Exploring and Expanding the Fatty-Acid-Binding Protein Superfamily in Fasciola Species

: The liver ﬂ ukes Fasciola hepatica and F. gigantica infect livestock worldwide and threaten food security with climate change and problematic control measures spreading disease. Fascioliasis is also a foodborne disease with up to 17 million humans infected. In the absence of vaccines, treatment depends on triclabendazole (TCBZ), and overuse has led to widespread resistance, compromising future TCBZ control. Reductionist biology from many laboratories has predicted new therapeutic targets. To this end, the fatty-acid-binding protein (FABP) superfamily has proposed multifunctional roles, including functions intersecting vaccine and drug therapy, such as immune modulation and anthelmintic sequestration. Research is hindered by a lack of understanding of the full FABP superfamily complement. Although discovery studies predicted FABPs as promising vaccine candidates, it is unclear if uncharacterized FABPs are more relevant for vaccine formulations. We have coupled genome, transcriptome, and EST data mining with proteomics and phylogenetics to reveal a liver ﬂ uke FABP superfamily of seven clades: previously identi ﬁ ed clades I − III and newly identi ﬁ ed clades IV − VII. All new clade FABPs were analyzed using bioinformatics and cloned from both liver ﬂ ukes. The extended FABP data set will provide new study tools to research the role of FABPs in parasite biology and as therapy targets.


■ INTRODUCTION
The trematode liver flukes, Fasciola hepatica and F. gigantica, are the causative agents of fasciolosis, a foodborne zoonotic disease affecting grazing animals and humans worldwide.Liver fluke causes economic losses of over US$ 3 billion worldwide per annum to livestock via mortality, reduction in host fecundity, susceptibility to other infections, decrease in meat, milk, and wool production and condemnation of livers. 1 Liver fluke disease of livestock is increasing worldwide, 2 with a number of potential contributing factors: climate change (warmer winters and wetter summers supporting larger intermediate mud snail host populations); fragmented disease management (only treating sheep not cattle and limiting veterinary interaction); encouragement of wetlands; livestock movement; or failure/resistance of chemical control treatments in the absence of commercial vaccines. 3,4−7 Furthermore, worldwide livestock movement is providing new opportunities for the introduction of pathogenic isolates. 4ontrol of liver fluke is currently via anthelmintics.The benzimidazole (BZM)-derivative, triclabendazole (TCBZ), is the drug most extensively used against Fasciola.Unlike other fasciolicides, TCBZ shows activity against both juvenile flukes, which are responsible for the damage to the liver of acute fasciolosis, and the mature flukes, which cause the debilitation of chronic fasciolosis. 8−12 Following end of patent protection, generic forms of TCBZ will likely lead to wider application, potential misuse and exacerbation of the spread of resistance.
The mode of action of TCBZ at the molecular level has yet to be resolved.Laboratories report a variety of biological effects of TCBZ on liver fluke (for reviews see refs 9 and 13).As well as increased efflux, 14 enhanced biotransformation and metabolism of TCBZ has also been hypothesized to play a major role in detoxification and resistance.Anthelmintic resistance can arise from efficient detoxification via Phase II and Phase III conjugation, sequestration, and efflux mechanisms.To this end, a type I fatty-acid-binding protein (FABP), Fh15 protein, with sequestration potential is significantly upregulated on TCBZ exposure in resistant adult liver fluke, 15 and an FABP from Schistosoma japonicum has been reported to play a role in Praziquantel drug binding. 16Moreover, increased expression of FABP mRNA occurs in drug-resistant strains of Anopheles gambiae during permethrin insecticide exposure. 17Thus, increased expression of FABP may be a generic invertebrate response to drug challenge and a potential resistance marker.
The soluble super family of FABPs are small ( 14,15 kDa) proteins that bind or sequestrate hydrophobic ligands such as anthelmintics. 16The precise function of each FABP type has remained imperfectly understood since subspecialization of functions was suggested.At least nine distinct types of cytoplasmic FABPs have been identified in mammals, each showing a characteristic pattern of tissue distribution. 18FABPs isolated from the same tissue of different vertebrate species show sequence identities of 70% and higher, whereas FABPs isolated from different tissues of a single species have sequence similarity as low as 20%.However, their tertiary structure is remarkably conserved, consisting of 10 antiparallel β-strands comprising a βbarrel (containing conserved amino acid residues that are involved in ligand binding) and a helix-turn-helix cap. 16mportantly, for vaccine candidature and drug sequestration, FABP is also an abundant component of the soluble tegumental proteome of adult liver fluke. 19,20Vaccination against liver fluke remains in the research stage. 21,22Previous studies have placed cytosolic FABPs among the major potential vaccine candidates. 23accine trials suggest that both native and recombinant F. hepatica FABPs induce significant levels of protection in different animal models against infection with F. hepatica and crossprotection against S. mansoni and S. bovis with antifluke, antifecundity, and antipathology effects. 23he recent discovery of FABPs in the cargo of exosome-like vesicles released from adult F. hepatica, 24 suggests new roles for parasite FABPs within host cells.In support of this hypothesis, in vitro assays demonstrate that FABP I (Fh12) alters the behavior of monocyte-derived macrophages, with increased arginase expression/activity and an increase in Chitinase-3-like protein. 25ABP I also down-regulated nitric oxide production and the expression of nitric oxide synthase in interacting cells, exhibiting a potent anti-inflammatory effect in inducing the production of alternatively activated macrophages. 25Furthermore, FABP I has been shown to suppress inflammatory cytokines in a model of septic shock potentially delivering its effect via binding to CD14 coreceptors. 26espite growing evidence supporting the importance of FABPs for the establishment of liver fluke in the vertebrate host, our knowledge of the superfamily complement is fragmented even with many reductionist studies.We report the mining of the genome, transcriptomic and EST data sets, supplemented with proteomics and phylogenetics, to systematically reveal the complexity and novelty within the liver fluke FABP superfamily.Furthermore, we have identified which of these FABPs are recognized by the immune system for further vaccine discovery.

■ EXPERIMENTAL PROCEDURES Fluke Collection and Cultures
Adult F. hepatica were recovered from naturally infected ovine livers immediately postslaughter from a local abattoir in Mid-Wales, U.K. Fluke were washed several times in PBS at 37 °C to remove host material by regurgitation of gut contents, as previously described. 27Adult flukes were transported to the laboratory and maintained in Fasciola saline (FS; Dulbecco's modified Eagle's medium (DMEM) [w/o NaHPO 3 and PO 4 ] plus 2.2 mM Ca [C 2 H 3 O 2 ], 2.7 mM MgSO 4 , 61.1 mM glucose, 1 μM serotonin, 5 μg mL −1 gentamycin, 15 mM N-2hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES), pH 7.4) at 37 °C for 2 h.Metacercariae were purchased from Ridgeway Research and excysted following the physiological method outlined by Dixon 28,29 and previously described. 30tosol Preparations and 2DE F. hepatica extracts were obtained by homogenization of frozen fluke at 4 °C in lysis buffer containing 20 mM potassium phosphate, pH 7.4, 0.1% (v/v) Triton-X100, and protease inhibitors (Roche, Complete-Mini, EDTA-free).After homogenization, samples were centrifuged at 100 000g for 1 h at 4 °C, and the supernatant was termed the cytosolic fraction.Cytosolic protein extracts were precipitated with an equal volume of icecold 20% TCA in acetone (w/v) and washed twice in ice-cold acetone before solubilization into isoelectric focusing buffer (IEF buffer) consisting of 8 M urea, 2% w/v CHAPS, 33 mM DTT, and 0.5% carrier ampholytes v/v (Biolyte 3-10, Bio-Rad) as previously described. 31For preliminary 2DE, a total of 100 μg was passively in-gel rehydrated for 16 h and isoelectrically focused on 7 cm linear pH 3−10 IPG strips (BioRad) for 10 000 Vh.For analytical 2DE a total of 500 μg of each replicate sample was passively in-gel rehydrated for 16 h and isoelectrically focused on 17 cm pH 4.7 to 5.9 or 7−10 IPG strips (BioRad) to 60 000 Vh.All IEF was conducted on a Protean IEF Cell (Bio-Rad).
Gel image quantitative differences between protein spots were analyzed via Progenesis PG220 software, version 2006 (Nonlinear Dynamics).Spots were manually detected on gels with normalization performed using total spot volume multiplied by 100.Quantitative analysis was based on average gels created from four biological replicates.

Protein Identification
Protein spots were manually excised and tryptically digested and prepared for mass spectrometry as previously described. 32,33andem mass spectrometry (MS/MS) was performed according to the method of Moxon et al., 34 followed by data processing for database searching.Samples prepared for liquid chromatography−tandem mass spectrometry (LC−MS/MS) were analyzed using electrospray ionization (ESI), as previously reported. 34Peptide mixtures from trypsin-digested gel spots were separated using a LCPackings Ultimate nano-HPLC System.Sample injection was via an LC Packings Famos autosampler, and the loading solvent was 0.1% formic acid.The precolumn used was a LCPackings C18 PepMap 100, 5 mm, 100 Å, and the nano HPLC column was a LC Packings PepMap C18, 3 mm, 100 Å.The solvent system was: solvent A (2% ACN with 0.1% formic acid) and solvent B (80% ACN with 0.1% formic acid).The LC flow rate was 0.2 μL/min with a gradient employed using 5% solvent A to 100% solvent B in 1 h.The HPLC eluent was sprayed into the nano-ES source of a Waters Q-TOFμMS via a New Objective Pico-Tip emitter.The MS was operated in positive ion mode, and multiply charged ions were detected using a data-directed MS/MS experiment.Collisioninduced dissociation (CID) MS/MS mass spectra were recorded over the mass range m/z 80−1400 Da with scan time 1 s.MassLynx v 3.5 (Waters, U.K.) ProteinLynx suite of tools was used to process raw fragmentation spectra.Each spectrum was combined and smoothed twice using the Savitzky−Golay method at ±3 channels with background noise subtracted at polynomial order 15 and 10% below curve.Monoisotopic peaks were centered at 80% centroid setting.Sequest compatible (.dta) file peak mass lists for each spectrum were exported, and spectra common to each 2DE spot were merged into a single MASCOT generic format (.mgf) file using the online Peak List Conversion Utility available at www.proteomecommons.org.Merged files were submitted to a MASCOT MS/MS ions search within a locally installed Mascot server (www.matrixscience.com)to search an "in-house" database constructed from 6260 (858 763 residues) F. hepatica EST sequences were downloaded and translated from the Sanger Institute (ftp://ftp.sanger.ac.uk/pub/ project/pathogens/Fasciola/hepatica/ESTs/).Search parameters were as described in Morphew et al. 33

Western Blotting
Western blotting samples were prepared and separated according to the method described above.Proteins were transferred to HybondTM-C Extra nitrocellulose membrane (GE Healthcare) using a Trans-blot Cell (Bio-Rad) at 20 V overnight.Transfer was carried out in 192 mM glycine (Sigma), 25 mM Tris base, and 20% (v/v) methanol according to the method of Towbin et al. 35 The membrane was stained for 1 min with 0.1% (w/v) amido black 10B (Naphthol Blue Rack, Sigma) in 10% (v/v) acetic acid and 25% (v/v) isopropanol and destained in 10% (v/ v) acetic acid and 25% (v/v) isopropanol to evaluate transfer efficiency.
NC membranes were blocked in TTBS (0.1 M Tris base, pH 7.5, 0.9% (w/v) NaCl, 1% (v/v) TweenR 20 (Acros)) containing 5% (w/v) skimmed milk powder for a minimum of 4 h.Membranes were then washed in TTBS for 10 min before being incubated for 1 h with the primary antibody diluted at 1:5000 in TTBS containing 1% (w/v) skimmed milk powder.In all Western blots pooled, bovine whole sera from Fasciola hepaticachallenged and nai ̈ve cattle were used as previously described. 36embranes were washed three times for 5 min in TBS (0.1 M Tris base, pH 7.5, 0.9% (w/v) NaCl, 1% (v/v)) to remove residual primary antibody and incubated for 1 h with antibovine IgG (whole molecule) conjugated to alkaline phosphatase (AKP, Sigma) diluted at 1:30 000 or antibovine IgM (whole molecule) conjugated to AKP (Novus Biologicals) diluted at 1:1000 both in TTBS containing 1% (w/v) skimmed milk powder.The membranes were again washed three times for 5 min in TBS.Visualization of the resulting immunocomplexes was achieved by developing the membranes in 5-bromo-4-chloro-3-indolyl phosphate (BCIP; Sigma) and nitro blue tetrazolium (NBT; BDH, VWR Life Sciences) according to the manufacturer's instruction.

Bioinformatics: Transcripts and Phylogenetics
All sequence alignments were carried out using ClustalW multiple alignment 37 via BioEdit Version 7.0.9.0 (6/27/07). 38reviously published FABP sequences from both Fasciola species were retrieved from the Genbank database (http://www.ncbi.nlm.nih.gov/) and aligned with ESTs and transcripts from available sources 39 (ftp://ftp.sanger.ac.uk/pub/project/ pathogens/Fasciola/hepatica/ESTs/; EBI-ENA archive ERP000012: an initial characterization of the F. hepatica transcriptome using 454-FLX sequencing).Transcripts and ESTs that matched at least one of the already known FABPs in a BLAST analysis were included in alignments.From the alignment a phylogenetic tree was constructed in MEGA v 4.0 using a neighbor-joining method, 1000-replicate, bootstrapped tree.The amino acid data were corrected for a gamma distribution (level set at 1.0) and with a Poisson correction.

Bioinformatics: Analysis of Novel FABPs
Secondary structure analysis of novel FABP isoforms was conducted using PSIPRED available at http://bioinf.cs.ucl.ac.uk/psipred/.Motif analysis was conducted using InterPro 40 and phosphorylation predictions using NetPhos 2.0. 41Sequences from F. hepatica and F. gigantica encoding novel FABP isoforms were also subjected to epitope predictions using a Kolaskar and Tongaonkar Antigenicity prediction method, 42 available at http://tools.immuneepitope.org/tools/bcell/iedb_input,and signal peptide analysis using the Signal P 4.1 server and a cutoff point of 0.45. 43Intron−exon structures of novel F. hepatica FABP isoforms were also determined using Artemis 44 on sequences identified in the F. hepatica genome. 45

Recombinant Protein Production and Purification
FABP IV and V from both fasciolids were amplified from plasmids containing inserts with the addition of Nde-I and Not-I restriction enzyme sites for directional cloning into the pET28a (Novagen) expression vector.Recombinant FABP protein was expressed via the Escherichia coli BL21(DE3) and pET28a expression vector system (Novagen) and purified using nickelaffinity via a C-terminal polyhistadine tag as previously described. 31Purity was assessed by ESI mass spectrometry and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).

Characterization of FABP from F. hepatica Ontogeny
To effectively resolve the known FABP isoforms present in F. hepatica, we subjected both adult and newly excysted juveniles (NEJs) to preliminary 2D SDS-PAGE analysis using broad-range IPG strips to localize FABP representatives.On the basis of previous 2D SDS-PAGE investigations into the FABP superfamily of Fasciola, 15,19,36 the comparison of NEJ and adult 2D arrays revealed a dramatic reduction of FABP isoforms in NEJs (Figure 1a).This absence of FABPs from the NEJ proteome included the slightly acidic FABPs, isoforms I and II (Figure 1a (i)), and the basic isoform, FABP III (Figure 1a (ii)).As a result of these preliminary proteomes, further FABP isoform delineation was to be conducted upon adult samples using selective IPG strips (pH 4.7 to 5.9 for acidic FABPs and pH 7−10 for basic FABPs) to provide maximum resolution.

FABP Subproteome of Adult F. hepatica
Separation of adult F. hepatica somatic homogenates on narrow and micro range IPG strips and subsequent SDS-PAGE resulted in identification of FABP isoforms belonging to the type I, II, and III subfamilies.A total of 15 protein spots were consistently identified via Progenesis analysis lying below the 20 kDa marker and occupying the recognized FABP "zones" (Figure 1b).MSMS data from the peptides excised from 4.7 to 5.9 micro range gels (Figure 1c) confirmed that FABP II isoforms appear in four different locations on the gel, varying in pI as well as in molecular weight (Table 1).FABP I (Fh15) isoform was found in three locations of the same gel analysis.A discrepancy in pI and molecular weight between values calculated by Progenesis (5.3-5.6/17.4-20.5 kDa) and those predicted from Genbank entries (5.93/14936.04Da for FABPII, Q7M4G1 and 5.91/14712.08Da for Fh15, Q7M4GO) was observed.To account for multiple FABP locations, we investigated potential sites of posttranslational phosphorylation and identified five serine and threonine residues for FABP I (Fh15) and three for FABP II predicted with high significance to be phosphorylated.
MSMS data from the peptides excised from pH 7−10 narrow range gels (Figure 1d) confirmed that FABP III is the most abundant FABP isoform and was identified in five different locations on the gel, varying principally in pI (Table 1).FABP II was identified in two further locations and in both cases colocalized with FABP III.Of interest was the solitary identification of FABP III colocalized with a serine/threonine protein kinase.As with FABP I and II, FABP III potential sites of post-translational phosphorylation were investigated, identifying three serine and one threonine residues predicted with high significance to be phosphorylated.Only two protein spots identified did not contain FABP isoforms.Circled spots correspond to those proteins consistently present on averaged gels, and their putative protein identifications can be found in Table 1.Spectra from mass spectrometry were subjected to MSMS ion searches using MASCOT (Matrix Science) searching an in house EST database.Significant hits, at P = 5%, have a MASCOT score of 25 or greater.All significant EST hits were then subjected to BLAST analysis against GenBank to assign an identity to matching ESTs.Therefore, all reported accession numbers are from GenBank.NS: Not significant.

Potential Immune Recognition of FABP Isoforms
Having identified the isoforms of FABP present in adult F. hepatica, somatic homogenate samples were then subjected to 1D Western blotting using a pooled bovine infection sera time course to reveal the dynamics of the FABP superfamily members and the host IgG response.Somatic homogenate samples were probed with time-course sera from week 0, nai ̈ve animals, to week 14 post-infection (excluding week 10 post-infection) to look for Those from F. gigantica were taken from the study of Young et al. 39 and those from F. hepatica were taken from Young et al. 39 or transcripts produced by the University of Liverpool (EBI-ENA archive ERP000012: an initial characterization of the F. hepatica transcriptome using 454-FLX sequencing).Sequences from F. gigantica NEJs were sequenced in house.
FABP interaction.Data showed that an IgG response to proteins of the appropriate sizes could be observed from as early as 2 weeks post-infection (Figure S1A), with two bands visualized.The extent of this IgG recognition persisted until 14 weeks postinfection.To determine if any of the FABP isoforms (FABP I, II, or III) were responsible for the IgG recognition, we subjected all samples to 2DE profiling, as described previously followed by Western blotting.
The 2DE array, spanning pH 4.7 to 5.9, of the cytosolic fraction of adult liver fluke homogenate was transferred to NC membranes and assayed with sera from infected and nai ̈ve cattle.Pooled sera from week-8 post-infection produced the strongest IgG response and thus selected to probe 2DE Western blot arrays.The transfer of somatic samples to membranes for antibody testing was assessed using amido black staining (Figure S1b).Once probed with week 8 sera only very weak IgG responses were seen.This recognition was primarily to spots containing FABP I (Figure S1c, spots 2, 4, and 5 in Figure 1c) and extremely weakly to those spots containing FABP II (Figure S1c, spots 6−8 in Figure 1c).Importantly, the strongest, albeit weak, IgG response was to spot 2, which contained both FABP I and a translation initiation inhibitor (Figure S1c, spot 2 in Figure 1c).Thus, it is likely that the immune recognition of FABPs I−III is extremely low.Importantly, nai ̈ve sera (week 0) showed no IgG response (Figure S1d).IgG responses were also highlighted in an unresolved section of the gel at the pI 5.9 side of the gel.
The 2-DE array (pH 7−10) of the cytosolic fraction of adult liver fluke homogenate was transferred to NC membranes and assayed with sera from infected and nai ̈ve cattle.As with acid FABP isoforms, week-8 infection sera was chosen to screen the basic putative FABPs for immunoreactivity.However, no antibody recognition could be detected for any of the basic FABP isoforms or indeed any proteins lower than the 30 kDa marker (Data not shown).

Transcript and Phylogenetic Analysis of FABP Isoforms
Available transcript and EST data sets from both F. hepatica and F. gigantica and the F. hepatica genome were searched for representative FABP superfamily members.All identified members representing known FABP isoforms I, II, and III and potential new FABP isoforms were aligned with members of the known vertebrate FABP classes.From the multiple alignment a phylogenetic tree was constructed to ascertain the Fasciola FABP clade structure.The previously identified Fasciola FABP I, II, and III clades were well-conserved and most related to clades from the known vertebrate Heart (H, B, A, T, and M-FABP) and Intestinal (I and K-FABP) FABP groups (Figure 2).The FABP type I clade comprises the recombinant FABP I (Fh15) from F. hepatica and reveals a clear separation between representatives of both Fasciola species into two subclades.This subclade structure is not replicated in clades II or III because there is a lack of F. gigantica FABP II representatives, and the sequences of FABP III from both species are 100% identical at the amino acid level.A fourth novel clade (Bootstrap support 90%), termed FABP V, clustered near the known three isoforms but much closer to the known vertebrate classes, in particular, the vertebrate keratinocyte and intestinal FABP isoforms.Interestingly, the FABP Sm14 from S. mansoni clustered close to Fasciola FABP clade V. Three further novel clades of Fasciola FABPs were identified with strong bootstrap support, termed FABP IV (99%), VI (91%), and VII (96%; Figure 2).All three of these putative novel FABP clades, as with FABP clade I, also showed a distinct separation between representatives of F. hepatica and F. gigantica.Both FABP clades VI and VII were most related to the vertebrate liver and ileal forms of vertebrate FABPs with FABP clade IV to a lesser extent.
Representatives of the novel four FABP clades from both F. hepatica and F. gigantica were cloned and sequenced.This revealed proteins of 132 (FABP IV), 134 (FABP V), 162 (FABP VI), and 166 (FABP VII) amino acids in length.The average amino acid sequence similarity to the previously identified Fasciola FABPs ranged from as low as 22 to 48%.Specifically, FABP IV was most similar to FABP I (24 to 25%), FABP V to FABP III (48%), FABP VI to FABP II (24 to 25%), and FABP VII to FABP III (22%).Initially, all sequencing was performed on adult cDNA however expression of all four novel FABPs was confirmed in NEJ cDNA (data not shown).All F. hepatica and F. gigantica clones for the novel FABP classes are included in Figure 2.

Confirmation of Novel FABP Isoforms
All four newly identified putative FABP sequences were confirmed as FABPs using bioinformatics.As an initial comparison, the gene structures of the four new isoforms were compared with those of FABPs I, II, and III (Figure 3A).This analysis revealed that only the structure of FABP IV matched those of I−III.FABP V differed by only 6 bp (3 bp in exon 2 and 3 bp in exon 4).Of note was the dramatic increase in size of exon 4 in both FABP VI and VII.
All novel FABP isoforms were then subjected to secondary structure prediction analysis to identify the characteristic ten antiparallel β-strands comprising a β-barrel and the helix-turnhelix cap.All novel FABP isoforms from both F. hepatica and F. gigantica were predicted to have the FABP characteristic structure (10 β-strands and 2 α-helices; Figure 3B).All four novel isoforms were also subjected to InterPro sequence analysis for FABP domain predictions.In all cases, both IPR011038 Calycin-like and IPR012674 Calycin were predicted for both Table 2. Fatty-Acid-Binding Signature of Novel Fatty-Acid-Binding Proteins a signature requirement a Fatty-acid-binding signatures of all four novel FABPs from both F. hepatica and F. gigantica are highly conserved, especially for FABP V. Residues in bold show 100 suitability to the signature requirements.A percentage match is given for the 18 residue motif.
Fasciola species isoforms.Furthermore, three of the four isoforms (excluding FABP IV) were predicted to have IPR000463 cytosolic fatty-acid-binding properties.Of note was a wellconserved cytosolic fatty-acid-binding signature in all novel FABP isoforms, in particular FABP V with 94.4% sequence identity to the recognized motif signature (Table 2; GKWKLV-DSRDFDKVMVEL).All four novel sequences were also subjected to a SVMProt prediction to identify functional characterization.Novel FABPs were all identified as lipid-binding proteins, further confirming their status as FABPs, with the exception of FABP VI.In this instance, FABP VI was identified as a zinc-binding protein.As expected, all novel FABP isoforms lacked a signal peptide, further confirming them as cytosolic FABP isoforms.Importantly, both N-terminal residues G6 and W8 are conserved in all novel FABP isoforms.

Expression of FABP Isoforms IV and V
Because of their similarity to the three previously known FABP isoforms, both FABP IV and V from F. hepatica and F. gigantica were chosen for expression in E. coli for immunological characterization.Both proteins from both species were purified using Ni 2+ affinity chromatography and assessed for purity using 1D and 2D SDS-PAGE in conjunction with ESI−MS for an accurate mass assessment.All proteins were purified to a high level, as assessed by both methods (Figure 4).Both FABP IV (predicted molecular weight 15 073 Da and 16 351 Da with the Histag) and V (predicted molecular weight 15 372 Da and 16 650 Da with the Histag) had commonly observed Na + adduct formation (mass shifts of 22 Da between peaks) and a shift of 63 Da from the expected molecular weight relating to ammonium formate (NCOONH 4 ) adduct formation during ESI−MS analysis (Figure 4a).FABP V was observed as a more basic isoform from 2D SDS-PAGE (approximate pI 8.4; Figure 4c) than that of FBAP IV (Figure 4b).Minor shifts from the predicted molecular weights were seen for both isoforms, which may indicate lipid binding during purification as no delipidation steps were incorporated.Both expressed FABP isoforms were assessed for serum antibody responses using a time course of Fasciola-hepaticachallenged and nai ̈ve bovine sera (weeks 0−14 post infection).FABP IV showed no IgM or IgG responses throughout a 14week infection (data not shown).However, FABP V from both F. hepatica and F. gigantica had observable IgM and IgG responses, as assessed by Western blotting (Figure 5).IgM responses toward FABP V were visualized strongly at week 4−6, whereas IgG responses were visualized earlier at weeks 2−4.
A total of 41 Fasciola individuals were subjected to an analysis of FABP V to identify any potential sequence variation within the population (data not shown).This included 33 adult F. hepatica from three different populations (South Gloucester, Camarthen and Llanidloes; all U.K.) and eight adult F. gigantica (North and South India).Upon analysis, only a few single amino acid polymorphisms (SAAPs) were identified between all Fasciola spp.FABP V sequences.This was limited to just four of the individuals sequenced with a total of four SAAP sites.A single SAAP site was localized in two individual F. hepatica from the Llanidloes population (Q66L and Y130N), a further SAAP was identified in a Northern India F. gigantica sample (Q66R), and finally two SAAPs in one F. gigantica sequence from the South of India (V69I and K87R).Interestingly, three of the five SAAPs identified (those at sites 66 and 69) fall within a predicted epitope that spans amino acid residues 63−70.

Absence of Novel FABP Isoforms in Exosome-Like Vesicles
Because both FABP II and III have been previously identified in F. hepatica exosome-like vesicles 24 proteomic data were searched for all novel FABP isoforms.Interestingly, none of the novel FABP isoforms were identified in exosome-like vesicles (personal communication, Marcilla, June 2014).

■ DISCUSSION
Reductionist studies have proposed diverse roles in parasitism for FABP isoforms, including uptake/transport of fatty acids (FAs), as there is limited synthesis in helminth worms, immune modulation via their fatty acid ligands, and the sequestration of anthelmintics as a resistance mechanism.However, to effectively investigate the function of the FABPs, it is imperative to first resolve and fully delineate the complexity and novelty within the superfamily.Hence, we have utilized a polyomics approach incorporating genomics, transcriptomics, and proteomics to reveal an FABP superfamily from the pathogenic liver fluke parasite F. hepatica with seven distinct isoforms, four more than previously discovered.
Our previous observations revealed a dominance of FABP abundance in adult liver fluke rather than the more pathogenic NEJs, where FABPs were significantly reduced in abundance.Relatively low levels of FABPs in juvenile liver fluke compared with adults is replicated in other parasitic platyhelminths and thus is not simply related to adult survival in the liver/bile environment.For example, in schistosomes, an increase in FABPs is observed from newly developed schistosomula through lung-stage schistosomula and with the greatest abundance of FABPs found in adults, 46,47 suggesting that FABPs are important for development.The differential expression of FABPs from F. hepatica may be explained by functional roles, namely, intracellular transport and detoxification.Intracellular FABPs function as FA transport proteins. 16,18Therefore, in the absence of FA synthesis in adult parasitic flatworms, high levels of FABPs may transport FAs that are taken up via uncharacterized tegumental mechanisms from the host environment. 48However, within NEJs there is a high level of preformed/stored lipid to support initial host survival 49,50 and likely a reduced requirement for FA uptake and FABP transporters and hence their dramatic reduction observed in 2DE arrays (Figure 1A).The greater abundance of adult liver fluke FABPs may be related to adult feeding patterns.At the onset of blood feeding it has been suggested that FABPs are essential for the uptake of FAs from host blood. 51Furthermore, FABPs sequestrate and remove haem, a toxic byproduct of blood feeding responsible for hydrogen peroxide (H 2 O 2 ) production if in the free form. 52In support of this theory, antibodies directed toward F. hepatica FABPs can be observed around 2−4 weeks post-infection (ref 53 and the present study) coinciding with migration through the liver and the onset of blood feeding behavior.
Delineating the adult FABP proteome revealed the three known Fasciola FABP isoforms, reported in the recent publication of the F. hepatica genome, 45 in multiple locations within the 2DE arrays (three locations for FABP I, six for FABP II, and five5 for FABP III), an observation often causing confusion over true isoform identification. 54So why are the F. hepatica FABP isoforms resolving into numerous protein spots, and will this impact upon future diagnostics or vaccines?Multiple resolved versions of FABP isoforms have been attributed to posttranslational modification, especially phosphorylation, 55 and tyrosine phosphorylation is known to inhibit/modulate the binding of FAs. 56In mammals, phosphorylation of FABPs appears low, >1%, 54 but bioinformatics predicts multiple phosphorylation sites in Fasciola FABPs, tentatively suggesting a difference in FA transport regulation between host and parasite.In contrast, multiple resolved versions of FABP isoforms could also be attributable to irreversible ligand or FA binding. 54esolving the function of each FABP isoform is compounded by the complex diversity, tissue, and temporal specificity and ligand preferences of each FABP isoform. 16For example, FgFABP I and FgFABP III from F. gigantica have some overlapping roles, yet FgFABP I supports the male reproductive system, and FgFABP III supports the female reproductive system. 57It is likely that this distribution of FABPs I and III will be replicated within F. hepatica tissues.In the current study, F. hepatica FABP isoforms I−III were assessed for host immune recognition.Mirroring the study of Chunchob et al., 57 no immune recognition was observed to FABP isoform III using bovine sera.However, recognition of a protein spot containing FABP isoform I was seen, unlike for FgFABP I.However, the strongest, albeit weak, immune recognition identified was to a protein spot containing both FhFABP I and translation initiation inhibitor (TII).Given the lack of immune recognition to other FhFABP I protein spots in the array, the primary response seen may relate to TII.This is especially pertinent with the identification of TII in F. hepatica tegument preparations, both S2SS and UTCS fractions, of the study by Wilson et al., 20 and in surface preparations, both SPF and IPF protein fractions, from the study of Harcariz et al. 58 It is likely that TII is exposed to the Representative Western blots of rFhFABP V samples looking for both IgM and IgG responses rFhFABP V were run on 14% acrylamide SDS-host immune system and thus elicit the immune response identified in the current study.
We performed a bioinformatic analysis of currently available transcript and genomic databases for F. hepatica and F. gigantica to probe for the first time the complexity of the liver fluke FABP superfamily.As with previous phylogenetic studies, FABP isoforms I, II, and III formed a distinct group close to the vertebrate H-and I-FABP groups.FABP I isoforms had a clear separation related to Fasciola species not seen in FABP II or FABP III isoform groupings due to limited sequence data availability and 100% amino acid sequence identity between fasciolids, respectively.
Further phylogenetic analysis identified four new FABP isoforms represented in both F. hepatica and F. gigantica and in both adults and NEJs, namely, isoforms IV−VII.All four of these novel isoforms were confirmed as FABPs using bioinformatics.Only FABP isoform VI did not conform to every bioinformatics analysis predicting an FABP; namely, a SVMProt prediction did not identify FABP VI as a lipid-binding protein, predicting a zincbinding protein instead.In this case, FABP VI was classified as an FABP on its gene structure (exons 1−3 matching known FABP exon structures), protein structure (10 β-strands and 2 αhelices), and its highly conserved cytosolic FA-binding domain signature (88.9% amino acid identity).Phylogenetically, isoforms IV, VI, and VII were split between F. hepatica and F. gigantica, as seen with isoform I. Using bioinformatic analysis, all novel sequences were confirmed as authentic FABP isoforms by analyzing gene structures, sequence motifs, and secondary structure prediction.
FABP isoform V was closely related to isoforms I−III but closer to vertebrate I-and K-FABPs.Vertebrate I-FABPs are generally known to bind FAs only rather than additional ligands.In addition, FAs bound to I-FABPs are also bound in a different conformation to the other vertebrate FABPs (FAs bind to I-FABPs in a bent conformation instead of a U-shaped conformation when bound to alternative vertebrate FABPs). 59herefore, it is possible that FABP V is expressed specifically for the uptake of FAs from the host, although binding similarities to I-FABPs will require confirmation as the ligand-binding residues of I-FABP (Y70, L72, A73, W82, Q115 and Y117) are not conserved in FABP isoform V. 59 The three other novel Fasciola FABP isoforms, IV, VI, and VII, clustered away from the previously known Fasciola FABPs and were located close to the vertebrate L-and IL-FABP groups.As with the FABP I, II, and III isoforms, we would expect to identify multiple versions upon a 2DE array.The potential of cysteine modification via glutathionylation or cysteinylation, related to the redox state of the host liver, may complicate distinguishing true FABP isoforms, as seen for vertebrate L-FABP. 54Vertebrate L-FABPs are distinctly different from other vertebrate FABP groups.First, they differ in the FA uptake mechanism, which in L-FABPs occurs via diffusion 16 rather than collision, as demonstrated for H-and I-FABPs.Furthermore, as a result of their large binding pocket, ligands of L-FABPs bind at a 2:1 molar stoichiometry ratio as opposed to 1:1 in all others examined. 60mportantly, vertebrate L-and IL-FABPs are capable of binding FAs as well as bulky ligands such as bile salts, cholesterol, and haem. 61While all of these abilities remain to be confirmed in the novel Fasciola FABP isoforms IV, VI, and VII, it would seem logical that these isoforms could well be adapted for a "life in bile" as a blood feeder.Interestingly, two of the novel FABP isoforms, VI and VII, had a 31 and 33 amino acid C-terminal extension, respectively.While the role of these extensions is currently unknown, they may well be involved in interactions or attachments to other proteins or for membrane association. 62t has been suggested that a reduced diversity of FABP isoforms in invertebrates compared with vertebrates represents a lower specificity for ligands but a larger repertoire of interactions within the cell. 16In contrast, the number of FABP isoforms in Fasciolids has expanded, suggesting potential specialization of FABP isoforms.The vertebrate groupings H-, I-, and L-FABPs reflect their specific binding abilities. 61Thus, it is likely that Fasciola FABPs may also segregate according to their binding abilities.The separation of FABP isoforms I, II, III, and V from isoforms IV, VI, and VII may represent two clusters based on their respective binding capabilities.
Disappointingly, none of the newly recognized FABP isoforms (IV−VII) were identified during proteomic analysis.This could potentially result from the quantity of protein in the identified protein spots, identified as FABPs I, II, and III, suppressing any recognition of the novel FABP isoforms.Alternatively, it may be that FABPs IV−VII may be of more importance in the juvenile stages of the parasite; such is the case with cathepsin proteases (cathepsin L1, 2, and 5 in adults, whereas cathepsin L3 and 4 with cathepsin B in juveniles). 27ABP isoforms IV and V were expressed as recombinant forms to further understand the immune responses directed to the FABP family for diagnostic or vaccine potential.Despite its absence from F. hepatica exosome-like vesicles, FABP isoform V shows potential as a diagnostic or vaccine candidate with strong IgM and IgG responses seen in pooled bovine infection sera.As diagnostics, Fasciola FABPs show promise with two studies by Allam and colleagues, suggesting FABP for the diagnosis of F. gigantica infections in both buffalo and man.63,64 In both cases FABPs were purified from crude adult worm extracts, likely to contain multiple FABP isoforms.Thus, FABP isoforms I−III are likely to dominate the preparation, but the presence of FABP IV−VII cannot be discounted.Furthermore, Hillyer et al. 53 also noted that antibodies to FABP I could also be observed at 2−4 weeks post-infection, highlighting the excellent potential of FABPs as diagnostics.However, to further improve diagnostic potential, the correct choice of which FABP isoform to target is essential.For example, from the evidence presented in the current study, poor recognition to FABP I−III in natural infections and strong IgM and IgG responses to FABP V from weeks 2−4 FABP isoform V may be a potential choice as a diagnostic.However, with the expanded FABP family, from three to seven members, each isoform must be investigated specifically to gauge the best choice for a diagnostic.
As vaccine candidates, FABPs have also been studied in depth.Early studies using FhFABPs looked promising, with trials in mice revealing 69−78% protection.Unfortunately, these protection rates did not translate well into cattle trials with two studies giving 31 and 55% protection (vaccine trials reviewed by ref 65).Poor translation into target animals was also observed in trials for protection against F. gigantica.For example, in buffalo, a 35.8% reduction in worm burdens with associated reductions in liver enzymes (AST) was observed. 66However, this trial also demonstrated a high antifecundity effect. 66As with other candidate vaccines based on protein superfamilies, the choice of which isoform to vaccinate with is key decision ensuring plasticity of the target is low. 27Therefore, Fasciola FABP isoform V may be a strong candidate for novel vaccine trials, perhaps as part of a combination therapy.A potential combination therapy could incorporate FABP V and an immune suppressor component, such as FhGST-S1. 19,67FABP potential could also be enhanced by conducting similarity studies between those isoforms eliciting an immune response (FABP isoforms I, II, and V), which may reveal the structural details that are responsible for triggering this immune response in the host organism.This could lead to the rational design of protein complexes that may prove to be more effective in vaccine trials than current vaccine candidates.

■ CONCLUSIONS
A polyomics approach has successfully revealed the cytosolic FABP superfamily complement expressed in F. hepatica adults.These have been classified into seven isoforms, types I−VII, with a potential clustering of isoforms into two groups: group one consisting of isoforms I, II, III, and V and group two consisting of isoforms IV, VI, and VII.These two groupings may reflect their binding dynamics.Importantly, FABP isoform V from both F. hepatica and F. gigantica shows promise as a new diagnostic antigen or as a vaccine candidate.

* S Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jproteome.6b00331.
Figure S1.IgG antibody responses to known F. hepatica FABP isoforms using Western blotting.(PDF) Table S1.Full coverage of the putative protein identifications of F. hepatica fatty-acid-binding proteins using MASCOT.(XLS)

Notes
The authors declare no competing financial interest.

■ ACKNOWLEDGMENTS
This work was supported by the Biotechnology and Biological Sciences Research Council; Grant BB/H009256/1: Developing a "validation portfolio" to exploit key virulence proteins in Fasciola species for parasite control.

Figure 1 .
Figure 1.Representative global protein arrays of F. hepatica FABPs.F. hepatica ontogeny 2D SDS-PAGE protein arrays of (a) newly excysted juvenile and (b) adult F. hepatica.Circled areas within the arrays localize known FABP isoforms I, II, and III.Therefore, narrow range (pH 7−10) and microrange (pH 4.7 to 5.9) IPG strips were chosen as the most suitable range to effectively resolve the F. hepatica FABP members (denoted by dashed lines).Both arrays were loaded with 250 μg of cytosolic protein and run on linear pH 3−10 IPG strips and on 14% acrylamide SDS-PAGE.Both arrays were Coomassie blue stained.(c) Representative microrange (pH 4.7 to 5.9) and (d) narrow range (pH 7−10) protein arrays of adult F. hepatica somatic samples for FABP location and identification.Both panels c and d were loaded with 1 mg of cytosolic protein, run on 14% acrylamide SDS-PAGE, and Coomassie blue stained.Circled spots correspond to those proteins consistently present on averaged gels, and their putative protein identifications can be found in Table1.

Figure 2 .
Figure 2. Phylogenetic analysis of vertebrate and Fasciola FABPs to determine the clade structure of FABP isoforms present in F. hepatica and F. gigantica.Neighbor-joining phylogenetic tree constructed using amino acid sequences through MEGA v 4.0 with 1000 bootstrapped support and a Poisson correction.All reported accession numbers are from Genbank.Where sequences were identified in silico, only contig numbers are reported.Those from F. gigantica were taken from the study of Young et al.39 and those from F. hepatica were taken from Young et al.39 or transcripts produced by the University of Liverpool (EBI-ENA archive ERP000012: an initial characterization of the F. hepatica transcriptome using 454-FLX sequencing).Sequences from F. gigantica NEJs were sequenced in house.

Figure 3 .
Figure 3. Characterization of novel FABP isoforms.(A) Bioinformatic characterization of F. hepatica fatty-acid-binding protein gene structures.All three currently identified FABP isoforms (I, II, and III) were identified within the F. hepatica genome 45 and their intron-exon structures identified.These were with those of the four novel FABP isoforms (IV, V, VI, and VII).Exons shaded in gray indicate deviation from those structures identified within isoforms I, II, and III.Reported exon sizes are in nucleotide bp.(B) Fatty-acid-binding protein secondary structure prediction.A multiple alignment of all four novel FABP isoforms from both F. hepatica and F. gigantica was subjected to secondary protein structure prediction to identify the FABP characteristic structure containing 10 β-strands and 2 α-helices.Predictions were carried out using PsiPred Version 3.2. 68Each β-strand or α-helix is boxed and numbered.The extended C-terminal of isoforms VI and VII is apparent with additional β-strand or α-helix predictions shown in gray shaded dashed boxes or open dashed boxes, respectively.The three domains that constitute the lipocalin-binding domain are boxed in blue.Arrowed are the starting and ending residues of the IPR000463 cytosolic fatty-acid-binding domain signature within domain 1.The GXW triplet domain in domain 1 is underlined in green.All predicted phosphorylation sites (S, T, and Y) are boxed in red.

Figure 4 .
Figure 4. Expression of recombinant protein forms of novel FABP isoforms.(a) Purification of FABP IV from F. hepatica with purity revealed via ESI mass spectrum of the Ni 2+ affinity purified rFhFABP IV showing the molecular weight at 16287.7 ± 0.49 Da (with six sodium adduct peaks).Data for FABP V not shown.(b) Purity was also assessed with a 2D SDS-PAGE protein array run on 14% SDS-PAGE and Coomassie blue stained.(c) Purification of FABP V from F. hepatica assessed with a 2D SDS-PAGE protein array run on 14% SDS-PAGE and Coomassie blue stained.

Figure 5 .
Figure 5. Antibody responses to novel Fasciola FABP isoform V.Representative Western blots of rFhFABP V samples looking for both IgM and IgG responses rFhFABP V were run on 14% acrylamide SDS-PAGE, electro-transferred to membranes, and Western blotted with pooled bovine infection sera from week 0, nai ̈ve sera, to week 14 postinfection as the primary antibody with antibovine IgM (a) or antibovine IgG (b) as the secondary antibody.A F. hepatica somatic sample (S) was also included and as a positive control using sera from week 8 postinfection.White arrows indicate the start of immune recognition.Black arrow highlights the F. hepatica cathepsin L proteases recognized by bovine IgG.36