Establishment Of Protein Fingerprint Database Of Common Clinical Gram Positive Cocci

Objective: To establish a protein fingerprint database of common clinicalGram positive cocci including Staphylococcus aureus, Staphylococcusepidermidis, Staphylococcus haemolyticus, Enterococcus faecium andEnterococcus faecalis, and to provide a practicable tool for rapididentification of bacteria.Methods:1.185clinical bacterial isolates including Staphylococcus aureus,Staphylococcus epidermidis, Staphylococcus haemolyticus, Enterococcusfaecium and Enterococcus faecalis were collected and studied.2. All of thestrains were identified by16S rRNA gene sequencing.3. Bacterial proteinswere detected by surface enhanced laser desorption&ionizationtime-of-flight mass spectrometry (SELDI-TOF MS). Protein fingerprintswere analyzed by ProteinChip and Biomarker Wizard software.4. Thereproducibility of SELDI-TOF MS was analyzed by comparing the proteinfingerprints from different spots of AU chip.5. The stability of the proteinexpression of bacteria from each species were analyzed based on theprotein fingerprints which acquired from different strains belong to the same species.6. The influence on protein fingerprint under different culturetimes were analyzed.7. All isolates were assigned into training set whichincluded20strains of each species and testing set which included the reststrains. The protein profiles of all isolates were detected by SELDI-TOFMS. The common protein peaks of each species from training set wereselected by Biomarker Wizard software and mean MW of these peaks werecalculated. The data were imported into self-constructed Fingerwavesoftware. Then a small protein fingerprint database was built, whichincluding S. aureus, S. epidermidis, S. haemolyticus, E. Faecium and E.Faecalis.8. The ability of the database to determine these five species wasthen established using85additional testing strains for these five species.The tested strains were identified based on a similarity between the data ofprotein peaks of unknown strain and those of strains used to construct thedatabase. The species to which the unknown strain is considered to belongis the one for which the best similarity was obtained.Results:1.16S rDNA fragments of expected were obtained for clinicallyisolated bacteria. After compared with GenBank database, the sequences ofall strains matched perfectly with corresponding type strains (matching rate≥98.5%).2. Bacteria of different species had various protein fingerprintsfrom3KD to20KD.3. The protein profiling of the same bacterium hadgood reproducibility.4. The different strains from the same species hadsimilar protein fingerprints and the CV of main protein biomarkers in molecular weight was less than0.5%.5. No significant difference wasfound between protein fingerprint of different culture times.6. In thispreliminary research, a small protein fingerprint database was developedfor five common Gram positive cocci.7. All the tested strains using thedatabase were successfully identified with a result perfectly consistent withthose obtained by using automated microbiology analysis system andmolecular biological method.Conclusion: The technology of SELDI-TOF MS coupled withself-constructed Fingerwave software can identify the bacteria accurately.Thus, construction of a large-scale protein fingerprint database is desired,which will provide a practicable tool for rapid identification of bacteria.