It Comes With The Ter­roir

by Roberto

There is some­thing very spe­cial about wine, at least for my taste. I do, how­ever, ad­mit to the pos­si­bility that many of our read­ers will won­der what's the big deal about the bev­er­age, it's just fer­mented grape juice af­ter all. But whether you ab­solutely love the thought of en­joy­ing that trea­sured bot­tle that you've been ag­ing in your cel­lar for years or you sim­ply can't palate the stuff, there's no deny­ing that wine has been an im­por­tant com­po­nent of hu­man cul­ture for mil­len­nia. We've dis­cussed the fact that some of the ear­li­est hu­man set­tle­ments show ev­i­dence of wine pro­duction here in STC. And it should have been no sur­prise that we chose to re­play John Ingraham's eno­log­i­cal clas­sic on the dis­cov­ery of bac­te­ria re­spon­si­ble for mal­o­lac­tic fer­men­ta­tion in wine a cou­ple of weeks ago as the end-of-year cel­e­bra­tions come nearer. Fol­low­ing in that spirit, I write to you to­day about how, per­haps not sur­pris­ingly, the metage­nomics craze has pen­e­trated the cul­ture of wine. But first a lit­tle per­sonal back­ground.

I do love wine. Very much. But my knowl­edge of wine is rudi­men­tary at best, at least com­pared to that of some of my clos­est friends. I know that Caber­net Sauvi­gnons are red and Sauvi­gnon Blancs are white. But I have been know to fail to iden­tify the wine when blind­folded and served a Pinot Noir and a Pinot Blanc side by side and at the same tem­per­a­ture. Just so you know. When it comes to some of my fa­vorite wines, Mal­becs from Men­doza (Ar­gentina) are the first to come to mind. And there is no doubt in my mind that Mal­becs from Ca­hors (France) do not even come close. Apo­­logies to those of you who love the French Mal­becs, though you prob­a­bly dis­like the ones from Ar­gentina. But are those dis­par­i­ties I per­ceive in the wines from those two re­gions due to dif­fer­ences in the lo­ca­tion where the grapes were grown or are they due to the ways the grapes are grown and the wines pro­duced? From my per­spec­tive, it's all about how the viti­cul­tur­ist cares for the vines and how the enol­o­gist pro­ceeds in mak­ing the wine. "No, no, no!" scream the wine cognoscenti, "viti­cul­tur­ists and enol­o­gists aside, it's all about the ter­roir!" Ter­roir?

Fig­ure 1. Ter­roir is said to come mostly from cli­mate and soil. Source

What on Earth is  ter­roir? The pun is in­tended as earth is terra in Latin from where Eng­lish gets such words as ter­rain and ter­ri­tory. But there is no sim­ple Eng­lish trans­lation for the French word ter­roir. Ter­roir em­com­passes many fac­tors of any given lo­ca­tion that af­fect grapes and there­fore the wine. Mostly it's about the cli­mate and the soil, but ba­si­cally it's any­thing that pro­duc­ers feel is be­yond hu­man con­trol. Ter­roir is "big time" in the wine in­dustry as it is used to des­ig­nate wine ap­pel­la­tions, the le­gal de­f­i­n­i­tions of spe­cific wine grow­ing re­gions. Thus, the char­ac­ter­is­tics of dif­fer­ent ter­roirs are thrown around among wine pro­fes­sion­als and afi­ciona­dos alike. Im­por­tantly, con­nois­seurs are said to be able to taste the ter­roir in a wine and thus iden­tify a wine's ori­gin. Lucky for them! From all the wine I've drunk I've yet to taste ter­roir! But still, I find it ex­tremely in­trigu­ing. Is there some­thing real about the lo­cal ter­roir that makes its way into the fi­nal prod­uct to yield re­gion­ally dis­tinct wine charac­ter­is­tics? Peo­ple have been ar­gu­ing about this for a long time with­out reach­ing agree­ment. And, as of lately, mi­crobes have come into the pic­ture.

That mi­crobes would be part of ter­roir makes sense, doesn't it? Af­ter all, mi­crobes are part of eve­ry en­vi­ron­ment and their ac­tiv­i­ties help shape their en­vi­ron­ment. The first inklings in­di­cat­ing that this is in­deed is the case came from a won­der­ful pa­per by Nicholas Bokulich and col­leagues from the group of David Mills at UC Davis, a place with a long tra­di­tion of out­stand­ing aca­d­e­mic eno­lo­gy. The au­thors ap­plied in full force a cul­ture-in­de­pen­dent ap­proach to as­sess the mi­cro­bial com­munity com­po­si­tion of di­verse grape musts (destemmed crushed grapes al­ready at the win­ery, ready for fer­men­ta­tion to be­gin). Their ap­proach it­self was not novel: high-through­put short am­pli­con se­quenc­ing fol­lowed by the req­ui­site con­stel­la­tion of sta­tis­ti­cal analy­ses that are the bread and but­ter of metage­nomics. What was novel was the sub­ject mat­ter — they car­ried out their study us­ing nearly 300 sam­ples of grape musts of dif­fer­ent grape va­ri­eties from four of the ma­jor grape grow­ing re­gions of Cal­i­for­nia of the 2010 and 2012 vin­tages. Aside from the long list of "who's there" (un­avoid­able in these sorts of stud­ies) their re­sults hinted at the fact that re­gional, site-spe­­cific, and grape va­ri­ety fac­tors shaped the fun­gal and bac­te­r­ial con­sor­tia present in grape musts. They were also able to cor­re­late the mi­cro­bial com­mu­nity com­po­si­tions with spe­cific cli­matic feat­ures.

Fig­ure 2. Grape must bac­te­r­ial com­mu­ni­ties de­mon­strate dis­tinct re­gional pat­terns. Weighted UniFrac dis­tance den­dro­gram com­par­ing bac­te­r­ial com­mu­ni­ties of Chardon­nay musts from across Cal­i­for­nia. Branches are col­ored by the grow­ing re­gions they rep­re­sent, white branches en­com­pass two or more re­gions. Pie charts rep­re­sent av­er­age phy­lum-level tax­o­nomic com­po­si­tions of all sam­ples from each site. Adapted from source

These re­sults led them to con­clude that vine­yard envi­ronments and mi­cro­bial com­mu­nity are linked and that this non-ran­dom dis­tri­b­u­tion sug­gests the ex­is­tence of a "mi­cro­bial ter­roir" that could un­der­lie the re­gional varia­tion among wine grapes. Quite ex­cit­ing that mi­crobes might be in part con­tribut­ing to ter­roir! The au­thors go on to spec­u­late on the im­pli­ca­tions of their re­sults: "These ex­ploratory find­ings pose a par­a­digm shift in our under­stan­d­ing of food and agri­cul­tural sys­tems be­yond grape and wine pro­duc­tion, wherein mi­cro­bial com­mu­ni­ties play ac­tive roles in prod­uct qual­ity char­ac­ter­is­tics." I agree on the pos­si­ble im­pli­ca­tions of their re­sults but "par­a­digm shift" sug­gests a change in world­view akin to the Coper­ni­can rev­o­lu­tion, which their re­sults are not go­ing to gen­er­ate — ex­cit­ing as they might be. Per­haps they should re-read Thomas Kuhn. Quib­bling aside, I do be­lieve this is a land­mark pa­per in the field of grape metage­nomics.

Bokulich and col­leagues fol­lowed that ini­tial re­port with a per­haps even more strik­ing pa­per. In their se­quel, they showed a cor­re­la­tion be­tween the com­mu­nity com­po­si­tion of grape must with the metabolome in fin­ished wines. They were even able to pre­dict metabo­lite abun­dances in fin­ished wine based on the mi­cro­bial com­mu­nity com­po­si­tion of the start­ing prod­uct. As is the case for most metage­nomic stud­ies, their com­bined re­sults pro­vide cor­re­la­tions, but causal­ity — whether mi­cro­bial ac­tiv­i­ties them­selves are re­spon­si­ble in part for as­pects of the ter­roir — is a dif­fer­ent mat­ter. As the au­thors say in their clos­ing re­marks: "We iden­tify nu­mer­ous as­so­ci­a­tions be­tween the wine mi­cro­biome and metabolome, and fu­ture stud­ies are ne­cessary to es­tab­lish causative links be­tween the mi­cro­bial con­sor­tia, wine metabo­lites, and sen­sory char­ac­ter­is­tics."

In those two pa­pers, the au­thors de­cide to use grape must as a proxy for the grapes them­selves. The jus­ti­fi­ca­tion for this choice is that grape must "rep­re­sents the junc­tion be­tween grape sur­face and fer­men­ta­tion mi­cro­biota, when all grape-de­rived mi­cro­bial in­puts are present but in­oc­u­la­tion and the on­set of fer­men­ta­tion has not yet al­tered this com­mu­nity." Un­der­stood. But still, I've al­ways reck­oned that a lot hap­pens from the mo­ment the grapes are har­vested to when the must is pro­duced. From per­sonal ex­pe­ri­ence I know that af­ter har­vest­ing, the grapes be­gin to get crushed un­der their own weight. In ad­di­tion, their their tem­per­a­ture in­creases dur­ing trans­porta­tion which is of­ten sev­eral hours. Then the grapes are sorted, destemmed and crushed at the win­ery, which is en­dowed with its own mi­cro­biome as well. All of these facts are likely to lead to changes in the mi­cro­bial com­mu­ni­ties present. No doubt that car­ry­ing out mi­cro­bial com­mu­nity com­po­si­tion analy­ses on the grapes them­selves, im­me­di­ately af­ter they are cut from the vine, would be a worth­while study. That is why I was so pleased to read a pa­per, by Va­le­rio Mez­za­salma and collea­gues from Mas­simo Labra's group at the Uni­ver­sity of Mi­lan, look­ing di­rectly at what's on the grape and how those com­mu­ni­ties evolve as the grapes are turned into must and then into wine.

Fig­ure 3. Rel­a­tive abun­dance of bac­te­r­ial or­ders (A) and fungi fam­i­lies (B) re­cov­ered in G, iM, and eM sam­ples at each Sar­din­ian sam­pling lo­cal­ity. Source

This pa­per de­scribes a study based in the Ital­ian is­land of Sar­dinia. (Wish I had been in­vited to be part of the study!) The au­thors chose the lo­ca­tion based, in part, on the fact that the con­di­tions there are such that the lo­cal grape va­ri­ety, Can­nonau, is grown in dif­fer­ent ar­eas of the is­land, each with its own ter­roir. (For you wine afi­ciona­dos, Can­nonau is known as Gar­nacha or Grenache else­where.) They chose four ar­eas Al­gero (ALG), Mores (MOR), San­tadi (SAN) and Mamoiada (MAM) and sam­pled grapes (G), ini­tial must (iM) and end must (eM). To con­trol for win­ery-to-win­ery vari­a­tion, they took all grapes col­lected and processed them at a sin­gle win­ery. Their re­sults cor­rob­o­rated my sense that changes in com­mu­nity com­po­si­tion can and do oc­cur be­tween grapes on the vine and ini­tial must. The taxa abun­dance chart shown in the fig­ure brings this point home. No­tice how the pat­terns dif­fer be­tween the G and iM sam­ples for all four lo­ca­tions for both bac­te­r­ial and fun­gal taxa. The au­thors went one step fur­ther, which yielded an­other strik­ing re­sult from their work. When they looked at the bac­te­r­ial taxa found on the grapes, they dis­cov­ered that each lo­ca­tion had its own pri­vate taxon sig­na­ture! Now, if there is a pos­si­bil­ity for unique con­tri­bu­tions to fla­vor and aroma de­vel­op­ment as the grapes ma­ture and then are turned into wine, it might come from those taxa that are unique to each re­gion. That would be a won­der­ful gain in our un­der­stand­ing of the mi­cro­bial con­tri­bu­tion to ter­roir! But alas, these au­thors did not tell us any­thing about the metabolome of the fin­ished wine. Not even a de­scrip­tion of how the wine tasted, that is, could they still de­tect each in­di­vid­ual ter­roir? I afraid we'll have to await the re­sults of fu­ture stud­ies. Nonethe­less, it is nice to see that we are mak­ing great progress in our un­der­stand­ing of the sub­tler roles that mi­crobes play in the mak­ing of that bev­er­age that has ac­com­pa­nied hu­man­ity for so much of its his­tory. A votre santé!