Pathogens boosted by food additive

Epidemic strains of the bacterium Clostridium difficile have now been found to grow on unusually low levels of the food additive trehalose, providing a possible explanation for C. difficile outbreaks since 2001.
Jimmy D. Ballard is in the Department of Microbiology and Immunology, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma 73190, USA.

Search for this author in:

Between 2001 and 2006, epidemic strains of the bacterium Clostridium difficile, which can inhabit the bowel and cause dangerous diarrhoea, unexpectedly emerged in the United States, Canada and several European countries1,2. Most of these strains originated from a single lineage of C. difficile known as ribotype 027 (RT0272), which has now spread around the world3. Of particular concern has been the correlation between RT027 and a dramatic increase in deaths related to C. difficile4. The mystery of why this ribotype and a second one, RT078, became so prevalent apparently out of thin air has remained largely unsolved5. On page 291, Collins et al.6 raise the possibility that the seemingly harmless addition of a sugar called trehalose to the food supply contributed to this disease epidemic.

Collins and colleagues first explored how RT027 and RT078 grow, by comparing carbon-source preferences between strains of C. difficile. They noted a peculiar property of these two lineages — they can use low concentrations of trehalose as a sole source of carbon. Next, the authors analysed the genomes of RT027 and RT078, and discovered that each encodes unusual sequences that might explain their ability to grow in low levels of trehalose.

The researchers showed that RT027 carries a single-nucleotide genetic variant that changes an amino-acid residue in the protein TreR from leucine to isoleucine. TreR is a transcriptional repressor that is inhibited by trehalose. When active, TreR prevents expression of the gene treA, which encodes a phosphotrehalase enzyme involved in metabolizing trehalose into glucose and glucose derivatives. Thus, trehalose is metabolized only when its levels are high enough to inhibit TreR. Collins et al. propose that the mutation in RT027 changes TreR’s affinity for trehalose and allows it to be repressed by substantially lower levels of the sugar than normal. This frees the TreA protein to metabolize trehalose and allows RT027 to grow on low levels of the sugar (Fig. 1).

Figure 1 | Increased virulence of the bacterium Clostridium difficile. Two lineages of C. difficile, RT027 and RT078, have become widespread since the early 2000s. Collins et al.6 have demonstrated that different mutations have arisen in each strain to improve the microbes’ ability to grow on low concentrations of the sugar trehalose, which has been added to foods since 2001. RT078 has acquired four genes, including one that encodes the protein PtsT, which transports trehalose into C. difficile cells. In RT027, mutation of the protein TreR increases the protein’s affinity for trehalose, which in turn inhibits TreR’s ability to bind to DNA and repress transcription of the gene treA. TreA protein, expressed when TreR is repressed, metabolizes trehalose to glucose and derivatives, enabling cell growth at low trehalose concentrations.

By contrast, RT078 has adapted to grow on low amounts of trehalose by acquiring four genes involved in trehalose uptake and metabolism. The genes encode second copies of TreR and TreA, a trehalose transporter protein dubbed PtsT that helps cells take up the sugar, and another enzyme, TreX, involved in trehalose metabolism. Unexpectedly, RT078 does not share the genetic alteration in TreR that is found in RT027. As Collins and colleagues point out, it therefore seems that two epidemic strains of C. difficile have optimized trehalose metabolism in unrelated ways.

The investigators next provided evidence that trehalose metabolism directly relates to enhanced virulence of RT027 in vivo. First, they showed that deleting treA in RT027 and thereby preventing trehalose metabolism markedly reduced the virulence of this strain in mice. Second, adding trehalose to the diet of mice infected with RT027 increased the animals’ risk of death. However, the bacterial load of RT027 was not higher in mice fed trehalose than in those on a trehalose-free diet, indicating that increased risk of death is not simply due to the presence of more bacteria. Rather, the authors found that improved trehalose metabolism enables RT027 to produce higher levels of a C. difficile toxin.

Turning to RT078, Collins et al. demonstrated that just one of the four acquired proteins, the trehalose transporter PtsT, was responsible for the strain’s increased ability to grow on low levels of trehalose (Fig. 1). The authors showed that PtsT confers a competitive growth advantage over other lineages in the presence of trehalose.

Finally, Collins and colleagues investigated the relevance of their observations in humans. Experimental infection would be difficult in people, so the researchers instead collected fluid from the small intestine of three participants on a normal diet. The fluid contained levels of trehalose sufficient to promote expression of treA in RT027 but not in other strains, supporting the potential for human relevance.

RT027 was first isolated in 1985, from a person infected with C. difficile. But this ribotype was not associated with hospital outbreaks, increased death rates or epidemics until the early 2000s. Similarly, RT078 lineages isolated before the C. difficile epidemics carry the genetic information for enhanced trehalose metabolism, but were of little consequence to the epidemiology of this disease. Why did these ribotypes suddenly emerge at epidemic levels only 15 years ago?

Collins and colleagues propose a surprising answer. Before 1995, high production costs made trehalose untenable as a food additive. But manufacturing innovations7 reduced the cost of trehalose production more than 100-fold8, and the US Food and Drug Administration and European agencies approved the sugar as a safe food additive in 2000 and 2001, respectively (see; Trehalose is now added to a variety of food products, including pasta, ice cream and minced beef. The authors provide a timeline (see Fig. 6 of the paper) to illustrate how supplementing the food supply with trehalose preceded the C. difficile outbreaks caused by RT027 and RT078. They therefore suggest that the addition of trehalose to the food supply might have increased the sugar in the human bowel to levels high enough to enable growth of these ribotypes.

The study’s findings raise several avenues for future research. For instance, the connection between trehalose metabolism and toxin production, and how this is linked to increased death rates in people infected with RT027, will require further analysis. Whether trehalose in the human colon, where disease occurs, reaches high enough levels to affect RT027 and RT078 virulence is also unknown. The authors tested fluid from the small intestine, thus bypassing the colon, where the complex complement of gut microbes might break down trehalose.

Despite these concerns, the correlative findings of Collins and colleagues’ study are compelling. It is impossible to know all the details of events surrounding the recent C. difficile epidemics, but the circumstantial and experimental evidence points to trehalose as an unexpected culprit.

Nature 553, 285-286 (2017)

doi: 10.1038/d41586-017-08775-4
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up


  1. 1.

    Bartlett, J. G. Ann. Intern. Med. 145, 758–764 (2006).

  2. 2.

    McDonald, L. C. et al. N. Engl. J. Med. 353, 2433–2441 (2005).

  3. 3.

    He, M. et al. Nature Genet. 45, 109–113 (2013).

  4. 4.

    Hunt, J. J. & Ballard, J. D. Microbiol. Mol. Biol. Rev. 77, 567–581 (2013).

  5. 5.

    Kuijper, E. J., van Dissel, J. T. & Wilcox, M. H. Curr. Opin. Infect. Dis. 20, 376–383 (2007).

  6. 6.

    Collins, J. et al. Nature 553, 291–294 (2018).

  7. 7.

    Maruta, K. et al. Biosci. Biotechnol. Biochem. 59, 1829–1834 (1995).

  8. 8.

    Higashiyama, T. Pure Appl. Chem. 74, 1263–1269 (2002).

Download references