A novel phospholipase from Trypanosoma brucei

Summary

Phospholipase A₁ activities have been detected in most cells where they have been sought and yet their characterization lags far behind that of the phospholipases A₂, C and D. The study presented here details the first cloning and characterization of a cytosolic PLA₁ that exhibits preference for phosphatidylcholine (GPCho) substrates. Trypanosoma brucei phospholipase A₁ (TbPLA₁) is unique from previously identified eukaryotic PLA₁ because it is evolutionarily related to bacterial secreted PLA₁. A T. brucei ancestor most likely acquired the PLA₁ from a horizontal gene transfer of a PLA₁ from Sodalis glossinidius, a bacterial endosymbiont of tsetse flies. Nano-electrospray ionization tandem mass spectrometry analysis of TbPLA₁ mutants established that the enzyme functions in vivo to synthesize lysoGPCho metabolites containing long-chain mostly polyunsaturated and highly unsaturated fatty acids. Analysis of purified mutated recombinant forms of TbPLA₁ revealed that this enzyme is a serine hydrolase whose catalytic mechanism involves a triad consisting of the amino acid residues Ser-131, His-234 and Asp-183. The TbPLA₁ homozygous null mutants generated here constitute the only PLA₁ double knockouts from any organism.

Introduction

Phospholipase A₁ (PLA₁) (EC 3.1.1.32) specifically hydrolyses the sn-1 acyl esters from phospholipids (PLs) producing free fatty acids (FAs) and lysophospholipids (lysoPLs). PLA₁ activities have been described in numerous cell types and tissues from a diverse range of organisms, but despite their apparent ubiquity and diversity relatively very few PLA₁s have been cloned and characterized.

A minority of PLA₁s cloned thus far possess a cytosolic subcellular localization, all three of which are homologues of one another and specifically utilize phosphatidic acid (GPA)¹ as substrate (Higgs et al., 1998; Tani et al., 1999; Nakajima et al., 2002). PLA₁ enzymes whose primary amino acid sequences exhibit homology with neutral lipase sequences can be categorized into a separate subgroup; these PLA₁, like their lipase homologues, are secreted from their cell (Soldatova et al., 1993; Hoffman, 1994; King et al., 1996; Sato et al., 1997; Watanabe et al., 1999; Sonoda et al., 2002). The preferred substrates observed for this group of enzymes so far are GPA (Sonoda et al., 2002) and phosphatidylserine (GPSer) (Sato et al., 1997). Secreted PLA₁ from bacteria that are not homologous to either lipases or mammalian PA-PLA₁ have also been cloned (Givskov et al., 1988; Schmiel et al., 1998). Finally, two unrelated sequences cloned from Arabidopsis thaliana code for PLA₁s localized to organelles (Ishiguro et al., 2001; Noiriel et al., 2004); a third (Kato et al., 2002) is homologous to mammalian PA-PLA₁ (Higgs et al., 1998).

The biological significance of any of the various PLA₁ forms is not well understood and has not been explored in depth. Mounting evidence, however, suggests that some PLA₁ act to generate bioactive lipid molecules, a role normally thought to be reserved for the phospholipases A₂, C and D. PS-PLA₁ secreted from platelets selectively forms the lipid mediator lysoGPSer (Sato et al., 1997) which can effectively stimulate histamine release from mast cells (Hosono et al., 2001). Also, a PLA₁ from A. thaliana has been shown to catalyse the release of linoleic acid, the initial and regulatory step in jasmonic acid biosynthesis, a multifunctional growth regulator in plants (Ishiguro et al., 2001). Furthermore, two GPA-utilizing PLA₁s are suspected to play a role in the regulation of intracellular membrane transport (Tani et al., 1999; Nakajima et al., 2002; Shimoi et al., 2005). PLA₁ have also been theorized to co-ordinate with other phospholipases to form lipid messengers such as lysoGPA, which has been shown to be produced by a PLD/PA-PLA₁ concerted mechanism and to activate the cellular receptor EDG7/LPA₃ in Sf9 cells (Sonoda et al., 2002).

Likewise, it has been postulated for decades that sn-2 FA cleavage from PLs may sometimes be mediated by concerted sequential PLA₁/LysoPLA activities (van den Bosch, 1980; Arthur, 1989; Burgoyne and Morgan, 1990; Badiani and Arthur, 1991; Pete et al., 1996; Pete and Exton, 1996; Wang et al., 1997; Ridgley and Ruben, 2001; Gauster et al., 2005). A relevant consequence of this type of reaction could be, for example, the discharge of arachidonic acid (AA), which can act as a second messenger and precursor of eicosanoids (Gijon et al., 2000).

The PLA₁/LysoPLA route towards the release of unsaturated FAs has been supported indirectly by in vitro studies using soluble subcellular fractions from bovine brain (Pete et al., 1996) and from the kinetoplastid protozoan parasite Trypanosoma brucei, the infective agent responsible for African Sleeping Sickness (Ridgley and Ruben, 2001). Both of these studies observed the requirement of sn-1 acyl cleavage prior to the release of AA from sn-1-palmitoyl-sn-2-arachidonoyl-sn-glycero-3-phosphatidylcholine [GPCho(16:0/20:4)]. In support of this, a soluble PLA₁ and a soluble LysoPLA, the latter of which selectively deacylates lysoGPCho(−/20:4), have been chromatographically purified from bovine brain (Pete et al., 1994; Pete and Exton, 1996). Moreover, LysoPLA activities in guinea pig microsomes have been observed to show preference for lysoGPCho(−/20:4) and lysoGPEth(−/20:4) (Arthur, 1989; Badiani and Arthur, 1991). Sequence identification of the enzymes responsible for the PLA₁ and LysoPLA activities from any of the examples outlined above has not been made.

The fate of unsaturated FAs liberated from lipids in T. brucei is not certain, but free AA has been implicated in regulating calcium mobilization in the cells (Eintracht et al., 1998; Catisti et al., 2000). Alternatively, the acid has been shown to serve as a precursor for prostaglandin biosynthesis in T. brucei (Kubata et al., 2000). Though studies on the utilization of unsaturated FAs in T. brucei have been initiated, none of the phospholipases responsible for FA release from PLs have been identified. However, sn-1 FA cleavage from GPCho by T. brucei homogenates has been observed to be robust and optimal at pH 6.0–8.5; and this activity was attributed to PLA₁ activity in this organism (Tizard et al., 1978a; Hambrey et al., 1981; Opperdoes and van Roy, 1982). The homogenates possessed PLA₁ activity that was 10-fold greater than its LysoPLA activity, though both activities appeared to originate from the same soluble enzyme fraction (Sage et al., 1981). A 380-fold enrichment in PLA₁ activity was achieved and the enzyme was purported to be 26 kDa (Hambrey et al., 1984). Fractions enriched in PLA₁ activity were unable to cleave myristate from the glycosylphosphatidylinositol anchor of the variant surface glycoprotein (Hambrey et al., 1986).

To date, the only cloned and characterized phospholipase in T. brucei is GPI-PLC, which is a phospholipase C implicated in the cleavage of the variant surface glycoprotein from its glycosylphosphatidylinositol anchor (Hereld et al., 1988; Carrington et al., 1998; Gruszynski et al., 2006). This report details the identification and characterization of a cytosolic PLA₁ from T. brucei that deacylates GPCho containing long polyunsaturated and highly unsaturated FAs.

Results

Identification and cloning of a T. brucei PLA₁ gene that has a prokaryotic origin

In order to identify putative PLA₁ genes in the T. brucei genome, a comprehensive search for phospholipases and lipases was carried out using the T. brucei genome database (GeneDB, Sanger Institute Pathogen Sequencing Unit, strain 927). Thirteen genes encoding putative lipolytic enzymes were identified, most of which contained the lipase consensus pattern [LIV] – {KG} – [LIVFY] – [LIVMST] – G – [HYWV] – S – {YAG} – G – [GSTAC] in their deduced amino acid sequence. The predicted protein sequence of each gene was subjected to local alignment analysis. Results revealed that one peptide sequence from T. brucei showed sequence similarity to PhlA, a secreted PLA₁ from the bacterium Serratia liquefaciens (Givskov et al., 1988). The T. brucei phospholipase A₁ (TbPLA₁) was only homologous to bacterial PLA₁ isoforms, including YplA, a PhlA homologue from Yersinia enterocolitica (Schmiel et al., 1998).

Very interestingly, the homologue with the greatest overall similarity to TbPLA₁ is a putative PLA₁ from the proteobacterium Sodalis glossinidius (morsitans strain) (Toh et al., 2006), an intracellular endosymbiont of Glossina (tsetse) flies, the insect vector host of T. brucei. There are no eukaryotic homologues of TbPLA₁, including in two other kinetoplastid members Trypanosoma cruzi and Leishmania spp., the aetiologic agents of Chagas disease and Leishmaniasis respectively. However, orthologues of TbPLA₁ were found in existing GeneDB coding sequences from the livestock-infective Trypanosoma congolense and Trypanosoma vivax. TbPLA₁ does not share similarity to neutral lipases nor other previously reported eukaryotic PLA₁ sequences, apart from possessing a lipase consensus motif.

Using sequence-specific primers, putative TbPLA₁ was amplified from T. brucei genomic DNA (Lister 427 strain) by PCR and the 903 bp open reading frame (ORF) sequenced (GenBank™ accession number AJ716154). An alignment of the TbPLA₁ predicted amino acid sequence with those of the other Trypanosoma spp. and bacterial homologues is shown (Fig. 1). The deduced primary sequence of TbPLA₁ shares 21.5%, 18% and 18% identity with its homologues from S. glossinidius, S. liquefaciens and Y. enterocolitica respectively; whereas T. congolense and T. vivax possess 41% and 37% identity with TbPLA₁ respectively.

Fig. 1.

Multiple sequence alignment of trypanosome and bacterial PLA₁. Amino acid sequences were aligned with MUSCLE (Edgar, 2004) and edited using the BOXSHADE algorithm. Amino acid residues below circles represent those which were mutated to reveal active-site components (closed circles) and non-active-site residues (open circles). The lipase consensus pattern is underscored by a double line and the signal sequence utilized for PhlA secretion (Givskov et al., 1988) is underlined. TcPLA₁, Trypanosoma congolense PLA₁ (product of GeneDB Systematic name congo50h08.q1k_9); TbPLA₁, Trypanosoma brucei PLA₁ (Accession Number CAG29794); TvPLA₁, Trypanosoma vivax PLA₁ (product of GeneDB Systematic name Tviv1355g06.p1k_2); SgPLA₁, Sodalis glossinidius PLA₁ (Accession Number Q2NQT2); YplA, Yersinia enterocolitica PLA₁ (Accession Number O85477); PhlA, Serratia liquefaciens PLA₁ (Accession Number P18952).

The bacterial sequences possess a signal sequence used in targeted secretion that appears to be absent or very altered and unrecognizable (by the SIGNALP algorithm) in the trypanosome sequences. This finding was the first suggestion of an intracellular localization for TbPLA₁. All in all, these results provide the first evidence that TbPLA₁ is a unique eukaryotic PLA₁ because it is solely related to PLA₁ from bacteria. Furthermore, the high similarity of the TbPLA₁ sequence to that of a bacterial endosymbiont of the vector responsible for transmission of T. brucei parasites to mammals strongly suggests that a T. brucei ancestor acquired the PLA₁ gene through horizontal gene transfer (HGT) during cohabitation with the bacterium inside an infected tsetse fly.

Expressed and purified recombinant TbPLA₁ exhibits sn-1 exclusivity

To determine whether TbPLA₁ could cleave FAs esterified to the sn-1 position of PLs, a source of pure and active protein with which to carry out in vitro analysis was sought using recombinant protein expression techniques. The entire TbPLA₁ ORF was subcloned into the expression vector pET15b. Subsequent transformation of Escherichia coli and induction of TbPLA₁ expression by IPTG produced de novo synthesized soluble TbPLA₁ with an N-terminal His·tag (Fig. 2A, lane 3). The soluble enzyme was purified to homogeneity using nickel affinity chromatography followed by anion exchange (Fig. 2A, lanes 6 and 7). The pure His·tag-cleaved protein consistently migrated as a larger molecular weight protein by SDS-PAGE than expected from its predicted 34.80 kDa size, though peptide mass fingerprinting and MALDI-TOF mass spectrometry analysis shows that the protein is indeed TbPLA₁ with a nominal molecular weight of 34.98 kDa respectively (data not shown).

Fig. 2.

Purification and regiospecificity of recombinant TbPLA₁.

A. SDS-PAGE analysis of proteins from TbPLA₁-overexpressing E. coli cells during various purification steps as compared with molecular weight standards (lane 1). The various fractions include insoluble proteins (lane 2), the soluble protein fraction containing His·tagged recombinant TbPLA₁ (lane 3), nickel column flow through (lane 4), nickel affinity eluate (lane 5), purified His·tagged recombinant TbPLA₁ after anion exchange (~3 μg) (lane 6), purified His·tag-cleaved TbPLA₁ after thrombin digestion and anion exchange (~3 μg) (lane 7).

B. Phospholipase activity was measured at various pH values using the fluorescent BODIPY®C₁₁-PC assay at a mole fraction of 0.018 with a final concentration of substrate at 0.075 mM. A universal buffer containing 50 mM Tri-sodium citrate, 50 mM Tris and 50 mM NaCl was used to obtain the various pH conditions. Data points represent the activity measured relative to the maximal pH-activity value of pH 7.0. Data represent the average from two experiments performed in triplicate.

C. Metabolism of GPCho(16:0/[³H]18:0) after a 5 min reaction is shown as a function of varying amounts of recombinant TbPLA₁. Formation of lysoGPCho(−/[³H]18:0) appears to be dose-dependent, whereas [³H]18:0 acid cleavage from the sn-2 position is undetectable.

The pH dependence for recombinant TbPLA₁ was assessed using the Bis-BODIPY® FL-C₁₁-PC assay. Results demonstrated that the enzyme shows highest specific activity between pH 6.0 and 8.5, and that enzyme activity is abolished at low pH (Fig. 2B). The optimal activity was detected at pH 7.0 and this pH was used in all other subsequent measurements of TbPLA₁ activity.

The regiospecificity of cleavage for recombinant TbPLA₁ was initially examined by incubating it with radiolabelled GPCho(16:0/[³H]18:0). This substrate was presented to the enzyme in the form of TX-100/GPCho mixed micelles. Both the formation of free [³H]18:0 acid and lysoGPCho(−/[³H]18:0) was monitored as a function of TbPLA₁ concentration (Fig. 2C). Results clearly show that lysoGPCho(−/[³H]18:0) formation mirrors the metabolism of GPCho(16:0/[³H]18:0), indicating sn-1 activity. TbPLA₁ apparently has no LysoPLA₂ activity because the resultant product from the first reaction, lysoGPCho(−/[³H]18:0), cannot serve as a substrate for any subsequent reactions, based on the observation that [³H]18:0 acid from position sn-2 was not detected. When Ca²⁺, sometimes required as a cofactor for soluble PLA activity, was included in the assay, no [³H]18:0 acid was seen to accumulate and no enhancement of sn-1 activity was observed. Taken together, these results indicated that recombinant TbPLA₁ hydrolyses FA esters exclusively at the sn-1 position of the glycerol backbone of GPCho.

TbPLA₁ most efficiently hydrolyses GPCho molecules

The ability of recombinant enzyme to hydrolyse sn-1 hydrocarbon chains of neutral lipid substrates and PLs containing different headgroups was investigated by quantifying any free FA reaction products by gas chromatograph mass spectrometer (GC-MS) after their derivatization into methyl esters. A summary analysis of the reaction products from an assortment of lipid substrates indicated that purified TbPLA₁ deacylated a variety of choline-containing PLs with greater efficiency than PLs containing ethanolamine, inositol, serine, or just phosphorous at the sn-3 position (Table 1). When acyl chain length is held constant with palmitate, substrate preference was observed, in decreasing order, as GPCho(16:0/16:0)> GPSer(16:0/16:0)>>GPA(16:0/16:0)=GPEth(16:0/16:0), the latter two were hydrolysed at around 1/10 the rate of GPCho(16:0/16:0).

Table 1.

Hydrolysis of lipid substrates by TbPLA₁

Substrate	PLA1 activitya (μmol min−1 mg−1
GPCho(16:0/18:1)	120.2 ± 18.7
GPCho(18:0/20:4)	118.4 ± 5.4
GPCho(16:0/14:0)	114.3 ± 6.0
GPCho(14:0/18:0)	110.9 ± 2.0
GPCho(14:0/14:0)b	100.4 ± 8.7
GPCho(16:0/16:0)b	84.8 ± 7.4
GPCho(16:0/18:0)	78.5 ± 7.9
GPCho(18:0/18:0)b	54.0 ± 7.6
GPSer(16:0/16:0)b	49.2 ± 7.6
lysoGPCho(14:0/−)	16.1 ± 1.0
GPA(16:0/16:0)b	8.6 ± 0.7
GPEth(16:0/16:0)b	7.3 ± 2.4
Bovine liver GPInsb,c	0.7 ± 0.04d
DG(16:0/16:0)b	0.5 ± 0.06d
TG(18:0/18:0/18:0)	NDd
GPCho(1-0-16:0/16:0)	ND

^aAssays were conducted as described under Experimental procedures. Activity is expressed as micromoles of sn-1 FA released per min mg⁻¹ of protein. Values are means ± SD, n = 3 determinations, each performed in triplicate. ND, none detected.

^bRegiospecificity of FA hydrolysis detected cannot be discerned with this substrate.

^cThe only FAME detected was that from stearic acid (C18:0), which comprises nearly 50% of the FAs in bovine liver GPIns.

^d1.85 μg of enzyme was added instead of 0.1 μg, and the reaction was extended to 20 min. Less than 10% of substrate was turned over.

TbPLA₁ displayed some LysoPLA₁ activity towards lysoGPCho(14:0/−) that amounted to 14.5% and 16.0% of its PLA₁ activity towards GPCho(14:0/18:0) and GPCho(14:0/14:0) respectively. A variety of molecular species of GPCho were hydrolysed with average rates between 54.0 and 120.2 μmol min⁻¹ mg⁻¹. Diacylglycerol (DG) and bovine liver GPIns showed exceedingly poor substrate capabilities because their rate of hydrolysis was less than 1/100 the rate of GPCho even when 18.5-fold more enzyme was added. No fatty acyl hydrolysis from plasmalogen GPCho or from triacylglycerol (TG) could be detected. These results show that TbPLA₁ exhibits robust activity towards GPCho and that GPCho, and to a lesser extent GPSer, is the preferred substrate for the enzyme in vitro.

TbPLA₁ activity is mediated by a Ser-His-Asp catalytic triad

Although TbPLA₁ is not homologous to neutral lipases, its amino acid sequence contains a lipase consensus pattern that harbours a conserved GXSXG motif (Fig. 1), an indicator that TbPLA₁ may be a member of the serine hydrolase superfamily. The catalytic triad in these enzymes is usually occupied by a base residue (His), an acid (Asp) and a nucleophile (Ser), the latter of which is identifiable as the S in the GXSXG motif. The segments of sequence in which the base and acid residues lie are not as obvious due to the regions’ generally lower degree of conservation, but the acid is often located well between the nucleophile and base. Based on this information it was predicted that Asp-183 could be an acid residue involved in TbPLA₁ catalysis (Fig. 1).

Nevertheless, two conserved aspartic acid residues in recombinant TbPLA₁ (Asp-63 and Asp-183), along with a third (Asp-232), which is conserved in the trypanosomes, were mutated to alanine by site-directed mutagenesis in order to compare activity to recombinant wild-type (WT) protein. Mutant enzyme activity was also examined after replacing the T. brucei GXSXG serine (Ser-131) and the only conserved histidine residue (His-234) to alanine. All of the mutant TbPLA₁ proteins were expressed and purified at the same levels as WT TbPLA₁ with the exception of the D183A mutant, which produced fourfold less protein (Fig. 3). The activity of the D63A and D232A mutants maintained an average of 89% and 67% of recombinant WT activity respectively, implying that the presence of these residues is not obligatory for TbPLA₁ activity. The TbPLA₁ activity measured for the D183A, H234A and S131A mutants was less than 5% of the activity observed for the WT enzyme, but nearly equal to that of the no-enzyme control (Fig. 3). The negation of activity for three enzymes containing individual serine, histidine and aspartic acid residue mutations suggests that hydrolysis of substrate was prevented due to an interruption of a charge relay system that is known to be invoked by the Ser-His-Asp catalytic triads in serine hydrolases.

Fig. 3.

Probing for active-site residues. Mutant TbPLA₁ proteins were expressed and purified from E. coli in parallel as stated in the Experimental procedures and 10 μl of the final eluate was analysed by SDS-PAGE (above graph). The radiolabelled PLA₁ assay was used with 5 ng of purified WT and mutant TbPLA₁ to compare substrate hydrolysis after 10 min; a no-enzyme negative control was tested in parallel. Mutant activities are shown as relative percentages of WT activity (26 μmol min⁻¹ mg⁻¹).

Generation of TbPLA₁ mutants

As no PLA₁ has ever before been directly studied in vivo in its native cell, the next line of experiments sought to do just that in order to gain insight into TbPLA₁ function in the trypanosome itself. To set conditions for which to accomplish this, experiments which led to the creation of a double knockout of TbPLA₁ were performed. Promoterless deletion constructs (5′-puro^R-3′, 5′-bsd^R-3′) targeted to the 5′ and 3′ flanking regions of the TbPLA₁ ORF were generated and genes for antibiotic resistance were cloned between the flanking regions. These constructs were used to stably transform PCF stage ‘WT’ cell lines, which are in fact transgenic clones that constitutively express T7 RNA polymerase and the tetracycline repressor protein (Wirtz et al., 1994). TbPLA₁ is a single copy gene (data not shown) and therefore two successive rounds of gene replacement by UTR-targeted homologous recombination with 5′-puro^R-3′ and 5′-bsd^R-3′ were sufficient in this case to replace the two alleles of TbPLA₁. Evidence for successful generation of the null mutant cell line Δpla1 is shown (Fig. 4A).

Fig. 4.

Validation and characterization of TbPLA₁ mutant cell lines.

A. TbPLA₁ null mutants were generated in the procyclic form of T. brucei after sequential TbPLA₁ UTR-targeted homologous recombination with UTR-flanked puromycin and blasticidin resistance genes. Cell lines were analysed by Southern blot after digesting genomic DNA with EcoRV. EcoRV-digested T. brucei WT gDNA reveals one size of DNA fragment detectable by fluorescein-labelled TbPLA₁ ORF probe. In TbPLA₁ null mutant cells (Δpla1) both TbPLA₁ alleles are absent. A tetracycline (Tet)-inducible myc-tagged recombinant ectopic copy of TbPLA₁ (PLA1-myc) cloned into a phleomycin-resistant pLEW100 cassette was introduced into a different locus in WT gDNA to produce transgenic TbPLA₁ overexpression cell lines (ovexPLA1-myc), or introduced in the null mutant gDNA to produce a rescue cell line (Δpla1 rescPLA1-myc).

B. Northern blot analysis of TbPLA₁ mutants. Total RNA extracted from WT and TbPLA₁ mutant trypanosomes were hybridized with TbPLA₁ (top panels), loading controls are also presented (bottom panels).

C. Western blot analysis of cell lysates of TbPLA₁ mutants. Both native PLA₁ (theoretically 32.4 kDa) and tagged PLA1-myc (theoretically 33.9 kDa) proteins were detected by antibodies raised against the purified recombinant enzyme. * = background bands that show loading in the null mutant lanes.

Genetic transformation was also carried out that served to generate cell lines which overexpressed TbPLA₁ using the T. brucei tetracycline-inducible overexpression vector pLEW100 (Wirtz et al., 1999). pLEW100/PLA1-myc was engineered by cloning a recombinant form of TbPLA₁ into pLEW100. Upon tetracycline induction, trypanosomes transformed with pLEW100/PLA₁-myc produce an ectopic copy of TbPLA₁ with a C-terminal myc tag (PLA1-myc). In PCF cells, both the WT and Δpla1 cell lines were stably transformed with pLEW100/PLA1-myc (Fig. 4A). The respective ovexPLA1-myc overexpression and Δpla1 rescPLA1-myc rescue clones, under tetracycline induction, produced recombinant ectopic TbPLA₁ mRNA transcripts that were, on average, > 23-fold over the level of TbPLA₁ transcripts in both their WT and respective uninduced counterpart cells (Fig. 4B). In culture, the TbPLA₁ null mutants and tetracycline-induced TbPLA₁ overexpression cell lines have no apparent phenotype in morphology or in growth rate (data not shown).

The various overexpression clones were examined for their ability to translate the overexpressed TbPLA₁-myc mRNA. A Western blot probed with anti-TbPLA₁ antibodies revealed that overexpression of PLA1-myc was minimal when compared with the level of mRNA overexpression (Fig. 4C). Compared with uninduced ovexPLA1-myc cells, its induced counterpart cells produced, on average, only 2.2 times the amount of TbPLA₁ protein (native PLA1 + PLA1-myc). This observed level of regulation at the level of mRNA production to protein translation is not surprising considering the potential damage a membrane-lytic enzyme such as TbPLA₁ could invoke if expressed at uncontrollable levels in the cell.

TbPLA₁ synthesizes unsaturated lysoGPCho metabolites in vivo

To be able to capture metabolites created via PLA₁ action in vivo, a highly sensitive detection method which enabled identification and characterization of global changes in T. brucei PLs was needed. To this end, nano-electrospray ionization tandem mass spectrometry (nano-ESI-MS-MS) was employed to analyse individual PL classes present in WT, Δpla1, ovexPLA1-myc and Δpla1 rescPLA1-myc cells. In positive ion mode, collision-induced dissociation of PLs containing a choline head group results in the production of a unique fragment ion of m/z 184 (Kerwin et al., 1994). Precursor ion scanning for m/z 184 from total lipids extracted from WT cells selectively detected [M+H]⁺ ions of GPCho and sphingomyelin (Fig. 5A). Ions in the precursor m/z 184 ion scan spectra were annotated based on their [M+H-NMe₃⁽⁺⁾]⁺ daughter ion masses compared with that of their theoretical values.

Fig. 5.

TbPLA₁ functions to synthesize long-chain unsaturated lysoGPCho intermediates.

A. Positive ion ESI-MS-MS spectrum of m/z 184 lipid precursors in total lipid extracts from T. brucei WT cells. Identities of the major [M+H]⁺ ions are indicated as well as internal standard lysoGPCho [M+H]⁺ peaks.

B. Positive ion ESI-MS-MS short-range spectra of the parents of the m/z 184 ion from total lipid extracts from WT and TbPLA₁ mutant cell lines. The major sets of lysoGPCho [M+H]⁺ metabolites detected are boxed and annotated next to the m/z peak from which they are derived. The numbers in place of x:y refer to the total number of sn-2 FA carbon atoms (x) and their degree of unsaturation (y). Panel insets are similar short-range spectra from lipid extracts spiked with lysoGPCho internal standards used for quantitative purposes (Table 2).

The major peak areas in the m/z 184 precursor ion spectrum represent GPCho with the following number of FA carbons and their degrees of saturation (represented as their sum): m/z 834, C40:6; m/z 836, C40:5; m/z 838, C40:4; m/z 786, C36:2; m/z 788, C36:1; m/z 790, C36:0; m/z 774, alkylacyl C36:1 and/or alkenylacyl C36:0; m/z 772, alkylacyl C36:2 and/or alkenylacyl C36:1; m/z 810, C38:4; m/z 812, C38:3; m/z 814, C38:2.

Detection of lysoGPCho metabolites formed from either PLA₁ or PLA₂ activity was facilitated from m/z 184 precursor ion scanning of total lipids in a mass range (m/z 400–700) which excluded the dominating diacyl and alkylacyl/alkenylacyl GPCho species. Peak analysis of short range spectra from PCF WT cell lipids resulted in the identification of a set of [M+H]⁺ lysoGPCho metabolites with the following FA constituents: m/z 568, C22:6; m/z 570, C22:5; m/z 572, C22:4; m/z 520, C18:2; m/z 522, C18:1; m/z 524, C18:0; m/z 548, C20:2; m/z 546, C20:3; m/z 544, C20:4 (Fig. 5B, panel 1). The abundance of these long-chain mostly polyunsaturated and highly unsaturated lysoGPCho metabolites is markedly lower than nearly all of the diacyl GPCho species as shown by its relative proportion to the minor diacyl species GPCho(28:1) and GPCho(28:0).

To be able to quantify the amount of long-chain unsaturated lysoGPCho intermediates synthesized by TbPLA₁ activity, lysoGPCho series peak intensities were also compared with internal standard peak intensities of non-natural lysoGPCho(17:0/−) and lysoGPCho(24:0/−) added to each sample prior to lipid extraction (Fig. 5B, panel 1 inset). The calculated amounts of lysoGPCho metabolites from WT cellular lipids are tabulated as shown (Table 2). A permeating feature of the lysoGPCho molecules synthesized in T. brucei is that their acyl moiety is composed of a long and highly unsaturated chain of 22 carbons in nearly 50% of the metabolite population. This observation is consistent with the notion of the suspected presence of these FAs in some of the most abundant diacyl GPCho species (m/z 834–836), if the sn-1 position was occupied by C18:0 (Fig. 5A) (Patnaik et al., 1993).

Table 2.

Quantitative analysis of lysoGPCho metabolites in TbPLA₁ mutants.^a

		lysoGPCho seriesb
Cell line	Tet	(−/18:y)	(−/20:y)	(−/22:y)	Total
WT		53	31	72	156
Δ pla1		15	8	7	30
ovexPLA1-myc	−	56	22	81	159
	+	159	42	231	432
Δ plal rescPLAI-myc	−	29	14	23	66
	+	52	22	54	128

^aTotal lipid extracts from 10⁸ PCF T. brucei cells spiked with 500 pmol of each internal standard [lysoGPCho(17:0/−) and lysoGPCho(24:0/−)] were examined by ESI-MS-MS. lysoGPCho molecules were detected from precursor ion scanning for m/z 184 (see insets from Fig. 5B), and they were quantified by comparison with the internal standards. Values are from a typical analysis and are presented as pmol per 10⁸ cells. Tet, tetracycline.

^bIntegration of peaks to obtain peak areas was performed by grouping lysoGPCho into various series of peaks whereby one series comprises both the major and minor lysoGPCho molecules, and their isotopes, containing the same FA chain length but with varying degrees of unsaturation, represented by ‘y’.

Equivalent systematic ESI-MS-MS analysis of PCF Δpla1 cells revealed that the levels of lysoGPCho metabolites in the null mutants decreased on average by 81% when compared with WT levels (Table 2) (Fig. 5B, panel 2 and inset). There was a 90% decrease in the number of moles of lysoGPCho(−/22:y), whereas lysoGPCho(−/18:y) and lysoGPCho(−/20:y) levels decreased by 72% and 74% in the TbPLA₁ null mutants respectively. The variety and abundance of GPCho(28:1)/GPCho(28:0) (Fig. 5B, panel 2) and the other diacyl GPCho molecular species (data not shown) in the null mutant remained unaltered. Overall, these results suggest that in vivo lysoGPCho biosynthesis in T. brucei is largely dependent upon TbPLA₁ activity. Furthermore, due to the sn-1 exclusivity of TbPLA₁ (Fig. 2 and Table 1), the FA of each lysoGPCho is esterified at the sn-2 position of these metabolites. It is unknown whether the small amount of lysoGPCho species still present in the null mutants is due to enzymatic synthesis in vivo by an alternative PLA, or whether they represent background levels of fragment ions randomly and mechanically formed from diacyl GPCho during the sample preparation and/or the ionization process. However, the observation that lysates of Δpla1 cells could not produce any lysoGPCho supports the idea that these cells are most likely totally devoid of unsaturated lysoGPCho molecules (Fig. 6).

Fig. 6.

Phospholipase activity from lysates of T. brucei cells. PCF phospholipase activity from 10⁶ cell equivalents was examined after 20 min incubation with GPCho(16:0/[³H]18:0). The TLC autoradiograph shows the metabolism of GPCho(16:0/[³H]18:0) into lysoGPCho(−/[³H]18:0) by PLA₁ in WT and single knockout (Δpla1−/+) lysate. PLA₁ activity is absent in the cell-free control and the Δpla1 double knockout cell line. To stimulate any possible PLA₂ activity in the lysates 10 mM CaCl₂ was added to each assay. Even with the dominant PLA₁ activity being absent from Δpla1 clones, no PLA₂ activity was observed in their lysate. Free [³H]18:0 was used as a standard to monitor PLA₂ activity and its R_f is indicated. O, origin; F, front.

Confirmation that TbPLA₁ mediates lysoGPCho metabolite synthesis in vivo comes from GPCho analysis in ovexPLA1-myc cells. Tetracycline-induced overexpression of TbPLA₁ in these cells resulted in an elevated abundance of lysoGPCho compared with either WT or non-induced ovexPLA1-myc cells (Fig. 5B, panels 3 and 4 and their respective insets). The levels of lysoGPCho intermediates from the (−/18:y) and (−/22:y) series in the overexpression cell line reached 3.0 and 3.2 times higher than normal PCF levels of these metabolites respectively (Table 2). Interestingly, the (−/20:y) series intermediates only showed a modest (1.4-fold) increase after tetracycline induction.

TbPLA₁-mediated synthesis of T. brucei long-chain polyunsaturated and highly unsaturated lysoGPCho was also established using the PCF Δpla1 rescPLA1-myc rescue clones. The Δpla1 ablated lysoGPCho(−/18:y) metabolite levels were nearly restored (98%) to WT levels upon induction with tetracycline in Δpla1 rescPLA1-myc cells, whereas the non-induced cells continued to display the Δpla1 phenotype (Fig. 5B, panels 5 and 6 and their respective insets) (Table 2). The number of moles of lysoGPCho(−/20:y) and lysoGPCho(−/22:y) were brought back to 71% and 75% of normal levels in the PCF rescue cell line.

Data from ESI-MS-MS analysis of TbPLA₁ mutants collectively revealed a range of lysoGPCho metabolites whose production is at least partly regulated by the levels of TbPLA₁ in the cell. A comprehensive short range search for lysoGPEth, lysoGPSer, lysoGPIns and lysoGPA yielded negative results; this supports previous findings that lysoGPCho in T. brucei is the predominant lysoPL synthesized by this parasite (Dixon and Williamson, 1970; Patnaik et al., 1993). The robust activity of TbPLA₁ towards GPCho both in vitro and in vivo strongly suggests that GPCho is the preeminent substrate for this enzyme.

The global PL composition of the TbPLA₁ mutants generated in this study was also examined, but the high mass range spectrums of diacyl GPCho, GPEth, GPSer and GPIns from the TbPLA₁ mutants remained unchanged compared with the composition of these lipids in WT cells (data not shown).

TbPLA₁ is constitutively expressed in T. brucei and its gene product is located in the cytosol

The first empirical indication that TbPLA₁ may be a cytosolic protein came from observing that a high-speed soluble fraction (obtained from centrifuging lysate at 100 000 g for 1 h) from WT cells, but not lysate from TbPLA₁ null mutant cells, could quickly form lysoGPCho in our radiolabelled assay (data not shown). The observations that optimal TbPLA₁ activity occurs at a neutral pH and that sequence analysis suggests that this protein has lost its signal peptide also support the notion that TbPLA₁ resides in the cytosol. Therefore, to confirm the intracellular location of the native enzyme, the presence of TbPLA₁ in subcellular fractions of T. brucei was investigated. A Western blot of subcellular fractions probed with antibodies raised against the purified recombinant enzyme reveals that native TbPLA₁ is recognizable as a soluble protein that localizes entirely in the cytosolic protein fraction in both bloodstream (BSF) and PCF WT T. brucei (Fig. 7A, lanes 3 and 4). No TbPLA₁ is detected in the cytosolic fraction of PCF Δpla1 cells (Fig. 7A, lane 5), nor in the large granular or microsomal fractions of BSF WT cells (Fig. 7A, lanes 1 and 2).

Fig. 7.

Cellular distribution of TbPLA₁.

A. Subcellular fractions were prepared as described in Experimental procedures and used in immunoblotting with anti-TbPLA₁ antibodies. TbPLA₁ is constitutively expressed and was detected in the cytosolic fraction of both BSF and PCF WT cells (lane 3 and lane 4 respectively) but absent from PCF TbPLA₁ null mutants (lane 5). TbPLA₁ is also absent from the large granular and small granular (microsomal) fractions from BSF WT cells (lane 1 and lane 2 respectively).

B. PCF TbPLA₁ null mutant cells transformed with an ectopic copy of an N-terminal EGFP-PLA1 gene fusion (Δpla1 EGFP-PLA1-myc) or a C-terminal PLA1-EGFP gene fusion (Δpla1 PLA1-EGFP) were analysed after tetracycline induction by anti-TbPLA₁ (top panel) or anti-EGFP (lower panel) immunoblotting for the presence of expressed fusion protein (~60 kDa). WT cells and null mutants transformed with an ectopic copy of EGFP only (Δpla1 EGFP) serve as negative controls for fusion protein expression. * = background bands just smaller than native TbPLA₁ that show loading in the null mutant lane.

C. Analysis of the subcellular localization by immunofluorescence microscopy of both PLA1/EGFP fusion proteins (right panels) compared with EGFP only confirms a cytoplasmic distribution for soluble TbPLA₁. PCF transgenic mutants are also shown in phase contrast (left panels) and after DAPI staining (middle panels).

In a further gesture to pinpoint TbPLA₁ subcellular location, Δpla1 cells transformed with an altered pLEW100/PLA1-myc engineered to contain EGFP fused at either the N-terminus (Δpla1 EGFP-PLA1-myc) or C-terminus (Δpla1 PLA1-EGFP) of TbPLA₁ were generated. TbPLA₁ expressed as a fusion protein with EGFP upon tetracycline induction of Δpla1 PLA1-EGFP and Δpla1 EGFP-PLA1-myc cells. Evidence that full-length fusion proteins were produced by these cell lines was obtained by Western blot analysis. Anti-TbPLA₁ antibodies detected the fusion proteins in the expected 60 kDa range, whereas the WT TbPLA₁ was detected in its native 35 kDa range (Fig. 7B, upper panel). A copy of pLEW100 that contained only EGFP was also introduced into the null mutants as a control cell line (Δpla1 EGFP). Western blot analysis using anti-EGFP antibodies showed that Δpla1 EGFP cells expressed non-fused EGFP (27 kDa) after tetracycline induction (Fig. 7B, lower panel). Both Δpla1 EGFP-PLA1-myc and Δpla1 PLA1-EGFP induced cells expressing the PLA1/EGFP fusion proteins fluoresced upon UV exposure in a pattern very similar to cells expressing EGFP only, which is known to be a soluble protein localized to the cytosol of cells in which it is expressed (Fig. 7C). The localization studies taken together thus suggest that TbPLA₁ functions by hydrolysing sn-1 FAs of GPCho molecules exposed to the neutral environment of the cytoplasm.

Discussion

This report describes the identification and characterization of a PLA₁ from T. brucei, the first phospholipase of type A to be identified from this parasite, and the only PLA₁ cloned from any protist. TbPLA₁ should be considered a true PLA₁ because it possesses high and specific activity towards sn-1 esters of diacyl PLs, exhibits LysoPLA₁ activity that is a fraction of its PLA₁ activity, and shows very little activity towards neutral lipid substrates.

TbPLA₁ is distinctive from previously identified PLA₁ from other eukaryotes because it is only homologous to prokaryotic extracellular PLA₁ enzymes. Outside of its own Genus the TbPLA₁ peptide sequence most resembles a PLA₁ homologue from S. glossinidius, a recently established bacterial secondary endosymbiont of the Glossina (tsetse) fly (Aksoy et al., 1995). This finding together with an analysis of trypanosomatid phylogeny allows us to propose that a T. brucei ancestor acquired the S. glossinidius PLA₁ gene through HGT after/during its adaptation to a parasitic lifestyle in its tsetse host. The PLA₁ HGT is predicted to have occurred before the species in the ‘T. brucei’ clade of lineages diverged from one another, but after they diverged from the species in the clade containing T. cruzi, T. lewisi and T. theileri (Fig. 8). This hypothesis is supported by several observations and deductions. First, TbPLA₁ homologues are absent in T. cruzi or the Leishmania spp. These two kinetoplastids do not use Glossina as a host and it is unlikely that they acquired PLA₁ through vertical gene transfer from an early kinetoplastid ancestor and then subsequently lost it. Second, an S. glossinidius PLA₁ homologue is present in T. congolense and T. vivax, whose lineages diverge within the same clade as T. brucei. Third, T. brucei, T. congolense and T. vivax all infect Glossina flies. Last, both T. lewisi and T. theileri, which have been reported to possess very little PLA₁ activity (Tizard et al., 1977a; 1980), are rooted to a clade outside of T. brucei and T. congolense, trypanosomes that have high PLA₁ activity (this study) (Tizard et al., 1978b,c; Hambrey et al., 1980). Unlike prokaryote–prokaryote HGT, HGT between prokaryote and eukaryote is still regarded as more rare and atypical, and it is most common in protists (Boucher and Doolittle, 2000; Field et al., 2000; de Koning et al., 2000). On the whole, the probable HGT that created trypanosome PLA₁ from a proteobacteria contributed to the evolution of the T. brucei lineage, and it supports the idea of the coevolution of trypanosomes with their hosts.

Fig. 8.

Proposed HGT point for S. glossinidius PLA₁. The arrow indicates the proposed acquisition of S. glossinidius PLA₁ by an ancestor of T. brucei. The maximum likelihood tree was calculated using the Phylip v3.6 Dnamlk algorithm and based on a MUSCLE alignment of glycosomal glyceraldehyde 3-phosphate dehydrogenase (gGAPDH) genes from various kinetoplastids. Values at nodes are maximum likelihood bootstrap values (%) obtained from 100 replicates (Seqboot). The tree is rooted to the corresponding gene of Euglena gracilis. The tree phylogenies agree very well with the more detailed and exact gGAPDH trees previously published (Hamilton et al., 2004; 2005). The gGAPDH sequences were obtained from GenBank and their accession numbers are: Trypanosoma theileri, AJ620282; Trypanosoma lewisi, AJ620272; Trypanosoma cruzi, AJ620269; Trypanosoma rangeli, AF053742; Trypanosoma simiae, AJ620293; Trypanosoma congolense, AJ620291; Trypanosoma godfreyi, AJ620292; Trypanosoma brucei, AJ620284; Trypanosoma vivax, AJ620295; Crithidia fasciculata, AF047493; Leishmania major, AF047497; Euglena gracilis, L21903.

This study successfully employed both in vitro and in vivo analytical techniques to establish a correlation that GPCho is the preferred substrate class upon which TbPLA₁ acts. The results from this study also provide evidence that TbPLA₁ is an intracellular enzyme localized to the cytosol, which is in stark contrast to the fate of its secreted bacterial homologues. Therefore, in addition to its distinct evolution, TbPLA₁ is also a novel PLA₁ because it is the only cytosolic PLA₁ thus far discovered whose preferred substrates are GPCho molecules. In fact, of all the PLA₁ cloned and characterized across taxa, only the two organelle-associated PLA₁s from A. thaliana have been shown to be able to hydrolyse GPCho to some extent (Ishiguro et al., 2001; Noiriel et al., 2004). Just three other cytosolic PLA₁ enzymes have been cloned previously from other organisms, all of which are mammalian in origin and preferentially hydrolyse GPA (Higgs et al., 1998; Tani et al., 1999; Nakajima et al., 2002). TbPLA₁, on the other hand, shows very little activity towards GPA.

GPCho is the major PL in T. brucei, constituting approximately 57% of all the PL species present in T. brucei PCF parasites. The entire GPCho pool is not, however, a potential source of substrate for TbPLA₁; roughly a quarter of the GPCho pool consists of non-hydrolysable plasmanyl/plasmenyl species (Patnaik et al., 1993). Furthermore, in most eukaryotic cells GPCho is thought to be the major PL of the outer leaflet of the plasma membrane, a location which would render many GPCho substrates inaccessible to TbPLA₁. Though TbPLA₁ presumably has access to PLs on the inner leaflet of the plasma membrane, the presence of GPSer and GPEth, which are widely accepted to reside in large part on the inner leaflet of the plasma membrane in their plasmanyl/plasmenyl form, could also naturally limit to a certain extent the availability of GPCho to TbPLA₁ in this location. Though the amount of GPCho distributed along cytosolic surfaces in the cell is unknown, it is quite possible that the majority of TbPLA₁ substrates are part of the outer leaflets of organelles and vesicles, and this includes the ER, where GPCho is de novo synthesized.

An important component in this study was the desire to answer some fundamental questions concerning the nature of T. brucei PLA₁ activity. The absence of a crystal structure for any true PLA₁² enzyme precludes making deductions on their mechanisms of action without empirical evidence. As such, to better comprehend the structural basis for the functional characteristics observed for TbPLA₁, the identification of the catalytic residues involved in sn-1 ester hydrolysis were sought. Sitedirected mutagenesis of recombinant TbPLA₁ provided evidence that the catalytic mechanism employed by TbPLA₁ is similar to that of the serine hydrolases, in that its activity appears to rely on the concerted action of a catalytic triad composed of Ser-131, Asp-183 and His-234. In theory, loss of activity from any of the mutants could have been due to conformational changes or protein misfolding engendered by the substituted alanine during protein expression. However, due to the good understanding of the catalytic mechanisms used by hydrolases, and the residues involved, it is unlikely that the complete loss of activity from these soluble and well-expressed mutant proteins was caused by unintentional structural anomalies. Also, the orientation of the catalytic triad residues of TbPLA₁ is consistent with that of the other serine hydrolases where the residues appear in the order Ser-Asp-His. Accordingly, TbPLA₁ is the newest member of the serine hydrolase family.

The central discovery from this report is that TbPLA₁ mediates the synthesis of long-chain polyunsaturated and highly unsaturated lysoGPCho metabolites in T. brucei. Detecting TbPLA₁-dependent fluctuations of these metabolites was reliant upon creating TbPLA₁ double knockout and overexpression transgenic cells. This is the only PLA₁ double knockout from any organism, a poignant fact which explains why not much is understood about this class of enzymes. As such, there is no precedent study which directly detects and manipulates in vivo PLA₁ activity in its native cell as is described here. Biochemical analysis of the TbPLA₁ null mutants and various transgenic TbPLA₁ overexpression cell lines showed that depleted or increased levels of TbPLA₁ provoked lysoG-PCho metabolite ablation and elevation respectively, whereas alterations in other PL pools were not detected. Interestingly, the levels of the lysoGPCho(−/20:y) intermediates did not rise as much as the lysoGPCho(−/18:y) and lysoGPCho(−/22:y) intermediates during TbPLA₁ overexpression. Importantly, however, it should be acknowledged that although certain lysoGPCho species are more abundant than others in the mass spectra, this does not necessarily mean that they are synthesized at greater rates by TbPLA₁ because both LysoPLA and acyltransferases could show preference to further metabolize a particular series of lysoPL intermediate (Arthur, 1989; Pete and Exton, 1996). A consequence of this type of scenario would be that very little if any global increase in metabolites would be detected despite their increased synthesis.

Unsurprisingly, TbPLA₁ null mutants and overexpression cell lines maintained normal cell growth and revealed no apparent phenotype in culture. In Leishmania major, another trypanosomatid, the inability to de novo synthesize sphingolipids and ether PLs, principle components of membranes, were also viable and did not result in growth defects (Zufferey et al., 2003; Denny et al., 2004). Furthermore, the abundance of lysoGPCho in eukaryotic cells is very well regulated and normally very low (0.5–1.0% of total lipids), and their loss does not seem to affect viability in other cells (van den Bosch, 1980). The high specific activity fostered by TbPLA₁, coupled with its cytoplasmic localization, poses a threat to the cell if high levels of the enzyme were allowed to freely circulate and activate because it could potentially lyse the cell from within by producing excess FAs and lysoGPCho. Besides some protection provided by the plasmanyl/plasmenyl lipids present in T. brucei, tight regulation of lysoGPCho levels should be crucial for the normal functioning of T. brucei. It is shown here that tetracycline induction of both the overexpression and rescue cell lines resulted in very abundant overexpression of TbPLA₁ mRNA, but not substantial PLA₁ overexpression. This result suggests that lysoGPCho levels in T. brucei are at least in part regulated at the level of TbPLA₁ mRNA translation and protein synthesis. Regulation of lysoGPCho levels is also most likely mediated downstream of their formation by regulating the enzymes that can further metabolize lysoPLs. As part of a Lands cycle, a T. brucei acyltransferase could remodel lysoGPCho intermediates to form a new diacyl GPCho, or alternatively a LysoPLA could remove the remaining FA. A search of the T. brucei database does highlight several putative acyltransferase and LysoPLA sequences, whose activities may account for the T. brucei LysoPLA and acyltransferase activities observed by us and others (Tizard et al., 1977b; Samad et al., 1988; Bowes et al., 1993; Werbovetz and Englund, 1996).

One of the lysoGPCho metabolites detected in this study, lysoGPCho(−/20:4), has previously been shown to be an intermediate in a proposed sequential PLA₁/LysoPLA mechanism for the liberation of AA in T. brucei, an organism which is thought to lack PLA₂ activity (Ridgley and Ruben, 2001). Cytosolic PLA₂ enzymes, which have a marked specificity for sn-2 AA liberation from certain PLs, are quite identifiable at the sequence level by their C2 and α/β hydrolase domains (Six and Dennis, 2000), yet a search of the T. brucei genome database does not reveal any cytosolic PLA₂ homologues. It is worth mentioning that a soluble enzyme activity from T. brucei that may release sn-2 FAs has been purported to be due to a PLA₂ (Shuaibu et al., 2001), but it is curious that the properties of this activity are very similar to the properties presented here for recombinant TbPLA₁. Other authors have initially reported PLA₂ activity in T. brucei before later proposing that the activity was due to a concerted PLA₁/LysoPLA mechanism (Eintracht et al., 1998; Ridgley and Ruben, 2001). Thus, with long-chain PLA₂ activity possibly absent in T. brucei, the parasite potentially could have evolved a PLA₁/LysoPLA sn-2 deacylation system which could utilize its adapted TbPLA₁.

If TbPLA₁ is part of an alternative mechanism to release sn-2 FAs, its loss potentially could have been compensated for by other lipolytic activities. However, this study showed that redundancy at the level of lysoGPCho synthesis is not supported by T. brucei; both the lysate PLA₁ activity and ESI-MS-MS experiments indicated that lysoGPCho synthesis is severely and most likely totally compromised in TbPLA₁ null mutants. Redundancy in lysoGPCho formation could, however, exist at undetectable levels, levels that could be sufficient to supply the necessary levels of AA to the cell, if this is a critical activity. It is possible that a lysosomal PLA₁ activity is present in T. brucei (Opperdoes and van Roy, 1982), but this would most likely serve to degrade endocytosed PLs and not serve to recover lost cytosolic PLA₁ activity.

Though the results reported here bring us a step closer, it is still not proven that TbPLA₁ functions in a pathway to release AA from GPCho, nor if this AA release is essential for the cell. However, if redundancy is a reason why TbPLA₁ null mutants are still viable and show no growth phenotype, it most likely does not exist at the level of endogenous lysoGPCho(−/20:4) synthesis, but at the level of AA production and/or liberation from other lipids. In higher eukaryotes, both endogenous and exogenous linoleic acid (C18:2Δ^9,12) can be desaturated, elongated and desaturated again to produce AA (C20:4Δ^5,8,11,14). Both elongation and desaturation of FAs in trypanosomes has been observed (Dixon et al., 1971). But, it has recently been shown that T. brucei uses three of its elongases (elongases 1–3) in BSF and PCF parasites to elongate pre-existing FAs to myristate (C14:0) and stearate (C18:0) respectively (Lee et al., 2006); therefore, elongase 4, shown to elongate polyunsaturated FAs, may be the only elongase which may play a role in AA synthesis. Of course, AA can also be taken up directly by the cell from LDL-linked or albumin-linked lysoGPCho and non-esterified FAs in the blood, whose AA derivatives amount to 6.98% and 0.97% of the AA composition in human plasma respectively (Croset et al., 2000). T. brucei LysoPLA activity has been shown to efficiently degrade exogenous lysoGPCho (Samad et al., 1988; Werbovetz and Englund, 1996). Last, AA could conceivably be freed from neutral lipids by lipases.

It is evident that PL degradation by intracellular phospholipases and lysophospholipases is an exceedingly complex matter. This study has contributed significantly to the understanding of how TbPLA₁ functions at a biochemical and molecular level in T. brucei. Equally important is the fact that TbPLA₁ has emerged as a unique example of a PLA₁ enzyme.

Experimental procedures

Cloning of TbPLA₁

blast analysis identified a putative PLA₁ in the T. brucei genome. NdeI-containing forward (5′-atagtacggcatATGTCGGAATCCGTTGATCAGT-3′) and BamHI-containing reverse (5′-gacattaggatcctcTGAAGAACCGAACCCTCCG-3′) primers were used to PCR amplify TbPLA₁ from T. brucei genomic DNA using Pfu polymerase (small letters in primer sequences denote here and elsewhere an engineered restriction enzyme site and/or linker sequence). The PCR product was gel purified (Qiagen), ligated into a blunt-ended TOPO cloning vector (Invitrogen) and sequenced on both strands. Some purified PCR product was used as a probe in Southern blotting after fluorescein-labelling by random priming (Amersham Biosciences).

Expression and purification of TbPLA₁

TbPLA₁ excised from TOPO using NdeI and BamHI was subcloned into the pET15b expression vector (Novagen), resequenced and transformed into E. coli BL21 Gold cells (Novagen). The cells were grown at 37°C to an OD₆₀₀ of 0.8 in 3.0 l of Luria–Bertani broth containing 50 μg ml⁻¹ ampicillin, transferred to a 15°C shaking water bath, and induced overnight with 1.0 mM IPTG. Cells were harvested by centrifugation and resuspended in 40 ml of binding buffer (50 mM NaH₂PO₄ pH 8.4, 150 mM NaCl, 50 mM imidazole, 5% glycerol), lysed by a One-shot Cell Disrupter (Constant Systems), and the lysate was clarified by centrifugation at 10 000 g, 4°C for 25 min. The resultant supernatant was filtered (0.45 μm) and loaded onto a charged Hi-Trap nickel column (Amersham Biosciences) equilibrated with binding buffer. The column was washed extensively with binding buffer and recombinant protein was eluted in one step with 1.0 ml elution buffer (50 mM NaH₂PO₄ pH 8.4, 150 mM NaCl, 500 mM imidazole, 5% glycerol). Half of the eluate was immediately diluted 30-fold with anion exchange binding buffer (20 mM Tris-HCl pH 8.0, 5% glycerol) and loaded onto a Resource™ Q anion exchange column (Amersham Biosciences) equilibrated with binding buffer. The remaining half was digested with thrombin for 4 h at room temperature to cleave the His·tag from TbPLA₁ and then diluted and loaded onto the column. Gradient elution was achieved by using an AKTA Purifier pump (Amersham Biosciences) at 0.5 ml min⁻¹ using anion exchange elution buffer (20 mM Tris-HCl pH 8.0, 0.5 M NaCl, 5% glycerol), and aliquots from the major peak were snap frozen. This 1 day procedure yields greater than 15.0 mg l⁻¹ of highly active purified TbPLA₁. If needed, buffer exchange by dialysis was performed before certain experiments. Protein concentration was determined by the Bio-Rad protein assay using bovine serum albumin as a standard. His·tagcleaved TbPLA₁ was used in all experiments except where noted.

Radiolabelled PLA₁ assay

PLA₁ activity was determined by measuring the formation of lysoGPCho(−/[³H]18:0) from GPCho(16:0/[³H]18:0) (American Radiolabelled Chemicals, St Louis, MO) that was presented to the enzyme as mixed micelles. Mixed micelles were prepared as an 8× stock solution containing 1.2 mM GPCho(16:0/18:0) (specific activity 26.6 μCi μmol⁻¹) and 65.9 mM TX-100 (0.018 mole fraction) in 50 mM Tris-HCl, pH 7.0. The final assay mixture contained 50 mM Tris-HCl, pH 7.0, 75 mM NaCl, 150 μm GPCho substrate, and 0.63–60.0 ng of TbPLA₁ recombinant protein in a total volume of 40 μl. Enzyme was omitted from control samples. Reactions were initiated by the addition of substrate and incubated at 22°C for 5 min unless stated otherwise. TbPLA₁ reaction rates were linear up to 38 min. Lipids were extracted by the method of Bligh and Dyer (1959) and the lower phase was dried under nitrogen and redissolved in chloroform. Samples were applied to HPTLC Silica 60 plates (Merck) and developed in chloroform/methanol/water (10:10:3, v/v). The HPTLC plates were exposed to a BAS-TR 2040 imaging plate and the radiolabelled lipids detected with a FLA-5100 scanner (all from Fujifilm) and quantified using Advanced Image Data Analyzer (AIDA) densitometry software (Raytest).

Substrate specificity by GC-MS

An 80 μl reaction mixture containing 50 mM Tris-HCl pH 7.0, 100 mM NaCl and 3.75 mM substrate at a mole fraction of 0.036 with TX-100 was incubated with 100 ng enzyme at 22°C for 10 min. Reaction rates were linear with respect to time up to 35 min. Enzyme was omitted from control samples. Heptadecanoic acid (C17:0) internal standard (50 nmol) was added to each reaction tube before extraction of free FAs by the method of Dole (Dole and Meinertz, 1960). Fatty acid methyl esters (FAMEs) were prepared using diazomethane in diethyl ether and let to evaporate overnight before redissolving in 30 μl dichloromethane. Sample (1.0 μl) was injected onto a 30 m HP-5 column (J & W Scientific) and analysed using an Agilent 6890N GC-MS. Individual FAMEs were identified according to retention times and mass spectra of standards, and quantified by comparison of peak areas with that of the internal standard.

Site-directed mutagenesis

Five pairs of complimentary primers were synthesized to mutate selected amino acids from TbPLA₁ to alanine and they are the following: D63A, 5′-GAGGGACTTAAAAAAAGGGCCGAGAGTTGCGG-3′ and 5′-CCGCAACTCTCGGCCCTTTTTTTAAGTCCCTC-3′; D183A, 5′-GCCGAAGCGCTTGCTGTGTTCCAACGACATCCG-3′ and 5′-CGGATGTCGTTGGAACACAGCAAGCGCTTCGGC-3′; D232A 5′-CAGGCCACATTATATTGCTGCTCACCTCATAGAGAC-3′ and 5′-GTCTCTATGAGGTGAGCAGCAATATAATGTGGCCTG-3′; H234A, 5′-CATTATATTGATGCTGCCCTCATAGAGACAATTATTG-3′ and 5′-CAATAATTGTCTCTATGAGGGCAGCATCAATATA ATG-3′; S131A, 5′-GTTTCATTCACGGGGCATGCGCTTGGCGGTGGTTTG-3′ and 5′-CAAACCACCGCCAAGCGcATGCCCCGTGAATGAAAC-3′. The primers and the vector template, pET15b/TbPLA₁, were used with the QuickChange Site-directed Mutagenesis kit (Stratagene). Plasmids harbouring the altered TbPLA₁ were sequenced to confirm the mutation.

Creation of TbPLA₁ mutants

The 5′- and 3′-UTR sequences immediately upstream and downstream from the TbPLA₁ gene, respectively, were obtained from the database. 5′-UTR (270 bp) and 3′-UTR (410 bp) fragments were amplified from genomic DNA using the 5′-UTR sense primer (5′-ataagaatgcggccgcTCACCATTTGGCACGTGAGGGA-3′) and antisense primer (5′-cgtttaaacttacggaccgtcaagcttGATTCCTTATAAAATGATTCCACTTG-3′), and the 3′-UTR sense primer (5′-tgacggtccgtaagtttaaacggatccACTCATAACGCAAAA-3′) and antisense primer (5′-gtaagcggccgcTCACAAACAAATATTACAGACAGGAAT-3′). These two products, with 5′-NotI sites, were linked (via the complimentary underlined primer sequence) in a linking PCR reaction, amplified by PCR, NotI digested and ligated into the NotI site of pGEM5Zf (Promega). The linker region was excised by HindIII/BamHI digestion and the puromycin (puro^R) and blasticidin (bsd^R) drug resistance genes with HindIII/BamHI sticky ends were ligated individually into the cloning site. The resultant plasmids were NotI digested to release their 5′-puro^R-3′ and 5′-bsd^R-3′ DNA inserts, precipitated with ethanol, and redissolved in sterile water prior to being used in the electroporation and selection of T. brucei transgenic populations, performed as described elsewhere (Martin and Smith, 2006a,b).

Ectopic copies of TbPLA₁ were engineered into pLEW100/PLA1-myc by inserting the ORF of TbPLA₁ into the HindIII/PacI cloning site of pLEW100 (Wirtz et al., 1999) by using the HindIII-containing ORF forward (5′-atagtacaagcttATGTCGGAATCCGTTGATCAGT-3′) and PacI-containing reverse primers (5′-gatagcttaattaaTGAAGAACCGAACCCTCCG-3′). Mid-log cells were electroporated with 10 μg of NotI-linearized pLEW100/PLA1-myc plasmids in a total volume of 450 μl of cytomix buffer. The transfected parasites were selected in medium containing phleomycin (2.5 μg ml⁻¹). When tetracycline was added to the media during biochemical analysis, a final concentration of 1 μg ml⁻¹ was used. Cells were counted each day and were passaged only when the density was between 2 × 10⁶ and 3 × 10⁶ (normally every second day).

EGFP fused to TbPLA₁ was engineered in both a C- and N-terminal orientation to produce the plasmids pLEW100/EGFP-TbPLA1-myc (N-terminal) and pLEW100/TbPLA1–EGFP (C-terminal). N-terminal EGFP was cloned into the HindIII cloning site of pLEW100/PLA1-myc using HindIII-containing forward (5′-tagtacaagcttATGGTGAGCAAGGGCGAGG-3′) and reverse (5′-tagtacaagcttCTTGTACAGCTCGTCCATGCC-3′) primers. C-terminal EGFP was cloned into the PacI/BamHI cloning site (displacing the myc tag) of pLEW100/PLA1-myc using PacI-containing forward (5′-gatagcttaattaacATGGTGAGCAAGGGCGAGG-3′) and BamHI-containing reverse (5′-atagtacggatccaTTACTTGTACAGCTCGTCCATGCC-3′) primers. The fusion proteins were tetracycline induced overnight, washed twice in PBS, and fixed in 4% paraformaldehyde. The cells were mounted on polysine slides, washed and stained with DAPI before visualizing EGFP with exposure to UV light on a Zeiss Axiovert 200 imaging microscope.

Southern and Northern blotting

For Southern blotting, T. brucei genomic DNA (2 μg/lane) was digested with EcoRI and the fragments separated on a 0.7% agarose gel, transferred to nitrocellulose and UV-cross-linked. For Northern blotting, total RNA was extracted with an RNA Mini Kit (Qiagen) and 12 μg was separated on 1% MOPS/Formaldehyde, 0.7% agarose gel before transfer and linking to nitrocellulose. The blots were blocked with ULTRAhyb™ (Ambion) for 1 h at 42°C before hybridization with fluorescein-labelled TbPLA₁ (southern) or [³²P]-CTP-labelled TbPLA₁ probe (northern) overnight at 42°C. The blots were washed twice with 2× SSC plus 0.1% SDS for 20 min each, and washed twice with 0.5× SSC plus 0.1% SDS for 20 min each at 42°C. Northern blots were directly exposed to film whereas bound probe on the Southern blots were detected using horseradish peroxidase-conjugated antifluorescein antibody with the Gene Images CDP-Star kit (Amersham Biosciences). The bands on the Northern blot-transferred images were analysed by densitometry using AIDA.

Electrospray tandem mass spectrometry

Total lipids from 10⁸ mid-log PCF trypanosomes were extracted by the method of Bligh and Dyer (1959) and the lower chloroform phase was washed once with fresh upper phase. The resultant lower phase lipid extract was dried under nitrogen and redissolved in 20 μl of chloroform/methanol (1:2). For samples containing internal standards, lysoGPCho(17:0/−) and lysoGPCho(24:0/−) (0.5 nmol/standard) were added prior to lipid extraction. An aliquot of the samples was analysed with a Micromass Quattro Ultima triple quadrupole mass spectrometer equipped with a nano-electrospray source. Samples were loaded into thin-wall nanoflow capillary tips (Waters) and analysed by ESI-MS-MS in both positive (GPCho, GPSer and GPEth) and negative ion mode (GPIns, GPSer and GPEth) under capillary/cone voltages of 0.7 kV/50 V and 0.9 kV/50 V respectively. Tandem mass spectra were obtained using argon as the collision gas (~3.0 mTorr) with collision offset energies as follows: 35 V, positive ion precursors of m/z 184 for GPCho; 45 V, negative ion precursors of m/z 241 for GPIns; 25 V, neutral loss scanning of m/z 141 in positive mode for GPEth; 22 V, neutral loss scanning of m/z 185 in positive mode for GPSer; 50 V, negative ion precursors of m/z 196 for GPEth; and 28 V, neutral loss scanning of m/z 87 for GPSer. Each spectrum encompasses at least 50 repetitive scans of 4 s duration.

Subcellular fractionation and Western blotting

Log phase bloodstream or PCF form parasites [strain 427, MITat 1.2, cultured as previously described in SDM-79 media at pH 7.4 which contained 15% heat-inactivated fetal calf serum as a lipid source (Martin and Smith, 2006b)] were osmotically lysed at 10⁶ cell equivalents ml⁻¹ using cold H₂O containing 2 μg ml⁻¹ leupeptin and 0.1 mM TLCK. Cell lysis was confirmed by microscopy. Lysate was centrifuged at 1000 g to pellet nuclei and unbroken cells. The resultant supernatant was then centrifuged at 14 500 g for 10 min to pellet the large granular fraction, which was resuspended with PBS to an equivalent volume of aspirated supernatant. The 14 500 supernatant was centrifuged again at 140 000 g for 1 h. The resultant pellet, the small granular fraction, was resuspended in PBS after aspirating the supernatant, the cytosolic protein fraction. The proteins from 10 μl of each fraction were resolved on a 15% SDS-PAGE gel and analysed by Western blot. The blot was developed using an ECL detection kit according to the manufacturers’ instructions (Amersham Biosciences). Purified recombinant TbPLA₁ was used to obtain rabbit antiserum (Eurogentec). Anti-rabbit TbPLA₁ antibodies were affinity purified using a recombinant TbPLA₁-coupled CNBr-activated Sepharose column and following the protocol outlined by the manufacture (Amersham Biosciences). If needed, the bands on the Western blottransferred images were analysed by densitometry using AIDA.

Bis-BODIPY® FL-C₁₁-PC assay

Bis-BODIPY® FL-C₁₁-PC (Molecular Probes) was used as a GPCho fluorescent substrate at a 1:100 molar ratio with GPCho(16:0/16:0), and TX-100/lipid micelles were made. In its diacyl state, the proximity of the BODIPY fluorophores on adjacent PL acyl chains causes an intramolecular self-quenching of fluorescence. Hydrolysis and release of an acyl chain by a phospholipase A alleviates quenching and an increase in fluorescence results. The fluorophores are situated at the ends of two 11 carbon chains thus mimicking a GPCho with an approximate length near to that of GPCho(16:0/16:0).

The assay was developed for use in a 96 well plate. The assay was performed at a mole fraction of 0.018 [mole fraction = (lipid)/(lipid + detergent)]. Micelles were made as an 8× working solution which was then diluted accordingly for use in an 80 μl reaction per well. Each well initially contained 40 μl of buffer (50 mM Tris-HCl, pH 7.0) containing the micelle substrate (0.075 mM final). The reaction was started by the addition of 40 μl reaction buffer (50 mM Tris-HCl, 100 mM NaCl, pH 7.0) containing 400 pg (11.4 pmol) of recombinant TbPLA₁. The reaction was left to proceed at 22°C for 10 min before quenching with 170 μl of 100% methanol. Liberated acyl chains containing a fluorophore were excited at 485 nm and detected through a 520 nm centred bandpass filter using the FLX 800 Microplate Reader (Bio-Tek Instruments). All experiments were performed at least three times in triplicate and data points varied by less than one standard deviation. Velocities obtained were initial velocities.