- Email: [email protected]
Novel fluorescent sensing system for α-fructosyl amino acids based on engineered fructosyl amino acid binding protein
Biosensors and Bioelectronics 22 (2007) 1933–1938
Novel fluorescent sensing system for ␣-fructosyl amino acids based on engineered fructosyl amino acid binding protein Akane Sakaguchi, Stefano Ferri, Wakako Tsugawa, Koji Sode ∗ Department of Biotechnology, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan Received 20 April 2006; received in revised form 10 August 2006; accepted 14 August 2006 Available online 2 October 2006
Abstract A novel fluorescent sensing system for ␣-glycated amino acids was created based on fructosyl amino acid binding protein (FABP) from Agrobacterium tumefaciens. The protein was found to bind specifically to the ␣-glycated amino acids fructosyl glutamine (Fru-Gln) and fructosyl valine (Fru-Val) while not binding to -fructosyl lysine. An Ile166Cys mutant of FABP was created by genetic engineering and modified with the environmentally sensitive fluorophore acrylodan. The acrylodan-conjugated mutant FABP showed eight-fold greater sensitivity to Fru-Val than the unconjugated protein and could detect concentrations as low as 17 nM, making it over 100-fold more sensitive than enzyme-based detection systems. Its high sensitivity and specificity for ␣-substituted fructosyl amino acids makes the new sensing system ideally suited for the measurement of hemoglobin A1c (HbA1c), a major marker of diabetes. © 2006 Published by Elsevier B.V. Keywords: Diabetes; Hemoglobin A1c (HbA1c); Sensing system; ␣-Fructosyl amino acid; Periplasmic binding proteins (PBPs); Fructosyl amino acid binding protein (FABP)
1. Introduction Periplasmic binding proteins from Gram-negative bacteria (bPBPs) serve as essential primary receptors for the transport of carbohydrates, amino acids, anions, metal ions, and peptides (Wandersman et al., 1979; Oh et al., 1994; Powlowski and Sahlman, 1999; Jacobson et al., 1991; Davies et al., 1999). bPBPs are monomeric and consist of two domains linked by a hinge region. The ability of bPBP to undergo drastic conformational changes upon binding a target molecule has led to the design of several novel biosensing systems (Dwyer and Hellinga, 2004). A number of bPBPs have recently been successfully used as scaffolds for the development of fluorescent nanosensors by modification with appropriately positioned fluorescent probes (De Lorimier et al., 2002). bPBPs labeled with environmentally sensitive fluorophores undergo changes in fluorescence intensity or emission spectra when binding with their corresponding target molecules. This principle is not only limited to in vitro monitoring, but has also recently been applied to an in vivo FRET
∗
Corresponding author. Fax: +81 42 388 7072. E-mail address: [email protected] (K. Sode).
0956-5663/$ – see front matter © 2006 Published by Elsevier B.V. doi:10.1016/j.bios.2006.08.022
nanosensor based on a fluorescence protein fused to bPBP for glutamate (Okumoto et al., 2005). Although several attempts have been reported on the designing of on-demand substratebinding proteins, the development of bPBP-based sensing systems is strongly dependent on the natural diversity of available bPBPs that can be produced by recombinant technology. To develop novel diagnostic kits for monitoring glycemic control of diabetes patients, we have been engaged in the construction of novel biosensing systems for glycated proteins, such as the glycated hemoglobin HbA1c (Ladenson et al., 1985) and glycated albumin (Yatscoff et al., 1984), based on enzymes (Tsugawa et al., 2000, 2001; Ogawa et al., 2002; Sakaguchi et al., 2003) and molecularly imprinted polymers (Sode et al., 2001; Sode et al., 2003; Yamazaki et al., 2003). Protein glycation is a series of non-enzymatic reactions of reducing sugars with free amino groups on proteins (Hudge and Rist, 1953). Glycation starts with a reversible reaction to produce a Schiff base intermediate that can then undergo an essentially irreversible Amadori rearrangement to produce an Amadori product, such as fructosyl amine. The biochemical monitoring of protein glycation based on the measurement by fructosyl amine oxidase (FAOD) of fructosyl amino acids released by proteolytic digestion has recently been
1934
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
reported (Hirokawa et al., 2004; Sakurabayashi et al., 2003). FAODs, with Km values generally of 10−3 to 10−2 M, catalyze the oxidative degradation of fructosyl amino acids. Since glycated proteins, such HbA1c and glycated albumin, cannot be substrates for FAODs, these glycated proteins are subjected to proteolytic digestion prior to reaction with FAODs. The liberated the fructosyl amino acids, N␣ -substituted fructosyl-valine (FruVal) and N -substituted fructosyl-lysine (-Fru-Lys) released by protease digestion of HbA1c and glycated albumin, respectively, can then be detected enzymatically. An FAOD that reacts specifically with Fru-val might therefore be an ideal component of an HbA1c enzyme sensor. We reported an amperometric sensor system for HbA1c detection employing ␣-fructosyl amino acid specific FAOD (Sakaguchi et al., 2003). Unfortunately, a large quantity of FAOD is required for the construction of a biochemical monitoring kit for HbA1c due to the low concentration of target molecules (10−4 M in human blood) and the enzyme’s relatively high Km value. Considering that bPBPs generally show affinities (Kd generally 10−7 to 10−6 M) that are over 1000-fold higher than most enzymes, the application of bPBP for glycated protein sensing may provide a highly sensitive and effective system. We have recently succeeded in the identification and recombinant production of bPBP for fructosyl amino acid (Sakaguchi et al., 2005). This was the first report demonstrating bPBP conformational change resulting from binding to fructosyl amino acid, which was monitored by autofluorescence. Furthermore, this protein is the only fructosyl amino acid binding protein (FABP) that has been prepared by recombinant DNA technology. The current paper reports the construction of a novel fluorescent sensing system for fructosyl amino acid based on genetically engineered FABP. We found that the FABP is specific for ␣-glycated amino acids over -glycated amino acid. This selectivity is beneficial for applying FABP in HbA1c sensing to distinguish the Fru-Val produced from the degradation of HbA1c from the -Fru-Lys resulting from the degradation of glycated albumin. We also constructed a 3D structural model of our novel FABP to determine the ideal locations to introduce environmentally sensitive fluorophores to improve the sensitivity of the sensing system. 2. Materials and methods 2.1. Chemicals, enzymes, and bacterial strain Fru-Gln, Fru-Val, and -Fru-Lys were synthesized from d-glucose with l-glutamine, l-valine, and Na-Boc-l-lysine, respectively, by the method of Finot (Finot et al., 1969), with minor modifications. The purity of synthesized compounds was determined by TLC and by the alkaline-NBT method (Baker et al., 1993). PfuUltra High-Fidelity DNA polymerase was purchased from Stratagene (CA, USA). Restriction endonucleases were purchased from Takara (Kyoto, Japan). Acrylodan (6-acryloyl-2-dimethylaminonaphthalene) was from Molecular Probes Invitrogen Detection Technologies (CA, USA). All other chemicals were of reagent grade.
Fig. 1. Ribbon depiction of the closed (left) and open (right) forms of the predicted 3D structure of FABP based on Gln-BP (pdb:1WND). Tryptophan residues are shown in ball and stick style (pink); the upper residue corresponds to Trp256, which was selected for mutagenesis. The residues predicted to make up the substrate-binding domain (Met52, Ala56, Asp71, Phe96, Ala113, Ala114, Gly116, Arg121, Leu163, Leu200 and Asp201) are shown in wire frame style (yellow). Ile 166, selected for substitution to Cys, is shown in stick style (green). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)
2.2. Construction of FABP cysteine mutants A 3D structural model of fructosyl amino acid binding protein (FABP) from Agrobacterium tumefaciens EHA (Rhizobium radiobacter EHA) (Sakaguchi et al., 2005) was predicted by the molecular operating environment (MOE) software (Chemical Computing Group, Montreal, Canada) based on the open form structure of glutamine binding protein (Gln-BP, pdb:1WND, (Sun et al., 1998)) (Fig. 1). Based on the 3D model, residues Ile166 and Trp256 were selected for substitution to cysteine. Site-directed mutagenesis was carried out on the FABP expression vector pESA (Sakaguchi et al., 2005) by the QuikChange method (Strategene, CA, USA) using the following mutagenic primer pairs: (i) 5 -GCACGCTGCAGGAATGCTATGCCGAGAAG-3 and 5 -CTTCTCGGCATAGCATTCCTGCAGCGTGC-3 for the Ile166Cys mutation, and (Prendergast et al., 1983) 5 TGCAGGACGGCACCTGCAAGAAGCTTCATG-3 and 5 GAAGCTTCTTGCAGGTGCCGTCCTGCATGG-3 for the Trp256Cys mutation (mutation sites are underlined). The sequences of the created mutants were confirmed using the ABI Prism BigDye Terminator Cycle Sequencing Kit v3.0 on an ABI Prism 3100 Genetic Analyzer (Applied Biosytems, CA, USA). 2.3. Autofluorescence monitoring Wild-type (WT) and mutant (Ile166Cys and Trp256Cys) FABPs were expressed in E. coli and purified as described previously (Sakaguchi et al., 2005). The binding ability of purified FABPs were then analyzed by monitoring autofluorescence and dissociation constants (Kd ) were calculated by double reciprocal Klotz plot using the relative fluorescence intensity change
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
at each substrate concentration. The relative intensity changes (F/F0 ) were calculated as the percentage of the total intensity in the absence of substrate and used to evaluate the FABP autofluorescence changes. 2.4. Acrylodan modification of FABP mutants Purified mutants were modified with the environmentally sensitive fluorophore acrylodan as described in the Molecular Probes manual for thiol-reactive probes. Briefly, each purified FABP mutant was incubated at a concentration of 20 M in 10 mM potassium phosphate buffer, pH 7.0, containing 300 M acrylodan, for 8 h at 4 ◦ C. The unconjugated acrylodan was removed by using Microcon YM-10 centrifugal filter units (10 kDa MW cut-off, Millipore) at 5000 × g. 2.5. Characterization of fluorescently labeled FABP The fluorescence of acrylodan-conjugated FABP mutants was monitored using the same system as autofluorescence monitoring. Slits of excitation and emission monochromators were set at 5 and 3 nm, respectively. The excitation wavelength was 390 nm. Maximum intensity was measured after incubating acrylodan-conjugated FABP mutants at a final concentration of 0.6 M with each substrate for 2 min at 25 ◦ C. 3. Results and discussion 3.1. Substrate specificity of FABP The drastic conformational changes that FABP undergoes upon binding its ligand can be monitored by measuring tryptophan-related protein autofluorescence (Sakaguchi et al., 2005). The substrate specificity of FABP binding was investigated by measuring autofluorescence in the presence of different potential ligands. Based on both Kd values and maximum relative intensity changes, FABP binds preferentially to Fru-Gln, resulting in an 8.2% increase in autofluorescence intensity and a calculated Kd value of 0.2 M. FABP also binds to Fru-Val with very high affinity, resulting in a 3.9% autofluorescence increase with a Kd value of 0.6 M. Binding of glucose, fructose, glutamine, and valine could not be detected. Remarkably, incubation of FABP with -Fru-Lys also failed to produce any detectable conformational changes. The socA gene encoding FABP was originally identified as a component of the soc operon (Baek et al., 2003) in A. tumefaciens, responsible for the assimilation of the opine santhopine, a common name for Fru-Gln (Kim et al., 2001). Considering the putative metabolic role of FABP, the protein’s preference for FruGln is not surprising. Our results showing that FABP recognizes both Fru-Gln and Fru-Val with high affinity while not recognizing -Fru-Lys indicates that the protein is specific for ␣-fructosyl amino acids. Gerhardinger et al. previously reported the isolation of an -fructosyl amino acid-specific Amadori product binding protein from the fructosyl amino acid-utilizing soil bacterium Pseudomonas sp. (Gerhardinger et al., 1994). Our FABP is a unique binding protein that can recognize specifically Fru-Val
1935
and can discriminate -Fru-Lys, which is a very desirable property for a component an HbA1c monitoring system. 3.2. 3D model prediction of FABP Secretion into the periplasmic space leads to the cleavage of FABPs N-terminal 35-residue signal sequence to produce a monomeric peptide composed of 242 amino acids (Sakaguchi et al., 2005). The absence of cysteine residues in FABP allows the convenient and specific fluorophore modification via the formation of thiol bond by introducing a unique Cys residue. We created a 3D structural model of FABP to help us determine the ideal locations to introduce such cysteine substitutions. A previous amino acid sequence homology analysis indicated that FABP shares a conserved structural scaffold with cluster 3 binding proteins (Sakaguchi et al., 2005). Particularly high similarities were observed with polar amino acid binding proteins, such as histidine binding proteins (Oh et al., 1994; Wolf et al., 1995), lysine/arginine/ornithine binding protein (Kang et al., 1991) and Gln-binding protein (Sun et al., 1998) from E. coli and Salmonella typhimurium. FABP contains two signature sequences within the N-terminal domain, residues 69–80 and 101–115, that are characteristic of cluster 3 binding proteins (Tam and Saier, 1993). Furthermore, the residues predicted to be at the ligand recognition site are almost identical to those of other cluster 3 binding proteins. SocA is therefore expected to have a 3D structure similar to those of other cluster 3 binding proteins. In addition to the high similarity between FABP and Gln-BP (27% identical, 40% similar), secondary structure predictions also show that the hinge region of FABP is of the same length as Gln-BP and shorter than those of the other polar amino acid-BPs. Gln-BP was therefore used as a template to create a 3D structural model of FABP. As other cluster 3 binding proteins, the 3D model of FABP shows two distinct structural domains with the ligand-binding site located at their interface (Fig. 1). Met52, Ala56, Asp71, Phe96, Ala113, Ala114, Gly116, Arg121, Leu163, Leu200 and Asp201 were predicted by homology alignment with GlnBP to be residues interacting with the ligand. Because modification of these residues would likely adversely affect ligand binding, they were avoided as candidate sites for fluorophore labeling. The autofluorescence from FABPs two Trp residues, Trp256 and Trp263, changes upon binding Fru-Gln in a concentrationdependent manner (Sakaguchi et al., 2005). These Trp residues were therefore considered potential locations for introducing an environmental sensitive fluorophore. Trp263 was avoided because the predicted 3D model shows that it is located near the substrate-binding site. Trp256 is predicted to be located in the protein scaffold of one of the two major domains and was therefore selected for substitution to Cys for subsequent fluorophore modification. Ile166, located close to the active site without contributing to ligand binding, was also chosen for substitution to Cys and fluorophore modification. 3.3. Amino acid substitution of FABP We created the Ile166Cys and Trp256Cys mutants of FABP and expressed them in E. coli. As with the wild-type protein,
1936
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
Fig. 2. Effect of Fru-Val binding on the autofluorescence emission spectra of FABP mutants. The fluorescence spectra of purified Trp256Cys and Ile166Cys (1.6 M) in the absence (dotted line) and presence of 1.5 M Fru-Val. The measurements were carried out in 10 mM MOPS buffer (pH 7.0) at 25 ◦ C. The results are expressed as arbitrary units (arb.U).
these engineered FABP mutants were recovered in soluble form from the periplasmic space. Spectroscopic analysis of Ile166Cys and Trp256Cys FABPs both showed autofluorescence maxima of 335 nm with excitation at 295 nm (Fig. 2), as was observed with the wild-type protein. The fluorescence intensity of Ile166Cys was approximately 53 arbitrary units (arb.U), almost identical to that of the wild-type protein. The fluorescence intensity of Trp256Cys (approximately 11 arb.U) was greatly reduced to less than 25% of the wild-type level, which is not surprising considering that we are measuring Trp-related autofluorescence. Both mutant proteins showed similar changes in autofluorescence upon addition of an equimolar amount of the ligand Fru-Val (Fig. 2). Addition of higher concentrations of FruVal resulted in an almost 3% increase in autofluorescence of Ile166Cys and a 15% increase that of Trp256Cys (Fig. 3). The wild-type protein had previously shown an intermediate relative intensity increase of 6% (Sakaguchi et al., 2005). The fluorescence intensity of non-ligand-binding Trp256Cys FABP is fourto five-fold lower than the wild type and Ile166Cys mutant, indicating that Trp256 was the major contributor to autofluorescence. However, the absolute changes in fluorescence upon ligand binding are 2.5 arb.U, 2.2 arb.U, and 1.7 arb.U for wildtype, Trp256Cys, and Ile166Cys, respectively (data not shown). This suggests that the major contribution of Trp256 to autoflu-
Fig. 3. Correlation between Fru-Val concentration and autofluorescence. The maximal autofluorescence intensity changes were determined at different FruVal concentrations under the same conditions as Fig. 2.
orescence changes only slightly, if at all, upon ligand binding. This may explain why Trp256Cys undergoes such a large relative fluorescence change upon addition of Fru-Val, although we cannot rule out other effects such as unexpected structural changes. From the dependence of relative autofluorescence change on ligand concentration (Fig. 3), Kd values for Fru-Val binding to Ile166Cys and Trp256Cys FABPs were calculated to be 2.8 and 1.4 M, respectively. Although these values are two- to fourfold higher than for Fru-Val binding to wild-type FABP, they still retain their ligand-induced bending motion and a ligand binding ability that is several hundred times higher than enzymes, thus remaining ideal biosensor elements for HbA1c. These results also support the validity of our structural models. 3.4. Characterization of fluorescence-labeled FABP Ile166Cys and Trp256Cys FABPs were then modified with the environmentally sensitive fluorophore acrylodan to investigate their potential as elements for fluorescent sensors. Acrylodan fluorescence emission is highly sensitive to its local environment when bound to protein, and exhibits changes in both intensity and emission wavelength that reflect the effective dielectric constant of the environment around the fluorophore. The emission spectrum of acrylodan-conjugated Ile166Cys (Ile166Cys-Ac) shows a single emission peak at 520 nm with excitation at 390 nm (Fig. 4). Acrylodan-conjugated Trp256Cys (Trp256Cys-Ac) showed two emission peaks, at 456 and 472 nm (data not shown). However the fluorescence rapidly decreased after fluorophore modification and we were unable to obtain a stable fluorescence spectrum for Trp256Cys-Ac to investigate its ligand binding ability. The fact that Ile166Cys-Ac shows a stable fluorescence signal during the spectroscopic analysis indicates that the fluorescence character of acrylodan itself is not responsible for the instability of Trp256Cys-Ac fluorescence under our assay conditions. Furthermore, considering that unconjugated Trp256Cys can be prepared as a pure soluble protein in similar quantities as wild-type and Ile166Cys FABPs, the amino acid substitution has no detectable effect on the protein’s structural stability. Although Trp256Cys is a stable protein,
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
Fig. 4. Effect of Fru-Val binding on the fluorescence emission spectra of acrylodan-modified FABP. The fluorescence spectra of acrylodan-modified Ile166Cys (1.6 M) in the absence (thin solid line) or presence of 0.08 M (thin dotted line), 0.15 M (thick dotted line), and 2.3 M (thick solid line) Fru-Val.
the attachment of acrylodan at Cys256 may have resulted in a structurally unstable protein. Trp256Cys-Ac was therefore ignored for further analysis and we focused on Ile166Cys-Ac. 3.5. FA fluorescent measurement using acrylodan-conjugated Ile166Cys FABP As with the autofluorescence of unconjugated Ile166Cys, the spectral properties of Ile166Cys-Ac also changed as a result of binding to the ligand Fru-Val (Fig. 4), presumably as a result of the protein’s conformational changes. A progressive increase in the fluorescence intensity at 520 nm was observed with increasing Fru-Val concentration between 17 nM and 2.5 M and becoming saturated above 2.5 M (Fig. 5). The relative intensity change of Ile166Cys-Ac in the presence of 2.3 M Fru-Val FABP was about 16-fold higher than that of Trp-related autofluorescence at 336 nm of the unconjugated Ile166Cys FABP. No changes in fluorescence intensity were detected with the addition
1937
of up to 3.7 M -Fru-Lys (Fig. 5), indicating that Ile166CysAc maintained the specificity for ␣-substituted fructosyl amino acids that was found in the wild-type protein. The sensitivity of acrylodan conjugated-FABP was eight times higher than Trp-related autofluorescence, allowing us to detect very low Fru-Val concentrations via differences in fluorescence intensity. The minimum Fru-Val concentration detectable using Ile166Cys-Ac was approximately 17 nM, which is over 100-fold more sensitive than enzyme-based detection systems (Tsugawa et al., 2000, 2001; Ogawa et al., 2002; Sakaguchi et al., 2003). Its high sensitivity and specificity for ␣-substituted fructosyl amino acids makes Ile166Cys-Ac ideally suited for the measurement of HbA1c, a major marker of diabetes. HbA1c is a hemoglobin molecule in which the N-terminal valine residue of the  subunit is glycated by glucose. The conventional biochemical measurement systems for HbA1c involve an initial proteolytic digestion to liberate Fru-Val, which is subsequently measured by the enzyme fructosyl amine oxidase (FAOD) (Tsugawa et al., 2000, 2001; Ogawa et al., 2002; Sakaguchi et al., 2003; Hirokawa et al., 2004; Sakurabayashi et al., 2003). However, currently isolated FAODs are unable to completely discriminate Fru-Val from -Fru-lys, which is the major fructosyl amino acid derived from glycated albumin. The application of FAOD for HbA1c measurement therefore requires the removal of serum proteins prior to the enzyme digestion process. The ability of Ile166Cys-Ac to completely discriminate between Fru-Val and -Fru-lys makes such a separation process unnecessary. The application of Ile166Cys-Ac as a novel fluorescent sensor element may provide a new approach to the diagnostic monitoring of HbA1c. We are now further investigating FABPs potential applications in HbA1c diagnosis, optimizing the proteolytic digestion conditions for HbA1c detection from real whole blood as well as looking at the possibility to detect undigested HbA1c. FABP-based sensors for HbA1c in direct and homogenous sensing systems could be constructed by combining several different techniques and methods, such as fluorophore-modification, fluorescence resonance energy transfer, optical fibers, quartz crystal microbalance, and surface plasmon resonance. 4. Conclusion We created a 3D structural model of FABP and used it to carry out protein engineering and conjugation with an environmentally sensitive fluorophore. The resulting FABP-based sensor is over 100-fold more sensitive than enzyme-based detection systems and is capable of detecting very low Fru-Val concentrations (approx. 17 nM). The conjugated FABP is also able to completely discriminate between Fru-Val and -Fru-lys. The high sensitivity and specificity for ␣-substituted fructosyl amino acids demonstrated for the fluorophore-modified FABP makes the new sensing system ideally suited for the measurement of HbA1c that is a major marker of diabetes.
Fig. 5. Correlation between FA concentration and fluorescence intensity of acrylodan-modified FABP. The maximal fluorescence intensity changes of purified acrylodan-modified Ile166Cys (0.6 M) were determined at different FruVal (䊉) and -Fru-Lys () concentrations.
Acknowledgement This work is partly supported by the Grant-in-Aid for the 21st Century COE “Future Nano-Materials” from the Ministry
1938
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
of Education, Culture, Sports, Science and Technology (MSXT) of Japan. References Baek, C.H., Farrand, S.K., Lee, K.E., Park, D.K., Lee, J.K., Kim, K.S., 2003. J. Bacteriol. 185, 513–524. Baker, J.R., Zyzak, D.V., Thorpe, S.R., Baynes, J.W., 1993. Clin. Chem. 39, 2460–2465. Davies, T.G., Hubbard, R.E., Tame, J.R., 1999. Protein Sci. 8, 1432–1444. De Lorimier, R.M., Smith, J.J., Dwyer, M.A., Looger, L.L., Sali, K.M., Paavola, C.D., Rizk, S.S., Sadigov, S., Conrad, D.W., Loew, L., Hellinga, H.W., 2002. Protein Sci. 11, 2655–2675. Dwyer, M.A., Hellinga, H.W., 2004. Curr. Opin. Struct. Biol. 14, 495–504. Finot, P.A., Viani, R., Bricout, J., Mauron, J., 1969. Experientia 25, 134–135. Gerhardinger, C., Taneda, S., Marion, M.S., Monnier, V.M., 1994. J. Biol. Chem. 269, 27297–27302. Hirokawa, K., Nakamura, K., Kajiyama, N., 2004. FEMS Microbiol. Lett. 235, 157–162. Hudge, J.E., Rist, C.E., 1953. J. Am. Chem. Soc. 75, 316–322. Jacobson, B.L., He, J.J., Vermersch, P.S., Lemon, D.D., Quiocho, F.A., 1991. J. Biol. Chem. 266, 5220–5225. Kang, C.H., Shin, W.C., Yamagata, Y., Gokcen, S., Ames, G.F., Kim, S.H., 1991. J. Biol. Chem. 266, 23893–23899. Kim, K.S., Baek, C.H., Lee, J.K., Yang, J.M., Farrand, S.K., 2001. Mol. Plant Microbe Interact. 14, 793–803. Ladenson, J.H., Chan, K.M., Kilzer, P., 1985. Clin. Chem. 31, 1060–1067. Ogawa, K., Stollner, D., Scheller, F., Warsinke, A., Ishimura, F., Tsugawa, W., Ferri, S., Sode, K., 2002. Anal. Bioanal. Chem. 373, 211–214.
Oh, B.H., Kang, C.H., De Bondt, H., Kim, S.H., Nikaido, K., Joshi, A.K., Ames, G.F., 1994. J. Biol. Chem. 269, 4135–4143. Okumoto, S., Looger, L.L., Micheva, K.D., Reimer, R.J., Smith, S.J., Frommer, W.B., 2005. Proc. Natl. Acad. Sci. USA 102, 8740–8745. Powlowski, J., Sahlman, L., 1999. J. Biol. Chem. 274, 33320–33326. Prendergast, F.G., Meyer, M., Carlson, G.L., Iida, S., Potter, J.D., 1983. J. Biol. Chem. 258, 7541–7544. Sakaguchi, A., Tsugawa, W., Ferri, S., Sode, K., 2003. Electrochemistry 71, 422–445. Sakaguchi, A., Ferri, S., Sode, K., 2005. BBRC 336, 1074–1080. Sakurabayashi, I., Watano, T., Yonehara, S., Ishimaru, K., Hirai, K., Komori, T., Yagi, M., 2003. Clin. Chem. 49, 269–274. Sode, K., Takahashi, Y., Ohta, S., Tsugawa, W., Yamazaki, T., 2001. Anal. Chim. Acta 435, 151–156. Sode, K., Ohta, S., Yanai, Y., Yamazaki, T., 2003. Biosens. Bioelectron. 18, 1485–1490. Sun, Y.J., Rose, J., Wang, B.C., Hsiao, C.D., 1998. J. Mol. Biol. 278, 219– 229. Tam, R., Saier Jr., M.H., 1993. Microbiol. Rev. 57, 320–346. Tsugawa, W., Ishimura, F., Ogawa, K., Sode, K., 2000. Electrochemistry 68, 869–871. Tsugawa, W., Ogawa, K., Ishimura, F., Sode, K., 2001. Electrochemistry 69, 973–975. Wandersman, C., Schwartz, M., Ferenci, T., 1979. J. Bacteriol. 140, 1–13. Wolf, A., Shaw, E.W., Oh, B.H., De Bondt, H., Joshi, A.K., Ames, G.F., 1995. J. Biol. Chem. 270, 16097–16106. Yamazaki, T., Ohta, S., Yanai, Y., Sode, K., 2003. Anal. Lett. 36, 73– 87. Yatscoff, R.W., Tevaarwerk, G.J., Macdonald, J.C., 1984. Clin. Chem. 30, 446–449.
Novel fluorescent sensing system for ␣-fructosyl amino acids based on engineered fructosyl amino acid binding protein Akane Sakaguchi, Stefano Ferri, Wakako Tsugawa, Koji Sode ∗ Department of Biotechnology, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan Received 20 April 2006; received in revised form 10 August 2006; accepted 14 August 2006 Available online 2 October 2006
Abstract A novel fluorescent sensing system for ␣-glycated amino acids was created based on fructosyl amino acid binding protein (FABP) from Agrobacterium tumefaciens. The protein was found to bind specifically to the ␣-glycated amino acids fructosyl glutamine (Fru-Gln) and fructosyl valine (Fru-Val) while not binding to -fructosyl lysine. An Ile166Cys mutant of FABP was created by genetic engineering and modified with the environmentally sensitive fluorophore acrylodan. The acrylodan-conjugated mutant FABP showed eight-fold greater sensitivity to Fru-Val than the unconjugated protein and could detect concentrations as low as 17 nM, making it over 100-fold more sensitive than enzyme-based detection systems. Its high sensitivity and specificity for ␣-substituted fructosyl amino acids makes the new sensing system ideally suited for the measurement of hemoglobin A1c (HbA1c), a major marker of diabetes. © 2006 Published by Elsevier B.V. Keywords: Diabetes; Hemoglobin A1c (HbA1c); Sensing system; ␣-Fructosyl amino acid; Periplasmic binding proteins (PBPs); Fructosyl amino acid binding protein (FABP)
1. Introduction Periplasmic binding proteins from Gram-negative bacteria (bPBPs) serve as essential primary receptors for the transport of carbohydrates, amino acids, anions, metal ions, and peptides (Wandersman et al., 1979; Oh et al., 1994; Powlowski and Sahlman, 1999; Jacobson et al., 1991; Davies et al., 1999). bPBPs are monomeric and consist of two domains linked by a hinge region. The ability of bPBP to undergo drastic conformational changes upon binding a target molecule has led to the design of several novel biosensing systems (Dwyer and Hellinga, 2004). A number of bPBPs have recently been successfully used as scaffolds for the development of fluorescent nanosensors by modification with appropriately positioned fluorescent probes (De Lorimier et al., 2002). bPBPs labeled with environmentally sensitive fluorophores undergo changes in fluorescence intensity or emission spectra when binding with their corresponding target molecules. This principle is not only limited to in vitro monitoring, but has also recently been applied to an in vivo FRET
∗
Corresponding author. Fax: +81 42 388 7072. E-mail address: [email protected] (K. Sode).
0956-5663/$ – see front matter © 2006 Published by Elsevier B.V. doi:10.1016/j.bios.2006.08.022
nanosensor based on a fluorescence protein fused to bPBP for glutamate (Okumoto et al., 2005). Although several attempts have been reported on the designing of on-demand substratebinding proteins, the development of bPBP-based sensing systems is strongly dependent on the natural diversity of available bPBPs that can be produced by recombinant technology. To develop novel diagnostic kits for monitoring glycemic control of diabetes patients, we have been engaged in the construction of novel biosensing systems for glycated proteins, such as the glycated hemoglobin HbA1c (Ladenson et al., 1985) and glycated albumin (Yatscoff et al., 1984), based on enzymes (Tsugawa et al., 2000, 2001; Ogawa et al., 2002; Sakaguchi et al., 2003) and molecularly imprinted polymers (Sode et al., 2001; Sode et al., 2003; Yamazaki et al., 2003). Protein glycation is a series of non-enzymatic reactions of reducing sugars with free amino groups on proteins (Hudge and Rist, 1953). Glycation starts with a reversible reaction to produce a Schiff base intermediate that can then undergo an essentially irreversible Amadori rearrangement to produce an Amadori product, such as fructosyl amine. The biochemical monitoring of protein glycation based on the measurement by fructosyl amine oxidase (FAOD) of fructosyl amino acids released by proteolytic digestion has recently been
1934
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
reported (Hirokawa et al., 2004; Sakurabayashi et al., 2003). FAODs, with Km values generally of 10−3 to 10−2 M, catalyze the oxidative degradation of fructosyl amino acids. Since glycated proteins, such HbA1c and glycated albumin, cannot be substrates for FAODs, these glycated proteins are subjected to proteolytic digestion prior to reaction with FAODs. The liberated the fructosyl amino acids, N␣ -substituted fructosyl-valine (FruVal) and N -substituted fructosyl-lysine (-Fru-Lys) released by protease digestion of HbA1c and glycated albumin, respectively, can then be detected enzymatically. An FAOD that reacts specifically with Fru-val might therefore be an ideal component of an HbA1c enzyme sensor. We reported an amperometric sensor system for HbA1c detection employing ␣-fructosyl amino acid specific FAOD (Sakaguchi et al., 2003). Unfortunately, a large quantity of FAOD is required for the construction of a biochemical monitoring kit for HbA1c due to the low concentration of target molecules (10−4 M in human blood) and the enzyme’s relatively high Km value. Considering that bPBPs generally show affinities (Kd generally 10−7 to 10−6 M) that are over 1000-fold higher than most enzymes, the application of bPBP for glycated protein sensing may provide a highly sensitive and effective system. We have recently succeeded in the identification and recombinant production of bPBP for fructosyl amino acid (Sakaguchi et al., 2005). This was the first report demonstrating bPBP conformational change resulting from binding to fructosyl amino acid, which was monitored by autofluorescence. Furthermore, this protein is the only fructosyl amino acid binding protein (FABP) that has been prepared by recombinant DNA technology. The current paper reports the construction of a novel fluorescent sensing system for fructosyl amino acid based on genetically engineered FABP. We found that the FABP is specific for ␣-glycated amino acids over -glycated amino acid. This selectivity is beneficial for applying FABP in HbA1c sensing to distinguish the Fru-Val produced from the degradation of HbA1c from the -Fru-Lys resulting from the degradation of glycated albumin. We also constructed a 3D structural model of our novel FABP to determine the ideal locations to introduce environmentally sensitive fluorophores to improve the sensitivity of the sensing system. 2. Materials and methods 2.1. Chemicals, enzymes, and bacterial strain Fru-Gln, Fru-Val, and -Fru-Lys were synthesized from d-glucose with l-glutamine, l-valine, and Na-Boc-l-lysine, respectively, by the method of Finot (Finot et al., 1969), with minor modifications. The purity of synthesized compounds was determined by TLC and by the alkaline-NBT method (Baker et al., 1993). PfuUltra High-Fidelity DNA polymerase was purchased from Stratagene (CA, USA). Restriction endonucleases were purchased from Takara (Kyoto, Japan). Acrylodan (6-acryloyl-2-dimethylaminonaphthalene) was from Molecular Probes Invitrogen Detection Technologies (CA, USA). All other chemicals were of reagent grade.
Fig. 1. Ribbon depiction of the closed (left) and open (right) forms of the predicted 3D structure of FABP based on Gln-BP (pdb:1WND). Tryptophan residues are shown in ball and stick style (pink); the upper residue corresponds to Trp256, which was selected for mutagenesis. The residues predicted to make up the substrate-binding domain (Met52, Ala56, Asp71, Phe96, Ala113, Ala114, Gly116, Arg121, Leu163, Leu200 and Asp201) are shown in wire frame style (yellow). Ile 166, selected for substitution to Cys, is shown in stick style (green). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)
2.2. Construction of FABP cysteine mutants A 3D structural model of fructosyl amino acid binding protein (FABP) from Agrobacterium tumefaciens EHA (Rhizobium radiobacter EHA) (Sakaguchi et al., 2005) was predicted by the molecular operating environment (MOE) software (Chemical Computing Group, Montreal, Canada) based on the open form structure of glutamine binding protein (Gln-BP, pdb:1WND, (Sun et al., 1998)) (Fig. 1). Based on the 3D model, residues Ile166 and Trp256 were selected for substitution to cysteine. Site-directed mutagenesis was carried out on the FABP expression vector pESA (Sakaguchi et al., 2005) by the QuikChange method (Strategene, CA, USA) using the following mutagenic primer pairs: (i) 5 -GCACGCTGCAGGAATGCTATGCCGAGAAG-3 and 5 -CTTCTCGGCATAGCATTCCTGCAGCGTGC-3 for the Ile166Cys mutation, and (Prendergast et al., 1983) 5 TGCAGGACGGCACCTGCAAGAAGCTTCATG-3 and 5 GAAGCTTCTTGCAGGTGCCGTCCTGCATGG-3 for the Trp256Cys mutation (mutation sites are underlined). The sequences of the created mutants were confirmed using the ABI Prism BigDye Terminator Cycle Sequencing Kit v3.0 on an ABI Prism 3100 Genetic Analyzer (Applied Biosytems, CA, USA). 2.3. Autofluorescence monitoring Wild-type (WT) and mutant (Ile166Cys and Trp256Cys) FABPs were expressed in E. coli and purified as described previously (Sakaguchi et al., 2005). The binding ability of purified FABPs were then analyzed by monitoring autofluorescence and dissociation constants (Kd ) were calculated by double reciprocal Klotz plot using the relative fluorescence intensity change
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
at each substrate concentration. The relative intensity changes (F/F0 ) were calculated as the percentage of the total intensity in the absence of substrate and used to evaluate the FABP autofluorescence changes. 2.4. Acrylodan modification of FABP mutants Purified mutants were modified with the environmentally sensitive fluorophore acrylodan as described in the Molecular Probes manual for thiol-reactive probes. Briefly, each purified FABP mutant was incubated at a concentration of 20 M in 10 mM potassium phosphate buffer, pH 7.0, containing 300 M acrylodan, for 8 h at 4 ◦ C. The unconjugated acrylodan was removed by using Microcon YM-10 centrifugal filter units (10 kDa MW cut-off, Millipore) at 5000 × g. 2.5. Characterization of fluorescently labeled FABP The fluorescence of acrylodan-conjugated FABP mutants was monitored using the same system as autofluorescence monitoring. Slits of excitation and emission monochromators were set at 5 and 3 nm, respectively. The excitation wavelength was 390 nm. Maximum intensity was measured after incubating acrylodan-conjugated FABP mutants at a final concentration of 0.6 M with each substrate for 2 min at 25 ◦ C. 3. Results and discussion 3.1. Substrate specificity of FABP The drastic conformational changes that FABP undergoes upon binding its ligand can be monitored by measuring tryptophan-related protein autofluorescence (Sakaguchi et al., 2005). The substrate specificity of FABP binding was investigated by measuring autofluorescence in the presence of different potential ligands. Based on both Kd values and maximum relative intensity changes, FABP binds preferentially to Fru-Gln, resulting in an 8.2% increase in autofluorescence intensity and a calculated Kd value of 0.2 M. FABP also binds to Fru-Val with very high affinity, resulting in a 3.9% autofluorescence increase with a Kd value of 0.6 M. Binding of glucose, fructose, glutamine, and valine could not be detected. Remarkably, incubation of FABP with -Fru-Lys also failed to produce any detectable conformational changes. The socA gene encoding FABP was originally identified as a component of the soc operon (Baek et al., 2003) in A. tumefaciens, responsible for the assimilation of the opine santhopine, a common name for Fru-Gln (Kim et al., 2001). Considering the putative metabolic role of FABP, the protein’s preference for FruGln is not surprising. Our results showing that FABP recognizes both Fru-Gln and Fru-Val with high affinity while not recognizing -Fru-Lys indicates that the protein is specific for ␣-fructosyl amino acids. Gerhardinger et al. previously reported the isolation of an -fructosyl amino acid-specific Amadori product binding protein from the fructosyl amino acid-utilizing soil bacterium Pseudomonas sp. (Gerhardinger et al., 1994). Our FABP is a unique binding protein that can recognize specifically Fru-Val
1935
and can discriminate -Fru-Lys, which is a very desirable property for a component an HbA1c monitoring system. 3.2. 3D model prediction of FABP Secretion into the periplasmic space leads to the cleavage of FABPs N-terminal 35-residue signal sequence to produce a monomeric peptide composed of 242 amino acids (Sakaguchi et al., 2005). The absence of cysteine residues in FABP allows the convenient and specific fluorophore modification via the formation of thiol bond by introducing a unique Cys residue. We created a 3D structural model of FABP to help us determine the ideal locations to introduce such cysteine substitutions. A previous amino acid sequence homology analysis indicated that FABP shares a conserved structural scaffold with cluster 3 binding proteins (Sakaguchi et al., 2005). Particularly high similarities were observed with polar amino acid binding proteins, such as histidine binding proteins (Oh et al., 1994; Wolf et al., 1995), lysine/arginine/ornithine binding protein (Kang et al., 1991) and Gln-binding protein (Sun et al., 1998) from E. coli and Salmonella typhimurium. FABP contains two signature sequences within the N-terminal domain, residues 69–80 and 101–115, that are characteristic of cluster 3 binding proteins (Tam and Saier, 1993). Furthermore, the residues predicted to be at the ligand recognition site are almost identical to those of other cluster 3 binding proteins. SocA is therefore expected to have a 3D structure similar to those of other cluster 3 binding proteins. In addition to the high similarity between FABP and Gln-BP (27% identical, 40% similar), secondary structure predictions also show that the hinge region of FABP is of the same length as Gln-BP and shorter than those of the other polar amino acid-BPs. Gln-BP was therefore used as a template to create a 3D structural model of FABP. As other cluster 3 binding proteins, the 3D model of FABP shows two distinct structural domains with the ligand-binding site located at their interface (Fig. 1). Met52, Ala56, Asp71, Phe96, Ala113, Ala114, Gly116, Arg121, Leu163, Leu200 and Asp201 were predicted by homology alignment with GlnBP to be residues interacting with the ligand. Because modification of these residues would likely adversely affect ligand binding, they were avoided as candidate sites for fluorophore labeling. The autofluorescence from FABPs two Trp residues, Trp256 and Trp263, changes upon binding Fru-Gln in a concentrationdependent manner (Sakaguchi et al., 2005). These Trp residues were therefore considered potential locations for introducing an environmental sensitive fluorophore. Trp263 was avoided because the predicted 3D model shows that it is located near the substrate-binding site. Trp256 is predicted to be located in the protein scaffold of one of the two major domains and was therefore selected for substitution to Cys for subsequent fluorophore modification. Ile166, located close to the active site without contributing to ligand binding, was also chosen for substitution to Cys and fluorophore modification. 3.3. Amino acid substitution of FABP We created the Ile166Cys and Trp256Cys mutants of FABP and expressed them in E. coli. As with the wild-type protein,
1936
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
Fig. 2. Effect of Fru-Val binding on the autofluorescence emission spectra of FABP mutants. The fluorescence spectra of purified Trp256Cys and Ile166Cys (1.6 M) in the absence (dotted line) and presence of 1.5 M Fru-Val. The measurements were carried out in 10 mM MOPS buffer (pH 7.0) at 25 ◦ C. The results are expressed as arbitrary units (arb.U).
these engineered FABP mutants were recovered in soluble form from the periplasmic space. Spectroscopic analysis of Ile166Cys and Trp256Cys FABPs both showed autofluorescence maxima of 335 nm with excitation at 295 nm (Fig. 2), as was observed with the wild-type protein. The fluorescence intensity of Ile166Cys was approximately 53 arbitrary units (arb.U), almost identical to that of the wild-type protein. The fluorescence intensity of Trp256Cys (approximately 11 arb.U) was greatly reduced to less than 25% of the wild-type level, which is not surprising considering that we are measuring Trp-related autofluorescence. Both mutant proteins showed similar changes in autofluorescence upon addition of an equimolar amount of the ligand Fru-Val (Fig. 2). Addition of higher concentrations of FruVal resulted in an almost 3% increase in autofluorescence of Ile166Cys and a 15% increase that of Trp256Cys (Fig. 3). The wild-type protein had previously shown an intermediate relative intensity increase of 6% (Sakaguchi et al., 2005). The fluorescence intensity of non-ligand-binding Trp256Cys FABP is fourto five-fold lower than the wild type and Ile166Cys mutant, indicating that Trp256 was the major contributor to autofluorescence. However, the absolute changes in fluorescence upon ligand binding are 2.5 arb.U, 2.2 arb.U, and 1.7 arb.U for wildtype, Trp256Cys, and Ile166Cys, respectively (data not shown). This suggests that the major contribution of Trp256 to autoflu-
Fig. 3. Correlation between Fru-Val concentration and autofluorescence. The maximal autofluorescence intensity changes were determined at different FruVal concentrations under the same conditions as Fig. 2.
orescence changes only slightly, if at all, upon ligand binding. This may explain why Trp256Cys undergoes such a large relative fluorescence change upon addition of Fru-Val, although we cannot rule out other effects such as unexpected structural changes. From the dependence of relative autofluorescence change on ligand concentration (Fig. 3), Kd values for Fru-Val binding to Ile166Cys and Trp256Cys FABPs were calculated to be 2.8 and 1.4 M, respectively. Although these values are two- to fourfold higher than for Fru-Val binding to wild-type FABP, they still retain their ligand-induced bending motion and a ligand binding ability that is several hundred times higher than enzymes, thus remaining ideal biosensor elements for HbA1c. These results also support the validity of our structural models. 3.4. Characterization of fluorescence-labeled FABP Ile166Cys and Trp256Cys FABPs were then modified with the environmentally sensitive fluorophore acrylodan to investigate their potential as elements for fluorescent sensors. Acrylodan fluorescence emission is highly sensitive to its local environment when bound to protein, and exhibits changes in both intensity and emission wavelength that reflect the effective dielectric constant of the environment around the fluorophore. The emission spectrum of acrylodan-conjugated Ile166Cys (Ile166Cys-Ac) shows a single emission peak at 520 nm with excitation at 390 nm (Fig. 4). Acrylodan-conjugated Trp256Cys (Trp256Cys-Ac) showed two emission peaks, at 456 and 472 nm (data not shown). However the fluorescence rapidly decreased after fluorophore modification and we were unable to obtain a stable fluorescence spectrum for Trp256Cys-Ac to investigate its ligand binding ability. The fact that Ile166Cys-Ac shows a stable fluorescence signal during the spectroscopic analysis indicates that the fluorescence character of acrylodan itself is not responsible for the instability of Trp256Cys-Ac fluorescence under our assay conditions. Furthermore, considering that unconjugated Trp256Cys can be prepared as a pure soluble protein in similar quantities as wild-type and Ile166Cys FABPs, the amino acid substitution has no detectable effect on the protein’s structural stability. Although Trp256Cys is a stable protein,
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
Fig. 4. Effect of Fru-Val binding on the fluorescence emission spectra of acrylodan-modified FABP. The fluorescence spectra of acrylodan-modified Ile166Cys (1.6 M) in the absence (thin solid line) or presence of 0.08 M (thin dotted line), 0.15 M (thick dotted line), and 2.3 M (thick solid line) Fru-Val.
the attachment of acrylodan at Cys256 may have resulted in a structurally unstable protein. Trp256Cys-Ac was therefore ignored for further analysis and we focused on Ile166Cys-Ac. 3.5. FA fluorescent measurement using acrylodan-conjugated Ile166Cys FABP As with the autofluorescence of unconjugated Ile166Cys, the spectral properties of Ile166Cys-Ac also changed as a result of binding to the ligand Fru-Val (Fig. 4), presumably as a result of the protein’s conformational changes. A progressive increase in the fluorescence intensity at 520 nm was observed with increasing Fru-Val concentration between 17 nM and 2.5 M and becoming saturated above 2.5 M (Fig. 5). The relative intensity change of Ile166Cys-Ac in the presence of 2.3 M Fru-Val FABP was about 16-fold higher than that of Trp-related autofluorescence at 336 nm of the unconjugated Ile166Cys FABP. No changes in fluorescence intensity were detected with the addition
1937
of up to 3.7 M -Fru-Lys (Fig. 5), indicating that Ile166CysAc maintained the specificity for ␣-substituted fructosyl amino acids that was found in the wild-type protein. The sensitivity of acrylodan conjugated-FABP was eight times higher than Trp-related autofluorescence, allowing us to detect very low Fru-Val concentrations via differences in fluorescence intensity. The minimum Fru-Val concentration detectable using Ile166Cys-Ac was approximately 17 nM, which is over 100-fold more sensitive than enzyme-based detection systems (Tsugawa et al., 2000, 2001; Ogawa et al., 2002; Sakaguchi et al., 2003). Its high sensitivity and specificity for ␣-substituted fructosyl amino acids makes Ile166Cys-Ac ideally suited for the measurement of HbA1c, a major marker of diabetes. HbA1c is a hemoglobin molecule in which the N-terminal valine residue of the  subunit is glycated by glucose. The conventional biochemical measurement systems for HbA1c involve an initial proteolytic digestion to liberate Fru-Val, which is subsequently measured by the enzyme fructosyl amine oxidase (FAOD) (Tsugawa et al., 2000, 2001; Ogawa et al., 2002; Sakaguchi et al., 2003; Hirokawa et al., 2004; Sakurabayashi et al., 2003). However, currently isolated FAODs are unable to completely discriminate Fru-Val from -Fru-lys, which is the major fructosyl amino acid derived from glycated albumin. The application of FAOD for HbA1c measurement therefore requires the removal of serum proteins prior to the enzyme digestion process. The ability of Ile166Cys-Ac to completely discriminate between Fru-Val and -Fru-lys makes such a separation process unnecessary. The application of Ile166Cys-Ac as a novel fluorescent sensor element may provide a new approach to the diagnostic monitoring of HbA1c. We are now further investigating FABPs potential applications in HbA1c diagnosis, optimizing the proteolytic digestion conditions for HbA1c detection from real whole blood as well as looking at the possibility to detect undigested HbA1c. FABP-based sensors for HbA1c in direct and homogenous sensing systems could be constructed by combining several different techniques and methods, such as fluorophore-modification, fluorescence resonance energy transfer, optical fibers, quartz crystal microbalance, and surface plasmon resonance. 4. Conclusion We created a 3D structural model of FABP and used it to carry out protein engineering and conjugation with an environmentally sensitive fluorophore. The resulting FABP-based sensor is over 100-fold more sensitive than enzyme-based detection systems and is capable of detecting very low Fru-Val concentrations (approx. 17 nM). The conjugated FABP is also able to completely discriminate between Fru-Val and -Fru-lys. The high sensitivity and specificity for ␣-substituted fructosyl amino acids demonstrated for the fluorophore-modified FABP makes the new sensing system ideally suited for the measurement of HbA1c that is a major marker of diabetes.
Fig. 5. Correlation between FA concentration and fluorescence intensity of acrylodan-modified FABP. The maximal fluorescence intensity changes of purified acrylodan-modified Ile166Cys (0.6 M) were determined at different FruVal (䊉) and -Fru-Lys () concentrations.
Acknowledgement This work is partly supported by the Grant-in-Aid for the 21st Century COE “Future Nano-Materials” from the Ministry
1938
A. Sakaguchi et al. / Biosensors and Bioelectronics 22 (2007) 1933–1938
of Education, Culture, Sports, Science and Technology (MSXT) of Japan. References Baek, C.H., Farrand, S.K., Lee, K.E., Park, D.K., Lee, J.K., Kim, K.S., 2003. J. Bacteriol. 185, 513–524. Baker, J.R., Zyzak, D.V., Thorpe, S.R., Baynes, J.W., 1993. Clin. Chem. 39, 2460–2465. Davies, T.G., Hubbard, R.E., Tame, J.R., 1999. Protein Sci. 8, 1432–1444. De Lorimier, R.M., Smith, J.J., Dwyer, M.A., Looger, L.L., Sali, K.M., Paavola, C.D., Rizk, S.S., Sadigov, S., Conrad, D.W., Loew, L., Hellinga, H.W., 2002. Protein Sci. 11, 2655–2675. Dwyer, M.A., Hellinga, H.W., 2004. Curr. Opin. Struct. Biol. 14, 495–504. Finot, P.A., Viani, R., Bricout, J., Mauron, J., 1969. Experientia 25, 134–135. Gerhardinger, C., Taneda, S., Marion, M.S., Monnier, V.M., 1994. J. Biol. Chem. 269, 27297–27302. Hirokawa, K., Nakamura, K., Kajiyama, N., 2004. FEMS Microbiol. Lett. 235, 157–162. Hudge, J.E., Rist, C.E., 1953. J. Am. Chem. Soc. 75, 316–322. Jacobson, B.L., He, J.J., Vermersch, P.S., Lemon, D.D., Quiocho, F.A., 1991. J. Biol. Chem. 266, 5220–5225. Kang, C.H., Shin, W.C., Yamagata, Y., Gokcen, S., Ames, G.F., Kim, S.H., 1991. J. Biol. Chem. 266, 23893–23899. Kim, K.S., Baek, C.H., Lee, J.K., Yang, J.M., Farrand, S.K., 2001. Mol. Plant Microbe Interact. 14, 793–803. Ladenson, J.H., Chan, K.M., Kilzer, P., 1985. Clin. Chem. 31, 1060–1067. Ogawa, K., Stollner, D., Scheller, F., Warsinke, A., Ishimura, F., Tsugawa, W., Ferri, S., Sode, K., 2002. Anal. Bioanal. Chem. 373, 211–214.
Oh, B.H., Kang, C.H., De Bondt, H., Kim, S.H., Nikaido, K., Joshi, A.K., Ames, G.F., 1994. J. Biol. Chem. 269, 4135–4143. Okumoto, S., Looger, L.L., Micheva, K.D., Reimer, R.J., Smith, S.J., Frommer, W.B., 2005. Proc. Natl. Acad. Sci. USA 102, 8740–8745. Powlowski, J., Sahlman, L., 1999. J. Biol. Chem. 274, 33320–33326. Prendergast, F.G., Meyer, M., Carlson, G.L., Iida, S., Potter, J.D., 1983. J. Biol. Chem. 258, 7541–7544. Sakaguchi, A., Tsugawa, W., Ferri, S., Sode, K., 2003. Electrochemistry 71, 422–445. Sakaguchi, A., Ferri, S., Sode, K., 2005. BBRC 336, 1074–1080. Sakurabayashi, I., Watano, T., Yonehara, S., Ishimaru, K., Hirai, K., Komori, T., Yagi, M., 2003. Clin. Chem. 49, 269–274. Sode, K., Takahashi, Y., Ohta, S., Tsugawa, W., Yamazaki, T., 2001. Anal. Chim. Acta 435, 151–156. Sode, K., Ohta, S., Yanai, Y., Yamazaki, T., 2003. Biosens. Bioelectron. 18, 1485–1490. Sun, Y.J., Rose, J., Wang, B.C., Hsiao, C.D., 1998. J. Mol. Biol. 278, 219– 229. Tam, R., Saier Jr., M.H., 1993. Microbiol. Rev. 57, 320–346. Tsugawa, W., Ishimura, F., Ogawa, K., Sode, K., 2000. Electrochemistry 68, 869–871. Tsugawa, W., Ogawa, K., Ishimura, F., Sode, K., 2001. Electrochemistry 69, 973–975. Wandersman, C., Schwartz, M., Ferenci, T., 1979. J. Bacteriol. 140, 1–13. Wolf, A., Shaw, E.W., Oh, B.H., De Bondt, H., Joshi, A.K., Ames, G.F., 1995. J. Biol. Chem. 270, 16097–16106. Yamazaki, T., Ohta, S., Yanai, Y., Sode, K., 2003. Anal. Lett. 36, 73– 87. Yatscoff, R.W., Tevaarwerk, G.J., Macdonald, J.C., 1984. Clin. Chem. 30, 446–449.