Uridine Diphosphogalactose Research Articles

Diagnosis of Inherited Disorders of Galactose Metabolism

Galactose metabolism occurs through an evolutionarily conserved pathway in which galactose and uridine diphosphoglucose are converted to glucose-1-phosphate and uridine diphosphogalactose through the action of three sequential enzymes: galactokinase (GALK, EC 2.7.1.6), galactose-1-phosphate uridyltransferase (GALT, EC 2.7.7.12), and uridine phosphogalactose 4'-epimerase (GALE, EC 5.1.3.2). Inborn errors of galactose metabolism occur with impaired activity for each of the enzymes. Classical galactosemia is the most common and the most severe of these diseases and is caused by deficiency of the GALT enzyme, affecting from approximately 1 in 10,000 to 1 in 30,000 live births. Deficiency of GALE is the rarest of the three diseases. Assays for galactitol and galactose-1-phosphate and methods for assaying enzyme activities of GALT, GALK, and GALE are provided here. Interpretation of diagnostic results for screen-positive newborns or symptomatic patients, as well as therapeutic interventions based on biochemical phenotype and molecular genotype, are also included as decision trees.

Processive lipid galactosyl/glucosyltransferases from Agrobacterium tumefaciens and Mesorhizobium loti display multiple specificities

The glycosyltransferase family 21 (GT21) includes both enzymes of eukaryotic and prokaryotic organisms. Many of the eukaryotic enzymes from animal, plant, and fungal origin have been characterized as uridine diphosphoglucose (UDP-Glc):ceramide glucosyltransferases (glucosylceramide synthases [Gcs], EC 2.4.1.80). As the acceptor molecule ceramide is not present in most bacteria, the enzymatic specificities and functions of the corresponding bacterial glycosyltransferases remain elusive. In this study, we investigated the homologous and heterologous expression of GT21 enzymes from Agrobacterium tumefaciens and Mesorhizobium loti in A. tumefaciens, Escherichia coli, and the yeast Pichia pastoris. Glycolipid analyses of the transgenic organisms revealed that the bacterial glycosyltransferases are involved in the synthesis of mono-, di- and even tri-glycosylated glycolipids. As products resulting from their activity, we identified 1,2-diacyl-3-(O-beta-D-galacto-pyranosyl)-sn-glycerol, 1,2-diacyl-3-(O-beta-D-gluco-pyranosyl)-sn-glycerol as well as higher glycosylated lipids such as 1,2-diacyl-3-[O-beta-D-galacto-pyranosyl-(1-->6)-O-beta-D-galacto-pyranosyl]-sn-glycerol, 1,2-diacyl-3-[O-beta-D-gluco-pyranosyl-(1-->6)-O-beta-D-galacto-pyranosyl]-sn-glycerol, 1,2-diacyl-3-[O-beta-D-gluco-pyranosyl-(1-->6)-O-beta-D-gluco-pyranosyl]-sn-glycerol, and the deviatingly linked diglycosyldiacylglycerol 1,2-diacyl-3-[O-beta-D-gluco-pyranosyl-(1-->3)-O-beta-D-galacto-pyranosyl]-sn-glycerol. From a mixture of triglycosyldiacylglycerols, 1,2-diacyl-3-[O-beta-D-galacto-pyranosyl-(1-->6)-O-beta-D-galacto-pyranosyl-(1-->6)-O-beta-D-galacto-pyranosyl]-sn-glycerol could be separated in a pure form. In vitro enzyme assays showed that the glycosyltransferase from A. tumefaciens favours uridine diphosphogalactose (UDP-Gal) over UDP-Glc. In conclusion, the bacterial GT21 enzymes differ from the eukaryotic ceramide glucosyltransferases by the successive transfer of up to three galactosyl and glucosyl moieties to diacylglycerol.

Metabolism of 13C galactose by lymphoblasts from patients with galactosemia determined by NMR spectroscopy

In order to assess the pathways by which galactose is metabolized by galactose-1-phosphate uridyltransferase (GALT) deficient cells, lymphoblasts from 10 galactosemic patients with defined genotypes (six Q188R homozygotes, two S153L homozygotes, and two with homozygous deletions) were incubated with 1 mM 1- or 2- 13C galactose for 2.5 and 5 h. The 13C-labeled metabolites were identified and quantified using nuclear magnetic resonance and the results were compared to that obtained with cells from eight normal individuals. Cells from galactosemic patients formed two to three times the galactose-1-phosphate (Gal-1P) in normal cells, no difference being observed between the various genotypes. Galactitol formation was not significantly different from normal cells. No labeled galactonate was detected. Cells with the Q188R and S135L mutations formed both labeled uridine diphosphogalactose (UDPgal) and uridine diphosphoglucose (UDPglu), but to a lesser extent than normals, whereas cells with the GALT deletion did not. The pattern of 13C enrichment of the ribose carbons of adenosine monophosphate upon incubation of the normal cells with 1- 13C galactose paralleled that found for incubations with 1- 13C glucose, which is consistent with galactose disposition through the Leloir pathway to glucose and its subsequent metabolism to ribose. Cells with the GALT deletion formed no detectable labeled ribose, whereas cells from a patient homozygous for Q188R mutation formed labeled ribose in a pattern similar to normal albeit with lower enrichment. The results suggest that there is residual GALT activity and function of the Leloir pathway in the presence of the Q188R as well as S135L mutation.

Galactose metabolism in normal human lymphoblasts studied by (1)H, (13)C and (31)P NMR spectroscopy of extracts.

The development of tools to follow and quantitate the fate of galactose in mammalian cells is crucial to the study and understanding of the inherited disorders of galactose metabolism. In this study we incubated normal human lymphoblasts with 1- or 2-(13)C galactose for 2.5 or 5 h and prepared TCA extracts of the cells. The various galactose metabolites were identified and quantified using a combination of proton, carbon and phosphorus NMR spectra. Galactose-1-phosphate (gal-1P), uridine diphosphogalactose, uridine diphosphoglucose and galactitol were present in the extracts. Average levels for gal-1P were around 10 nmol/mg protein and for uridine diphosphoglucose, uridine diphosphogalactose and galactitol in the range of 0.5-2 nmol/mg protein. Galactonate was never found in any conditions. Percentage labeling could be estimated for gal-1P and for the ribose carbons of AMP. The labeling agrees with a conversion of galactose to glucose through the Leloir pathway.

Changing the Donor Cofactor of Bovine α1,3-Galactosyltransferase by Fusion with UDP-galactose 4-Epimerase: MORE EFFICIENT BIOCATALYSIS FOR SYNTHESIS OF α-Gal EPITOPES

Two fusion enzymes consisting of uridine diphosphogalactose 4-epimerase (UDP-galactose 4-epimerase, EC ) and alpha1, 3-galactosyltransferase (EC ) with an N-terminal His(6) tag and an intervening three-glycine linker were constructed by in-frame fusion of the Escherichia coli galE gene either to the 3' terminus (f1) or to the 5' terminus (f2) of a truncated bovine alpha1, 3-galactosyltransferase gene, respectively. Both fusion proteins were expressed in cell lysate as active, soluble forms as well as in inclusion bodies as improperly folded proteins. Both f1 and f2 were determined to be homodimers, based on a single band observed at about 67 kDa in SDS-polyacrylamide gel electrophoresis and on a single peak with a molecular mass around 140 kDa determined by gel filtration chromatography for each of the enzymes. Without altering the acceptor specificity of the transferase, the fusion with the epimerase changed the donor requirement of alpha1, 3-galactosyltransferase from UDP-galactose to UDP-glucose and decreased the cost for the synthesis of biomedically important Galalpha1,3Gal-terminated oligosaccharides by more than 40-fold. For enzymatic synthesis of Galalpha1,3Galbeta1,4Glc from UDP-glucose and lactose, the genetically fused enzymes f1 and f2 exhibited kinetic advantages with overall reaction rates that were 300 and 50%, respectively, higher than that of the system containing equal amounts of epimerase and galactosyltransferase. These results indicated that the active sites of the epimerase and the transferase in fusion enzymes were in proximity. The kinetic parameters suggested a random mechanism for the substrate binding of the alpha1, 3-galactosyltransferase. This work demonstrated a general approach that fusion of a glycosyltransferase with an epimerase can change the required but expensive sugar nucleotide to a less expensive one.

The Role of Ligand Density in the Enzymatic Glycosylation of Carbohydrates Presented on Self-Assembled Monolayers of Alkanethiolates on Gold.

The biological activity of immobilized carbohydrates can show a dramatic dependence on the density of carbohydrate. This is the result of investigations with self-assembled monolayers that present N-acetylglucosamine groups as a model substrate for glycosylation by bovine β-1,4-galactosyltransferase (GalTase; see picture). Surface plasmon resonance spectroscopy and carbohydrate-binding lectins were used to characterize the reaction at the interface. UDP-Gal=uridine diphosphogalactose.

Uridine diphosphoglucose content of human erythrocytes: Assessment by conversion to uridine diphosphoglucuronate

To settle the ongoing controversy regarding differential uridine diphosphoglucose (UDPG) and uridine diphosphogalactose (UDPGal) content of erythrocytes, which may be important in evaluating the metabolic abnormality in patients with galactosemia, we derived a combined enzymatic-high-performance liquid chromatography (HPLC) assay. Uridine diphosphoglucuronate (UDPGA), the unique product of UDPG dehydrogenase activity, was separated and quantified by HPLC in extracts of human erythrocytes. The quantity of UDPGA produced in cell filtrates incubated with the enzyme corresponds to the amount of UDPG directly determined by HPLC. The amount of UDPGA produced was independent of the enzyme purity or activity used. On the other hand, the amounts of UDPG estimated by fluorometric measurement of the production of reduced nicotinamide adenine dinucleotide varied with the enzyme purity and activity. The combined enzymatic-HPLC method confirms the direct determinations of UDPG content of normal erythrocytes. The results indicate that, under appropriate conditions, the fluorometric-based assay will give accurate estimates of UDPG, but the direct HPLC method yields consistent and correct UDPG and UDPGal determinations.

31P NMR spectra of intact red blood cells: quantitation of UDPGlucose and UDPGalactose.

The combined levels of uridine diphosphogalactose (UDPGal) and uridine diphosphoglucose (UDPGlu) were measured directly by 31P NMR spectroscopy in intact fully oxygenated erythrocytes. Quantitative analysis was obtained using a sealed capillary of methylene diphosphonate (MDP) calibrated with standard solutions of UDPGlu and UDPGal of known concentration prepared in KRB/BSA. The combined peaks of UDPGlu and UDPGal were integrated after subtraction of the underlying broad component originating from membrane phospholipids. The average concentration of 27.16 +/- 5.18 nmole/ml or 8.08 +/- 1.36 mumole/100 g hemoglobin obtained for these metabolites correlated well with their total determined by HPLC of trichloracetic acid (TCA) extracts of the same samples.

Target sizes of galactosyltransferase, sialyltransferase, and uridine diphosphatase in Golgi apparatus of rat liver.

Target inactivation analysis was used to measure the functional size of uridine diphosphogalactose: N-acetylglucosamine beta(1,4)galactosyltransferase (galactosyltransferase), cytidine monophospho-N-acetyl-neuraminic acid: beta-galactoside alpha(2,6) sialytransferase (sialyltransferase), and uridine diphosphatase (UDPase) in Golgi membranes isolated from rat liver. The size of nucleoside diphosphatase (NDPase), an enzyme similar to UDPase but localized in rat liver endoplasmic reticulum, was also estimated by target inactivation analysis. The related enzymes, UDPase and NDPase, have target sizes of 96 +/- 4 and 77 +/- 3 kDa, while galactosyltransferase and sialyltransferase have target sizes of 97 +/- 10 and 130 +/- 20 kDa, respectively. The target inactivation sizes of galactosyltransferase and of sialyltransferase are about twice the monomer molecular weights of these enzymes obtained from sedimentation studies of the solubilized membranes as well as those predicted from previously reported cDNA sequences. We conclude from our studies that galactosyltransferase and sialyltransferase probably function as dimers in the Golgi membrane.

31P NMR analysis of red blood cell UDPGlucose and UDPGalactose: Comparison with HPLC and enzymatic methods

The levels of uridine diphosphogalactose (UDPGal) and uridine diphosphoglucose (UDPGlu) in trichloroacetic acid extracts of human red blood cells (RBC) were measured by 31P NMR spectroscopy. Individual determinations were compared to results obtained by enzymatic and high-pressure liquid chromatographic (HPLC) methods. The characteristic doublet of the P β resonance signals of both UDPGal and UDPGlu were detected in proton-decoupled spectra of extracts. Quantitative analyses were obtained by employing a standard, methylene diphosphonate, in an external capillary tube during data acquisition for periods of 14 to 24 h using an “inverse-gated” pulse sequence. The ratio of the integrated area of each of the uridine sugar nucleotide doublets to the area of the external reference peak was linear with concentrations between 0.03 and 0.50 m m. There was no difference between the mean value obtained by 31P NMR of 6.6 ± 1.4 μmol UDPGlu/100 g Hgb or 2.1 ± 0.6 μmol UDPGal/100 g Hgb and the corresponding levels determined enzymatically or by HPLC in identical RBC extracts. When analyzed as paired data, only UDPGlu by NMR was found to be lower than the value obtained by HPLC. As a quantitative analytical tool, NMR spectrometry validated both the enzymatic and HPLC methods used for measurement of uridine sugar nucleotides in our laboratories.

Characterization of a novel biochemical abnormality in galactosemia: Deficiency of glycolipids containing galactose or N-acetylgalactosamine and accumulation of precursors in brain and lymphocytes

Classic galactosemia, an inborn error of human galactose metabolism, is characterized by a deficiency of the enzyme galactose-1-phosphate uridyltransferase (GALT). The current model for the pathophysiology of this disease ascribes most of its symptoms to the toxicity of intracellular galactose-1-phosphate (Gal-1-P), one of the substrates of GALT which accumulates in the untreated disease state. Recently, a reduction in the intracellular concentration of UDP-Gal (uridine diphosphogalactose), one of the products of GALT, has been described in treated galactosemic patients. We investigated whether galactosemic patients might also have reduced amounts of those macromolecules that depend on UDP-Gal for their biosynthesis. We report a reduction in glycolipids that contain either galactose or its derivative N-acetylgalactosamine and an accumulation of the precursors to these compounds in the brain of a neonate with galactosemia. We also found an imbalance in glycolipids in galactosemic lymphoblasts. This novel biochemical abnormality observed in galactosemic patients is not addressed by dietary galactose-restriction therapy and could explain some of the chronic neurologic and other complications of galactosemia.

The concentration of red blood cell UDPGlucose and UDPGalactose determined by high-performance liquid chromatography

We have developed a sensitive method that employs high-performance liquid chromatography to separate and quantitate uridine diphosphogalactose (UDPGal) and uridine diphosphoglucose (UDPGlu) in human red blood cells. The trichloracetic acid extracts of red blood cells were chromatographed using a Dionex CarboPac anion-exchange resin and a 20–40% potassium phosphate buffer, pH 4.5, in a gradient-elution program. UDPGal and UDPGlu were detected spectrophotometrically at 254 nm. Recoveries of UDPGal and UDPGlu ranged from 96 to 106%. Under these conditions, there was exceptionally good reproducibility in stored specimens, and the method was sensitive in the low picamole range. The mean values and standard deviations in adults were 2.9 ± 0.4 and 7.8 ± 0.8 μmol/100 g Hgb for UDPGal and UDPGlu, respectively. The values in children were 4.5 ± 1.2 and 10.2 ± 1.7 μmol/100 g Hgb for UDPGal and UDPGlu, respectively. Values determined by the HPLC method are in excellent agreement with those obtained by enzyme analysis.

The role of murine cell surface galactosyltransferase in trophoblast: Laminin interactions in vitro

Implantation of the mouse embryo involves the invasion of the secondary trophoblast giant cells of the ectoplacental cone (EPC) into the uterine decidua. The mechanisms of this event are poorly understood. The putative substrate molecules found in the decidua which could support trophoblast invasion include laminin, fibronectin, and collagen type IV. EPCs dissected from Day 7.5 embryos were cultured on all three matrices. Galactosyltransferase (GalTase) was localized by immunolabeling on trophoblast cell surfaces when grown on laminin but not the other matrices. Perturbations of the enzyme:substrate complex with α-lactalbumin, uridine diphosphogalactose, anti-GalTase, and pregalactosylation of the matrix did not affect rates of EPC attachment. However, decreased rates of migration or altered morphologies of spreading cells were observed. Laminin, and not fibronectin or collagen type IV, could be galactosylated with both exogenous GalTase or EPC outgrowths. Digests of galactosylated laminin produced a glycoconjugate substrate with a molecular weight of <10K. The results suggest that invasive secondary trophoblast cells possess a GalTase-mediated migration system that is functional on laminin.

Alteration in isoenzyme patterns of serum galactosyltransferase activity in ovarian cancer patients: preliminary results

Isoelectric focusing on agarose gel was used to separate the isoenzymes of serum galactosyltransferase (uridine diphosphogalactose: N-acelylglucosaminyl galactosyltransferase, EC 2.4.1.22) from 8 healthy women, and 11 ovarian cancer patients of whom 4 were in clinical remission. In all cases, we found 7 major peaks with isoelectric points ranging from 4.0–5.4. The most acidic peaks were preferentially evelated in the tumor-bearing patients, particularly the peak with pI 4.44.

Galactose metabolism in human ovarian tissue.

Galactose metabolism was studied in human ovarian tissue obtained from 14 women controls between 21 and 72 y of age, and one 21-y-old galactosemic patient with hypergonadotrophic hypogonadism. Tissue slices were incubated with 1-14C-galactose, and labeled intermediates were analyzed by anion-exchange column chromatography. Activities of enzymes related to the galactose pathway: galactokinase, transferase, epimerase, uridine diphosphoglucose (UDPGlc) and uridine diphosphogalactose pyrophosphorylases, and UDPGlc and uridine diphosphogalactose pyrophosphatases were measured in ovarian homogenates using radioisotopic, spectrophotometric, and fluorometric techniques. Incorporation of carbon label from 1-14C-galactose into various galactose and glycolytic intermediates, as well as carbon dioxide and TCA-insoluble materials was demonstrated in samples from non-galactosemic controls. In tissue from the galactosemic individual, no labeled carbon dioxide was produced and very little incorporation into TCA-insoluble material was found. Labeled galactose-1-phosphate was elevated. In normal ovarian tissue, specific activities of galactokinase, transferase, epimerase, and UDPGlc pyrophosphorylase are much higher than those found in the red cells and in testes. UDPGlc pyrophosphorylase activity is about 50 times that of transferase, suggesting that uridine nucleotide sugars have an important role in the normal development and function of the ovary. It is hypothesized that premature ovarian failure, often observed in patients with galactosemia, is due to interference with nucleotide sugar metabolism and the synthesis of galactose containing glycoproteins and glycolipids consequent to the enzymatic defect in the major pathway of galactose metabolism.

The galactosyltransferase activity of hepatic nodules during rat liver carcinogenesis

Galactosyltransferase has been isolated from putative preneoplastic hepatocyte nodules generated in the resistant hepatocyte model by the procedure of Solt et al. (Am. J. Pathol., 88 (1977) 595–609). The following observations have resulted from these studies: (a) the specific activity of galactosyltransferase isolated from hepatocyte nodules by affinity chromatography was reduced to about 1 3 that of the enzyme in control and in liver tissue surrounding the nodules; (b) the galactosyl-transferase activity from normal rat serum eluted from the α-lactalbumin affinity column as a single peak (spec. act. = 1.57 nmol/min per mg) while that from the serum of nodule-bearing rats eluted in two distinct peaks (spec. act. = 2.49 and 0.49 nmol/min per mg protein); (c) the elution profile of the enzyme from hepatocyte nodules was broad compared to that from normal liver, surrounding liver or serum; (d) the K m for N-acetyl-D-glucsamine (GlcNAc) was lower in all four independent batches of nodules compared to the K m for GlcNAc from control and surrounding liver; (e) the K m for uridine diphosphogalactose (UDP-Gal) was higher for the enzyme from nodules compared to that from control tissue. These data suggest that the hepatocyte nodule produces several glycoforms of galactosyl-transferase the kinetic properties of which differ from those of the enzyme from control liver.

Agarose gel isoelectrofocusing of UDP-galactose pyrophosphorylase and galactose-1-phosphate uridyltransferase. Developmental aspect of UDP-galactose pyrophosphorylase

The uridine diphosphogalactose pyrophosphorylase activity has been determined in human adult and fetal tissues as well as blood of various ages by measurement of UDP-galactose production from gal-1-p and UTP. The highest activity was found from adult liver in which the specific activity was about 5% of the gal-1-p uridyltransferase activity. In general adult tissues had a somewhat higher activity than the corresponding fetal tissues except erythrocytes in which fetuses had a 5–10 times higher activity than adults. From normal blood the pyrophosphorylase activity in erythrocytes decreased with age, but in the case of galactosemia the decrease with age was not distinct. According to agarose gel isoelectrofocusing studies, at least two isozyme forms for UDP-galactose pyrophosphorylase exist with the activity bands between pH 6.0–6.15. The patterns of AGIF bands of pyrophosphorylase varied according to the age of the samples, suggesting the development of the isozyme forms of pyrophosphorylase to be age-dependent. Uridyltransferase, on the other hand, resolved into multiple bands between pH 5.1–5.6 on agarose gels and the patterns varied according to the variants but not to the age. Significance of the decrease in the pyrophosphorylase activity in erythrocytes with age as well as of the difference in AGIF bands between normal and the galactosemic were discussed with regard to the pathology of classical galactosemia.

Mechanism of glycosylation in the Golgi apparatus.

A major role of the Golgi apparatus in liver is the terminal glycosylation of secreted serum proteins and of plasma membrane glycoproteins. Galactosyltransferase is a membrane-bound Golgi enzyme that transfers galactose directly from uridine diphosphogalactose (UDP-Gal) to terminal N-acetylglucosamine groups of N-asparagine-linked glycoproteins during secretion. Sialytransferase then transfers sialic acid from cytidine monophosphosialic acid (CMP-NAN) to the newly added terminal galactose of the glycoprotein. In the cell, the transfer reaction must occur on the lumen side of the Golgi membrane. UDP-Gal is synthesized mainly in the cytoplasm and CMP-NAN is synthesized in the nucleus in liver. An important question for understanding the mechanism is, how do these nucleotide sugars gain access to the transferases? A second question involves uridine diphosphate (UDP), a highly inhibitory product of galactosyltransferase. How is UDP removed from the lumen of the Golgi fast enough to prevent product inhibition of the galactosyltransferase? We have shown that isolated Golgi, although vesiculated, retains its original orientation. The vesicles are oriented with greater than 90% of both galactosyltransferase and sialyl-transferase on the luminal side of the vesicles. Using intact vesicles, we can show that UDP-Gal is taken up via a saturable carrier system present in the Golgi membrane. During galactosylation in vitro, UDP formed in the lumen of Golgi vesicles is rapidly converted to UMP by a nucleoside diphosphatase in the lumen. Uridine monophosphate, which is much less inhibitory to the galactosyltransferase than UDP, is then transported out of the lumen by a second carrier and is broken down further to uridine by 5'-nucleotidase on the cytoplasmic side of the Golgi vesicles. The transport of nucleotides appears unique to the Golgi membranes, since neither rough endoplasmic reticulum nor plasma membrane vesicles from rat liver accumulate these nucleotides.

Galactose transfer to endogenous acceptors within Golgi fractions of rat liver.

The distribution of galactosyl transferase was studied using trans and cis Golgi fractions isolated by a modification of the Ehrenreich et al. procedure (1973. J. Cell Biol. 59:45-72) as well as an intact Golgi fraction isolated by a new one-step procedure. Two methods of assay were used. The first method analyzed the ability of Golgi fractions to transfer galactose (from uridine diphosphogalactose [UDP-gal] substrate) to the defined exogenous acceptor ovomucoid. The second method assessed the transfer of galactose from UDP-gal substrate to endogenous acceptors (endogenous glycosylation). The trans Golgi fraction (Golgi light) was highly active by the first method but revealed only low activity by the second method. Golgi fractions enriched in central and cis elements (the Golgi intermediate, heavy and especially the intact Golgi fraction) were highly active in both methods of assay. The endogenous glycosylation approach was validated by gel fluorography of the endogenous acceptors. For all Golgi fractions, transfer of galactose was revealed to secretory glycopeptides. It is concluded that galactosyl transferase activity in vivo occurs primarily in central and cis Golgi elements but not trans Golgi vesicles.

The nucleotide content of rat liver Golgi vesicles

The nucleotide profile of rat liver Golgi vesicles isolated using sucrose gradients has been determined by high-pressure liquid chromatography. The nucleotide composition of this Golgi preparation, probably modified by osmotic shock, differs from that of liver supernatant fraction and from isolated rough microsomes. The major nucleotides present in the Golgi have been tentatively identified as uridine diphosphate and a peak containing uridine monophosphate plus cytidine monophosphate at 1.6 and 0.5 nmol/mg protein, respectively. In order to minimize osmotic shock, we have modified the isolation of Golgi using D 2O-sucrose gradients. Intact Golgi from these gradients were extracted directly and analyzed. Higher levels of nucleotides were found in the unshocked preparation, and the profile was also altered, although it was still distinct from that of liver supernatant. Four major peaks were found, tentatively identified as uridine monophosphate plus cytidine monophosphate, adenosine monophosphate, UDP, and uridine diphosphogalactose plus uridine diphosphoglucose, at 6.4, 6.4, 6.1, and 3.3 nmol/mg protein. These results indicate that the membrane of the Golgi apparatus is not freely permeable to nucleotides but that selective mechanisms exist for the uptake or exclusion of specific nucleotides from this organelle. The fact that UDP is selectively retained in shocked Golgi vesicles may indicate the presence of a binding protein which would prevent interference of Golgi function by UDP, a highly inhibitory product of galactosyltransferase.