Infections with antibiotic resistant bacteria are increasing at an alarming rate. Antibiotic resistance is a particular problem in patients with blood stream infections or sepsis, due to the mortality that rises exponentially as effective treatment is delayed. Bacterial species identification can often be used to guide antibiotic therapy while true antibiotic susceptibility test results are still pending. In fact, a few types of molecular assays have recently begun to offer rapid detection of the bacteria responsible for blood stream infections. However, existing methods are cumbersome, expensive and are limited to detection of a small minority of possible pathogens. We have developed a new paradigm for multiplex bacterial species identification. Our approach replaces the idea of using one molecular probe to identify one DNA target with the concept of using a mixture of a small number of molecular probes to act together and identify a very wide range of DNA target sequences. We will use this approach to design a real-time PCR assay capable of detecting virtually all pathogenic bacteria and common contaminants within a single PCR assay well. Key to this assay is the development of novel """"""""sloppy molecular beacons (SMB), which can tolerate substantial sequence mismatches with their DNA target and yet still fluoresce brightly in the presence of these targets. In this proposal, we will combine our novel multiplex detection approach and our SMB technology with an additional form of multiplexing which we call a """"""""virtual array"""""""" to make a broadly useful """"""""molecular blood culture"""""""". The molecular blood culture will identify virtually all pathogenic bacteria and common contaminants in a rapid single-well PCR assay. We will then develop this assay for direct species identification first on positive patient blood cultures, and later on blood samples drawn directly from patients with bacteremia and sepsis.
The specific aims of this proposal are: 1) to develop a model assay that identifies virtually all pathogenic bacteria in a single PCR tube or well in under two hours. 2) To further improve assay performance by creating a virtual array within the assay tube or well, enabling the incorporation of additional assay targets. 3) To develop a library of hybridization patterns for simple and multiple infections by serial analysis of clinical strains and identify the """"""""Tm space"""""""" of each bacterial species. 4) To test and optimize the detection system on clinical blood cultures and directly drawn patient blood samples and then prospectively test the assay.
This proposal will develop methods to very rapidly identify the bacteria that cause infections in a sick patient's blood stream or in other parts of the body that are normally sterile. Rapid identification will make it possible to treat patients with the correct antibiotics more rapidly. This will save patient lives, reduce days in the hospital and permit more rational use of antibiotics.
