Second Generation Inhibitors

It is clearly too early to tell whether or not the aforementioned first-generation methyltransferase inhibitors will have chemotherapeutic value alone or in combination chemotherapy (29,31). Even so, two new directions in the inhibition of methyltransferases are being explored. Both first-generation inhibitors of DNA methyltransferase (32) and second-generation inhibitors of methyl-transferase action are under study (37-41). The use of second-generation inhibitors should avoid the side effects produced by the necessity of random incorporation into DNA required for decitabine action because second-generation inhibitors are DNA analogs that act as direct inhibitors of the methyltransferase (38,40,42,43).

Although the second-generation inhibitors do not use ordered protein arrays or bionanotechnology per se, they are potential components of the third-generation inhibitors, as described in Subheading 3.4. Thus, it is valuable to review their design properties.

3.3.1. Electronic Structure Activity Relations-Identified Targets and DNA Methyltransferase Inhibitor Design

Electronic structure activity relations (ESAR) have successfully predicted the effects of the mechanism-based inhibitors FdC (11), dU (44), and 2-pyrimidinone (45) on bacterial enzyme. Thus, it appears to be of general utility in determining the effects that a candidate inhibitor will have on methyltransferase action. An algorithm for analyzing candidate targets is depicted in Fig. 4. The results of calculations using the algorithm for a series of known and likely inhibitors of methyltransferases are depicted graphically in Fig. 5. In light of the known parameters of the human DNA (cytosine-5)methyltransferase, generally termed Dnmtl, the data indicate that the candidate second-generation inhibitors of the enzyme should have the following characteristics:

1. They should generally present a duplex region (that may contain mispairs) comprising about 28 bp of duplex in a hairpin formed of a single strand (46,47) or duplex formed from two fully or partially complementary oligodeoxynucleotides (48).

2. The partial duplex should contain nonproductive (44) or weakly productive three-nucleotide motifs (38,48). As defined in the cited references, a three-nucleotide recognition motif consists of a cytosine or preferably a 5-methylcytosine on one strand that is paired with a guanine or an inosine residue that is 3' to the targeted base. In short, this motif defines the targeted base. It is depicted in Fig. 3 (second-generation panel).

3. These nonproductive motifs should target the enzyme to a nonproductive or base-modified nucleotide in which the reference absolute value of the frontier orbital energy difference (see Fig. 4) for nucleophilic attack is >0.73797 eV, or >7.39287 eV for methyltransfer based on the data given in (43).

The hairpin molecules described in refs. 38, 41, 46, and 49 satisfy these criteria with frontier orbital energy differences for methyl transfer exceeding the requirement in point no. 3 for a normally paired nucleotide analog.

3.3.2. Multiply Mistargeted Bases

Saturation curves for a productive substrate (dCS1) and a multiply mistargeted nonproductive substrate (NdCS2) are depicted in Fig. 5. The sequences are as follows:

1. dCS1: CG-methylated oligo targeting normally paired dC

Oligo CG-methyl 5' GTC-CAC-CAG-ATC-CGG-GCT-ACC-TGG-CCT-CGA 3' Oligo no-methyl 3! CAG-GTG-GTC-TAG-GCC-CGA-TGG-ACC-GGA-GCT 5'

2. NdCS2: non-CG-methylated oligo targeting normally paired dC, dT, dA, and dG

Oligo non-CG-methyl 5' GTC-CAC-CAG-ATC-CGG-GCT-ACC-TGG-CCT-CGA 3' Oligo no-methyl 3' CAG-GTG-GTC-TAG-GCC-CGA-TGG-ACC-GGA-GCT 5'

in which x indicates a mistargeted base, * indicates an appropriately targeted base, and m indicates a targeting methyl group.

Since previous work with substrate dCS1 (50,51) has shown that tritium incorporation is confined almost exclusively to the methyl-targeted cytosine (*), tritium incorporation into substrate NdCS2is likely to be directed to the two productive sites in the substrate or other cytosine residues in the oligo-deoxynucleotide. Based on the ESAR, 10 of the 12 targeted sites will produce nonproductive binding of the enzyme to sites targeting dT, dG, or dA where LUMO energies exceed that required for nucleophilic attack. The human enzyme has a footprint in the range of 28 bp (48), suggesting that it would generally bind only once to an oligodeoxynucleotide of this type with a probability of 10/12 (0.833) of binding to a nonproductive site. Thus, the low Vm and low Km observed for this substrate strongly suggest that the bulk of the interaction with the substrate is at nonproductively targeted dA, dG, and dT sites. Uniform methylation also raises the Tm of the oligodeoxynucleotide from 74 to 80°C (data not shown), suggesting that the enzyme may also be able to bind at methylated targeting sites but be slowed by the additional energy required to disrupt the base-stacking interactions as the bases are flipped out of the helix.

Given that the kinetics monitor tritium incorporation into either substrate, the kinetics are described by the reaction scheme given in Fig. 5A, left. The general equations describing this reaction (52) are as follows:

v = (Vm1[S1]/Km1 + Vm2[S2]/Km2)/(1 + [S1]/Km1 + [S2]/Km2) (1)

CO Qo

CO Qo

Fig. 4. ESAR for second-generation methyltransferase inhibitors. (A) ESAR algorithm for deciding whether nucleophilic attack or methyltransfer will be affected in inhibition. (B) Frontier orbital picture comparing models of targets that permit nucleophilic attack at a cytosine model compound (acceptor attacked) with models of targets that do not permit nucleophilic attack (acceptor not attacked): (I) 1-methyl-5-pyrimidinone cation (N3 protonated); (II) 1-methyl-5-fluorocytosine cation (N3 protonated); (III) 1-methyl-5-azacytosine cation (N3 protonated); (IV) 1-methylcytosine cation (N3 protonated); (V) 1-methyl-4-thiouracil; (VI) 1-methyl-3-bromouracil; (VII) 1-methyl-2-pyrimidinone; (VIII) 1-methyl-5-fluorouracil; (IX) 1-methyluracil; (X) 1-methylpseudouracil (y); (XI) 1-methylthymine; (XII) 1-methyl-5-azacytosine; (XIII) 1-methylcytosine; (XIV) 9-methyl-8-oxoguanine; (XV) 9-methyladenine; (XVI) 9-methylguanine; (XVII) 9-methyl-7-deazaguanine; (XXII) N3, N5 dimethyl cysteine thiolate anion.

Fig. 4. ESAR for second-generation methyltransferase inhibitors. (A) ESAR algorithm for deciding whether nucleophilic attack or methyltransfer will be affected in inhibition. (B) Frontier orbital picture comparing models of targets that permit nucleophilic attack at a cytosine model compound (acceptor attacked) with models of targets that do not permit nucleophilic attack (acceptor not attacked): (I) 1-methyl-5-pyrimidinone cation (N3 protonated); (II) 1-methyl-5-fluorocytosine cation (N3 protonated); (III) 1-methyl-5-azacytosine cation (N3 protonated); (IV) 1-methylcytosine cation (N3 protonated); (V) 1-methyl-4-thiouracil; (VI) 1-methyl-3-bromouracil; (VII) 1-methyl-2-pyrimidinone; (VIII) 1-methyl-5-fluorouracil; (IX) 1-methyluracil; (X) 1-methylpseudouracil (y); (XI) 1-methylthymine; (XII) 1-methyl-5-azacytosine; (XIII) 1-methylcytosine; (XIV) 9-methyl-8-oxoguanine; (XV) 9-methyladenine; (XVI) 9-methylguanine; (XVII) 9-methyl-7-deazaguanine; (XXII) N3, N5 dimethyl cysteine thiolate anion.

CO CO

Fig. 4. (continued)

se o

Fig. 5.

3.3.3. Singly Mistargeted Mispaired Bases

Although the methyltransferase reaction involves both AdoMet and DNA as substrates, at constant AdoMet concentrations, the DNA saturation kinetics are hyperbolic for either the highly productive or weakly productive substrate. When a constant concentration of the productive substrate is present and varying amounts of the weakly productive substrate are added, inhibition by the weakly productive substrate follows the kinetics predicted for a two-DNA substrate system. In many cases, incorporation into weakly productive competing substrates is very low. This makes it difficult to estimate Km2 from direct saturation experiments. In these cases, we have found it useful to arrive at an upper limit for Vm2 and determine an average value for Km2 from Eq. 2.

In general, we have noted that multiply mistargeted second-generation inhibitors are more potent than singly mistargeted inhibitors based on side-by-side comparisons. An inhibition curve for the multiply mistargeted inhibitor NdCS2 is provided in Fig. 5B.

Was this article helpful?

0 0

Post a comment