Bryan Merrill and Sandra Burnett, Department of Microbiology and Molecular Biology
The deadly honey bee disease American Foulbrood (AFB) is caused by the bacterium Paenibacillus larvae. The widespread use of antibiotics to prevent or treat AFB has led to many drug-resistant strains of P. larvae. Bacteriophages can target and destroy specific strains of bacteria and can be used to treat bacterial infections. This study accomplished two main goals. We determined the efficacy of previously isolated P. larvae phages as a treatment for AFB by applying a phage spray to infected beehives. We also isolated P. larvae bacteria from the infected hives we treated and identified phages to better counteract a wider variety of local P. larvae strains.
Materials and Methods
Local field isolate PL2 was used to culture phage Fawkes, and isolate PL6 was used to culture phages Emery and Abouo. Large scale phage liquid cultures were prepared by inoculating 500 mL of 0.25x LB broth in a 2 L flask with 10 mL of overnight culture of bacteria, 10 mL of 1 M CaCl2, and 10 mL of hightiter phage culture (at least 108 plaque forming units per mL). Flasks were incubated at 35°C in a shaker rotating at 150 rpm for 24 hours. Liquid cultures were pelleted, passed through a 0.45 μm filter, and stored at 4°C. The titer of each liquid culture was tested using a spot-test dilution series to ensure that some clearing occurred at a dilution of 10-8.
Each treatment consisted of 25 mL of phage Fawkes, 25 mL of phages Emery/Abouo, and 275 mL of a 1:1 mixture of sugar and distilled water. Hives were treated by spraying ~8 mL of treatment onto each side of each frame in the bottom two brood boxes. Each infected hive was treated three times with three days between each treatment.
The amount of AFB infection in each hive was assessed before and after the phage treatment regimen. We counted the number of frames with brood (eggs, uncapped larvae, or capped larvae) and assigned each brood frame a value based on the number of cells with an AFB-infected larvae (0 = no infected cells, 1 = 0-10 infected cells, 2 = 10-100 infected cells, 3 = 100+ infected cells). Total infection scores were calculated by dividing the sum of the frame scores by the total number of brood frames.
Additional P. larvae field isolates were gathered from infected hives by sampling ropey larval remains, suspending them in 30% DMSO, and storing them at 4°C until they were processed. A portion of each sample was diluted in 50 uL of 1X PBS (pH = 7.2) for 10 minutes and then streaked on MYPGP or PLA media which are selective for P. larvae. Streaked samples were incubated for 48-72 hours at 35°C.
Hive debris samples were gathered from beekeepers throughout Utah and used for phage isolation. Hive debris was ground in a mortar and pestle and added to a mixture of 5 different P. larvae isolates.
Five groups of hives were treated throughout the course of this study. Group 1 consisted of four hives that received all three phage treatments. The active AFB infection persisted and reached the advanced stages even after tylosin was applied a month prior. Immediately following the first treatment a hard freeze caused a considerable drop in the population in each hive. However, inspection during the second treatment revealed a large new batch of eggs and young larvae that the queen had laid. Unfortunately, the hives were not able to overcome the infection and the population loss. By the third treatment all four hives were either dead or too weak to survive the winter.
Group 2 consisted of three hives where two were treated and the other was not. The treatment for the first hive was applied incorrectly and the bees absconded. The second treated hive recovered and survived the winter. The untreated hive died before winter set in.
Group 3 consisted of 12 treated hives and 3 untreated hives. Each of the treated hives received all three treatments. One untreated hive died before treatment could begin, and during the treatment process the other two untreated hives also died. Of the 12 treated hives, 10 recovered and began increasing the bee population while two died. Of the 10 that recovered, two still had recurrent infection and were too weak going into winter to survive. Of the eight overwintering hives, six survived (although one was vandalized and died just before spring) and two died. One of the surviving hives showed AFB symptoms in the spring, was given a phage treatment, and was well prepared going into winter yet again.
Group 4 consisted of three hives. One hive was treated for an active infection and two hives that were treated preventatively. The treated hive saw a decrease in the AFB infection score during the course of the treatment (from 1.06 to 0.75) but after the treatment stopped the infection score rose to 0.78. Of the hives treated preventatively, one became infected with AFB while the other one did not.
Group 5 consisted of a single hive that was infected with AFB and was treated three times. The AFB infection score dropped following the first treatment from 0.54 to 0.33 but rose steadily, reaching near pretreatment levels.
We made several key observations during the treatment of AFB-infected hives with phages. First, following the first treatment there was a significant increase in brood production. Figure 1 illustrates a before (1a) and after (1b) picture of the same frame from treatment group 3 that is indicative of the results seen in the other 10 hives that were successfully treated. Second, hives in advanced stages of infection or were weak did not survive despite treatment. Third, our treatment is not effective against all strains of P. larvae in infected hives that we tested, indicating a need to add more phages to the treatment mixture.
We gathered additional samples from healthy and infected beehives throughout the state. From infected hives throughout the state of Utah we gathered and cultured 45 different P. larvae isolates from ropey larval remains. The 16S rRNA gene of 38 of those isolates was successfully sequenced and used to confirm that the bacteria was indeed P. larvae. The remaining seven strains also appear to be P. larvae based on negative catalase activity and appearance similar to the other strains. Of 57 bee debris samples gathered, 28 were enriched in P. larvae bacteria and phages were identified in 13 of them. Cross-infectivity assays and genome sequencing will soon be done to identify phages with broad host ranges that are lytic and lack known virulence genes for use in an expanded safety and treatment study during 2015.
Phage therapy is a good option for treating AFB in honeybees. Although this initial study treated hives with only three phages, all hives showed improvement after applying the treatment. The identification of many more bacterial strains and phages will allow us to select phages that will perform better at treating infected hives where our current treatment fell short. This pivotal study will pave the way for future research that will test a treatment containing more phages on a larger number of infected hives.