EQUINE VETERINARY JOURNAL Equine vet. J.

(2002) 34 (4) 349-358

349

Molecular characterisation of carbohydrate digestion and absorption in equine small intestine

J. DYER, E. FERNANDEZ-CASTAO MEREDIZ, K. S. H. SALMON, C. J. PROUDMAN, G. B. EDWARDS and S. P. SHIRAZI-BEECHEY* Epithelial Function and Development Group, Department of Veterinary Preclinical Sciences, University of Liverpool, Brownlow Hill and Crown Street, Liverpool L69 7ZJ and Equine Division, The Philip Leverhulme Large Animal Hospital, The University of Liverpool, Leahurst, Neston, Wirral CH64 7TE, UK.
Keywords: horse; small intestine; dietary carbohydrate; SGLT1; disaccharidases

Summary Dietary carbohydrates, when digested and absorbed in the small intestine of the horse, provide a substantial fraction of metabolisable energy. However, if levels in diets exceed the capacity of the equine small intestine to digest and absorb them, they reach the hindgut, cause alterations in microbial populations and the metabolite products and predispose the horse to gastrointestinal diseases. We set out to determine, at the molecular level, the mechanisms, properties and the site of expression of carbohydrate digestive and absorptive functions of the equine small intestinal brush-border membrane. We have demonstrated that the disaccharidases sucrase, lactase and maltase are expressed diversely along the length of the intestine and D-glucose is transported across the equine intestinal brushborder membrane by a high affinity, low capacity, Na+/glucose cotransporter type 1 isoform (SGLT1). The highest rate of transport is in duodenum > jejunum > ileum. We have cloned and sequenced the cDNA encoding equine SGLT1 and alignment with SGLT1 of other species indicates 8589% homology at the nucleotide and 8487% identity at the amino acid levels. We have shown that there is a good correlation between levels of functional SGLT1 protein and SGLT1 mRNA abundance along the length of the small intestine. This indicates that the major site of glucose absorption in horses maintained on conventional grass-based diets is in the proximal intestine, and the expression of equine intestinal SGLT1 along the proximal to distal axis of the intestine is regulated at the level of mRNA abundance. The data presented in this paper are the first to provide information on the capacity of the equine intestine to digest and absorb soluble carbohydrates and has implications for a better feed management, pharmaceutical intervention and for dietary supplementation in horses following intestinal resection. Introduction The horse is a nonruminant herbivore and fermentation of plant fibre in the equine large bowel, by resident microflora, results in the formation of short chain fatty acids, which provide a large part
*Author to whom correspondence should be addressed.

of the horses energy requirement. Todays horse, however, is subjected to an increasingly artificial diet that contains high levels of concentrates (grain) which, when digested and absorbed in the small intestine, can provide a substantial fraction of metabolisable energy (Alexander 1955; Argenzio and Hintz 1972). However, if levels and components of these diets exceed the ability and capacity of equine small intestine to digest and absorb them, they reach the large bowel, cause alterations in microflora and the fermentation products, and predispose the horse to gastrointestinal disturbances. Any gut dysfunction in horse is threatening, because it can precede colic, the main cause of equine mortality (Hintz and Cymbaluk 1994). In nonruminant species, digestible dietary carbohydrates are hydrolysed in the intestinal lumen by pancreatic -amylase and brush-border membrane disaccharidases, sucrase, maltase and lactase, to the monosaccharides, D-glucose, D-fructose and D-galactose. These monosaccharides are absorbed across the enterocyte brush-border membrane by specific monosaccharide transporters (Shirazi-Beechey 1995). Very little information is available on the mechanisms and intestinal sites of carbohydrate digestion and absorption in the equine intestinal tract. This knowledge is critical for assessing the ability and capacity of the equine intestine to digest and absorb carbohydrates, in order to better understand the pathogenesis of intestinal disease and improve feed management. The majority of studies relating to carbohydrate digestion and absorption to date have been based on whole animal feed experiments (Argenzio and Hintz 1972; Roberts 1975). While these studies have improved our understanding of carbohydrate metabolism in the horse, the complexity of whole body balance studies would not reveal the processes operative at the molecular level. In this study, we have set out, for the first time, to determine the digestive and absorptive capabilities of equine small intestine at the cellular and molecular levels. Materials and methods Collection of tissue Intestinal tissue, from 23 mature horses (age >4 years) maintained on a grass-based diet, was collected from the local abattoir within 15 min of slaughter; none were geriatric. Sections of small

350

Molecular characterisation of carbohydrate digestion

intestine, approximately 20 cm in length, were removed from the proximal (30 cm distal to the pylorus), mid (halfway along the small intestine) and distal (30 cm proximal to the ileocaecal junction) intestine. Sections were opened longitudinally, rinsed in ice-cold 0.9% (w/v) NaCl, pH 7.4, and blotted with paper towels to remove excess mucus. A fresh sample of intestine was fixed in 4% (w/v) phosphate buffered paraformaldehyde, for histology, and the mucosa of the remainder of the same tissue was then removed by scraping. The mucosal scrapings were wrapped in aluminium foil and frozen immediately in liquid nitrogen. Following transportation to the laboratory, frozen tissue samples were stored at -80C until use. Mucosal scrapings from tissues that were shown to be intact, i.e. with epithelial cells attached to the muscularis mucosa, as revealed by histology, were used. Isolation of brush-border membrane vesicles Brush-border membrane vesicles (BBMV) were prepared from equine small intestinal mucosal scrapings using a method based on that of Shirazi-Beechey et al. (1988, 1990). Briefly, mucosal scrapings were thawed in a solution (100 mmol/l mannitol, 2 mmol/l HEPES/Tris, pH 7.1), containing a cocktail of protease inhibitors (Complete protease inhibitor cocktail tablets; D-glucose UV test Kit)1 (Rowell et al. 1997). The mucosal scrapings were homogenised for 2 x 1 min in a Waring blender at the highest setting. The volume was made up to 100 ml with the same buffer, MgCl2 was added to a final concentration of 10 mmol/l and the homogenate was stirred for 20 min, on ice. The homogenate was centrifuged at 5000 g (3000 rpm, Sorvall SS 34 rotor) for 10 min in a Sorvall RC 5C+ centrifuge (Sorvall RC 5C+ centrifuge)2. The pellet was discarded and the supernatant centrifuged at 30,000 g (16,000 rpm, Sorvall SS 34 rotor) for 30 min. The resulting pellet was resuspended in 35 ml solution containing 100 mmol/l mannitol, 2 mmol/l HEPES/Tris, pH 7.4, 0.1 mmol/l MgSO4, using a Potter/Elvejhem hand-held homogeniser and centrifuged at 30,000 g (16,000 rpm, Sorvall SS 34 rotor) for 45 min. The final pellet, containing purified brush-border membranes, was resuspended in a small volume of isotonic buffer (300 mmol/l mannitol, 20 mmol/l HEPES/Tris, pH 7.4, 0.1 mmol/l MgSO4, 0.02% [w/v] NaN3) by passing through a 27 gauge needle several times. The BBMV were divided into aliquots and stored in liquid nitrogen, until use. All steps in the procedure were carried out at 4C. Estimation of protein Protein concentration was determined by its ability to bind Coomassie Blue, according to the Bio-Rad assay technique (Tarpey et al. 1995). Bovine -globulin (1100 g protein) was used as the standard. Enzyme assays Sucrase, maltase and lactase activities were measured at 38C (equine body temperature) in the original cell homogenates and final brush-border membrane fractions, as described previously (Dahlqvist 1964; Shirazi-Beechey et al. 1989, 1990). D-glucose, released as a result of the hydrolysis of disaccharide substrates (sucrose, lactose and maltose, 28 mmol/l final concentration), was measured using a commercially available kit1 according to the manufacturers instructions. p-chloromercuri-benzoate (pCMB) was included in the lactase assay to inhibit any potential

contamination from lysosomal -galactosidase. Alkaline phosphatase was measured at pH 10 in the presence of 5 mmol/l MgCl2 and 0.25 mmol/l CaCl2 using p-nitrophenyl phosphate as the substrate (Shirazi et al. 1981). The potential presence of any basolateral and organelle membranes was determined by assessing the activities of the following marker enzymes; K+-activated phosphatase (basolateral membrane), -mannosidase (Golgi), Tris-resistant -glucosidase (endoplasmic reticulum), acid phosphatase (lysosomes) and succinate dehydrogenase (mitochondria), as described previously (Shirazi-Beechey et al. 1989, 1990). Assay of Na+-dependent D-glucose transport Na+-dependent D-glucose transport into BBMV was measured at 38C in a solution containing either 100 mmol/l NaSCN or 100 mmol/l KSCN and 100 mmol/l mannitol, 20 mmol/l HEPES/Tris, pH 7.4, 0.1 mmol/l MgSO4 and 0.02% (w/v) NaN3. Incubation was stopped at specific time points (3 s for the initial rate and 3 s, 10 s, 1, 10 and 30 min for the time-course measurements) by the addition of 1 ml ice-cold buffer containing 0.1 mmol/l phlorizin (Shirazi-Beechey et al. 1989). Aliquots of this solution were then filtered rapidly through 0.2 m cellulose acetate/nitrate filters (0.2 m cellulose acetate/nitrate filters)3, using the filtration stop technique, as described previously (Shirazi-Beechey et al. 1990). All initial rate measurements were taken after 3 s incubation period as transport was determined to be linear up to 4 s (data not shown). Uptakes were measured in duplicate or triplicate. For the kinetic studies D-glucose concentration was 0.012.5 mmol/l. Na+-dependent uptake was calculated as the difference in rates measured in the presence and absence of Na+. Competition studies were carried out by determining the initial (3 s) rate of uptake of 0.1 mmol/l D-glucose in the presence of 1 mmol/l competitor (Hediger and Rhoads 1994). Cloning and sequencing of equine SGLT1 Oligonucleotide primers, derived initially from the consensus nucleotide sequences of ovine and human SGLT1 cDNA and subsequently from the equine SGLT1 sequence, were used in reverse transcription-polymerase chain reaction (RT-PCR) against total RNA isolated from equine jejunal mucosal scrapings. RNA samples were prepared using a commercially available kit (RNeasy Mini Kit)4 according to the manufacturers instructions. First strand synthesis was performed using 200 u Superscript II5 RNase H- reverse transcriptase and random primers (pGEM-T)6 at 42C for 60 min. cDNA was purified on a nucleospin column (Machery-Nagel Nucleospin Extract)7 and amplified by PCR with 2.5 u Pfu DNApolymerase8 and the following primer pairs: (1) Sense Ls20, nt 484 (5-AGATCTCGGCAGACATCTTT-3), antisense Has2, nt 952 (5-GCAGCCACCCTTCACGTGAGA-3) (2) Sense s891, nt 891 (5-ACAGACCAGGTCATCGTGCAGC-3) antisense as1599, nt 1599 (5-GGCAAAGAGGATGATGGCAAAG-3) (3) sense s1406, nt 1406 (5-ACCACCCATTGCAGCTGTCTTC-3) antisense as2025, nt 2025 (5-CAGCAGAAAGCAGGACTCAGGC-3) The reaction conditions were as follows: 1 x reaction buffer (10 mmol/l KCl, 10 mmol/l (NH4)2SO4, 20 mmol/l Tris-HCl, pH 8.75, 0.1% (v/v) Triton X-100, 1 mg/ml BSA), 200 mol/l

J. Dyer et al.

351

each dNTP, 50 pmol each primer, 3 mmol/l MgCl2, 1 g cDNA. PCR amplification was performed at 94C, 20 s; 57C, 20 s; 72C, 150 s for 30 cycles. A final extension of 7 min at 72C was included. The RT-PCR products were tailed with adenosine using Thermoprime plus DNA polymerase (AB gene), cloned into pGEM-T6 and custom sequenced. Equine specific SGLT1 sense and anti-sense primers were synthesised based on these sequences and used in subsequent reactions to clone the 5- and 3-ends of equine SGLT1 (eSGLT1). The 5-UTR was cloned using rapid amplification of cDNA ends (RACE) using a commercially available kit5 and the following nested gene-specific anti-sense primers; Las18, nt 681 ( 5 -AAGCAAACCCAGTCAGGATG-3) and Has7, nt 482 (5-CTTGGTGAAAATGTAGAGCA-3). The 3-UTR was cloned by 3 RACE. First strand synthesis was as described for RT-PCR, using the 3 RACE primer (5-GGCCACGCGTCGACTAGTACT17-3). A 400 bp product was amplified from the cDNA using Pfu DNA polymerase with the eSGLT1 specific sense primer s1767, nt 1767 (5-ACACCTTGGAGGAGAGCATTGA-3) and the 3 RACE primer. The product was tailed, cloned into pGEM-T as described and custom sequenced. The GenBank accession number for the equine SGLT1 mRNAsequence is AJ292081. Western blot analysis The abundance of SGLT1 protein in the equine intestinal BBMV was determined by western blotting, as described previously (Dyer et al. 1997a). The protein contents of BBMV were separated on 8% polyacrylamide gels containing 0.1% (w/v) SDS and were electrotransferred to nitrocellulose membrane (TransBlot nitrocellulose)9. Samples were blotted with an antibody raised to a synthetic nonodecapeptide (STLFTMDIYTKIRKKASEK) corresponding to a highly conserved region of the SGLT1 amino acid sequence in other species (Hirayama and Wright 1992; Shirazi-Beechey 1995; Dyer et al. 1997a; Wood et al. 2000), including eSGLT1 (Fig 7). The 75 kDa immunoreactive band was blocked when antibodies were pre-incubated with the immunising peptide indicating the specificity of the immunoreaction. The membranes were developed using the ECLsystem10 and exposed to film (BioMax MR film)11. The intensities of the immunoreactive bands, detected in the BBMV, were quantified using scanning densitometry (Phoretix 1D Quantifier)12 and reported as arbitrary units. Lactase abundance was determined using a monoclonal antibody (mlac6) to human lactase as described previously (Dyer et al. 1997b). Northern blot analysis Total RNA was isolated from duodenal, jejunal and ileal mucosal scrapings of equine small intestine using a commercially available kit4. RNA samples, 3 g per lane, were separated on 1% agarose gels containing 0.66 mol/l formaldehyde and transferred to nylon membrane (Hybond N)10. RNA integrity and equality of loading were determined by methylene blue staining. The nylon membrane was prehybridised for 3 h at 42C in a buffer containing 50% (v/v) deionised formamide, 5 x SSC, 6 x Denhardts reagent, 0.2% (w/v) SDS, 10% (w/v) dextran sulphate, 2.5 mmol/l sodium pyrophosphate, 25 mmol/l MES, pH 6.5 and 0.01% (v/v) antifoam B13. Equine SGLT1 cDNA (25 ng)

TABLE 1: Disaccharidase activity in BBMV isolated from the intestine of grass-fed horses (mean s.e.) Specific activity (mol/min/mg protein) Duodenum Jejunum Ileum Sucrase Maltase Lactase 0.353 0.088 (n = 7) 0.880 0.118 (n = 8) 0.068 0.025 (n = 7) 0.424 0.075 (n = 7) 0.998 0.155 (n = 7) 0.101 0.041 (n = 7) 0.221 0.037 (n = 7) 0.742 0.108 (n = 8) 0.029 0.011 (n = 7)

corresponding to nucleotides 14062025, was labelled with [32P]-dCTP by random priming and hybridised with the membrane for 16 h at 42C. Washes were performed at 55C for 15 min once with 5 x SSC, 0.5% (w/v) SDS, 0.25% (w/v) sarkosyl and 3 times with 0.1 x SSC, 0.1% (w/v) SDS. The membrane was apposed to x-ray film (BioMax MS)11, with a single intensifying screen, at -80C for 6 h. The intensity of the transcripts was quantified using scanning densitometry (Phoretix 1D Quantifier)12. Statistics Data are expressed as mean s.e. and statistical comparisons are made using one-way analysis of variance, = 0.01. Results The intestinal tissue used was obtained from grass-fed horses, slaughtered by conventional methods at the local abattoir. On occasion, histology of the tissue revealed that there had been shedding of the epithelium, as has been reported previously in intestinal tissue of other species obtained from animals slaughtered by conventional methods (Fell 1961; Shirazi-Beechey et al. 1989). Therefore, care was taken to use only tissues shown to be intact by histology. Furthermore, a cocktail of protease inhibitors was included in the buffers used to prepare BBMV and this appeared to be essential for maintaining intact equine intestinal brush-border membrane proteins. The specific activities of the brush-border marker enzymes alkaline phosphatase and disaccharidases were 2025-fold enriched in BBMV over the original cellular homogenate. The activities/abundances of marker proteins characteristic of organelle/basolateral membranes were negligible in BBMV, indicating that the membrane vesicles were derived from the brush-border membrane and were purified. Equine small intestinal disaccharidases In order to determine the ability of the equine small intestine to hydrolyse disaccharides, the activities of sucrase, maltase and lactase were determined. The results are presented in Table 1. Sucrase, maltase and lactase are expressed in equine small intestine. Sucrase specific activities were highest proximally (duodenum = jejunum > ileum), while maltase activity was similar in all 3 regions. The horse appeared to express lactase postweaning (see Table 1). This was confirmed by western blot analysis using a monoclonal antibody specific to the brush-border membrane lactase. The antibody cross-reacts with a single protein with an apparent molecular mass of 150 kDa in equine intestinal BBMV, a similar size to lactase reported in other species

352

Molecular characterisation of carbohydrate digestion

a)

500 b) 400

TABLE 2: Comparison of sequence data for SGLT1 from various species No. cDNA Species amino acids homology (%) Horse Man Mouse Rabbit Rat Sheep 662 664 665 662 665 664 100 88.8 84.7 85.4 84.6 88.6 Amino acid Amino acid identity (%) similarity (%) 100 87.0 86.8 84.6 86.5 86.2 100 91.4 91.8 90.6 92.0 89.7

150 kD

300 200 100 0 D J I D J I

Fig 1: Lactase expression in equine small intestinal brush-border membrane vesicles (BBMV). BBMV (15 g protein/lane) prepared from equine duodenum (D), jejunum (J) and ileum (I) were separated on 8% polyacrylamide gels, electrotransferred to nitrocellulose membrane and western blotted using an anti-lactase monoclonal antibody (mlac6), as described in Materials and methods. a) The results of a typical western blot indicating the presence of the 150 kDa lactase protein. b) Densitometric analysis of western blots carried out using Phoretix 1D12. Data are expressed as mean s.e. (n = 6) in arbitrary units. 1500

The sequence comparisons are relative to the horse sequence. 250

200

150 1250 1000 100 750 500 250 0 0.0 0 50

0.5

1.0 1.5 8 10 1225 30 35 Time (min)

Duodenum

Jejunum

Ileum

Fig 2: Time course of Na+-dependent D-glucose transport by equine duodenal brush-border membrane vesicles. The uptake of 0.1 mmol/l Dglucose into equine duodenal BBMV was measured at 38C in the presence of an inwardly directed (out>in) concentration gradient of either 100 mmol/l NaSCN ( ) or KSCN ( ), under zero trans conditions as described in Materials and methods. All uptakes were measured in duplicate and values are presented as mean s.e. (n = 3).

Fig 3: Initial rate of Na+-dependent D-glucose transport in equine small intestinal brush-border membrane vesicles. The initial (3 s) rate of 0.1 mmol/l D-glucose uptake into BBMV(0.1 mg protein) was measured at 38C in the presence of 100 mmol/l NaSCN, as described in Materials and methods. Results are presented as mean s.e. (n = 7).

(Schlegel-Haueter et al. 1972; Mantei et al. 1988; Dyer et al. 1997b). The results are presented in Figure 1. Densitometric analyses of western blots for lactase in the intestinal brush-border membranes of several horses are shown in Figure 1b. As can be seen, lactase expression was highest in the duodenal and jejunal BBMV, with a marked reduction in levels detected in BBMVfrom the distal small intestine, P= 0.065. The large s.e. seen for jejunal lactase activity and abundance reflect the fact that some horses showed the highest levels of lactase expression in jejunum, whereas others had little lactase expressed in this region. However, all horses expressed higher levels of lactase in the duodenum than the ileum. A similar distribution pattern was seen for lactase activity reported in Table 1, with higher levels of activity in the duodenum and jejunum than in the ileum. Monosaccharide transport in equine small intestine Time-course and initial rates of D-glucose transport: The ability of equine small intestine to transport D-glucose was measured in BBMV and results are presented in Figure 2. D-glucose was

transported into the membrane vesicles in a Na+-dependent manner. The uptake exhibited a classical overshoot, i.e. a transient increase in intravesicular glucose concentration above that of the incubation medium. When Na+ in the incubation medium was replaced by K+ the overshoot was abolished and glucose uptake rates were negligible. Na+-dependent glucose transport was inhibited when 0.1 mmol/l phlorizin was included in the incubation medium. Phlorizin is a specific inhibitor of Na+-dependent D-glucose transport (Wright et al. 1994; ShiraziBeechey 1995). The initial rates of the Na+-dependent transport of D-glucose in BBMV isolated from the duodenal, jejunal and ileal mucosa of grazing horses are presented in Figure 3. Transport rates were highest in the duodenum (200 22 pmol/s/mg protein), lower in the jejunum (116 33 pmol/s/mg protein) and lowest in the ileum (38 16 pmol/s/mg protein), (P= 0.0014). Kinetics and substrate specificity of Na+-dependent D-glucose transport: The effect of substrate concentration on D-glucose transport in BBMV was investigated and the results, using duodenal BBMV, are presented in Figure 4. Initial rate (3 s) of D-glucose uptake was measured at glucose concentrations of

J. Dyer et al.

353

1500 a)

1000
S/V 0.005
0.004

b)

500

0.003 0.002 0.001

-1

1 [S] (mmol/l)

0.0

0.5 1.0 [D-glucose] (mmol/l)

Fig 4. Effect of substrate concentration on the initial rate of D-glucose uptake by equine small intestinal brush-border membrane vesicles. The initial rate of D-glucose transport by equine intestinal BBMV was measured as a function of D-glucose concentration under zero trans conditions as described in Materials and methods. Results are expressed as (a) MichaelisMenten plot of D-glucose concentration vs. initial (3 s) rate of D-glucose uptake for duodenal BBMV and (b) a Hanes plot of linear regression analysis for duodenal (q), jejunal (s) and ileal (v) BBMV. Uptakes were carried out in duplicate and are expressed as mean s.e. (n = 2).

0.012.5 mmol/l D-glucose. Uptake conformed to MichaelisMenten kinetics (Fig 4a) and linear-regression analysis (Fig 4b) indicated the presence of a single transport system for D-glucose, with a Km of 0.49 0.06 mmol/l and a Vmax of 1312 141.9 pmol/s/mg protein. Km values obtained using jejunal and ileal BBMV were not significantly different. Vmax values, however, were 918 73.3 and 698 61 pmol/s/mg protein in the jejunal and the ileal BBMV, respectively. Kinetic studies indicate that the aboral decline in Na+-dependent D-glucose transport is due to decrease in the protein abundance (Vmax) rather than the affinity of the transporter for glucose. The Km of 0.49 mmol/l for Dglucose suggests that the Na+/glucose cotransporter in equine small intestine is the high affinity isoform SGLT1. SGLT1 has an affinity for D-glucose in the range 0.050.8 mmol/l in other species (Malo and Berteloot 1991; Shirazi-Beechey et al. 1991a; Hediger and Rhoads 1994; Hirayama et al. 1996). The sugar substrate selectivity of the transporter was investigated by measuring the initial rate of 0.1 mmol/l D-glucose uptake in the presence of 1 mmol/l known substrates of SGLT1 (Hopfer 1987; Shirazi-Beechey 1995). Results are shown in Figure 5. These monosaccharides inhibited D-glucose uptake into equine BBMV in the order of D-glucose (82.6%) > -methyl-Dglucose (75.2%) = D-galactose (72.3%) >> 3-O-methyl-Dglucose (29.5%) >D-xylose (16.6%). This pattern of substrate selectivity is similar to that shown for SGLT1 in other species (Wright 1993; Hediger and Rhoads 1994; Diez-Sampedro et al. 2001) and further suggests that the equine intestinal Na+/glucose cotransporter is the SGLT isoform 1 (see also eSGLT1 sequence). Cloning and sequencing of equine intestinal SGLT1 cDNA Using the techniques of reverse transcription-polymerase chain reaction (RT-PCR) and rapid amplification of cDNA ends (RACE),

initially with homologous primers designed against consensus sequences of SGLT1 in other species and subsequently with specific equine SGLT1 primers, 2.17 kb equine SGLT1 cDNA was cloned and sequenced. Translation of the overlapping contiguous sequences indicates that the equine SGLT1 cDNA encodes for a 662 amino acid protein. Hydropathy analysis of the peptide sequence (Vector NTi Suite)14 suggests a secondary structure with 12 putative membrane-spanning regions and both the N- and Ctermini located intracellularly, which is in agreement with models based on hydropathy analysis for SGLT1 in other species (Hediger and Rhoads 1994). Alignment of the eSGLT1 amino acid sequence with those of SGLT1 in other species is shown in Figure 6 with conserved areas shaded grey. Sequence homology data is presented in Table 2. Equine SGLT1 has between 8985% homology with SGLT1 from other species at the nucleotide level and between 9092% similarity and 8487% identity at the amino acid level. SGLT1 has been shown to be N-glycosylated in rabbit, human and sheep at Asn248 (Wright 1993; Shirazi-Beechey 1995). This residue, indicated by an asterisk in Figure 6, is conserved in eSGLT1. Polyclonal antibodies have been raised against a nonodecapeptide (STLFTMDIYTKIRKKASEK) corresponding to amino acids 402-420, a conserved intracellular loop region of SGLT1 in other species. This region is boxed in Figure 6. As can be seen, equine SGLT1 shares a high degree of homology with other species over this region, with 17/19 amino acid identity with SGLT1 from human, mouse, rat, rabbit and sheep. Equine SGLT1 shares the same hypervariable regions as SGLT1 in other species. These are most notably amino acids 6-22 at the N-terminus and amino acids 576600 at the C-terminus. As can be seen in Figure 6, eSGLT1 does not have an extra amino acid at position 579 as do rat and mouse; however, eSGLT1 has the 2 amino acid deletion at positions 592593, seen previously only in rabbit SGLT1. How these changes may influence the function of equine SGLT1 awaits analysis by expression in heterologous systems. SGLT1 protein abundance in equine small intestinal BBMV In the equine intestinal BBMV samples, the antibody to SGLT1 cross-reacted with a single band of 75 kDa in all 3 regions of the
400

300

200

100

Competitor Fig 5: Sugar specificity of equine SGLT1. The initial rate of the Na+dependent uptake of 0.1 mmol/l D-glucose into equine duodenal BBMV was measured at 38C in the presence of 1 mmol/l of the indicated competitor. Results are expressed as mean s.e. (n = 3).

354

Molecular characterisation of carbohydrate digestion

1 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (1) (1) (1) (1) (1) (1) (1) 1 51 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (51) (51) (51) (51) (51) (51) (51) 2 101 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (101) (101) (101) (101) (101) (101) (101) 3 151 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (151) (151) (151) (151) (151) (151) (151) 4 201 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (201) (201) (201) (201) (201) (201) (201) 5 251 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (251) (251) (251) (251) (251) (251) (251) 6 301 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (301) (301) (301) (301) (301) (301) (301) 7

50

100

150

200

250

*
300

350

Fig 6: Alignment of the deduced amino acid sequence of eSGLT1 with SGLT1 from various species. Sequence alignment was carried out by Clustal W analysis using the alignment programme of Vector NTi14. Homologous regions are shaded. The boxed area indicates the region to which the anti-peptide antibody was raised. The conserved glycosylation site at Asn248 is indicated (*). Putative membrane spanning regions (Hediger and Rhoads 1994) are indicated by lines below the sequence and are numbered 1 to 12.

J. Dyer et al.

355

351 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (351) (351) (351) (351) (351) (351) (351) 8 401 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (401) (401) (401) (401) (401) (401) (401) 9 451 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (451) (451) (451) (451) (451) (451) (451) 10 501 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (501) (501) (501) (501) (501) (501) (501) 12 551 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (551) (551) (551) (551) (551) (551) (551) 601 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (598) (600) (601) (600) (598) (600) (601) 651 Horse SGLT1 Human SGLT1 Mouse SGLT1 Rabbit SGLT1 Rat SGLT1 Sheep SGLT1 Consensus (648) (650) (651) (648) (651) (650) (651) 665 11

400

450

500

550

600

650

Fig 6 continued: For details see previous page.

356

Molecular characterisation of carbohydrate digestion

a)

750 b)

a)

1500 1250

b)

75 kDa

500

7 kb 4.4 kb

1000 750 500 250

250

2.4 kb

0 D J I
D J I

Fig 7: SGLT1 protein abundance in BBMV isolated from equine small intestine. BBMV (15 g protein/lane) prepared from equine duodenum (D), jejunum (J) and ileum (I) were separated on 8% polyacrylamide gels, electrotransferred to nitrocellulose membrane and western blotted using an anti-SGLT1 polyclonal antibody, as described in Materials and methods. a) The results of a typical western blot indicating the presence of the 75 kDa SGLT1 protein. b) Densitometric analysis of western blots carried out using Phoretix 1D12. Data are expressed as mean s.e. (n = 7) in arbitrary units.

Fig 8: SGLT1 mRNA expression in equine small intestinal mucosa. Total RNA (3 g per lane) prepared from equine duodenal (D), jejunal (J) and ileal (I) mucosal scrapings was separated on 1% agarose denaturing gels, transferred to nylon membrane and northern blotted using 32P-labelled eSGLT1 cDNA probe, as described in Materials and methods. a) The results of a typical northern blot indicating the presence of 3 SGLT1 transcripts. b) Densitometric analysis of northern blots carried out using Phoretix 1D12. Data are expressed as mean s.e. (n = 7) in arbitrary units.

small intestine (Fig 7a), a molecular mass consistent with the size of SGLT1 in other species (Lescale-Matys et al. 1993; ShiraziBeechey 1996; Dyer et al. 1997b; Wood et al. 2000). Densitometric analyses of immunoreactive bands measured in BBMV from the duodenum, jejunum and ileum of several grazing horses are shown in Figure 7b. The profile of SGLT1 protein expression along the length of equine small intestine correlated with Na+-dependent D-glucose transport rates, with highest levels in the duodenum decreasing distally. These data further support that the proximal to distal decline is due to decrease in the level of functional Na+/glucose cotransporter protein in the brush-border membrane. Northern blot analysis of SGLT1 mRNA in equine intestine equine SGLT1 cDNA (nucleotides 14062025) hybridised to 3 transcripts of 7, 4.4 and 2.4 kb in equine intestinal RNAsamples (Fig 8a), with the 7 and 2.4 kb transcripts being the most abundant. Densitometric analyses of the total area of all transcripts indicate that SGLT1 mRNAlevels were highest in the duodenum and decreased distally. This pattern of expression correlates with the SGLT1 activity and protein abundance. The presence of multiple transcripts for SGLT1 is usual, with many species having more than one and some ruminant species having up to 5 (Wood et al. 2000). This difference in size appears to be due to variation in the length of the 3 untranslated region (Peng and Lever 1995; Tarpey et al. 1995; Wood et al. 2000). The results of these studies indicate collectively that SGLT1 is expressed in the equine small intestine and that SGLT1 expression is highest in the proximal intestine and decreases distally. The high degree of correlation between SGLT1 protein function and abundance with SGLT1 mRNA level suggests that, in mature grazing horses, SGLT1 expression is regulated at the level of mRNAabundance. Discussion Little information has been available on the mechanisms and membrane proteins involved in the digestion and absorption of dietary carbohydrates in equine small intestine. Equally, little is known about the sites of carbohydrate digestion and absorption
32P-labelled

along the equine intestinal tract. Alexander (1955) showed that the absorption of glucose takes place in the small intestine of the horse and that this nutrient can fulfil some basal energy requirements. Alexander and Chowdhury (1958) detected the activity of lactase, maltase and amylase in the ileal juice sampled from one pony with an ileal cannula. In the present study, we have determined the level of activity and the abundance of disaccharidases measured in the intestinal brush-border membranes of a large number of mature horses. We have shown that sucrase, maltase and lactase are expressed in the equine intestinal brush-border membrane. Sucrase activity is highest in the proximal small intestine of the horse and levels are similar to those reported in the intestine of other nonruminant species (Kessler et al. 1978; Shirazi et al. 1981; Shirazi-Beechey et al. 1989; Dyer et al. 1997b). Maltase is expressed all along the equine small intestine. Maltase activity is similar in proximal, mid and distal regions and is extremely high compared with other species; 1 mol/min/mg protein in the horse (this study) vs. 0.013-0.054 mol/min/mg wet weight tissue in man (Dahlqvist 1964; Dyer et al. 1997b) and 0.014 mol/min/mg protein in sheep (Shirazi-Beechey et al. 1989, 1991b). The brush-border membrane lactase is the enzyme responsible for the breakdown of the milk sugar, lactose. While the majority of animals and human subjects can digest lactose in suckling life, large populations of non-Caucasians and some animals become lactase deficient in adulthood (Henning et al. 1994). Consumption of lactose by these populations leads to intestinal disturbances and diarrhoea. The expression of lactase is preprogrammed (genetically controlled) and is not diet-induced (Duluc et al. 1991; Shirazi-Beechey et al. 1991b; Henning et al. 1994; Dyer et al. 1997b). In the studies carried out by Roberts (1975), it was reported that, in horses older than age 3 years, lactose feeding did not produce an increase in plasma glucose levels and induced passing of loose faeces. Based on these studies, he concluded that mature horses do not tolerate lactose. In spite of Roberts conclusions, some horse-owners feed milk pellets to mature horses, particularly in an attempt to improve the weight and condition of show animals. Our studies indicate clearly that functional lactase is expressed in the intestinal tissue obtained from mature horses. This has been shown by direct measurements of the activity and abundance of lactase in the

J. Dyer et al.

357

purified brush-border membranes isolated from proximal, mid and distal regions of the intestine, indicating that these horses would have the ability to digest lactose. Our data support the conclusion made by Argenzio (1975) that the incidence of disaccharidase deficiency in horses is rare. Analysis of brushborder membrane vesicles prepared from the intestine of weaned, immature (age <4 years), horses demonstrated the presence of much higher lactase levels (data not shown) indicating that, although lactase expression in equine small intestine does decrease in maturity, it is not lost altogether. Further work is underway in our laboratory to investigate the developmental changes in equine small intestinal digestive and absorptive function. We have shown that glucose and galactose are transported in the equine intestine by a Na+/glucose cotransporter isoform 1 (SGLT1). cDNA encoding SGLT1 has been cloned from the intestinal tissue of various species (Hediger et al. 1987; Hediger and Rhoads 1994; Shirazi-Beechey 1995; Tarpey et al. 1995; Turk and Wright 1997) and the amino acid sequences have been deduced. The comparison of SGLT1 sequences in various species indicates that this protein is conserved throughout evolution (Wright 1993; Shirazi-Beechey 1996). Despite the high percentage of identity in the amino acid sequences, there are some minor differences. These differences in the amino acid sequence have been shown to have a profound effect on the kinetic properties of the transporters. The differences in the amino acid sequence of SGLT1 in various species not only provide information on the relationship between structure and function, but also an insight into the differences in physiological environment encountered by SGLT1 in diverse species. We have shown that SGLT1 in the intestine of horses consuming grass-based diets is expressed with highest abundance in the proximal intestine and the protein has a high affinity and a low capacity for the sugar substrates. This has physiological significance in the functioning of SGLT1 in the small intestine. The identification of the regions of the gut involved in the digestion and absorption of dietary sugars will be important, particularly to animals having undergone gastrointestinal surgery when they are maintained subsequently on diets containing soluble digestible carbohydrates. Small intestinal resection (SIR) is commonly performed in the surg i c a l treatment of equine colic (Tate et al. 1983), but little is known about the long-term effects of SIR on horse and ponies. Knowledge of the regional functional characteristics of the equine intestine will allow a more accurate prediction of the potential dietary complications the animal may experience during recovery and ways in which the diet can be supplemented during this critical period. The data presented in this paper provide baseline information for a better understanding of the pathogenesis of dietary-induced intestinal disease and improved feed management in the horse. Acknowledgements The financial support of the Home of Rest for Horses and the Horserace Betting Levy Board is gratefully acknowledged. We thank Dr S. Wood for his constructive comments. We are grateful to Professor D. Swallow for the gift of the antibody to lactase. A preliminary report of this work was presented as an oral communication at the British Equine Veterinary Association Congress 2000, Birmingham, UK.

Manufacturersaddresses
Lewes, East Sussex, UK. Laboratory Products Ltd., Bishops Stortford, Hertfordshire, UK. 3 Millipore UK Ltd., Watford, Hertfordshire, UK. 4Qiagen Ltd., Crawley, West Sussex, UK. 5Invitrogen Life Technologies Ltd., Paisley, UK. 6Promega UK Ltd., Southampton, UK. 7Helena Biosciences, Sunderland, Tyne and Wear, UK. 8Stratagene (Europe) Ltd., Amsterdam, The Netherlands. 9BioRad Laboratories Ltd., Hemel Hempstead, Hertfordshire, UK. 10Amersham-Pharmacia Biotechnology Lrd., Little Chalfont, Buckinghamshire, UK. 11Kodak Ltd., Hemel Hempstead, Hertfordshire, UK. 12Non-linear Dynamics Ltd., Newcastle Upon Tyne, Tyne and Wear, UK. 13Sigma-Aldrich Co. Ltd., Poole, Dorset, UK. 14Informax Inc., Oxford, Oxfordshire, UK.
2Kendro 1Boehringer-Mannheim,

References
Alexander, F. (1955) Factors affecting the blood sugar concentration in horses. Quart. J. expt. Physiol. 40, 24-31. Alexander, F. and Chowdhury, A.K. (1958) Enzymes in the ileal juice of the horse. Nature 181, 190. Argenzio, R.A. (1975) Functions of the equine large intestine and their interrelationships in disease. Cornell Vet. 65, 303-330. Argenzio, R.A. and Hintz, H.F. (1972) Effect of diet on glucose entry and oxidation rates in ponies. J. Nutr. 102, 879-892. Dahlqvist, A. (1964) Method for assay of intestinal disaccharidases. Analyt. Biochem. 7, 18-25. Diez-Sampedro, A., Eskandari, S., Wright, E.M. and Hirayama, B.A. (2001) Na +-tosugar stoichiometry of SGLT3. Am. J. Physiol. Renal Physiol. 49, F278-F282. Duluc, I., Boukamel, R., Mantei, N., Semenza, G., Raul, F. and Freund, J.N. (1991) Sequence of the precursor of intestinal lactase-phlorizin hydrolase from fetal rat. Gene 103, 275-276. Dyer, J., Barker, P.J. and Shirazi-Beechey, S.P. (1997a) Nutrient regulation of the intestinal Na +/glucose co-transporter (SGLT1) gene expression. Biochem. Biophys. res. Comm. 230, 624-629. Dyer, J., Hosie, K.B. and Shirazi-Beechey, S.P. (1997b) Nutrient regulation of human intestinal sugar transporter (SGLT1) expression. Gut 41, 56-59. Fell, B. (1961) Cell shedding in the epithelium of the intestinal mucosa-fact and artifact. J. path. Bact. 81, 251-254. Hediger, M.A. and Rhoads, D.B. (1994) Molecular physiology of sodium-glucose cotransporters. Physiol. Rev. 74, 993-1026. Hediger, M.A., Coady, M.J., Ikeda, T.S. and Wright, E.M. (1987) Expression cloning and cDNAsequencing of the Na +/glucose co- transporter. Nature 330, 379-381. Henning, S., Rubin, D. and Shulman, R. (1994) Ontogeny of the intestinal mucosa. In: Physiology of the Gastrointestinal Tract, Ed: L. Johnson, Raven Press, New York. pp 571-610. Hintz, H.F. and Cymbaluk, N.F. (1994) Nutrition of the horse. Ann. rev. Nutr. 14, 243267. Hirayama, B.A. and Wright, E.M. (1992) Glycosylation of the rabbit intestinal brush border Na+/glucose cotransporter. Biochim. Biophys. Acta 1103, 37-44. Hirayama, B.A., Lostao, M.P., Panayotova-Heiermann, M., Loo, D.D., Turk, E. and Wright, E.M. (1996) Kinetic and specificity differences between rat, human and rabbit Na+- glucose cotransporters (SGLT-1). Am. J. Physiol. 270, G919-G926. Hopfer, U. (1987) Membrane transport mechanisms for hexose and amino acids in the small intestine. In: Physiology of the Gastrointestinal Tract, Ed: L. Johnson, Raven Press, New York. pp 499-1526. Kessler, M., Acuto, O., Storelli, C., Murer, H., Muller, M. and Semenza, G. (1978) A modified procedure for the rapid preparation of efficiently transporting vesicles from small intestinal brush border membranes. Their use in investigating some properties of D-glucose and choline transport systems. Biochim. Biopys. Acta 506, 136-154. Lescale-Matys, L., Dyer, J., Scott, D., Freeman, T.C., Wright, E.M. and ShiraziBeechey, S.P. (1993) Regulation of the ovine intestinal Na+/glucose cotransporter (SGLT1) is dissociated from mRNA abundance. Biochem. J. 291, 435-440. Malo, C. and Berteloot, A. (1991) Analysis of kinetic data in transport studies: new insights from kinetic studies of Na+-D-glucose cotransport in human intestinal brush-border membrane vesicles using a fast sampling, rapid filtration apparatus. J. membr. Biol. 122, 127-141. Mantei, N., Villa, M., Enzler, T., Wacker, H., Boll, W., James, P., Hunzicker, W. and

358

Molecular characterisation of carbohydrate digestion

Semenza, G. (1988) Complete primary structure of human and rabbit lactasephlorizin hydrolase: implications for biosynthesis, membrane anchoring and evolution of the enzyme. Embo. J. 7, 2705-2713. Peng, H. and Lever, J.E. (1995) Regulation of Na+-coupled glucose transport in LLCPK1 cells. Message stabilization induced by cyclic AMP elevation is accompanied by binding of a M(r) = 48,000 protein to a uridine-rich domain in the 3-untranslated region. J. biol. Chem. 270, 23996-24003. Roberts, M.C. (1975) Carbohydrate digestion and absorption studies in the horse. Res. vet. Sci. 18, 64-69. Rowell, A., Dyer, J., Hofmann, R.R. and Shirazi-Beechey, S.P. (1997) Expression of Na+ /glucose cotransporter in the intestinal brush-border membrane of ruminants with different feeding habits. Biochem. Soc. Trans. 25, 482S. Schlegel-Haueter, S., Hore, P., Kerry, K.R. and Semenza, G. (1972) The preparation of lactase and glucoamylase of rat small intestine. Biochim. Biophys. Acta 258, 506-519. Shirazi, S.P., Beechey, R.B. and Butterworth, P.J. (1981) Potent inhibition of membrane-bound rat intestinal alkaline phosphatase by a new series of phosphate analogues. Biochem. J. 194, 797-802. Shirazi-Beechey, S.P. (1995) Molecular biology of intestinal glucose transport. Nutr. res. Rev. 8, 27-41. Shirazi-Beechey, S.P. (1996) Intestinal sodium-dependent D-glucose co-transporter: dietary regulation. Proc. Nutr. Soc. 55, 167-178. Shirazi-Beechey, S.P., Gorvel, J.-P. and Beechey, R.B. (1988) Phosphate transport in intestinal brush-border membrane. J. Bioenerg. Biomembr. 20, 273-288 Shirazi-Beechey, S.P., Kemp, R.B., Dyer, J. and Beechey, R.B. (1989) Changes in the functions of the intestinal brush border membrane during the development of the ruminant habit in lambs. Comp. Biochem. Physiol. [B] 94, 801-806.

Shirazi-Beechey, S.P., Hirayama, B.A., Wang, Y., Scott, D., Smith, M.W. and Wright, E.M. (1991a) Ontogenic development of lamb intestinal sodium-glucose cotransporter is regulated by diet. J. Physiol. (Lond.) 437, 699-708. Shirazi-Beechey, S.P., Smith, M.W., Wang, Y. and James, P.S. (1991b) Postnatal development of lamb intestinal digestive enzymes is not regulated by diet. J. Physiol (Lond.) 437, 691-698. Shirazi-Beechey, S.P., Davies, A.G., Tebbutt, K., Dyer, J., Ellis, A., Taylor, C.J., Fairclough, P. and Beechey, R.B. (1990) Preparation and properties of brushborder membrane vesicles from human small intestine. Gastroenterol. 98, 676685. Tarpey, P.S., Wood, I.S., Shirazi-Beechey, S.P. and Beechey, R.B. (1995) Amino acid sequence and the cellular location of the Na+-dependent D- glucose symporters (SGLT1) in the ovine enterocyte and the parotid acinar cell. Biochem. J. 312, 293-300. Tate Jr., L.P., Ralston, S.L., Koch, C.M. and Everitt, J.I. (1983) Effects of resection of the small intestine in the pony. Am. J. vet. Res. 44, 1187-1191. Turk, E. and Wright, E.M. (1997) Membrane topology motifs in the SGLT cotransporter family. J. membr. Biol. 159, 1-20. Wood, I.S., Dyer, J., Hofmann, R.R. and Shirazi-Beechey, S.P. (2000) Expression of the Na +/glucose co-transporter (SGLT1) in the intestine of domestic and wild ruminants. Pflugers Archiv. European J. Physiol. 441, 155-162. Wright, E.M. (1993) The intestinal Na+/glucose cotransporter. Ann. rev. Physiol. 55, 575-589. Wright, E.M., Hirayama, B.A., Loo, D.D., Turk, E. and Hager, K. (1994) Intestinal sugar transport. In: Physiology of the Gastrointestinal Tract, 3rd edn. Ed: L.R. Johnson. Raven Press, New York. pp 1751-1772.

Received for publication: 8.1.01 Accepted: 26.6.01