Title: DNA ligaseI is recruited to sites of DNA replication by an interaction with proliferating cell nuclear antigen: identification of a common targeting mechanism for the assembly of replication factories
Abstract: Article1 July 1998free access DNA ligase I is recruited to sites of DNA replication by an interaction with proliferating cell nuclear antigen: identification of a common targeting mechanism for the assembly of replication factories Alessandra Montecucco Alessandra Montecucco Istituto di Genetica Biochimica ed Evoluzionistica CNR, Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Search for more papers by this author Rossella Rossi Rossella Rossi Istituto di Genetica Biochimica ed Evoluzionistica CNR, Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Dipartimento di Genetica e Microbiologia 'A.Buzzati Traverso' Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Search for more papers by this author David S Levin David S Levin Department of Molecular Medicine, Institute of Biotechnology, The University of Texas Health Science Center at San Antonio, San Antonio, TX, 78245 USA Search for more papers by this author Ronald Gary Ronald Gary Life Sciences Division, Los Alamos National Laboratory, Los Alamos, NM, 87545 USA Search for more papers by this author Min S Park Min S Park Life Sciences Division, Los Alamos National Laboratory, Los Alamos, NM, 87545 USA Search for more papers by this author Teresa A Motycka Teresa A Motycka Department of Molecular Medicine, Institute of Biotechnology, The University of Texas Health Science Center at San Antonio, San Antonio, TX, 78245 USA Search for more papers by this author Giovanni Ciarrocchi Giovanni Ciarrocchi Istituto di Genetica Biochimica ed Evoluzionistica CNR, Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Search for more papers by this author Antonello Villa Antonello Villa Dipartimento di Farmacologia, CNR and B.Ceccarelli Centers and DIBIT Scientific Institute S.Raffaele, via Olgettina 60, Università di Milano, 20132 Milano, Italy Search for more papers by this author Giuseppe Biamonti Corresponding Author Giuseppe Biamonti Istituto di Genetica Biochimica ed Evoluzionistica CNR, Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Search for more papers by this author Alan E Tomkinson Alan E Tomkinson Department of Molecular Medicine, Institute of Biotechnology, The University of Texas Health Science Center at San Antonio, San Antonio, TX, 78245 USA Search for more papers by this author Alessandra Montecucco Alessandra Montecucco Istituto di Genetica Biochimica ed Evoluzionistica CNR, Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Search for more papers by this author Rossella Rossi Rossella Rossi Istituto di Genetica Biochimica ed Evoluzionistica CNR, Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Dipartimento di Genetica e Microbiologia 'A.Buzzati Traverso' Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Search for more papers by this author David S Levin David S Levin Department of Molecular Medicine, Institute of Biotechnology, The University of Texas Health Science Center at San Antonio, San Antonio, TX, 78245 USA Search for more papers by this author Ronald Gary Ronald Gary Life Sciences Division, Los Alamos National Laboratory, Los Alamos, NM, 87545 USA Search for more papers by this author Min S Park Min S Park Life Sciences Division, Los Alamos National Laboratory, Los Alamos, NM, 87545 USA Search for more papers by this author Teresa A Motycka Teresa A Motycka Department of Molecular Medicine, Institute of Biotechnology, The University of Texas Health Science Center at San Antonio, San Antonio, TX, 78245 USA Search for more papers by this author Giovanni Ciarrocchi Giovanni Ciarrocchi Istituto di Genetica Biochimica ed Evoluzionistica CNR, Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Search for more papers by this author Antonello Villa Antonello Villa Dipartimento di Farmacologia, CNR and B.Ceccarelli Centers and DIBIT Scientific Institute S.Raffaele, via Olgettina 60, Università di Milano, 20132 Milano, Italy Search for more papers by this author Giuseppe Biamonti Corresponding Author Giuseppe Biamonti Istituto di Genetica Biochimica ed Evoluzionistica CNR, Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy Search for more papers by this author Alan E Tomkinson Alan E Tomkinson Department of Molecular Medicine, Institute of Biotechnology, The University of Texas Health Science Center at San Antonio, San Antonio, TX, 78245 USA Search for more papers by this author Author Information Alessandra Montecucco1, Rossella Rossi1,2, David S Levin3, Ronald Gary4, Min S Park4, Teresa A Motycka3, Giovanni Ciarrocchi1, Antonello Villa5, Giuseppe Biamonti 1 and Alan E Tomkinson3 1Istituto di Genetica Biochimica ed Evoluzionistica CNR, Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy 2Dipartimento di Genetica e Microbiologia 'A.Buzzati Traverso' Università di Pavia, via Abbiategrasso 207, 27100 Pavia, Italy 3Department of Molecular Medicine, Institute of Biotechnology, The University of Texas Health Science Center at San Antonio, San Antonio, TX, 78245 USA 4Life Sciences Division, Los Alamos National Laboratory, Los Alamos, NM, 87545 USA 5Dipartimento di Farmacologia, CNR and B.Ceccarelli Centers and DIBIT Scientific Institute S.Raffaele, via Olgettina 60, Università di Milano, 20132 Milano, Italy *Corresponding author. E-mail: [email protected] The EMBO Journal (1998)17:3786-3795https://doi.org/10.1093/emboj/17.13.3786 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info In mammalian cells, DNA replication occurs at discrete nuclear sites termed replication factories. Here we demonstrate that DNA ligase I and the large subunit of replication factor C (RF-C p140) have a homologous sequence of ∼20 amino acids at their N-termini that functions as a replication factory targeting sequence (RFTS). This motif consists of two boxes: box 1 contains the sequence IxxFF whereas box 2 is rich in positively charged residues. N-terminal fragments of DNA ligase I and the RF-C large subunit that contain the RFTS both interact with proliferating cell nuclear antigen (PCNA) in vitro. Moreover, the RFTS of DNA ligase I and of the RF-C large subunit is necessary and sufficient for the interaction with PCNA. Both subnuclear targeting and PCNA binding by the DNA ligase I RFTS are abolished by replacement of the adjacent phenylalanine residues within box 1. Since sequences similar to the RFTS/PCNA-binding motif have been identified in other DNA replication enzymes and in p21CIP1/WAF1, we propose that, in addition to functioning as a DNA polymerase processivity factor, PCNA plays a central role in the recruitment and stable association of DNA replication proteins at replication factories. Introduction From bacteriophages to mammals, DNA replication is carried out by a complex multiprotein machinery. In mammalian cells, this process occurs at discrete sites on the nuclear matrix within a defined phase of the cell cycle. Progress in the study of the enzymology of mammalian DNA replication was greatly facilitated by the development of an in vitro assay in which bi-directional DNA replication is initiated at the SV40 replication origin in the presence of T antigen (Li and Kelly, 1984). Using this assay, mammalian replication proteins have been purified and the molecular mechanisms of DNA replication have been investigated in reconstitution experiments with replication protein A (RP-A), replication factor C (RF-C), proliferating cell nuclear antigen (PCNA), DNA polymerase δ, DNA polymerase α/primase, RNase H, FEN-1, DNA ligase I and SV40 T antigen (Waga and Stillman, 1994; Waga et al., 1994). It has been suggested that after origin firing, synthesis of the leading strand is initiated by DNA polymerase α/primase and then extended by DNA polymerase δ (or ϵ), which is tethered to the DNA by an interaction with a PCNA trimer that encircles the template and acts as a sliding clamp. In contrast, synthesis of the lagging strand requires the repeated sequential action of two DNA polymerases to generate the Okazaki fragments that subsequently are joined together. After the initial RNA–DNA fragment is laid down by DNA polymerase α/primase, it is extended up to the 5′ phosphate of the adjacent Okazaki fragment by a complex of DNA polymerase δ (or ϵ) and PCNA. The involvement of PCNA in both leading- and lagging-strand DNA synthesis and its interactions with the clamp loader RF-C (Fotedar et al., 1996; Cai et al., 1997; Uhlmann et al., 1997), DNA polymerase δ (Jonsson et al., 1995), FEN-1 (Li et al., 1995; Wu et al., 1996) and DNA ligase I (Levin et al., 1997) suggest that PCNA plays a central role in co-ordinating leading- and lagging-DNA synthesis at the replication fork. To ensure that the genome is duplicated faithfully once and only once per cell cycle, DNA replication is regulated by a complex set of control mechanisms. Different regions of the genome, identified by their characteristic higher order chromatin structure, are duplicated in a precisely defined temporal order during S phase (Pardue and Gall, 1970). This appears to reflect the sequential firing of different origins of replication, but these origins and their regulatory mechanisms are as yet poorly understood. When proliferating cells are pulse-labelled with bromodeoxyuridine (BrdU), discrete sites of DNA replication, termed replication foci, are detectable within the nucleus by indirect immunofluorescence with anti-BrdU antibodies (Nakayasu and Berezney, 1989; Hozak et al., 1994). The co-localization of DNA replication proteins and enzymes such as DNA polymerase α (Hozak et al., 1993), PCNA (Bravo and MacDonald Bravo, 1987; Hozak et al., 1993), DNA ligase I (Montecucco et al., 1995) and RP-A (Cardoso et al., 1993) with sites of BrdU incorporation has been demonstrated. In addition, other proteins such as cyclin A-dependent kinase (cdk) (Cardoso et al., 1993) and DNA-(cytosine-5)-methyltransferase (MCMT) (Leonhardt et al., 1992), which are involved in the regulation of DNA synthesis and methylation of newly synthesized DNA, respectively, also co-localize at the sites of DNA synthesis. Based on these kinds of observations, it has been proposed that, in either late G1 or early S phase, proteins involved in DNA replication are assembled at discrete sites termed replication factories (Hozak et al., 1993), whose intranuclear location is maintained by binding to an underlying nuclear matrix. A prediction of the model described above is that some, if not all, of the replication proteins should contain a sequence that allows them to be targeted to replication factories. Regions of MCMT (Leonhardt et al., 1992) and DNA ligase I (Montecucco et al., 1995; Cardoso et al., 1997) that appear to mediate the recruitment of these enzymes to replication factories have been identified. The N-terminal region of DNA ligase I (residues 1–115) is necessary for subnuclear targeting (Montecucco et al., 1995) but is not required for catalytic activity in vitro (Tomkinson et al., 1990). Interestingly, this same fragment of DNA ligase I has been shown recently to bind to PCNA (Levin et al., 1997). In this study, we have examined the relationship between the PCNA-binding and replication factory targeting functions that reside within the N-terminal 115 amino acids of DNA ligase I. We have identified a 20 amino acid sequence that is necessary and sufficient for interaction with PCNA in vitro and replication factory targeting in vivo. Since similar changes in this sequence disrupt both functions, it appears that the interaction of DNA ligase I with PCNA is required for the association of this enzyme with replication factories in S-phase cells. In addition, we have found that the N-terminus of the large subunit of RF-C, which is homologous to the 20 amino acid sequence of DNA ligase I, also functions as a replication factory targeting sequence (RFTS) and binds to PCNA. Results Identification and characterization of the DNA ligase I replication factory targeting sequence DNA ligase I is composed of a catalytic C-terminal domain (residues 217–919) and an N-terminal domain (residues 1–216) that is not required for enzyme activity but is essential for in vivo function (Tomkinson et al., 1990; Kodama et al., 1991; Petrini et al., 1995). In order to identify the protein determinants that are necessary and sufficient for the stable association of DNA ligase I with replication factories, a series of epitope-tagged deletion mutants were constructed as shown in Figure 1. After transient transfection of the recombinant plasmids into the human DNA ligase I-deficient fibroblast cell line 46BR.1G1 (Barnes et al., 1992), the expression and subnuclear localization of the epitope-tagged polypeptides were evaluated by indirect immunofluorescence with the HUC1.1 monoclonal antibody (mAb) that is specific for the epitope tag. To determine whether the tagged proteins co-localized with sites of DNA synthesis, the transfected cells were incubated with the nucleotide analogue BrdU prior to fixation and co-stained with anti-BrdU antibodies. In a typical experiment, ∼50% of the cells in the whole population were actively replicating their DNA as determined by BrdU incorporation. A similar proportion of cells expressing epitope-tagged polypeptides were in S phase. The behaviour of the epitope-tagged DNA ligase I polypeptides is summarized in Figure 1. Tagged proteins recruited to the replication factories co-localized with BrdU incorporation sites in at least 85% of the S-phase cells. Figure 1.Structure and subnuclear distribution of epitope-tagged versions of human DNA ligase I.Epitope-tagged full-length DNA ligase I (Lig-Tag wt) and versions of this polypeptide containing the indicated deletions were constructed as described in Materials and methods. All the deletions occur within the N-terminal domain (residues 1–216), while the catalytic domain (residues 217–919) is intact. The DNA ligase I NLS (residues 119–131), which is present in all the epitope-tagged polypeptides, is indicated by the arrowhead. The ability of the DNA ligase I polypeptides to associate with replication factories was assayed by indirect immunofluoresence microscopy and scored as + (targeting proficient) or − (targeting deficient). Download figure Download PowerPoint A nuclear localization signal (NLS), which is required for active transport into the nucleus through nuclear pores, has been identified in DNA ligase I at residues 119–131 (Montecucco et al., 1995). Deletion of all the residues in the N-terminal domain that extend beyond the NLS (Δ132–216 mutant) had no effect on the recruitment of DNA ligase I catalytic domain to replication factories (Figure 2A). Thus, the N-terminal 131 residues of DNA ligase I are sufficient for both nuclear localization and the specific association of the tagged catalytic domain with replication factories. Subnuclear targeting but not nuclear localization of the tagged polypeptide to the factories was abolished by deletion of residues 2–6 (Δ2–6 mutant) (Figure 2B). In contrast, the polypeptides encoded by the deletion mutants L1–30/NLS, L1–20/NLS (data not shown) and L1–11/NLS (Figure 2C), which contained only the first 30, 20 or 11 residues and the NLS (residues 119–131) from the N-terminal domain, were both located in the nucleus and correctly targeted to the sites of DNA synthesis. These results suggest that the first 11 residues of DNA ligase I are essential for the recruitment of this enzyme to replication factories. Figure 2.Intracellular localization of epitope-tagged DNA ligase I polypeptides by indirect immunofluorescence. Human fibroblasts were transfected with plasmids encoding the following epitope-tagged versions of DNA ligase I that are described in Figure 1: (A) residues 132–216 deleted (Δ132–216); (B) residues 2–6 deleted (Δ2–6); (C) residues 1–11 plus the NLS [L(1–11/NLS)] and then pulse-labelled with BrdU prior to methanol fixation. Cells were co-stained with HUC1.1 and FITC-conjugated anti-BrdU mAbs. Epitope-tagged DNA ligase I polypeptides were detected with a rhodamine-conjugated sheep anti-mouse secondary antibody. Antigen–antibody complexes were visualized by confocal laser scanning microscopy. For a better visualization of the ability of the tagged proteins to co-localize with sites of BrdU incorporation, mid-late S-phase nuclei (characterized by a reduced number of large BrdU foci) are shown. Download figure Download PowerPoint If the first 11 amino acids of DNA ligase I constitute an authentic RFTS, then this sequence should be necessary and sufficient to direct the recruitment of an unrelated polypeptide to sites of DNA replication. To test this, we constructed a plasmid in which the 11 amino acid putative RFTS was fused in-frame to the N-terminus of green fluorescent protein (GFP). Since GFP is small enough to enter the cell nucleus passively (Silver, 1991), we expected that the 11 residues would trigger GFP association with replication factories even in the absence of an NLS. Although this fusion protein was detectable in the nucleus of transfected cells, it was not recruited to sites of DNA synthesis (Figure 3A) and its subcellular distribution was indistinguishable from that of GFP alone (data not shown). However, association of GFP with replication factories was mediated efficiently by the first 20 residues of DNA ligase I as shown in Figure 3B. In S-phase cells, the polypeptide encoded by L(1–20)GFP was also found to co-localize with PCNA (data not shown), a previously identified component of replication factories (Bravo and MacDonald Bravo, 1987; Hozak et al., 1993). Figure 3.Identification of the DNA ligase I RFTS by immunofluorescence. Human fibroblasts were transfected with plasmids encoding the following fusion proteins: (A) residues 1–11 of DNA ligase I fused to GFP [L(1–11)GFP]; (B) residues 1–20 of DNA ligase I fused to GFP [L(1–20)GFP]; (C) residues 1–11 and the NLS of DNA ligase I fused to GFP [L(1–11/NLS)] and fixed 48 h later with 4% paraformaldehyde. The autofluorescent signal of GFP was visualized by confocal laser scanning microscopy. About 30% of cells expressing either L(1–20)GFP or L(1–11/NLS) showed a punctate nuclear pattern as exemplified in (B) and (C). The images shown in (A), (B) and (C) are equivalent exposures. (D) Cells transfected with the plasmid encoding residues 1–11 and the NLS of DNA ligase I fused to GFP [L(1–11/NLS)] were fixed and stained with anti-PCNA (PC10) mAb. PCNA was detected with a rhodamine-conjugated sheep anti-mouse Ig secondary antibody. The staining pattern of PCNA and the autofluorescent signal of GFP in the same cell are shown. Cells in S phase were identified by their punctate staining pattern after incubation with the PCNA antibody. In 80% of S-phase cells that were also expressing the GFP fusion protein, the distribution of PCNA and fusion protein was coincident. The absence of such a relationship in the remaining fraction of cells appeared to be due to the high level of expression of the fusion protein. (E) Human fibroblasts were transfected with a plasmid encoding an epitope-tagged version of DNA ligase I, in which residues 12–20 were deleted (Δ12–20), and 48 h later pulse-labelled with BrdU prior to methanol fixation. Cells were co-stained with HUC1.1 and FITC-conjugated anti-BrdU mAb. The epitope-tagged DNA ligase I polypeptide was detected with a rhodamine-conjugated sheep anti-mouse secondary antibody. Antigen–antibody complexes were visualized by confocal laser scanning microscopy. Download figure Download PowerPoint Residues 12–20 were not required for replication factory targeting of an epitope-tagged version of DNA ligase I, in which the first 11 amino acids, the NLS (residues 119–131) and the catalytic domain (residues 217–919) were fused (Figure 2C). In contrast, residues 12–20, in addition to the first 11 amino acids of DNA ligase I, were required to direct GFP to the sites of DNA synthesis (Figure 3A and B). Since residues 12–20 and the NLS of DNA ligase I both contain a high proportion of charged amino acids, we reasoned that the DNA ligase I NLS may be able to substitute for residues 12–20 if it is positioned adjacent to residues 1–11. To test this idea, we constructed a fusion protein with residues 1–11 of DNA ligase I followed immediately by the DNA ligase I NLS (residues 119–131) and GFP at the C-terminus. After transient transfection with this construct, a punctate nuclear pattern was observed in a subset of the transfected cells that was consistent with the association of the GFP fusion protein with the replication factories (Figure 3C). To confirm that this was the case, the transfected cells were stained with a mAb to PCNA. The co-localization of GFP protein with PCNA in S-phase cells (Figure 3D) demonstrates that the L(1–11/NLS)GFP fusion protein is targeted to DNA replication factories (Bravo and MacDonald Bravo, 1987; Hozak et al., 1993). To directly address the role of DNA ligase I residues 12–20 in replication factory targeting, we constructed an epitope-tagged version of full-length DNA ligase I with just these residues deleted. This version of DNA ligase I entered the nucleus but, in contrast to epitope-tagged full-length DNA ligase I, did not associate with replication factories (Figure 3E). Thus, residues 12–20 are not required for nuclear localization but constitute an integral part of the DNA ligase I RFTS that cannot be substituted by other sequences rich in basic residues, such as the DNA ligase I NLS, unless the positively charged sequence is immediately adjacent to residues 1–11. Based on these results, we conclude that the DNA ligase I RFTS is situated between residues 1 and 20, and suggest that the RFTS can be subdivided into two distinct regions: box 1 (residues 1–11) that is rich in hydrophobic amino acid residues and box 2 (residues 12–20) that is rich in positively charged amino acid residues (Figure 4A). Figure 4.Identification of an RFTS in the large subunit of RF-C: analysis of RFTS function by amino acid substitution. (A) Alignment of sequences at the N-terminus of DNA ligase I from vertebrates and at the N-terminus of the RF-C large subunit from eukaryotes. The positions of box 1 and box 2 within the RFTS are outlined. Conserved hydrophobic amino acids (in bold) are also indicated. Basic residues in box 2 are underlined. h, human; m, mouse; x, Xenopus laevis; d, Drosophila melanogaster; sc, Saccharomyces cerevisiae. Human fibroblasts were transfected with plasmids encoding the following fusion proteins; (B) residues 1–24 of human p140 RF-C subunit fused to GFP [R(1–24)GFP]; (C) residues 1–11 of human p140 RF-C subunit fused to GFP [R(1–11)GFP]; (D) residues 1–11 of human p140 RF-C subunit and the SV40 NLS fused to GFP [R(1–11/NLS)] and fixed 48 h later with 4% paraformaldehyde. The autofluorescent signal of GFP was detected by confocal laser scanning microscopy. About 30% of cells expressing either R(1–24)GFP or R(1–11/NLS) showed a punctate nuclear pattern as exemplified in (B) and (D). The images shown in (B), (C) and (D) are equivalent exposures. (E) Human fibroblasts were transfected with a plasmid encoding an epitope-tagged version of DNA ligase I, L(F/G8–9), in which the phenylalanine residues at positions 8 and 9 were replaced by glycine residues, and 48 h later pulse-labelled with BrdU prior to methanol fixation. Cells were co-stained with HUC1.1 and FITC-conjugated anti-BrdU mAb. The epitope-tagged DNA ligase I polypeptide was detected with a rhodamine-conjugated sheep anti-mouse secondary antibody. Antigen–antibody complexes were visualized by confocal laser scanning microscopy. Download figure Download PowerPoint In a recent study (Cardoso et al., 1997) it was reported that targeting of DNA ligase I to replication factories is mediated by a bipartite sequence composed of residues 1–28 and 111–179. Since these authors identified this bipartite sequence by its ability to direct an unrelated polypeptide that normally is excluded from the nucleus, it is conceivable that residues 111–179, which contain the previously described DNA ligase I NLS (119–131) (Montecucco et al., 1995), allow the fusion protein to enter the nucleus, whereas residues 1–28, which contain the RFTS defined in this study (1–20), direct the subnuclear localization of the fusion protein. A homologous RFTS at the N-terminus of the large subunit of replication factor C Having identified the DNA ligase I RFTS, we used this sequence to search the SwissProt data bank using the BLITZ program (Smith and Waterman, 1981). Although this search identified several proteins containing sequences exhibiting homology with box 1 of the DNA ligase I RFTS, only in the case of the large subunit of the RF-C was this sequence positioned at the N-terminus and flanked by a region rich in positively charged amino acids (Figure 4A). Next we sought to determine whether this motif, whose position and amino acid sequence are conserved among eukaryotic RF-C large subunits, functions as an RFTS. Plasmids expressing fusion proteins with either the first 11 (box 1 of the putative RFTS) or the first 24 residues (boxes 1 and 2 of the putative RFTS) of the human RF-C p140 subunit at the N-terminus of GFP were constructed. After transient transfection of the recombinant plasmids into the human fibroblast cell line 46BR.1G1, we observed that the GFP fusion protein with the intact putative RFTS was targeted efficiently to replication factories (Figure 4B) whereas the GFP fusion protein with only box 1 of the putative RFTS was not (Figure 4C). Interestingly, box 2 could be substituted for efficiently by PKKKRKV, the NLS of SV40 T antigen (Figure 4D), providing additional experimental evidence for the idea that the overall positive charge, not the primary sequence, of box 2 is critical for the function of the RFTS. These results are consistent with those obtained with the homologous RFTS of DNA ligase I and suggest that DNA ligase I and RF-C are targeted to replication factories by the same mechanism. From the alignment of the DNA ligase I RFTS and the RFTS of the RF-C large subunit, it is apparent that the sequence IxxFF within box 1 is the most conserved sequence within the targeting motif (Figure 4A). Presumably the conservation of the hydrophobic residues reflects the critical roles that these amino acids play in the function of the RFTS. The inactivation of the DNA ligase I RFTS by the deletion of residues 2–6 (Figure 2) is consistent with the notion that the conserved isoleucine residue is essential for RFTS function, in particular since residues corresponding to amino acids 3 and 4 of DNA ligase I are not present in the RFTS of the RF-C large subunit (Figure 4A). To investigate the role of the adjacent phenylalanine residues in the RFTS, we constructed a DNA ligase I mutant in which phenylalanines 8 and 9 were both substituted with glycine residues (F/G8–9). Replacement of the conserved phenylalanines with glycine residues resulted in no co-localization of the tagged DNA ligase I L(F/G8–9) polypeptide with sites of DNA replication (Figure 4E), demonstrating that the phenylalanine residues are essential for the function of the RFTS. The RFTS motif mediates the interaction of DNA ligase I with PCNA Previously we have shown that DNA ligase I binds to PCNA and that this interaction with trimeric PCNA, which is topologically linked to duplex DNA, tethers DNA ligase I to the DNA molecule (Levin et al., 1997). Since the binding to PCNA is mediated by residues within the N-terminal 118 amino acids of DNA ligase I (Levin et al., 1997), we have examined the relationship between the replication factory targeting and PCNA binding functions that reside within the same region of DNA ligase I. Using a pull-down assay, we found that PCNA bound specifically to GST fusion proteins containing either residues 1–118 (Figure 5A, lanes 4 and 9) or 1–19 (Figure 5A, lanes 2 and 7) of DNA ligase I. Since similar quantities of PCNA bound to the glutathione beads with either of the GST–DNA ligase I fusion proteins as the ligand, it appears that residues 20–118 of DNA ligase I do not contribute to the formation of a stable interaction between PCNA and DNA ligase I. Figure 5.The PCNA-binding activity of N-terminal fragments of DNA ligase I and the RF-C large subunit: analysis of PCNA binding by amino acid substitution. (A) Glutathione beads were incubated with cleared lysates from bacteria expressing GST (lane 1), a GST fusion protein containing residues 1–19 of human DNA ligase I (lane 2), GST–Lig I1–19 fusion with an F8A/F9A double amino acid substitution (lane 3), GST fusion protein containing residues 1–118 of human DNA ligase I (lane 4), or GST–Lig I1–118 fusion with an F8A/F9A double amino acid substitution (lane 5). An equal aliquot of lysate from bacteria expressing human PCNA was added to each assay. After washing, protein complexes were denatured and separated by SDS–PAGE. Polypeptides were detected by staining with Coomassie Blue (lanes 1–5) or by immunoblotting with PCNA antibody (lanes 6–10). The positions of PCNA (PCNA), GST (GST), GST–Lig I1–118(118) and GST–Lig I1–19 (19) are indicated. The additional bands in lanes 4 and 5 are probably due to degradation of the N-terminal fragment of DNA ligase I which is susceptible to proteolysis (Tomkinson et al., 1990). (B) PCNA (35 pmol) was pre-incubated with a peptide corresponding to the first 23 residues of human DNA ligase I (WT) or a version of this polypeptide in which the phenylalanine residues at positions 8 and 9 were replaced by alanine residues (M). The molar ratio of the peptide to trimeric PCNA is in