Brettanomyces
Brettanomyces is a common defect in wine, but controversy surrounds the subject. Jamie Goode sifts the available data and canvasses the views of world-famous winemakers to determine how widespread the problem is, and whether brett can ever be a good thing. (Originally published in Harpers Wine and Spirit Weekly, 18 April 2003, p 42-46)
When Brian Fletcher, chief winemaker at Calatrasi in Sicily, found
out I was writing this feature, he couriered me a bottle of red
wine. Labelled simply as ‘Brettanomyces’, it was a sample from
Puglia that Brian had recently been sent by a producer there. So I
opened it and poured a glass. Immediately, I got a whiff of animal
sheds with some savoury, cheesy character. The palate was similarly
animal-like, with a thin metallic edge. Very rustic. Not
undrinkable, but getting there, and a textbook example of a
Brettanomyces-infected wine.
For those scratching their head wondering what on earth
Brettanomyces is, let me explain. It’s a yeast – that is a
unicellular type of fungus, not a bacterium – that is a common
spoilage organism in winemaking. The goal of this article is to
assess how much of a problem it is, what its effects are and how it
can be prevented. Finally, I’ll look at the controversial issue of
whether low levels of ‘brett’, as it is widely known, can ever be a
good thing, adding complexity to certain sorts of wines.
First, a dull but necessary paragraph to clear up a potential
confusion. The name Dekkera is often used interchangeably with
Brettanomyces. They are actually the same genus (this is the
taxonomic group just above ‘species’), with Dekkera being used for
the ascospore-forming (sporogenous) form of this yeast, and
Brettanomyces used for the non-spore forming type. There are
currently five recognized species of Brettanomyces/Dekkera: B.
nanus, B. bruxellensis, B. custersianus and B. naardenensis, with a
range of synonyms in common use. Of these, research indicates that
B. bruxellensis is the most relevant to wine.
The microbiology of wine production is a complex business, and it is
beyond the scope of this feature to go into too much detail. But let
me try to give you a feel for the concepts involved. Think of the
plants growing on the slopes up the side of a mountain. At the
bottom there are hundreds of different types, with the pattern of
vegetation changing and progressively decreasing in diversity with
altitude (and a corresponding drop in temperature). It’s a bit like
that with fermenting wine, except that here the variation is
temporal and not spatial -- it is a gradually changing environment
for yeasts. In freshly crushed grape must there are many different
yeast species present, including those normally found on grapes.
These rapidly disappear as fermentation starts and alcohol rises.
The environment becomes more and more inhospitable, and after a
while the only significant yeast species present is Saccharomyces
cerevisiae. As alcoholic fermentation finishes, the S. cerevisiae
population decreases significantly. If by this stage the sugar and
nutrient supplies are exhausted, that’s the end of things and the
wine is stable. But if they aren’t, this leaves the way open for
spoilage bugs to develop; brett is one of the worst culprits here.
What does bretty wine taste and smell like?
Volatile phenols and fatty acids are the key molecules responsible
for the olfactory defects in wines affected by brettanomyces.
According to Peter Godden, of the Australian Wine Research
Institute, ‘The anecdotal dogma in this area is that 4-ethyl-phenol,
isovaleric acid and 4-ethyl-guiacol are the key molecules, in order
of sensory importance’. But he adds that he has seen variations in
brett character in different bottles of the same wine. 4-ethy-phenol
is the most prominent molecule in bretty wines, giving aromas of
stables, barnyards and sweaty saddles (apparently, but I must admit
to never having smelled one). Its presence in wine is an almost
certain indicator of a brett infection, and this is what most
diagnostic labs test for to indicate the presence of brett.
4-ethyl-guiacol is a little more appealing, known for its smoky,
spicy aromas. Isovaleric acid, a volatile fatty acid, is known for
its rancid, horsey aroma, and as yet there is no analytical
technique that picks it out: in gas chromatography/mass spectrometry
(GCMS) another compound elutes at the same time, which masks it.
Godden emphasizes that this is a complex area of study: ‘There is
not much of a relationship between overall brett character and
4-ethyl-phenol levels, and there are synergistic effects between the
three most important sensory compounds.’
As with other volatile odorants, people differ widely in their
sensitivity to these molecules, and each individual shows a range of
different thresholds (for example, the threshold for detecting an
odorant differs from the threshold for recognition of the same
odorant). Godden suggests that a useful sensory threshold to use for
4-ethyl-phenol is 420 micrograms/litre. At this concentration and
beyond, a wine will typically be noticeably bretty. Below this
concentration, the character of the wine may be changed but people
won’t, on average, recognize that this is due to 4-ethyl-phenol.
Because the threshold for 4-ethyl-phenol drops when 4-ethyl guiacol
is also present -- and in brett-infected wine they always occur
together in a ratio of about 10:1 -- this threshold is calculated
for a 10:1 mixture of 4-ethyl-phenol and 4-ethyl-guiacol.
How common is brett?
The short answer is that brett is highly prevalent, and represents
an increasing problem, even in new world countries such as
Australia. ‘We first started raising this as an issue four years
ago’, says Peter Godden of the Australian Wine Research Institute, ‘
and on 1st July we’re planning to start a major project looking at
Brettanomyces’. As a scientist, he feels that for such an important
issue, this is a relatively under-researched area. ‘There is a lot
of conjecture: anecdotal observations are very important but we have
to be careful with them because they can skew people’s opinions.’
Although brett can and does occur with whites, it is predominantly a
red wine problem. This is because red wines are far higher in
polyphenol content, and generally have a higher pH, both factors
which encourage brett development for reasons which outlined below.
With rising standards of winemaking worldwide, I was a little
surprised to hear that brett is on the increase. There seem to be
two contributing factors to this rise. First, there is the current
trend for ‘natural’ wines. ‘Minimalist winemaking is a perfect
recipe for bretty wine’, says Godden. ‘It’s probable that the
increase in brett in the 1990s can be traced back to the winemaking
fad to stop adding sulphur at crushing’. Indeed, the most effective
way of preventing brett is to maintain an adequate concentration of
free sulphur dioxide (SO2). Randall Grahm of California’s Bonny Doon
comments, ‘If one is ideologically committed to no sulfitage at the
crusher, this increases one’s chances of brett dramatically.
Likewise, if one uses low or no SO2 in the elevage of the wines,
this greatly increases the risk of brett’. Preliminary studies by
the AWRI show that there is a lot of genetic variability among
Brettanomyces strains. This makes the correct use of sulphur even
more important. If it is added in small, regular doses, winemakers
might unintentionally be selecting for SO2-resistant strains of
brettanomyces, or to put it another way super-brett strains that are
then even harder to eliminate. So timing and magnitude of SO2
additions are important as well as the actual concentrations: the
best way to get rid of brett seems to be large SO2 additions at
strategic intervals.
Second, there is the move towards ‘international’ styles of red
wine, made in an extracted style from super-ripe grapes. ‘These are
higher in pH and are richer in polyphenols’, explains Grahm. pH is
important, likely through its role in modulating the effectiveness
of SO2 additions. The higher the pH, the less effective SO2 is and
the more likely that brettanomyces will grow. Polyphenol content is
important because these compounds are the precursors for the
volatile phenols largely responsible for bretty odours.
A vital risk factor is the presence of residual sugars and nitrogen
sources left over at the end of fermentation. With the gradual rise
in alcohol levels over the last 20 years, the last bit of sugar
commonly isn’t being metabolised by the yeast. Godden suggests that
one solution is to try to keep the wines warm while they are being
pressed. As well as sugar, a nitrogen source is needed for brett to
grow. In fermenting wine, S. cerevisiae uses amino acids as a
nitrogen source. A recent winemaking trend has been to add
diammonium phosphate (DAP) as a supplementary nitrogen source for
yeasts, to reduce the risk of stuck fermentations. However, fewer
than half of musts need actually use this additive, and DAP has been
described as ‘junk food’ for yeasts – they’ll use this in preference
to amino acids, leaving them in the wine as a nitrogen source that
encourages the growth of brett.
Old barrels are frequently touted as the main culprits of brett, but
Randall Grahm adds, ‘The received wisdom about old barrels, old
foudres being the great repository of brett I think is somewhat
mythical and simplistic: dirty barrels, dirty wines, q.e.d.’ Grahm
adds that, ‘Since brett is largely ubiquitous, a rampant brett
infection is often more of a function of a large inoculum coming in
on the grapes.
To gauge the extent of the current brett problem, Godden and his
colleagues recently completed a survey of Cabernet Sauvignon wines
in five major regions of Australia. He’s unable to give the actual
results, because these are sensitive, and he thinks that the samples
size, around 170 bottles, isn’t big enough to allow him to draw a
firm conclusion. ‘But if a consumer were to go out and buy a mixed
dozen,’ he told me, ‘several bottles would have more than 425
micrograms/litre 4-ethyl-phenol: if you drink wine regularly, you’ll
have come across a lot of brett.’
Before the 1990s, brett was common in Bordeaux. The wines of several
well known classed growths were well known for their distinctive
‘stink’. This was almost certainly because of brett infections, but
without the data – and most properties would understandably be
reluctant to own up to this – I can’t name any names. Since the
early 1990s, however, brett has become much rarer, and this is
mainly due to the groundbreaking work of Dr Pascal Chatonnet. In
1993 Chatonnet carried out a survey of 100 French wines, and showed
that a staggering third of those tested had levels of volatile
phenols above the perception threshold.
The conclusion seems to be that brettanomyces is widespread, and
virtually every barrel of red wine has the potential to go bretty.
Create the right environment for it, and you’ll have a brett
infection. Thus the key objective for winemakers isn’t to create a
sterile winery, which will never happen, but to make sure that their
barrels aren’t a receptive environment for brett to grow in.
Brett, Mourvèdre or terroir? A case study
Brettanomyces is a favoured discussion topic among wine geeks,
who’ll often enter into lengthy discussions about whether a certain
wine is bretty or not. One wine that keeps cropping up in this
context is Château de Beaucastel, the highly regarded Châteauneuf du
Pape estate. To some, the distinctive earthy, slightly animal-like
characteristics of many past vintages of Beaucastel have reflected
an expression of terroir, or even the higher than average Mourvèdre
content of this wine. Others think it’s because of brett infection.
Who is right?
Back in early 1998, Charles Collins, an American wine collector,
became so frustrated with the endless wine geek discussions about
Beaucastel and brett that he decided to find out for himself. He got
hold of some scientific papers on the subject and read up about the
subject. ‘I realised that the presence of the compound
4-ethyl-phenol is a virtually certain indicator of the presence of a
brett infection’, recalls Collins. He contacted a lab who does
testing for 4-ethyl-phenol and sent them some Beaucastel from his
cellar. ‘I opted to test two of the most famous vintages, the 1989
and 1990’, Collins told me. ‘These wines are supposed to represent
what great Beaucastel is all about.’ He prepared the samples for
shipment in sterilized glass 375 ml bottles and used fresh corks to
seal them. The wines were labelled so that the lab had no clue as to
their identity.
The results? According to Collins, ‘they showed indisputable
evidence that significant brett infections occurred in both the 1989
and 1990 vintages of Beaucastel.’ Microscan and plating tests showed
only small amounts of mostly dead brett cells, but the
4-ethyl-phenol levels were 897 micrograms/litre for the 1989 and a
whopping 3330 micrograms/litre for the 1990. Collins concludes, ‘if
you personally like the smell of brett, then none of this should you
dissuade you from buying and cellaring Beaucastel. You should,
however, give up the myth that the odd flavours are due to
terroir—they aren’t.’ I would add that while I’ve detected what I’ve
always assumed, in the absence of data, to be high levels of brett
in some vintages of Beaucastel—the 1991 springs to mind as one of
the brettiest wines I’ve ever encountered—in vintages since the
mid-1990s I haven’t encountered any. But, of course, unlike Collins,
I haven’t done the lab tests that would be needed to verify this.
‘We believe in natural winegrowing and winemaking, and I must admit
that this has led us to have serious debates with scientists
spanning three generations’, responds Beaucastel’s Marc Perrin. ‘In
the mid-1950s, for instance, our grandfather, Jacques Perrin,
decided to stop using chemical pesticides or herbicides on the
vineyard. At that time, when scientists were recommending the use of
such chemicals for productivity or lobby reasons, that seemed crazy
and impossible. Now, it seems that people have changed their mind
and more and more vineyards are turning organic. I could quote many
more examples of opposition between a scientific vision of wine and
our traditional/terroir oriented philosophy of wine, and the subject
of Brettanomyces is just one more’, he explained. ‘There are
certainly some Brettanomyces in every natural wine, because
Brettanomyces is not a spoilage yeast (as many people think) but one
of the yeasts that exist in winemaking. Some grapes, like Mourvèdre,
are richer in 4-ethyl-phenol 'precursors' than others and we have a
high percentage of these grapes in our vineyard. Of course, you can
kill all natural yeasts, then use industrial yeast to start the
fermentation, saturate the wine with SO2 and then strongly filtrate
your wine. There will then be no remaining yeasts, but also no taste
and no typicity. That is the difference between natural wine and
industrial wine, between craftsmanship and mass-market product.’
Adding complexity?
Beaucastel has been widely acknowledged as one of the world’s great
wines over recent decades. Yet from Collins’ limited sampling
coupled with individual tasters’ experiences, it seems likely that
some of the most successful past vintages of this wines have been
marked by high levels of brett. This leads us to a critical—and
fascinating—question: is brett ever a good thing? In small
quantities, can it have a positive influence on certain styles of
red wines?
If surveys such as those of Chatonnet and Godden are to be
extrapolated across all wines, it is likely that many wines with
above-threshold levels of brett have received critical acclaim and
have been enjoyed by countless consumers. This leads to the
conclusion that while most people won’t enjoy a really stinky wine,
low levels of brett might not be a problem—indeed, a bit of brett
might even add complexity to certain robust styles of wines.
Bob Cartwright, senior winemaker of Leeuwin Estate in Western
Australia’s Margaret River region, acknowledges that ‘a lot of
winemakers like to have some as a complexing character—the question
is how much is too much?’. Randall Grahm is undecided. ‘I suppose
this could theoretically add some complexity to a wine. The problem
is that for now, this is not easily controllable’.
Pascal Chatonnet is opposed. He sees the problem of brett as a lack
of fruit and loss of typicity. ‘If brett is able to grow in all the
red wines of the planet—and this is the case—then all the wines will
have the same odour, which is a pity’.
Godden is another who isn’t keen on the idea. ‘My view is that if we
could eliminate it altogether we would’, but he stressed that he
wouldn't go so far as to say it is always negative. Godden cites
some results from the PhD thesis of Phil Spillman, now winemaker
with Villa Maria in New Zealand. In one study Spillman did some
sensory analyses. The strongest relationship he found was an inverse
correlation between levels of 4-ethyl-phenol and wine preferences.
‘I’ve not been able to find an Aussie winemaker who doesn’t find 100
micrograms/litre negative’, adds Godden. ‘In tests where brett
character has been added, it has a severe adverse effect on the
palate. 4-ethyl-guiacol can be interesting and complexing and
doesn’t have the negative palate effect of 4-ethyl-phenol, but with
brett infection you get 10 times as much 4-ethyl-phenol than
4-ethyl-guiacol.’
Randall Grahm has an novel suggestion, though: ‘It would be very
interesting if we could isolate a strain of brett that worked in
wine, depleting nutrients but producing very low levels of 4-ethyl
phenol. In this way, one could inoculate one’s wine with brett, much
the same way as one inoculates one’s wine with malolactic bacteria,
thus depleting nutrients and rendering the wine safe from further microbial
degradation.’ Now there’s a project for the microbiologists. Any takers?