These ARE The Yeast You Are Looking For

A few weeks ago I attempted to isolate wild Brettanomyces from the soil, based on some papers which identified Brettanomyces as part of the rhizosphere (microorganisms associated with plant roots) of several plant species. This first attempt simply plated water which I had used to rinse the roots of bean plants onto agar selective for Brettanomyces. This attempt did not work, and instead I grew some pretty – but utterly useless for fermentation – filamentous fungi.

I thought this would be it for the year – it was already well into fall, and most plants in our garden were on their way to dormancy or death. But, thanks to some late season warm weather, I was able to make a second attempt at it. This time using an enrichment approach prior to plating.

So did it work?

This is a monster post, so here’s a quick TOC for anyone who wants to jump around:

• Enriching for Wild Brettanomyces – my strategy for this experiment
• The Enrichment in Practice – the details of the experiment itself
• Results – what I found
• Overview of Isolated Cultures – a summary of the 12 pure cultures I generated
• Conclusions – what I found
• Limitations – caveat emperor
• What’s next – yes, there will be a Part III, IV and V to this series

Other Parts to this Series:

• Part I: These Are Not The Yeast You Are Looking For
• Part II: These ARE The Yeast Your Are Looking For (this post)
• Part III: Wild Brettanomyces – a Needle in a Haystack
• Part IV: Fermentation characteristics

Enriching for Wild Brettanomyces

My plan was to “grow-out” any wild Brettanomyces that were present, while also suppressing the growth of unwanted yeast, fungi, and bacteria. Microbiologically speaking, this is a selective approach that relies on several properties to enrich for Brettanomyces while suppressing the unwanted organisms.

Growth Medium: I chose to use YPD medium instead of the WLN medium I used last time. This is a simpler medium that is nutritionally rich and supports the growth of Brettanomyces. It’s also much cheaper than WLN, and is much easier to prepare, which were the real reasons for the switch.

Inhibiting Bacteria: I added the antibiotics ampicillin and chloramphenicol, both at 50 mg/L, to act as a broad-spectrum anti-bacterial agent.

Inhibiting Most Fungi and Yeast: I used the same “special sauce” as last time – 50 mg/L cycloheximide* – to inhibit the growth of most yeast and fungi.

*Cycloheximide is highly toxic. Please do not try to replicate this project unless you know how to work with it safely.

Selecting for wild Brettanomyces: This is where my plan differs the most from last time. To help enrich for Brettanomyces I took advantage of two properties of Brett which non-yeast fungi lack. The first property is a relatively strong ability to grow in liquid culture compared to filamentous fungi. As a general rule (meaning there are exceptions), yeast will dominate in a liquid culture during early time-points, with filamentous fungi becoming dominant later on. To take advantage of this, I transfered a small amount of the cultures started from plant roots into fresh medium after 7 days (a process called “sub-culturing”). This inoculated the fresh medium at a time when Brettanomyces should be at or near peak growth, but before filamentous fungi become dominant. In theory, this should give Brettanomyces a good chance of dominating the sub-cultures.

The second property is the one that should strongly select for wild Brettanomyces – the ability to grow in the absence of oxygen. I created an anaerobic (oxygen-free) environment by layering mineral oil over the liquid culture. All filamentous fungi are obligate aerobes, meaning they cannot survive in the absence of oxygen, while Brettanomyces has the ability to ferment and grow in the absence of oxygen. Combined with the cycloheximide, there is perhaps two or three genera of yeast which will grow under these anaerobic conditions…including Brettanomyces.

Detecting Brettanomyces: I used the same approach as last time to detect Brettanomyces. Two drops (~100 uL) of the cultures were plated on solid medium (this time YPD-agar) containing the same antibiotics and cycloheximide as the liquid medium used above, plus 22 mg/L of bromocresol green. The antibiotics + cycloheximide suppress bacteria and most yeast/fungi, while the bromocresol green allows for the identification of Brettanomyces. Yeasts will take up this dye, but cannot process it, leading to blue/green colonies. Brettanomyces can break down the dye, producing white colonies. While this does not provide a definitive identification, it is a good enough of an identification to justify the cost and time that it takes to perform a full work-up of the putative Brettanomyces.

From a microbiological perspective, these agar plates are termed “selective and differential media” (hence the meme in the header image). The medium is selective as it allows the preferential growth of a limited range of organisms, and it is differential as the colour-change of the medium allows you to differentiate between certain species.

The Enrichment In Practice

I’ve depicted the enrichment process in the diagram on the left. I collected a mixture of roots from the last of the living plants in my garden – dill, purple beans, chives, parsley, and a bit of crab grass. I then used the chopped roots to inoculate two tubes of highly oxygenated medium.

48 hours later I mixed the contents of the two tubes to re-oxygenate and equalize the microbial contents. I then transferred a portion of the medium back to the emptied tube such that both tubes had the same volume. I placed 1 mL of mineral oil on top of one cultures to deprive it of oxygen (anaerobic culture). The other tube was left open to allow for continued air exchange (aerobic culture).

Five days later (1 week after the initial inoculation) I inspected both cultures in my microscope, looking for yeast-like (and preferably Brettanomyces-like) cells. None were observed in the aerobic culture, but a few were found in the anaerobic culture. As such, I plated a small amount of the anaerobic culture.

At this point I sub-cultured both the aerobic and anaerobic cultures, even though the former lacked any obvious Brettanomyces-like cells. The rationale for sub-culturing the aerobic culture was that Brettanomyces may still be present, but at too small a number to be identified by microscopy. I sub-cultured the aerobic culture by transferring 1 mL of this culture into two tubes of freshly made YPD + antibiotics + cycloheximide (e.g. 0.5 mL into each of the new tubes). I immediately topped one of the new tubes with mineral oil, and left the other exposed to air. I then did the same with the anaerobic culture. This gave me a total of four sub-cultures: anaerobic and aerobic sub-cultures of the initial aerobic culture, and anaerobic and aerobic sub-cultures of the initial anaerobic culture.

For simplicity I’ll refer to the sub-cultures as “initial-subculture”, so the aerobic sub-culture of the initial anaerobic culture would be called “anaerobic-aerobic”.

I allowed the sub-cultures to grow for another week. At this point I inspected them by microscopy and plated all four sub-cultures.

Results

Initial cultures

To setup the initial cultures I harvested roots from a variety of plants which were still green and living from my garden. This included beans, dill, chives, parsley, and crab grass. The roots were finely chopped, mixed, and divided evenly between two well-oxygenated tubes of media. Two days later there was visible evidence of growth, in the form of a slight haze to the medium (see the third image from the left, above). At this point I mixed the contents of the two tubes together, in order to even out any microbial differences between the tubes, and then separated the mixture back into two separate tubes. I layered oil over one of these cultures to create an anaerobic environment and allowed the tubes to grow for an additional 5 days.

On the final day placed a single drop of the aerobic and anaerobic cultures onto a slide and imaged the suspensions on my microscope. My intent here was two-fold. Firstly, I was looking for the presence of yeast-like cells. Secondly, if yeast-like cells were present I wanted to get a rough count of how many were present. The reason for the latter was to allow me to dilute the sample properly in order to plate it out on my selective and differential medium.

The aerobic culture (image to the left, above) had a thin layer of mould on the top of the tube, and unsurprisingly, under the microscope I found large numbers of highly elongated and vacuolated cells typical of moulds. Surprisingly, I also found a large number of bacteria. While some antibiotic resistant bacteria are to be expected, the numbers present were much higher than is reasonable. I suspect that there was some metabolism of the antibiotics by the moulds, thus allowing the bacteria to grow. Despite a lot of time spent searching, I found little evidence of yeast. Because of this I didn’t bother plating this sample.

My anaerobic culture (image to the right, above) was, at first glance, a little dismaying. Unlike the aerobic culture, there was very little growth visible in the sample, with most of the objects in the sample appearing to be dust and other debris. However, after a lot of searching, I was able to find several clumps of cells that had the prototypical appearance of yeast cells. Several of these also had interlinked cells in a growth pattern similar to that of Brettanomyces. Because Brettanomyces-like cells were present – though-be-it at a low density – I plated out a few drops (~200 ul) of the undiluted culture onto an agar plate.

The plate was incubated at room temperature (currently around 19C) for 4 days, at which point hundreds of colonies were visible on the plate (image to left). Of these, none were green, consistent with the expected decolourization of the bromocresol green by Brettanomyces. That said, this wasn’t a total success as the unexpectedly high density of colonies essentially prevents me from picking pure colonies from the plate. As such, I will only pick colonies from the sub-cultures, which may limit the diversity of Brettanomyces species I can recover.

I also prepared sub-cultures of both the aerobic and anaerobic cultures at the same time that I was inspecting the cultures on the microscope. To prepare these cultures, I transfered 0.5 mL of the initial culture to a fresh tube containing 5 mL of medium + antibiotics + cycloheximide. This was then incubated at room temperature until visible growth had occurred.

Sub-cultures

The subcultures grew much more quickly than I had expected, with all four sub-cultures showing a notable pellet of yeast on the bottom of the tube after just three days (image to the left). So contrary to my enrichment plan, I ended up imaging and plating out these cultures after 3 days, instead of the planned seven.

As with the initial cultures, I first inspected the sub-cultures on my microscope (slideshow, below). The aerobic sub-cultures were not usable – as expected, the aerobic-aerobic sub-culture had a lot of filamentous fungi plus some bacteria, while the aerobic-anaerobic had more bacteria than yeast. The fungi and bacteria far out umbered the yeast in both cultures, making recovery of pure yeast cultures difficult. This may also indicate an issue with the plate I prepared from the original anaerobic culture – clearly, the bacteria had survived the anaerobic culture (how else would they have emerged in the aerobic sub-culture?). This may mean that many of the colonies on the agar plate prepared from the original anaerobic culture may be bacteria rather than yeast. As such, I decided to not sample from that plate.

The anaerobic sub-cultures were much more promising (images in the slide show, above) – both showed a range of yeast-like cells of varying morphologies, with no obvious bacterial contamination. The yeast cells also had a range of morphologies. This may represent multiple strains/species of wild Brettanomyces, or may indicate one (or a few) strains of Brettanomyces at different stages of growth. There is also the possibility that these may be non-Brettanomyces yeast. The dose of cycloheximide I used (50 mg/L) is on the edge of what both Debaryomyces housenii (also known as Candida famata) and Kluyveromyces marxianus can survive. While the morphology of the cells in the anaerobic cultures are not consistent with K. marxianus, D. housenii is morphologically similar to Brettanomyces. I had based my 50 mg/L on what is standard for Brettanomyces isolation from wine and fruits. Debaryomyces do grow poorly at this concentration – one paper (which I cannot find again), claimed that only a few Debaromyces strains can grow over 30 mg/L, and none at 100 mg/L cycloheximide.

Given the high density of yeast in these cultures, I first used a serial dilution to prepare 1:10,000, 1:100,000 and 1:1,000,000 dilutions of the original cultures. If you don’t know how to do a serial dilution, this is covered in my video on preparing wet mounts for microscopy. I then plated 2 drops (~0.1 mL) of these highly diluted cultures onto YPD + antibiotics + cycloheximide plates. The plates were incubated at room temperature until colonies had formed – which only took 2 days!

Much to my surprise, all of the aerobic plates had fungal colonies – even the anaerobic-aerobic plates. These appeared as star-like or “spikey” colonies (see image above). I confirmed these were filamentous fungi by microscopy (not shown). The fungal colonies also have a slight blue-grey hue, suggesting that they are taking up the bromocresol green.

Scattered among the fungal colonies were beautiful yeast colonies. Even better, only yeast colonies were present on the anaerobic-anaerobic plates! Even more exciting, with only one exception (out of over 500 colonies), these yeast colonies were stark-white. While references and information is thin, at least one paper suggests that D. housenii (called C. famata in the paper) will have blue/green pigmented colonies on bromocresol green agar, and as mentioned above, I have yet to see a cell with the appearance of K. marxianus. This means that these white colonies have a good chance of being wild Brettanomyces!

I picked 4 colonies from each of the 3 dilution plates from the anaerobic-subculture of the initial anaerobic culture, placing each colony on a grid plate, as shown in this old video of mine. I also cultured the only blue yeast colony I found, as I was curious what it was. These were allowed to grow for 2.5 days, producing large colonies. Once larger colonies has formed, cell morphology was assessed by microscopy. These findings are summarized below.

Overview of Isolated Cultures

I isolated a total of 12 pure colonies for further analysis. All but one colony grew to the desired large size in just 2.5 days. While I was focused on putative Brettanomyces isolates, I also included the only blue yeast colony that I identified. This blue colony can be seen in position A4 in the image to the left.

Much to my surprize, colony C4 also turned blue despite coming from a white “mother colony”. This colony also struggled to grow, relative to the others, suggesting that it was more sensitive to cycloheximide and therefore not likely a species of Brettanomyces. To further the identification process, I imaged all 12 colonies as wet-mounts, imaging all colonies at the maximum possible magnification to resolve the cell structure as much as possible. I focused this imaging on finding dividing cells, as dividing cells are the most useful for species identification.

Colony	Image	Notes & Putative ID
A-1		Putative ID: Brettanomyces Confidence: Medium Notes: Cells show classical Brettanomyces budding pattern and elongated cell morphology.
A-2		Putative ID: Unknown Confidence: Low Notes: Cells are apiculate, which is unusual for Brettanomyces.
A-3		Putative ID: Brettanomyces Confidence: Low Notes: Occasional cells show classical Brettanomyces budding pattern and elongated cell morphology.
A-4		Putative ID: Schizosaccharomyces Confidence: Low Notes: This is one of the blue colonies, meaning it is highly unlikely to be Brettanomyces. In addition, the cells appear to divide by binary fission, which is not something done by Brettanomyces.
B-1		Putative ID: Brettanomyces Confidence: Medium Notes: Cells show classical Brettanomyces budding pattern.
B-2		Putative ID: Brettanomyces Confidence: Low Notes: Cell morphology is consistent with Brettanomyces but dividing cells were not observed and some cells appear apiculate.
B-3		Putative ID: Unknown Confidence: Low Notes: Cells are apiculate, which is unusual for Brettanomyces.
B-4		Putative ID: Brettanomyces Confidence: Medium Notes: Cells show classical Brettanomyces budding pattern and elongated cell morphology.
C-1		Putative ID: Brettanomyces Confidence: Medium Notes: Cells show classical Brettanomyces budding pattern and elongated cell morphology.
C-2		Putative ID: Brettanomyces Confidence: Low Notes: Cells show classical Brettanomyces budding pattern but lack an elongated cell morphology.
C-3		Putative ID: Unknown Confidence: Low Notes: Cells are apiculate, which is unusual for Brettanomyces.
C-4		Putative ID: Brettanomyces Confidence: Low Notes: Cells show classical Brettanomyces budding pattern and elongated cell morphology. However, the colony also accumulated some blue dye, and grew poorly on cycloheximide, both of which are not typical of Brettanomyces.

Conclusions – Did I find wild Brettanomyces?

Clearly I have isolated several strains that I have putatively identified as wild Brettanomyces. This identification is based on five lines of evidence:

1. Rapid growth under anaerobic conditions, producing sufficient yeast to form a visible pellet. This indicates that the yeast are highly fermentative and capable of tolerating the presence of alcohol.
2. Growth in the presence of a high concentration of cycloheximide. This is a capability of only a few species of yeast.
3. Absence of blue colouration when grown on bromocresol green-containing medium. Brettanomyces is one of the only yeast species capable of degrading this dye.
4. Colony morphology. The cell shape (ovoid-to-irregular) and pattern of cell division (cells bud from the termini of the mother cell, not the sides) is consistent with Brettanomyces.
5. Vacuole. The cells have a large central vacuole that is easily resolved by microscopy. Again, this is consistent with Brettanomyces.

None of the above characteristics provide, on their own, a definitive identification. However, this particular combination of characteristics is unique to Brettanomyces. Even so, I only rated my confidence as a maximum of “medium”, as I cannot be certain of my identifications without additional tests. So while I am confident that several colonies are wild Brettanomyces, I do not yet have all the evidence needed for a firm identification nor for identification of specific species.

None-the-less, based on these findings I am quite confident that I have isolated wild Brettanomyces from the plant rhizosphere! There are some interesting implications for this conclusion…but you’ll have to wait for part III of this experiment before I get into those.

Limitations

There are a few limitations to how I conducted this experiment that are worth discussing:

Issue 1: Bias Introduced by Enrichment: The first of these is the consequence of using liquid cultures to enrich for Brettanomyces. My enrichment process selects for Brettanomyces by taking advantage of its ability to grow relatively quickly under fermentative (low/no-oxygen) conditions in a liquid medium. The use of anaerobic conditions also served to eliminate non-yeast fungi and non-fermentative yeast, while cycloheximide suppressed the growth of all but a small handful of yeast species. While this approach appears to have worked very well – in that it did enrich for some Brettanomyces – it would not have enriched for all Brettanomyces.

In particular, this enrichment approach was based on the biology of Brettanomyces bruxellensis, and is unlikely to work for other Brettanomyces species. For example, B. claussenii and B. anomalus are not as quick to grow under fermentative conditions as is B. bruxellensis, and therefore may have been out-competed during sub-culturing. Likewise, B. custersianus, B. naardenensis, and B. nanus – at least by some reports – grow poorly in liquid culture and under low-oxygen conditions. Thus, my enrichment process may have eliminated – rather than enriched – these species. In other words, while this enrichment approach can be used to selectively enrich for strongly fermentative species of Brettanomyces, it would not be suitable for identifying the full range of Brettanomyces species present in a sample.

In fact, it is very well possible that I started off with dozens of unique strains of Brettanomyces spread across three or four species, and ended up selecting for a single strain from one species, merely because that strain was a stronger fermenter and outgrew everything else.

Issue 2: Incomplete Selection: A second limitation is the dose of cycloheximide I used. While this dose is higher than what is used in most Brettanomyces studies, most of those studies are isolating Brettanomyces from environments such as fermented wine – e.g. environments that already enrich for Brettanomyces and which lack significant microbial diversity. After I had started this project I found papers which indicated that to truly select for Brettanomyces I needed to double the concentration of cycloheximide. As such, there is a theoretical possibility that two other species may have been isolated instead of, or alongside, Brettanomyces. The first of these – Kluyveromyces marxianus – is not thought to be a soil organism, and rather is a lactose-fermenting yeast found on the udders of animals and in milk-derived foods (e.g. cheese & kefir). It is also morphologically distinct from Brettanomyces, with none of my isolates having a morphology similar to K. marxianus.

On the other hand, Debaryomyces housenii is a very common yeast that is known to live in soil, is cycloheximide-tolerant, is fermentative, has a similar cellular morphology to Brettanomyces, and divides in a similar fashion to Brettanomyces. The dose of cycloheximide I used (50 ug/mL) is above that recommended for isolating D. housenii (10 ug/mL), but may not be lethal to D. housenii. Information on whether D. housenii can decolourize bromocresol green was hard to find, with only one paper commenting on it indirectly – and seemingly indicating that D. housenii is incapable of decolourizing bromocresol green. But given the sparsity of information on D. housenii‘s ability to declourize bromocresol green, it remains a potential (if unlikely) organism isolated in this experiment. I also had one yeast species which appeared to divide by binary fission – if true, this would be a yeast which is only distantly related to either Brettanomyces, Debaryomyces, or Kluyveromyces (all of which are in the same Saccharomycetaceae family of yeasts). If this truly is a fission yeast, than it is likely a species of Schizosaccharomyces, from the Schizosaccharomycetaceae family of yeasts.

What’s Next for the Wild Brettanomyces project?

That is it for this post, but it’s not the end of this project. I appear to have 7 potential isolates of wild Brettanomyces, plus 5 interesting non-Brettanomyces isolates that are worth deeper investigation (especially the putative fission yeast).

I am in the process of performing a genetic identification of all 12 strains, which will give a firm identification of the species that I have purified. With luck, Part III of this series describing those results will be posted in one to two weeks. Once I have some firm identifications, I plan on performing some pure and mixed fermentations with any strains confirmed to be wild Brettanomyces (or otherwise of interest to brewers), to see what their flavour profile is like. Look for that post in December or early 2023.

Other Parts to this Series:

• Part I: These Are Not The Yeast You Are Looking For
• Part II: These ARE The Yeast Your Are Looking For (this post)
• Part III: Wild Brettanomyces – a Needle in a Haystack
• Part IV: Fermentation characteristics