Microscale Measurements of Michaelis-Menten Constants of Neuraminidase with Nanogel Capillary Electrophoresis for the Determination of the Sialic Acid Linkage.

Phospholipid nanogels enhance the stability and performance of the exoglycosidase enzyme neuraminidase and are used to create a fixed zone of enzyme within a capillary. With nanogels, there is no need to covalently immobilize the enzyme, as it is physically constrained. This enables rapid quantification of Michaelis-Menten constants (KM) for different substrates and ultimately provides a means to quantify the linkage (i.e., 2-3 versus 2-6) of sialic acids. The fixed zone of enzyme is inexpensive and easily positioned in the capillary to support electrophoresis mediated microanalysis using neuraminidase to analyze sialic acid linkages. To circumvent the limitations of diffusion during static incubation, the incubation period is reproducibly achieved by varying the number of forward and reverse passes the substrate makes through the stationary fixed zone using in-capillary electrophoretic mixing. A KM value of 3.3 ± 0.8 mM (Vmax, 2100 ± 200 μM/min) was obtained for 3'-sialyllactose labeled with 2-aminobenzoic acid using neuraminidase from Clostridium perfringens that cleaves sialic acid monomers with an α2-3,6,8,9 linkage, which is similar to values reported in the literature that required benchtop analyses. The enzyme cleaves the 2-3 linkage faster than the 2-6, and a KM of 2 ± 1 mM (Vmax, 400 ± 100 μM/min) was obtained for the 6'-sialyllactose substrate. An alternative neuraminidase selective for 2-3 sialic acid linkages generated a KM value of 3 ± 2 mM (Vmax, 900 ± 300 μM/min) for 3'-sialyllactose. With a knowledge of Vmax, the method was applied to a mixture of 2-3 and 2-6 sialyllactose as well as 2-3 and 2-6 sialylated triantennary glycan. Nanogel electrophoresis is an inexpensive, rapid, and simple alternative to current technologies used to distinguish the composition of 3' and 6' sialic acid linkages.

Sialic acids, the common name for N-acetylneuraminic acids, are the terminal monomer on the nonreducing end of glycans. Sialic acids and enzymes associated with their synthesis or catabolism are involved in a number of cellular processes[1] and are implicated in physiological dysfunctions including cancer,[2−5] antibody function,[6] and inflammation.[7] In proteins, asparagine-linked glycans contain a common core structure,[8] such that can be capped by sialic acids that are adjacent to a galactose residue. These sialic acid−galactose sequences are linked from carbon 2 on the sialic acid to carbon 3 or 6 on the adjacent monomer. The position of this linkage is relevant to cancer,[9] and as a result, the linkage chemistry is monitored. Although there is a critical need for routine sialic acid determinations, the challenges of linkage analysis, isomerization, and data interpretation make structural assignment difficult using current analytical technologies.

Oligosaccharides have been identified using benchtop sequencing with enzymes, but this requires a considerable amount of exoglycosidase and substantial incubation time.[10−14] The cost, stability, and sample preparation related to the use of the enzyme are limiting factors. Ultimately, separation strategies that reduce the sample handling and sample volume are critical. Microflow systems can efficiently screen chemical processes, are more stream-lined to develop and optimize reactions, and are used to predict the success of scaling up a method. There are a few barriers to rapidly assessing enzymatic processing on the microscale, which include the cost and lifetime of enzyme preparations. These barriers can be overcome with immobilized enzymes, which can have enhanced performance[15] and have increased stability. Methods of covalent enzyme immobilization require mild derivatization conditions and extensive optimization, and this has led to the development of new strategies to physically confine enzymes without covalent modification.[16,17]

To realize the full potential of enzymes in chemical assays, enzymes must be manipulated on the microscale without immobilization. In addition, the rate of enzyme catalysis must be established for different reaction conditions because the enzyme rate is specific for each substrate, and it is dependent upon the conditions used for the enzymatic reaction. Precise knowledge of enzyme activity can be quantified as Michaelis–Menten constants (KM), and it is important to harnessing enzymes for biomolecule recognition, sequencing, and assessment. With this information, enzyme-based microscale analyses have the potential to shed light on the relationship between the biomolecular structure of the substrate and enzyme function. However, new analytical tools are required that stabilize enzyme performance and consume small quantities of enzyme in faster analyses.

Capillary electrophoresis has been adapted to utilize enzymes to improve detection limits and separation specificity with electrophoretically mediated microanalysis, commonly referred to as EMMA.[18] This method relies upon differences in electrophoretic mobility of enzyme, substrate, and product for the analysis. EMMA is rapid, consumes nanoliter volumes of enzyme, and is reported extensively[19−22] as a method to enhance detection limits or provide indirect detection. While not as prevalent, the use of EMMA to determine KM is feasible if the enzyme remains in the native state, the enzyme does not adsorb to the capillary surface, and the electrophoretic differences of analyte, enzyme, and product facilitate separation. Significant optimization is required to ensure that the incubation conditions are compatible with the electrophoresis, and exquisite strategies to address this optimization have been described.[23−25] These barriers are overcome with the use of phospholipid nanogels, which self-assemble to form a thermally responsive material that is suitable as a replaceable gel to sieve DNA in capillary electrophoresis,[26,27] as a viscosity switch to close and open channels in microfluidics,[28,29] and as a viscous additive to improve capillary electrophoresis separations of oligosaccharides.[30−32] Nanogels are biocompatible materials that immobilize and pattern enzymes in microscale channels.

This paper outlines a new approach to EMMA that utilizes nanogels to physically constrain an enzyme in a separation capillary. Reconstitution of the neuraminidase enzyme in nanogel at low concentration results in a longer lifetime upon storage. This provides a means to cost-effectively use a single enzyme stock solution to deliver the subnanoliter volumes required for the capillary method. An electrophoretic mixing method is developed to circumvent diffusion limitation of static incubation. For the first time, nanogels are harnessed to precisely quantify the Michaelis–Menten constants of an enzyme with different degrees of stereospecificity in the presence of substrates with different linkage positions. The trisaccharide sialyllactose, which is O-(N-acetyl-α-neuraminosyl)-[2 → 3(or 2 → 6)]-O-β-d-galactopyranosyl-(1 → 4)-D-glucose, is used as a model substrate to characterize neuraminidase. The different catalysis rates obtained for sialic acid residues with 2–3 or 2–6 linkages are used to analyze the linkage position of a mixture of 2–3 and 2–6 sialyllactose and in the sialylated triantennary glycan substrate mixture that contained both Galβ(1–3)GlcNAc and Galβ(1–4)GlcNAc. Similar results are achieved when a specific and nonspecific enzyme are used. The approach extends the nanogel separation method to general as well as specific enzymes.

Materials and Methods

Preparation and Derivatization of Standards

Enzyme studies with an anticipated KM in the mM range require that oligosaccharide substrate concentrations in the mM range be used to obtain a Michaelis–Menten curve. For the KM studies, the substrate was labeled with a chromophore detected by UV–visible absorbance. The chromophore 2-aminobenzoic acid was used as a reagent to label the oligosaccharides as reported previously.[33] The reaction was performed with excess label to ensure 100% labeling efficiency to prevent any bias in the KM measurement. If the labeling reaction was not complete, then the concentration of the labeled substrate that was UV-absorbing would have been less than the total concentration of substrate (i.e., the combination of labeled and unlabeled substrate) and would confound the measurement of KM. A ratio of 1400 nmol 2-aminobenzoic acid:200 nmol sialyllactose (7:1) was achieved by reacting 1 μL volume of 0.2 M oligosaccharide dissolved in water with 1 μL volume of 1.39 M 2-aminobenzoic acid and 23 μL volume of 1 M sodium cyanoborohydride at 65 °C for 2 h in 0.5 M acetic acid (glacial acetic acid diluted with methanol). Once the reaction was complete, the reaction mixture was evaporated to dryness on a Savant SpeedVac concentrator (Thermo Scientific, Waltham, MA). Excess 2-aminobenzoic acid was removed from the labeling reaction using a Discovery DPA-6S solid phase extraction cartridge (50 mg packing material, Supelco, Bellefonte, PA) with slight modifications to a previously established literature procedure.[31] Specifically, once the sample was loaded in the extraction cartridge, the 2-aminobenzoic acid was eluted using 10 mL of 95% acetonitrile, 5% aqueous 1 mM triethylamine, and the retained sugars were eluted from the cartridge using 3 mL of aqueous 25 mM triethylamine. To ensure that all of the sialyllactose was labeled and recovered from the purification process, it was compared against a second reaction performed with excess sialyllactose as described in the Supporting Information.

Studies designed to distinguish the linkage position of sialyllactose or glycan were accomplished with substrate concentrations in the nM range. For these linkage position studies the substrate was labeled with a chromophore detected by fluorescence. The fluorescently conjugated oligosaccharides and glycans were labeled as previously described[34] with slight modifications. Glycan labeling was accomplished using 100 mM 8-aminopyrene-1,3,6-trisulfonic acid in 20% acetic acid for a reaction of 7 nmol glycan:250 nmol dye in a total reaction volume of 5 μL. The labeled glycan sample was purified using a 1 kDa molecular weight cutoff filter (#MCP001C41, Pall Corporation, Ann Arbor, MI). Sialyllactose labeling was accomplished using 100 mM 8-aminopyrene-1,3,6-trisulfonic acid in 20% acetic acid for a reaction of 5 nmol oligosaccharide:250 nmol dye in a total reaction volume of 5 μL. The labeled sialyllactose samples were purified using the DPA-6S column as described for labeling with 2-aminobenzoic acid. Once purified, the samples were dried in a SpeedVac concentrator before reconstituting to 100 μL in water and storing at −20 °C.

Neuraminidase in powder was reconstituted to a concentration of 33.6 mUnits/μL in 50 mM potassium phosphate with the pH adjusted to 5.2 using 1 M sodium hydroxide. The appropriate volume of master stock (i.e., less than or equal to 0.52 μL) was diluted with 10% nanogel to a final volume of 50 μL. For both neuraminidase enzymes, 1 Unit was defined as the amount of enzyme required to produce 1 μmol of methylumbelliferone in 1 min at 37 °C, pH 5.0 from 2′-(4-methyl-umbelliferyl)-alpha-d-N acetylneuraminic acid.[35,36]

Capillary Electrophoresis

Analyses were performed using a P/ACE MDQ or MDQ Plus (Sciex, Redwood City, CA) configured by the manufacturer with laser-induced fluorescence detection (3 mW air cooled argon ion laser or 3 mW solid state laser, with λex = 488 nm, λem = 520 nm) and a photodiode array or UV–visible absorbance detector (monitored at 214 nm). A 25 μm internal diameter, 360 μm outer diameter fused silica capillary (Polymicro Technologies, Phoenix, AZ) was used for separation. Each day capillaries were prepared as previously reported.[32] Unless otherwise noted, the background electrolyte was 100 mM 3-(N-morpholino)propanesulfonic acid buffered to pH 7. Phospholipids were prepared as described previously, aliquoted, and stored at −20 °C.[31,32,37] The phospholipid preparation, which was q = 2.5 (i.e., [DMPC]:[DHPC] = 2.5) and 10% phospholipid by weight, had low viscosity below the gel phase transition temperature and was easily introduced in the capillary at a temperature of 19 °C or lower. Prior to each run, the capillary was held at 19 °C and prepared as previously reported[31] with slight modification as described in the Supporting Information. Data collection and analysis were performed using 32 Karat Software version 7.0 (MDQ) or 10.2 (MDQ Plus). The Michaelis–Menten curves were fit using GraphPad Prism version 4.03 (San Diego, CA).

Results and Discussion

Patterning Enzyme in Capillary

The nanogel fixed enzyme zone was achieved by creating a pseudostationary enzyme plug in phospholipid nanogel, as depicted in Figure . Enzyme prepared in the nanogel was introduced into the capillary at 19 °C, which was a temperature that maintained fluid-like viscosity of the material.[28,29] At this temperature, the enzyme was positioned in the thermally controlled region of the capillary to ensure that the desired temperature was maintained during analyses. The temperature was then increased to 37 °C to form a viscous gel[28] that maintained the enzyme position within the capillary and supported the enzyme reaction. Two model substrates, sialyllactose with different linkages shown in Figure , were used. Once the fixed enzyme zone was patterned in the capillary, as seen in Figure , the substrate was electrokinetically driven into the enzyme zone. Following incubation, substrate was then separated from product (Figure ). The catalysis rate of neuraminidase was quantified by monitoring the conversion of substrate to the lactose product (Figure ). The separation of sialyllactose and lactose did not require the use of phospholipid nanogel to resolve these analyte peaks. Therefore, the fixed enzyme zone was embedded in an aqueous solution of 100 mM background electrolyte buffered to pH 7. The position of the enzyme zone within the capillary was established by migrating the substrate to a particular position, incubating substrate with enzyme, and then separating and quantifying the peak areas of the substrate and enzymatically generated product. The substrate positions that generated larger product peak area coincided with the position of the enzyme zone, as summarized in Figure S-1 in the Supporting Information. With the position of the enzyme zone established, the impact of nanogel on the enzyme catalysis rate was determined.

Figure 1

Depiction of electrophoretic migration of substrate in-capillary containing enzyme in a fixed zone. The 3′- or 6′-sialyllactose was incubated in enzyme and converted to lactose. The nonreducing end of the oligosaccharide was labeled with a chromophore (e.g., 2-aminobenzoic acid, for UV-absorbance detection).

Effect of Phospholipid Nanogel on Enzyme Performance

Proteins are reconstituted in a variety of additives to maintain structure and function.[38−41] These additives influence both stability and activity in a complex manner.[42] Phospholipids interact with proteins through different mechanisms.[43] Physical constraint of enzymes with lipids has been reported using edge stabilized phospholipid nanodiscs to study enzyme catalysis of membrane protein.[44,45] Although phospholipids are used to eliminate nonspecific adsorption of proteins to surfaces,[46−48] they are underexplored as an additive for soluble proteins.

To evaluate the effect of phospholipid nanogel on neuraminidase, the enzyme performance was monitored either in a traditional aqueous solution of 50 mM potassium phosphate adjusted to a pH of 5.2 with sodium hydroxide or in nanogel comprised of phospholipid dissolved in the same aqueous solution. The preparations of neuraminidase were diluted to a concentration of 350, 250, or 150 μUnits/μL, and the enzyme activity was evaluated by quantifying the rate of conversion of sialyllactose substrate to lactose. The rates summarized in Figure demonstrated that in the presence of nanogel the enzyme retained high activity for the 30-day period regardless of concentration. The stability of the enzyme diluted in the aqueous buffer was concentration-dependent. The 350 μUnits/μL did not change (Figure A). However, the 250 μUnits/μL (Figure B) and 150 μUnits/μL (Figure C) enzyme solutions decreased in activity until there was no detectable activity on days 3 and 1, respectively. The reaction velocity of 350, 250, and 150 μUnits/μL enzyme was 1100 ± 50, 830 ± 50, and 500 ± 40, respectively. At the highest concentration of 350 μUnits/μL, the reaction velocity in nanogel was 1.5 times higher than the reaction velocity in aqueous electrolyte (750 ± 60 μM lactose/min). For the lower enzyme concentrations, the reaction velocity in nanogel as compared to that obtained in aqueous electrolyte on day 1 was 1.7 (500 ± 10 μM lactose/min) and 2.5 (200 ± 10 μM lactose/min) times higher. These findings were in agreement with literature evidence that lipid monolayers improve the performance of exoglycosidase enzymes, including galactosidase[49−51] and neuraminidase.[52] The improvement in stability observed at the lower enzyme concentration may be due to molecular crowding. The compaction of the enzyme by molecules in the solution maintained enzyme in a folded state and impacted the enzyme rate if the effective concentration of active protein was increased.[53,54]

Figure 2

Plot of enzyme activity. Substrate was incubated in α2-3,6,8,9 neuraminidase suspended in phospholipid nanogel (10% lipid with [DMPC]/[DHPC] = 2.5 in 50 mM potassium phosphate pH adjusted to 5.2 with sodium hydroxide) or in the same aqueous solution devoid of phospholipid at an enzyme concentration of 350 μUnits/μL (2A), 250 μUnits/μL (2B), and 150 μUnits/μL (2C). Substrate (5.4 mM 3′-sialyllactose) was incubated in enzyme for 2 min at 37 °C and separated. Separations were performed at 37 °C in a 25 μm i.d. capillary, with an effective length of 30 cm and E = 500 V/cm (reverse polarity).

Optimizing Incubation Times in the Fixed Enzyme Zone

When enzyme catalysis was performed in-capillary using static incubations, the measured velocity (Table ) decreased with increasing incubation time. This diffusion-limited interaction between substrate and enzyme hindered the quantitative determination of enzyme performance (i.e., KM values) because enzyme rates must be consistent regardless of incubation time. In nanogel electrophoresis, the limitations of diffusion-based transport of substrate and product were overcome by electrophoretic mixing. In this electrophoretic mixing approach, the sustained contact between substrate and enzyme was achieved by alternating the polarity of the applied voltage to reversibly drive the substrate through the enzyme zone as depicted in Figure . The number of passes and length of each pass were selected to achieve the desired incubation time. The process of mixing in the fixed zone was feasible when the position of the enzyme was profiled in the capillary to determine the zone boundaries as summarized in Figure S-1 in the Supporting Information. The rate of enzyme catalysis increased with the electrophoretic velocity due to the increased probability of enzyme–substrate collision. An Ohm’s law plot was performed (Figure S-2 in the Supporting Information) to confirm that this was not due to Joule heating. As summarized in Table , electrophoretic mixing generated catalysis with a precision of 9% relative standard deviation for four different electrophoretic sweep times as compared to 30% relative standard deviation for four different static incubation times.

Table 1

Effect of Substrate Delivery on Ratea

Time, s	static,b μM/s	mixed,c μM/s
40	13.9 ± 0.6	9.9 ± 0.9
60	11.6 ± 0.3	9.5 ± 0.5
100	8.4 ± 0.3	8.5 ± 0.4
200	7.3 ± 0.1	8.4 ± 0.5
Aved	10 ± 3 (30%)	9.1 ± 0.7 (9%)

Data are averages (n = 3) using 5.4 mM 3′-sialyllactose labeled with 2-aminobenzoic acid and 336 μUnits/μL α2-3,6,8,9 neuraminidase in nanogel at 37 °C.

Incubation performed with E = 0 V/cm after driving the substrate to the center of the fixed enzyme zone.

Performed by electrophoresing substrate through the enzyme with E = 250 V/cm for multiple forward (F) and reverse (R) passes of: 40 s (20F-10R-10F), 60 s (20F-20R-20R), 100 s ([20F-20R]2-20F) ,or 200 s ([20F-20R]4-20F-10R-10F).

Data are averages of rates at four incubation times.

Figure 3

Conceptual diagrams of multipass electrophoretic mixing for control of the incubation time in the enzyme. The mixing duration was determined by the total time the substrate passes through the zone. Mixing may require that passes of different times are combined as in the case of a 40 s incubation. The positions of the arrows in the zone are conceptual and do not imply that the substrate traversed from the top of the zone to the bottom with each successive pass. The electropherograms were obtained using substrate of 5.4 mM 3′-sialyllactose and 336 μUnits/μL of α2-3,6,8,9 neuraminidase that was suspended in phospholipid nanogel. Experimental conditions are as described in Figure .

Determination of Michaelis–Menten Constants for Neuraminidase

The in-capillary enzymatic method was well-suited to characterize the relationship between the substrate concentration and the performance of the rate of enzyme production. The velocity of product formation was calculated as product concentration/incubation time and then plotted to determine the Michaelis–Menten constant. An accurate measurement of enzyme performance required that the classical rules derived for KM determinations were followed.[55] In particular, no more than 10% of the substrate could be converted to product because the presence of product inhibits the rate of reaction. An additional consideration was that the amount of product formed must fall within the linear range of quantification for the method of detection. A KM curve fit using nonlinear regression required 2 points at or near saturation and an additional 3 points distributed across the region where the velocity changes significantly with substrate concentration.

The application of nanogel for in-capillary determination of enzyme activity was demonstrated with a neuraminidase that cleaves sialic acid with an α 2-3, 2-6, 2-8, or 2-9 linkage. Each determination utilized nanogel enzymolysis with mixing at an enzyme concentration of 336 μUnits/μL prepared in nanogel. Incubations were performed at 37 °C. Electrophoretic mixing was performed during the incubation period to obviate the time-related dependence of velocity on incubation time observed in static incubations. A set of five electropherograms was obtained with 3′-sialyllactose at concentrations of 0.40, 1.25, 3.4, 5.4, and 7.4 mM (Figure A). The peak area for the lactose product obtained in each separation was quantified (Figure A inset). The rate of enzymatic conversion was calculated as the concentration of lactose produced over the incubation time. These determinations were performed in triplicate; 15 values were plotted as shown in Figure B and fit using nonlinear regression to obtain the KM of 3.3 mM with a standard deviation of 0.8 mM.

Figure 4

(A) Electropherograms of 3′-sialyllactose substrate and the lactose product generated after enzymatic reaction. The traces were offset in time by 0.1 min for visualization from the lower to upper traces at the following 3′-sialyllactose concentrations: 0.40 mM, 1.25 mM, 3.4 mM, 5.4 mM, and 7.4 mM. The increase in the area of lactose with increasing 3′-sialyllactose concentration is depicted in the inset. The Michaelis–Menten curve (part B) was generated by plotting the substrate concentration versus the rate of product formation for 3′-sialyllactose with 336 μUnits/μL α2-3,6,8,9 neuraminidase in nanogel. Experimental conditions are as described in Figure .

Literature values have been reported for neuraminidase from Clostridium perfringens with related sialyllactose substrate. A KM value of 2.4 with unlabeled 3′-sialyllactose was obtained using potassium acetate buffered at pH 4.5, 37 °C, quantifying the product concentration with thiobarbituric-facilitated colorimetric detection with the alkali-Ehrlich method.[56] A KM value of 2.2 ± 0.3 mM with unlabeled 3′-sialyl-N-acetyllactosamine was obtained using 50 mM potassium phosphate at a pH adjusted to 5.16 with sodium hydroxide, 37 °C, quantifying the product with anion exchange chromatography coupled to electrochemical detection.[57] KM values are difficult to compare given that slight differences exist in the hydrolysis reaction conditions. In addition, the substrate used in this work was labeled with 2-aminobenzoic acid. In light of these differences in reactions, the experimental KM value was similar, as it was approximately 1.5 times higher than these literature values.

The applicability of nanogel for enzyme characterization was extended to study changes in the rate of catalysis for different substrates and different neuraminidase enzymes. The activity and specificity of α2-3,6,8,9 neuraminidase on 6′-sialyllactose, a substrate with different linkage positions, were examined using the nanogel capillary electrophoresis. The results demonstrated the utility of the method to screen enzymes of different specificity for the conversion of different substrate molecules. The values in Table S-1 in the Supporting Information were obtained under identical pH, ionic strength, substrate concentration, and enzyme concentration. The KM value from the Michaelis–Menten curve for the general neuraminidase acting on 6′-sialyllactose was quantified as 2 ± 1 mM (Figure S-3 in the Supporting Information). A similar KM value of 1.2 mM was reported for general neuraminidase and 6′-sialyllactose with slightly different reaction conditions (i.e., 100 mM sodium/potassium phosphate pH adjusted to 5.4, 37 °C).[58] The KM value was 3 ± 2 mM for the specific neuraminidase cleaving sialic acid from 3-sialyllactose, as shown in Figure S-4 in the Supporting Information. Incubation of the α2-3 neuraminidase with 6′-sialyllactose resulted in no production of lactose, confirming that the specific enzyme was effective at cleaving only 3′ sialic acid residues.

The KM value was independent of enzyme concentration, while Vmax was dependent upon neuraminidase concentration. Curves were obtained at the same enzyme concentration of 336 μUnits/μL. For this enzyme, the manufacturer reported that the rate was greater for 2-3 than 2-6.[59] The Vmax values (Table S-1 in the Supporting Information), which describe the rate of enzyme catalysis when fully saturated by substrate, indicated that the catalytic rate of the general enzyme (i.e., α2-3,6,8,9 neuraminidase) was five times faster for 2-3 substrate when compared to the 2-6 substrate. Furthermore, for conversion of 2-3 sialic acid substrate, the general enzyme was twice as fast as the 3′ specific enzyme. These findings were utilized to apply neuraminidase to rapidly distinguish the sialic acid linkage.

Differentiating the Sialic Acid Linkage with Neuraminidase

With knowledge of the enzyme catalysis, the application of a fixed zone of neuraminidase to distinguish sialyllactose linkage position was demonstrated. A nanogel separation (Figure trace A) of the 6′ and the 3′-sialyllactose in the absence of enzyme revealed that the 2-6 and the 2-3 sialyllactose linkages were separated. When a mixture of 2-3 and 2-6 sialyllactose was subjected to a fixed enzyme zone of 8 mUnits/μL neuraminidase specific for 2-3 sialic acid linkage, only the 2-3 sialyllactose was converted to lactose (Figure trace B). Triplicate runs with and without the fixed enzyme zone generated peaks that had a relative standard deviation in area of 4% and in time of 1%. These studies were performed using fluorescence detection. Data found in Figure S-5 in the Supporting Information expanded on these analyses by incorporating a fixed enzyme zone of 0.6 mUnits/μL α2-3 neuraminidase to confirm 3′-sialyllactose was converted to lactose and not just shifted in time.

Figure 5

Electropherograms of 3′ and 6′-sialyllactose to demonstrate the use of α2-3 neuraminidase to determine substrate linkages. The separation of 3′- and 6′-sialyllactose in trace A was obtained in the absence of enzyme. Trace B incorporated a fixed zone of α2-3 neuraminidase (loaded at 69 kPa for 7 s) suspended in phospholipid nanogel (10% lipid with [DMPC]/[DHPC] = 2.5 in 50 mM potassium phosphate pH adjusted to 5.2 with sodium hydroxide) at a concentration of 8 mUnits/μL to distinguish 2-3 from 2-6 sialyllactose by cleaving all the sialic acid on the 3′-sialyllactose. Separations and incubations were performed at 37 °C in a 25 μm i.d. capillary filled with nanogel, with an effective length of 50 cm and E = 400 V/cm (reverse polarity).

A similar strategy to identify the sialic acid linkage was applied to the mixture of trisialylated triantennary complex N-glycan shown in Figure . The N-glycan determinations were accomplished using the fixed nanogel enzyme zone embedded in capillary filled with an enzyme-free nanogel. A nanogel separation (Figure trace A) of the N-glycan in the absence of enzyme revealed that the N-glycan was fully sialylated and contained both Galβ(1–3)GlcNAc and Galβ(1–4)GlcNAc. These isomers were resolved electrophoretically when the capillary was filled with nanogel.[30] The separation in trace B was obtained with a fixed zone of 80 mUnits/μL α2-3 neuraminidase, resulting in complete cleavage of the 2-3 linked sialic acid. A higher concentration and larger fixed enzyme zone increased the relative standard deviation in area to 11% and in time to 2% (n = 3), but was required to desialylate the trisialylated N-glycan. No product peaks were generated that contain two sialic acids. A product peak associated with asialo N-glycan (Figure trace A) comprised <1% of the area of the N-glycan peaks. The peaks obtained in these traces indicated that two-thirds of the sialic acid linkages were 2-3, while one-third were 2-6.

Figure 6

Electropherograms of 0.15 nM trisialylated triantennary complex N-glycan incubated in α2-3 neuraminidase to determine linkage position. Trace A was obtained without enzyme, while trace B was obtained using 80 mUnits/μL α2-3 neuraminidase suspended in phospholipid nanogel. The fixed zone of α2-3 neuraminidase (loaded at 69 kPa for 35 s) generated an electropherogram of N-glycan devoid of all 2-3 linked sialic acid. The peak marked with the asterisk was present in the reaction blank. Experimental conditions were the same as those used in Figure .

To demonstrate that different catalytic rates of general neuraminidase for 2-3 vs 2-6 sialic acid linkages could be harnessed to distinguish linkage position, the experiment was repeated using general neuraminidase. The separation in trace A of Figure was obtained without enzyme, while the separation in trace B was obtained with a fixed enzyme zone of 4 μUnits/μL general neuraminidase. This resulted in complete cleavage of the 2-3 linked sialic acid. Triplicate runs with the fixed zone of general enzyme generated peaks that had a relative standard deviation in area of 10% and in time of 2%. No product peaks were generated that contained two sialic acids. The separation in trace C of Figure was obtained with a fixed zone of 2.4 mUnits/μL general neuraminidase, resulting in complete cleavage of sialic acid. The same sialic acid linkage composition obtained with specific enzyme was obtained using general enzyme by controlling the concentration of the general neuraminidase.

Figure 7

Electropherograms of 0.15 nM trisialylated triantennary complex N-glycan incubated in 2-3′,6′,8′,9′ neuraminidase to determine the 3′ versus 6′ sialic acid composition. The separation in trace A was obtained in the absence of enzyme. The separation in trace B was obtained using 4 μUnits/μL α2-3,6,8,9 neuraminidase suspended in phospholipid nanogel (loaded at 69 kPa for 21 s) and generated an electropherogram of N-glycan devoid of all 2-3 linked sialic acid. The fixed zone of 2.4 mUnits/μL α2-3,6,8,9 neuraminidase used in trace C (loaded at 69 kPa for 21 s) generated an electropherogram of N-glycan devoid of all sialic acid. The experimental conditions are the same as those used in Figure .

Conclusions and Future Directions

Nanogels are a biocompatible separation additive. At a concentration of 150 μUnits/μL, enzyme reconstituted in aqueous electrolyte has a rate that is approximately half of what is obtained when it is reconstituted in nanogel made in the same aqueous solution. Furthermore, when the preparation does not contain nanogel, the activity decreases dramatically at an enzyme concentration of 150 μUnits/μL such that product was not detectable on the second day of measurement. In contrast, neuraminidase reconstituted in nanogel maintained a rate of 500 ± 40 μM lactose/min, as seen in the generation of product measured throughout a 30-day period. A single enzyme stock in nanogel is a cost-effective means to deliver the subnanoliter volumes required for the capillary method.

Nanogel preparations are inexpensive, costing $0.09 for 5 μL.[60,61] The nanogel enzyme analyses provide an inexpensive, rapid, and simple means to analyze and quantify the linkage composition of oligosaccharides and are an alternative technology to current methods that rely on derivatization or benchtop digestion. Although the specific enzyme provides greater confidence in the glycan linkage composition, the same information is achievable with the general enzyme that has preferential catalytic specificity for one substrate linkage over another, such as the α2-3,6,8,9 neuraminidase. This is particularly useful in cases where a specific enzyme is not readily available (e.g., α2-8 or α2-9 specific sialidases) or is prohibitively expensive. The method will be harnessed to determine the linkage composition in mixtures of complex N-glycans. Future efforts will also center on adapting the method for different enzymes.