This paper describes a novel protein, Itt1p, involved in the control of translation termination in yeast. Two lines of evidence support this conclusion: (i) overexpression of Itt1p enhances the readthrough of UAA, UAG and UGA nonsense codons; (ii) Itt1p interacts with both eRF1 and eRF3 polypeptide chain release factors. It is noteworthy that in contrast to deletions of SUP45 and SUP35 the knockout of ITT1 is not lethal and does not noticeably influence the nonsense codon readthrough. This suggests that Itt1p is not essential for the release of completed polypeptide chains from the ribosome.
The suppressor effect of Itt1p was observed at increased Itt1p to eRF1 ratio, but did not occur when this ratio was normal, including the case when both proteins were overexpressed. This suggests that eRF1 experiences quantitative, rather than qualitative alteration. The simplest explanation of these and other data is that the binding of Itt1p to eRF1 makes it inactive in translation termination. It is important that Itt1p is likely to be expressed at lower level than eRF1. According to published estimates [19
], the difference in CAI of eRF1 (0.334) and Itt1p (0.127) could mean that eRF1 is expressed at about 3-fold higher level than Itt1p. The role of eRF3 in the suppressor effect of Itt1p is less clear. The suppressor effect of excess Itt1p was not affected by overexpression of eRF3, but was reduced in the presence of N-terminally truncated eRF3, which does not interact with Itt1p. To explain this, it is possible to suggest that the interaction of Itt1p with eRF3 is weaker than with eRF1, but it strengthens the binding of Itt1p to eRF1/eRF3 complex, thus enhancing the suppressor effect of Itt1p. However, the effect of the N-terminally truncated eRF3 can also be explained by observation that it promotes translation termination better than complete protein [8
]. It may be difficult to distinguish these two mechanisms and we consider it likely that both take place.
It is not clear whether the inhibitory effect of Itt1p represents its main function, or whether it is a consequence of recruiting eRFs for some function different from the translation termination. The second opportunity looks more appealing, and it is supported by some structural features of Itt1p. Itt1p belongs to a unique family of zinc finger proteins called TRIAD. It contains three double zinc finger elements, one of which belongs to a RING class [17
]. The similarity of Itt1p with its homologues from other eukaryotes is not high (20–30%), but it spans the whole length of Itt1p with the cysteines and histidines of zinc finger elements being highly conserved. This suggests a functional similarity of Itt1p to at least some TRIAD proteins.
The alignment of Itt1p with its closest homologues (Figure ) and other TRIAD proteins (not shown) reveals notable similarity of the second and third double zinc finger elements, manifested in similar spacing of cysteines and some conserved residues. This is in contrast to the earlier assignment of the third finger as belonging to a RING class C3
]. One cause of this difference is that the histidine of the proposed RING signature is poorly conserved (marked $ in Figure ). On the other hand, highly conserved histidine and cysteine residues (marked 7 and 8) were disregarded previously [17
]. Intriguingly, in our version the third element contains an odd number of conserved residues, nine. Either one of the residues is unimportant for binding zinc (marked &), or there could be alternative configurations for zinc binding by this element.
Many of RING finger-containing proteins bind ubiquitin-conjugating enzymes and are the substrates for E2-dependent ubiquitination [20
]. It was proposed that this mechanism can be used to target the RING-containing protein or associated proteins for degradation in a regulated manner. If so, the excess Itt1p could cause suppression by accelerating degradation of eRF3 and eRFl. However, this is not the case, since the overexpression of Itt1p did not affect the levels of eRF3 and eRF1. It is also known that RING finger proteins can be transcriptional factors [21
] and nuclear localization was predicted for many of the TRIAD proteins [17
]. However, only few of these proteins were functionally characterized. Two of them are human androgen receptor activator ARA54 (Figure ) and rat protein kinase C-associated protein [22
]. For both proteins it may be suggested that they function in cytoplasm and nucleus and play a role in transcription. Itt1p may also be involved in transcription regulation since two-hybrid analysis has shown that it interacts with the Snf11p transcription factor [24
]. This, together with the fact that Itt1p contains a putative nuclear import signal (Figure ), allows to speculate that Itt1p may perform coupling of translation termination with transcription of certain genes.
Itt1p is not the only protein that could link translation termination with other cellular processes in yeast. At present, two such proteins are known. The first is Upf1p, which is involved in the control of NMD pathway and stimulates translation termination probably by binding to eRF3 and eRF1 release factors [13
]. Strikingly, this protein and its partners, Upt2p and Upf3p, are also required to control the total accumulation of large number of mRNAs in addition to their role in RNA surveillance, though mechanisms of such control are unknown [25
]. The second is Mtt1p, a homologue of Upf1p, which interacts with both release factors, but is not involved in the NMD control and inhibits translation termination [16
]. This protein has 5'→ 3' DNA-dependent helicase activity and is thought to be involved in chromosome replication [26
]. However, the question is still open, whether such proteins can mediate the interdependence of these cellular processes.