HAL Id: hal-00929663https://hal.archives-ouvertes.fr/hal-00929663

Submitted on 1 Jan 1999

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.

Kishk - a dried fermented milk / cereal mixture. 3.Nutritional composition

Adnan Y. Tamime, Margaret N.I. Barclay, David Mcnulty, Thomas P.0’Connor

To cite this version:Adnan Y. Tamime, Margaret N.I. Barclay, David Mcnulty, Thomas P. 0’Connor. Kishk - a driedfermented milk / cereal mixture. 3. Nutritional composition. Le Lait, INRA Editions, 1999, 79 (4),pp.435-448. �hal-00929663�

Lait (1999) 79, 435-448© Inra/Elsevier, Paris

435

Original article

Kishk - a dried fermented milk/cereal mixture.3. Nutritional composition

Adnan Y. Tamime-", Margaret N.l. Barclay", David Mclvulty",Thomas P. O'Connor-

a Food Standards & Product Technology Department, SAC - Auchincruive,Ayr KA6 5HW, Scotland, United Kingdom

b Biomathematics & Statistics Scotland, University of Edinburgh, James Clerk Maxwell Building,The Kings Building, Edinburgh EH9 3JZ, Scotland, United Kingdom

C Department of Food Chemistry, University College Cork, Cork, Ireland

(Received 6 March 1998; accepted 24 November 1998) •

Abstract - An investigation of the nutritional properties of 25 commercial samples of Lebanese Kishkwas undertaken. Profiling of the carbohydrate-based nutrients (g·100 g-I on dry matter basis [DMBDin the samples gave the following ranges: fibre 7-12, phytic acid 0.7-1.6, and ~-glucan 0.1-0.6.Some Kishk samples contained appreciable amounts of polyunsaturated fatty acids, while the contentsof monounsaturated fatty acids of most of the samples were considerably lower than those present inmilk and other dairy products. Ali the Kishk sampi es contained appreciable quantities (mg- 100 s'[DMB]) of the major minerais (K 495, P 397, Ca 243 and Mg 123), and such product was a good sourceof Fe and Mn which originated from the Burghol. Sodium was present in high amounts (-1657 mg- 100s'[DMB D. The amino acids composition of the protein from Kishk was good. Vitamins C, pyridoxineand ~-carotene were not detected in the Kishk samples, and approximately half of these samplesdid not contain œ-tocopherol. The thiamin and riboflavin contents of Kishk were in the range ofwhat has been reported in the Iiterature. Kishk has a limiting vitamin factor and is not considered agood dietary source. The selenium content of the majority of the Kishk samples was good and sucha product may represent a potentially good dietary source. © Inra/Elsevier, Paris.

Kishk / ~-glucan / amino acid / mineraI / vitamin / selenium

Résumé - Le kishk - un mélange lait/céréales fermenté et séché: 3. Composition nutritionnelle.Les propriétés nutritionnelles de 25 échantillons de kishk libanais du commerce ont été étudiées.Les profils de nutriments glucidiques (en g pour 100 g de matière sèche) étaient compris dans lesgammes suivantes: fibre 7 à 12, acide phytique 0,7 à 1,6 et ~-glucane 0,1 à 0,6. Certains échan-tillons de kishk contenaient une quantité appréciable d'acides gras polyinsaturés, tandis que les

* Correspondence and reprints. a.tamime@au.sac.ac.uk

436 A.Y. Tamime et al.

teneurs en acides gras mono-insaturés de la plupart des échantillons étaient nettement inférieures àcelles présentes dans le lait et les produits laitiers. Tous les échantillons de kishk contenaient des quan-tités appréciables des principaux éléments minéraux (K 495, P 397, Ca 243 et Mg 123 mg-I 00 g-l dematière sèche), et constituaient de bonnes sources de Fe et Mn apportées par le Burghol. Le sodiumétait présent en forte quantité (-1657 mg·lOO g-l de matière sèche). La composition en acides ami-nés des protéines du kishk était bonne. La vitamine C, la pyridoxine et le l3-carotène n'étaient pas détec-tés dans les échantillons de kishk et environ la moitié des échantillons ne contenaient pas d'rx-toco-phérol. Les teneurs en thiamine et riboflavine du kishk étaient dans la gamme rapportée dans lalittérature. Le kishk avait une teneur en vitamines limitée et de ce fait n'est pas considéré comme étantune bonne source en diététique. La teneur en sélénium de la majorité des échantillons de kishk étaitbonne ce qui en fait une source intéressante dans le régime alimentaire. © InralElsevier, Paris.

kishk / l3-glucane / acide aminé / minéraux / vitamine / sélénium

1. INTRODUCTION

Fermented milk products have alwaysbeen considered a highly nutritious foodcontaining significant concentrations of pro-tein, fat, minerais and micronutrients, andhas been extensively investigated andreviewed by many research groups [2,24-27,36,37]. Sorne data are available onthe nutritional properties of Kishk, whichhas been recently reviewed by Tamime andO'Connor [45] and Tzanetakis [49]. Kishkis a dried yoghurt and cereal (Burghol) mix-ture which is a good source of protein[21-23,30,46,47]. The yoghurt makes upfor the amino acids limited in cereal (lysineand methionine). However, Burghol (par-boiled cracked wheat) is a good source ofcertain minerals (Fe, Cu and Mn) that aredeficient in milk [Il, 33].

Certain cereals may contain componentsof interest to human nutritionists including:1) dietary fibre, which is considered a goodvehicle to control blood cholesterol levelsand may also reduce the incidence of coloncancer; 2) ~-glucan and other soluble fibres,which may reduce cholesterol absorptionfrom the intestinal tract, and also reduce theformation of low density lipoprotein (LDL);3) phytic acid, which has the potentialtoexert negative nutritional effect and bindsdivalent minerals (Ca, Zn and Fe) [12,17,29,41,42,44,50]; and 4) selenium, which

has been demonstrated to exhibit an antiox-idant role, is involved in thyroid metabolismand antimutagenic activity [8]. An associa-tion between cardiovascular death, myocar-dial infraction and serum selenium levelshas been reported. The selenium content of700 UK foods has been recently determinedby Barclay and MacPherson [5] and Barclayet al. [6]. Limited data are available on thesecomponents and the vitarnin content of com-mercial Kishk samples, and only niacin lev-els are appreciably high when comparedwith skimmed milk powder. However, thiscould be attributed to the metabolic activityof the starter culture, and Burghol is also agood source of niacin [45].

The objectives of this study were toinvestigate in detail the chemical composi-tion, species of origin of milk protein, nutri-tional composition and microbiological qual-ities of 25 Lebanese Kishk samples obtainedfrom different outlets. The results of thenutrition al composition of Kishk are givenin detail here.

2. MATERIALS AND METHODS

2.1. Kishk samples

Twenty-five samples of commercial Kishk(-1 kg each) were obtained from different retai!outlets in Lebanon [48].

Kishk - nutritional composition

2.2. Analytical methods

2.2.1. Minerais

The Ca, P, Mg, K, Na,Zn, Fe and Mn con-tents in Kishk were determined using a nitric-perchloric acid digestion procedure according tothe method described by MAFF [31] followedby induction coupled plasrna-I 00 (lCP) emis-sion spectroscopy (Thermo Electron Ltd., Birch-wood, Warrington, UK).

The analysis was performed at the followingflow rates: 1) main argon at 15 Lmirr-'. 2) neb-ulizer argon at 0.5 L·min-1, and 3) sample at0.8 rnl.i-rnirr '. The minerai eluates were moni-tored at wavelengths of 317.9 nm - Ca; 214.9nm - P; 285.2 nm - Mg; 766.5 nm - K; 589.6 nm- Na; 213.9 nm - Zn; 238.2 nm - Fe; 324.8 nm-Cu and 257.6 nm - Mn.

2.2.2. Chemical analysis

Phytic acid, p-glucan and selenium contentsin the Kishk samples were determined accord-ing to the methods described by AOAC [4],McCleary and G1ennie-Holmes [32] and Hersheyand Oostdyke [18], respectively.

2.2.3. Vitamin assays

Retinol (vitamin A), œ-tocopherol (vitamin E)and p-carotene were analysed by the high per-formance Iiquid chromatography (HPLC) methodbased on the procedure described by Bieri et al.[7]. Kishk (1 g) was weighed out accurately intoa test tube, and extracted by the method of Butrissand Diplock [JO]. Following extractions, theresidues were reconstituted with 200 ul, of abso-lute alcohol.

An in-house reference sample consisting ofinfant baby food powder containing knownamounts of retinol, œ-tocopherol and p-carotenewas also analysed with the Kishk sampi es. HPLCanalysis was conducted using a Shimadzu LC-6Apump, a Rheodyne 7125 syringe loading inje ctor,and a Machery-Nagel Nucleosil 5 CI8 columnunder the following conditions:

437

Vitamin C (ascorbic acid + dehydroascorbicacid) was measured using a HPLC method basedon a modification of that of Speek et al. [40].After extraction, ascorbic acid (AA) was oxi-dised enzymatically by ascorbic acid oxidase todehydroascorbic acid (DHAA). DHAA was con-densed with o-phenylenediamine (OPDA) to itshighly fluorescent quinoxaline derivatives. Thesederivatives are separated on a reverse-phaseHPLC column and detected fluorometrically.DHAA can be determined separately by the sameprocedure with the omission of the enzymaticoxidation. The difference in DHAA measuredin the presence or absence of the ascorbic acidoxidase is related to the ascorbic acid content ofthe sample.

HPLC analysis was carried out within 12 husing a Shimadzu LC-6A pump, a Rheodyne7125 syringe-Ioading inje ctor, a Machery-NagelNucleosil 5 CI8 column and a pre-column of sim-ilar composition. A Waters Guard-Pak filter unitwas fitted to protect the column from beingclogged by particulates. The mobile phase wasmethanol (20 % v/v), 0.08 mol-L':' KH2P04buffer, pH 6.6 (1 :4; v/v) at a flow rate of1 mls-mirr '. DHAA was detected by measuringfluorescence of the eluate using a ShimadzuRF-535 fluorescence detector (excitation wave-length 365 nm and emission wavelength 418 nm).

Thiamin (vitamin B,) and vitamin B6 (pyri-doxine, pyridoxal and pyridoxamine) wereextracted and quantified as total vitamin B6 fromKishk samples by the method of Brubacher etal. [9]. Thiamin was oxidised to thiochrome usinga procedure based on AOAC [3]. The thiochromewas measured directly by HPLC. A ShimadzuLC-6A pump, a Rheodyne 7125 syringe-load-ing inje ctor and a Shimadzu RF 535 fluorescencedetector were used. A Machery-Nagel Nucleosil5 C'8 column and precolumn were also used. Themobile phase was methanol-water (95:5; v/v) ata flow-rate of 1 ml.irnirr '. Thiochrome was mea-sured by fluorescence detection, using an exci-tation wavelength of 365 nm and an emissionwavelength of 435 nm.

Riboflavin (vitamin B2) was extracted fromthe Kishk sampi es as outlined earlier for thiamin

Lamp Â. Solvent Flow rate (ml.irnirr")

œ-tocopherol D 292nm MeOH/Hp 97:3 2Retinol D 325 nm MeOH/H20 95:5 2p-carotene W 450nm MeOH/Hp/acetonitrile 47:6:47 2.2

438 A.Y. Tamime et al.

(vitamin BI); however, the oxidation step wasomitted. After filtration, the filtrate was diluted1:2 with a methanol-water solution (70:30;v/v),filtered through a 0.45-~m filter and injectedonto the column. An in-hou se reference sampleof infant baby food powder containing a knownconcentration of riboflavin was also analysed.A similar HPLC system and column was usedas outlined earlier for thiamin. The mobile phasewas methanol-water (70:30) at a flow rate of1 mlvrnirr '. Detection was by fluorescence usingexcitation and emission wavelengths of 453 and521 nm, respectively.

2.2.4. Amino acids

The concentrations of amino acids in 25 sam-pIes of Lebanese Kishk were determined by themethod of Spakman et al. [39] using model 3A29 automatic amino acid analyser (Carlo ErbaScience, Milan, Italy).

A sample of Kishk (100 mg) was weighed ina universal bottle and dissolved in 3 mL of6 mol-L'! HCI solution. Later, nitrogen was bub-bled through the sample to remove oxygen, thensealed and placed in an oyen at 110°C for 24 h.The sample was filtered into 100 mL volumetrieflask, diluted with distilled water and made-up to100 mL. Ten mL filtrate was dried and theresidue resuspended in Na-citrate buffer at pH2.2. Amino acids were quantified by measuringthe peak area, and the elution gradient was con-trolled electronically to neutralise the amino acidsin different pH buffers. The flow rate of thebuffers was 30 mL·h-1 for 20 min at pH 3.3,15 min at pH 4.13 and 40 min at pH 6.8. Theresults of amino acids contents were expressedas mg-g" N on a wet matter basis (WMB).

2.3. Statistical analysis

The data were analysed by univariate (analy-sis of variance), and multivariate (principal com-ponent analysis [PCA]) techniques using theGenstat computer pro gram as described byTamime et al. [48].

3. RESULTS AND DISCUSSION

3.1. Carbohydrate-based nutrients

The average concentration of carbohy-drate content in 25 samples of Kishk were

detailed by Tamime et al. [48]; the phyticacids, and ~-glucan contents (g.IOO g-l[DMB]) ranged between 0.71 and 1.62, and0.14 and 0.61, respectively (table I). Anappreciable amount of fibre (-9.3 g.100 g-l[DMB]) was found in the Kishk samples[48]. At present, there is widespread con-sensus on the benefits of dietary fibre [41].The source of fibre in the Kishk samples isthe Burghol, and the variation in the fibrecontent of the se samples could be attributedto: 1) different varieties of wheat used dur-ing the production of Burghol; 2) the effi-ciency of de-husking stage; and 3) the dif-ferent ratio of Burghol to yoghurt usedduring Kishk-making [45, 46]. High fibrediets increase the bulk of the faeces andreduce the transit time through the largeintestine and may progressively lower therisk of large bowel cancer. Furthermore,dietary fibre is the most effective means oftreating chronic constipation, diverticulardisease, obesity and diabetes [13]. The fibrecontent (g·100 g-l [DMB]) of 'all bran',brown flour, white flour and wholemealflour averaged 25.3, 7.3, 3.6 and 10.5,respectively [20]. Thus, eight Kishk sam-pies (numbers 4, 5, 11, 13, 15, 17-719)exarnined can provide a level of fibre z thanwholemeal flour.

Phytic acid (the hexaphosphoric ester ofinositol) was found in the 25 samples ofKishk at an average level of 0.94 g.100 g-l(DMB). The source of phytic acid is theBurghol rather than the yoghurt [46]. How-ever, the Phosphoinositol molecule is 88.8 %phosphate, and the phosphorus bound asphytate phosphorous was calculated (table J,figures in parentheses). The results suggestthat many Kishk samples have high levels ofphytate phosphate which may pre vent theabsorption of divalent cations (Ca, Fe andZn) in the body [34].

The ~-glucan content of the 25 samplesof Kishk averaged 0.33 g·100 g-l (DMB)which originated primarily from the Burghol.This level is rather low because the ~-glucancontent in wheat Burghol is 0.24 g·100 g-l

Kishk - nutritional composition 439

Table I.The nutritional components of different commercial samples of Kishk.Tableau I.Composition nutritionnelle de différents échantillons de kishk du commerce.

Components Minimum Maximum Mean

Carbohydrates-based (g·IOO g-I).Dietary fibre 6.51 12.24 9.32Phytic acid 0.71 (49)b 1.62 (> 100) 0.94l3-glucan 0.14 0.61 0.33

Main classes of fatty acids (% w/w)CSaturated 55.40 81.02 70.07Monounsaturated 15.93 29.84 23.25Polyunsaturated 4.42 22.40 7.15

Amino acids (mg-g :' N)Serine 158 567 268Proline 242 675 450Aspartic acid 189 763 397G1utamic acid 904 2769 1309Alanine + glycine 92 381 219Histidine 83 189 134Arginine 226 704 347Valine 48 234 143Tyrosine 157 406 234Phenylalanine 165 405 253Lysine' 100 322 204Leucine 247 765 424Iso-leucine 52 175 110Threonine 162 472 243

Selenium (ug-I 00 g-l)" 2.5 26.0 10.2Vitamins

Retinol } (llg.g-1)" 0.001 1.027 0.165œ-tocopherol } NOd 3.32 0.44Thiamin } (mg- 100 g-l)" NO 0.292 0.151Ribotlavin } 0.034 0.193 0.079

a Results are the average of two determinations perfonned on each sample; data were computed on dry matter basis;b figures in parentheses represent phosphate as phytate phosphate; C data were calculated on weight of fat;ct ND: not detected.a Les résultats sont la moyenne de deux déterminations par échantillons; les données étaient calculées par rapportà la teneur en matière sèche. b Les chiffres entre parenthèses représentent le phosphate en tant que phytate phos-phate. C Données calculées par rapport au poids de la matière grasse. ct ND : non détecté.

(DMB) when compared with barley Burghol(3.51 g·IOO g-l [DMB]) and oat Burghol(3.75 g·100 s' [DMBD [46]. In view of theCUITentnutritional value of ~-glucan, Kishkhas been manufactured in our laboratorywhere the wheat Burghol was replaced byeither barley or oats Burghol in order to

increase the ~-glucan content in Kishk, andthe results have been reported by Tarnime etal. [47].

In order to visualise the similaritiesamong the 25 samples of Kishk and the rela-tionships between the carbohydrate-basednutrients (fibre data taken from [48], phytic

440 A.Y. Tamime et al.

acid and ~-glucan components), a PCA wasperformed and principal component (PC)biplot was produced (figure 1). The mainfeatures of interest within a biplot are thelayout of the variables (vectors), clusters ofsamples (points), samples with high load-ings on particular variables and outliers fromclusters of points. The percentage varianceof a biplot is an indicator of how weil thedata are summarised infigure 1; 84.6 % ofthe correlation matrix of phytic acid, ~-glu-can and fibre is accounted for. The vectorsfor phytic acid and ~-glucan are almost co-incident, indicating that such componentsare strongly correlated. Conversely, the vec-tor for fibre is at right angles to the other twocomponents, indicating that fibre is uncor-related with both phytic acid and ~-glucan.None of the samples exhibit any obviousclustering; however, samples 5, 15, 18,25and possibly 19 are outliers and thus of inter-est. These samples have relatively high load-ings on fibre, and hence it is conjecturedthat these Kishk samples (5, 15, 18 and 19)were probably made with a higher propor-tion of Burghol than the others, or theBurghol used contained different amounts of

fibre reflecting the efficiency of de-husk-ing during the production of Burghol. Thephytic acid and ~-glucan contents of sample25 are similar to sample 18, but the formersample has an average fibre content.

3.2. Nutritional value of fatty acids

Profiling of the fatty acid content of dif-ferent commercial samples of LebaneseKishk has been reported elsewhere [48].Table 1 shows the total contents of saturated,mono- and polyunsaturated fatty acids pre-sent in the Kishk samples tested. In humansthe most essential fatty acid is arachidonic(CZO:4' n-ô) which is readily formed fromlinoleic acid (CI8:Z' n-ô), and hence it is the.main dietary source [34]; however, milk anddairy products are usually low in such essen-tial fatty acid while cereals are much higher[20].

Kishksamples 13, 17, 19and25containan appreciably higher amount of polyun-saturated fatty acids (22.4, 16.9, 13.9 and15.9 %, respectively; for details refer to[48]) th an the rest of the Kishk samples(- 5.2 %). Such marked increase in the

1.0

-0.5

1 1

-0.5 0.0 0.5Principal component 1 (55.2%)

~rt-: 0.5~N

1:alCaQ,

EauiiiQ,

'üc.;:a,

Fibre*-

\

15

• .19

5 4 111~13 .3• •• or .... 2

18 . 2O~.61"24

~

• ..22.9p. h.ytic 6.5.7acld .8• • •25 21 • 10jJ-glucan 23 ·14

0.0Figure 1. A principal corn-panent biplot of sorne car-bohydrates-based nutrientsin Kishk samples. (Forsample identification, referta [48]).Figure 1. Analyse en com-posantes principales dequelques nutriments gluci-diques dans les échantillonsde kishk. Voir Tamime etal. [48] pour l'identifica-tion des échantillons.

1.0

lKishk - nutritional composition

1.0

Mono-unsaturated

~ *0

~18

..,. 0.5r-:....~C'I- e 7 16l: e eCP 25 20~e1 Saturatedl: e 21e24 5100 0.0 *Co 14. 22823'12E

19/e

se e60 25<>17/e

e3iii e e2Co ;N"Cil: -0.5.;:

*0.Poly-unsaturated

polyunsaturated fatty acid contents may beattributed to: 1) using a higher level ofBurghol during the preparation of the Kishk,and/or 2) possible conversion of saturatedfatty acids into an unsaturated counterpartduring the secondary fermentation stage (see[45]).

Monounsaturated fatty acids are themajor constituent of human adipose tissue[34], and these fatty acids are the majorcomponent of olive oil which is used widelyin the Mediterranean diet. Monounsaturatedfatty acid contents of milk and other dairyproducts is considerably higher than thosereported for Kishk (table /) and, in particu-lar, Kishk samples 2, 3 and 6 (15.9, 17.1and 17.9 %, respectively) are well below theaverage of most dairy products (for detailsrefer to [48]).

Saturated fatty acids are more stable thanunsaturated fatty acids to oxidation, but gen-eraIly the nutritional advice is to decreasethe daily intake. Kishk samples 13, 17 and25 have the lowest levels (52, 58.8 and53.4 %, respectively), and again this mayonly be a reflection of using higher cerealcontent in Kishk-making [48].

441

A PCA was performed on the correlationmatrix of the saturated, mono- and polyun-saturated fatty acid data of the Kishk sam-ples, and a PC biplot was plotted (figure 2).Since the three variables are constrained toadd to 100 %, any one variable may be cal-culated from the other IWo,and hence a two-dimensional solution summarises table IIin its entirety. PC 1 accounted for 82.1 %of the total variation and contrasted the maindifference between samples of Kishk, whichwas due to the ratio of saturated to unsatu-rated fatty acids. PC 2 accounted for a fur-ther 17.4 % of the variation and separatesthe Kishk samples by the ratio of mono- topolyunsaturated fatty acids. Most of the datais contained in a single cluster; however,several samples form jet streams from themain cluster. Kishk samples 2, 3, 5 and 6 arenotable for their high proportion of satu-rated fatty acid, and sample 18 is high inmonounsaturates while samples 13, 17 and19 are high in polyunsaturates. Sample 25 iswell isolated (figure 2) because it has a highunsaturated fatty acids content (for sampleidentification, see [48]).

Figure 2. A principal com-ponent biplot of saturated,monounsaturated and poly-unsaturated fatty acid con-tents in Kishk samples.(For sample identification,refer to [48]).Figure 2. Analyse encomposantes principalesd'acides gras saturés, mono-insaturés et pol y-insaturésdes échantillons de kishk.Voir Tamime et al. [48]pour l'identification deséchantillons.

-0.5 0.0_1.01-..--- ......--.....&----.1.----1

-1.0 1.0Principal component 1 (82.1%)

0.5

442

Ser

Pro

Asp

Glu

Ala+G1yHis

't>U..oc:'E<

A.Y. Tamime et al.

Arg

ValTyr

PheLys

Leu

IleThr

>D------< *>-{Il---. * *

>--[}--1****o-{I)-< * *

4

'i[]---<

'11<0[]--. * * *

0/1}<

ofi--< * *œ-'il}-<

>{[]---o *([J-.

œ-.

13 .,QIQ.EIII.,

.><.c:.,~

14,15,16,17 0e.!:!5o

16,3,4

3,15,17,16

13,14

• 10,2.

o 1000 30002000Concentration of amino acid (mg g.1 N)

Figure 3. Box plot of concentration of amino acids of the Lebanese Kishk samples. (For sampleidentification, refer to [48].) The vertical line inside the box represents the median and the vertical endsof the box represent the 1st and 3rd quartiles. Asterisks represent outliers which are further out than1.5 x inter-quartile range from the 1st and 3rd quartiles, respectively.Figure 3. Concentrations en acides aminés des échantillons de kishk libanais, avec les intervallesd'incertitude. Voir Tamime et al. [48] pour l'identification des échantillons. Note: les lignes verti-cales à l'intérieur des boîtes représentent la valeur médiane et les extrémités verticales des boîtes, le1er et le 3e quartile. Les astérisques représentent les échantillons extrêmes qui sont au-delà de 1,5 foisla variation inter-quartile entre le 1er et le 3e quartile.

3.3.1. Amino acids

3.3. Micronutrients in Kishk

The average protein content of 25 sam-pies of Kishk was 17.8 g·IOO g-l (OMB)[48], which is considered very high, and thevariations in the different amino acid con-centrations (mg-g " N) is shown in table 1;however, a box plot (figure 3) illustrates thedistribution of the variable(s). Il is evidentthat there was extensive variation of theamino acid contents, in particular when com-pared with the commercial sample reportedby Jamalian and Pellett [28]. Sorne of theessential amino acids (Thr, Leu, Lys, Phe,Tyr and Val) were abundant in most of theKishk samples studied.

Tryptophan and sulphur-containingarnino acid contents are low in Kishk, whichis not really a good source [33]; however, the

analytical procedures are cumbersome andfor this reason they were not determined.The loss or decomposition of these aminoacids could be attributed to processing con-ditions such as the fermentation of milk fol-lowed by sun-drying. Nevertheless, the sameauthors reported that Kishk is a good sourceof amino acids when compared with the pro-visional pattern suggested by FAO/WHO[16], although Try was a limiting factor.The spectrum of amino acids in milk pro-teins is better than wheat flour [19,34, 35],and the data shown in table 1 suggest thatmilk protein in Kishk supplements the aminoacid deficiencies in the Burghol.

The PC biplots of the ami no acidsdemonstrate that the distribution of aminoacids is concentrated on the positive halfplane of the first PC and uniformly dis-tributed on the second Pc. Hence, PC 1

Kishk - nutritional composition 443

relates to the abundance of the amino acidswhile the second PC differentiates the sam-pIes. The PC biplot of the amino acids ofthe Kishk samples is shown in figure 4A.The Kishk samples form two major groupsplus a small group consisting of samples 3,5 and 10. The first group is characterisedby sampIes (13 ~ 17) having a higher-than-average content of Arg, Ser, Glu, Asp andAla + Gly. Within the second group, sam-pIes with relatively high Pro and Leu con-tents are plotted towards the top, while thebottom of this group is characterised by sam-pIes with high Ile. The third group consistsof samples 3, 5 and JO which have relativelyhigh levels of Thr and Tyr. Introducing theamino acid contents of Burghol andskimmed milk powder (SMP) into the dataset reverses the direction of the Ile and Valaxis infigure 4B with respect tofigure 4A.SMP is cIearly an outlier in the plot, butremarkably does not change the configura-tion of the rest of figure 4A.

Little published data are available on theamino acid contents of Kishk, and the dataon a commercial Lebanese product [28] andtwo commercial Cypriot Trahana products[14,45] have been used for comparativepurposes. A PC biplot was produced (fig-ure 4C) to highlight any differences betweenthe two main groups of the 25 samples ofKishk shown infigure 4A and the three com-mercial products. Adding the commercialsamples to the data set has a more profoundeffect on the configuration of samples.

Figure 4. Principal component biplots of aminoacids present in the Kishk samples (A), vis-a-visin skimmed milk powder (SMP) and Burghol(B) and in two Trahana and one Lebanese Kishkproducts (C). (For sample identification, refer to[48)).Figure 4. Analyses en composantes principalesdes acides aminés dans les échantillons de kishk(A) de poudre de lait écrémé et de burghoI (B) etdans deux trahana et deux kishk libanais (C).Voir Tamime et al. [48] pour l'identification deséchantillons.

1.0

i:"1 0.5=.N

E~a 0.0

~u

"[Üc -0.5;f

A

-1.0-1.0 -0.5 0.0 0.5 1.0

Principal componenll (32.7%)

1.0

B eSMP

1f!. LeuN 0.5ËN

ECl>c&. 0.0E0uc;;Q.

'ü.5 -0.5li.

-1.0-1.0 -0.5 0.0 0.5 1.0

Principal componenll (32.9%)

1.0 ------------ ....C

i:"': 0.5...N

C

~&. 0.0

~u

"['üc -0.5&:

-1.0 ... -- ... -- ... -- .... -- .....-1.0 -0.5 0.0 0.5

Principal componenll (31.9%)1.0

444 A.Y. Tamime et al.

Specifically the Tyr, Thr, and Ile, Val axesare reversed infigure 4C with respect to fig-ure 4A, while the vectors representing theremaining arnino acids remain fairly sirnilar.The location of the commercial samples infigure 4C is distinct from the other threegroups previously identified. This may sug-gest that the nutritional properties of pre-sent-day Kishk have improved over the past2 decades possibly due to the use of a higherratio of fermented milk: Burghol, such as4: 1 rather than 2: 1.

3.3.2. Vitamins

Vitamins C and ~-carotene were notdetected in any of the Kishk samples; how-ever, vitamins A, E, BI and B2 contents areshown in table J. The concentration of vitaminA ranged between 0.001 and 0.324 ug-g :'(DMB) in the majority of Kishks exceptsample 25 (1.027 ug-g' [DMB]). Nearlyhalf the Kishk samples did not contain anyœ-tocopherol, and the rest of the samplesaveraged -0.85 ug-g ' (DMB). Since theKishk was made from low fat milk, this isreflected in the low content of fat-solublevitamins present. Incidentally, vitamins Aand E contents of SMP averaged 3.5 and2.7 ug-g ", respectively [19]. The thiaminand riboflavin contents of these samplesaveraged 0.164 and 0.079 mg- 100 s:'(OMB), which appear to be in the range ofwhat has been published previously [45,49], and skimmed milk [19]. lt is evidentthat Kishk might be a limiting factor as abeneficial source of certain vitamins.

The PC biplot of the vitamin content ofthe Kishk sampI es provided limited infor-mation regarding the correlation matrix (datanot shown). However, there are two axeswithin the plot defined by the opposition ofretinol to œ-tocopherol, and at right angles toit, riboflavin and thiamin are juxtaposed.The majority of samples lie on the riboflavinand thiamin axis, and form a homogeneousgroup. Outliers occur on the retinol and a-tocopherol axis with sample 25 high in

retinol, and samples 8 and 15 high inœ-tocopherol.

3.3.3. Selenium

The selenium content (ug-Iûû g-l[DMB]) of the Kishk samples rangedbetween 2.5 and 26.0 (table /) while themean value was 10.3, which is similar tothe value of yoghurt (9.411g·100 g-l [DMB])reported by Barclay et al. [6], and higherthan skimmed or who le milk powder (7.2or 8.311g·100 g-l [OMB], respectively [19]).The distribution of selenium in the Lebanesesamples of Kishk is shown infigure 5. Theselenium content of Burghol was 8.811g·1Oüs'(DMB) which compares favourably withwheat flours available in the United King-dom (range between 2.5 and 6.5 Ilg.IOO g-l[OMB]) [6].

Assuming thatthe Kishk sarnpl es arecomposed of 80 % yoghurt to 20 % Burghol,the Lebanese yoghurts or most likely thestrained yoghurt used must represent a bettersource of selenium than the UK counterpart.However, the effect of selenium on the dietis skewed by the consideration of dry matter

0.4 10

...,oC

4i 0.3c-c:ot:a 0.22l1.

<Il8 al

Ci.E.,<Il

6 '0ë:::J

4 80.1

2

0.0 7.5 15.0 22.5 30.0

Selenium content (JI.g100g.')

Figure S. A histogram illustrating the distributionof selenium content in the Kishk sampI es.Figure S. Histogramme illustrant la distributiondu sélénium dans les échantillons de kishk.

Kishk - nutritional composition 445

Table II. Mineral analysis (mg- 100 g-I)a of different Lebanese Kishk sampI es, skimmed milk pow-der (SMP) and Burghol.Tableau II. Analyse des minéraux (mg- 100 g-I de matière sèche) de différents échantillons de kishklibanais, de poudre de lait écrémé et de burghol.

a Results are the average of two determinations performed on each sample; data were computed on dry matter basis;b SMP: skimmed milk powder; data compiled from [19]; C data (for coarse Burghol) compiled from [46]; d Tr: trace.a Les résultats sont la moyenne de deux déterminations par échantillon; les données étaient calculées par rapportà la teneur en matière sèche. b SMP, poudre de lait écrémé: données issues de Holland et al. [19]. C données(pour le burghol broyé) issues de Tamime et al. [46]. d Tr, trace.

figures rather than the fresh weight. Forexample, a UK yoghurt of 16 g-I 00 g' totalsolids content may have a selenium contentof 1.5 ug-I 00 g-l (DMB). Kishk sample 25(26 Ilg'lOO g-l [DMB]) represents a con-siderable dietary potential of selenium.

3.3.4. Minerais

The concentration of minerais in 25 sam-pies of Kishk is shown in table Il, and sam-pie 25 had high concentrations (mg- 100 g-l[DMB]) of Mg 260, Fe 18.5 and Cu 0.6,whilst in sample 3 the concentration of Fewas 12.8. Such high concentrations of min-erals, in comparison with the other Kishksamples, had dominated the PC biplot ofthe PCA, and they were therefore excludedfrom table Il and statistical analysis. How-ever, for comparative purposes, data of theconcentration of minerals of SMP andBurghol (coarse type) are also included intable Il. It is evident that the sodium con-

tent is very high due to added salt during thepreparation of Kishk. According to Tamimeet al. [48], the added salt in these Kishk sam-ples, which was calculated from chloride bytitration, averaged 2.8 g NaCl·IOO g-l(DMB). Although such a level of salt inKishk may have possible dietary implica-tions (such as a contributory factor to highblood pressure), the salt level will be reducedwhen the Kishk is prepared as a hot gruel'dish'.

The Lebanese Kishk samples containappreciable quantities of K, P and Ca (495,397 and 243 mg·IOO g-l [DMB], respec-tively), but the y are below the contentsfound in SMP. This is attributed to theBurghol where the concentration of theseelements is -30 % below that found in SMP(table Il). However, concentration of Mg(123 g-l Oüs' [DMB]) is similar to that inthe milk powder and Burghol (100 % reten-tion) with the exception of Kishk sampi es

446 A.Y. Tamime et al.

1.0 r---"""T---,......--~---,

<!-"1 0.5~N

'é"l:8. 0.0Eou

i'ü.5 -0.5s:

. .A

13 •• 5.-.21

1.0

/~.,.//

Fe *

B

<!-C'! 0.5E-N

'é"l:8. 0.0E8i'ü.5 -0.5s:

-1.0 L- __ -.J. "-- __ ...... ....

-1.01.0 -0.5 0.0 0.5Principal componen! 1 (65.6%)

1.0

Figure 6. Principal component biplots of minerai contents in the Kishk samples (A). and of P, Ca andFe of the same samples compared with three commercial and two laboratory-made Kishks (B). (Forsample identification, refer to [48]).Figure 6. Analyse en composantes principales des minéraux des échantillons de kishk (A) et de P, Caet Fe des mêmes échantillons comparés avec trois kishks du commerce et deux kishks fabriqués enlaboratoire (B). Voir Tamime et al. [48] pour l'identification des échantillons.

21 (low) and 2, 3, 10, II and 25 (high). Theconcentration of Zn, Mg and Cu (3.1, 2.1and 0.42 mg- 100 g-l [DMB], respectively) issimilar to Burghol, but much higher in Mgand Cu when compared with SMP (table Il).An appreciable quantity of Fe (5.8 mg- 100g'[DMB]) is present in the Kishk samples.

A PCA was performed and a PC biplotwas produced with Kishk samples 3 and 25excluded. In this analysis,figure 6A sum-marises 60.6 % of the correlation matrixbetween the concentration of minerais inthe Kishk samples. The variables did notshow any obvious clustering between thesampi es of Kishk, but samples 18 and 19are outliers. The two clusters of the Kishksamples are approximately at right anglesdenoting independence (figure 6A). Thereare three large positive (Cu, Mn, Zn) andthree large negative Ioadings (Ca, K, Na).Of the remaining mineraIs, Mg and Fe donot fit into the pattern of distribution of otherminerais while P is distinct in that it has ahigh loading on PC 1 and none on PC 2. It issurmised that PC 1 reflects the abundance of

the mineraIs while PC 2 measures the elec-trochemical property of the soil.

Only the P, Ca and Fe contents of thethree commercial and two laboratory-madeKishk samples have been reported [1, 15,33, 38], and these have been compared withour 23 samples of Kishk. A further PCAwas performed on the correlation matrix ofthe three available mineraIs. The first PC(figure 6B) accounted for 65.6 % of the totalvariation and sirnply represented the mean ofthe three mineraIs. The positions of sam-pIes along PC 1 are similar to the samplesshown infigure 6A. PC 2 accounted for afurther 21.2 % of the variation, and suchmapping of the Kishk samples is more akinto PC 3 (not shown) for the data infigure 6Athan its PC 2. Samples which were high in Pare positioned towards the top of the plot,while those that were high in Fe are found atthe bottom. Both the commercial and lab-oratory-made Kishk samples had Iow P, Caand Fe contents when compared withLebanese Kishk samples, and only one com-mercial sample was the only true outlier.

Kishk - nutritional composition

4. CONCLUSION

It is evident from the current results thatKishk has considerable dietary potential asa source of fibre, amino acids, minerais andselenium contents, but is deficient in cer-tain vitamins.

ACKNOWLEDGEMENTS

The authors thank Mr. R. Brown of the Uni-versity of Reading for his assistance with theamino acid analysis, and Dr. J. Dixon and Mr.T. McCreath for skilled technical assistance.SAC and BioSS receive financial support fromthe Scottish Office of Agriculture, Environmentand Fisheries Department (SOAEFD).

REFERENCES

[1] Abou-Donia S.A., Egyptian fresh fermentedmilk products, N. Z. J. Dairy Sei. Technol. 19(1984) 7-18.

[2] Anonymous, Fermented Milks and Health, Pro-ceedings of a Workshop held on 25-26 Septem-ber 1989 at Arnhem in the Netherlands, NIZO,Ede, 1989.

[3] AOAC, Official Methods of Analysis, Helrich K.(Ed.), 14th edn., Association of Official Ana-Iytical Chemists, Inc., Virginia, USA, 1984,pp. 836-837.

[4] AOAC, Official Methods of Analysis, Helrich K.(Ed.), Vol. 2, 15th edn., Association of OfficialAnalytical Chemists, Inc., Virginia, USA, 1990,pp. 800-801.

[5] Barclay M.N.I., MacPherson A., Selenium con-tent of wheat for bread making in Scot land andthe relationship between glutathione peroxidase(EC 1.11.1.9) levels in whole blood and breadconsumption, Br. J. Nutr. 68 (1992) 261-270.

[61 Barclay M.N.I., MacPherson A., Dixon J., Sele-nium content of a range of UK foods, J. FoodCompos. Anal. 8 (1995) 307-318.

[7] Bieri J.G., TolliverT.J., Catignani G.L., Simul-taneous determination of œ-tocopherol andretinol in plasma or red cells by high pressureliquid chromatography, Am. J. Clin. Nutr. 32(1979) 2143-2149.

[8] Bronzetti G., Antimutagens in food, Trends FoodSci. Technol. 5 (1994) 390-395.

[9] Brubacher, G., Miller-Mulot W., South-gate D.A.T., Methods for Determination of Vita-mins in Food, Elsevier Applied Science, London,1985, pp. 129-140.

447

[10] Butriss J.L., Diplock A.T., High performanceliquid chromatographie methods for vitamin E intissues, in: Packer L.E. (Ed.), Methods in Enzy-mol ogy, Academie Press, New York, 1984,pp. 131-138.

[II] Cadina M.A., Robinson R.K., The acceptabilityof yoghurt--cereal mixtures to a rural commu-nit y in Mexico, Eco!. Food Nutr. 8 (1979)169-174.

[12] Cummings J.H., Frelich W. (Eds.), Metabolicand Physiological Aspects of Dietary Fibre inFood - Dietary Fibre Intakes in Europe, A sur-vey conducted by members of the managementcommittee of COST 92, Commission of theEuropean Communities, Luxembourg, 1993.

[13] Cummings J.H., Bingham S.A., Heaton K.W.,Eastwood M.A., Fecal weight, colon cancer riskand dietary intake of nonstarch polysaccharide(dietary fibre), Gastroenterology 103 (1992)1783-1789.

[14] Economidou P.L., Steinkraus K.H., Greek tra-hanas, Symposium on Indigenous FermentedFoods - Thailand, 1977, cited by Steinkraus,1983.

[15] FAO, Food Composition Tables for the NearEast, Food and Nutrition Paper No. 26, Foodand Agriculture Organisation of the UnitedNations, Rome, Italy, 1982, pp. 100-103,226,231.

[16] FAO/WHO, Energy and Protein Requirements,Technical Report Series No. 724, World HealthOrganisation, Geneva, Switzerland, 1985,pp. 64-66.

[17] Frëlich W., Nyman M., Minerais, phytate anddietary fibre in different fractions of oat grain,J. Cereal Sei, 7 (1988) 73-82.

[18] Hershey J.W., Oostdyke T., Determination ofarsenic and selenium in environmental and agri-cultural samples by hydride generation atomicabsorption spectrometry, J. Assoc. Off. Anal.Chem. 71 (1988) 1090-1093.

[19] Holland B., Unwin 1.0., Buss D.H., Milk Prad-ucts and Eggs, The Fourth Supplement toMcCance and Widdowson's - The Composi-tion of Foods, The Royal Society of Chemistry,Cambridge, UK, 1989, pp. 16-19.

[20] Holland B., Welch A.A., Unwin 1.0., Buss D.H.,Paul A.A., Southgate D.A.T., McCance andWiddowson' s - The Composition of Foods, 5threvised and extended edition, The Royal SocietyofChemistry, Cambridge, UK, 1991, pp. 24-25.

[21 ] Ibanoglu S., Ainsworth P., Wilson G.,Hayes GD.,Effect of formulation conditions on the nutri-ents and acceptability of trahana, Food Chem. 53(1995) 143-147.

[22] Ibanoglu S., Ainsworth P., Wilson G.,Hayes GD.,Effect of formulation on protein breakdown, invitro digestibility, rheological properties andacceptability of Trahana, a traditional Turkishcereal food, Int, J. Food Sci. Technol. 30 (1995)579-585.

448 A.Y. Tamime et al.

[23] Ibanoglu S., Ainsworth P., Hayes G.D., The in sensory properties of skimmed milk and Rayebvitro protein digestibility and content of thiamin Kishk, Acta Aliment. 21 (1992) 67-76.and riboflavin in extruded trahana, a traditional [39] Spakman D.H., Stein W.H., Moore S., Auto-Turkish cereal food, Food Chem. 58 (1997) matie recording apparatus for use in the chro-141-144. matography of amino acids, Anal. Chem. 30

[24] IDF, Cultured Dairy Foods in Human Nutrition, (1958) 1190-1206.Bulletin No. 159, International Dairy Federa- [40] Speek AJ., Schrijver J., Schreurs W.H.P., Flu-tion, Brussels, Belgium, 1983. orometric determination of total vitamin C in

[25] lOF, Nutrition and Metabolism, Bulletin No. foodstuffs and beverages by high performance166, International Dairy Federation, Brussels, liquid chromatography with pre-column deriva-Belgium, 1983. tization, J. Agric. Food Chem. 32 (1984)

[26] lOF, Role of Milk Protein in Human Nutrition, 352-355.

Bulletin No. 253, International Dairy Federa- [41] Spiller G.A. (Ed.), Dietary Fibre in Humantion, Brussels, Belgium, 1990. Nutrition, 2nd edn., CRC Press, Boca Raton,

[27] lOF, Cultured Dairy Products in Human Nutri- FL,1993.

tion - Dietary Calcium and Health, Bulletin No. [42] Spiller GA, Kay R.M. (Eds.), Medical Aspects255, International Dairy Federation, Brussels, of Dietary Fibre, Plenum Medical Book, NewBelgium, 1991. York, 1980.

[28] Jamalian J., Pellett P.L., Nutritional value of [43] Steinkraus K.H. (Ed.), Handbook of IndigenousMiddle Eastern foodstuffs - IV. Amino acid Fermented Foods, Marcel Dekker, New York,composition, J. Sci. Food Agric. 19 (1968) 1983,pp.244-299.378-382. [44] Swain J.F., Rouse LL., Curley C.B., Sacks F.M.,

[29] Lapré J.A., van der Meer R., Dietary modulation Comparison of the effects of oat bean and low-on colon cancer risk: the roIes of fat, fibre and fibre wheat on serum lipoprotein levels andcalcium, Trends Food Sei, Technol. 3 (1992) blood pressure, N. Engl. J. Med. 322 (1990)320-324. 147-152.

[30] Lazos E.S., Aggelousis G., Bratakos M., The [45] Tamime AY., O'Connor T.P., Kishk - a driedfermentation of Trahanas: a milk-wheat flour fermented milklcereal mixture, Int. Dairy J. 5combination, Plant Foods Hum. Nutr. 44 (1993) (1995) 109-128.45-62. [46] Tamime A.Y., Muir D.D., Barclay M.N.J.,

[31] MAFF, The Analysis of Agricultural Material Khaskheli M., McNulty D., Laboratory-madeReference Book, No. 427, Her Majesty's Sta- Kishk from wheat, oat and barley: 1. Produc-tionery Office, London, UK, 1986. tion and comparison of chemical and nutritional

[32] McCleary B.V., Glennie-Holrnes M., Enzymic composition of Burghol, Food Res. Int. 30

quantification of (1~ 3) (1~4 )-~-D-glucan in (1997) 311-317.

barley and malt, 1. Inst. Brew. 91 (1985) [47] Tamime A.Y., Muir D.D., Barclay M.N.I.,285-295. Khaskheli M., McNulty D., Laboratory-made

[33] Morcos S.R., Hegazi S.M., EI-Damhougy S.T., Kishk from wheat, oat and barley: 2. Compo-

Fermented foods in common use in Egypt. sition quality and sensory properties, Food Res.

1. The nutritive value of Kishk, J. Sci. Food Int. 30 (1997) 319-326.

Agric. 24 (1973) 1153-1156. [48] Tamime A.Y., Barclay M.N.J., Amarowicz A.,

[34] Passmore R., Eastwood M.A., Davidson and McNulty D., Kishk - a dried fermented

Passmore Human Nutrition and Dietetics, 8th milklcereal mixture. 1. Composition of gross

edn., Churchill Livingstone, Edinburgh, 1986, components, carbohydrates, organic acids and

pp. 40-53, 103-131. fatty acids, Lait 79 (1999) 317-330.

[35] Paul A.A., Southgate D.A.T., Russell J., First [49] Tzanetakis N.M., Lait fermentés traditionnels

Supplement to McCance and Widdowson's- produits en Grèce, in: Zervas N.P., Hatzimina-

The Composition of Foods, The Royal Society glou J. (Eds.), The Optional Exploitation of

of Chemistry, Cambridge, UK, 1980, pp. 14-15. Marginal Mediterranean Areas by ExtensiveRuminant Production Systems, Proceedings of

[36] Renner E., Milk and Dairy Products in Human an International Symposium (Thessaloniki,Nutrition, Volkswirtschaftlicher Verlag, Munich, 1994) by HSAP and EAAP and sponsored by1983. EU (DGVI), FAO and CIHEAM EAAP Publi-

[37] Renner E. (Ed.), Micronutrients in Milk and cation No. 83, Thessaloniki, Greece, 1996,Milk-Based Food Products, Elsevier Applied pp. 393-399.Science Publishers, London, 1989. [50] Wood P.J., Oat ~-glucan-physicochemical prop-

[38] Salama A.A., Damir AA., Mohammed M.S., erties and physiological effects, Trends FoodEffect of cooking on nutrients, microbial and Sci. Technol. 2 (1991) 311-314.