The ribosomal RNA gene promoter and adjacent cis-acting DNA sequences govern plasmid DNA partitioning and stable inheritance in the parasitic protozoan Leishmania

Abstract

Detailed analysis of the Leishmania donovani ribosomal RNA (rRNA) gene promoter region has allowed the identification of cis-acting sequences involved in plasmid DNA partitioning and stable plasmid inheritance. We report that plasmids bearing the 350 bp rRNA promoter along with the 200 bp region immediately 3′ to the promoter exhibited a 6.5-fold increase in transformation frequency and were transmitted to daughter cells as single-copy molecules. This is in contrast to what has been observed for plasmid molecules in this organism so far. Moreover, we show that these low-copy-number plasmids displayed a remarkable mitotic stability in the absence of selective pressure. The region in the vicinity of the RNA pol I transcription initiation site, and also in the adjacent 200 nt, displays a complex structural organization and shares sequence similarity to the yeast autonomously replicating consensus sequence and centromere DNA elements. Deletion analyses indicated that these elements were necessary but not sufficient for plasmid DNA partitioning and stable inheritance, and that the rRNA promoter region was required for optimal function. These results suggest an interplay between RNA pol I transcription, DNA replication, DNA partitioning and mitotic stability in trypanosomatids. This is the first example of defined DNA elements for plasmid partitioning and stable inheritance in the protozoan parasite Leishmania.

INTRODUCTION

DNA sequence elements participating in chromosome replication, segregation and mitotic stability in Trypanosomatidae remain largely unknown. Currently, the analysis of the complete nucleotide sequence of several Trypanosoma brucei and Leishmania major Friedlin chromosomes (http://www.genedb.org) has not revealed any homologies to known consensus sequences such as those involved in site-specific initiation of replication or in centromere function in yeast (1). Studies of extrachromosomal circular amplicons that are frequently observed in drug-resistant Leishmania (2,3), failed to isolate origins of replication. This may be due to the multiple copy number of these episomes that may favor initiation of replication at more than one position. To date, the mitochondrial minicircle and maxicircle kinetoplastid DNA serve as the best model for examining the structure of origins of replication in these organisms [reviewed in (4,5)]. Replication of minicircles is initiated on the complementary strand to the universal minicircle sequence (GGGGTT GGTGTA) (6) and proceeds unidirectionally as a theta (θ) structure (4). Replication of the maxicircle DNA initiates at the variable region that is characterized by highly repetitive sequences (7,8) and proceeds unidirectionally as a θ structure similarly to the minicircle DNA (9).

Using methylation-sensitive restriction endonucleases to distinguish between input and replicated DNA, we have previously shown that plasmid DNA could autonomously replicate in Leishmania and that replication occurred even in the absence of any leishmanial sequence (10). Replicating plasmids in Leishmania often become joined to each other through the formation of multimers (3,10,11). Multimers probably occur as a result of defective replication termination or by recombination of monomers. Although it has not been demonstrated experimentally, it is possible that plasmid molecules in these parasites bear weak or adventitious origins of replication. A plasmid molecule with a weak origin may be able to replicate at a rate of at least once per cell cycle by increasing the number of sites at which initiation of replication can occur, thereby explaining plasmid multimerization in Leishmania. Evidence from Plasmodium falciparum argues that unstable plasmids can change to stable replicating forms following the formation of large concatemeric structures (12).

Extrachromosomal linear and circular amplicons described in many Leishmania species selected for resistance to various drugs are generally lost after 100–200 cell divisions in the absence of selective pressure (3,11,13,14). To avoid being lost from dividing cells, plasmids carry partitioning systems which ensure that at least one copy of the plasmid segregates into each daughter cell during cell division. Mechanisms that allow circular DNA molecules to be efficiently transmitted generally involve their association with segregated nuclear structures (15–18). In bacteria, low-copy-number plasmids encode a cis-acting centromere-like DNA sequence and two partitioning (par) genes. Together, these form a nucleoprotein complex which is required for faithful segregation of the plasmid before cell division [reviewed in (19)]. In many yeasts studied so far, replicating extrachromosomal DNA circles carrying only replication origins are generally mitotically unstable and display a strong maternal inheritance bias (MIB). The molecular explanation of MIB is unclear since the nuclear volume is shared approximately equally between mother cell and daughter bud. However, it may result from the failure of DNA circles to be liberated from replication sites, since linear plasmids do not exhibit MIB (20). A variety of different sequences that overcome MIB have been described, including centromeric sequences, telomere-associated repeats and a specific system encoded by the endogenous 2 µm circle plasmid (20–22). The successful propagation of the 2 µm circle present in most common strains of Saccharomyces cerevisiae at a copy number of ∼60 per cell is accomplished via a partitioning system and an amplification system [reviewed in (23)]. In Leishmania, extrachromosomal amplicons and replicons generaly demonstrate a strong segregation bias as they are gradually lost from the majority of daughter cells in the absence of drug selection (3,11).

In this study, we show that extrachromosomal vectors bearing the Leishmania donovani ribosomal RNA (rRNA) gene promoter and cis-acting DNA sequences immediately 3′ to the promoter are effectively segregated as single-copy plasmids to daughter cells and are stably inherited in contrast with the majority of episomal vectors in Leishmania. Moreover, we provide new insights on cis-acting DNA elements that play a role in plasmid DNA partitioning and stable inheritance in these parasites.

MATERIALS AND METHODS

Plasmid constructs

The GFP-S65T gene version (24) derived from vector phGFP-S65T (Clontech) was subcloned as an HindIII–XbaI fragment into pGEM7Zf-NEO (25) to yield vector pNEO-GFP. Signals essential for trans-splicing and polyadenylation of the transcripts encoding for the neomycin phosphotransferase (NEO) and green fluorescent protein (GFP) are present within the α-tubulin intergenic region which exists as two copies in the vector pNEO-GFP (26). Vectors p1.2NEO-GFP and prev1.2NEO-GFP were made by subcloning a 1.2 kb L.donovani genomic fragment bearing the 350 bp rRNA gene promoter region (27) and ∼850 bp flanking sequences until the start of the 18S rRNA gene into the SmaI site of vector pNEO-GFP in either orientation. The 1.2 kb genomic fragment was amplified by PCR with oligonucleotides 5′-GAGGTCTGCGATTGACGTAG-3′ and 5′-TCAACCAGATCTCAATTGTG-3′. In vector pNEO-GFP1.2, the same 1.2 kb fragment was inserted downstream of the GFP gene into the XbaI site of vector pNEO-GFP. The pBT1-1.2GFPNEO-BT1 vector initially used to inactivate the BT1 gene of L.donovani has been previously described (28). To generate p950NEO-GFP, p750NEO-GFP, p550NEO-GFP, p450NEO-GFP and p350NEO-GFP expression vectors, DNA fragments corresponding to the first 950 or 750 or 550 or 450 and/or 350 bp of the rRNA-promoter-bearing 1.2 kb region were amplified by PCR and cloned into the SmaI site of pNEO-GFP. These PCR products were created with the following oligonucleotides: 5′-GAGGTCTGCGATTGACGTAG-3′ (for all 5′ ends) and (950) 5′-ACAACTCAGCTTTTTGGGTG-3′, (750) 5′-TTCGGCGCAGCGCTCTCTAC-3′, (550)5′-GGACACACGAGCAGAGGTCA-3′, (450) 5′-CTTTCCAACGATACAGCGCG-3′and (350) 5′-ACAAACACGGCATCCACGCA-3′ (for annealing the 3′ ends of the DNA). Vectors p850NEO-GFP and p850Δ200NEO-GFP were made by subcloning the PCR amplified fragments corresponding to the last 850 and/or 650 nt of the 1.2 kb fragment lacking the rRNA promoter region into the SmaI site of pNEO-GFP. The following oligonucleotides were used for PCR amplification: 5′-GCTACACACAAGCAAAGGC-3′ (5′ end of 850), 5′-CGTATCCCCGGGACCTGCGG-3′ (5′ end of 650) and 5′-TCAACCAGATCTCAATTGTG-3′ (3′ end of both). The 200 bp region located immediately 3′ to the 350 bp rRNA promoter was amplified by PCR using oligonucleotides 5′-CAAAGGCAACAACGCAAGCGC-3′ and 5′-GGACACACGAGCAGAGGTACA-3′. Subcloning of this fragment into the SmaI site of pNEO-GFP resulted in the vector p200NEO-GFP.

Cell culture and transfections

The L.donovani donovani Sudanese 1S2D strain used in this study has been previously described (29). Leishmania cells were grown in SDM-79 medium supplemented with 10% fetal bovine serum (Multicell, Wisent Inc.) and 5 mg/ml of hemin (Sigma). Approximately 15 µg of plasmid DNA was used for transfections by electroporation as described previously (25). Transfectants were selected with 40 µg/ml of G-418 (Geneticin, Gibco-BRL) on liquid SDM medium. To calculate the transformation frequency following transient transfections, an equal molar DNA concentration from vectors pNEO-GFP and p550NEO-GFP was transfected by electroporation into Leishmania and the percentage of plasmid-bearing cells (positive for GFP fluorescence) was calculated 48 h post-transfection by flow cytometry. Stable Leishmania transfectants were grown on G-418-containing medium for ∼2–3 weeks. The transformation frequency of these transfected cells was calculated by dividing the number of G-418 resistant clones by the total number of clones grown on non-selective media.

Flow cytometry

The levels of GFP fluorescence in living recombinant Leishmania cells were quantified by an Epic XL flow cytometer (Beckman–Coulter) tuned to 488 nm (excitation) and 525 nm (emission). To evaluate stable plasmid inheritance in dividing cells, recombinant Leishmania were cultured in the absence of G-418 selection for several cell generations. Given that all the extrachromosomal vectors express the GFP gene, the percentage of plasmid-bearing fluorescent parasites was quantitatively measured by fluorescence flow cytometry through an Epic Elite ESP (Beckman–Coulter) apparatus. Cells were studied during mid-logarithmic phase in SDM culture media and 10 000 live parasites were analyzed per sample.

DNA and RNA manipulations

Total DNA from Leishmania was isolated using DNAzol (Gibco-BRL). Southern blotting, radiolabelling of DNA probes, hybridizations and washing conditions were performed following standard protocols (30). The NEO gene-specific probe was made by PCR. The GFP gene probe was generated by HindIII–XbaI digestion of vector phGFP-S65T. The trypanothione reductase (TR) gene (31) amplified by PCR was used as a control single-copy gene probe. The copy number of the different extrachromosomal vectors was estimated by comparative Southern blot hybridization. Quantitation of TR- and NEO-containing DNA fragments was done by densitometric analysis using a PhosphorImager with the ImageQuant3.1 software.

Nuclear run-on assays

Nuclear run-on assays using nuclei freshly isolated from L.donovani transfectants were performed as reported previously (29). Transcripts were labeled with [³²P]dUTP with or without α-amanitin (Sigma), an inhibitor of RNA polymerase II (pol II) transcription, at 200 µg/ml. Approximately 10 µg of the GFP, α-tubulin and 18S rRNA gene probes were transferred to a nylon membrane and hybridized with the labeled nascent RNA.

RESULTS

Position and orientation effects of the L.donovani rRNA promoter on downstream gene transcription

The rRNA gene promoter has been characterized in many Leishmania species (27,32–34). The RNA polymerase I (pol I) transcription initiation site of the L.donovani, Leishmania chagasi and Leishmania amazonensis rRNA unit was mapped to ∼1020 nt upstream of the 18S rRNA gene (27,32,33). The sequences essential for promoter function have been characterized in L.donovani and correspond to a 349 bp region from –179 to +170 with respect to the transcription initiation site (27). We have previously reported that the insertion of reporter genes downstream of the L.donovani rRNA promoter resulted in high levels of expression (35). We have also shown that the rRNA pol I promoter was active into RNA pol II transcription units of Leishmania when inserted in the reverse orientation relative to the chromosomal endogenous transcription (28). In this study, we have evaluated the effect of the rRNA promoter spacing and/or orientation on downstream gene transcription. The L.donovani rRNA promoter region (∼350 bp) was cloned as part of a 1.2 kb genomic fragment (see Fig. 1) in either orientation upstream of the NEO and GFP genes. As expected, the GFP and NEO genes on vector p1.2NEO-GFP are transcribed by RNA pol I, which is resistant to α-amanitin (Fig. 2B and data not shown). Transcription of the GFP gene on vector p1.2NEO-GFP is increased by 262-fold compared with the GFP RNA pol II transcription on vector pNEO-GFP (Fig. 2A and Table 1). The effect of the rRNA promoter activity on downstream GFP transcription was estimated by measuring GFP fluorescence by flow cytometric analysis as indicated in Materials and Methods. We show that the activity of the rRNA promoter is much higher when the promoter is placed at the vicinity of the genes to be transcribed. This is the case for episomal vectors p1.2NEO-GFP and pBT1-1.2NEOGFP-BT1 (Table 1). A distant localization (∼3.5 kb) of the rRNA promoter with respect to the NEO and GFP genes results in 10-fold decrease in GFP transcription (Table 1) on vector pNEO-GFP1.2 (Fig. 1). RNA pol I-mediated transcription on episomal vectors is strand independent, as there is no significant difference in GFP fluorescence levels between vector p1.2NEO-GFP and pBT1-1.2NEOGFP-BT1 in which the promoter region and downstream NEO and GFP genes are inserted into the biopterin transporter 1 gene (BT1) (28) in the opposite strand of the vector (Fig. 2B and Table 1). However, positioning of the rRNA promoter in the reverse orientation relative to NEO and GFP gene transcription abolishes RNA pol I activity. Indeed, for vector prev1.2NEO-GFP (Fig. 1), the NEO and GFP genes are transcribed by an α-amanitin sensitive RNA pol II as indicated by run-on studies (Fig. 2D). In Leishmania, no RNA pol II promoters have yet been described (36), and RNA pol II transcription can be initiated fortuitously on both DNA strands from several non-specific sites within the plasmid (37,38) and only 5′- and 3′-processing signals are needed for genes to be expressed (26,39).

Figure 1.

Schematic representation of episomal vectors containing the L.donovani rRNA gene promoter region in different configurations relative to the transcribed genes on the plasmid. Vector pNEO-GFP co-express the neomycin phosphotransferase gene (NEO) and the gene encoding for the green fluorescent protein (GFP). The 5′ and 3′ end processing of the NEO and GFP mRNAs is mediated by polypyrimidine-rich sequences within the α-tubulin intergenic region (IR) (26). One rRNA unit on chromosome 27 of L.donovani contains ∼39 copies of a 64 bp repeat, the 350 bp rRNA RNA pol I promoter region and the subunits 18S, 5.8S and 28S rRNAs (27). The 1.2 kb genomic fragment subcloned in vector p1.2NEO-GFP contains the 350 bp promoter region and 850 bp flanking sequences upstream of the 18S rRNA gene. This fragment was also subcloned downstream of the NEO and GFP genes in vector pNEO-GFP1.2 or in the opposite orientation relative to gene transcription on plasmid prev1.2NEO-GFP or finally in the non-coding strand of plasmid pBT1-1.2GFPNEO-BT1, a vector made for targeting the BT1 gene of L.donovani (28).

Figure 2.

RNA pol I- or RNA pol II-mediated transcription on various extrachromosomal vectors. Nuclear run-on assays were performed on recombinant L.donovani strains in the absence (–) or presence (+) of α-amanitin as indicated in Materials and Methods. The α-tubulin gene was used as a control for the α-amanitin-sensitive RNA pol II transcription, whereas the 18S rRNA gene was the control for α-amanitin-resistant RNA pol I transcription. (A) The GFP gene on vector pNEO-GFP lacking the rRNA gene promoter is transcribed by RNA pol II. (B) The GFP gene on vector p1.2NEO-GFP is under the control of the rRNA promoter and is therefore transcribed by RNA pol I. (C) In pNEO-GFP1.2, the rRNA promoter is positioned ∼3.5 kb upstream of the NEO-GFP cassette and genes are transcribed by RNA pol I. (D) In prev1.2NEO-GFP, the rRNA promoter region is placed in the reverse orientation relative to the NEO-GFP transcription unit, and these genes are transcribed by RNA pol II. (E) Genes on vector p850NEO-GFP, which lacks the 350 bp rRNA promoter region, are transcribed by RNA pol II.

Table 1. GFP fluorescence of recombinant L.donovani carrying episomal vectors with the rRNA promoter region in various contexts.

Recombinant Leishmania	Plasmid copies	GFP fluorescencea	Ratio of GFP fluorescence to plasmid copy	Efficiency of the rRNA promoter (%)
L.d. pNEO-GFP	51	39b	0.76 (1)c	–d
L.d. p1.2NEO-GFP	1	199	199 (262)	100
L.d. pNEO-GFP1.2	1	20	20 (26.3)	10
L.d. pBT1-1.2GFPNEO-BT1	1	160	160 (210.5)	87.6
L.d. prev1.2NEO-GFP	52	15	0.29 (0.38)	–d

^aAverage of four independent experiments.

^bValues of GFP fluorescence relative to background.

^cValues in parenthesis correspond to the fold increase in GFP fluorescence relative to pNEO-GFP vector control.

^dThe GFP gene on these episomal vectors is transcribed by RNA pol II.

A region of ∼100 bp immediately 3′ to the rRNA promoter assures active partition of plasmid DNA to dividing daughter cells

While analyzing the DNA content of recombinant GFP-expressing Leishmania by Southern blot hybridization (Fig. 3B) we observed that vectors bearing the 1.2 kb rRNA-promoter-containing region, such as p1.2NEO-GFP, pNEO-GFP1.2 and pBT1-1.2GFPNEO-BT1 were present at approximately one copy per diploid genome, in contrast to the promoter-lacking vector pNEO-GFP which was found at ∼50 copies per parasite genome (Figs 3B and 4B). This is intriguing because extrachromosomal plasmids in this organism are generally found in multiple copies which often multimerize (10). As mentioned in the Introduction, low-copy-number plasmids in other systems encode partitioning genes that are required for faithful segregation of the plasmid before cell division (19). However, vectors p1.2NEO-GFP, pNEO-GFP1.2 and pBT1-1.2GFPNEO-BT1 (Fig. 1) do not encode such genes or sequences related to them. In order to identify the sequences involved in plasmid DNA partitioning control, which ensures that at least one copy of the plasmid segregates into each daughter cell during Leishmania cell division, we made a series of deletion mutants from either side of the 1.2 kb rRNA-promoter-bearing fragment that we cloned into vector pNEO-GFP (Fig. 3A). Deletions encompassing the last 750 nt of the 1.2 kb fragment did not affect the low-copy-number distribution of vectors p950NEO-GFP, p750NEO-GFP, p550NEO-GFP and p450NEO-GFP (Fig. 3A) in dividing cells (Fig. 3). However, deletion of the next 100 nt located immediately 3′ to the rRNA promoter abolishes faithful segregation of the plasmid, as indicated by Southern blot hybridization (Fig. 3B) and FACS analysis (Fig. 4A). In fact, a variable number of p350NEO-GFP copies was observed in dividing cells (Fig. 3B) and often this plasmid was integrated into the rDNA genomic locus (data not shown), thereby supporting that sequences within the 100 bp region are involved in plasmid partition control to dividing daughter cells.

Figure 3.

A 100 bp region immediately 3′ to the rRNA promoter is responsible for plasmid DNA partitioning. (A) Schematic representation of a series of deletion constructs spanning the 1.2 kb rRNA promoter-containing region. The region designated as a 350 bp box corresponds to sequences essential for rRNA promoter function in L.donovani (27) with the transcription initiation site in the middle of the box. Dashed lines represent the deleted sequences in the different constructs. Vector p950NEO-GFP lacks the last 250 nt of the 1.2 kb fragment; p750NEO-GFP lacks the last 450 nt; p550NEO-GFP lacks the last 650 nt; p450NEO-GFP lacks the last 750 nt and p350NEO-GFP contains only the 350 bp rRNA promoter region. Vector p850NEO-GFP lacks the 350 bp promoter region and contains the downstream 850 bp. Vector p200NEO-GFP contains the region of 200 bp immediately downstream from the rRNA promoter, and vector p850Δ200NEO-GFP lacks the 350 bp promoter region and the adjacent 200 bp region. The effect of each of these deletions on GFP expression was evaluated by flow cytometry. Indicated values are relative to the p1.2NEO-GFP vector. ¹The vector bearing only the promoter region showed variable copy-number results and was often integrated into the rRNA genomic locus (not shown). The subregion of 100 bp immediately downstream of the 350 bp promoter region was identified as the region responsible for plasmid DNA partitioning allowing faithful segregation between mother and daughter cells at division. Plasmid copy number in the various recombinant L.donovani strains was estimated by FACS and hybridization studies. (B) Comparative Southern blot analysis of total genomic DNA digested with PstI (cuts outside the NEO and TR coding regions) and hybridized to NEO and trypanothione reductase gene probes (TR, single-copy gene used for normalization of the data) as well as densitometric quantitative evaluation of the hybridization signals (not shown) were used to calculate values presented in (A).

Figure 4.

A region of ∼100 bp adjacent to, but distinct from, the plasmid DNA partitioning determinant plays a role in stable plasmid inheritance. (A) As all plasmids express the GFP gene here, plasmid mitotic stability was measured by flow cytometry in cells grown in the absence of selective pressure for various parasite cell divisions as explained in Materials and Methods. Episomal vectors are as described in Figures 1 and 3A. L.donovani cells divide approximately every 10–12 h under our experimental conditions. The minimum sequence requirement for ensuring plasmid mitotic stability over >1000 cell generations in non-selective conditions of growth corresponds to the 550 bp region bearing the rRNA promoter and the adjacent 200 nt. (B) The stable maintenance of the various episomal vectors transmitted to daughter cells during cell division was also confirmed by Southern blot hybridization studies. The numbers on the top of the blots (0–1000) correspond to the number of cell generations in the absence of G-418 selection. TR and NEO probes are the same as those used in Figure 3.

Cis-acting DNA sequences adjacent to but distinct from the plasmid DNA partitioning determinant are implicated in stable plasmid inheritance

To determine whether extrachromosomal replicating plasmids capable of faithful segregation during cell division were stably inherited from cell to cell, we examined the distribution of these plasmids in cells grown under non-selective conditions. We made use of flow cytometric analyses (Fig. 4A) and Southern blot hybridization studies (Fig. 4B) to determine the percentage of fluorescent parasites bearing an episomal vector and also the copy number of these plasmids per parasite cell. Given that all plasmid molecules express GFP, loss of fluorescence can be a reliable marker for the loss of the episome. Single-copy plasmids bearing, in addition to the 350 bp rRNA promoter region, the adjacent 200 nt (p550NEO-GFP) or 400 nt (p750NEO-GFP) or 600 nt (p950NEO-GFP) or 850 nt (p1.2NEO-GFP) (Fig. 3A) were maintained uniformly in >99.9% of the FACS-sorted parasites in the absence of drug selection for >1000 cell divisions (Fig. 4). In fact, the growth was tested over a period of 2 years. However, episomal vectors p350NEO-GFP and p450NEO-GFP bearing only the promoter region or the promoter with the proximal 100 nt shown to be involved in plasmid DNA partitioning (Fig. 3) were not sufficient to ensure plasmid mitotic stability, as they were rapidly lost (<100 cell generations) in the absence of drug selection (Fig. 4). By deletion analyses, we showed that the requirements for plasmid maintenance were conferred by sequences located between nucleotides 450 and 550 immediately 3′ to the DNA partitioning determinant (Figs 3A and 4).

The rRNA gene promoter acts in conjunction with adjacent cis-acting DNA elements to ensure plasmid DNA partitioning and stable plasmid inheritance

In this study, we have identified cis-acting DNA elements that play an essential role in plasmid DNA partitioning and stable maintenance in dividing Leishmania cells (Figs 3 and 4). Given that these DNA elements are located immediately 3′ to the rRNA promoter (Fig. 5A), it is possible that their function in the faithful segregation and mitotic stability of the plasmid may require factors or cis-acting sequences interacting with the promoter. To determine if the rRNA promoter and RNA pol I transcription played a role in these processes, we made additional deletion constructs and tested their capacity for being effectively segregated and stably inherited. Deletion of the 350 bp rRNA promoter region in vector p850NEO-GFP (Fig. 3A) significantly increased the copy number of plasmid molecules that were transmitted to dividing cells from one to >10 copies. Moreover, it affected their mitotic stability because after 50 generations <25% of the parasites contained the plasmid, and the plasmid was completely lost after ∼300 cell divisions in the absence of drug pressure (Figs 3 and 4). Similar results were obtained with vector p200NEO-GFP (Fig. 3 and data not shown). Thus, in the absence of the upstream rRNA promoter region, neither of the downstream DNA elements (region of 200 bp) involved in plasmid partitioning and plasmid maintenance functioned. Deletion of both the promoter region and the 200 bp downstream region in vector p850Δ200NEO-GFP further induced poor segregation (>50 copies/parasite cell) and rapid loss of this plasmid from dividing cells (100–150 cell divisions) (Figs 3 and 4). Furthermore, vector prev1.2NEO-GFP bearing an inactive promoter (Fig. 2D) placed in the reverse orientation (Fig. 1) was present at multiple copies in daughter cells and was highly unstable (Fig. 4). Clearly, our data indicate that RNA pol II transcription does not have the same effect on plasmid DNA partitioning and stable inheritance than RNA pol I, suggesting that sequences within the rRNA promoter or proteins bound to it may be involved in plasmid segregation and mitotic stability. Taken together, these findings suggest that RNA pol I transcription in coordination with downstream cis-acting DNA elements plays an important role in regulating plasmid partitioning, hence ensuring stable inheritance of the plasmid DNA from cell to cell during cell division.

Figure 5.

The region encompassing the rRNA promoter and adjacent cis-acting DNA elements involved in plasmid DNA partitioning and stable inheritance displays a complex sequence organization and structure. (A) Schematic organization of the rRNA promoter and the adjacent 200 bp region bearing two distinct subregions with partition control and stable maintenance elements. (B) Detailed sequence organization of the 550 bp promoter-containing region. This region displays high levels of complexity with several sequence motifs and superstructures (indicated by arrows), including direct (DR 1–18) and inverted repeats (IR 1–6) of variable size, sequences with dyad symmetries (DS) and kinkable TG and CA dinucleotides situated between pyrimidine/purine tracks (in color). The region from nucleotides 150 to 200 at the vicinity of the RNA pol I transcription initiation start site (nucleotide 169) shares similarity with the ARS consensus (A/TTTTATA/GTTTA/T) in the budding yeast S.cerevisiae (44) and with the GGGGTTGGTGT universal minicircle sequence involved in replication initiation of the mitochondrial minicircle DNA in trypanosomatids (4). The region from nucleotides 332 to 414 shares sequence similarity with the CDEI (PuTCACPuTG) and CDEIII (TGTTT(T/A)TGNTTTCCGAA(A/C)NNNAAAAA) centromere DNA elements of S.cerevisiae (40). (C) Predicted secondary structure of the first 550 bp rRNA promoter-bearing region by the mfold algorithm. The position of the ARS-like consensus sequence relative to the initiation site of RNA pol I transcription is indicated. The region in the vicinity of the transcription initiation start is thermodynamically instable. The position of the CDEI- and CDEIII-like DNA elements as well as the most important dyad symmetries and inverted repeats are indicated by arrows.

The region encompassing the rRNA promoter and adjacent cis-acting DNA elements involved in plasmid DNA partitioning and stable inheritance displays a complex sequence organization and structure

Careful sequence analysis of the 550 bp region containing the rRNA promoter and adjacent cis-acting sequences involved in plasmid DNA partitioning and stable plasmid inheritance (Fig. 5A) revealed interesting features. The sequence from nucleotides 332 to 414 demonstrates a significantly higher dAT content (56%) than the average 35–37% content for the Leishmania genome and shares sequence similarity to the centromere (CEN) DNA elements CDEI (PuTCACPuTG) and CDEIII (TGTTT(T/A)TGNTTTCCGAA(A/C)NNNAAAAA) of S.cerevisiae and Kluyveromyces lactis [reviewed in (40, 41)]. As in yeast (40), the CDEIII-like sequence (332–357) is composed of two regions of outer and inner dyad symmetries, and the left region of dyad symmetry is flanked by a TG dinucleotide (Fig. 5B). This dinucleotide has been found to be essential for centromere function in yeast (42). The dyad symmetries within the CDEIII-like element could potentially fold into a hairpin (Fig. 5C). The sequence from nucleotides 408 to 414, which shares sequence homology with the CDEI element, is followed by a sequence with alternating G and T residues (Fig. 5B and C). This is similar to what has been found in yeast (40,41). Although the overall size and sequence composition of these Leishmania sequences resembles that of S.cerevisiae CEN DNA, their order is reversed (CDEIII/CDEI instead of CDEI/CDEIII) (Fig. 5B). In addition, a complex pattern of several direct repeats (DR), inverted repeats (IR) and regions of dyad symmetry (DS) were detected within this centromere-like sequence and throughout the 550 bp rRNA-promoter-containing region (Fig. 5B). Direct repeats and dyad symmetries are known to be part of centromeric sequences in the yeast Yarrowia lipolytica (43) and the budding yeast S.cerevisiae (40). Interestingly, a predicted structure of the 550 bp region by the mfold algorithm indicates that several of the inverted repeats or regions with dyad symmetry fold into various hairpins (Fig. 5C), thereby suggesting that these complex sequence motifs and superstructures (e.g. symmetries) may be important for centromere-binding proteins.

The T-rich strand from nucleotides 150 to 168 immediately 5′ to the RNA pol I transcription initiation start shares similarity with the autonomously replicating sequence (ARS) consensus (A/TTTTATA/GTTTA/T) in S.cerevisiae (44) (Fig. 5B). In addition, a sequence homologous to the GGGGTTGGTGT universal minicircle sequence involved in the initiation of replication of the mitochondrial minicircle DNA in trypanosomatids (4) was found several nucleotides downstream of the transcription initiation start site (Fig. 5B and C). The region from nucleotides 143 to 203 that contains the ARS-like sequence and the universal minicircle sequence exhibits helical instability, as predicted by the mfold algorithm (Fig. 5C). A DNA unwinding element whose helical instability facilitates origin activity has often been detected immediately 3′ to the essential ARS consensus in yeast (45,46). Furthermore, a large number of kinkable DNA sites rich in dinucleotides such as TG and CA were found throughout the 550 bp region (Fig. 5B). TG and CA dinucleotides situated between pyrimidine/purine tracks are often part of recognition signals for many proteins involved in recombination and/or replication (47,48).

Cis-acting DNA elements involved in replication priming were shown to promote high frequency of transformation and extrachromosomal maintenance of plasmid DNA (49,50). Using transient and stable transfections, we tested whether the transformation frequency of Leishmania by plasmid DNA containing the rRNA promoter region could be enhanced. We indeed observed a 6.5-fold increase in the transformation frequency of Leishmania transfected with the rRNA promoter-region-containing vector p550NEO-GFP in comparison with vector pNEO-GFP which lacks the promoter and the 200 bp downstream region (Table 2). A dramatic increase in the transformation frequency of T.brucei procyclics has also been reported for plasmids bearing the PARP gene RNA pol I promoter but not the rRNA promoter region (51). Taken together, these results support that the 350 bp rRNA promoter region and the downstream cis-acting DNA elements (200 bp) display a complex sequence organization and structure that shares some common features with the ARS consensus sequence and centromere DNA elements in yeast.

Table 2. High-frequency transformation of L.donovani promastigotes by plasmids containing the cloned rRNA promoter region.

Transfection	Episomal vector	Number of transfectantsa		Transformation frequencyb	Relative transformation increase
		–G-418	+G-418
Stable	pNEO-GFP	19.7 × 106	270	1.37 × 10–3	1
	p550NEO-GFP	33.6 × 106	2990	8.9 × 10–3	6.5
Transient	pNEO-GFP			0.26 × 10–2	1
	p550NEO-GFP			1.6 × 10–2	6.15

^aThese numbers were obtained by plating Leishmania transfectants on solid media and/or by measuring GFP fluorescence of transient transfectants by flow cytometry 48 h following transfection. The average of three independent experiments is shown.

^bThe transformation frequency for the stable transfectants was calculated by dividing the number of G-418-resistant cells by the total number of parasite cells grown in the absence of selective pressure.

DISCUSSION

In this paper, we report that plasmids bearing the L.donovani rRNA gene promoter and adjacent 200 bp region were transmitted at low copy number (one copy per cell on average) to daughter cells and displayed a remarkable mitotic relative stability with <0.1% loss per generation. Moreover, we show that the 200 bp region immediately 3′ to the rRNA promoter contains two distinct regions with cis-acting DNA elements that are involved in plasmid DNA partitioning and stable inheritance. No such elements have been reported previously in protozoan parasites. Low-copy-number plasmids, like chromosomes, have evolved mechanisms to ensure that efficient replicon partitioning occurs following DNA replication. DNA replication and partition appear to be no longer compartmentalized into separate stages but are parallel processes that proceed concomitantly through a cell cycle continuum (52). Plasmid copy number control is tightly associated with DNA replication (53,54) as episomes usually multimerize in the absence of strong origin of replication (10,13). In this study, we describe plasmids that segregate in a similar manner to that of the Leishmania chromosomes, e.g. at one copy per dividing cell (Fig. 3 and Table 1) in contrast with the majority of extrachromosomal vectors in these parasites, which are poorly segregated at high copy numbers between mother and daughter cells [Figs 3 and 4, Table 1 and (10,13)]. An extensive comparative sequence analysis of these low-copy-number plasmids depicted cis-acting DNA elements that share homology with sequences known to function as origins of DNA replication in other systems (Fig. 5B and C). These sequences are located at the vicinity of the RNA pol I transcription initiation start site and include the autonomously replicating sequence (ARS) consensus in S.cerevisiae (44,49, 50) and the universal minicircle sequence GGGGTTGGTGT of the mitochondrial minicircle DNA in trypanosomes and related parasitic protozoa (4,6). Interestingly, the region surrounding the RNA pol I transcription initiation start site demonstrates a predicted low helical stability (Fig. 5C). Helical instability facilitates eukaryotic replication origin activity, and a DNA unwinding element has often been detected at the vicinity of the ARS consensus (45,46,55). In our model system, RNA pol I transcription could promote plasmid replication from replication origins located near the promoter as a result of the negative supercoiling introduced by the movement of RNA polymerase as indicated in humans (56) and seen in bacteria (57). From human (58–60) to protozoa (61) and yeast (62–65), cis-acting DNA fragments involved in replication initiation were found within the highly transcribed rDNA locus.

The data here also suggest that sequences in the vicinity of the RNA pol I transcription initiation site (Fig. 5) might serve as the template for initiation of replication of single-copy plasmids in Leishmania as L.donovani cells transfected with these plasmids demonstrated a 6.5-fold increase in their transformation frequency (Table 2). Generally, sequences that function as replication origins on plasmids permit high-frequency transformation of cells (66). Putative origins of replication, including sequences within the PARP gene RNA pol I promoter region in the related parasite T.brucei, have been shown to control single-copy episome replication and to increase parasite transformation frequency (51). Considering that any DNA sequence, even a non-leishmanial sequence, could function as a replicator in Leishmania (10), a 6.5-fold increase in the transformation frequency is highly significant and suggests that a strong origin of replication, possibly involving the ARS-like consensus or the universal minicircle sequence, may be present at the vicinity of the RNA pol I transcription initiation site. The RNA pol I rRNA promoter in Leishmania has been characterized in many Leishmania species (27,32–34), but there is currently no evidence for a role in DNA replication in these parasites. Whether these sequences function as replication origins on plasmids remains to be determined. However, plasmids lacking the rRNA promoter undergo uneven segregation, thereby supporting the involvement of promoter sequences and possibly factors that bind the promoter in efficient replicon partitioning following DNA replication.

In yeast, high-copy-number plasmids carrying only replication origins show MIB and are poorly segregated to daughter cells (20). Generally, episomal vectors in Leishmania also demonstrate a strong segregation bias as they are transmitted at high copy number to daughter cells and are gradually lost from the majority of dividing cells in the absence of drug selection [Figs 3 and 4, and (3,11,13,14)]. In contrast, plasmids bearing the rRNA promoter and cis-acting DNA elements encompassing the 200 nt immediately 3′ to the promoter are transmitted as monomeric single-copy plasmids and are mitotically stable over 1000 cell divisions in the absence of drug pressure (Figs 3 and 4). The stable maintenance of these episomes is clearly contrasted by the general situation of insufficiently replicated or poorly segregated plasmids in these parasites. This supports the possibility of an effective segregation control between mother and daughter cells at division similarly to chromosomal DNA. In many yeasts, the presence of CEN on plasmids improved their mitotic stability and decreased their copy number (67,68). CEN may represent the only genetic determinant that allows autonomously replicating plasmids to attach to the mitotic apparatus to avoid being lost. A centromere could act as a stabilizing sequence that prevents excessive segregation and allows correct transmission of the genetic material to the daughter cells, like STB sequences in S.cerevisiae 2-µm plasmids (69). Segregation of low-copy plasmids in bacteria is mediated by a plasmid centromere-like site where plasmid-specified partition proteins bind to promote segregation [reviewed in (70)]. Also, the insertion of a CEN sequence on an ARS-containing plasmid helps to overcome MIB in S.cerevisiae (20–22). Only centromeres can supply the partition system required for ARS function in the yeast Y.lipolytica (43). Interestingly, in this study, we show that the 200 bp region immediately 3′ to the rRNA promoter contains sequences that share sequence similarity with the CDEI and CDEIII elements (Fig. 5B) of the S.cerevisiae CEN DNA (22,71) and that deletion of these sequences completely abolishes plasmid maintenance (Fig. 4). The S.cerevisiae centromere is characterized by three DNA elements: CDEI (8–9 bp), CDEII [78–86 bp (A+T)-rich spacer] and CDEIII (25–26 bp) (40). In spite of their sequence similarity to the yeast CDEI and CDEIII boxes, the order of the cis-acting DNA elements within the 200 bp region is reversed in Leishmania. In addition, the CDEII AT-rich spacer is shorter and contains lower dAT content than in yeast. Further experiments are needed to fully address the possibility that this region may correlate with a centromere-like function. However, additional levels of complexity were found within the 350 bp rRNA promoter and adjacent 200 bp region, including numerous direct and inverted repeats, dyad symmetries, kinkable TG and CA dinucleotides and alternating homopurine (mainly AAAn or GGGn) and homopyrimidine (mainly TTTn) elements (Fig. 5B). It is noteworthy that structures of similar high complexity have been found in Y.lipolytica, in S.cerevisiae and Candida glabrata centromere DNA elements (CDE) (42,43,72) and also in higher eukaryotes (48,73).

Our data indicate that cis-acting DNA elements located within the 200 bp region immediately 3′ to the rRNA promoter are necessary but not sufficient for plasmid DNA partitioning and stable inheritance, and that the rRNA promoter is required to ensure their function (Figs 3 and 4). The importance of the rRNA promoter in ‘regulating’ these processes was demonstrated experimentally by either deleting the promoter region or by placing it in the reverse orientation with respect to GFP gene transcription on the plasmid. In both cases, plasmids transcribed by RNA pol II lost their ability to be faithfully segregated and stably inherited (Figs 3 and 4, and Table 1). We show here that RNA pol II transcription has no effect on plasmid DNA replication, segregation and stable maintenance. In contrast, the rRNA RNA pol I promoter played an important role in these processes probably because of more active transcription and the ability of transcription factors to remodel chromatin when bound to the promoter, hence stimulating DNA replication as shown in yeast for RNA pol II and pol III transcription factors (74). Plasmid DNA bearing the rRNA promoter is transcribed by RNA pol I (Fig. 2B and C) and it should in principle co-localize with the chromosomal encoded rRNA at the nucleolus. The localization of the promoter-bearing plasmid at the nucleolus may favor interactions with factors involved in rDNA replication and transcription and its attachment to a cellular entity that divides equally between mother and daughter cells, as in the case of the rDNA. In T.brucei, RNA pol I promoters such as the rRNA (75) and the procyclic acidic repetitive protein (PARP) (51) promoters have been associated with mitotic stability or suggested to act as putative origins of replication (37,51), respectively. A link between DNA replication and transcription has been demonstrated in many systems. Also, transcription factors have been shown to play various roles in viral and cellular DNA replication origin function [reviewed in (76–78)]. In fact, the majority of the metazoan replication origins map to a transcription promoter or enhancer [reviewed in (79)]. Proteins binding to transcriptional regulatory elements such as a promoter could stimulate the formation of the replication initiation complex by engaging specific interactions with proteins of the initiation complex and/or by modulating the repressive chromatin structure around origins of replication (77,80–82).

In this study, we report that the L.donovani RNA pol I rRNA promoter region associated with adjacent cis-acting DNA elements can ensure plasmid DNA replication, partitioning and mitotic stability. Our data also support a model where RNA pol I transcription, DNA replication and segregation may share common features in trypanosomatid parasites and thus open numerous possibilities for future research in these areas. However, several questions remain to be answered. Of most interest is whether strong promoters, such as the RNA pol I rRNA promoter, are needed for origin activation and whether an associated centromeric sequence is also required.