Astringency, a textural defect in dairy products · 218 L Lemieux, RE Simard INTRODUCTION ln a...

HAL Id: hal-00929393https://hal.archives-ouvertes.fr/hal-00929393

Submitted on 1 Jan 1994

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Astringency, a textural defect in dairy productsL Lemieux, Re Simard

To cite this version:L Lemieux, Re Simard. Astringency, a textural defect in dairy products. Le Lait, INRA Editions,1994, 74 (3), pp.217-240. �hal-00929393�

https://hal.archives-ouvertes.fr/hal-00929393

https://hal.archives-ouvertes.fr

Lait (1994) 74, 217-240© Elsevier/INRA

217

Review

Astringeney, a texturai deteet in dairy produets

L Lemieux, RE Simard

Centre de recherche STELA, Département de sciences et technologie des aliments,Pavillon Paul-Comtois, Université Laval, Sainte-Foy (Québec), Canada G1K 7P4

(Received 24 September 1993; accepted 9 February 1994)

Summary - Like bitterness, astringency is a defect liable to affect dairy products, but it has notbeen studied as extensively. This review article atternpts first to clarify the confusing use in theIiterature of the term 'astringent' to describe some associated texturai defects of dairy products andto better define astringency. Although astringency in dairy products might originate from two knowncauses: compounds resulting from heat treatment and proteolytically induced peptides (probably C-terminal breakdown products of 13-casein), the presence of phenolic compounds could also be aplausible cause. Astringent compounds reported to be present in many dairy products are listed, asweil as some methods of extraction and isolation.

astringency 1 dairy products 1 peptide

Résumé - L'astringence, un défaut de texture dans les produits laitiers. Parmi les défautssusceptibles d'être rencontrés dans les produits laitiers, l'amertume est celui qui a été le plusétudié. Toutefois, l'astringence mérite également que l'on y prête attention. Aussi, cet article tente,en premier lieu, d'éclairer la confusion qui existe sur l'utilisation du mot astringent dans ladescription de certains défauts de texture rencontrés dans les produits laitiers. Bien quel'astringence des produits laitiers soit attribuable principalement à des composés résultant de l'effetdes traitements thermiques et de la protéolyse, notamment peptides dérivés de la caséine {3, descomposés phénoliques pourraient également être impliqués dans ce défaut. Cette revue desynthèse traite des différents produits lactés dans lesquels l'astringence est rencontrée, ainsi quedes méthodes d'extraction et d'isolation de ces composés astringents.

astringence 1 produit laitier 1 peptide

218 L Lemieux, RE Simard

INTRODUCTION

ln a previous report (Lemieux and Simard,1991) on defects in dairy products, the fac-tors Iikely to influence the development ofbitterness, mainly in cheeses, werereviewed. Bitter peptides originating fromcaseins - their formation, isolation andidentification, structure, masking and inhi-bition - were reviewed in an exhaustivesecond report (Lemieux and Simard, 1992).Although the exact role of different proteo-Iytic enzyme systems in the developmentof bitterness is not fully understood, us1-casein has been recognized as al waysinducing more bitterness than ~-casein. So,when extracting peptides, more bitter pep-tides originated from ~-casein than from us1-casein are detected. However, the resis-tance of ~-casein toward the enzymesinvolved in the cheese ripening process ishigher than that of uS1-casein. So, whenextracting peptides, more bitter peptidesoriginated from ~-casein than from us1-casein are detected. This particularity cou Idexplain the tendency to associate ~-caseinwith bitterness.

As stated previously (Lemieux andSimard, 1992), there are five primary tastes:sweet, sour, salt y, bitter and umami (whichmeans 'agreable'). Not only the primary butalso the secondary tastes, such as astrin-gency, metallicity, and hotness, contribute tothe characteristic taste of foods. Mouthfeelproperties that are commonly defined in

Preparation variables

terms of body texture or consistency aredetected by the sense of touch or feel ratherthan taste. Combined with flavour, they areessential components of organoleptic qual-ity and so influence consumer acceptabil-ity. Viscosity, texture and astringency arethe principal criteria related to tactile prop-erties. However, there is confusion in the Iit-erature around the term 'astringency'.Whereas bitterness is perceived primarilyon the back of the tongue, astringency isdetectable throughout the oral cavity(Rouseff, 1990). Although astringency hasnot been studied as extensively as bitter-ness, it is typically encountered in reconsti-tuted powdered milk and sterilized milk prod-ucts.

Detectability, intensity, quality and reac-tions are the four classes of sensory tech-niques applied to foods and beverages. Thelast two of these are respectively presentedin figure 1 by several specific sensoryattributes and preferences (Land, 1983).Moreover, the quality of food consists ofsafety, nutritional value and acceptability,ail of which are interrelated.

The need to develop a rational systemand nomenclature for describing and trans-lating texturai qualities into precisely definedmeasurable properties was noted manyyears aga by Szczesniak (1963). She haspointed out several inherent problems withthe nomenclature used to describe texturaicharacteristics in terms of simple words.Thus, a number of terms may be used to

Seve rai Specilic Sensory Acceptance level

AlIributes Preferences

\1 SightEsweet-........

Production ~Smell E

E sally...........,

and storage Taste __ -variables -- E bitter;1 Touch -.......E

Physico-chemical Sound E astringent ,properties

etc.

Fig 1. Dairy products. Influence 01 different variables on their sensory quality.Produits laitiers. Influence de différentes variables sur leur évaluation sensorielle.

Astringency in dairy products

describe the same characteristic or thesame term may be used to describe sev-eral characteristics. Moreover, the sa meterm may have different meanings for dif-ferent people. In order to supply the foodscientists with a rational tool for a scientificdescription of food texture, Szczesniak(1963) proposed a classification of texturaicharacteristics. Accordingly, astringency isa mouthfeel sensation, a texturai charac-teristic that cannot be easily resolved on thebasis of mechanical (pressure exerted onthe teeth, tongue, and palate during eating)and geometrical (appearance of the foodproduct) properties. The same descriptiveterms are used to describe different flavours.

With good panel training, and by settingthe descriptive terminology of the test formand the scoring scale to detect the finestpossible differences, one should improvethe sensitivity of the flavour test for foodproducts and beverages (Clapperton andHarwood, 1983). However, dairy qualityjudging methods differ from the descriptiveanalysis techniques used for sensory eva-luation of many other food products. In fact,descriptive terms in dairy quality judgingoften refer to root causes of defects ratherthan simple sensory experiences. Manydefeet oriented terms are complex concepts,since a single root cause may result in thedevelopment of multiple, distinct f1avourattributes (Claassen and Lawless, 1992).For example, the general astringency termmay include sensory experiences describedas metallic, chalky, green f1avour, rancid andoxidized (Thurston et al, 1935; St-Laurent,1989).

Besides reporting about the term 'astrin-gency' used to describe different defects indairy products, this review will cover thework done in this field and also show itsimportance. There are at least two weil do-cumented different sources of astringency,heat treatments (Forss,1969; Harper andHall, 1976) and proteolytically induced pep-tides (Harwalkar et al, 1989). A third source,

219

the presence of phenolic compounds, isalso proposed.

COMPLEXITY OF ASTRINGENCYDEFINITION

The taste mechanism, with its three distinctparts, gustatory (tongue), tactile (touch),and olfactory (smell), depends primarilyupon chemical stimulation in establishingthe composite sensory response termedflavour. The tongue, with receptors locatedprimarily on its sides and base, serves asthe major organ of taste. Papillae of varioustypes can be noted chiefly at the tip, alongthe sides, and at the base of the tongue.The taste buds, with which the sapid sub-stance (in Iiquid form) must make contactbefore a taste sensation occurs, are locatedin many of these papillae (Bodyfelt et al,1988; Hughes, 1992). Astringency, a sen-sation associated with trigeminal ganglionsystems, cannot be easily defined, and isconsidered a tactile f1avour defect describedas a dry, puckery sensation in the mouth(Joslyn and Goldstein, 1964; Harwalkar,1972a; Singleton and Noble, 1976). Astrin-gency has been suggested by Bate-Smith(1954) to be a sensation of touch or pres-sure instead of taste. This sensation hasalso been characterized by the terms chalky,rough, dry, mouth coating or powderybecause it suggests finely divided insolu-ble particles (Boudreau, 1980; Bassette etal, 1986). Many workers have tried to under-stand the mechanism by which astringentcompounds impart a sensation in the oralcavity. Schiffman et al (1992b) showed thatin rodents, the signais for astringent com-pounds are transduced by the chorda tym-pani nerve (a taste nerve) rather than bythe trigeminal nerve. They later (Schiffmanet al, 1992a) found regional differences insensitivity to astringent compounds in theoral cavity of humans. Indeed, the glos-sopharyngeal nerve may be more sensitiveto astringent stimuli than the nerves inner-


vating the anterior tongue. According toFischer et al (1992), the salivary flow ratewould affect the temporal perception ofastringency. Subjects having a low salivaflow-rate have thus been found to take alonger time to reach maximum intensity forastringency, and have a longer duration ofthis taste. These subjects are also able torate the maximal astringency intensity ofwine higher than subjects having a highsaliva flow-rate.

A carry-over effect is encountered in thesensory assessment of intensely astringentsamples, and it appears to increase the per-ceived astringency of the next sam pie. Inorder to overcome this problem, a reversed-order paired comparison tasting procedurehas been devised (Arnold, 1983). The panelwas divided into two groups and they tastedtheir pair of samples in reversed order. Tocompare a set of 'm' samples, it was thusnecessary to have [m(m-1 )/2] sessions togive a complete replication of ail possiblepairs, as within a session the same pair ofsamples was assessed by ail testers, halfin each order (Guinard et al, 1986). Thisprocedure enabled any carry-over effect tobe evaluated and eliminated, thus givingmore precise estimates of parameters forcomparison.

One physical effect of the presence ofastringent substances is the destruction ofthe natural lubricant property of the salivavia precipitation of proteins andmucopolysaccharides in the mucous secre-tions of the salivary glands. The mucousmembranes of the palate and/or tongue tendto shrink (pucker). Since its perception inthe mouth is not instantaneous, astringencyis th us considered as a delayed effect. Dif-ferent groups of true astringents have beenlisted by Kudale (1970): 1) salts of multiva-lent metallic cations (aluminum, chromium,zinc, lead, calcium, magnesium, etc); 2)vegetable tannins (eg gallotannic acid); 3)dehydrating agents (eg ethyl alcohol, ace-tone, glycerine); and 4) minerai acids.

Thus reference solutions of aluminumpotassium sulphate (alum, 1.5 mmol/I), and0.08% w/v and 0.11 % w/w alum in waterwere respectively used to characterizeastringency in acid or cottage cheese whey,infant formulas (Mc Gugan et al, 1979;Malcolmson and McDaniel, 1980), and plainyogurt (Harper et al; 1991). A referencestandard solution of alum was also used forthe detection of astringency (900 ml springwater + 1.2 9 alum) in flavoured milk beve-rages (Lederer et al, 1991).

Since astringency is only detected wh enand if the sample is placed in the mouth, it istraditionally Iisted as a flavour defeet. Flavourmay be a sensory perception, a combina-tion of taste and aroma, which are perceivedrespectively by the papillae on the tongueand olfactory epithelium in the nose (Schultzet al, 1967; Joglekar and Gupta, 1981).Flavour is also defined as being the neteffect of several physiological reactionsinvolving various combinations of taste,odour, mouthfeel, pain, sensibility to tem-perature, and kinesthetic sensationsattributed to the muscular effort of chewing(Gardner, 1966). Moreover, the term 'f1avour'also includes the astringent sensory property(Tobias, 1990).

More reeently, Watson (1992) suggestedthat flavour is the additive effeet of taste andsmell while texture is a measurement of thephysical characteristics; it usually involvesthe senses of sight, touch and sound. Table1 shows the sense of gustation and its asso-ciated sensations (Watson, 1992).

SENSORY DEFECTS INCLUDEDUNDER ASTRINGENCY

Shipe et al (1978) Iisted defects associatedwith astringency in miscellaneous along withchalkiness, while metallic off-flavour wasassociated with an oxidized defect althoughit is frequently differentiated from this defect.The metallic defect is characterized by an

·Astringency in dairy products 221

Table 1.The sense of gustation and its associated sensations.Le goût et ses perceptions associées.

Receptor Stimulus Sensation

Tongue/mouth Water solubleChemicals taste (sweet, salt, acid, bitter and metallic)

feeling (astringency)

astringent, rough, puckery mouthfeel, anda metallic sensation similar to that observedwhen an iron nail or metal foil is placed in themouth, and is due to lactates resulting fromthe solubilization of various metals (iron,copper and copper alloys) in lactic acid solu-tions. It is also generally associated with theearly stages of metal-induced oxidation(Hunziker et al, 1929; Whitfield et al, 1936).Unhomogenized or cream-Iine milk is sub-stantially more susceptible to the develop-ment of the metallic oft-flavour th an homo-genized milk for reasons that are not c1earlyunderstood. However, since stainless steelhas replaced monel (white metal) in milkhandling and processing equipment, themetallic defect has substantially decreasedas a problem. However, the presence ofrust or copper in equipment or water sup-plies has to be avoided.

-,

The metallic off-flavour is quite frequentlyencountered in dairy products such as but-termilk, sour cream, yogurt, cottage cheeseand cream cheese, but may also be found inother dairy products, including butter, certaintypes of cheese and ice cream (Bodyfelt etal, 1988). The metallic off-f1avour can beimitated by adding 1-octene-3-one to milkor cream (Shipe et al, 1978; Bassette et al,1986).

Chalkiness is a texturai (consistency)defect which is usually detected as numer-ous, extremely fine, undissolved particleswithin the masticated product. This tactiledefect is described by Charalambous (1980)as being similar, if not identical, to astrin-gency. A chalky defect is frequently encoun-

tered in skim milk samples. The composi-tion of skim milk appears to favour occur-rence of this defect (off-flavour); it probablystems from the ratio of proteins to milkfatfound in skim milk (Bodyfelt et al, 1988).Mechanical and thermal treatments of milkcause the formation of aggregates contain-ing protein, fat, and lactose as weil as inor-ganic salts of varying composition. Depend-ing on their size, specific weight and electriccharge, these aggregates either sedimentor clump together on the surface of UHT orin-bottle sterilized milk, imparting chalkinessor astringency (Blanc et al, 1981; Schrôder,1983). According to Burton (1988), increas-ing homogenization pressure would help toreduce the occurrence or the intensity ofthe chalky defect in UHT milk.

Calcium fortification of cottage cheesehas caused chalkiness (Puspitasari et al,1991). However, as the report in questiondid not define the chalky defect or report asandy mouthfeel, it is difficult to know ifastringency was involved.

Another off-flavour, the 'green flavour',due to the formation of relatively high con-centrations of acetaldehyde by the cultureused in the preparation of cultured milk prod-ucts (buttermilk, sour cream, kefir, Bulgarianbuttermilk) may suggest a somewhat 'astrin-gent' character. This off-flavour can beattributed to the selection of the wrongmicrobial strain, incubation at incorrect tem-peratures, and/or overincubation (Bodyfelt etal,1988).

An 'astringent', puckery feeling perceivedat the base of the tongue and upper throat


may also describe the 'rancidity' noted incottage cheese and other dairy productssuch as cream, milk and butter (Bodyfelt etal, 1988). This defect is due to the action ofthe enzyme lipase on milkfat and can beprevented, provided the raw milk and creamsupplies were free of this defect, by properpasteurization of ail milk products used inmaking cottage cheese.

Various stages may be differentiated inthe 'oxidized' off-f1avour. A very slight inten-sity encountered in ice cream or ice milkmakes the product f1avour seem fiat or miss-

, ing whereas for a moderate intensity or upona longer holding in the mouth, the samplemay be described more accurately as'astringent' metallic, or puckery. Upon incu-bation for 3 months at ambient room tem-perature, non-dairy and cow's milk pow-dered creamers developed an astringentoff-flavour. This defect has been related toan oxidation phenomenon by Jolly andKosikowski (1974). Heat processing of milkhas an effect on its flavour; generally, themore extensive the heat treatment, the moresubstantial the f1avour change (Joglekar andGupta, 1981). However, the release or acti-vation of sulphydryl groups during high heattreatment of milk (> 7rC) provides a pro-tection against oxidized flavour (Tobias,1976; Joglekar and Gupta, 1981; Bassetteet al, 1986; Morr and Ha, 1991). Marketsurveys done by Jensen and Poulsen (1992)have shown that one of the f1avour defectsdeveloped by retail packed cream within thestated period of shelf-life is the oxidizedf1avour. Resulting from oxidation of the unsat-urated fatty acids in milkfat, the oxidizedflavour defect may be described by the termscardboardy and metallic. Copper-inducedoxidations affect the lipid fraction of milk; inthis case the phospholipids of the fat globulemembrane are attacked, leading to a card-boardy flavour. As skim milk is not free fromphospholipids, a copper-induced oxidizedflavour development is therefore possible(Schrôder, 1983). A scoring guide for astrin-

gency and its flavour attributes in dairy prod-ucts is suggested in table II.

MECHANISM OF THE ASTRINGENTSENSATION

More work is necessary before the mecha-nism of the astringent sensation can bedefined and the significance of the sensa-tion fully understood. Astringency compari-son of two or more stimuli presented sequen-tially is difficult. Moreover, bitterness andastringency are 'twin sensations': oral astrin-gency may indeed be closely linked to bit-terness (Singleton and Noble, 1976; Clifford,1986); untrained taste panellists may thusconfuse astringency with bitterness (Lea andArnold, 1978). In an attempt to determine ifthe tactile component (dryness) and asso-ciated taste (bitterness) component of astrin-gency were perceptually separable, Lymanand Green (1990) found that the perceptionof mouth dryness increased with repeatedexposure to an astringent stimulus, and thatincreased salivary volume reduced astrin-gency. However, this last observation wouldnot support the suggestion of Joslyn andGoldstein (1964), relating astringency to areduction in, or inhibition of salivary flow.Scores for astringent and bitter off-flavourin milk are listed in table III (Shipe, 1980).

From their work on English ciders Leaand Arnold (1983) found that both bitter-ness and astringency were present andwere due to phenolic procyanidins. Whilethe sensation of bitterness was predomi-nant in the oligomeric procyanidins (peakingwith the tetramer), that of astringency cameout top with the polymers.

THEORIES ABOUT ASTRINGENCYIN MILK AND MILK PRODUCTS

Flavour defects in milk may result from manycauses. According to Shipe et al (1978),

Table II. A suggested scoring guide for flavour of milk, cream, cultured milk products (buttermilk, sour cream, creamed cottage cheese and culturedcream butter) and dry milks.Guide numérique proposé pour l'évaluation des différents défauts de saveur rencontrés dans le lait, la crème, les produits laitiers de culture (lait debeurre, crème sure, fromage cottage, beurre de culture) et dans les laits séchés.

}>

Intensity Flavour defect a!:!l.S·

of defect lCœMilk and cream Cuitured milk product Drymilks ::>o

'<

Astringent Metallic Rancid Astringent Green Metallic Chalky Rancid Astringent Chalky Metallic RancidS·a.l!>~."0

Slight 8 5 4 7 8 6 8 4 8 8 4 5 aa.c:

Definite 7 3 1 5 7 4 5 2 6 6 2 3 $

Pronounced 6 1 0 3 6 2 2 0 0-4 0-4 0 0-1

a A numerical score of '10' is assigned ta a product Iree 01 undesirable f1avours (no criticism) while an assigned score 01 zero (0) is indicative 01 an unsalable product.

NNU)


Table III. Scoring of astringent and bitter off-flavour in milk.Evaluation numérique des défauts d'astringenceet d'amertume dans le lait.

Off-flavour intensity a Off-flavour

Astringent Bitter

SlightDefinitePronounced

875

531

,a A score 01 10 is given to a sam pie that has a 'typicalIresh' taste and is Iree 01undesirable Ilavours. Sampleswith off-Ilavours are scored lower.

seven categories of f1avour defects havebeen defined, and are listed in table IV.Astringency is allotted to miscellaneous, acategory which includes the flavour defectsthat either cannot be attributed to a specifie

cause or specifically defined in sensoryterms.

During World War Il, a reconstituted be-verage prepared with whole milk powderwas supplied to the US armed forces. How-ever, this product was qualified as astrin-gent and because of this defect, the recon-stituted beverage was generally rejected(Patton and Josephson, 1952). This gaverise to research into the intricate chemicalmechanisms involved in texturai changes,particularly astringency, as this defect hadbeen noted in raw and heated milk prod-ucts (Josephson, 1954). A chalky sensa-tion was also reported for homogenized milkshortly after its commercial introduction.However, this defect might have beencaused by improper pasteurization-homog-enization procedures (Shipe et al, 1978). ltwas thus imperative, in order to control andprotect the characteristic properties of dairyfoods, to know the good practices of milktransformation and maintain them in the fac-tories.

Table IV. Categories of flavour defects in milk.Classification des défauts de flaveur du lait et identification de leur origine.

Flavour alterations - descriptive or associative terms Causes

Cooked or sulphurous, heated or rich, Heat treatmentcaramelized, scorched

Oxidized (light-induced), burnt, activated, sunlight Exposure to various forms of radiant energy

Rancid, goaty, soapy, butyric Milk lipase catalysed hydrolysis of milk fattriglycerides

Acid, malty, fruit y, unclean, bitter, putrid Accumulation of the products ofbacterial metabolisrn

Oxidized, cardboard, cappy, metallic, tallowy Reaction between molecular oxygen and lipids(oxidation)

Feed, weed, cowy, barny Transmission (passage of substances from thecow's feed or surroundings to milk while it isin the udder)

Absorbed, astringent, bitter, chalky, chemical, Miscellaneousfiat, foreign, lack freshness, salty.


Hea~astringencyandheatrreatmenŒ

Milk is heat processed to eliminatepathogenic microorganisms which may bepresent in the raw milk and to increase theshelf-life. While heat processing can pre-serve the natural flavour by careful inacti-vation of the enzymes, treatments may alsoalter f1avour in a more or less desirable man-ner during concentration and preservationstages.

Data reported by Hutton and Josephson(1951) suggested that development ofastringency in heated whey was associatedwith heat denaturation and dehydration ofthe whey proteins. Milk salts were also pre-sumed to be involved in the astringentdefect. Raw skim milk and raw whey samplesalso developed 'roughness' when heatedmomentarily (10 min) at 95°C (Patton andJosephson, 1952). Following ultracentrifu-gation in a Sharples' super-centrifuge at 35000 rpm, the heat coagulable substancesfrom rennet whey were easily removed andfound to be responsible for the astringency.However, as these substances were recov-ered with difficulty from heated skim milk,Patton and Josephson (1952) suggestedthat case in might exert a protective colloidalaction toward the heat-coagulable whey pro-teins. Their observations have since beensupported by Josephson et al (1967) withelectron microscopy. Among the whey pro-teins, p-Iactoglobulin was presumed to bea significant contributor to the astringentdefect (Hutton and Josephson, 1951). Asthe astringency in dry milk was primarilyattributed to the formation of insoluble cal-cium-protein aggregates as a result of heattreatment and drying, Josephson (1954)proposed to overcome this defect bysequestering of calcium ion activity andremoval of calcium ions from the milk and/orlow temperature processing of the dry milk.Tessier and Rose (1958) have also reportedthat heating skim milk (82°C, 30 min) causedprecipitation of calcium phosphate and low-

225

ered [Ca2+]; skim milk was also found tocontain from 2.5 to 3.4 mmol Ca2+/1.

Modler and Emmons (1978) have founda way to utilize acid whey obtained fromcottage, bakers or cream cheese, whichwould otherwise be a serious disposai prob-lem. Although neutralization of this productwith an aqueous suspension of calciumoxide (CaO) 1 molli has helped in reducingthe sticking problem on dryer box andcyclone while spray drying, the concentratedwhey was chalky and astringent (Modlerand Emmons, 1977). Removal of insolublecalcium salts improved flavour and yieldeda final product with characteristics close tothose of sweet whey.

The increments in astringency obtainedby Josephson et al (1967) on lenghteningthe heat treatment of skim milk and rennetwhey at 90°C from 10 to 30 min are shownin table V. Josephson et al (1967) alsoobserved that, for a given heat treatment,a higher level of astringency was producedin rennet whey than in skim milk. Further-more, the astringent components in wheyexhibited considerably larger particle sizes,as evidenced by their sedimentation at amuch lower force (3 000 9 for 30 min) thanthat required for heated skim milk (144 OOOifor 2 h). In order to relate the size and struc-

Table V. Astringency intensity of heated skimmilks.Intensité de l'astringence chez des laits écrémésayant subi différents traitements thermiques.

Sample description Astringency intensity

1. Pasteurized skim milk(63°C-30 min) No astringency

2. Heated skim milk(90°C-10 min) Slight astringency

3. Heated skim milk(90°C-30 min) Distinct astringency


ture of aggregate particles to the lever ofastringency, Josephson et al (1967) con-ducted electron microscopie studies onheated whey systems: fractions free ofastringency exhibited no distinct particulatecomponents, while upon flash heating at80°C, spherically shaped electron-denseparticles within the 30-300 urn range wereformed. These particles were similar inappearance to caseinate micelles in skimmilk. As the degree of heat treatment wasincreased (80°C, up to 30 min), aggregatesin whey systems appeared larger and morenumerous, and the lever of astringency wasmore pronounced. It appears that particlesize and composition, rather than shape,accounted for astringency in heated wheysystems (Bodyfelt et al, 1988).

While Josephson et al (1967) concludedthat astringency in milk systems is primarilycaused by heat-altered whey proteins and tosome extent milk salts, both of which areassociated by interaction with or adsorptionon the caseinate micelles, Kratzer et al(1967) reported the presence of astringencyin raw milk sampi es stored for 48 h at 4.4°C.As a matter of tact, astringency had the sec-ond highest mean occurrence in raw milksamples (table VI).

The studies of Kratzer et al (1967)allowed the identification of many important

Table VI. Frequency of some flavour typesdetected in raw milk.Fréquence de détection de quelques types deflaveur dans le lait cru.

Flavour Mean occurrence (%) a

FeedAstringentCowyOxidizedFlat

88.412.711.06.14.1

a The percentage sum is more than 100 because morethan one f1avour was detected in many samples.

sources of variation in milk flavour (envi-ronmental factors, stage of lactation andage of the cow at calving, genetic effectsand panellists for the milk evaluation). Heatalone did not account for the occurrence ofastringency in raw or pasteurized milks anddry milk powders. It was therefore imperativeto isolate and characterize the astringentcomponents in raw milk and milk products toknow their source.

Astringency and casein breakdownproducts (proteolysis)

Harwalkar (1969) was the first to report onthe isolation of astringent compounds fromCheddar cheese. The isolated astringentfraction had the characteristics of a com-plex of associated peptides with an unusu-ally high content of hydrophobic amino acidresidues (leucine, tyrosine, valine, phenyl-alanine, etc) and an isoelectric point at pH6.5. Some other physical and chemical prop-erties of the astringent fraction are listed intable VII. Absorption in the 200 to 300 nmrange had a maximum in the region of 276nm. The low minerai values obtained for thisfraction led Harwalkar (1972a) to concludethat minerais were not involved in Cheddarcheese astringency. Strong astringency wasdetectable at acid or neutral pH but not athighly alkaline pH (>pH 10). This loss ofastringency at pH 10.5 coincided with theionization of the phenolic groups. Thechanges resulting in loss of astringency werefound to be reversible, since astringencywas restored by lowering the pH. The stateof ionization of phenolic groups was thusfound to be important for the astringent taste(Harwalkar, 1972a).

Methods for reducing heat-induced astrin-gency in skim milk were tested by Kudale(1970). His results showed that althoughtreatment with ultrasonic vibrations waseffective in reducing this defect, an objec-tionable 'bumt protein' flavour was produced.

Astringency in dairy products 227

Table VII. Some physical and chemical properties of the astringent fraction.Quelques propriétés physiques et chimiques de la fraction astringente.

Property Values

Absorption spectra in the 200 to 300 nm range

Sedimentation velocity, S20,w

Estimated molecular mass a

Mineral content: calciummagnesiumphosphorus

Carbohydrate

276nm

1.25

9 000 - 12 000 Da

0.04%< 0.02%~ 0.05%

Undetectable

a Obtained by gel filtration chromatography on Sephadex (Harwalkar and Ellioll, 1971; Harwalkar, 1972 a) and by sizeexclusion HPLC (Lemieux et al, 1989).

Proteolytic enzymes (commercial bovinepancreatic enzyme preparations, crystallinetrypsin and pancreatin), when used at properconcentrations, were found to be very effec-tive in reducing heat-induced astringencyof skim milk. These findings thus supportedthe hypothesis that astringency in heatedskim milk can be reduced by dissociationof aggregates and/or complexes of milk con-stituents formed as a result of heat treat-ment.

As reported by Demott (1971), astrin-gency may also be produced wh en milk isfortified with Iron salts, especially ferroussalts (ferrous sulphate, ferrous gluconate,or ferrous lactate). Shipe et al (1972) havereported that incubation of glass bead immo-bilized trypsin (2.5 9 in each of four paddlecompartments) in bulk milk samples withcopper added helped to reduce the oxidizedflavour. Moreover, astringency was detectedif the level of immobilized trypsin exceeded2.5 9 per cham ber. An excess of enzymewould th us increase proteolysis and causeastringency.

Later on, while investigating the influenceof pH on astringent compounds and their

extractability from cheese with a mixture ofchloroform and methanol (C/M; 2/1, v/v),Harwalkar and Elliott (1971) and Harwalkar(1972a, b, c) found that only the fractionsextracted from an acid Cheddar cheese orskim milk culture were astringent, whilethose from neutral or slightly alkaline sam-pies were bland. The requirement for acidpH (pH 4 or 3) for the extraction of astringentcomponents would thus indicate that theyare involved in a salt-type linkage. Thus atneutral or alkaline pH, astringent compo-nents would be anionic and accordingly formsalts with various cations, especially cal-cium, similar to the situation in milk proteins.Presumably, these salts are slightly solublein the C/M solvent. Solubility of astringentpeptides may possibly be influenced by theionization state; these components seemedto be less tightly bound to the curd as theionic strength was reduced, and washingthe curd with pH-adjusted distilled waterappeared to release astringent componentsfrom the curd. As the pH ls lowered thebound calcium salt dissociates, giving rise toionic calcium and peptides free from saltlinkage. Furthermore, a low pH hinders ioni-


zation of free carboxyl groups, resulting in anincrease in apparent hydrophobicity, andmay thereby facilitate extraction of polypep-tides by moderately polar (polarity <l>= 5.0)sol vents such as chloroform-methanol mix-tures (<l>CHCI3 = 4.3; <l>MeOH = 6.6; <l>CHCI3 :MeOH; 2:1 = 5.0).

Amino acids found in the methanolicphase (astringent fraction) of the six driedfractions (untreated original milk: whole milkor 10% non-fat dry milk; acidified milk; whey;unwashed curd; washed curd and wash-ings) extracted with the C/M solvent werepredominantly hydrophobie residues (Har-walkar, 1972c); this is a typical character-istic of y-casein. Moreover, polyacrylamidegel electrophoretic (PAGE) patterns for theextracts of unwashed curd, which wereastringent, showed at least three faint bandsin the region of whole y-casein. It is thereforetempting to speculate that residual proteo-Iytic activity or its reactivation is involved inthe formation of astringent components instored non-fat dry milk or pasteurized milk.This possibility seems plausible from thefact that B-casetns result from proteolyticbreakdown of ~-casein (Gordon et al, 1972;Gordon and Groves, 1975; Reimerdes,1978). Since then, Harwalkar et al (1993)have confirmed this relationship betweenastringency in pasteurized skim milk treatedwith psychrotroph proteinase and plasminand the production of breakdown products of~-casein. Indeed, they were able to show,using FPLC (fast protein liquid chromato-graphy) and urea-PAGE (polyacrylamidegel electrophoresis) techniques, that astrin-gent compounds from treatment of purified~-casein with plasmin were y-casein com-ponents (fragments of ~-casein consisting ofresidues 29-209 (Yd, 106-209 (Y2-) and108-209 (Y3-)' They have thus been able tolink astringency with y-caseins. Moreover,these product levels were higher in extractsfrom astringent samples.

According to taste-panel results forcheese grading, aseptically manufactured

Gouda cheeses have been found by Visser(1977b, c) to contain astringent compounds.However, they were completely lacking incheese f1avour (Visser, 1977a). This mayexplain why Visser (1977c) was cautiousabout the possible contribution, as sug-gested by Harwalkar and Elliott (1971), ofastringency to cheese flavour.

Studies on heat-treated milk sampleshave been done by Jaddou et al (1978).They observed that astringent off-f1avour inmilks heat-treated for 3 s (UHT) decreas-ing with storage time (slight astringency after7 d to no astringency after 84 dl, while itincreased Iinearly du ring storage of milksheat-treated for 90 s (sterilized) at 140°C(astringent after 1 month to very astringentafter 112 d storage). These observationshave allowed Jaddou et al (1978) to explainthe latter astringency by a physical reason.Although this milk was free of astringencydu ring the first several weeks of storage, itdid show signs of gelation after 2 months,implying changes in its colloidal structure.Harwalkar (1991) has also reported aboutdefects in UHT-sterilized milk particularlywhen stored at higher th an ambient tem-peratures. From studies conducted on UHTmilks (homogenized whole UHT milks (3.5%fat)) stored in the dark for 4 months, partly at20°C and partly at 1°C, Mottar (1981) hasobserved, even after very intensive UHTtreatments, residual proteolytic and Iipolyticactivities. When compared to indirectlyheated UHT milk, directly heated UHT milkpresented a more important loss in taste(drop in rating). Moreover, directly heatedUHT milk stored at 20°C showed a higherdegree of proteolysis and Iipolysis than whenstored at 1°C. It follows that the taste patternof UHT milk stored at 20°C is highly depen-dent on the quality of the raw milk, moreparticularly on the concentration of ther-moresistant proteases and lipases. Prote-olysis appears to affect the taste of UHTmilk stored uncooled more negatively th anlipolysis. Although ultra-high temperature


treatment strongly restrains the enzymaticactivities, it does not stop them completely.These observations would thus suggest agreater involvement for proteolysis than forlipolysis in astringency.

Some peptides isolated from caseintreated by Pseudomonas fluorescens andAlcaligenes enzyme preparations have beenfound to be astringent and astringent-sour.From these findings Kostyra (1983) has sug-gested the following: 1) the sensation ofastringency may depend on the hydroxyland amide groups present in the peptides.The hydroxyl groups would be responsiblefor the formation of hydrogen bonds andelectrostatic interactions with relevant oralcavity proteins, while the amide groupswould react with nerve endings; 2) depend-ing on the proportlon.ol sour amino acidsto their amides, the peptides may show pre-dominance of sour over astringent taste orvice versa. Alcafigenes enzyme prepara-tions also liberated bitter peptides, markedby the astringent effect of other peptides.As already mentioned, oral astringency maybe closely Iinked to bitterness which is itselfrelated to the average hydrophobicity of thepeptides (Clifford, 1986).

As suggested by Charalambous (1980)there may be more than one combinationof components that elicit an astringentresponse. Although, as previously men-tioned, astringency has been associatedwith milk products that have been processedat high tempe rature and with proteolysis(Harwalkar et al, 1989), the presence ofphenolic compounds should also be takeninto consideration.

Astringency and phenofic compounds

Chemicals causing astringency are collee-tively known astannins, the larger naturalpolyphenols (Hughes, 1992). The proan-thocyanidins or condensed tannins, whichcharacteristically bind and precipitate pro-

229

teins, are phenolic polymers synthesizedby many plants (Bate-Smith, 1954). Thecow eats a lot of these plants to producemilk. Astringency results from the interac-tion of phenols with proteins of the mouthand saliva (Singleton and Noble, 1976). Pro-teins and polypeptides with a high prolinecontent have a high affinity for tannin whileglycoproteins in general seem to have lowaffinity for astringent phenols (Hagermanand Butler, 1981). However, casein, whichis a glycoprotein found in milk, is relativelyrich in proline residues when compared tothe whey proteins. Thus the proline contentof milk proteins varies from = 1.6 to =16.8%(caseins: ~-: 35 Pro/209 residues, 16.75%;as2-: 10 Pro/207 residues, 4.83%; as1-: 17Pro/199 residues, 8.54%; J(-: 20 Pro/169residues, 11.83%; whey proteins: ~-Iact: 8Pro/162 residues, 4.94%; œ-lact: 2 Pro/123residues, 1.62%; BSA: 28 Pro/582 residues,4.81%). However, proteins are precipitatedby proanthocyanidins most efficiently at pHvalues near their isoelectric points, whereprotein-protein electrostatic repulsion is mini-mized. Thus proteins with acidic isoelectricpoints like BSA (pl 4.9) have greater affini-ties for tannin at pH 4.9 than at pH 7.8, andbasic proteins like lysozyme have higheraffinities at the higher pH (table VIII).

Although the specificity of the interactionis a function of the size, conformation andcharge of the protein molecule, ~-caseinshould have a high affinity for the tannins atacidic pH due to ils proline content. Harwalkar(1972a) has been able to detect astringencyat acid or neutral pH but not at highly alka-line pH (> pH 10). The changes in flavour atalkaline pH were reversible when the pHwas lowered. Proanthocyanidin-proteln inter-actions could thus be affected by the pH(Barbeau and Kinsella, 1983). Harwalkar(1972a) has observed absorption spectralchanges, a shift in the maxima from 280 to290 nm, for the astringent fraction from acidpH to alkaline pH. According to Ribéreau-Gayon (1968), the UV absorption wave-


Table VIII. Milk proteins, their isoelectric point and their relative affinity for tannins.Protéines du lait, leur point isoélectrique et leur affinité relative pour les tannins.

Whey proteins Caseins

BSA a-lect fi-Iact 1\-fi-

Isoelectric point a 5.14 5.13 4.24-4.76 4.6-5.1 5.3-5.84.9

Relative affinity for tannins BSA > o-lact ~ ~-Iact ~- > 1(-

a Eigel et al, 1984.

length of the simple phenols is around 270nm. Upon neutralization at alkaline pH, phe-nolic compounds become 'phenolates' andth us have different spectral characteristics.Preliminary studies conducted by Guinard etal (1986) on acidity-astringency interactionsin model solutions and wines have indicatedthat acidity significantly increased the astrin-gency of the solutions tested. Moreover,since the ability of condensed tannins toprecipitate proteins decreased sharply atpH 8.0, Guinard et al (1986) suggested thatit was due to a modification of the phenolionization of the tannins, making them there-fore unavailable for hydrogen bonding. Thusan acidic pH, by raising the percentage oftannins in the phenol form, can therebyincrease the likelihood of hydrogen bond-ing between dihydroxyphenol groups of milktannins and ketoimide groups of mouth pro-teins. Besides, Harwalkar (1972b) has con-cluded that astringent components shouldbe extracted at acid pH.

Table IX. Changes in milk flavour upon oxidationVariations dans la saveur du lait au cours de l'oxydation.

It has been observed, for English eiders,that oxidation during juice extraction wasaccompanied by a marked lowering in thelevel of the polymerie procyanidins, withconsequent reduction in astringency (Leaand Timberlake, 1974). During oxidation ofmilk, the flavour defect changes from fiat tooily (table IX).

According to Bruhn et al (1976), a goodquality of green feeds is needed to reduceIiability to oxidation as they have a high con-tent of the antioxidant tocopherol. Phenoliccompounds in green feeds possess manyOH groups which, in aqueous solution,induce many hydrogen bonds to formbetween the tannin molecules, resulting inaggregates of various sizes.

The phenolic compounds present in milkcould, upon oxidation, change the flavourof the product to an oxidized flavour andthus cause a decrease in astringency.Indeed, astringency has been shown to

Flat ---> Metallic --->AstringentPuckery

Oxidized --->PaperyCardboardy

OilyTallowyPainty

Fishy


decrease with time in wine and cider (Leaand Timberlake, 1974; Charalambous,1980). In Cheddar cheese the astringentfraction also decreased upon proteolysisduring the maturation period (Lemieux etal, 1989). Gamma-caseins are proteolyticfragments of p-casein. Moreover, f1avourdevelopment and stability to oxidation areinfluenced by proteolysis of the p-casein,which is dependent on the storage condi-tions (Reimerdes, 1978).

Ali these observations should thus sug-gest that phenolic compounds could alsobe responsible for astringency in dairy prod-ucts.

THE DEFECT OF ASTRINGENCYIN MILK AND MILK PRODUCTS

Following the preceding where researchhas been done to try to understand andexplain astringency in milk and milk prod-ucts, some other works are Iisted in Iiteraturewhere the authors report about this defect.This section will be concerned about theseobservations and ways to reduce astrin-gency.

Fresh, eoneentrated, driedand UHT milks

Identification of compounds responsible forthe stale f1avour defect in sterilized con-centrated milk was performed by Arnold etal (1966). They found that astringency wasabsent from fresh whole milk, slightly pre-sent in the control sterile concentrate andmore marked in stale sterile concentrate.

The flavour stability of evaporated milkproduced by three different methods of pro-cessing:1) the conventional method; 2) thehigh-temperature short-time (HTST)method; and 3) the aseptic method, wasdetermined on subsequent storage at 10and 2YOC(Sundararajan et al, 1966). A sig-

231

nificant effect of processing on the initialf1avour score was observed: aseptic pro-cess> HTST > conventional process. How-ever, regardless of the method of manu-facture, the flavour of the productsdeteriorated more rapidly at 27 than at 10°Cand showed defects such as astringencyand puckery.

ln a study of the effect of preheat treat-ment and storage of whole milk powdermanufactured from New Zealand milk,astringency, which was defined as thedegree to which the reconstituted milk left adrying, chalky feel in the mouth after swal-lowing, was evaluated. This texturai char-acteristic (astringency) slowly increased upto 15 months, and then decreased after 18months' storage; moreover, it was not sig-nificantly affected by the preheating condi-tions (Baldwin et al, 1991).

Harwalkar et al (1989) have come to theconclusion that the astringent compoundsin UHT-sterilized milk, analysed by PAGE,are y-casein-like breakdown products ofcasein. Although heating UHT-milk bydirect-steam injection would also produceastringency, it is still not known which of they-caseins are astringent and if there is anoptimum ratio required to elicit the astrin-gency response (Harwalkar et al, 1993). Forthis purpose the amine acid compositionand sequence of the astringent compoundsare required.

lee eream and iee milk

Ice cream and ice milk prepared by addinga 12% corn sweetener solution showed anastringent off-flavour. Sulphur dioxide (asbisulfite ion), 5-hydroxymethylfurfural (asweetener impurity), ethanal (acetaldehyde),methylpropanal (isobutyraldehyde) and 3-methylbutanal (isovaleraldehyde), the f1avourcomponents of the corn sweeteners, havebeen found responsible for this flavourdefect (Oison, 1970).


Cheese

ln the early seventies cheese makers foundthat kasal, a blend of phosphate salts (pri-rnarily sodium aluminum phosphate, basic)could be used as an emulsifying salt up to alevel of 3%. Besides improving melt, appear-ance, texture and other characteristics ofprocessed cheese, cheese spreads andcheese foods, kasal added no astringentoff-f1avours of its own (Anonymous, 1971).

ln order to evaluate the role of microbiallyderived Strecker-type volatile compoundsin commercial Cheddar cheeses, Dunn andLindsay (1985) added phenylacetaldehyde(500 ppb), an authentic Strecker-type volatilealdehyde, to c1ean-flavoured mild Cheddarcheese. They observed that phenylac-etaldehyde contributed astringent and bit-ter sensations to the cheese flavours.

Yogurt

Out of 12 sensory descriptors used todescribe plain yogurt flavour-by-mouth eva-luation, astringency has been found byHarper et al (1991) to be significantly dlt-ferent among samples (P:::; 0.001). Peopleare more conscious nowadays about theirhealth and want low-calorie products. Mistryand Hassan (1990,1992) have developed adelactosed, high milk protein powder(HM PP) which serves as a stabilizer in non-fat yogurts to improve body and to reducewhey separation. They have thus been ableto produce low-fatlnon-fat yogurts from skimmilk fortified with HMPP. However, whenthe protein content exceeds 5.6% (Hassanand Mistry, 1991) yogurts are too firm andhave an astringent off-f1avour.

Traditional cow milk yogurts sold in Brazilhave a protein content of 3.6%. However,the protein content of a yogurt analogue pre-pared with dry soymilk, whey solids and skimmilk powder could not exceed 3.1% due to avery marked astringent off-flavour. Paolielo

et al (1987) have observed that by addingcalcium sulphate in small quantities (0.104%,w/v), they could improve overall acceptabil-ity and attain the normal protein level of 3.6%without adversely affecting consumer accept-ability. Chien and Snyder (1983) havealready shown that the addition (10-50%v/v) of skimmed cow's milk was effective indecreasing the astringency of soymilk. Theopposite has also been done. Soybean isknown as an inexpensive protein and excel·lent energy source. For that reason soybeanmilk has been used to replace milk in sornefermented products. Cheng et al (1990)have thus prepared sogurt, a yogurt-like soy-bean product. However, sogurt, a 100%soymilk product, had an astringent tasteintensity higher than that of yogurt.

Whey protein concentrate

Although whey protein concentrate (WPC) isgenerally described as bland, flavourattributes have been associated with il.Thus, personnel tasting solutions at 5% pro-tein of 50 WPC samples with greater than80% protein attributed milky, sweet, soapy,bitter and astringent off-flavours to ail prod-ucts whereas the chalky defect was uncom-mon (Huffman and Marks, 1990).

Hydrolysis of raw defatted soy flour byDenapsin (an acid proteolytic enzyme) gavea product which had an excellent nutritionalvalue without an astringent off-flavour(Moretti, 1978). Later on, this hydrolysatewas used together with recovered cheesewhey protein in a proportion of 3:7 (wheyprotein: hydrolysed soy protein) for thepreparation of a soluble protein mixture neu-tral in flavour.

Flavoured milks and formulas

High carbonation level (1.42 CO2 volumes)increased the sourness and astringency of


strawberry, raspberry, peach and root beerflavoured milks (Lederer et al, 1991). Sincethe taste perception of sourness often stim-ulates the feeling of astringency (Moncrieff,1971) and since some acids are perceivedas more astringent than sour, it is suggestedthat the panellists could have been con-fused in the astringency evaluation andcould have then responded to increasedacidity in the flavoured milks. Thus, raisingacidity has previously been suggested toincrease the intensity of astringency in wine(Guinard et al, 1986). Moreover, the rate ofsaliva production would have th us beenaffected (it would have increased) by thechemical composition of the beverage,affecting the sensory magnitude of the vis-cosity (Szczesniak, 1979).

ln a study which evaluated the sensoryproperties of formulas commonly replacinghuman milk (commercially prepared formu-las based on cow's milk, home-made for-mulas prepared using whole and evapo- ,rated milk), Malcolmson and McDaniel(1980) found them to be ail astringent, but invarying intensities. Only the whole milk for-mulas were found to be significantly lessastringent th an a reference sample madeof a 0.08% alum solution and consideredextremely astringent.

EXTRACTION AND ISOLATIONOF ASTRINGENT COMPONENTS

This section is a summary of the differentmethods dealing with the extraction and iso-lation of astringent components from dairyproducts. Aqueous suspensions or solu-tions have been tasted for astringency bya trained panel or experienced judges. Itcan be observed that technologies used forthe study of astringent components becomemore and more sophisticated with years.This way, from aggregate particles throughastringent peptides, the information obtained

233

is getting close to the identification of theastringent components.

According to Josephson et al (1967)selected milk systems, skim milk, whey andultrafiltrate systems, have been heated atvarying levels (skim milk: 82°C, 1 h or 80°C,30 min; whey: 80°C, flash to 30 min; ultra-filtrate systems: boiling bath: 10-30 minor 80°C, 30 min) to produce astringency.The astringent components were then sed-imented by ultracentrifugation at differentforces. Gel filtration (G-100 and G-200) onSephadex, polyacrylamide gel elec-trophoresis (PAGE), electron microscopyand chemical analyses were carried out tostudy changes in the size, shape and com-position of protein-salt aggregate particlespresent (Josephson et al, 1967).

The astringent components can beextracted from raw or pasteurized milk,UHT-milk and non-fat dry milk by means ofthe method proposed for isolating bitter andastringent compounds from Cheddar cheese(Harwalkar and Elliott, 1971). The majormodification is the adjustment of the milk topH 5.2 and freeze-drying before extractionby C/M (Harwalkar, 1972b, c). The astrin-gent components in these extracts havebeen partly purified by isoelectric precipita-tion at pH 7.0 and by size-exclusion chro-matography on Sephadex G-50 columnsusing water as a mobile phase. They havebeen further analysed by fast-protein liquidchromatography (FPLC) using a Mono-Qanion-exchange column and PAGE.

Ali the studied samples of aseptically (nor-mal, starter-free, rennet-free and rennet- andstarter-free) manufactured Gouda cheesescontained astringent compounds which havebeen isolated by Visser (1977b, c). The elu-tion profile of the astringent fraction obtained,according to the method of Harwalkar andElliott (1971), from a normal aseptic cheeseafter molecular sieving on Sephadex G-50with 0.01 N acetic acid, showed two frac-tions with a molecular mass of approximately7 000 and 3 000 Da, respectively.


ln order to analyse the chemical compo-sition of Swiss cheese and to relate it to itscharacteristic taste, Swiss cheese sampleshave been fractionated into oil-soluble,water-soluble-volatile and water-soluble-non-volatile components. This last fractionwas found to contain astringent componentswhich were further separated according to amodification of the procedure of Kirimura etal (1969). This process consisted of passingthe water-soluble-non-volatile fractionthrough 5 successive Oowex 50 columns(20 x 200 mm) with decreasing degree ofcross-linkage (x16, x12, x8, x4 and x2,respectively) to allow the separation ofamino acids and peptides of different sizes.The average chain length of the astringentpeptides (APL) was found to increase from1.66 to 6.06 with the reduced degree of resincross-linkage (Biede and Hammond, 1979)and it was calculated according to the for-mula (Takeuchi et al, 1967):

average peptide chain Total nitrogenlenqth (APL) or average = ------number of amino acids amino-nitrogenin the peptides

Astringent fractions have been isolatedaccording to the method of Harwalkar andElliott (1971) from Cheddar cheese madewith added Lactobacillus strains to accel-erate ripening. Size-exclusion high-perfor-mance liquid chromatography (HPSEC) ona TSK-G2000 SW column of these astrin-gent fractions has aliowed the estimation oftheir molecular masses at approximately84 000-18 000 Da (Lemieux et al, 1989).Oinakar et al (1989) have also used themethod of Harwalkar and Elliot (1971) toextract astringent components from Cheddarcheese prepared from cow or buffalo milkwith either calf rennet or a soluble partiallypurified enzyme from Withania coagulans.Upon tasting, the extract from control cheesewas found to be more astringent th an bit-ter.

Knowing that the appearance of off-flavours in UHT milk is partly due to the pro-teolysis, L6pez-Fandiiio et al (1993) haveused reversed-phase HPLC (RP-HPLC) toinvestigate the degree of protein breakdownduring storage of UHT milk. As measuredby SOS-PAGE, comparison of indirectlyheated UHT milk samples with directlyheated UHT milk samples showed a highlevel of casein degradation and gave riseto peptides which eluted below or at a con-centration of 35% solvent B (0.1 % trifluo-roacetic acid in acetronitrile). The peptidesso obtained were derived from the actionof Pseudomonas f1uorescens strain B52proteinase on casein and eluted before theproteose-peptone, a-Iactalbumin and 13-lac-toglobulin. According to L6pez-Fandiiio etal (1993), this RP-HPLC method could helpto predict the risk of deterioration on stor-age of freshly prepared milks and wouldthrow out a hint that the origin of astringencyin UHT milk might not be due to whey pro-teins.

ASTRINGENT PEPTIDES

Kirimura et al (1969), in their study on thecontribution of peptides and amino acids tothe taste of foodstuffs, found that dipeptidesinvolving the y-COOH group of glutamic acid(Glu) in the peptide linkage are astringent.The assumption is supported by Nishimuraand Kato (1988) who established that thepeptides listed in table X are astringent.

Astringency has been found to be oneof the eight flavour characteristics of acidor cottage cheese whey. In order to increasethe utilization of whey in foods, McGuganet al (1979) have mixed it with skim milk invarious proportions. Since astringency wasfound to be present at different levels, theyhave thus reported that change in astrin-gency cou Id result from changes in the ion-ized forms of astringent peptides.


Table X. Dipeptides tasted in 0.2% aqueous solu-tion and evaluated by a trained panel as beingastringentDipeptides ayant été évalués à une concentra-tion de 0,2% en solution aqueuse et identifiéscomme étant astringents par des panellistesentœînés.

Y-L-Glu-L-Valy-L-Glu-L-Leuy-L-Glu-L-lley-L-Glu-L-Trp

y-L-Glu-L-Tyrby-L-GIU-L-Pheb

y-L-Glu-L-Aspay-L-Glu-L-Glua

y-L-Glu-Glyay-L-Glu-L-Alaay-L-Glu-L-Sery-L-Glu-L-Thry-L-Glu-L-Pro

a Accompanied by sournessb Accompanied by sourness and bilterness.

CONCLUSION

Seing handled extensively between pro-duction and consumption, milk cornes intocontact with many plant surface equipmentsand it is thus prone to many potential sen-sory changes. These are the main causes ofpotential sensory changes in milk: i) acci-dentai contamination with chemicals; ii) milktreatments, in particular thermal process-ing; iii) and reactions during storage due to:a) microbial growth; b) enzymatic or oxida-tive activity.

Knowing that a successful sensory ana-Iysis depends more on people than on meth-ods, experimental designs and computers,a good panel training and a weil defineddescriptive terminology of the test form andof the scoring scale are th us necessary.

To describe sorne associated texturaidefects of dairy products, there is no statedconsensus among the sensory evaluatorsof dairy products in their understanding anduse of the attributes or terms for ratings.Thus regarding the confusion around theuse of the term 'astringent' a consensus indairy judging becomes imperative.

From the results mentioned above itappears that astringency arises in treated

235

milk from at least three different sources:1) from heat-induced association of wheyproteins and calcium phosphate with thecaseinate system (Josephson et al, 1967);2) they may be proteolytically induced pep-tides. In non-fat dry milk samples the astrin-gent components have bee'n shown to bepolypeptides associated with the casein frac-tion, particularly y-casein (Harwalkar, 1972c).Moreover, minerai content (calcium, mag-nesium and phosphorus) has notbeenshown to be involved, and heat alone can-not account for this defecl. Astringent com-ponents can be extracted only from acidi-fied milk; and 3) they may involve phenoliccompounds which are always present in thefeeds given to cows.

Ultrasonic vibrations and treatment withproteolytic enzymes (trypsin and pancre-atin) have been found to reduce astringencyin heated skim milk by the dissociation ofaggregates and lor complexes of milk con-stituents formed as a result of heat treat-ment (Kudale, 1970). One could establishpoints of similarity between Kudale's obser-vations and cheese proteolysis, where theastringency is reduced through the degra-dation of proteins (caseins) to free aminoacids. Indeed, proteolytic breakdown dur-ing the cheese ripening process (rennet,starter strain, bacteria, etc) implies thedegradation of high molecular masspolypeptides (astringent fraction, with Mrof'" 84000 to 18000 Da) to peptides and freeamino acids (Lemieux et al, 1989) with aloss of astringency.

Although the role of peptides contributingto the bitter taste in cheese has been stud-ied in detail, further research is necessary toexplain the contribution of peptides to foodtaste. The consumer is the final and deci-sive judge of eating quality. Acceptance ofa food is influenced to a great extent by thetactile properties exhibited by il. Producersshould th us supply food products with noassigned defecl. They th us should try tominimize the effects of processing. More


interest should also be directed towards theequilibrated sensation to make a dairy prod-uct appetizing and satisfying so that the con-sumer will appreciate and become fond of il.For example, for a frozen dairy dessert toretain its refreshing and pleasing charac-teristics, the mouthfeel should not be exces-sively cold, dry or astringent (Bodyfelt et al,1988).

Astringency should stimulate more inter-est among workers involved in the field ofdairy products. More effort should thus beinvolved in the astringent defect in dairyproducts to try to understand it and find theexact causes.

ACKNOWLEDGMENT

We would Iike to thank Dr VR Harwalkar for hishelpful suggestions during the preparation of thismanuscript.

REFERENCES

Anonymous (1971) Phosphate blend saves up to$6000 per million lb processed cheese. FoodProcess, 32, 20-21

Amold GM (1983) A tasting procedure for assess-ing bitterness and astringency. In: Sensoryquality in foods and beverages: de finition,measurement and control (Williams AA, AtkinRK, eds) Ellis Horwood Limited, Chichester,England,109-114

Arnold RE, Libbey LM, Day EA (1966) Identifi-cation of components in the stale flavor frac-tion of sterilized concentrated milk. J Food Sci31,566-573

Baldwin AJ, Coope HR, Palmer KC (1991) Effectof preheat treatment and storage on the prop-erties of whole milk powder. Changes in sen-sory properties. Neth Milk Dairy J 45, 97-116

Barbeau WE, Kinsella JE (1983) Factors affectingthe binding of chlorogenic acid to fraction 1leaf protein. J Agric Food Chem 31 , 993-998

Bassette R, Fung DYC, Mantha VR (1986) Off-f1a-vors in milk. CRC Crit Rev Food Sci Nutr24,1-52

Bate-Smith EC (1954) Astringency in foods. FoodProcess Packag 23, 124-127, 135

Biede SL, Hammond EG (1979) Swiss cheeseflavor. 1. Chemical analysis. J Dairy Sci 62,227-237

Blanc B, Odet G, Adda J, Bosset JO, Rüegg M(1981) New monograph on UHT milk. Appear-ance, flavor and texture aspects: recent devel-opments. Int Dairy Fed Bull 133,25-48

Bodylelt FW, Tobias J, TroutGM (1988) The sen-sory evaluation of dairy products. Reinhold,Van Nostrandreinhold, New York, 598 p

Boudreau JC (1980) Taste and the taste of foods.A review and a report on a Symposium. Natur-wissenschaften 67,14-20

Bruhn JC, Franke AA, Gobie GS (1976) Factorsrelating to development of spontaneous oxi-dized flavor in raw milk. J Dairy Sei 59, 828-833

Burton H (1988) Ultra-high-temperature pro-cessing of milk and milk products. Elsevier,Barking, 354 p

Charalambous G (1980) The analysis and controlof less desirable fla vors in foods and bever-ages. Academie Press, New York, 358 p

Cheng YJ, Thompson LD, Brittin HC (1990)Sogurt, a yogurt-Iike soybean product: devel-opment and properties. J Food Sci 55, 1178-1179

Chien JT, Snyder HE (1983) Detection and con-trol of soymilk astringency. J Food Sci 48,438-440

Claassen M, Lawless HT (1992) Comparison ofdescriptive terminology systems for sensoryevaluation of f1uid milk. J Food Sci 57, 596-600,621

Clapperton JF, Harwood JB (1983) Flavourdescription and f1avour classification. In: Sen-sory quality in foods and beverages: definition,measurement and control (Williams AA, AtkinRK, eds) Ellis Horwood Limited, Chichester,England, 82-88

Clifford MN (1986) Phenol-protein interactionsand their possible significance for astringency.ln: Interactions of food components (BirchGG, Lindley MG, eds) Elsevier, London, 143-163

Demott BJ (1971) Effects on flavor of fortifyingmilk with Iron and absorption of the Iron fromintestinal tract of rats. J Dairy Sci 54, 1609-1614


Dinakar P, Mathur MP, Datta Roy D (1989) Dif-ferences in proteolytic behaviour in Cheddarcheese prepared with calf and vegetable ren-net. Indian J Dairy Sei 42,792-796

Dunn HC, Lindsay RC (1985) Evaluation of therole of microbial Strecker-derived aroma com-pounds in unclean-type flavors of Cheddarcheese. J Dairy Sei 68,2859-2874

Eigel WN, Butler JE, Ernstrom CA, Farrell HM,Harwalkar VR, Jenness R, Whitney R Mcl R(1984) Nomenclature of proteins of cow's milk:fifth revision. J DairySei67, 1599-1631

Fischer U, Boulton R.B, Noble AC (1992) Effect ofsalivary flow rate on temporal perception ofbitterness and astringency. Chem Senses 16,521-522

Forss DA (1969) Flavors of dairy products: areview of recent advances. J Dairy Sei 52,832-840

Gardner WH (1966) Food acidulants. AlliedChemical Corp, New York, 185 p

Gordon WG, Graves Ml (1975) Primarysequence of beta, gamma, and minor caseins.J Dairy Sei 58, 574-582

Gordon WG, Groves Ml, Greenberg R, JonesSB, Kalan EG, Peterson RF, Townend RE(1972) Probable identification of y-, TS-, R-and S- caseins as fragments of ~-casein.J Dairy Sei 55, 261-264

Guinard JX, Pangborn RM, Lewis MJ (1986) Pre-Iiminary studies on acidity-astringency. Inter-actions in model solutions and wines. J SeiFood Agric 37, 811-817

Hagerman AE, Butler lG (1981) The specificity ofproanthocyanidin-protein interactions. J BiolChem 256, 4494-4497

Harper WJ, Hall CW (1976) Dairy technologyandengineering. The AVI Publishing Co Ine, West-port, Connecticut, 631 p

Harper SJ, Barnes Dl, Bodyfelt FW, Mc DanielMR (1991) Sensory ratings of commercialplain yogurts by consumer and descriptivepanels. J DairySci74, 2927-2935

Harwalkar VR (1969) Isolation and partial char-acterization of an astringent tasting compo-nent from Cheddar cheese. J Dairy Sei 52,885

Harwalkar VR (1972a) Characterization of anastringent !lavor fraction from Cheddar cheese.JDairySci55,735-741

237

Harwalkar VR (1972b) Influence of hydrogen ionconcentration on extractability and flavor ofbitter and astringent flavor components fromCheddar cheese and cultured milk. J DairySei 55, 742-743

Harwalkar VR (1972c) Isolation and partial char-acterization of an astringent fraction from milkand nonfat dry milk. J Dairy Sei 55, 1400-1404

Harwalkar VR (1991) Storage defects in UHT-sterilized milk. J Dairy Sei 74 (suppl 1),81

Harwalkar VR, Elliott JA (1971) Isolation of bit-ter and astringent fractions fram Cheddarcheese. J Dairy Sei 54,8-11

Harwalkar VR, Boutin-Muma B, Cholette H,McKellar RC, Emmons DB, Klassen G (1989)Isolation and partial purification of astringentcompounds from ultra-high temperature ster-ilized milk. J Dairy Res 56,367-373

Harwalkar VR, Cholette H, McKellar RC, EmmonsDB (1993) Relation between proteolysis andastringent off-f1avor in milk. J Dairy Sei 76,2521-2527

Hassan HN, Mistry VV (1991) Production of lowfat yogurt from a high milk protein powder.J DairySci74 (suppl 1), 96

Huffman lM, Marks PK (1990) Characteristicflavours of whey protein concentrates. In: BriefCommunications and Abstracts of Posters ofthe 23rd Int Dairy Congr (FIL). Montreal,Canada, 511 p

Hughes N (1992) The tasting phenomenon. NutrFood Sei 23, 4-6

Hunziker OF, Cordes WA, Nissen BH (1929)Metals in dairy equipment. Metallic corrosionin milk products and its effect on f1avor. J DairySei 12,140-181

Hutton JT, Josephson DV (1951) A tactual fla-vor defect of dried milk. J Dairy Sei 34, 488

Jaddou HA, Pavey JA, Manning DJ (1978) Chem-ical analysis of flavour volatiles in heat-treatedmilks. J Dairy Res 45, 391-403

Jensen GK, Poulsen HH (1992) Sensory aspects.Bulllnt Dairy Fed 271, 26-31

Joglekar S, Gupta SK (1981) Effect of heat pro-cessing on the flavour of milk and milk prod-ucts. Indian Dairyman, 33,143-147

Jolly RC, Kosikowski FV (1974) Fat characteris-tics and stability of nondairy and dairy pow-dered creamers. J Agric Food Chem 22, 295-298


Josephson DV (1954) Flavor origins - Review ofchemical mechanisms affecting flavor accept-ability of dairy products. J Agric Food Chem 2,1182-1185

Josephson RV, Thomas EL, Morr CV, CoulterST (1967) Relation of heat-induced changesin protein-salt constituents to astringency inmilk systems. J Dairy Sci 50, 1376-1383

Joslyn MA, Goldstein JL (1964) Astringency offruits and fruit products in relation to phenoliccontent. Adv Food Res 13, 179-217

Kirimura J, Shimizu A, Kimizuka A, Ninomiya T,Katsuya N (1969) The contribution of peptidesand amino acids to the taste of foodstuffs.J Agric Food Chem 17, 689-695

Kostyra H (1983) Characteristics of astringent,astringent-sour, and bitter peptides Iiberatedfrom casein by Pseudomonas and Alcaligenesenzyme preparation. Zesz Nauk Akad RolnTech Olszt, Technol Zywn 18, 5-32

Kratzer DO, Foreman CF, Freeman AE, Bird EW,Rosenberger WS, Nelson FE (1967) Impor-tant sources of variation in milk flavor. J DairySci50, 1384-1389

Kudale JM (1970) An investigation of methodsfor reducing astringency in heated skimmilk.Ph 0 thesis, University of Minnesota, U-M-IDissertation Services, 70-20209

Land DG (1983) What is sensory quality. In: Sen-sory quality in foods and beverages: definition,measurement and control (Williams AA, AtkinRK, eds) Ellis Horwood Limited, Chichester,England, 15-29.

Lea AGH, Timberiake CF (1974) The phenolics ofeiders. 1. Procyanidins. J Sci Food Agric 25,1537-1545

Lea AGH, Arnold GM (1978) The phenolics ofeiders: Bittemess and astringency. J Sci FoodAgric 29, 478-483

Lea AGH, Arnold GM (1983) Bitterness, astrin-gency and the chemical composition of eiders.ln: Sensory quality in foods and beverages:definition, measurement and control (WilliamsAA, Atkin RK, eds) Ellis Horwood Limited,Chichester, England, 203-211

Lederer CL, Bodyfelt FW, McDaniel MR (1991)The effect of carbonation level on the sensoryproperties of f1avored milk beverages. J DairySci74,2100-2108

Lemieux L, Simard RE (1991) Bitter f1avour indairy products. 1. A review of the factors Iikely

to influence its development mainly in cheesemanufacture. Lait 71 , 599-636

Lemieux L, Simard RE (1992) Bitter flavour indairy products. II. A review of bitter peptidesfrom caseins: their formation, isolation andidentification, structure, masking and inhibi-tion. Lait 72, 335-382

Lemieux L, Puchades R, Simard RE (1989) Size-exclusion HPLC separation of bitter and astrin-gent fractions from Cheddar cheese madewith added Lactobacillus strains to acceler-ate ripening. J Food Sci 54, 1234-1237

L6pez-Fandiiio R, Olano A, San José C, RamosM (1993) Application of reversed-phase HPLCto the study of proteolysis in UHT milk. J DairyRes 60, 111-116

Lyman BJ, Green BG (1990) Oral astringency:Effects of repeated exposure and interactionswith sweeteners. Chem Senses 15,151-164

Malcolmson LJ, McDaniel MR (1980) Magnitudeestimation of infant foods. II. Taste, textureand odor of infant formulas. Can Inst FoodSci Technol J 13, 56-63

Mc Gugan WA, Larmond E, Emmons DB (1979)Sorne observations on the flavor of acid whey.Can Inst Food Sei Technol J 12, 32-35

Mistry VV, Hassan HN (1990) Manufacture of lowfat yogurt using a high protein powder. In:Brief Communications and Abstracts ofPosters of the 23rd Int Dairy Congr (FIL). Mon-treal, Canada. 518

Mistry VV, Hassan HN (1992) Manufacture ofnonfat yogurt from a high milk protein pow-der. J Dairy Sci75, 947-957

Modler HW, Emmons DB (1977) Spray dryingacid whey treated with caleium. J Dairy Sci60 (suppl 1), 38-39

Modler HW, Emmons DB (1978) Calcium as anadjuvant for spray-drying acid whey. J DairySci 61,294-299.

Moncrieff RW (1971) The sour taste. Flavour Ind2,84-86

Moretti RH (1978) Production of aeid soluble soyprotein by enzymatic hydrolysis. In: Abstracts5th Congress - Intemational Congress of FoodScience and Technology. Kyoto, Japan, 233

Morr CV, Ha EYW (1991) Off-flavors of whey pro-tein concentrates: a literature review. Int DairyJ1,1-11

Mottar J (1981) Heat resistant enzymes in UHTmilk and their influence on sensoric changes


during uncooled storage. Milchwissenschaft36,87-91

Nishimura T, Kato H (1988) Taste of free aminoacids and peptides. Food Rev Int 4, 175-194

Oison RL (1970) Removal and identification ofsome f1avor constituents from corn sweeten-ers. Diss Abstr Int - B - Sei Eng 31, 2045

Paolielo MMB, Reddy KV, Da Silva RSF (1987)Calcium sulphate as organoleptic coadjuvantin the formulation of soy-whey yoghurt.Lebensm Wiss Technol20, 155-157

Patton S, Josephson DV (1952) Observations onthe tactual flavor qualities of heated milk.J Dairy Sci35, 161-165

Puspitasari NL, Lee K, Greber JL (1991) Calciumfortification of Cottage cheese with hydrocol-loid control of bitter flavor defects. J Dairy Sci74,1-7

Reimerdes EH (1978) The enzymatic degrada-tion of ~-casein and its technological signifi-canee. In: Abstracts 5th Congress - Interna-tional Congress of Food Science andTechnology. Kyoto, Japan. 232

Ribéreau-Gayon P (1968) Les composés phéno-liques des végétaux. Dunod, Paris. 254 p

Rouseff RL (1990) Bitterness in Foods and Bev-erages - Developments in Food Science 25.Elsevier, New York, 356 p

Schiffman SS, Simon SA, Graham BG (1992a)Regional differences in sensitivity to astrin-gent compounds in the oral cavity of humans.Chem Senses 16, 575

Schiffman SS, Simon SA, Suggs MS, SostmanAL (1992b) Comparison of chorda tympaniand trigeminal nerve responses to astringentcompounds in redents. Chem Senses 16, 576

Schrôder MJA (1983) Aspects of sensory qualityin milk and unfermented milk products. In:Sensory quality in foods and beverages: def-inition, measurement and control (WilliamsAA, Atkin RK, eds) Ellis Horwood Limited,Chichester, England, 401-411

Schultz HW, Day EA, Libbey LM (1967) Sympo-sium on Poods: The ehemistry and physiologyof flavours. AVI Publishing Co Inc, Westport,Connecticut. 535 p

Shipe WF (1980) Analysis and control of milk fla-vor. In: The analysis and control of less desir-able fla vors in foods and beverages (Char-alambous G, ed) Academie Press, New York.201-239

239

Shipe WF, Senyk G, Weetall HH (1972) Inhibi-tion of oxidized flavor development in milkby immobilized trypsin. J Dairy Sci 55, 647-648

Shipe WF, Bassette R, Deane DO, Dunkley WL,Hammond EG, Harper WJ, Kleyn OH, Mor-gan ME, Nelson JH, Scanlan RA (1978) Off-f1avors of milk: nomenclature, standards andbibliography. J Dairy Sci 61, 855-869

Singleton VL, Noble AC (1976) Wine Flavor andPhenolic Substances. In: phenolic, suif ur, andnitrogen compounds in food fla vors. Am ChemSoc (ACS) Symp Ser 26 (Charalambous G,Katz l, eds) Washington, USA. 47-70

St-Laurent AM (1989) Le goût de craie du lait:un lien avec l'alimentation hivemale. Prod LaitMars, 13

Sundararajan NR, Muck GA, Whitney R MeL,Tobias J (1966) Flavor of evaporated milk.II.Changes in f1avor on storage of evaporatedmilk made by three processes. J Dairy Sei 49,166-172

Szczesniak AS (1963) Classification of texturaicharacteristics. J Food Sci 28, 385-389

Szczesniak AS (1979) Classification of mouth-feel characteristics of beverages. In: Food tex-ture and rheology(Sherman P, ed) AcademiePress, London, 1-20

Takeuchi T, Yoshii T, Yoshii H (1967) Peptides inmiso and soy sauce. IV. Relation between theCIN ratio of the raw mate rials and the distri-bution patterns of peptides in miso. HakkoKogaku Zasshi (J Ferment Technol) 45,29-33

Tessier H, Rose 0 (1958) Calcium ion concen-tration in milk. J Dairy Sci 41, 351-359

Thurston LM, Carson Brown W, Dustman RB(1935) Oxidized flavor in milk. 1. The proba-ble relation of lecithin to oxidized f1avor.J DairySci 18, 301-306

Tobias J (1976) Organoleptic properties of dairyproducts. In: Dairy technology and engineer-ing (Harper WJ, Hall CW, eds) The AVI Pub-Iishing Co Inc, Westport, Connecticut. 75-140

Tobias J (1990) Compositional elements affectingf1avorof dairy foods: Formulated foods. J DairySci73,3657-3663

Visser FMW (1977a) Contribution of enzymesfrom rennet, starter bacteria and milk to pro-teolysis and f1avour development in Goudacheese. 2. Development of bitterness andcheese flavour. Neth Milk Dairy J 31, 188-209


Visser FMW (1977b) Contribution of enzymesfrom rennet, starter bacteria and milk to pro-teolysis and flavour development in Goudacheese. 3. Protein breakdown: analysis of thesoluble nitragen and amino acid nitragen frac-tions. Neth Milk DairyJ31 , 210-239

Visser FMW (1977c) Contribution of enzymesfrom rennet, starter bacteria and milk to pro-teolysis and flavour development in Gouda

cheese. 5. Some observations on bitterextracts from aseptically made cheeses. NethMilk Dairy J 31 , 265-276

Watson M (1992) Sensory characteristics of food.Nutr Food Sci 23, 4-6

Whitfield BH, Davis HP, Downs PA (1936) Theeffect of milk upon metals and metals uponmilk. J Dairy Sci 19,27-28

Astringency, a textural defect in dairy products · 218 L Lemieux, RE Simard INTRODUCTION ln a...

Documents

Transcript of Astringency, a textural defect in dairy products · 218 L Lemieux, RE Simard INTRODUCTION ln a...