 4, pp. 442–452. © Pleiades Publishing, Ltd., 2008. Original Russian Text © L.S. Koroleva, V.E. Kuz’min, E.N. Muratov, A.G. Artemenko, V.N. Sil’nikov, 2008, published in Bioorganicheskaya Khimiya, 2008, Vol. 34, No. 4, pp. 495–505.
Artificial Ribonucleases: Quantitative Analysis of the Structure– Activity Relationship and a New Insight into the Strategy of Design of Highly Efficient RNase Mimetics L. S. Korolevaa,1, V. E. Kuz’minb, E. N. Muratovb, A. G. Artemenkob, and V. N. Sil’nikova a
Institute of Chemical Biology and Fundamental Medicine, Siberian Branch, Russian Academy of Sciences, pr. Akademika Lavrent’eva 8, Novosibirsk, 630090 Russia b Department of Molecular Structure, Bogatskii Physicochemical Institute, National Academy of Sciences of Ukraine, Ukraine Received July 18, 2007; in final form, November 26, 2007
Abstract—The dependence of hydrolytic activity of artificial ribonucleases toward an HIV-I RNA fragment, a 21-mer oligonucleotide, and tRNAAsp on the structure of the RNase mimetic was analyzed. The quantitative structure–activity relationship (QSAR task) was determined by the method of simplex representation of the molecular structure where the amounts of four-atom fragments (simplexes) of fixed structure, symmetry, and chirality served as descriptors. Not only the types of atoms participating in simplexes, but also their physicochemical properties (e.g., partial charges, lipophilicities, etc.) were taken into account. This allowed the determination of the relative role of various factors affecting the interaction of molecules under study with the corresponding biological target. The 2D QSAR models obtained by the method of projection to latent structures have quite satisfactory statistical indices (R2 = 0.82–0.96; Q2 = 0.73–0.89), which help predict the activities of new compounds. The electrostatic properties of ribonuclease atoms were shown to contribute significantly to the manifestation of the hydrolytic activity of ribonucleases in the case of the 21-mer oligonucleotide and tRNA. In addition, the structural fragments that most greatly contribute to the alteration of the hydrolytic activity of RNases were identified. The models obtained were used for the virtual screening and molecular design of new highly efficient RNase mimetics. Key words: artificial ribonucleases, QSAR, simplex descriptors DOI: 10.1134/S1068162008040080
INTRODUCTION Ribonucleases from various organisms are promising preparations in the treatment of some viral and oncological diseases and present an alternative of the generally accepted chemotherapy.2 At present, the antiviral and antitumor effects of some RNases has been confirmed by clinical trials. The most known among these are RNase A and onconase, an RNase from the oocytes of the frog Rana pipiens. The application of ribonucleases as medicinal drugs is being studied intensively, in particular, in the context of directed mutagenesis and enhancement of the contribution of molecular determinants responsible for the selective injury of cancer cells [1–3]. At the same time, the shortcomings inherent in natural RNases, such as a high molecular weight, the presence of effective inhibitors, low stability, and bioacces1
Corresponding author; phone: +7 (383) 333-3762; fax: +7 (383) 333-3677; e-mail:
[email protected] 2 Abbreviations: QSAR, quantitative structure–activity relationship; Hia, histamine; PLS, projection to latent structures; PASS, prediction of activity spectra for substances.
sibility, as well as difficulties in producing homogeneous preparations, stimulate studies aimed at designing low-molecular-weight organic catalysts of RNA hydrolysis that imitate the functions of natural RNases [4]. In the last ten years, we in our laboratory have synthesized a wide spectrum of compounds of different structure that are capable of catalyzing, to a variable degree, the hydrolysis of phosphodiester bonds in RNA [5–8]. The RNA hydrolysis in the presence of these artificial ribonucleases is a directed process: the most efficient cleavage occurs at Pyr–Pur sites. It was shown that various RNA substrates have as a rule two to five sites at which the cleavage of about 90% of total phosphodiester bonds occurs [5–8]. RESULTS AND DISCUSSION One strategy to design artificial ribonucleases is a structural and functional modeling of the catalytic sites of natural enzymes. This strategy involves the introduction into the structure of RNase mimetics of amino acid residues (or their side functional groups) that play a key role in the functioning of the catalytic sites of natural
442
ARTIFICIAL RIBONUCLEASES: QUANTITATIVE ANALYSIS
443 His12
O
His12
Base1
O
O :N
NH2 O
Gln11
Lys41
O– H3N
P O
+
NH2
N
O
P
+
H3N
P
OH
Lys41
O–
O HO
N O
O
Base2
H
HN His119
NH
+
O
Gln11 O
Base1
O
O
O H +
O
O
NH
HN
Base2
O
HN His119
O–
O
OH
O Fig. 1. Scheme of the classical mechanism of the hydrolysis of phosphodiester bonds in ribonucleic acids by RNase A [4]. The location and functions of amino acid residues in the active site of RNase A are schematically shown.
O H3C
(CH2)13
+
+
N
(CH2)k
N
H N
C
R C H
CH2
L
A
N
B R – H, COOH, COOCH3
NH
C
k = 1, 3, 4, 5
Fig. 2. An example of the realization of the design strategy for artificial ribonucleases from four functional domains: a lipophilic fragment (A), an RNA-binding fragment (B), a linker (L), and a catalytic domain (C).
ribonucleases and are spatially arranged in the way optimal for the realization of the catalytic process (Fig. 1). By analogy with the terminology accepted in protein chemistry, the structural fragments of artificial ribonuclease molecules that are functionally analogous to particular domains of natural fragments are called domains. An artificial ribonuclease involves a catalytic domain (C), an RNA-binding domain (B), a linker group binding these domains (L), and a hydrophobic fragment (A) (Fig. 2). It has been shown earlier that the presence of an extended alkyl fragment A in the structure of artificial ribonucleases is an important prerequisite for attaining a high hydrolytic activity [9], and its absence markedly reduces the activity. This strategy was used in the synthesis of several series of artificial ribonucleases: • structural and functional analogues of the catalytic site of ribonuclease A: dipeptides [5] and tripeptides RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
(1)–(6) [10] containing histidine or Hia and positively charged amino acids (Lys and Arg): His-Lys-Arg (1) Lys-Thr-Hia (4) Lys-His-Arg (2) Arg-Lys-Hia (5) Lys-Asn-Hia (3) Asp-Lys-Hia (6) • structural and functional analogues of the catalytic site of ribonuclease T1 [6]: (1) Short peptides (7)–(12), which contain residues Arg, Glu, Ser, Thr, Lys, and Phe in different combinations:
Vol. 34
Arg-Thr-Glu-Phe-OC8H17 (7) Lys-Ser-Glu-Phe-OC8H17 (8) Lys-Thr-Glu-Phe-OC8H17 (9) Thr-Lys-Glu-Phe-OC8H17 (10) Glu-Thr-Lys-Phe-OC8H17 (11) Glu-Ser-Lys-Phe-OC8H17 (12) No. 4
2008
444
R
KOROLEVA et al. N+
where X = Gly (13), (14); βAla (15), (16); 4-aminobutanoic acid (17), (18); 6-aminohexanoic acid (19), (20); and p-aminobenzoic acid (21), (22). · RNase mimetics containing the catalytic active group, which is linked to the polycationic fragment based on 1,4-diazabicyclo[2,2,2]octane by linker alkyl groups of different length [6, 7, 9] (Fig. 2). · Two residues of quaternized 1,4-diazabicyclo[2,2,2]octane with alkyl substituents of different lengths [8], linked by different linkers (Fig. 3). In the last series, the mechanism of cleavage of RNA is based on the acceleration of its spontaneous hydrolysis due to the distortion of RNA structure by polycationic molecules (Fig. 4). The length of the rigid linker group of RNases of this series is comparable with the distance between adjacent phosphate groups of RNA. The interaction of these cationic molecules with RNA leads to the optimization of the geometry of the complex (a decrease in distance d and an increase in angle α), which facilitates the RNA hydrolysis via an “in line” mechanism [8, 11]. The catalytic activity of most RNase mimetics has been studied earlier in reactions with different tRNAs (tRNAAsp, [9], tRNALys [5, 9], and tRNAPhe [7, 12]). Ribonucleotides, synthetic (di- [13, 14] and oligomers of different length [13–15]) and natural (HIV I 96-mer
N+ X N+
N+
X:
R:
R
CH3 C4H9 C6H13 C12H25
* n* n = 3, 4, 5
Fig. 3. Artificial ribonucleases based on the residues of quaternized 1,4-diazabicyclo[2,2,2]octane with alkyl substituents of different lengths.
(2) Tetrapeptides (13)–(22), which contain negatively charged (Glu) and positively charged amino acids (Lys/Arg) separated by linkers of different length: Glu-X-Arg-Gly-OC10H21 (13), (15), (17), (19), (21); Glu-X-Lys-Gly-OC10H21 (14), (16), (18), (20), (22),
HO
U
O
2'
O
O H
P O
O O
P
O
H
O–
5' O
O–
O U
O N +
N
+
L i n k e r
2'
O O
N +
N +
O
O
7.1–8.2 Å
P
d α–
O
O–
H
O
5'
O
O H
P O–
O
A
O
O
H+
H
2'
O
O
P 5'
H O
O
O–
Fig. 4. Hypothetical mechanism of the hydrolytic cleavage of phosphodiester bonds by synthetic ribonucleases based on bisquaternary salts of 1,4-diazabicyclo[2,2,2]octane. RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
Vol. 34
No. 4
2008
ARTIFICIAL RIBONUCLEASES: QUANTITATIVE ANALYSIS
RNA fragment [6], an mRNA fragment of influenza virus protein M2 [15], and a tRNA-like fragment of turnip yellow mosaic virus RNA [5]), were also used as targets to determine how the structure of RNA substrates influences the efficiency of hydrolysis of phosphodiester bonds. The diversity of RNA substrates and the conditions of testing the catalytic activity of the artificial ribonucleases synthesized makes a direct comparison of the experimental data obtained for these substrates impossible. Considering the structural diversity of the compounds studied, this makes the establishment of general structure-catalytic activity relationships difficult. This work is the first attempt to determine the molecular structural characteristics responsible for the hydrolytic activity of RNase mimetics. The main tool to study the quantitative structure-activity relationship was the QSAR approach based on the simplex representation of the molecular structure [16, 17]. In all, 64 compounds (see formulas above and Fig. 5) exhibiting the hydrolytic activity toward the HIV I RNA fragment, a 21-unit oligonucleotide, and tRNAAsp were considered and included into subsets (Table 1). QSAR Method Based on Simplex Representation of Molecular Structure In the QSAR method used, molecular structures are described by simplex descriptors equal to the number of four-atom fragments (simplexes) of the fixed structure, symmetry, and chirality [17]. The atoms in simplexes can be differentiated according to different characteristics, in particular: · the symbol characterizing the chemical nature of the atom (designation of a chemical element or a more clearly defined type of the atom that takes its chemical surrounding into account); · partial charge on the atom [18, 19] (see Fig. 6) (reflects the electrostatic properties); · lipophilicity of the atom [20] (reflects the hydrophobic properties); · refraction of the atom (reflects to some extent the dispersion properties); and · donor/acceptor properties of the hydrogen atom in the potential H-bond.3 The use of different variants of the differentiation of tops of simplexes (atoms) is a crucial feature of the simplex approach4. An example of generation of simplex 3 Atoms
are divided into three groups: A, acceptor; D, hydrogen donor in the H-bond; and I, indifferent atom. 4 We believe that the detailing of atoms according to their nature (e. g., C, N, O), which is realized in many QSAR methods, constrains the possibilities of identification of pharmacophoric fragments. Thus, if the NH-group is chosen as a fragment that determines the possibility of hydrogen bonding, we can overlook H-bond donors, such as the OH-group. RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
445
descriptors for an alanine molecule using different modes of differentiation of atoms is given in Fig. 6. Simplex descriptors are actually fragmentary parameters, which describe not the entire molecule but its different parts. In order to reflect the structural features of the molecule as a whole (i.e., to describe its integral properties), the probabilities of the manifestation of different modes of the biological activity were also involved in this work as descriptors; they are calculated by the program PASS [21], which has proven useful in solving some QSAR tasks [22]. Statistical dependences were constructed using the PLS method [23], which has been used with advantage in processing great files of structural data since it is based on the conversion of a great number of structural parameters into a small number of latent variables. The search for an optimal set of structural parameters in the PLS was carried out using some procedures such as the exclusion of mutually correlating and constant parameters, the genetic algorithm [24], and the method of eliminating the noise descriptors [17]. The PLS equation can be represented as [16]: N
Y = b0 +
∑b x , i i
(1)
i=1
where Y is the activity under study, bi is the regression coefficients of PLS, xi is the value of the i-th descriptor, and N is the total number of descriptors. By using this equation in the framework of the simplex representation of the molecular structure, it is rather easy to solve the inverse problem (interpretation of the QSAR model). The contribution of each atom to the molecule can be defined as the ratio of the sum of regression coefficients (bi) of all simplexes containing this atom to the number of atoms in the simplex. Based on this information, it is possible to identify structural fragments (combinations of atoms) that contribute to, or prevent the manifestation of the activity studied. In the framework of the simplex approach, it is also possible to determine the relative influence of various physical and chemical factors on the interaction of molecules with the biological target. To do this, it is necessary to compare the percentage of modules of normalized contributions (bi) involved in the model of simplexes for each of the properties according to which atoms in simplexes were differentiated (charge, lipophilicity, refraction, etc.) (see Fig. 6). As a result, using the QSAR method based on the simplex representation of the molecular structure, we obtained three independent statistical models (I, II, and III), which describe the relationship between the structure of RNases and their hydrolytic activity toward: the HIV RNA fragment (I), a 21-unit oligonucleotide (II), and tRNAAsp (III). The quality of these models is wholly satisfactory (Table 2). The analysis of the contributions of bi [equation (1)] of simplexes involved in the model indicated that the
Vol. 34
No. 4
2008
RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
C12H25
N+
NH2
N+
H2N
O
O
O
N H
Vol. 34
No. 4
2008
(66)
N+
Cl
N H
+
C12H25 C12H25
C14H29 N
X=
N H O
n
N+
+
N+
(67)
N H
O
O
O
N
+
H N
N+
N
NH
O
R1 N+
C12H25
X:
a
N+
O
X
N
R
N COOCH3
N H
N+
N+ R2
O
(51)
N
NH
(53)–(64)
N+
NH
N+
(47)
N+
NH
O
(53) (54) (55) (56) (57) (58) (59) (60) (61) (62) (63) (64)
No.
(52)
H N CHCH2
N+
C14H29 N+
H N
(36)–(46)
(48)
N+
NH
O N+ (CH2)k C
H2N
N C2H5 N+
N+
O (50)
N+
N H
(65)
NH
O
C15H31
N3
N+
N+
(27)
n = 1 (23) n = 2 (24) n = 3 (25) n = 5 (26)
H3C (CH2)n
O
n
(CH2)4
O
(31)
O
(35)
N H
N
O
N H
N H
n = 1 (28) n = 2 (29) n = 3 (30)
O C10H21
O C10H21
O
X=
X=
(36) (37) (38) (39) (40) (41) (42) (43) (44) (45) (46)
a a a a a a a a a a a a
X
O
meta meta meta meta ortho ortho ortho ortho para para para para
Isomer
N+
13 13 13 13 13 13 13 13 13 13 1
n
N COOCH3 No.
N+
H N
Fig. 5. Structures of artificial ribonucleases (23)–(64) examined in the study and ribonucleases (65)–(67) proposed on the basis of QSAR models.
N+
+
N
(34)
NH2
N
NH
NH2
O
O
X=
H N
NH
H N
O NH2
HN
NH
NH HN
N+ (49)
X
X
(CH2)4
NH
(CH2)4
N H
HO O
O
N H
O
(32) R = H (33) R = COOMe
H2N
N+ Br–
N+ Br–
NH2
NH
N
(H2C)4
O
R
HN
C14H29 N+ Br–
C14H29 N+ Br–
HO O
O
H2N
R
N
CH3 C12H25 C4H9 C6H13 CH3 C12H25 C4H9 C6H13 CH3 C12H25 C4H9 C6H13
R2
NH
COOCH3
COOH H H COOCH3 H COOCH3 COOH H COOCH3 COOH COOH
CH3 C12H25 C4H9 C6H13 CH3 C12H25 C4H9 C6H13 CH3 C12H25 C4H9 C6H13
R1
N H
3 1 3 3 4 4 4 5 5 5 3
k
NH
446 KOROLEVA et al.
ARTIFICIAL RIBONUCLEASES: QUANTITATIVE ANALYSIS
447
NH2 HO O Alanine
Differentiation according to atom type
Differentiation according to charge
Description of the step Calculation of atomic characteristics
H 0.07
H 0.07
0.02
N –0.21 O 0.24 C 0.06 H 0.02 H C C –0.04 0.12 H 0.02 O –0.18 H 0.02 –0.21
H
A ≤ –0.1 –0.1 < B ≤ –0.05 –0.05 < C ≤ –0.01 –0.01 < D ≤ 0.01 0.01 < E ≤ 0.05 0.05 < F ≤ 0.1 0.1 < G H N H
O
O H H
O
C+ O
O
N +
C O
C
H
C
C O
H +
C
+ H
G
+ …
A
G
G
E +
G
F
+
A E
F
C E
A E +
Generation of simplex descriptors F +
G A
E F
A
F
E
E + C
A
H
C
E A
H N
A
G
C
H
H +
H
C
C
Setting of new atom labels
F
H
Differentiation of atoms into groups
E A
F
+3
F
F C
+…
E
Fig. 6. Generation of simplex descriptors in the framework of the simplex representation of the molecular structure, as exemplified by an alanine molecule.
major contribution in the case of the 21-unit oligonucleotide is made by electrostatic (49%) and hydrophilic/hydrophobic (38%) interactions, and the effect of the nature of atoms accounts for only 13%. In the case of tRNAAsp, electrostatic interactions make the major contribution (40%). In this case, the nature of the atom and the hydrophilic/hydrophobic interactions also play a great role (Fig. 7). In the case of the HIV RNA fragment, integral descriptors characterizing the molecule as a whole have a greater effect on the cleavage (84%), RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
and the contribution of the nature of atoms accounts for only 16%. By the solution of the inverse task (interpretation of QSAR models), we determined the molecular fragments of RNase that promote and prevent the manifestation of the activities under study (Table 3). In the case of the 21-unit oligonucleotide and tRNAAsp, particular regions of the molecule contribute markedly to the manifestation of the ribonuclease activity. In the case of the HIV RNA 96-mer fragment, not so much the presence or absence of particular structural fragments is of
Vol. 34
No. 4
2008
448
KOROLEVA et al.
Table 1. Observable hydrolytic activity of compounds toward RNA substrates Degree of hydrolysis of P–O bonds, %* Compound no.
HIV RNA fragment
21-unit oligonucleotide
tRNAAsp
(1)
–
55
–
(2)
–
31
(3)
–
(4)
Degree of hydrolysis of P–O bonds, %* Compound no.
HIV RNA fragment
21-unit oligonucleotide
tRNAAsp
(33)
–
13
–
–
(34)
–
99
–
2
–
(35)
–
70
–
–
10
–
(36)
74
–
–
(5)
–
31
–
(37)
–
–
8
(6)
–
1
–
(38)
–
–
20
(7)
31
–
–
(39)
–
–
70
(8)
25
–
–
(40)
–
–
26
(9)
15
–
94
(41)
–
–
45
(10)
32
–
–
(42)
–
–
85
(11)
34
–
–
(43)
–
–
76
(12)
35
–
–
(44)
–
–
90
(13)
37
–
–
(45)
–
–
64
(14)
13
–
–
(46)
–
–
2
(15)
49
–
–
(47)
–
–
9
(16)
15
–
–
(48)
–
–
1
(17)
50
–
–
(49)
–
–
2
(18)
25
–
–
(50)
–
–
1
(19)
71
–
–
(51)
–
–
51
(20)
32
–
–
(52)
–
–
13
(21)
21
–
–
(53)
–
1
–
(22)
29
–
–
(54)
–
91
71
(23)
62
–
–
(55)
–
10
–
(24)
77
–
–
(56)
–
28
–
(25)
70
–
–
(57)
–
1
–
(26)
40
–
–
(58)
–
88
–
(27)
79
–
–
(59)
–
8
–
(28)
89
–
–
(60)
–
30
–
(29)
49
–
–
(61)
–
4
–
(30)
82
–
–
(62)
–
95
–
(31)
7
–
–
(63)
–
15
–
(32)
–
67
–
(64)
–
33
–
* Quantitative data on total RNA depolymerization obtained from electrophoregrams are presented. The total RNA depolymerization was RNA int⎞ determined by the formula: ⎛ 1 – ------------------ × 100%, where RNAint is the amount of intact RNA after incubation, and RNAtot is the initial ⎝ RNA tot⎠ amount of RNA. Dash denotes the lack of data (no experiment was carried out). RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
Vol. 34
No. 4
2008
ARTIFICIAL RIBONUCLEASES: QUANTITATIVE ANALYSIS
449
Table 2. Statistical characteristics of selected QSAR models* Model no.
Substrate type
R2
Q2
M
N
I
HIV RNA fragment
0.836
0.788
26
24 (4 + 0 + 0 + 0 + 0 + 20)
II
21-Unit oligonucleotide
0.959
0.892
22
21 (3 + 8 + 10 + 0 + 0 + 0)
III
tRNAAsp
0.823
0.732
17
58 (19 + 24 + 14 + 1 + 0 + 0)
* R2 is the determination coefficient, Q2 is the determination coefficient calculated under the conditions of sliding mode control; M is the amount of molecules in the training subset; N is the number of descriptors for each type of simplex differentiation (atom type, charge, lipophilicity, refraction, ability to form hydrogen bond) and the number of integral descriptors (program PASS).
importance for the manifestation of activity as the structure of the molecule as a whole. This indicates the prevalence of integral characteristics (PASS parameters) over local (simplex) ones (Fig. 7). The calculated data are consistent with the experimental results. It was shown that, on going from extended RNAs to short models, the hydrolytic activity of all artificial ribonucleases decreases [13]. In the case of the 21-unit oligonucleotide, the difference in cleavage efficiency between the compounds of one series is considerably less than in the case of the cleavage of the HIV RNA 96-mer fragment by these compounds. As exemplified by the 21-mer oligonucleotide, the functional groups forming the catalytic sites of natural
enzymes [guanidinium (1) and hydroxyl groups (6) and (25), as well as the group of the amide fragment (21)] have an unfavorable effect on the catalytic activity (Table 3). In addition, the influence of the imidazole fragment (4) on the catalytic activity is ambiguous. Aromatic fragments 18-20 also decrease the activity of the compounds toward the 21-unit oligonucleotide. Hydrophobic fragments 14–16 and alkyl fragments containing the amide bonds (10–12, 29, 30) contribute to the manifestation of the catalytic activity (toward all substrates except HIV RNA). Fragments 2, 3, and 13, which contain the quaternary salts of 1,4-diazabicyclo[2,2,2]octane contribute insignificantly to the
Influence, % 90 80 1 2 3 4 5
70 60 50 40 30 20 10 0
tRNAAsp
21-mer Oligonucleotide
HIV RNA fragment
Substrate Fig. 7. Relative influence of some physicochemical factors on the change in the catalytic activity of RNases: (1), characteristics of individuality of atoms (atom type); (2), hydrophilic/hydrophobic interactions; (3), electrostatic interactions; (4), dispersion interactions; (5), integral parameters. RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
Vol. 34
No. 4
2008
450
KOROLEVA et al.
Table 3. Relative contributions of fragments of RNases to changes in their hydrolytic activity (∆, %)* Fragment no. 1
Fragment –NH–C(=NH)NH2
HIV RNA fragment 21-Unit oligonucleotide +2
–11 U
2
N+
N+
+3
3
N+
N+
+4
NH
4
tRNAAsp
+4
+6
+1
U
U
–2 –4 –5 – – – – –
–6 –6 – +3 +28 +16 +22 +24
+1 +1 – – – – – –
+2
+3
+2
– – – 0 – – – 0 – +2 – – +1 –
+38 +7 +15 U –13 –15 –17 –8 –7 –8 –5 –28 – U
+6 +5 +6 +2 – – – – +1 – +2 – +7 +1
0
U
+8
0
+12
+8
0
+13
+11
– 0 – –
U U – –
+6 +1 +4 –1
N
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
–O–CH2–CH2– >CH–CH2–OH >CH–CH(CH3)–OH –O–Cl–C6H5 –NH–CO–O–t-Bu –CH2–CO–NH–(CH2)4– –NH–(CH2)4–CH(NH)– –NH–CO–CH(NH2)–(CH2)4–NH– N+
N+
≥N–C12H25 ≥N–C4H9 ≥N–C6H13 –(CH2)3–NH2 (m) –CH2–C6H4–CH2– (o) –CH2–C6H4–CH2– (p) –CH2–C6H4–CH2– >CH–CO–NH–CH< >CH–C(=O)–O–CH3 –(CH2)3–NH–C(=NH)–NH2 –CH2–CO–NH2 –CH2–CHOH–CH3 ≥N–C14H29 ≥N–C2H5 CH2 CH NH CO (CH2)3 CH2 CH NH CO (CH2)4 CH2 CH NH CO (CH2)5
–CH2–CH2–NH–CO–(CH2)3– –COOH –CO–C15H31 –N3
* Sign + denotes an enhancement of catalytic properties; sign –, a weakening of the catalytic properties; a dash designates a lack of data (structures with this fragment are not involved in the training subset for the property); U denotes an ambiguous influence on the catalytic activity (the direction of influence substantially depends on the surroundings of the fragment). RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
Vol. 34
No. 4
2008
ARTIFICIAL RIBONUCLEASES: QUANTITATIVE ANALYSIS Table 4. (A) Experimentally observed and (B) predicted values of the hydrolytic activity (degree of hydrolysis of P–O bonds in RNA, %) for the most promising compounds*
Structure
HIV RNA fragment
21-mer Oligonucleotide
tRNAAsp
451
ACKNOWLEDGMENTS This work was supported by the Russian Foundation for Basic Research (project no. 07-04-00990-a), the PharmaMed RUXO-008-N0-06, the program of basic research of the Russian Academy of Sciences “Origin and evolution of the biosphere,” a grant of the administration of Novosibirsk region for young scientists, and partially by the Ukrainian Research Center (grant STCU no. 3147).
A
B
A
B
A
B
(19)
71
52
–
>95
–
81
(24)
77
68
–
>95
–
>95
REFERENCES
(25)
70
68
–
>95
–
>95
(27)
79
62
–
>95
–
>95
(28)
89
82
–
>95
–
>95
(30)
82
82
–
>95
–
>95
(42)
–
84
–
75
85
74
(44)
–
73
–
82
90
80
(54)
–
>95
91
91
–
>95
(58)
–
>95
88
89
–
>95
(62)
–
>95
95
87
–
>95
(65)
–
>95
–
94
–
72
(66)
–
>95
–
>95
–
>95
(67)
–
>95
–
>95
–
>95
1. Makarov, A.A. and Ilinskaya, O.N., FEBS Lett., 2003, vol. 540, pp. 15–20. 2. Matousek, J., Comp. Biochem. Physiol. C: Toxicol. Pharmacol., 2001, vol. 129, pp. 175–191. 3. Benito, A., Ribo, M., and Vilanova, M., Mol. BioSyst., 2005, vol. 1, pp. 294–302. 4. Artificial Nucleases, Zenkova, M.A., Ed. in Nucleic Acids and Molecular Biology, vol. 13, Berlin: Springer, 2004. 5. Zhdan, N.S., Kuznetsova, I.L., Vlassov, A.V., Sil’nikov, V.N., Zenkova, M.A., and Vlassov, V.V., Bioorg. Khim., 1999, vol. 25, pp. 723–732. 6. Koroleva, L.S., Donina, A.A., Tamkovich, N.V., Kovalev, N.A., Zenkova, M.A., and Sil’nikov, V.N., Izv. Akad. Nauk, Ser. Khim., 2005, no. 11, pp. 2596–2605. 7. Konevetz, D.A., Beck, I.E., Beloglazova, N.G., Sulimenkov, I.V., Sil’nikov, V.N., Zenkova, M.A., Shishkin, G.V., and Vlassov, V.V., Tetrahedron, 1999, vol. 55, pp. 503–512. 8. Burakova, E.A. and Silnikov, V.N., Nucleosides, Nucleotides & Nucleic Acids, 2004, vol. 23, pp. 915–920. 9. Konevetz, D.A., Mironova, N.L., Beck, I.E., Zenkova, M.A., Shishkin, G.V., Vlassov, V.V., and Sil’nikov, V.N., Bioorg. Khim., 2002, vol. 28, pp. 367–378; Rus. J. Bioorg. Chem., 2002, vol. 28, pp. 331–341. 10. Kuznetsova, I.L., Zhdan, N.S., Zenkova, M.A., Vlassov, V.V., and Sil’nikov, V.N., Izv. Ross. Akad. Nauk, Ser. Khim., 2004, no. 2, pp. 435–442. 11. Soukur, G.A. and Breaker, R.R., RNA, 1999, vol. 5, pp. 1308–1325. 12. Beloglazova, N.G., Sil’nikov, V.N., Zenkova, M.A., and Vlassov, V.V., FEBS Lett., 2000, vol. 481, pp. 277–280. 13. Beloglazova, N.G., Mironova, N.L., Konevets, D.A., Petyuk, V.A., Sil’nikov, V.N., Vlassov, V.V., and Zenkova, M.A., Mol. Biol. (Moscow), 2002, vol. 36, pp. 1068–1073. 14. Zenkova, M.A., Chumakova, N.L., Vlassov, A.V., Komarova, N.I., Ven’yaminova, A.G., Vlassov, V.V., and Sil’nikov, V.N., Mol. Biol. (Moscow), 2000, vol. 34, pp. 456–460. 15. Beloglazova, N., Vlassov, A., Konevetc, D., Sil’nikov, V., Zenkova, M., Giege, R., and Vlassov, V., Nucleosides Nucleotides, 1999, vol. 18, pp. 1463–1465. 16. Kuz’min, V.E., Artemenko, A.G., Lozitsky, V.P., Muratov, E.N., Fedtchouk, A.S., Dyachenko, N.S., Nosach, L.N., Gridina, T.L., Shitikova, L.I., Mudrik, L.M., Mescheriakov, A.K., Chelombitko, V.A., Zheltvay, A.I., and Vanden Eynde, J.-J, Acta Biochim. Polon., 2002, vol. 49, pp. 157–168.
* Dash denotes the absence of data.
hydrolysis of the HIV RNA 96-mer fragment and tRNAAsp. The effect of aromatic fragments can be oppositely directed, depending on their surroundings. The hydrophobic aromatic fragment 8, which is located at the end of the structure, makes a positive contribution to the activity, whereas the aromatic residues 18-20, which are localized inside the structure, prevent the increase in the hydrolytic activity. The models obtained were used for the virtual screening and molecular design of efficient RNase mimetics (Table 4). By the screening of the catalytic activity, 11 compounds were selected [(19), (24), (25), (27), (28), (30), (42), (44), (54), (58), (62)], and three novel compounds [(65), (66), and (67)] with a high potential hydrolytic activity were designed. In further studies, it is intended to test all these compounds. The results will be published in ensuing communications. RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
Vol. 34
No. 4
2008
452
KOROLEVA et al.
17. Kuz’min, V.E., Artemenko, A.G., Lozitsky, R.N., Fedtchouk, A.S., Lozitsky, V.P., Muratov, E.N., and Mescheriakov, A.K., SAR and QSAR in Env. Res., 2005, vol. 16, pp. 219–230. 18. Jolly, W.L. and Perry, W.B., J. Am. Chem. Soc., 1973, vol. 95, pp. 5442–5450. 19. Kuz’min, V.E. and Berestetskaya, E.L., Zh. Strukt. Khim., 1983, vol. 24, pp. 131–132. 20. Wang, R., Fu, Y., and Lai, L., J. Chem. Inf. Comput. Sci., 1997, vol. 37, pp. 615–621.
21. Poroikov, V.V., Filimonov, D.A., Karaicheva, E.I, et al., Eksp. Klin. Farmakol., 1995, no. 2, pp. 56–62. 22. Filimonov, D.A., Akimov, D.V., and Poroikov, V.V., Khim.-Farm. Zh., 2004, no. 1, pp. 21–24. 23. Rannar, S., Lindgren, F., Geladi, P., and Wold, S., J. Chemometrics, 1994, vol. 8, pp. 111–125. 24. Hasegawa, K., Miyashita, Y., and Funatsu, K., J. Chem. Inf. Comput. Sci., 1997, vol. 37, pp. 306–310.
RUSSIAN JOURNAL OF BIOORGANIC CHEMISTRY
Vol. 34
No. 4
2008
