CO-INOCULATION OF SELECTED WINE BACTERIA What is co-inoculation? Co-inoculation is the pactice of inoculating selected wine bacteia at the beginning of the winemaking pocess shotly afte yeast inoculation. This techniue is gaining in populaity because not only will it secue the malolactic fementation (MLF), but also because thee ae definite advantages that ae ecognized by winemakes and pofessionals. Malolactic fementation, the enzymatic decaboxylation of L-malic acid to L-lactic acid and cabon dioxide, is the impotant seconday fementation conducted by wine bacteia (Vesai et al., 999). Thee ae diffeent timing of inoculation possibilities with selected wine bacteia (figue ), such as co-inoculation which is the inoculation of wine bacteia at the beginning of alcoholic fementation (AF) shotly afte yeast addition, inoculation at 2/3 of the alcoholic fementation (ealy inoculation) and inoculation afte the completion of AF (seuential inoculation). DENSITY 9 8 7 6 5 4 3 2 99 Co-inoculation: addtition of bacteia 24 48 h hous afte yeast inoculation Ealy inoculation (at 2/3 of the AF) Seuential Ealy inoculation (duing (afte AF) 5 5 2 25 3 35 TIME (days) Figue : Diffeent timing of inoculation of selected wine bacteia How does it wok? Co-inoculation, whee bacteia ae inoculated biefly afte yeast inoculation gives the selected wine bacteia a moe favoable medium, mainly lowe ethanol concentations and a bette nutient availability. Since yeast gows moe vigoously, ML bacteia activity will be suppessed duing active AF, but the selected bacteia will acclimatize slowly to the inceasing alcohol levels. Bacteia tansition fom the lag to the logaithmic phase of gowth in a mixed cultue with yeast coinciding with the stat of the death phase of the yeast. This phenomenon may bing essential bacteial nutients to the system as a esult of yeast death and autolysis. Inoculation in the middle of alcoholic fementation vey often esults in a moe significant die-off of the selected ML bacteia, caused by the poduction of yeast-deived toxic compounds othe than ethanol and sulfu dioxide duing this highly active stage of AF. The most intense levels of yeast-induced antagonism by metabolites such as decanoic acid may be encounteed at this stage. Howeve unde low ph conditions (< ph 3.5) inoculation at /3 d of alcoholic fementation could be moe favoable, because at this stage all sulphu dioxide added at cush will be bound and less active against the selected wine bacteia. Most compatible yeast stains fo ealy inoculation stategies ae low poduces of SO 2, with a low to medium nitogen demand and modeate fementation kinetics.
Managing acetic acid poduction When talking about the pactice of co-inoculation, it is impotant to addess the concen of possible poduction of acetic acid by the lactic acid bacteia. Inoculation of wine with malolactic state cultues was taditionally pacticed afte the end of the alcoholic fementation, when all fementable sugas have been consumed by yeast and esidual sugas ae unde 2 g/l so as to avoid the possible poduction of acetic acid and D-lactic acid, a situation which is efeed to as piûe lactiue" (Ribéau-Gayon et al., 975). Howeve, inoculation of wine bacteia with selected yeast has been advocated in the USA since the ealy 98 s because it was felt the bacteia had a bette chance of gowing and acclimatizing in the absence of ethanol. The bacteia will not suffe fom a shotage of nutients no will they be exposed to the toxic effects of alcohol. Pevious expeiments (Semon et al., 2; Rosi et al., 23; Jussie et al., 26) showed that significant amount of acetic acid will not be poduced out of sugas duing gowth of MLB and active MLF. The tials conducted using simultaneous inoculation of bacteia with yeast (co-inoculation) always had no significant diffeence in the final acetic acid concentation. Moe ecently, in a study done by Zappaoli et al., (29) in high alcohol wines showed that in Covina and Rondinella vaieties, which ae used in the poduction of Amaone wine, the acetic acid levels wee simila, o even lowe in co-inoculation situation compaed to seuential inoculation. Fo example,.9 g/l of acetic acid wee measued in co-inoculation wheeas.2 g/l wee measued in seuential inoculation. What ae the benefits Secue and time saving One of the moe obvious advantages of co-inoculation is a bette contol ove the winemaking pocess in tems of time management and secuity of MLF completion. Jussie et al., (26) obseved a significant eduction in time fo MLF fom Chadonnay at a ph of 3.53 and ethanol ove 3% (v/v) when co-inoculation was induced with espect to seuential AF/MLF. Unde high alcohol conditions (6% and above), an impotant stess in winemaking, in the study of Zappaoli et al., (29), they not only showed that the MLF was completed successfully unde difficult conditions, but that it was completed ealie than seuential inoculation (7 days, vesus 2 days). What ae the benefits Sensoy impact Recent studies investigated the impact of co-inoculation on the wine sensoy uality. It was shown (Knoll et al., 22, Costello et al., 22, Batowsky et al., 2, Azzolini et al., 2) that selected wine bacteia have the potential to influence the aoma pofile of wines by the poduction of volatile seconday metabolites and modify the gape o yeast deived metabolites such as ethyl estes, acetate estes, acids and alcohols. These sensoy compounds ae stongly influenced by the stain of wine bacteia used fo MLF, as well as the timing of wine bacteia inoculation is vey impotant fo the wine aoma and flavo. What ae the benefits Risk management The time between the end of alcoholic fementation and the onset of malolactic fementation is a citical peiod. Unstabilized wine is still at isk fo aomatic deviations. Coinoculation with selected Oenococcus oeni can help avoid the poduction of potential spoilage compounds by fist educing the isk of spontaneous MLF duing alcoholic fementation by suppessing wild bacteia, and at the same time conducting a moe contolled MLF. This is especially impotant in ed wine with a high ph whee spontaneous MLF may occu duing AF, causing stuck AF and ise in volatile acidity (VA) (Van de continued on next page
What ae the benefits Risk management Mewe et al., 26). Duing co-inoculation, the micobiological activity of both yeast and bacteia helps to limit development of contaminating micooganisms such as heteofementative Lactobacillus species, Pediococcus, and Bettanomyces. Conseuently, the poduction of volatile phenols may be avoided. In a study done by Gebaux et al., (29), it was shown that ealy inoculation of selected wine bacteia did not allow fo the gowth of Bettanomyces, even when it was intentionally inoculated in Pinot Noi wines fom Bugundy (Fance). Pof Maet du Toit A WORD FROM OUR EXPERT Apat fom alcoholic fementation, malolactic fementation (MLF) is a seconday fementation conducted by lactic acid bacteia (LAB), fistly to educe the acidity of wine and secondly to contibute to wine aoma. Maet is cuently the Depatment head of Viticultue and Oenology, and Institute fo Wine Biotechnology at Stellenbosch Univesity (South Afica). She heads the eseach goup focusing on the ole of lactic acid bacteia (LAB) in winemaking, especially on the contibution of malolactic fementation (MLF) on wine aoma, using lactobacilli as state cultues as well as cetain spoilage mechanisms associated with wine LAB. She is the autho of 62 pee-eviewed scientific papes, 3 book chaptes, 9 pesentations at both intenational and national confeences, and gaduated 33 maste and 7 doctoal students. Oenococcus oeni is cuently still the best adapted state cultue fo MLF, especially fo low ph and high ethanol conditions and its contibution to wine aoma is well undestood. MLF state cultues can be inoculated at two stages of fementation, namely seuential inoculation, but with highe alcohol levels due to climate changes, the pessue on the stains to pefom unde these conditions is becoming challenging. This has led to inoculation at anothe stage of fementation, the co-inoculation of yeast and bacteia at the beginning of alcoholic fementation. It is impotant that co-inoculation is done within 24 hous afte yeast inoculation, othewise alcohol and the competition fom the actively fementing yeast impacts on the inoculated MLF state. A cucial facto is to ensue that the yeast and bacteia ae compatible; theefoe yeast selection needs to be consideed caefully. The biggest uestion with egads to this technology is the potential poduction of acetic acid fom sugas in the must. Howeve, in the last 7 yeas of being involved in co-inoculation eseach it was neve expeienced that co-inoculation yielded significantly highe levels of acetic acid. Co-inoculation has a numbe of advantageous. Fistly, the must contains all the necessay nutients needed by the bacteia and theefoe the addition of exta nutients is not necessay. Secondly, the completion of MLF is faste compaed to seuential inoculation, which means that SO 2 can be added soone and the potential of micobial spoilage is educed. Futhemoe, with co-inoculation esults in bette implantation and out-competing of the natual LAB floa, which means the stain inoculated is the one that will dominate MLF. The othe cucial facto is that thee is no o limited alcohol pesent in the must which ensue highe suvival ates and vitality of the inoculated stains. Wines made with a co-inoculation stategy has a diffeent aoma pofile than wines made with seuential inoculation, they ae peceived as moe fuity, balanced and a fulle body. Afte MLF the wines ae also bette integated and in hamony at such an ealy stage. Co-inoculation is a tool that can be used to ensue poblems nomally associated with some seuential inoculations ae no longe pat of the euation, as well as to divesify you wine style though the poduction of diffeent aoma compounds o atios of aomas in the final wine. This technology has also opened the oppotunity fo othe wine LAB, such as Lactobacillus plantaum to be used in the futue as MLF state cultues, as the matix and challenges ae much less compaed to seuential inoculation.
THE RESULTS. MLF Length and Reliability Co-inoculation will shoten significantly the length of MLF compaed to seuential o even moe so, spontaneous MLF. In seveal studies, those esults wee consistently epeated. Fo example, figue 2 shows the esults of vaious tials caied out in diffeent vaietals, ove diffeent vintages and conditions, as well as with diffeent selected wine bacteia compaing co-inoculation to spontaneous MLF. In all cases, the length is significantly educed. Not only co-inoculation shotens MLF but is also vey eliable in a vaiety of situations Time (days afte yeasting) Figue 3 shows the esults of MLF length and completion unde limiting conditions in a 26 Amaone wine made out of patially died gapes (ph 3.3, Alcohol 5.5% v/v, Total SO 2 5 mg/l). Zappaoli and Tossi (26), could successfully achieve malolactic fementation using co-inoculation techniues (bacteia inoculation day afte yeast) with VP4 MBR cultue compaed to seuential inoculation and spontaneous MLF, which had not stated afte 9 days. Malic acid, L-lactic acic (g/l) 8 7 6 5 4 3 2 2.5 2.5.5 m m m m m m m m m m m m m 8 6 24 32 4 48 56 64 72 8 88 96 Figue 3: melot Co-inoculated Spontaneous MLF gamay syah gamay 2 cab-fanc cab-fanc 2 Time (days) Malic acid consumption (solid symbols) and L-lactic acid poduction (hollow symbols) detemined in tials co-inoculated with yeasts and bacteia ( o ), inoculated with bacteia post AF ( ) and not inoculated with bacteia ( l m ). Continuous and dotted aows indicate the bacteia inoculation times befoe and afte AF, espectively. Values ae means ± standad deviation (bas) of thee independent tials. cot melot 2 Figue 2 : Length of MLF in diffeent vaietals and diffeent vintages with co-inoculation with seleted wine bacteia melot 3 Expeiments done in collaboation with the Univesity Catolica in Chile investigated the inteest of co-inoculation techniues in high ph musts with highe alcohol potentials. In the 25 vintage in Chile using 8 diffeent lots of high ph (3.5 to 3.9) gapes (Caménèe, Syah, Melot, Cabenet Sauvignon and Petit Vedot) with alcohol potential between 4 to 5 % vol. Co-inoculation of the bacteia 24 hous afte yeast inoculation was compaed to the contol wines without bacteia inoculation. Again the total length of malolactic fementation was significantly educed by co-inoculation of bacteia, clealy indicating the dominance of the selected wine bacteia. Total duation of MLF fom the time of inoculation was about two times faste in the co-inoculated tanks compaed with the spontaneous contols. The eduction in MLF time and eliability of completion is an impotant advantage since it will educe significantly the necessity to heat the cella, which would be necessay when using seuential inoculation since it would happen late in the season, and conseuently, the cella (and wines) would need to be wamed up to stat up the MLF. Anothe advantage is the possibility to have the wines ealie stabilized, which means that they ae commecially eady faste compaed to, fo example if seuential o spontaneous MLF was used 2. Sensoy impact It has been obseved that wines that have undegone simultaneous AF/MLF tend to be less buttey and ae fuitie (Henick-Kling 993; Batowsky et al., 22; Jussie et al., 26; Kiege 26; Massea et al., 29, Batowsky et al., 2). In a study done by Knoll et al., (22), it was shown that in Riesling wines with seuential MLF had the lowest concentations of acetate estes and ethyl estes, most notably due to lowe concentations of acetic acid phenylethyleste, acetic acid 3-methylbutyleste, butyic acid ethyleste, lactic acid ethyleste and succinic acid diethyleste. This might potentially esult in deceased fuitiness sensation in wines with seuential MLF. The wines with the co-inoculation, on the othe hand, had the highest concentation of fuity ethyl estes. In addition, changes in the este concentations wee also affected by the bacteial stain used. O. oeni Lalvin VP4 seemed to poduce highe concentations of vaious fuity estes, such as popionic acid ethyleste, butyic acid ethy-leste o lactic acid ethyleste, associated with fuitiness, milky notes and mouthfeel, espectively (Figue 4).
THE RESULTS (a) Concentation [μg/l] (b) Concentation [μg/l] 25 2 5 5 25 2 5 5 Figue 4: Aveage concentation of total ethyl este and total acetate estes in (a) wines inoculated with VP4 and (b) PN4 at diffeent inoculation time fo the MLF. Co-inoculation of selected yeast and MLB also has impotant stylistic implication in tems of diacetyl poduction. Ou studies have shown that co-inoculation often esults in moe fuit-diven wine styles as opposed to lactic, buttey, nutty styles that esult when MLF stats upon completion of alcoholic fementation (seuential inoculation). Fo example, Figue 5 shows diacetyl concentations in a 2 Chadonnay fom Val de Loie (Fance). The selected bacteia Beta poduces significantly less diacetyl in coinoculation (48h) than in ealy inoculation (2/3 AF) o seuential inoculation (post AF). The impact of the ML stain on diacetyl poduction is not as stong in co-inoculation since the wines will show epeatedly low level of diacetyl with this techniue, no matte which wine bacteia is used. Diacetyl (mg L-) Wine w/o MLF Wine w/o MLF.6.4.2.8.6.4.2 total ethyl estes total acetate estes 24h 4% AF 6% AF End AF total ethyl estes total acetate estes 24h 4% AF 6% AF End AF Beta Beta 2/3d AF Beta post AF co-inoculation (48 h) Figue 5: Diacetyl concentation in a 2 Chadonnay (Val de Loie) with diffeent timing of inoculation fo MLF with Beta. 3. Management of undesiable compounds and undesiable indigeneous floa In co-inoculation stategies, it was found that significantly less biogenic amines and no histamine and tyamine wee poduced compaed to inoculation afte the end of alcoholic fementation in a tial done in collaboation with Stellenbosch Univesity (du Toit et al., 27, van de Meve et al., 26) (Figue 6). The low concentation of putescine and cadaveine also found in the wines with coinoculation oiginate fom the gape must. Biogenic amines mg/l 25 2 5 5 O.oeni co-inoculation O.oeni 2 co-inoculation O.oeni seuential O.oeni 2 seuential Histamine Tyamine Putescine Cadaveine Figue 6: Biogenic amine levels in a 26 Cabenet Sauvignon (South Afica) femented with yeast Lalvin ICV D254 compaing co-inoculation with ML state cultue vesus seuential inoculation. Selected wine bacteia have always been sceened duing the selection pocedues using genetic techniues to assue that the genes coding fo the enzymes histidine decaboxylase o onithine decaboxylase, which ae esponsible fo the fomation of biogenic amines, ae not expessed. It was assumed that fo inoculation post alcoholic fementation (seuential), the spontaneous bacteia floa was esponsible fo the poduction of highe biogenic amine levels analyzed in these teatments. The implantation contol using PCR (RAPD) with pimes M3 techniues confimed ou findings. In high ph conditions we could achieve % implantation fo the co-inoculated bacteia and wine yeast teatments, wheeas in all the selected bacteia post alcohol fementation teatments othe bacteia DNA pofiles wee also found. Co-inoculation can also be a a useful tool to pevent fomation of the unwanted volatile phenols 4-ethylphenol and 4-ethylguaiacol. The elimination, o dastic eduction, of these compounds esults in supeio wines. If the timing of alcoholic and malolactic fementation is good, malolactic fementation can be achieved immediately afte alcoholic fementation and if a gap between the end of AF and
THE RESULTS the stat of MLF is avoided, Bettanomyces can also be avoided, because the wine is stabilized ealie. Co-inoculation can become an efficient tool to pevent Bettanomyces development, and conseuently, moe winemakes ae using this techniue to fight against this contamination. Figue 7 shows the esults of a Cabenet Fanc tial fom Fance, whee the inoculation with MLB dastically educed the population of Bettanomyces as well as the levels of volatile phenols in the wines. Co-inoculation will not only help contol Bettanomyces, but it will also limit the development of othe undesiable species such as Pediococcus and Lactobacillus, especially in wines with ph highe than 3.5. The use of co-inoculation allows fo an ealie wine stabilization which pevents the development of contaminants and esults in cleane and moe aomatic wines. 8 7 6 5 4 3 2 Co-inoculated wine Non-inoculated contol wine 66 35 5 Bettanomyces population afte AF (UFC/mL) 68 Volatile phenols afte MLF (µg/l) Figue 7: 26 Cabenet fanc: Analysis of Bettanomyces contamination and volatile phenols. A QUICK SUMMARY The pactice of co-inoculation is becoming moe popula. In Fance and Spain fo example, close to 5% of MLF is now done via co-inoculation. The advantages ae numeous, such as ensuing a faste moe secue pocess and educing time fo the MLF. Co-inoculation is an impotant modulato in sensoy development, and it helps limit the development of spoilage micooganisms and thus limits off flavo compound poductions. Fo example, a wine bacteia like the Enofem Beta can poduce highe levels of diacetyl duing seuential inoculation. Co-inoculation on the othe hand, will educe the poduction of diacetyl and conseuently einfoces the fuity chaacte of white wines. Timing of inoculation, inteaction with yeast, the pesence of pecusos that pomote the poduction of aomatic molecules, ph and tempeatue conditions ae all citeia that modulate aomatic expession in wines. Choosing a wine bacteia has become a paamete to take into consideation fo developing a specific wine pofile. Ou next topic: Acetaldehyde management duing winemaking Lallemand Austalia Pty Ltd Jason Amos 23-25 Eudina Ave Edwadstown, 539, SA Ph : +6 8 8276 2 jamos@lallemand.com www.lallemandwine.com