Aquatic organisms and, in particular, filter feeders, such as mussels, are continuously exposed to toxicants dissolved in the water and, presumably, require adaptations to avoid the detrimental effects from such chemicals. Previous work indicates that activity of ATP-binding cassette (ABC) transporters protects mussels against toxicants, but the nature of these transporters and the structural basis of protection are not known. Here we meld studies on transporter function, gene expression, and localization of transporter protein in mussel gill tissue and show activity and expression of two xenobiotic transporter types in the gills, where they provide an effective structural barrier against chemicals. Activity of ABCB/MDR/P-glycoprotein and ABCC/MRP-type transporters was indicated by sensitivity of efflux of the test substrate calcein-AM to the ABCB inhibitor PSC-833 and the ABCC inhibitor MK-571. This activity profile is supported by our cloning of the complete sequence of two ABC transporter types from RNA in mussel tissue with a high degree of identity to transporters from the ABCB and ABCC subfamilies. Overall identity of the amino acid sequences with corresponding homologs from other organisms was 38–50% (ABCB) and 27–44% (ABCC). C219 antibody staining specific for ABCB revealed that this transporter was restricted to cells in the gill filaments with direct exposure to water flow. Taken together, our data demonstrate that ABC transporters form an active, physiological barrier at the tissue-environment interface in mussel gills, providing protection against environmental xenotoxicants.
- ATP-binding cassette transporters
- multixenobiotic resistance
aquatic organisms are continuously exposed to a multitude of natural and anthropogenic toxicants dissolved in the water. Cellular toxicant defense systems are therefore extremely important, and efflux pump mechanisms have been proposed to play a major role as a first line of cellular defense. These efflux pumps are members of the ATP-binding cassette (ABC) transporter family and recognize a vast array of chemicals as substrates. Because of elimination of multiple toxicants from the cell, this protective mechanism is referred to as multixenobiotic resistance (MXR) (8, 28). MXR is supposedly ubiquitous in aquatic invertebrates, and indications for expression and activity of MXR pumps were found in cells of sponges (Porifera) (30), innkeeper worms (Echiurida) (46), sea stars and sea urchins (Echinodermata) (19, 41), and bivalves and snails (Mollusca) (6, 31, 44). The presence of MXR pumps in cells of these organisms is supported by functional assays showing increased accumulation of putative substrates in the presence of P-glycoprotein (P-gp) transporter inhibitors and cross-reactivity with pan-P-gp antibodies. Other ABC subfamilies provide protection against toxicants (33), and recent work on sea urchin physiology and genomics has provided the first example of action of these other ABC transporter types in an aquatic invertebrate (19).
Various studies have indicated expression and activity of putative MXR pumps in gill tissue of bivalves by functional assays using transporter substrates and inhibitors (6, 13, 28, 29, 44), binding of the antibody C219 directed against an epitope of the mammalian P-gp to protein fractions from gill tissue, and homology of partial cDNA sequence fragments with ABC transporter genes (6, 11–13, 25, 36, 37, 39, 44). However, a lack of extended sequence data precludes exact identification of the transporter types in bivalves, and there has been no evidence for activity of a transporter type other than a putative P-gp pump in bivalve gills.
Here we examine the nature of efflux transport in bivalve gills by quantifying efflux transporter activities with specific pharmacological inhibitors, by cloning of complete ABC transporter genes and analysis of respective protein sequence data for homologies and structural characteristics, and by identification of tissue areas with transporter expression. We provide evidence for activity and expression of an ABCB and an ABCC transporter in the gills and present full-length transporter cDNA sequences, enabling their classification. Transporter expression was found in regions of the gills with high environmental exposure, establishing a structural mechanism for cellular xenobiotic defense and a physiological barrier in the gills. Taken together, our data support the significance of ABC transporters as an active physiological, selective-permeability barrier in bivalve gills that antagonizes uptake of putative toxicants from the environment.
MATERIALS AND METHODS
California mussels (Mytilus californianus Conrad, 1837) (valve length = 68–80 mm) were collected from the rocky intertidal zone at Hopkins Marine Station. Animals used for efflux transporter activity assays were maintained in tanks with running seawater (∼15°C) for ≥24 h and ≤2 wk before experiments. Animals used for RNA extractions or immunohistochemistry were processed immediately after collection.
Calcein-AM (Ca-AM) was obtained from Molecular Probes (Eugene, OR) and MK-571 from Cayman Chemicals (Ann Arbor, MI). PSC-833 was a gift from Novartis (Basel, Switzerland). P-gp-specific antibody C219 was purchased from Signet Laboratories (Dedham, MA) and goat anti-mouse FITC-conjugated antibody from Sigma (St. Louis, MO). All other chemicals were obtained from Sigma.
Efflux Transporter Activity in Cells of Mussel Gills
Ca-AM, a substrate of ABCB and ABCC transporter proteins (9, 22), was used to record transporter activity in mussel gill cells. Ca-AM is nonfluorescent, but once inside cells, it is hydrolyzed by cytosolic esterases and forms green, fluorescent calcein. Compared with Ca-AM, which diffuses through cell membranes and is actively effluxed by transporter proteins, translocation of calcein from the cytosol to the outside of the cell by diffusion through cellular membranes or by active transport is negligible (9, 22). Inasmuch as calcein is instantly generated from Ca-AM in the cell, the amount of calcein that accumulates in the cell resembles almost the entire Ca-AM influx. ABC efflux transporters strongly antagonize Ca-AM influx; thus the amount of calcein in the cell is a measure of transporter activity: high activity of transporters is indicated by a weaker calcein signal, whereas a stronger fluorescence signal corresponds to lower transporter activity.
Activity of transporters in gill cells was assayed according to the procedure reported previously (35) with some modifications. Gill tissue was excised from live mussels and placed in filtered seawater (FSW). Biopsy punches (Acuderm, Ft. Lauderdale, FL) were used to excise equally sized tissue pieces (5 mm diameter) from gills. The tissue disks were flushed with FSW to remove mucus and kept in FSW until use.
For testing chemicals to modulate Ca-AM uptake of mussel gills, stocks of the test compounds dissolved in DMSO were added to FSW with 0.25 μM Ca-AM to the desired concentrations. All solutions were adjusted to 0.1% DMSO along with a DMSO control. DMSO at this concentration had no effect on transport activity. We found that the test compounds PSC-833 and MK-571 precipitate at higher concentrations in seawater; therefore, the maximum tested concentrations were 20 μM for PSC-833 and 50 μM for MK-571. Incubations were performed in polystyrene dishes in a volume of 5 ml of test solution. Three disks per treatment were exposed to test solutions for 90 min in the dark with gentle rocking, washed twice in FSW to remove dye from the tissue surface, frozen on dry ice, and stored at −20°C. For quantification of accumulated calcein, each tissue disk was sonicated in 200 μl of hypotonic lysis buffer (10 mM KCl, 1.5 mM MgCl2, and 10 mM Tris·HCl, pH 7.4), the sonicates were centrifuged at 13,000 g for 5 min, and the supernatant was transferred to a black 96-well microplate (Whatman, Middlesex, UK). The quantity of dye in the supernatant was measured with a fluorescence microplate reader (CytoFluor II, PerSeptive Biosystems, Framingham, MA; 530-nm emission, 485-nm excitation). Background fluorescence signals of lysis buffer and gill tissue with no calcein were negligible. Transporter activity assays were repeated at least three times for each concentration of test compound on a different mussel for each experiment.
Mean and SD of arbitrary fluorescence values from replicate experiments were calculated for each concentration of test compounds and untreated controls. To determine whether differences between calcein fluorescence values in extracts of inhibitor-treated and untreated tissue disks were statistically significant, ANOVA and Dunnett's test were applied. Differences were regarded as significant if P < 0.05. For statistical analyses, JMP version 5.1 (SAS Institute, Cary, NC) was used.
PCR and Analyses of Transporter Sequences
Total RNA was extracted from ∼30 mg of gill tissue that had been excised from a freshly collected specimen of M. californianus and vigorously washed in ice-cold FSW. For extraction of total RNA, the Nucleospin RNA II kit (Macherey-Nagel, Düren, Germany) or TRIzol reagent (Invitrogen, Carlsbad, CA) was used according to the manufacturer's instructions.
Full-length cDNA sequences of MXR transporters from mussel gill were obtained with rapid amplification of cDNA ends (RACE). For RACE reactions, the SMART RACE cDNA amplification kit (Clontech, Palo Alto, CA) was used.
For the construction of primers for the RACE reactions, partial cDNA sequences of abcb and abcc orthologs were obtained using PCR. The primers used in these initial reactions were directed against sequences of transporter genes from Mytilus edulis (36, 37) [National Center for Biotechnology Information (NCBI) accession nos. AF159717 and AF39714]. Respective amino acid sequences of these partial cDNAs show a high degree of identity with ABCB and ABCC orthologs from other organisms. The primer pairs AGA GGT TCT ATG ACC CAG ATG CAG GAC/GAA GAA CAA CCT TGG TGA TAG CAC ACC (partial abcb) and ATG AAG CTG GAG CCG TTT GAT GAA TAC/TCT AAT AGT AAA TTG GTC GGT GAG TCG (partial abcc) generated 441- and 395-bp PCR products, respectively, that were highly similar to the corresponding M. edulis sequences (data not shown). Homology of the translated sequences with translated abcb and abcc cDNAs from other organisms was confirmed with NCBI tblastx (data not shown).
3′- and 5′-RACE of the abcb ortholog.
In nested reactions, the primers GCG AGG AGC TCA GTT ATC AGG AGG GCA G and CCA TTG CCA GGG CTT TGA CTA GAG ACC C were used for 3′-RACE (1,283-bp product), and the primers GGG TCT CTG ATC AAA GCC CTG GCA ATG G and CTG CCC TCC TGA TAA CTG AGC TCC TCG C were used for 5′-RACE (1,854-bp product). An additional 3′-RACE reaction was performed using nested primers directed against a sequence farther downstream: GCT GAC AAA GAA GAG GAG CCT GAG GAG C/GGC ATC ATG TGT TGC TGG ATG TAC CAT GCC (2,459-bp product).
3′- and 5′-RACE of the abcc ortholog.
A nested 3′-RACE reaction was performed with the primers GTT CTC GAT GAA GCA ACC GCT GCC GTA G and CAG ATG ACC TCA TAC AGA CCA CGA TCA G (345-bp product). For 5′-RACE reactions, the primers CTG ATC GTG GTC TGT ATG AGG TCA TCT G/CTA CGG CAG CGG TTG CTT CAT CGA GAA C (649-, 1,609-, and 2,150-bp products), CAC TGA TGA TCC AGT CCG TCT TTT AAA CC/GAG GGC TGT CCA AAT ATC TTC GTT AGA G (454-bp product), GCA CGT GCC AAA CTG ACT CTT TGC TTT TG/CCC TTT TCA CCA ATT TCT GTC TGA TCC (1,417-bp product), and GGA TAG CGA GTA TGT AAC CTT TCC AG (1,132-bp product) were used.
PCR was performed for 40 cycles at 94°C (5 s), 65°C (10 s), and 72°C (3 min) or in a touchdown mode. RACE PCR products were gel purified, cloned into the vectors PCRII-TOPO (Invitrogen) and PGEM-T-easy (Promega), and sequenced on an ABI 3100 sequencer using standard cycle sequencing protocols. Partial cDNA sequences were edited and assembled using Sequencher version 4.5. Sequencing coverage was at least three times for all sequences.
For PCR amplification of the full-length cDNAs of mussel, abcb and abcc primers directed against parts of the assumed 3′- and 5′-untranslated regions (UTRs) of the genes were used. Primers were GGA TTT ATA CAG GTC GGC ATA TCC C (forward) and CCT GAA AAT GAG GTC AAG TGA ACT TAC TA (reverse) for abcb and GGA GGT ATG TTC ACA CTT TAT C (forward) and CAT TTG ACA TAA AAT CAT ATA TCA AAG AG (reverse) for abcc. PCR was performed for 35 cycles at 94°C (10 s), 58°C (10 s), and 72°C (5 min) using Advantage 2 polymerase mix (Clontech).
Sequences were analyzed using tools at the Expert Protein Analysis System (Expasy) Proteomics Server (http://expasy.org/) and NCBI basic alignment search tool (BLAST) search engines. The Polyphobius algorithm (http://phobius.sbc.su.se/poly.html) was used to identify transmembrane helices within the amino acid sequences of the mussel transporters. This algorithm provides homology-supported predictions of the transmembrane topology of peptides (24). Predictions made by the Polyphobius algorithm for the transmembrane topologies of human ABC transporters corresponded well with models in the literature (3, 7); therefore, we chose the Polyphobius algorithm for the analyses of the transporter sequences from mussel. Multiple sequence alignments and determinations of identity rates between amino acid sequences of ABC transporters from different organisms were performed using Clustal X version 2.0. MEGA version 4 was used for molecular phylogenetic analyses of ABC transporter sequences from different organisms (45).
Immunostaining of ABCB transporter protein was performed on cryosections of mussel gill tissue. Freshly excised gills were fixed in 4:1 methanol-DMSO for 3 h at −20°C. The tissue was washed three times in PBS (5 min each) and then soaked in optimal cutting temperature (OCT) compound at 4°C overnight. Cryosections (7–9 μm) were cut from a frozen block of tissue in OCT compound and transferred to poly-l-lysine-covered slides, and sections were washed several times with PBS. For antibody staining, sections were first covered with a blocking solution (5% BSA and 0.05% Tween in PBS), then with anti-ABCB1 antibody C219 (1:20–1:50 dilution), and, finally, after three 10-min washes with PBS, with goat anti-mouse FITC antibody (1:200 dilution). Antibodies were diluted with blocking solution. Blocking and antibody exposures were performed in a moist chamber at 37°C for 1 h. After three 10-min washes with PBS, sections were mounted and viewed under a fluorescence microscope (Axiovert S-100, Zeiss). We controlled for unspecific fluorescence or unspecific binding of secondary antibody by examining sections that had been subjected to the identical staining procedure, except buffer free of C219 was used at the step of exposure to primary antibody.
Effects of PSC-833 and MK-571 on Transporter Activity in Mussel Gills
The cyclosporin derivate PSC-833, an inhibitor of ABCB (38), and MK-571, which blocks ABCC-type pumps (14), increased calcein fluorescence in a dose-dependent manner, indicating inhibition of Ca-AM efflux transporters (Fig. 1). A slight effect of PSC-833 was seen at 0.1 μM, and effects were significant at ≥5 μM (P < 0.05), where the effect was saturated. MK-571 inhibited Ca-AM efflux at higher concentrations, with significant effects at ≥20 μM (P < 0.05). The concentration at which the MK-571 effect was saturated could not be determined, since the inhibitor precipitated from solution at >50 μM.
Figure 1 also shows the effect on activity when the two inhibitors are combined. If MK-571 (20 μM) is added to gill tissue treated with a concentration of PSC-833 at which the effect is saturated (20 μM), fluorescence (inhibition) increases beyond the maximum-effect levels when either compound is applied alone. This indicates that both inhibitors act on different targets, supporting the existence of at least two transporters: one maximally inhibited by PSC-833, most likely an ABCB-type activity, and another inhibited by MK-571, most likely an ABCC-type activity.
These data also indicate that saturation of the PSC-833 effect at 5 μM is not a result of some unspecific inhibition by PSC-833 of the esterase that cleaves Ca-AM into fluorescent calcein. If this were the case, there would be no further increase in fluorescence when MK-571 is added, since esterase inhibition would have prevented formation of the fluorescent product.
Transporter Sequence Data
Sequences from RACE reactions were assembled and resulted in two full-length transcripts of ABC transporters from M. californianus gill tissue. In addition, 172 and 790 bp of 5′- and 3′-UTRs from abcb and 105 and 133 bp of 5′- and 3′-UTRs from abcc were identified. Sequences of full-length cDNAs obtained by assemblies of RACE products were confirmed by RT-PCR using primers directed against sequences of the 3′- and 5′-UTRs of the genes. The PCR products had the expected sizes of 4,080 bp (abcb) and 4,580 bp (abcc), respectively (Fig. 2), and were consistent with the sequences of the assembled RACE products.
Open reading frames are 3,936 bp for mussel abcb (accession no. EF521414) and 4,497 bp for mussel abcc (accession no. EF521415) orthologs.1 Sizes of the respective peptide structure products are 144.5 kDa (1,311 aa) and 177.6 kDa (1,498 aa).
Analyses of the amino acid sequences with Prosite (http://ca.expasy.org/prosite/) revealed the structural organization of full ABC transporters in two subunits, each with a membrane-spanning domain (MSD) and a nucleotide-binding domain (NBD). The NBDs contain typical and highly conserved motifs of ABC transporters: the Walker A, Walker B, and ABC signatures and so-called A loops upstream of the Walker A regions (Figs. 3 and 4) (1). Further analyses of transmembrane domains with the Polyphobius algorithm indicated 12 and 17 putative transmembrane helices for mussel ABCB and ABCC, respectively (Figs. 3 and 4). The transmembrane helices of mussel ABCB can be associated with two MSDs, each containing six transmembrane helices; this finding is consistent with the typical organization of other full transporters from the ABCB family. Mussel ABCC probably contains a third MSD in the NH2-terminal region of the protein comprising five transmembrane helices, whereas the two following MSDs each contain six transmembrane helices. This structural organization is in accordance with the topology of the human “long” ABCC transporters ABCC1, ABCC2, ABCC3, ABCC6, and ABCC7 (7) (Fig. 3). Typically, the NH2-terminal end in these proteins is located extracellularly, whereas it is oriented vs. the cytosol in the ABC transporter proteins with only two MSDs, such as mammalian ABCB1, ABCB4, and ABCB11 and ABCC4, ABCC5, ABCC8, and ABCC9 (7) (Fig. 3).
N-glycosylation of proteins is indicated by sequon Asn-X-The/Ser (where X can be any amino acid except proline) (18, 26). Three N-glycosylation sequons are in the NH2-terminal MSD at positions Asn91-Gly92-Ser93, Asn100-Ala101-The102, and Asn106-Ala107-The108 of mussel ABCB, corresponding to the extracellular loop closest to the NH2-terminal end (Figs. 3 and 4). Glycosylation most likely explains the ∼30-kDa discrepancy between the size predicted for mussel ABCB from the amino acid sequence and the size indicated for this protein by Western blot (170 kDa) (6). Human ABCB1, which shows a high degree of identity with mussel ABCB, is glycosylated and contains N-glycosylation sites in positions similar to those in the mussel protein (3) (Fig. 4).
ABCC from mussel contains two N-glycosylation sequons at positions Asn969-Gly970-The971 and Asn973-Glu974-The975, corresponding to the extracellular loop at the NH2-terminal end of MSD2 (Figs. 3 and 4). In contrast, homologous human ABCC1 has two N-glycosylation sequons in MSD1 and one in MSD2 (21) (Fig. 4). Since glycosylation has no effect on functional properties of ABCC1 (40), these differences in glycosylation are not indicative of differences in function of the two protein homologs.
NCBI Blast2 Protein Database Query (blastp) showed close matches (expect values of 0.0) of the mussel sequences with multidrug-resistance (MDR)/P-gp/ABCB protein or multidrug resistance protein (MRP)/ABCC protein from various other organisms.
Identities of mussel ABCB are 38–50% with ABCB1 (P-gp, MDR1), ABCB4 (P-gp, MDR3), ABCB5, and ABCB11 [sister of P-gp, BSEP (bile salt export pump)] from various vertebrates and related invertebrate orthologs (see the table in Fig. 7). However, mussel ABCB cannot be associated with any specific vertebrate ABCB subtype. Thus ABCB1, ABCB4, ABCB5, and ABCB11 and related subtypes show closer relationships with each other than with ABCB from mussel and other invertebrates (Fig. 5).
Identities of mussel ABCC are 27–44% with human ABCC1, ABCC2, ABCC3, ABCC4, ABCC5, ABCC6, and ABCC7 and ABCC orthologs from invertebrates. The extent of identity is greatest with ABCC1, ABCC2, and ABCC3 from human and with MRP-like proteins from invertebrates Drosophila melanogaster and Caenorhabditis elegans (40–44%) (see the table in Fig. 7). This comparatively close relationship of mussel ABCC with a cluster of human ABCC1, ABCC2, and ABCC3 and related invertebrate ABCC orthologs is also indicated in the phylogenetic tree (Fig. 5).
When similarities of different sections of the transporter sequences from different species are compared, degrees of identity are higher for NBDs than for MSDs. For mussel ABCB and human ABCB1, identities are 46% for MSD1, 45% for MSD2, and 64% for NBD1 and NBD2. Comparison of mussel ABCC with human ABCC1 showed identities of 21% for MSD0, 42% for MSD1, 41% for MSD2, 63% for NBD1, and 72% for NBD2.
Immunohistochemical Localization of ABCB in Mussel Gills
For the localization of transporter expression, transverse sections of gill tissue were immunostained with anti-ABCB antibody C219. Mussel ABCB contains the C219 epitope VQEALD; therefore, it is assumed that antibody binding was specific and would indicate the presence of ABCB.
Gills in mussels are lobe-like tissues that comprise filaments, long and thin structures with a dorsal-ventral orientation. Each filament is organized in a frontal and an abfrontal zone with ciliated cells and an intermediate zone with nonciliated endothelial cells (Fig. 6). The ciliated cells produce the water stream, which flows along the surface of the gill tissue. Each filament contains a blood vessel, which in transverse sections of the gills is visible as a large oblong cavity surrounded by a thin layer of the different cell types (Fig. 6) (16).
Antibody staining was observed in the apical membranes of cells in the frontal zone of the gill filaments, which is the part of the tissue directly facing the water flow (Fig. 6). Thus expression of ABCB in mussel gills was concentrated in tissue areas with the highest exposure to the environment.
Using pharmacological inhibitors, data on expression and localization of efflux pumps in the gills, and analyses of transporter cDNA sequences from mussel, we have consolidated information from studies on transporter function in mussel gill tissue. We identified ABC transporter types in the gill, which, on the basis of similarities with xenobiotic transporters from other organisms, are also assumed to act as a cellular defense against toxicants. Furthermore, our data indicate that the MXR defense in mussel gill is mediated by at least two ABC transporter types. Thus, in addition to a P-gp activity, which had been suggested previously, we also found activity and expression of an ABCC/MRP-type transporter. This is also the first report on full-length cDNA sequences of ABC transporter genes in molluscs.
Efflux Transporter Activities in Mussel Gills
Earlier studies had indicated the activity of a P-gp-like protein in bivalve gills. Thus the efflux of rhodamine dyes, which are substrates of this transporter, was sensitive to verapamil, vinblastine, trifluoperazine, emetine, quinidine, forskolin, pentachlorophenol, and cyclosporin A, all known competitive substrates or inhibitors of P-gp (6, 13, 44). Our studies with PSC-833, an effective inhibitor of human ABCB1 (38), extend this list and, with the assumption of specificity of PSC-833 also for mussel ABCB, confirm the hypothesis of P-gp-like transport in mussel gills. Furthermore, we found that efflux of Ca-AM in mussel gills responded to MK-571, a specific inhibitor of ABCC transporters in mammals (14). Given that MK-571 shows similar specificity in mussels, this transporter type is also active in mussel gills. This assumption is also supported by the cloning of highly related genes from mussel tissue (see below) and the augmented inhibition by MK-571 in gill tissue when the PSC-833-sensitive activity had been completely inhibited (Fig. 1) (34).
Calcein accumulation in mussel gills was less sensitive to MK-571 than to PSC-833 (Fig. 1). This finding is similar to results from mammalian cell systems, where the concentration of MK-571 required to exert an effect on MRP activity was also >5 μM, whereas PSC-833 was effective at much lower levels (2). The comparatively high concentrations of MK-571 required to exert an effect on MRP activity probably result from low membrane permeability of the compound (7).
Homologies and Aspects of Substrate Recognition of Mussel Transporters
Consistent with the data from the dye efflux assays, we identified transcripts of two ABC transporters in mussel gills that, on the basis of sequence homologies, are associated with the ABCB and ABCC subfamilies. Similarities of mussel sequences to representatives of those subfamilies, which act as efflux pumps of xenobiotics or other hydrophobic, organic compounds, are particularly high.
Mussel ABCB shows high degrees of identity with ABCB1, ABCB4, and ABCB11, a group of closely related P-gps in vertebrates. Whereas function of mammalian ABCB4 and ABCB11 is mainly restricted to transport processes in hepatocytes (15, 47), ABCB1 recognizes numerous chemically nonrelated compounds as substrates, is expressed in many tissues (5), can confer resistance against a broad array of xenotoxicants as substrates (17), and might also have physiological roles in normal metabolism (23, 42).
Degrees of identity of mussel ABCC were greatest with mammalian ABCC1, ABCC2, and ABCC3, which can be involved in cellular resistance against toxicants (7, 27). In fact, from the ABCC family, ABCC1 and ABCC2 are the most important “genuine” xenobiotic ABCC exporters in vertebrates (27, 33).
On the basis of sequence identities, phylogenetic relationships of the mussel transporters with xenobiotic transporters from other organisms can be clearly established. However, when identities of 40–50% (which reversely imply differences in 50–60% of amino acid positions) are considered, it seems noticeable that respective homologs from different species appear to have similar substrate specificities. As mentioned above, cellular efflux activity in mussel gills acts on numerous substrates and is modulated by inhibitors of mammalian P-gp, and mussel ABCC appears to be sensitive to MK-571, which inhibits mammalian MRPs. As also mentioned above, P-gp- and MRP-type efflux activities, characterized by broad substrate recognition and similar substrate spectra, are found in a wide range of organisms. These functional resemblances of homologous transporters from different organisms seem even more remarkable when we consider that sequences of MSDs, which contain the sites for substrate binding, have comparatively low degrees of conservation.
Similarities in substrate spectra of corresponding xenobiotic transporters from different organisms indicate that there may be evolutionary pressure to conserve functional properties; that is, the toxicants, such as plant or microbial metabolites encountered by the organisms in different environments, may share chemical characteristics.
Closely related homologs with different substrate spectra show that recognition of multiple and similar substrates is not an inherent feature of homologous transporters. For instance, human ABCB4 or MDR3 is a P-gp that shows high (77%) sequence similarity with human ABCB1, whereas ABCB1 is a highly specialized transporter that translocates only certain phospholipids (42).
Significance of Different Transporters and Location of ABCB
Substrate recognition of vertebrate ABCB1 and ABCC1 overlaps, but there are also considerable differences between the substrate spectra of both transporters. ABCC pumps transport conjugated compounds, such as GSH-, glucuronate-, and sulfate-conjugated organic anions; these are not recognized as substrates by the ABCB-type transporters (4). Since there is high activity of phase II metabolism enzymes, such as glutathione S-transferase, in bivalve gills (31, 32), it can be assumed that mussel ABCC functions as an efflux mechanism of resulting conjugated metabolites. Thus coexpression of ABCB and ABCC in the bivalve gills broadens the spectrum of compounds that may be directly effluxed or transported as metabolites. This could be an adaptation to the wide range of natural compounds that might be encountered by mussels (10).
The localized expression of ABCB at the tissue-environment interface in the gill supports its participation as a selective barrier that hinders entrance of xenobiotics into the tissue and also eliminates endogenous metabolic end products. It will be of interest to see if other ABC transporters share a similar location.
ABC Transporters as Part of an Active Barrier System at the Gill-Environment Interface
Besides providing new data on the molecular and functional role of ABC transporters in mussel gills, the present study also enlarges a view of the bivalve gill as an important external barrier between the environment and the organism.
In addition to the active efflux of putatively detrimental molecules, using high-affinity influx transporters, such as Na-dependent amino acid uptake, bivalve gills also function to bring in essential compounds from the environment (48). Thus import and export processes establish a highly selective permeability of the gill tissue, enabling simultaneous accumulation of vital compounds and expulsion of detrimental compounds.
The gill-environment interface in the mussel is then structurally comparable to blood-tissue barriers in vertebrates, such as the blood-brain barrier, which maintain homeostasis of sensitive tissues by controlling influx of nutrients and prevent entry of xenobiotics by ABC transporters (43). An interesting difference is that blood-tissue barriers are, by definition, internal; the mussel gill can now be viewed as a similar tissue barrier, but one that controls entry of nutrients and xenobiotics from the external environment.
Perspectives and Significance
Efflux transporter activity has been described in a wide range of aquatic invertebrate taxa (6, 19, 30, 31, 44, 46), with substrate profile and immunologic identification indicating a P-gp-type activity. As shown in the present study with a marine bivalve, this “typical” P-gp activity is seen, despite considerable interspecies sequence differences of the P-gp gene products. Similarly, other xenobiotic efflux transporters, such as the ABCCs, appear to also be functionally conserved, despite differences in sequences. This may indicate that although the toxicants that are normally encountered by these organisms share similar physicochemical properties, the sequence changes might also have some adaptive value. This could be adaptation to other environmental variables, such as temperature. Different temperatures would result in different membrane lipids (20), with resultant modification of membrane-spanning proteins, and this is where the major variations are seen.
Another difference could be in regulation of titer of efflux transporters. Flexibility in response to toxicants seems to correlate with the probability of encountering environmental toxicants. For example, mussels in organically rich coastal areas and estuaries show higher inherent transporter activity than bivalves from pristine, low-nutrition freshwater environments (44). These questions can be resolved with additional knowledge of the toxicants naturally encountered by these organisms, along with information on membrane composition and relation of gene structure to regulation of gene activity.
This publication was supported in part by the Deutsche Forschungsgemeinschaft, the National Sea Grant College Program of the US Department of Commerce National Oceanic and Atmospheric Administration under grant R/CZ-182 through the California Sea Grant College Program, and the California State Resources Agency.
We are greatly indebted to Dr. Tsung-Han Lee, who provided cryosections of mussel gills and performed the immunostainings. The schematic drawings of mussel gills are based on designs by Chris Patton. We thank Dr. G. Albrecht Luckenbach for comments on the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society