|Values are valid only on day of printing.|
Click CC box for captions; full transcript is below.
Published: February 2012Print Record of Viewing
Matrix-assisted laser desorption ionization time of flight (MALDI TOF) mass spectrometry offers a safe, cost-effective, and adaptable system for rapid identification of bacteria and fungi. Recent studies have shown that MALDI TOF may outperform conventional identification assays. This presentation provides an explanation of this technology and our experience using it in the clinical laboratory.
Presenter: Robin Patel, MD
Welcome to Mayo Medical Laboratories' Hot Topics. These presentations provide short discussions of current topics and may be helpful to you in your practice.
Our presenter for this program is Dr. Robin Patel, Chair of the Clinical Microbiology Division in the Department of Laboratory Medicine and Pathology at Mayo Clinic in Rochester, Minnesota. Dr. Patel discusses how matrix-assisted laser desorption ionization time of flight (MALDI TOF) mass spectrometry works for bacterial identification and Mayo Clinic’s experience with the technology in the clinical laboratory. Thank you, Dr. Patel.
Thank you, Sharon, for that introduction.
The objectives of this presentation are to explain how matrix-assisted laser desorption ionization time of flight (also referred to as MALDI TOF) mass spectrometry works, to summarize our experience with matrix-assisted laser desorption ionization time of flight mass spectrometry for bacterial identification, and to explain the role of matrix-assisted laser desorption ionization time of flight mass spectrometry in the clinical laboratory.
Mass spectrometry measures particles based on their mass-to-charge ratio. To do this, a sample (in the described method, the whole organism) is exposed to an ion source, and its particles (in the described method, proteins) ionized, separated based on their mass-to-charge ratio, detected and then the generated mass spectrum compared to a library of mass spectra.
Mass spectrometry requires an ion source, mass analyzer, and detector. There are multiple possible ion sources. In the past, ionization required molecules in the gas phase, limiting analysis to volatile compounds or those that could be rendered volatile. Large nonvolatile polar molecules, such as proteins, were not easily analyzed and, therefore, mass spectrometry was not used for protein analysis. With the arrival in the late 1980s of matrix-assisted laser desorption ionization, mass spectrometry based on microbial proteomics became possible. Matrix-assisted laser desorption ionization is a soft ionization technique allowing molecules to remain relatively intact during ionization. Proteins can be measured as little protein fragmentation occurs. Following ionization, the ions are separated, enabling measurement of mass. Using the approach covered in today’s presentation, ions are separated by time of flight in a flight tube.
I will first go over how matrix-assisted laser desorption ionization time of flight mass spectrometry is practically done in the clinical laboratory, starting from a colony. There are several systems available; I will overview the Bruker Daltonics system, the system with which I have the most experience. A colony may be picked directly from a plate and then smeared onto a target plate. Then, 1 to 2 mcL of a “matrix” consisting, for example, of alpha-cyano-4-hydroxycinnamic acid dissolved in acetonitrile and trifluoroacetic acid, is added and dried on the plate.
The target plate is placed into the plate chamber of the mass spectrometer, the plate chamber closed, and analysis performed. A 24-spot target plate can be prepared and analyzed in under an hour. This includes spotting the colonies, adding matrix, and analyzing the spectra, equating to ~2 to 3 min/sample.
Let us look at the details. The sample is mixed with the matrix and cocrystalized onto the target plate, the "matrix-assisted" component of matrix-assisted laser desorption ionization. The matrix "buffers" the sample, preventing its decomposition, and enabling transformation of laser light into heat.
A laser is applied, the "laser" component of matrix-assisted laser desorption ionization.
The matrix absorbs energy from the laser, releasing it into the sample as heat. This causes the sample to desorb and form singly charged ions, the "desorption and ionization" component of matrix-assisted laser desorption ionization.
Next the mass of the ions is analyzed. This is accomplished using a flight tube, the lighter ions traveling faster and, therefore, being detected earlier than the heavier ions. In the described method, particles are typically singly charged. The net result is generation of a mass spectrum in which the mass-to-charge is plotted against signal intensity. Only highly abundant proteins that are of low mass and ionize readily are detected. These are typically ribosomal proteins, although the specific nature of the analyzed proteins is not part of the analysis.
The mass profile is used as a fingerprint or mass spectrum (as shown here) to compare with those of well-characterized organisms in a database. The spectrum typically includes genus- and species-specific peaks, so that with a comprehensive library of spectra, the genus and often the species of the organism may be determined using bioinformatics.
Multiple commercial platforms are available. The most studied system is that from Bruker Daltonics, and includes a mass spectrometer along with the "Biotype"” software and database. Another system uses a Shimadzu Axima Assurance mass spectrometer with Launchpad software and the AnagnosTec GmbH (SARAMIS) database. This system was recently acquired by bioMérieux and is being redeveloped and called "VITEK MS." Yet another system, Andromas, is used mostly in Europe. The last 2 systems exclusively use cell lysis on the target plate (without off-plate extraction).
Using the Bruker system, some organisms may require preparatory extraction in order to generate spectra of sufficient quality to enable microbial identification. A general approach to extraction is shown. The isolate of interest is placed into a microcentrifuge tube with 70% ethanol, mixed, and centrifuged at 20,800 g. The supernatant is decanted and the pellet dried. 70% formic acid and acetonitrile are added and the mixture vortexed and centrifuged at 20,800 g for 2 minutes. The supernatant is deposited onto a target plate, dried and overlaid with matrix.
The number of publications on microbial identification by matrix-assisted laser desorption ionization time of flight mass spectrometry has skyrocketed over the last couple of years because of high identification rates combined with fast turnover time and low costs for consumables compared with conventional methods. This new technology has performed excellently; some studies have even shown that matrix-assisted laser desorption ionization time of flight mass spectrometry outperforms conventional identification assays.
Matrix-assisted laser desorption ionization time of flight mass spectrometry was initially evaluated for single organisms or organism groups, but more recently has been shown to apply to all types of bacteria and even fungi. The majority of studies have been from outside of the United States, particularly from Europe, where matrix-assisted laser desorption ionization time of flight mass spectrometry is already in widespread use in clinical microbiology laboratories.
We have recently published 2 studies evaluating matrix-assisted laser desorption ionization time of flight mass spectrometry for bacterial identification. The first compared the Bruker mass spectrometry system to the BD Phoenix Automated Microbiology System for identification of Gram-negative bacilli. A Microflex LT mass spectrometer was used in conjunction with Bruker Biotyper MALDI Automation Control software version 2.0 and a Biotyper library containing 3740 reference spectra. The system generates a score; criteria used to attain species- and genus-level identifications are shown. A score of ³2.0 indicates acceptable species identification (assuming that there is at least a 10% score difference between the top match and different genera or species with closely related spectra). A score of 1.7 to 1.999 indicates acceptable genus identification. If the score is <1.7, the system does not reliably identify the organism. These score cutoffs were specified by Bruker Daltonics. Discrepant results were resolved by conventional biochemicals and/or 16S ribosomal RNA gene or gyrB sequencing.
We studied 440 Gram-negative bacilli. 89% of 440 isolates generated a high confidence score (> 2.0) with direct colony testing. The 11% that did not generate a score of > 2.0 were extracted. Using the BD Phoenix system, 83% were identified to the genus level and 75% to the species level. Using mass spectrometry, 93% were identified to the genus level and 82% to the species level. Statistically, mass spectrometry outperformed the BD Phoenix for identification of Gram-negative bacilli to the genus and species levels.
The 440 isolates included 308 common Gram-negative bacilli, defined as organism types seen more than once a week in our laboratory. This included organisms such as Escherichia coli and Pseudomonas aeruginosa. Both systems identified 93% of common isolates to the species level, and the BD Phoenix and mass spectrometry systems identified 95% and 96% of common isolates, respectively, to the genus level. There was no statistical difference between the systems for identification of common Gram-negative bacilli.
The 440 isolates included 132 uncommon Gram negative bacilli. Mass spectrometry identified 85% of the uncommon organisms to the genus level and 56% to the species level. The BD Phoenix system identified 52% of the uncommon organisms to the genus level and 34% to the species level. Mass spectrometry outperformed the BD Phoenix for identification of the uncommon Gram-negative bacilli.
There were limitations of the mass spectrometry system evaluated. Shigella species were not included in the studied database and Shigella species was identified as Escherichia coli. These 2 organisms are so closely related that they are not readily distinguished by this technique. In the same vein, closely related species were not differentiated from one another, including the species comprising the Enterobacter cloacae complex and those comprising the Klebsiella pneumoniae complex. Finally, the database contained some confusing synonyms, variously identifying Stenotrophomonas maltophilia as Stenotrophomonas maltophilia, Pseudomonas hibiscicola, or Pseudomonas beteli for example.
Incorrect genus identification was provided for 58 (13%) and 10 (2%) isolates using the BD Phoenix and Bruker mass spectrometry systems, respectively. Correct genus but incorrect species identification was provided for 10 (2%) and 19 (4%) of the isolates for the BD Phoenix and Bruker mass spectrometry systems, respectively.
We followed-up our first study by a second study of 305 clinical isolates representing species of staphylococci, streptococci and related genera listed in Koneman’s Color Atlas & Textbook of Diagnostic Microbiology including 12 Staphylococcus species, 14 Streptococcus species, 8 Enterococcus species, and 14 "related genera" including Rothia, Pediococcus, Micrococcus, Granulicatella, Abiotrophia, Aerococcus, Gemella, Macrococcus, Kocuria, Helcococcus, Arthrobacter, Lactococcus, Leuconostoc, and Facklamia species.
Isolates were cultured overnight on 5% sheep blood agar at 35ºC, except Granulicatella and Abiotrophia species, which were cultured on chocolate agar. Results were compared to phenotypic testing and partial 16S ribosomal RNA gene sequencing. Seven isolates were excluded as they were not present in the Bruker Biotyper database, leaving 298 isolates. Discrepant results were resolved by additional biochemical analysis and/or partial 16S ribosomal RNA gene sequencing. Colonies were tested directly and subjected to extraction.
298 isolates were tested following extraction and by direct colony testing. Following extraction, 95% were identified to the genus level and 69% to the species level. With direct colony testing, 56% were identified to the genus level and 20% to the species level.
Subgroup analysis was performed for the staphylococci, streptococci and enterococci and the "related genera." Staphylococci, streptococci and enterococci subjected to extraction yielded a genus-level identification in 98% of cases and a species-level identification in 79% of cases. Staphylococci, streptococci, and enterococci tested directly from colonies yielded a genus-level identification in 63% of cases and a species-level identification in 26% of cases. The "related genera" subjected to extraction yielded a genus-level identification in 88% of cases and a species-level identification in 4% of cases. The "related genera" tested directly yielded a genus-level identification in 40% of cases and a species-level identification in 5% of cases.
There were some limitations. Four of 10 Streptococcus mitis species group isolates were misidentified as Streptococcus pneumoniae using extraction. All 10 Streptococcus pneumoniae were correctly identified. This finding was previously reported. The difficulty in distinguishing between closely related species may be due to the similar composition of ribosomal proteins, the main target for matrix-assisted laser desorption ionization time of flight mass spectrometry. Two Enterococcus casseliflavus/gallinarum were misidentified as Enterococcus phoeniculicola using extraction and one by direct colony testing. Enterococcus phoeniculicola has been isolated from birds and it is not known to cause human disease. This misidentification was previously reported. In one report, correct identification of Enterococcus casseliflavus was reached when the library was updated. We tested one of our isolates using preparatory extraction after implementing an update to our library and also confirmed Enterococcus casseliflavus.
In summary, matrix-assisted laser desorption ionization mass spectrometry is automated, green, doesn't require specific expertise in mass spectrometry, and has a rapid turnaround time and high throughput capability. It only requires a single colony and is associated with a low exposure risk due to sample inactivation. Although not covered in today's presentation, this approach is cost effective and has demonstrated high interlaboratory reproducibility. It has broad applicability (covering all types bacteria including anaerobes, and fungi). The system is open and adaptable by the user.
There are limitations to matrix-assisted laser desorption ionization time of flight mass spectrometry. No susceptibility information is provided and the technology is not generally useful for direct testing of clinical specimens. Some organisms require repeat analyses, and additional processing. The acceptable score cutoffs vary between studies. Some closely related organisms are not differentiated, and sporulation, not covered in today’s presentation, may challenge identification. The available databases could use improvement. Comparison of data from different companies’ instruments is not feasible. Laboratories acquiring the needed equipment will suffer financial loss on existing equipment. Finally, the available systems are not approved by the United States Food and Drug Administration.
This slide illustrates bacteriology workflow today, including the use of rapid biochemicals, an automated phenotypic identification system, long tubed biochemicals, and 16S ribosomal RNA gene sequencing.
With matrix-assisted laser desorption ionization mass spectrometry, colonies growing on plates can be directly identified, without the need for many other identification tools used today.
I would like to thank Ryan Saffert, PhD, Scott Cunningham, Nancy Wengenack, PhD, Adnan Alatoom, MD, PhD, the Clinical Bacteriology and Mycology Laboratory staff, Bruker Daltonics and bioMérieux for their valued input into this presentation.