ACA-23-3971Rev.Highlighted: Fluorescent Parallel Electrophoresis Assay of Enzyme Inhibition

Abstract

Background.

Enzyme inhibitors comprise the largest class of pharmaceutical compounds. The discovery and development of new enzyme inhibitor drug candidates depends on sensitive tools to quantify inhibition constants, K_i, for the most promising candidates. A high throughput, automated, and miniaturized approach to measure inhibition is reported. In this technique enzyme inhibition occurs within a 16 nL nanogel reaction zone that is integrated into a capillary. The reaction and electrophoresis separation are completed in under 10 minutes. The nanoliter enzyme reaction zones are easily positioned inside a standard separation capillary by pseudo-immobilizing enzymes within a thermally reversible nanogel.

Results.

This report optimizes and validates a capillary nanogel electrophoresis reaction and separation with a multi-capillary array instrument. Inhibitor constants are determined for the neuraminidase enzyme to quantify the effect of the transition state analog, 2,3-dehydro-2-deoxy-N-acetylneuraminic acid (DANA), as well as the inhibitor Siastatin B. With the multi-capillary array assay replicate K_i values are determined to be 5.6 ± 0.3 μM (n = 3) and 9.2 ± 0.3 μM (n = 3) for DANA and Siastatin B, respectively. The enzyme reaction in each separation capillary converts the substrate to a product in real time. The nanogel is used under suppressed electroosmotic flow, sustains enzyme function, and is easily filled and replaced by changing the capillary temperature. Using laser-induced fluorescence allows the determination to be achieved with substrate concentrations well below the Michaelis-Menten constant, making the method independent of the substrate concentration and therefore a more easily implemented assay.

Significance.

A lower measurement cost is realized when the reaction volume is miniaturized because the amounts of enzyme, substrate and inhibitor are reduced. Fast enzyme reactions are possible because of the small reaction volume. With a multi-capillary array, the inhibition assay is achieved in a fraction of the time required for traditional methods. The separation-based assay can even be applied to labeled substrates not cleaned up following the labeling reaction.

Graphical Abstract

1. INTRODUCTION

Enzyme inhibitors comprise a major class of pharmaceutical compounds [1], making inhibiting enzymes a key strategy for pharmaceutical research [2]. Neuraminidases, also known as sialidases, are implicated in numerous diseases, and inhibitors of these enzymes that cleave sialic acids from oligosaccharides are prominent therapeutic targets [3]. Thus far, inhibitors have been commercialized for influenza A, but research and discovery in new compounds to address a broader range of neuraminidases is ongoing. Indeed, inhibitor screening and quantification of the effect of an inhibitor on enzyme activity is critical to therapeutic discovery and drug development [4]. Before inhibitor candidates can be evaluated with in vivo models, extensive testing in vitro is required to screen and validate inhibitor performance. This drives the demand for rapid, high-throughput assays [2].

Existing high-throughput technologies are based on 96-well plate assays and chromogenic or fluorogenic dyes [4]. While these methods are high throughput, they require multistep processing and consume large amounts of enzymes and other reagents. Additionally, more sensitive enzymatic assays for quantification are still required for the validation and confirmation of the mechanism for inhibition which involves the determination of the inhibitor constant (K_i) [5]. The sensitivity of fluorometric or colorimetric reactions carried out and analyzed in a well plate suffer from background interferences from the substrate or inhibitors [6]. Capillary electrophoresis-based enzyme activity assays overcome these drawbacks because the substrate and reaction products are isolated prior to detection [7–9]. Moreover, this separation step enhances the assay sensitivity by separating fluorogenic reagents and contaminants from substrates, which mitigates interferences from auto-fluorescence and quenching.

Miniaturization of chemical assays reduces the sample and reagent consumption, making it an appealing strategy for methods used to screen for enzyme inhibition or to quantitatively determine the K_i [5]. This is especially relevant when the total number of assays increases, for example in high throughput parallel assays. Miniaturization is also beneficial in separation-based assays where an immobilized enzyme is integrated into a channel or capillary to facilitate reaction monitoring by eliminating manual processing following the reaction. Although reactions with immobilized enzymes typically require a chemical conjugation [7–9], enzymes have been non-covalently immobilized in phospholipid nanogel positioned within an electrophoresis capillary [10–14]. These nanogel preparations used to reconstitute the enzyme and as the separation medium are considered biocompatible because when the enzyme is prepared in the nanogel medium the enzyme activity is improved and the enzyme remains active for several weeks, which is in contrast to enzyme preparations in aqueous solutions [12]. The nanogel forms through the self-assembly of a phospholipid bilayer, resulting in structures that form nanodisks or ribbons. The preparation has a thermally reversible viscosity, forming a gel above 23°C and a liquid below 23°C [15–17]. This fluid is easily introduced and patterned in a capillary at cold temperatures, while at warmer temperatures it is locked in place as a gel. When the analysis is completed, it is expelled from the capillary at low temperatures. The nanogel is critical because the eight capillaries within a commercial parallel separation cartridge cannot be replaced by the user. With nanogel, the immobilized enzyme can be repeatedly integrated into all of the separation capillaries through a single automated process of parallel patterning prior to the separation. The net cost of this approach is low because the method uses such small enzyme volumes. This is critical for costly enzymes. Moreover, the nanogel is consumed at a cost of pennies per run. Fundamental to this approach the gel stabilizes the enzyme, maintaining function for more than 30 days as compared to 1 day when formulated in aqueous solution [12]. Lastly, the nanogel is combined with a semi-permanent phospholipid coating that suppresses the electroosmotic flow enabling separations under reversed polarity [18–20].

In this report, a sialyllactose substrate tagged with a fluorescent dye is cleaved by neuraminidase and the substrate and remaining product are detected at nanomolar levels with laser-induced fluorescence. The enzyme reaction is performed in the presence of an inhibitor and the amount of product formed depends upon the inhibitor concentration. A plot of the enzyme activity and substrate concentration is used to determine the half maximal inhibitory concentration (IC₅₀), which is, in turn, used to calculate the K_i according to Equation 1, where K_m is the Michaelis-Menten constant [14].

$$K_i={\text{IC}}_{50}/\left(1+\left[\text{S}\right]/K_m\right)$$ (Equation 1)

As this substrate concentration is significantly lower than the Michaelis-Menten constant of 3.3 mM, the K_i determination becomes independent of the substrate identity and concentration. This makes the method both practical and appealing because the exact concentration of labeled substrate is no longer needed for the K_i calculation. As a result, simple and fast sample preparation methods can be used. Notably, reducing the substrate to nanomolar concentrations requires less enzyme to maintain the initial reaction rate assumptions inherent to enzyme kinetic analyses [21].

Thermally reversible nanogels have been used previously to immobilize enzymes within a narrow bore separation capillaries (i.e. 25-micrometer inner diameter) [10–14]. Still, they have not been successfully adapted to the larger 50-micrometer inner diameter capillaries preferred for pharmaceutical analyses that are commercially available for the 8-capillary array instrument. This report presents an optimized patterning approach that consumes 16 nL of 3.4 μU/mL enzyme (9 g/U), corresponding to less than 500 picograms per run. A parallel separation in an 8-capillary array is achieved for two known neuraminidase inhibitors in less than 10 minutes. The total analysis time was reduced more than 8-fold because a blank run (no inhibitor) was completed simultaneously with seven various inhibitor concentrations to obtain a K_i curve. This is in contrast to a single capillary system where blank runs must be included repeatedly throughout the duration of the analysis to use as a reference value to normalize the percent enzyme activity remaining following inhibition. Remarkably, each inhibition curve consumed only 4 nanograms of enzyme, with triplicate measurements requiring only 12 nanograms of enzyme. This report outlines a universal approach for automated and high throughput inhibition measurements with a simple nanogel electrophoresis additive enabling reversed polarity separations and thermally immobilized enzyme in-line reaction zones. The method is designed to be readily adapted to a variety of inhibitors and other enzyme systems.

2. MATERIALS AND METHODS

2.1. Reagents.

3’-sialyllactose sodium salt (OS04397) and 6’-sialyllactose sodium salt (OS04398) were purchased from Carbosynth (Biosynth Carbosynth Ltd, San Diego, CA, USA). 3’-N-Morpholino propane sulfonic acid (MOPS, M0707) was purchased from TCI American (Portland, OR, USA). Acetonitrile (34851), 8-aminopyrene-1,3,6-trisulfonic acid trisodium salt (A7222), 2,3-dehydro-2-deoxy-N-acetylneuraminic acid (DANA, D9050), methanol (646377), Neuraminidase from Clostridium perfringens (C. welchii) powder (N2876), Siastatin B (S8063), sodium acetate, anhydrous (7510-OP) sodium cyanoborohydride in tetrahydrofuran (156159), and triethylamine (471283) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Sodium hydroxide pellets were from Thermo Scientific Chemicals (A16037.36, Thermo Fisher Scientific, Waltham, MA, USA). The phospholipids, 1,2-dihexanoyl-sn-glycero-3-phosphocholine (DHPC, 850305P) and 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC, 850345P) are purchased from Avanti Polar Lipids (Alabaster, AL, USA). All aqueous solutions are made with purified deionized water (18 MΩ cm) from an Elga Purelab Prima reverse osmosis system coupled to an Elga-Veolia Chorus I water purifier (Woodbridge, IL, USA).

2.2. Derivatization of oligosaccharides.

The reducing end of 6’-sialyllactose and of 3’-sialyllactose was derivatized with the APTS fluorophore to enable optical detection. The conjugation reaction was carried out as previously described [12]. Briefly, a 5X molar ratio of ATPS to sugar is combined in 20% aqueous acetic acid. Sodium cyanoborohydride in tetrahydrofuran is added to catalyze the conjugation reaction to a final reaction concentration of 0.5 M. Unless otherwise noted, reactions are incubated at 65 °C.

2.3. Removal of excess APTS reagent.

After the reaction, the solvent is evaporated using a Savant SpeedVac concentrator (Thermo Scientific, Waltham, MA). When desired, excess APTS reagent is removed via solid phase extraction using a Discovery DPA-6S solid-phase extraction cartridge (50 mg packing material, Supelco, Bellefonte, PA). The dried sample is reconstituted in 100 μL of deionized water and 10 μL is added to acetonitrile such that the final solution is 95% (v/v) acetonitrile:water is loaded onto the cartridge. The sample is then washed 10X with 1 mL of 95 % (v/v) acetonitrile:water containing 25 mM triethylamine to elute the excess APTS. Finally, the retained oligosaccharide is eluted with 3 mL of 25 mM triethylamine in water followed by 1 mL of 300 mM triethylamine in water. Fractions are recombined, dried, and stored at −20°C until ready for analysis. The dried sample is reconstituted in 1.5 mM MOPS buffered at pH 7 prior to analysis.

2.4. Nanogel and enzyme preparation.

Long-to-short-chain phospholipid molar ratios, q = DHPC/DMPC, are prepared as previously reported [11, 22] for q = 0.5 and 2.5 at 5 % and 10 %, respectively, in 50 mM sodium acetate buffer, pH 5. Neuraminidase powder is reconstituted in 50 mM sodium acetate buffer, pH 5, to a concentration of 33.6 mU/μL. A secondary stock is prepared by adding 1.0 μL of master to 100 μL of 10 % nanogel (q = 2.5). Finally, a working concentration of 3.4 μU/μL is made by adding 1.04 μL of secondary stock to 100 μL of 10% nanogel (q = 2.5). The working concentration should be the concentration of enzyme that results in 10% conversion of substrate to product.

2.5. Capillary electrophoresis.

Capillary electrophoresis separations were carried out using a P/ACE MDQ Plus (Sciex, Redwood City, CA) equipped with a laser-induced fluorescence detection system using a 50 μm i.d./360 μm o.d. fused-silica capillary (Polymicro Technologies, Phoenix, AZ). Capillary array separations were carried out using a Biophase 8800 (Sciex, Redwood City, CA) equipped with a laser-induced fluorescence detection system and an eight capillary array cartridge consisting of 50 μm i.d./360 μm o.d. fused-silica capillaries (L_t = 30 cm; L_d = 20 cm). Each day, the capillary is conditioned as previously reported [10]. Charges on the capillary surface are passivated using a semi-permanent coating composed of 5% phospholipid nanogel (q = 0.5). Passivating the surface charges suppresses the electroosmotic flow which enables the separation of anionic analytes under reverse polarity. The background electrolyte is 50 mM MOPS, pH 7. Substrate, enzyme, and nanogel coats are stored at 4 °C in a thermally regulated unit of the instrument. In the Biophase 8800 array instrument, inhibitors are also included in the refrigerated tray. Capillary patterning, including enzyme and inhibitor introduction to the capillary, is carried out at 19 °C. After patterning, the temperature is raised to 37 °C. The sample is introduced using electrokinetic injections at −2.5 kV 4 s and electrophoresis is used to drive the substrate through the reaction zone and the subsequent separation with an applied voltage of −5 kV (167 V/cm).

2.6. K_i curve fitting.

Dose-response curves are produced using an inhibitor dissolved in 50 mM MOPS, pH 7 where remaining enzyme activity is plotted against log[inhibitor]. GraphPad Prism version 9.1.2 (Dotmatics, San Diego, CA) is used to fit the curve as a sigmoidal dose-response four-parameter curve constrained to converge at 0 and 100% of normalized activity.

3. RESULTS AND DISCUSSION

3.1. Neuraminidase activity assay.

N-acetylneuraminic acid (i.e. sialic acid) is a negatively charged saccharide which caps many physiologically relevant oligosaccharides. In asparagine linked glycans (N-glycans), these terminal sialic acids most frequently form linkages from carbon 2 on the sialic acid to carbon 3 or carbon 6 on the penultimate galactose residue. As a result, sialic acids are critical ligands for biomolecular recognition [23]. Neuraminidase cleaves terminal sialic acid monomer from oligosaccharides. The goal of this study is to evaluate neuraminidase inhibition using a separation-based method that integrates the enzyme reaction with electrophoresis. Enzyme assays performed without a separation step rely on the use of a special label that undergoes a spectroscopic shift upon cleavage. A common example is 2’-(4-methylumbelliferyl)-α-D-N-acetylneuraminic acid, commonly referred to as MUNANA, for which 4-methylumbelliferyl (4-MU) generates a fluorescent signal at 460 nm once cleaved [4, 24, 25]. Challenges of this type of assay are the weak fluorescent signal of the 4-MU and the false-positive results generated from quenching of the fluorescence signal [6]. A separation-based assay provides exceptional flexibility in evaluating enzyme activity because it allows the user to distinguish the product from an intact substrate labeled with a more robust fluorophore. Using a separation-based approach also enables the investigation of the enzyme with any substrate, including saccharides with different sialic acid linkages on a wide array of glycoform ligands. Moreover, fluorescent detection at low substrate concentrations reduces the amount of enzyme (e.g. 100-fold) required for these analyses. The sialyllactose ligands contain the sialic acid-galactose saccharide sequence found on N-glycans, making them a relevant and readily available model of cellular systems. The action of neuraminidase is quantified by monitoring either the released sialic acid or the de-sialylated residue that remains. When the fluorescently labeled sialyllactose substrate is cleaved by the neuraminidase, the de-sialylated residue retains the fluorescent label making it easily detected. Integrating the separation with the reaction in capillary electrophoresis streamlines the quantification of both the product and the substrate.

3.2. Enzyme activity assays with fluorescently labeled sialyllactose.

An advantage of fluorescent detection is that the substrate can be easily conjugated to labels with high quantum efficiency that are matched to intense monochromatic light sources. For example, sensitive detection can be achieved by using 8-aminopyrene-1,3,6-trisulfonic acid (APTS) as the fluorescent labeling reagent because it has a high quantum yield (i.e., Φ≈0.95) that is insensitive to pH between pH 4–10 [26, 27], and is compatible with the common laser line of 488 nm. Protocols for bench top labeling of APTS at the reducing end of different saccharides are well established [28, 29] and the structures of the labeled substrate and reaction product with sialyllactose are shown in Figure 1A. With an APTS-labeled substrate, laser-induced fluorescence detection provides low detection limits of both the unconverted labeled substrate and the de-sialylated product. Additionally, there is flexibility in the fluorescent labeling protocol used to prepare the substrate prior to analyzing it with the capillary nanogel electrophoresis assay. For example, when the APTS labeling of 6’-sialyllactose for 5 min at 65°C was compared to labeling at 37°C, the product peak area of the 5-min reaction was 80% of that obtained overnight (see Figure S2, Table S2 in the Supporting Information). At the same time, spontaneous de-sialylation was not significantly different. While the decreased labeling time observed at 37°C may be favorable for specific applications, the reduced labeling efficiency does not adversely affect the electrophoresis assay because an exact sample concentration is not necessary when the substrate concentration is small enough to make the [S]/K_m ratio in Equation 1 insignificant to the calculation (i.e. (1 + [S]/K_m ~ 1). In this case, the determination of K_i is approximated as the half maximal inhibitory concentration, where K_i = IC₅₀.

Figure 1.

A) Structures of 6’-sialyllactose (SL) substrate and lactose (L) product labeled with 8-aminopyrene-1,3,6-trisulfonic acid (APTS, λ_ex= 425nm, λ_em= 503 nm) used to evaluate sialidase inhibition using fluorescence detection. B) Conceptual depiction of a separation-based enzyme activity assay. C) Separations of APTS-conjugated SL and L under reversed polarity and suppressed electroosmotic flow. The blue trace is achieved using a background electrolyte composed of nanogel while the black trace contains the same electrolytes, but is devoid of nanogel. Original traces and figures of merit are in Figure S1 and Table S1 in the supporting information.

The separation capillary was passivated with lipid to suppress the electroosmotic flow, allowing the separation to occur under reversed polarity (i.e. anodic reservoir at the detection site). A zone of enzyme, which was reconstituted in nanogel, was introduced into the separation capillary using pressure. This zone was then further pressure-pushed into the separation capillary with background electrolyte to ensure that it was located in the thermally controlled region of the capillary cartridge (see Fig 1B). By doing this, the temperature of the enzyme reaction was regulated, which is important because enzyme reactions are temperature dependent. Additionally, the pH of the reaction zone (pH 5) and background electrolyte (pH 7) was different to leverage the pH-sensitive nature of enzyme conversion. Accumulation of enzyme in areas outside of the enzyme reaction zone (e.g. on the capillary wall or at the zone interface) would not result in lactose product because of the pH. When the background electrolyte on either side of the reaction zone was comprised of an aqueous electrolyte, the sialyllactose substrate and lactose were not fully resolved in a 50-micrometer inner diameter separation capillary operated under reversed polarity with suppressed electroosmotic flow (see Fig 1C lower trace). To improve this resolution, the aqueous background electrolyte was modified to include nanogel. Nanogel is more viscous than aqueous electrolytes and is compatible with the semi-permanent lipid coating that self-assembles on the silica surface to mask the negative charge. As a result, the electroosmotic flow is more effectively suppressed by using a nanogel-based background electrolyte [30]. As shown in the upper trace in Figure 1C, when the enzyme zone was loaded into a separation capillary filled with nanogel the sialyllactose-lactose peak resolution was improved. In fact, the resolution of the 6’-sialyllactose and lactose peaks increased from 2.6 to 4.1 in the absence and presence of nanogel, respectively.

3.3. Separation before and after the reaction

The enzyme reaction was pushed into the thermally controlled region of the separation channel. This subjected the injected sample to an electrophoretic separation prior to the enzyme conversion. As the injected sample was separated prior to the reaction, it was not necessary for the substrate to be pure. For sialylated compounds, an additional advantage of a pre-reaction separation step was that it resolved any de-sialylated sialyllactose resulting from thermal or chemical degradation of the sample. This is beneficial to neuraminidase assays because it allows the user to quantify and account for non-enzymatic contributions to de-sialylation. The labeled sialyllactose used in this research contained 2% of lactose contaminant. If this non-enzymatically formed lactose contaminant was not accounted for following the enzyme reaction, the conversion that was measured would have been higher than the true value. In this separation-based assay, the trace lactose contaminant was separated from the sialyllactose substrate prior to the enzyme zone, and the lactose that was generated enzymatically within the reaction zone was separated from both the sialyllactose substrate and the lactose contaminant in the region of the separation channel following the reaction zone.

The effective separation length for the pre-reaction separation of the sample and the post-reaction separation of the substrate and product was determined by the position of the reaction zone within the separation capillary. This concept is illustrated in Figure 2A,B, in which the enzymatically generated lactose product peak migrates after the 6’-sialyllactose substrate but before the contaminant peak. As shown in the three electropherograms in Figure 2C, when the enzyme zone was pushed deeper into the capillary, the enzymatically generated lactose product peak shifted closer to the substrate peak and was baseline resolved from the contaminant. The electropherograms demonstrate that positioning the enzyme zone further into the separation channel improved the resolution between lactose product and contaminant (see Table 1). The separation efficiency of the lactose product decreased as the enzyme zone was pushed further. When longer pre-separation lengths were used to bracket the pH 5 nanogel zone within a pH 7 buffer, the longitudinal diffusion increased from the increased duration of the pressure-driven laminar flow. Based on these results, the enzyme reaction was positioned 5.2 cm in the capillary using a 25 s pressure push time to obtain the best peak resolution of the product peak from both the substrate peak and from the contaminant peak.

Figure 2.

6’-sialyllactose substrate containing a small percent of de-sialylated lactose is injected into a capillary patterned with nanogel. As the reaction zone is pushed deeper into the separation capillary the resolution of the remaining substrate, product, and non-enzymatically generated lactose contaminant are resolved. Original traces are in Figure S3 in the supporting information.

Table 1.

Effect of Enzyme Position on Separation

Enzyme position (cm)	RS,P	RP,C	NS×103	NP×103
3.1	1.7	0.8	140	50
5.2	1.5	1.0	80	50
7.3	1.4	1.1	70	40

Enzyme is positioned by applying 17 kPa (2.5 psi) of pressure to push the zone into the capillary. The time of the applied pressure is 15 s, 25 s, and 35 s for 3.1, 5.2, and 7.3 cm zones, respectively.

^*Resolution (R_s) calculated as the difference in migration time (t_m) divided by the average width at the peak base defined by 4σ (w_b). The plate count (N) calculated as 16 times the squared value of t_m divided by w_b. R_S,P and R_P,C are the resolution of the substrate (S) from product (P), and of the product from contaminant, respectively.

3.4. Inhibitor introduction and Ki validation of the reaction zone patterning.

The patterning geometry placed the 0.8 cm enzyme zone 5.2 cm within the capillary. A study was carried out to develop a method to bracket the enzyme with inhibitors as an alternative to directly spiking the inhibitor into an enzyme preparation. This was beneficial because it reduced the extensive enzyme consumption and method preparation that would have been required if several enzyme aliquots had been prepared to create a set containing each specific inhibitor concentration needed to construct the K_i curve. For this method the length of the inhibitor zones bracketing the reaction zone was only 4.8 cm and 0.8 cm on the injection and detection sides of the reaction zone, respectively. The inhibitor was reconstituted in aqueous electrolyte which made the method of inhibitor preparation simpler. Thus, 5.6 cm of the total 30 cm capillary length contained aqueous electrolyte rather than nanogel. As long as the nanogel was included on the detection side of the 0.8 cm aqueous inhibitor zone, the product, substrate, and contaminant peaks were sufficiently resolved (see Figs 3, 4, 5). Enzymatic hydrolysis of APTS-labeled 6’-sialyllactose was performed in the presence and absence of an inhibitor in the enzyme preparation using a reaction zone bracketed with inhibitor zones for both cases. Electropherograms and results are given in Figure S4, Tables S3, S4 in the Supporting Information. There was no statistical difference (students t-test, n = 3 each, ρ = 0.05) in the amount of product formed with inhibitor present or absent from the enzyme nanogel zone for either DANA or Siastatin B. Finally, a K_i curve was run using DANA on an MDQ Plus single capillary instrument using APTS-labeled 6’-sialyllactose. The doseresponse curves are given in Figure S5 in the Supporting Information. The K_i was determined to be 5.0 ± 0.3 μM which is consistent with previous literature values for DANA and sialidase from Clostridium perfringens reported as 5.0 ± 0.9 μM [14]. A student’s t-test confirms no significant difference (n = 3 each, ρ = 0.05).

Figure 3.

Concept diagram of patterning with nanogel, enzyme, and inhibitors using an 8-capillary array. Inhibitor concentration is varied across 8 capillaries including a blank, where no inhibitor is added, and 7 other concentrations for a dose-response curve. Inhibitor regions are given in blue with increasing concentrations corresponding to darker shades. Aqueous inhibitor zones are sandwiched around enzyme and the nanogel.

Figure 4.

K_i determination of DANA using a BioPhase 8800 8-capillary array. The electropherograms for each of the 8 capillaries are stacked and zoomed in to emphasize the increase in the lactose product peak with decreasing DANA concentrations. The dose-response curves obtained with only 3 runs produced a K_i of 5.6 ± 0.3 μM.

Figure 5.

K_i determination of Siastatin B using a BioPhase 8800 8-capillary array. The electropherograms for each of the 8 capillaries are stacked and zoomed in to emphasize the increase in the lactose product peak with decreasing Siastatin B concentrations. The dose-response curves obtained with only 3 runs produced a K_i of 9.2 ± 0.3 μM.

3.5. Inhibitor patterning for multi-capillary nanogel electrophoresis.

An 8-capillary nanogel electrophoresis array was used to rapidly quantify inhibitor performance against two different known Clostridium perfringens sialidase inhibitors (see Fig 3). Instead of completing a set of individual electrophoresis runs of different inhibitor concentrations performed in series, these runs were completed simultaneously in parallel to produce an entire dose-response curve in a single determination. Each inhibitor must be prepared as a standard and introduced into the sample tray. With the 8-capillary cartridge, all 8 capillaries are patterned at the same time. Since the inhibitor is introduced during the capillary patterning step and flushed out between runs, the analyst can easily switch inhibitors and rapid K_i determinations can be realized. As illustrated in Figure 3, one capillary contained no inhibitor and was used to normalize the percent activity remaining when an inhibitor was present. The other seven capillaries included the inhibitor at equally spaced concentrations on a log scale to produce the dose-response curve. When dose-response curves were obtained using a single capillary instrument, the analysis was completed with up to 9 concentrations; although, only 4 points between 20 % and 80 % inhibition were critical to quantify the IC₅₀ value [14].

3.6. K_i determinations of sialidase inhibitors in parallel using a capillary array.

The 8-capillary nanogel patterning approach was used to quantify the DANA and Siastatin B bacterial sialidase inhibitors (see Tables S5–S10 in the Supporting Information). The 8-capillary instrument is configured by the manufacturer to provide temperature regulation throughout the capillary cartridge. The instrument also provides temperature control in the compartment that stores the sample tray. The separation and quantification of the enzyme reaction was completed in under 7 minutes as shown in Figure 4 and Figure 5. The reproducibility in migration time from capillary-to-capillary and run-to-run was < 1 % RSD. The average (n = 3) K_i determined for DANA (Fig 4) and Siastatin B (Fig 5) was 5.6 ± 0.3 μM and 9.2 ± 0.3 μM, respectively, with a 5 % RSD. These values are in agreement (students t-test, n = 3 each, ρ = 0.04) with the K_i values obtained with single capillary analyses of DANA (5.0 ± 0.9 μM) and Siastatin B (11 ± 1 μM) which used 2-aminobenzoic acid-labeled 3’-sialyllactose as the substrate to quantify sialidase inhibition [14]. By using the APTS labeling for fluorescent detection, a lower concentration of substrate (i.e. 100 nM) was used to quantify the inhibition than was required with the 2-aminobenzoic acid labeling (i.e. 3 mM) for absorbance detection. These results demonstrate the effectiveness of the parallel approach and fluorescent labeling for a rapid inhibitor quantification assay.

3.7. Effects of sialyllactose linkage and excess APTS reagent on K_i determination.

The enzyme catalysis was screened against the substrates 3’-sialyllactose and 6’-sialyllactose conjugated to APTS with different labeling protocols. The effect on enzyme inhibition was established with a single determination with the substrate concentration equal to the K_i as this produces 50% inhibition of the enzyme catalysis. Inhibition was screened using DANA at a concentration of 5.0 μM based on the K_i determination. The following substrate preparations were investigated: 3’-sialyllactose and 6’-sialyllactose reacted with APTS overnight at 37°C; 6’-sialyllactose reacted with APTS for 5 min at 65°C; and 6’-sialyllactose reacted with APTS for 5 min with no solid phase extraction clean-up procedure for removal of excess dye. All four labeling procedures resulted in nearly 50% inhibition (see Fig S6, Tables S11–S14 in the Supporting Information) suggesting that under conditions where substrate concentration is << K_m, the determination of K_i for this enzyme is independent of the sialic acid linkage position of sialyllactose and that accurate K_i quantitation is not contingent upon complete labeling efficiency. Furthermore, since the substrate is diluted to nanomolar concentrations prior to analysis, it is not necessary to remove excess dye after the reaction is complete. These findings indicate that lengthy fluorescent conjugation and solid-phase extraction protocols for labeling and removal of excess dye are not required for inhibitor performance investigations.

4. CONCLUSIONS AND FUTURE DIRECTIONS

A capillary nanogel electrophoresis method with fluorescence detection for quantifying inhibition was validated using a model neuraminidase system. This new separation-based assay is appealing for the discovery of improved inhibitors because it is both quantitative and high throughput. The increased detection sensitivity with the fluorescent labeling reduces the amounts of enzyme, substrate, and inhibitor required for the assay. Moreover, using substrate at levels well below the K_m value makes the method independent of substrate concentration. As a result, the approach is compatible with different labeling protocols that do not achieve high labeling efficiency. Because the injected sample can be separated before passing through the enzyme reaction zone, it is possible to use samples that have not been processed to reduce the excess labeling reagent that remains after the labeling reaction. Additionally, the APTS labeling strategy described in this report can be performed with a wide range of substrates.

By completing the capillary patterning and separations in parallel with a multi-capillary array the time for analysis is decreased more than 8-fold because the processing of capillary filling, the temperature equilibration prior to the separation, and all the 7 min separations are performed simultaneously. Moreover, when the determination is completed in series with a single capillary instrument, additional analyses are needed because each inhibition result must be normalized against the amount of enzyme conversion obtained in the absence of an inhibitor. These enzyme conversion reference values must be repeated intermittently throughout the study when runs are done in series in case the enzyme performance fluctuates throughout the study. While the method requires access to the multi-capillary electrophoresis instrument, it can be completed with a more traditional single-capillary electrophoresis instrument. In this case, the benefits of a separation-based fluorescence assay are still realized. As presented in this report the quantification of a K_i value for a single inhibitor performed in triplicate is achieved with 24 individual electrophoresis runs, consuming a total of only 12 nanograms of enzyme. While a neuraminidase enzyme from Clostridium perfringens and known inhibitors (DANA, Siastatin B) are used as a model for method development and validation, the patterning approach can be easily adapted to other enzymes and potential inhibitors. Additionally, the fluorescent labeling can be completed in 5 minutes, adapted to other substrates, and analyses can be performed with or without post-reaction processing to remove the excess labeling reagent. A separation-based enzyme assay that is automated, sensitive, microscale, and high-throughput, better enables the discovery of pharmaceutically relevant inhibitors.

Highlights:

A 16-nanoliter nanogel enzyme reaction zone is integrated simultaneously in each separation channel in a capillary array.

In-line inhibition reactions are complete in seconds and quantified in 7-min parallel electrophoresis runs.

The substrate is rapidly conjugated to a fluorescent label, and K_i determinations are independent of the substrate concentration.

Abbreviations:

APTS

8-aminopyrene-1,3,6-trisulfonic acid

DANA

2,3-dehydro-2-deoxy-N-acetylneuraminic acid

DHPC

1,2-dihexanoyl-sn-glycero-3-phosphocholine

DMPC

1,2dimyristoyl-sn-glycero-3-phosphocholine

IC₅₀

half maximal inhibitory concentration

K_i

inhibitor constant

K_m

Michaelis-Menten constant

MOPS

3’-N-Morpholino propane sulfonic acid

4-MU

4-methylumbelliferyl

MUNANA

2’-(4-methylumbelliferyl)-α-D-N-acetylneuraminic acid