Analytical Biochemistry 381 (2008) 107–112

Contents lists available at ScienceDirect

Analytical Biochemistry

journal homepage: www.elsevier .com/ locate/yabio

A fluorescence polarization assay to quantify biotin and biotin-bindingproteins in whole plant extracts using Alexa-Fluor 594 biocytin

Harry Martin *, Colleen Murray, John Christeller, Tony McGhieThe Horticulture and Food Research Institute of New Zealand Limited, HortResearch, Palmerston North 4474, New Zealand

a r t i c l e i n f o

Article history:Received 2 May 2008Available online 20 June 2008

Keywords:Fluorescence polarizationBiotinAvidinAlexa-Fluor 594 biocytinTobacco leaf

0003-2697/$ - see front matter � 2008 Elsevier Inc. Adoi:10.1016/j.ab.2008.06.025

* Corresponding author. Fax: +64 6 354 6731.E-mail address: hmartin@hortresearch.co.nz (H. M

a b s t r a c t

A high-throughput fluorescence polarization assay has been developed for the detection of biotin andbiotin-binding proteins in whole leaf extracts. Various groups are investigating the insecticidal propertiesof avidin and other biotin-binding proteins expressed in leaves of transgenic plants. The methods com-monly used to quantify biotin and avidin in leaf extracts are enzyme-linked immunosorbent assay(ELISA) and Western blotting. Here we describe a homogeneous fluorescence polarization (FP) methodthat quantifies transgenic avidin in whole leaf extract by the simple addition of the fluorescent avidinligand Alexa-Fluor 594 biocytin (AFB). The FP assay exploits the fact that AFB excites and emits in regionsof the spectrum that are relatively free of background fluorescence in leaf extract. Transgenic leaf avidincan be quantified within 1–2 h by the FP method, in comparison with 1–2 days for ELISA and Westernblotting. The FP method can also measure the amount of biotin in control leaves, not expressing avidin.Functional avidin levels of 1.54 lM (26.1 lg/g leaf tissue) were detected in tobacco leaves expressing vac-uole-targeted avidin. Control leaves had biotin levels of around 0.74 lM (�0.18 lg/g leaf tissue). Reagentcosts are minimal: typically AFB is used at concentrations of 1–10 nM, avidin is used at 1–100 nM, andsample volumes are 20 lL in 384-well microplates.

� 2008 Elsevier Inc. All rights reserved.

Biotin is a water-soluble vitamin that is required for normal cel-lular metabolism and growth [1,2]. It functions as a carboxyl car-rier in carboxylation, decarboxylation, and transcarboxylationreactions, and biotin deficiency diseases have been documentedin humans [3,4]. Although biotin has traditionally been viewed asan essential cofactor of carboxylases, there have also been long-standing suggestions of a role for biotin in the regulation of geneexpression [5–8]. For example, biotin has been reported to stimu-late the synthesis of hepatic glucokinase and to repress phospho-enolpyruvate carboxykinase mRNA in rat liver [5,9]. Biotin hasalso been reported in the nuclei of cells; as long ago as 1963,Dakshinamurti and Mistry [10] showed that a large proportion ofradioactive biotin, injected into chicks and rats, localized to the nu-clear fraction. Recently, a potential role for biotin in gene regula-tion has been demonstrated by specific biotinylation of histones.All five histone classes extracted from human peripheral bloodmononuclear cells contain biotin [11–13]. Recent data collectedby Mock et al. show that marginal biotin deficiency occurs in a sub-stantial proportion of pregnant women [14]. A study conducted inmice demonstrated that marginal biotin deficiency was teratogenic[15,16]. These studies emphasize the neglected status of biotin asan important vitamin for which adequate intake and daily require-

ll rights reserved.

artin).

ment data are inadequate [17,18], and the need for a rapid, reliableassay for biotin content in food material.

Biotin is an essential growth factor for insects. The insecticidalproperties of avidin, a biotin-sequestering protein, have beenknown since 1959, when it was shown that avidin was insecticidalwhen included in the diet of housefly larvae [19,20] and subse-quently against a wide range of insects [21–31]. Avidin is a tetra-meric protein of approximate molecular weight 68 kDa. Avidincontains four biotin-binding sites with the remarkable affinity forbiotin of 10�15 M�1. Because of their insecticidal properties, avidinand streptavidin have been expressed in a variety of agriculturallyimportant plant species, for example, tobacco, apple, maize, andrice [23,26,29,32]. The safety of avidin in transgenic rice [29] andmaize [23] has been demonstrated, as has the loss of most avidinactivity by heat denaturation on cooking at 95 �C for 5 min [29].Furthermore, avidin has an added advantage over conventionalinsecticides in that, as a component of the stored crop, it is notwashed away during processing and continues to act as an insecti-cide during storage. In addition, unlike insecticidal chemicalsprays, avidin should have a minimal effect as an environmentalpollutant because it is a biodegradable natural protein, found inthe egg whites of birds and reptiles. Biotin is a natural antidoteto avidin.

Murray and co-workers [32] have expressed avidin in tobaccoplants using an N-terminal vacuolar targeting sequence. The plantsare phenotypically normal and have no detectable changes in

108 Assay of biotin and avidin in leaves by fluorescence polarization / H. Martin et al. / Anal. Biochem. 381 (2008) 107–112

biotin metabolism. The intracellular separation of the expressedavidin (in the vacuoles) from the leaf biotin (in the cytoplasm,chloroplasts, mitochondria, and nucleus) allows the plant to thrivewhile being protected from insect pests. The avidin is released onlyfrom its stores and therefore able to bind plant biotin following cellmaceration by the insect pest during feeding. Typically, the level ofavidin expression in transgenic plants required to suppress larvaldevelopment is around 4 lM [32].

Quantification of the levels of biotin in wild-type leaves is usu-ally achieved by a competitive ELISA1 method [33], and avidinexpression can be estimated by Western blot and ELISA [32,34].Schray and co-workers [35] described the use of the fast, homoge-neous fluorescence polarization (FP) method for quantifying pico-gram levels of biotin and low nanogram levels of avidin inphosphate-buffered saline (PBS) within a period of 10 min. FP relieson the fact that rapid tumbling of small molecule fluors causes emit-ted fluorescence to be depolarized, whereas the emitted light ofslow-spinning fluors bound to larger molecules, such as proteinreceptors, is more highly polarized. The ratiometric nature of FPhas the additional advantage that it greatly reduces experimentalnoise. For example, FP copes well with murky samples and withsample volume variation, which constitute major problems for fluo-rescence intensity-dependent assays. Kada and co-workers [36] re-port the great utility of the biotin-4-fluorescein probe in generalFP and fluorescence-quenching assays for avidin and biotin measure-ment in samples such as egg white. In this article, we show that forthe quantification of biotin-binding proteins and biotin in whole leafhomogenates, Alexa-Fluor 594 biocytin (AFB) is ideal, because itsfluorescence emission peak falls in a region of low background fluo-rescence of the leaf extract. Biocytin is a form of biotin covalentlybound to lysine and this covalent modification of certain carboxylaselysines allows the enzymes to function. Avidin binds the fluorescentform of biocytin (AFB) with high affinity, and competitive inhibitionof AFB binding to avidin by biotin is the basis of the FP assay.

Materials and methods

Materials

Black, shallow 384-well untreated assay plates were purchasedfrom Nunc Company Ltd, Rochester, NY, USA. Biotin 5-fluoresceinand PBS tablets were purchased from Fluka. Bovine serum albumin(BSA), avidin agarose, 3-(N-morpholino)propanesulfonic acid(MOPS), diethyldithiocarbamate (DEDTC) and insoluble polyvinyl-polypyrrolidone (PVPP) were supplied by Sigma–Aldrich. AFB wassupplied by Invitrogen Corporation, Carlsbad, CA, USA. Plant tissuewas obtained from Nicotinia tabacum ‘Samsung’ transformed usingstandard methods [37]. Avidin was expressed as a vacuole-tar-geted chimera using the pGreen II/pSoup binary vector system[38]. Leaf tissue from transgenic tobacco expressing a heterologouschitinase was used as a control.

Methods

All fluorescence readings were performed using the Tecan Sa-fire2 fluorescence microplate reader (Tecan, Austria) at 22 �C, in avolume of 20 lL. For measurement of FP, excitation/emissionwavelengths of 470/525 and 590/625 nm were used for fluoresceinand AFB, respectively. In each case, a bandwidth of 10 nm was set.

1 Abbreviations used: ELISA, enzyme-linked immunosorbent assay; FP, fluorescencepolarization; mP, millipolarization; PBS, phosphate-buffered saline; BSA, bovineserum albumin; AFB, Alexa-Fluor 594 biocytin; MOPS, 3-(N-morpholino)propanesul-fonic acid; DEDTC, diethyldithiocarbamate; PVPP, insoluble polyvinylpolypyrroli-done; LOD, limit of detection.

For fluorescence intensity wavelength scans, a fixed gap of 40 nmwas set between excitation and emission with a 10-nm bandwidth.

Leaf samples were ground with a mortar and pestle inextraction buffer—50 mM MOPS, 1 mM, DEDTC, 5% PVPP, pH7.0—at a ratio of 1 g leaf:5 mL buffer. Homogenates were thenchilled on ice and centrifuged for 5 min at 11,000g to removeinsoluble material. Homogenates were stored at �20 oC untiluse.

For assay, leaf samples and fluorescent probes were diluted inPBS containing 0.1 mg/mL BSA. Biotin and avidin are both stableunder these conditions without the use of inhibitors. Avidin is nat-urally proteinase resistant, a prerequisite for a protein that seques-ters a dietary vitamin. As the aim of this method was to measurefunctionally active avidin-binding sites in the leaf homogenate(i.e., excess of avidin over biotin), no attempt was made to regen-erate biotin-binding capacity by dissociating biotin already boundto avidin.

Calculations and statistical analysis

Millipolarization (mP) values were calculated with theequation

millipolarizationðmPÞ ¼ 1000½ðIS—IPÞ=ðISþ IPÞ�

where IS is the parallel emission intensity and IP is the perpendic-ular emission intensity. The above equation assumes that light istransmitted equally well through both parallel and perpendicularoriented polarizers. In practice, this is not true and a correctionmust be made to measure the absolute polarization state of themolecule. This correction factor is called the ‘‘G factor.” In our as-says, the IP value was adjusted by a G factor of 1.176 that generatedan mP value of 27 for free fluorescein.

All statistical analyses (sigmoidal curve fitting and IC50 esti-mates) were performed with Origin software Version 7.5 (Origin-Lab, Northampton, MA, USA).

Of the assortment of equations available for fitting in the Origingraphic package, the logistic equation of the form

y ¼ A2þ ðA1� A2Þ=ð1þ ðx=x0ÞpÞ

was empirically chosen because it fit well to the experimental datasets. Each data point was the average of at least two determinations.Data presented are representative of at least two independentexperiments.

Results

To gauge the feasibility of FP measurements in leaf whole ex-tracts, a sample of homogenized leaf was scanned to generate afluorescence profile and the scan was compared with the fluores-cence peaks of biotin–fluorescein and AFB, two avidin ligands.The result (Fig. 1) shows that although at equimolar concentra-tions, the fluorescence signal from fluorescein is five times strongerthan that from AFB; the leaf background fluorescence at the emis-sion maximum of the AFB ligand is 20-fold lower. This means thatin plant extract, the signal-to-noise ratio of AFB is around fourfoldhigher than that of biotin–fluorescein. Therefore it is clear thatgreater sensitivity will be achieved using AFB.

To determine whether the presence of avidin expression intransgenic leaves would be detectable in whole leaf extracts, twodilution series of homogenates, one from control leaves and onefrom leaves expressing avidin, were compared with respect tothe ability to bind AFB. Binding of the fluorescent AFB ligand to avi-din should cause an increase in FP (mP units) and, as the sample isdiluted, polarization units should decline as avidin concentrationdecreases. Ten microliters of leaf extract was diluted fivefold inPBS containing 0.1 mg/mL BSA and added to 10 lL of 20 nM AFB.

0

2000

4000

6000

8000

500 600 700 800

emission wavelength (nm)

fluor

esce

nce

inte

nsity

_

control leaf extract

biotin-fluorescein

Alexa fluor 594 biocytin

Fig. 1. Scanned fluorescence profiles of 2 diluted control tobacco leaf extract (s),10 nM biotin–fluorescein (j), and 10 nM Alexa-Fluor 594 biocytin (AFB) (4). Thegap between excitation and emission was fixed at 50 nm.

Fig. 3. Estimate of the avidin concentration in avidin-expressing tobacco leaf extract:A dilution series of avidin-expressing leaf extract (j) is compared with an avidinstandard curve in PBS (N) and in control leaf extract (s). PBS avidin standardcurve EC50 = 4.3 ± 0.1 nM avidin (monomer), control leaf avidin standard curveEC50 = 4.3 ± 0.1 nM avidin (monomer), transgenic leaf EC50 (fractional concentra-tion) = 0.00278 ± 0.00013, and avidin (monomer) concentration in transgenicleaf = 1.54 lM.

Assay of biotin and avidin in leaves by fluorescence polarization / H. Martin et al. / Anal. Biochem. 381 (2008) 107–112 109

To confirm that high mP values in the avidin-expressing plant weredue to the presence of avidin, increasing concentrations of biotinwere mixed with the leaf sample. Fig. 2 shows the clear presenceof avidin in homogenates of avidin-expressing leaves, as the mPvalues of the AFB probe were high and reduced by increasing biotinconcentration.

Wilbur and co-workers [39] reported that AFB has a roughlythreefold slower association rate than biotin and has a very slowdissociation rate. To estimate the amount of avidin in the leaf ex-tract, the sample was compared with a standard curve of avidinin PBS and a standard curve in a control leaf extract. No extraneousbiotin was used in this assay, which measured the amount of avi-din necessary to bind 50% of the AFB. The AFB was present at a finalassay concentration of 5 nM. The control leaf extract had been trea-ted with avidin–agarose (40 ll agarose beads in 1 mL of leaf extractdiluted 1:25 with PBS, 0.1 mg/mL BSA, 30 min agitation) to removeendogenous biotin that would interfere with binding of the AFB

Fig. 2. Comparison of fluorescence polarization (mP) of 10 nM AFB in the presenceof (20 diluted, avidin-expressing tobacco leaf extract (j) and 20 diluted, control leaf(s) at increasing biotin concentration.

probe to avidin. This step was necessary because the FP assay isa functional assay that measures only excess, unoccupied biotinbinding sites on avidin. The results illustrated in Fig. 3 confirm thatthe polarization value of avidin-bound fluor was unaffected by thepresence of small amounts of leaf extract (at the top concentration,the leaf extract was a 1/20 dilution of leaf homogenate, droppingtwofold at each dilution step). The X axis in Fig. 3 represents thefunctional concentration of avidin in terms of avidin monomers.The functional concentration of avidin in whole leaf was estimatedto be 1.54 ± 0.07 lM. Estimates of avidin concentration were de-rived simply by assuming that the concentrations of avidin at theIC50 of the standard curve and the unknown (transgenic leaf) sam-ple were identical. The avidin concentration at the standard curveIC50 was then multiplied by the dilution factor of the leaf sample toestimate the actual functional avidin concentration in the originalleaf extract. The same procedure was applied to estimating biotinconcentration in leaves from a biotin standard curve IC50. This ap-proach was taken because producing a complete dose–responsecurve will reveal problems with the method. Interfering factorssuch as contamination of the signal, which can occur at higher leafconcentrations, are detected by this approach. However, for normalapplications, where fluorescence intensity scans confirm that theAFB signal is unaffected by background fluorescence, it is sufficientto fit mP data onto the standard curve. A full dose–response curvefor each sample is unnecessary. This has been confirmed by ourown, now routine, sample screening.

Unlike measurement of avidin in leaf homogenate, which can beperformed in a simple direct assay of AFB binding to avidin, biotinmeasurement requires the combined use of biotin and biocytin in acompetitive binding assay. Avidin, biotin, and biocytin were mixedtogether to determine whether the order of reagent addition af-fected the outcome. Fig. 4 illustrates the result of an experimentin which a constant amount of avidin (6 nM final monomer con-centration) was mixed with a constant amount of biocytin (1 nMfinal concentration) and a dilution series of biotin ranging from333 to 0.07 nM (final concentration). The first two reagents weremixed at room temperature for 1 h followed by addition of the finalreagent. The result confirms there was no advantage, in terms ofsensitivity, to adding the biotin first, followed by the biocytin

Fig. 4. Effect of the order of addition of reagents on mP. Avidin (18 nM), biotin, andAFB (3 nM) were mixed together in equal volumes with one of these reagents beingadded last. The first two reagents were allowed to stand for 1 h before addition ofthe last reagent. Last reagent: (s) biotin, (h) avidin, AFB (d). Final concentrationswere: avidin: 6 nM avidin monomer, biotin: as shown on X axis, AFB: 1 nM

110 Assay of biotin and avidin in leaves by fluorescence polarization / H. Martin et al. / Anal. Biochem. 381 (2008) 107–112

(i.e., the same IC50 was generated by the simultaneous addition ofbiotin and biocytin to avidin); this may reflect the faster on-rate ofbiotin over AFB and the displacement of AFB by biotin observed byWilbur and co-workers [39]. Interestingly, if biocytin is added tothe avidin first and incubated for 1 h before the addition of biotin,then a substantial minority (approximately 30%) of the biocytin isdisplaced from the avidin, whereas the remainder of the biocytinappears to be irreversibly bound. Conceivably this may indicatethat avidin occupied at four sites may undergo cooperative, tighterbinding than avidin in which fewer sites are occupied, as the pro-portion of irreversibly bound AFB remained constant irrespectiveof increasing biotin concentration. In this regard, Feltus and co-workers [40] have demonstrated cooperative binding of biotin byimmobilized avidin.

To estimate the concentration of biotin in a control tobacco leafextract not expressing avidin, the homogenate was adsorbed withavidin–agarose to remove any biotin, as described previously. Thebiotin-cleared extract was then compared with the same controlleaf extract, not exposed to avidin–agarose, for its ability to inhibitthe binding of AFB. These data (Fig. 5a) show that, at a concentra-tion of around 3 nM avidin monomer, 1 nM AFB, and a 50-fold dilu-tion of leaf extract, avidin clearly had inhibitory activity because ofthe presence of biotin. This combination of reagents, at a similarconcentration, was therefore used to quantify the amount of biotinin the leaf extract, by comparison with a standard curve of biotin(Fig. 5b).

Fig. 5. (A) Detection of free biotin in control tobacco leaf by FP. Biocytin was addedto control leaf extract (j) or control leaf extract preadsorbed with avidin–agaroseto remove free biotin (s). These samples were then mixed with increasingconcentrations of avidin to reveal the avidin-blocking effect of endogenous leafbiotin. The final reagent concentrations were: 10 nM AFB, 1/60 diluted leaf extract,avidin (monomer) as shown. (B) Estimate of biotin concentration in control tobaccoleaf. Control leaf extract was untreated (d) or exposed to avidin–agarose to removebiotin (4). After addition of AFB, avidin was added. A standard curve of biotin wasprepared with the same AFB and avidin reagents (h). Final reagent concentrationswere: 1 nm AFB and 1.2 nM avidin (monomer). Arrowed samples 1–3 are describedin (C). (C) Interference of the AFB fluorescence signal by leaf extract fluorescence atthe highest leaf extract concentration. Selected samples from (B) were scanned forfluorescence intensity (50-nm gap between excitation and emission). The samplecorresponding to the highest leaf extract of 1.6% (arrow 1, B) (d) is compared with afourfold dilution of that sample containing 0.4% leaf extract (4) (arrow 2, B) andwith a saline control (j) (arrow 3, B).

"

Leaf extract was prepared at a dilution of 1:5 (w/v). This samplewas further diluted 10-fold in PBS/0.1% BSA buffer so that its finalfractional leaf concentration was 0.083. A small volume of 200 nMAFB was added to the sample to bring the AFB concentration to2 nM. From this sample, a dilution series was prepared in 2 nMAFB. Ten microliters of each dilution was then mixed with 10 L of2.4 nM avidin (monomer). The final concentrations of AFB and avi-

Assay of biotin and avidin in leaves by fluorescence polarization / H. Martin et al. / Anal. Biochem. 381 (2008) 107–112 111

din in each microplate well then became 1 nM AFB and 1.2 nM avi-din (monomer). After a 60-min equilibration period, the mP valuesgenerated were compared with a standard curve of biotin in PBS/BSA prepared in the same way. The EC50 for the standard curvewas 1.03 nM biotin, whereas the 50% binding was achieved withleaf sample at a fractional concentration of 0.0011. This wouldindicate that the biotin concentration in the leaf was 0.94 nM.

It was, however, evident that the lower mP values in the mostconcentrated samples did not achieve the same low mP values ofthe standard curve (Fig 5b, arrow 1). This was because at the high-est concentrations of leaf sample, the fluorescence signal and,therefore, the EC50 were both affected by the background fluores-cence of the leaf extract (see Fig. 5c showing the fluorescenceintensity scans of selected samples). At the highest concentrationsof leaf extract, there was a small but real interference with the mPsignal coming from fluorescent materials in the leaf extract. Whenthe Origin EC50 curve-fitting software was given the true minimumvalue 45 mP (from the PBS standard curve), this estimate of biotinconcentration changed to 0.75 nM. To confirm that this approach isvalid (i.e., using the PBS minimum mP value in place of the leaf ex-tract minimum value), the leaf extract was cleared of biotin usingavidin–agarose and used as a diluent to prepare a biotin standardcurve. At a final leaf extract dilution of 1/400, no effect was ob-served on the biotin EC50 compared with a standard curve pre-pared in PBS (data not shown). These results confirmed that thefluorescent background of the leaf did not compromise the mP val-ues in the region of the IC50 and also supported the choice of AFBover biotin–fluorescein in FP assays of leaf extracts. Standardcurves prepared in PBS are suitable for estimating leaf biotin. Theuse of the PBS standard curve can be validated by measurementof the fluorescence of the diluted sample in the presence and ab-sence of AFB to confirm that the background emission, in the ab-sence of AFB, is a small fraction of the AFB signal fluorescence.The background fluorescence (in perpendicular and parallelplanes) can then be subtracted from the mP calculation (FP instru-ments will perform this blank subtraction automatically).

Discussion

The speed, simplicity, and accuracy of FP for quantifying biotinconcentration in biological samples were reported by Schray andco-workers [35] using biotin–fluorescein. However, because ofthe high fluorescence of leaf extracts in the fluorescein range ofexcitation and emission (480–550 nm), fluorescein is not ideal forthe analysis of leaf samples. Here we have shown that the proper-ties of AFB make it well suited to the analysis of leaf extracts forbiotin and avidin measurement. The fluorescence intensity curvein Fig. 1 shows the emission peak of AFB fitting into a region ofthe leaf fluorescence profile with very low background levels.The relative intensities of fluorescein and AFB were not optimizedand were only compared at pH 7.4 and using the ionic strength ofPBS; the profile may therefore change in different media. On stor-age of leaf extract at –20oC for several weeks, a green precipitate ofchlorophyll-containing material formed, which could be removedby centrifuging the sample. Although this changed the backgroundso that the peaks at 700 nm were reduced (compare Fig. 1 withFig. 5c), the high fluorescence around 500–550 nm was not af-fected by the chlorophyll precipitation, nor were the estimated ex-pressed avidin or biotin concentration values.

In addition to FP, Kada and co-workers [36] used fluorescencequenching as a method for estimating the binding of biotin–fluo-rescein to avidin. We found that the fluorescence intensity offluorescein and AFB bound to avidin was reduced by about50%. However, the ratiometric nature of FP gives it a greatadvantage over fluorescence quenching as a method for measur-

ing avidin binding. The fluorescence intensity fluctuations in ourdata were very large compared with the polarization values(data not shown).

The partial reversibility of biocytin binding (Fig. 4) was surpris-ing (about 30% of the AFB could be displaced by 1 h of incubationwith high biotin concentrations, but no further AFB displacementoccurred even after overnight incubation). This may indicate thatthat a subpopulation of the avidin molecules in solution had a low-er affinity for biocytin, whereas another portion of the avidinbound the biocytin in an essentially irreversible way, perhaps be-cause of some degradation or denaturation during productionand preparation of the avidin sample. Another possibility is thatthis result was evidence of cooperative binding as reported byFeltus and co-workers [40]. The molar biocytin concentration inthese experiments was near the molar concentration of avidinbinding sites; some of the biocytin molecules may be weakly andreversibly bound, whereas others may have bound biocytin irre-versibly at all four binding sites. In avidin molecules where all foursites are occupied, the affinity may be higher than in avidin occu-pied at fewer sites by biocytin. The avidin occupied by AFB at fewerthan four sites may be more susceptible to displacement by biotin.

We have not quoted a limit of detection (LOD) for this proce-dure because the LOD will be dependent on the particular samplein use. We have prepared leaf extracts from various plants, forexample, tobacco, apple, kiwifruit, Arabidopsis, and found consider-able variation between and within plants with respect to leaf ex-tract color intensity, which varies depending on factors includingthe age of the plant, the season, and the location of the leaf onthe plant. The LOD is also influenced by the instrumentation avail-able. Using the Tecan Safire2 FP instrument and AFB in PBS, wefound the polarization dose–response curve to be stable using bio-cytin at concentrations as low as 100 pM (data not shown).

FP allows the quantification of only functional avidin, as bindingsites occupied by endogenous biotin cannot bind AFB. Comparisonof the three methods—FP, ELISA, and Western blotting—in ourhands, revealed very close agreement between functional ELISA(i.e., ELISA that detects avidin by its ability to bind to immobilizedbiotin) and FP for avidin detection. Although we do not give thesedata here, we have published many studies of both avidin and bio-tin measurement in plants by Western blotting and ELISA[21,32,33,41,42].

In certain instances where transgenic avidin expression is veryweak, there remains an excess of biotin in the homogenate. Inthese cases, FP cannot detect avidin, but it can detect a reductionin the concentration of biotin. Western blotting or ‘‘antigenic”rather than ‘‘functional” ELISA is necessary to quantify levels of avi-din expression that are lower than the endogenous plant biotinconcentration.

The FP procedure allows the rapid quantification of any biotin-binding proteins where the avidin concentration is in excess overendogenous biotin. In addition to avidin, two other biotin-bindingproteins [43,44] have been expressed transgenically in plants,streptavidin [32] and bradavidin (Murray, unpublished). It is verylikely that the FP assay described here will be suitable for quanti-fying all these biotin-binding proteins in whole extracts of plants.From the aspect of quantifying biotin concentrations in plants, theAFB FP method allows the measurement of free biotin and biotin inthe form of biotinylated protein.

Because of its simplicity, the FP procedure is well suited forhigh-throughput, automated screening, requiring only dilution ofthe plant sample and addition of the fluorescent probe in 20-l wellsof a 384-well microplate. We found that at concentrations around1 nM biotin and �1.2 nM avidin monomer, an equilibration periodof around 1 h at room temperature was necessary, after which nofurther mP changes occurred. At biotin and avidin concentrationsabove 25 nM, the equilibration occurred within minutes. We esti-

112 Assay of biotin and avidin in leaves by fluorescence polarization / H. Martin et al. / Anal. Biochem. 381 (2008) 107–112

mate the time required from leaf processing to avidin or biotinmeasurement is less than 2 h for a single operator, and half thistime is taken by the equilibration of avidin with its ligands. The re-agent costs are negligible and, with the recent increase in under-standing of the role of biotin in health and physiology, theavailability of a procedure for direct measurement of biotin andavidin in whole homogenates of plant foliage is useful.

References

[1] C. Alban, D. Job, R. Douce, Biotin metabolism in plants, Annu. Rev. Plant Physio:Plant Mol. Biol. 51 (2000) 17–47.

[2] J. Shellhammer, D. Meinke, Arrested embryos from the bio1 auxotroph ofArabidopsis thaliana contain reduced levels of biotin, Plant Physiol. 93 (1990)1162–1167.

[3] P.V. Attwood, J.C. Wallace, Chemical and catalytic mechanisms of carboxyltransfer reactions in biotin-dependent enzymes, Acc. Chem. Res. 35 (2002)113–120.

[4] W.L. Nyhan, Multiple carboxylase deficiency, Int. J. Biochem. 20 (1988) 363–370.[5] K. Dakshinamurti, Biotin: a regulator of gene expression, J. Nutr. Biochem. 16

(2005) 419–423.[6] Y.I. Hassan, J. Zempleni, Epigenetic regulation of chromatin structure and gene

function by biotin, J. Nutr. 136 (2006) 1763–1765.[7] R. Rodriguez-Melendez, J. Zempleni, Regulation of gene expression by biotin

[review], J. Nutr. Biochem. 14 (2003) 680–690.[8] A. Vilches-Flores, C. Fernandez-Mejia, Effect of biotin upon gene expression

and metabolism, Rev. Invest. Clin. 57 (2005) 716–724.[9] G. Romero-Navarro, G. Cabrera-Valladares, M.S. German, F.M. Matschinsky, A.

Velazquez, J. Wang, C. Fernandez-Mejia, Biotin regulation of pancreaticglucokinase and insulin in primary cultured rat islets and in biotin-deficientrats, Endocrinology 140 (1999) 4595–4600.

[10] K. Dakshinamurti, S. Mistry, Tissue and intracellular distribution of biotin-C14

OOH in rats and chicks, J. Biol. Chem. 238 (1963) 294–296.[11] T.D. Ballard, J. Wolff, J.B. Griffin, J.S. Stanley, S. van Calcar, J. Zempleni,

Biotinidase catalyzes debiotinylation of histones, Eur. J. Nutr. 41 (2002) 78–84.[12] N. Kothapalli, G. Camporeale, A. Kueh, Y.C. Chew, A.M. Oommen, J.B. Griffin, J.

Zempleni, Biological functions of biotinylated histones, J. Nutr. Biochem. 16(2005) 446–448.

[13] J. Zempleni, Uptake, localization, and noncarboxylase roles of biotin, Annu.Rev. Nutr. 25 (2005) 175–196.

[14] D.M. Mock, J.G. Quirk, N.I. Mock, Marginal biotin deficiency during normalpregnancy, Am. J. Clin. Nutr. 75 (2002) 295–299.

[15] D.M. Mock, Marginal biotin deficiency is teratogenic in mice and perhapshumans: a review of biotin deficiency during human pregnancy and effects ofbiotin deficiency on gene expression and enzyme activities in mouse dam andfetus, J. Nutr. Biochem. 16 (2005) 435–437.

[16] D.M. Mock, N.I. Mock, C.W. Stewart, J.B. LaBorde, D.K. Hansen, Marginal biotindeficiency is teratogenic in ICR mice, J. Nutr. 133 (2003) 2519–2525.

[17] D.A. Bender, Optimum nutrition: thiamin, biotin and pantothenate, Proc. Nutr.Soc. 58 (1999) 427–433.

[18] H.M. Said, Biotin: the forgotten vitamin, Am. J. Clin. Nutr. 75 (2002) 179–180.[19] C.A. Benschoter, Effect of dietary biotin on reproduction of the house fly, J.

Econ. Entomol. 60 (1967) 1326–1328.[20] K. Miura, T. Takaya, K. Koshiba, The effect of biotin deficiency on the

biosynthesis of the fatty acids in a blowfly, Aldrichina grahami, duringmetamorphosis under aseptic conditions, Arch. Int. Physiol. Biochim. 75(1967) 65–76.

[21] E.P. Burgess, L.A. Malone, J.T. Christeller, M.T. Lester, C. Murray, B.A. Philip,M.M. Phung, E.L. Tregidga, Avidin expressed in transgenic tobacco leavesconfers resistance to two noctuid pests, Helicoverpa armigera and Spodopteralitura, Transgenic Res. 11 (2002) 185–198.

[22] S.G. Cooper, D.S. Douches, E.J. Grafius, Insecticidal activity of avidin combinedwith genetically engineered and traditional host plant resistance againstColorado potato beetle (Coleoptera: Chrysomelidae) larvae, J. Econ. Entomol.99 (2006) 527–536.

[23] K.J. Kramer, T.D. Morgan, J.E. Throne, F.E. Dowell, M. Bailey, J.A. Howard,Transgenic avidin maize is resistant to storage insect pests, Nat. Biotechnol. 18(2000) 670–674.

[24] L.A. Malone, E.P. Burgess, C.F. Mercer, J.T. Christeller, M.T. Lester, C. Murray,M.M. Phung, B.A. Philip, E.L. Tregidga, J.H. Todd, Effects of biotin-bindingproteins on eight species of pasture invertebrates, N.Z. Plant Protect. 55 (2002)411–415.

[25] N.P. Markwick, J.T. Christeller, L.C. Docherty, C.M. Lilley, Insecticidal activity ofavidin and streptavidin against four species of pest Lepidoptera, Entomol. Exp.Appl. 98 (2001) 59–66.

[26] N.P. Markwick, L.C. Docherty, M.M. Phung, M.T. Lester, C. Murray, J.L.Yao, D.S. Mitra, D. Cohen, L.L. Beuning, S. Kutty-Amma, J.T. Christeller,Transgenic tobacco and apple plants expressing biotin-binding proteinsare resistant to two cosmopolitan insect pests, potato tuber moth andlightbrown apple moth, respectively, Transgenic Res. 12 (2003) 671–681.

[27] S. Meiyalaghan, M.F.G. Takla, O.O. Jaimess, Y. Shang, M.M. Davidson, P.A.Cooper, P.J. Barrell, J.M.E. Jacobs, S.D. Wratten, A.J. Conner, Evaluation oftransgenic approaches for controlling tuber moth in potatoes, Commun. Agr.Appl. Biol. Sci. 70 (2005) 641–650.

[28] S.C. Saxena, S. Kaul, Qualitative vitamin requirements of Oryzaephilus mercatorfauvel and their deficiency effects on the survival and growth of F1-progeny,Arch. Int. Physiol. Biochim. 82 (1974) 49–54.

[29] K. Yoza, T. Imamura, K.J. Kramer, T.D. Morgan, S. Nakamura, K. Akiyama, S.Kawasaki, F. Takaiwa, K. Ohtsubo, Avidin expressed in transgenic rice confersresistance to the stored-product insect pests Tribolium confusum and Sitotrogacerealella, Biosci. Biotechnol. Biochem. 69 (2005) 966–971.

[30] Y.C. Zhu, J.J. Adamczyk Jr., S. West, Avidin, a potential biopesticide andsynergist to Bacillus thuringiensis toxins against field crop insects, J. Econ.Entomol. 98 (2005) 1566–1571.

[31] T.D. Morgan, B. Oppert, T.H. Czapla, K.J. Kramer, Avidin and streptavidin asinsecticidal and growth-inhibiting dietary proteins, Entomol. Exp. Appl. 69(1993) 97.

[32] C. Murray, P.W. Sutherland, M.M. Phung, M.T. Lester, R.K. Marshall, J.T.Christeller, Expression of biotin-binding proteins, avidin and streptavidin, inplant tissues using plant vacuolar targeting sequences, Transgenic Res. 11(2002) 199–214.

[33] J.T. Christeller, M.M. Phung, Changes in biotin levels in the leaves of two applecultivars during the season, N.Z. J. Crop Hortic. Sci. 26 (1998) 39–43.

[34] J.T. Christeller, N.P. Markwick, J. Poulton, M. O’Callaghan, Binding of aninsecticidal transgene product to soil: biological activity of soil-bound avidinand the effects of time and microbial activity, Soil Biol. Biochem. 38 (2006)2043.

[35] K.J. Schray, P.G. Artz, R.C. Hevey, Determination of avidin and biotin byfluorescence polarization, Anal. Chem. 60 (1988) 853–855.

[36] G. Kada, H. Falk, H.J. Gruber, Accurate measurement of avidin and streptavidinin crude biofluids with a new, optimized biotin–fluorescein conjugate,Biochim. Biophys. Acta 1427 (1999) 33–43.

[37] R. Horsch, J. Fry, N. Hoffman, D. Eichholtz, S. Rogers, R. Fraley, A simple andgeneral method for transferring genes into plants, Science 227 (1985) 1299–1301.

[38] R. Hellens, A. Edwards, N. Leyland, S. Bean, P. Mullineaux, pGreen: a versatileand flexible binary Ti vector for Agrobacterium -mediated planttransformation, Plant Mol. Biol. 42 (2000) 819–832.

[39] D.S. Wilbur, P.M. Pathare, D.K. Hamlin, M.B. Frownfelter, B.B. Kegley, W.-Y.Leung, K.R. Gee, Evaluation of biotin–dye conjugates for use in an HPLC assayto assess relative binding of biotin derivatives with avidin and streptavidin,Bioconj. Chem. 11 (2000) 584–598.

[40] A. Feltus, S. Ramanathan, S. Daunert, Interaction of immobilized avidin with anAequorin–biotin Conjugate: an Aequorin-linked assay for biotin, Anal.Biochem. 254 (1997) 62–68.

[41] J.T. Christeller, L.A. Malone, J.H. Todd, R.M. Marshall, E.P.J. Burgess, B.A. Philip,Distribution and residual activity of two insecticidal proteins, avidin andaprotinin, expressed in transgenic tobacco plants, in the bodies and frass ofSpodoptera litura larvae following feeding, J. Insect Physiol. 51 (2005) 1117–1126.

[42] E.P.J. Burgess, B.A. Philip, J.T. Christeller, N.E.M. Page, R.K. Marshall, M.W.Wohlers, Tri-trophic effects of transgenic insect-resistant tobacco expressing aprotease inhibitor or a biotin-binding protein on adults of the predatorycarabid beetle Ctenognathus novaezelandiae, J. Insect Physiol. 54 (2008) 518–528.

[43] N.M. Green, Avidin and streptavidin, Methods Enzymol. 184 (1990) 51–67.[44] H.R. Nordlund, V.P. Hytonen, O.H. Laitinen, M.S. Kulomaa, Novel avidin-like

protein from a root nodule symbiotic bacterium, Bradyrhizobium japonicum, J.Biol. Chem. 280 (2005) 13250–13255.