|Home | About | Journals | Submit | Contact Us | Français|
In a recent report published in this journal, Y. Qian and M.-X. Guan reported on their studies on interaction of aminoglycosides with human mitochondrial 12S rRNA carrying the A1555G deafness-associated mutation (15). Using RNA oligonucleotides representing the small subunit's rRNA decoding A site (i.e., the drug binding pocket) and carrying the A1555G mutation, the authors used chemical footprinting by dimethyl sulfate (DMS) to study the interaction of aminoglycosides with 12S rRNA. On the basis of their findings, the authors suggest that the A1555G mutation alters the binding properties of aminoglycosides at the A site of rRNA and leads to conformational changes in 12S rRNA with the A1555G mutation. I agree with the conclusion in general, which is also supported by other available evidence (6, 7). However, I am afraid that the authors' argumentation is somewhat compromised as it is only weakly supported by the experimental design and outcome.
(i) In the authors' footprinting studies, there is no correlation or any relationship between the defined chemical structure of an aminoglycoside compound and the drug-induced change in chemical modification for RNA oligonucleotides carrying the wild-type versus the deafness-associated sequence. This is surprising, as a vast body of evidence at the genetic, biochemical, and structural levels indicates that the interaction of aminoglycosides with the small-subunit rRNA can be categorized along the main structural elements of these compounds: i.e., (i) 4,5 versus 4,6 compounds and (ii) the 6′ OH versus 6′ NH2 substituent (1, 4, 8, 14).
Hamasaki and Rando are frequently referred to for having demonstrated specific binding of aminoglycosides to a human rRNA construct based on the A1555G polymorphism (5). Similar to Qian and Guan, these investigators made various RNA constructs prepared according to mitochondrial 12S rRNA and bacterial 16S rRNA sequences. It was found that the mitochondrial A1555G mutant construct stoichiometrically bound aminoglycosides, while the wild-type mitochondrial sequence did not bind aminoglycosides at all. However, what is frequently neglected is that an important control was at fault, putting the experimental design into doubt: i.e., a wild-type 16S rRNA bacterial construct bound aminoglycosides stoichiometrically, while a G1491A mutant bacterial 16S rRNA was unable to specifically bind aminoglycosides. The latter observation is at odds with the observation that the G1491A alteration only little affects aminoglycoside susceptibility in complete bacterial ribosomes (9, 10).
(ii) Streptomycin does not bind to the part of the A site rRNA used in Qian's and Guan's study (1). It is thus hard to understand that the authors observe streptomycin-induced changes in DMS reactivity for their RNA oligonucleotides similar to those of aminoglycosides targeting the A site (e.g., paromomycin and tobramycin).
(iii) As pointed out previously (6), there is a frequent misalignment of rRNA residues in the hearing field. This misperception presumably dates back to an early report by Hutchin et al. (12). Mitochondrial 12S rRNA position A1555 is not equivalent to bacterial 16S rRNA position 1491 but is homologous to that of bacterial 16S rRNA position 1490 (13) (Fig. (Fig.1).1). For the sake of the A1555G mutation, reference to studies of aminoglycoside susceptibility in bacterial ribosomes with alterations of the C1409-G1491 interaction (2, 3) is thus misleading. Both, the mitochondrial 12S rRNA wild type and A1555G mutants have a C-C opposition at corresponding E. coli positions 1409 to 1491 (Fig. (Fig.1).1). Investigations on the C1409-G1491 interaction in bacteria can hardly serve as a model for the mitochondrial 12S rRNA A1555G alteration.
In an effort to resolve the problems with mitochondrial 12S rRNA oligonucleotides and their interaction with aminoglycosides, we have recently used a genetic strategy to replace in bacterial ribosomes the small subunit's rRNA decoding A site in H44 (i.e., the aminoglycoside binding site) with that of various eukaryotic ribosomes (6, 7, 11). Toward this end, we constructed hybrid ribosomes in which the bacterial small ribosomal subunit's decoding A site has been replaced with the A site of wild-type mitochondrial ribosomes and with that of mitochondrial ribosomes carrying the A1555G deafness-associated mutation. Subsequent studies demonstrated that the A1555G deafness-associated mutation increased binding of aminoglycosides to the drug binding pocket and resulted in hypersusceptibility to aminoglycoside-induced miscoding. Finally, drug affinity and aminoglycoside susceptibility of the A1555G deafness-associated hybrid ribosomes were found to be similar to those of bacterial ribosomes, with a C1409-C1491 opposition corroborating the alignment given in Fig. Fig.11.
In response to the comments addressed by Dr. Erik Böttger, we are aware that he generally agreed with the conclusion that we made in our recent report. In particular, the works of Böttger's group have also shown that the deafness-associated A1555G mutation in human mitochondrial 12S rRNA altered binding properties with aminoglycoside antibiotics, even though the bacterial models for their experiments were not same as the human mitochondrial 12S rRNA. However, we do not agree with some of his comments about the experimental design and outcome in our study.
The A1555G and C1494T mutations in human mitochondrial 12S rRNA have been associated with both aminoglycoside-induced and nonsyndromic hearing loss in many families worldwide (1, 3, 8, 11, 16). In fact, the A1555G and C1494T mutations are located at the highly conserved A site of mitochondrial 12S rRNA (3, 4, 11, 16). In the wild-type version of the rRNA, A1555 and C1494 are in apposition to each other but do not form a base pair (3, 16). It is anticipated that the A1555G or C1494T mutation creates a new G-C or A-U base pair, making the human mitochondrial ribosome more like that of bacteria and, consequently, altering binding sites for aminoglycosides. The proposed structures of the A site of small rRNA between human mitochondria and bacteria including Escherichia coli and Mycobacterium smegmatis, as shown in Fig. Fig.1,1, are similar but not identical. It is worthwhile to note that there is a difference in the A site sequences of the 16S rRNA between E. coli and M. smegmatis that Dr. Böttger used in his studies (5-7). Thus, the structure and function of bacterial 16S rRNAs should not be fully applied to that of human mitochondrial 12S rRNA. Therefore, it is not appropriate to make a statement that we misaligned the human mitochondrial and bacterial small rRNA residues.
In the bacterial system, many lines of evidence showed the interaction between bacterial 16S rRNA and aminoglycosides via the main structural elements (2, 9, 15). The structure of the A site of bacterial 16S rRNA and their sites for the interaction with aminoglycosides were well known (14, 15). On the other hand, the structure of human mitochondrial 12S rRNA, especially in the A site spanning the A1555G and C1494T mutations, remains to be resolved. In our recent report, we focused on understanding the interaction of aminoglycosides with human mitochondrial 12S rRNA carrying the A1555G mutation by employing an RNA-directed chemical modification approach (12). Our data demonstrated that the A1555G mutation increased binding of six aminoglycosides to 12S rRNA, and these bound drugs caused conformation changes in the A site of 12S rRNA over its wild-type counterpart. In particular, the patterns of chemical modification in the RNA oligonucleotides carrying the A1555G mutation by dimethyl sulfate were distinct from those of the wild-type version in the presence of aminoglycosides. Unlike other studies (2, 14, 15), the RNA-directed chemical modification approach was not good enough to define the structure of the A site in 12S rRNA and determine how the interaction between bacterial 16S rRNA and aminoglycosides occurs via the main structural elements (10, 13).
Furthermore, we showed that in the RNA analogue carrying the A1555G mutation, the reduced reactivity toward DMS occurred in base G1555 as well as bases C1556 and A1553 in the presence of paromomycin, neomycin, gentamicin, kanamycin, tobramycin, and streptomycin. In particular, base G1555 exhibited marked but similar levels of protection in the presence of 0.1 μM to 100 μM neomycin, gentamicin, and kanamycin. In contrast, the levels of protection in base G1555 appeared to be correlated with the concentrations of paromycin, tobramycin, and steptomycin. In addition, the increasing reactivities toward these probes in the presence of these aminoglycosides were observed in bases A1492, C1493, C1494, and A1557 in the RNA analogue carrying the A1555G mutation. These data were not full agreement with those in the bacterial systems observed by other groups, including Dr. Böttger's (5, 13). Moreover, the reduced reactivity toward DMS in the RNA oligonucleotides carrying the A1555G mutation occurred in bases G1555, A1556, and A1553 in the presence of 0.1 μM, 1 μM, 10 μM, and 100 μM streptomycin. This correlates with the observation that subjects carrying the A1555G mutation experienced hearing loss after exposure to the conventional dose of streptomycin (1, 3, 8, 11, 16). Thus, human mitochondrial 12S rRNA carrying the A1555G, unlike the E. coli 16S rRNA, may interact with the streptomycin at the A site of this rRNA. However, further work needs to be done to test this hypothesis. In summary, despite evolutionary conservation of small ribosomal rRNAs among various species, there are some unique features in human mitochondrial 12S rRNA. Further studies need to be done to elucidate the molecular mechanism of aminoglycoside ototoxicity.