Tastes, Structure and Solution Properties of

D-Glucono-1,5-lactone

Sneha A. Parke, Gordon G. Birch, Douglas B. MacDougall and David A. Stevens1

Department of Food Science and Technology, University of Reading, Whiteknights, PO Box 226, Reading RG66AP, UK

1On leave from Frances L. Hiatt School of Psychology, Clark University, 950 Main St, Worcester, MA 01610, USA

Correspondence to be sent to: Professor G. G. Birch, Department of Food Science and Technology, University ofReading, Whiteknights, PO Box 226, Reading, RG6 6AP UK

Abstract

D-glucono-1,5-lactone differs from D-glucopyranose only in that it has a C=O group instead of CHOH group at carbon

atom number one. The molecule therefore possesses an intact 3,4a-glycol group and is sweet. However, it

autohydrolyses in water solution at room temperature, forming D-gluconic acid and D-glucono-1,4-lactone. As the

solution pH falls it becomes sweet-sour and eventually almost completely sour as the generated hydronium ions

dominate both the solution properties and the taste perceptions elicited. It is shown that the ratio of generated

hydronium ions to unchanged lactone accords with anticipated taste quality during the first 28 min of

autohydrolysis. Changes in both apparent specific volume and apparent isentropic compressibility illustrate increasing

solute-solvent interaction and increasing disturbance of water structure during the course of autohydrolysis. These

changes are consistent with the concurrent sweet to sour change, but do not explain the weak bitterness which also

accompanies them. Chem. Senses 22: 53-65, 1997.

Introduction

The autohydrolysis of D-glucono-1,5-lactone incurs con-current structural and 'sapophore' changes in the solute.Such changes are measurable by solution property andpsychophysical parameters which should be mutuallyinterpretable (Shamil et al., 1987). The molecule is thereforeunique as a model for exploring the mechanisms of tastechemoreception. This paper reports changes in the tastequalities in relation to solution composition and solutebehaviour of D-glucono-1,5-lactone over the course of time.

D-Glucono-1,5-lactone (also known as glucono-delta-lactone or GDL) is an interesting sapid molecule with a

'glucophore' or AH,B system (Shallenberger and Acree,1967) identifiable at its 3,4a-glycol function. It differs fromD-glucopyranose only in the absence of the hemiacetalcentre, which does not, in any case, affect its taste (Birch etal, 1986). The molecule therefore predictably tastes sweet,but even when crystals are placed within the mouth, theinitial sweetness is soon accompanied by a sour sensation.The lactone is autohydrolysed by either water or oral fluid,producing D-gluconic acid, which is a mild acidulant.

In the organism, GDL is a normal metabolite of glucosedegradation by the pentose phosphate pathway. The

© Oxford University Press

54 I &A. Parkeetal

FAOAVHO Joint Expert Committee on Food Additives(JECFA, 1986) has given GDL a non-specified AcceptableDaily Intake. GDL has also been given its GRAS status in1986 by the Food and Drug Administration (FDA) in theUSA (Anon., 1989).

Gluconic acid, its salts and lactones are mild,non-corrosive, non-toxic compounds. Their ability to formchelates with metal ions in caustic solutions make themimportant to industry. Their physiological compatibilitymeans that they can be ingested without risk (Rajakylae,1981) and thus are ideal molecules for taste research.

GDL is an internal ester of gluconic acid. It is producedby the fermentation of glucose, followed by crystallizationof the lactone.

[ a ] D = lOOaJcL (1)

bio-oraditkm

C 6 H 1 2 O 6 -» C 6 H 1 2 O 7

H2O

C6H10O6

glucoie gluconic acid H2O glucono-l,5-lmcto!K

Because the taste sensation of aqueous solutions of GDLis complex, initially sweet and bitter, changing to sour overtime, GDL provides a means of studying the relationshipsamong chemical variables and taste quality. Accordingly,solution property measurements were carried out onsolutions of GDL in water over time, and the intensities oftheir taste qualities determined.

where a = observed rotation (°); c = concentration of thesolution (%); and L = length of polarimeter tube (dm).

Density and sound velocity values were determined usingan Anton Paar Density Sound Analyser (DSA 48) fromPaar Scientific Ltd, Raynes Park, London, UK. Temper-ature was maintained at 20 ± 0.1 °C. The density of thesample was measured from the period of oscillation of anoscillating U-tube. The sound velocity was calculated fromthe propagation speed of ultrasonic pulses in a knowndistance within the sample in the measuring cell. Theinstrument was calibrated using air and distilled water.Density and sound velocity measurements were accurate to±1 x 1O » g/cm3 and ±1 m/s respectively.

Apparent molar volumes (<|>v cm3/mol) and apparentspecific volumes (ASV cm3/g) were calculated from densityvalues using equations (2) and (3):

= 1000(4,- M2/d (2)

where do = density of water at one temperature (g/cm3); d -

density of solution at the same temperature (g/cm3); m =molality of the solution (mol/kg of water); and Mi =molecular weight of solute (g/mol):

ASV = <j>v/M2 (3)

Materials and methods

Chemical analysesChemicals used in solution measurements and psycho-physical studies were reagent grade and were obtained fromBDH, Lutterworth, Leicestershire, UK. Glucono-1,5-lactone was obtained from Sigma Chemical Co., Poole,Dorset, UK. Water used for solution studies was HPLCgrade. Solutions were made up w/w and all measure-ments were carried out at 20°C. Analyses were duplicatedto increase reliability. As no differences were foundbetween analyses, the results of the first replicate arereported.

Optical rotations were measured using an automaticdigital polarimeter (POLAAR 20, Optical Activity Ltd,Huntingdon, Cambridgeshire) at 589.3 nm. Specificrotations ([a]D°) were calculated using the followingequation:

Isentropic apparent molar compressibilities

cm3/mol.bar) were calculated from both density and sound

velocity values using the equation:

(4)

where ps = isentropic compressibility coefficient of solution(bar"1); and Pso = isentropic compressibility coefficient ofwater (bar"1).

Isentropic compressibility coefficients were calculatedfrom:

ps = 100/iA/ (5)

where u = sound velocity of solution (m/s).Partial molar volume, partial specific volume and isen-

tropic partial molar compressibility values were obtained atinfinite dilution, by extrapolating the best fit to the curve tozero concentration.

Tastes, Structure and Solution Properties of D-Glucooo-l,5-lactone I 55

Sensory analyses

Experiment I : changes in taste qualities ofGDL over time

SubjectsTen men and five women, ranging in age from 20 to 60 years(median age = 22 years), served as volunteer assessors.

ProcedureThe subjects were screened for sensitivity to sweet (fructose),sour (citric acid), salty (sodium chloride) and bitter (quininesulphate) materials, and their ability to identify thesequalities. Each person was given samples of the tastants inincreasing concentrations. They reported when a tastedifferent from water was detected, and when the quality wasidentifiable. All subjects showed normal sensitivity. Someconfusion between sour and bitter was found. When thisoccurred, the subjects were given practice in identifyingthose qualities. A week later, screening for the ability todiscriminate taste intensities was done. For this the subjectswere asked to rank three concentrations presented inrandom order, of each of the sweet, sour and bittermaterials, and identify their qualities. All subjects did thiswithout error.

A 10 % solution of GDL was tasted at 2, 7, 14, 21 and 28min after introduction of bottled water (Highland SpringLtd, Scotland) to the crystals. Samples were 10 ml and allmaterial was spat out after tasting. The mouth was rinsedwith bottled water before and after each tasting. Sweetness,sourness and bitterness were evaluated at different sessions,each a week apart, and then the series was replicated, againwith weekly session. Thus a total of six test sessions wereheld. Three per cent fructose, 0.25% citric acid and 0.003%quinine sulphate were given as reference materials prior tothe tasting of GDL in the respective sessions. For thereference materials and each tasting of GDL, the magnitudeestimates of the sweetness, sourness or bitterness wasindicated by panellists by marking an open-ended line; eachestimate was made with respect to the reference solution.

Experiment II: estimates of the sweetness ofGDL using tip of the tongueThe intensities reported by the panellists in experiment Iwere undoubtedly affected by mixture suppression, thereduction of intensity of tastes produced when two or moretaste qualities are experienced simultaneously (e.g.

Pangborn, 1961). Therefore the sweetness of GDL shouldbe determined under conditions designed to minimizemixture suppression. Pre-exposure to one of the twotastants in a mixture can release the suppression resultingfrom that tastant (Kroeze, 1979; Lawless, 1979). Pilot workwas conducted by holding citric acid in the mouth prior totasting GDL to determine if its sourness would adaptsufficiently to remove the effects of mixture suppression.However, evidence of potentially harmful effects of keepingthe acid in the mouth indicated this technique could not beused with assessors, and another method was sought.

Generally, the intensity of sourness and bitterness,relative to sweetness, of the appropriate tastants should beless when only the anterior tongue is stimulated than withwhole mouth stimulation (Collings, 1974; Pfaffman et al,1969; Sandick and Cardello, 1981). Pilot tastings of GDLsupported this expectation. Accordingly, to reduce thepossible effects of mixture suppression, the sweetness ofGDL was determined when only the anterior of the tonguewas stimulated.

SubjectsFifteen men and four women, with a median age of 24 years(range = 20-35) served as panellists. None had priorlaboratory experience in the scaling of taste intensities.

ProcedureThe method of magnitude estimation was described to theparticipants who then practised estimating the intensity ofthe hue of five samples of coloured water. When allunderstood the method and were comfortable makingmagnitude estimations, they were given labelled 25 mlsamples of sweet (5% glucose), sour (0.1% citric acid), salty(0.5% sodium chloride) and bitter-tasting (0.003% quininesulphate) solutions to remind them of those tastes. Thesamples were presented in 30 ml plastic cups (clear medicinecups, ART 1030, Roundstone Catering Equipment Limited,Wiltshire, UK). The assessors were told to extend theirtongue, press the lips around it and then insert the tongueinto the cup in order to taste the solution. They were told tonote the tastes while the tongue was in the solution, and toignore any tastes experienced during the subsequent waterrinse. All subsequent tastings were also done in this manner.

The participants were then given a modulus solution of8% glucose and told to report the number that bestrepresented the intensity of its sweetness. Any other taste,should there be one, was to be ignored. Every 2.5 min

56 I S.A. Parkert al

AH.B

glucono-1,5-lactone

ASV: 0.61 cmJ/g

Swect/billcr

H

HO

H

H

COOH

OH

H

OH

OH

glucono-1,4-Uctone

ASV : 0.63 cm'/g

Sweet/bitter

CHjOH

gluconic acid

ASV. 0 51 cmVg

Sour

Figure 1 D-Glucono-1, 5-lactone, D-gluconic acid and D-glucono-1, 4-lactone system.

thereafter, one of 17 samples was tasted and the intensity ofits sweetness estimated. Every third sample was a modulusand so identified. The panellists were told that its intensityshould be rated independently of their earlier experienceswith the modulus, and that the intensities of the othersamples should be rated relative to the last modulus tasted.The two samples presented after the modulus presentationswere randomly selected (without replacement) from fourconcentrations of glucose (4.0, 7.0, 10.0, 13.0%), citric acid(0.05, 0.1, 0.2, 0.3%) and 10% GDL solutions at ages 2, 7,12and 17 min since mixed with water. Because of constraintsdue to the age-times of the GDL solutions, they could begiven only on the 1st, 5th, 11th, 13th and 17th presentationsfollowing the initial modulus sample. The mouth was rinsedbefore and after each tasting. All material was spat out.

Results and discussion

Chemical analysesGDL is a molecule which changes both in structure andtaste in solution over the course of time. The changingchemical features of the GDL molecule are measured as

solution properties and these provide vital informationabout its hydrostatic packing characteristics within thethree-dimensional hydrogen-bonded structure of water.These chemical changes were then related to the change intastes as determined by psychophysical determinations.

On dissolution in water, glucono-1,5-lactone rapidlyautohydrolyses to gluconic acid (Figure 1) which in turneventually generates glucono-1,4-lactone. The concentrationof the acid increases to a maximum at ~3 h. Thiscorresponds to 87% of total solute as gluconic acid. Whengluconic acid is dissolved, on the other hand, it lactonizes tothe 1,4- and 1,5-forms. The higher the acid concentration,the greater is the rate of conversion of the acid toglucono-1,5-lactone and to glucono-1,4-lactone. The latteris a furanose form of lactone. Conversion of the acid intothe furanose form (1,4-lactone) is more favoured than intothe pyranose structure (1,5-lactone). Depending on the pHof the solution, direct interconversion of the lactones alsooccurs (Jermyn, 1960 ; Takahashi and Mitsumoto, 1963).

The pKa of gluconic acid has been determined to be 3.60(K^ = 2.51 x lO^4) at 20°C. Gluconic acid ionizes in solutionto gluconate ions and hydronium ions (Figure 2)

Autohydrolysis of GDL is slow at room temperature, with

Tastes, Structure and Solution Properties of D-Ghicono-l,5-lactoDe I 57

H"

HO"

H"

H-

COOH

"OH

COO"

H,O

-OH

-OH

CH,OH

gluconic acid

H-

HO-

H-

H-

OH

H

OH

OH

HjO'

hydronium ton

CHiOH

gtuconate ion

Figure 2 lonization of gluconic acid.

the formation of ~83% gluconic acid (a small proportion ofwhich is ionized), leaving 12% glucono-l,5-lactone and5% glucono-l,4-lactone after 3 days. The equilibriumproportions of the constituents vary with temperature,concentration, solvent and pH (Isbell and Frush, 1963). Thereaction is fastest in the first 3 h of hydrolysis as shown byHPLC (Combes and Birch, 1988) and by optical rotationplots (Figure 3).

Solution measurements carried out on a 10% solution ofGDL clearly show the changes taking place in the solution.

Density values initially increase as GDL (mol. wt 178.14g/mol) is converted to gluconic acid (mol. wt 196.16 g/mol),then decrease again as glucono-l,4-lactone is formed(Figure 4). Figure 4 also shows a similar trend in soundvelocity values with a rise as the number of ions in solutionincreases.

Molar and specific volumes and compressibilities were thencalculated from the density and sound velocity data at differenttimes during the initial hydrolysis of GDL (Table 1).

Apparent molar volume (<{>v) and apparent specific volume(ASV) are expected to increase as the pyranose ringstructure of GDL is converted to the acyclic form ofgluconic acid. Since the free acid form of gluconic acid fitspoorly with water structure, we would expect the packingefficiency of the molecules in water to be reduced as theconcentration of undissociated gluconic acid rises insolution. However, dissociation of the free acid to ionswould increase interaction with water and both <{>v and ASVmight be expected to decrease. This phenomenon probablydominates in the solution, therefore an overall decrease in <{)vand ASV is observed. This latter point directly relates to thesteady increase in sour taste with H+ since the freehydronium ions are the principal originators of that tastequality (Ganzevles and Kroeze, 1987). ASVs fall from 0.609cm3/g (the sweet range) to 0.582 cm3/g during the first 1.5 hof hydrolysis as the solution becomes more acidic (Figure 5).

Initial hydrolysis takes place so rapidly in solution that,

while the solute is being dissolved, some conversion into theacid form has already taken place. Therefore the ASV of thesolution at the start is estimated by extrapolating the initialpart of the ASV versus time curve to time = 0. This suggeststhat the ASV of GDL (0.61 cm3/g) lies in the sweet rangedespite the fact that the molecule tastes sweet/bitter by thetime that the solution is tasted (Shamil et ai, 1987).

Isentropic apparent molar compressibilities (,K^) can beregarded as the extent to which water of hydration aroundthe solute molecule can be compressed (Galema andHoiland, 1991). The following model briefly explains theconcept of compressibility:

Large positive

Large negative

cm /mol.bar)

Apolar, hydrophobic solutes

Carbohydrates (intermediate, negativeA (S) values)

Ions (K*,) = -30 to -50 x 10"*cm3/mol.bar)

Solutions become less compressible as A^s) values becomemore negative. Assuming that the solutes themselves areincompressible, it follows that water is most compatible withitself and has a large positive A^f) value. Moving down themodel, compatibility of the solute with water structuredecreases. In the case of carbohydrate molecules, the waterstructure is slightly disturbed by the hydrogen-bondednetwork around the solute which firmly holds the wateraround the solute, therefore ^ s ) decreases. Ions, with theirelectrostrictive forces, cause water structure to collapsearound the solute. This water of hydration is tightly heldto the ions, reducing compressibility even further.Compatibility with water is thus reduced (even thoughinteraction of the solute molecules with water molecules isincreased), and water structure is disturbed.

Change in isentropic apparent molar compressibility andhence the structure of water is easily explained in the case ofGDL (Figure 6). The lactone itself has an intermediate A^s)

value of ~-1.3 x 10~3 cm3/mol.bar. As gluconic acid isformed and the proportion of ions in solution increases, afall in isentropic apparent molar compressibility is observedfrom -1.307 x 10"3 to -2.072 x 1O-3 cm3/mol.bar from 3 to105 min of hydrolysis, indicating the disturbance of waterstructure caused by the ions. This results in a correspondingdecrease in isentropic apparent specific compressibility from-7.337 x 10-*to-1.155 x 10"5 cm3/g.bar.

The whole concept of compressibility fits in well with that

58 I S. A. Parkc <rf oi

Figure 3 Plots of observed and specific rotations of GDL solution over time.

1.04*0

1.0420 -

1.0416

1.0405

1 fMfrfl

1.0395

1.0390 -I

f\\• * • - -

' * • MAM •" • * • *

_

• - Dnty

- • — Sound valodty

i

• - • - .

-

• u - -

»•»

• • « • -

• •

1824

1523

1522

1521

1520

1519 '

1518

1517

Figure 4 Plots of density and sound velocity of GDL solution over time.

of apparent molar and apparent specific volumes (Figure 7).Low A^S), low <t>v and low ASV (as in the case of ionicstructures which impart saltiness and/or sourness) all showthat the solute molecule is heavily hydrated, thereforehydrophilic. This in turn implies that there is extensivehydrogen bonding and associated electrostrictive forcesbetween solute and solvent molecules, and therefore the

molecule can be more easily and rapidly transported to thereceptors. It can also, for that matter, be transported todeeper layers of the lingual epithelium where the appro-priate receptors lie (Birch et al., 1993). There is physiologicalevidence that salt and sour receptor sites lie deeper in thelingual epithelium than sweet and bitter ones (Green andFrankmann, 1987).

Tastes, Structure and Solution Properties of D-Gtucooo-l,5-lactoae I 5 9

Table 1 Solution property measurements on GDL solution

Time(min)

37

14

2128354249637791105

Average mol. wt(g/mol)

178.15178.15178.17178.20178.23178.30178.37178.45178.78178.97179.32179.43

Density(g/cm3)

1.03891.03901.03931.03971.04011.04051.04081.04111.04161.04191.04221.0425

Sound velocity(nVs)

1518.001518.141518.341518.811519.101519.591519.971520.451521.031521.201521.591521.98

Apparentmolar volume(cm /mol)

108.55108.39107.92107.28106.66106.07105.64105.23104.74104.44104.30103.91

Apparentspecificvolume(cm3/g)

0.6090.6080.6060.6020.5980.5950.5920.5900.5860.5840.5820.579

103 x Isentropic 106x Isentropicapparent molar apparent specificcompressibility compressibility(cm /mol.bar) (cm /g.bar)

-1.307-1.332-1.388-1.481-1.558-1.650-1.730-1.798-1.900-1.945-2.004-2.072

-7.337-7.479-7.792-8.314-8.740-9.255-9.463

-10.08-10.63-10.87-11.17-11.55

Initial glucono-1,5-lactone concentration = 0.6264 mol/l.

Sensory analyses

Experiment 1: Changes in taste qualities of GDLover timeThe magnitude estimates were normalized for each subjectseparately for each session by dividing the estimates forGDL by that given for that session's reference material.Thus, the normalized estimate of the intensity of a GDLsolution was proportional to the appropriate referencesolution. A normalized estimate of 1.0 indicates theintensity was equal to that of the reference solution for theindividual; an estimate of 1.5 indicates that the intensity wasone and a half times more intense than the referencesolution, etc.

Separate repeated-measures ANOVAs were performed forthe estimates of sweetness, sourness and bitterness of GDLaveraged over the two replications. Source tables for theseanalyses are given in Table 2, and plots of the meanestimates are shown in Figure 8. There were significantchanges in the perceived intensity of sweetness, sourness andbitterness as age of GDL increased. For sweetness, theintensity decreased about one-third as GDL aged from 2 to28 min. For the other two tastes, intensity increased ~4.5times for sourness and 2-fold for bitterness, from that at age2 min to age 28 min. Many of the panellists commented onthe sourness of the GDL solution as being extremelyunpleasant after 21-28 min.

Although the reference solutions were chosen to be

approximately equal on the basis of pilot tasting by theauthors, equality cannot be assumed for the individuals inthe panel, and meaningful comparisons of intensity cannotbe made between the taste qualities. However validcomparisons of the overall direction of change can be made.

The sweetness intensity of glucono-l,5-lactone through-out the first experiment was found to be lower than thesweetness of the 3% fructose reference supplied. Theintensity of bitterness at time = 25 min was the same as thatof the 0.003% quinine sulphate. For sourness, GDL time =17 min had the same intensity as that of the 0.25% citricacid, i.e. a mean intensity response of 1.0 (Figure 8). Thesefigures may well understate the intensity experienced ifmixture suppression were absent.

Experiment II: Estimates of the sweetness of GDLusing tip of the tongueEach individual's estimates were normalized by multiplyingthem by the quotient of 10.0 divided by the mean of theestimates for the moduli. Thus the sample estimates were allrepresented relative to each individual's estimate of 8%glucose.

While tasting with only the anterior tongue reduced thesourness of the taste of the GDL, some sourness remained.Comments by participants and inspection of the datasuggested that people differed in their ability to discriminatesweet and sour tastes with tongue immersion. Accordingly,for the analyses the participants were classified on the basis

60 I S-A-ParkertoA

109

109

108

10S

| 107

I 107| 106

106

106

108

104

\

\

•X

s

j

i

- • • Apparent Voter Volum*

• Apparent SpacHIc Volume

. . . . .

0.810

0.606

0.600

0.SS6

0.890 i

0.5S5

osrs10 20 30 40 60 80 70 80 90 100 110

Tkiw(inhB)

Figure 5 Plots of apparent molar and apparent specific volumes of GDL solution over time.

-14C-03

-2.1E-OJ

N

•

•ue-06

-ue-cw

-1.1E-08 s

-1.1&O8 1

-1JE-06

-1JE-06

40 •0

TkTW(mko)

100 110

Figure 6 Plots of isentropic apparent molar and apparent specific compressibilities of GDL solution over time.

of the ratio of their mean estimate of the sweetness ofglucose to that of citric acid. Those who reported therelative sweetness of glucose to be >4 times that of citricacid were classed as relative discriminators (n = 10), and theremaining as relative non-discriminators (n = 9). Thus thedata were analysed by separate split-plot ANOVAs for eachtastant. The main effects were age of GDL solution anddiscrimination class; for citric acid and glucose they were

concentration and discrimination class. Table 3 presents thesource tables for the analyses. Figure 9(a) shows the relativesweetness intensities of GDL as a function of age of GDLsolution, grouped by discrimination class, and Figure 9(b,c)shows the sweetness intensities of glucose and citric acid asa function of concentration, with participants grouped byrelative discrimination of sweetness in all three figures.

For the estimates of the sweetness of GDL, a significant

Tastes, Structure and Solution Properties of D-Gtucono-l,5-lactooe I 61

Tute Onalhv

Sab

Sour

Swce

Billet

ASV l im'/ i l

<O33

0 33 - 0.52

0.52-0.71

0.71-0.93

MorMore hydro phiBc molecules

More interactive whh water itnicture

I Receptor the* tie deeper in the

epithelium

"ftble 2 ANOVA source table, experiment I

Source of variation df MS

Figure 7 Apparent specific volumes and taste quality. (Source: Shamil eta/., 1987.)

interaction of age of solution and discrimination class wasfound. For those individuals classed as discriminators, thesweetness of GDL decreased as the GDL solution aged, andthe concentration of the lactone decreased. However,the non-discriminators, who were apparently confusingsourness with sweetness, reported a general increase inintensity as the solution aged, and the concentration ofgluconic acid and thus H+ increased.

For the estimates of sweetness of glucose, there wereno differences between the discriminators and non-discriminators; both showed very similar increases insweetness as concentration of glucose increased. For theestimates of the sweetness intensity of citric acid there wasno interaction of concentration of solution andclassification, but, as expected, there was a significant maineffect for classification. This difference in results for glucoseand citric acid supports the interpretation that theclassification into discriminators and non-discriminatorswas indeed based on a difference in response to a non-sweetquality; when only sweetness was involved, the groups didnot differ.

Since all three sweetness functions were obtained from thesame test sessions using a common modulus, comparisonscan be made between them. The linear regression functionfor glucose, using log-transformed data from the 19panellists, was used to determine the glucose concentrationproducing the sweetness of a 2 min old GDL solution. Thefunction was found to be y = 0.6368.x + 0.4451, r2 = 0.96,where y is the log magnitude estimate of sweetness and x thelog concentration of glucose (% w/v). Thus the sweetness ofthe 2 min old GDL was about that of 3% glucose.

Glucono-l,5-lactone and glucono-l,4-lactone have ASVsof 0.61 and 0.63 cm3/g respectively and they both tastesweet/bitter (Shamil et al., 1987). They possess the AH,Bsites on their molecules which is believed to cause sweetness(Figure 1). The bitterness may originate in the C=O groupwhich is always present in solution but increasing

SweetnessIndividualsAge of GDLResidual

SournessIndividualsAge of GDLResidual

BitternessIndividualsAge of GDLResidual

154

60

154

60

154

60

2.1340.4610.083

1.6559.6000.147

2.3621.4340.178

5.53 <0.001

65.39 <0.001

8.04 <0.001

concentration of gluconic acid may also elevate bitternessover the course of time.

Gluconic acid has an ASV of 0.51 cm3/g and this clearlyplaces it in the sour range of taste quality (Figure 7). Itdissociates in solution (pKt = 3.60) (Figure 2). As the pH ofthe solution falls during auto hydrolysis (Figure 12), theintensity of sourness increases. Even though a fall in pHfrom 3.82 to 2.5 causes the proportion of dissociation ofgluconic acid in solution to fall from 62.4 to 7.2% after 105min, there is still a large concentration of the acid already insolution which supplies the necessary hydrogen ionconcentration to maintain sourness (Table 4).

The concentration of lactone at time = 28 min (Table 4)suggests that there are enough molecules in solution to causesweetness; ~99.5 % of the solution is the lactone. But thesignificant fall in sweet sensation reported by panellists isprobably the result of mixture suppression by the sournessof gluconic acid. Hydrogen ion concentration increased by afactor of five from 2 to 28 min and this explains theexponential increase in sourness intensity. The relativelyweak intensity of sourness at the beginning of hydrolysiswas possibly due to the hydrogen ion concentration at time= 2 min (-1.5 x 10"4 M) being below the threshold forsourness.

Regression analyses at times 2, 7, 14, 21 and 28 min showthat pH and H+ concentration are highly correlated topercentage sourness (Table 5). We can therefore assume thatsourness is mainly due to protons in solution.

Regression analyses show that the sweetness of GDLfollows the molar concentration of lactone, and sournessfollows the concentration of H+. For experiment I, y =

62 I S.A. Parke et al

Figure 8 Experiment I (whole mouth tasting): mean intensity estimates for the sweetness, bitterness and sourness of GDL solution over time

•feble 3 ANOVA source table

Source of variation

GDLIndividualsDiscrimination classIndividual within class

GDL ageGDL age x discrimination classGDL age within class

GlucoseIndividualsDiscrimination classIndividual within classGlucose concentrationGlucose concentration x discrimination classGlucose concentration within class

Citric acidIndividualsDiscrimination classIndividual within classCitric acid concentrationCitric acid concentration x discrimination classCitric acid concentration within class

df

181

1733

51

181

1733

51

181

1733

51

MS

119.587541.15694.7897.355

44.69815.379

19.36313.91419.684

216.26214.14316.727

41.107421.402

18.73627.96226.14314.593

F

5.709

0.4782.906

0.707

12.9290.846

22.491

1.9161.791

P

<0.03

NS<0.05

NS

<0.001NS

<0.001

NSNS

100.07[lactone] - 55.61, r2 = 0.98, and for the discriminatorsin experiment II: \y = 2155.4[lactone] - 1208.7, r2 = 0.95. Theregression of sourness intensity judgements in experiment Iover H+ concentration produced y = 2077.3[H+] + 0.006, r2

= 0.96. The difference in slopes for the two sweetness

functions is likely to be due to differences in procedure(whole mouth versus anterior tongue tasting) and in effectsof mixture suppression. The slope was steeper for experi-ment II, where suppression was expected to be reduced bylocalisation of stimulus than in experiment I. This is the

Tastes, Structure and Solution Properties of D-Glucooo-l^-lacttme I 6 3

1818141210B6420

is

121086

(a) GDL

& °

(b) GLUCOSE

3.00 8.00 7.00 S.00Concentration (% w/v)

11.00 13.00

181614121 0 •

86 •

42

(c) CITRIC ACID-Dtscrimliutof*

' Non-tflscrtminatoni

0.03 0.08 0.13 0.18 0.23

Concentration (% w/v)0.28

Figure 9 Experiment II (tip of the tongue tasting): Sweetness intensities,relative to 8% glucose, for GDL, glucose and citric acid, grouped by subjects'discrimination of 'sweetness' vs sourness of citric acid. For GDL the resultsare shown as a function of age of solution. Table 4 gives the relevantconcentrations of lactone and gluconic acid. For glucose and citric acid theresults are shown directly as a function of concentration.

Table 5 Relationship between pH and hydrogen ion concentration torelative sourness of GDL solution

Time (min) Meansourness

PHa [H+ ]b

(mol/l)

27142128

V = 0.9345.

V = 0.9802.

0.310.530.771.161.69

3.843.693.363.233.12

1.445 x2.042 x4.365 x5.888 x7.586 x

1(T*10"4

10"4

10"4

10"4

expected result of reduced mixture suppression sincesuppression is related to concentration.

Conclusion

The solution chemistry of D-glucono-l,5-lactone and itshydrolysis products was explored in terms of solutehydration, hydrostatic packing characteristics and com-pressibilities. These parameters allowed the interpretationof some of the taste changes which accompany theautohydrolysis of the molecule. The autohydrolysis ofD-glucono-l,5-lactone is accompanied by a marked increasein sourness, which is explained by the increase in hydroniumion concentration. The small, but significant drop insweetness cannot, however, be attributed to any significant

Table 4 pH and concentration of constituents as GDL autohydrolyses over time

Time(min)

37142128354249637791105

PH

3.823.693.363.233.122.972.892.822.652.592.512.49

[H+](mol/l)

1.514E-042.042E-044.365E-045.888E-047.586E-041.072E-031.288E-031.514E-032.239E-032.570E-033.090E-033.236E-03

[Gluconic acid]total(mo|/l)

2.426E-043.701 E-041.195E-031.969E-033.049E-035.642E-037.895E-031.063E-026.042E-012.887E-024.111E-024.492E-02

[Lactone]as GDL(mol/l)

0.6260.6260.6250.6240.6230.6210.6180.6160.6040.5970.5850.581

% Lactoneremaining

99.9699.9499.8199.6999.5199.1098.7498.3096.4695.3993.4492.83

%aciddissociated

62.4055.1636.5329.9024.8818.9916.3214.2310.098.907.527.20

Initial glucono-1,5-lactone concentration = 0.6264

64 I S.A.Parkertoi

20 30 40 SO 60 70 60 SO 100 110

Figure 10 Change in pH and percentage dissociation of gluconic acid with time

change in lactone concentration but is presumably causedby mixture suppression. The results of this experimentconceptualize the use of combined solution effects forexplaining the taste qualities of sweeteners. For example,

molecules such as sodium saccharin and acesulfame Kpresent combined ionic solutes to the receptors which mightbe interpretable in a similar manner to D-glucono-1,5-lactone. Such sweeteners are now under investigation.

ACKNOWLEDGEMENTS

We thank the European Community (EC-AIR PL-94-2107) and the BBSRC (45/F02510) for funding in support of this work.

REFERENCES

Anon. (1989) Glucono-delta-lactone in cheese-making. Eur. DairyMag., 2, 61 -66.

Birch, G.G., Karim, R., Chavez, A.L and Morini, G. (1993)Sweetness, structure and specific volume. In Mathlouthi, M.,Kanters, J.A. and Birch, G.G. (eds). Sweet-tasteChemoreception. Elsevier, Essex, pp. 129-139.

Birch, G.G., Shamil, S. and Shepherd, Z. (1986) Role of theanomeric centre in sugar sweetness. Experientia, 42,1232-1234.

Cesaro, A. (1986) Thermodynamics of carbohydrate monomers andpolymers in aqueous solution. In Hinz, H.-J. (ed.),Thermodynamic Data for Biochemistry and Biotechnology.Springer Verlag, Berlin, pp. 177-207.

Collings, V.B. (1974) Human taste response as a function of locus ofstimulation on the tongue and soft palate. Percept. Psychophys.,16, 169-174.

Combes, C.L and Birch, G.G. (1988) Interaction of D-glucono-1,5-lactone with water. Food Chem., 27, 283-298.

Galema, S.A. and Hoiland, H. (1991) Stereochemical aspects ofhydration of carbohydrates in aqueous solutions. 3. Density andultrasound measurements./ Phys. Chem., 95, 5321-5326.

Galema, S.A., Howard, E., Engberts, J.B.F.N. and Grigera, J.R.(1994) The effect of stereochemistry upon carbohydratehydration. A molecular dynamics simulation ofp -D-galactopyranose and (ct,p)-D-talopyranose. CarbohydrateRes., 265, 215-225.

Ganzevles, P.GJ. and Kroeze, J.H.A. (1987) The hydrogen ion andthe undissociated acid as sour agents. Chem. Senses, 12,563-576.

Green, B.G. and Frankmann, S.R (1987) The effect of cooling thetongue on the perceived intensity of taste. Chem. Senses, 12,609-619.

Tastes, Structure and Solution Properties of D-Ghicono-l^-lactone I 6 5

Hoiland, H. (1986) Partial molar volumes of biochemical modelcompounds in aqueous solution In Hinz, H.-J. (ed.), Thermo-dynamic Data for Biochemistry and Biotechnology. SpringerVerlag, Berlin, pp. 17-44.

Hoiland, H. (1986) Partial molar compressibilities of organic solutesin water In Hans-Jurgen Hinz, (ed.), Thermodynamic Data forBiochemistry and Biotechnology. Springer Verlag, Berlin, pp.129-147.

Isbell, H.S. and Frush, H.L (1963) Lactonization of aldonic acids InWhistler, R.L. and Wolfrom, M.L (eds), Methods in Carbo-hydrate Chemistry. Academic Press, London, Vol. II, pp. 16-19.

Jermyn, M.A. (1960) Studies on the glucono-5-lactonase ofPseudomonas fluorescens. Biochim. Biophys. Ada, 37, 78-92.

Kroeze, J.H.A. (1979) Masking and adaptation of sugar sweetnessintensity. Physiol. Behav., 22, 347-351.

Lawless, H.T. (1979) Evidence for neural inhibition in bittersweettaste mixtures. J. Comp. Physiol. Psychol., 93, 538-547.

McCormick, R.D. (1983) The pH factor: choosing the optimumacidulant. Processed Prepared Food, 154, 106-113.

Moncrieff, R.W. (1967) The Chemical Senses. Leonard Hill, London.

Morris, V.J. and Chilvers, G.R. (1984) Cold setting alginate-pectinmixed gels./ Sci. FoodAgric, 35, 1370-1376.

Moskowitz, H.R. (1971) Ratio scales of add sourness. Percept.Psychophys., 9, 371-374.

Moskowitz, H.R. (1974) Sourness of acid mixtures./ Exp. Psychol.,102,640-647.

Pangborn, R.M. (1961) Taste interrelationships. II. Suprathresholdsolutions of sucrose and citric add. J. Food Sci., 26, 648-655.

Pangborn, R.M. (1963) Relative taste intensities of selected sugarsand organic acids. J. Food Sci., 28, 726-733.

Pfaffman, C, Fisher, G.L.and Frank, M.K. (1969) The sensory andbehavioral factors in taste preferences. In Hayashi, T. (ed.),011'action and Taste. Pergamon Press, Oxford, Vol. II, pp.361-381.

Rajakylae, E. (1981) HPLC analysis of adds formed in biochemical orcatalytic oxidations of glucose, Kemia-Kemi, 8, 811-812.

Sandick, B. and Cardello, A.V. (1981) Taste profiles from singlecircumvallate papillae: Comparison with fungiform profiles.Chem. Senses, 6, 197-204.

Sawyer, D.T. and Bagger, J.B. (1959) The lactone-acid-salt equilibriafor D-glucono-5-lactone and the hydrolysis kinetics for thislactone.V. Am. Chem. Soc., 81, 5302-5306.

Serpelloni, M., Lefevre, P. and Dusautois, C. (1990) Glucono-delta-lactone in milk ripening. Dairy Ind. Int., 55, 35-39.

Shallenberger, R.S. and Acree, T.E. (1967) Molecular theory of sweettaste. Nature, 216, 480-482.

Shamil, S., Birch, G.G., Mathlouthi, M. and Clifford, M.N. (1987)Apparent molar volumes and tastes of molecules with morethan one sapophore. Chem. Senses, 12, 397-409.

Takahashi, T. and Mitsumoto, M. (1963) Transformation andhydrolysis of D-glucono-yand 5-lactone. Nature, 199, 765-767.

Received on May 14, 1996; accepted on July 24, 1996