Salmonella remains an extremely important bacterial pathogen to the swine industry. It is a significant pathogen affecting swine health and also represents one of the most important foodborne pathogens affecting people. A key aspect of Salmonella control is the use of cultivation methods in the laboratory. Samples must be cultured to isolate specific Salmonella strains, and this step is necessary to understand the spread of Salmonella throughout the swine production system. Unfortunately, conclusions about Salmonella transmission are highly dependent on the performance characteristics of these cultivation methods, and recently, we quantified a disturbing fact: the probability of detecting a specific Salmonella strain in a sample might have very little to do with its concentration in the sample but more to do with its ability to compete in the cultivation media and with the specific mixture of Salmonella strains present in the sample.

The overall objective of this project was to characterize the bias that cultivation media has on Salmonella detection and enumeration. To accomplish this, we conducted a series of competition experiments using Salmonella serovars and strains from the swine production system. We had the following specific aims:
1) Establish whether four Salmonella strains isolated from the swine production system exhibit heterogeneity in growth and competitive fitness during cultivation in broth media.
2) Determine which genes are either up or down regulated in these Salmonella strains in the presence of different cultivation broth media and during competition with other Salmonella strains.

To conduct our experiments, we used 4 different Salmonella enterica serovars originally isolated from swine: S. Agona, S. Derby, S. Mbandaka, and S. Typhimurium.  When grown individually in different media that are routinely used to culture Salmonella from swine samples, we found the following. In one broth, S. Derby exhibited the fastest growth and therefore appeared to have the potential to outcompete the other strains.  However, in a second broth at 37°C, S. Derby did not grow at all.  In this broth at 37°C, S. Typhimurium grew the fastest. In a third medium at 37°C, S. Agona grew the fastest.  S. Derby was able to grow slightly in the early hours of the growth curve.  When the serovars were grown in a Most Probable Number format, there were again major differences among strains.  In general, the serovars had a more difficult time growing at 42°C than at 37°C, even though 42°C is often used.  S. Derby again did not grow at all.  These results began to demonstrate that important Salmonella serovars could be entirely missed using standard protocols.
 
When the strains were competed against each other in 2, 3 or 4-way competitions, there was a clear ordering of competitive ability among the strains, but this order depended on the media and temperature.  Once again, S. Derby did not grow well and was never detected in any competition. 

Finally, we identified specific genes in these strains that might be contributing to the differential growth characteristics.  An understanding of the genetic basis of these differences could help us design a more appropriate cultivation protocol (with more appropriate media) to accurately culture all the Salmonella strains that might be present in swine samples.

This study confirmed what we had previously found in a pilot study: that the probability of detecting a specific Salmonella strain in a sample might have very little to do with its concentration in the sample but more to do with its ability to compete in the cultivation media and with the specific mixture of Salmonella strains present in the sample.  We will now proceed to identify optimal strategies for cultivating Salmonella from swine samples. 

Dr. Randall Singer, DVM, MPVM, PhD
Associate Professor of Epidemiology
Department of Veterinary and Biomedical Sciences
300A Veterinary Sciences Building
1971 Commonwealth Ave., St. Paul, MN  55108
612-625-6271""
[email protected]