Monocarboxylate-H+ cotransporters, such as monocarboxylate transporter (MCT) SLC16A, have been suggested to mediate transruminal fluxes of short-chain fatty acids, ketone bodies, and lactate. Using an RT-PCR approach, we demonstrate expression of MCT1 (SLC16A1) and MCT2 (SLC16A7) mRNA in isolated bovine rumen epithelium. cDNA sequence from these PCR products combined with overlapping expressed sequence tag data allowed compilation of the complete open reading frames for MCT1 and MCT2. Immunohistochemical localization of MCT1 shows plasma membrane staining in cells of the stratum basale, with intense staining of the basal aspects of the cells. Immunostaining decreased in the cell layers toward the rumen lumen, with weak staining in the stratum spinsoum. Immunostaining in the stratum granulosum and stratum corneum was essentially negative. Since monocarboxylate transport will load the cytosol with acid, expression and location of Na+/H+ exchanger (NHE) family members within the rumen epithelium were determined. RT-PCR demonstrates expression of multiple NHE family members, including NHE1, NHE2, NHE3, and NHE8. In contrast to MCT1, immunostaining showed that NHE1 was predominantly localized to the stratum granulosum, with a progressive decrease toward the stratum basale. NHE2 immunostaining was observed mainly at an intracellular location in the stratum basale, stratum spinosum, and stratum granulosum. Given the anatomic localization of MCT1, NHE1, and NHE2, the mechanism of transruminal short-chain fatty acid, ketone body, and lactate transfer is discussed in relation to a functional model of the rumen epithelium comprising an apical permeability barrier at the stratum granulosum, with a cell syncitium linking the stratum granulosum to the blood-facing stratum basale.
- stratified epithelium
it is well recognized that the rumen of sheep and cattle mediates nutrient uptake in the form of large quantities of short-chain fatty acids (SCFA), such as acetate, propionate, and butyrate, generated by intraruminal microbial fermentation (6, 7, 14, 16, 20, 21, 23). The rumen is covered by a stratified epithelium, which consists of leaflike papillae. These papillae, which may reach a length of 10–15 mm in cattle, greatly increase the absorptive surface area of the rumen (8, 22). Interpretation of transruminal fluxes of isolated rumen mounted in Ussing chambers has relied on a simple mucosal model that incorporates a rumen-facing apical membrane and a blood-facing basolateral membrane (14). However, this does not reflect the stratified structure of the epithelium. From the lumen surface, four distinct cell layers can be distinguished: stratum corneum, stratum granulosum, stratum spinosum, and stratum basale (8, 22). Recently, we confirmed and further developed a three-compartment model (22) for the bovine rumen epithelium from data showing the cellular distribution of key protein elements of epithelial function (8): 1) integral membrane tight junction proteins, such as claudin-1, were associated with stratum granulosum cells, where zonae occludentes were present in electron-microscopic sections, and 2) the Na+-K+-ATPase was present at high density only in cells of the stratum basale, where the highest subcellular density was present at the basal (blood-facing) infoldings adjacent to the basement membrane. A functional syncitium is formed by gap junctions, since cells of the stratum granulosum and stratum spinosum and the apical pole of the stratum basale show strong connexin-47 immunoreactivity (8).
In vivo a considerable SCFA concentration gradient exists between rumen lumen and blood plasma. In vitro studies of SCFA in isolated bovine rumen epithelium show that the absorption mechanism is complex; for SCFA with no concentration gradients, butyrate showed the greatest net absorption, whereas acetate and propionate showed a small net secretion (21). The kinetic characteristics of absorption most likely indicate multiple processes impacting on transruminal fluxes. Passive, nonionic diffusion of these relatively lipophilic substrates is likely to coexist with absorption mediated via specific transporters, such as the monocarboxylate transporter (MCT). In addition, there is considerable intraepithelial metabolism of 14C-labeled SCFA, with 78% and 95% metabolism of propionate and butyrate, respectively, when presented from the rumen lumen and 37–38% metabolism, respectively, when presented from the blood side in glucose-containing media (21). The fate of metabolized SCFA is not solely CO2, since ketone bodies, lactate, and minor components such as amino acids may be labeled (21).
In nonruminants, generation of luminal SCFA from microbial fermentation and transepithelial SCFA absorption is a feature of colonic cells. Cuff et al. (5) reported that MCT1 is primarily responsible for butyrate absorption in human colon cells across the apical brush-border membrane. In Caco-2 cells, butyrate uptake was blocked by α-cyano-4-hydroxycinnamate, propionate, and l-lactate (9). Multiple isoforms of MCT were detected by PCR, but MCT1 was found to be the most abundant isoform by RNase protection assay (9). Antisense treatment for MCT1 also reduced butyrate uptake (9).
Recently, MCT1 expression by RT-PCR and by immunocytochemistry in adult sheep rumen and calf rumen has been reported (12, 16). In cultured cells from sheep rumen, an intracellular acidification following exposure to d-β-hydroxybutyrate, acetoacetate, or lactate, consistent with SCFA-H+ symport, was observed (16). The primary purpose of the present study was to investigate the expression of MCT in adult bovine rumen epithelium and to correlate the cell-specific expression of MCT protein with the specific cell types comprising the rumen epithelium.
Since SCFA may induce an increase in transepithelial Na+ absorption in isolated rumen mounted in Ussing chambers and rumen epithelial cells are known to express high levels of Na+/H+ exchange (NHE) activity (6, 7, 14, 20), cytosolic acid load induced by uptake of SCFA via H+-SCFA transport or nonionic diffusion could provide a functional coupling to increased NHE activity (16). Since NHE5 is expressed only in mammalian brain and sperm and NHE6 and NHE7 are present only intracellularly (26), we investigated expression of NHE1, NHE2, NHE3, NHE4, and NHE8 to determine whether they might maintain acid extrusion in the face of cytosolic acidification. The mucosal model derived from SCFA transport in the colon would predict MCT expression at the “apical” plasma membrane of the stratum granulosum facing the rumen, together with NHE2, NHE3, NHE4, or NHE8, as in other gastrointestinal epithelia (26).
Tissue collection and preparation.
Rumen samples were obtained postmortem from commercial slaughterhouses after humane slaughter in accordance with United Kingdom legislation. Samples were excised from the intestines of cattle within 15–40 min of slaughter. An ∼100-cm2 section of rumen from the ventral sac ∼10 cm from the left longitudinal groove was immediately washed and then transported in ice-cold organ transplantation preservation solution [in mM: 140 sucrose, 42.3 Na2HPO4, and 26.7 NaH2PO4 (pH 7.4)]. Arterial forceps, scissors, and scalpel were used to strip the epithelial mucosae from the muscle layers. For total RNA preparation, samples of rumen were stored at 4°C in RNAlater (Ambion) or snap frozen in liquid nitrogen and stored at −80°C (see below). For immunocytochemistry, full-thickness mucosae or rumen papillae isolated directly using scissors were immersion fixed in −20°C methanol or acetone or ice-cold 3% paraformaldehyde in PBS for ≥4 h or overnight. Tissue/papillae were then incubated for 24 h in 30% sucrose in PBS at 4°C overnight before they were embedded in optimal temperature cutting compound (Tissue-Tek, Miles Laboratories, Naperville, IL) and cryosectioned.
Protein samples were prepared by homogenization of ∼1 g of tissue per 5 ml of homogenization buffer (0.3 M sucrose, 25 mM imidazole, and 1 mM EDTA) at setting 6 on a Polytron (model CH-6010, Kinematica, Kreis-Lucerne, Switzerland). Homogenates were centrifuged at 4,000 g for 10 min at 4°C for removal of nuclei and cell debris. Protein concentration was determined by Bradford assay (3). Homogenates were then solubilized in 4× SDS sample buffer (40 mM Tris·HCl, 4 mM EDTA, 10% SDS, 40% glycerol, 0.04% bromphenol blue, and 8% β-mercaptoethanol) to give a final protein concentration of 0.5 μg/μl.
Slices of kidney material from a single lobe of the multipapillary bovine kidney were transported, and homogenate was processed as described for the rumen.
RNA isolation and RT-PCR.
The acid-phenol extraction reagent RNAzol B (Biogenesis) was used to extract total RNA from bovine rumen epithelium and underlying smooth muscle. Tissue was ground in liquid nitrogen, and RNAzol B (2 ml/0.1 g of tissue) was added. The suspension was homogenized using a Polytron (model CH-6010, Kinematica). Chloroform (0.2 ml/ml of RNAzol) was added, and the sample was shaken vigorously for 15 s before it was incubated on ice for 15 min. After centrifugation at 20,000 g for 30 min at 4°C, the aqueous phase was transferred to a clean tube. An equal volume of isopropanol was added, and the sample was incubated at 4°C for 15 min before centrifugation at 20,000 g for 60 min at 4°C. The solvent was decanted, and the pellet was washed in 85% ethanol before final centrifugation at 20,000 g for 15 min at 4°C. The pellet was air-dried at room temperature and finally dissolved in 10 mM Tris·HCl (pH 7.5). Purity and yield of RNA were assessed by absorbance at 260 and 280 nm.
Omniscript reverse transcriptase (Qiagen) was used according to the manufacturer's protocol for reverse transcription of RNA (2 μg). Negative controls, where reverse transcriptase was omitted from the reaction to discount false-positive results due to DNA contamination, were included for each RNA.
Hot-start Taq DNA polymerase (Qiagen) was used according to the manufacturer's protocol for amplification of 2.5 μl of reverse-transcribed RNA in a 25-μl reaction. For controls, cDNA was replaced with deionized water for each primer pair to discount contamination in the reagents. Mg2+ concentration of 1.5 mM, primer concentration of 0.5 μM, and 1.25 enzyme units were used per reaction. The annealing temperature was 54–60°C, as appropriate for primer pairs. Amplification was carried out over 35 cycles. Two approaches were used for MCT (SLC16) gene family primer design. 1) Before the existence of an annotated bovine genome, Clustal W analysis of human, mouse, and rat sequences were performed, and a dendogram was constructed to place sequences in SLC16 families by homology for human SL16A1–8. Oligonucleotide gene-specific primers, including appropriate redundancies where appropriate, were designed from the published human sequences, and regions of greatest similarity within family groups were chosen wherever possible. Table 1 shows the forward and reverse primers and their predicted product sizes. 2) In the confirmatory approach, partial or complete bovine SLC16A1–8 sequences (Table 1) were used to design alternative bovine-specific PCR SLCA16A1–8 primers. For NHE family members (NHE1, NHE2, NHE3, NHE4, and NHE8), bovine-specific PCR primers were designed from partial or complete bovine NHE sequences (Table 2).
PCR products were analyzed using agarose gel electrophoresis of PCR mixtures, and ethidium bromide fluorescence was used to visualize products. Products for bovine SLC16A1 and SLC16A7 bands were excised from the gel and cloned using the TA cloning vector pCR2.1 TOPO (Invitrogen). The identity of the cloned PCR products was determined by automatic bidirectional sequencing via fluorescent dideoxy dye termination on an automated sequencer (model 377, ABI Prism). At least four separate clones were sequenced for bovine SLC16A1 and SLC16A7. Identity of the cloned PCR fragments was compared with bovine expressed sequence tags (EST) using BLASTN (www.ncbi.nlm.nih.gov) to identify EST that overlap complete full protein coding sequences for bovine SLC16A1 and SLC16A7 (1). Clustal W (www.ebi.ac.uk/clustalw) was used for multiple amino acid sequence comparison with human, mouse, and rat sequences (11). BLAT (http://genome.ucsc.edu/cgi-bin/hgBlat?command=start), from the March 2005 Bos taurus draft assembly (Btau_2.0, Baylor College of Medicine Human Genome Sequencing Center), was used for comparison and confirmation of bovine SLC16A1 and SLC16A7 sequences with the bovine genomic sequence. Positive expression data for bovine SLC16A1 and SLC16A7 by RT-PCR were also confirmed using bovine-specific primers at the 5′ and 3′ regions (A and X primer pairs for SLC16A1 and SLC16A7, respectively; Table 1). Complete bovine SLC16A7 nucleotide sequence data were deposited at the National Center for Biotechnology Information (accession no. DQ387003).
Sections (5 μm) were washed three times with PBS, permeabilized with 0.1% saponin in PBS for 20 min, and then blocked with 10% serum (species dependent on host animal used for secondary antibody production) in PBS. Sections were incubated overnight at 4°C with primary antibody (typically diluted 1:50 in 10% serum-PBS), washed three times, and incubated for 1 h at room temperature with an appropriate Alexa Fluor-conjugated secondary antibody (Molecular Probes). Typically, sections were counterstained with ethidium homodimer I (Molecular Probes) to highlight the nuclei. Appropriate controls were obtained by omission of primary antibody or by preabsorption of primary antibody by peptide immunogen according to the manufacturer's data sheets. Immunofluorescence was detected by confocal laser scanning microscopy (model TCS-NT, Leica, with Kr-Ar laser). All immunocytochemistry was repeated for material obtained from at least three animals.
Protein samples were loaded at 10–20 μg protein per well on 4–20% polyacrylamide gels (Criterion, Bio-Rad). After electrophoresis, proteins were transferred to 0.45-μm nitrocellulose membranes and blocked with 5% milk powder in PBS with 0.1% Tween 20 (Blotto) for 1 h at room temperature. Membranes were probed with primary antibodies overnight at 4°C, washed three times with PBS-0.1% Tween 20, and then incubated for 1 h at room temperature with an appropriate horseradish peroxidase-labeled secondary antibody. The membranes were washed three more times, and signals were detected by enhanced chemiluminescence. Western analysis was representative of data from at least three animals.
Rabbit polyclonal antibodies against MCT1 SLC16A1 (MCT13A, affinity-purified IgG) and MCT2 SLC16A7 (MCT22A, affinity-purified IgG) were obtained from Alpha Diagnostics (San Antonio, TX). Antibodies were raised against a KHL-coupled 15-amino acid peptide sequence derived from the COOH terminus of the rat sequences for MCT1 and within the COOH terminus for human MCT2. The MCT1 and MCT2 peptides were obtained to act as a blocking peptide to test for antibody specificity. For MCT1, comparison of the rat COOH-terminal sequence with the bovine sequence gives 7 of 15 identities. In addition to MCT13A, immunocytochemistry for MCT1 in bovine rumen was also performed with MCT12A (Alpha Diagnostics), a rabbit purified IgG against a human COOH-terminal 19-mer peptide giving 14 of 19 identities between the bovine and human sequences. For MCT2, the exact sequence was not available (see results). For NHE1 and NHE2, rabbit polyclonal antibodies (NHE11A and NHE21A affinity-purified IgG, respectively) were obtained from Alpha Diagnostics. NHE11A and NHE21A were raised against rat COOH-terminal cytoplasmic 22- and 20-amino acid peptide sequences, respectively, which were also used as a blocking peptide for antibody specificity in immunohistochemistry. For NHE11A and NHE21A, the COOH-terminal sequence is conserved across species [human, mouse, rat, pig, and rabbit, including bovine NHE1 (GI:2498047) for NHE11A, and human and mouse for NHE21A; see manufacturer's data sheet].
A rabbit polyclonal antibody against CD147 (catalog no. sc13976, Santa Cruz Biotechnology, distributed by Autogen Bioclear, Wiltshire, UK) was raised against a recombinant protein comprising the first 200 amino acids of the NH2 terminus of human CD147.
For immunocytochemistry, Alexa Fluor (Molecular Probes)-conjugated secondary antibodies were used at a dilution of 1:20–1:50. For Western blotting, horseradish peroxidase-conjugated antibodies were used at a dilution of 1:100,000 (goat anti-rabbit IgG; Sigma) or 1:250,000 [goat anti-mouse IgG (Sigma) and donkey anti-goat IgG (Santa Cruz Biotechnology)].
In Fig. 1, A and B, bovine rumen epithelial or smooth muscle cDNA was used as template to demonstrate PCR products corresponding to SLC16A1 and SLC16A7 (MCT1 and MCT2). All other primer pairs were negative (not shown). This pattern of SLC16A expression (MCT1 and MCT7 positive; MCT2, MCT3, MCT4, MCT5, MCT6, and MCT8 negative) was confirmed using bovine-specific primer pairs (Table 1, Fig. 1D). Cloning and bidirectional sequencing of multiple clones (≥4) of the PCR products shown in Fig. 1A by basic local alignment search tool (BLAST) (1) confirm that the identities of the ∼843- and ∼588-bp products are bovine SLC16A1 and SLC16A7, respectively (86% and 83% identical, respectively, at the nucleotide level with their human homologs). Figures 2 and 3 show multiple alignments for the bovine PCR products and human, mouse, and rat SLC16A1 and SLC16A7 amino acid sequences. It is clear that the bovine amino acid sequences for SLC16A1 and SLC16A7 PCR products are similar to their human homologs (84% and 81% identical, respectively). BLAST search of the EST databases with the PCR product sequence revealed six bovine EST clones [CK778726 and BE479063 (extended by sequencing of the insert) and CB458284 for SLC17A1; CF931033, CN791524, and CK976735 for SLC16A7], with identities that partially overlap the cloned bovine PCR-derived partial sequences or each other. Contigs comprising cDNAs for the complete protein coding sequences for bovine SLC16A1 and SLC16A7 (Figs. 2 and 3) were confirmed by BLAST search of bovine genomic sequence data. Sequence-specific primer pairs were designed for the 3′ and 5′ regions of the cDNA for bovine SLC16A1 and SLC16A7 with appropriate-sized RT-PCR products observed from bovine rumen cDNA (Fig. 1C). The SLC16A1 sequence is identical to accession no. NM_001037319 (National Institutes of Health Mammalian Genome Collection) derived from bovine colon (12). The SLC16A7 sequence has been submitted to GenBank (accession no. DQ387003).
Complete primary sequences of bovine SLC16A1 and SLC16A7 are compared with human mouse and rat protein sequences in Figs. 2 and 3. The amino acid sequence was 88% identical to the human sequence for bovine SLC16A1 and 78% identical to the human sequence for bovine SLC16A7. The area of greatest divergence resides in the large cytoplasmic loop between transmembrane (TM) segments 6 and 7 for SLC16A1 and SLC16A7. The majority of the amino acid substitutions are conservative. Comparison of the bovine COOH terminus with human, rat, and mouse sequences is important, since specific MCT antibodies have been raised against COOH-terminal peptides (Figs. 2 and 3). For SLC16A1, there is conservation of the terminal four amino acid residues between species (ESPV); however, there is poor homology between MCT2 bovine and rat/mouse sequences and a better match to human MCT2 (SERETNI; Fig. 3).
MCT1 immunostaining of bovine rumen epithelium with the MCT13A antibody is shown in Fig. 4. Staining was marked at the plasma membrane of stratum basale cells, with concentration at the blood-facing side. The membrane infoldings at the basal pole of the stratum basale cells give rise to an abundance of Na+-K+-ATPase at this location (8). MCT1 immunostaining is diffuse in the overlying cells of the stratum spinosum and stratum granulosum but negative in the stratum corneum. The specificity of immunostaining was confirmed by peptide preabsorption of the MCT13A antibody, which resulted in negative staining in all cell layers. Similar results were obtained using MCT12A antibody (see methods). Despite a discernable protein band at ∼45 kDa in the Western blot for MCT2, only weak MCT2 immunostaining was discernable in the epithelium; staining was diffuse in all epithelial layers below the stratum corneum (not shown).
Since plasma membrane localization and function of MCT1 are dependent on a tight association with CD147 (also known as OX-47 and basigin) (13), we determined the expression and localization of CD147 by immunostaining. CD147 immunostaining (Fig. 5) is present mainly in the cells of the stratum basale, with less expression in the stratum spinosum and stratum granulosum. Immunostaining in the stratum corneum was negative. The expression and localization of MCT1 and CD147 mainly in stratum basale cells are inconsistent with the mucosal model for SCFA permeation envisaged for transmucosal transport in the colon of nonruminants.
Next, using RT-PCR, we investigated the expression of NHE family members in bovine rumen. Figure 6 shows expression of NHE1, NHE2, NHE3, and NHE8 in bovine rumen. In Western blots using the NHE1-specific antibody (Fig. 7), two polypeptides (∼90 and 110 kDa) are recognized in bovine tissue homogenates. Whereas the ∼90-kDa protein predominates in bovine kidney protein homogenates, the 110-kDa protein predominates in rumen epithelium (Fig. 7). Immunocytochemistry demonstrates plasma membrane staining of cells in the stratum basale and stratum spinosum and the most intense staining at the rumen lumen-facing side of the stratum granulosum cells (Fig. 7). Thus NHE1 is concentrated at the cellular location of the morphological permeability barrier within the stratified rumen epithelium. In contrast, NHE2 immunostaining is present mainly in an intracellular location in stratum basale cells but is also present in overlying cells of the stratum spinosum and stratum granulosum (Fig. 8).
The present data showing that SLC16A1 (MCT1) and SLC16A7 (MCT2) are expressed in bovine rumen epithelium are consistent with rumen mediation of SCFA absorption. These present data confirm and extend the observation that MCT1 is expressed in sheep rumen epithelium (16) and calf intestine (12). Sequence analysis confirms the identities of RT-PCR products as MCT1 and MCT2. Furthermore, using overlapping EST data combined with direct sequencing of BE479063, we were able to build the complete cDNA sequences for the bovine proteins. The monocarboxylate-proton cotransporter family, including MCT1 and MCT2, possesses a plasma membrane topography of 12 TM segments (10). In addition, a number of conserved motifs and residues have been identified in the MCT family. Two motifs have been identified as being highly conserved (10): [D/E] G [G/S] [W/F] [G/A]W appears as DGGWGW in the bovine MCT2 sequence, and YF-K[R/K][R/L]-LA-[G/A]-A-AG at the beginning of helix 5 appears critical for MCT1 function. This latter motif (indicated in bold) is, as expected, retained in the bovine MCT1 sequence (yfykrrplangLAMAG) and in the MCT2 sequence (yfynkrpvangLAMAG) with an asparagine at position 4 and a valine at position 8. Compared with the human sequence, TM 8 for the bovine sequence is conserved; thus critical residues (e.g., D302 and R313 in the human sequence) involved in substrate binding/recognition are unaltered (10). Hydrophobic residues in TM 10 may alter substrate specificity; e.g., F362 of bovine MCT1 matches the human, rather than the mouse or rat, sequence (V355). The two extracellular-facing lysine residues in the loop between helix 7 and helix 8, which render MCT1 resistant to covalent DIDS modification, are also retained in the bovine sequence (10).
For immunolocalization studies of MCT1 expression in sheep rumen, Muller et al. (16) employed an MCT1-specific antibody that used the COOH-terminal region of rat MCT1 as immunogen (LQNSSGDPAEEESPV), similar to the MCT1-specific antibody used in the present study. Comparison of the rat COOH-terminal sequence with the bovine sequence gives 7 of 15 identities (Fig. 2), and the localization of MCT1 is consistent with that in sheep rumen, with immunostaining concentrated in the cells of the stratum basale toward the basal (blood-facing) pole. The distribution is also very similar to that described for MCT1 in the rumen of young calves (12) and is therefore not altered when rumen fermentation is established. The distribution of MCT1 reported here is very similar to that previously observed for the Na+-K+-ATPase (8). Basal infoldings observed by electron microscopy serve to amplify membrane area in this region of the stratum basale cell; therefore, membrane transporter proteins such as MCT1 and the Na+-K+-ATPase appear concentrated in this area. We have used an anti-human MCT2 antibody, inasmuch as homology is closest between bovine and human MCT2. Rat and mouse MCT2 are dissimilar within the COOH terminus compared with the bovine sequence. Compared with MCT1, MCT2 immunostaining was weak and diffuse in all epithelial cell layers, except the stratum corneum.
CD147 is a member of the IgG superfamily but is membrane bound by a single conserved TM span (17). CD147 shows extensive alternative splicing, with recognition of multiple transcripts (17). CD147 is tightly associated with MCT1 and is important for traffic of MCT1 and MCT4, but not MCT2, to the cell surface (13). The MCT1-CD147 association is seen in several epithelial cell types, including retinal pigment and thyroid epithelial cells (17, 19, 25). The distribution of CD147 was investigated in rumen epithelium and compared with that of MCT1; both are concentrated in stratum basale cells; however, in addition to the plasma membrane staining for MCT1, there is extensive cytoplasmic immunostaining for CD147.
Taken together, the present data show that although MCT1 and MCT2 are expressed by rumen epithelium, neither MCT1, CD147, nor MCT2 is concentrated at the outer permeability barrier of the stratum granulosum. Although we cannot eliminate the possibility that MCT2- and, therefore, SCFA-H+ symport contributes to uptake of SCFA across the outer barrier, it will comprise only a minor fraction compared with nonionic diffusion (see below).
The presence of MCT1 on the blood-facing membrane of stratum basale rumen epithelial cells must be correlated to the physiology of the organ. As discussed by Halestrap and Meredith (10), the general role of MCT1 is uptake or release of lactate from anoxia-exposed cells to maintain glycolysis, lactate uptake for gluconeogenesis and lipogenesis, or ketone body uptake for oxidative phosphorylation. In organs where multiple MCT isoforms are expressed, separate cells may express a different MCT isoform (e.g., brain glia and neurons or red and white muscle fibers). Thus metabolic cooperation and lactate cycling are established (10). Since we have demonstrated MCT2 expression in underlying smooth muscle, the existence of a local metabolic cooperation between cells containing a high mitochondrial density in the stratum basale and underlying connective tissue/smooth muscle in the rumen wall cannot be excluded. However, the main function envisaged for MCT in epithelial tissues, including the rumen, is nutrient (SCFA and lactate) absorption from microbial fermentation in the lumen. In colonic epithelial cells in nonruminants, MCT1 is thought to be responsible for SCFA and ketone body uptake across the apical membrane (5). The present data show that MCT1 protein expression in the stratum corneum and stratum granulosum (the site of the likely permeability barrier) is negative, thus excluding the possibility that MCT1 functions in SCFA/lactate uptake at the apical outer-facing membrane of rumen. What then is the mechanism of uptake of SCFA such as acetate, butyrate, and propionate at this outer-facing barrier? Transepithelial measurements of lactate transport in bovine epithelium indicate that, at >2 mM SCFA, transepithelial transport displays first-order kinetics and that net (active) flux is essentially zero (15). In nongradient conditions, acetate and propionate showed a small net secretion, whereas butyrate showed net absorption (20, 21). Thus a secondary active process for SCFA cannot be entirely excluded. At the concentrations of SCFA and lactate used in these in vitro measurements, MCT function would be saturated. However, with the likely concentration gradient for SCFA established between the rumen and blood in vivo [and with concentrated feeds that contain digestible carbohydrates generating high (up to 40 mM) rumen lactate in severe acidosis (18)], cellular uptake into the stratum granulosum and deeper layers may proceed by diffusion/nonionic diffusion (23). Given the location of MCT1 at the basal pole of stratum basale cells in bovine and ovine rumen, we hypothesize that MCT1 will mediate lactate, SCFA, and ketone body exit to blood (16). Further studies are required to determine the exact contribution of MCT2 and whether alternative transporters that may mediate SCFA-anion exchange or Na+-dependent uptake are expressed in bovine rumen and localized to the stratum granulosum.
Uptake of SCFA into cultured rumen cells is associated with a decrease in cytosolic pH (16). Nonionic diffusion (permeation of the conjugate acid SCFA) across the outer barrier would deliver protons to the cytosol of these cells. The stimulation of net Na+ transport across rumen epithelium by SCFA indicates that an NHE is present at the outer-facing membrane (6, 14, 20, 21). In bovine rumen epithelium, the SCFA-mediated increase in short-circuit current is sensitive to amiloride inhibition, consistent with this model (6). Accordingly, we have investigated the cellular expression of NHE family members in rumen epithelium. RT-PCR indicates expression of NHE1, NHE2, NHE3, and NHE8. Western blotting of bovine kidney and rumen samples for NHE1 indicates two protein bands of ∼90 and 110 kDa: the ∼110-kDa band predominates in the rumen and the ∼90-kDa band predominates in bovine kidney. According to genomic analysis, the human NHE1 gene produces, by alternative splicing, seven types of transcripts, which are predicted to encode seven distinct proteins (24). However, the major transcripts encode ∼93.5-kDa proteins of 841 amino acid residues. In intact tissues, glycosylation of this core protein results in an apparent ∼110-kDa molecule. Coupaye-Gerrard et al. (4) hypothesize that glycosylation determines delivery of mature NHE1 to its final location at the basolateral membrane in A6 cells. In the rumen epithelium, the distribution of NHE1 confirms expression concentrated at the outer-facing membrane of cells associated with the formation of the permeability barrier, namely, the stratum granulosum. In contrast, NHE2 is present mainly in a cytosolic location, in the cells of the stratum basale, stratum spinosum, and stratum granulosum. In mouse skin, which has a cellular architecture similar to that of the rumen, NHE1 is expressed at the stratum granulosum (2), where it controls the acidification of extracellular microdomains as measured directly by fluorescent half-time imaging of extracellular pH indicators (2). Thus, despite the near-neutral pH of the skin surface of the stratum corneum, NHE1 maintains local extracellular acidity adjacent to the stratum granulosum (2). Use of NHE1−/− animals confirms the central role of NHE1 in generation of the localized extracellular acidic pH (2). In these animals, disruption of the permeability barrier shows that NHE1 function is essential in maintaining normal skin integrity (2). In the rumen, SCFA absorption is not markedly influenced by bulk rumen pH, but the existence of a local area of acidity maintained by the presence of NHE1 at the stratum granulosum would facilitate formation of nonionized SCFA and promote diffusional influx into the stratum granulosum. Since the cells of the stratum granulosum are functionally coupled to those of the stratum spinosum and stratum basale, where MCT1 is localized, MCT1 would mediate exit of SCFA or metabolites to blood. The presence of NHE2 and other NHEs in recycling endosomes within stratum basale cells could form a functional reserve (26), together with other NHEs, to facilitate pH homeostasis in the presence of accumulation of SCFA and cytosolic acidification due to nonionic diffusion from the rumen lumen.
In conclusion, the present study provides morphological data to support a three-compartment model for transcellular transport of SCFA and lactate absorption by rumen epithelium (7) in which nonionic diffusion is the primary mechanism for SCFA uptake at the outer permeability barrier of the stratum granulosum. We have demonstrated expression of NHE1 at the apical stratum granulosum and propose that this NHE1 expression generates localized extracellular acidity to promote nonionic diffusional uptake of SCFA to the cytosol of these cells. The imposed H+ load and cytosolic acidification will further activate NHE1. Protons will shuttle across the outer-facing membrane of the stratum granulosum at the expense of Na+ influx. Diffusional flow of SCFA through the functional syncitium will deliver SCFA to the mitochondria-rich stratum basale cells, where metabolism (oxidation and formation of ketone bodies) may occur, with exit to blood being mediated via MCT1. Since cotransport of H+ will occur via MCT1, net acid extrusion from the system may occur and, therefore, contribute to regulation of intracellular pH. Further studies of pH-homeostatic mechanisms, especially in response to cytosolic alkalinization, are required.
This work was funded by Pfizer Global Research.
We thank Dr. K. Gration, who was instrumental in initiation of this project, and Dr. J. Aschenbach for helpful discussions on rumen function.
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 © 2007 the American Physiological Society