Many microbiologists begin their workday by opening up the incubator shaker and taking out a culture flask; they then swirl it around to examine the progression of cell growth. They can examine the turbidity of the broth to tell if cells are reproducing as planned. Opening up the lid and sniffing the culture will tell seasoned microbiologists if they’re growing what they think they’re growing, because many bacterial species produce signature aromas. But what microbiologists cannot see are minor contaminants lurking within: a nightmare for controlled studies. 

Shrestha et al. recently flushed a contaminant from their culture flask. They were working with a particular strain of Geobacter sulfurreducens, strain DL1. Their main interest was in using experimental evolution to improve this strain’s natural ability to conduct an electrical charge. In their experimental setup, G. sulfurreducens DL1 would take electrons from acetate and donate them to a graphite stick anode with a potential of -400 mV, producing electrical current. They inoculated the device and let it incubate for five months, with regular additions of fresh medium. The outcome? At the end of the experiment, the “evolved” strain was now better at producing electrical currents, and was better able to form biofilms on the graphite anode.  

The next step was sequencing the genome of the evolved strain, referred to as KN400. The researchers discovered that the genome of the KN400 evolved strain was drastically different from that of the inoculated strain DL1. Over 27,000 single nucleotides differed between the two strains’ genomes, and KN400 had over 100 genes not found in DL1. Given these genomic differences, the researchers estimated that these strains must have diverged at least 1,000 years ago, and that such extensive divergence was unlikely to occur during the five months of incubation with a graphite anode.

At first they were stumped. The DL1 cultures they were working with had been meticulously carried though the microbiological ritual of single colony re-streaking for strain purity. They also used the same culturing techniques for DNA isolation and complete genome sequencing, from which they found no sign of the contaminating KN400 (even with an average of 80-fold coverage of each nucleotide of DL1’s genome). Where did KN400 come from? At this point it may have been easy to blame one of the overworked graduate students or postdocs who performed the experiments. Luckily, these researchers investigated further.

It turns out that KN400 was always lurking in their cultures, at a very low level. By amplifying and sequencing millions of copies of one gene, they were able to detect KN400 in the DL1 inoculum at about 1/1x105 copies. This drastically low abundance was quite stable, retained through single colony re-streaking. Inoculating the two strains at equal densities always resulted in KN400 decreasing to this frequency, and the only way to lure KN400 out of hiding was to incubate the mixed culture in the experimental evolution conditions of -400 mV, an environment where they later learned DL1 died off within two transfers.  Eventually the researchers were able to obtain a pure culture of DL1, by repeatedly diluting liquid cultures.

The implications for their research are not that profound. They still have a strain that can conduct electricity very efficiently, and now they have rid their DL1 culture of its contaminant. But what about the implications for microbiology? Low-level contamination is probably much more common that we think. And this study has shown that whole genome sequencing doesn’t always uncover those contaminants lurking within. The important question is whether these contaminants contribute to the phenotypic properties of “pure” cultures. The answer to this question has far-reaching implications for microbiology, past and present.