| Glycobiology | Pages |
A high-throughput microscale method to release N-linked oligosaccharides from glycoproteins for matrix-assisted laser desorption/ionization time-of-flight mass spectrometric analysis
Introduction
Results
Discussion
Materials and methods
Acknowledgments
Abbreviations
References
A high-throughput microscale method to release N-linked oligosaccharides from glycoproteins for matrix-assisted laser desorption/ionization time-of-flight mass spectrometric analysis
This report describes a convenient method for the rapid and efficient release of N-linked oligosaccharides from low microgram amounts of glycoproteins. A 96-well MultiScreen assay system containing a polyvinylidene difluoride (PVDF) membrane is employed to immobilize glycoproteins for subsequent enzymatic deglycosylation. Recombinant tissue-type plasminogen activator (rt-PA) is used to demonstrate the deglycosylation of 0.1-50 µg of a glycoprotein. This method enabled the recovery of a sufficient amount of N-linked oligosaccharides released enzymatically with peptide N-glycosidase F (PNGaseF) from as little as 0.5 µg rt-PA for subsequent analysis by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry. The immobilization of rt-PA to the PVDF membrane did not sterically inhibit the PNGaseF-mediated release of oligosaccharides from rt-PA as determined by tryptic mapping experiments. Comparison of the oligosaccharides released from 50 µg of rt-PA by either the 96-well plate method or by a standard solution digestion procedure showed no significant differences in the profiles obtained by high-pH anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD). Both neutral and sialylated oligosaccharide standards spiked into wells were recovered equally as determined by HPAEC-PAD. One advantage of this approach is that reduction and alkylation can be performed on submicrogram amounts of glycoproteins with easy removal of reagents prior to PNGaseF digestion. In addition, this method allows 60 glycoprotein samples to be deglycosylated in 1 day with MALDI-TOF or HPAEC-PAD analysis being performed on the following day. Key words: matrix-assisted laser desorption ionization time-of-flight mass spectrometry/N-linked oligosaccharide/polyvinylidene difluoride/recombinant tissue-type plasminogen activator
Introduction
Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF) has been demonstrated to be useful for both structural characterization (Stahl et al., 1994; Sutton et al., 1994) and quantitation (Harvey, 1993; Field et al., 1996) of oligosaccharides. Generally, underivatized neutral oligosaccharides can be detected at the 50-80 fmol level, and underivatized acidic oligosaccharides, those containing sialic acid, are detectable at the 10-25 fmol level (Papac et al., 1996).
The inability to easily manipulate sub-micromolar amounts of protein samples has often precluded the exploitation of the low detection limits obtainable with MALDI-TOF. In response to this challenge, many investigators have developed a variety of ingenious sample-handling approaches that allow them to take advantage of the exquisite sensitivity offered by MALDI-TOF. One general strategy is to concentrate the analyte from a dilute solution onto the MALDI-TOF target (Hutchens and Yip, 1993; Courchesne and Patterson, 1997). The principal problem encountered during the handling of submicrogram quantities of protein is adsorptive losses that occur when a protein comes into contact with a surface (e.g., microfuge tubes, chromatographic supports, etc.). An approach that minimizes adsorptive losses is the immobilization of the protein to a surface that is compatible with subsequent chemical or enzymatic digestion, and will also allow buffer exchanges to be performed without protein loss. Polyvinylidene difluoride (PVDF) is one such surface which was first introduced successfully to extend the useful range of Edman sequencing to the low picomole range (Matsudaira, 1987), and is now used routinely for blotting proteins from gels for sequencing by either Edman chemistry or mass spectrometry (Gharahdaghi et al., 1996). Weitzhandler and coworkers successfully applied this approach to the analysis of glycoproteins first separated by either sodium dodecyl sulfate polyacrylamide (Weitzhandler et al., 1994) or isoelectric focusing (Andersen et al., 1994) gel electrophoresis. Sufficient monosaccharides or oligosaccharides could be released from the PVDF for subsequent analysis by high pH anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD). Weitzhandler's work suggested a similar approach could enable oligosaccharides to be released and recovered from submicrogram amounts of a glycoprotein for analysis by MALDI-TOF.
Within our group, a significant need existed to develop a sample preparation strategy that would enable enzymatic release of N-linked oligosaccharides from a large number of glycoprotein samples in one day. A second criterion was to develop a method capable of efficiently alkylating and reducing low microgram quantities of glycoproteins for subsequent release of N-linked oligosaccharides suitable for glycoform profiling using MALDI-TOF. The availability of a 96-well plate lined with PVDF offered the opportunity to develop a high-throughput multistep method capable of isolating the N-linked oligosaccharides from low microgram quantities of glycoproteins. We will demonstrate using the 96-well MultiScreen-IP plate that submicrogram amounts of recombinant tissue-type plasminogen activator (rt-PA) can be deglycosylated to release enough of the N-linked oligosaccharides for subsequent glycoform profiling by MALDI-TOF.
Results
Development of the 96-well plate assay
We have used the MultiScreen assay system to develop a high-throughput microscale method for oligosaccharide release. The MultiScreen assay system offers a convenient 96-well plate format with a wide variety of membrane support choices. Plates made with the high protein-binding membrane, PVDF (Immobilon P), were chosen, because the first step of this method requires that the glycoproteins are efficiently captured from solution. One advantage of using PVDF to capture proteins from solution is that the membrane acts as a concentration device; therefore, large volumes of dilute solutions can be applied to the wells to capture proteins. Integral to this system is the vacuum manifold that allows unbound components to be washed from the wells simply by applying a gentle vacuum to the plate. The underside of the plate has a fitted rubber gasket with a small aperture that directs the effluent from each well straight down when a vacuum is applied. This design allows samples to be drained from individual wells, and collected into a standard 96-well plate seated in the vacuum manifold underdrain if desired. However, we have found that higher oligosaccharide recoveries are attained with less potential for chemical contamination, if the oligosaccharides are recovered by pipette and transferred to microfuge tubes.
Due to the hydrophobic nature of PVDF membranes, the first step required for protein adsorption is to wet the membrane with an organic solvent miscible with water. Initially, methanol is used to wet the membrane (Figure 1). Following repeated washings with water, the protein is added and then reduced and carboxymethylated. Between washes, the well is drained completely before addition of the next wash solution. By reducing and carboxymethylating the protein, detergents that are often required for complete enzymatic deglycosylation can be avoided. This is advantageous since some detergents are deleterious to the crystallization and ionization steps in MALDI-TOF (Chait and Roepstorff, 1993). Prior to addition of the enzyme, the membrane is blocked with polyvinylpyrrolidone (PVP) 360 to prevent the adsorption of the endoglycosidase (Henzel et al., 1994). The membrane was washed extensively with water following the blocking step to remove excess PVP 360 that could interfere with MALDI-TOF analysis.
Figure 1. Flow-scheme for 96-well plate assay. Reduction and carboxymethylation (RCM) buffer is described in the methods section. DTT, dithiothreitol; IAA, iodoacetic acid; PVP (spiky circles). polyvinylpyrrolidone.
Optimization of enzymatic deglycosylation conditions
The peptide N-glycosidase F (PNGaseF) digestion was performed in 10 mM Tris-acetate (pH 8.3) instead of the manufacturer recommended incubation buffer (20 mM sodium phosphate pH 7.5 containing 50 mM EDTA). The combination of phosphate buffer and subsequent desalting with cation-exchange resin leads to the formation of lactones that complicate the MALDI-TOF spectra, and decrease the signal-to-noise achieved (Papac et al., 1996). The formation of lactones is indicated by the appearance of peaks 18 Da less than the expected molecular weight for sialylated oligosaccharides. Fortunately, the use of cation-exchange resin to desalt oligosaccharides released in Tris-acetate does not promote the formation of lactones. The PNGaseF prepared by Oxford GlycoSystems is preferred because it is formulated in 20 mM Tris-HCl.
The concentration of PNGaseF, and incubation time necessary to achieve complete deglycosylation of rt-PA was determined by comparing the absolute signal from the MALDI-TOF mass spectra obtained for the released N-linked oligosaccharides. Twenty micrograms of rt-PA were applied to each well, and the concentration of PNGaseF added in 50 µl of digestion buffer was varied from 3.1 to 50 U/ml. Complete deglycosylation was achieved in 18 h with as little as 6.3 U of PNGaseF per ml of digestion buffer (data not shown). Using 12.5 U/ml of PNGaseF, no change in the absolute response of the MALDI-TOF mass spectra was observed with digestion times of 2, 4, or 24 h, suggesting that complete digestion had occurred within the first two h. These observations have led to adoption of the following digestion conditions (25 U/ml of PNGaseF for 3 h), thereby insuring complete deglycosylation (Figure 1).
Tryptic maps of rt-PA obtained either before (Figure 2A) or after deglycosylation with PNGaseF (Figure 2B) were examined to determine the extent of deglycosylation. Both rt-PA samples used in the tryptic mapping experiments were derived from 50 µg of rt-PA applied to wells. The deglycosylated rt-PA was obtained using the procedure outlined in Figure 1. The region of the tryptic map shown contains more than 13 peptides (Figure 2). The two labeled peaks are two of the three glycopeptides expected for rt-PA. The T11 glycopeptide which contains the high mannose oligosaccharides is shown to be completely deglycosylated (compare Figure 2, A and B) (Vehar et al., 1986). Because two other peptides coelute with the T17 glycopeptide, a similar conclusion cannot be drawn directly from the chromatograms (Figure 2A,B). Hence, fractions spanning the expected elution time of the three glycopeptides (T45, T17, and T11) were collected and then analyzed by MALDI-TOF. No ions corresponding to the expected m/z for the three glycopeptides were detected (data not shown).
Figure 2. Tryptic maps of 50 µg of recombinant tissue-type plasminogen activator (A) before and (B) after digestion with peptide N-glycosidase F.
Recovery of N-linked oligosaccharides from wells
The recovery of oligosaccharide standards spiked into wells was investigated for two reasons (Table I). First, poor recovery may raise the limit of detection. Secondly, oligosaccharides may be recovered differentially based upon their monosaccharide composition. Typically, no less than 48 µl of the total 50 µl of digestion buffer were recovered from a well, and no more than 3 µl of buffer was lost in the cation-exchange resin. No preferential recovery of either the neutral or acidic oligosaccharides was observed throughout the experiment (Table I).
Desialylation of the acidic oligosaccharide was detected when the samples from the recovery experiment were analyzed by HPAEC-PAD (data not shown). Roughly 15% of the disialylated diantennary oligosaccharide was desialylated to the monosialylated and nonsialylated form. Therefore, the di-, mono- and nonsialylated forms were summed to calculate recovery (Table I). To eliminate acetate ions which adversely affected the chromatographic separation, the samples were dried in a vacuum centrifuge. This drying process was found to promote desialylation of the disialylated oligosaccharide. If the samples spiked into the wells were directly injected onto the HPAEC-PAD after acetic acid treatment and cation-exchange chromatography without drying, only 3% of the disialylated oligosaccharide lost sialic acid. Samples that were not treated with acetic acid prior to desalting with cation-exchange resin also experienced a 3% desialylation suggesting that the cation-exchange resin was responsible for this low level of desialylation.
Table I
| Sample treatment | Peak area percenta | |
| Neutral | Acidic | |
| Untreated (control) | 100 ± 2 | 100 ± 2 |
| Only cation-exchange | 90 ± 2 | 89 ±4 |
| Acetic acid + cation-exchange | 90 ± 2 | 90 ±3 |
| Microplate + acetic acid + cation-exchange | 73 ± 4 | 76 ± 6 |
Comparison of the 96-well method to the standard solution method
We assessed the accuracy of the 96-well plate method for recovery of oligosaccharides by comparing it to the method performed in solution. Figure 3 compares the HPAEC-PAD profiles of the N-linked oligosaccharides released from 50 µg of rt-PA using either the standard solution conditions (Figure 3A) or the 96-well plate method (Figure 3B). All of the major peaks corresponding to known oligosaccharide structures are observed in relatively comparable proportions. The major difference between the two chromatograms is that the sample from the well produced half the absolute response as the same amount of protein from the solution method. This difference is attributed primarily to the binding capacity of the membrane being exceeded when 50 µg of rt-PA is applied.
Figure 3. HPAEC profile of the N-linked oligosaccharides released from 50 µg of recombinant tissue-type plasminogen activator digested with PNGaseF in (A) solution and (B) a well.
A prerequisite for obtaining an accurate profile of the oligosaccharides by either HPAEC-PAD or MALDI-TOF is acetic acid treatment of the released oligosaccharides prior to cation-exchange chromatography (Figure 1). The acid treatment is necessary to convert the glycosylamine form of the oligosaccharide into the reducing form (Hardy and Townsend, 1994). If acetic acid pretreatment is omitted, less than 50% of the oligosaccharides are recovered from the cation-exchange resin based upon the HPAEC-PAD profile (data not shown). Furthermore, a preferential loss of the neutral species over the sialylated oligosaccharides is observed in both the HPAEC-PAD chromatogram and the MALDI-TOF mass spectra. Interestingly, when acetic acid pretreatment is omitted, the MALDI-TOF data suggest that neutral oligosaccharides that contain a core fucose on the reducing end are recovered with higher efficiencies than high mannose oligosaccharides that do not contain core fucose (data not shown).
MALDI-TOF mass spectrometry of the released oligosaccharides
Figure 4 illustrates a typical negative-ion mass spectrum obtained for N-linked oligosaccharides released from 5 µg of rt-PA applied to a well. The acidic, sialic acid containing, oligosaccharides are preferentially observed as (M-H)- ions, since the spectrum is acquired in the negative-ion mode (Papac et al., 1996). To detect the neutral oligosaccharides (e.g. high mannose) as the (M+Na)+ ions, the spectra need to be acquired in the positive-ion mode (Figure 5). The high mannose oligosaccharides can be detected in the negative-ion mass spectrum as citrate adducts (M+191)- (Figure 4, ions labeled with *). The spectra shown in Figure 4 and 5 were acquired with only 1% of the material recovered from a well. This corresponds to analyzing the oligosaccharides released from [sim]850 fmol of rt-PA. Even the minor hybrid species, Hyb 4 (Table II), is observed with a signal-to-noise of greater than 10:1. The mass accuracy for the smallest signal labeled, Hyb 4, is better than 0.05%. This mass accuracy combined with previously obtained oligosaccharide structural information is sufficient to infer the structures shown in Table II and III (Spellman et al., 1989). All structures reported by Spellman et al. (1989) are observed in the two mass spectra (Figures 4 and 5). No ions corresponding to losses of H2O are observed in Figure 4 indicating that lactone formation of sialylated oligosaccharides does not occur during isolation.
Figure 4. Negative-ion MALDI-TOF mass spectrum of 1% of the N-linked oligosaccharides released from 50 µg of recombinant tissue-type plasminogen activator. The labels are defined in Table II. The '*" indicates citrate adducts of neutral oligosaccharides. The spectrum was acquired using THAP as the matrix and was smoothed with a 19 point Savitsky-Golay function.
Figure 5. Positive-ion MALDI-TOF mass spectrum of 1% of the N-linked oligosaccharides released from 50 µg of recombinant tissue-type plasminogen activator. The labels are defined in Table III. The spectrum was acquired using sDHB as the matrix and was not smoothed.
Table II
One of the most useful aspects of this method is the ability to manipulate sub-microgram quantities of glycoproteins for subsequent MALDI-TOF mass spectrometric analysis. Figure 6 demonstrates the capabilities of this method to release the N-linked oligosaccharides from 0.1-1 µg of rt-PA for analysis by negative-ion MALDI-TOF. A similar series of spectra are obtained in the positive-ion mode with 1 µg, 0.5 µg, and 0.25 µg of rt-PA (data not shown). To ensure that the amount of protein loaded into each well was the expected concentration based upon serial dilution (5 mg/ml starting concentration), the concentration of the final dilution used to load the wells was verified by amino acid analysis. The concentration of rt-PA determined by amino acid analysis was about 15% lower than that estimated by absorbance at 280 nm.
Figure 6. Negative-ion MALDI-TOF mass spectrum of 2.5% of the N-linked oligosaccharides released from (A) 1 µg, (B) 0.5 µg, and (C) 0.1 µg of recombinant tissue-type plasminogen activator. The labels are defined in Table II. The spectra were acquired using THAP as the matrix and were smoothed with a 19 point Savitsky-Golay function.
One change to the standard protocol that enabled the acquisition of MALDI-TOF spectra for 0.1-1 µg of rt-PA was reducing the PNGaseF digestion volume from 50 µl to 20 µl. With only 0.5 µg of rt-PA applied to a well, all the oligosaccharide species labeled in Figure 6 were detectable with a signal-to-noise of greater than 5:1. Even when 0.1 µg of rt-PA was applied to the well, the five most abundant oligosaccharide species were detected (Figure 6C). In all three spectra, only 2.5% of the released oligosaccharides were analyzed. For 0.1 µg of rt-PA, this corresponds to analyzing the oligosaccharides released from [sim]42 fmol of protein. Notice in Figure 6 the quantitative nature of the MALDI-TOF response first noted by Harvey et al. (1993). Compared to the 1.0 µg load of rt-PA, the 0.5 µg load gives approximately half the signal and the 0.1 µg load gives about a tenth of the signal. This same trend in the absolute response of the MALDI-TOF signal was observed for rt-PA loadings of 2.5, 5.0, 10, and 20 µg (data not shown).
Discussion
PVDF-lined Multiscreen Assay plates enabled the development of a high-throughput method for release of N-linked oligosaccharides from low microgram quantities of glycoproteins. The 96-well format is a familiar approach often utilized to streamline analytical assays. By incorporating a PVDF-lined 96-well plate into the sample preparation strategy for release of N-linked oligosaccharides, we can now efficiently release the oligosaccharides from 60 glycoprotein samples in 1 day. Furthermore, the 96-well format offers the potential for automation of the assay. We have chosen rt-PA as the model glycoprotein for development of this assay, because rt-PA contains high mannose, complex, and hybrid oligosaccharides (Spellman et al., 1989)
When using this method as a high-throughput assay, 50 µg of each glycoprotein sample is typically loaded into an individual well. According to the manufacturer, the capacity of an individual well is 25 µg for bovine serum albumin, and 47 µg for an immunoglobulin. Both HPAEC-PAD and MALDI-TOF demonstrated that loading 50 µg of rt-PA into an individual well exceeded the protein binding capacity of the well for rt-PA. The response observed in the HPAEC-PAD chromatogram for the N-linked oligosaccharides released from 50 µg of rt-PA loaded into a well was almost half of that expected (Figure 3). In addition, the MALDI-TOF signal obtained for the oligosaccharides released from 40 µg rt-PA was the almost same as the signal obtained from 20 µg. We have found that the strength of the signal from the MALDI-TOF spectrum provides a good correlation to the amount of material loaded (Figure 6). Based upon the MALDI-TOF and HPAEC data, the capacity of a well for rt-PA is [sim]25 µg. In spite of these observations made for rt-PA, 50 µg of glycoprotein has been chosen as the standard loading size, except in instances with limited sample, to maximize the amount of glycans recovered. Although not demonstrated, exceeding the membrane capacity may enable the PVP 360 blocking step to be eliminated. A caveat to exceeding the membrane capacity is that preferential binding of under-glycosylated forms of the glycoprotein could occur. Fortunately, evidence of preferential binding was not observed for rt-PA by MALDI-TOF when the total protein binding capacity of a well was exceeded.
One of the intended uses of this assay is to release N-linked oligosaccharides from large numbers of various glycoprotein samples so that the glycoform profile of recombinantly expressed glycoprotein therapeutics can be monitored during the development and manufacturing process. This use dictates that the oligosaccharides released and recovered accurately reflect their distribution found on the glycoprotein. A potential problem that could alter the observed distribution of oligosaccharides would be if site- or structure-based preferential release of oligosaccharides occurred. This concern is most easily eliminated by using conditions that result in the complete release of all the N-linked oligosaccharides. Several aspects of the method were investigated and incorporated into the approach to facilitate the complete release of the N-linked oligosaccharides.
The first step incorporated into the method to ensure complete deglycosylation was to denature, reduce, and alkylate the protein. This procedure unfolds the protein minimizing steric hindrance of the endoglycosidase to the glycosylation sites, thereby enabling complete deglycosylation. Although often used, detergents are not a generally applicable method for unfolding glycoproteins for digestion with PNGaseF, since this enzyme is sensitive to the presence of ionic detergents (O'Neill, 1996). In some instances, higher concentrations of denaturing detergents (e.g., sodium dodecyl sulfate) than PNGaseF can tolerate may be required for complete denaturation of the glycoprotein being digested. Furthermore, detergents can interfere with subsequent MALDI-TOF analysis (Chait and Roepstorff, 1993).
A second aspect investigated that could influence the extent of deglycosylation was the amount of enzyme and time of incubation required. An effective means to assess the completeness of deglycosylation was to monitor the absolute MALDI-TOF signal of the released oligosaccharides. Although the absolute response of a signal in MALDI-TOF can fluctuate for many analytes, oligosaccharides have been shown to behave somewhat more predictably (Harvey, 1993). A 5 U/ml concentration of PNGaseF has been reported adequate for complete deglycosylation of several glycoproteins except [alpha]1-acid glycoprotein (Hirani et al., 1987) . Even though complete deglycosylation of reduced and alkylated rt-PA was achieved with as little as 6 U/ml of PNGaseF, 25 U/ml has been adopted to ensure complete deglycosylation with the short incubation times used.
Table III
Another feature of our incubation conditions which favors complete deglycosylation is the low ionic strength of the incubation buffer used (10 mM Tris-acetate, pH 8.3). The use of a low ionic strength digestion buffer has been shown to increase the rate of PNGaseF-mediated deglycosylation (Gosselin et al., 1992). There are two advantages of using a low ionic strength digestion buffer. First, complete deglycosylation can be achieved within two h as determined by the absolute response of the oligosaccharide signal in the MALDI-TOF mass spectra. Second, there is less Tris-acetate in the digestion buffer that could interfere with crystallization and ionization during MALDI-TOF analysis.
An attribute of the 96-well approach is that tryptic maps can also be generated to determine the extent of deglycosylation or to investigate O-linked glycosylation sites. In this example with rt-PA, all three tryptic glycopeptides were verified by MALDI-TOF as being completely deglycosylated. Even the high mannose-containing glycopeptide (T11), which is the most resistant to deglycosylation with PNGaseF, was completely deglycosylated (unpublished observation).
An additional aspect of the sample preparation procedure, which required careful scrutiny to ensure that the oligosaccharides recovered accurately reflected their distribution on the glycoprotein, was recovery of the released oligosaccharides. Besides influencing the ultimate sensitivity of this approach, recovery could impact the accuracy of this approach if there was a preferential recovery. We did observe a 75% recovery for the two standard oligosaccharides; 10% was attributed to loss within the cation-exchange resin, the other 15% was probably lost in the well as unrecovered digestion buffer and oligosaccharide entrained within the membrane (Table I). The oligosaccharide was probably not bound to the PVDF membrane (Weitzhandler et al., 1993). This 25% loss of material did not result in any noticeable preferential loss as evidenced by similar recoveries for a neutral and an acidic oligosaccharide (Table I). Moreover, the similarity of the HPAEC-PAD chromatograms of oligosaccharides released using either standard solution conditions or from the 96-well method further supports a uniform recovery. A second concern addressed by the recovery experiment was the extent of desialylation. Desialylation would alter the distribution of glycoforms thus overestimating the proportion of oligosaccharides not terminating with sialic acid. A significant amount of desialylation was observed in the recovery experiments when samples were dried prior to HPAEC-PAD analysis. Fortunately, eliminating the drying step prior to HPAEC-PAD analysis minimized desialylation to about 3%. We are currently investigating several strategies to completely eliminate desialylation.
In addition to increasing sample throughput, use of the 96-well plate lined with PVDF decreased the amount of a glycoprotein necessary to obtain a MALDI-TOF oligosaccharide map. The immobilization of the glycoprotein to a solid support enabled all sample preparation steps to be performed in one reaction vessel, thereby minimizing adsorptive losses that occur during sample handling (e.g., buffer exchanges, enzyme digestion, alkylation). Immobilization of the glycoprotein to PVDF was the primary factor that allowed reproducible MALDI-TOF oligosaccharide maps to be obtained with 0.5 µg of rt-PA (Figure 6B).
In addition to immobilizing the glycoprotein to PVDF, two additional steps were taken to obtain a MALDI-TOF mass spectrum for N-linked oligosaccharides released from sub-microgram amounts of rt-PA. The first step, was to reduce the sample digestion volume from 50 µl to 20 µl. The spectra shown in Figure 6 were acquired with only 2.5% of the collected sample. Concentration of the sample potentially offers a means to lower the working range of this assay. Unfortunately, concentrating the oligosaccharides by drying the sample and reconstituting the dried oligosaccharides did not improve the signal-to-noise of the spectra and can result in acid-catalyzed desialylation.
PNGaseF is an amidase that releases N-linked oligosaccharides as glycosylamines (Tarentino et al., 1985). Failure to treat the samples with acetic acid prior to cation-exchange chromatography leads to both preferential recovery of the sialylated oligosaccharides and poor mass accuracy that is readily observed in the mass spectra and HPAEC-PAD chromatograms (data not shown). The preferential loss of neutral oligosaccharides over sialylated oligosaccharides can be partially attributed to the negative charge found on the sialic acid which probably prevents strong binding of the sialylated oligosaccharide to the cation-exchange resin. A second reason may be attributed to the observation that both neutral and acidic oligosaccharides containing core fucose were better recovered. Greater than 90% of the sialylated oligosaccharides contain core fucose, whereas approximately 90% of the neutral oligosaccharides do not contain fucose. This suggests that fucose attached [alpha]1,6 to the reducing end may hasten conversion of the glycosylamine to the free reducing sugar. Based upon the recovery experiments, the acetic acid treatment does not promote acid-catalyzed desialylation.
Previously we have demonstrated that the limit of detection for a disialylated diantennary oligosaccharide containing core fucose (2122) is [sim]25 fmol (applied to the MALDI-TOF target) with a signal-to-noise of 5:1 (Papac et al., 1996). Our ability to analyze 2.5% of the oligosaccharides released from 0.1 µg of rt-PA equates to 50 fmol of the most abundant species applied to the target (2122, Table II). The disialylated diantennary oligosaccharide with core fucose, 2122, is detected with a signal-to-noise ratio of >10:1 (Figure 6C). This suggests that a high recovery of oligosaccharides is achieved even at the sub-microgram level. Using the 96-well plate format for oligosaccharide release, we can now exploit the sensitivity of MALDI-TOF. With this method, the minimum amount of protein necessary for detection of the major sialylated oligosaccharides found on rt-PA using MALDI-TOF as the detector is 100 ng of rt-PA. For the neutral oligosaccharides, the minimum amount of protein needed is 250 ng of rt-PA. The greater amount of protein required for the neutral oligosaccharides is due to the method of detection not being as sensitive for neutral oligosaccharides. Obviously, the minimum amount of protein required will also be dependent on the extent of glycosylation of the glycoprotein under investigation.
Competition for ionization and nonuniform distribution of sample deposition are two factors which could adversely affect the MALDI-TOF analysis. We have taken significant number of steps to control these processes and can demonstrate that MALDI-TOF is a quantitative technique. A detailed description of the quantitative capabilities of MALDI-TOF mass spectrometry will be presented in a separate article.
A major factor which stimulated the development of this methodology was the increasing development of glycoprotein pharmaceuticals using recombinant DNA technology. The potential, afforded by the high sensitivity of MALDI-TOF mass spectrometry, to obtain detailed information on glycan structure could not be realized without improved sample preparation methods. The use of the PVDF membranes to manipulate small samples is anticipated to have wide applicability. This article has focused on sample preparation for glycan analysis and we have begun to employ this method in two main areas: investigation of the metabolic clearance of glycoproteins, and the development of cell culture conditions and recovery process methods for glycoprotein pharmaceuticals. For pharmacokinetic studies, we have found that it is possible to recover submicrogram amounts of an immunoadhesin from circulation by affinity chromatographic methods, and directly assess the role of the glycans in clearance mechanisms without needing to increase the dose to suprapharmacological levels (unpublished observations). For product development purposes, the ability to evaluate the glycan profiles from glycoprotein products (i.e., monoclonal antibodies and immunoadhesions) on small samples has wide-ranging benefits such as initial clone selection and optimization of cell culture parameters. The high throughput capability allows a rigorous assessment of many candidate clones or many cell culture parameters at small scale. This avoids the significant efforts needed to scale up to obtain larger quantities of material (required by previous sample preparation and analysis methods) before key decisions can be made. This can obviously be extended to monitoring product quality during cell culture optimization and scale-up from research (at milliliter and liter scales) to production (at kiloliter scale and beyond).
Broader applications of arrays of samples immobilized on PVDF membranes could be envisioned. As noted above, the released glycans could be analyzed by chromatographic methods (with or without derivatization) or exoglycosidases could be used in place of the PNGaseF used this example. These applications should be generally useful in studying glycoprotein structure/function and in providing a detailed understanding of control of glycosylation phenomena. The ability to process many samples in parallel without significant adsorptive losses should also find general utility in the peptide mapping of proteins.
Materials and methods
Chemicals and biochemicals
The matrix 2[prime],4[prime],6[prime]-trihydroxyacetophenone monohydrate, 2,5-dihydroxybenzoic acid, and 5-methoxysalicylic acid were purchased from Aldrich (Milwaukee, WI). The oligosaccharide standards were purchased from Oxford GlycoSystems (Rosedale, NY). HPLC grade acetonitrile was obtained from Burdick and Jackson (Muskegon, Wisconsin). The ammonium citrate (dibasic), iodoacetic acid, dithiothreitol, and polyvinylpyrrolidone 360 were acquired from Sigma Chemical (St. Louis, MO), and the trifluoroacetic acid was purchased from Pierce Chemical (Rockford, IL). The Ultra-Pure Tris base used was acquired from J. T. Baker (Phillipsburg, NJ). The sodium hydroxide used for HPAEC-PAD was analytical reagent grade purchased from Mallinckrodt (Paris, KY). All water used was from a Milli-Q water purification system (Millipore, Bedford, MA).
Release of N-linked oligosaccharides from rt-PA on 96-well plate (Figure 1)
The 36 outer perimeter wells of a MultiScreen-IP plate (pore size 0.45 µm, Millipore) were filled with deionized water to minimize evaporative losses during sample incubation in a heated environment. All 96 wells can be used and evaporative losses can be minimized during sample incubation, if the plate is placed in a ziplock bag containing a moist paper towel. To prepare the remaining 60 interior wells for adsorption of the glycoprotein, the wells were wetted with 100 µl of methanol that was drawn through the PVDF membrane using a house vacuum applied to the MultiScreen vacuum manifold (Millipore). The PVDF membrane was next washed three times with 300 µl of water, and once with 50 µl of RCM buffer (8 M urea containing 360 mM Tris, pH 8.6, and 3.2 mM EDTA). Between all washing steps, the wells were allowed to drain completely. The desired amount of glycoprotein, rt-PA (Genentech, Inc., So. San Francisco, CA) was loaded into a well already containing 10 µl of RCM buffer. The total volume of the rt-PA solution was brought to no less than 50 µl with additional RCM buffer, and was drawn through the membrane with a gentle vacuum.
After applying the glycoprotein to the PVDF membrane, the wells were washed twice with 50 µL aliquots of RCM buffer. Reduction was performed by addition of 50 µl of 0.1 M dithiothreitol in RCM buffer and incubation at 37°C for 1 h. Following reduction, a vacuum was applied to the wells to remove the dithiothreitol, and then the wells were washed three times with 300 µl of deionized water. Carboxymethylation of the cysteine residues was performed by addition of 50 µl of 0.1 M iodoacetic acid in RCM buffer, followed by incubation for 30 min at room temperature in the dark. The excess iodoacetic acid was removed by gentle vacuum, and then the wells were washed three times with 300 µl of deionized water.
To prevent adsorptive losses of the endoglycosidase, the PVDF membrane was blocked by addition of 100 µl of a 1% aqueous solution of polyvinylpyrrolidone 360 and incubation at room temperature for 60 min. The blocking agent was then removed with a gentle vacuum, and the wells washed three times with 300 µl of deionized water.
N-linked oligosaccharides were released from the reduced and carboxymethylated rt-PA by addition of 1.25 U of peptide-N-glycosidase F (Oxford GlycoSystems) diluted in either 50 µl or 20 µl of 10 mM Tris-acetate (pH 8.3). Digestion was performed at 37°C for 3 h with the wells of the MultiScreen 96-well plate covered with Parafilm to prevent evaporation of the digestion buffer.
Following digestion, the Tris-acetate buffer containing the released N-linked oligosaccharides was transferred by pipette to 0.5 ml polypropylene tubes. The released oligosaccharides are then incubated 3 h at room temperature with 150 mM acetic acid. The PNGaseF, Tris, and any residual sodium or potassium ions were removed prior to MALDI-TOF analysis by desalting the samples with a 0.3 ml bed of cation-exchange resin (AG50W-X8 resin, hydrogen form, 100-200 mesh; Bio-Rad, Hercules, CA) packed into compact reaction columns (Amersham, Cleveland, OH).
Release of N-linked oligosaccharides from rt-PA in solution
Recombinant tissue-type plasminogen activator was reduced with 42 mM dithiothreitol in RCM buffer for 10 min at 37°C. Carboxymethylation was achieved by bringing the concentration of iodoacetic acid to 50 mM and incubating at 37°C for 30 min in the dark. The reaction was quenched by addition of excess dithiothreitol. The alkylated rt-PA was dialyzed four times in 1 l of 75 mM sodium phosphate, 5 mM EDTA (pH 8.0) using a 6000-8000 molecular weight cutoff dialysis membrane (Spectrum, Houston, TX). Following dialysis, the protein concentration was estimated by reading the UV absorbance at 280 and 320 nm([epsis]280 = 1.7 absorbance units/mg/ml of protein).
The N-linked oligosaccharides were released from 200 µg of reduced and carboxymethylated rt-PA by incubation with 1 U of recombinant peptide-N-glycosidase F (Oxford GlycoSystems) for 18 h at 37°C. The released oligosaccharides were recovered after precipitation of the protein with 75% ethanol. Following drying of the recovered supernatant, the oligosaccharides were redissolved in 13 mM acetic acid and incubated at room temperature for 2 h. The acid-treated samples were dried in a Savant Speed Vac and then dissolved in deionized water to a concentration of 1 mg/ml (based on the starting protein content of the samples) prior to HPAEC-PAD analysis.
Recovery of N-linked oligosaccharides from wells
To assess recovery, 500 pmol of a disialylated diantennary oligosaccharide containing core fucose and 500 pmol of high mannose oligosaccharide were spiked into wells containing 25 U/ml of PNGaseF in 50 µl of 10 mM Tris-acetate (pH 8.3). The wells were treated with the same reduction, carboxymethylation, and blocking steps as wells containing intact rt-PA (Figure 1). Following a 2 h incubation at 37°C, the digestion buffer containing the oligosaccharide standards was acidified with acetic acid to mimic conditions normally used (Figure 1). After a 2 h incubation at room temperature, the sample was desalted on cation-exchange resin, dried in a vacuum centrifuge and reconstituted with 50 µl of deionized water prior to HPAEC-PAD analysis. The solution used to spike the wells was analyzed directly by HPAEC-PAD to determine the response expected for 100% recovery.
Tryptic mapping analysis of rt-PA
The PVDF membrane from wells containing 50 µg of either PNGaseF-treated or untreated rt-PA was punched out into 1.5 ml polypropylene tubes, and suspended in 75 µl of 100 mM ammonium bicarbonate containing 10% acetonitrile. The PVDF-bound rt-PA was digested with sequencing-grade modified trypsin (Promega, Madison, WI) at an enzyme to substrate ratio of 1:20 on a per weight basis. Incubation was performed overnight at 37°C, and stopped by addition of 75 µl of 4% trifluoroacetic acid and 100 µl of 0.1% trifluoroacetic acid.
Separation of the tryptic peptides was achieved by injecting 125 µl of the digest onto a HP-1090 Series II liquid chromatograph (Hewlett Packard, Palo Alto, CA) equipped with a Vydac (2.1 × 250 mm) C-18 column (Hesperia, CA). The column temperature was maintained at 40°C and the flow-rate was 0.2 ml/min. Mobile phase A contained 0.1% trifluoroacetic acid in water, and mobile phase B contained 0.1% trifluoroacetic acid in acetonitrile. The gradient program began by increasing mobile phase B linearly to 25% B in 50 min after initially holding the mobile phase at 0% B for 5 min. The gradient was then ramped to 60% B in 35 min, and finally the mobile phase was increased to 90% B in 10 min. The absorbance was monitored with a diode-array detector at 214 nm.
High-pH anion-exchange chromatography
The high-pH anion-exchange chromatography with pulsed amperometric detection was performed with a DX-500 system (Dionex, Sunnyvale, CA) equipped with a CarboPac PA-100 anion-exchange column (4 × 250 mm, Dionex) and an ED-40 electrochemical detector (Dionex). Separation of the oligosaccharides was achieved with a gradient program at a flow-rate of 1 ml/min operated at ambient temperature. Solvent A consisted of 0.1 N NaOH and solvent B consisted of 0.1 N NaOH containing 0.5 M sodium acetate. The column was equilibrated with 2.5% solvent B for 3 min. Using a linear gradient, the oligosaccharides were eluted by increasing solvent B to 40% in 50 min.
Matrix and sample preparation
The 2[prime],4[prime],6[prime],-trihydroxyacetophenone matrix (THAP) was prepared by dissolving 2 mg of 2[prime],4[prime],6[prime],-trihydroxyacetophenone in 1 ml of acetonitrile/13.3 mM ammonium citrate (dibasic) 1:3 (v/v). The 2,5-dihydroxybenzoic acid matrix (sDHB) was prepared by dissolving 2 mg of 2,5-dihydroxybenzoic acid + 0.1 mg of 5-methoxysalicylic acid in 1 ml of ethanol/10 mM aqueous sodium chloride 1:1 (v/v).
Typically, 0.5 µl of analyte was applied to a polished stainless steel target, and 0.3 µl of matrix was then added to the analyte. All samples were dried under vacuum (50 × 10-3 Torr), and then allowed to absorb moisture from the atmosphere (Papac et al., 1996).
MALDI-TOF analysis
The MALDI-TOF mass spectrometer used to acquire the mass spectra was a Voyager Elite (PerSeptive Biosystems, Framingham, MA) equipped with delayed extraction. All samples were irradiated with ultraviolet light (337 nm) from a N2 laser. The instrument was operated in the linear configuration (2m flight-path). Both negative and positive ions were accelerated to 20 kV after an 80 ns delay.
A two-point external calibration using oligosaccharide standards was used for mass assignment of the ions. Generally, a mass accuracy of <0.1% was obtained using external calibration.
Spectra from 256 laser shots were summed to obtain the final spectrum. This instrument utilized a multichannel plate detector. Therefore to prevent matrix ions from saturating the detector, the detector voltage was held below threshold until ions of mass to charge greater than 500 m/z could strike the detector.
Acknowledgments
We thank Michael Molony for performing the amino acid analysis, and Laura Lerner for helpful discussions.
Abbreviations
DNA, deoxyribonucleic acid; DTT, dithiothreitol; EDTA, ethylenediamine tetraacetic acid; HPAEC-PAD, high pH anion-exchange chromatography with pulsed amperometric detection; IAA, iodoacetic acid; MALDI-TOF, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; PNGaseF, peptide N-glycosidase F; PVDF, polyvinylidene difluoride; PVP, polyvinylpyrrolidone; RCM, reduction and carboxymethylation; rt-PA, recombinant tissue-type plasminogen activator; sDHB, 2,5-dihydroxybenzoic acid matrix; THAP, 2[prime],4[prime],6[prime]-trihydroxyacetophenone matrix
References
1Present address: NPS Pharmaceuticals, Inc., 420 Chipeta Way, Salt Lake, City, UT 84108, USA
2To whom correspondence should be addressed
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 2 Apr 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. Mille, P. Bobrowicz, P.-A. Trinel, H. Li, E. Maes, Y. Guerardel, C. Fradin, M. Martinez-Esparza, R. C. Davidson, G. Janbon, et al. Identification of a New Family of Genes Involved in {beta}-1,2-Mannosylation of Glycans in Pichia pastoris and Candida albicans J. Biol. Chem., April 11, 2008; 283(15): 9724 - 9736. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Kainz, A. Gallmetzer, C. Hatzl, J. H. Nett, H. Li, T. Schinko, R. Pachlinger, H. Berger, Y. Reyes-Dominguez, A. Bernreiter, et al. N-Glycan Modification in Aspergillus Species Appl. Envir. Microbiol., February 15, 2008; 74(4): 1076 - 1086. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. P Jongen, G. J Gerwig, B. R Leeflang, K. Koles, M. L. Mannesse, P. H. van Berkel, F. R Pieper, M. A Kroos, A. J. Reuser, Q. Zhou, et al. N-glycans of recombinant human acid {alpha}-glucosidase expressed in the milk of transgenic rabbits Glycobiology, June 1, 2007; 17(6): 600 - 619. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J.S. Jones, D. I. Papac, E. H. Chin, R. Keck, S. A. Baughman, Y. S. Lin, J. Kneer, and J. E. Battersby Selective clearance of glycoforms of a complex glycoprotein pharmaceutical caused by terminal N-acetylglucosamine is similar in humans and cynomolgus monkeys Glycobiology, May 1, 2007; 17(5): 529 - 540. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. O. Beenhouwer, E. M. Yoo, C.-W. Lai, M. A. Rocha, and S. L. Morrison Human Immunoglobulin G2 (IgG2) and IgG4, but Not IgG1 or IgG3, Protect Mice against Cryptococcus neoformans Infection Infect. Immun., March 1, 2007; 75(3): 1424 - 1435. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Kanda, T. Yamada, K. Mori, A. Okazaki, M. Inoue, K. Kitajima-Miyama, R. Kuni-Kamochi, R. Nakano, K. Yano, S. Kakita, et al. Comparison of biological activity among nonfucosylated therapeutic IgG1 antibodies with three different N-linked Fc oligosaccharides: the high-mannose, hybrid, and complex types Glycobiology, January 1, 2007; 17(1): 104 - 118. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Morelle, C. Flahaut, J.-C. Michalski, A. Louvet, P. Mathurin, and A. Klein Mass spectrometric approach for screening modifications of total serum N-glycome in human diseases: application to cirrhosis Glycobiology, April 1, 2006; 16(4): 281 - 293. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Ferrara, F. Stuart, P. Sondermann, P. Brunker, and P. Umana The Carbohydrate at Fc{gamma}RIIIa Asn-162: AN ELEMENT REQUIRED FOR HIGH AFFINITY BINDING TO NON-FUCOSYLATED IgG GLYCOFORMS J. Biol. Chem., February 24, 2006; 281(8): 5032 - 5036. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Sturiale, R. Barone, A. Fiumara, M. Perez, M. Zaffanello, G. Sorge, L. Pavone, S. Tortorelli, J. F. O'Brien, J. Jaeken, et al. Hypoglycosylation with increased fucosylation and branching of serum transferrin N-glycans in untreated galactosemia Glycobiology, December 1, 2005; 15(12): 1268 - 1276. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Geysens, T. Pakula, J. Uusitalo, I. Dewerte, M. Penttila, and R. Contreras Cloning and Characterization of the Glucosidase II Alpha Subunit Gene of Trichoderma reesei: a Frameshift Mutation Results in the Aberrant Glycosylation Profile of the Hypercellulolytic Strain Rut-C30 Appl. Envir. Microbiol., June 1, 2005; 71(6): 2910 - 2924. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Bobrowicz, R. C. Davidson, H. Li, T. I. Potgieter, J. H. Nett, S. R. Hamilton, T. A. Stadheim, R. G. Miele, B. Bobrowicz, T. Mitchell, et al. Engineering of an artificial glycosylation pathway blocked in core oligosaccharide assembly in the yeast Pichia pastoris: production of complex humanized glycoproteins with terminal galactose Glycobiology, September 1, 2004; 14(9): 757 - 766. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Suzuki and Y. C. Lee Site-specific N-glycosylation of chicken serum IgG Glycobiology, March 1, 2004; 14(3): 275 - 292. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Koles, P. H. C. van Berkel, F. R. Pieper, J. H. Nuijens, M. L.M. Mannesse, J. F.G. Vliegenthart, and J. P. Kamerling N- and O-glycans of recombinant human C1 inhibitor expressed in the milk of transgenic rabbits Glycobiology, January 1, 2004; 14(1): 51 - 64. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Choi, N. Tomiya, J. H. Kim, J. M. Slavicek, M. J. Betenbaugh, and Y. C. Lee N-glycan structures of human transferrin produced by Lymantria dispar (gypsy moth) cells using the LdMNPV expression system Glycobiology, July 1, 2003; 13(7): 539 - 548. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Callewaert, E. Schollen, A. Vanhecke, J. Jaeken, G. Matthijs, and R. Contreras Increased fucosylation and reduced branching of serum glycoprotein N-glycans in all known subtypes of congenital disorder of glycosylation I Glycobiology, May 1, 2003; 13(5): 367 - 375. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B.-K. Choi, P. Bobrowicz, R. C. Davidson, S. R. Hamilton, D. H. Kung, H. Li, R. G. Miele, J. H. Nett, S. Wildt, and T. U. Gerngross Use of combinatorial genetic libraries to humanize N-linked glycosylation in the yeast Pichiapastoris PNAS, April 29, 2003; 100(9): 5022 - 5027. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. M. El-Fasakhany, K. Ichihara-Tanaka, K. Uchimura, and T. Muramatsu N-Acetylglucosamine-6-O-Sulfotransferase-1: Production in the Baculovirus System and Its Applications to the Synthesis of a Sulfated Oligosaccharide and to the Modification of Oligosaccharides in Fibrinogen J. Biochem., March 1, 2003; 133(3): 287 - 293. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Shields, J. Lai, R. Keck, L. Y. O'Connell, K. Hong, Y. G. Meng, S. H. A. Weikert, and L. G. Presta Lack of Fucose on Human IgG1 N-Linked Oligosaccharide Improves Binding to Human Fcgamma RIII and Antibody-dependent Cellular Toxicity J. Biol. Chem., July 19, 2002; 277(30): 26733 - 26740. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Callewaert, S. Geysens, F. Molemans, and R. Contreras Ultrasensitive profiling and sequencing of N-linked oligosaccharides using standard DNA-sequencing equipment Glycobiology, April 1, 2001; 11(4): 275 - 281. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. E. Idusogie, P. Y. Wong, L. G. Presta, H. Gazzano-Santoro, K. Totpal, M. Ultsch, and M. G. Mulkerrin Engineered Antibodies with Increased Activity to Recruit Complement J. Immunol., February 15, 2001; 166(4): 2571 - 2575. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T.S. Raju, J. B. Briggs, S. M. Borge, and A. J. S. Jones Species-specific variation in glycosylation of IgG: evidence for the species-specific sialylation and branch-specific galactosylation and importance for engineering recombinant glycoprotein therapeutics Glycobiology, May 1, 2000; 10(5): 477 - 486. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Shields, A. K. Namenuk, K. Hong, Y. G. Meng, J. Rae, J. Briggs, D. Xie, J. Lai, A. Stadlen, B. Li, et al. High Resolution Mapping of the Binding Site on Human IgG1 for Fcgamma RI, Fcgamma RII, Fcgamma RIII, and FcRn and Design of IgG1 Variants with Improved Binding to the Fcgamma R J. Biol. Chem., February 23, 2001; 276(9): 6591 - 6604. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Lee, S. Sundaram, N. L. Shaper, T. S. Raju, and P. Stanley Chinese Hamster Ovary (CHO) Cells May Express Six beta 4-Galactosyltransferases (beta 4GalTs). CONSEQUENCES OF THE LOSS OF FUNCTIONAL beta 4GalT-1, beta 4GalT-6, OR BOTH IN CHO GLYCOSYLATION MUTANTS J. Biol. Chem., April 20, 2001; 276(17): 13924 - 13934. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||