Green fluorescent protein-propidium iodide (GFP-PI) based assay for flow cytometric measurement of bacterial viability

Background: Several staining protocols have been developed for flow cytometric analysis of bacterial viability. One promising method is dual staining with the LIVE/DEAD BacLight bacterial viability kit. In this procedure, cells are treated with two different DNA-binding dyes (SYTO9 and PI), and viability is estimated according to the proportion of bound stain. SYTO9 diffuses through the intact cell membrane and binds cellular DNA, while PI binds DNA of damaged cells only. This dual-staining method allows effective separation between viable and dead cells, which is far more difficult to achieve with single staining. Although SYTO9-PI dual staining is practical for various bacterial viability analyses, the method has a number of disadvantages. Specifically, the passage of SYTO9 through the cell membrane is a slow process, which is significantly accelerated when the integrity of the cell membrane is disrupted. As a result, SYTO9 binding to DNA is considerably enhanced. PI competes for binding sites with SYTO9 and may displace the bound dye. These properties diminish the reliability of the LIVE/DEAD viability kit. In this study, we investigate an alternative method for measuring bacterial viability using a combination of green fluorescent protein (GFP) and PI, with a view to improving data reliability. Methods: Recombinant Escherichia coli cells with a plasmid containing the gene for jellyfish GFP were stained with PI, and green and red fluorescence were measured by FCM. For comparison, cells containing the plasmid from which gfp was removed were stained with SYTO9 and PI, and analyzed by FCM. Viability was estimated according to the proportion of green and red fluorescence. In addition, bioluminescence and plate counting (other methods to assess viability) were used as reference procedures. Results: SYTO9-PI dual staining of bacterial cells revealed three different cell populations: living, compromised, and dead cells. These cell populations were more distinct when the GFP-PI combination was used instead of dual staining. No differences in sensitivity were observed between the two methods. However, substitution of SYTO9 with GFP accelerated the procedure. Bioluminescence and plate counting results were in agreement with flow cytometric viability data. Conclusions: in bacterial viability analyses, the GFP-PI combination provided better distinction between current viability stages of E coli cells than SYTO9-PI dual staining. Additionally, the overall procedure was more rapid. No marked differences in sensitivity were observed. (C) 2004 Wiley-Liss, Inc.