The group has now published a paper in PeerJ following their investigation into what went wrong. Ronald tells us the new paper, titled “The Xanthomonas Ax21 protein is processed by the general secretory system and is secreted in association with outer membrane vesicles,”
…provides experimental details on some of the experiments we carried out during the 18 month process of investigating the function of Ax21 and also describes additional experiments we conducted to decipher the function of this protein.
We also asked Ronald about criticisms — voiced in comments here at Retraction Watch and elsewhere — that in a narrative about the retractions, she had not cited two papers that had allegedly found her group’s work irreproducible. Those two papers — in PNAS (Danna et al) and the Plant Journal (Mueller et al) — were not cited in the PeerJ paper either. She explains:
The Mueller et al paper challenges the Danna et al paper – not the retracted papers from my lab. My two rice/Xanthomonas papers that were eventually retracted were not challenged by any group. We discovered the mistakes ourselves when reproducing our own experiments. The experiments described in the Danna et al paper were carried out in Arabidopsis in Fred Ausubel’s lab.
We did not cite the Danna et al and Mueller et al papers in our PeerJ publication because the experiments are unrelated. Yes, they are both in the field of plant pathology, and yes, the Danna et al experiments were initiated because of our Science publication, but that is where the similarity ends.
In the Peer J paper, we examined the function of the Ax21 protein in the bacterial pathogen Xanthomonas oryzae pv. oryzae, which infects rice. The Danna et al paper studied effects of “Ax21-derived” peptides on immune responses of Arabidopsis. The Danna et al paper uses a different plant species, different assays, and different pathogen species. Also it, addresses a different question. Danna et al investigated whether Ax21 or “Ax21-derived (also called A1 peptide)” peptides could trigger Arabidopsis FLs2-mediated resistance to Pseudomonas. Because rice Xa21 and Arabidopsis FLs2 are both stuctural similarities, Danna wanted to test if any of the 40+ receptor kinases in Arabidopsis also could recognize peptides related to Ax21.
The only overlap with our retracted Science paper is that Ausubel’s group examined a set of alanine scanning mutants that my lab had tested in rice (see Fig. S4A of the the retracted science paper). In Danna’s experiments, Ax21 did not trigger immunity Arabidopsis; only the A1 peptide was active in their experiments.
Mueller et al proposed that flg22 contamination in the A1 peptide preps could explain the results of Danna. This was a question that our team considered in the original publication. To address this question, Danna et al resynthesized the A1 peptide and carried out mass spectrometry analysis and flg22 dose–response experiments, which did not reveal any obvious contamination. However, contamination is still a possibility. To resolve the question, the Mueller et al and the Danna et al labs will need to carry out the same experiments with the same reagents.
For more details, please see: “FLS2-Mediated Responses to Ax21-Derived Peptides: Response to the Mueller et al. Commentary” in Plant Cell. [This commentary includes Danna, Ronald, and Ausubel as authors.]
In the PeerJ paper we discuss the more relevant publications- e.g. the paper from an Irish lab that reproduced the results of our retracted PLOS ONE paper.