Project Summary

Bacteriophage Involvement in the Dissemination of Multi-drug Resistance by Salmonella

Principle Investigator(s):
Todd R. Callaway
U.S. Department of Agriculture, Agricultural Research Service
Completion Date:
May 2008



Bacteriophage are viruses that prey on bacteria and are commonly found in the gut of cattle and other mammals.  Lysogenic phage insert their DNA into the bacterial chromosome where it can remain as an inert “stowaway” genetic sequence for generations.  When the host bacterium is stressed, a hidden (or latent) phage can excise itself from the genome and take over the cellular machinery to produce daughter phage.  When this occurs, phages can inadvertently carry genes from one bacteria to another, a process known as transfection, which is widely used in laboratories to genetically-engineer bacteria.  A natural phage transfection transferred the genes that code for Shiga toxin 1 and 2 from Shigella dysenteriae into E. coli O157:H7 during the past 50,000 years.  Unfortunately, this was not a one-time event; gene transfer occurs with a variable and unknown frequency in the gastrointestinal tract of cattle.  Genes transferred by phage can range from relatively innocuous to dangerous, including genes that encode antimicrobial resistance.

Multi drug-resistant Salmonella are an increasing threat to the safety of beef.  Some of these MDR strains can appear spontaneously or be transmitted along over geographic or temporally.  Genes encoding antimicrobial resistance can be transferred between Salmonella species or from other bacterial species.  This natural genetic transfer can occur by DNA absorption, “bacterial sex”, or transfection via phage.  However, the mechanism of how antimicrobial resistance genes move among bacteria in the animal has never been adequately explained.

Recently our research group, in cooperation with the NCBA, found that lytic phage are more widespread in cattle in the real world than had been previously thought.  However the incidence of lysogenic phage (those than insert themselves into the bacterial genome) is unknown. Thus, a question arose; do lyosgenic phage provide a natural mechanism to spread antibiotic genes within Salmonella in the animal gut?

To answer this question we sampled feces from cattle from both beef cattle and dairy cattle.  Dairy cattle were included in this study because 50% of the ground beef is produced from dairy cattle.

The stated objectives for this work were: 

  1.  We will collect 120 fecal samples from each of the above locations (480 total samples).  Each sample will be assayed for populations of Salmonella spp., multi drug-resistant (MDR) Salmonella, phage, and phage active against Salmonella spp. 
  2. Feces will be stressed with UV light to cause excision of lysogenic phage present in the population 
  3. Salmonella isolates will be separately stressed through application of UV light to determine how many of these isolates contain latent phages capable of excising and carrying genes to new hosts. 
  4. We will combine phage-susceptible Salmonella strains with MDR Salmonella from our cattle samples (that represent different sero-groups) in continuous flow culture (to simulate the intestinal tract).  Phages active against both the MDR strain and the susceptible strain will be added to the mixed culture.  We will collect samples from the continuous flow culture over a period of 7 d and plate a 10-fold dilution series on 3 types of BGA plates each supplemented with an antibiotic that the MDR Salmonella is resistant to, but not the susceptible Salmonella.  Isolates from each dilution will be identified by sero-group to determine if the susceptible Salmonella strains acquired resistance from the MDR Salmonella.   

Technical procedures:  Salmonella species were isolated using the two plate method of Edrington et al. 2007.  Antimicrobial susceptibility was determined using the Sensititre automated antimicrobial susceptibility system according to the manufacturer’s directions (Trek Diagnostic Systems, Westlake, OH).  Escherichia coli ATCC 25922, E. coli ATCC 35218, and Enterococcus faecalis ATCC 29212 were used as quality control organisms.   All bacteriophage-caused plaques purified from all of the initial plates were assessed for their ability to form subsequent plaques on a range of intestinal bacteria and several strains of Salmonella.  Phage isolates were screened for activity against: S. derby, S. typhimurium, S. dublin, S. enteriditis, S. cholerasuis, S. montevideo, S. mbandaka.    Latent phage from within bacteria and Salmonella were obtained by treating the bacteria or fecal samples with 15 μg/ml of mitomycin C to cause the induction of a bacterial “SOS response” causing the phage to “abandon ship”.  Following incubation, supernatant samples were spotted onto an agar lawn of E. coli B (an indicator strain that is highly subject to phage colonization) and plaques were indicative of phage presence.   

Chemostats (New Brunswick, Bioflow 3000) contained Viande-Levure (VL) broth, previously used in our laboratory to simulate intestinal conditions in cattle.  Co-cultures of MDR Salmonella Enteritidis and a susceptible Salmonella Kentucky were established via normal methods in our laboratory.  Phage were be added to achieve a 10:1 phage to total Salmonella populations ratio initially.  Samples were diluted daily to determine populations of Salmonella species and of phage (via serial dilution and plating onto E. coli B lawns).  Plating onto supplemented antibiotic plates indicated if Salmonella could grow on the Gen and Tet plates, and isolates were examined on the highest dilution plate to verify the serogroup differences between the Group C and D serotypes used.


From the 560 total fecal samples from feedlot and dairy cattle, only 22 (3.9%) contained any Salmonella.  Serotypes isolated included: 4 Cerro, 3 Mbandanka, 3 Heidelberg, 2 Kentucky, 2 Anatum, 2 Montevideo, Thompson, Javiana, Typhimurium, Orion, Oranienburg, Give, and Senftenberg.  Of these, none were found to be multi drug-resistant Salmonella.  Bacteriophage that lysed E. coli B were found in the fecal samples of 239 of the fecal samples collected indicating that phage are widespread in cattle as has been previously demonstrated. Of these 239 phage isolates only 2 isolates demonstrated lytic activity specifically against any Salmonella spp.    

Mitomycin C was used because UV light could not penetrate consistently through the opaque fecal material.  Mitomycin stresses bacterial cells and causes phage inserted in the genome to excise themselves, replicate and move on to new hosts that are not in immediate danger of dying.   

When cattle feces were treated with mitomycin C and resulting supernatants spotted against E. coli B.  Clearing zones (plaques) on the E. coli were created by 18 fecal samples that did not contain lytic phages (as determined in result 1 above).  These phage could be tentatively identified as probably having existed as a lysogen in the feces.  Thus these phage isolates were maintained for further study in subsequent research herein.

When the 22 Salmonella isolates were treated with mitomycin C. only 3, a Salmonella Enteritidis, Cerro and Javiana, contained lysogenic phage within its genome.  Under several attempts both using UV light in pure culture and using mitomycin C, these were the only cattle Salmonella isolates that demonstrated it was host to a lysogenic phage.  The isolated phages from Enteritidis and Cerro lysed the serotypes Typhimurium, Kentucky and Enteritidis when spotted onto lawns of these serotypes, but did not lyse Derby, Anatum or Cholerasuis.  The Javiana phage isolate did not create zones of clearing on any Salmonella species.   

Phages are widely distributed in cattle and this result was confirmed by the presence of generic phage in > 40% of the cattle fecal samples.  However, the low number of Salmonella-killing phages isolated in the present study corresponds with previous research from our group that found few Salmonella-killing phage in swine and poultry.    

Overall, our results indicate that lysogenic phage do increase genetic transfer of antibiotic resistance genes, at least in test-tube situations.  However, the low incidence of these phages in the intestinal tract of cattle may point to a limitation in their real-world role as genetic dissemination mechanism.  The fact that nearly 5% of the isolated Salmonella (albeit from a small n of 22) carried lysogenic phage suggests that these phage could play a role in future natural transformation events via carriage of antibiotic resistance or other virulence genes.  However, the ability of these transformed Salmonella to survive and compete in a mixed microbial ecosystem may not be as great as has been hypothesized in the absence of direct selective pressures.   

Further research is underway to examine the lysogenic phage as well as some of the apparently lytic phages that show some lysogenic features for their potential role as vectors for genetic transfer.  The chemostat study is being repeated currently with a selective pressure in the form of gentamycin treatment being added concurrently with the addition of the phage.  


Our data indicate that the incidence of these “hidden” phage is quite low (< 5%) in cattle.  Furthermore, the number of Salmonella isolates from cattle was very small, and only one of these Salmonella carried a hidden phage.   When these hidden (lysogenic) phages were added at artificially high concentrations to cultures of Salmonella, they did increase the rate of transfer of antibiotic genes in very specialized conditions, but in an environment more similar to the animal gut the gene transfer was less successful, and the resistant bacteria were quickly eliminated.  Bacteriophage can be involved in the dissemination of antimicrobial genes, however their incidence appears to be quite low and do not increase the risk of the development of MDR Salmonella.