Reexamination of the Akp3−/− mouse intestine showed that, despite the lack of intestinal alkaline phosphatase (IAP), the Akp3−/− gut still had considerable alkaline phosphatase (AP) activity in the duodenum and ileum. This activity is due to the expression of a novel murine Akp6 gene that encodes an IAP isozyme expressed in the gut in a global manner (gIAP) as opposed to duodenum-specific IAP (dIAP) isozyme encoded by the Akp3 gene. Phylogenetically, gIAP is similar to the rat IAP I isozyme. Kinetically, gIAP displays a 5.7-fold reduction in catalytic rate constant (kcat) and a 30% drop in Km, leading to a 4-fold reduction kcat/Km compared with dIAP, and these changes in enzymatic properties can all be attributed to a crucial R317Q substitution. Western and Northern blot analyses document the expression of Akp6 in the gut, from the duodenum to the ileum, and it is upregulated in the jejunum and ileum of Akp3−/− mice. Developmentally, Akp3 expression is turned on during postnatal days 13–15 and exclusively in the duodenum, whereas Akp6 and Akp5 are expressed from birth throughout the gut with enhanced expression at weaning. Posttranslational modifications of gIAP have a pronounced effect on its catalytic properties. Given the low catalytic efficiency of gIAP, its upregulation during fat feeding, its sequence similarity with rat IAP I, and the fact that rat IAP I has been implicated in the upregulation of surfactant-like particles during fat intake, it appears likely that gIAP may have a role in mediating the accelerated fatty acid intake observed in Akp3−/− mice fed a high-fat diet.
- alkaline phosphatases
- site-directed mutagenesis
- computer modeling
- fat absorption
- intestinal physiology
intestinal alkaline phosphatase (IAP), as the name implies, is expressed in the small intestine of many species. Lymph and serum levels of IAP increase after a fatty meal and the lymphatic triglyceride output is directly correlated with the lymphatic IAP output (13, 23), whereas IAP levels are dramatically decreased upon starvation (16). IAP is found associated with the brush border of the intestinal epithelium and enriched in surfactant-like particles (SLP) (1, 2, 8, 9, 14). After fat feeding, an IAP-containing membrane surrounds the lipid droplet in the enterocyte (36, 37). These findings suggest a role for SLP in fat absorption involving the induction of IAP by fat, which in turn results in IAP-induced SLP secretion.
In humans, alkaline phosphatases (APs) are encoded by four genes traditionally named after the tissues where they are predominantly expressed; the tissue-nonspecific AP (TNAP) gene (ALPL), located on chromosome 1, is expressed at highest levels in liver, bone, and kidney (hence the alternative name “L/B/K” AP), in the placenta during the first trimester of pregnancy, and at lower levels in numerous other tissues. The other three isozymes, i.e., placental (PLAP), placental-like or germ cell (GCAP), and IAP, show a much more restricted tissue expression; hence the general term tissue-specific APs. These isozymes are encoded by three genes (ALPP, ALPP2, and ALPI, respectively) clustered on human chromosome 2, bands q34–q37 (for review, see Ref. 24) and are closely related to one another, showing 90 and 87% identical nucleotide and amino acid sequences, respectively. The orthologous TNAP gene in mice is called Akp2 and is located on mouse chromosome 4. The mouse Akp3 gene encodes the IAP isozyme and the mouse Akp5 gene encodes the embryonic AP (EAP) isozyme that appears to be related to both the human PLAP and GCAP isozymes. In addition, mice have an inactive pseudo-AP gene Akp-ps1 (21). Although the human and mouse genomes appear to harbor a single IAP gene, the rat genome encodes two IAP genes, i.e., Alpi1 and Alpi2 (11, 20, 29), that produce mRNAs of 2.7 kb (IAP I) and 3.0 kb (IAP II) (10). Bos taurus in turn have an unprecedented level of complexity in their IAP genes since the bovine gut has been shown to coexpress up to seven IAP-like genes (22, 33).
Inactivation of the mouse IAP gene (Akp3) gives rise to apparently healthy mice under routine laboratory conditions. However, challenging these Akp3−/− mice with a long-term high-fat diet or with short-term administration of corn oil leads to an accelerated rate of fat absorption and elevation of plasma triglycerides, and these mice gain weight faster than wild-type (WT) littermate controls (26). These data indicate that IAP has a functional role in a rate-limiting step during fat absorption. Interestingly, we found residual AP activity in the gut of Akp3−/− mice (25). This activity was not due to expression of either the Akp2 or Akp5 genes. A search for homologous genes in the Mouse Genome databases revealed the presence of a new locus that appears to encode an IAP-like gene that we have named Akp6. The mouse Akp-ps1, Akp5, Akp6, and Akp3 loci, in this spatial order, are clustered in the same region of mouse chromosome 1. Here, we report that the Akp6 encodes an AP isozyme expressed through the small intestine, which we have named global IAP (gIAP) to distinguish it from the product of the Akp3 gene that is expressed in a duodenum-specific manner (dIAP). Although gIAP is catalytically fivefold less efficient that dIAP, it is upregulated in the gut of Akp3−/− mice during high-fat feeding.
MATERIALS AND METHODS
To obtain a full length Akp6 cDNA, total RNA isolated from the small intestine of Akp3−/− mice was reverse transcribed, and a 2.3-kb fragment was amplified by PCR using primers (forward primer: 5′-ATGCAGGGAGACTGGGTGCTGCTGTTGT-3′, reverse primer: 5′-TAGCCTTCTGGAAATCAGTTGGTACCTCC-3′) chosen from the sequence XM_129951, which was predicted from the annotated genomic sequence (NT_039173) by the GNOMON gene prediction algorithm. The PCR reactions were carried out using platinum Pfx DNA polymerase (Invitrogen, Carlsbad, CA) and Taq Extender PCR additive (Stratagene, La Jolla, CA). After confirming the entire sequence, we subcloned the 1.6-kb coding region of the 2.0 kb into an mammalian expression vector pCMV Script to express native Akp6 protein (gIAP) in vitro. To express a FLAG-tagged soluble form of gIAP, a DNA sequence encoding the FLAG peptide and a stop codon was added after the corresponding site of codon 525 (7). We used a forward (5′-CTCAGGCACACCTGGTGCACG G-3′) and a reverse primer (5′-TGCAAGGAGGGCCCTTA CTT GTC ATC GTC GTC CTT GTA GTC TGTGGTGGTGG-3′) to introduce an ApaI restriction site (italic), a stop codon, and the FLAG epitope (underlined).
To obtain a coding region of Akp3 cDNA, total RNA isolated from small intestine of WT mice was reverse transcribed, and a 1.6-kb fragment was amplified by PCR using primers chosen form the cDNA sequence NM_007432, which is derived from a genomic sequence, M61705 (forward primer: 5′-CCAGCCATGCAGGGACCCTGGGTGCTGC-3′; reverse primer: 5′-TAGATGTGCTGGAGTTTAGGACTCCGCA-3′). The sequence of the amplified 1.6-kb fragment was confirmed and the fragment was subcloned into the pCMV Script to express native Akp3 protein (dIAP). A DNA sequence encoding the FLAG peptide and a stop codon was added after the corresponding site of codon 532 by using the 5′-GAGCCAGCGGGCCCTACTTATCGTCATCGTCTTTGTAGTCGTTATGGACC-3′ oligonucleotide as forward primer and the 5′-GCCAGCCAGCTCACTAGTGAACGC-3′ as reverse primer to introduce an ApaI site (italic), a stop codon, and a FLAG epitope (underlined).
The full-length Akp5 cDNA, BC052874, was available as an IMAGE clone, 30033693, and was purchased from ATCC (Manassas, VA). The 2.1-kb insert was subcloned into the pCMV Script to express native Akp5 protein (EAP), and a DNA sequence encoding the FLAG peptide and a stop codon was added after the corresponding site of codon 506 using 5′-TGCCAACAGGTACCGCAAGGTTCACTTGTCGTCATCGTCCTTGTAATCGCTCTGGCCA-3′ as forward primer and 5′-GGCTGATGTCACTGAGGAAGAGAG-3′a 3′ as reverse primer to introduce a KpnI site (italic), a stop codon, and a FLAG epitope (underlined).
Expression of recombinant enzymes.
Each expression vector was transfected into COS-1 cells by standard electroporation, and after 24 h the FCS-supplemented culture medium was replaced by serum-free medium, OptiMax I (Invitrogen, Carlsbad, CA). Supernatant was collected after 60 h of further culture and concentrated ∼15 times with Centriplus (Millipore, Bedford, MA). We also transfected Chinese hamster ovary cells with the expression vectors and obtained clones stably expressing each of the mouse AP genes after 2 wk of selection with 400 μg/ml G418.
All kinetic measurements were performed in triplicate at 25°C. For the measurement of relative catalytic activities, microtiter plates were coated with the M2 antibody (depending on the goals, ranging from 0.2 to 1 μg/ml). For the calculation of catalytic rate constants kcat, we used PLAP-FLAG with known kcat = 460 s−1 as a reference for each microtiter plate. Recombinant APs were incubated with the plates for 2 h at room temperature, after which plates were washed with Tris-buffered saline, containing 0.008% Tween 80. The AP activity of bound enzymes or enzyme mutants was then measured at 405 nm as a function of time, using p-nitrophenylphosphate (pNPP, 0.05. 20 mM) as a substrate, either at pH 9.8 in 1.0 M diethanolamine buffer, containing 1 mM MgCl2 and 20 μM ZnCl2 or at pH 7.5 in 1 M Tris·HCl, containing 1 mM MgCl2 and 20 μM ZnCl2, as indicated. Recordings for kinetic computation were selected from those parts of the curve, where absorbance at 405 nm (A405 nm) vs. time was linear and reaction rates were recorded as A405 nm/time, taking into account an extinction coefficient of 10.08 × 103 M−1 cm−1 for the reaction product p-nitrophenol. Thus initial rates were calculated over a time interval up to 2 h, excluding the first 5 min from evaluation, because of non-steady-state conditions. Data were linearized in 1/velocity vs. substrate. Michaelis-Menten concentration plots and linear regression were performed, calculating slopes ± SD and intercepts ± SD in GraphPad Prism version 3.0a (GraphPad Software, San Diego, CA).
Protein structure analysis and alignment using Laser Gene software indicated that peptide 208–231 from the deduced Akp6 sequence was potentially antigenic due primarily to its location on the 3D model of the enzyme and also based on the number of dissimilarities with the corresponding sequences of the Akp3 and Akp5 cDNAs. Two rabbits were immunized with the corresponding gIAP synthetic peptide, and the antiserum was affinity purified on an affinity column using solid-phase peptide. Peptide synthesis, immunization, and affinity purification were processed by QCB (Hopkinton, MA).
Tissue collection, histology, and immunohistochemistry.
All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of the Burnham Institute for Medical Research (Animal Usage Form # 06–106, approved on 6/23/2006). The production and characterization of Akp3−/− mice has been described previously (26). The intestine was fixed with 10% buffered formalin, washed with increasing sucrose concentrations, and embedded in optimal cutting temperature compound as described (26). Serial sections (8 μm-thick) were used for immunohistochemistry. Alternatively, some of the formalin-fixed tissues were processed through standard protocols for paraffin sections and used for hematoxylin and eosin staining. To examine IAP expression, a Vectastain ABC Kit (Vector Laboratories, Burlingame, CA) was used according to the manufacturers' instruction. The antisera were diluted in blocking solution and incubated with the tissue for 2 h at room temperature. Normal rabbit serum was used as the negative control.
Western blot analysis.
Mouse tissues were homogenized in lysis buffer (150 mM NaCl, 1% NP-40, 0.1% SDS, 50 mM Tris·HCl, pH 7.4), supplemented with 1 mM PMSF and 1% proteinase inhibitor cocktail (Sigma, St. Louis, MO). Protein concentration was determined by using the BCA assay kit (Pierce, Rockford, IL). Protein samples were denatured in 2% SDS and 0.025% β-mercaptoethanol and loaded on 8–16% acrylamide Tris-glycine gel (Invitrogen, Carlsbad, CA). Gels were transferred to Optitran nitrocellulose membranes (Schleicher & Shuell, Dassel, Germany), and blotted membranes were stained with Ponceau S to ensure efficiency of transfer. Washed membranes were blocked with SuperBlock reagent (Pierce) prior to incubation with primary antibodies, which were raised against peptide 208–231 of Akp5 or Akp6 in rabbits. Peroxidase-labeled goat anti-rabbit IgG antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was used as secondary antibody, and a positive reaction was visualized by ECL+ plus kit (Amersham Pharmacia, Piscataway, NJ). Since standard control genes, such as β-actin, GAPDH, or cytochrome c, showed a segment-specific variability of expression in the small intestine, we used endogenous mouse IgG as internal control for protein loading. In case of recombinant enzyme samples derived from COS-1 or CHO cells, a mouse monoclonal anti-FLAG antibody (Sigma) was used as primary antibody and peroxidase-labeled goat anti-mouse IgG as second antibody (Calbiochem, San Diego, CA).
Postnatal mouse small intestines were divided into four segments and homogenized in Tris·HCl buffer (pH 8.9) containing 0.1% Triton X-100. the homogenates were emulsified with an equal volume of n-butyl alcohol, and the water phase was separated from the butyl alcohol phase after repeated centrifugation. Protein concentration of the extracts was measured by BCA assay kit, and samples containing 1.5 mg/ml protein were incubated on 96-well ELISA plates coated with anti-gIAP antibody. The wells were saturated with recombinant gIAP protein and incubated with serial concentrations of pNPP (20, 10, 5, 2.5, 1.25, and 0 mM), and the kinetic reaction was measured at A405 nm via a microplate reader (Molecular Devices, Sunnyvale, CA) to calculate Km values. For endo-β-galactosidase treatment, butanol extracts of intestinal segment 1 from 2-day-old mice (1.5 mg/ml protein × 50 μl per well) were incubated on 96-well plate coated with the anti-gIAP antibody to trap gIAP. After washing, the wells were treated with 0.003 units of endo-β-galactosidase from Escherichia freundii (Seikagaku) overnight, followed by washing and enzyme immunoassay.
Low- and high-fat diets.
Gender-, age-, and genotype-matched mice (3–5 in each group) were fed either a 4.5% or an 11% fat diet (mouse chow no. 5015, Purina Mills, St. Louis, MO) as previously reported (26). The high-fat diet was given to the Akp3−/− mice and littermate WT control mice for 2 mo. Adult mice were also fed 500 μl of corn oil by gavage following an overnight fast. The intestines were immediately immersed in RNAlater (Ambion, Austin, TX), divided into segments, which were opened up to remove the ingesta, and processed for RNA isolation.
Northern blot analysis.
Total RNA was isolated by a standard protocol (5). Ten micrograms of RNA was loaded in each lane. One set of samples was used for ethidium bromide staining. The other set of samples was used for blotting. Specific probes were prepared from 3′ UTR region of Akp6: a 284-bp ApaI-KpnI fragment and Akp3: a 236-bp BamHI fragment. The entire Akp5 cDNA was used as a probe owing to relatively low level of Akp5 expression compared with Akp3 and Akp6. The specificity of each probe was tested by Southern blot hybridization.
Identification of Akp6 at the genomic and mRNA level.
Previous data (26) had shown that Akp3−/− mice had a normal birth rate, normal life span, and no obvious abnormalities unless challenged with a high-fat diet, either via forced fat loading by gavage or long-term high-fat feeding, in which case the mice became obese as a consequence of an accelerated transport of fatty acids through the gut. Reexamination of the Akp3−/− mouse intestine showed that, despite the lack of IAP, the Akp3−/− gut still had considerable AP activity in duodenum and ileum (25). Upon examining the expression of the known AP genes along the length of the small intestine by RT-PCR, we found no evidence of upregulation of the Akp2 or Akp5 genes or of expression of the Akp-ps1 pseudogene (21). Similarly, we examined the intestines of the Akp5−/− mice and found no upregulation of Akp2, Akp3, or the Akp-ps1 pseudogene (not shown).
Akp6 and Akp3 sequence comparison.
Therefore, we searched for the possible existence of a novel AP gene that might explain the persistent AP activity in duodenum and the upregulated activity in the ileum. We used the BLAT tool at the http://genome.ucsc.edu to query the mouse genome database, since this software is designed to quickly find sequences of 95% and greater similarity of length 40 bases or more and DNA BLAT works by keeping an index of the entire genome in memory. Using the mouse Akp3-derived protein sequence as query, this search revealed the presence of a predicted new locus, based on spliced mouse ESTs deposited by the RIKEN Consortium (XM_129951 RIKEN Database) and also by The Genome Institute of Novartis Research Foundation using the Affimetrix U74B chip. We have named this new locus Akp6 and the isozyme it encodes gIAP to reflect its nonsegmental, generalized expression in the gut and to distinguish it from the dIAP encoded by the Akp3 gene and expressed exclusively in the duodenum, as we demonstrate in this paper. As shown on Fig. 1, all four mouse tissue-specific AP loci (Akp3, Akp5, Akp6, and the Akp-ps1 pseudogene) are closely linked at the 1C5 site in chromosome 1, with each gene occupying ∼3.5 kb and composed of 11 exons and 10 introns. In contrast, the Akp2 gene that is located at 4D3 in chromosome 4 and stretches for 55 kb and consists of 12 exons and 11 introns including an alternative exon (exon 1b), located ∼30 kb downstream of exon 1a (30, 31). The direction of the Akp3 gene and the Akp-ps1 pseudogene is opposite to that of Akp5 and Akp6 genes. We also extended our search for novel AP genes by probing the databases, this time with the less stringent BLAST search, using the Akp2, Akp3, Akp5, and Akp-ps1 DNA sequences as queries and found no further evidence of additional loci besides the Akp6 locus. Furthermore, with the caveat that negative data are seldom conclusive, to date we have found no confirmation in the Mouse Genome Databases that the Akp1 or Akp4 loci actually exist. These loci were postulated based on hybrid maps and biochemical data (34) but no sequences exist for either of these two proposed loci.
We then proceeded to cloning and sequencing the full-length Akp6 cDNA from Akp3−/− intestine RNA. The 2.2 kb obtained by PCR was sequenced and its coding sequence was identical to the predicted cDNA, XM_129951. Table 1 shows all amino acid differences between gIAP and dIAP. Notably absent from that table are all the active site residues, i.e., D42, H153, S155, E311, D316, H320, D357, H358, H360, H432, that are perfectly conserved between gIAP and dIAP. However, the Zn1 and Mg neighboring H317, a residue involved in the indirect control of enzyme catalysis in human PLAP (18) and preserved in EAP, was differentially substituted for Gln in gIAP and for Arg in dIAP (Table 1). gIAP and dIAP display 87% sequence identity, with the largest degree of substitution in gIAP occurring in the crown domain of the molecule, which consistently displays the largest degree of dissimilarity among AP isozymes and that confer unique biological properties to the isozymes (3). The NH2 terminus of the molecule, recently shown to be involved in a structural stabilization of APs (17), is very well conserved, showing two amino acid substitutions only.
Structure of the gIAP protein.
Phylogenetic comparison of the gIAP sequence with those of the human, mouse, rat, and bovine AP genes showed that the gIAP sequence is closely related to mouse dIAP and also to rat IAP I (Fig. 2). This is significant since in the rat IAP I mRNA can be found in both the duodenal and jejunal mucosa, but IAP II mRNA is only expressed in the duodenum (4). It would appear that gIAP corresponds in expression and perhaps function to that of rat IAP I, whereas dIAP corresponds more to rat IAP II (15).
We then modeled the structure of gIAP on the basis of the crystallographic coordinates of human PLAP (1EW2; http://pdbbeta.rcsb.org/pdb/) (19). Our a priori hypothesis was that this murine isozyme would have the same fold, the same quaternary structure, and nearly all the same amino acid orientations as the human PLAP structure. We did the modeling using the Swiss Model server via Deep View as the local interface. Waters and nonamino acids were removed from the template structure. The returned model had only 0.09 Å RMS on backbone atoms. These were observed to be due to mostly very small shifts (∼0.15 Å, up to 0.26 Å) in the backbone positions around amino acid differences. There is 77% sequence identity and 87% sequence similarity in the mature gIAP protein with respect to PLAP across the modeled domain. The quaternary structure is based on notes in PDB submission and also from the EBI PQS server analysis which states that the “oligomer state appears to be correct” (4,168.7 Å2 of surface area is buried via the formation of the dimer in the template structure). Superposing the gIAP model and the crystal structure of PLAP showed complete preservation of the backbone structure. Figure 3A shows the modeled gIAP structure with active site metal atoms displayed. Equally shown is Q317 in the vicinity of the active site. Figure 3B shows the microenvironment of Q317 in larger detail and superimposed onto the detailed microenvironment in the PLAP structure, harboring H317. In PLAP, H317 is positioned so as to be a direct ligand to an active site water molecule and an indirect ligand to the Mg2+ ion and the noncovalently bound phosphate group (18). As can be observed in Fig. 3B, the H317Q substitution brings residue 317 closer to the Zn1 and away from the Mg2+ ion and this movement opens space to accommodate the D273E substitution.
Kinetic characterization of the gIAP isozyme.
In view of potentially important structural differences between gIAP and dIAP, we characterized the enzymatic properties of this novel murine AP isozyme in more detail. To that effect we introduced a FLAG epitope at the COOH terminus of the Akp3, Akp5, and Akp6 cDNAs (at the location predicted to harbor the GPI anchor based on its homology to PLAP) and we expressed each plasmid and recovered and purified the recombinant EAP, gIAP, and dIAP isozymes. The flagged enzymes were bound to anti-FLAG antibody-coated microtiter plates and the bound activity was measured kinetically to determine kcat and Km at pH 9.8 by using pNPP as a substrate. The data shown in Table 2 demonstrate that gIAP has kinetic characteristics quite distinct from those of dIAP and EAP. Thus gIAP displays a 5.7-fold reduction in kcat and a 30% drop in Km. These changes lead to a 4-fold reduction in catalytic efficiency (kcat/Km) of gIAP compared with dIAP. The catalytic efficiency of gIAP is comparable to that of EAP, although EAP displays a 40-fold lower kcat. However, the 8-fold drop in Km leads to a kcat/Km ratio equivalent in bothgIAP and EAP.
Next we investigated whether differences in catalytic efficiencies (Table 2) could be associated with amino acid substitutions of critical active site residues. Since all amino acid residues involved in active site metal coordination were identical, with only residue 317 in the active site vicinity differing between gIAP, dIAP, and EAP (Table 1), this residue was mutagenized. In PLAP, H317 is indirectly involved in Mg2+ ion stabilization through coordination with a water molecule (18), and mutations at this position have been shown to affect kcat and to modulate the enzyme activity. Mutagenizing R317A, R317H, and R317Q in dIAP leads to different effects: whereas the former two substitutions enhance kcat more than threefold, the R317Q substitution reduced the turnover rate to that equivalent of the gIAP isozyme (Fig. 4). The catalytic upregulation induced by the R317A and R317H mutations indicates that it is not the introduction of a His residue at this position that is responsible for the enhanced catalysis, but rather the loss of the Arg and its coordination. The poor EAP enzymatic efficiency therefore depends on residues other than H317, conserved between PLAP and mouse EAP.
Expression of gIAP in the gut.
Using the amino acid sequence information (Table 1) and the three-dimensional model (Fig. 3), we could identify three putative regions useful for the generation of anti-peptide antibodies, specific to gIAP. Because of its favorable location in the easily accessible shoulder region of the gIAP modeled structure (Fig. 3A), we chose as epitope the peptide C208FPKGTPDPEYPSDSNQSGTRLDD231Q of the mature gIAP polypeptide to generate an antiserum. This peptide, with a Cys residue added at the NH2 terminus to facilitate conjugation was used to produce rabbit polyclonal anti-peptide antisera to gIAP. The antiserum reacted strongly with gIAP but also cross-reacted with dIAP, but not with EAP. Immunohistochemical staining of the small intestine of Akp3−/− (devoid of dIAP) mice reveals strong expression of gIAP in the intestinal mucosa and also of what appear to be neuroendocrine cells scattered at low frequency throughout the length of the intestine (Fig. 5). The antiserum was also useful in Western blot analyses to confirm the presence of gIAP in WT and Akp3−/− intestines, extending from the duodenum to the ileum of WT but especially Akp3−/− intestines. In addition, whereas gIAP is expressed along the intestinal tract, this new isozyme is upregulated in the jejunum and ileum of Akp3−/− mice (Fig. 6).
Northern blot analysis (Fig. 7), using Akp3-, Akp5-, and Akp6-specific cDNA probes, detected bands in the range of 2.5–2.3 kb for each of these three mRNAs. The intestines for this analysis were collected and cut into four segments. In agreement with the Western blots, Akp3 expression was only found in the duodenum, whereas, in addition to its presence in duodenum, Akp6 was expressed throughout the gut, and it was particularly upregulated in Akp3−/− mice (Fig. 7). We then examined the expression of Akp3, Akp5, and Akp6 in WT as well as Akp3−/− mice fed a high-fat diet. For the gavage experiment, mice were fasted 16 h and 0.5 ml of corn oil was administered via a feeding needle into the stomach. Intestine samples were collected 5 h later. For high-fat feeding, mice were housed with 11% high-fat diet for 8 wk before analysis of their intestines. Northern blot analysis of the four intestinal segments show that Akp6 is indeed expressed along the entire length of the intestinal tract, after fat feeding, both in WT and Akp3−/− mice, but that its expression is downregulated in the duodenum and upregulated in the ileum of mice when fed a high-fat diet either by gavage or by long-term feeding.
Given the differential expression of Akp3, Akp5, and Akp6 in the adult intestine, it became of interest to probe for the putative differential regulation of these genes at earlier developmental stages, i.e., neonatal and postnatal stages bracketing the time of weaning, since the nutritional status of mice changes dramatically at weaning time. As shown on Fig. 8, Akp3 is not expressed in the early postnatal stages. Akp3 expression only becomes visible at day 15 and full expression is reached after day 17. In contrast, both Akp5 and Akp6 are expressed from birth, but their expression also increases after weaning.
Posttranslational modulation of the catalytic properties of gIAP.
The Western blot data shown both in Figs. 4 and 6 clearly indicated that the recombinant gIAP used as a control for the antibody detection had a lower molecular mass than the native gut-derived gIAP. This is likely the result of underglycosylation at one or more of the four possible sites available for N-linked glycosylation, present in dIAP, gIAP, and EAP, by the cells used to express the recombinant enzymes. Although a single gIAP species is present in the adult intestine, we wanted to confirm that such would also be the case in the neonatal and postnatal gut. To our surprise, intestinal samples from 2-day-old and 10-day-old mice showed a larger band (∼70–75 kDa) and a smaller band (∼55 kDa), which corresponds to the predicted molecular mass of unmodified GPI-anchored gIAP polypeptide (54,526 Da) (Fig. 9). The larger band observed in intestinal segment 4 appears to be the same size as gIAP detected in adult gut. To examine the catalytic properties of these modified gIAP, intestinal segments from 2-day-old WT mice were homogenized in Tris buffer (pH 8.9) containing 0.1% Triton X-100 and extracted with n-butyl alcohol. Extracts (1.5 mg/ml protein concentration) were incubated on 96-well plates coated with the anti-gIAP antibody. Enzymatic activity of the specifically bound gIAP protein was measured with serial concentrations of substrate (20, 10, 5, 2.5, 1.25, and 0 mM pNPP). Intestinal segments 2 and 3 (corresponding to jejunum) showed lower Km values (0.67 ± 0.18 mM and 0.77 ± 0.20 mM, respectively) than gIAP from segment 1 (duodenum; 0.86 ± 0.13 mM) or segment 4 (ileum; 1.00 ± 0.44 mM). Enzyme activity was also lower in segments 2 and 3 (Fig. 9). To assess whether the change in molecular mass was associated with N-linked glycosylation particularly by polylactosamines (12), butanol extracts of intestinal segment 1 from 2-day-old mice were bound onto anti-gIAP-coated 96-well plates. After washing, the wells were treated with 0.003 units of endo-β-galactosidase overnight. Endo-β-galactosidase specifically cleaves β-galactosidic linkage in polylactosamines (12). We found that this enzymatic treatment reduced the activity of gIAP to levels comparable to those present in segments 2 and 3, suggesting that it is a change in polylactosamines that is associated with modulation of catalytic properties of gIAP. A similar change was observed for EAP, indicating that segments 2 and 3 of the neonatal gut are unable to fully glycosylate these glycoproteins in these intestinal segments before weaning, consistent with the developmental expression of galactosyl-transferases in the postnatal gut (27).
The finding that Akp3−/− mouse intestines still express alkaline phosphatase activity with the characteristics of IAP has constituted the basis for our search for the existence of an unknown IAP isozyme in the mouse. The present work confirms the existence in the mouse of an IAP gene, which we denominated Akp6, and which shows homologies but also clear dissimilarities to the previously reported Akp3 gene. As for mouse Akp3, Akp5, and the Akp-ps1 pseudogene, Akp6 is closely linked at the 1C5 site in chromosome 1. Its size equals ∼3.5 kb and it contains 11 exons and 10 introns.
In the presence of normal biliary flow, fat is digested and absorbed primarily in the proximal small intestine (duodenum and jejunum) with minor amounts absorbed in the distal ileum (35). The ileum retains a good capacity for fat absorption but under normal circumstances does not receive much dietary lipid, because the jejunum is so efficient in fat absorption. We found by Western and Northern blot analysis that Akp3 is exclusively expressed in the mouse duodenum, compatible with a partial role for dIAP in fat absorption. Similarly, when analyzed via Western blotting, we found that the novel Akp6 gene is expressed from the duodenum to the ileum of WT and Akp3−/− mice, fed a normal chow. Owing to its lower molecular size, the expression in various parts of the gut could be monitored accurately via Western blotting, an analysis revealing that gIAP was similarly present in the duodenum of WT and Akp3−/− mice, but that the absence of Akp3 triggered upregulation of areas more distal than the duodenum, down to the ileum. Northern blot analysis of various segments of the gut similarly demonstrated that Akp6 was expressed in the duodenum of WT and Akp3−/− mice, but that distally gIAP was expressed at higher levels in the Akp3−/− mice. Our study also revealed that Akp6 is expressed in the entire postnatal small intestine and can be detected as early as embryonic day 18.5, whereas the duodenum-restricted expression of Akp3 starts only prior to weaning.
Our finding that high-fat feeding upregulates the expression of Akp6 even further in Akp3−/− mice, over the entire gut, and that in WT mice, corn oil as well as high-fat feeding induce a strong upregulation of Akp6 expression in all parts of the gut, raises the question of the functional role for both gIAP and dIAP. Akp3−/− mice fed a normal diet show normal growth curves, despite absent dIAP and upregulated gIAP. However, under stressed feeding conditions, Akp3−/− mice show a rapid weight gain, compared with that in WT mice (26). It would be tempting to associate the weight gain in Akp3−/− mice with a role for the upregulated Akp6 in either accelerated fat transport or increased capacity for fat absorption, or both. Recent studies (25) showed that high-fat diet feeding in Akp3−/− mice resulted in liver steatosis but that the fat intake did not differ in these mice. Under conditions of high-fat feeding (15% fat) rodents develop steatorrhea (28). It seems likely that the Akp3−/− animals were able to absorb triacylglycerol fat more efficiently than WT, leading to increased weight and liver fat, whereas dietary intake did not change. Hence, these results confirm an association of dIAP with proximal fat absorption but implicate gIAP with enhanced fat absorption in the distal intestine. The mechanism by which dIAP regulates production of gIAP is not clear. The data imply a modulatory role of dIAP in controlling fat transport or fat absorption, absent in the Akp3−/− mice.
Analysis of the kinetic properties of gIAP and EAP vs. those of dIAP shows that they are very inefficient enzymes, at least when assayed with artificial substrates. In fact, this conclusion is to some extent valid for all mouse and human intestinal-like AP isozymes compared with those in the calf intestine that have the highest turnover number of any AP known to date (∼6,000 s−1) (22). It is tempting to speculate that these distinct kinetic properties could be related to the different substrate specificity, perhaps related to the different dietary habits of cattle vs. mice and humans, i.e., a purely vegetarian diet vs. a fat-containing diet. Alternatively, these data may also indicate that the catalytic activity of the mouse and human enzymes do not necessarily reflect their biological role(s). The highest degree of structural dissimilarity was found in the crown domain of gIAP and dIAP, an area known to vary considerably from one AP isozyme to another. These large differences both kinetically and structurally therefore suggest that the functions of gIAP and dIAP are different, an interpretation further supported by their differential expression regulation in WT and Akp3−/− mice. Other novel observation made in this study are that two molecular forms of EAP and gIAP, which appear to differ in the N-glycosylation state, are found in the neonatal intestine in a segment-restricted manner and that the degree of glycosylation has an effect on the catalytic properties of these isozymes. dIAP is not affected because it is not expressed at these early neonatal stages and also not expressed in those regions of the gut that have differential galactosyltransferase activity (27). Thus, if EAP and gIAP have a catalytic role in the gut, posttranslational modification plays a modulatory role on their function.
Except for Q317, all active site residues were preserved in gIAP compared with dIAP. This residue was shown before to modulate AP catalysis (18). We presently have found that the single mutation in dIAP of R317G sufficed to eradicate most of the enzyme activity, a finding supported by active site modeling, which clearly showed that substitution of H317 in the crystal model of the PLAP active site for Q317 resulted in a microenvironment in which the new Q317 points toward the active site Zn1 metal. It is not hard to imagine that such an exchange interferes with catalysis, thus generating an almost inactive IAP mutant.
In view of the similarities between gIAP and rat IAP I on the one hand and dIAP and rat IAP II on the other, it is tempting to speculate that the function of gIAP and that of rat IAP I are similar. The mechanism whereby an enzyme anchored in the apical membrane of the enterocyte could enter the lymph and serum is not obvious. IAP is anchored to the external surface of the apical brush border via a glycosyl-phosphatidylinositol anchor, and the absence of a transmembrane sequence made especially puzzling the appearance of IAP in the serum. By using tannic acid as such a preservative, membranes were found on the surface of the enterocyte that were enriched in IAP (6). These membranes appeared to have morphological characteristics similar to those of pulmonary surfactant, namely occurring in swirls but without cross-linking features. Indeed, membranes isolated from the rat small intestine were found to contain surfactant proteins B and D, whereas those membranes from rat and human small intestine also contained surfactant protein A (8). In the rat, IAP II mRNA increased much more after fat feeding. Transfection of Caco-2 cells with cDNA encoding rat IAP I increased cellular IAP by twofold, but the increase in secreted IAP was much greater, averaging 32-fold (32). Thus increased IAP production appears to drive the production of SLP in Caco-2 cells. The secreted IAP is membrane associated, and these membranes have the properties of SLP (6). The yield of SLPs is highest in the proximal half of the rat intestine (8), similar to the area of maximal lipid absorption, and coinciding with the mucosa that expresses both IAP I and IAP II. Hence, if a similar function would exist in the mouse, gIAP would be primarily involved in the production of surfactant following high-fat feeding, especially in the absence of dIAP, when gIAP is upregulated.
In conclusion, we describe a novel murine intestinal phosphatase, gIAP, homologous to dIAP, but with structural and enzymatic properties that underscore a functional distinction between both isozymes. Furthermore, their differential expression in the gut and regulation during high-fat feeding is suggestive of different biological functions for gIAP and dIAP.
This work was supported by Grant DE12889 from the National Institute of Dental Research.
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