Comparative genomics: a must for winemaking

In 2004, archaeologists uncovered the earliest evidence for winemaking in the world when a biochemical analysis of organic residues on an ancient ceramic jar at a 9000-year-old Neolithic village in Georgia yielded a suite of compounds characteristic of red wine.

The analysis also detected traces of bacterial preservatives — the Neolithic winemakers appear to have deliberately added tree resins to their brew to extend the wine’s life after fermentation, a practice that prefigured the addition of pine resin as a preservative and flavoring agent in latter-day Greek retsina wines.

Dr Anthony Borneman, a principal research scientist at of the Australian Wine Research Institute in Adelaide, is using whole-genome techniques to study the yeasts and bacteria involved in winemaking. He is an invited speaker at this year’s Australasian Genomic Technologies Association annual conference in the Hunter Valley from 11-14 October.

He said the archeological evidence from Neolithic times attests to winemaking being man’s earliest biotechnological endeavour.

Winemaking, breadmaking and beer brewing all rely on the same friendly yeast to perform the fermentation process: Saccharomyces cerevisiae.

Although it is the major agent in wine fermentation, Dr Borneman said S. cerevisiae is by no means a lone player. The ‘magic’ of grapes spontaneously fermenting after crushing actually involves a complex microbiological progression of hundreds of species of fungi, yeasts and bacteria that proliferate and die in successive waves of microbial boom and bust.

He said S. cerevisiae is generally a latecomer to the party (its presence is almost undetectable on intact grape berries), but once its numbers increase, the increasing concentration ethanol it produces kills all but a few other select species, such that S. cerevisiae completely dominates the later stages of alcoholic fermentation. However, modern approaches to winemaking have looked to short-circuit what can be a long, drawn-out process that could go very wrong (vinegar anyone?).

“From the 1970s Australian wineries began using commercially available, purified starter yeast strains of S. cerevisiae,” Borneman said. “Yeast-supply companies sell them as freeze-dried products that are added to the must after the grapes are crushed. They provide a reliable and predictable ferment with little risk of microbiological spoilage.

“The winemaker doesn’t have to wait very long for fermentation to proceed as they bypass the normal lag-time in S. cerevisiae growth. The downside is that the complex microbiological progression of a uninoculated ferment is all but replaced by a monoculture of a single strain of S. cerevisiae.

“Given the huge efficiency advantages provided by commercial yeast, the vast majority of Australian wineries quickly started to take advantage of these strains.

“Most wineries still do. But while some strains offer to impart different flavour characteristics, many wine experts believe the inoculation monoculture yields a far less complex, somewhat one-dimensional wine. So there are now the beginnings of a backlash, where winemakers are revisiting the past, as has happened with raw-milk cheeses and sourdough bread.”

Dr Borneman said a venturesome band of Australian winemakers — especially those aiming at the top end of the market — have gone back to nature. They are experimenting with so-called ‘wild’ ferments in pursuit of novel flavours and complex mouthfeel, hoping to give their wines a marketing edge over those fermented with commercial yeast cultures.

Wild yeast strains in uninoculated ferments take time to kick off — “You’re never really sure what’s in there after you crush the grapes,” he said. “It’s a leap of faith and there’s always a risk that undesirable microbes will take hold while the numbers of wild wine yeasts are building up, ruining the wine.”

While ‘going feral’ would seem a risky business, Dr Borneman said some winemakers have become adept at wrangling wild ferments to produce predictable results.

Their enterprise provides fresh fields for inexpensive, high-throughput genomics technologies to explore, to identify the genetic sources of novel characteristics that wine drinkers will appreciate.

He said genomic studies of wild yeast and bacterial species could help the wine industry identify the genetic sources of novel characters that could enhance the distinctively Australian characters that made our wines so successful two decades ago, while taking much of the risk out of wild fermentation.

So, now the wild local yeasts’ potentially unique contributions to the fermentation process can be added to the traits that stem from a wine’s terroir — the effect of local soils, climate and other environmental factors on the composition of the grape juice.

Given the massive investment in time and money required for classical microbiological research, Dr Borneman said there has been relatively little study of the regional yeast species and strains that colonise Australian grapes.

“We’re currently involved in a collaborative sequencing project with Yalumba, Treasury Wine Estates and the Ramaciotti Centre for Genomics at UNSW, using high-throughput genomics to study wild ferments,” he said.

“We extract the total DNA from samples of the ferment and amplify and sequence a particular variable region that serves as a molecular barcode to differentiate fungal species. We also count the copy number of the various barcodes to see how much each wild yeast species contributes to the ferments.

“Think of the must (the freshly pressed grape juice, including seeds, stems and skin of the fruit) as a sort of palette that will determine the eventual character of the wine after fermentation and ageing.

“The environment, including the weather, can influence the kinds of yeast that colonise the fruit — competition from botrytis and other fungal diseases also has an influence on yeast composition and the eventual flavour of the wine. That’s where we’re heading with metagenomics.”

Metagenomics involves sampling every microbe present in the early stages of the ferment, which Dr Borneman describes as “a complex soup of different organisms”.

“As a wild ferment kicks off, S. cerevisiae starts from a very low base and eventually wipes out every other microbe as it increases the alcohol content. But early in the ferment, the lesser players are making a range of different metabolites that give the wine greater complexity.

“Conventional microbiological assays are extremely labour intensive, so you can’t afford to analyse the wild microbe community in large numbers of samples in any detail.

“But the cost of barcode-sequencing a particular fermentation sample has recently dropped well below $100 — closer to $50, in fact, and it’s still coming down.

“We’re still in the data-gathering phase, getting in more samples of wild ferments from around Australia, that will allow us to begin correlating the wild yeasts’ contributions to the distinctive characters associated with particular winemaking regions of Australia.

“We hope the study will tell us if certain winemaking regions have their own ‘special’ yeast species. As we accumulate more data, we want to provide winemakers with information that will make wild ferments more predictable and customisable. We might even be able to predict from early samples of the must how the ferment is likely to proceed, given different winemaking interventions.”

That capability would require inexpensive, real-time sequencing technology. Dr Borneman said one company, Oxford Nanopore, is already marketing an early version of a portable sequencer the size of a chocolate bar, costing around US$1000, that plugs into a laptop to upload its data for rapid analysis.

“You blink, and the technology has advanced,” he said. “Oxford Nanopore’s sequencer sucks DNA through a membrane studded with protein pores and can sequence DNA fragments on the fly.

“If it continues to develop as promised, it would be a game changer, not only for the wine industry, but for medical and agricultural applications — for example, it could allow a doctor to rapidly diagnose a bacterial infection from a blood sample, or a farmer could detect and diagnose a rust disease on his wheat crop and hit it with the right fungicide before it does serious damage.

“Once the technology matures, we’re talking about a medical-type approach to monitoring the ferment. It would give winemakers the ability to check how the yeasts and other microbiota are going, and provide early warning of undesirable microbes that could contaminate the ferment with off-flavours, so they could use appropriate interventions to suppress them.”

The “other microbes” that Dr Borneman refers to include the malign yeast species Brettanomyces (Dekkera) bruxellensis, better known to winemakers simply as “brett”, and the bacterium Oenococcus oeni.

Oenococcus is a key player in malolactic fermentation, with its ability to soften the taste of a wine by converting tart-tasting malic acid to lactic acid.

Dr Borneman has recently published research papers on his genomic investigations of strain variation in B. bruxellensis (see below) and O. oenii.

In their Oenococcus paper, published in BioMedCentral Genomics in 2012, and titled Comparative analysis of the Oeoncoccus genome reveals genetic diversity in industrially relevant pathways, Dr Borneman and his AWRI colleagues note that the lactic acid bacterium O. oeni is one of a rare few bacteria species that not only survive but actively proliferate at the high alcohol concentrations in wine ferments.

Unlike other bacteria present in wine, Oenococcus is positively beneficial to wine quality because of its ability to metabolise tart-tasting malic acid and transform it to lactic acid — hence the term “malolactic fermentation”.

He said little is known about the genetic diversity of Oenococcus in Australia. “Australia’s grape vines were brought out from Europe, which makes it hard to determine what was already here, as opposed to what wine-associated microbes hitched a ride,” he said.

Using Illumina sequencing, the AWRI team obtained whole-genome sequences from 11 strains of O. oeni, collected from several wineries and wild sources around Australia, to assess the extent of genetic variation in the species. Three other strains sequenced in an earlier study brought the sampled population to 14 strains.

The authors say O. oeni has an unusually compact genome for a free-living bacterium — just 1.8 megabases. They speculate that it has undergone genomic ‘streamlining’ during its adaptation to the challenging, narrow ecological niche of fermenting grape juice and wine.

Each individual strain’s genome consists of around 1800 genes, but the collective analysis of the isolates identified a total of 2800 open reading frames that comprise the pan genome of the species.

Of this figure, the species’ core genome — the subset of genes common to all isolates — consists of fewer than 1200 genes.

The authors say the data they have obtained on the genetic variation across the 14 O. oeni isolates is vital for harnessing the phenotypic variation present in economically important bacteria involved in fermenting grape juice to wine.

The expansion of the pan genome, according to the AWRI paper, is partly due to an accumulation of DNA sequences scattered around the genome that help the bacterium to defend itself against bacteriophage attack.

The remainder consists of a variety of genes of potential importance to industrial fermentation systems, including genes involved in cell-wall polysaccharide synthesis, sugar transport and utilisation, and amino acid biosynthesis.

Dr Borneman said there is evidence that, in addition to its primary role in converting malic acid to lactic acid, the growth of O. oeni in wine ferments has other effects on flavour, aroma and mouthfeel.

If comparative genomics can link these phenotypes back to the presence or absence of certain genes in identified strains, it should be possible to develop new pure cultures of strains that will help winemakers reliably produce a range of wine styles by malolactic fermentation.

Some wineries already carry out microbiological assays on samples of O. oeni and S. cerevisiae strains to check for the presence of desired strains in the ferment, but Dr Borneman said the procedure is laborious and time-consuming. Using genomic tools to directly sequence different strains of O. oeni or S. cerevisiae is now cheaper, faster and more accurate.

In an article for Current Opinion in Biotechnology in 2013, Dr Borneman and AWRI colleagues Dr Isak Pretorius and Dr Paul Chambers described comparative genomics as “a revolutionary tool for winemaking”.

They concluded: “With the growing accessibility and affordability of genome sequencing, we are witnessing the birth of a new era in industrial microorganism strain development; comparative genomics in industrial strains is providing a richer and deeper understanding of the genetic composition and variation of these crucial microbes.

“New genomic technologies are providing us with the means of rapid identification of genetic loci that shape industrially important traits.

“This will enable the development of new wine yeast strains that offer improved fermentation performance, and a means of tailoring wine sensory properties to meet consumer demand.”

A diagram showing the diversity of the microbial biota in a wine ferment.

A beast of a yeast

The yeast Brettanomyces (Dekkera) bruxellensis is a survivor. While almost everything else is dying around it, it weathers the storm produced by the growth of S. cerevisiae, waiting patiently for it to die due to a lack of nutrients and a buildup of its own waste (alcohol).

In this highly hostile environment of finished wine, Brettanomyces flourishes. Unfortunately for winemakers, as this yeast slowly grows by taking advantage of leftover complex nutrients, which S. cerevisiae does not have the metabolic weaponry to consume, it produces a number of volatile compounds such as ethyl-phenols that impart aromas reminiscent of a barnyard floor or even the medicinal character of band-aids. Such descriptive terms do not lend themselves to the production of fine wine, such that Brettanomyces spoilage is a real and present concern for winemakers worldwide.

In their Brettanomyces paper, Dr Borneman showed that four Australian isolates of B. bruxellensis have a core diploid genome that is sufficient for their survival, but two were triploids: their cells contain an extra haploid genome (a full set of unpaired chromosomes).

He determined that the sequences of the haploid genomes were highly divergent between the two strains.

Dr Borneman said some Saccharomyces yeast species, such as the ubiquitous S. pastorianus, which produces all of the world’s lager beer, possess a similar allotriploid genome that allows them to survive in a wider range of environmental conditions than their diploid relatives.

He suggested the triploid Brettanomyces strains arose through multiple hybridisation events between different strains or species.

Remarkably, he said, the triploid strains account for 92% of Brettanomyces isolates from across the Australian wine industry.

Their near-total replacement of diploid strains in Australian wineries is evidence of strong positive selection pressure for the triploid forms because of their ability to survive in a broader range of environmental conditions than diploid strains.

“Interestingly, this correlates with the triploid strains being significantly more resistant to sulfite, the main control measure used by winemakers against Brettanomyces spoilage,” Dr Borneman said. It suggests that Brettanomyces, much like antibiotic-resistant bacteria, is fighting back against the measures used for its control.

The study therefore provided a snapshot of the microbe in the act of evolving resistance to sulfite. Forewarned, winemakers have time to find other preservatives to suppress Brettanomyces and delay the development of sulfite resistance.

Image: WRI wine genomics expert Dr Anthony Borneman.