Fibrosis quística en imágenes. Clasificación de Bhalla para la tomografía computarizada en pacientes pediátricos / Cystic fibrosis in images: the Bhalla scoring system for computed tomography in paediatric patients

Para el seguimiento de la evolución del daño pulmonar en los pacientes con fibrosis quística (FQ) se han desarrollado distintos sistemas de clasificación o puntuación mediante la radiografía de tórax y la tomografía computarizada de alta resolución (TCAR). La TCAR es una técnica más sensible que la radiografía de tórax para evaluar la anatomía, sin embargo, en pacientes pediátricos el uso de la tomografía computarizada (TC) debe reducirse al mínimo necesario y utilizar sistemas de protección radiológica y reducción de dosis. Uno de los sistemas de clasificación mediante TCAR más empleados es el propuesto por Bhalla en 1991, que facilita la evaluación de la gravedad y la evolución de estos pacientes en función de los distintos hallazgos radiológicos. Presentamos diferentes ejemplos de estos criterios para la TCAR, observados al revisar los estudios de un grupo de 48 pacientes (AU)

To monitor the course of lung damage in patients with cystic fibrosis (CF) using the different chest X-ray and high resolution computed tomography (HRCT) scoring systems that have been developed. The HRCT technique is more sensitive than chest radiography to evaluate the anatomy. However, in paediatric patients, the use of CT should be kept to a minimum, and guidelines for radiation protection and dose reduction should be applied. One of the most used classification systems for HRCT is the one proposed by Bhalla in 1991, which helps in the assessment of the severity and course of the disease in these patients depending on the different imaging findings. We present various examples of these criteria for HRCT, observed while reviewing a group of 48 paediatric patients (AU)

[Cystic fibrosis in images: the Bhalla scoring system for computed tomography in paediatric patients]. / Fibrosis quística en imágenes. Clasificación de Bhalla para la tomografía computarizada en pacientes pediátricos.

To monitor the course of lung damage in patients with cystic fibrosis (CF) using the different chest X-ray and high resolution computed tomography (HRCT) scoring systems that have been developed. The HRCT technique is more sensitive than chest radiography to evaluate the anatomy. However, in paediatric patients, the use of CT should be kept to a minimum, and guidelines for radiation protection and dose reduction should be applied. One of the most used classification systems for HRCT is the one proposed by Bhalla in 1991, which helps in the assessment of the severity and course of the disease in these patients depending on the different imaging findings. We present various examples of these criteria for HRCT, observed while reviewing a group of 48 paediatric patients.

Glutaminase activity is confined to the mantle of the islets of Langerhans.

Glutamatergic signalling plays an important role in the coordination of hormone secretion from the endocrine pancreas. Thus, glutamate production must be a process exquisitely regulated to ensure a proper transmitter function. Recently we have reported that the endocrine pancreas co-expresses two isoforms of the protein glutaminase (GA), denoted as kidney-type (KGA) and liver-type (LGA). However, how GA activity, and therefore glutamate production, is regulated in the islets represents a critical issue that remains unresolved. Since the purification of these enzymes from rat islets is a daunting task, in order to characterize each isoform we have taken advantage of the spatial segregation of these isoenzymes in pancreas. To assist us with this goal, we have developed a new procedure that enables us to assay GA activity in situ. The assay is highly specific for GA as indicated by its dependence on glutamine and orthophosphate. Surprisingly, LGA, which is abundantly expressed by beta-cells, did not show detectable activity under the assay conditions. All the GA activity detected in pancreatic islets was attributed to KGA and was confined to the mantle of the islets. Double labelling analyses strongly suggested that alpha-cells should be regarded as the site of glutamate production in the endocrine pancreas.

Overexpression, purification, and characterization of glutaminase-interacting protein, a PDZ-domain protein from human brain.

A human brain cDNA clone coding for a novel PDZ-domain protein of 124 amino acids has been previously isolated in our laboratory. The protein was termed GIP (glutaminase-interacting protein) because it interacts with the C-terminal region of the human brain glutaminase L. Here we report the heterologous expression of GIP as a histidine-tagged fusion protein in Escherichia coli cells. The induction conditions (temperature and isopropyl beta-d-thiogalactopyranoside concentrations) were optimized in such a way that GIP accounted for about 20% of the total E. coli protein. A simple and rapid procedure for purification was developed, which yielded 17 mg of purified GIP per liter of bacterial cell culture. The apparent molecular mass of the protein by SDS-PAGE was 16 kDa, whereas in native form it was determined to be 28 kDa, which suggests dimer formation. The nature and integrity of the recombinant protein were verified by mass spectrometry analysis. The functionality of the GIP protein was tested with an in vitro activity assay: after being pulled down with glutathione S-transferase-glutaminase, GIP was revealed by Western blot using anti-GIP antibodies. Furthermore, the glutaminase activity in crude rat liver extracts was inhibited by the presence of recombinant purified GIP protein.

The C-terminus of human glutaminase L mediates association with PDZ domain-containing proteins.

The enzyme glutaminase in brain is responsible for the synthesis of neurotransmitter glutamate. We used the two-hybrid genetic selection system in yeast to look for interactors of glutaminase in human brain. We have identified two proteins containing PDZ domains, alpha1-syntrophin and a glutaminase-interacting protein, named GIP, that showed association with human glutaminase L, as deduced from specificity test of the two-hybrid system. The complete GIP cDNA clone has 1315 nucleotides with a 372-base open reading frame encoding a 124-amino acids protein. Glutaminase associates with both PDZ proteins through its C-terminal end; mutagenesis of single amino acids revealed the sequence -ESXV as essential for the interaction. These data suggest the possibility that PDZ domain-containing proteins are involved in the regulation of glutaminase in brain.

Identification of two human glutaminase loci and tissue-specific expression of the two related genes.

Glutaminolysis is initiated by either of two isoforms, K- and L-types, of the enzyme phosphate-activated glutaminase. The chromosomal localization, genomic organization, and the tissue-specific expression of the genes have been investigated in the human by using isoform-specific cDNA probes. Results obtained from radiation hybrid mapping experiments assigned the K-glutaminase gene to human Chromosome (Chr) 2, and a second locus for L-glutaminase in Chr 12 was identified. Southern blot analysis with the L-cDNA probe showed hybridization to a single restriction fragment, while four to seven fragments were found to hybridize to the K-cDNA probe. The distribution of human glutaminase expression was also investigated: the L-cDNA probe detected a single band of 2.4 kb in liver, brain, and pancreas, whereas a single transcript of approximately 4.4 kb was detected in kidney, brain, heart, placenta, lung, and pancreas by using the K-cDNA probe. This work provides evidence that the human liver and kidney glutaminase isozymes are encoded by separate genes located on different chromosomes; furthermore, the expression pattern in human tissues revealed for both isoenzymes differs notably from the paradigm based upon the isoenzyme distribution in rats.

Changes in mRNAs for enzymes of glutamine metabolism in the tumor-bearing mouse.

Changes in the relative mRNA levels of phosphate-activated glutaminase (PAG) and glutamine synthetase (GS) in the liver and kidney of mice bearing a highly malignant strain of Ehrlich ascites tumor cells were determined at different days after tumor transplantation. Kidney glutaminase mRNA steadily increased, reaching maximum values at day 10 of tumor growth, while those of glutamine synthetase did not change, resulting in a sustained decrease of the GS/PAG ratio in the kidneys of tumor-bearing animals compared with controls. However, the GS/PAG ratio in the liver significantly increased, mainly due to a strong decrease in PAG, whereas GS mRNA levels remained almost unaffected. These results, combined with those previously reported on enzymatic activities and glutamine concentrations in the host-tumor system, suggest a long-term regulation of the host glutaminase enzymes in order to increase the circulating glutamine levels needed for tumor growth.

Inhibition of glutaminase expression by antisense mRNA decreases growth and tumourigenicity of tumour cells.

Phosphate-activated glutaminase has a critical role in tumours and rapidly dividing cells and its activity is correlated with malignancy. Ehrlich ascites tumour cells transfected with the pcDNA3 vector containing an antisense segment (0.28 kb) of rat kidney glutaminase showed impairment in the growth rate and plating efficiency, as well as a shortage in the glutaminase protein and activity. The C-terminal segment used is well conserved in all glutaminase sequences known. The transfected cells, named 0.28AS-2, displayed remarkable changes in their morphology compared with the parental cell line. The 0.28AS-2 cells also lost their tumourigenic capacity in vivo. Control mice developed an ascitic tumour, with a lifespan of 16+/-1 days, when inoculated with 10(7) cells/mouse; on the contrary, animals inoculated with transfected cells up to 2.5 times the cell numbers of control mice did not develop tumours and behaved as healthy animals. The ability to revert the transformed phenotype of antisense-transfected cells confirms the relevance of glutaminase in the transformation process and could provide new ways for the study of gene therapy.

Molecular cloning, sequencing and expression studies of the human breast cancer cell glutaminase.

Phosphate-activated glutaminase (GA) is overexpressed in certain types of tumour but its exact role in tumour cell growth and proliferation is unknown. Here we describe the isolation of a full-length cDNA clone of human breast cancer ZR75 cells, by a combination of lambdagt10 cDNA library screening and the rapid amplification of cDNA ends ('RACE') technique. The cDNA of human GA is 2408 nt with a 1806-base open reading frame encoding a 602-residue protein with a predicted molecular mass of 66309 Da. The deduced amino acid sequence contains a putative mitochondrial import presequence of 14 residues at the N-terminal end. Heterologous expression and purification in Escherichia coli yielded a product of the expected molecular size that was recognized by using antibodies against the recombinant human GA. Sequence analyses showed that human GA was highly similar to the rat liver enzyme. Northern gel analysis revealed that the gene is present in human liver, brain and pancreas, in which a major transcript of 2.4 kb was demonstrated, but not in kidney, heart, skeletal muscle, lung or placenta. These results strongly suggest that the first human GA cloned, the GA from ZR-75 breast cancer cells, and presumably those from human liver and brain, are liver-type isoenzymes, in sharp contrast with the present view that considers the kidney type as the isoform expressed in all tissues with GA activity, with the exception of postnatal liver.

Upregulation of glyceraldehyde-3-phosphate dehydrogenase mRNA in the spleen of tumor-bearing mice.

The influence of tumor implantation on glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels and stability was determined in the spleen of tumor-bearing mice. While GAPDH mRNA levels were not altered in skeletal muscle, kidney and liver from tumor-bearing mice, tumor implantation led to a 5.6-fold increase in the levels of splenic GAPDH mRNA. An enhanced message stability was observed in splenocytes from tumor-bearing animals, suggesting the involvement of post-transcriptional mechanisms in the selective GAPDH mRNA accumulation after tumor implantation. The GAPDH activity/glycolytic flux ratio was 18.5 in the spleen of healthy mice. Therefore, the three-fold increase in the glycolytic flux observed after tumor implantation could hardly justify the necessity for the upregulation of GAPDH.

Involvement of essential cysteine and histidine residues in the activity of isolated glutaminase from tumour cells.

The pH dependence of the phosphate-activated glutaminase isolated from Ehrlich tumour cells suggests a functional role for two prototropic groups with apparent pKa of 9.3 and 7.7 at the active site of the protein; these pKa values are compatible with cysteine and histidine residues, respectively. This possibility was investigated by chemical modification studies of the purified enzyme. N-Ethylmaleimide fully inactivated the purified glutaminase; the reaction order was very close to 1.0, suggesting that N-ethylmaleimide modifies glutaminase at a single essential site. Spectrophotometric studies of the isolated protein treated with diethyl pyrocarbonate indicate that two histidine residues are modified. Since glutaminase is loosely associated to the inner mitochondrial membrane, modification experiments were also carried out using mitochondrial membrane fractions. N-Ethylmaleimide and diethyl pyrocarbonate gave similar results in mitochondria membrane-bound enzyme to those obtained with purified enzyme. Glutamate, which behaves as a competitive inhibitor of the enzyme, partially protected the inactivation caused by N-ethylmaleimide in membrane-bound experiments. The results suggest the existence of a critical histidine residue(s) in the tumour glutaminase, and strongly support the notion that a cysteine residue, which is located at (or near) the active site, is involved in the catalytic mechanism as well.

Early differential expression of two glutaminase mRNAs in mouse spleen after tumor implantation.

The influence of progressive tumor growth on phosphate-activated glutaminase (PAG) expression in splenocytes from mice bearing Ehrlich ascites carcinoma cells was investigated. Implantation of Ehrlich ascites tumor cells in the peritoneal cavity of mice led to a 2.3-fold stimulation of spleen PAG activity 48 h later. Four days after tumor implantation the glutaminase activity had returned to nearly basal value and remained at this level throughout the tumor development. Northern blot analysis indicated that two species of glutaminase mRNA were expressed in the spleen, which showed a differential expression pattern during the first 2 days after tumor implantation. The abundance of the transcript of higher electrophoretic mobility (approximately 3 kb) constantly increased over the first 2 days of tumor growth. The mRNA of lower electrophoretic mobility (approximately 6 kb) peaked at 12 h after tumor implantation and returned to control values at 48 h. These results demonstrate that tumor has the capability of altering glutaminase expression in the host spleen.

Identification and characterization of two distinct intracellular GLUT4 pools in rat skeletal muscle: evidence for an endosomal and an insulin-sensitive GLUT4 compartment.

In skeletal muscle, acute insulin treatment results in the recruitment of the GLUT4 glucose transporter from intracellular vesicular structures to the plasma membrane. The precise nature of these intracellular GLUT4 stores has, however, remained poorly defined. Using an established skeletal-muscle fractionation procedure we present evidence for the existence of two distinct intracellular GLUT4 compartments. We have shown that after fractionation of crude muscle membranes on a discontinuous sucrose gradient the majority of the GLUT4 immunoreactivity was largely present in two sucrose fractions (30 and 35%, w/w, sucrose; denoted F30 and F35 respectively) containing intracellular membranes of different buoyant densities. Here we show that these fractions contained 44+/-6 and 49+/-7% of the crude membrane GLUT4 reactivity respectively, and could be further discriminated on the basis of their immunoreactivity against specific subcellular antigen markers. Membranes from the F30 fraction were highly enriched in transferrin receptor (TfR) and annexin II, two markers of the early endosome compartment, whereas they were significantly depleted of both GLUT1 and the alpha1-subunit of (Na++K+)-ATPase, two cell-surface markers. Insulin treatment resulted in a significant reduction in GLUT4 content in membranes from the F35 fraction, whereas the amount of GLUT4 in the less dense (F30) fraction remained unaffected by insulin. Immunoprecipitation of GLUT4-containing vesicles from both intracellular fractions revealed that TfR was present in GLUT4 vesicles isolated from membranes from the F30 fraction. In contrast, GLUT4 vesicles from the F35 fraction were devoid of TfR. The aminopeptidase, vp165, was present in GLUT4 vesicles from both F30 and F35; however, vesicles isolated from F30 contained over twice as much vp165 per unit of GLUT4 than those isolated from F35. The biochemical co-localization of vp165/GLUT4 was further substantiated by double-immunogold labelling of ultrathin muscle sections. Overall, our data indicate the presence of at least two internal GLUT4 pools: one possibly derived from an endosomal recycling compartment, and the other representing a specialized insulin-sensitive GLUT4 storage pool. Both pools contain vp165.

Submitochondrial localization and membrane topography of Ehrlich ascitic tumour cell glutaminase.

The intramitocondrial localization of the phosphate-activated glutaminase from Ehrlich cells has been examined by a combination of techniques, including: mitochondria subfractionation studies, chemical modification with sulfhydryl group reagents of different permeability, enzymatic digestion in both sides of the inner mitochondrial membrane, and immunological studies. Using alkaline extraction at high ionic strength, hypoosmotic shock and freezing-thawing cycle techniques, the enzyme was found in the particulate fraction. On the contrary, glutaminase activity was labile when subfractionation was carried out by digitonin/lubrol method; Western blot analysis localized the inactive enzyme in the matrix fraction. In addition, glutaminase was fully inactivated when mitoplasts were incubated with phospholipase A2 and phospholipase C. The enzyme also showed a non-linear Arrhenius plot with a break at 24 degrees C. The membrane-impermeant thiol reagents mersalyl and p-chloromercuriphenylsulfonic acid do not inhibit glutaminase activity in freeze-thawed mitochondria and mitoplasts, but N-ethylmaleimide, which is membrane permeant, strongly inhibited the enzyme. However, mersalyl and p-chloromercuriphenylsulfonic acid were effective inhibitors when the alkylation was performed on the matrix side of mitoplasts or using detergent-solubilized enzyme. Furthermore, trypsin digestion of mitoplasts was only effective inactivating glutaminase when the proteolysis was carried out on the matrix side of the vesicles. Enzyme-linked immunosorbent assay of the soluble and membrane fractions obtained in the preparation of submitochondrial particles, revealed that most of the enzyme was solubilized, but in the inactive form. Phase separation with Triton X-114 rendered most of the protein in the aqueous phase. These results taken together discard a transmembrane localization for the protein, whereas they are consistent with anchorage of glutaminase on the matrix side of the inner mitochondrial membrane, the matrix portion of the enzyme being relevant for its function.

Proteolytic cleavage of cellubrevin and vesicle-associated membrane protein (VAMP) by tetanus toxin does not impair insulin-stimulated glucose transport or GLUT4 translocation in rat adipocytes.

Acute insulin stimulation of glucose transport in fat and skeletal muscle occurs principally as a result of the hormonal induced translocation of the GLUT4 glucose transporter from intracellular vesicular stores to the plasma membrane. The precise mechanisms governing the fusion of GLUT4 vesicles with the plasma membrane are very poorly understood at present but may share some similarities with synaptic vesicle fusion, as vesicle-associated membrane protein (VAMP) and cellubrevin, two proteins implicated in the process of membrane fusion, are resident in GLUT4-containing vesicles isolated from rat and murine 3T3-L1 adipocytes respectively. In this study we show that proteolysis of both cellubrevin and VAMP, induced by electroporation of isolated rat adipocytes with tetanus toxin, does not impair insulin-stimulated glucose transport or GLUT4 translocation. The hormone was found to stimulate glucose uptake by approx. 16-fold in freshly isolated rat adipocytes. After a single electroporating pulse, the ability of insulin to activate glucose uptake was lowered, but the observed stimulation was nevertheless nearly 5-fold higher than the basal rate of glucose uptake. Electroporation of adipocytes with 600 nM tetanus toxin resulted in a complete loss of both cellubrevin and VAMP expression within 60 min. However, toxin-mediated proteolysis of both these proteins had no effect on the ability of insulin to stimulate glucose transport which was elevated approx. 5-fold, an activation of comparable magnitude to that observed in cells electroporated without tetanus toxin. The lack of any significant change in insulin-stimulated glucose transport was consistent with the finding that toxin-mediated proteolysis of both cellubrevin and VAMP had no detectable effect on insulin-induced translocation of GLUT4 in adipocytes. Our findings indicate that, although cellubrevin and VAMP are resident proteins in adipocyte GLUT4-containing vesicles, they are not required for the acute insulin-induced delivery of GLUT4 to the plasma membrane.

Analyses of the co-localization of cellubrevin and the GLUT4 glucose transporter in rat and human insulin-responsive tissues.

We have investigated the subcellular distribution and association of cellubrevin, a low molecular weight protein implicated in the process of membrane fusion, with intracellular membranes containing the insulin-sensitive GLUT4 glucose transporter from rat adipocytes, rat skeletal muscle and human skeletal muscle. SDS-PAGE and immunoblot analyses of subcellular fractions of adipocytes and skeletal muscle indicated a positive correlation between the distribution of GLUT4 and cellubrevin in intracellular membrane fractions tested from all tissues. The identity of the polypeptide reacting with antiserum against cellubrevin was further confirmed on the basis of its susceptibility to proteolysis by tetanus toxin. Immunoisolation of GLUT4-containing vesicles from a microsomal fraction enriched with GLUT4 and cellubrevin revealed that cellubrevin could be coprecipitated with GLUT4 vesicles from adipocytes. In contrast, intracellular GLUT4 vesicles isolated from both rat and human skeletal muscle were devoid of any detectable immunoreactivity towards cellubrevin. The observation that cellubrevin does not colocalise with intracellular GLUT4 in skeletal muscle from two different species, rat and human, would strongly suggest that it is unlikely to participate in the insulin-induced delivery of GLUT4 to the cell surface in skeletal muscle.

Sedimentation and immunological analyses of GLUT4 and alpha 2-Na,K-ATPase subunit-containing vesicles from rat skeletal muscle: evidence for segregation.

In skeletal muscle insulin induces the translocation of both the GLUT4 glucose transporter and the alpha 2 subunit of the Na,K-ATPase from an intracellular membrane (IM) compartment to the plasma membrane (PM). Fractionation studies of rat skeletal muscle using a discontinuous sucrose gradient have indicated that the insulin-induced loss of both proteins occurs from a fraction containing intracellular membranes (IM) of common density. This raises the possibility that both proteins may be colocalized in a single intracellular compartment or are present in separate membrane vesicles that are of similar buoyant density. In this study we report the membrane vesicles from the insulin-responsive IM fraction can in fact be separated on the basis of differences in their sedimentation velocities; immunoblot analyses of fractions collected from a sucrose velocity gradient revealed the presence of two separate peaks for GLUT4 and the alpha 2 subunit of the Na,K-ATPase. One of these peaks representing a fast sedimenting population of vesicles (with a sedimentation coefficient of 2697 +/- 57 S) reacted against antibodies to the alpha 2 subunit of the Na,K-ATPase, whereas, the second peak contained a population of much slower sedimenting vesicles (with a sedimentation coefficient of 209 +/- 4 S) were practically devoid of the alpha 2-subunit. By contrast, the slow sedimenting vesicles were enriched by approximately 32-fold in GLUT4 relative to the starting IM fraction when the fractional protein content was taken into account. Immunoprecipitation of GLUT4-containing vesicles from the insulin-sensitive IM fraction revealed that no immunoreactivity towards either the alpha 2 or the beta 1 subunits of the Na,K-ATPase could be observed, signifying that the insulin-responsive subunits of the Na,K-ATPase and GLUT4 were present in different membrane vesicles and that it was unlikely, therefore, that the insulin-induced redistribution of these proteins to the PM occurs from a common intracellular pool.

Rab4, but not the transferrin receptor, is colocalized with GLUT4 in an insulin-sensitive intracellular compartment in rat skeletal muscle.

The role of Rab4, a small molecular weight GTP binding protein implicated in endosomal/plasma membrane (PM) recycling, in the translocation of the GLUT4 transporter in rat skeletal muscle was studied. Muscle membranes, prepared by subcellular fractionation of control and insulin treated rat skeletal muscle, were subjected to SDS/PAGE and immunoblot analyses. Insulin treatment caused an increase in GLUT4 in a plasma membrane (PM) enriched fraction from an intracellular membrane (IM) fraction. Immunoprecipitation of GLUT4 vesicles from the IM compartment revealed that Rab4 could be coprecipitated. Importantly, however, and unlike in adipocytes, immunoisolated GLUT4 vesicles from rat skeletal muscle contained no detectable immunoreactivity towards the transferrin receptor, suggesting that Rab4 was present in a GLUT4 IM pool that was not characteristic of early endosomes. The involvement of Rab4 in GLUT4 translocation was further supported by the finding that insulin treatment resulted in a significant (43%) reduction in Rab4 in the IM compartment. Our results suggest (i) that insulin induces the loss of both GLUT4 and Rab4 from the same IM compartment, (ii) that Rab4 may be involved in GLUT4 translocation based on its coprecipitation with the transporter from the insulin-sensitive pool and (iii) that Rab4 can be localized to intracellular membranes which appear not to be of early endosome origin.

The GLUT5 hexose transporter is also localized to the basolateral membrane of the human jejunum.

The intestine is a major site of expression of the human GLUT5 hexose transporter, which is thought to be localized exclusively to the brush border membrane (BBM) where its major role is likely to be in the absorption of fructose. In this study we present novel biochemical and morphological evidence showing that the GLUT5 transporter is also expressed in the basolateral membrane (BLM) of the human intestine. BBM and BLM were isolated by fractionation of human jejunum. BBM were enriched with alkaline phosphatase activity by over 9-fold relative to a crude jejunal homogenate and contained immunoreactive sucrase-isomaltase and GLUT5 proteins. By contrast the BBM fraction was substantially depleted of immunoreactive a1 subunits of the Na,K-ATPase and GLUT2 glucose transporters which were abundantly present in the BLM fraction. This BLM fraction was enriched by over 11-fold in potassium-stimulated phosphatase activity relative to the crude homogenate; BLM also reacted to immunological probes for GLUT5 but showed no observable reactivity with antibodies directed against sucrase-isomaltase. Quantitative immunoblotting revealed that the BBM and BLM contained near equal amounts of GLUT5 per mg of membrane protein. Immunogold localization of GLUT5 on ultrathin sections of human jejunum showed that GLUT5 was present in both apical BBM and BLM. This gold labelling was absent when antiserum was pre-incubated with the antigenic peptide corresponding to a specific C-terminal sequence of human GLUT5. Quantitative analyses of the number of gold particles per unit length of BBM and BLM indicated that the mean density of gold labelling was marginally greater in the BBM (0.399 gold particles/micrometer) than in the BLM (0.293 gold particle/micrometer). The localization of GLUT5 in the BLM of the human jejunum may suggest that it specifically participates in the transfer of fructose across the basal membrane of the enterocyte.