Despite their sequence similarity, it is clear from in vivo studies that Pif1 helicases do not have identical functions. Indeed, in the case of the two baker's yeast proteins, their functions appear to be non-overlapping, even though they affect the same DNA targets. Therefore, one possibility is that the Pif1 family sequence similarity reflects their acting at common DNA targets (). The three fungal enzymes are predicted to have both mitochondrial and nuclear forms, and all three proteins affect telomeres. However, the three proteins do not affect telomeres and mtDNA in the same way, and the effects of Rrm3p, ScPif1p and Pfh1p on chromosomal replication are not limited to telomeres.
The basis for the recruitment of Pif1-like helicase to common DNA targets is unclear. One can imagine that genomic loci like rDNA, tRNA genes, telomeres or mitochondrial DNA form particular DNA structures upon replication that need to be resolved by structure-specific helicases for replication to progress efficiently. Another possibility is that the sequence similarity reflects common mechanism(s) of action. The four Pif1 family members that have been characterized are all rather non-processive 5′–3′ DNA helicases. In addition, ScPif1p has the more unusual properties of preferentially unwinding RNA/DNA hybrids and of displacing the telomerase catalytic subunit from DNA (20
). The role of Rrm3p in moving replication past protein–DNA complexes suggests that it too might be able to displace proteins from DNA (39
). In addition, the involvement of Pif1 helicases in Okazaki fragment processing suggests that they could function in 5′ flap extension, possibly in combination with other helicases. It is possible that Pif1 family helicases share some or all these mechanistic properties. These alternative mechanistic models are summarized in .
The fact that Pif1 family members have different cellular function is not surprising. Although these proteins have >30% similarity in all pairwise combinations over an ~400 amino acid regions that contains the helicase motifs (6
), their N- and C-terminal regions are not similar in sequence or even size. For Rrm3p, the 249 N-terminus is essential for its in vivo
role in promoting DNA replication and also acts as a negative regulator of protein abundance (42
). In addition, the Rrm3p amino terminus contains a putative PCNA interaction motif and interacts with PCNA by two-hybrid criteria (43
). Thus, the N- and C-terminal regions of Pif1 family helicases may promote protein–protein interactions important to recruit the helicase to specific sites of action. These motifs might contribute to or even dictate the cellular functions of the helicases (51
). In addition, the helicase domains of Pif1 family members are similar but not identical, and these domains are also likely to contribute to the in vivo
specificity of the helicases. For example, a hybrid protein having the N-terminus and C-terminus of Rrm3p fused to the helicase domain of Pif1p does not provide Rrm3p function, even though the fusion protein is stably expressed (42
). The core helicase might contain determinants of substrate specificity or preference as well as requirements for loading and strand recognition, thereby specifying most of the biochemical properties that can be studied in vitro
. One clear objective is to develop in vitro
studies aimed at determining the substrate preferences of these helicases and their interacting partners. These types of analysis will allow us to link the numerous genetic observations made on the Pif1 helicases to their biochemical properties and to understand how, at the mechanistic level, these conserved helicases operate in the cell.