Are Biocontrols for Plant Diseases Safe to Humans?

Posted on by Paul Vincelli

I recently received a phone call from a citizen sounding quite frightened about an unusual bacterial infection she was experiencing.  Her physician had diagnosed the infection to be due to the bacterium, Pantoea agglomerans.  I did my best to answer her questions, primarily informing her how widespread and common this bacterium can be on plant surfaces and in soil.

As I thought about this case, the thing that most struck me was that P. agglomerans (which has been known by several scientific names) has been widely studied as a potential biocontrol agent for plant diseases.  As I dug into the scientific literature, I learned about many cases of P. agglomerans causing opportunistic infections in humans, often (though not always) associated with immunosuppression.

“Wait a minute!” I thought.  You mean, this natural biocontrol agent, which we commonly assume is nothing but beneficial, could actually cause harm to humans?  Most field plant pathologists are well-aware of the mistaken—but widely held—assumption that synthetic chemicals are harmful while natural chemicals are safe [1].  This isn’t the first time I have wondered about the safety of microbial biocontrol agents.  However, this was the first time I had ever “drilled down” into the scientific literature on this topic.

There was more there than I expected.  A number of bacterial species that can provide some biological control of plant disease have indeed been shown to be opportunistic human pathogens.  The ones I read about are listed below, with citations to some of the medical literature associating them with human infections:

  • P. agglomerans [2, 3];
  • Stenotrophomonas maltophilia [4];
  • Bacillus cereus [5];
  • Bacillus subtilis [6];
  • Lysobacter enzymogenes [7].

My literature search was not at all exhaustive.  Therefore, there could even be additional examples of biocontrol bacteria with “alter-egos” as opportunistic pathogens in humans.

There is a significant caveat.  Even though a bacterial species may be reported both in human infections and as a plant colonist, the human strains and the plant strains may not necessarily come from the same populations [8].  But I don’t take much comfort in that, for the following reasons:

  1. Relatively recent studies suggest that some of the human-infecting and plant-colonizing strains of P. agglomerans seem worryingly similar [9, 10].
  2. Even if they are from different populations, the classification of the human-infecting and plant-colonizing strains as the same species indicates that they have some phenotypic overlap.

To be honest, my discoveries concern me much more than traces of synthetic pesticides in my food.  And to think that farm workers sometimes apply biocontrol agents with no safety equipment!

So the take-away is, our concern for food safety should be in proportion to scientific risk, and not based on assumptions.  Biocontrol agents for plant disease control, sometimes assumed to be safe because they derive from Nature herself, may not be as safe as we think.  Likewise, pesticides—whether natural or synthetic—should not be assumed to be safe.

We do a good job evaluating pesticide safety in the USA, through intensive and increasingly sophisticated scientific and regulatory procedures.  Our pesticides undergo such extensive safety testing that, while I wash my produce well, I don’t concern myself with parts-per-billion—or parts-per-trillion—of pesticides in my food [11].  But my safety “radar” is up now on biocontrol agents.  Some questions worth asking:

  • What kind of safety testing do biocontrol agents undergo?
  • How good is our knowledge of exposure routes and doses?
  • Are they tested against immunocompromised mammals?
  • Are they tested for chronic effects?
  • What else do we need to know about biocontrol agents in order to assure the safety of our food supply, as we seek to reduce the use of synthetic pesticides by using alternative, “natural” materials?

 

Literature cited

  1. Ames, B. N., Profet, M. and Gold, L. S., Dietary pesticides (99% all natural). PNAS, 1990, Vol. 87, p. 7777-7781. Available from: http://www.pnas.org/content/87/19/7777.full.pdf
  2. Dutkiewicz, J., Mackiewicz, B., Lemieszek, M. K., Golec, M. and Milanowski, J., Pantoea agglomerans: a marvelous bacterium of evil and good .Part I. Deleterious effects: Dust-borne endotoxins and allergens – focus on cotton dust. Ann Agric Environ Med, 2015, Vol. 22, p. 576-88, DOI: 10.5604/12321966.1185757. Available from: http://www.ncbi.nlm.nih.gov/pubmed/26706959
  3. Cruz, A. T., Cazacu, A. C. and Allen, C. H., Pantoea agglomerans, a plant pathogen causing human disease. J Clin Microbiol, 2007, Vol. 45, p. 1989-92, DOI: 10.1128/JCM.00632-07. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17442803
  4. Brooke, J. S., Stenotrophomonas maltophilia: an emerging global opportunistic pathogen. Clin Microbiol Rev, 2012, Vol. 25, p. 2-41, DOI: 10.1128/CMR.00019-11. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22232370
  5. Bottone, E. J., Bacillus cereus, a volatile human pathogen. Clin Microbiol Rev, 2010, Vol. 23, p. 382-98, DOI: 10.1128/CMR.00073-09. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20375358
  6. Oggioni, M. R., Pozzi, G., Valensin, P. E., Galieni, P. and Bigazzi, C., Recurrent septicemia in an immunocompromised patient due to probiotic strains of Bacillus subtilis. Journal of Clinical Microbiology, 1998, Vol. 36, p. 325–326. Available from: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC124869/
  7. Dempsey, K. E., Riggio, M. P., Lennon, A., Hannah, V. E., Ramage, G., Allan, D. and Bagg, J., Identification of bacteria on the surface of clinically infected and non-infected prosthetic hip joints removed during revision arthroplasties by 16S rRNA gene sequencing and by microbiological culture. Arthritis Res Ther, 2007, Vol. 9, p. R46, DOI: 10.1186/ar2201. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17501992
  8. Bonaterra, A., Badosa, E., Rezzonico, F., Duffy, B. and Montesinos, E., Phenotypic comparison of clinical and plant-beneficial strains of Pantoea agglomerans. 2014, Vol. 17, p. 81-90, DOI: 10.2436/20.1501.01.210. Available from: http://www.ncbi.nlm.nih.gov/pubmed/26418852
  9. Volksch, B., Thon, S., Jacobsen, I. D. and Gube, M., Polyphasic study of plant- and clinic-associated Pantoea agglomerans strains reveals indistinguishable virulence potential. Infect Genet Evol, 2009, Vol. 9, p. 1381-91, DOI: 10.1016/j.meegid.2009.09.016. Available from: http://www.ncbi.nlm.nih.gov/pubmed/19800991
  10. Rezzonico, F., Smits, T. H., Montesinos, E., Frey, J. E. and Duffy, B., Genotypic comparison of Pantoea agglomerans plant and clinical strains. BMC Microbiol, 2009, Vol. 9, p. 204, DOI: 10.1186/1471-2180-9-204. Available from: http://www.ncbi.nlm.nih.gov/pubmed/19772624
  11. Winter, C. K. and Katz, J. M., Dietary exposure to pesticide residues from commodities alleged to contain the highest contamination levels. J Toxicol, 2011, Vol. 2011, DOI: 10.1155/2011/589674. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21776262

 

8 thoughts on “Are Biocontrols for Plant Diseases Safe to Humans?”

  1. Thank you for sharing this! From my own side project, some endophytic bacterial isolates from the Bacillus cereus clade express enterotoxin genes when grown in pure culture. I think that a portion of the community have fallen into the “its natural, so safe” mentality. Radon is natural, but we know of all the issues it causes. In my job as an editor, I occasionally see manuscripts screening known human pathogens for immunocompromised individuals or even species that might be mycotoxin producers screen as biological control agents! I think the first screening should be to determine species and then tossing anything known to be a potential problem for human health.

    Like

    Reply

    1. Thanks for your comment. I honestly don’t know how to design a screening program for potential biocontrol agents, but you approach seems reasonable: screen first for potential pathogens, and discard those candidates.

      Like

      Reply

    2. I am still puzzling over this topic. Figure 1 in the following paper shows that Bacillus strains may possess both desirable genetics (cry protein conferring biopesticidal properties) and one or two virulence plasmids (Genomic insights into the taxonomic status of the Bacillus cereus group, http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4571650/). This seems consistent with our shared concern over potential virulence of biocontrol strains.

      Like

      Reply
  2. Pingback: 047 Genetic Engineering, Extension and Communication | Talking Biotech Podcast

  5. This update to this post is based on consideration of additional information.

    The biggest weakness in the post relates to microbial strains. In the post, I noted that biocontrol strains may differ from human-infecting strains of the same nominal species. It is worth affirming this, as advanced genomics techniques can sometimes show substantial differences between strains that may share the same species name but which may, in fact, merit distinct species names. This is more likely to be the case if the taxonomic designation of the biocontrol agent doesn’t account for the latest scientific findings.

    I also want to acknowledge that colleagues who do research on biocontrol take seriously the possibility that a biocontrol strain might also pose a health risk. This article wasn’t written for those researchers but, rather, it was an account of my own surprise at discovering this potential risk. Certainly, other scientists not necessarily in the field of biocontrol have been surprised at the issue as I was. Members of the general public may share that surprise, so the post remains relevant.

    Like

    Reply

  6. Here is a reply from Dr. Gywn Beattie, Professor & Robert Earle Buchanan Distinguished Professor of Bacteriology, Iowa State University, posted with permission:
    “I understand that you are trying to make the point that we need to be equally vigilent in evaluating the risks of biocontrol agents as we are for chemical controls, and I agree with this. The article has an implicit assumption, however, that we are not as vigilent for biocontrol agents that are used commercially as we are for chemicals that are used commercially. Is that true?

    “Another issue that I think was omitted was the predictive information that we now have on strains because of the availability of genome sequence information. Your presentation is a little garbled because you refer to strains as being from the same or different populations, but a strain is a distinct genetic line that itself forms a population, so this terminology of strains coming from the different populations doesn’t make sense to me (and still doesn’t if we take population to mean “community”). You wrote “Even though a bacterial species may be reported both in human infections and as a plant colonist, the human strains and the plant strains may not necessarily come from the same populations [8].“ I think technically this would be more accurate if written as “…, the human strains and the plant strains are generally distinct strains, that is, they are genetically distinct.” And whereas our taxonomy indicates they are similar enough to be in the same species (points 1 and 2), the development of commercial biocontrol strains these days is almost always accompanied by generating the complete genome sequence of the biocontrol agent, which provides critical data for risk analysis. The presence of genetic determinants that are known to be associated with human pathogenic (or animal pathogenic) strains can be identified. Of course this is not a guarantee of safety, but the better our database and understanding of the genetic basis of human pathogens is, the better we can predict the risk. Realistically, our understanding of human pathogen genetics is increasing by leaps and bounds due to the genome sequencing of large numbers of strains that are and are not pathogens.

    “Lastly, while a few of the strains found to exhibit biocontrol activity are close enough in their classification to human or animal pathogens to benefit from comparative genomic analyses, a large number of biocontrol agents are so distantly related to any known human or animal pathogens that they pose a very low risk of causing harm as human or animal pathogens.”

    Like

    Reply

Leave a comment