This paper reports on the use of a digital microfluidic platform to perform multiplex automated genetic engineering (MAGE) cycles on droplets containing Escherichia coli cells. Bioactivated magnetic beads were employed for cell binding, washing, and media exchange in the preparation of electrocompetent cells in the electrowetting-on-dieletric (EWoD) platform. On-cartridge electroporation was used to deliver oligonucleotides into the cells. In addition to the optimization of a magnetic bead-based benchtop protocol for generating and transforming electrocompetent E. coli cells, we report on the implementation of this protocol in a fully automated digital microfluidic platform. Bead-based media exchange and electroporation pulse conditions were optimized on benchtop for transformation frequency to provide initial parameters for microfluidic device trials. Benchtop experiments comparing electrotransformation of free and bead-bound cells are presented. Our results suggest that dielectric shielding intrinsic to bead-bound cells significantly reduces electroporation field exposure efficiency. However, high transformation frequency can be maintained in the presence of magnetic beads through the application of more intense electroporation pulses. As a proof of concept, MAGE cycles were successfully performed on a commercial EWoD cartridge using variations of the optimal magnetic bead-based preparation procedure and pulse conditions determined by the benchtop results. Transformation frequencies up to 22% were achieved on benchtop; this frequency was matched within 1% (21%) by MAGE cycles on the microfluidic device. However, typical frequencies on the device remain lower, averaging 9% with a standard deviation of 9%. The presented results demonstrate the potential of digital microfluidics to perform complex and automated genetic engineering protocols.

Species of Candida frequently cause life-threatening infections in neonates, transplant and intensive care unit (ICU) patients, and others with compromised host defenses. The successful management of systemic candidiasis depends upon early, rapid diagnosis. Blood cultures are the standard diagnostic method, but identification requires days and less than half of the patients are positive. These limitations may be eliminated by using real-time polymerase chain reaction (PCR) to detect Candida DNA in the blood specimens of patients at risk. Here, we optimized a PCR protocol to detect 5–10 yeasts in low volumes of simulated and clinical specimens. We also used a mouse model of systemic candidiasis and determined that candidemia is optimally detectable during the first few days after infection. However, PCR tests are often costly, labor-intensive, and inconvenient for routine use. To address these obstacles, we evaluated the innovative microfluidic real-time PCR platform (Advanced Liquid Logic, Inc.), which has the potential for full automation and rapid turnaround. Eleven and nine of 16 specimens from individual patients with culture-proven candidemia tested positive for C. albicans DNA by conventional and microfluidic real-time PCR, respectively, for a combined sensitivity of 94%. The microfluidic platform offers a significant technical advance in the detection of microbial DNA in clinical specimens.

COS-7 cells expressing 1,360 residues from the amino terminus of porcine submaxillary mucin were used to determine whether this region, containing the D1, D2, and D3 domains, is involved in forming mucin multimers. Analysis of the proteins immunoprecipitated from the medium of transfected cells by reducing SDS-gel electrophoresis showed a single N-glycosylated protein with no indication of proteolytically processed forms. Without prior reduction, only two proteins, corresponding to monomeric and disulfide-linked trimeric species, were observed. The expressed protein devoid of N-linked oligosaccharides also formed trimers, but was secreted from cells in significantly less amounts than glycosylated trimers. Pulse-chase studies showed that the disulfide-linked trimers were assembled inside the cells no earlier than 30 min after protein synthesis commenced and after the intracellular precursors were N-glycosylated. Trimer formation was inhibited in cells treated with brefeldin A, monensin, chloroquine, or bafilomycin A1, although only brefeldin A prevented the secretion of the protein. These results suggest that trimerization takes place in compartments of the Golgi complex in which the vacuolar H+-ATPase maintains an acidic pH. Coexpression in the same cells of the amino-terminal region and the disulfide-rich carboxyl-terminal domain of the mucin showed that these structures were not disulfide-linked with one another. Cells expressing a DNA construct encoding a fusion protein between the amino- and carboxyl-terminal regions of the mucin secreted disulfide-linked dimeric and high molecular weight multimeric species of the recombinant mucin. The presence of monensin in the medium was without effect on dimerization, but inhibited the formation of disulfide-linked multimers. These studies suggest that disulfide-linked dimers of mucin are subsequently assembled into disulfide-linked multimers by the amino-terminal regions. They also suggest that the porcine mucin forms branched disulfide-linked multimers. This ability of the amino-terminal region of mucin to aid in the assembly of multimers is consistent with its amino acid identities to the amino-terminal region of human von Willebrand factor, which also serves to form disulfide-linked multimers of this protein.

The complete structure of the DNA encoding the polypeptide chain of porcine submaxillary mucin has been determined. The polypeptide is composed of distinct domains. A large central domain containing tandem repeats of 81 residues each is flanked by much shorter domains with sequences similar to the tandem repeats. Four disulfide-rich domains, three at the amino terminus and one at the carboxyl terminus, complete the chain. The disulfide-rich domains have significant sequence identity to those of other mucins and prepro-von Willebrand factor. The coding region of the mucin gene is highly polymorphic, and three alleles were identified in a single animal that encoded different numbers of the 81-residue tandem repeats. A single large exon devoid of introns encodes the tandem repeat domains. The largest allele with 135 tandem repeats encoded 13,288 amino acids to give a polypeptide with M r = 1,184,106. The other two alleles contained 99 and 125 tandem repeats, respectively. Each allele also showed different restriction fragment length polymorphisms, which is consistent with the different patterns seen in individual animals. Fragment length polymorphism was also seen within two different families of animals, indicating that the polymorphism observed occurs in a single generation.

COS-7 cells transfected with three different expression vectors encoding the 240-amino acid residue, disulfide-rich domain at the carboxyl terminus of porcine submaxillary mucin have been used to determine the possible function of the domain in forming higher oligomers of the mucin polypeptide chain. The domain is expressed as a disulfide-bonded dimer, as shown by SDS-gel electrophoretic analysis of the immunoprecipitated domain in the presence and absence of reducing agent and the cross-linking agent bis(sulfosuccinimidyl) suberate. Molecular weight determination by gel filtration on agarose columns in 6 M guanidine HCl confirmed dimer formation. However, the domain expressed is heterogeneous as the result of different extents of glycosylation. Pulse-chase studies with the 35S-labeled domain show that dimer formation and secretion from cells occur very rapidly. Moreover, dimer formation is not dependent on the N-linked oligosaccharides on the domain. Evidence is presented that dimer formation most likely occurs in the endoplasmic reticulum before complex-type oligosaccharide synthesis is completed. Neither brefeldin A nor tunicamycin interferes with the rate of dimer formation. These studies suggest that the disulfide-rich domain acts to form dimers of the polypeptide chain of mucin. This role of the domain is consistent with its amino acid sequence similarity to the disulfide-rich domain of human prepro-von Willebrand factor, which also serves to form dimers of this blood coagulation factor.

Addition of N-acetylgalactosamine to threonine and serine is the first step in the synthesis of O-glycosidically linked oligosaccharides. A UDP-N-acetyl-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase (EC 2.4.1.41) from porcine submaxillary glands was recently purified to electrophoretic homogeneity, and polyclonal antibodies against the purified transferase were raised. Immunoblots of porcine, bovine, and ovine submaxillary gland extracts with the anti-transferase antibodies gave a single band and the antibodies reacted equally well with the purified glycosylated and N-glycanase-treated transferase. Immunoelectron microscopic localization of the transferase was achieved in Lowicryl K4M thin sections and frozen-thawed thin sections of porcine and bovine submaxillary gland by using the protein A-gold technique. Specific gold particle labeling was observed in the cis Golgi apparatus and smooth-membraned vesicular structures in close topological relation with it. Labeling was undetectable in the rough endoplasmic reticulum, its transitional elements, and smooth-membraned structures close to them, the trans Golgi apparatus, mucin droplets, and the plasma membrane. The onset of labeling for peptide-bound GalNAc as detected with Vicia villosa isolectin G4 mirrored the transferase immunolocalization as directly shown by double labeling and extended into the trans Golgi apparatus and mucous droplets. Apomucin immunolabeling was found throughout the endoplasmic reticulum and the intermediate compartment and partially overlapped the region of transferase labeling in the Golgi apparatus as demonstrated by double immunolabeling. Thus, the initial step of UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase-mediated O-glycosylation in porcine and bovine submaxillary gland cells occurs in the cis Golgi apparatus. The possible involvement of the intermediate compartment remains to be clarified.

The acceptor substrate specificity of a pure polypeptide N-acetylgalactosaminyltransferase has been examined with synthetic polypeptides with sequences identical, or similar to those found in porcine mucin or human erythropoietin. The sequences adjacent to either threonine or serine markedly influence the formation of GalNAc-O-Thr and GalNAc-O-Ser. Examination of the mucin-like peptide VLGXXAV, where X is Thr, Ser, or Ala, shows only Thr-containing peptides to be acceptors. The best substrate is formed when XX is TT. Peptides with XX as either AT or TA are less effective and those with XX as either ST or TS are much less effective acceptors. The amino acids adjacent to serine in the peptide formed by residues 121-131 in human erythropoietin, PPDAASAAPLR, also markedly influence the formation of GalNAc-O-Ser. Thus, PPDASSSAPLR and PPDVVSVVPLR are about 5- and 30-fold, respectively, less active than the erythropoietin peptide. The peptide PPDGGSGGPLR is inactive. The shorter peptide DAASAAPL is also about 5-fold less active than the full-length peptide, but the peptide AASAA is inactive. These studies indicate that one transferase can form both GalNAc-O-Ser and GalNAc-O-Thr residues when the sequences adjacent to the glycosylated residue are of the proper kind. Thus, in contrast to earlier suggestions, there is no evidence that different transferases form GalNAc-O-Ser and GalNAc-O-Thr. Examination of tissue homogenates from various tissues confirms this conclusion.

A UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase from porcine submaxillary glands was purified to electrophoretic homogeneity. IgG prepared from antisera against the pure enzyme immunoprecipitated the transferase in Triton X-100 extracts of submaxillary glands. The submaxillary transferase is a membrane-bound enzyme in contrast to the pure bovine colostrum enzyme, which is soluble in the absence of detergents. Both transferases have similar properties but also differ significantly. Examination of the acceptor substrate specificity of the submaxillary gland transferase showed that it specifically transferred N-acetylgalactosamine from UDP-GalNAc to the hydroxyl group of threonine and was devoid of transferase activity toward serine-containing peptides. These results imply that more than one transferase is involved in forming the GalNAc-threonine and the GalNAc-serine linkages found in O-linked oligosaccharides in glycoproteins. The amino acid sequence adjacent to glycosylated threonine residues may influence the rate of glycosylation by the pure transferase. For example, the second threonine residue in the sequence, Thr-Thr, appears to be glycosylated about twice as fast as the first and more rapidly than single, isolated threonine residues. However, no unique consensus sequence for glycosylation of threonine residues is evident, and any accessible threonine residue appears to be a potential acceptor substrate.

Glycoproteins generally contain carbohydrate structures attached via either of two linkages. They may be N-linked to the amide nitrogen of asparagine or they may be O-linked glycosidically to the hydroxyl groups of threonine or serine residues. O-linked oligosaccharides are characteristically found on mucins or cell surface mucin-like molecules, which are often heavily glycosylated, containing up to 65–85% carbohydrate by weight. A number of problems are frequently encountered in structural studies of heavily glycosylated glycoproteins. Although it is desirable to analyze intact glycopeptides, such glycoproteins are frequently resistant to proteolytic cleavage and high yield purification of glycopeptides may be hampered by microheterogeneity of the extent of glycosylation at each serine or threonine residue. Sequencing methods using standard gas phase reagents and polybrene carrier for peptide immobilization are generally unsatisfactory for glycopeptides. Sequencer cycle blanks are usually observed at the attachment site of the large relinked carbohydrate units; however, the interpretation of HPLC chromatograms of O-glycosylated amino acid phenylthiohydantoin derivatives may be much more deceptive. This chapter describes a methodology for sequencing glycopeptides and identifying O-glycosylated threonine residues.

The sequence of a 3.65-kilobase cDNA encoding a large portion of the polypeptide chain of porcine submaxillary mucin (apomucin) has been completed. The encoded polypeptide contains 1150 residues with the carboxyl-terminal 240 residues forming a globular domain that is rich in half-cystine, but deficient in sites for oligosaccharide attachment. The remaining 910 residues preceding the half-cystine-rich domain appear devoid of secondary structures, but they are rich in serine and threonine to which the O-linked oligosaccharides are bound. The first 391 residues of apomucin contain several tandemly repeated, identical sequences of 81 residues. Blots of genomic DNA partially digested with restriction nucleases show that at least 25 of these identical repeats are present in apomucin. The amino acid composition of apomucin isolated in the absence of protease inhibitors was shown earlier (Eckhardt, A. E., Timpte, C. S., Abernethy, J. L., Toumadje, A., Johnson, W. C., Jr., and Hill, R. L. (1987) J. Biol. Chem. 282, 11339-11344) to be devoid of half-cystine. In contrast, the amino acid composition of mucin purified in the presence of protease inhibitors contains half-cystine in amounts predicted by the cDNA sequence and also suggests that this mucin has about 25 tandem repeats. Thus, apomucin contains at least 2800 amino acid residues. Moreover, immunoblots of apomucin prepared in the presence or the absence of protease inhibitors, with antibodies specific for the half-cystine-rich domains or the tandem repeat sequences, show that the half-cystine-rich domain is absent in apomucin unless protease inhibitors are present throughout. Both types of mucin, however, contain the highly repetitive sequences. The molecular weight of undegraded apomucin has not been established exactly, but gel filtration in 6 M guanidine hydrochloride suggests that it is considerably higher than 250,000. RNA blot analysis shows that apomucin mRNA is large and polydisperse in accord with the message size necessary to synthesize the large apomucin polypeptide. These structural features of apomucin suggest a model for the structure of the mucin molecule that correlates well with its reported properties.

Antibodies prepared against enzymatically deglycosylated porcine submaxillary gland mucin (apomucin), which were unreactive with native mucin and its partially deglycosylated derivatives, were used to immunolocalize apomucin in situ. Electron microscopy of sections of Lowicryl K4M-embedded tissue reacted successively with antibodies and protein A-gold complexes showed apomucin exclusively in mucous cells within the rough endoplasmic reticulum, transitional elements of the endoplasmic reticulum, and vesicles at the cis side of the Golgi apparatus. The Golgi apparatus, forming mucous droplets, and mucous droplets contained no apomucin. Although the rough endoplasmic reticulum contained most of the apomucin in mucous cells, some cisternae of the endoplasmic reticulum and the nuclear envelope were devoid of apomucin. Examination of tissue sections treated with the glycosidases used to prepare apomucin revealed immunolabel for apomucin throughout the secretory pathway. Colloidal gold coated with Helix pomatia lectin was used to detect nonreducing N-acetylgalactosamine residues. In mucin-producing cells lectin-gold was found in the mucous droplets, the forming mucous droplets, and throughout the Golgi apparatus but mostly in the cis portion of this organelle. In tissue sections reacted successively with lectin-gold and anti-apomucin/protein A-gold, both types of gold complex could be found in the cis side of the Golgi apparatus. These data indicate that the O-glycosylation of mucin is a posttranslational event that occurs in the Golgi apparatus and begins in the cis side of the Golgi apparatus.

A lambda gt11 cDNA library, prepared from porcine submaxillary gland mRNA, was screened with anti-apomucin IgG, and five antibody-reactive phage were isolated. The phage with the largest cDNA insert, designated lambda PSM103, was further characterized. Its fusion protein reacted with anti-apomucin IgG and was used to affinity purify antibodies that specifically reacted with apomucin, indicating that the protein shares antigenic determinants with apomucin. The nucleotide sequence of 1510 bases in the 3.7-kilobase cDNA insert of lambda PSM103 has been established, thereby giving a deduced amino acid sequence of 503 residues in apomucin, or about 45% of the molecule. The deduced sequence of the apomucin polypeptide was found to contain 4.8 tandemly repeated, identical sequences of 81 residues each. The presence of these uniquely repeated sequences was confirmed by restriction endonuclease digestion of DNA derived from lambda PSM103. The repeat sequence was also confirmed in apomucin by the isolation of an 81-residue tryptic peptide with an amino acid composition and an amino-terminal amino acid sequence (up to 44 residues) identical to those of the tandem repeat. Moreover, the peptide was isolated in 760% yield, indicating that the tandem repeat occurs at least eight times in apomucin. The presence of such a long repetitive region in the gene for apomucin raises the possibility for considerable polymorphism in the gene and a corresponding size heterogeneity of apomucin. The predicted secondary structure of the 503 residues confirms the earlier proposal that apomucin is an extended, nonglobular polypeptide. Although the sequences around 192 serine and threonine residues have been established in apomucin, a recognition sequence for the N-acetylgalactosaminyltransferase that initiates glycosylation of apomucin is not evident, except that the glycosylated residues occur in turns.

Porcine submaxillary gland mucin was deglycosylated with a mixture of pure glycosidases to give apomucin containing less than 1% carbohydrate. The resulting apomucin freed of glycosidases was found to contain nine amino acids: threonine, serine, glutamic acid, proline, glycine, alanine, valine, isoleucine, and arginine. Serine, threonine, glycine, and alanine comprise 77% of the composition. The molecular weight of apomucin was 96,500 as determined by gel filtration in guanidine hydrochloride. Its Stokes radius was greater than 68.6 A, a far larger value than expected for a globular protein with Mr = 96,500. Circular dichroism spectroscopy of apomucin suggests that it contains 42% aperiodic or "other" structure, 40% beta-turns, 10% antiparallel pleated sheet, and 8% helical structures. The predicted secondary structure of a 50-residue peptide from ovine submaxillary gland mucin resembles the circular dichroism predictions, being dominated by turns that would lead to an extended nonglobular structure. Analysis for the secondary structure of a 36-residue tryptic peptide derived from porcine submaxillary gland apomucin predicts a similar structure. It is concluded that apomucin is likely devoid of traditional secondary structure and serves as a scaffold upon which oligosaccharides are added in O-glycosidic linkage. When sufficient sialic acid is present in the oligosaccharides, native highly viscous mucin containing about two-thirds carbohydrate by weight is obtained.

A small-scale affinity chromatographic procedure was developed to screen for the presence of fucose and mannose/N-acetylglucosamine-binding lectins in small amounts of rat tissues. Of all tissues examined, only the liver contained the fucose-binding lectin, whereas both liver and blood serum contained the mannose/N-acetylglucosamine lectin. By means of immunocytological methods using antibodies to hepatic lectins, the fucose lectin was shown to be uniquely present in Kupffer cells and absent in all other types of rat macrophages examined. The binding and uptake of different neoglycoproteins by nonparenchymal cell fractions of liver indicated that the fucose-binding lectin was either not responsible for the uptake or that more than one lectin was acting. With the identification of another lectin (Mr = 180,000) by the above screening procedure for hepatic lectins and the results of studies in the following paper (Haltiwanger, R.S., and Hill, R. L. (1986) J. Biol. Chem. 261, 7440-7444) two lectins appear to be involved. A small amount of the hepatic mannose/N-acetylglucosamine lectin was found by the above screening procedure to have a higher affinity for L-fucosyl-bovine serum albumin-Sepharose than the majority of the lectin in hepatocytes. This lectin, called the high affinity form, was purified and its properties examined. On a weight basis the high affinity form bound 7-12 times more ligand than the normal form. Its Ka for L-fucosyl-bovine serum albumin was 2.3 X 10(9) M-1 compared to 3.5 X 10(8) M-1 for the normal form. Moreover, the concentrations of monosaccharides required to inhibit the high affinity form were about 3 times less than those required to inhibit binding of the normal form. The two forms, however, have identical molecular weights (32,000) under reducing and nonreducing conditions, bind anti-lectin antibodies in the same way, and give identical peptide maps after V-8 protease digestion. The structural basis for the different binding affinities of the two forms remains unknown.

A rapid, simple, and inexpensive method has been developed for preparing UDP- N -acetylgalactosamine in amounts sufficient for several thousand assays of enzymes that employ this nucleotide sugar as substrate. The UDP- N -acetylglucosamine-4-epimerase in extracts of porcine submaxillary glands was used to convert UDP- N -acetylglucosamine to an equilibrium mixture of UDP- N -acetylglucosamine and UDP- N -acetylgalactosamine (molar ratio, 77:23). The two nucleotide sugars were separated from components in the extract by ion-exchange chromatography and then separated from one another by affinity chromatography on a column of Griffonia simplicifolia lectin I bound to agarose. The UDP- N -acetylgalactosamine was obtained in pure form by ion-exchange chromatography in an overall yield of 91% from the equilibrium mixture. The separation of the two nucleotide sugars by affinity chromatography also provides a rapid assay for the UDPGlcNAc-4-epimerase, which is more accurate and less time consuming than earlier published assays.

An a1 + 3 mannoside 61 + 2 N-acetylglucosaminyltransferase from porcine liver designated transferase I, has been purified 4,800-fold in a 10% yield to a specific activity of 4.8 pmol/min/mg by affinity chromatography on UDP-hexanolamine-Sepharose.Sodium dodecyl sulfate gel electrophoresis of the purified enzyme showed a major species with an apparent molecular weight of 59,000 and a minor species with an apparent molecular weight of 52,000.The transferase is specific for a1 -* 3-linked mannose residues in the sequence Manal 3 3Manj31 --f 4GlcNAc.The best acceptor (K,