Formation of linear inverted repeat amplicons following targeting of a
http://www.100md.com
《核酸研究医学期刊》
Division of Molecular Biology and Centre of Biomedical Genetics, The Netherlands Cancer Institute Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands 1Faculty of Medicine, Department of Medical Biology, Laval University, Infectious Disease Research Center, RC709, CHUL Research Center (CHUQ) 2705 blvd Laurier, Quebec, Canada G1V 4G2
*To whom correspondence should be addressed. Tel: +31 020 512 2880; Fax: +31 020 669 1383; Email: p.borst@nki.nl
ABSTRACT
Attempts to inactivate an essential gene in the protozoan parasite Leishmania have often led to the generation of extra copies of the wild-type alleles of the gene. In experiments with Leishmania tarentolae set up to disrupt the gene encoding the J-binding protein 1 (JBP1), a protein binding to the unusual base ?-D-glucosyl-hydroxymethyluracil (J) of Leishmania, we obtained JBP1 mutants containing linear DNA elements (amplicons) of 100 kb. These amplicons consist of a long inverted repeat with telomeric repeats at both ends and contain either the two different targeting cassettes used to inactivate JBP1, or one cassette and one JBP1 gene. Each long repeat within the linear amplicons corresponds to sequences covering the JBP1 locus, starting at the telomeres upstream of JBP1 and ending in a 220 bp sequence repeated in an inverted (palindromic) orientation downstream of the JBP1 locus. We propose that these amplicons have arisen by a template switch inside a DNA replication fork involving the inverted DNA repeats and helped by the gene targeting.
INTRODUCTION
Leishmania sp. is a unicellular protozoan parasite belonging to one of the oldest eukaryotic lineages, the Kinetoplastida. Leishmania is known for its genomic plasticity. Changes in the karyotype of Leishmania, such as triploidy for one chromosome or genome-wide polyploidy, have been reported following attempts to inactivate essential genes (1–3). DNA amplification can also be induced in this parasite by selection with a variety of drugs including methotrexate (4–7), oxyanions (antimony, arsenite) (8–11), mycophenolic acid (12), tunicamycin (13) and vinblastine (14). Amplicons have even been found in unselected laboratory isolates (15).
The structure of the amplicons in Leishmania varies: they can be circular or linear, and can contain direct or inverted DNA repeats. They are usually the product of a conservative amplification with no alteration in the source chromosome and are frequently lost in the absence of the selection that induced the amplification (16). The preferred model of DNA amplification in Leishmania resulting in inverted DNA repeats involves the self-annealing of inverted (palindromic) DNA repeats in a region of single-stranded DNA being replicated, followed by the self-replication of one of the DNA strands and the consecutive synthesis of its complementary strand (5,9,17). Whereas in mammalian cells DNA amplicons are often the product of many successive rearrangements (18), DNA amplification in Leishmania normally occurs in one step. Hence, the amplicons found in Leishmania usually have a simple structure, which makes this parasite a good model for the study of gene amplification (16).
Base J or ?-D-glucosyl-hydroxymethyluracil is a unique hypermodified base only present in the nuclear DNA of kinetoplastid parasites and Euglena (19–21). It replaces 0.5% of thymine in the genome of Trypanosoma, mostly in repetitive sequences (22), including the telomeric repeats (20). The enzymes involved in J biosynthesis have not been isolated yet. Hence, we have not been able to produce parasites without J and the function of J remains a matter of speculation. We isolated a J-binding protein, however, named JBP1 for J-binding protein 1 (23), which binds to J-containing duplex DNA with high specificity (24,25). JBP1 null trypanosomes are viable in vitro and in mice, and have no obvious defect in the stability of their DNA repeats or in gene expression. Interestingly, they have 20-fold less J, suggesting that JBP1 is involved in the maintenance of J levels in T.brucei (26).
To get more insight into the function of JBP1, we tried to inactivate the JBP1 gene by gene targeting (via homologous recombination) in the promastigote (insect) form of Leishmania tarentolae, a parasite of lizards. We could easily inactivate one allele, but all attempts to make a JBP1 null Leishmania resulted in cell lines maintaining a wild-type allele, suggesting that JBP1 is essential in this parasite. Some of these lines were found to contain new linear DNA plasmids, harboring the targeting constructs used to inactivate JBP1. We present here the structure of these plasmid amplicons and a model for their generation during the targeting of the JBP1 locus.
MATERIALS AND METHODS
Culture cell lines and transfections
The L.tarentolae TarIIWT cell line (5) was cultured in SDM 79 medium (27). Transfections were done as described previously (6). The cells were selected at the following drug concentrations: 20 μg puromycin (Sigma) per ml; 20 μg neomycin (Gibco) per ml; 100 μg hygromycin B (Roche) per ml; 200 μg paromomycin (Sigma) per ml. The frequency of transfection was determined by serial dilution of the cells after transfection. The reversion experiment was done by culturing the amplicon-containing cell lines in the absence of drug pressure. The DNA measurement by flow-cytometry was done as described in Munoz-Jordan and Cross (28).
Cloning procedures and inactivation constructs
The cloning of the L.tarentolae JBP1 gene was already described in Cross et al. (23). A cosmid containing Ltar JBP1 was isolated by screening a cosmid library reported in Brochu et al. (29), and a 7.2 kb fragment containing JBP1 was digested with HindIII–ClaI and cloned in pBluescript (Stratagene) giving the construct Ltar JBP1-pBluescript. To make the KO NEO and KO HYG inactivation constructs, the NEO and HYG markers, both preceded by a 90 bp polypyrimidine stretch (Y), were first digested out of the PSPY-NEO and PSPY-HYG vectors (30) by a XbaI–BglII digest, and cloned in PSL1180 (Amersham) giving, respectively, the PSLY-NEO and PSLY-HYG vectors. A 2.4 kb ApaI fragment covering the 3'-untranslated region (3'-UTR) part of Ltar JBP1 was isolated by an ApaI digest of Ltar JBP1-pBluescript, and cloned downstream of the NEO or HYG markers by insertion in the ApaI site of PSLY-NEO or PSLY-HYG. A 2 kb HindIII–XbaI fragment covering the 5'-UTR of Ltar JBP1 was then cloned into the HindIII–SpeI digested PSLY-NEO + 3'-UTR or PSLY-HYG + 3'-UTR vectors, leading to the Ltar JBP1 KO NEO and KO HYG constructs. The cassettes were taken out of the PSL1180 backbone by a HindIII–BstbI double digest prior to transfection. Note that the NEO marker used by us has a point mutation at position 406 (bp) (C A mutation, Gln Lys). To construct the KO PUR inactivation cassette, the gene coding for the puromycin acetyl transferase was HindIII–ClaI cloned after the second -tubulin intergenic region (containing processing signals) of pGEM 7Zf -NEO- (6), resulting in the construct pGEM 7Zf -NEO--PUR. The -PUR cassette was then subcloned into the EcoRI–ClaI sites of pSP72 (Promega), and finally integrated between the 5'-UTR and 3'-UTR sequences of JBP1 by an XbaI–BglII digest of the Ltar JBP1 KO NEO construct (which removes the YNEO cassette). The Ltar JBP1 KO PUR construct was linearized with NcoI prior to transfection. The rearrangement point of the JBP1 KO NEO amplicon was cloned by ligating 2.0 kb EcoRV fragments and 3.0 kb BamHI fragments of digested genomic DNA of JBP1 KO NEO into pBluescript (i.e. fragments of the size of the rearrangement point for these digests). The clones of interest were isolated by screening with the probe 9. Similar experiments were done in order to clone the rearrangement point of the JBP1 KO HYG:KO NEO and KO PUR:KO HYG amplicons. The sequences corresponding to the rearrangement point in the TarIIWT cell line were isolated following similar procedures.
DNA extraction, blotting and hybridization
DNA was gel-extracted using a QIAGEN gel-extraction kit. DNA electrophoresis and blotting were performed using standard conditions (31). Membranes were hybridized with radioactively labeled probes using the Prime-it random primer labeling kit (Stratagene) at 42°C in a formamide buffer , 50 mM NaPi (Na2HPO4, NaH2PO4), 5 mM EDTA, 0.1% SDS and 0.25 mg salmon sperm DNA per ml]. Membranes were washed with 3x SSC, 0.1% SDS and 0.1x SSC and 0.1% SDS. The JBP1 3'-UTR probe corresponds to the 2.4 kb ApaI fragment described in the Cloning section. The XbaI–BglII YNEO and YHYG cassettes (see Cloning section) were used as probes to check for the site of integration of the constructs. The telomeric probe was described in van Leeuwen et al. (20). Hybridization with the telomeric probe was done at 65°C in a water-based buffer . Washes were done with 6x SSC and 0.1% SDS.
Chromosome separation and amplicon extraction
DNA plugs were made as described in van der Ploeg et al. (32). Chromosomes were separated using a CHEF-DRII apparatus (Bio-rad) under the following conditions: 24 h, 6 V/cm, initial switch time: 35 s, final switch time: 120 s. Amplicons were extracted from low melting point agarose gels by agarase (New England Biolabs) treatment following the indications of the manufacturer.
Sequencing
The cosmid covering the JBP1 locus was sequenced by shot-gun sequencing by GATC Biotech AG (Jakob-Stadler Platz 7, D-78467 Konstanz, Germany). The rearrangement point and the corresponding wild-type sequence were obtained by traditional sequencing, using an ABI Prism 3700 DNA Analyzer (Applied Biosystems). The sequence data have been submitted to the GenBank/EMBL/DDBJ database under the accession numbers: AY842844, AY842845, AY842846, AY842847, AY842848 and AY842849.
Design and isolation of the probes scattered on the JBP1 chromosome
Leishmania major Friedlin JBP1 was localized at 18 kb downstream of one of the ends of chromosome 9 by a computer-based screen (BLAST search). Oligonucleotides were designed in order to make 500 bp (or 1 kb for probe 7) PCR fragments adjacently located or spaced by 500 bp or 1 kb, starting at 20 kb downstream of JBP1 (going towards the distant telomere). PCR fragments were made using genomic DNA of L.major Friedlin as a template. Probe 1 was isolated by PCR using primers that were designed based on the sequence of a cosmid covering the JBP1 locus and adjacent sequence of L.tarentolae. The PCR products were purified, labeled and hybridized on digested genomic DNA of L.tarentolae. The list of oligonucleotides used during this study is available upon request.
Quantitation of the copy number of the amplicon
The copy number of the amplicons in the various mutants was determined by quantitative autoradiography with a FLA-3000 apparatus (Fuji) using the Bas reader version 3.14 and Tina version 2.09 softwares. The signal obtained by hybridization on digested genomic DNA of TarIIWT was compared with the one of the cell lines with amplicons, using probes located in the amplicon. Probes located upstream of the rearrangement point (i.e. not present in the amplicon) were used in order to correct for DNA loading.
RESULTS
JBP1 null mutants cannot be generated in L.tarentolae
In order to inactivate JBP1 in Leishmania, targeting constructs were made containing a selection marker (NEO, HYG or PUR) flanked by the 5'-UTR and 3'-UTR sequences of the L.tarentolae (Ltar) JBP1 gene, and these were transfected into the Leishmania TarIIWT cell line. Integration of the cassettes into the JBP1 locus was verified by hybridization, as schematically shown in Figure 1A. Transfection of the KO HYG and KO PUR constructs resulted in integration into the JBP1 locus as shown by the appearance of new bands hybridizing with the JBP1 3'-UTR probe of, respectively, 3.4 and 7.0 kb in addition to the 4.7 kb wild-type band in the KO HYG and KO PUR cell lines (Figure 1B, lanes 2 and 3). The inactivation of the second JBP1 allele by transfection of the KO PUR construct into the KO HYG mutant failed, however, and resulted in a cell line containing two inactivated alleles, but maintaining a JBP1 wild-type allele (Figure 1B, lane 4). All other attempts to obtain JBP1 null mutants resulted in cell lines maintaining a wild-type allele. Failure to obtain a double gene inactivation suggests that JBP1 is an essential gene in Leishmania, as also reported for the DHFR (dihydrofolate reductase) gene (1) and for the TR (trypanothione reductase) gene (3). Our mapping data indicate that the wild-type allele in the JBP1 KO HYG:KO PUR mutant is present in a supernumerary JBP1-containing chromosome, since no DNA rearrangement was detected in the vicinity of the JBP1 locus. Moreover, no ectopic insertion of JBP1 was found into other chromosomes and the overall DNA content (plo?dy) of the JBP1 KO HYG:KO PUR mutant was normal as determined by flow-cytometry (data not shown).
Figure 1 Targeting of the JBP1 locus and analysis of the mutant cell lines obtained. (A) Map of the JBP1 locus before and after integration of the inactivation cassettes. Sequences flanking the JBP1 locus were cloned upstream and downstream of selection markers. Sequences Y (polypyrimidine stretch) (30) and TUB (intergenic region of the alpha tubulin array of Leishmania enriettii) (49) were used for mRNA processing. The location of the NcoI sites and the JBP1 3'-UTR probe are depicted. The drawing is approximately to scale. (B) Southern blot of genomic DNA digested with NcoI. The blot was hybridized with the JBP1 3'-UTR probe. Lanes: 1, TarIIWT; 2, JBP1 KO HYG; 3, JBP1 KO PUR; 4, JBP1 KO HYG:KO PUR; 5, JBP1 KO NEO; 6, JBP1 KO HYG:KO NEO; 7, JBP1 KO PUR:KO HYG; and 8, JBP1 KO HYG:KO PUR:KO NEO. The alleles present in the different strains are indicated above the lanes. The asterisk identifies the amplified alleles. The wild-type 4.7 kb band can be used as a loading control, except in lane 5 where the loading can be checked by comparison of the amplified 2.7 kb band with lane 6.
Targeting of the JBP1 gene of L.tarentolae results in unexpected allelic ratios in some JBP1 mutants
While trying to inactivate JBP1, we also generated mutants with a higher copy number of the inactivated alleles or the wild-type allele (compare the wild-type and inactivated alleles in lanes 1–4 versus lanes 5–8 in Figure 1B). Such unexpected allelic ratios were observed in the JBP1 KO NEO, KO HYG:KO NEO, KO PUR:KO HYG cell lines and in a JBP1 KO HYG:KO PUR:KO NEO cell line in which we tried to inactivate the remaining wild-type allele of the JBP1 KO HYG:PUR mutant using the KO NEO construct. Most mutants have two different amplified alleles, one with the marker inserted when the amplification occurred and the other already integrated in the genome prior to the transfection that induced the amplification.
Presence of small linear amplicons in some JBP1 mutants
A chromosomal size fractionation by pulsed-field gradient electrophoresis, which allows the separation of all chromosome-sized DNA molecules of Leishmania (10,17,33,34), revealed the presence of new additional small linear chromosomes of 100 kb only in the cell lines with an abnormal JBP1 allelic ratio (Figure 2). Hybridization of a blot of the pulsed-field gel showed that the amplified markers were present in these small chromosomes (amplicons) (Table 1). The marker in the inactivation cassette used in the transfection that resulted in the appearance of the amplicons was always found in the amplicons, and in two out of four mutants was also in the JBP1-containing chromosome, as summarized in Table 1. The linear amplicons were only stable in the presence of drug selection and they were rapidly diluted out when we cultured the cell lines in the absence of drug pressure (data not shown). In the presence of drug selection, the copy number of the amplicons is 4–5 copies per cell as determined by quantitative autoradiography.
Figure 2 Pulsed-field gel stained with ethidium bromide showing separation of chromosome-sized DNA of L.tarentolae and the amplicon bands present in the JBP1 mutants. The arrow is pointing at the region containing the linear amplicons. Lanes M: Saccharomyces cerevisiae marker; 1, TarIIWT; 2, JBP1 KO HYG; 3, JBP1 KO PUR; 4, JBP1 KO HYG:KO PUR; 5, JBP1 KO NEO; 6, JBP1 KO HYG:KO NEO; 7, JBP1 KO PUR:KO HYG; and 8, JBP1 KO HYG:KO PUR:KO NEO.
Table 1 Genotype of the cell lines with a linear amplicon
The JBP1 KO NEO amplicon consists of a long inverted repeat
The fact that some amplicons contain two copies of the JBP1 locus, each with a different marker or in one case with a marker gene and a wild-type JBP1 allele (see Table 1), suggested that the amplicon contains a DNA duplication covering the JBP1 locus. To verify this, the JBP1 KO NEO amplicon was extracted from a pulsed-field gradient gel as in Figure 2 and roughly mapped (data not shown). The sum of the size of all the bands arising from several digests fell short by almost half the size of the amplicon estimated by pulsed-field electrophoresis (100 kb) (data not shown), indicating that the JBP1 KO NEO amplicon contained long sequences that were duplicated. The digested amplicon was hybridized with a telomeric probe, which recognized two bands in all the digestions tested, suggesting that the linear amplicons had telomeres at both ends (Figure 3A). A probe located 13 kb upstream of the beginning of the Ltar JBP1 locus (named probe 1, see Figure 4), (see Materials and Methods for a description of how probe 1 was isolated), recognized the same fragments that hybridized with the telomeric probe in some digests, indicating that the location of JBP1 is subtelomeric (see for example the EcoRV digest in lane 2 in Figure 3A and B). These data also suggested that the two ends of the amplicon were located upstream of JBP1, implying that the JBP1 KO NEO amplicon had an inverted repeat structure. The differences in the size of the fragments recognized by the telomeric probe and probe 1 in the BamHI digest (lane 1 in Figure 3A and B) can be explained by the presence of restriction sites between the JBP1 upstream probe (probe 1) and the telomeres, and by the presence of the KO NEO cassette inserted into the JBP1 locus. The presence of two telomeric fragments is probably due to a difference in the number of telomeric repeats at the two ends of the amplicon, as the number of telomeric repeats is known to be highly variable in kinetoplastid parasites (35,36).
Figure 3 Analysis of the telomeric ends of the JBP1 KO NEO amplicon. Southern blot on digested JBP1 KO NEO amplicon isolated from the pulsed-field gel. The blots were hybridized with a telomeric probe (A) and with probe 1 (see Figure 4), located 13 kb upstream of the JBP1 locus (B). Lane 1, BamHI; and lane 2, EcoRV.
Figure 4 Schematic representation of the results of the hybridization experiments done to detect the rearrangement point in the JBP1 KO NEO amplicon. The small rectangles represent the approximate location in the JBP1 chromosome (Chr 9) of the different probes that were generated by PCR on genomic DNA of L.major Friedlin. The white rectangles correspond to the probes that are not present in the amplicon, the black rectangles depict the probes that are present in the amplicon. The gray rectangles correspond to the probes that detect the rearrangement point. The bars indicate the approximate locations of the EcoRV sites around the JBP1 locus in L.tarentolae. The small arrows indicate the location of the RS1, RS2 and RS3 repeats. Their orientation is indicated by the direction of the arrow. The triangles represent the telomeric repeats. The JBP1 gene is depicted by the yellow rectangle. The drawing is approximately to scale.
The inversion point in the JBP1 KO NEO amplicon corresponds to a repetitive sequence located downstream of the JBP1 locus
In order to determine the inversion point of the long inverted repeat structure, we generated 500 bp probes by PCR, scattered along the JBP1-containing chromosome of L.major Friedlin, the strain of reference for Leishmania sp., which has been completely sequenced (www.genedb.org) (see Figure 4 for the location of the different probes). L.major JBP1 is located in chromosome 9, 18 kb upstream of the right end of the chromosome, consistent with the subtelomeric location of JBP1 in L.tarentolae. The probes were used one by one on digested genomic DNA of TarIIWT and JBP1 KO NEO. Hybridization of L.major Friedlin probes with genomic DNA of L.tarentolae is possible due to the high homology between the sequences of both species (80–85% overall identity). Most probes hybridized to the same set of restriction fragments in the TarIIWT and JBP1 KO NEO, but more intensively with the DNA of the JBP1 KO NEO mutant (see Supplementary Data), implying that they are located in the amplicon, but not in the vicinity of the rearrangement point (Figure 4). Probes 10–13 hybridized with the same intensity with the DNA of TarIIWT and JBP1 KO NEO (see Supplementary Data), showing that they are located outside the rearrangement point (i.e. not in the amplicon) (Figure 4). However, a few probes located at respectively 35 kb (probe 4) and 41 kb (probes 7–9) downstream of the end of the JBP1 gene hybridized with a new band only present in the JBP1 KO NEO lane (see Supplementary Data), indicating that they are located at the rearrangement point of the amplicon (Figure 4). The extra bands hybridizing with both probe 4 and probe 9 in the JBP1 KO NEO mutant were cloned and sequenced. The corresponding sequences in TarIIWT were also cloned and sequenced in order to facilitate the identification of the rearrangement point (see Materials and Methods).
Comparison of the sequence of the rearrangement point of the JBP1 KO NEO amplicon with the corresponding sequences in TarIIWT revealed that the junction of both arms of the amplicon corresponded to a perfect repetitive sequence of 222 bp, located three times in a direct or reverse (palindromic) orientation downstream of the JBP1 locus. We name these repeats RS1 to RS3 for repeated sequence 1 to 3. Figure 4 presents the approximate location of these perfect repeats, which are respectively located at 25, 35 and 41 kb downstream of the end of the JBP1 gene, and Figure 5 shows the alignment of the sequences bordering the RS2 and RS3 repeats with the rearrangement point of the JBP1 KO NEO amplicon. We infer that the amplicon could have the structure shown in Figure 6A, and that it was formed by a DNA rearrangement involving the RS2 and RS3 repeats which are located, respectively, 35 and 41 kb downstream of the JBP1 locus (according to the genome project of L.major Friedlin) and oriented in a reverse (palindromic) fashion.
Figure 5 Sequence of the rearrangement point of the JBP1 KO NEO amplicon. Sequence alignment of the inversion point of the JBP1 KO NEO amplicon with the corresponding sequence in the JBP1 wild-type chromosome (i.e. sequences bordering the RS2 and RS3 repeats). The red capital letters depict the perfect homology between the inversion point of the amplicon (KO NEO line) and the RS2 and RS3 repeats. The capital letters show the homology between the sequence surrounding the rearrangement point and the corresponding sequences on the JBP1 chromosome. The orientation of the sequence is the one used in the bottom panel of Figure 6A, where the 5' end of the sequence corresponds to the left-hand side of the schematic (i.e. the RS2 sequence is presented in an inverted way compared to the wild-type orientation). Note the presence of a small mismatch inside the homologous sequence of RS2 and RS3. We have not determined whether this is a polymorphism between RS2 and RS3 or a point mutation introduced during our cloning procedure.
Figure 6 Structure of the JBP1 KO NEO amplicon. (A) Schematic representation of the right end of chromosome 9 of L.tarentolae (top panel) and of the JBP1 KO NEO amplicon (bottom). The amplicon is one long inverted repeat (depicted by the long broken arrows) with an RS repeat located at the rearrangement point (shown by the yellow star). The location of the A, B, C and D primers and their orientation is shown by the colored triangles. The small black triangles correspond to the telomeres. The RS repeats are depicted by the small arrows, the JBP1 and NEO genes by the colored boxes. We have not determined on which arms each gene lays. (B) PCR products obtained using two primers (A and B) oriented in the same way towards the RS2 and RS3 repeats (left panel) or using a combination of primers (C and D) located on both sides of the RS3 repeat (right panel). Lanes: 1, Marker; 2, TarIIWT; 3, JBP1 KO NEO; 4, JBP1 KO NEO revertant; 5, JBP1 KO NEO purified amplicon; and 6, TarIIWT + JBP1 KO NEO amplicon.
Confirmation of the structure of the JBP1 KO NEO amplicon
To verify that the amplicon has the structure summarized in Figure 6A, a PCR was done on genomic DNA of TarIIWT, JBP1 KO NEO and on the purified KO NEO amplicon using primers A and B oriented in the same direction as JBP1, and located at respectively 410 bp upstream of the RS2 repeat and 556 bp downstream of the RS3 repeat (Figure 6A). These two primers should be spaced by 1200 bp (counting the 222 bp RS repeat) on each side of the rearrangement point of the KO NEO amplicon if it has the structure presented in the Figure 6A. A PCR product having the expected size was obtained only using DNA from the JBP1 KO NEO cell line and the purified amplicon (alone or mixed with TarIIWT genomic DNA) (Figure 6B, left panel, lanes 3, 5 and 6). The PCR product was sequenced and was found to correspond to the predicted rearrangement point. No PCR product was obtained using a JBP1 KO NEO revertant cell line that lost the amplicon by culturing in the absence of drug selection (Figure 6B, left panel, lane 4). No PCR product was obtained either on the purified amplicon using primers C and D located on both sides of RS3 showing that at least one of these primers (primer D) is located outside the rearrangement point (Figure 6B, right panel, lane 5). Note that a PCR product is obtained with primers C and D on the genomic DNA of the JBP1 KO NEO mutant since this cell line still has a wild-type chromosome (Figure 6B, right panel, lane 3).
Structure of the JBP1 KO HYG:KO NEO and JBP1 KO PUR:KO HYG amplicons
The experiments done with the JBP1 KO NEO amplicon were repeated for the JBP1 KO HYG:KO NEO and JBP1 KO PUR:KO HYG cell lines and showed that these amplicons have a similar inverted repeat structure (data not shown). The inversion points in the JBP1 KO HYG:KO NEO and JBP1 KO PUR:KO HYG amplicons correspond to the same repeated sequence found at the junction of the KO NEO amplicon. However, in the former amplicons, the rearrangement happens to involve the RS1 and the RS2 repeats, also oriented in a reverse (palindromic) fashion, rather than RS2 and RS3 (see Supplementary Data). It should be noted that the length of the homology between the RS1 and RS2 repeat is 210 bp instead of 222 bp. Although we did not map the JBP1 KO HYG:KO PUR:KO NEO amplicon, we suspect that its rearrangement involved another repeat, called RS4, present more downstream of JBP1 than RS3 in L.major Freidlin (see also Discussion). This would explain its longer size compared to the other amplicons as shown by pulsed-field gradient electrophoresis (compare lane 8 versus lanes 5–7 in Figure 2).
The amplicons cannot be generated by drug pressure on a cell line having one marker integrated in the JBP1 locus
Generation of a linear DNA amplicon following a targeting event has not been described in Leishmania before, or in any other organism. To determine whether the amplification was caused by the transfection itself or was due to a highly recombinogenic property of the JBP1 locus, we tried to generate amplicons by increasing the drug pressure on the JBP1 KO HYG mutant, which has one copy of the HYG marker integrated in the JBP1 wild-type chromosome. In Leishmania, this procedure often results in DNA amplification of genes involved in the resistance to the drug used (5,16). We increased the hygromycin B concentration in which JBP1 KO HYG was cultured stepwise by a factor two until it reached 2 mg/ml (40 times the normal concentration at which L.tarentolae is cultured). After culture for over three months at this concentration, we found no amplicons by hybridization on digested genomic DNA and pulsed-field gel electrophoresis (data not shown), suggesting that the integration of the inactivation cassette facilitated the formation of the amplicon.
DISCUSSION
In our attempts to disrupt both alleles of JBP1, an essential gene in Leishmania, we obtained mutants that appear to have a supernumerary JBP1-containing chromosome, as exemplified by the JBP1 KO HYG:KO PUR mutant and mutants containing linear plasmids (amplicons) with a long inverted duplication (Figure 6A). Whereas mutants with supernumerary chromosomes have been generated before in Leishmania by attempts to knock out essential genes (1–3), the linear plasmids are new and their unusual structure provides detailed information about the DNA rearrangement that generated them.
Leishmania is known to readily amplify DNA segments if challenged, be it by increasing drug pressure (5,10,16,37), by attempts to target essential genes (1–3), by nutrient stress or by subcloning (38). Amplification is even frequently seen in unselected laboratory stock (15). Linear DNA amplicons with an inverted repeat structure similar to the one described here have been described before in Leishmania in response to drug selection (39) and in the absence of selection (40,41). The generation of circular DNA amplicons by the circularization of the -tubulin locus has also been reported following attempts to inactivate this essential gene in Leishmania (42). There are no previous reports, however, of linear amplicons with an inverted repeat structure, having different markers in the two arms generated following a targeting event. We attribute the rarity of this type of amplicon to four factors (see Figure 7): (i) the targeting of an essential gene; (ii) the subtelomeric location of JBP1, which may facilitate formation of stable linear amplicons; (iii) the inferred presence of a DNA replication origin upstream of JBP1; and (iv) the presence of inverted repeats not far downstream of JBP1 facilitating strand switch in the replication fork.
Figure 7 Postulated model to explain the generation of the JBP1 KO NEO amplicon. We propose that replication starts upstream of JBP1. Two copies of JBP1 are generated after the replication fork has passed the JBP1 locus. One of these copies is targeted by the inactivation construct, which replaces JBP1 by NEO (A). The perturbation caused by the targeting could cause a stalling of the replication fork. Reversal of the fork would enable the RS3 repeat of the leading strand to anneal with the not yet replicated RS2 repeat of the lagging strand (B). Replication would restart using the other strand as a template followed by the synthesis of the complementary strand (C). Pealing of the misreplicated section would lead to an amplicon having an inverted repeat structure and two single-stranded DNA having on one strand JBP1 and on the other the NEO marker (D). Annealing of the strands would create a heteroduplex that would be repaired (E). The newly introduced marker would end up in the JBP1 chromosome only in half of the events, depending on which strand is repaired (F). The leading strand of the replication fork is shown in orange, the lagging strand in blue. The RS and telomeric repeats, as well as the JBP1 and NEO genes, are depicted as in Figure 6.
Model for linear amplicon formation
Figure 7 presents our preferred model for the formation of the amplicons. In this model, the inactivation cassette integrates into one arm of a replication fork that recently passed through the JBP1 locus. Insertion of the cassette might cause the replication fork to stall . Reversal of the fork would allow the RS3 repeat of the leading strand to anneal to the single-stranded RS2 repeat of the lagging strand (Figure 7B). This template strand switch is necessary to explain the presence of two different markers at the two ends of most amplicons. Re-initiation of DNA synthesis in the replication fork, indicated very schematically by the broken line in Figure 7C, would result in an inverted repeat amplicon having in one arm a marker already integrated in the genome of the parasite and in the other the newly transfected marker (Figure 7C and D). Pealing of the amplicon would allow the parental strands to snap back together, giving rise to a partial heteroduplex that would need to be repaired, for instance by gene conversion (Figure 7E). If the choice of the strand to be repaired is random, half the targeting events will result in a new marker in the targeted chromosome. The other half will leave the original gene in place. This is what we find as shown in Table 1. The alleles that get amplified and end up in the linear plasmids are the one in the inactivation cassette and the one that was targeted by the cassette. Similar models involving a template switch in a DNA replication fork have been proposed to explain the formation of amplicons bearing an inverted duplication in mammalian cell lines (44–46).
Our model is based on the following assumptions. The fact that the two arms of the amplicons contain different marker genes (or a wild-type gene and a marker) suggests that the targeting construct must have hit replicating DNA (see Figure 7). Hence, we infer the presence of an origin of replication located upstream of JBP1. The presence of a short inverted repeat at the inversion point of the linear plasmid indicates that the long inverted repeat structure of the plasmid arose probably from a template switch of the leading strand in a replication fork, as depicted in Figure 7. We propose that the integration of the inactivation constructs affected the progression of the replication fork allowing the DNA strand switch at the origin of the amplicon formation, but this is speculative.
It is obvious that the time window available for the generation of these plasmids must be narrow: targeting must take place after the replication fork has passed through the JBP1 gene, but if the fork has progressed too far beyond the RS repeats, the strand switch is probably not possible anymore. The distance between JBP1 and the RS repeats is 40 kb. Assuming a similar rate of DNA replication fork progression in Leishmania and yeast, in which it is 4 kb/min (47), the time window for initiating plasmid formation is roughly 10 min, i.e. 1/36th of the 6 h cell cycle. Following targeting, the complex process of plasmid formation still has to start. It is therefore not surprising that plasmid formation is not a frequent event. Once the replication fork is far beyond the RS repeats, mutants can only arise by missegregation of the JBP1 chromosome resulting in a cell triploid for a single chromosome. The same result will be obtained after targeting of a pre-existing single-chromosome triploid (followed by the transmission of the three chromosomes to the daughter cells). Since our Leishmania strain is basically diploid, the frequency at which these triploids arise must be low. Indeed, we have found that mutants triploid for the JBP1 chromosome were generated at rates 100- to 1000-fold lower than the rate at which the single JBP1 mutants arose (results not shown).
Other models of DNA amplification
Besides our ‘template-switch in a replication fork’ model shown in Figure 7, three other possibilities can be considered that could explain the generation of plasmids. Reversal of the replication fork might generate single newly synthesized strands, which can partially snap back on themselves by means of the RS repeats. If this u-turn is followed by self-copying and replication of the duplex formed, a plasmid with the same marker on both arms will arise, as exemplified by the JBP1 KO HYG:KO NEO plasmid (which cannot be explained by our strand-switch model) (see Table 1). This mutant has NEO as the only marker present in the amplicon and might have been formed by an auto-annealing of the branch of the replication fork bearing the NEO marker. Formation of amplicons by auto-annealing of palindromic repeats followed by self-copying of one DNA strand is a preferred model to explain the formation of amplicons in Leishmania (5,9,17).
A second possibility to generate the amplicons is to cut off the parental DNA at the replication fork, as suggested by Nalbantoglu and Meuth (44). What argues against this alternative is that the replicating chromosome from which the plasmid arises remains intact (data not shown). The broken chromosome could repair itself, however, by copying homologous DNA . Since the intact chromosome contains either of the markers present in the amplicon and not the marker present in the other chromosome, this repair mechanism remains implausible, as it would have to involve the newly made plasmid.
A third popular model of DNA amplification involves the over-replication of a portion of the genome causing an ‘onion-like’ structure, followed by the resolution or recombination of the over-replicated segment (16,18,46). This model cannot easily explain our observation, however, that the inactivation construct that caused amplification is always found in the amplicon and only in 2 out of 4 events in the original JBP1 chromosome (Table 1). The ‘onion-skin’ model would rather predict that this marker is always present in the amplicon and not in the original chromosome, in the case where the construct would integrate in the over-replicated sequence, or that it is found in both the amplicon and the JBP1 chromosome, if the cassette integrated prior to over-replication and formation of the amplicon. It is nevertheless possible that both events occurred in our case, but we find that unlikely.
Amplicon copy number
In all mutants analyzed, the amplicons are present in multiple copies, even though one would expect that one copy of the marker replacing the JBP1 gene should be enough to confer resistance to the drugs at the concentration used. Several factors probably contribute to this plasmid multiplicity: one factor is the use of a NEO gene in our experiments that contains a mutation (see Materials and Methods), which gives a diminished activity towards geneticin (G418). This results in a marginal resistance if a single NEO copy is present per cell (data not shown). In the generation of the JBP1 KO NEO mutant, which has also one NEO copy in the JBP1 chromosome, we obviously used a concentration of G418 that was too high for the cells to survive without gene amplification. The mutant NEO gene confers high-level resistance to paromomycin, however, and when the JBP1 KO NEO cells were transferred to paromomycin they lost the plasmid. The multiplicity of linear plasmids can also be explained by the fact that our plasmids do not contain centromeres and are rapidly lost in the absence of selection, like other linear plasmids in Leishmania. Since cells without a single copy of the plasmid will die in the presence of selection (since most amplicons contain markers that are not present in the wild-type chromosome, see Table 1), there must be a strong selection in the mutant population for clones that can transmit at least one plasmid copy to each daughter cell. This must drive up the copy number and we think that this partially explains why our mutants contain on average 4–5 plasmid copies per cell. If plasmid segregation is random, a cell with 4 copies (before replication) will generate less than 1% of daughter cells without a single plasmid copy. Such a low loss should not measurably affect the growth of the population.
Implication of repetitive sequences in the generation of the linear amplicons
At the inversion point of the three amplicons that we mapped, we found repeats that we hold responsible for the template switch in a stalled replication fork. The tendency of autoannealing of repeats RS1-3 is limited as we were unable to generate amplicons after increasing the drug pressure, a process that often yields amplicons in Leishmania (5,16). This suggests that the integration of the targeting cassette facilitates the template switch at the origin of the amplicon formation, e.g. by stalling the replication fork. Repeats (direct or inverted) are often associated with DNA amplification in Leishmania. For instance, repeats were proposed to be responsible for the formation of circular and linear amplifications of the H locus arising from arsenite or methotrexate selection (9,10). Interestingly, a repetitive DNA sequence having a moderate homology (50%) with the sequence found at the rearrangement point of our amplicons (RS repeats) is also present at the same location in the chromosome 9 of L.major Friedlin where the JBP1 gene is located. A computer-based genome screen of the repetitive element found in chromosome 9 of L.major Friedlin revealed that this sequence is found in many other chromosomes (data not shown). Moxon et al. (48) have proposed that ‘mutation rates vary among sites in the genome and that this variation is adaptive, because it promotes evolutionary flexibility in the face of environmental change, without necessarily increasing the overall load of deleterious mutations’. These highly mutable loci are called contingency genes, because they ‘facilitate the efficient exploration of phenotypic solutions to unpredictable aspects of the host environment’ (48). Leishmania can respond to some environmental challenges with a specific type of mutation, i.e. gene amplification, and the repeats associated with this amplification fulfill a clear contingency function (9,16). Possibly, the RS repeats studied here fulfill a similar role. It could be worth looking at other already described amplicons in Leishmania to see whether the formation of some of them could be explained by these RS repeats. The contingency genes that might be amplified through these RS repeats remain to be identified.
SUPPLEMENTARY MATERIAL
Supplementary Material is available at NAR Online.
ACKNOWLEDGEMENTS
We are grateful to Cristiane Bentin Toaldo, Dr Jero Calafat, Hans Janssen, Dr Jos Jonkers, Rudo Kieft, Dr Juliette Moyersoen, Dr Hein te Riele, Dr Robert Sabatini, Dr Sebastian Ulbert and Zhong Yu for input and discussion, as well as to Dr Marc Ouellette for the generous gift of numerous plasmids and the cosmid bank. P.A. Genest is an NSERC (Natural Sciences and Engineering Research Council of Canada) PGS-B scholar. This work was supported in part by a grant from NWO-CW (99015). Funding to pay the Open Access publication charges for this article was provided by the above-mentioned organizations.
REFERENCES
Cruz, A.K., Titus, R., Beverley, S.M. (1993) Plasticity in chromosome number and testing of essential genes in Leishmania by targeting Proc. Natl Acad. Sci. USA, 90, 1599–1603 .
Mottram, J.C., McCready, B.P., Brown, K.G., Grant, K.M. (1996) Gene disruptions indicate an essential function for the LmmCRK1 cdc2-related kinase of Leishmania mexicana Mol. Microbiol., 22, 573–583 .
Dumas, C., Ouellette, M., Tovar, J., Cunningham, M.L., Fairlamb, A.H., Tamar, S., Olivier, M., Papadopoulou, B. (1997) Disruption of the trypanothione reductase gene of Leishmania decreases its ability to survive oxidative stress in macrophages EMBO J., 16, 2590–2598 .
Beverley, S.M., Coderre, J.A., Santi, D.V., Schimke, R.T. (1984) Unstable DNA amplifications in methotrexate-resistant Leishmania consist of extrachromosomal circles which relocalize during stabilization Cell, 38, 431–439 .
White, T.C., Fase-Fowler, F., van Luenen, H., Calafat, J., Borst, P. (1988) The H circles of Leishmania tarentolae are a unique amplifiable system of oligomeric DNAs associated with drug resistance J. Biol. Chem., 263, 16977–16983 .
Papadopoulou, B., Roy, G., Ouellette, M. (1992) A novel antifolate resistance gene on the amplified H circle of Leishmania EMBO J., 11, 3601–3608 .
Grondin, K., Kundig, C., Roy, G., Ouellette, M. (1998) Linear amplicons as precursors of amplified circles in methotrexate-resistant Leishmania tarentolae Nucleic Acids Res., 26, 3372–3378 .
Detke, S., Katakura, K., Chang, K.P. (1989) DNA amplification in arsenite-resistant Leishmania Exp. Cell. Res., 180, 161–170 .
Ouellette, M., Hettema, E., Wust, D., Fase-Fowler, F., Borst, P. (1991) Direct and inverted DNA repeats associated with P-glycoprotein gene amplification in drug resistant Leishmania EMBO J., 10, 1009–1016 .
Grondin, K., Papadopoulou, B., Ouellette, M. (1993) Homologous recombination between direct repeat sequences yields P-glycoprotein containing amplicons in arsenite resistant Leishmania Nucleic Acids Res., 21, 1895–1901 .
Grondin, K., Haimeur, A., Mukhopadhyay, R., Rosen, B.P., Ouellette, M. (1997) Co-amplification of the gamma-glutamylcysteine synthetase gene gsh1 and of the ABC transporter gene pgpA in arsenite-resistant Leishmania tarentolae EMBO J., 16, 3057–3065 .
Wilson, K., Collart, F.R., Huberman, E., Stringer, J.R., Ullman, B. (1991) Amplification and molecular cloning of the IMP dehydrogenase gene of Leishmania donovani J. Biol. Chem., 266, 1665–1671 .
Detke, S., Chaudhuri, G., Kink, J.A., Chang, K.P. (1988) DNA amplification in tunicamycin-resistant Leishmania mexicana. Multicopies of a single 63-kilobase supercoiled molecule and their expression J. Biol. Chem., 263, 3418–3424 .
Henderson, D.M., Sifri, C.D., Rodgers, M., Wirth, D.F., Hendrickson, N., Ullman, B. (1992) Multidrug resistance in Leishmania donovani is conferred by amplification of a gene homologous to the mammalian mdr1 gene Mol. Cell. Biol., 12, 2855–2865 .
Beverley, S.M. and Coburn, C.M. (1990) Recurrent de novo appearance of small linear DNAs in Leishmania major and relationship to extra-chromosomal DNAs in other species Mol. Biochem. Parasitol., 42, 133–141 .
Beverley, S.M. (1991) Gene amplification in Leishmania Annu. Rev. Microbiol., 45, 417–444 .
Papadopoulou, B., Roy, G., Ouellette, M. (1993) Frequent amplification of a short chain dehydrogenase gene as part of circular and linear amplicons in methotrexate resistant Leishmania Nucleic Acids Res., 21, 4305–4312 .
Stark, G.R. and Wahl, G.M. (1984) Gene amplification Annu. Rev. Biochem., 53, 447–491 .
Gommers-Ampt, J.H., van Leeuwen, F., de Beer, A.L., Vliegenthart, J.F., Dizdaroglu, M., Kowalak, J.A., Crain, P.F., Borst, P. (1993) Beta-D-glucosyl-hydroxymethyluracil: a novel modified base present in the DNA of the parasitic protozoan T.brucei Cell, 75, 1129–1136 .
van Leeuwen, F., Taylor, M., Mondragon, A., Moreau, H., Gibson, W., Kieft, R., Borst, P. (1998) Beta-D-glucosyl-hydroxymethyluracil is a conserved DNA modification in kinetoplastid protozoans and is abundant in their telomeres Proc. Natl Acad. Sci. USA, 95, 2366–2371 .
Dooijes, D., Chaves, I., Kieft, R., Dirks-Mulder, A., Martin, W., Borst, P. (2000) Base J originally found in kinetoplastida is also a minor constituent of nuclear DNA of Euglena gracilis Nucleic Acids Res., 28, 3017–3021 .
van Leeuwen, F., Kieft, R., Cross, M., Borst, P. (2000) Tandemly repeated DNA is a target for the partial replacement of thymidine by ?-D-glucosyl-hydroxymethyluracil in Trypanosoma brucei Mol. Biochem. Parasitol., 109, 133–145 .
Cross, M., Kieft, R., Sabatini, R., Wilm, M., de Kort, M., van der Marel, G.A., van Boom, J.H., van Leeuwen, F., Borst, P. (1999) The modified base J is the target for a novel DNA-binding protein in kinetoplastid protozoans EMBO J., 18, 6573–6581 .
Sabatini, R., Meeuwenoord, N., van Boom, J.H., Borst, P. (2002) Recognition of base J in duplex DNA by J-binding protein J. Biol. Chem., 277, 958–966 .
Sabatini, R., Meeuwenoord, N., van Boom, J.H., Borst, P. (2002) Site-specific interactions of JBP with base and sugar moieties in duplex J-DNA. Evidence for both major and minor groove contacts J. Biol. Chem., 277, 28150–28156 .
Cross, M., Kieft, R., Dirks-Mulder, A., Chaves, I., Borst, P. (2002) J-binding protein increases the level and retention of the unusual base J in trypanosome DNA Mol. Microbiol., 46, 37–47 .
Brun, R. and Schonenberger, M. (1979) Cultivation and in vitro cloning of procyclic culture forms of Trypanosoma brucei in a semi-defined medium Acta Trop., 36, 289–292 .
Munoz-Jordan, J.L. and Cross, G.A. (2001) Telomere shortening and cell cycle arrest in Trypanosoma brucei expressing human telomeric repeat factor TRF1 Mol. Biochem. Parasitol., 114, 169–181 .
Brochu, C., Haimeur, A., Ouellette, M. (2004) The heat shock protein HSP70 and heat shock cognate protein HSC70 contribute to antimony tolerance in the protozoan parasite Leishmania Cell. Stress. Chap., 9, 294–303 .
Papadopoulou, B., Roy, G., Ouellette, M. (1994) Autonomous replication of bacterial DNA plasmid oligomers in Leishmania Mol. Biochem. Parasitol., 65, 39–49 .
Sambrook, J., Fritsch, E., Maniatis, T. Molecular Cloning: A Laboratory Manual, (1989) Cold Spring Harbor, NY Cold Spring Harbor Press .
van der Ploeg, L.H., Schwartz, D.C., Cantor, C.R., Borst, P. (1984) Antigenic variation in Trypanosoma brucei analyzed by electrophoretic separation of chromosome-sized DNA molecules Cell, 37, pp. 77–84 .
van der Ploeg, L.H., Cornelissen, A.W., Barry, J.D., Borst, P. (1984) Chromosomes of kinetoplastida EMBO J., 3, 3109–3115 .
Hightower, R.C., Bliska, J.B., Cozzarelli, N.R., Santi, D.V. (1989) Analysis of amplified DNAs from drug-resistant Leishmania by orthogonal-field-alternation gel electrophoresis. The effect of size and topology on mobility J. Biol. Chem., 264, 2979–2984 .
Bernards, A., Michels, P.A., Lincke, C.R., Borst, P. (1983) Growth of chromosome ends in multiplying trypanosomes Nature, 303, 592–597 .
Sunkin, S., Kiser, P., Myler, P., Stuart, K. (2000) The size difference between Leishmania major Friedlin chromosome one homologues is localized to sub-telomeric repeats at one chromosomal end Mol. Biochem. Parasitol., 109, 1–15 .
Grondin, K., Roy, G., Ouellette, M. (1996) Formation of extrachromosomal circular amplicons with direct or inverted duplications in drug-resistant Leishmania tarentolae Mol. Cell. Biol., 16, 3587–3595 .
Rovai, L., Tripp, C., Stuart, K., Simpson, L. (1992) Recurrent polymorphisms in small chromosomes of Leishmania tarentolae after nutrient stress or subcloning Mol. Biochem. Parasitol., 50, 115–125 .
Dubessay, P., Ravel, C., Bastien, P., Lignon, M.F., Ullman, B., Pages, M., Blaineau, C. (2001) Effect of large targeted deletions on the mitotic stability of an extra chromosome mediating drug resistance in Leishmania Nucleic Acids Res., 29, 3231–3240 .
Navarro, M., Liu, J., Muthui, D., Ortiz, G., Segovia, M., Hamers, R. (1994) Inverted repeat structure and homologous sequences in the LD1 amplicons of Leishmania spp Mol. Biochem. Parasitol., 68, 69–80 .
Ortiz, G. and Segovia, M. (1996) Characterisation of the novel junctions of two minichromosomes of Leishmania major Mol. Biochem. Parasitol., 82, 137–144 .
Curotto de Lafaille, M.A. and Wirth, D.F. (1992) Creation of null/+ mutants of the -tubulin gene in Leishmania enriettii by gene cluster deletion J. Biol. Chem., 267, 23839–23846 .
Michel, B., Grompone, G., Flores, M.J., Bidnenko, V. (2004) Multiple pathways process stalled replication forks Proc. Natl Acad. Sci. USA, 101, 12783–12788 .
Nalbantoglu, J. and Meuth, M. (1986) DNA amplification-deletion in a spontaneous mutation of the hamster aprt locus: structure and sequence of the novel joint Nucleic Acids Res., 14, 8361–8371 .
Hyrien, O., Debatisse, M., Buttin, G., de Saint Vincent, B.R. (1988) The multicopy appearance of a large inverted duplication and the sequence at the inversion joint suggest a new model for gene amplification EMBO J., 7, 407–417 .
Stark, G.R., Debatisse, M., Giulotto, E., Wahl, G.M. (1989) Recent progress in understanding mechanisms of mammalian DNA amplification Cell, 57, 901–908 .
Reynolds, A.E., McCarroll, R.M., Newlon, C.S., Fangman, W.L. (1989) Time of replication of ARS elements along yeast chromosome III Mol. Cell. Biol., 9, 4488–4494 .
Moxon, E.R., Rainey, P.B., Nowak, M.A., Lenski, R.E. (1994) Adaptive evolution of highly mutable loci in pathogenic bacteria Curr. Biol., 4, 24–33 .
Laban, A., Tobin, J.F., Curotto de Lafaille, M.A., Wirth, D.F. (1990) Stable expression of the bacterial neor gene in Leishmania enriettii Nature, 343, 572–574 .(Paul-André Genest, Bas ter Riet, Carole )
*To whom correspondence should be addressed. Tel: +31 020 512 2880; Fax: +31 020 669 1383; Email: p.borst@nki.nl
ABSTRACT
Attempts to inactivate an essential gene in the protozoan parasite Leishmania have often led to the generation of extra copies of the wild-type alleles of the gene. In experiments with Leishmania tarentolae set up to disrupt the gene encoding the J-binding protein 1 (JBP1), a protein binding to the unusual base ?-D-glucosyl-hydroxymethyluracil (J) of Leishmania, we obtained JBP1 mutants containing linear DNA elements (amplicons) of 100 kb. These amplicons consist of a long inverted repeat with telomeric repeats at both ends and contain either the two different targeting cassettes used to inactivate JBP1, or one cassette and one JBP1 gene. Each long repeat within the linear amplicons corresponds to sequences covering the JBP1 locus, starting at the telomeres upstream of JBP1 and ending in a 220 bp sequence repeated in an inverted (palindromic) orientation downstream of the JBP1 locus. We propose that these amplicons have arisen by a template switch inside a DNA replication fork involving the inverted DNA repeats and helped by the gene targeting.
INTRODUCTION
Leishmania sp. is a unicellular protozoan parasite belonging to one of the oldest eukaryotic lineages, the Kinetoplastida. Leishmania is known for its genomic plasticity. Changes in the karyotype of Leishmania, such as triploidy for one chromosome or genome-wide polyploidy, have been reported following attempts to inactivate essential genes (1–3). DNA amplification can also be induced in this parasite by selection with a variety of drugs including methotrexate (4–7), oxyanions (antimony, arsenite) (8–11), mycophenolic acid (12), tunicamycin (13) and vinblastine (14). Amplicons have even been found in unselected laboratory isolates (15).
The structure of the amplicons in Leishmania varies: they can be circular or linear, and can contain direct or inverted DNA repeats. They are usually the product of a conservative amplification with no alteration in the source chromosome and are frequently lost in the absence of the selection that induced the amplification (16). The preferred model of DNA amplification in Leishmania resulting in inverted DNA repeats involves the self-annealing of inverted (palindromic) DNA repeats in a region of single-stranded DNA being replicated, followed by the self-replication of one of the DNA strands and the consecutive synthesis of its complementary strand (5,9,17). Whereas in mammalian cells DNA amplicons are often the product of many successive rearrangements (18), DNA amplification in Leishmania normally occurs in one step. Hence, the amplicons found in Leishmania usually have a simple structure, which makes this parasite a good model for the study of gene amplification (16).
Base J or ?-D-glucosyl-hydroxymethyluracil is a unique hypermodified base only present in the nuclear DNA of kinetoplastid parasites and Euglena (19–21). It replaces 0.5% of thymine in the genome of Trypanosoma, mostly in repetitive sequences (22), including the telomeric repeats (20). The enzymes involved in J biosynthesis have not been isolated yet. Hence, we have not been able to produce parasites without J and the function of J remains a matter of speculation. We isolated a J-binding protein, however, named JBP1 for J-binding protein 1 (23), which binds to J-containing duplex DNA with high specificity (24,25). JBP1 null trypanosomes are viable in vitro and in mice, and have no obvious defect in the stability of their DNA repeats or in gene expression. Interestingly, they have 20-fold less J, suggesting that JBP1 is involved in the maintenance of J levels in T.brucei (26).
To get more insight into the function of JBP1, we tried to inactivate the JBP1 gene by gene targeting (via homologous recombination) in the promastigote (insect) form of Leishmania tarentolae, a parasite of lizards. We could easily inactivate one allele, but all attempts to make a JBP1 null Leishmania resulted in cell lines maintaining a wild-type allele, suggesting that JBP1 is essential in this parasite. Some of these lines were found to contain new linear DNA plasmids, harboring the targeting constructs used to inactivate JBP1. We present here the structure of these plasmid amplicons and a model for their generation during the targeting of the JBP1 locus.
MATERIALS AND METHODS
Culture cell lines and transfections
The L.tarentolae TarIIWT cell line (5) was cultured in SDM 79 medium (27). Transfections were done as described previously (6). The cells were selected at the following drug concentrations: 20 μg puromycin (Sigma) per ml; 20 μg neomycin (Gibco) per ml; 100 μg hygromycin B (Roche) per ml; 200 μg paromomycin (Sigma) per ml. The frequency of transfection was determined by serial dilution of the cells after transfection. The reversion experiment was done by culturing the amplicon-containing cell lines in the absence of drug pressure. The DNA measurement by flow-cytometry was done as described in Munoz-Jordan and Cross (28).
Cloning procedures and inactivation constructs
The cloning of the L.tarentolae JBP1 gene was already described in Cross et al. (23). A cosmid containing Ltar JBP1 was isolated by screening a cosmid library reported in Brochu et al. (29), and a 7.2 kb fragment containing JBP1 was digested with HindIII–ClaI and cloned in pBluescript (Stratagene) giving the construct Ltar JBP1-pBluescript. To make the KO NEO and KO HYG inactivation constructs, the NEO and HYG markers, both preceded by a 90 bp polypyrimidine stretch (Y), were first digested out of the PSPY-NEO and PSPY-HYG vectors (30) by a XbaI–BglII digest, and cloned in PSL1180 (Amersham) giving, respectively, the PSLY-NEO and PSLY-HYG vectors. A 2.4 kb ApaI fragment covering the 3'-untranslated region (3'-UTR) part of Ltar JBP1 was isolated by an ApaI digest of Ltar JBP1-pBluescript, and cloned downstream of the NEO or HYG markers by insertion in the ApaI site of PSLY-NEO or PSLY-HYG. A 2 kb HindIII–XbaI fragment covering the 5'-UTR of Ltar JBP1 was then cloned into the HindIII–SpeI digested PSLY-NEO + 3'-UTR or PSLY-HYG + 3'-UTR vectors, leading to the Ltar JBP1 KO NEO and KO HYG constructs. The cassettes were taken out of the PSL1180 backbone by a HindIII–BstbI double digest prior to transfection. Note that the NEO marker used by us has a point mutation at position 406 (bp) (C A mutation, Gln Lys). To construct the KO PUR inactivation cassette, the gene coding for the puromycin acetyl transferase was HindIII–ClaI cloned after the second -tubulin intergenic region (containing processing signals) of pGEM 7Zf -NEO- (6), resulting in the construct pGEM 7Zf -NEO--PUR. The -PUR cassette was then subcloned into the EcoRI–ClaI sites of pSP72 (Promega), and finally integrated between the 5'-UTR and 3'-UTR sequences of JBP1 by an XbaI–BglII digest of the Ltar JBP1 KO NEO construct (which removes the YNEO cassette). The Ltar JBP1 KO PUR construct was linearized with NcoI prior to transfection. The rearrangement point of the JBP1 KO NEO amplicon was cloned by ligating 2.0 kb EcoRV fragments and 3.0 kb BamHI fragments of digested genomic DNA of JBP1 KO NEO into pBluescript (i.e. fragments of the size of the rearrangement point for these digests). The clones of interest were isolated by screening with the probe 9. Similar experiments were done in order to clone the rearrangement point of the JBP1 KO HYG:KO NEO and KO PUR:KO HYG amplicons. The sequences corresponding to the rearrangement point in the TarIIWT cell line were isolated following similar procedures.
DNA extraction, blotting and hybridization
DNA was gel-extracted using a QIAGEN gel-extraction kit. DNA electrophoresis and blotting were performed using standard conditions (31). Membranes were hybridized with radioactively labeled probes using the Prime-it random primer labeling kit (Stratagene) at 42°C in a formamide buffer , 50 mM NaPi (Na2HPO4, NaH2PO4), 5 mM EDTA, 0.1% SDS and 0.25 mg salmon sperm DNA per ml]. Membranes were washed with 3x SSC, 0.1% SDS and 0.1x SSC and 0.1% SDS. The JBP1 3'-UTR probe corresponds to the 2.4 kb ApaI fragment described in the Cloning section. The XbaI–BglII YNEO and YHYG cassettes (see Cloning section) were used as probes to check for the site of integration of the constructs. The telomeric probe was described in van Leeuwen et al. (20). Hybridization with the telomeric probe was done at 65°C in a water-based buffer . Washes were done with 6x SSC and 0.1% SDS.
Chromosome separation and amplicon extraction
DNA plugs were made as described in van der Ploeg et al. (32). Chromosomes were separated using a CHEF-DRII apparatus (Bio-rad) under the following conditions: 24 h, 6 V/cm, initial switch time: 35 s, final switch time: 120 s. Amplicons were extracted from low melting point agarose gels by agarase (New England Biolabs) treatment following the indications of the manufacturer.
Sequencing
The cosmid covering the JBP1 locus was sequenced by shot-gun sequencing by GATC Biotech AG (Jakob-Stadler Platz 7, D-78467 Konstanz, Germany). The rearrangement point and the corresponding wild-type sequence were obtained by traditional sequencing, using an ABI Prism 3700 DNA Analyzer (Applied Biosystems). The sequence data have been submitted to the GenBank/EMBL/DDBJ database under the accession numbers: AY842844, AY842845, AY842846, AY842847, AY842848 and AY842849.
Design and isolation of the probes scattered on the JBP1 chromosome
Leishmania major Friedlin JBP1 was localized at 18 kb downstream of one of the ends of chromosome 9 by a computer-based screen (BLAST search). Oligonucleotides were designed in order to make 500 bp (or 1 kb for probe 7) PCR fragments adjacently located or spaced by 500 bp or 1 kb, starting at 20 kb downstream of JBP1 (going towards the distant telomere). PCR fragments were made using genomic DNA of L.major Friedlin as a template. Probe 1 was isolated by PCR using primers that were designed based on the sequence of a cosmid covering the JBP1 locus and adjacent sequence of L.tarentolae. The PCR products were purified, labeled and hybridized on digested genomic DNA of L.tarentolae. The list of oligonucleotides used during this study is available upon request.
Quantitation of the copy number of the amplicon
The copy number of the amplicons in the various mutants was determined by quantitative autoradiography with a FLA-3000 apparatus (Fuji) using the Bas reader version 3.14 and Tina version 2.09 softwares. The signal obtained by hybridization on digested genomic DNA of TarIIWT was compared with the one of the cell lines with amplicons, using probes located in the amplicon. Probes located upstream of the rearrangement point (i.e. not present in the amplicon) were used in order to correct for DNA loading.
RESULTS
JBP1 null mutants cannot be generated in L.tarentolae
In order to inactivate JBP1 in Leishmania, targeting constructs were made containing a selection marker (NEO, HYG or PUR) flanked by the 5'-UTR and 3'-UTR sequences of the L.tarentolae (Ltar) JBP1 gene, and these were transfected into the Leishmania TarIIWT cell line. Integration of the cassettes into the JBP1 locus was verified by hybridization, as schematically shown in Figure 1A. Transfection of the KO HYG and KO PUR constructs resulted in integration into the JBP1 locus as shown by the appearance of new bands hybridizing with the JBP1 3'-UTR probe of, respectively, 3.4 and 7.0 kb in addition to the 4.7 kb wild-type band in the KO HYG and KO PUR cell lines (Figure 1B, lanes 2 and 3). The inactivation of the second JBP1 allele by transfection of the KO PUR construct into the KO HYG mutant failed, however, and resulted in a cell line containing two inactivated alleles, but maintaining a JBP1 wild-type allele (Figure 1B, lane 4). All other attempts to obtain JBP1 null mutants resulted in cell lines maintaining a wild-type allele. Failure to obtain a double gene inactivation suggests that JBP1 is an essential gene in Leishmania, as also reported for the DHFR (dihydrofolate reductase) gene (1) and for the TR (trypanothione reductase) gene (3). Our mapping data indicate that the wild-type allele in the JBP1 KO HYG:KO PUR mutant is present in a supernumerary JBP1-containing chromosome, since no DNA rearrangement was detected in the vicinity of the JBP1 locus. Moreover, no ectopic insertion of JBP1 was found into other chromosomes and the overall DNA content (plo?dy) of the JBP1 KO HYG:KO PUR mutant was normal as determined by flow-cytometry (data not shown).
Figure 1 Targeting of the JBP1 locus and analysis of the mutant cell lines obtained. (A) Map of the JBP1 locus before and after integration of the inactivation cassettes. Sequences flanking the JBP1 locus were cloned upstream and downstream of selection markers. Sequences Y (polypyrimidine stretch) (30) and TUB (intergenic region of the alpha tubulin array of Leishmania enriettii) (49) were used for mRNA processing. The location of the NcoI sites and the JBP1 3'-UTR probe are depicted. The drawing is approximately to scale. (B) Southern blot of genomic DNA digested with NcoI. The blot was hybridized with the JBP1 3'-UTR probe. Lanes: 1, TarIIWT; 2, JBP1 KO HYG; 3, JBP1 KO PUR; 4, JBP1 KO HYG:KO PUR; 5, JBP1 KO NEO; 6, JBP1 KO HYG:KO NEO; 7, JBP1 KO PUR:KO HYG; and 8, JBP1 KO HYG:KO PUR:KO NEO. The alleles present in the different strains are indicated above the lanes. The asterisk identifies the amplified alleles. The wild-type 4.7 kb band can be used as a loading control, except in lane 5 where the loading can be checked by comparison of the amplified 2.7 kb band with lane 6.
Targeting of the JBP1 gene of L.tarentolae results in unexpected allelic ratios in some JBP1 mutants
While trying to inactivate JBP1, we also generated mutants with a higher copy number of the inactivated alleles or the wild-type allele (compare the wild-type and inactivated alleles in lanes 1–4 versus lanes 5–8 in Figure 1B). Such unexpected allelic ratios were observed in the JBP1 KO NEO, KO HYG:KO NEO, KO PUR:KO HYG cell lines and in a JBP1 KO HYG:KO PUR:KO NEO cell line in which we tried to inactivate the remaining wild-type allele of the JBP1 KO HYG:PUR mutant using the KO NEO construct. Most mutants have two different amplified alleles, one with the marker inserted when the amplification occurred and the other already integrated in the genome prior to the transfection that induced the amplification.
Presence of small linear amplicons in some JBP1 mutants
A chromosomal size fractionation by pulsed-field gradient electrophoresis, which allows the separation of all chromosome-sized DNA molecules of Leishmania (10,17,33,34), revealed the presence of new additional small linear chromosomes of 100 kb only in the cell lines with an abnormal JBP1 allelic ratio (Figure 2). Hybridization of a blot of the pulsed-field gel showed that the amplified markers were present in these small chromosomes (amplicons) (Table 1). The marker in the inactivation cassette used in the transfection that resulted in the appearance of the amplicons was always found in the amplicons, and in two out of four mutants was also in the JBP1-containing chromosome, as summarized in Table 1. The linear amplicons were only stable in the presence of drug selection and they were rapidly diluted out when we cultured the cell lines in the absence of drug pressure (data not shown). In the presence of drug selection, the copy number of the amplicons is 4–5 copies per cell as determined by quantitative autoradiography.
Figure 2 Pulsed-field gel stained with ethidium bromide showing separation of chromosome-sized DNA of L.tarentolae and the amplicon bands present in the JBP1 mutants. The arrow is pointing at the region containing the linear amplicons. Lanes M: Saccharomyces cerevisiae marker; 1, TarIIWT; 2, JBP1 KO HYG; 3, JBP1 KO PUR; 4, JBP1 KO HYG:KO PUR; 5, JBP1 KO NEO; 6, JBP1 KO HYG:KO NEO; 7, JBP1 KO PUR:KO HYG; and 8, JBP1 KO HYG:KO PUR:KO NEO.
Table 1 Genotype of the cell lines with a linear amplicon
The JBP1 KO NEO amplicon consists of a long inverted repeat
The fact that some amplicons contain two copies of the JBP1 locus, each with a different marker or in one case with a marker gene and a wild-type JBP1 allele (see Table 1), suggested that the amplicon contains a DNA duplication covering the JBP1 locus. To verify this, the JBP1 KO NEO amplicon was extracted from a pulsed-field gradient gel as in Figure 2 and roughly mapped (data not shown). The sum of the size of all the bands arising from several digests fell short by almost half the size of the amplicon estimated by pulsed-field electrophoresis (100 kb) (data not shown), indicating that the JBP1 KO NEO amplicon contained long sequences that were duplicated. The digested amplicon was hybridized with a telomeric probe, which recognized two bands in all the digestions tested, suggesting that the linear amplicons had telomeres at both ends (Figure 3A). A probe located 13 kb upstream of the beginning of the Ltar JBP1 locus (named probe 1, see Figure 4), (see Materials and Methods for a description of how probe 1 was isolated), recognized the same fragments that hybridized with the telomeric probe in some digests, indicating that the location of JBP1 is subtelomeric (see for example the EcoRV digest in lane 2 in Figure 3A and B). These data also suggested that the two ends of the amplicon were located upstream of JBP1, implying that the JBP1 KO NEO amplicon had an inverted repeat structure. The differences in the size of the fragments recognized by the telomeric probe and probe 1 in the BamHI digest (lane 1 in Figure 3A and B) can be explained by the presence of restriction sites between the JBP1 upstream probe (probe 1) and the telomeres, and by the presence of the KO NEO cassette inserted into the JBP1 locus. The presence of two telomeric fragments is probably due to a difference in the number of telomeric repeats at the two ends of the amplicon, as the number of telomeric repeats is known to be highly variable in kinetoplastid parasites (35,36).
Figure 3 Analysis of the telomeric ends of the JBP1 KO NEO amplicon. Southern blot on digested JBP1 KO NEO amplicon isolated from the pulsed-field gel. The blots were hybridized with a telomeric probe (A) and with probe 1 (see Figure 4), located 13 kb upstream of the JBP1 locus (B). Lane 1, BamHI; and lane 2, EcoRV.
Figure 4 Schematic representation of the results of the hybridization experiments done to detect the rearrangement point in the JBP1 KO NEO amplicon. The small rectangles represent the approximate location in the JBP1 chromosome (Chr 9) of the different probes that were generated by PCR on genomic DNA of L.major Friedlin. The white rectangles correspond to the probes that are not present in the amplicon, the black rectangles depict the probes that are present in the amplicon. The gray rectangles correspond to the probes that detect the rearrangement point. The bars indicate the approximate locations of the EcoRV sites around the JBP1 locus in L.tarentolae. The small arrows indicate the location of the RS1, RS2 and RS3 repeats. Their orientation is indicated by the direction of the arrow. The triangles represent the telomeric repeats. The JBP1 gene is depicted by the yellow rectangle. The drawing is approximately to scale.
The inversion point in the JBP1 KO NEO amplicon corresponds to a repetitive sequence located downstream of the JBP1 locus
In order to determine the inversion point of the long inverted repeat structure, we generated 500 bp probes by PCR, scattered along the JBP1-containing chromosome of L.major Friedlin, the strain of reference for Leishmania sp., which has been completely sequenced (www.genedb.org) (see Figure 4 for the location of the different probes). L.major JBP1 is located in chromosome 9, 18 kb upstream of the right end of the chromosome, consistent with the subtelomeric location of JBP1 in L.tarentolae. The probes were used one by one on digested genomic DNA of TarIIWT and JBP1 KO NEO. Hybridization of L.major Friedlin probes with genomic DNA of L.tarentolae is possible due to the high homology between the sequences of both species (80–85% overall identity). Most probes hybridized to the same set of restriction fragments in the TarIIWT and JBP1 KO NEO, but more intensively with the DNA of the JBP1 KO NEO mutant (see Supplementary Data), implying that they are located in the amplicon, but not in the vicinity of the rearrangement point (Figure 4). Probes 10–13 hybridized with the same intensity with the DNA of TarIIWT and JBP1 KO NEO (see Supplementary Data), showing that they are located outside the rearrangement point (i.e. not in the amplicon) (Figure 4). However, a few probes located at respectively 35 kb (probe 4) and 41 kb (probes 7–9) downstream of the end of the JBP1 gene hybridized with a new band only present in the JBP1 KO NEO lane (see Supplementary Data), indicating that they are located at the rearrangement point of the amplicon (Figure 4). The extra bands hybridizing with both probe 4 and probe 9 in the JBP1 KO NEO mutant were cloned and sequenced. The corresponding sequences in TarIIWT were also cloned and sequenced in order to facilitate the identification of the rearrangement point (see Materials and Methods).
Comparison of the sequence of the rearrangement point of the JBP1 KO NEO amplicon with the corresponding sequences in TarIIWT revealed that the junction of both arms of the amplicon corresponded to a perfect repetitive sequence of 222 bp, located three times in a direct or reverse (palindromic) orientation downstream of the JBP1 locus. We name these repeats RS1 to RS3 for repeated sequence 1 to 3. Figure 4 presents the approximate location of these perfect repeats, which are respectively located at 25, 35 and 41 kb downstream of the end of the JBP1 gene, and Figure 5 shows the alignment of the sequences bordering the RS2 and RS3 repeats with the rearrangement point of the JBP1 KO NEO amplicon. We infer that the amplicon could have the structure shown in Figure 6A, and that it was formed by a DNA rearrangement involving the RS2 and RS3 repeats which are located, respectively, 35 and 41 kb downstream of the JBP1 locus (according to the genome project of L.major Friedlin) and oriented in a reverse (palindromic) fashion.
Figure 5 Sequence of the rearrangement point of the JBP1 KO NEO amplicon. Sequence alignment of the inversion point of the JBP1 KO NEO amplicon with the corresponding sequence in the JBP1 wild-type chromosome (i.e. sequences bordering the RS2 and RS3 repeats). The red capital letters depict the perfect homology between the inversion point of the amplicon (KO NEO line) and the RS2 and RS3 repeats. The capital letters show the homology between the sequence surrounding the rearrangement point and the corresponding sequences on the JBP1 chromosome. The orientation of the sequence is the one used in the bottom panel of Figure 6A, where the 5' end of the sequence corresponds to the left-hand side of the schematic (i.e. the RS2 sequence is presented in an inverted way compared to the wild-type orientation). Note the presence of a small mismatch inside the homologous sequence of RS2 and RS3. We have not determined whether this is a polymorphism between RS2 and RS3 or a point mutation introduced during our cloning procedure.
Figure 6 Structure of the JBP1 KO NEO amplicon. (A) Schematic representation of the right end of chromosome 9 of L.tarentolae (top panel) and of the JBP1 KO NEO amplicon (bottom). The amplicon is one long inverted repeat (depicted by the long broken arrows) with an RS repeat located at the rearrangement point (shown by the yellow star). The location of the A, B, C and D primers and their orientation is shown by the colored triangles. The small black triangles correspond to the telomeres. The RS repeats are depicted by the small arrows, the JBP1 and NEO genes by the colored boxes. We have not determined on which arms each gene lays. (B) PCR products obtained using two primers (A and B) oriented in the same way towards the RS2 and RS3 repeats (left panel) or using a combination of primers (C and D) located on both sides of the RS3 repeat (right panel). Lanes: 1, Marker; 2, TarIIWT; 3, JBP1 KO NEO; 4, JBP1 KO NEO revertant; 5, JBP1 KO NEO purified amplicon; and 6, TarIIWT + JBP1 KO NEO amplicon.
Confirmation of the structure of the JBP1 KO NEO amplicon
To verify that the amplicon has the structure summarized in Figure 6A, a PCR was done on genomic DNA of TarIIWT, JBP1 KO NEO and on the purified KO NEO amplicon using primers A and B oriented in the same direction as JBP1, and located at respectively 410 bp upstream of the RS2 repeat and 556 bp downstream of the RS3 repeat (Figure 6A). These two primers should be spaced by 1200 bp (counting the 222 bp RS repeat) on each side of the rearrangement point of the KO NEO amplicon if it has the structure presented in the Figure 6A. A PCR product having the expected size was obtained only using DNA from the JBP1 KO NEO cell line and the purified amplicon (alone or mixed with TarIIWT genomic DNA) (Figure 6B, left panel, lanes 3, 5 and 6). The PCR product was sequenced and was found to correspond to the predicted rearrangement point. No PCR product was obtained using a JBP1 KO NEO revertant cell line that lost the amplicon by culturing in the absence of drug selection (Figure 6B, left panel, lane 4). No PCR product was obtained either on the purified amplicon using primers C and D located on both sides of RS3 showing that at least one of these primers (primer D) is located outside the rearrangement point (Figure 6B, right panel, lane 5). Note that a PCR product is obtained with primers C and D on the genomic DNA of the JBP1 KO NEO mutant since this cell line still has a wild-type chromosome (Figure 6B, right panel, lane 3).
Structure of the JBP1 KO HYG:KO NEO and JBP1 KO PUR:KO HYG amplicons
The experiments done with the JBP1 KO NEO amplicon were repeated for the JBP1 KO HYG:KO NEO and JBP1 KO PUR:KO HYG cell lines and showed that these amplicons have a similar inverted repeat structure (data not shown). The inversion points in the JBP1 KO HYG:KO NEO and JBP1 KO PUR:KO HYG amplicons correspond to the same repeated sequence found at the junction of the KO NEO amplicon. However, in the former amplicons, the rearrangement happens to involve the RS1 and the RS2 repeats, also oriented in a reverse (palindromic) fashion, rather than RS2 and RS3 (see Supplementary Data). It should be noted that the length of the homology between the RS1 and RS2 repeat is 210 bp instead of 222 bp. Although we did not map the JBP1 KO HYG:KO PUR:KO NEO amplicon, we suspect that its rearrangement involved another repeat, called RS4, present more downstream of JBP1 than RS3 in L.major Freidlin (see also Discussion). This would explain its longer size compared to the other amplicons as shown by pulsed-field gradient electrophoresis (compare lane 8 versus lanes 5–7 in Figure 2).
The amplicons cannot be generated by drug pressure on a cell line having one marker integrated in the JBP1 locus
Generation of a linear DNA amplicon following a targeting event has not been described in Leishmania before, or in any other organism. To determine whether the amplification was caused by the transfection itself or was due to a highly recombinogenic property of the JBP1 locus, we tried to generate amplicons by increasing the drug pressure on the JBP1 KO HYG mutant, which has one copy of the HYG marker integrated in the JBP1 wild-type chromosome. In Leishmania, this procedure often results in DNA amplification of genes involved in the resistance to the drug used (5,16). We increased the hygromycin B concentration in which JBP1 KO HYG was cultured stepwise by a factor two until it reached 2 mg/ml (40 times the normal concentration at which L.tarentolae is cultured). After culture for over three months at this concentration, we found no amplicons by hybridization on digested genomic DNA and pulsed-field gel electrophoresis (data not shown), suggesting that the integration of the inactivation cassette facilitated the formation of the amplicon.
DISCUSSION
In our attempts to disrupt both alleles of JBP1, an essential gene in Leishmania, we obtained mutants that appear to have a supernumerary JBP1-containing chromosome, as exemplified by the JBP1 KO HYG:KO PUR mutant and mutants containing linear plasmids (amplicons) with a long inverted duplication (Figure 6A). Whereas mutants with supernumerary chromosomes have been generated before in Leishmania by attempts to knock out essential genes (1–3), the linear plasmids are new and their unusual structure provides detailed information about the DNA rearrangement that generated them.
Leishmania is known to readily amplify DNA segments if challenged, be it by increasing drug pressure (5,10,16,37), by attempts to target essential genes (1–3), by nutrient stress or by subcloning (38). Amplification is even frequently seen in unselected laboratory stock (15). Linear DNA amplicons with an inverted repeat structure similar to the one described here have been described before in Leishmania in response to drug selection (39) and in the absence of selection (40,41). The generation of circular DNA amplicons by the circularization of the -tubulin locus has also been reported following attempts to inactivate this essential gene in Leishmania (42). There are no previous reports, however, of linear amplicons with an inverted repeat structure, having different markers in the two arms generated following a targeting event. We attribute the rarity of this type of amplicon to four factors (see Figure 7): (i) the targeting of an essential gene; (ii) the subtelomeric location of JBP1, which may facilitate formation of stable linear amplicons; (iii) the inferred presence of a DNA replication origin upstream of JBP1; and (iv) the presence of inverted repeats not far downstream of JBP1 facilitating strand switch in the replication fork.
Figure 7 Postulated model to explain the generation of the JBP1 KO NEO amplicon. We propose that replication starts upstream of JBP1. Two copies of JBP1 are generated after the replication fork has passed the JBP1 locus. One of these copies is targeted by the inactivation construct, which replaces JBP1 by NEO (A). The perturbation caused by the targeting could cause a stalling of the replication fork. Reversal of the fork would enable the RS3 repeat of the leading strand to anneal with the not yet replicated RS2 repeat of the lagging strand (B). Replication would restart using the other strand as a template followed by the synthesis of the complementary strand (C). Pealing of the misreplicated section would lead to an amplicon having an inverted repeat structure and two single-stranded DNA having on one strand JBP1 and on the other the NEO marker (D). Annealing of the strands would create a heteroduplex that would be repaired (E). The newly introduced marker would end up in the JBP1 chromosome only in half of the events, depending on which strand is repaired (F). The leading strand of the replication fork is shown in orange, the lagging strand in blue. The RS and telomeric repeats, as well as the JBP1 and NEO genes, are depicted as in Figure 6.
Model for linear amplicon formation
Figure 7 presents our preferred model for the formation of the amplicons. In this model, the inactivation cassette integrates into one arm of a replication fork that recently passed through the JBP1 locus. Insertion of the cassette might cause the replication fork to stall . Reversal of the fork would allow the RS3 repeat of the leading strand to anneal to the single-stranded RS2 repeat of the lagging strand (Figure 7B). This template strand switch is necessary to explain the presence of two different markers at the two ends of most amplicons. Re-initiation of DNA synthesis in the replication fork, indicated very schematically by the broken line in Figure 7C, would result in an inverted repeat amplicon having in one arm a marker already integrated in the genome of the parasite and in the other the newly transfected marker (Figure 7C and D). Pealing of the amplicon would allow the parental strands to snap back together, giving rise to a partial heteroduplex that would need to be repaired, for instance by gene conversion (Figure 7E). If the choice of the strand to be repaired is random, half the targeting events will result in a new marker in the targeted chromosome. The other half will leave the original gene in place. This is what we find as shown in Table 1. The alleles that get amplified and end up in the linear plasmids are the one in the inactivation cassette and the one that was targeted by the cassette. Similar models involving a template switch in a DNA replication fork have been proposed to explain the formation of amplicons bearing an inverted duplication in mammalian cell lines (44–46).
Our model is based on the following assumptions. The fact that the two arms of the amplicons contain different marker genes (or a wild-type gene and a marker) suggests that the targeting construct must have hit replicating DNA (see Figure 7). Hence, we infer the presence of an origin of replication located upstream of JBP1. The presence of a short inverted repeat at the inversion point of the linear plasmid indicates that the long inverted repeat structure of the plasmid arose probably from a template switch of the leading strand in a replication fork, as depicted in Figure 7. We propose that the integration of the inactivation constructs affected the progression of the replication fork allowing the DNA strand switch at the origin of the amplicon formation, but this is speculative.
It is obvious that the time window available for the generation of these plasmids must be narrow: targeting must take place after the replication fork has passed through the JBP1 gene, but if the fork has progressed too far beyond the RS repeats, the strand switch is probably not possible anymore. The distance between JBP1 and the RS repeats is 40 kb. Assuming a similar rate of DNA replication fork progression in Leishmania and yeast, in which it is 4 kb/min (47), the time window for initiating plasmid formation is roughly 10 min, i.e. 1/36th of the 6 h cell cycle. Following targeting, the complex process of plasmid formation still has to start. It is therefore not surprising that plasmid formation is not a frequent event. Once the replication fork is far beyond the RS repeats, mutants can only arise by missegregation of the JBP1 chromosome resulting in a cell triploid for a single chromosome. The same result will be obtained after targeting of a pre-existing single-chromosome triploid (followed by the transmission of the three chromosomes to the daughter cells). Since our Leishmania strain is basically diploid, the frequency at which these triploids arise must be low. Indeed, we have found that mutants triploid for the JBP1 chromosome were generated at rates 100- to 1000-fold lower than the rate at which the single JBP1 mutants arose (results not shown).
Other models of DNA amplification
Besides our ‘template-switch in a replication fork’ model shown in Figure 7, three other possibilities can be considered that could explain the generation of plasmids. Reversal of the replication fork might generate single newly synthesized strands, which can partially snap back on themselves by means of the RS repeats. If this u-turn is followed by self-copying and replication of the duplex formed, a plasmid with the same marker on both arms will arise, as exemplified by the JBP1 KO HYG:KO NEO plasmid (which cannot be explained by our strand-switch model) (see Table 1). This mutant has NEO as the only marker present in the amplicon and might have been formed by an auto-annealing of the branch of the replication fork bearing the NEO marker. Formation of amplicons by auto-annealing of palindromic repeats followed by self-copying of one DNA strand is a preferred model to explain the formation of amplicons in Leishmania (5,9,17).
A second possibility to generate the amplicons is to cut off the parental DNA at the replication fork, as suggested by Nalbantoglu and Meuth (44). What argues against this alternative is that the replicating chromosome from which the plasmid arises remains intact (data not shown). The broken chromosome could repair itself, however, by copying homologous DNA . Since the intact chromosome contains either of the markers present in the amplicon and not the marker present in the other chromosome, this repair mechanism remains implausible, as it would have to involve the newly made plasmid.
A third popular model of DNA amplification involves the over-replication of a portion of the genome causing an ‘onion-like’ structure, followed by the resolution or recombination of the over-replicated segment (16,18,46). This model cannot easily explain our observation, however, that the inactivation construct that caused amplification is always found in the amplicon and only in 2 out of 4 events in the original JBP1 chromosome (Table 1). The ‘onion-skin’ model would rather predict that this marker is always present in the amplicon and not in the original chromosome, in the case where the construct would integrate in the over-replicated sequence, or that it is found in both the amplicon and the JBP1 chromosome, if the cassette integrated prior to over-replication and formation of the amplicon. It is nevertheless possible that both events occurred in our case, but we find that unlikely.
Amplicon copy number
In all mutants analyzed, the amplicons are present in multiple copies, even though one would expect that one copy of the marker replacing the JBP1 gene should be enough to confer resistance to the drugs at the concentration used. Several factors probably contribute to this plasmid multiplicity: one factor is the use of a NEO gene in our experiments that contains a mutation (see Materials and Methods), which gives a diminished activity towards geneticin (G418). This results in a marginal resistance if a single NEO copy is present per cell (data not shown). In the generation of the JBP1 KO NEO mutant, which has also one NEO copy in the JBP1 chromosome, we obviously used a concentration of G418 that was too high for the cells to survive without gene amplification. The mutant NEO gene confers high-level resistance to paromomycin, however, and when the JBP1 KO NEO cells were transferred to paromomycin they lost the plasmid. The multiplicity of linear plasmids can also be explained by the fact that our plasmids do not contain centromeres and are rapidly lost in the absence of selection, like other linear plasmids in Leishmania. Since cells without a single copy of the plasmid will die in the presence of selection (since most amplicons contain markers that are not present in the wild-type chromosome, see Table 1), there must be a strong selection in the mutant population for clones that can transmit at least one plasmid copy to each daughter cell. This must drive up the copy number and we think that this partially explains why our mutants contain on average 4–5 plasmid copies per cell. If plasmid segregation is random, a cell with 4 copies (before replication) will generate less than 1% of daughter cells without a single plasmid copy. Such a low loss should not measurably affect the growth of the population.
Implication of repetitive sequences in the generation of the linear amplicons
At the inversion point of the three amplicons that we mapped, we found repeats that we hold responsible for the template switch in a stalled replication fork. The tendency of autoannealing of repeats RS1-3 is limited as we were unable to generate amplicons after increasing the drug pressure, a process that often yields amplicons in Leishmania (5,16). This suggests that the integration of the targeting cassette facilitates the template switch at the origin of the amplicon formation, e.g. by stalling the replication fork. Repeats (direct or inverted) are often associated with DNA amplification in Leishmania. For instance, repeats were proposed to be responsible for the formation of circular and linear amplifications of the H locus arising from arsenite or methotrexate selection (9,10). Interestingly, a repetitive DNA sequence having a moderate homology (50%) with the sequence found at the rearrangement point of our amplicons (RS repeats) is also present at the same location in the chromosome 9 of L.major Friedlin where the JBP1 gene is located. A computer-based genome screen of the repetitive element found in chromosome 9 of L.major Friedlin revealed that this sequence is found in many other chromosomes (data not shown). Moxon et al. (48) have proposed that ‘mutation rates vary among sites in the genome and that this variation is adaptive, because it promotes evolutionary flexibility in the face of environmental change, without necessarily increasing the overall load of deleterious mutations’. These highly mutable loci are called contingency genes, because they ‘facilitate the efficient exploration of phenotypic solutions to unpredictable aspects of the host environment’ (48). Leishmania can respond to some environmental challenges with a specific type of mutation, i.e. gene amplification, and the repeats associated with this amplification fulfill a clear contingency function (9,16). Possibly, the RS repeats studied here fulfill a similar role. It could be worth looking at other already described amplicons in Leishmania to see whether the formation of some of them could be explained by these RS repeats. The contingency genes that might be amplified through these RS repeats remain to be identified.
SUPPLEMENTARY MATERIAL
Supplementary Material is available at NAR Online.
ACKNOWLEDGEMENTS
We are grateful to Cristiane Bentin Toaldo, Dr Jero Calafat, Hans Janssen, Dr Jos Jonkers, Rudo Kieft, Dr Juliette Moyersoen, Dr Hein te Riele, Dr Robert Sabatini, Dr Sebastian Ulbert and Zhong Yu for input and discussion, as well as to Dr Marc Ouellette for the generous gift of numerous plasmids and the cosmid bank. P.A. Genest is an NSERC (Natural Sciences and Engineering Research Council of Canada) PGS-B scholar. This work was supported in part by a grant from NWO-CW (99015). Funding to pay the Open Access publication charges for this article was provided by the above-mentioned organizations.
REFERENCES
Cruz, A.K., Titus, R., Beverley, S.M. (1993) Plasticity in chromosome number and testing of essential genes in Leishmania by targeting Proc. Natl Acad. Sci. USA, 90, 1599–1603 .
Mottram, J.C., McCready, B.P., Brown, K.G., Grant, K.M. (1996) Gene disruptions indicate an essential function for the LmmCRK1 cdc2-related kinase of Leishmania mexicana Mol. Microbiol., 22, 573–583 .
Dumas, C., Ouellette, M., Tovar, J., Cunningham, M.L., Fairlamb, A.H., Tamar, S., Olivier, M., Papadopoulou, B. (1997) Disruption of the trypanothione reductase gene of Leishmania decreases its ability to survive oxidative stress in macrophages EMBO J., 16, 2590–2598 .
Beverley, S.M., Coderre, J.A., Santi, D.V., Schimke, R.T. (1984) Unstable DNA amplifications in methotrexate-resistant Leishmania consist of extrachromosomal circles which relocalize during stabilization Cell, 38, 431–439 .
White, T.C., Fase-Fowler, F., van Luenen, H., Calafat, J., Borst, P. (1988) The H circles of Leishmania tarentolae are a unique amplifiable system of oligomeric DNAs associated with drug resistance J. Biol. Chem., 263, 16977–16983 .
Papadopoulou, B., Roy, G., Ouellette, M. (1992) A novel antifolate resistance gene on the amplified H circle of Leishmania EMBO J., 11, 3601–3608 .
Grondin, K., Kundig, C., Roy, G., Ouellette, M. (1998) Linear amplicons as precursors of amplified circles in methotrexate-resistant Leishmania tarentolae Nucleic Acids Res., 26, 3372–3378 .
Detke, S., Katakura, K., Chang, K.P. (1989) DNA amplification in arsenite-resistant Leishmania Exp. Cell. Res., 180, 161–170 .
Ouellette, M., Hettema, E., Wust, D., Fase-Fowler, F., Borst, P. (1991) Direct and inverted DNA repeats associated with P-glycoprotein gene amplification in drug resistant Leishmania EMBO J., 10, 1009–1016 .
Grondin, K., Papadopoulou, B., Ouellette, M. (1993) Homologous recombination between direct repeat sequences yields P-glycoprotein containing amplicons in arsenite resistant Leishmania Nucleic Acids Res., 21, 1895–1901 .
Grondin, K., Haimeur, A., Mukhopadhyay, R., Rosen, B.P., Ouellette, M. (1997) Co-amplification of the gamma-glutamylcysteine synthetase gene gsh1 and of the ABC transporter gene pgpA in arsenite-resistant Leishmania tarentolae EMBO J., 16, 3057–3065 .
Wilson, K., Collart, F.R., Huberman, E., Stringer, J.R., Ullman, B. (1991) Amplification and molecular cloning of the IMP dehydrogenase gene of Leishmania donovani J. Biol. Chem., 266, 1665–1671 .
Detke, S., Chaudhuri, G., Kink, J.A., Chang, K.P. (1988) DNA amplification in tunicamycin-resistant Leishmania mexicana. Multicopies of a single 63-kilobase supercoiled molecule and their expression J. Biol. Chem., 263, 3418–3424 .
Henderson, D.M., Sifri, C.D., Rodgers, M., Wirth, D.F., Hendrickson, N., Ullman, B. (1992) Multidrug resistance in Leishmania donovani is conferred by amplification of a gene homologous to the mammalian mdr1 gene Mol. Cell. Biol., 12, 2855–2865 .
Beverley, S.M. and Coburn, C.M. (1990) Recurrent de novo appearance of small linear DNAs in Leishmania major and relationship to extra-chromosomal DNAs in other species Mol. Biochem. Parasitol., 42, 133–141 .
Beverley, S.M. (1991) Gene amplification in Leishmania Annu. Rev. Microbiol., 45, 417–444 .
Papadopoulou, B., Roy, G., Ouellette, M. (1993) Frequent amplification of a short chain dehydrogenase gene as part of circular and linear amplicons in methotrexate resistant Leishmania Nucleic Acids Res., 21, 4305–4312 .
Stark, G.R. and Wahl, G.M. (1984) Gene amplification Annu. Rev. Biochem., 53, 447–491 .
Gommers-Ampt, J.H., van Leeuwen, F., de Beer, A.L., Vliegenthart, J.F., Dizdaroglu, M., Kowalak, J.A., Crain, P.F., Borst, P. (1993) Beta-D-glucosyl-hydroxymethyluracil: a novel modified base present in the DNA of the parasitic protozoan T.brucei Cell, 75, 1129–1136 .
van Leeuwen, F., Taylor, M., Mondragon, A., Moreau, H., Gibson, W., Kieft, R., Borst, P. (1998) Beta-D-glucosyl-hydroxymethyluracil is a conserved DNA modification in kinetoplastid protozoans and is abundant in their telomeres Proc. Natl Acad. Sci. USA, 95, 2366–2371 .
Dooijes, D., Chaves, I., Kieft, R., Dirks-Mulder, A., Martin, W., Borst, P. (2000) Base J originally found in kinetoplastida is also a minor constituent of nuclear DNA of Euglena gracilis Nucleic Acids Res., 28, 3017–3021 .
van Leeuwen, F., Kieft, R., Cross, M., Borst, P. (2000) Tandemly repeated DNA is a target for the partial replacement of thymidine by ?-D-glucosyl-hydroxymethyluracil in Trypanosoma brucei Mol. Biochem. Parasitol., 109, 133–145 .
Cross, M., Kieft, R., Sabatini, R., Wilm, M., de Kort, M., van der Marel, G.A., van Boom, J.H., van Leeuwen, F., Borst, P. (1999) The modified base J is the target for a novel DNA-binding protein in kinetoplastid protozoans EMBO J., 18, 6573–6581 .
Sabatini, R., Meeuwenoord, N., van Boom, J.H., Borst, P. (2002) Recognition of base J in duplex DNA by J-binding protein J. Biol. Chem., 277, 958–966 .
Sabatini, R., Meeuwenoord, N., van Boom, J.H., Borst, P. (2002) Site-specific interactions of JBP with base and sugar moieties in duplex J-DNA. Evidence for both major and minor groove contacts J. Biol. Chem., 277, 28150–28156 .
Cross, M., Kieft, R., Dirks-Mulder, A., Chaves, I., Borst, P. (2002) J-binding protein increases the level and retention of the unusual base J in trypanosome DNA Mol. Microbiol., 46, 37–47 .
Brun, R. and Schonenberger, M. (1979) Cultivation and in vitro cloning of procyclic culture forms of Trypanosoma brucei in a semi-defined medium Acta Trop., 36, 289–292 .
Munoz-Jordan, J.L. and Cross, G.A. (2001) Telomere shortening and cell cycle arrest in Trypanosoma brucei expressing human telomeric repeat factor TRF1 Mol. Biochem. Parasitol., 114, 169–181 .
Brochu, C., Haimeur, A., Ouellette, M. (2004) The heat shock protein HSP70 and heat shock cognate protein HSC70 contribute to antimony tolerance in the protozoan parasite Leishmania Cell. Stress. Chap., 9, 294–303 .
Papadopoulou, B., Roy, G., Ouellette, M. (1994) Autonomous replication of bacterial DNA plasmid oligomers in Leishmania Mol. Biochem. Parasitol., 65, 39–49 .
Sambrook, J., Fritsch, E., Maniatis, T. Molecular Cloning: A Laboratory Manual, (1989) Cold Spring Harbor, NY Cold Spring Harbor Press .
van der Ploeg, L.H., Schwartz, D.C., Cantor, C.R., Borst, P. (1984) Antigenic variation in Trypanosoma brucei analyzed by electrophoretic separation of chromosome-sized DNA molecules Cell, 37, pp. 77–84 .
van der Ploeg, L.H., Cornelissen, A.W., Barry, J.D., Borst, P. (1984) Chromosomes of kinetoplastida EMBO J., 3, 3109–3115 .
Hightower, R.C., Bliska, J.B., Cozzarelli, N.R., Santi, D.V. (1989) Analysis of amplified DNAs from drug-resistant Leishmania by orthogonal-field-alternation gel electrophoresis. The effect of size and topology on mobility J. Biol. Chem., 264, 2979–2984 .
Bernards, A., Michels, P.A., Lincke, C.R., Borst, P. (1983) Growth of chromosome ends in multiplying trypanosomes Nature, 303, 592–597 .
Sunkin, S., Kiser, P., Myler, P., Stuart, K. (2000) The size difference between Leishmania major Friedlin chromosome one homologues is localized to sub-telomeric repeats at one chromosomal end Mol. Biochem. Parasitol., 109, 1–15 .
Grondin, K., Roy, G., Ouellette, M. (1996) Formation of extrachromosomal circular amplicons with direct or inverted duplications in drug-resistant Leishmania tarentolae Mol. Cell. Biol., 16, 3587–3595 .
Rovai, L., Tripp, C., Stuart, K., Simpson, L. (1992) Recurrent polymorphisms in small chromosomes of Leishmania tarentolae after nutrient stress or subcloning Mol. Biochem. Parasitol., 50, 115–125 .
Dubessay, P., Ravel, C., Bastien, P., Lignon, M.F., Ullman, B., Pages, M., Blaineau, C. (2001) Effect of large targeted deletions on the mitotic stability of an extra chromosome mediating drug resistance in Leishmania Nucleic Acids Res., 29, 3231–3240 .
Navarro, M., Liu, J., Muthui, D., Ortiz, G., Segovia, M., Hamers, R. (1994) Inverted repeat structure and homologous sequences in the LD1 amplicons of Leishmania spp Mol. Biochem. Parasitol., 68, 69–80 .
Ortiz, G. and Segovia, M. (1996) Characterisation of the novel junctions of two minichromosomes of Leishmania major Mol. Biochem. Parasitol., 82, 137–144 .
Curotto de Lafaille, M.A. and Wirth, D.F. (1992) Creation of null/+ mutants of the -tubulin gene in Leishmania enriettii by gene cluster deletion J. Biol. Chem., 267, 23839–23846 .
Michel, B., Grompone, G., Flores, M.J., Bidnenko, V. (2004) Multiple pathways process stalled replication forks Proc. Natl Acad. Sci. USA, 101, 12783–12788 .
Nalbantoglu, J. and Meuth, M. (1986) DNA amplification-deletion in a spontaneous mutation of the hamster aprt locus: structure and sequence of the novel joint Nucleic Acids Res., 14, 8361–8371 .
Hyrien, O., Debatisse, M., Buttin, G., de Saint Vincent, B.R. (1988) The multicopy appearance of a large inverted duplication and the sequence at the inversion joint suggest a new model for gene amplification EMBO J., 7, 407–417 .
Stark, G.R., Debatisse, M., Giulotto, E., Wahl, G.M. (1989) Recent progress in understanding mechanisms of mammalian DNA amplification Cell, 57, 901–908 .
Reynolds, A.E., McCarroll, R.M., Newlon, C.S., Fangman, W.L. (1989) Time of replication of ARS elements along yeast chromosome III Mol. Cell. Biol., 9, 4488–4494 .
Moxon, E.R., Rainey, P.B., Nowak, M.A., Lenski, R.E. (1994) Adaptive evolution of highly mutable loci in pathogenic bacteria Curr. Biol., 4, 24–33 .
Laban, A., Tobin, J.F., Curotto de Lafaille, M.A., Wirth, D.F. (1990) Stable expression of the bacterial neor gene in Leishmania enriettii Nature, 343, 572–574 .(Paul-André Genest, Bas ter Riet, Carole )