|Home | About | Journals | Submit | Contact Us | Français|
Matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry is a rapid, accurate method for identifying bacteria and fungi recovered on agar culture media. We report herein a method for the direct identification of bacteria in positive blood culture broths by MALDI-TOF mass spectrometry. A total of 212 positive cultures were examined, representing 32 genera and 60 species or groups. The identification of bacterial isolates by MALDI-TOF mass spectrometry was compared with biochemical testing, and discrepancies were resolved by gene sequencing. No identification (spectral score of <1.7) was obtained for 42 (19.8%) of the isolates, due most commonly to insufficient numbers of bacteria in the blood culture broth. Of the bacteria with a spectral score of ≥1.7, 162 (95.3%) of 170 isolates were correctly identified. All 8 isolates of Streptococcus mitis were misidentified as being Streptococcus pneumoniae isolates. This method provides a rapid, accurate, definitive identification of bacteria within 1 h of detection in positive blood cultures with the caveat that the identification of S. pneumoniae would have to be confirmed by an alternative test.
In the early 1970s the first semiautomated blood culture system, the radiometric Bactec system, was introduced into the clinical microbiology laboratory. In subsequent years this system was refined with the development of fully automated, closed, continuously monitoring systems for the detection of microbial growth. Recently, a commercial real-time PCR system (LightCycler SeptiFast; Roche Molecular Systems) was introduced with the hope that culture-based systems could be replaced with this technology; however, initial reports documented that this system can be used as a complement to but not a replacement for the current generation of automated systems (16, 17, 19). Because culture-based systems will be used in the near future, accurate, rapid identification methods are still needed. As with blood culture systems, a transition in identification systems began in the early 1970s with the introduction of commercial biochemical strips and panels and then with the rapid development and refinement of automated instruments that inoculate, incubate, interpret, and report microbial identifications. Currently, most bacteria and fungi can be identified with these systems in a few hours to 1 to 2 days, with slow-growing or metabolically inert organisms requiring additional time or supplementary tests. The identification of organisms recovered in blood culture broths requires an initial subculture of the broth and overnight incubation to obtain isolated colonies for further testing or the concentration of the bacteria or fungi by centrifugation before further processing. In general, the approach of concentrating organisms by centrifugation and then identification by rapid biochemical tests (1a), fluorescent in situ hybridization (FISH) (4, 6, 15, 18), or commercial biochemical systems is accurate, although the limitations of incubation delays, the need for supplementary tests to identify fastidious organisms, and recommendations that confirmatory identification should be performed with subcultured isolates are not eliminated.
Although not widely used in clinical microbiology laboratories at this time, proteomic profiling is an alternative to biochemical and genome-based identification schemes. Analysis of bacterial or fungal colonies by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry produces reproducible, species-specific spectral patterns that can be used to identify microorganisms at the species level. A broad spectrum of organisms has been identified, including Gram-positive and Gram-negative bacteria, nocardia, mycobacteria, yeasts, and molds (1a-3, 5, 7-10, 12, 13). Compared to standard biochemical identification methods, this technique is rapid and reproducible with minimal consumable costs and has an accuracy comparable to that of genome sequencing (14). For laboratories currently using MALDI-TOF mass spectrometry for bacterial and fungal identifications, it is attractive to extend this technology to other applications. MALDI-TOF mass spectrometry has not been used previously for the identification of organisms in positive blood culture broths because a high organism density is necessary for an acceptable spectrum and hemoglobin and serum proteins can interfere with interpretations of the spectral patterns. In this work we present a technique using MALDI-TOF mass spectrometry for the identification of bacteria recovered from positive blood culture broths.
All blood cultures were processed with the Bactec 9240 blood culture system (Becton Dickinson, Franklin Lakes, NJ) by using Standard 10 Aerobic/F and 10 Anaerobic/F blood culture media. A total of 179 positive cultures from patient samples as well as an additional 33 broths cultures inoculated with less commonly isolated bacterial species were analyzed. All positive bottles were removed from the blood culture instrument between 7 a.m. and midnight and held at room temperature until processed for MALDI-TOF analysis at approximately 10 a.m. and 3 p.m. Inocula for the seeded cultures were prepared by suspending a single colony from a culture grown overnight in 10 ml of sterile saline and inoculating the blood culture bottle with 100 μl of the bacterial suspension and 10 ml of human blood. Inoculated bottles became positive between 6 and 30 h after placement into the blood culture instrument.
Samples were prepared by drawing blood from positive cultures into two 4-ml Vacuette Z serum separator tubes with a clot activator (Greiner Bio-One, Monroe, NC) and performing a series of five 1- to 2-min washing/centrifugation steps to remove red blood cells (RBCs) and proteins from the blood culture broths. Specifically, the serum separator tubes were initially centrifuged at 8,500 rpm for 2 min in a StatSpin Express 2 centrifuge (Iris Sample Processing, Westwood, MA). A small pellet of bacteria was visible at the surface of the polymeric gel, and the majority of the RBC component was pelleted beneath the gel layer. All of the liquid was removed except for approximately 500 μl that was used to gently resuspend the bacterial pellet, taking care not to disrupt the gel layer. If lysed blood and RBC debris remained above the gel layer, then all of the liquid was removed and 500 μl of fresh tryptic soy broth (TSB) was added to resuspend the bacteria. The bacterial suspensions from the two serum separator tubes were combined and inoculated into a 1.5-ml plastic microtube (Sarstedt, Nümbrecht, Germany). A low-speed spin was carried out with an Eppendorf 5415D centrifuge at 1,000 rpm for 1 min to pellet any remaining RBCs. The liquid portion was carefully transferred into a new 1.5-ml tube and recentrifuged at 13,000 rpm for 1 min to pellet the bacteria. The bacterial pellet was resuspended in 1 ml of RBC lysis solution (4.15 g NH4Cl, 0.84 g NaHCO3, and 0.186 g EDTA in 495 ml of cell culture-grade water) and incubated at 35°C for 10 min, and the bacterial cells were then pelleted at 13,000 rpm for 1 min, washed in 1 ml sterile water, centrifuged at 13,000 rpm for 1 min, and resuspended in 1 ml 70% ethanol.
The mixture of bacteria and ethanol was centrifuged in an Eppendorf 5415D centrifuge at 13,000 rpm for 2 min. The liquid was removed, and the pellet was briefly respun followed by the removal of residual ethanol and then resuspended in 50 μl of 70% formic acid. Fifty microliters of acetonitrile was added, and the sample was vortexed briefly. The mixture was centrifuged for 2 min at 13,000 rpm, and the supernatant transferred into a clean microtube. Samples were either analyzed immediately or stored at −20°C until MALDI-TOF mass spectrometry analysis.
The MALDI-TOF α-cyano-4-hydroxycinnamic acid matrix was prepared daily as a saturated solution in 50% acetonitrile and 2.5% trifluoroacetic acid (TFA). The sample to be analyzed was warmed to room temperature, and 1 μl was spotted onto a steel target plate (Bruker Daltonics Inc., Billerica, MA) and gently mixed with 2 μl of matrix solution.
Samples were evaluated by use of an UltraFlex I TOF-TOF apparatus (Bruker Daltonics Inc., Billerica, MA) in linear positive-ion mode across the m/z range of 2,000 to 20,000 with gating of ions below m/z 400 and a delayed extraction time of 450 ns. Each spot was measured by using 1,000 laser shots at 25 Hz in groups of 50 shots per sampling area of the spot. The data sampling rate was 0.5 GHz. Each plate was calibrated by using a mixture of Protein Standards I and Peptide Standards II (Bruker Daltonics). Spectra were analyzed by using MALDI BioTyper software (v 2.0) (BioTyper Library v 2.0.4; Bruker Daltonics), a proprietary algorithm for spectral pattern matching resulting in a logarithmic score from 0 to 3. Previous work using discrete bacterial colonies determined that a score of >1.9 indicates species identification, a score of 1.7 to 1.9 indicates genus identification, and a score of <1.7 indicated no identification (14). The applicability of these criteria for the identification of broth culture isolates was evaluated as part of this study. All MALDI-TOF mass spectrometry results were compared to identification by standard biochemical methods, and discrepancies were resolved by 16S rRNA gene sequencing.
To assess the use of MALDI-TOF mass spectrometry for the identification of bacteria recovered from blood cultures, the identity and spectral scores were determined for bacteria recovered from 212 positive blood cultures (Table (Table1).1). Spectral scores of <1.7 are generated most commonly when too few bacteria are in a sample to produce a sufficient number of peaks for reliable identification. Bacteria in 42 (19.8%) cultures produced a spectral score of <1.7 and could not be identified by this direct method. Twelve of the 42 isolates were Propionibacterium acnes isolates with spectral scores consistently of <1.7, an observation previously reported for tests performed by using isolated colonies (14). Although a specific identification by MALDI-TOF would be desirable, the Gram stain morphology of P. acnes is sufficiently characteristic that a presumptive identification from the initial positive broth should not pose a problem. All of the other bacteria not identified directly in broth cultures were identified accurately when subcultured colonies were tested.
Of the remaining 170 cultures, 24 had a score of between 1.7 and 1.9 and 146 had a score of >1.9. A total of 162 (95.3%) of the 170 bacterial isolates were identified correctly with a score of ≥1.7. All 8 isolates of Streptococcus mitis had spectral scores of >1.9 and were identified incorrectly as being Streptococcus pneumoniae isolates. The inability of the BioTyper software to classify S. mitis correctly has been recognized by other investigators (M. Kostrzewa, Bruker Diagnostics, personal communication). Five additional isolates of S. pneumoniae were correctly identified. We recommend that all organisms identified by MALDI-TOF mass spectrometry as being S. pneumoniae isolates should be confirmed by performing a bile solubility test directly on the blood culture broth (11).
We did not observe any difference in the identifications of isolates at the species level with spectral scores of 1.7 to 1.9 and >1.9. With the exception of two isolates of Lactobacillus (spectral scores of 1.7 to 1.9), one Ochrobacterium isolate (spectral score of 1.7 to 1.9), and one Salmonella isolate (spectral score of >1.9), all organisms were identified to the species level. Based upon previous work from our laboratory as well as others (Kostrzewa, personal communication), species within the Enterobacter cloacae complex cannot be separated definitively. Additionally, although four cultures of Acinetobacter baumannii were correctly identified in this work, we do not believe that Acinetobacter baumannii and Acinetobacter genospecies 3 are consistently separated by this method. The finding that spectral scores for bacteria in blood cultures were generally lower than scores for bacteria tested from culture plates is not surprising because the score is influenced by both the species of bacteria and number of organisms tested. Using a series of dilution experiments, we determined that excellent spectra were obtained when a minimum of 106 CFU were spotted onto the target plate (our unpublished observations), a concentration easily obtained from culture plates but at the threshold of what is normally present in a positive blood culture bottle. Although the spectral score could be improved by processing a larger volume of broth or preincubating the broth before MALDI-TOF mass spectrometry analysis, we believe that this is unnecessary because a spectral score of ≥1.7 could accurately identify 75.8% of the isolates to the species level.
Ten of the positive blood cultures were polymicrobic. At least one organism from 9 of 10 cultures was identified with a spectral score of >1.7. While it is encouraging that mixtures of multiple organisms did not produce an incorrect identification and the predominant organism was identified in all but one culture, not all organisms present in polymicrobial blood cultures can be reliably detected by examining the MALDI-TOF mass spectrometry profiles.
The experiments presented in this report were performed with Bactec standard 10 Aerobic/F and 10 Anaerobic/F blood culture media. Although experiments with other medium formulations were not performed, we have observed uniformly consistent spectra and reproducible identifications with bacteria and fungi grown on a variety of media. The only variable that appears to affect obtaining satisfactory spectra is the number of organisms present in the test sample.
In summary, the use of MALDI-TOF mass spectrometry for the identification of blood culture isolates directly from positive broth cultures was accurate and rapid. Although inadequate spectra were produced for approximately 20% of the organisms, most commonly due to insufficient numbers of bacteria in the blood culture broth, all identifications with spectral scores of ≥1.7 were accurate except for S. mitis isolates. Misidentifications of these isolates could be rapidly resolved with a direct bile solubility test. We believe that the direct identification of blood culture isolates is a useful application of mass spectrometry for laboratories using this technology for routine microbial identification.
This research was supported by the Intramural Research Program of the NIH Clinical Center, Departments of Laboratory Medicine and Critical Care Medicine.
Published ahead of print on 2 December 2009.