Quorum sensing paper retracted when new study suggests compounds weren’t what they seemed

plos oneThe authors of a paper on quorum sensing — in simple terms, how bacteria “talk” to one another — have retracted it after another group’s findings led them to discover that the mixture they used weren’t what they thought.

The refreshingly detailed retraction notice in PLOS ONE explains:

The authors request the retraction of the manuscript, “Stereochemical Insignificance Discovered in Acinetobacter baumannii Quorum Sensing,” as the main conclusion reporting a case of stereochemical insignificance in A. baumannii is not supported by more recent data generated by our and another group.

Subsequent to our publication, another laboratory reported the synthesis and biological evaluation of the acylhomoserine lactones (AHLs), N-((R)-3-hydroxydodecanoyl)-L-homoserine lactone (3a) and N-((S)-3-hydroxydodecanoyl)-L-homoserine lactone (3b), described in our manuscript (see Stacy, D. M.; Welsh, M. A.; Rather, P. N.; Blackwell, H. E. Attenuation of Quorum Sensing in the Pathogen Acinetobacter baumannii Using Non-native N-Acyl Homoserine Lactones. ACS Chem. Biol. 2012, 7, 1719-1728.). In our case, these AHLs were synthesized using CBS reduction methodology followed by HPLC purification to yield what we believed to be the pure diastereomers. During our biological characterization of these compounds, we found the AHLs to exhibit very similar autoinducer activities in Acinetobacter baumannii. In the report in ACS Chem. Biol., the authors prepared these same AHLs via Noyori reduction and in their hands, the 3b diastereomer exhibited a 40-fold decrease in autoinducer activity; thus, we sought to repeat our biological analyses to confirm our results. The diastereomeric AHLs prepared for our publication no longer existed in the laboratory and when we attempted to repeat the reported synthetic and purification methods, pure compounds could not be obtained, instead a mixture of 3a and 3b co-eluted as confirmed using chiral HPLC. Pure 3-hydroxy AHLs were obtained using the Noyori reduction conditions as reported by Blackwell and colleagues. To examine the effect of mixtures of 3a and 3b on autoinducer activity in Acinetobacter baumannii, we measured the EC50 values of pure AHLs prepared from the Noyori reduction methodology and varying mixtures of these diastereomers (1:1-1:9). Upon examination, little difference was found to exist between the EC50 values of pure 3a and a 1:1 mixture of 3a and 3b (0.173 mM and 0.498 mM, respectively; 2.9-fold difference). Moreover, our reported values of 0.67 mM and 0.82 mM for 3a and 3b, respectively, align with a ~1:1-1:2 mixture of the compounds.

In the light of the results obtained, it is likely that the preparations employed for the work described in the PLOS ONE article included a mixture of the 3a and 3b diastereomers; the novel data do not support the conclusion that stereochemistry is not significant for the signaling activity of the two compounds.

The attached figure reports the results of our further analyses showing how the stereochemistry and the mixture of the two stereoisomers affect the biological activity.

No competing interests declared.

Citation: Garner AL, Kim SK, Zhu J, Struss AK, Watkins R, et al. (2012) Stereochemical Insignificance Discovered in Acinetobacter baumannii Quorum Sensing. PLoS ONE: http://dx.doi.org/10.1371/annotation/1345f9b0-016e-4123-9e72-6ccc4fa17ba2

Kudos to the authors for coming forward about the problem, and correcting the scientific record.

The paper has been cited twice, according to Thomson Scientific’s Web of Knowledge. Of note, the last author of the paper, Kim Janda, had an unrelated retraction last year for duplication.

2 thoughts on “Quorum sensing paper retracted when new study suggests compounds weren’t what they seemed”

    1. Larry

      This is bugging me.

      So, there was a conference abstract and afterwards the work was published. Some do it the other way around, they wrangle the publication out just in time for when the conference begins.

      I accept what the authors say in terms of mixing up mixtures – but that certainly doesn’t give me a good feeling regarding the simple labelling of mixtures in a laboratory – how difficult can it be? It should also match the laboratory notebook records so can be double checked – either it is different or it is not.

      Having different results from another laboatory is not surprising – it happens all the time – I suspect in this case there may have been one sloppy worker who got results they liked and they went with it. A good explanation would be to say just that if that’s what happened but I expect they saved the embarrassment of the sloppy one.

      But, it is very nice to see them come clean, even if no apology was forthcoming.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.