Which antibiotics will kill bacteria but allow yeast to grow?

December 20, 2010 2:58 AM

Is there a broad spectrum antibiotic that will kill / slow down most bacteria, but will not affect yeast at all?

I am looking for a way to select for yeast in soil samples. I have found that most of my cultures are taken over by bacteria instead of yeast, even if I make sure that yeast is present (such as by adding yeast to the sample). This antibiotic must not affect the yeast at all; for example, yeast can survive chloramphenicol but they grow much, much more slowly in the presence of it.

Does such a thing exist? Or any ideas on where I should search? The problem is finding out if a given antibiotic affects yeast or not. The literature (that I have found) is not helpful.

posted by Peter Petridish to Science & Nature (13 answers total)

Generally speaking, most antibiotics shouldn't affect fungal growth.

That said, chloramphenicol is problematic in that it has such a broad spectrum of activity that it can be considered toxic to non-bacteria. It is rarely prescribed for humans for this reason. Indeed, wikipedia mentions an study where it is used to treat a fungal infection in frogs. Which probably explains why yeast don't like it either.

I would probably stick to a beta-lactam antibiotic like ampicillin or penicillin. They are actually fungal in origin, and so shouldn't affect the growth of yeast.

Its probably going to be a case of trying several from the beta-lactam family to see what works.
posted by TheOtherGuy at 3:35 AM on December 20, 2010

Gah. Here is the link the to the frog fungus treatment.
posted by TheOtherGuy at 3:37 AM on December 20, 2010

I think a lot of the protein synthesis inhibitors (tetracycline, chloramp, pretty much all the ones that attack 30S/50S subunits...) might cause problems for fungi, even though you'd not think they'd be able to bind to the eukaryotic ribosome. Same with agents that attack DNA gyrase / topoisomerase IV. I reckon TheOtherGuy's beta-lactam idea is good.
posted by nicolas léonard sadi carnot at 3:56 AM on December 20, 2010

Good answers so far. You can also consider the concentration you're using, and mix a bunch of cheap antibiotics to increase the spectrum of bacteria inhibition. Combining a beta-lactam with a reasonable dose of an aminoglycoside (e.g. 50 mg/L gentamicin) and a tetracycline (probably get the most effect out of ~5 mg/L doxycycline) will inhibit virtually all bacteria but at least in the case of common model yeasts used in labs, won't give you any trouble.
posted by rxrfrx at 4:11 AM on December 20, 2010

This paper (PDF) separated Pediococcus and Saccharomyces by centrifugation. You could conceivably re-suspend the yeast pellet and run the procedure again to better wash away the bacteria.
posted by exogenous at 4:54 AM on December 20, 2010

Response by poster: Aminoglycosides make me a bit nervous, as my yeast is susceptible to G418. Streptomycin does attack fungi, but (from what I can find out), would there be any problems using gentamicin?
posted by Peter Petridish at 5:26 AM on December 20, 2010

gentamicin affects protein synthesis, along the lines as nicolas suggested.

It is also an aminoglycoside. which makes you 'nervous'

Is there any reason why you can't use a beta-lactam?
posted by TheOtherGuy at 5:54 AM on December 20, 2010

Maybe a low-concentration mix of a polymyxin (colistin or something) and daptomycin? They both act against the bacterial cell membrane: I bet polymixin won't hurt yeasts, since it's quite specific against the lipopolysaccharides in gram negative bacteria; I'm not as sure on daptomycin, since it has a more complex mode of action, but you do need something to attack gram positives, which it does admirably.
posted by nicolas léonard sadi carnot at 6:23 AM on December 20, 2010

Yeast grows on amp plates just fine. It wasn't fun to make a big batch of plates after someone had made a loaf of bread earlier that morning, and then come back to see them all contaminated with yeast in the fridge and having to remake them this time without the yeasty person.
posted by koolkat at 7:33 AM on December 20, 2010

I'm pulling this one straight out of my butt, but hops are used for this purpose in beer. I don't know if that is helpful at all, but it is true.
posted by Aizkolari at 10:13 AM on December 20, 2010

Can we ask questions about your experimental design? Is there a reason you can't run a side-by-side positive control, such that you can account for any inhibition of yeast growth by the antibiotic? After all, if you're counting on the antibiotic not affecting the yeast growth, you're going to need such a control if you plan to publish.
posted by endless_forms at 10:51 AM on December 20, 2010

Response by poster: I am looking for the presence of yeast in environmental samples using culturing techniques, which for my environments are more effective than molecular techniques. So, it is not a question of how much the antibiotics are reducing yeast growth. Instead, its a matter of if I am able to detect rare yeast. So, I can publish given that I define the sensitivity and repeatability the method.

I think a good approach is to use more than one antibiotic, so probably a beta-lactam and another antibiotic. Who knows which microbes will be in soil, which has me leaning towards an antibiotic cocktail.
posted by Peter Petridish at 2:13 PM on December 20, 2010

as a recovery control, to define whether you can still get rare yeast out of a treated soil, you should include 'spiked' samples, where you add yeast at low numbers to identical soil samples, treat with antibiotic or not (replicates of each) and then count the actual fungi cfus and species. you probably need to do a range of 'spikes' from relatively high numbers, to below the numbers you expect to see for 'rare yeast.'
posted by Tandem Affinity at 5:14 PM on December 30, 2010 [1 favorite]