The FASEB Journal express article 10.1096/fj.02-0031fje. Published online November 15, 2002.
The role of indigenous microflora in the development of murine intestinal fucosyl- and sialyltransferases N. Nanda Nanthakumar,* Dingwei Dai,*,† David S. Newburg,‡ and W. Allan Walker* *Developmental Gastroenterology Laboratory, Combined Program in Pediatric Gastroenterology and Nutrition, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts; †Shanghai Institute for Pediatric Research, Xinhua Hospital and Shanghai Second Medical University, China; ‡Program in Glycobiology, Shriver Center for Mental Retardation, Waltham, Massachusetts Corresponding author: W. Allan Walker, Developmental Gastroenterology Laboratory, Combined Program in Pediatric Gastroenterology and Nutrition, Massachusetts General Hospital, 114 16th Street (114-3503), Charlestown, MA 02129, USA. E-mail:
[email protected] ABSTRACT Most enteric bacteria use intestinal brushborder glycoconjugates as their target host cell receptors. It has been postulated that resident microbes regulate specific glycosyltransferases that are responsible for synthesizing brushborder glycoconjugates. To investigate this hypothesis, we measured glycosyltransferase enzyme activities in intestine from different regions of maturing conventional (CONV), germ-free (GF), and ex-germ-free (XGF) mice and compared them to general enzyme markers of gut development, for example, disaccharidases. High α2,3/6Sialyltransferase (ST) activity and low α1,2-fucosyltransferase (FT) activities were detected from duodenum to colon in suckling CONV mice, but the relative levels of these activities reversed during the third postnatal wk, rapidly reaching adult levels by the fourth wk. These agerelated enzyme changes were significantly attenuated in GF mice, maintaining an immature pattern well past 3 wk. Introduction of gut microflora in GF mice rapidly initiated maturation of glycosyltransferase activity but had no significant affect on developmental programming of dissacharidases. Therefore, in mice, intestinal glycosyltransferase activities are under tissue and developmental control and microflora play a major role in their specific ontogeny but not in overall development. These findings may help explain the regional specificity of commensal bacteria and of enteric pathogens and may also relate age-related changes in microflora to susceptibility to enteropathogens. Key words: luminal bacteria • germ-free mice • postnatal development • brush border membrane glycoconjugates • extrinsic regulation
C
arbohydrates on cell-surface glycoproteins and glycolipids play key roles in cell-cell recognition, as antigenic determinants, and in interactions with the extracellular environment. They serve as binding determinants for a variety of endogenous ligands
(e.g., hormones and onco-developmental antigens) as well as exogenous ligands (e.g., microbial adhesins, bacterial toxins, and viral hemagglutinins) (1–5). During postnatal development, the sialic acid/fucose ratio of surface carbohydrates in the rodent intestinal epithelium reverses from a predominance of sialic acid to a predominance of fucose (8, 9). This inversion during the third postnatal wk has been attributed to a reciprocal change in sialyltransferase and fucosyltransferase activities (8–10). This diminished sialylation and increased fucosylation causes the suckling and adult gut to express different isoforms of membrane glycoproteins and glycolipids. However, these developmental changes also coincide with weaning (11), which is accompanied by changes in luminal microbes. We have been investigating whether colonization by commensal bacteria could contribute to the poorly understood underlying mechanisms controlling enzyme activity in the gut, especially with regard to regional specificity, and whether this regulation is luminal microbial specific or generalized. To elucidate the interactions, that is, the “cross-talk,” between microflora and host intestinal epithelial cells, the intestine’s microbial ecology must be controlled and manipulated. Mice raised under germ-free (GF) conditions, that is, without any resident microorganisms, provide a critical technology for manipulating and simplifying this ecosystem (2, 17, 19). Comparisons between mice with normal microbiota and mice maintained in a GF state indicate that maturation of the intestine is affected by interactions with its resident microbiota (3, 18, 19). Without microflora, the rate of epithelial cell renewal in the small intestine is reduced, the cecum becomes massively enlarged, and the composition and spatial organization of the diffuse gutassociated lymphoid tissue (GALT) are altered (5, 6, 19, 29). Colonization of GF mice with a mixture of fecal microorganisms obtained from species-matched conventional (CONV) mice led to alteration in luminal glycoconjugates (20) and induction of GDP-fucose asialo-GM1 α1,2fucosyltransferase in the epithelium of the small intestine; levels of fucosyl asialo-GM1 increased whereas asialo- GM1 decreased (21). Colonization of GF mice with Bacteriodes thetaiotamicron restored FT mRNA and the expression of fucosylated glycoconjugates in the small intestinal epithelium to normal levels (3, 17). Colonization with segmented filamentous bacteria (a subset of indigenous intestinal microbes) recapitulated these biochemical changes in the host and also restored some features of the diffuse GALT of mice with complete CONV microflora (3, 18, 19). However, very little is known about whether indigenous microflora specifically affect the ontogeny of other glycosyltransferase activities or the regional specificity of such control and whether this affect is specific or part of a generalized effect on developmental patterns. Accordingly, the goals of this study were to investigate tissue specificity in the regulation of intestinal and colonic glycosyltransferase activities by microflora and compare this effect to modifications of classic developmental markers, for example, disaccharidases (11, 22). The activities of α2,3/6-sialyltransferase (CMP-NeuAc: βGal α2,3/6-sialyltransferase) and FT (GDPFuc: βGal α1,2-fucosyltransferase) were measured along with sucrase and lactase in the duodenum, jejunum, ileum, and colon of CONV, GF, and ex-germ-free (XGF) mice during development. MATERIALS AND METHODS Chemicals
Ultra-pure D-glucose, sucrose, lactose, and bovine serum albumin (BSA) were obtained from the Sigma (St. Louis, MO). Protein concentrations were measured using a BCL kit (Pierce, Rockford, IL) against BSA standards in a colorimetric assay according to the manufacturer’s protocol. All other chemicals were either of reagent- or molecular-grade from Sigma. Glucose was measured using the Trinder 100 kit from Sigma as described previously (24). All reagents for lectin-flourescence, labeled UEA-1, SNA, MAK II lectins, and anti-Fade were purchased from Vector Lab (Burlington, CA). All radiolabeled precursors for enzyme assays were purchased from New England Nuclear (Boston, MA). Experimental animals CONV female Black Swiss mice were purchased from Taconic Farms (Germantown, NY) at 16– 18 days of gestation or at desired ages. The date of birth of pups was designated as day 0. Animals were housed with their dams under conventional conditions with a 12-h-light/12-h-dark cycle and access to mouse chow and water ad libitum. GF mice of the same strain were purchased from the same vendor at the desired ages. All GF mice were shipped in the morning in a GF isolator and arrived that afternoon at our facility. After an overnight stay in the isolator, the animals were killed immediately after removal from the GF isolator. There was no difference in all assays when the GF animals killed on the same day of arrival or the day after. Conventionalized GF mice (XGF) were produced by removing 4 and 6-wk-old GF mice from their GF environment, inoculating them with intestinal contents from age-matched CONV mice as previously described (19, 25), and keeping them in the same cage as CONV mice. They, along with age-matched GF and CONV mice, were killed 2 wk later. All animals were killed between 12 noon and 3 p.m. Preparation of membrane fractions The entire small intestine and colon were removed and thoroughly flushed with ice-cold 0.9% NaCl for preparation of enzyme fractions as previously described (8, 22). In brief, the small intestine was divided into duodenum, jejunum, and ileum as follows: The small intestine from the stomach to the ligament of Treitz was defined as duodenum, and the proximal and distal halves of the remaining small intestine were defined as jejunum and ileum, respectively. The intestine was placed on a glass plate maintained at 4°C and cut open; the mucosa was harvested by scraping with a microscope glass slide. All subsequent procedures were performed at 4°C. A 10% mucosal homogenate in 0.1 M Tris-HCl buffer (pH 7.4) was centrifuged at 1000g for 15 min to remove nuclei and cellular debris. The supernatant was then centrifuged at 105,000g for 1 h, resulting in the sedimentation of a membrane fraction and soluble cell fluid. The resulting pellets were resuspended in the same buffer used for homogenization, aliquoted, frozen, and stored at –80°C or used immediately for the enzyme assay. Protein determination Protein was determined using a BCA protein assay (Pierce, Rockford, IL) modified for use in 96well microtiter plates according to the manufacturer’s protocol. To each 50 µl protein sample, 200 µl of working reagent was added followed by incubation at 37°C for 30 min. Absorbence at
560 nm was measured on a microtiter plate reader (BT 2000 Microkinetics, Fisher, Pittsburgh, PA). The concentration of each sample was calculated using a standard curve produced with BSA. Sialyltransferase assay α2,3/6-Sialyltransferase activity was measured using asialofetuin as an exogenous acceptor. The reaction mixture, in a total volume of 0.1 ml, contained 800 µg asialofetuin, 12.5 µmol sodium phosphate buffer (pH 6.8), 0.5% Triton X-100, 55 nmol CMP-[14C]NeuAc (0.1 µCi, spec. act. = 1.8 mCi/mmol) (New England Nuclear, Boston, MA) and an appropriate amount of enzyme solution (50–100 µg protein). Incubation was carried out at 37°C for 1 h and terminated by the addition of 0.5 ml 15% trichloroacetic acid containing 5% phosphotungstic acid. The precipitate was collected and washed by rapid vacuum filtration, and the radioactive products were measured as previously described (8, 22). Assays were performed in duplicate with acceptordeficient blanks used to subtract background counts due to any endogenous acceptor. CMPNeuAc was present at saturation concentrations, and product formation was linear for 1 h of incubation time and up to 150 µg enzyme protein under these conditions. The specific activity was expressed as nmol CMP-[14C]NeuAc accepted by asialofetuin/h/mg of protein. Fucosyltransferase assay α1,2-Fucosyltransferase enzyme activity was assessed using phenyl-β-D-galactoside as the exogenous acceptor as previously described with minor modification (8–11). In brief, the reaction mixture for each assay contained, in a total volume of 0.1 ml, 25 mM phenyl-β-Dgalactoside, 20 mM sodium phosphate buffer (pH 6.1), 10 mM fucose, 5 mM ATP, 20 mM MgCl2, 50 mM NaCl, 0.5% Triton X-100, 10 nmol GDP-[14C]fucose (0.1 µCi, spec. act. = 11 mCi/mmol; New England Nuclear), and appropriate amounts of enzyme solution (50–100 µg protein). The concentration of GDP-fucose was at the saturation level, and product formation was linear for 2 h of incubation time for up to 100 µg of enzyme protein in a reaction volume of 0.1 ml. Assays were performed at 37°C and terminated after 2 h by the addition of 100 µl of ethanol followed by dilution with 1 ml of 4°C H2O. The mixture was then centrifuged at 15,000g for 5 min. The supernatant was collected and applied to C-18 Bond Elute cartridges (500 mg) that had previously been washed with 6 ml of acetonitrile followed by 6 ml water. After application of the sample, the cartridges were washed with 5 ml water to remove the radiolabeled precursor. The product, [14C]fucosylphenyl-β-D-galactoside, was eluted with 1.5 ml 50% acetonitrile directly into scintillation vials. Five milliliters of scintillation cocktail (Ready Safe, Beckman, Fullerton, CA) was added to each vial to obtain a clear solution, and the radioactivity was determined by scintillation counting. The specific activity was expressed as nmol GDP[14C]fucose accepted by phenyl-β-D-galactoside/h/mg of protein. Disaccharidase activities The small intestine was divided into duodenum, jejunum, and ileum, and the lumen of each section was rinsed with sterile cold phosphate-buffered saline (PBS) before freezing at –20oC for subsequent analysis. On the day of analysis, the tissues were homogenized in 9 vol of 0.154 M KCl using a Potter-Elvehjem Teflon-glass homogenizer. The tissue homogenate was then
assayed in duplicate using sucrose and lactose as substrates as described previously (24). The amount of glucose liberated by these disaccharidases was quantitated using a glucose-oxidase method. Protein concentrations were measured using a BCL kit (Pierce) against BSA standards in a colorimetric assay according to the manufacturer’s protocol. Enzyme activity was expressed as micromoles of substrate hydrolyzed per hour per milligram of total protein. Analysis of glycoconjugate expression by lectin-fluorescence Analysis of α1,2-fucosyl-glycoconjugates was performed using Ulex europaeus agglutinin (UEA-1), α2,6-sialyl-glycoconjugates using FITC-conjugated Sambucus nigra agglutinin (SNA), and α2,3-sialyl-glycoconjugates using Maackia amurensis lectin II (MAK II) on frozen tissue sections as previously described (22). In brief, 1 cm tissue from each region of the intestine was fixed for 2 h in 4% paraformaldehyde at 4oC, washed in ice-cold PBS containing 30% sucrose overnight at 4oC, and embedded in OCT. Frozen sections (6–7 µm thick) were blocked with PBS containing 2% BSA, then stained with labeled lectins for 1 h following the manufacturer’s suggested concentration and recommended buffer. Tissue sections were then washed three times in cold PBS, mounted using Anti-Fade, and analyzed by fluorescent and confocal microscopy when necessary. Statistical analysis Results are expressed as the mean ±SE. Effects of age and treatment on enzyme activities were analyzed by two-way ANOVA. After overall significance was confirmed, post hoc tests for individual variables were performed by a two-tailed unpaired t test. Differences of P
