A novel fluorescent sensor protein for detecting changes in airway surface liquid glucose concentration

Both lung disease and elevation of blood glucose are associated with increased glucose concentration (from 0.4 to ~4.0 mM) in the airway surface liquid (ASL). This perturbation of ASL glucose makes the airway more susceptible to infection by respiratory pathogens. ASL is minute (~1 μl/cm2) and the measurement of glucose concentration in the small volume ASL is extremely difficult. Therefore, we sought to develop a fluorescent biosensor with sufficient sensitivity to determine glucose concentrations in ASL in situ. We coupled a range of environmentally sensitive fluorophores to mutated forms of a glucose/galactose-binding protein (GBP) including H152C and H152C/A213R and determined their equilibrium binding properties. Of these, GBP H152C/A213R–BADAN (Kd 0.86 ± 0.01 mM, Fmax/F0 3.6) was optimal for glucose sensing and in ASL increased fluorescence when basolateral glucose concentration was raised from 1 to 20 mM. Moreover, interpolation of the data showed that the glucose concentration in ASL was increased, with results similar to that using glucose oxidase analysis. The fluorescence of GBP H152C/A213R–BADAN in native ASL from human airway epithelial cultures in situ was significantly increased over time when basolateral glucose was increased from 5 to 20 mM. Overall our data indicate that this GBP is a useful tool to monitor glucose homoeostasis in the lung.


INTRODUCTION
The respiratory tract is covered by a thin layer of fluid that lines the luminal surface of the epithelium [airway surface liquid (ASL)], which is important for lung defence against infection. In the upper airway, glucose concentrations are normally ~0.4 mM, 12.5 times lower than in plasma at ~5 mM, which serves to deprive bacteria of a potential food source and thus limits bacterial growth [1][2][3][4]. There is accumulating evidence that glucose concentrations are raised in patients (up to 4 mM) during airway inflammation (e.g. in respiratory disease), when blood glucose levels are increased (e.g. in diabetes mellitus) and, more potently, when both are present. Similar changes can also be detected in the ASL of in vitro models of airway epithelium in response to basolateral hyperglycaemia and proinflammatory stimuli [5][6][7]. The increase in ASL glucose concentrations makes the airways more susceptible to infection from pathogens such as methicillin-resistant Staphylococcus aureus and Pseudomonas aeruginosa.
Measurement of glucose concentrations in the airways and distal lungs of patients, in in vivo models of airways disease and in vitro models of airway epithelium is difficult because of limited accessibility, the thinness of the ASL (~7 μm) and low concentrations of glucose obtained when dilution techniques are used (e.g. brochoalveolar lavage and exhaled breath condensate). Standard techniques such as glucose oxidase analysis or HPLC are concentration-limited and time-consuming respectively. Moreover, neither technique enables real-time dynamic changes in ASL glucose concentration in response to hyperglycaemic/pro-inflammatory challenge to be measured. The development of biosensors to measure glucose in the airway in vitro and in vivo will be a useful tool to better understand ASL glucose homoeostasis, how it is changed with inflammation and hyperglycaemia and how it modifies the innate immunity in the ASL.
Recently, the development of molecular glucose sensors for less invasive continuous monitoring of plasma glucose in patients has received significant interest in the field of diabetes research. One of these biosensors employs glucose/galactose-binding protein (GBP) which alters its conformation when glucose binds. This protein has been covalently linked to environmentally sensitive fluorophores which allow glucose binding to be detected as a change in fluorescence [8,9]. Native bacterial GBP has a very low dissociation constant (K d 2 μM) [8] which is not appropriate for ASL glucose measurement. However, in the present study we have used mutated forms of the GBP with altered binding constants (GBP H152C and GBP H152C/A213R), covalently bound to environmentally sensitive fluorophores. The molecular glucose sensors were characterized in terms of affinity and dynamic range, and their potential for measurement of glucose in ASL in situ was assessed.

Expression and purification of GBP proteins
The expression vectors pET303-GBP H152C and pET303-GBP H152C/A213R were used for the production of the GBP mutants [8]. GBP proteins were overexpressed in E. coli BL21(DE3) Gold. Cells were grown at 37 °C and expression was induced overnight at 20 °C in the presence of 0.5 mM IPTG. Cells were pelleted by centrifugation and processed as previously described [9]. Briefly, cells were resuspended in 50 mM NaH 2 PO 4 , 300 mM NaCl, 10 mM imidazole and 5 mM tris-(2-carboxyethyl)phosphine (TCEP), pH 8, supplemented with Complete EDTA-free protease inhibitor cocktail (Roche). The suspension was incubated on ice with 1 mg·ml −1 lysozyme for 30 min before lysis by sonication (VibraCell, Jencons PLS), clarified by centrifugation and purified by Nichromatography on an ÄKTA Purifier system (GE Healthcare) at 4 °C. The 5 ml Ni-NTA column was equilibrated with 50 mM NaH 2 PO 4 , 300 mM NaCl, 10 mM imidazole and 5 mM TCEP, pH 8. The purified protein was eluted in buffer containing 300 mM imidazole and stored at −80 °C. Purity of the eluted fractions was determined by SDS/PAGE.

Measurement of protein concentration
Protein concentration was determined using a Nanodrop 1000 spectrophotometer (Thermo Scientific) with a molar absorption coefficient (ε o ) of 37,930 M −1 ·cm −1 , calculated from the amino acid composition [13].
Protein labelling using thiol-reactive dyes (IANBD, BADAN, mero53, mero62 and mero87) For labelling with UV-excitable fluorophores, GBP proteins (100 μM) were incubated in the presence of 10-fold excess of dye (IANBD or BADAN) overnight at 4 °C in PBS and 5 mM TCEP, pH 7.4 [14]. The excess of dye was removed by gel filtration using a PD-10 column (GE Healthcare). For labelling with merocyanine dyes, 100 μM of protein were mixed with 5-fold excess of dye for 2 h at room temperature in PBS, pH 7.4. The excess of dye was removed by extensive dialysis in PBS, pH 7.4 at 4 °C (Slide-a-Lyser 10 kDa cut-off, Pierce).

Protein labelling using click chemistry for oxazine derivatives (Blue Oxazine, Red Oxazine, Nile Blue) and chromis dyes (Chromis 630, Chromis 678)
First, GBP-alkyne was obtained by incubating 100 μM GBP with 10-fold excess of iodoacetamide alkyne in PBS and 2.5 mM TCEP, pH 7.4, for 2 h at room temperature. The excess of iodoacetamide alkyne was removed by extensive dialysis in PBS, pH 7.4, at 4 °C (Slide-a-Lyser 10 kDa cut-off, Pierce). The second step of the process consisted of labelling the GBP-alkyne with the dye azide using the Click-iT ® Protein Reaction Buffer kit following the manufacturer's recommendations. Briefly, 40 μM GBP-alkyne was incubated with 2.5-fold excess of dye in 100 μM reaction buffer containing 25 μl of copper sulfate reagent and 25μl of kit additive 1, for 1 h at room temperature (V t = 500 μl). The excess of dye was removed by extensive dialysis in PBS, pH 7.4, at 4 °C (Slide-a-Lyser 10 kDa cutoff, Pierce).

Equilibrium glucose binding
Equilibrium glucose-binding curves were obtained by continuous titration followed by corrections for dilution and photobleaching. To determine the fluorescence dynamic range (F max /F 0 ), the co-operativity (n) and the dissociation constant (K d ) of the labelled GBP proteins, glucose affinity assays were performed using an automated syringe pump (ALADDIN 1000, WPI). Labelled GBP proteins (40-100 nM) were titrated continuously with D-glucose at 10-20 μl/min flow rate in a stirred 3 ml cuvette, in PBS, pH 7.4. Fluorescence was measured at fluorophore excitation and emission wavelength peaks (Table  1)
Results are presented as means ± S.E.M., unless otherwise stated.

Equilibrium glucose binding to GBP H152C and H152C/A213R
GBP proteins were highly purified on a single-step affinity chromatography column ( Figure  S1), allowing accurate concentration measurements by spectrophotometry. One litre of culture yielded in 200-250 mg of purified protein.
Glucose equilibrium binding for labelled GBP H152C and H152C/A213R showed that using UV-excitable environmentally sensitive dyes such as BADAN and IANBD offered the highest fluorescence dynamic range (F max /F 0 ) ranging from 2.7 to 3.6 ( Table 1). GBP labelled with long-wavelength excitable dyes showed a maximum fluorescence change upon glucose binding of 80 %for GBP H152C coupled to merocyanines. The dissociation constant (K d ) for GBP H152C and GBP H152C/A213R ranged from 0.046 ± 0.002 mM to 0.68 ± 0.02 mM and from 0.86 ± 0.01 mM to 15.2 ± 0.3 mM, respectively ( Figure 1 and Table 1). For most labelled glucose-binding proteins, glucose binding did not occur with cooperativity as determined by analysis of the Hill coefficient (n). All the parameters obtained from the glucose titration curves are summarized in Table 1. GBP H152C and GBP H152C/ A213R labelled with Red Oxazine and Chromis 630 did not show a significant fluorescence change upon glucose binding (results not shown).

Specificity of GBP H152C/A213R-BADAN (GBP-BADAN) for D-glucose
The specificity of glucose induced fluorescence of GBP-BADAN was analysed in PSS by comparing fluorescence in the presence of D-glucose, its isomer L-glucose and D-fructose. Fluorescence did not change with increasing D-fructose or L-glucose concentration but exhibited a similar dose response to D-glucose as described above (Figure 2A). There was a significant difference in the mean values between 0.01 and 0 mM D-glucose (i.e. in the presence of L-glucose) (P = 4.6944×10 −5 , n = 3). This indicated that the limit of detection was below 0.01 mM. However, we did not explore this further because values below 0.01 mM were not relevant to our study system. This was also the case for glucose concentrations above 50 mM. The calculated value for precision (3.3×S.D./slope for the linear part of the graph shown in Figure 1) was 0.27 mM.

Comparison of GBP-BADAN fluorescence with glucose oxidase measurement of glucose concentration
Glucose oxidase analysis showed similar D-glucose specificity, with mM values obtained for increasing concentrations of L-glucose and D-fructose being unchanged from 0 ( Figure  2B). Interestingly, GBP-BADAN exhibited changes in fluorescence at low concentrations of D-glucose (from 0.01 mM; Figure 2A). Changes in low glucose concentrations were less discernable by glucose oxidase analysis. Furthermore, changes in fluorescence could be detected at concentrations up to 50 mM where no accurate and repeatable readings could be determined at this concentration by glucose oxidase analysis calibrated for standard glucose measurement (glucose oxidase data not shown). Therefore, GBP-BADAN appears to be a more sensitive probe than glucose oxidase and can measure glucose over a broader range of concentration.

GBP-BADAN to measure hyperglycaemia-induced changes in Calu-3 ASL glucose concentration
Serous and mucous cells of the airway submucosal glands make a significant contribution to the volume, composition and consistency of the ASL in vivo. Calu-3 cells, a heterogenous cell line derived from human submucosal glands, secrete a protein-rich fluid containing mucins, antimicrobials and enzymes [15,20,21]. Calu-3 monolayers were used to determine whether the protein-rich ASL would affect the stability of GBP-BADAN to accurately measure glucose content. Fluid was added to the apical surface and removed after 1 and 24 h, and glucose concentrations measured by glucose oxidase and GBP-BADAN fluorescence. Hyperglycaemia was mimicked by increasing the D-glucose concentration of the basolateral fluid (from 1 to 20 mM), as previously described [6].
GBP-BADAN fluorescence in ASL increased significantly when basolateral D-glucose was elevated from 1 to 5 mM and above (P < 0.05, n = 4) with maximal values occurring when basolateral glucose was 15-20 mM (P < 0.01, n = 4). ASL glucose concentrations were higher in samples obtained after 24 h compared with 1 h after elevation of basolateral Dglucose ( Figure 3A).
A similar pattern was observed when glucose concentration in the samples of ASL was analysed by glucose oxidase method. At 1 h, glucose concentration in ASL was very low (0.007 ± 0.002 mM) and at the limit of detection when basolateral D-glucose was 1-2 mM but significantly increased when basolateral glucose was raised to 5 mM (0.3 ± 0.06 mM in ASL; P < 0.01, n = 4) and was highest when basolateral D-glucose was increased to 15 mM (1.1 ± 0.3 mM; P < 0.01, n = 4). ASL glucose further increased after 24 h rising to 0.6 ± 0.17 and 5.0 ± 1.2 mM when basolateral D-glucose was 5 and 15 mM, respectively ( Figure  3B). Importantly, ASL glucose concentration remained lower than basolateral glucose at any concentration tested.
Airway epithelial monolayers maintained their transepithelial resistance when glucose in the basolateral chamber was replaced with fructose, as previously reported [22]. As fructose did not bind to GBP-BADAN, ASL from cells grown in basolateral fructose was used to obtain ASL in which glucose was not present and correct for background fluorescence. Then, using the standard curve ( Figure 2C), ASL glucose concentration can be determined. Importantly, the interpolated ASL glucose concentrations were similar to those determined by glucose oxidase analysis at 0.2 ± 0.1 and 0.8 ± 0.4 mM when basolateral D-glucose was 5 mM and 2.0 ± 0.8 and 4.8 ± 1.1 mM when basolateral D-glucose was 15 mM at 1 and 24 h respectively ( Figure 3C). These data indicate that GBP-BADAN fluorescence can be used to accurately measure ASL glucose at similar sensitivity to glucose oxidase analysis.

GBP-BADAN to measure dynamic changes in HBEC ASL glucose concentration in response to hyperglycaemic challenge in situ
Having shown that GBP-BADAN could be used to measure changes in ASL glucose concentrations collected from Calu-3 monolayers, we investigated whether the GBP assay could be used to detect dynamic changes in situ, in response to elevated basolateral glucose, by applying GBP-BADAN directly to the apical surface of primary HBECs. Applying even small volumes of fluid to the apical surface of HBECs induces fluid absorption [23]. Therefore, to control for the effect of ASL volume changes on glucose concentration, we used Texas Red-labelled dextran together with GBP-BADAN and calculated the change in fluorescence of GBP-BADAN/Texas Red ratio. Using this methodology, we found that fluorescence increased in the ASL of all HBECs after 1 h and reached peak fluorescence after 3 h (Figure 4). Although no significant difference in GBP-BADAN fluorescence was observed in the ASL of HBECs bathed in 10 mM basolateral glucose compared with 5 mM control, increasing basolateral glucose to 20 mM significantly enhanced fluorescence after 2 h (P < 0.05, n = 4). The increase in GBP-BADAN fluorescence was maintained throughout the 5 h time-course and was also significantly greater than cells bathed in 5 mM D-glucose plus 15 mM L-glucose (from 3 h onwards; P < 0.05, n = 4), consistent with a glucoseinduced change in fluorescence rather than one induced by a difference in osmotic gradients.
The fluorescence of GBP-BADAN was monitored over 7 h at 37 °C and no significant change in fluorescence was observed indicating that the labelled protein was stable over this time course ( Figure S2).

DISCUSSION
The world burden of diabetes is increasing both as an independent disease and as comorbidity with other chronic conditions. In the lung, evidence is accumulating that hyperglycaemia associated with Type 1 and Type 2 diabetes results in increased glucose concentration in the ASL which promotes infection, particularly when associated with underlying respiratory disease [1,5,6,[24][25][26][27]. However, measurement of real-time changes of ASL glucose in vitro and in vivo has not previously been possible.
Evidence indicates that glucose concentrations in the ASL range from approximately 0.4 to 4.0 mM in vivo and in vitro [1,6,7,24]. In order to efficiently monitor glucose concentrations in ASL in vivo, the biosensors should have a K d between 0.1 and 10 mM and a high fluorescence dynamic range upon glucose binding. We initially disregarded GBP H152C (which contains a mutation to enable covalent labelling of fluorophores via the cysteine amino acid) because when it was bound to BADAN the K d was 2 μM, somewhat below our ideal range [8,28]. However, our data showed that when bound to other fluorophores, particularly IANBD, a shift in the K d to higher, more useful, values in the millimolar range could be achieved. Similar increases in K d were also observed when these fluorophores were attached to GBP H152C/A213R. These data indicate that the molecular structure of the fluorophore bound to these proteins can significantly modify affinity for the ligand, consistent with the findings of others [9,29]. Unfortunately, linking the double mutant to IANBD increased the K d to ~32 mM (results not shown), making it unlikely to be suitable for ASL glucose measurement, although it gave the next best dynamic range after BADAN. Interestingly, all the longer wavelength fluorophores, which were designed for in vivo imaging, were less sensitive to environmental changes and had lower fluorescence dynamic ranges. These fluorophores were based on optimization of both environmental sensitivity and brightness, as these factors together determine sensitivity in living cells [30]. Of these dyes, GBP H152C/A213R-mero62 was the most promising for further development with a K d of 1.4 mM and a fluorescence dynamic range of 1.7.
GBP H152C/A213R-BADAN (GBP-BADAN) gave the most useful K d (0.86 ± 0.01 mM) and the largest fluorescence dynamic range (F max /F 0 3.5) and exhibited clear specificity for D-glucose. Moreover, we found that it was a better sensor over the range of glucose concentrations required than using the glucose oxidase method. This was in part due to technical limitations of the sensing machine. That is, in order to obtain readings at concentrations below 1 mM, the Glucox analyser has to be recalibrated differently from settings used to obtain blood glucose readings (2-20 mM). This meant that glucose concentrations in the higher range >10 mM did not give a consistent output value. When measuring glucose in samples of ASL removed from the surface of cells, GBP-BADAN and glucose oxidase analysis showed similar changes with increasing basolateral glucose. Glucose oxidase analysis gave absolute glucose concentrations whereas we obtained fluorescence intensity for GBP-BADAN. We utilized the finding that the presence of basolateral D-fructose (which maintains cell survival) is not sensed by GBP-BADAN to obtain background fluorescence readings for ASL with no glucose. This enabled us to convert the mean fluorescence data into glucose concentration (mM) by interpolation of the standard curve. Glucose concentration values obtained by fluorescence and the glucose oxidase data were similar, providing support for the use of GBP-BADAN to quantify glucose in ASL. Most importantly, the data we gained from these experiments showed that glucose concentration in ASL increased with raised basolateral glucose but concentrations remained significantly lower in ASL, consistent with our previous findings in vitro and in vivo [6,7,24].
Work by others has shown that highly technical glucose sensor technology using ZnOcoated microelectrodes or a phosphorescence sensing material [crystalline iridium(III)-Zn(II) co-ordination polymers], which utilize glucose oxidase production of H 2 O 2 to quantify glucose, can be useful in measuring glucose in small fluid volumes such as ASL and over similar ranges of glucose (0.1-6 mM) [26,31,32]. However, the benefit of a fluorescent molecular glucose sensor that is easy to produce and can be used to rapidly measure real-time (and reversible) changes in glucose concentration in ASL overlying cells in situ is particularly attractive. In a unique experiment, we used GBP-BADAN applied directly to the ASL of HBECs in situ and showed that glucose concentration rises faster and to higher levels over a time course of 5 h when basolateral glucose concentration was increased from 5 to 20 mM. These data support our model of glucose homoeostasis where raised basolateral glucose concentration increases the driving force for glucose diffusion into ASL. The resultant ASL glucose concentration is the net balance between diffusion and glucose uptake by the epithelial cells. Thus, if uptake remains the same, increased diffusion would result in elevated ASL glucose [33,34].
These experiments now open the way for development of fluorophore-labelled GBPs as biosensors, for use in vitro and in vivo, to further understand how hyperglycaemia and respiratory disease change ASL glucose concentration in the proximal and distal lung. We now intend to use GBP-BADAN applied directly to ASL of proximal and distal airway cells in vitro to explore factors regulating glucose homoeostasis. In vivo, we have previously used glucose oxidase sticks to detect glucose in human nasal secretions. We have shown that changes in glucose concentration in nasal secretions are representative of the airway as a whole [24,35,36]. However, these sticks are no longer available, colour change interpretation was subjective and dynamic changes were difficult to measure. Therefore, immobilization of GBP-BADAN and its incorporation into a small fibre-optic probe (<1 mm diameter) for insertion into the nasal cavity, could prove a more accurate and minimally invasive biosensor for detection of glucose in nasal secretions [37]. During our in vitro experiments we did not observe any changes in transepithelial resistance which indicated GBP-BADAN toxicity, but this will need to be tested further. Nevertheless, we hope such tools will be helpful to find new ways to reduce glucose in ASL and help combat respiratory infections.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.  Table 1.