Lactic Acid Bacteria

Transcription

1 APPLIED AD EVIROMETAL MICROBIOLOGY. June 1989, p /89/616177$2.O/O Copyright 1989, Amerian Soiety for Mirobiology Vol. 55. o. 6 Freon 11 Extration of Volatile Metabolites Formed by Certain Lati Aid Bateria RALPH P. TRACEY',* AD TREVOR J. BRITZ Vitiultural an d Oenologial Resear(h Inistitute, Stellenibos(h, antid Departmlenut of' Mi(robiology, University of 'tlie OIrantige Free State, 93 Bloel,nfotmtein,2 Soidtli Afiri(a Reeived 7 ovember 1988/Aepted 3 Marh 1989 The volatile metabolites formed by 18 lati aid bateria, representing three genera, were extrated from a omplex medium by using a Freon 11 extration method. The Freon extrats were then analyzed by apillary gas hromatography, and ertain extrats were analyzed by gas hromatographymass spetrometry. A total of 35 major peaks, of whih 2 were positively identified, were used to differentiate between the various strains. On the basis of the results obtained, it was possible to differentiate between the members of the genera Latobaillus, Pedioous, and Leuonosto, as well as between various speies within the genus Leuonosto. Of the 1 Leuonosto oenos strains inluded in this study, 9 yielded similar results, but it was still possible to differentiate between the various strains. L. oenos B66 differed from the other L. oenos strains. Use of the Freon 11 extration tehnique to determine volatile metabolites formed by lati aid bateria was shown to be highly reproduible and of great value. Furthermore, ertain ompounds not previously known to be formed by lati aid bateria were found. The use of end produt analysis by gasliquid hromatography for the haraterization and identifiation of bateria is well doumented. In the eighth edition of Bergev 's M,iailitil of Determinatiie Bataeriology (3), as well as in Bergevy's Maniual ofsystemati Bateriology (7, 13), a large number of generi desriptions inlude metaboli end produts as riteria for the inlusion of an organism in a partiular genus. Most gasliquid hromatography methods suffer from the disadvantage that a pretreatment of the sample is often neessary and that this is sometimes followed by a derivatization proedure (14). Even though Thornhill and Cogan (14) regard pretreatment and derivatization as a disadvantage, it must be borne in mind that pretreatment or derivatization of samples an ontribute greatly to the analyses and, in so doing, an inrease the value of the analyses and subsequent grouping or lassifiation of speies or strains. Diret injetion of samples to measure the fermentation produts formed by bateria has only a limited appliation, beause in most ases only a few ompounds are determined. Beause only the main ompounds of fermentation are determined, organisms may be grouped together when, in reality, their relationship to eah other is not very high. It is therefore neessary that additional tests be arried out to onfirm the relationships. Reently a reproduible tehnique, involving the use of a Freon 11 extration method, was published to determine the amounts and identities of terpenes in grape juie and wine by apillary gas hromatography (1). The use of Freon 11 for the extration of volatiles has also gained wide aeptane in the field of wine flavor researh. It has been used to determine the amounts and identities of volatile yeast metabolites in fermenting grape musts, for whih the gashromatographi analysis of Freon 11 extrats revealed over 72 esters and other metabolites (1). Furthermore, Freon 11 has various other advantages: it is relatively heap, easily obtainable, noninflammable, and nontoxi (11). Two disadvan * Corresponding author. t Present address: Department of Mirobiology. University of the Orange Free State, P.O. Box 339, 93 Bloemfontein. South Afria. tages, however, are that Freon 11 does not extrat ethanol, whih is an important metabolite formed by some bateria, and that beause Freon has a boiling point of 23.6 C, all the work must be arried out below this temperature in aironditioned rooms. Sine ethanol is not one of the major metabolites formed during gluose atabolism by lati aid bateria, it is possible that the Freon 11 metabolite extration tehnique will be of value in the differentiation of speies or even strains of the lati aid bateria. The purpose of this study was to determine whether this metabolite extration tehnique ould be used in the differentiation of lati aid bateria and, more speifially, strains of Leuonosto OeUIOS whih are of importane to the wine industry. MATERIALS AD METHODS Baterial strains. The strains used in this study inlude the type strains Latobaillus asei CIB 81, Pedioous dalumltiosius DSM 2331, Leuonosto relnoris DSM 2346, L. Dextraniulm DSM 2484, L. latis DSM 222, L. mesenteroides DSM 2343, L. palraimeseniteroides DSM 2288, and L. oenios DSM 2252, as well as referene strains Latobaillius pl(uiitairin CIB 6376, L. oenos DSM 2255, and L. oecios DMS L. oenios B66, 37D, 19, and 36 were seleted as representative entrotype strains of four lusters obtained during a numerial taxonomi study on malolati bateria from the main wineproduing areas in the world (15). Three L. oenios strains (L5x, 3, and 7B) were seleted on the basis of having the lowest similarity to the entrotype strain (15) of eah of the three partiular lusters. The strains were preserved by lyophilization on filter paper disks (6). Culture onditions. For eah strain, 5 ml of enrihed tomato juie broth (15) was inoulated with a filter paper disk and inubated at 25TC for 48 h. A 1ml sample of this ulture was then used to inoulate 3 ml of the same medium. Exellent growth was found after inubation at 25 C for 4 days, whereafter the ells were removed by entrifugation for 2 min at 18, x g. The supernatant was reovered and Downloaded from on Deember 18, 218 by guest 1617

2 1618 TRACEY AD BRITZ Appt. EVIRO. MICROBIOL. immediately ooled to C for extration with Freon 11 and subsequent gashromatographi analysis. Extration proedure and onentration of the extrat. The proedure and extration apparatus as desribed by Marais (1) were used. The internal standard, 2ethyl1hexanol (8 p.g/liter) was added to 25 ml of the ooled sample in a measuring flask and arefully mixed. A 2ml portion of Freon 11 and the ooled sample ontaining the internal standard was arefully poured into the extration apparatus desribed previously (1). This was then installed with its bottom in 5 mm of ie to prevent the formation of an emulsion. A olleting funnel and ondenser, through whih ethanol (at 5 C) was irulated, were fitted to the extration unit. A 25ml pearshaped flask ontaining 2 ml of Freon was fitted to the extration unit, and the fitted flask was then immersed in a water bath at 35 C. The extration was performed for 2 h at a ontrolled temperature of 2 C. To onentrate the extrat, a Vigreux frationing olumn (27 by 2 mm) and an air ondenser (55 by 13 mm) were fixed onto the flask ontaining the extrat to failitate reflux. The flask was plaed in a water bath at 35 C, with the room temperature at about 2 C, and the extrat was onentrated to 2 ml. The flask was then plaed in solid, to freeze out any traes of water from the extrat. The onentrated extrat was then transferred to a 3ml pearshaped flask with a Pasteur pipette. A small ondenser (22 by 8 mm), fitted with a spiral made from Teflon, was fitted to the flask, and onentration was ontinued under partial reflux to approximately.1 ml. The extrats were then stored at 12 C prior to analysis. Gas hromatography onditions. The gashromatographi analyses were arried out on a Hewlett Pakard 588A gas hromatograph equipped with a hydrogen flame ionization detetor and dual integrators. A Carbowax 2M fusedsilia apillary olumn (5 m by.31 mm [inner diameter]) was used under the following operating onditions: injetion temperature, 2 C; detetor temperature, 25 C; olumn temperature, 6 C for 1 min, then inreased by 1 C/min to 19 C and held at 19 C for 3 min; flow rate of arrier gas (helium), 1.5 ml/min; split flow rate, 12 ml/min; split ratio, 9:1; septum purge, 6 ml/min; hydrogen flow rate, 3 ml/min; air flow rate, 3 ml/min; injetion volume, 1 [Ll. Identifiation of the major peaks. A Finnigan gas hromatographmass spetrometer (model 45) equipped with a Carbowax 2M fusedsilia olumn (5 m by.31 mm [inner diameter]) and gashromatographi operating onditions, as desribed above, were used to identify the major peaks formed by the L. oeilos strains on the basis of retention time and mass spetra. The identities of the various hemial ompounds were onfirmed by omparing their mass spetra with those of authenti standards analyzed under similar onditions. The following operating onditions were used: interfae temperature, 19 C; manifold temperature, 1 C; olumn flow rate, 1.4 ml/min; split ratio, 48:1, eletron multiplier voltage, 1,6 V; sanning rate, 35 to 45 amu/ehs with a.1s delay between sans. Determination of the relative onentrations. The relative onentrations of the major ompounds were alulated by using the internal standard alibration method: relative onentration (b) = [area (b) x onentration (al)]/area (a), where a is the internal standard and b is an unknown ompound. Reproduibility. The reproduibility of the tehnique was determined by analyzing the Freon 11 extrats of four separately grown ultures of eah of Laztobaillits asei CIB 81 and L. ocelos DMS The mean, standard deviation, and oeffiient of variation for eah separate peak was alulated from the relative onentrations of eah peak. RESULTS AD DISCUSSIO Reproduibility. The data for the four separately grown ultures of Latowbaillits asei CIB 81 and L. oenios DSM 2252 are given in Table 1. The oeffiients of variation were very low, and it was therefore possible that all the peaks ould be used as an aid in the differentiation of the strains. The high level of reproduibility also indiates that this tehnique is reproduible when determining volatile metabolites formed by lati aid bateria. Basal medium. The omposition of the Freon extrat of the basal medium (enrihed tomato juie broth) is given in Table 2. The relative amounts of volatile produts were found to be similar for eah bath of basal medium used. A hromatogram showing the volatiles present in the basal medium is presented in Fig. IA. With the use of the Freon 11 extration method, 12 ompounds were found in the basal medium. This omplex medium was speifially used beause it was previously found to ontain all the required amino aids and other growth fators (15). Metabolites produed by the various lati aid bateria. The relative onentrations of the major peaks of all the strains are given in Table 2. A total of 35 major volatiles ould be differentiated, of whih 2 were positively identified. By using the Freon 11 extration tehnique, it is possible to identify several previously unknown ompounds formed by lati aid bateria. o previous literature on the formation of these ompounds by lati aid bateria ould be found. These ompounds inlude 3methyl2butanol, ethyl2hydroxypropionate, benzaldehyde, 3(methylthio)1 propanol, (x,dimethylbenzenemethanol, and benzenemethanol. Of these ompounds, two were present in the basal medium but at muh lower onentrations than produed by the organisms used in the study. It is possible that by applying this tehnique to other bateria, even more ompounds formed by bateria will be disovered. These unique metabolites ould play a vital role in differentiating between bateria and thus possibly ontribute to an even better lassifiation of bateria. A fat that must be taken into onsideration is that a rih omplex basal medium was used in the study, and this will greatly influene the type and onentration of metabolites produed. Differentiation between the various lati and bateria. After visual omparison of the hromatograms of the various strains on the basis of the presene or absene and/or the relative onentrations of the ompounds, we used the following ompounds to differentiate between the various strains (peak numbers also given): isobutanol (peak 1), isoamyl alohol (peak 2), aetoin (peak 3), 3methyl2 butanol (peak 4), ethyl2hydroxypropionate (peak 5), aeti aid (peak 6), butanoi aid (peak 9),,xdimethylbenzenemethanol (peak 13), hexanoi aid (peak 14), heptanoi aid (peak 17), otanoi aid (peak 19), nonanoi aid (peak 21), deanoi aid (peak 22), dodeanoi aid (peak 25), tetradeanoi aid (peak 31), and the unidentified ompounds in peaks 8, 1, 18, 23, 24, 27, 28, 33, and 34. The various strains of lati aid bateria (Table 2: Fig. 1B) ould easily be differentiated. Latobailulis asei (Fig. ib) ould be separated from the other strains on the basis that it formed the largest amount of ethyl2hydroxypropionate (peak 5) of all the strains studied (Table 2) and also produed large amounts of aeti aid (peak 6), (xdimethylbenzenemethanol (peak 13), benzenemethanol (peak 15), Downloaded from on Deember 18, 218 by guest

3 VOL FREO 11 EXTRACTIO OF VOLATILE METABOLITES 1619 TABLE 1. Relative onentrations, means, standard deviations, and oeffiients of variation for the four Freon 11 extrats of eah of Latobaillus asei ICB 81 and L. oenos DSM 2252 Latobailluds asei L. oenios Compound Peak Retention no. time (min) Mean Mean onn oen ± SD Coeffiient of Mean onn Coeffiient of variation" K) MaonSD variation e( Isobutanol ± ± Isoamyl alohol ± ± Aetoin ± Methyl2butanol ± ± Ethyl2hydroxypropionate ,122.5 ± ± Aeti aid ± ± Benzaldehyde ± ± Unidentified ± Butanoi aid ± ± Unidentified ± Pentanoi aid ± ± (Methylthio)1propanol ± ± (x,odimethylbenzenemethanol ± ± Hexanoi aid ± ± Benzenemethanol ± ± Phenylethanol ± ± Heptanoi aid ± ± Unidentified ± ± Otanoi aid ± ± Unidentified ± ± onanoi aid ± ± Deanoi aid ± ± Unidentified ±.42.9 Dodeanoi aid ± ± Unidentified ± Tetradeanoi aid ± ± Unidentified ± ± " Coeffiient of variation = (standard deviation/mean) x 1. nonanoi aid (peak 21), deanoi aid (peak 22), and tetradeanoi aid (peak 31). The formation of the highest onentrations of hexanoi aid (peak 14) and dodeanoi aid (peak 25) by the Latobaillus planttariin strain and the formation of high onentrations of aetoin (peak 3), otanoi aid (peak 19), and deanoi aid (peak 22) and a small amount of aeti aid (peak 6), as well as the absene of two unidentified ompounds (peaks 8 and 1) and nonanoi aid (peak 21), ould be used to differentiate this strain from the other strains inluded in this study (Table 2). The P. dainnltlosl(s strain was differentiated from the others on the basis that it formed the largest amount of isoamyl alohol (peak 2) of all the strains studied; it also formed high onentrations of ethyl2hydroxypropionate (peak 5), and it formed two unidentified ompounds (peaks 8 and 1). The abovementioned differenes for these organisms an also be omplemented with the use of morphology and ell wall omposition as previously desribed (15). When the various Lelionosto speies were ompared with eah other, it was found that although they generally yielded similar profiles, they ould still be separated (Table 2). L. rzemo)ris was separated mainly on the basis of the high onentrations of 2phenylethanol (peak 16) and dodeanoi aid (peak 25) formed and the presene of high onentrations of the unidentified peaks 26 and 33. This strain was the only organism that produed the unidentified ompound in peak 33. The L. devtroaniiiimi strain ould be differentiated on the basis of the amounts of butanoi aid (peak 9), nonanoi aid (peak 21), deanoi aid (peak 22), and an unidentified ompound (peak 34) formed. The high onentration of 3(methylthio)1propanol (peak 12) and the low onentrations of (x,o.dimethylbenzenemethanol (peak 13) and heptanoi aid (peak 17) formed were used to separate L. latis from the other strains. Aording to Cogan et al. (4), L. latis forms aetoin (a fat that was onfirmed in this study). even though only at low onentrations (Table 2). L. iiesenteroides was differentiated from the other strains on the basis of the formation of the largest amount of aetoin (peak 3), as well as the relatively high onentrations of 3methyl2butanol (peak 4), the presene of the unidentified ompounds in peaks 18 and 24, and the absene of aeti aid (peak 5) and the unidentified ompound in peak 1 (Table 2). The formation of the largest amounts of o,adimethylbenzenemethanol (peak 13), heptanoi aid (peak 17), and the unidentified ompounds in peaks 18, 2. 27, and 28 made it possible to separate L. pt1e(lra1ieseniteroides from all the other strains. This was the only strain to form the unidentified ompound 28, no ethyl2hydroxypropionate (peak 5), no aeti aid (peak 6), and no unidentified ompounds 8 and 1. and it harateristially produed only low onentrations of butanoi aid (peak 9), hexanoi aid (peak 14). otanoi aid (peak 19), and tetradeanoi aid (peak 31). Garvie (5) found that it was diffiult to separate L. paiwt/iteseniteioides from the nondextranforming strains of L. mieseniteroides, whereas this study (Table 2) showed that these speies an easily be distinguished. The formation of aeti aid (peak 6) an also be used to separate them from various other Leuonosto speies, sine the L.,nesetitenroides and L. p(iriilnesetlteroiles ultures used in this study did not form any aeti aid (peak 6). Downloaded from on Deember 18, 218 by guest

4 162 TRACEY AD BRITZ O C4 loc ' rc tr C e4 \'Ct oc finoci k C] In 1x o m I n r' kl O In l kr O C4 OInInC InInCf C' Inkr O7 kc D In' C ti r l In UO In In r _ Inm..* ~~1f In ZC b el In <; ~t O e ee "C n In C I IC t Itn4 r e C _ C _ C C O It o C It r In " C In C 6 _ rl~ _ "It It 1 'C C It WI It a O kf xo I nk O m C ' C C ' b In O Ot e^,i OIC So OC,It l. 'I 4 In O I In OtO^m In e' C' o n _~In t t e^ ee, eln,o 'IC O I e n e x = I aizt rq k:c 4 e4 Cnk r ZOOZm O In ' t In C 'IC IC x l C O O C~ ~t l ltiot 'I x In O er F,lI: C4 CV 'r In O4 t Z 'IC In C': In ' InC' t m Io t t C t l e I 1ZC In O I 4 I. C _ tz) _ tcr C,, ~I Ot^ O C l O C:t t x *. ~t e4r O o _l o r x e'o t e~ t O 1. O. l. t F In t e^ o _ O In *r Z eo.e k l O u o Z.~C.g ZC ~t q (< C etc O ax^ C. _ O e~ t e4 O In In X, e.$ C1 O S) U m :i E 1 C r x in.in.) C.C'. C. a C..).'C a...., CC O. E E Z.o.= a ue E _~ _ ~4 e~ In4 ' 'C e4 On " InOO 1 " 5 OC en 1 S Oe1S ~l ' 4 :: l~ InIn~4 1 l r ' O 4 In mf O Z C In b ri e^ o C I )n o O t ' e4 4mCC O In' ec^ e^ e4 t w o o In o O' I OO = " I e' = = r o x Crf 4 li Co en C O r ( ri r ii t IC " li = lt = r m C In f Int In i In a) li IC r i ; r 11 ~~~~~~~~~~~~~~~ In 4 en Co O n CI C~ " rn 4~ r 1t In~ n o) I V m all, r In) S 'Cr ioo i ol 1 V) a, n ', 'I C: In t t In In O 6 C" 7Cn CC I In Inet In ^ o o e O O O C C C m oo ut In In oo Or _ O t 1 1 I r It ^ m I l 'I a In In un _C = en ltin 'C ooaoc mil_ 'InC OO (r)c'ic In O Cnllt 1 C.C.C.C.C.C.C.C.C.e. en en f)f) _OU l lac ' O ') U s > O~V>n= O tqnn OQQQQ q CZ oq) Q Z,E Q o > ; _ n = _ E CZ* o _. C' m..l o O b _ o O Cy O C rrol O z On ol 11 O ^t o m knv kr) '" V),' O all O~ I O In CZ) ZtZ _ DO ) tt6^ ~ ~ ^ O 6mOZ; DD ad S D <O 4 C) m li APPL. EVIRO. MICROBIOL., 1.4 = = = = = C = Downloaded from on Deember 18, 218 by guest

5 VOL. 55, 1989 FREO 11 EXTRACTIO OF VOLATILE METABOLITES i= _ w n ofo > CD, ) C) r P. P= _A C~~~~~~~~~~~C CD~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C m 3 A (1) Z 3 1 4~~~~~~~~~~~~~~~~~ ~~~~ CDo V Q ) 1 )~~~~~~~~~. ~~~ )~r C)~~~~~~~~~~~~~~~~~ 2t ZrI n Downloaded from on Deember 18, 218 by guest C ) ADg~~~~

6 1622 TRACEY AD BRITZ APPL. EVIRO. MICROBIOL. Strain uster" TABLE 3. Peaks used to differentiate between the various L. oenos strains Relative onns" of peak no.': it) t) 35 DSM DSM DSM B66 A (entrotype) L5x A D B (entrotype) B C (entrotype) D (entrotype) B D Strains not designated as entrotype straiins are the strains seleted as having the lowest similarity to the entrotype strain. " Relative onentrations are expressed as follows:, t); +, 1 to 5(); 2+, 51 to 1i)(); 3+, 11)1 to 2); 4+, 2111 to 3t)(); 5+, 3t)1 to 4)1); 6+, 4t1 to 5t); 7+, 5()1 to 75); 8 +, > 75). Peak identities arle given in Table 2. Garvie (5) redued the speies in the genus Leiwonosto from six to four by inluding L. de.vtranium and L. (remnoris as subspeies of L. /nesenteoidies. The results from this study (Table 2) generally onfirm that L. remoris and L. dextrani(um are losely related. However, L. Inesenteroites does not appear to be losely related to them. The L. mesenteroides strain appeared to be more losely related to the L. paralnesenteroides strain than to the L. dexrwanium and L. remnoris strains. More representative strains from these four speies (or subspeies, as proposed by Garvie [5]) will have to be studied to onfirm the above observations. In a previous study (15), it was shown that the differentiation between the leuonostos and pediooi an be problemati and that ell wall analysis and morphology are important riteria in differentiating the leuonostos from other genera. Morphologially, Latohaillus an easily be distinguished from the ooid genera Leionosto and Pedioous. Separation of the three genera Lato bailluts, Pedioous, and Leitonosto by using volatile metabolites was omparatively easy, espeially when omparing the differenes in the amounts of isoamyl alohol (peak 2), aetoin (peak 3), ethyl2hydroxypropionate (peak 5), aeti aid (peak 6), and hexanoi aid (peak 14) formed. The presene or absene of several other metabolites is also of value. However, the results obtained in this study, with the exeption of results for the L. oe)cos speies, were based on only a single representative of eah speies and must therefore be arefully interpreted before any general onlusions or reommendations an be made. One positive onlusion that an be made is that the Freon 11 tehnique an be of great value in differentiating between lati aid bateria. Differentiation of the L. oenos strains. Overall, the majority of L. oenios strains yielded similar metaboli profiles, but the relative amounts of eah ompound formed differed markedly. Of the 35 important ompounds deteted in this study, only 13 yielded major differenes that ould be used to differentiate between the L. oeuos strains (Tables 2 and 3). The use of these 13 ompounds in the separation of the L. oecuos strains used in this study is shown in Table 3. These inlude aetoin (peak 3), ethyl2hydroxypropionate (peak 5), and the unidentified ompounds in peaks 8, 1, 24, and 27 (Table 3). Although the L. oeiios strains ompared favorably with the L. oenos referene strains regarding the presene or absene of ompounds 1 to 22, marked differenes were found for the remaining ompounds. Strain B66 (Table 2) was separated from the other strains sine this organism formed the highest onentrations of the unidentified ompounds in peaks 8, 1, 34, and 35 of all the strains studied. Several other differenes are also evident. Strain B66 is the entrotype strain of luster A (15), and the volatile metabolites formed by this strain is a further onfirmation that it differs from the other strains inluded in this study. In ontrast to this, strain LSx did not form or formed muh less of the unidentified ompounds in peaks 8, 1, 34, and 35 (Table 3). L. oenos 37D and 3 were seleted as representatives of the largest phenotypi luster (luster B) from a previous taxonomi study (15). The volatilemetabolite profiles of these strains differed to a ertain degree, and they ould be separated on the basis of the presene or absene and/or onentrations of ompounds 5, 6, 9, 27, 29, 31, and 34 (Tables 2 and 3). Strain 19 ould be separated, mainly beause it was the only L. oenios strain to form large onentrations of aetoin (peak 3) and was the only strain to form the unidentified ompound in peak 23. Further profile harateristis of this organism were the absene of ethyl2hydroxypropionate (peak 5) and the unidentified ompounds in peaks 8 and 1 (Table 3). This strain had a different overall volatilemetabolite profile from those of the other L. oenios strains. It was the entrotype strain of luster C and was phenotypially fairly losely related to luster B (15). The two representative strains from luster D (36 and 7B) were separated from the other strains on the basis of the formation of less isobutanol (peak 1) and isoamyl alohol (peak 2). They ould also be separated from eah other by using aetoin (peak 3) and the unidentified ompounds in peaks 8, 1, 24, and 27 (Table 3). The overall volatilemetabolite profiles of these two strains ompared favorably with those of the other L. oenios strains. One again, it must be stressed that the data obtained and used in the differentiation of the different L. oenos strains are based on single representatives of the phenotypi lusters (15) and that any onlusions made must be seen in this light. Further work is also required on these and other strains that are to be used to indue malolati fermentation. It has been shown that L. oenos strains are the most reliable organisms whih an be used for induing malolati fermentation (2, 8, 9). Therefore, to find the L. oenos strain that is best suited for induing malolati fermentation and that will have the least negative effet on the wine, the strains will have to be tested individually in wine, and these wines must subsequently then undergo a sensory evaluation and analysis by the Freon extration method. Downloaded from on Deember 18, 218 by guest

7 Vot 55, 1989 FREO 11 EXTRACTIO OF VOLATILE METABOLITES 1623 Utility and value of the Freon I I extration tehnique. Although the Freon extration tehnique is more timeonsuming and elaborate than previously published gashromatographi tehniques for the study of volatile metabolites produed by bateria, we believe that the amount of work and time taken for the extration are duly rewarded in the results obtained, sine a wider range of metabolites are determined than by onventional diretinjetion hromnatography. Furthermore, it is possible to differentiate between speies of a partiular genus and between strains of a partiular speies. When organisms are used industrially, as with the lati aid bateria, this tehnique ould help to determine the ontribution of the organisms to the flavor of the various produts. In 1965, Pilone and Kunkee (12) showed that flavor differenes in wines in whih malolati fermentation had been indued by bateria were small. The use of different and newer strains thus warrants a new look at the possible ontribution of the miroorganisms to the flavor of wine. It is also important to try and orrelate the sensory evaluation of wines whih have undergone malolati fermentation with hemial analyses by the tehnique desribed in this paper. ACKOWLEDGMETS The tehnial assistane of Helena Kritzinger and Isabeau Bronn is aknowledged. Johan Marais is thanked for his ritial review of the paper. LITERATURE CITED 1. Akhtar, M., R. E. Subden, J. D. Cunningham, C. Fvfe, and A. Meiering Prodution of volatile yeast metabolites in fermenting grape musts. Can. Inst. Food Si. Tehnol. J. 18: Beelman, R. B., A. Gavin, and R. M. Keen A new strain of Leuionosto oeno.s for indued malolati fermentation in Eastern wines. Am. J. Enol. Viti. 28: Buhanan, R. E., and. E. Gibbons (ed.) Bergevs*s manual of determinative bateriology. The Williims & Wilkins Co.. Baltimore. 4. Cogan, T. M., M. O'Dowd, and D. Mellerik Effets of ph and sugar on aetoin prodution froni itraite by Leuonosto( Iulalis. AppI. Environ. Mirobiol. 41: Garvie, E Genus Luuouo.so. p In P. H. A. Sneath.. S. Mair. M. E. Sharpe. and J. G. Holt (ed.). Bereevs manual of systemati bateriology. vol. 2. The Williams & Wilkins Co.. Baltimore. 6. Joubert, W. A., and T. J. Britz A simple and inexpensive method for the longterm preservation of mirobial ultures. J. Mirobiol. Methods 7: Krieg,. R., & J. G. Holt (ed.) Bergey's manuail of systemati bateriology. vol. 1. The Williams & Wilkins Co.. Baltimore. 8. Kunkee, R. E Malolati fermentattion. Adv. AppI. Mirobiol. 9: LafonLafourade, S., E. Carre, and P. RibereauGavon Ourrene of lati aid bateria during the different stages of vinifiation and onservation of wines. Appl. Environ. Mirobiol. 46: Marais, J A reproduible apillary gias hromnatographi tehnique for the determination of speifi terpenes in grape juie and wine. S. Afr. J. Enol. Viti. 7: Marais, J., and A. C. Houtman Quantitative gas hroniatographi determination of speifi esters 'and higher alohols in wine using Freon extration. Am. J. Enol. Viti. 3: Pilone, G. J., and R. E. Kunkee Sensory haraterization of wines fermented with several malolaiti strains of bateria. Am. J. Enol. Viti. 16: Sneath, P. H. A.,. S. Mair, M. E. Sharpe, and J. G. Holt (ed.) Bergey's manual of systemati bateriology. vol. 2. The Williams & Wilkins Co.. Baltimore. 14. Thornhill, P. j., and T. M. Cogan Use of gasliquid hromatography to determine the end produts of growth of lati aid bateria. AppI. Environ. Mirobiol. 47:125( Traey, R. P.. and r. J. Britz A numerial taxononli study of Leoiubonsw oenos strains from wine. J. Appl. Balteriol. 63: Downloaded from on Deember 18, 218 by guest