Portal de Periódicos da CAPES

A mitogen-activated protein (MAP) kinase homologue of Leishmania mexicana is essential for parasite survival in the infected host

1998; Springer Nature; Volume: 17; Issue: 9 Linguagem: Inglês

10.1093/emboj/17.9.2619

1460-2075

Martin Wiese,

Venomous Animal Envenomation and Studies

Article1 May 1998free access A mitogen-activated protein (MAP) kinase homologue of Leishmania mexicana is essential for parasite survival in the infected host Martin Wiese Martin Wiese Max-Planck-Institut für Biologie, Abteilung Membranbiochemie, Corrensstrasse 38, D-72076 Tübingen, Germany Search for more papers by this author Martin Wiese Martin Wiese Max-Planck-Institut für Biologie, Abteilung Membranbiochemie, Corrensstrasse 38, D-72076 Tübingen, Germany Search for more papers by this author Author Information Martin Wiese1 1Max-Planck-Institut für Biologie, Abteilung Membranbiochemie, Corrensstrasse 38, D-72076 Tübingen, Germany The EMBO Journal (1998)17:2619-2628https://doi.org/10.1093/emboj/17.9.2619 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The parasitic protozoon Leishmania mexicana undergoes two major developmental stages in its life cycle exhibiting profound physiological and morphological differences, the promastigotes in the insect vector and the amastigotes in mammalian macrophages. A deletion mutant, Δlmsap1/2, for the secreted acid phosphatase (SAP) gene locus, comprising the two SAP genes separated by an intergenic region of ∼11.5 kb, lost its ability to cause a progressive disease in Balb/c mice. While in vitro growth of promastigotes, invasion of host cells and differentiation from promastigotes to amastigotes was indistinguishable from the wild-type, the mutant parasites ceased to proliferate when transformed to amastigotes in infected macrophages or in a macrophage-free in vitro differentiation system, suggesting a stage-specific growth arrest. This phenotype could be reverted by complementation with 6 kb of the intergenic region of the SAP gene locus. Sequence analysis identified two open reading frames, both encoding single copy genes; one gene product shows high homology to mitogen-activated protein (MAP) kinases. Complementation experiments revealed that the MAP kinase homologue, designated LMPK, is required and is sufficient to restore the infectivity of the Δlmsap1/2 mutant. Therefore, LMPK is a kinase that is essential for the survival of L.mexicana in the infected host by affecting the cell division of the amastigotes. Introduction Infection with Leishmania, a parasitic protozoon, represents a major health problem in the tropics and subtropics, with ∼380 000 cases annually and 367 million people at risk of infection (Ashford et al., 1996). Depending on the species, Leishmania causes a spectrum of diseases ranging from self-healing cutaneous lesions to lethal visceral forms (Alexander and Russell, 1992). During their life cycle, the parasites undergo profound morphological changes. The spindle-shaped, flagellated procyclic promastigotes, which proliferate in the gut of the sandfly, differentiate into non-dividing infective metacyclic cells. When the insect takes a blood-meal, the metacyclics are transmitted into the skin of the mammal. Once in the mammal, the parasites enter macrophages where they reside in phagolysosomes. In this low-pH, hydrolytic environment, they transform into the oval, non-motile amastigotes and proliferate. Morphologically, amastigotes are smaller than promastigotes and have a rudimentary flagellum buried in the flagellar pocket. Ultrastructurally, Leishmania mexicana amastigotes have large lysosome-like organelles called megasomes, which are absent in promastigotes. Moreover, they differ in their surface architecture and in their secretory products. The surface of the promastigotes is composed of lipophosphoglycan (LPG), the surface protease gp63 and glycoinositolphospholipids (GIPLs) as dominant surface markers (McConville and Ferguson, 1993). At least in L.mexicana, amastigotes lack LPG and gp63, having only GIPLs and host-derived glycosphingolipids on their surface (Bahr et al., 1993; Winter et al., 1994). It is likely that the changes in the biochemistry and morphology of Leishmania during its life cycle are the result of programmed changes in gene expression in response to the changes in the external environment of the parasite. However, the signal transduction pathways involved in promastigote-to-amastigote differentiation and vice versa, and in life stage-specific gene expression are not known. Through analogy with other eukaryotic cells (Treisman, 1996), protein phosphorylation can be expected to be important for differentiation and proliferation of Leishmania. Therefore, protein kinases and phosphatases, and their regulation, are likely to be critical for the parasite's development. In fact, changes in phosphoprotein abundance and in the overall phosphorylation pattern in Leishmania and other kinetoplastids, such as Trypanosoma brucei and Trypanosoma congolense, have been documented throughout their life cycles (Mukhopadhyay et al., 1988; Aboagye-Kwarteng et al., 1991; Parsons et al., 1991, 1993, 1995; Dell and Engel, 1994), but their significance for parasite differentiation remains unknown. In a recent review, Boshart and Mottram (1996) summarized the knowledge of protein phosphorylation and protein kinases in trypanosomatids. They presented a collection of cloned genes of various protein phosphatases and kinases: protein phosphatases 1 and 2A from T.brucei and 2C from Leishmania chagasi; cdc2-related kinases from L.mexicana, T.brucei and Crithidia fasciculata; a NIMA-related, a polo-like and a potential mitogen-activated protein (MAP) kinase from T.brucei; protein kinase A from L.major and T.brucei; MAP kinase kinases from L.chagasi and Leishmania major (Boshart and Mottram, 1996, and references therein). Here, I report the identification of a L.mexicana gene coding for a protein with strong homology to MAP kinases of yeast and higher eukaryotes that is required for amastigote survival and proliferation in macrophages. Therefore, this gene is essential for the infection of mammals. Results A non-pathogenic deletion mutant of Leishmania mexicana In a previous publication (Wiese et al., 1995), we reported the construction of gene deletion mutants in L.mexicana encompassing parts of the secreted acid phosphatase (SAP) gene locus (Figure 1A and B), in order to analyse the composition and structure of the unique filamentous complex formed by this enzyme (Ilg et al., 1991). Two closely related genes, lmsap1 and lmsap2, were found to be tandemly arranged. The open reading frames are separated by an intergenic region of ∼11.5 kb. Three mutants were derived lacking either both alleles of lmsap1 (Δlmsap1), lmsap2 (Δlmsap2) or both genes including the intergenic region (Δlmsap1/2; Figure 1). Owing to the speculation that the SAPs may contribute to the pathobiology in Leishmaniasis by eliciting a humoral immune response (Chang et al., 1990), we tested all mutants for their infection potential in Balb/c mice. Both single-gene deletion mutants showed lesion development comparable with control mice infected with wild-type L.mexicana (data not shown). In contrast, promastigotes of the third deletion mutant, Δlmsap1/2, did not cause lesion development in mice infected with promastigotes. Figure 1.Genomic organization of the SAP gene locus, lmsap deletion mutants and complementation strategy. (A) Genomic organization of the SAP genes lmsap1 and lmsap2. The open reading frames are marked by arrows indicating the direction of transcription. Approximate sizes of DNA regions covered by the mRNAs are indicated below the respective genes leaving an intervening region of ∼6 kb. p1 and p2 are gene-specific probes used to map the transcribed regions in a Northern-blot analysis. (B) DNA regions replaced by resistance markers in the deletion mutants Δlmsap1, Δlmsap2 and Δlmsap1/2. (C) DNA regions cloned into pX63PHLEO or pX63polPHLEO and used in complementation experiments of the null mutant Δlmsap1/2. Download figure Download PowerPoint In culture, proliferation of Δlmsap1/2 promastigotes was indistinguishable from wild-type promastigotes. Mutant promastigotes of late logarithmic growth phase could be transformed in the absence of macrophages into amastigote-like cells by pH and temperature shifts to pH 5.5 and 34°C, respectively (Figure 2A). Starting at cell densities of either 1×106 or 4×106 promastigotes/ml, the cell counts increased over 4 days to ∼5×106 and 3×107 cells/ml, respectively, and then remained constant. In contrast, the wild-type reached a cell density of 3.5×107 cells/ml in either 4 or 6 days, irrespective of the size of the inoculum, indicating growth to stationary phase. After 4 weeks at 34°C and pH 5.5 with intermittent supply of fresh medium, an elevation of the pH to 7.0 and a reduction of the temperature to 26°C, which are the optimal conditions for L.mexicana promastigote growth in vitro, resulted in some of the cultivated Δlmsap1/2 mutant amastigote-like parasites re-transforming into promastigotes, indicating their viability and a reversibility of differentiation independent of the deletion. Infection of peritoneal macrophages derived from Balb/c mice revealed that the Δlmsap1/2 mutant retained its ability to enter host cells and to differentiate into the amastigote form. However, in the course of 6 days after the infection, the ratio of infected to uninfected macrophages dropped from 0.9 on the first day to 0.09 on day 6, indicating that the parasites ceased to proliferate and were cleared by the macrophages (Figure 2B). In contrast, the wild-type reached densities up to 25 amastigotes per infected macrophage within 3 days post-infection (results not shown). In summary, the Δlmsap1/2 mutant is impaired in the ability to grow axenically as amastigotes, and it cannot proliferate in macrophages or mice. Figure 2.Growth of wild-type and Δlmsap1/2 null mutant amastigotes. (A) Growth curves of in vitro transformed promastigotes starting at 1×106 and 4×106 cells/ml, respectively. Transformation was achieved by pH and temperature shifts as described in Materials and methods. (▴) L.mexicana wild-type, (□) Δlmsap1/2. (B) Infection of peritoneal macrophages with the mutant Δlmsap1/2. Macrophages were grown on glass coverslips overnight, infected with promastigotes at a ratio of 20:1, coverslips were removed at the indicated time points and processed for inspection as described in Materials and methods. Three times, 100 macrophages were randomly inspected, the average taken and the macrophages grouped according to their parasite load. (□) 1–3 amastigotes/macrophage, (▴) 4–6 amastigotes/macrophage, (○) 7–9 amastigotes/macrophage, (♦) >10 amastigotes/macrophage. Download figure Download PowerPoint The intergenic region is responsible for the Δlmsap1/2 phenotype The deletion of both SAP genes in the mutant Δlmsap1/2 did not impair the parasites with regard to their uptake by macrophages or their differentiation to amastigotes. Therefore, it appeared very likely that in L.mexicana the SAP is not required for the initial steps of the infection, but may be necessary to establish a progressive infection later on. However, we have not been able to detect any material reactive with an anti-protein mAb specific for SAP (mAb LT8.2; Ilg et al., 1993) using immunofluorescence in macrophages infected with wild-type promastigotes in vitro, or using immunoelectron microscopy of infected macrophages in mouse lesions (M.Wiese, T.Ilg and Y.–D.Stierhof, unpublished results). This could be due to a modification or loss of the mAb LT8.2 epitope located on the exposed C-terminus after the highly O-glycosylated repetitive region of SAP (Wiese et al., 1995; Y.-D.Stierhof, M.Wiese, T.Ilg, P.Overath, M.Häuer and U.Aebi, in preparation), preventing the binding of the antibody, or simply to the absence of the protein despite the presence of a lmsap mRNA (Wiese et al., 1995). If the SAP is not required to establish an infection, it appears likely that there is at least one additional gene located within the 11.5 kb intergenic region of the SAP gene locus, which has been removed concomitantly with the SAP genes, lmsap1 and lmsap2, in the null mutant, Δlmsap1/2. To clarify this situation, two plasmids were designed and introduced into the deletion background. First, the transcribed regions of the two SAP genes were mapped roughly in a Northern blot experiment using gene-specific probes (p1 and p2; Figure 1A) that hybridized to the 3′-flanking DNA of either lmsap2 or lmsap1 (data not shown). This defined a 6 kb portion of the lmsap intergenic region that could carry additional genes (Figure 1A) responsible for the mutant phenotype. A suitable DNA fragment carrying most of this portion was subcloned into the leishmanial expression vector pX63PHLEO (Freedman and Beverley, 1993) resulting in construct 1 (Figure 1C). Construct 2 was obtained by the addition of the lmsap2 gene and its flanking regions to the first construct, thereby forming a plasmid encompassing lmsap2, the intergenic region and a truncated lmsap1 gene (Figure 1C). These two constructs were introduced into the Δlmsap1/2 mutant background. Lesion development after injection of 1×106 mutant or wild-type promastigotes into hind footpads of Balb/c mice was monitored for 70 weeks (Figure 3). Leishmania mexicana wild-type promastigotes, included as a control, gave rise to the most progressive lesion development. Both plasmid constructs conferred the potential to cause the disease, however, with a lag of ∼10 weeks, indicating that the prolonged in vitro culture used to obtain the mutants may have resulted in a decrease in their infectious potential. The Δlmsap1/2 null mutant did not lead to a significant swelling of the inoculated footpads. The overlapping bars for the standard deviations show clearly that there is no significant difference in the course of the infection in the Δlmsap1/2 mutant complemented with a construct carrying or lacking lmsap2. The slower kinetics of lesion development compared with the wild-type could be due to the episomal location of the essential gene that could either lead to a heterogeneous expression of the corresponding gene product in the amastigotes resulting in protein levels too low to survive in some cells, or the plasmid could be lost in some parasites during division, resulting in the death of the segregants, and thereby slowing down the progression of the disease. This shows that this plasmid-based complementation system can only yield qualitative information when compared with the wild-type. Despite the slower progression of lesion development in the complemented Δlmsap1/2 null mutants the final lesion size in infected mice was comparable with wild-type infections resulting in metastasis and finally death of the mice. Therefore, the SAP of L.mexicana is not required either for the initial steps of the infection, for the differentiation of promastigotes to amastigotes or for their survival and proliferation in the mammalian host. This result is not unexpected, as a different pathogenic species, L.major, does not secrete an acid phosphatase (Lovelace and Gottlieb, 1986). Figure 3.Footpad infection experiment of SAP null mutant Δlmsap1/2 complemented with constructs 1 and 2. (□) L.mexicana wild-type, (○) Δlmsap1/2 + construct 2, (▿) Δlmsap1/2 + construct 1, (⋄) Δlmsap1/2; compare Figure 1C. Bars represent standard deviations in each group of five infected mice. Download figure Download PowerPoint Identification of lmpk and orf2 As the inability of the mutant Δlmsap1/2 to develop lesions in infected mice was not due to the loss of the SAP, it was very likely that at least one other gene, which was essential for amastigote survival in the infected host cell, was located within the 6 kb of intergenic region of the SAP gene locus sufficient to complement the null mutant. Sequence analysis of this region led to the identification of two open reading frames, orf1 and orf2 (Figure 4A). Orf1 consists of 1074 bp coding for a protein of 358 amino acids, with a calculated molecular mass of 41 kDa. A homology search in the DDBJ/EMBL/GenBank database revealed significant homology to MAP kinase genes, e.g. 55.6% amino acid identity to the KFR1 MAP kinase homologue of T.brucei (Hua and Wang, 1994). Figure 5 shows an alignment of the L.mexicana kinase homologue encoded by orf1 with other protozoan kinases (T.brucei KFR1 kinase, Dictyostelium discoideum MAP kinase, Plasmodium falciparum MAP kinase homologue) and one representative each of yeast (Saccharomyces cerevisiae KSS1), plant (Arabidopsis thaliana MAP kinase), and mammalian (Rattus norvegicus ERK2 kinase) kinases. The sequence displays the typical 11 major conserved subdomains (I–XI) and the sequence motifs characteristic for proline-directed serine/threonine protein kinases. These motifs are the phosphate anchor ribbon for ATP binding, Gly21-Ser22-Gly23-Ala24-Tyr25-Gly26, at the N-terminus (consensus Gly-X-Gly-X-X-Gly), the potential regulatory phosphorylation sites at Thr176 and Tyr178 in the phosphorylation lip (Leu158-Arg185), the P+1 specificity pocket (Ala180-Thr181-Arg182-Trp183-Tyr184-Arg185) and the catalytic site residues Lys43, Arg57, Arg60, Glu61, Arg136, Asp137, Lys139, Asn142, Asp155 (Mg2+ ligand), Arg160 and Thr181 (Zhang et al., 1994; Wang et al., 1997). Because of the high homology and the presence of all the typical MAP kinase features, orf1 was designated lmpk for the L.mexicana protein kinase gene. The corresponding protein is termed LMPK. Figure 4.Southern and Northern analysis of orf1 (lmpk) and orf2. (A) Genomic organization of orf1 (lmpk) and orf2 in the SAP gene locus showing all relevant restriction sites and splice addition sites (sas). Restriction sites in bold italics produce hybridizing bands in the Southern blot experiment. The open reading frames are boxed, the arrows indicate the direction of transcription. (B) Southern analysis of genomic DNA from L.mexicana. DNA (5 μg) was cleaved with the indicated restriction enzymes, separated on a 0.7% agarose gel, blotted onto a nylon membrane and probed with DIG-labelled, gene-specific probes for either orf1 or orf2 as indicated. (C) Northern analysis of total RNA from L.mexicana promastigotes (P) and amastigotes (A). Total RNA (20 μg) was separated on a 1% denaturing formaldehyde agarose gel, blotted onto a positively charged nylon membrane and probed with the respective radiolabelled, gene-specific probes. Numbers indicate the size of DNA and RNA standards (kb). Download figure Download PowerPoint Figure 5.Alignment of LMPK with various MAP kinase amino acid sequences. Leish LMPK, L.mexicana MAP kinase homologue; Tryp KFR1, KFR1 kinase from T.brucei (Hua and Wang, 1994; accession No. l10997); Dict MPK, D.discoideum MAP kinase (Segall et al., 1995; accession No. l33043); Plas MPK, P.falciparum MAP kinase (accession No. x82646); Arab MPK, A.thaliana MAP kinase (Mizoguchi et al., 1993; accession No. d21840); Rat ERK2, R.norvegicus ERK2 kinase (Boulton et al., 1991; accession No. m64300) and Yeast KSS1, S.cerevisiae KSS1 kinase (Courchesne et al., 1989; accession No. m26398). I–XI indicate MAP kinase typical domains. Diamonds mark potential regulatory phosphorylation sites at Thr176 and Tyr178, arrows indicate catalytic site residues, brackets depict the phosphate anchor ribbon (Gly21-Gly26) and the P+1 specificity pocket (Ala180-Arg185) and asterisks mark the C-terminus of the protein. Dashes indicate gaps introduced to optimize the alignment; dots represent identical amino acid residues. Numbering corresponds to LMPK. Download figure Download PowerPoint Orf2 showed no significant homology to the sequences in the database. It codes for a putative protein of 418 amino acids with a calculated molecular mass of 45 kDa. There is no potential signal sequence for import into the endoplasmic reticulum. Interestingly, the protein is rich in Arg (11 mole per cent), has an overall charge of 21 and an isoelectric point of 11.66. Moreover, it contains a number of phosphorylation motifs for cAMP-dependent protein kinase, casein kinase II and protein kinase C as revealed by a motif search using the GCG DNA analysis program (Devereux et al., 1984) and two potential MAP kinase phosphorylation sites matching the consensus Pro-X-Ser/Thr-Pro where X is any amino acid (Ruderman, 1993). Southern-blot analysis of both genes with a selection of restriction enzymes, cleaving inside and next to the open reading frames using gene-specific probes, resulted in a pattern of hybridizing bands corresponding to the mapped SAP gene locus (Figure 4A and B), indicating that both genes are present in one copy per haploid genome. Moreover, both genes are transcribed in promastigotes and amastigotes with a bias to higher transcript levels in the mammalian amastigote stage (Figure 4C). The sizes of the mRNAs are 2.6 and 2.0 kb for lmpk and orf2, respectively. Therefore, these mRNAs occupy nearly all of the 6 kb of the, as yet uncharacterized, intergenic region of the SAP gene locus. In Kinetoplastidae, mature mRNAs are formed from a polycistronic precursor by transfer of the so-called mini-exon carrying a cap structure to the 5′-end of protein coding RNAs, in a trans-splicing reaction using an AG dinucleotide splice-acceptor site (Laird, 1989). Trans-splicing is accompanied by polyadenylation on the 3′-end of the RNA. Whereas the splice-acceptor site is determined by the AG dinucleotide, there is no consensus for poly(A) addition. In contrast, poly(A) site selection of a given gene has been shown to be dependent on the location of the splice-acceptor site of the next gene located further downstream (LeBowitz et al., 1993). Splice-acceptor sites for lmpk and orf2 were determined using total RNA in a reverse transcriptase polymerase chain reaction (RT–PCR) with gene internal and mini-exon primers (see Materials and methods). Amplified fragments were cloned and sequenced. A single AG splice-acceptor site for lmpk was located 320 bp upstream of the ATG translation initiation codon. Orf2 was found to use at least four adjacent splice-acceptor sites in distances of 383, 379, 335 and 313 bp upstream of the first likely ATG translation initiation codon. The genomic organization of the genes in the SAP gene locus is in accordance with the already described tight packing of genes in the Leishmania genome (LeBowitz et al., 1993). DNA from representatives of every Leishmania species was tested in a Southern blot experiment with gene-specific probes for lmpk and orf2. Specifically hybridizing fragments revealed that both genes are present in at least one copy per haploid genome of L.amazonensis, L.braziliensis, L.tropica, L.major, L.aethiopica and L.donovani (data not shown). Despite ∼58% nucleotide identity between lmpk and kfr1, corresponding to 55.6% amino acid identity (Figure 5), no hybridization was observed with T.brucei DNA, showing that hybridization and washing conditions were stringent enough to detect only highly homologous DNA sequences. lmpk is essential for amastigote proliferation The first complementation experiment showed clearly that the defect of the mutant Δlmsap1/2 is due to the lack of the intergenic region. Therefore, a second set of complementation experiments was performed to discriminate whether one of the two genes of the intergenic region, or both, are essential for the survival of the parasites in macrophages or mice. Either lmpk or orf2, including the respective 3′-untranslated regions (3′-UTR) (Figure 1C, constructs 3 and 4), was subcloned into pX63polPHLEO and reintroduced into the null mutant, Δlmsap1/2. Peritoneal macrophages from Balb/c mice infected with promastigotes from the resulting mutants Δlmsap1/2+lmpk or Δlmsap1/2+orf2 at a ratio of 20:1 were inspected microscopically for parasite load at various time points after infection (Figure 6). Both mutant parasite lines differentiated into the typical ovoid amastigote form. The initial infection rate reached by both cell types was essentially the same, starting with 1–3 parasites per infected macrophage. However, at later time points, only those parasites complemented with lmpk revealed a progressive infection indicated by the increasing number of macrophages harbouring 10 or more amastigotes, while the number of macrophages with a parasite load of 1–3 decreased. Therefore, lmpk alone enabled the parasites to proliferate in the macrophages, whereas orf2 is dispensable. Footpad infection of Balb/c mice with promastigotes from the mutants Δlmsap1/2, Δlmsap1/2+lmpk or Δlmsap1/2+orf2 corroborated the in vitro result (Figure 7). Only mice infected with the mutant Δlmsap1/2+lmpk began to develop lesions 10 weeks post-infection. Lesion size increased steadily over 60 weeks, indicating a progressive infection in the mice. The mutant containing orf2 showed no lesion development in Balb/c mice. Figure 6.Infection of peritoneal macrophages with the complemented SAP null mutant, Δlmsap1/2 (compare with Figure 2B). (A) Δlmsap1/2+orf1 (lmpk), (B) Δlmsap1/2+orf2, (□) 1–3 amastigotes/macrophage, (▴) 4–6 amastigotes/macrophage, (○) 7–9 amastigotes/macrophage, (♦) >10 amastigotes/macrophage. Download figure Download PowerPoint Figure 7.Footpad infection experiment of SAP null mutant Δlmsap1/2 complemented with constructs 3 and 4. (□) Δlmsap1/2 + construct 3 (lmpk), (●) Δlmsap1/2 + construct 4 (orf2); (▵) Δlmsap1/2; compare with Figure 1C. Bars represent standard deviations in the group of five infected mice. Download figure Download PowerPoint Expression of lmpk A polyclonal antiserum against a C-terminal peptide of LMPK (anti-CLMPK) was raised in rabbits, affinity-purified using the peptide and employed to detect LMPK in total cell lysates. On an immunoblot, the antiserum recognized a band with a molecular mass corresponding to the calculated molecular mass of 41 kDa in both L.mexicana promastigotes and amastigotes (Figure 8, lanes 1 and 4). As expected, this reaction was absent in a lysate from the mutant Δlmsap1/2 (Figure 8, lane 2). The additional band seen in amastigote lysates may be derived from contaminating mouse lesion material or an amastigote protein interacting with secondary antibodies. Freshly prepared Δlmsap1/2+lmpk lesion amastigotes were positive for LMPK (Figure 8, lane 5), as well as Δlmsap1/2+lmpk promastigotes either freshly transformed from mouse lesion-derived amastigotes or from recombinant parasites that had not been passaged through the mouse (Figure 8, lanes 6 and 3). Both types of promastigotes were kept under selective pressure by the presence of the antibiotic phleomycin in the culture medium, whereas the lesion-derived Δlmsap1/2+lmpk amastigotes maintained the lmpk gene without selective pressure for >1 year (compare with Figure 7) and expressed the protein in amounts similar to that of wild-type L.mexicana amastigotes. The high level of lmpk expression in promastigotes freshly transformed from lesion-derived amastigotes, which were then cultivated under antibiotic selection, suggested that lmpk is still plasmid encoded and not integrated into the genome by a recombination event. This was confirmed by Southern blot experiments using total DNA from these promastigotes cleaved by a set of restriction enzymes specific for the identification of plasmid-encoded lmpk, and by re-isolation of the plasmid by transformation of Escherichia coli using total promastigote DNA, followed by restriction analysis of 23 independent plasmid clones (data not shown). All the plasmids showed the same restriction pattern as the original plasmid that was used to transfect the promastigotes of the null mutant Δlmsap1/2, indicating the stability of the plasmid construct as an extrachromosomal element in complemented Δlmsap1/2 amastigotes for >14 months without selective pressure. Figure 8.Immunoblot of LMPK from promastigote and amastigote total cell lysates with a polyclonal antiserum against a C-terminal peptide. Lane 1, wild-type promastigotes; lane 2, Δlmsap1/2 promastigotes; lane 3, Δlmsap1/2 + lmpk promastigotes; lane 4, wild-type amastigotes from mouse lesions; lane 5, Δlmsap1/2 + lmpk amastigotes from mouse lesions; lane 6, Δlmsap1/2 + lmpk promastigotes freshly transformed from lesion-derived amastigotes. The molecular mass of standard proteins in kDa is indicated. Download figure Download PowerPoint Discussion A deletion mutant (Δlmsap1/2) in L.mexicana lacking both alleles of the two tandemly arranged lmsap genes including the intergenic region, that was originally developed to analyse the structure and function of the filamentous SAP complex, was found to have lost its potential to develop lesions in Balb/c mice and to grow in macrophages. However, mutant promastigotes showed the same initial infection rate of cultured macrophages as wild-type promastigotes, suggesting that the defect of the Δlmsap1/2 mutant does not impair their potential to invade macrophages and to differentiate into the amastigote form. This differentiation was independent of the presence of macrophages and could also be induced by temperature and pH shifts from 26°C and pH 7.0 to 34°C and pH 5.5 in the culture medium. In contrast to wild-type cells, the Δlmsap1/2 mutant ceased to proliferate after a 4-fold increase in cell number after 4 days without reaching stationary growth phase. This result is interpreted as follows. A late-logarithmic promastigote culture is not synchronized with respect to the differentiation state containing some non-dividing metacyclic cells. After triggering differentiation, the metacyclic cells will be able to differentiate into dividing amastigotes immediately; however, most of the c