| Glycobiology | Pages |
Analysis of the N-acetylneuraminic acid and N-glycolylneuraminic acid contents of glycoproteins by high-pH anion-exchange chromatography with pulsed amperometric detection (HPAEC/PAD)
Introduction
Results
Discussion
Materials and methods
Abbreviations
References
Analysis of the N-acetylneuraminic acid and N-glycolylneuraminic acid contents of glycoproteins by high-pH anion-exchange chromatography with pulsed amperometric detection (HPAEC/PAD)
Introduction
Sialic acids (e.g., N-acetylneuraminic acid (Neu5Ac) and N-glycolylneuraminic acid (Neu5Gc)) occupy terminal positions in many glycoprotein and glycolipid oligosaccharides. The structures, biosynthesis, and functions of sialic acids have been reviewed (Schauer, 1991; Varki, 1992). When a mammalian glycoprotein loses sialic acid residues, the exposed galactose residues are recognized by the hepatic asialoglycoprotein receptor and the glycoprotein is removed from circulation (Ashwell and Morrell, 1974). In some cases the loss of sialic acid causes reduced activity (Takeuchi et al., 1989). Neu5Gc is found in all mammals except humans and apes. The importance of sialic acids to a glycoprotein's serum half-life, and possibly its activity, emphasize the need to determine the sialic acid content of a glycoprotein when assaying its function and/or its efficacy as a pharmaceutical.
Total sialic acids can be determined colorometrically using the periodate-thiobarbituric acid assay (Warren, 1959). After derivatization, individual sialic acids can be determined by gas chromatography (Reuter and Schauer, 1994) or by reversed-phase HPLC using either UV or fluorescence detection. The periodate-thiobarbituric acid assay has been adapted to reversed-phase HPLC to determine individual sialic acids, improve sensitivity, and reduce interferences (Powell and Hart, 1986). Using the fluorophore 1,2-diamino-4,5-methylenedioxybenzene (DMB), 57-192 fmol of Neu5Ac was detected in 10 µl of sample (Hara et al., 1989). Recently, ortho-phenylenediamine was shown to be a useful derivatization reagent for sialic acid analysis (Anumula, 1995). Sialic acids containing base-labile O-acetyl modifications were determined after derivatization with DMB, separation by reversed-phase HPLC, and detection using electrospray-ionization mass spectrometry (Klein et al., 1997). This technique is most useful for determining labile sialic acids present in low abundance. If sensitivity is not important, underivatized Neu5Ac and Neu5Gc can be detected by UV absorbance at 205 nm (Xia and Gilmer, 1994).
Neu5Ac and Neu5Gc from glycoproteins (Hermentin and Seidat, 1991; Seppala et al., 1991; Grollman, 1993), glycopeptides (Rohrer et al., 1993), a sialylated glycoconjugate (Jorge and Abdul-Wajid, 1995), equine serum (Manzi et al., 1990), and cytoplasmic extracts (Potvin et al., 1995) have been analyzed using HPAEC/PAD. Using HPAEC/PAD, Neu5Ac and Neu5Gc can be analyzed without derivatization. Hermentin and Seidat (1991) validated HPAEC/PAD to quantify the Neu5Ac content of [alpha]1-acid glycoprotein after 0.1N H2SO4 hydrolysis. The anion-exchange column (CarboPac PA1) commonly used with HPAEC/PAD was used to separate and quantify O-acetylated sialic acids at neutral pH (Manzi et al.,1990). These sialic acids can not be analyzed at high pH due to de-O-acetylation. The Neu5Ac content of a sialylated glycoconjugate used as an investigational vaccine for cancer patients was quantified using 0.1 N H2SO4 hydrolysis and HPAEC/PAD (Jorge and Abdul-Wajid, 1995). They used 2-keto-3-deoxyoctulosonic acid (KDO) as an internal standard in protein-containing samples to monitor differences in hydrolysis, sample handling, and PAD response. A recent report described a HPAEC/PAD assay for KDO after acid hydrolysis of a lipopolysacccharide (Kiang et al., 1997).
Here we show HPAEC/PAD analysis of Neu5Ac and Neu5Gc using a new anion-exchange column that causes Neu5Gc to elute earlier while keeping Neu5Ac well separated from the void volume. We determined the linearity, sensitivity, and reproducibility of the method, and evaluated its accuracy and reproducibility for analyzing Neu5Ac and Neu5Gc from acid and neuraminidase digests of glycoproteins. The present work improves the reproducibility and ruggedness of HPAEC/PAD sialic acid analysis. We and others have found that HPAEC/PAD sialic acid can be complicated by a need for equilibration at the start of an experiment until a stable peak area response is obtained. Additionally, over time (days to weeks) a progressive decrease in peak area response for sialic acids that correlates with electrode recession has been observed. To correct for these changes we evaluated KDN as a post-hydrolysis internal standard. We found that KDN was an effective internal standard. To understand the relationship between working electrode recession and Neu5Ac and Neu5Gc electrochemical responses, we evaluated the effect of working electrode recession on linearity, sensitivity, reproducibility, and sample analysis. We also measured Neu5Ac and Neu5Gc with eight randomly chosen used electrodes that had different levels of working electrode surface recession. As electrode recession increased, sensitivity decreased and the linear range shifted. Regardless of the amount of electrode recession, Neu5Ac, KDN, and Neu5Gc exhibited linear responses between 10 and 500 pmol. Use of the KDN internal standard permitted accurate analysis of the Neu5Ac and Neu5Gc contents of a glycoprotein with either a new or a recessed working electrode.
Results
Separation of Neu5Ac, KDN, and Neu5Gc
We evaluated gradient and isocratic separations of Neu5Ac, KDN, and Neu5Gc on the CarboPac PA10 column (Table I). We chose the gradient separation shown in Figure 1. With these separation conditions, Neu5Ac is separated from the void volume by 4.3 min, the internal standard KDN elutes between Neu5Ac and Neu5Gc, Neu5Gc elutes in less than 10 min, and 27 min are required between injections. If the CarboPac PA1 column is used, Neu5Ac and KDN elute at approximately the same time, but Neu5Gc elutes about a min later (data not shown).
Figure
Table I
| Separation type | Gradient time | [NaOH] (mM) | Initial [NaOAc] (mM) | Final [NaOAc] (mM) | Neu5Ac retention time | KDN retention time | Neu5Gc retention time |
| Gradient | 20 | 100 | 20 | 180 | 11.5 | 14.3 | 20.8 |
| Gradient | 20 | 100 | 50 | 180 | 8.8 | 11.8 | 19.2 |
| Gradient | 15 | 100 | 50 | 180 | 8.3 | 10.8 | 16.8 |
| Gradient | 10 | 100 | 50 | 180 | 7.8 | 9.7 | 14.4 |
| Gradient | 15 | 100 | 70 | 200 | 7.0 | 9.0 | 14.8 |
| Gradient | 15 | 100 | 70 | 250 | 6.8 | 8.7 | 13.0 |
| Gradient | 15 | 100 | 70 | 300 | 6.2 | 8.1 | 11.7 |
| Gradient | 15 | 100 | 70 | 350 | 6.1 | 7.8 | 10.8 |
| Gradient | 10 | 100 | 70 | 300 | 5.9 | 7.3 | 10.0 |
| Gradient | 15 | 100 | 100 | 200 | 5.5 | 7.2 | 13.0 |
| Gradient | 15 | 100 | 100 | 250 | 5.2 | 7.1 | 11.5 |
| Isocratic | - | 50 | 150 | 150 | 3.7 | 4.7 | 8.6 |
| Isocratic | - | 100 | 100 | 100 | 5.8 | 8.5 | 22.2 |
| Isocratic | - | 100 | 120 | 120 | 4.8 | 6.8 | 15.7 |
| Isocratic | - | 100 | 150 | 150 | 3.7 | 5.1 | 10.1 |
| Isocratic | - | 100 | 200 | 200 | 3.0 | 3.7 | 6.1 |
Sensitivity and linearity
We found that the limits of detection (signal [ge] 3× noise) of Neu5Ac, KDN, and Neu5Gc were dependent on working electrode wear. Using a working electrode whose surface was 350 microns recessed from the plastic block surface, we found the limits of detection of Neu5Ac, KDN, and Neu5Gc were 2, 5, and 2 pmol, respectively. This working electrode had not been reconditioned after analyzing monosaccharides and sialic acids for more than one year using the waveform described in Materials and methods. The Figure 1 inset shows a separation of 1 pmol of each analyte when a new working electrode was used for the analysis. Here the limits of detection of Neu5Ac, KDN, and Neu5Gc were 1, 2, and 0.5 pmol, respectively.
Figure 2 shows the linearity of this method between 10 and 2000 pmol of each analyte. The response of Neu5Ac and Neu5Gc is not linear beyond 2000 pmol and the response of KDN is linear to 10,000 pmol (r2 = 0.9997, data not shown). The above linearities were determined with the recessed working electrode used for the sensitivity studies. With a new working electrode, the linear range was 2-500 pmol (Neu5Ac r2 = 0.9985, KDN r2 = 0.9995, Neu5Gc r2 = 0.9972). The response of KDN was linear to 10,000 pmol (r2 = 0.9956).
Figure
ffect of electrode wear on sialic acid response
In an earlier study (prior to using a KDN internal standard) we assessed the effect of working electrode condition on the responses of Neu5Ac and Neu5Gc. Figure 3 shows the peak area responses to 250 pmol of Neu5Ac and Neu5Gc for eight working electrodes recessed from the electrode block to different extents. For both sialic acids, the greater the electrode recession, the lower the electrochemical response. Neu5Gc response is particularly sensitive to electrode recession, which is consistent with the observation that the Neu5Gc sensitivity is increased more than the Neu5Ac sensitivity when switching from a recessed to a new working electrode. The recessed electrode used in the sensitivity studies (not one of the eight electrodes in Figure 3) was recessed 350 µm. To evaluate the response differences between reconditioned (as described in Materials and methods) working electrodes, we measured the linearity of Neu5Ac and Neu5Gc between 10 and 2000 pmol after reconditioning six of the eight (randomly chosen) electrodes. The inset of Figure 3 shows the results of this study. The RSDs of the average areas of 2000 pmol Neu5Ac and Neu5Gc are 2.7% and 7.2%, respectively. For 250 pmol injected, the area RSDs for Neu5Ac and Neu5Gc are 3.7% and 5.1%, respectively. Therefore, this experiment also suggests that Neu5Gc is more sensitive than Neu5Ac to the condition of the working electrode.
Figure
Reproducibility Figure 4A shows the reproducibility of analyzing 200 pmol each of Neu5Ac, KDN, and Neu5Gc using a working electrode that was recessed 350 µm. An injection was made every 27 min for 48 h. Figure 4B shows that area reproducibility is not affected by interspersing acid hydrolyses and neuraminidase digestions of glycoproteins. Statistics from the experiments in Figure 4 and a second trial of each experiment are reported in Table II. In all four experiments, the area RSDs of Neu5Ac and Neu5Gc are improved when their areas are adjusted to reflect changes in the internal standard (KDN) response. The response for each sialic acid decreased from the first experiment without samples to the last experiment with interspersed samples (compare Figure 4A to 4B) which is consistent with a working electrode that is receding with use. Retention times were stable (RSDs < 0.5%) and unaffected by sample injections. Neither areas nor retention times showed increasing or decreasing trends during the 48 h analyses. When these experiments were repeated with a new working electrode (data not shown), area RSDs were slightly higher (2-3% without samples and 4-5% with samples) and retention time RSDs were <0.5%. The new electrode had approximately two times the electrochemical response for 200 pmol of each analyte compared to the recessed electrode (350 µm). Adjustment with the internal standard lowered area RSDs to 1-2% without samples and 2-3% with samples.
Figure
Analysis 1
Analysis 2
Analysis with interspersed samples 1
Analysis with interspersed samples 2
Average Neu5Ac area
857,306
750,399
695,410
673,250
Average KDN area
627,279
546,101
514,358
514,380
Average Neu5Gc area
1,212,326
1,069,418
997,346
972,178
Neu5Ac area RSD (%)
2.1
1.5
2.5
1.6
KDN area RSD (%)
1.5
1.2
2.1
1.3
Neu5Gc area RSD (%)
2.0
1.6
2.1
1.3
Neu5Ac area RSD (%)
after internal standard correction1.0
1.4
1.8
1.3
Neu5Gc Area RSD (%)
after internal standard correction1.1
1.2
1.4
0.8
Average Neu5Ac
retention time (min)5.86
5.76
5.82
5.80
Average KDN
retention time (min)7.19
7.07
7.11
7.09
Average Neu5Gc
retention time (min)9.90
9.77
9.82
9.80
Neu5Ac retention time RSD (%)
0.45
0.39
0.40
0.26
KDN retention time RSD (%)
0.39
0.21
0.35
0.26
Neu5Gc retention time RSD (%)
0.34
0.19
0.39
0.21
Average Neu5Ac (mol/mol protein)
Average Neu5Gc (mol/mol protein)
Sample
Experiment 1
Experiment 2
% Difference
Experiment 1
Experiment 2
% Difference
Bovine fetuin HClb
17 + <1
17 ± <1
0
0.51 ± <0.01
0.54 ± 0.05
+5.9
Bovine fetuin neuraminidase
16 ± <1
17 ± <1
+6.3
0.53 ± <0.01
0.59 ± 0.04
+11.3
Bovine transferrin HCl
1.1 ± <0.1
1.1 ± 0.1
0
1.8 ± <0.1
1.9 ± <0.1
+5.6
Bovine transferrin neuraminidase
1.2 ± <0.1
1.2 ± <0.1
0
2.3 ± <0.1
2.2 ± <0.1
-4.3
Human transferrin HCl
3.7 ± 0.1
3.8 ± <0.1
+2.7
Not detected
Not detected
-
Human transferrin neuraminidase
3.8 ± 0.1
3.7 ± <0.1
-2.6
Not detected
Not detected
-
Determination of Neu5Ac and Neu5Gc contents of glycoproteins
We evaluated the precision and accuracy of this method for determining the Neu5Ac and Neu5Gc compositions of bovine fetuin, bovine transferrin, and human transferrin. Sialic acids were released from each glycoprotein by 0.1 N HCl hydrolysis and by neuraminidase treatment. The Neu5Ac and Neu5Gc compositions of these three glycoproteins are reported in Table III. These compositions are in agreement with published values (Richardson et al., 1973; Spik et al., 1975; Spiro et al., 1974; Townsend et al., 1989). the small differences between the values of samples analyzed on different days show that the precision of this method is good. There is also agreement between the values from neuraminidase treatment and HCl hydrolysis. Comparison of the KDN peaks in the neuraminidase digests to those in the HCl hydrolyses revealed that the KDN peak is smaller in the HCl hydrolyses. The KDN peaks in the neuraminidase digests had the same areas as the standard (data not shown). Because the KDN is added after the digest has been dried, it is unlikely that the KDN has been degraded in the acid hydrolysate. In our calculations, the areas of Neu5Ac and Neu5Gc in the acid hydrolyses were adjusted to reflect the reduction of KDN electrochemical response observed between the standard and the sample. When Neu5Ac, KDN , and Neu5Gc were added individually to blank acid hydrolyses, their electrochemical responses were each reduced 10%. Experiments designed to elucidate the cause of the reduced electrochemical response of sialic acids in HCl hydrolysates revealed that the response of KDN was less suppressed in hydrolysates dried more than once and in samples with higher protein concentrations (Rohrer, unpublished observations). To evaluate whether electrode condition impacts the accuracy of sample analysis, we analyzed HCl hydrolyses and neuraminidase digests of bovine fetuin with both a recessed and a new electrode. The data in Table IV show that the recession of the working electrode does not impact the accuracy of sample analysis provided that standards are used.
Table IV| New electrode | New electrode | Old electrode | Old Electrode | |
| Preparation | Average Neu5Ac | Average Neu5Gc | Average Neu5Ac | Average Neu5Gc |
| 0.1 N HCl | 17 ± < 1 | 0.49 ± 0.01 | 18 ± < 1 | 0.48 ± 0.01 |
| Neuraminidase | 19 ± < 1 | 0.56 ± 0.02 | 19 ± < 1 | 0.53 ± 0.02 |
Discussion
We evaluated the separation of Neu5Ac, KDN, and Neu5Gc using the CarboPac PA10 column, a column that also separates neutral and amino sugar monosaccharides (Weitzhandler et al., 1996). The retention times of Neu5Ac and KDN were similar to those observed with the CarboPac PA1 column, while Neu5Gc eluted a bit earlier using the CarboPac PA10 column. The method we evaluated requires 27 min. If faster separations are needed, the flow rate can be increased to 1.5 ml/min and/or a steeper gradient can be used. We chose this separation because Neu5Ac was well separated from the void volume, reducing the possibility of interference from nonbound or early-eluting components in some samples. Isocratic separations using the CarboPac PA1 have been reported (Hermentin and Seidat, 1991; Seppala et al., 1991; Grollman et al., 1993; Jorge and Abdul-Wajid, 1995). Using a 0.1 M NaOH 0.15 M NaOAc eluent at a flow rate of 1 ml/min, Neu5Ac, KDN, and Neu5Gc are eluted at 3.7, 5.1, and 10.1 min from the CarboPac PA10 column. Using an isocratic method, peak width increases with retention time, and therefore; it may be difficult to accurately determine Neu5Gc in the same injection as Neu5Ac when Neu5Gc is present in a low concentration relative to Neu5Ac. Because of possible immunogenicity, quantification of small amounts of Neu5Gc may be important (Noguchi et al., 1995).
The linearity and minimum detection limits of this method were dependent on the working electrode recession. Using either a new or a deeply recessed (350 µm) working electrode, response is linear between 10 and 500 pmol. This is similar to the range (2-450 pmol) reported for o-phenylenediamine labeling (Anumula, 1995), but narrower than the range (2 pmol to 20 nmol) reported for the HPLC adaptation of the thiobarbituric acid assay (Powell and Hart, 1986). The sensitivity of electrochemical detection was best with a new working electrode (0.5-2 pmol), but not as high as the 57-196 fmol reported for 1,2-diamino-4,5-methylenedioxybenzene labeling (Hara et al., 1986).
The method presented here was reproducible over 48 h. Retention times of standards varied <0.5% and peak areas <5%, even when glycoprotein hydrolysate samples were analyzed between injections of standards. As reported by Anumula, we found that 2-4 h (4-8 injections) were occasionally required to achieve stable response (data not shown). Variation in peak area can be reduced by adjusting with the KDN internal standard. KDN is an effective internal standard for this method because it elutes between Neu5Ac and Neu5Gc, its response is linear in the linear ranges of Neu5Ac and Neu5Gc, and its response corrects variations in Neu5Ac and Neu5Gc responses. In a recent study using a CarboPac PA1 column, KDO was reported to be an effective internal standard for a HPAEC/PAD analysis of Neu5Ac and a sialylated glycoconjugate (Jorge and Abdul-Wajid, 1995). In that study, KDO was added prior to hydrolysis and therefore corrected for hydrolysis differences as well as matrix and electrode differences. In our separation, KDO coelutes with KDN and was not evaluated as an internal standard because we believe KDN is more structurally similar to Neu5Ac and Neu5Gc than KDO. We added KDN after hydrolysis to evaluate differences in response due to the state of the electrode and the sample matrix. KDN has recently been shown to be present on mammalian tissues and human lung carcinoma cells (Inoue et al., 1996). Therefore, when analyzing a new glycoprotein, at least one sample should be prepared without the addition of KDN. Electrochemical response is known to vary with temperature, and it is therefore possible that peak area variation could be further reduced with temperature control of the electrochemical cell. Because retention times did not decrease during these experiments, we do not believe a column wash is needed after every injection when analyzing sialic acids in the presence of small amounts (<5 µg) of protein.
We have shown that the method presented here can accurately determine the Neu5Ac and Neu5Gc contents of three glycoproteins. Accuracy was unaffected by working electrode condition despite the doubling of Neu5Ac and Neu5Gc peak area responses when switching from a recessed (350 µm) to a new working electrode. The electrochemical responses of KDN, Neu5Ac, and Neu5Gc were reduced in the 0.1 N HCl hydrolyses compared to neuraminidase digests, but the reduction in internal standard response can be used to correct for the reductions in Neu5Ac and Neu5Gc responses. Diminished electrochemical response was also found when 0.1 N TFA was substituted for 0.1 N HCl but was eliminated if 2 N acetic acid (80°C for 3 h) was used (data not shown). We did not evaluate the use of acetic acid to accurately release sialic acids from glycoproteins. Because repeated sample drying lowers the amount of response reduction, we believe that residual chloride and trifluoroacetate are at least in part responsible for the reduction in electrochemical response.
Materials and methods
Materials
Sodium acetate was obtained from Fluka (Ronkonkoma, NY). acetic acid and 50% NaOH were purchased from Fisher (Pittsburgh, PA). N-Acetylneuraminic acid (Neu5Ac) and N-glycolylneuraminic acid (Neu5Gc) were obtained from Pfanstiehl (Waukegan, IL).
3-Deoxy-d-glycero-d-galacto-2-nonulosonic acid (KDN) and [alpha]-2-keto-3-deoxyoctonate (KDO) were purchased from Toronto Chemicals (Toronto, Canada). Hydrochloric acid (6 N) was purchased from Perkin Elmer-ABI (Foster City, CA). Trifluoroacetic acid was obtained in 1 ml ampoules from Pierce (Rockford, IL). Bovine fetuin was purchased from GIBCO-BRL (Grand Island, NY). Bovine transferrin (apo) was obtained from Calbiochem (San Diego, CA). Human transferrin (partial Fe) and Arthrobacter ureafaciens neuraminidase were purchased from Boehringer Mannheim (Indianapolis, IN). Autosampler vials, caps, and septa were obtained from Sun Brokers (Wilmington, NC; part numbers 500-114, 200-288, 200-396).
Chromatography
A DX-500 chromatography system (Dionex, Sunnyvale, CA) consisting of a GP40 pump and an ED40 electrochemical detector and cell outfitted with a Au working electrode was used for all experiments. Injections were made by an AS3500 autosampler (Thermo Separation Products, San Jose, CA) equipped with a tefzel rotor seal, a stainless steel needle, and a 50 µl sample loop. The entire system was controlled by PeakNet chromatography software (Dionex). Neu5Ac, KDN, and Neu5Gc were separated on a CarboPac PA10 column (Dionex) and its guard. One hundred millimolar NaOH and 100 mM NaOH 1 M sodium acetate (0.2 µM-filtered) were installed as eluents A and B, respectively. The separation gradient was 7-30% B from 0 to 10 min, 30% B from 10 to 11 min, and 30-7% B from 11 to 12 min at a flow rate of 1 ml/min. The autosampler cycle time was 27 min. To detect sialic acids with the ED40 we used the 'carbohydrate waveform" (E1 = 0.05V, t1 = 400 ms; E2 = 0.75V, t2 = 200 ms; E3 = -0.15V, t3 = 400 ms) that is included in the PeakNet software.
Preparation of primary standards and samples
Neu5Ac and Neu5Gc were dried in a vacuum desiccater over NaOH pellets for 64 h and used to prepare solutions of Neu5Ac (9.68 mM) and Neu5Gc (2.52 mM). A 3.58 mM primary solution of KDN was prepared from a new bottle. Protein (51 µg bovine fetuin, 83 µg human transferrin, and 88 µg bovine transferrin ) was dissolved in 400 µl 0.1 M HCl, heated for 1 h at 80°C, dried in a SpeedVac (Savant), and dissolved in 450 µl deionized glass-distilled water and 50 µl 0.1 mM KDN prior to analysis. The same amounts of each protein were dissolved in 200 µl 0.1 M NaOAc pH 5 containing 1 mU of neuraminidase, mixed, and incubated for 18 h at 37°C. Prior to analysis, 250 µl water and 50 µl 0.1 mM KDN were added to each digest. Protein concentrations were estimated using the BCA Protein Assay (Pierce, Rockford, IL). We also used 0.1 N trifluoroacetic acid (80°C for 1 h) and 2 N acetic acid (80°C for 3 h) to release sialic acids.
Linearity, minimum detection limits, and reproducibility
From the primary standards we prepared and made dilutions of a 400 µM mixture of Neu5Ac, KDN, and Neu5Gc. To measure linearity and minimum detection limits we made five 25 µl injections of each solution (400, 200, 80, 40, 20, 8, 4, 2, 0.8, 0.4, 0.2, 0.08, 0.04, 0.02 µM). To ascertain reproducibility we prepared a mixture of Neu5Ac and Neu5Gc (0.011 mM each) and a 0.1 mM KDN solution. KDN solution (50 µl) was added to 450 µl of the Neu5Ac/Neu5Gc mixture, and the resulting mixture was used for reproducibility studies. We repeatedly analyzed 200 pmol (20 µl) of the standard for 48 h (107 injections). To determine the effect of sample analysis on the reproducibility of standard analysis, we alternated standard and glycoprotein hydrolysate sample injections for 48 h. The 0.1 mM KDN solution added to the standard mix was also added to the samples. We used a working electrode that was more than a year old and then repeated the experiments with a new working electrode.
valuation of the relationship of electrode wear to response
To determine the effect of electrode wear on sialic acid response, we evaluated eight working electrodes that had been used for different, but undetermined, amounts of time. We measured the depth of recession of the Au surface from the electrode block under a microscope using a series of wires of known widths. Each electrode was evaluated with six injections of a Neu5Ac and Neu5Gc mixture (250 pmol each). After this evaluation, six of the eight electrodes (chosen randoming) were reconditioned. To recondition, the electrode block was sanded with 600 grit silicon carbide sand paper until the block was even with the electrode surface. A microscope was used to evaluate the quality of reconditioning. Then each electrode was polished as described in the electrochemical detector manual and used to evaluate Neu5Ac and Neu5Gc linearity (10-2000 pmol).
Abbreviations
DMB, 1,2-diamino-4,5-methylenedioxybenzene; HPAEC, high pH anion-exchange chromatography; KDN, 3-deoxy-d-glycero-d-galacto-2-nonulosonic acid; KDO, 2-keto-3-deoxyoctulosonic acid; Neu5Ac, N-acetylneuraminic acid; Neu5Gc,N-glycolylneuraminic acid; PAD, pulsed amperometric detection.
References
1To 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: 18 Jan 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Suriano, S. K Ghosh, D. Chaudhuri, A. Mittelman, A. Banerjee, and R. K Tiwari Sialic acid content of tissue-specific gp96 and its potential role in modulating gp96-macrophage interactions Glycobiology, December 1, 2009; 19(12): 1427 - 1435. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Higuchi, P. Xin, M. S. Buckley, D. R. Erickson, and V.P. Bhavanandan Characterization of the rabbit homolog of human MUC1 glycoprotein isolated from bladder by affinity chromatography on immobilized jacalin Glycobiology, July 1, 2000; 10(7): 659 - 667. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||