Were the DIY CRISPR kits contaminated?
Dear all,
We would like to give you an update on the situation regarding the possible contamination of the DIY Bacterial Gene Engineering CRIPSR kit distributed by The Odin online shop. We believe it is an issue that should be openly discussed as it can potentially affect the DIY biology community as a whole.
In March 2017, the authorities for health and food safety (LGL) of Bavaria, in Germany, issued a press release revealing their results of an investigation on the DIY Bacterial Gene Engineering CRIPSR kit supplied by The Odin, a company selling biotechnology reagents over the web. The LGL stated that the bacterial strain supplied in the kit was contaminated with potentially pathogenic species, such as Enterobacter species and Klebsiella pneumoniae. Based on these findings, the kit was subsequently banned from being used in Germany.
The press release came shortly after Rüdiger Trojok organized an event entitled "CRISPR kitchen". LGL representatives requested the organizers from the Institute for Technology Assessment of Karlsruhe (KIT) to hand out a vial of the bacteria obtained from an Odin CRISPR kit, there on display. At the time, the LGL did not specify that they had already found a contamination in another sample.
After the press release, some people involved in activities of the DIYbio community of Europe felt that a proactive attitude was needed to address the issue in the most transparent manner and organized a volunteer committee.
It has been difficult to gather information from both the LGL and the Odin. The LGL press release did not contain any detailed information and we had only access to a leaked report containing a summary of their results. We managed to get some additional information about the methodology used from LGL, but we could not access any of the raw data. LGL researchers told us they would not release the data until their investigation is completed, and will only publish their results in a peer-reviewed journal at a later point.
The Odin produced blast mapping of the E.coli 16S sequence to their bacterial stock and showing 99% identity. They also showed that the strain was negative for lactose fermentation. Both these results were in contradiction with the findings described in the LGL report.
It was decided that gathering more information was necessary to evaluate the situation independently and start an open discussion on how we can promote safe and lawful use of these and other kits offered in and to the DIYbio community. The goal was to evaluate whether contamination was indeed a systematic problem or an isolated accident.
We gathered 3 samples independently from LGL and analyzed them. Two of these samples came from the incriminated CRISPR kit and the last one was an Escherichia coli HME63 strain identical to that included in the kits, all of them being sold by The Odin and shipped to European customers. To obtain high-quality DNA sequencing data, we secured a collaboration to sequence the samples at the European Molecular Biology Laboratory (EMBL) Genomics Core Facility in Heidelberg, Germany. We discussed with the authorities in Berlin (LAGESO, the LGL equivalent) how to safely handle the samples and shipped them to EMBL for analysis.
EMBL Genomics Core Facility collaborators were able to revive only one of the three samples, which showed bacterial growth both in liquid culture and on agar plate. From the 16 colonies grown on the agar plate, 10 colonies were chosen for library preparation and DNA sequencing on an Illumina platform (MiSeq). The sequencing reads have been uploaded at ENA database under the following ID PRJEB23486. We encourage everyone to go and analyze the data if interested. The Core Facility is analyzing the data, but we wanted to share with you some preliminary results:
The E.coli MG1655 strain is the closest strain relative to the E.coli HME63 supplied with the kit and whose complete genome is available. Therefore, we first aligned the sequencing reads to the E.coli MG1655 genome, but noticed that only 30% of the reads are mapping, which wouldn't be expected if the colonies contained E.coli HME63.
We then proceeded to assemble the unmapped reads in large sequences (>50 kb) and checked to which bacteria species they were matching. The analysis returned various Enterobacter species (aerogenes, cloacae and hormaechei), as well as Escherichia coli and Salmonella enterica. However, given the relative phylogenic proximity of these species, with this approach it was not possible to pin down the exact bacteria in the colonies. Nevertheless, it gave an indication of the possible bacteria identity.
Given the recurrence of Enterobacter species in the results, we aligned the reads this time to the genome of Enterobacter hormaechei subsp. Steigerwaltii and we could now see that 83% of the reads were mapping and corresponded to an expected good mapping rate on a specific genome. Notably, the presence of Enterobacter species was also mentioned in the LGL report.
Moreover, as The Odin made available to us 16S sequence of the E.coli bacteria stock in their possession, we sought to determine if that sequence could be found in our reads. Unfortunately, the reads encompassing the 16S region showed again a much better match to Enterobacter hormaechei rather than E.coli.
This suggests that the bacteria in the colonies are most probably an Enterobacter species and not E.coli HME63. The 30% of reads mapping to E.coli are probably the result of cross mapping reads between E.coli and Enterobacter genomes.
We would like to mention a few words of caution. First, the methodology we used to obtain these results (whole genome sequencing) is different from the one employed by LGL (metabolic and phenotypic analyses). Second, the sample we analyzed was also different from theirs and a single occurrence could again be anecdotic. Third, we cannot guarantee that the samples were stored appropriately at all times before we collected them and ran the sequencing analysis. However, the fact still remains: both our results and the LGL investigation seem to confirm a contamination of the E. coli strain provided with the kit, possibly with the same bacterial species.
Even though the European Centre for Disease Prevention and Control (ECDC) has pointed out that the risk of infection using such contaminated kit is low, we believe it is not acceptable to have laypeople, often not skilled in laboratory techniques, as it may happen in DIYbio, unknowingly handling potentially harmful biological materials. We believe it is our responsibility to act responsibly, something that was explicitly stated in the DIYbio code of conduct and thus learn from this incident and inform the whole community about it. A failure to handle biomaterials appropriately is detrimental to the goals of the community to democratize the knowledge and technology of life sciences.
Authorities in the EU have to follow generally stricter rules than other parts of the world, including the US. Genetic engineering is particularly heavily regulated and only very few and specific educational experiments can be done outside of authorized laboratories. However, we want to stress here that our findings are not discussing the meaningfulness of European GMO laws. The problem we are concerned with is the shipping of living biological samples, which could contain unwanted bacteria or even potentially pathogenic ones due to insufficient quality assurance procedures. This problem not only unnecessarily endangers the customers, it also calls into question the ability of the DIYbio kit manufacturers.
We believe the value of educational kits needs to be stressed out and we want to support their safe distribution and use. We would like to start a discussion involving DIYbio practitioners, as well as the kit producers, academics, the authorities and the public, to collaboratively develop guidelines, or if necessary develop appropriate and efficient methods and procedures to guarantee the safety of the kits and their users.
We believe this should be a community effort in the spirit of the values of the DIYbio code of conduct and so this is also a call to all of you to get involved. If you have any questions regarding the analysis, please engage in the open discussion here on the google DIYbio mailing list.
Finally we would like to acknowledge the help provided by the EMBL Genomics Core Facility team, Vladimir Benes, Jonathan Landry and Anja Telzerow. Without their work and supportive attitude towards the community this work would have not been possible.
Best Regards,
Rüdiger Trojok
Marco R Cosenza
Julian Chollet
Luc Henry
