Multiple Mitochondrial DNA Polymerases in Trypanosoma brucei

Share Embed


Descrição do Produto

Molecular Cell, Vol. 10, 175–186, July, 2002, Copyright 2002 by Cell Press

Multiple Mitochondrial DNA Polymerases in Trypanosoma brucei Michele M. Klingbeil, Shawn A. Motyka, and Paul T. Englund1 Department of Biological Chemistry Johns Hopkins Medical School Baltimore, Maryland 21205

Summary Kinetoplast DNA (kDNA), the unusual mitochondrial DNA of Trypanosoma brucei, is a network containing thousands of catenated circles. Database searching for a kDNA replicative polymerase (pol) revealed no mitochondrial pol ␥ homolog. Instead, we identified four proteins (TbPOLIA, IB, IC, and ID) related to bacterial pol I. Remarkably, all four localized to the mitochondrion. TbPOLIB and TbPOLIC localized beside the kDNA where replication occurs, and their knockdown by RNA interference caused kDNA network shrinkage. Furthermore, silencing of TbPOLIC caused loss of both minicircles and maxicircles and accumulation of minicircle replication intermediates, consistent with a role in replication. While typical mitochondria contain one DNA polymerase, pol ␥, trypanosome mitochondria contain five such enzymes, including the previously characterized pol ␤. Introduction Kinetoplastid protozoa include parasites that cause tropical diseases. They are also important as one of the earliest diverging eukaryotic groups that contain a mitochondrion (Sogin and Silberman, 1998). Given their phylogenetic distance from higher eukaryotes, it is not surprising that kinetoplastids display unusual biological properties. One example is their unique mitochondrial DNA, known as kinetoplast DNA (kDNA), which is composed of several thousand DNA circles catenated into a single network. The network is condensed into a diskshaped structure in the matrix of the cell’s single mitochondrion near the flagellar basal body. In Trypanosoma brucei, the parasite causing African sleeping sickness, the kDNA network contains several dozen maxicircles (23 kb) and several thousand minicircles (1 kb). Maxicircles resemble other mitochondrial DNAs in that they encode rRNAs and a few proteins including subunits of respiratory complexes. Minicircles encode guide RNAs that control editing of maxicircle transcripts, a process involving insertion or deletion of uridine residues to form an open reading frame (Estevez and Simpson, 1999). The structural complexity of kDNA dictates an unusual replication mechanism (reviewed in Klingbeil et al., 2001). Much of our knowledge regarding the replication process has come from studying kDNA replication proteins purified from a related parasite, Crithidia fasciculata. These proteins include universal minicircle sequence binding protein (UMSBP, which binds the minicircle replication origin) (Tzfati et al., 1995), primase 1

Correspondence: [email protected]

(Li and Englund, 1997), topoisomerase II (topo II) (Melendy and Ray, 1989), and structure-specific endonuclease (SSE1, which has RNase H activity) (Engel and Ray, 1998). Surprisingly, the only known kinetoplastid mitochondrial DNA polymerase (pol) is a pol ␤. This abundant enzyme is 33% identical to mammalian pol ␤ (Torri and Englund, 1995), an enzyme involved in nuclear base excision repair. The C. fasciculata enzyme was the first example of a pol ␤ found in mitochondria and is thought to play a role in gap filling during minicircle replication (Torri et al., 1994). No proofreading replicative mitochondrial polymerase has been identified from kinetoplastids or any other protozoan. A notable feature of trypanosomes is that their mitochondrial replication proteins are situated in specific regions surrounding the kDNA disk. Topo II (Melendy et al., 1988), pol ␤ (Ferguson et al., 1992), and SSE1 (Engel and Ray, 1999) localize in two antipodal sites flanking the kDNA disk. In contrast, UMSBP (Abu-Elneel et al., 2001) and primase (Li and Englund, 1997) are found in the kinetoflagellar zone (KFZ), a region of the mitochondrial matrix between the face of the kDNA disk and the membrane nearest the flagellar basal body. The KFZ is important because kDNA replication involves the vectorial release of minicircles from the network into this zone where DNA synthesis initiates (Drew and Englund, 2001). Along with UMSBP and primase, a replicative polymerase for minicircles is likely to be found in this region. As replication proceeds (and may even be completed) in the KFZ, the advanced replication intermediates (or progeny molecules) migrate to the antipodal sites where SSE1 and pol ␤ presumably remove primers and fill most of the gaps. The progeny are then attached to the network periphery by topo II (Wang and Englund, 2001). We expected that a kDNA replicative pol would be related to pol ␥, the only one of at least 19 eukaryotic DNA pols implicated in mitochondrial replication (Hu¨bscher et al., 2002). Pol ␥ is a family A DNA polymerase, a group including bacterial pol I, the phage T7 polymerase, and eukaryotic pol ␪. Although these proteins have diverse roles such as gap filling and primer removal, replication of mitochondrial and phage DNA, and DNA repair, all contain the conserved family signature, the pol A domain, in the C-terminal region. We made major efforts to identify a kDNA replicative polymerase in trypanosomes. Initial attempts, involving fractionation of C. fasciculata mitochondrial extracts or PCR using degenerate primers, were unsuccessful in identifying a putative pol ␥. We then turned to a genomics approach, searching for candidate pol genes in the T. brucei genome databases. Searches for a pol ␥ homolog revealed no candidates, consistent with our PCR results. Instead, we found multiple sequence fragments related to bacterial pol I. This observation led to our characterization of a pol I-like gene family. T. brucei has four genes that are members of this family, and remarkably all four produce proteins that localize to the mitochondrion. As for the functions of the multiple mitochondrial polymerases, we provide genetic evidence that at least one, which we showed to be enzymatically

Molecular Cell 176

Figure 1. Domain Organization of Selected Family A DNA Polymerases Domains were identified using the Pfam database (accession numbers are in parentheses). POL A, family A pol domain (PF00476); 5⬘ EXO, 5⬘ exonuclease (PF02739); 3⬘ EXO (Ellipse), 3⬘ exonuclease (PF01612); 3⬘ EXO (Hexagon), 3⬘ exonuclease (PF00929); Helicase, DEAX box helicase domain (PF00270); C, helicase C-terminal domain (PF00271). Sequences are grouped by ascribed and presumed functions. Abbreviations: T5, Bacteriophage T5; Ec, Escherichia coli; Dm, Drosophila melanogaster; Hs, Homo sapiens; Tp, Treponema pallidum; and Tb, T. brucei.

active, is involved in kDNA replication. Together with the previously characterized pol ␤, there are at least five DNA polymerases in trypanosome mitochondria. Results Identification of a DNA Polymerase I-like Gene Family in T. brucei Family A DNA polymerases, with Escherichia coli pol I as the prototype, have a conserved pol domain near the C terminus (Figure 1). Taking advantage of the nearly complete sequence information in the T. brucei genome databases, we searched for sequences related to the mitochondrial replicative pol ␥ and other family A members. A TBLASTN search did not identify a pol ␥ homolog. However, when using the E. coli pol I sequence as a query, we unexpectedly found four different groups of sequence fragments that shared significant similarity with bacterial pol I polymerase domains (E ⫽ 1.4 ⫻ 10⫺3 to 8.1 ⫻ 10⫺13). To sequence the pol I-like genes, we screened a T. brucei 927 genomic P1 bacteriophage library. For each gene, we obtained an average of six clones, consistent with the estimated 5-fold coverage of the P1 library. We selected a P1 clone for each gene, and sequencing revealed large ORFs without introns (TbPOLIA, 2874 bp; TbPOLIB, 4212 bp; TbPOLIC, 4947 bp; and TbPOLID, 4887 bp). These genes encode predicted proteins with molecular masses of 104.9, 158.7, 182.0, and 182.8 kDa, respectively. Each gene is expressed in procyclic T. brucei with transcript sizes large enough to accommodate the ORFs (see below, Figure 5). The GenBank accession numbers are AF445376, AF445377, AF445378, and AF445379, respectively. The Family A Polymerase Domain Sequence alignment of the four candidate T. brucei polymerases with other family A members indicated signifi-

cant similarity only in the C-terminal pol domain (the T. brucei pol domains are 25%–35% identical to their prokaryotic counterparts). The pol domain of TbPOLIA is most similar to that of Thermus thermophilus, while the pol domains of TbPOLIB, -IC, and -ID are most similar to those of Treponema and Sepulina. All the T. brucei pol domains contain the A, B, and C motifs (Sousa, 1996); see Figure 2A for an alignment of these motifs from selected family A members. Several amino acid residues are conserved among all family A members, including the four T. brucei sequences (Figure 2A). These residues, which have been implicated in polymerase function based on mutational analyses of E. coli and T. aquaticus pol I, include: Asp and Glu in motif A; Arg, Lys, Tyr, and Gly in motif B; and His and Asp in motif C. The two conserved Asp residues (motifs A and C) coordinate two Mg2⫹ ions that stabilize the transition state and facilitate phosphoryl transfer (Steitz, 1998). The conserved Arg and Lys in motif B interact with the phosphates of the incoming dNTP (Patel et al., 2001), and a single Glu residue (motif A) is involved in the selectivity of dNTPs over rNTPs (Astatke et al., 1998). The conservation of these essential residues in the trypanosome sequences suggests that they are functional polymerases. Phylogenetic Analysis of Family A DNA Polymerases As one of the earliest diverging eukaryotes with a mitochondrion, T. brucei occupies a unique position in evolution. The evolutionary divergence of trypanosomes from higher eukaryotes and the probable absence of a pol ␥ from their genome suggested that at least one of the T. brucei pol I-like genes could be ancestral to pol ␥ of higher eukaryotes. If so, a pol I-like protein could be a kDNA replicative polymerase. To test this hypothesis, we conducted phylogenetic analyses to determine the evolutionary relationship among the T. brucei and other

Mitochondrial DNA Polymerases in Trypanosomes 177

Figure 2. Relationship of Family A DNA Polymerases (A) Alignment of conserved sequence motifs A, B, and C of selected family A DNA polymerases. Identical residues are shaded black, and similar residues are shaded gray where 12/18 residues are identical or similar. (B) Unrooted maximum parsimony phylogenetic tree of the family A pol domain. For clarity, some closely related pol I sequences were omitted from the tree. Numbers indicate bootstrap values at internal nodes calculated from 1000 random replica. Functions are superimposed on the phylogenetic tree. Nodes a and b represent alternative placements of TbPOLIB, -IC, and -ID using PUZZLE. Abbreviations are defined in Figure 1 and as follows: Aa, Aquifex aeolicus; Bs, Bacillus stearothermophilus; Ce, Caenorhabditis elegans; Mt, Mycobacterium tuberculosis; Rm, Rhodothermus marinus; Rp, Rickettsia prowazekii; Ta, Thermus aquaticus; Sc, Saccharomyces cerevisiae; Ss, Synechocystis species; and Xl, Xenopus laevis. Accession numbers: Aa pol I (AE000765); Bs pol I (P52026); Ce pol ␪ (U50184); Dm pol ␪ (L76559); Dm pol ␥ (U62547); Ec pol I (AE000461); Hs pol ␪ (NM_006596); Hs pol ␥ (P54098); Mt pol I (Q07700); Rm pol I (AAD28505); Rp pol I (AJ235273); Sc pol ␥ (P15807); Ss pol I (BBA10748); Ta pol I (P19821); Tp pol I (AE001195); and Xl pol ␥ (S68258).

family A DNA polymerases. Our analyses expanded on previous work of this type (Braithwaite and Ito, 1993). Because of the multidomain structure of family A members, we used only the conserved pol domain for phylogenetic analyses. Phylogenetic reconstructions revealed three clearly defined clades of bacterial pol I, replicative mitochondrial pol ␥, and repair pol ␪ with similar tree topologies using both maximum parsimony and PUZZLE analyses. The placement of the T. brucei sequences was not monophyletic. For example, using parsimony (Figure 2B), TbPOLIA grouped with the pol ␪ subfamily and was basal to the other eukaryotic sequences. The three other T. brucei paralogs formed a group basal to the pol ␥ sequences with high bootstrap values. There are limits to our interpretation of these data. For example, long branch attraction can cause misinterpretation of maximum parsimony results (Felsenstein, 1978; Huelsenbeck, 1997). In addition, although our PUZZLE analysis revealed a placement of the three T. brucei paralogs at node a (Figure 2B), there was also an alternative placement at node b (Figure 2B), within the repair clade. Despite these limitations, our data are consistent with the possibility that TbPOLIB, -IC, or -ID could be mitochondrial kDNA replicative proteins.

Other Kinetoplastid Organisms Contain Multiple Pol I-like Genes Searching other parasite genome databases for pol I-like sequences revealed a similar gene family in Leishmania major and possibly in T. cruzi (for which the database is much less complete). Three large ORFs from L. major (CAC37107, CAB91834, and CAC04270) exhibited 48%, 49%, and 63% sequence identity with TbPOLIB, -IC, and -ID, respectively. A fourth incomplete sequence, AL449623, exhibited 52% identity to the C-terminal 2/3 of TbPOLIA. Sequence fragments of putative homologs for TbPOLIA, -IC, and -ID were also detected in the T. cruzi EST and genome databases. Of the four pol I-like sequences in T. brucei and L. major, only POLIB and -ID contained a predicted multidomain structure. In addition to the conserved pol domain, both contain a putative 3⬘ exonuclease domain (Figure 1). Although the POLID exonuclease domain shares similarity with those of pol I and pol ␥, the POLIB exonuclease domain is markedly different. The domain structure of POLIB will be discussed elsewhere. Four Pol I-like Proteins Target to the Mitochondrion Most mitochondrial proteins are nuclear encoded, and their subsequent import is usually directed by N-terminal

Molecular Cell 178

Figure 3. Mitochondrial Targeting of the Four Pol I-like Proteins (A) N-terminal sequences of selected kinetoplastid nuclear-encoded mitochondrial proteins. Capital letters indicate known cleaved presequences. Abbreviations: Cf, Crithidia fasciulata; Lt, Leishmania tarentolae; LipDH, dihydrolipoamide dehydrogenase; MP48, mitochondrial RNA ligase. (B) Stably transfected cell populations expressing IA-GFP, IB-GFP, IC-GFP, and IDGFP were established, and fluorescence of DAPI and the corresponding fusion proteins were detected in live parasites. k, kDNA disk; n, nucleus. In control experiments (data not shown), fluorescence of free GFP was distributed throughout the cytoplasm and nucleus, and La9-GFP fusion protein localized exclusively to the nucleus (Marchetti et al., 2000). (C) Live T. brucei procyclic cells (YTat1.1) labeled with MitoTracker Red. (B and C) Cells were incubated with 5 ␮g/ml DAPI or 1 ␮M MitoTracker Red (Molecular Probes) for 30 min, collected by low speed centrifugation, washed, and resuspended in Cytomix at 108 cells/ml. Due to rapid cell movement and fading of Mitotracker fluorescence, it was impossible to obtain merged GFP/DAPI/Mitotracker images. Bar, 4 ␮m.

targeting sequences. Some kinetoplastid targeting sequences are unusually short (often 9 amino acids) (Hausler et al., 1997), while others resemble mitochondrial targeting sequences in other eukaryotes (Neupert, 1997). TbPOLIC and -ID have three to four residues at the N terminus similar to those of known kinetoplastid targeting sequences (Figure 3A), and the algorithm PSORT II predicted mitochondrial targeting for both. To investigate the subcellular localization of these proteins, we constructed C-terminal GFP fusions for all four ORFs using the expression vector pXSGFPM3FUS (Marchetti

et al., 2000). All four transgenic cell lines had doubling times similar to that of the parental cell line (9 hr), indicating that fusion protein overexpression did not adversely affect cell growth. Remarkably, fluorescence microscopy of live cells stably expressing IA-GFP, IB-GFP, ICGFP, and ID-GFP all revealed a fluorescence pattern typical of the trypanosome branched tubular mitochondrion (Figure 3B). The same localization pattern was observed after staining live wild-type cells with Mitotracker Red (Figure 3C). These data indicate that all four pol I-like proteins are mitochondrial.

Mitochondrial DNA Polymerases in Trypanosomes 179

Figure 4. Immunolocalization of T. brucei Pol I-like Proteins (A) Colocalization of TbPOLIA and LipDH (a mitochondrial matrix marker). (B) Immunofluorescence of TbPOLIB. (C) Immunofluorescence of TbPOLIC. (D) Colocalization of TbPOLID and LipDH. (E) Enlargement of one cell from merged image in (B). (F) Enlargement of one cell from merged image in (C). For (B) and (C), the immunofluorescence appears on the flagellar side of the kDNA disk as determined by merging fluorescence and phase images. k, kDNA disk; n, nucleus. Bar, 2.5 ␮m.

Immunolocalization of the Pol I-like Proteins To confirm the GFP localization and to determine whether the endogenous pol I-like proteins are positioned near the kDNA disk (like all other known kDNA replication proteins), we performed immunofluorescence experiments using antibodies generated against C-terminal peptides. Immunofluorescence using TbPOLIA- and TbPOLID-specific antibodies revealed uniform staining of the branched tubular mitochondrion as observed with the GFP fusion proteins, suggesting that both proteins are distributed throughout the mitochondrial matrix. As expected, they colocalized with the mitochondrial matrix enzyme lipoamide dehydrogenase (LipDH) (Figures 4A and 4D). The localization of TbPOLIB and -IC was different. Instead of uniform mitochondrial staining, most was in

one or two spots localized to the flagellar side of the kDNA disk (Figures 4B, 4C, 4E, and 4F). This region is the kinetoflagellar zone where minicircle replication initiates. POLIB localization always appeared as two distinct foci. However, the localization of TbPOLIC usually appeared as a single elongated zone or as two barely resolved spots adjacent to the kDNA disk. The fact that the corresponding GFP fusion proteins localized throughout the mitochondrial matrix is likely due to overexpression. RNA Interference Studies RNA interference (RNAi) is a powerful method for assessing gene function. Our laboratory developed an inducible RNAi system in T. brucei that allows study of potentially lethal phenotypes (Wang et al., 2000). With

Molecular Cell 180

Figure 5. Effect of RNAi (A) Growth curves of clonal RNAi cell lines were determined by culturing the cells in the absence (solid squares) or presence (open circles) of 1 ␮g/ml tetracycline. Cell density is plotted as the product of cell number and total dilution. Insets: Northern analysis of mRNA levels from uninduced (⫺) and induced cells expressing dsRNA for 40 hr (⫹). Sizes of the mRNAs are: TbPOLIA, 4.2 kb; TbPOLIB, 4.4 kb; TbPOLIC, 6.1 kb; and TbPOLID, 5.3 kb. Percent reduction of mRNA levels induced by RNAi was estimated by densitometry of autoradiograms. The dsRNA was detected on all the Northerns except for that of TbPOLID (data not shown). (B and C) Cells were treated with tetracycline for the indicated days of RNAi induction, washed, and resuspended in Cytomix at 108 cells/ml. More than 200 cells at each time point were scored by eye for the size of the DAPI-stained kDNA disk. (B) Effect of TbPOLIB RNAi on kDNA morphology. Solid circles, normal appearing kDNA; solid squares, abnormally small kDNA; and open triangles, kDNA apparently composed of multiple DAPI-staining spots. (C) Effect of TbPOLIC RNAi on kDNA morphology. Solid circles, normal appearing kDNA; solid squares, abnormally small kDNA; open circles, no kDNA; and open triangles, multiple DAPI-staining spots. (D) Examples of kDNA morphology induced by TbPOLIC RNAi. Cells with multiple DAPI-staining spots on occasion showed a morphology similar to that of ancillary kDNA (not shown) (Miyahira and Dvorak, 1994).

this method, a fragment of the gene of interest is inserted into pZJM, a vector that produces dsRNA upon induction with tetracycline. We found that RNAi of TbPOLIA led to ⬎95% reduction in TbPOLIA mRNA after 40 hr without any subsequent effect on trypanosome growth, raising the possibility that TbPOLIA is not essential under normal growth conditions (Figure 5A). There was also no growth phenotype after induction of TbPOLID dsRNA, but since there was no effect on the TbPOLID mRNA level, we can draw no conclusions about the function of this protein (Figure 5A). In studies of other T. brucei genes using pZJM, we found examples where the mRNA level was not reduced after tetracycline induction (Wang et al., 2000 and our unpublished data).

In contrast, induction of TbPOLIB and -IC dsRNA led to a ⬎95% and 85% loss of target mRNA within 40 hr, respectively, and inhibition of cell growth within 7 days (Figure 5A). In the case of TbPOLIB, the decline in cell growth was accompanied by changes in the size and morphology of the kDNA network as visualized by DAPI staining. By day 10, about 65% of the cells contained normal-sized kDNA, but 20% contained abnormally small kDNA or multiple DAPI-staining spots (Figure 5B). The remaining 15% of the cell population showed unequal kDNA division and other abnormal kDNA morphology (data not shown). The kDNA phenotype that accompanied TbPOLIC RNAi was more striking. As early as 3 days after induction, only 55% of the cell population still

Mitochondrial DNA Polymerases in Trypanosomes 181

contained normal-sized kDNA, while the remaining cells contained abnormally small kDNA (30%), no kDNA (5%), or multiple DAPI-staining spots heterogeneous in size (10%) (Figure 5C). The largest spot occupied the characteristic location of kDNA with smaller spots appearing to have fragmented from the kDNA (Figure 5D). In some cases, a smaller spot was detected on the opposite side of the nucleus (anterior end of the cell) (data not shown). After 8 days of RNAi, ⬎65% of the cells contained small kDNA, and at day 15 only small DAPI staining spots, if any, were detected. TbPOLIC RNAi Affects Minicircle Replication Although the shrinking kDNA phenotype suggested a defect in kDNA replication, another possibility existed. The networks could simply be decatenated, with minicircles and maxicircles dispersed throughout the mitochondrion and therefore difficult to detect by DAPI staining. However, if the phenotype were due to a replication block, we would predict a progressive decrease in the minicircle and maxicircle copy number. To distinguish between these possibilities, we induced TbPOLIC RNAi, purified total DNA at different times, and analyzed samples for minicircles, maxicircles, and a nuclear gene by dot-blotting. We found a decrease in copy number of both minicircles and maxicircles relative to nuclear DNA (Figure 6A), thereby suggesting that kDNA network shrinkage was due to a replication defect. We next examined the effects of TbPOLIC silencing on free minicircle replication intermediates. The kDNA replication process involves release of covalently closed minicircles from the network, unidirectional ␪ structure replication, and subsequent reattachment of newly synthesized gapped progeny. Free minicircle replication intermediates are easily resolved by gel electrophoresis and have been well characterized in C. fasciculata (Birkenmeyer and Ray, 1986; Kitchin et al., 1985), and T. equiperdum (Ryan and Englund, 1989). We therefore fractionated DNA samples on an agarose gel and detected free minicircle replication intermediates by Southern hybridization (Figure 6B). Intact networks do not enter the gel and are inefficiently transferred to membranes. However, they appear to decline in concentration consistent with Figure 6A. Free minicircle replication intermediates normally constitute only a small percentage of network minicircles in log phase cells. In contrast, the concentration of free minicircle replication intermediates actually increases during RNAi of TbPOLIC. Lane 0 shows the free minicircle population from uninduced cells in which covalently closed and gapped minicircles are the major species. During the course of RNAi, there was a progressive increase in covalently closed (up 1.7-fold, day 6) and gapped circles (up 2-fold, day 6) and a more pronounced accumulation of multiply gapped and oligomeric species. These changes in the population of free minicircle replication intermediates provide additional evidence that TbPOLIC plays a role in kDNA replication. TbPOLIC Is an Active DNA Polymerase Although the T. brucei pol I-like proteins contain all the residues known to be essential for pol I activity, we wanted to formally show that at least one pol I-like protein was active. Recombinant TbPOLIC containing a

C-terminal hemagglutinin (HA) tag was expressed in yeast under control of the galactose-inducible GAL1 promoter. Using SDS-PAGE and Western blotting with a monoclonal HA antibody, we detected a protein near the predicted size (ⵑ200 kDa) in extracts of galactoseinduced cells but not from uninduced cells (Figure 6C). Some of the 200 kDa protein could be precipitated using HA antibody coupled to protein A-Sepharose, but none precipitated with beads alone. We assayed the precipitated protein for DNA polymerase activity using an oligonucleotide primer-template (Figure 6D; Table 1). PAGE of the resulting products revealed extension products (lane 4) similar to those obtained using Klenow pol I (lanes 1 and 2), indicating that TbPOLIC is an active DNA polymerase. No extension products were detected in precipitates from beads alone (lane 5) or from precipitates of uninduced cells (lane 6). Discussion To better understand kDNA replication, we initiated a project to identify a kDNA replicative polymerase. Using a genomics approach, we discovered four candidate polymerases in T. brucei with similarity to bacterial pol I. We also found similar sequences in the genome databases of the related species, L. major and T. cruzi, indicating that multiple pol I-like proteins are characteristic of trypanosomatids. Most other eukaryotes, in contrast, have only a single pol I-like enzyme, pol ␪, which is involved in nuclear DNA repair (Tosal et al., 2000). Amazingly, localization studies using GFP fusions (Figure 3B) and immunofluorescence (Figure 4) showed that all four candidate polymerases localize in the cell’s single mitochondrion. There is also another polymerase localized in trypanosomatid mitochondria, a pol ␤ (Ferguson et al., 1992). We initially purified and characterized this pol ␤ from C. fasciculata (Torri and Englund, 1995) and recently identified a homolog in T. brucei (70% identity with C. fasciculata pol ␤, T. Saxowsky and P.T.E., unpublished data). Furthermore, fractionation of T. brucei mitochondrial extracts indicated that they contain a pol ␤ as well as other DNA polymerases (Fuenmayor et al., 1998). Therefore, in striking contrast to other eukaryotes that apparently have just one mitochondrial DNA polymerase, pol ␥, T. brucei has at least five mitochondrial polymerases. While writing this paper, we discovered yet another pol I-like ORF in the T. brucei database. This ORF resembles pol ␪, and we are now characterizing that gene and its product. So far, we have demonstrated polymerase activity only for recombinant TbPOLIC (Figure 6D), but the conservation of residues known to be essential for pol I activity suggests that the other pol I-like proteins will also be active. Attempts to express the others are in progress. The finding of multiple pol I-like proteins in trypanosomes raises questions regarding the evolution of family A DNA polymerases. Of foremost interest is whether pol I-like proteins in trypanosomes are intermediates in the evolution of the highly diverged mitochondrial pol ␥. Genome sequencing of numerous organisms shows that prokaryotes possess a single family A member, pol I, and that most eukaryotes possess pol ␥ and pol ␪ (S. cerevisiae lacks pol ␪). Phylogenetic analyses of the

Molecular Cell 182

Figure 6. Function of TbPOLIC (A) Loss of minicircles and maxicircles after TbPOLIC RNAi induction. Total DNA (5 ⫻ 105 cell equivalents/spot) was purified and analyzed by dot-blotting with probes to minicircles, maxicircles, and the nuclear hexose transporter gene, THT1 (Wang and Englund, 2001). Hybridization was measured by phosphorimaging and reported as a percentage of that in uninduced cells. To correct for unequal loading, the amount of minicircle and maxicircle DNA was normalized to THT1. (B) Effect on free minicircle replication intermediates. Total DNA (106 cell equivalents/ lane) from the same induction as in (A) was purified and fractionated on a 1.5% agarose gel (20 ⫻ 25 ⫻ 0.5 cm, 70 V, 18 hr) including ethidium bromide (Wang and Englund, 2001). After transfer to a membrane, kDNA was detected with a minicircle probe. Lane M, kDNA markers of covalently closed (I), gapped and nicked (II), and linearized (III) free minicircles. Lanes 0–8, DNA samples from the indicated days of RNAi induction. Asterisk, crosshybridization to nuclear DNA. Oligomeric and multiply gapped species resemble those previously identified in T. equiperdum, a related species (Ryan and Englund, 1989; Shapiro, 1994). (C) Immunoprecipitation of TbPOLIC-HA from yeast. Cell equivalent amounts of immunoprecipitate pellets (P) and supernatants (S) were analyzed by Western blotting with HA epitope antibodies followed by chemiluminescence. (D) DNA synthesis by TbPOLIC-HA. Products were resolved by 20% denaturing PAGE and visualized by autoradiography. Cell equivalent amounts of precipitates were assayed. Lane 1, Klenow pol I (80 ␮U); lane 2, Klenow pol I (40 ␮U); lanes 3 and 7, 5⬘-[32P]oligonucleotides B1 and B3; lane 4, immunoprecipitate (␣-HA beads) from galactose induced cells; lane 5, immunoprecipitate (beads only) from galactose-induced cells; and lane 6, immunoprecipitate (␣-HA beads) from uninduced cells. For lanes 4–6, comparable amounts of beads were used.

family A pol domain suggest that a single gene duplication event occurred early in eukaryotic evolution (presumably soon after the mitochondrial endosymbiosis) that gave rise to the pol ␪ repair and pol ␥ replicative clades. The branching pattern among the T. brucei sequences also implies the existence of two clades. One or more of the T. brucei sequences in the replicative clade could represent an ancestral form of pol ␥ that may now be trypanosome specific and required for replication of their unique kDNA. The lack of a pol ␥ homolog in any trypanosomatid genome database supports this hypothesis. Alternatively, trypansomatids could have lost a pol ␥ homolog and simply expanded the pol ␪ family. Both scenarios are possible at this point because we have no information regarding family A pols in other early branching eukaryotes. Nonetheless, the expansion of a pol I-like gene family in trypanosomes was an important event for generating diversity and new functions. It is interesting that Arabidopsis thaliana also appears to lack pol ␥, but does have a pol ␪ and two other pol

I-like sequences. The latter may function in mitochondrial or chloroplast DNA replication (Burgers et al., 2001). Furthermore, rice has a pol I-like enzyme that is localized to a plastid subcellular fraction and may function in plastid maturation (Kimura et al., 2002). All previously studied kDNA replication proteins localize to discrete sites around the kDNA disk (see Introduction). However, using GFP fusions and immunolocalization, we found that TbPOLIA and -ID distribute throughout the mitochondrial matrix. The localization of TbPOLIB and -IC differed between the GFP fusion and the endogenous protein visualized by specific antibodies. Whereas most of the endogenous proteins localized near the kDNA disk (Figures 4B and 4C), the GFP fusions were distributed uniformly throughout the mitochondrial matrix (with an enrichment often seen near the disk) (Figure 3B). This discrepancy could be explained by the fact that the GFP fusions were overexpressed from the strong procyclin promoter and that excess protein saturated the localization sites near the kDNA disk. Overex-

Mitochondrial DNA Polymerases in Trypanosomes 183

Table 1. Oligonucleotides Primera

Sequenceb

POLIA-gfpS POLIA-gfpA POLIA-rnaiS POLIA-rnaiA

GCTAGCTCGTTACATGGTCATCCGc GCTAGCACTCGGCACCCGCAGTCC GAGTTTCTCGAGGTTTGCGGACTATGACAAGC GAGTTTAAGCTTTCCTCAGGGTTTCTTCTGAC

POLIB-gfpS POLIB-gfpA POLIB-rnaiS POLIB-rnaiA

TTAAGCTTATGCGGCTAAATAGCTGCTGG TTAAGCTTTTCTACACTGTCAGTAACTTCC CGGAAACTCGAGCGACAACAAGGACAAGATG GTGGAAAAGCTTGCGACGAGGCGGTTCAAAG

POLIC-gfpS POLIC-gfpA POLIC-rnaiS POLIC-rnaiA POLIC-haS POLIC-haA

TTAAGCTTATGCGACGTACTTTCAGTCGC TTAAGCTTCCTGGACAACTCCCCTAG TGGTGGCTCGAGTCAACAATACGGCTATTGTGG GAGTTTAAGCTTTGCGGGAGGACACGCCGGC CGGTTTGTATTACTTCTTATTCAAATGTAATAAAAGTATCAACTCGAGATGCGACGTACTTTCAGTCGCTACG ATAGGGATAGCCCGCATAGTCAGGAACATCGTATGGGTAAAAGATGCGCCTGGACAACTCCCCTAGTGATGGTCC

POLID-gfpS POLID-gfpA POLID-rnaiS POLID-rnaiA

TAGATGAGTTAAGCTTATGCTGCGGCGGCTCTTCAAG TGAGAGAAGTAAGCTTCAGTGTCTCCTCAATGACAACGGG GGGAAACTCGAGGATGGGACAATAAGTAAG GAGGAAAAGCTTCTTGAATGCCTCTGAGGTCG

B1-primer B3-template

AGCTACCATGCCTGCACGAA AGCTATGACCATGATTACGAATTGCTTAGTTCGTGCAGGCATGGTAGCT

a

S and A indicate sense and antisense primers. Restriction sites in primer used in subsequent cloning are underlined and start codons are bold. c The start codon is provided by the vector and is directly upstream of the NheI site. Therefore, the GFP fusion protein has two extra amino acids, Ala and Ser, at the N terminus. b

pression could also explain why IA-GFP fusion (but never endogenous TbPOLIA visualized with specific antibody) localized to the nucleus as well as to the mitochondrion in 7% of the cells (data not shown). What could be the functions of the multiple mitochondrial DNA polymerases? They could have redundant functions, or more likely, each could have a specific role in kDNA replication or repair. We used RNAi as a genetic tool to begin dissecting the roles of these proteins. RNAi of TbPOLIA caused nearly complete degradation of its mRNA without affecting growth. This finding raises the possibility that TbPOLIA could be involved in kDNA repair, and further studies are needed to address this hypothesis. RNAi did not clarify a role for TbPOLID, as there was no decrease in mRNA levels. The localization of TbPOLIB and -IC provides some clue to their function. Based on immunofluorescence, these proteins are positioned between the face of the kDNA disk and the mitochondrial membrane nearest the flagellar basal body, the kinetoflagellar zone. This is the precise region where free minicircle replication initiates and intermediates are detected. TbPOLIB and -IC are thus ideally positioned for a role in kDNA replication. There was variation in TbPOLIB and -IC localization among different cells (e.g., localization in one or two sites, with variation in spacing or position) (Figures 4B and 4C and data not shown), and this variation could be due to changes in localization during the cell cycle. There is ample precedent for such variation with other kDNA replication proteins (topo II and pol ␤ [Johnson and Englund, 1998], SSE1 [Engel and Ray, 1999], and UMSBP [Abu-Elneel et al., 2001]). Further evidence concerning the function of TbPOLIB and -IC came from RNAi studies. Consistent with a role in replication, silencing of these genes resulted in growth

inhibition and shrinking of the kDNA network (Figure 5). Network shrinking was previously observed following RNAi of T. brucei topo II, a protein that attaches newly replicated minicircles to the network (Wang and Englund, 2001). The more subtle network shrinking following TbPOLIB silencing will require further study. However, RNAi of TbPOLIC provided stronger evidence for a kDNA replication defect. The more striking kDNA shrinkage was not due to just decatenation of the network but to progressive loss of both minicircles and maxicircles (Figure 6A). Minicircle loss appears to result from a direct effect on their replication (see below). Maxicircle loss could also be a direct effect or could be due to indirect effects from the loss of minicircles. The decrease in kDNA content due to TbPOLIC silencing was accompanied by dramatic changes in the population of free minicircle replication intermediates (Figure 6B). RNAi caused accumulation of all types of intermediates with the most striking increases observed for multiply gapped and oligomeric species. Therefore, TbPOLIC is involved in the minicircle replication process, but this finding does not point to a precise role for this enzyme (e.g., whether it is involved in leading or lagging strand synthesis). Interpretation of these data is complicated by the fact that TbPOLIC was silenced relative to the multiple other mitochondrial polymerases (including TbPOLIB, which also localizes in the kinetoflagellar zone) and the possibility that silencing of the TbPOLIC activity was incomplete. With regard to the latter point, there was ⬎85% reduction in specific mRNA after 40 hr of RNAi, but we have no information regarding the protein stability or the minimal number of TbPOLIC molecules required for the replication process. If a few TbPOLIC molecules were still synthesized during RNAi, there might be enough to allow replication to occur at

Molecular Cell 184

a greatly reduced rate, causing a buildup of some replication intermediates. In any case, the localization of the TbPOLIC in the kinetoflagellar zone (the site for initiation of minicircle replication), the fact that it is required for maintenance of the kDNA network, and that fact that its knockdown perturbs the population of free minicircle replication intermediates provide strong evidence for a function in kDNA replication. Further studies are underway to determine the precise role of this enzyme. Experimental Procedures Polymerase Gene Identification, Library Screening, and Sequencing Using sequences of several bacterial pol I proteins as queries, TBLASTN searches of the T. brucei genome databases (http:// www.tigr.org/tdb/mdb/tbdb/ and http://www.sanger.ac.uk/Projects/ T_brucei/) revealed four different sequence groups with similarity to family A pol domains. We amplified these sequences by PCR from T. brucei 427 genomic DNA and used the PCR products to probe a bacteriophage P1 library of T. brucei 927 genomic DNA on a high-density filter (from S. Melville, University of Cambridge) using the recommended conditions (http://parsun1.path.cam.ac.uk/libs2.htm). Individual P1 clones (19E2, TbPOLIA; 8A12, TbPOLIB; 21A10, TbPOLIC; and 10A1, TbPOLID) were sequenced on both strands at the DNA Analysis Facility (R. Ingersoll, Johns Hopkins Medical School), and primer walking was used to generate contigs containing each ORF. Consensus sequences were generated using Sequencher Software (Gene Codes Corp.). Sequence Analysis The T. brucei polymerase ORFs were used for additional database searches at the NCBI blast server (http://www.ncbi.nlm.nih.gov/ BLAST) and the parasite genome blast server (http://www.ebi.ac.uk/ blast2/parasites.html). Protein domains were predicted using the Pfam protein families database (http://pfam.wustl.edu), and organellar targeting was predicted with the PSORTII program (http://psort. nibb.ac.jp). Other protein sequences were downloaded from public databases. Due to the significant sequence divergence among the family A polymerases, a protein profile alignment was first generated for bacterial pol I sequences using PILEUP (GCG) with default settings. All other family A sequences were subsequently aligned to the profile. The Pfam (version 6.2) protein alignment for the family A pol domain was used to refine the alignment, and manual adjustments were made using SEQLAB (GCG). For the phylogenetic analysis of family A polymerases, only the pol domain was used. Large inserts and ambiguous sites were excluded (alignment is available upon request). Phage sequences were not used for the two phylogenetic reconstructions. Parsimony analysis used the PHYLIP PROTPARS program (http://evolution.genetics.washington.edu/phylip/progs. data.prot.html) with 1000 replicates and branch swapping to search for the shortest trees. A maximum-likelihood analysis used the program PUZZLE (Strimmer and von Haeseler, 1996) with 1000 iterations, the Jones, Thorton, and Taylor substitution matrix, and eight rate categories. Plasmid Constructs To generate GFP fusion proteins, the coding regions of TbPOLIA, -IB, -IC, and -ID were amplified by PCR (high-fidelity Taq polymerase, Roche Diagnostics, Inc.) using P1 clones as template and primers containing either NheI or HindIII linkers (Table 1). The PCR products were digested with the appropriate restriction enzymes and ligated into the corresponding sites of pXSGFPM3FUS (Marchetti et al., 2000) (from C. Tschudi, Yale University) to generate the inframe C-terminal fusion proteins IA-GFP, IB-GFP, IC-GFP, and IDGFP. The IA and IC constructs contained full-length coding sequence. Because of internal HindIII sites, the IB and ID constructs contained the N-terminal 1053 aa (75%) and 1086 aa (67%) of the coding sequence, respectively. Prior to transfection, the IA-GFP construct was linearized with PpuM I, while IB-, IC-, and ID-GFP

constructs were linearized with BstXI for insertion into the chromosomal tubulin locus. For RNAi experiments, nonhomologous fragments of TbPOLI(A–D) coding sequence (each 500 bp) were PCR amplified using P1 clones as template and primers containing XhoI or HindIII linkers (Table 1). The PCR products were then ligated into the RNAi vector, pZJM (Wang et al., 2000). Constructs were linearized with NotI prior to transfection to allow integration into the chromosomal nontranscribed rDNA spacer region. All constructs were verified by DNA sequencing. Trypanosome Growth and Transfection For GFP experiments, T. brucei procyclic strain YTat1.1 (from E. Ullu, Yale University) was cultured at 27⬚C in Cunningham’s medium containing 20% fetal bovine serum (Fantoni et al., 1994). For stable transfections, cells (2–3 ⫻ 107 ) were washed three times and then resuspended in 0.5 ml of cytomix (van den Hoff et al., 1992). Cells were electroporated with 25 ␮g of linearized GFP expression constructs in a chilled 4 mm gapped cuvette using two pulses (10 s apart) from a BTX electroporator (Model ECM 600) set for peak discharge at 1.6 kV and resistance-timing mode R2 (24 ⍀). After electroporation, cells were transferred to 9.5 ml of Cunningham’s medium. Selection was applied after 24 hr by adding G418 (50 ␮g/ ml). Cells were analyzed for GFP expression as early as 24 hr posttransfection. For RNAi experiments, T. brucei procyclic strain 29-13 (from G. Cross, Rockefeller University) was cultured in SDM-79 medium containing 15% fetal bovine serum, G418 (15 ␮g/ml), and hygromycin (50 ␮g/ml). The drugs maintain expression of the T7 RNA polymerase and the tetracycline repressor, respectively (Wirtz et al., 1999). Stable transfections with pZJM constructs were performed as described (Wang et al., 2000), and selection was applied with phleomycin (2.5 ␮g/ml) 24 hr later. Stable cell lines were cloned by limiting dilution in SDM-79 medium (containing G418, hygromycin, and phleomycin) using 96-well tissue culture plates at 1 cell/ml. Culturing under 5% CO2 at 27⬚C (cloning method from W. Gibson, University of Bristol) resulted in clonal cell lines with plating efficiencies ranging from 30%–60%. Cell densities were determined using a Coulter Counter (model Z1, Coulter Electronics), and cultures were diluted 1:10 when the density reached 3–5 ⫻ 106 cells/ml. Production of Antibodies The following C-terminal peptides were synthesized with an extra cysteine at the N terminus for coupling: pol IA, LGDLEEWTVD HALGLRVPS; pol IB, DALIPELSKEVTDSVEITV; pol IC, KAEIHIGPSL GELSR; and pol ID, DVVAGENMGALRPVVIEETL. All peptides were coupled to maleimide-activated keyhole limpet hemocyanin or to Sulfo-link beads (both from Pierce Chemical Co.) to elicit production of polyclonal sera in rats. Immunizations followed standard protocols at Cocalico Biologicals (Reamstown, PA). Sera were collected after six boosts. Immunofluorescence Immunofluorescence was as described (Wang and Englund, 2001). Slides were incubated 90 min with primary antibody diluted in blocking buffer, 1:25 for anti-pol IA, -pol IB, -pol IC, and -pol ID. No immunofluorescence signal was detected with preimmune sera. The primary antibody, rabbit anti-LipDH (from L. Krauth-Siegel, Universita¨t Heidelberg), was diluted 1:500 for colocalization studies. Slides were incubated with secondary antibody for 45 min as follows: 1:250 dilution of AlexaFluor 488 goat anti-rat IgG (Molecular Probes) or 1:250 dilution of AlexaFluor 594 goat anti-rabbit IgG (Molecular Probes). Fluorescence Microscopy Cells were examined with an Axioskop microscope (Carl Zeiss, Inc.). Images were captured with a Sensys CCD camera (Photometrics Ltd.) using IP Lab software v3.1 (Scanalytics). For GFP analyses, live cell images were collected using fluorescein and DAPI single channel filter sets (chroma 41004 and 31000, respectively; Chroma Technology Corp.). Imaging of fixed cells for immunofluorescence used a multichannel emission filter equipped with a computer-driven filter wheel containing individual excitation filters for DAPI, fluores-

Mitochondrial DNA Polymerases in Trypanosomes 185

cein, and Texas Red (chroma 61002; Chroma Technology Corp.). Registration was confirmed using 0.56 ␮m TetraSpeck microspheres (Molecular Probes). Virtually no bleed-through of fluorescein and rhodamine channels was detected. RNA Purification and Analysis Total RNA was purified from 5 ⫻ 107 parasites using the Purescript RNA Isolation Kit (Gentra System) and fractionated on a 1.5% agarose/7% formaldehyde gel. To ensure that equal amounts of RNA were loaded per lane, rRNA was estimated by ethidium bromide staining. RNA was then transferred to GeneScreen Plus membrane (NEN). Specific mRNAs were detected by Northern hybridization using a 32P-labeled probe (random primed) made from the same PCR fragment used in the corresponding pZJM construct (Table 1). Hybridization and washing conditions were as described (Wang et al., 2000). Recombinant TbPOLIC Expression and Immunoprecipitation TbPOLIC coding sequence was PCR amplified from P1 clone 21A10. The primer contained sequences homologous to the GAL1-URA3 promoter and the HA epitope in pHS15 (Sesaki and Jensen, 1999) (from H. Sesaki, Johns Hopkins Medical School). The PCR product and XhoI-NotI-digested pHS15 were cotransformed into yeast strain JB811 (MATa leu2 trp1 ura3-52 prb1122 pep4-3 prc1-407, from J. Boeke, Johns Hopkins Medical School), and pPOLIC-HA was formed by homologous recombination. JB811 cells carrying pPOLIC-HA were grown in raffinose medium (OD600, 1.0), centrifuged, and resuspended (OD600, 0.5) in galactose medium to induce TbPOLIC-HA expression (4 hr, 30⬚C). Uninduced cells were grown in raffinose medium for 4 hr. Immunoprecipitation was performed at 4⬚C. Cells (10 OD600 units) were lysed with glass beads in 1 ml IP buffer (50 mM Tris-HCl [pH 7.5], 5 mM EDTA, 50 mM NaCl, 1% Triton X-100, and protease inhibitors: 1 mM PMSF, 1 ␮g/ml each of aprotinin, pepstatin, and leupeptin). Extracts were centrifuged (12,500 ⫻ g, 10 min), and 75 ␮l of a 1:1 slurry of protein A-Sepharose (beads only) was added to the supernatants. Samples were gently agitated for 30 min, and the beads were cleared by centrifugation (12,500 ⫻ g, 1 min). Then 75 ␮l of a 1:1 slurry of 12CA5 monoclonal HA antibody coupled to protein A-Sepharose (␣-HA coupled beads) or beads only was added to each supernatant. The tubes were gently agitated for 4 hr, and beads were collected by centrifugation and supernatants saved. Pellets were washed three times with IP buffer containing 1 mM EDTA and resuspended in 500 ␮l of 50 mM Tris-HCl (pH 8.0), 0.1 mM EDTA, 50 mM KCl, 2 mM DTT, 10% glycerol, and protease inhibitors as above. DNA Polymerase Assay Reaction mixtures (20 ␮l) contained 50 mM Tris-HCl (pH 8.0), 5 mM MgCl2, 100 mM KCl, 1 mM DTT, 100 ␮g/ml bovine serum albumin, 25 ␮M each of dCTP and dGTP, 5 ␮M of B1/B3 primer-template (see Table 1), and 50 nM each of [␣-32P]dATP and [␣-32P]dTTP (both 3000 Ci/mmol). Reactions were incubated at 37⬚C for 90 min and stopped by adding 2⫻ formamide gel loading buffer. Acknowledgments We dedicate this paper to the memory of Viiu Klein. We appreciate comments on the manuscript by Drs. D. Robinson, H. Sesaki, T. Shapiro, B. Sollner-Webb, and members of our lab. M.M.K. was supported by NRSA (5F32AI09789). This work was supported by the National Institutes of Health (GM27608). Received: November 21, 2001 Revised: May 7, 2002 References Abu-Elneel, K., Robinson, D.R., Drew, M.E., Englund, P.T., and Shlomai, J. (2001). Intramitochondrial localization of universal minicircle sequence-binding protein, a trypanosomatid protein that binds kinetoplast minicircle replication origins. J. Cell Biol. 153, 725–734. Astatke, M., Ng, K., Grindley, N.D., and Joyce, C.M. (1998). A single

side chain prevents Escherichia coli DNA polymerase I (Klenow fragment) from incorporating ribonucleotides. Proc. Natl. Acad. Sci. USA 95, 3402–3407. Birkenmeyer, L., and Ray, D.S. (1986). Replication of kinetoplast DNA in isolated kinetoplasts from Crithidia fasciculata. Identification of minicircle DNA replication intermediates. J. Biol. Chem. 261, 2362–2368. Braithwaite, D.K., and Ito, J. (1993). Compilation, alignment, and phylogenetic relationships of DNA polymerases. Nucleic Acids Res. 21, 787–802. Burgers, P.M., Koonin, E.V., Bruford, E., Blanco, L., Burtis, K.C., Christman, M.F., Copeland, W.C., Friedberg, E.C., Hanaoka, F., Hinkle, D.C., et al. (2001). Eukaryotic DNA polymerases: proposal for a revised nomenclature. J. Biol. Chem. 276, 43487–43490. Drew, M.E., and Englund, P.T. (2001). Intramitochondrial location and dynamics of Crithidia fasciculata kinetoplast minicircle replication intermediates. J. Cell Biol. 153, 735–744. Engel, M.L., and Ray, D.S. (1998). A structure-specific DNA endonuclease is enriched in kinetoplasts purified from Crithidia fasciculata. Nucleic Acids Res. 26, 4733–4738. Engel, M.L., and Ray, D.S. (1999). The kinetoplast structure-specific endonuclease I is related to the 5⬘ exo/endonuclease domain of bacterial DNA polymerase I and colocalizes with the kinetoplast topoisomerase II and DNA polymerase ␤ during replication. Proc. Natl. Acad. Sci. USA 96, 8455–8460. Estevez, A.M., and Simpson, L. (1999). Uridine insertion/deletion RNA editing in trypanosome mitochondria—a review. Gene 240, 247–260. Fantoni, A., Dare, A.O., and Tschudi, C. (1994). RNA polymerase IIImediated transcription of the trypanosome U2 small nuclear RNA gene is controlled by both intragenic and extragenic regulatory elements. Mol. Cell. Biol. 14, 2021–2028. Felsenstein, J. (1978). Cases in which parsimony and compatibility methods will be positively misleading. Syst. Zool. 27, 401–410. Ferguson, M., Torri, A.F., Ward, D.C., and Englund, P.T. (1992). In situ hybridization to the Crithidia fasciculata kinetoplast reveals two antipodal sites involved in kinetoplast DNA replication. Cell 70, 621–629. Fuenmayor, J., Zhang, J., Ruyechan, W., and Williams, N. (1998). Identification and characterization of two DNA polymerase activities present in Trypanosoma brucei mitochondria. J. Eukaryot. Microbiol. 45, 404–410. Hausler, T., Stierhof, Y.D., Blattner, J., and Clayton, C. (1997). Conservation of mitochondrial targeting sequence function in mitochondrial and hydrogenosomal proteins from the early-branching eukaryotes Crithidia, Trypanosoma and Trichomonas. Eur. J. Cell Biol. 73, 240–251. Hu¨bscher, U., Maga, G., and Spadari, S. (2002). Eukaryotic DNA polymerases. Annu. Rev. Biochem. 71, 133–163. Huelsenbeck, J.P. (1997). Is the Felsenstein zone a fly trap? Syst. Biol. 46, 69–74. Johnson, C.E., and Englund, P.T. (1998). Changes in organization of Crithidia fasciculata kinetoplast DNA replication proteins during the cell cycle. J. Cell Biol. 143, 911–919. Kitchin, P.A., Klein, V.A., and Englund, P.T. (1985). Intermediates in the replication of kinetoplast DNA minicircles. J. Biol. Chem. 260, 3844–3851. Kimura, S., Uchiyama, Y., Kasai, N., Namekawa, S., Saotome, A., Ueda, T., Ando, T., Ishibashi, T., Oshige, M., Furukawa, T., et al. (2002). A novel DNA polymerase homologous to Escherichia coli DNA polymerase I from a higher plant, rice (Oryza sativa L.). Nucleic Acids Res. 30, 1585–1592. Klingbeil, M.M., Drew, M.E., Liu, Y., Morris, J.C., Motyka, S.A., Saxowsky, T.T., Wang, Z., and Englund, P.T. (2001). Unlocking the secrets of trypanosome kinetoplast DNA network replication. Protist 152, 255–262. Li, C., and Englund, P.T. (1997). A mitochondrial DNA primase from the trypanosomatid Crithidia fasciculata. J. Biol. Chem. 272, 20787– 20792.

Molecular Cell 186

Marchetti, M.A., Tschudi, C., Kwon, H., Wolin, S.L., and Ullu, E. (2000). Import of proteins into the trypanosome nucleus and their distribution at karyokinesis. J. Cell Sci. 113, 899–906. Melendy, T., and Ray, D.S. (1989). Novobiocin affinity purification of a mitochondrial type II topoisomerase from the trypanosomatid Crithidia fasciculata. J. Biol. Chem. 264, 1870–1876. Melendy, T., Sheline, C., and Ray, D.S. (1988). Localization of a type II DNA topoisomerase to two sites at the periphery of the kinetoplast DNA of Crithidia fasciculata. Cell 55, 1083–1088. Miyahira, Y., and Dvorak, J.A. (1994). Kinetoplastidae display naturally occurring ancillary DNA-containing structures. Mol. Biochem. Parasitol. 65, 339–349. Neupert, W. (1997). Protein import into mitochondria. Annu. Rev. Biochem. 66, 863–917. Patel, P.H., Suzuki, M., Adman, E., Shinkai, A., and Loeb, L.A. (2001). Prokaryotic DNA polymerase I: evolution, structure, and “base flipping” mechanism for nucleotide selection. J. Mol. Biol. 308, 823–837. Ryan, K.A., and Englund, P.T. (1989). Synthesis and processing of kinetoplast DNA minicircles in Trypanosoma equiperdum. Mol. Cell. Biol. 9, 3212–3217. Sesaki, H., and Jensen, R.E. (1999). Division versus fusion: Dnm1p and Fzo1p antagonistically regulate mitochondrial shape. J. Cell Biol. 147, 699–706. Shapiro, T.A. (1994). Mitochondrial topoisomerase II activity is essential for kinetoplast DNA minicircle segregation. Mol. Cell. Biol. 14, 3660–3667. Sogin, M.L., and Silberman, J.D. (1998). Evolution of the protists and protistan parasites from the perspective of molecular systematics. Int. J. Parasitol. 28, 11–20. Sousa, R. (1996). Structural and mechanistic relationships between nucleic acid polymerases. Trends Biochem. Sci. 21, 186–190. Steitz, T.A. (1998). A mechanism for all polymerases. Nature 391, 231–232. Strimmer, K., and von Haeseler, A. (1996). Quartet puzzling: a quartet maximum likelihood method for reconstructing tree topologies. Mol. Biol. Evol. 13, 964–969. Torri, A.F., and Englund, P.T. (1995). A DNA polymerase beta in the mitochondrion of the trypanosomatid Crithidia fasciculata. J. Biol. Chem. 270, 3495–3497. Torri, A.F., Kunkel, T.A., and Englund, P.T. (1994). A beta-like DNA polymerase from the mitochondrion of the trypanosomatid Crithidia fasciculata. J. Biol. Chem. 269, 8165–8171. Tosal, L., Comendador, M.A., and Sierra, L.M. (2000). The mus308 locus of Drosophila melanogaster is implicated in the bypass of ENU-induced O-alkylpyrimidine adducts. Mol. Gen. Genet. 263, 144–151. Tzfati, Y., Abeliovich, H., Avrahami, D., and Shlomai, J. (1995). Universal minicircle sequence binding protein, a CCHC-type zinc finger protein that binds the universal minicircle sequence of trypanosomatids. Purification and characterization. J. Biol. Chem. 270, 21339– 21345. van den Hoff, M.J., Moorman, A.F., and Lamers, W.H. (1992). Electroporation in ‘intracellular’ buffer increases cell survival. Nucleic Acids Res. 20, 2902. Wang, Z., and Englund, P.T. (2001). RNA interference of a trypanosome topoisomerase II causes progressive loss of mitochondrial DNA. EMBO J. 20, 4674–4683. Wang, Z., Morris, J.C., Drew, M.E., and Englund, P.T. (2000). Inhibition of Trypanosoma brucei gene expression by RNA interference using an integratable vector with opposing T7 promoters. J. Biol. Chem. 275, 40174–40179. Wirtz, E., Leal, S., Ochatt, C., and Cross, G.A. (1999). A tightly regulated inducible expression system for conditional gene knock-outs and dominant-negative genetics in Trypanosoma brucei. Mol. Biochem. Parasitol. 99, 89–101.

Lihat lebih banyak...

Comentários

Copyright © 2017 DADOSPDF Inc.