Secretion of a Novel, Developmentally Regulated Fatty Acid-binding Protein into the Perivitelline Fluid of the Parasitic Nematode, Ascaris suum *

Early development of the parasitic nematode, Ascaris suum , occurs inside a highly resistant eggshell, and the developing larva is bathed in perivitelline fluid. Two- dimensional gel analysis of perivitelline fluid from infective larvae reveals seven major proteins; a cDNA en- coding one of these, As-p18, has been cloned, sequenced, and protein expressed in Escherichia coli . The predicted amino acid sequence of As-p18 exhibits similarities to the intracellular lipid-binding protein (iLBP) family in- cluding retinoid- and fatty acid-binding proteins (FABP). As-p18 is unusual in that it possesses a hydro- phobic leader that is not present in the mature protein, the developmental regulation of its expression, and in terms of its predicted structure. Recombinant As-p18 both The present study was initiated to identify major proteins in the perivitelline fluid of A. suum infective larvae, specifically those potentially involved with the maintenance of the highly resistant lipid layer. Toward this end, we have identified a novel fatty acid-binding protein (FABP), designated As-p18, in the perivitelline fluid surrounding the infective L2. Based on sequence and structure analysis, and its fatty acid binding methods to analyze results of competition experiments using the program O (26). Three insertions were added using the lego-loop option, selecting by inspection the loop with the smallest root-mean-square deviation of base residue positions that had no significant Van der Waals overlap of atoms. Following energy minimization of the entire molecule, these loops were subjected to 2000 iterations (1 ps) of two-step Verlet molecular dynam-ics at 300 K; 2000 iterations at 200 K and 2000 at 100 K. This was followed by restrained energy minimization to convergence of the entire molecule and finally unrestrained energy minimization to convergence. Molecular simulation calculations were performed using the X-PLOR program and the CHARMM force field (27).

Parasitic nematodes cause medical and economic damage on a global scale, in part the result of the longevity and environmental robustness of their dispersal stages, their eggs. Arguably the most successful and widely distributed nematode parasite of humans is Ascaris lumbricoides, which infects over 1 billion people. A morphologically indistinguishable species, Ascaris suum, occurs in pigs and has become the best understood nematode in biochemical terms. The eggs contaminate soil, and the embryos develop to infective, second-stage larvae (L2s) 1 within the eggs. The L2 undergoes developmental arrest until ingestion by the host, and may survive for up to 7 years, but little is known of the biochemical basis for such long term survival. The problems faced by the L2 include maintaining both the egg's impermeability to water and dissolved ions (they can survive in dilute formalin and sulfuric acid) and its permeability to oxygen, as well as the potential accumulation of toxic products of lipid peroxidation. The ascarid L2 develops in the perivitelline fluid enclosed within the highly resistant, chitinous eggshell (1). The innermost lipid layer of the eggshell accounts for the extreme impermeability and resistance and is composed of three unusual, structurally related glycosides, the ascarosides (1)(2)(3). Perturbation of this ascaroside layer appears to be involved in the initiation of the hatching process upon infection of a new host (4 -6). The organism, therefore, has particular need of a mechanism to transport and store hydrophobic compounds between the larva and the eggshell.
The present study was initiated to identify major proteins in the perivitelline fluid of A. suum infective larvae, specifically those potentially involved with the maintenance of the highly resistant lipid layer. Toward this end, we have identified a novel fatty acid-binding protein (FABP), designated As-p18, in the perivitelline fluid surrounding the infective L2. Based on sequence and structure analysis, and its fatty acid binding function, As-p18 is an unusual member of lipid-binding protein (LBP) family in having a hydrophobic leader sequence, being under strong developmental regulation, and in terms of its predicted structure.
Preparation of Samples for Electrophoresis-L2s were artificially hatched, as modified from Urban et al. (7). Briefly, L2s were treated with 5.25% sodium hypochlorite for 10 min at room temperature. Eggshells were disrupted with glass beads by gentle stirring at 37°C in Earle's balanced salt solution with protease inhibitor mixture A (20 g/ml pepstatin A, 20 g/ml chymostatin, 20 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 2 mM EDTA, 2 mM EGTA). The suspension was centrifuged at 200 ϫ g for 10 min, and the resulting supernatant was centrifuged at 155,000 ϫ g for 1 h at 4°C. The hatching fluid was concentrated and exchanged by ultrafiltration (Centricon 10, Amicon) into 20 mM MOPS, pH 7.2, containing protease inhibitor mixture A.
Larvae were disrupted by treatment in a French pressure cell at 20,000 p.s.i. in 20 mM MOPS, pH 7.2, containing protease inhibitor mixture B (mixture A plus 20 g/ml aprotinin, 20 g/ml soybean trypsin inhibitor, 40 g/ml antipain, and 20 g/ml caproic acid). All adult tissues were homogenized in 20 mM MOPS, pH 7.2, with protease inhibitor mixture B. Larval and adult homogenates were brought to 1% with Triton X-100 on ice for 30 min and centrifuged at 10,000 ϫ g for 30 min at 4°C before concentration of the supernatant by ultrafiltration, as outlined above. Protein concentrations were estimated by the Bradford method (Bio-Rad). All samples were aliquoted, lyophilized, and stored at Ϫ80°C until use.
Amino Acid Sequencing-Proteins separated by two-dimensional gel electrophoresis were transferred to a Problott membrane (Applied Biosystems) in 10 mM CAPS buffer, pH 11, with 10% methanol according to the manufacturer's instructions. Proteins of interest were excised from the membrane and the NH 2 -terminal amino acid sequences were analyzed with a model 477A Protein Sequencer (Applied Biosystems).
cDNA Cloning and Sequencing-The full-length cDNA of As-p18 was cloned from cDNA prepared with a Marathon kit (Clontech). Briefly, cDNAs were synthesized from total RNA isolated from day 5 A. suum eggs, and the double-stranded cDNAs then were adapted with adapter primer 1 (AP-1) and primer 2 (AP-2). With these adapted doublestranded cDNAs as a template, a degenerate sense primer (primer A, 5Ј-GA(C/T) AA(A/G) TT(C/T) (C/T)T(N) GG(N) AC(N) TT(C/T) AA-3Ј) designed from NH 2 -terminal amino acid sequence of As-p18 (DKFL-GTFK) and the AP-1 primer were used for 3Ј-RACE. The 3Ј-RACE PCR products were analyzed on a 1% agarose gel and a weak band at about 600 base pairs was purified from the gel by GENO-BIND (Clontech). The purified DNA was used as template for nested PCR with primer A and AP-2 primer. The nested PCR products were analyzed on a 1% agarose gel. The DNA at about 600 base pairs was purified from the gel by GENO-BIND (Clontech), cloned into pNoTA/T7 vector (5Ј-3Ј, Inc.), and both strands sequenced by the dideoxynucleotide chain termination method of Sanger et al. (13) using Sequenase 2.0 (U.S. Biochemical Corp.). According to the sequence of the nested PCR product, two complementary gene-specific primers (primer B, 5Ј-TCC TTG TGA TGC GTG TCT TTC-3Ј; primer C, 5Ј-GAA AGA CAC GCA TCA CAA GGA-3Ј) were designed. Primer B and primer C were used with AP-1 for 5Ј-RACE and 3Ј-RACE, respectively, with the adapted day 5 cDNA. These 5Ј-RACE and 3Ј-RACE products were fused to prepare the full-length cDNA, which was both strands sequenced as described above.
Northern Blotting-Total RNAs from different developmental stages and adult tissues were isolated by acid guanidinium thiocyanate-phenol-chloroform extraction according to Chomczynski et al. (14). Total RNA (20 g) was separated on a 1% agarose formaldehyde denatured gel and transferred to a Nytran Plus positively charged nylon membrane (Schleicher & Schuell), and the membrane was baked in a vacuum oven at 80°C. The full-length As-p18 cDNA was used as a probe after labeling with [ 32 P]dCTP and purification on a NICK column (Pharmacia). The membrane was hybridized and autoradiographed at Ϫ80°C.
Construction of Expression Vector for As-p18 -A cDNA encoding the full-length mature As-p18 protein was expressed in the pQE-30 vector (QIAGEN), which is designed to produce a recombinant protein bearing a 6xHis tag at the NH 2 terminus. A BamHI site was added to the 5Ј end immediately before the nucleotide (nt 72) coding for the first amino acid of the mature protein, and a PstI site was added to the 3Ј-untranslated region at nt 603-608 by PCR with primer D (5Ј-G GTT GGA TCC AAG ACT CTA CCC GAC-3Ј) and primer E (5Ј-GC AAT CTG CAG TAA CAG ATG ATC AG-3Ј), respectively. The PCR fragment at about 550 base pairs was cloned into pCR II vector (Invitrogen). The 532-base pair BamHI/PstI fragment digested from pCR II/As-p18 plasmid DNA was cloned into pQE-30 vector via BamHI/PstI sites. This pQE-30/As-p18 plasmid was transformed into M15 (pREP4) strain of Escherichia coli (QIAGEN) and both strands sequenced, as described above, to verify the insert sequence.
Expression and Purification of Recombinant Protein-E. coli with the pQE-30/As-p18 were cultured in LB medium (10 g of tryptone, 5 g of yeast extract, and 5 g of NaCl/liter) until the absorbance at 600 nm reached 0.6 units. Expression of recombinant As-p18 (rAs-p18) was induced by 1 mM isopropyl-thio-␤-D-galactopyranoside for 4 h at 37°C. rAs-p18 was purified by affinity chromatography on Ni-NTA-agarose (QIAGEN) under nonreducing conditions. Briefly, bacteria were sonicated on ice in 50 mM NaH 2 PO 4 , 300 mM NaCl, pH 7.8 and the lysate was centrifuged at 12,000 ϫ g for 20 min at 4°C. Triton X-100 (0.5% final concentration) was added and the solution was gently stirred with Ni-NTA-agarose for 60 min on ice. The resin was then loaded into a column, washed with 50 mM NaH 2 PO 4 , 300 mM NaCl, 10% glycerol, pH 6.0, and eluted with a gradient of 0 -300 mM imidazole. The purified rAs-p18 was concentrated in 10 mM Tris-HCl, pH 7.5, by ultrafiltration (Centricon 10, Amicon) and stored at Ϫ80°C in small aliquots. This procedure provided approximately 16 mg of rAs-p18/liter of culture. Recombinant ABA-1 (rABA-1) allergen from A. lumbricoides was produced by similar methods, as will be described in detail elsewhere. 2 Antibody Generation and Immunoblotting-Purified rAs-p18 was separated on 12% SDS-polyacrylamide gels and used to immunize rabbits as described previously (15). Antibodies against rAs-p18 were affinity-purified using protein bound to CNBr-activated Sepharose 4B (Sigma) according to Kent (16). For immunoblotting, samples were separated on 12% SDS-polyacrylamide gels and transferred to nitrocellulose membranes. Membranes were blocked with 1% bovine serum albumin (BSA), incubated with primary antibody and alkaline phosphatase-conjugated goat anti-rabbit IgG secondary antibody, and developed with nitro blue tetrazolium and 5Ј-bromo-4-chloro-3-indolyl phosphate.
Spectrofluorimetry and Fluorescence-based Ligand Binding-Fluorescence measurements were made at 20°C with a Spex FluorMax spectrofluorimeter (Spex Industries), using 2 ml samples in a silica cuvette. Raman scattering by the solvent was corrected for where necessary using appropriate blank solutions. The fluorescent fatty acid analogs 11-((5-dimethylaminonaphthalene-1-sulfonyl) amino)undecanoic acid (DAUDA), and dansyl-DL-␣-aminocaprylic acid (DACA) were obtained from Molecular Probes and Sigma, respectively. All-transretinol and oleic acid were also obtained from Sigma. The dansylated fatty acids were stored as stock solutions (ϳ1 mg ml Ϫ1 in ethanol) in the dark at Ϫ20°C. They were freshly diluted to 1 M with phosphatebuffered saline (PBS; 171 mM NaCl, 3.35 mM KCl, 10 mM Na 2 PO 4 , 1.8 mM KH 2 PO 4 ) pH 7.2, before use in the fluorescence experiments. Competitors of fluorescent fatty acid binding were prepared as stock solutions (ϳ10 mM in ethanol) and diluted in PBS before use. Retinol binding to proteins was tested by adding retinol (6 l of a freshly prepared 180 M solution in ethanol) directly to a cuvette containing protein in PBS. Reference proteins BSA, ␤-lactoglobulin (bovine), ribonuclease A, ovalbumin (chicken), and transferrin (bovine) were obtained from Sigma and prepared as 10 mg ml Ϫ1 stock solutions in PBS. For estimation of the dissociation constant, successive 5-or 10-l samples of rAs-p18 (at a monomer concentration (assuming monomeric dispersion) of 26.9 M) were added to 2 ml of 0.7 M DAUDA, and the fluorescence measured at 476 nm ( exc ϭ 345 nm). The concentration of the ethanol stock solution of DAUDA was checked by absorbance of a 1:10 dilution in methanol at 335 nm, using an extinction coefficient ⑀ 335 of 4400 M Ϫ1 cm Ϫ1 . The concentration of rAs-p18 was estimated by absorbance at 280 nm, using an extinction coefficient of ⑀ 280 ϭ 27310 M Ϫ1 cm Ϫ1 , based on the amino acid composition of the recombinant protein (17). Fluorescence data were corrected for dilution, and fitted by standard nonlinear regression techniques (using Microcal ORIGIN software) to a single noncompetitive binding model to give estimates of the dissociation constant (K d ) and maximal fluorescence intensity (F max ). Similar nonlinear regression methods were used to analyze results of competition experiments in which oleic acid was progressively added to DAUDA/rAs-p18 mixtures. A stock solution of oleic acid in ethanol was freshly diluted to 10.6 M in PBS, and increasing concentrations of oleic acid were added to a mixture containing 0.71 M DAUDA and 0.94 M rAs-p18. For all of the fluorescence-based experiments, residual detergent was removed from rAs-p18 solutions by passage through an Extracti-Gel D column (Pierce).
Circular Dichroism-Circular dichroism (CD) spectra were recorded at 20°C in a JASCO J-600 spectropolarimeter using quartz cells of path length 0.02 or 0.05 cm. The protein concentration was 0.3 mg ml Ϫ1 , as estimated by absorbance at 280 nm, in PBS. Molar ellipticity values were calculated using a value of 118 Da for the mean residue weight calculated from the amino acid sequence of rAs-p18 protein. Analysis of the secondary structure of the protein was performed using the CON-TIN procedure (18) over the range from 240 to 195 nm.
Structural Predictions-Protein sequence comparisons were carried out with the BLAST (19) program to search the protein sequence data base (blastp). MaxHom (20) through SwissProt and Wisconsin Sequence Analysis Package (GCG) were used for protein sequence alignment. Secondary structure predictions were made using the PHD (21-23) program at SwissProt and GOR (24). Molecular mass and theoretical pI values for As-p18 were calculated with the ProtParam program through the SwissProt and GCG packages.
The As-p18 structure was modeled on the structure of mouse adipocyte lipid-binding protein (ALBP, Protein Data Bank code 1ADL) determined at 1.6-Å resolution by Lalonde et al. (25). A sequence alignment based on that produced by the PHD program on the Predict-Protein server (23) was used to direct the mutation of residues in the mouse protein to those of As-p18 using the program O (26). Three insertions were added using the lego-loop option, selecting by inspection the loop with the smallest root-mean-square deviation of base residue positions that had no significant Van der Waals overlap of atoms. Following energy minimization of the entire molecule, these loops were subjected to 2000 iterations (1 ps) of two-step Verlet molecular dynamics at 300 K; 2000 iterations at 200 K and 2000 at 100 K. This was followed by restrained energy minimization to convergence of the entire molecule and finally unrestrained energy minimization to convergence. Molecular simulation calculations were performed using the X-PLOR program and the CHARMM force field (27).

RESULTS
The As-p18 Protein and Cloning of Its Encoding cDNA-Perivitelline fluid was recovered from infective A. suum eggs after mechanical disruption and analyzed by twodimensional gel electrophoresis and NH 2 -terminal sequencing of selected proteins (Fig. 1). Seven abundant proteins with apparent molecular weights of about 50, 43a, 43b, 18, 16, 13, and 12.5 kDa were discernible by this procedure, in addition to a number of minor components. The NH 2 -terminal amino acid sequence of the protein, which migrated at about 18 kDa and a pI of 7.8, exhibited similarity to a group of small, water-soluble iLBPs including fatty acid-and retinoid-binding proteins iden-tified from a variety of different organisms (Figs. 2 and 3). This protein was designated As-p18 and selected for further study.
The cDNA encoding a fragment of As-p18 was isolated by PCR using degenerate oligonucleotides designed from the amino-terminal amino acid sequence generated from material eluted from two-dimensional gels. Full-length sequence for the As-p18 cDNA was then obtained by 3Ј-and 5Ј-RACE from day 5 egg cDNAs prepared as described under "Experimental Procedures." The sequence contained 628 nucleotides with a portion of the 22-nucleotide spliced leader sequence (SL1) characteristic of many nematode mRNAs (5Ј-CCA AGT TTG AG-3Ј, Fig. 2) (28). Downstream of the SL1 sequence, no 5Ј-untranslated region was apparent, as the initiation ATG codon immediately followed the spliced leader, and an open reading frame of 492 nucleotides ended with the termination codon TAA. The 3Ј-untranslated region of 125 nucleotides contained a sequence identical to the consensus polyadenylation signal, AATAAA, and a short poly(A) tail after the polyadenylation signal. The open reading frame predicts a protein with a molecular mass of 19,077 Da including a hydrophobic leader peptide of 20 amino acids which was not observed at the amino terminus of the mature, egg-derived protein (Fig. 3). This leader peptide exhibited features characteristic of secretory signals identified in other organisms (29). The predicted molecular mass of the mature As-p18 was 16,911 Da, and its theoretical pI was 7.1 by the ProtParam program and 7.8 by the GCG packages, respectively.
Comparison of the As-p18 sequence with protein sequences available in the data bases indicated that it had greatest similarity to iLBPs from variety of other sources ranging from flatworms to mammals and it was most similar to proteins predicted from the Caenorhabditis elegans genome project (see Fig. 3 for alignment), which are presumed to be LBPs based on their alignment with LBPs from non-nematode parasitic helminths and mammals. At the amino acid level, As-p18 exhibited 43% identity to C. elegans LBP-1 (GenBank™ accession no. U40420, protein identification code 1065518), 42% to C. elegans LBP-2 (GenBank™ accession no. U40420, protein code 1065519), 33% to mouse ALBP, 32% to rabbit myelin P2 protein, 30% to Fasciola hepatica FABP, 29% to mouse brain FABP, 28% to Schistosoma mansoni FABP and Echinococcus granulosus FABP, and 27% to bovine heart FABP (25, 30 -35). The nematode sequences were the only ones arising from the data base searches which had hydrophobic leader peptides. The nematode sequences had many more residues in common than with the other iLBPs, although residues conserved across the entire array were apparent. The nematode proteins also had two insertions of three and four amino acids in similar posi-tions, which are absent in the other iLBPs. The cleavage site for the leader peptide is between Ala-20 and Lys-21 of As-p18, as defined by the NH 2 -terminal amino acid analysis of protein isolated from A. suum eggs, and this is also the most probable cleavage site identified by the Signalp program, based on neural networks trained on eukaryotic sequences (36). The Signalp program also predicts cleavage sites in the C. elegans proteins precisely aligning with that for As-p18.
The Expression of As-p18 Is Developmentally Regulated-Early A. suum larval stages and adult tissues were examined for As-p18 by immunoblotting with affinity-purified antiserum FIG. 2. Nucleotide and predicted amino acid sequence of As-p18. cDNA encoding As-p18 was cloned and sequenced as described under "Experimental Procedures." The 17-amino acid NH 2terminal sequence determined by direct sequencing of the As-p18 spot obtained from a two-dimensional gel is underlined, as is a consensus polyadenylation signal in the 3Ј-untranslated region.
FIG. 3. Predicted amino acid sequence of As-p18 aligned with iLBP sequences from other sources. Dots indicate identity with As-p18. The boxed area is the NH 2 -terminal sequence determined by direct sequencing of As-p18. The amino acids of As-p18 are numbered.
against rAs-p18. Immunoblotting analysis revealed that As-p18 was absent in day 0 unembryonated eggs, began to be synthesized at about day 3 of development, was abundant at day 10 (when the L1 is formed), and remained abundant in eggs containing the infective L2. However, As-p18 could not be detected either in the parasitic L3 or adult A. suum muscle, ovaries, testis, or intestine (Fig. 4A).
Northern blotting using the radiolabeled full-length cDNA as a probe identified a single mRNA species of about 0.7 kb in total RNA extracted from Day 3 eggs, L1, and L2 (Fig. 4B). Surprisingly, significant bands of this size were also observed in Northern blots of RNA extracted from unembryonated eggs, ovaries, and especially L3s, even though the As-p18 protein could not be demonstrated by immunoblotting of these stages. To confirm the identity of the mRNA observed in ovaries and L3s on Northern blots, As-p18-specific primers were designed and used for PCR of cDNA pools prepared from mRNA isolated from both ovaries and L3s. The sequences obtained from both sources were identical to that of full-length As-p18 (data not shown).
To localize As-p18 during early larval development, infective L2 were hatched artificially, and both hatched larvae and hatching fluid were analyzed by SDS-PAGE and immunoblotting with affinity-purified antiserum against rAs-p18. As expected, As-p18 was present primarily in the hatching fluid, but a weak signal was also observed from the hatched L2s rinsed free from hatching fluid (data not shown). Immunolocalization and in situ hybridization experiments are currently in progress to identify the cells responsible for As-p18 synthesis.
Structural Analysis-The sequence of As-p18 aligns with iLBPs whose structures are known from x-ray crystallographic studies (37). These iLBPs are ␤-sheet-rich proteins, which have 10 antiparallel ␤-sheets forming so-called ␤-barrel or ␤-clam structures, with one or two short helices closing one end of the barrel. The ligand is held within a cavity in the center of the barrel, where it is removed from solvent and thereby protected from oxidation. The PHD program predicts As-p18 to be ␤-rich (19.0% helix, 35.0% extended/␤-structures, and 46.0% remainder) based on multiple alignments assembled from the As-p18 sequence and similar mammalian proteins by the MaxHom routine, and GOR predicts 14.2% helix, 59.1% extended/␤structure. The predictions of predominance of ␤-sheet structures are borne out by circular dichroism of rAs-p18. The far-UV circular dichroism spectrum of rAs-p18 (Fig. 5) showed a strong ␤-signal, and analysis of the data obtained over the range 240 to 195 nm by the CONTIN procedure provided the following estimates of secondary structure: 62% ␤-sheet and less than 5% ␣-helical content. This estimate is close to the known content of ␤-structure from crystal structures of iLBPs (37)(38)(39)(40).
Modeling of the As-p18 sequence to mouse ALBP (25), the iLBP most similar to As-p18, demonstrated that the protein can be successfully modeled as a ␤-barrel with two short helices, although modifications have to be made to accommodate the amino acid insertions (Fig. 6A). These have been made in loop regions with the first insertion adding an extra turn of helix to the second ␣-helical region. The second insertion is close to a loop region. The third loop insertion is between ␤-strands 7 and 8 and has the effect of increasing the enclosure of the lipid-binding site, although a separate, currently unsuspected, function could still be attributed to this region.
The final model of As-p18 has 96.1% of residues in the two most-favored regions of the Ramachandran plot and an acceptable overall geometry, both determined with Procheck (41). There is a single major cavity in the protein interior as determined by the program Voidoo (42). The volume available to a probe of radius 1.4 Å in this cavity is 335 Å 3 , also determined using Voidoo. It is an artifact of the modeling process used that a cavity within a protein will contract during the energy minimization stage. The alternative is to model the protein with a ligand filling the cavity, thus partially predefining the results of the experiment. This was considered less acceptable than a small artificial contraction of the ligand-binding cavity. Nevertheless, the cavity remains of a suitable size to bind fatty acid ligands; oleic acid is approximately 300 Å 3 while retinol occupies 330 Å 3 , as calculated from the volume which these ligands occupy within the known crystal structures of their binding proteins. The cavity is shown in Fig. 6B. In mouse ALBP, Arg-126 is available to bind a fatty acid carboxylate group, while in As-p18 the corresponding Arg (Arg-158) is chargeneutral in combination with Asp-36, invariably a serine in other similar proteins, and so less available for carboxylate , testis (T), and intestine (I) were separated on a 12% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. The membrane was incubated with affinity-purified antiserum against rAs-p18 and developed as described under "Experimental Procedures." B, Northern blot of A. suum RNA isolated from different larval stages and adult tissues using a cDNA clone for As-p18 as probe. Total RNA (20 g) from larval stages and adult tissues as in panel A was separated on a formaldehyde denatured 1% agarose gel and transferred to a nylon membrane. The membrane was hybridized with a full-length 32 P-labeled As-p18 cDNA probe and exposed for 3 days at Ϫ80°C. binding. The remainder of the ligand-binding cavity is lined with either hydrophobic residues or with hydrogen-bonded hydrophilic side chains. The only exceptions to this are Arg-46, which has its side chain directed toward the cavity, and Lys-90. Both these anomalies can be explained from the differences between the consensus sequence and that of As-p18. Arg-46 probably replaces Arg-158 in binding the fatty acid carboxylate group, while Lys-90 forms a charge pair with Asp-100, making this region of the ligand binding pocket also charge-neutral.
Additional evidence for the accuracy of this model, especially the insertions, is found in the positioning of the nonpolar hydrophilic side chains in the ligand-binding cavity. The positioning of Tyr-160, Asn-67, Asn-80, His-88, and the backbone of Glu-101 creates a four link hydrogen-bond chain forming one "wall"' of the ligand-binding cavity.
Fatty Acid Binding Activity of Recombinant As-p18 -As-p18 FIG. 6. Structural modeling of As-p18. A, model of the structure of As-p18. Ribbon representation of the predicted structure, based on the crystal structure of mouse ALBP. The positions of the three insertions which occur in As-p18, but not in mouse ALBP or other similar proteins, are shown. B, the predicted structure of the ligand binding site of As-p18 and the space available for ligand occupancy. The labeled residues projecting into the cavity are responsible for the formation of the predominately hydrophobic cavity and maintenance of the charge distribution on the cavity surface. Tyr-160, Asn-67, Asn-80, His-88, and the backbone of Glu-101 create a four-link hydrogen-bond chain forming one face of the ligand-binding cavity.
was expressed in E. coli without its putative leader peptide but with a His 6 affinity tag at its amino terminus and six additional amino acids derived from the vector (see "Experimental Procedures"). rAs-p18 was soluble under nonreducing conditions and was purified to apparent homogeneity by chromatography on Ni-NTA-agarose. Under SDS-PAGE on 12% gels, its apparent molecular mass was 22 kDa, while the calculated molecular mass is 18,309 Da (data not shown).
The fatty acid binding capacity of rAs-p18 was examined using fluorescent lipid analogs, as described previously (43,44). The emission spectrum of DAUDA is altered upon entry into fatty acid-binding proteins (45) and binding to rAs-p18 was accompanied by a substantial increase of fluorescence intensity and a shift in the fluorescence emission peak from 543 nm to 498 nm (Fig. 7). To place this blue shift in context, rat liver FABP, BSA, and the ABA-1 allergen of A. suum produce shifts in DAUDA emission to 500, 495, and 475 nm, respectively (43,45). The degree of the shift in the dansyl fluorophore emission is considered to be a measure of the polarity of the binding site (46), and that brought about by rAs-p18 is similar to other iLBPs. Titration of rAs-p18 with DAUDA yielded an apparent dissociation constant (K d ) of (1.2 Ϯ 0.5) ϫ 10 Ϫ7 M (Fig. 8). The binding of a natural, nonfluorescent ligand (oleic acid) was then determined from its competitive effects on the fluorescence of the rAs-p18⅐DAUDA complex. The progressive addition of oleic acid to the rAs-p18⅐DAUDA complex caused a stoichiometric reversal of the fluorescence effect, presumably by the displacement of DAUDA from As-p18. Analysis of the oleic acid titration curve yielded an apparent K d (oleic) of Ϸ 2 ϫ 10 Ϫ8 M (Fig.  8), which is within the range of K d found for other iLBPs (47).
We also investigated binding of rAs-p18 to a fluorescent fatty acid, DACA, in which the dansyl fluorophore is attached at the ␣ carbon, rather than the hydrocarbon terminal, as is DAUDA. The DACA probe was found not to alter its fluorescence characteristics when mixed with rAs-p18, although it bound well to BSA and rABA-1 in control experiments (data not shown). This lack of DACA binding may indicate that the mechanism of binding of fatty acids is different from BSA and ABA-1, or merely that the dansyl group interrupts the hydrogen bonding network which anchors the carboxylate group in iLBPs (48).
Several of iLBPs bind retinol in addition to fatty acids, so As-p18 was also tested for retinol binding by fluorescence. Since retinol is highly unstable in water, the fatty acid binding assay was modified so that retinol was added directly to the rAs-p18 solution in the fluorescence cuvette. This assay yielded no evidence of binding by rAs-p18, although control proteins BSA, bovine lactoglobulin and rABA-1 all showed strong binding, as revealed by a dramatic increase in the intensity of fluorescence of retinol (data not shown).
Many proteins that bind fatty acids have a tryptophan residue in or near the binding pocket of the protein, the fluorescence emission of which is altered upon ligand binding, presumably through direct involvement in ligand binding, alteration in the environment of the tryptophan, or local changes in protein conformation. This has been observed for ␤-lactoglobulin and for serum albumin (43,49), and, in the case of the former, this effect has been used to measure the affinity of ligand binding (49). As-p18 has three tryptophan residues, and excitation of these at 287 nm produced an emission spectrum that peaked at 339 nm (data not shown). Although this is a compound spectrum from the emission from three tryptophan residues, the spectrum was simple, and there was a clear blue shift away from that of solvent-exposed tryptophan in denatured protein (355 nm, data not shown). Presumably this means that the tryptophans are removed from solvent water even in the apo-protein. For comparison, the peak emissions of the single tryptophans in ␤-lactoglobulin, BSA, and ABA-1 are 335, 344, and 308 nm, respectively (data not shown and Ref. 43). There was no indication of a change in the tryptophan fluorescence of As-p18 upon addition of oleic acid ligand (data not shown). This suggests that the tryptophans in As-p18 are not close to the ligand, and none are associated with ligand binding. DISCUSSION We have demonstrated that the perivitelline fluid surrounding the developing larva of the parasitic nematode, A. suum, has a relatively simple polypeptide composition and have identified from it a novel, abundant fatty acid-binding protein, As-p18. The predicted amino acid sequence of As-p18 exhibits significant similarity to iLBPs isolated from a variety of different organisms, ranging from flatworms to mammals.
All LBPs appear to have similar conformations with 8 or 10 anti-parallel ␤-strands and two short ␣-helices, even though some members of the group exhibit as little as 20% sequence identity (37-38, 48, 50). However, conserved positions have been identified. As-p18 exhibits a number of significant differences in these conserved regions and, with the exception of presumptive LBP-encoding sequences recently identified through the C. elegans genome project, is the only member of this group with a putative secretory signal. Of the 39 conserved polar and nonpolar positions in most iLBPs, 8 are not conserved in As-p18. For example, two invariant residues that correspond to Ser-55 and Phe-57 of the human ALBP are at a proposed portal region for ligand entry (37) and are not conserved in As-p18 (Glu-79 and Leu-81).
Sequence differences aside, however, As-p18 is clearly a functional fatty acid-binding protein. It has a high affinity for oleic acid, and there is strong evidence that it is structurally very similar to iLBPs, based on the preponderance of ␤-sheet as indicated by the CD analysis and computer-based predictions of its secondary structure. However, while the modeling study does show that the sequence conforms well to a 10-stranded iLBP, there are indications of modifications to the binding site which could relate to ligand specificity. In addition, the unusual extra loops exposed on the surface of the protein might be involved in unsuspected protein-protein interactions.
Another unusual feature of As-p18 is the tight developmental control of its expression, which is not common among LBPs, although a few cases have been reported (51,52). As-p18 is not present in unembryonated eggs, but begins to be synthesized at about day 3 and is abundant in the perivitelline fluid by the time the L1 is formed on day 10. The reasons for the apparent lack of As-p18 protein in these stages is unclear, but the data suggest that the maternal As-p18 mRNA in the unembryonated egg (53) may not be translated. For the L3, however, an alternative explanation may be that As-p18 is synthesized and immediately secreted, so it does not accumulate within the larva. As-p18 appears to be present in the perivitelline fluid prior to development of any of the secretory structures possessed by later stages of A. suum, and no As-p18 immunoreactivity has been detected in the duct leading to the secretory pore in infective L2. 3 As-p18 must therefore arise in the perivitelline fluid by an unexpected route, and studies are under way to characterize the cells responsible for As-p18 synthesis and secretion into the perivitelline fluid before the development of a functional secretory system. In view of the similarity between As-p18 and the putative LBPs from C. elegans, it would seem valuable to investigate the developmental control, function and site of their synthesis in this organism. It is conceivable that one or both homologues are important to the survival of the embryo within the eggshell.
Although the function of As-p18 is unclear, a number of interesting possibilities are apparent. The early development of A. suum from unembryonated egg to infective L2 takes place entirely within a highly resistant eggshell, and the quiescent L2 can remain viable for several years while it awaits ingestion by an appropriate host (54). Carbohydrate is utilized during the first 5 days of development and then resynthesized from stored triglycerides by the glyoxylate cycle (55). Stored triglycerides are the only energy source for quiescent L2s, and the appearance of As-p18 in the perivitelline fluid parallels the increasing utilization of triglycerides as development proceeds (55,56). As-p18 may therefore play a role by sequestering potentially toxic fatty acids and their peroxidation products accumulating around a larva enclosed within an impermeable eggshell. The resistance of the eggshell is thought to be due to a highly impermeable lipid layer comprised of unusual glycosides (ascarosides) secreted by the fertilized egg prior to the expression of As-p18 (3). We are currently testing the binding capacity of As-p18 for ascarosides and potential products of lipid peroxidation. As-p18 also may play a role in the altered permeability of the lipid layer that accompanies the onset of hatching. This altered permeability permits the diffusion of trehalose out of the perivitelline fluid, and this change in osmotic pressure is presumed to activate the quiescent larva whose subsequent movement may further mechanically disrupt the lipid layer (1,5).
FABPs have been isolated from a number of flatworms, including S. mansoni, E. granulosus, and F. hepatica and have received attention as being potentially important in helminth parasitism. Indeed, the flatworm FABPs appear to be important antigens released during the course of infection with these helminths and have been identified as promising vaccine candidates (31,(33)(34)57). In addition, an ethanolamine-binding protein recently has been identified as a major antigen released during the larval migration of the related dog ascarid, Toxocara canis (58). Finally, the ABA-1 allergen produced by A. suum also displays strong fatty acid binding activity, even though it is quite different in structure from the LBPs (43,59).
As-p18 is, therefore, a member of the ␤-barrel lipid-binding proteins, which is unusual in the possession of a leader peptide in its precursor, and its lack of retinol binding. It is also the first member of this family from nematodes to be characterized as an FABP, and makes it probable that the proteins identified by similarity from C. elegans indeed represent functional FABPs. It clearly possesses some unique features, which may reflect adaptation to a specific function within the egg.