Cloning, expression, and characterization of the acyl-CoA-binding protein in African trypanosomes.

African trypanosomes are shielded from their hosts' defenses by a coat of variant surface glycoprotein molecules, each of which is attached to the plasma membrane by a glycosylphosphatidylinositol anchor. During the later stages of glycosylphosphatidylinositol biosynthesis, myristic acid is incorporated into the anchor from the donor myristoyl-CoA by a series of unique fatty acid remodeling and exchange reactions. We have cloned and expressed a recombinant trypanosome acyl-CoA-binding protein that has a preference for binding relatively short chain acyl-CoAs and that has a high affinity for binding myristoyl-CoA (K(d) = 3.5 x 10(-10) M). This protein enhances fatty acid remodeling of glycosylphosphatidylinositol precursors in the trypanosome cell-free system. We speculate that the trypanosome acyl-CoA-binding protein plays an active role in supplying myristoyl-CoA to the fatty acid remodeling machinery in the parasite.

The protozoan parasite Trypanosoma brucei is responsible for causing African sleeping sickness in humans and Nagana in cattle. The parasite is spread between hosts by the tsetse fly vector and, in the bloodstream, possesses a dense surface coat of 10 million copies of variant surface glycoprotein (VSG) 1 (1,2). The VSG coat protects the parasite from lysis by the alternative complement pathway. During an infection the parasite undergoes a process called antigenic variation whereby different VSGs are expressed on the plasma membrane (reviewed in Ref. 3). Regardless of which VSG variant is expressed on the surface, each VSG molecule is attached to the outer leaflet of the plasma membrane by a conserved glycosylphosphatidylinositol (GPI) anchor (4). The C-terminal amino acid of the VSG protein is attached, via an ethanolamine-phosphate bridge, to a conserved moiety of Man␣1-2Man␣1-6Man␣1-4GlcN-myo-inositol-1-HPO 4 -3(sn-1,2-dimyristoylglycerol) (Man 3 GlcN-PI). The T. brucei GPI biosynthetic pathway was elucidated using a cell-free system containing washed trypanosome membranes (5,6). The pathway involves assembly of ethanolamine-P-Man 3 GlcN-PI (glycolipid AЈ) from PI followed by fatty acid remodeling whereby both fatty acids of the PI moiety are re-modeled to myristate (C14:0) through the sequential deacylation and reacylation of the glycerol backbone (Fig. 1). The product, a dimyristoylated GPI known as glycolipid A (8), is linked to newly synthesized VSG in the endoplasmic reticulum by a transamidation reaction (10 -12). Experiments using the cell-free system also showed that the donor for the myristoylation reactions of fatty acid remodeling was myristoyl-CoA (8).
In most, possibly all, eukaryotes it is known that acyl-CoAs are bound to acyl-CoA-binding proteins (ACBPs) that prevent their metabolism and allow the efficient shuttling of acyl-CoA to the cell acylation machinery (13). In this study we have expressed, purified, and characterized a recombinant trypanosome ACBP and studied the effect of this molecule on the fatty acid remodeling machinery of GPI biosynthesis.

EXPERIMENTAL PROCEDURES
Organisms and Reagents-Genomic DNA from T. brucei MIT at1.4 was isolated as described previously (14). Routine DNA manipulations were performed in Escherichia coli strain DH5␣. All chemicals were of the highest grade available from BDH or Sigma, and restriction enzymes were from Promega. DNA sequencing of double-stranded DNA was accomplished by the dideoxynucleotide chain termination method (15) by automated cycle sequencing using the dye terminator method (ABI PRISM big dye terminator kit, Perkin-Elmer). Trypanosome EST (accession No. W40086) was a gift from P. Majiwa (International Livestock Research Institute, Kenya) and recombinant bovine ACBP was a gift from J. Knudsen (Institute of Biochemistry, Denmark).
Cloning of the T. brucei acyl-CoA-binding Protein-BLASTn search of the GenBank TM Data base was performed using the gene encoding human ACBP (accession No. P07108). A cDNA fragment corresponding to the identified EST sequence (accession No. W40086) was amplified by polymerase chain reaction (PCR). Primers for amplification were 5Јggcatatgaggaaacgttg-3Ј (sense) and 5Ј-ccggatcctacaggttcga-3Ј (antisense). The sense primer contains a NdeI site and the antisense primer contains a BamHI site for cloning. The cDNA plasmid encoding the EST sequence was used as template. Amplification conditions were 94°C for 45 s, 50°C for 1 min, and 72°C for 1 min for 30 cycles. The PCR product was separated on an agarose gel, and the DNA of the expected size was isolated (Qiaex II kit, Qiagen). The PCR product was labeled with fluorescein-dUTP by random priming (Gene Images kit, Amersham Pharmacia Biotech) and used as a probe in a Southern blot of restriction-digested genomic T. brucei DNA. This probe was also used to screen a 5.5-kilobase EcoRI restriction-digested genomic T. brucei size selected bacteriophage library. The filters were washed at high stringency (0.5ϫ SSC, 0.1% SDS, 60°C), and one positive colony from the size selected library was confirmed by PCR. This bacterial colony, containing a genomic clone of the T. brucei ACBP gene, was replated and screened twice more prior to sequencing.
Overexpression of T. brucei ACBP-The ACBP gene was amplified by PCR using the sequenced genomic clone as template. The sense strand primer 5Ј-ggccatggtggaggaaacgcttgatgaaaag-3Ј contained a NcoI site (underlined) and an ATG initiation codon, and the antisense strand primer 5Ј-ggctcgagctacaggttcgacgggggcac-3Ј contained a XhoI site (underlined) and a stop codon. The PCR product was cloned using the NcoI and XhoI into pUC18 vector (Sureclone kit, Amersham Pharmacia Biotech). The insert was subsequently excised by digestion with NcoI and XhoI, and the gel-purified insert was ligated into the NcoI and XhoI cloning site of the pFastBac HTa expression vector (Life Technologies, Inc.). The pFastBac-ACBP plasmid was used in the Bac-to-Bac TM Bacu-lovirus Expression System (Life Technologies, Inc.). Briefly, the plasmid was transformed into competent DH10Bac E. coli cells, and recombinant bacmid DNA recovered after transposition. Sf9 insect cells were transfected with the bacmid DNA in the presence of CellFECTIN reagent and recombinant baculovirus particles recovered as stock virus. Sf9 cells were grown in suspension culture in Sf-900 II medium (Life Technologies, Inc.) containing 50 units/ml penicillin, 50 g/ml streptomycin, 5% fetal calf serum at 27°C. Optimal expression of recombinant protein was achieved using 4 ϫ 10 6 pfu/ml virus to infect 2 liters of Sf9 cells at 2 ϫ 10 6 cells/ml for 3 days at 27°C. Sf9 cells were pelleted (500 ϫ g, 5 min, 4°C) and resuspended in 200 ml of lysis buffer (20 mM Tris-HCl, pH 8, 100 mM KCl, 1 mM phenylmethanesulfonyl fluoride), and cells were disrupted by sonication (6 ϫ 45 s pulses interrupted with cooling on ice). Cell debris was removed by centrifugation at 10,000 ϫ g for 10 min prior to applying the supernatant to a Ni 2ϩ -resin column (1.6 ϫ 12 cm chelating Sepharose fast flow, Amersham Pharmacia Biotech). The histidine-tagged recombinant protein was eluted from the column with a linear gradient of imidizole from 10 mM to 1 M in 20 mM Bis-Tris propane, 20 mM Tris-HCl, pH 7.5, 300 mM NaCl at 3 ml/min. Fractions were concentrated and desalted with a Centricon plus-20 (Amicon) filtration unit, and the histidine tag was removed with rTEV protease (Life Technologies, Inc.) at 33 units/mg recombinant protein for 16 h at 4°C. The recombinant protein was further purified on a gel permeation column (1 ϫ 30 cm Superdex 75 HR, Amersham Pharmacia Biotech) in 50 mM Na-Hepes, pH 7.5, 300 mM KCl at 0.5 ml/min. The sample was concentrated to 1 mg/ml using a Centricon plus-20, and aliquots were either used immediately or snap-frozen in liquid nitrogen prior to storage at Ϫ80°C.
Electrospray-Mass Spectrometry Analysis-Trypanosome ACBP (3 M) was incubated in the presence or absence of 10 M myristoyl-CoA in 50 l of water for 20 min at room temperature. An equal volume of 50% methanol containing 0.2% formic acid was added, and the samples were applied to a Micromass Quattro electrospray mass spectrometer to acquire positive-ion mass spectra over the range m/z 600 -1800. Scans were averaged and processed using MassLynx software, and the mass of proteins was determined using MaxEnt.
Acyl-CoA Competition Binding Assay-This competition assay has been previously described (17). Briefly, 0.1 pmol of unlabeled acyl-CoA was mixed with 0.1 pmol of [ 3 H]myristoyl-CoA in 75 l of binding buffer (10 mM potassium phosphate buffer, pH 7.4). Recombinant trypanosome or bovine ACBP (0.05 pmol) was added in 25 l of binding buffer. The samples were mixed and incubated at 37°C for 30 min, chilled on ice for 10 min, and mixed with 0.2 ml of an ice-cold 50% slurry of Lipidex 1000 (Canberra Packard) in binding buffer. After 100 min on ice, the samples were centrifuged at 12000 ϫ g for 5 min at 0°C. The radioactivity in 100 l of the resulting supernatant was determined by scintillation counting. The assay was carried out in triplicate, and controls without ACBP were performed.

RESULTS
Identification and Cloning of Trypanosome ACBP Gene-We identified a cDNA fragment (GenBank TM accession number W40086) encoding a putative trypanosome ACBP by BLASTn search using the human ACBP gene as a query. The putative ACBP open reading frame present in this EST was amplified by PCR, gel-purified, labeled by random priming, and used as a probe.
A Southern blot of T. brucei genomic DNA digested with several restriction enzymes and hybridized with the ACBP probe indicated a single copy ACBP gene/haploid genome (Fig.  2). The identification of two restriction fragments when the genomic DNA was digested with MboI or BglII (Fig. 2, lanes 6  and 7) was consistent with the EST DNA sequence that predicts single MboI and BglII restriction sites within the putative ACBP open reading frame.
Based on the Southern blot (Fig. 2, lane 3), gel-purified 5.5-kilobase EcoRI fragments of genomic DNA were ligated into a pUC vector to produce a size selected library. The library was screened for the ACBP gene (19), and a clone containing a 279-base pair gene encoding a putative 93-amino acid trypanosome ACBP (predicted molecular mass of 10.69 kDa) was isolated. Alignment of the predicted amino acid sequence of the putative trypanosome ACBP with the human, bovine, and yeast ACBP sequences (20 -22) using Clustal V is shown (Fig.  3). The predicted trypanosome sequence showed 14% identity and 41% similarity to these other ACBP proteins. The putative trypanosome ACBP was 7 amino acids longer than the other homologs with a 6-amino acid extension close to the N terminus.
Overexpression and Purification of Recombinant Trypanosome ACBP-The putative trypanosome ACBP gene was cloned into the pFastBac plasmid that was used to produce a baculovirus encoding the recombinant his-tagged protein. This virus was used to infect Sf9 cells, and after optimization of virus titer levels and exposure times, expression of the recombinant protein at 1.5 mg/l was achieved (Fig. 4, lane 2). The presence of a N-terminal histidine tag sequence on our recombinant protein allowed rapid and simple purification using a Ni 2ϩ -resin column. The recombinant protein was eluted from this column with 400 mM imidizole (Fig. 3, lane 3). The Nterminal histidine tag sequence was removed by the protease rTEV (Fig. 3, lane 4), and the protein was further purified on a gel permeation column (Fig. 3, lanes 5 and 6).
Electrospray Mass Spectrometry Analysis of Trypanosome ACBP-An aliquot of the purified recombinant trypanosome ACBP was applied to an electrospray mass spectrometer, and spectra were collected in positive ion mode over the range m/z 600 -1800. Using MaxEnt in the MassLynx program to deconvolute the data, only one major protein species was detected with a measured average mass of 10,692 Da (Fig. 5A), 2 which is close to the predicted average molecular mass of putative trypanosome ACBP (10,690 Da). The ACBP protein was preincubated with myristoyl-CoA and reapplied to the electrospray mass spectrometer and spectra collected as before. This time two major species (10,692 and 11,670 Da) were detected (Fig.  5B) in a ratio of 4:1. The mass difference between the two peaks is 978 Da, which is the exact mass of myristoyl-CoA. Thus, the species at 11,670 Da is consistent with a complex of trypanosome ACBP and myristoyl-CoA. The high cone voltage of the electrospray mass spectrometer and the denaturing acidic conditions (25% methanol, 0.1% formic acid) used would be expected to dissociate all but high affinity protein-ligand interactions. Thus, the detection of some ACBP⅐myristoyl-CoA complex in the spectrum (Fig. 5B) suggests that the recombinant putative trypanosome ACBP has a high affinity for myristoyl-CoA and establishes the protein as an ACBP.
Dissociation Constant of Trypanosome ACBP for Myristoyl-CoA-The affinity of recombinant trypanosome ACBP for myristoyl-CoA was measured by equilibrium dialysis, and the data were analyzed using a Scatchard plot (Fig. 6). Based on the slope (-1/K d ), the dissociation constant for myristoyl-CoA is 3.5 Ϯ 0.4 ϫ 10 Ϫ10 M. The stoichiometry of interaction deduced from the data in Fig. 6 was 0.77, suggesting that some of the recombinant ACBP was unable to bind [ 3 H]myristoyl-CoA. The most likely explanation for this is that a small proportion of the purified recombinant ACBP was already occupied by acyl-CoA, 2 The minor species visible in Fig. 5A at 11,724 Da most likely represents oleoyl-CoA bound to the purified recombinant ACBP.   (20), bovine (21), and yeast (22) origin using the Clustal V alignment program. * and ϩ indicate sequence identity and similarity, respectively. most likely oleoyl-CoA. 2 Acyl-CoA Chain Length Specificity-To (a) determine the specificity of trypanosome ACBP for chain length acyl-CoA and (b) compare this with the specificity of bovine ACBP, we stud-ied the competition for binding [ 3 H]myristoyl-CoA by various unlabeled acyl-CoAs. Bovine ACBP showed broad specificity with the greatest relative affinity for stearoyl-CoA (Fig. 7). This is similar to the previously reported finding using a similar assay and an epr spectroscopy assay that measured the displacement of spin-labeled acyl-CoA by various acyl-CoAs (17). However, the trypanosome ACBP showed a very different acyl-CoA specificity with the greatest relative affinity for shorter chain acyl-CoAs such as lauroyl-CoA (Fig. 7).
The Effects of Trypanosome ACBP on GPI Fatty Acid Remodeling in the Cell-free System-Trypanosome membranes were incubated with UDP-GlcNAc and GDP-Man to generate in situ GPI substrates for fatty acid remodeling and subsequently with [ 3 H]myristoyl-CoA in the presence or absence of equimolar trypanosome ACBP. In the absence of trypanosome ACBP, incorporation of [ 3 H]myristate into lipids, including into mature GPI precursors glycolipid A and glycolipid C, was observed (Fig. 8A, lanes 1 and 2) as described previously (8). Incorporation was significantly greater at 100 nM than at 20 nM [ 3 H]myristoyl-CoA. The incorporation of [ 3 H]myristate into neutral lipids and phospholipids was unaffected by the presence of ACBP, but the incorporation of [ 3 H]myristate into glycolipids A and C was increased about 3-fold, as determined by densitometry (Fig. 8A, lanes 3 and 4). These data gave the first indication that trypanosome ACBP can facilitate the fatty acid remodeling and/or fatty acid exchange reactions. To study this further and to look specifically at the fatty acid remodeling reactions, the cell free system was pulse-labeled with GDP-[ 3 H]Man in the presence of UDP-GlcNAc to produce prelabeled glycolipid AЈ and (Fig. 8B, lane 1). In our hands, we find that fatty acid remodeling of these species to glycolipid AЉ and glycolipid A can be achieved over 15 min (though somewhat inefficiently) by the addition of ATP and CoA alone (Fig. 8B, lanes 2-4). Presumably, endogenous acyl-CoA activity can generate myristoyl-CoA in situ from some membrane-bound store of myristate. This conversion of glycolipids AЈ and to AЉ and A was greatly stimulated by the inclusion of recombinant trypanosome ACBP (Fig. 8B, lanes 5-7). Furthermore, the glycolipid A that was produced was further processed to glycolipid C, the inositolacylated version of glycolipid A, and the final product of the GPI biosynthetic pathway. These results show that trypanosome ACBP greatly facilitates the fatty acid remodeling and inositol-acylation of GPI intermediates in the trypanosome cellfree system and suggests that trypanosome ACBP is likely to be involved in supplying myristoyl-CoA to the fatty acid remodeling machinery in vivo. To access whether this affect was specific to trypanosome ACBP, the affect of recombinant trypanosome ACBP on fatty acid remodeling was directly compared with that of recombinant bovine ACBP (Fig. 8B, lanes 8  and 9). The bovine ACBP appears to be just as efficient as the trypanosome protein in stimulating fatty acid remodeling and inositol acylation. DISCUSSION We were interested in studying the fatty acid remodeling (8) and fatty acid exchange reactions (18,23) responsible for the exclusivity of myristic acid in the formation of the GPI anchor of trypanosome VSG (4). Studies using myristic acid analogs, such as O-11, suggest that their incorporation into the GPI anchor of VSG ultimately leads to the death of the parasites in culture (24,25). Previous studies using trypanosome cell-free system experiments have shown that the donor of myristate for fatty acid remodeling (8) and exchange (18,23) is myristoyl-CoA. However, other studies also suggested that the parasite was able to utilize myristoyl-lyso-phosphatidylcholine as a source of myristate (26 -28). This has been further investigated, and it appears that myristoyl-lyso-phosphatidylcholine is converted to myristoyl-CoA, which can be subsequently used in the fatty acid remodeling reactions (29). Thus, the consensus view is that myristoyl-CoA is the donor for myristic acid incorporation into trypanosome GPI anchors.
Based on the work of Knudsen and others (13, 30 -32), it is clear that the levels of free acyl-CoAs in the cytoplasm of most eukaryotic cells are very low because of the presence of ACBPs with high affinity for long chain acyl-CoAs. We wished to find the trypanosome ACBP homolog and to assess the effect of including this protein in the in vitro fatty acid remodeling assays. These assays use washed membrane preparations (the so-called cell-free system) that are likely to be devoid of endogenous (soluble) ACBP.
We searched GenBank TM for potential trypanosome ACBP homologs with the gene encoding human ACBP and were able to detect a trypanosome EST sequence (accession No. W40086). We subsequently cloned the full-length genomic sequence, which showed significant similarity to other ACBP homologs (Fig. 2). However, sequence conservation was lower for the putative trypanosome ACBP than that previously observed between the other ACBP homologs (13). Previous studies identified 13 conserved amino acids found in all ACBP homologs and suggested that all of these residues might be necessary for binding acyl-CoAs (13). In bovine ACBP these amino acids are Phe-5, Leu-15, Tyr-28, Lys-32, Gln-33, Gly-37, Pro-44, Lys-54, Trp-58, Gly-63, Ala-69, Tyr-73, and Leu-80. However, only six of those residues are absolutely conserved in trypanosome ACBP (Tyr-35, Gln-40, Gly-44, Pro-51, Lys-61, Ala-76) together with four conservative substitutions (Leu-6, Arg-65, Phe-80, Val-87) and three nonconservative substitutions (Thr-22, His-39, Lys-70). We considered that these differences from the other eukaryotic ACBPs studied to date might result in an altered specificity for binding acyl-CoAs. Having demonstrated that recombinant putative trypanosome ACBP was indeed a high affinity ACBP (Figs. 5 and 6), we tested the specificity of the trypanosome ACBP for acyl-CoA chain length using a competition assay (Fig. 7). These data showed that trypanosome ACBP had a distinct preference for shorter chain acyl-CoAs when compared directly to the bovine ACBP (Fig. 7).
We utilized the previously resolved three-dimensional NMR structure of the complex between bovine ACBP and palmitoyl-CoA (33) to predict the structure of trypanosome ACBP using the Collaborative Computational Project (34). This model predicts that the amino acids in the trypanosome ACBP at positions Thr-16, Lys-20, Ile-28, Lys-31, Leu-32, Tyr-35, Trp-38, Val-57, and Lys-61 are within 0.3 nm of the bound palmitoyl-  5-7, respectively). In a separate experiment, trypanosome membranes were pulse-labeled with GDP-[ 3 H]Man and chased with ATP/CoA for 15 min in the presence of trypanosome or bovine ACBP (lanes 8 and 9). CoA. Of these, Leu-32, Tyr-35, and Lys-61 are conserved; Lys-20 and Trp-38 are similar in bovine ACBP, whereas Thr-16, Ile-28, Lys-31, and Val-57 are not conserved between the two structures. These differences may have a role in defining the chain length specificity of the two ACBPs. We intend to investigate these issues using site-directed mutagenesis.
Previous studies have investigated the binding affinity of acyl-CoAs for ACBP using epr spectroscopy (17) and titration microcalorimetry (22,(35)(36)(37) ranging from 4 ϫ 10 Ϫ10 to 4.5 ϫ 10 Ϫ14 M depending on acyl chain length and the method used. Measurements made by direct titration microcalorimetry range from 2.4 ϫ 10 Ϫ7 and 1.7 ϫ 10 Ϫ8 M for octanoyl-CoA and dodecanoyl-CoA with bovine ACBP (35, 36) to 8.5 ϫ 10 Ϫ11 M for palmitoyl-CoA with yeast ACBP (22). In our studies, we used equilibrium dialysis to directly measure the K d value of trypanosome ACBP for myristoyl-CoA and obtained a figure of 3.5 Ϯ 0.4 ϫ 10 Ϫ10 M (Fig. 6). This K d value is consistent with the generally high affinity of acyl-CoAs for ACBPs. We were unable to obtain an accurate K d value for bovine ACBP and myristoyl-CoA, which was out of the range of our assay and must have a K d value of Ͻ 1 ϫ 10 Ϫ10 M. Previous titration studies showed that bovine ACBP has a preference for longer chain acyl-CoAs, with an optimal chain length of C18 -20 (17,35,36). The competition experiment (Fig. 7) is consistent with these data and show that whereas bovine ACBP has a relatively low affinity for shorter acyl-CoAs (ϽC12), the trypanosome homolog has a preference for shorter acyl-CoAs (ϽC20) and a remarkably high affinity for C10 and C12 acyl-CoAs.
Finally, we investigated the effect of trypanosome ACBP on the trypanosome cell-free system that has previously been used to study GPI biosynthesis. We anticipated that we might observe an effect because the washed membranes of the cell free system should be depleted of endogenous ACBP, whereas in vivo the majority of cellular myristoyl-CoA would be bound to ACBP. Initially, we investigated the direct effect on fatty acid remodeling and/or fatty acid exchange by adding exogenous [ 3 H]myristoyl-CoA Ϯ ACBP to the cell-free system. This showed that the presence of trypanosome ACBP substantially enhanced the labeling of glycolipids A and C with [ 3 H]myristate (Fig. 8A). We then further investigated the effect of ACBP on fatty acid remodeling by prelabeling glycolipids AЈ and in the cell-free system and subsequently generating myristoyl-CoA in situ by the addition of ATP and CoA. When ACBP was included, the efficiency of fatty acid remodeling and inositol-acylation was significantly increased. This strongly suggests that myristoyl-CoA is usually supplied to the fatty acid remodeling machinery via ACBP. This is consistent with the roles of ACBP in other eukaryotes where the ACBP is responsible for the transport of acyl-CoA directly to cell acylation machinery (22, 30 -32). Indeed, despite the limited similarity between the trypanosome and bovine ACBP, the latter seems to be able to supply myristoyl-CoA to the fatty acid remodeling machinery in the trypanosome cell-free system (Fig. 8B).