A nanobiosensor based on hppd for mesotrione detection

2106 IEEE SENSORS JOURNAL, VOL. 15, NO. 4, APRIL 2015

A Nanobiosensor Based on4-Hydroxyphenylpyruvate Dioxygenase

Enzyme for Mesotrione DetectionPâmela Soto Garcia, Alberto Luís Dario Moreau, Jéssica Cristiane Magalhães Ierich, Ana Carolina Araujo Vig,

Akemi Martins Higa, Guedmiller S. Oliveira, Fábio Camargo Abdalla, Moema Hausen, and Fábio L. Leite

Abstract— The herbicide residue from intensive agriculturalactivity provokes environmental disturbances and human healthinjuries. Among the enzymatic disruptor herbicides, mesotrioneis able to inhibit 4-hydroxyphenylpyruvate dioxygenase (HPPD),which plays a key role in the carotenoid synthesis. Therefore,enzyme-based sensors are innovative options for monitoringherbicides used in agriculture. Compared to the standard sensors,biosensors have assorted advantages, such as practicality, quickresponse, low cost, and high sensitivity. A nanobiosensor wasdeveloped herein based on HPPD for mesotrione detection.Theoretically, the molecular docking and molecular dynamicssimulation estimated the interacting regions of HPPD withmesotrione. Experimentally, the atomic force microscope tip func-tionalization with HPPD immobilized in self-assembled mono-layers was confirmed by fluorescence microscopy and atomicforce spectroscopy. The cross-linker N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride was responsible for properlypreserving the enzyme on the tip. The nanobiosensor proposedhere was successfully able to detect mesotrione molecules. Sucheffectiveness in the development of nanobiosensors promisesreliable, precise, and low-cost techniques, which apply to a broadrange of issues, from ecology to medicine.

Index Terms— AFM, AFS, chemical functionalization,nanobiosensors, molecular docking, molecular dynamicssimulation, mesotrione, 4-hydroxyphenylpyruvate dioxygenase.

Manuscript received June 3, 2014; revised September 23, 2014 andOctober 17, 2014; accepted November 2, 2014. Date of publicationNovember 20, 2014; date of current version January 29, 2015. Thiswork was supported in part by CNPq (CNPq/INCT, 573742/2008-1), in partby FAPESP (FAPESP/INCT, 2008/57859-5, 2007/05089-9, 2010/00463-2,2010/04599-6, 2013/09746-5, 2013/21958-8, 2011/17840-6, 2014/12082-4),in part by nBioNet, and in part by CAPES (PNPD/20131505). The associateeditor coordinating the review of this paper and approving it for publicationwas Dr. Chang-Soo Kim.

P. S. Garcia, J. C. M. Ierich, A. C. A. Vig, A. M. Higa,G. S. Oliveira, M. Hausen, and F. L. Leite are with the NanoneurobiophysicsResearch Group, Department of Physics, Chemistry and Mathematics,Federal University of São Carlos, São Carlos 18052-780, Brazil (e-mail:[email protected]; [email protected]; [email protected]; [email protected]; [email protected];[email protected]; [email protected]).

A. L. D. Moreau is with the Department of Physics, Federal Institute ofEducation, Science and Technology of Itapetininga, Itapetininga 18202-000,Brazil, and also with the Nanoneurobiophysics Research Group, Departmentof Physics, Chemistry and Mathematics, Federal University of São Carlos,São Carlos 18052-780, Brazil (e-mail: [email protected]).

F. C. Abdalla is with the Laboratory of Structural and Functional Biology,Department of Biology, Federal University of São Carlos, São Carlos18052-780, Brazil (e-mail: [email protected]).

Color versions of one or more of the figures in this paper are availableonline at http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/JSEN.2014.2371773

I. INTRODUCTION

SENSORS are devices used for detection and measurementof physical properties [1]. There are many different sorts

of sensors, and the understanding of their mechanisms requiresa multidisciplinary knowledge [2]. When biomolecules areused to measure relevant biological material, the sensoris defined as a biosensor [3], [4]. Those biomolecules areusually immobilized through chemical or physical transducers,creating a surface that makes possible the direct measurementof a specific molecule [5], [6]. If the biosensor functions onthe nanometric scale, it is classified as nanobiosensor [7].

A nanobiosensor is obtained by the functionalization ofAtomic Force Microscope (AFM) tips and canmeasure forcesat an atomic scale [8], [9]; the application of those tipson the development of nanobiosensors is known as AtomicForce Spectroscopy (AFS) [10]. To build a functionalizedAFM tip requires the chemical modification of its surface.Such procedure provides three essential benefits: (i) a highsensitivity device; (ii) detections at the molecular level; and(iii) simulation of a mimetic microenvironment [11], [12].

The functionalization process requires a previous study ofthe set of molecules involved, while the availability of theactive site and substrate orientation are important parametersfor the AFM measurements [11], [13], [14]. In this context,computer simulation is an easy and economical approach toestimate intermolecular interactions. Therefore, the combina-tion of theoretical and experimental methodologies providesboth macro and micro scale perspectives to the experiments.

Herbicides act on specific metabolic pathways in plants,as inhibitors of the synthesis of carotenoids and amino acids.This triggers photosynthesis failure, which leads to plantstarvation and death [15]. Herbicides have a deep impact onhuman health and the environment, and the monitoring of suchagrochemicals in order to minimize their impact and invitethe development of alternative weed control procedures inintensive agriculture is crucial. Plants and animals can sharetargets of homoplasic molecules, and the metabolization ofherbicide molecules can generate secondary metabolics. Thesecan be more harmful than the herbicide itself, in addition tothe surfactant products that compose the herbicide formula,which have been shown to be extremely toxic to animals byingestion, inhalation or contact [16]. Among assorted herbicidetypes, few act by inhibiting specific enzymes. This type of

1530-437X © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

GARCIA et al.: NANOBIOSENSOR BASED ON 4-HPPD ENZYME 2107

Fig. 1. Conversion of HPP to homogentisate, catalyzed by HPPD enzyme(depicted on green elipse) at the tyrosine pathway [adapted from literature 21].

inhibition process is a reaction between a molecule and anenzyme, which forms a complex that blocks the enzymefunction [16], [17].

The tyrosine catabolism is a common process inplants and animals and requires key enzymes, such as4-Hydroxyphenylpyruvate dioxygenase (HPPD) that catalyzeHydroxyphenylpyruvate (HPP) to homogentisate (Fig. 1),which participate in the energetic metabolism [18]–[20].In animals, the lack of such conversion can adversely affectdifferent metabolic pathways and cause severe metabolicdisorders, such as Tyrosinemia and Hawkisinuria [18].In plants, the inhibition of this sequence of metabolic eventsleads to lower levels of chlorophyll and carotenoids. Thelatter protect the chlorophyll from excess light. Becauseof those inhibitory events and the fact that weeds growmuch more quickly than the economic plant of interest, theherbicide causes leaves of the plant to blanch and, in a coupleof weeks, causes all photosynthesis to cease [21].

Several herbicides promote HPPD inhibition. Among them,the triketone herbicide mesotrione (C14H13NO7S) is usedon pre- and post-emergency crops. After application, it isabsorbed by leaves and roots and is readily translocated tothe plant vascular system [22]–[24]. The HPPD inhibitors areconsidered low risk pesticides.

Therefore, to contribute to advances in this researchfield, a nanobiosensor based on AFM functionalized tipswas developed using the techniques of AFS and AFM [3],[7], [11], [25] to perform mesotrione detection through theinhibition of the target enzyme, HPPD. In our research,the development of nanobiosensors focus on the study ofneurodegenerative diseases and on agrochemical detection,especially enzymatic inhibitor herbicides [15], [26]–[29].

Mesotrione identification by an AFM tip nanobiosensorfunctionalized with HPPD, studied hereunder controlledparameters, can improve the application of biotechnologyto environmental safety. All techniques and methodologiesdeveloped in this paper can be adapted to detect otherpotential contaminants, such as pesticides and trace metalsthat can harm the environment.

II. MATERIALS AND METHODS

A. Molecular Docking Parameters

The initial crystallography structure of the HPPD wasobtained from the online repository Protein Data Bank (PDB),ID: 3ISQ [30]. The PDB file was prepared for the dock-ing calculations. All missing residues were completed, andthe molecular structure was energy-minimized. Eight Mole-cular Dockings were used to determine the HPPD affinityregions with mesotrione by means of the AutoDock Tools,version 1.5.6 [31]. The grid boxes had two different sizes, inAngstrom (Å): grid boxes 1, 3, 5 and 7 had dimensions ofx = 56, y = 56 and z = 126, and grid boxes 2, 4, 6 and8 dimensions were x = 68, y = 66 and z = 66, following theprotocol used by Franca et al. [32]. This last group was anattempt of creating smaller boxes and obtaining more preciseresults. The Lamarckian Genetic Algorithm (LGA) was usedto fit the mesotrione molecule in the HPPD, which was setas a rigid structure, in a total of 100 runs for each docking.The program was also set to calculate the internal electrostaticenergy. The RMSD Cluster Tolerance was adjusted to a limitof 2.0 Å.

B. Molecular Dynamics SimulationFrom the eight Molecular Dockings obtained previously,

four systems were chosen based on their final scored energiesand cluster conformations on which to perform MolecularDynamics (MD). Nanoscale Molecular Dynamics – NAMDversion 2.9 [33] and Visual Molecular Dynamics – VMDversion 1.9.1 [34] were used to prepare the systems understudy. The main goal was to rearrange the atoms to minimizeand equilibrate the molecule energetically. The first step wasto balance the negative charge of the system. The final chargewas 9.66340 e−6 C. The next step was the minimizationof the system’s energies conducted in 100000 steps, total-ing 200 ps. The temperature adjustments were performed inthe NVT ensemble, at 298 K and 1.0 atm, using a Langevinthermostat. The cut-off distance of 16 Å was the sameused inOliveira et al. [35]. The equilibration of the systems followedthe minimization, with certain modified parameters in the NpTensemble. Pressure and temperature were controlled using theLangevin method at 1 atm and 310 K. The cut-off distanceof 12 Å was needed for accurate interaction contacts over thetrajectory. The electrostatic interactions were calculated withthe Particle Mesh Ewald (PME).

C. Experimental Procedure – MaterialsAccording to Orry and Abagyan, (2012) [36], the

homology protocol established a minimum of 25% sequencesimilarity between two molecules. In this case, the similaritybetween human and plant HPPD is 33% as performed onthe BLAST online platform [37]. Therefore, our modelpresents a moderate similarity that is feasible for use inthe mesotrione herbicide detection process. In addition,only the pure human HPPD enzyme was available forpurchase, while the plant one is commercially restricted.Consequently, in order to evaluate the nanobiosensorbehavior, the study was conducted using the human enzyme.


The pure human HPPD enzyme was acquired from AcrisAntibodies, Inc. (San Diego, CA), NCBI NP_001165464. Theenzyme was received lyophilized, and prior to experiments,it was reconstituted in Mili-Q water. The mesotrioneagrochemical was purchased from Chem Service (WestChester, PA) and diluted in pure-grade acetone, purchasedfrom Qhemis, Hexis Científica S/A (Indaiatuba, SP). Thefollowing pure-grade materials were purchased from Sigma-Aldrich (St. Louis, MO): monobasic sodium phosphate,dibasic sodium phosphate, sodium chloride, triethy-lamine (TEA), 3-aminopropyltriethoxysilane (APTES), andN-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochlo-ride (EDC). The glutaraldehyde (GLU) (as a 25% aqueoussolution) was purchased from Nuclear, CAQ (Diadema, SP),and the Casein (CAS) was obtained from skimmed milkpowder. The fluorochrome used was fluorescamine(C17H10O4; Sigma-Aldrich, USA). The AFM tips used weretriangular silicon nitride tips from NanoWorld-InnovativeTechnologies (Switzerland), model PNP-TR-20. The cantileverused had the following specifications: overall thickness:600 nm, length: 200 μm, width: 2 × 28 μm, resonancefrequency: 17 kHz, spring constant (k): 0.08 N/m, radius:<10 nm, coating (detector side): Cr/Au. The substrateused was a muscovite mica from Ted Pella. The AtomicForce Microscope was Veeco Multimode V with PicoForcemode coupled with a fluid cell. The AFM software wasNanoScope 7.0, and the Origin 6 software was usedfor statistical analysis. To obtain force curves, the AFMVeeco Multimode V with PicoForce mode package was used,in contact mode and in a liquid environment, using a fluid cell.To use this fluid cell, it was necessary to build a system on theAFM using a syringe with in and out pipes. The fluorescencemicroscope used to characterize the functionalized AFM tipswas a Leica DM 4000b, coupled with a digitalization imagesystem by a Leica DFC310 FX camera, and the softwareLeica FIM (Fluorescence Intensity Manager).

D. Experimental Procedure – Solution PreparationThe solutions were prepared as described: HPPD was

prepared using a dilution of 0.1 mg of enzyme in 5 mLof Mili-Q, then stored in small aliquots (300μL) atapproximately −20 °C; mesotrione was prepared by adilution of 0.0333 gin 50 mL of acetone PA ACS; phosphatebuffered saline (PBS) was prepared with monobasic sodiumphosphate, dibasic sodium phosphate and sodium chloride.The PBS solution buffer was at pH 7.2; PBS solutionwith HPPD was prepared with a solution of PBS solutionbuffer and HPPD enzyme (100 μL of enzyme in 5 mL ofPBS solution buffer); EDC with HPPD was prepared bydiluting 0.0013 g of EDC in 1 mL Mili-Q water, and 20 μLof this solution was added to 150 μL of HPPD enzyme insolution; casein solution was prepared by diluting 0.02 g ofskimmed milk powder in 1 mL of Mili-Q water.

E. Experimental Procedure – Methodologies for Nanobiosen-sor Assembly and Immobilization of Target Molecules

Based on Silva et al., 2013 [15] and Moreau, 2005 [38],the tip functionalization method used for the immobilization

of the HPPD enzyme by linker molecules was performedusing two distinct approaches. The first approach used APTES,TEA and GLU, and the second approach used APTES, TEA,EDC and CAS. The procedure described in this paper wasrelated to the second method, once the first method wasnot successful in the HPPD attachment to an AFM tip. Theprocedure steps can be summarized as follows: (i) the AFMtips were cleaned for 20 min using U.V. (240 nm, ProCleaner,Bioforce); (ii) the tips were exposed to the APTES andTEA vapors (40 μl) during 45 min in a reaction chamber(previously cleaned with nitrogen steam); (iii) the tips wereimmersed in a solution containing HPPD and EDC for 2 h at4 °C; (iv) three baths were prepared with fresh CAS solutionin Mili-Q water; (v) the AFM tips were immersed in theCAS solution and were incubated during 1 h at 37 °C; and(vi) a PBS bath solution injected in the AFM fluid cell formeasurements.

The substrate preparation procedure, which is made onmuscovite mica, was performed as follows: (i) the mica wascleaved with adhesive tape and cleaned in UV light (240 nm,ProCleaner, Bioforce) for 20 min; (ii) in a reaction chamber,cleaned with nitrogen, the mica was exposed to APTES andTEA at a ratio of 1:1 for 45 min; (iii) 200 μL of GLU wasdeposited on the mica surface for 10 minutes; (iv) after threeMili-Q baths, the herbicide mesotrione was added to the micasurface, during 25 min in a nitrogen chamber; (iv) the micawas immersed in a CAS solution at 37 °C during 1 h and thenimmersed in a Mili-Q bath; and (v) the mica was immersedin a PBS bath before being transferred to the AFM magneticholder for measurements.

F. Experimental Procedure – Atomic Force Spectroscopy

After the tip functionalization and mica placement in theAFM apparatus, the laser beam was adjusted and the tipapproached to the mica surface. The injected solutions andthe software Nanoscope Analysis were previously adjustedto carry out the force measurements. The nanobiosensorexperiments were performed in two steps: (i) nanobiosensormeasurement, using a PBS buffer solution injected in thefluid cell, and (ii) system inhibition, injecting PBS bufferwith HPPD enzyme in the fluid cell. When this solution withenzyme is injected in the system, it attaches to the herbicide atthe mica surface, avoiding the HPPD on the tip and interactingwith the sample, as shown in Fig. 2. This space can interactwith the HPPD on the tip, providing adhesion force values.The data were obtained from adhesion force measurements ofthe enzyme-herbicide interaction, performed on 10 differentsubstrate regions.

G. Experimental Procedure – Tip Characterization UsingFluorescence Microscopy

The fluorochrome-HPPD conjugation was measured as fol-lows: the functionalized tip was immersed for 5 min influorescamine-dimethyl sulfoxide solution; afterwards, the setwas bathed three times in Mili-Q water. The fluorescent imageswere obtained digitally and the software Leica Image Analysiswas used for processing and editing the images.


Fig. 2. (a) Herbicide-enzyme interaction; (b) enzyme being injected on thesystem; (c) enzyme-enzyme interaction and (d) enzyme-free spaces wherethere is no interaction with the herbicide molecules over the mica substrate.

Fig. 3. Four mesotrione molecules (in green) docked to the differentregions on the HPPD molecular structure in order to run MD simulation.Two cofactors are presented, Sodium and Chloride ions (in black and purple,respectively). The total energies (E values in kJ.mol−1) are represented foreach mesotrione position. The inset above shows structural formula of themesotrione.

III. RESULTS AND DISCUSSION

A. System Energy

The immobilization and stability of biomolecular systemson functionalized AFM tips is one concern in the design ofsensitive and selective biosensors [7], [15], [29]. As mentionedbefore, HPPD (PDB code: 3ISQ) [30] was chosen to act asa biologic sensor. To evaluate its behavior in an aqueoussolution, computational simulations (Molecular Docking andMolecular Dynamics - MD simulation) were performed tomonitor the HPPD fluctuations such as its interaction energieswith mesotrione herbicide. The most favorable dockedpositions scored are shown in Fig. 3. Four different confor-mations of the mesotrione on binding regions of the HPPDwere considered. By running MD simulations, the systems

TABLE I

INHIBITION COEFFICIENT (Ki) AND REFERENCE RMSD, AND TOTAL

ENERGIES (ET) FOR EACH SYSTEM OF THE MOLECULAR DOCKING

CALCULATIONS AFTER 5 ns OF MD SIMULATION

were energy minimized and energy equilibrated, in order toanalyze the fluctuations and mobility of the mesotrione-HPPDset in aqueous solution. The electrostatic and van der Waalsinteractions were estimated and calculated. The results after5 ns of MD simulation and the average energies are listed inTable I.

Energetically, the most interactive region of the HPPDenzyme was found in system 4, but this result cannot bedirectly related to experimental ones because the orientationof the HPPD on the tip must be considered, which depends onthe arrangement of multiple enzymes together. The goal of thiscomputational analysis is to provide an atomistic perspectiveon the binding regions of HPPD with mesotrione.

The parameters of inhibition coefficient (Ki) and root meansquare deviation (RMSD) were analyzed using MolecularDocking calculations for scoring HPPD binding sites withmesotrione. According to Franca et al. [32], the concentrationof the herbicide required to inhibit an enzyme activity isexpected to be lower for the most favorable binding region.In Table I, mesotrione has the lowest inhibition coefficient andfavorable interaction energies for systems 5 and 8. For thesesystems, the adhesion force for mesotrione was higher. Theanalysis of the Ki revealed that the inhibition coefficient washigh for systems 1 and 2; as a result, the clusters formed bythese systems were not suitable for biosensor requirements.Additionally, the RMSD results revealed a large value forsystems 1, 2 and 5 (5.99 Å, 4.16 Å and 3.37 Å, respectively)and a low value for system 8 (1.71 Å). Therefore, system 8 ismore interactive than the other ones, and the experimentalforce curve obtained with the AFM is strong when the HPPDis oriented to the substrate similar to system 8.

B. Root Mean Square Deviation – Protein Stability

Fig. 4 shows the computed RMSD for the 4 evaluatedsystems. The HPPD structural fluctuations were monitoredduring 5 ns of MD simulation in the presence of mesotrionemolecules. As shown, all systems have similar averageRMSD: 1.5 Å. The fluctuations can be attributed to the saltbridges and hydrogen bonds formed and broken over time.According to Franca et al. [32], charged amino acids such asARG, LYS, ASP and GLU located on the border can inducenew hydrogen bonds between water molecules and HPPD,causing small structural fluctuations. The amino acids that


Fig. 4. Structural fluctuation of the HPPD at the presence of mesotrioneherbicide during 5 ns.

TABLE II

NUMBER OF SALT BRIDGES AND HYDROGEN BONDS FORMED AFTER

5 ns OF MD SIMULATION

contribute more hydrogen bonds were ASP, GLU and LYS,while CYS, ILE and TRP do not have relevant contributionsto the total number of hydrogen bonds during the MDsimulation. Moreover, during all simulations, salt bridgenumbers from the HPPD were almost the same, and HPPDstructure was preserved. As shown in Table II, the numberof salt bridges remained constant during the MD simulation;thus, the enzymatic structure was not affected by the solvent.As a result, no effect was observed at the loops and sidechains, which have high RMSD values, and no denaturationwas detected on the HPPD enzyme structure.

Table II shows the number of salt bridges and hydrogenbonds at the beginning (0 to 2.5 ns) and at the end(2.5 to 5.0 ns) of the simulation. Both initial and late patternspresented similarities, which consequently revealed that theposition of the mesotrione has no influence on the HPPDenzyme structure. Finally, the computational results showedthat HPPD is stable enough to be used as a biosensor andhas specific interactive regions to mesotrione.

C. Support of Tip Functionalization byFluorescent Labeling

According to literature [39], fluorescence microscopy (FM)is usedto confirm the AFM tip functionalization. Thecombination of AFM and other techniques, such as confocallaser scanning microscopy and fluorescent imaging, provides abetter understanding of biological studies, enlarging the possi-bilities of investigation and giving more detailed information.

Fig. 5. The AFM tips observed by two microscopy techniques: brightfield (a, c) and fluorescent mode (b, d). The same tip is observed inboth techniques for the nonfunctionalized-control group (a, b) and for thefunctionalized one (c, d). The tips functionalized with HPPD and conjugatedwith fluorescamine (Si/HPPD-F) presented intense blue fluorescence while thenonfunctionalized ones remained dark in fluorescent mode (b). This qualitativeresult confirmed the HPPD presence on tips.

TABLE III

ADHESION FORCE (AF) OBTAINED WITH AFS EXPERIMENTS FOR

CONTROL (600 ADHESION FORCE CYCLES)

They may become important tools in medicine, detectingdiseases in early stages [40]. In this paper, the use of FMto detect HPPD confirmed the AFM tip functionalization.The images obtained by FM showed that the methodologywas effective in attaching the biomolecule to the tip (Fig. 5).Furthermore, other studies [41] suggest functionalizationevaluation by confocal microscopy and mediated by indirectfluorescent labeling to be an effective tool to scan and detectall labeling distribution on the tip surface at higher resolution.

D. Mesotrione Detection by AFM Tip Nanobiosensor

The first experimental data were obtained from controltips, organized as follows: (type 1) clean tips, without anyfunctionalization; (type 2) tips functionalized with APTES andTEA; (type 3) tips functionalized with APTES, TEA, EDCand CAS. These three control tips were used to perform forcemeasurements in the AFM liquid cell, over the sample withthe herbicide. The obtained adhesion force data were lowerthan expected, at values around 0.4 nN for measurements insolution [42], [43].

The force measurement characterizations of all control tipswere used as parameters for the nanobiosensor according tothe values shown in Table III. The type 2 tips showed highadhesion values, most likely due to the interaction between the


Fig. 6. Representative histograms for the sets of measurements to theNanobiosensor (AFnb) and System Inhibition (AFsi). The adjustment wasperformed by Gauss curve.

APTES and the herbicide. Compared to those with APTES, thetype 3 tips showed lower values because CAS favors the inhi-bition of the active sites of APTES, as recently reported [44].Limanskii [45] also performed a functionalization on siliconnitride AFM probes, using APTES vapors. In that work, theuse of the linking agent Disuccinimidyl suberate (DSS) wasfollowed by albumin attachment. The model proposed byLimanskii [45] and the one presented here share the use ofAPTES vapors to successfully induce modifications on the tip.

The nanobiosensor, developed through the functionaliza-tion of the HPPD enzyme on APTES, TEA, EDC andCAS, is expected to present a higher value of adhesioncompared to the control tests, specifically detecting the her-bicide mesotrione. The results presented here are in agree-ment with our previous experiments using nanobiosensorswith diclofop, atrazine and metsulfuron-methyl agrochem-icals [15], [26], [28]. The high adhesion values that wereobtained confirm the effectiveness of the functionalizationmethod under aqueous conditions. Therefore, the protocolestablished for the nanobiosensors is straightforward and canbe applied to assorted detections.

After obtaining the control data, the nanobiosensors weredeveloped and tested. First, a functionalization methodologywith APTES, TEA, GLU, and HPPD was tested (data notshown). However, this functionalization was not efficient anddid not provide good adhesion force values, which led to theconclusion that the biomolecule HPPD did not link properlyon GLU, and the active sites were most likely not in favorablepositions to link to the substrate.

Fig. 6 shows that the adhesion values ranged approximately1.5 nN and reached a recover of 63% to the nanobiosensorand 35% to the system inhibition. These values are includedin Table IV. Although the frequency changed, the adhesionforce remained the same, as expected. According to theproposed model, the system was probably inhibited due toHPPD linkage on mesotrione (Fig. 2d).

The method was evaluated by measuring the nanobiosensorforce value, which was two times higher than with thecontrol tips. This finding implies that the HPPD was properly

TABLE IV

ADHESION FORCE (AF) OBTAINED WITH AFS EXPERIMENTS FOR

NANOBIOSENSOR AND INHIBITION

orientated on the tip due to the EDC cross-linker, probablyexposing the interaction sites to mesotrione molecules.

All data sets presented demonstrate that the nanobiosensordeveloped here was effective for mesotrione detection. Theinhibition parameter is very informative as it verified thefidelity by the characterizing approach, while the FM directlyconfirmed the functionalization. The promising resultsobtained by our research group [7], [15], [29], [46] bringforward insights to the study of intermolecular detections.

IV. CONCLUSION

The combination of theoretical and experimental studiesidentified possible regions where the herbicide mesotrioneinteracts on the HPPD molecular structure. Additionally,the AFM adhesion measurements showed the accuracy ofthe functionalized HPPD nanobiosensor, which was alsocorroborated by the FM tip labeling. The next step of ourinvestigation is to compare the results from human HPPD withthe plant HPPD because the latter is more directly affected bymesotrione molecules. Finally, the originality of the biosensorproposed in this paper is based on AFS categorical detectionof herbicides for environmental monitoring.

ACKNOWLEDGMENTS

The authors of this paper would like to thankM. Castilho de Almeida Moura for the tip drawingsusing Corel Draw. They acknowledge the Post-GraduationProgram of Biotechnology and Environment Monitoring ofthe Federal University of São Carlos, Sorocaba.

REFERENCES

[1] J. Riu, A. Maroto, and F. X. Rius, “Nanosensors in environmentalanalysis,” Talanta, vol. 69, no. 2, pp. 288–301, Apr. 2006.

[2] J. R. Stetter, W. R. Penrose, and S. Yao, “Sensors, chemical sensors,electrochemical sensors, and ECS,” J. Electrochem. Soc., vol. 150, no. 2,pp. S11–S16, Feb. 2003.

[3] D. K. Deda et al., “Atomic force microscopy-based molecular recogni-tion: A promising alternative to environmental contaminants detection,”in Current Microscopy Contributions to Advances in Science and Tech-nology, vol. 5, A. Méndez-Vilas, Ed., 1st ed. Badajoz, Spain: FormatexResearch Center, 2012, pp. 1–30.

[4] A. F. Melo, “Desenvolvimento preliminar de um biossensor enzimáticopara determinação de taninos hidrolisáveis,” Univ. Fed. Rio de Janeiro,Rio de Janeiro, Brazil, Tech. Rep., 2008.

[5] A. Berquand et al., “Antigen binding forces of single antilysozyme Fvfragments explored by atomic force microscopy,” Langmuir, vol. 21,no. 12, pp. 5517–5523, Jun. 2005.

[6] O. H. Willemsen, M. M. Snel, A. Cambi, J. Greve, B. G. De Grooth,and C. G. Figdor, “Biomolecular interactions measured by atomic forcemicroscopy,” Biophys. J., vol. 79, no. 6, pp. 3267–3281, Dec. 2000.

[7] D. K. Deda et al., “The use of functionalized AFM tips as molecularsensors in the detection of pesticides,” Mater. Res., vol. 16, no. 3,pp. 683–687, Jun. 2013.


[8] C. Steffens, F. L. Leite, C. C. Bueno, A. Manzoli, and P. S. Herrmann,“Atomic force microscopy as a tool applied to nano/biosensors,” Sensors,vol. 12, no. 6, pp. 8278–8300, 2012.

[9] P. Hinterdorfer and Y. F. Dufrêne, “Detection and localization of singlemolecular recognition events using atomic force microscopy,” NatureMethods, vol. 3, no. 5, pp. 347–355, May 2006.

[10] F. Leite, “Theoretical models for surface forces and adhesion and theirmeasurement using atomic force microscopy,” Int. J. Molecular Sci.,vol. 13, no. 12, pp. 12773–12856, 2012.

[11] F. L. Leite and P. S. P. Herrmann, “Application of atomic force spec-troscopy (AFS) to studies of adhesion phenomena: A review,” J. Adhes.Sci. Technol., vol. 19, nos. 3–5, pp. 365–405, Jan. 2005.

[12] A. Noy, D. V. Vezenov, and C. M. Lieber, “Chemical force microscopy,”Annu. Rev. Mater. Sci., vol. 27, no. 1, pp. 381–421, 1997.

[13] B. Dordi, J. P. Pickering, H. Schönherr, and G. J. Vancso, “Invertedchemical force microscopy: Following interfacial reactions on thenanometer scale,” Eur. Polym. J., vol. 40, no. 5, pp. 939–947, May 2004.

[14] K. Wadu-Mesthrige, B. Pati, W. M. McClain, and G.-Y. Liu, “Disaggre-gation of tobacco mosaic virus by bovine serum albumin,” Langmuir,vol. 12, no. 14, pp. 3511–3515, Jan. 1996.

[15] A. C. N. da Silva et al., “Nanobiosensors based on chemically modifiedAFM probes: A useful tool for metsulfuron-methyl detection,” Sensors,vol. 13, no. 2, pp. 1477–1489, Jan. 2013.

[16] J. E. Casida and M. Tomizawa, “Insecticide interactions withγ -aminobutyric acid and nicotinic receptors: Predictive aspects ofstructural models,” J. Pesticde Sci., vol. 33, no. 1, pp. 4–8, 2008.

[17] N. Nugaeva, K. Y. Gfeller, N. Backmann, H. P. Lang, M. Düggelin, andM. Hegner, “Micromechanical cantilever array sensors for selective fun-gal immobilization and fast growth detection,” Biosensors Bioelectron.,vol. 21, no. 6, pp. 849–856, Dec. 2005.

[18] G. R. Moran, “4-hydroxyphenylpyruvate dioxygenase,” ArchivesBiochem. Biophys., vol. 433, no. 1, pp. 117–128, Jan. 2005.

[19] M. Kavana and G. R. Moran, “Interaction of (4-hydroxyphenyl)pyruvatedioxygenase with the specific inhibitor 2-[2-nitro-4-(trifluoromethyl)benzoyl]-1,3-cyclohexanedione,” Biochemistry, vol. 42,no. 34, pp. 10238–10245, Sep. 2003.

[20] S. Molchanov and A. Gryff-Keller, “Inhibition of 4-hydroxyphenylpyruvate dioxygenase by 2-[2-nitro-4-(trifluoromethyl)benzoyl]-1,3-cyclohexanedione,” Acta Biochim.Polonica, vol. 56, no. 3, pp. 447–454, 2009.

[21] K. Grossmann and T. Ehrhardt, “On the mechanism of action andselectivity of the corn herbicide topramezone: A new inhibitor of4-hydroxyphenylpyruvate dioxygenase,” Pest Manage. Sci., vol. 63,no. 5, pp. 429–439, May 2007.

[22] R. Martinazzo, D. P. Dick, M. M. Hirsch, S. B. Leite, andM. do Carmo Ruaro Peralba, “Sorption of atrazine and mesotrionein oxisols and estimation of contamination potential,” Química Nova,vol. 34, no. 8, pp. 1378–1384, Jan. 2011.

[23] P. H. Raven, Biologia Vegetal, vol. 9, G. Koogan, Ed., 6th ed., Rio deJaneiro, Brazil: Guanabara Koogan, 2001.

[24] G. Meazza et al., “The inhibitory activity of natural products on plantp-hydroxyphenylpyruvate dioxygenase,” Phytochemistry, vol. 60, no. 3,pp. 281–288, Jun. 2002.

[25] G. Binnig, C. F. Quate, and C. Gerber, “Atomic force microscope,” Phys.Rev. Lett., vol. 56, no. 9, pp. 930–933, Mar. 1986.

[26] B. B. Souza et al., “Modern trends in nanobiosensors using atomicforce microscopy,” presented at the Latin Amer. Conf. Metastable andNanostruct. Mater., São Carlos, Brazil, 2012.

[27] E. F. Franca, A. M. Amarante, and F. L. Leite, “Introduction to atomicforce microscopy simulation,” in Science, Technology, Applicationsand Education, vol. 4, A. Méndez-Vilas, Ed., 1st ed. Badajoz, Spain:Formatex Research Center, 2010, pp. 1338–1349.

[28] C. C. Bueno, “Desenvolvimento de um nanobiossensor para o moni-toramento da qualidade ambiental no setor agrícola,” M.S. thesis, Dept.Phys. Chem. Math., Univ. Federal São Carlos, São Paulo, Brazil, 2013.

[29] C. C. Bueno et al., “Nanobiosensor for diclofop detection based onchemically modified AFM probes,” IEEE Sensors J., vol. 14, no. 5,pp. 1467–1475, May 2014.

[30] E. S. Pilka et al., “Crystal structure of human 4-hydroxyphenylpyruvatedioxygenase,” Biochemistry, vol. 43, no. 32, pp. 10414–10423, 2004.

[31] G. M. Morris et al., “AutoDock4 and AutoDockTools4: Automateddocking with selective receptor flexibility,” J. Comput. Chem., vol. 30,no. 16, pp. 2785–2791, Dec. 2009.

[32] E. F. Franca, “Designing an enzyme-based nanobiosensor using molec-ular modeling techniques,” Phys. Chem. Chem. Phys., vol. 13, no. 19,pp. 8894–8899, 2011.

[33] L. Kalé et al., “NAMD2: Greater scalability for parallel moleculardynamics,” J. Comput. Phys., vol. 151, no. 1, pp. 283–312, 1999.

[34] W. Humphrey, A. Dalke, and K. Schulten, “VMD: Visual moleculardynamics,” J. Molecular Graph., vol. 14, no. 1, pp. 33–38, Feb. 1996.

[35] G. S. Oliveira, “Molecular modeling of enzyme attachment on AFMprobes,” J. Molecular Graph. Model., vol. 45, pp. 128–136, Sep. 2013.

[36] A. J. W. Orry and R. Abagyan, Homology Modeling—Methods andProtocols, vol. 857, 1st ed. New York, NY, USA: Humana Press, 2012.

[37] Basic Local Alignment Search Tool. [Online]. Available:http://blast.ncbi.nlm.nih.gov

[38] A. L. D. Moreau and M. A. Cotta. (Nov. 17, 2005).Processamento e Funcionalização de Pontas Para AplicaçõesBiologicas de Microscopia de Força Atomica. [Online]. Available:http://www.bibliotecadigital.unicamp.br/document/?code=vtls000385613,accessed Sep. 17, 2014.

[39] G. S. Lorite et al., “Surface physicochemical properties at the micro andnano length scales: Role on bacterial adhesion and Xylella fastidiosabiofilm development,” PLoS One, vol. 8, no. 9, p. e75247, Sep. 2013.

[40] B. J. Haupt, A. E. Pelling, and M. A. Horton, “Integrated confocaland scanning probe microscopy for biomedical research,” Sci. World J.,vol. 6, pp. 1609–1618, Dec. 2006.

[41] A. Li et al., “Molecular mechanistic insights into the endothelialreceptor mediated cytoadherence of plasmodium falciparum-infectederythrocytes,” PLoS One, vol. 6, no. 3, p. e16929, Mar. 2011.

[42] U. Dammer et al., “Specific antigen/antibody interactions measured byforce microscopy,” Biophys. J., vol. 70, no. 5, pp. 2437–2441, May 1996.

[43] R. De Paris, T. Strunz, K. Oroszlan, H.-J. Güntherodt, and M. Hegner,“Force spectroscopy and dynamics of the biotin-avidin bond stud-ied by scanning force microscopy,” Single Molecules, vol. 1, no. 4,pp. 285–290, Dec. 2000.

[44] M. Breitenstein, R. Höölzel, and F. F. Bier, “Immobilization of differentbiomolecules by atomic force microscopy,” J. Nanobiotechnol., vol. 8,no. 1, p. 10, May 2010.

[45] A. P. Limanskii, “Functionalization of amino-modified probes for atomicforce microscopy,” Biophysics, vol. 51, no. 2, pp. 186–195, Apr. 2006.

[46] A. C. N. da Silva et al., “Nanobiosensors exploiting specific interactionsbetween an enzyme and herbicides in atomic force spectroscopy,”J. Nanosci. Nanotechnol., vol. 14, no. 9, pp. 6678–6684, Sep. 2014.

Pâmela Soto Garcia was born in Sorocaba, Brazil,in 1984. From 2006 to 2007, she performed under-graduate research at the Biomonitoring Laboratory,Faculty of Technology at Sorocaba (FATEC-SO),Sorocaba. In 2008, at her second undergraduateresearch at the Dante Pazzanese Institute, São Paulo,Brazil, she studied devices for medical applications.She received the B.S. degree in health technologyfrom FATEC-SO, in 2008. In 2009, she studiedmicrobiology in public health at the Adolfo LutzInstitute, São Paulo. She received the M.Sc. degree

in biotechnology and environmental monitoring from the Federal Universityof São Carlos (UFScar), Sorocaba, in 2014. Since 2012, she has developedatomic force microscopy (AFM) tips nanobiosensors, and is a specialist inAFM and Nanotechnology with the Nanoneurobiophysics Research Group,UFScar, where she is currently pursuing the Ph.D. degree in nanobiosensors.

Alberto Luís Dario Moreau was born in São Paulo,Brazil, in 1977. He received the B.S. degree inphysics, and the M.Sc. and Ph.D. degrees from theState University of Campinas, Campinas, Brazil, in2003, 2005, and 2011, respectively. He is currentlya Professor and Coordinator of the Basic PhysicsLaboratory at the Federal Institute of Education,Science and Technology, Itapetininga, Brazil. Since2013, he has been with the NanoneurobiophysichsResearch Group, Federal University of São Carlos,Sorocaba, and has experience in biophysics with an

emphasis on functionalization and immobilization of biomaterials surfaces andinterfaces, force spectroscopy with atomic force microscopy (AFM), AFMtopographic analysis of biomaterials, carbon nanotubes and graphene, andbiosensors in semiconductor platforms.


Jéssica Cristiane Magalhães Ierich was born inSorocaba, Brazil, in 1991. She received the degree intechnology on biomedical systems from the Facultyof Technology at Sorocaba, Sorocaba, in 2011, andthe M.S. degree in biotechnology and environmentalmonitoring from the Federal University of São Car-los (UFSCar), Sorocaba, in 2014. In 2012, she hadthe opportunity to study 3-D structures of proteinsusing homology modeling and molecular dynam-ics simulation. Also, she has studied enzymaticinhibition process by herbicides for nanobiosensors

applications. She is currently pursuing the bachelor’s degree in biologicalsciences and the Ph.D. degree at UFSCar. Her Ph.D. study is focused on thedescription of antigen-antibody interaction by means of computational andtheoretical approaches.

Ana Carolina Araujo Vig was born in São Paulo,Brazil, in 1992. She is currently pursuing the bach-elor’s degree in chemistry at the Federal Universityof São Carlos (UFSCar), Sorocaba, Brazil. In 2010,she began teaching Chemistry for private students,preparatory courses at Corporative University, andtutoring at Aprendiz Reinforcement School. She alsotutored the students of UFSCar coursing Physics Iin 2012. In 2011, she joined the GNN ResearchGroup, functionalizing atomic force microscopy tipsfor the study of nanobiosensors. Currently, she has a

Scientific Initiation in Theoretical and Computational Chemistry, studying theIgG antibody, specially its binding site, and its relation to multiple sclerosis.

Akemi Martins Higa is currently pursuing thebachelor’s degree in biological sciences from theFederal University of São Carlos (UFSCar), Soro-caba, Brazil. She was born in São Paulo, Brazil, in1992. She joined the GNN Research Group in 2012,studying the immobilization of enzymes on atomicforce microscopy tips. Since 2013, she has studiedthe development of quantum dots functionalizationtechniques to cover them with biomolecules, suchas antibodies and antigens. The main purpose of herstudies with the group is to develop a nanobiosensor

that promotes an accurate and early diagnosis for multiple sclerosis disease.

Guedmiller S. Oliveira received the B.S. and M.S.degrees in physical chemistry from the Federal Uni-versity of Uberlândia, Uberlândia, Brazil, in 2006and 2009, respectively, and the Ph.D. degree inphysical chemistry from the Federal University ofSão Carlos (UFSCar), Sorocaba, Brazil, in 2013.Since 2007, he has worked with computer simulationproviding an atomistic point of view for experi-mental procedures. His expertise lies on quantummechanics theory, molecular dynamics simulation,and it combines results from experimental and theo-

retical analysis through statistical thermodynamics to improve comprehensionof the macromolecular phenomena. He currently holds a post-doctoral positionwith UFSCar.

Fábio Camargo Abdalla received the bachelor’sdegree in biological sciences from São Paulo StateUniversity, Rio Claro, Brazil, in 1996, the mas-ter’s degree in biological sciences with a minor inmolecular cellular biology from São Paulo StateUniversity and the University of Utrecht, Utrecht,The Netherlands, in 1999, the Ph.D. degree in bio-logical sciences with a minor in molecular cellularbiology from São Paulo State University and KeeleUniversity, Keele, U.K., in 2002, and the Post-Doctoral degree from São Paulo State University, in

2006. He is currently a Professor with the Federal University of São Carlos,Sorocaba, Brazil. He has experience in cell and molecular biology with anemphasis on structural and functional biology and chemical ecology.

Moema Hausen was born in Rio de Janeiro, Brazil,in 1977. She received the Ph.D. degree from theState University of Rio de Janeiro, Rio de Janeiro,in 2009. After three years performing her firstpost-doctoral assistance at the Brazilian Center forPhysics Research, Rio de Janeiro, she is currentlyinvolved in a second one, at the Biotechnologyand Environmental Monitoring Post-Graduation Pro-gram, Federal University of São Carlos, Sorocaba,Brazil. In 2000, she started in biomedical scientificlaboring in the following themes—cell biology, his-

tology, protozoology, transmission, and assorted state-of-the-art microscopytechniques, such as the scanning electron, transmission electron, fluorescence,and confocal ones. Her main goals actually are the application of high-endmicroscopy techniques to integrated approaches on materials and biologicalsciences.

Fábio L. Leite was born in Itanhaem, Brazil.He received the B.Sc. degree in physics from SãoPaulo State University, Rio Claro, Brazil, in 2000,and the M.Sc. and Ph.D. degrees in materials scienceand engineering from the University of São Paulo,São Carlos, Brazil, in 2002 and 2006, respec-tively. From 2007 to 2008, he was a Post-DoctoralResearcher with the Alan Graham MacDiarmidInstitute of Innovation and Business, Embrapa Agri-cultural Instrumentation (Embrapa), São Carlos,with Dr. O. N. de Oliveira, Jr., Dr. L. H. C. Mattoso

(Embrapa), and A. G. MacDiarmid, and was a recipient of the University ofPennsylvania Nobel Prize in Chemistry in 2000. His efforts at the MacDiarmidInstitute focused on conducting polymers, nanosensors, and atomic forcemicroscopy (AFM) with environmental applications. Since 2009, he has beenan Assistant Professor and a Researcher with the Federal University ofSão Carlos, Sorocaba, Brazil, and the Head of the NanoneurobiophysicsResearch Group. He has authored over 50 published papers, five books,10 book chapters, and holds two patents. His research interests are relatedto the development of nanobiosensors using AFM and computational nano-technological for application in the studies of a variety neurodegenerative andautoimmune diseases.