Applied Microbiology and Biotechnology

, Volume 93, Issue 3, pp 965–974

MALDI-TOF MS in microbiological diagnostics—identification of microorganisms and beyond (mini review)


  • Andreas Wieser
    • Max von Pettenkofer-Institut für Hygiene und Medizinische MikrobiologieLudwig-Maximilians-Universität, München
  • Lukas Schneider
    • Max von Pettenkofer-Institut für Hygiene und Medizinische MikrobiologieLudwig-Maximilians-Universität, München
  • Jette Jung
    • Max von Pettenkofer-Institut für Hygiene und Medizinische MikrobiologieLudwig-Maximilians-Universität, München
    • Max von Pettenkofer-Institut für Hygiene und Medizinische MikrobiologieLudwig-Maximilians-Universität, München

DOI: 10.1007/s00253-011-3783-4

Cite this article as:
Wieser, A., Schneider, L., Jung, J. et al. Appl Microbiol Biotechnol (2012) 93: 965. doi:10.1007/s00253-011-3783-4


Few developments in microbiological diagnostics have had such a rapid impact on species level identification of microorganisms as matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF MS). Conventional differentiation methods rely on biochemical criteria and require additional pre-testing and lengthy incubation procedures. In comparison, MALDI-TOF MS can identify bacteria and yeast within minutes directly from colonies grown on culture plates. This radically new, methodically simple approach profoundly reduces the cost of consumables and time spent on diagnostics. The reliability and accuracy of the method have been demonstrated in numerous studies and different systems are already commercially available. Novel applications of the system besides microbial species level identification are also being explored. This includes identification of pathogens from positive blood cultures or directly from patient samples, such as urine. Currently, intriguing MALDI-TOF MS developments are being made regarding the phenotypic detection of certain antibiotic resistance mechanisms, e.g., β-lactamases and carbapenemases. This mini review provides an overview of the literature in the field and also includes our own data and experiences gathered from over 4 years of routine MALDI-TOF MS use in a university hospital’s microbiological diagnostics facility.


MALDI-TOFMicrobiological diagnosticPathogen identificationBlood culturesAntibiotic resistanceStrain typing


Mass spectrometry is the analytic technique used to analyze the mass to charge ratio of various compounds. Different techniques based on various ionization and detection systems have been developed. The most widely used method to date for the analysis of biomolecules is matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF MS). It is based on the ionization of co-crystallized sample material by short laser pulses. The ions are accelerated and their time of flight is measured in a vacuum flight tube. MALDI-TOF MS has successfully been used in research to determine the mass of proteins and peptides in addition to identifying previously unknown proteins (Marvin et al. 2003). MALDI-TOF MS contributed to the diagnosis of tumors, rheumatoid arthritis, Alzheimer’s disease, and allergies through the identification of specific biochemical markers (Marvin et al. 2003).

The first attempts to identify microorganisms using mass spectrometry were performed as early as 1975 (Anhalt and Fenselau 1975). However, these experiments suffered from irreproducible results due to the variabilities caused by growth conditions and media. Only with the discovery of MALDI-TOF MS in the 1980s the analysis of relatively large biomolecules, including larger ribosomal proteins, became possible (Hillenkamp and Karas 1990). The latter are less influenced by culture conditions allowing MALDI-TOF MS to be consistently used to differentiate bacterial species (Claydon et al. 1996; Holland et al. 1996; Demirev et al. 1999; Fenselau and Demirev 2001; Krishnamurthy and Ross 1996; Krishnamurthy et al. 1996).

In recent years, MALDI-TOF MS has been implemented in routine laboratories and utilized as a completely new approach for the identification of bacteria and yeast (Fig. 1).
Fig. 1

Principle of MALDI-TOF MS identification of bacteria and yeast in schematic diagram. Laser impact causes thermal desorption of (ribosomal) proteins of bacteria/yeast embedded in matrix material and applied to the target plate (analytes shown as red, light blue, and orange spheres, the matrix is given as green spheres). In an electric field, ions are accelerated according to their mass and electric charge. The drift path allows further separation and leads to measurable differences in time of flight of the desorbed particles that are detected on top of the vacuum tube. From the time of flight, the exact mass of the polypeptides can be calculated

Procedural overview

Since MALDI-TOF MS is a very sensitive technique, only a small amount of microbial biomass is required for analysis (for bacteria 104 to 106 CFU). To identify a microorganism, the sample is mixed with 1 μL of matrix solution and placed on the steel surface of the target plate to dry. The matrix solution (cinnamic acid or a benzoic acid derivate) co-crystallizes with the sample on the target plate (Fig. 2). A typical target plate can hold between 16 and 384 samples. The loaded target plate is inserted into the machine where it is then transported to the measuring chamber. Within the mass spectrometer, a high vacuum has to be continuously maintained. However, upon insertion of the loaded target plate, air is introduced into the system and the vacuum must be reestablished before sample analysis can be performed. Once a sufficient vacuum has been created, the individual samples are exposed to short laser pulses. The laser’s energy vaporizes the microorganism together with the matrix, leading to ionization of the (ribosomal) proteins. An electromagnetic field, created by a potential of about 20 kV, accelerates the ions before they enter the flight tube. The time of flight (TOF) of the analytes to reach the detector at the end of the flight tube is precisely measured. The degree of ionization as well as the mass of the proteins determines their individual TOF. Based on this TOF information, a characteristic spectrum is recorded and constitutes a specific sample fingerprint, which is unique for a given species.
Fig. 2

Subcultured colonies are applied on a clean target (a). Samples are overlaid with matrix solution and air dried (b). Measurement in the MALDI-TOF MS is conducted (c); an analyte-specific mass spectrum is obtained (d); analyte identification through automatic matching of the generated mass spectrum with spectra in the database (e)

For species level identification, the size range generally used is between 2 and 20 kDa as it was found to be very stable and with a strong signal to noise ratio. Interestingly, this size range is dominated by ribosomal proteins which ionize well, provide accurate spectra, and are only minimally influenced by microbial growth conditions (Fig. 3). The computer software automatically compares the collected spectra with a reference databank containing a wide variety of medically relevant isolates (Fig. 4). The measured spectra are subject to method-inherent noise and therefore, will never be exactly identical for an individual isolate. The software which compares the spectra generates a numerical value (score value) based on the similarities between the observed and stored data sets (Fig. 4). This score value provides information about the validity of the identification. A score value above 2.0 is generally considered to be a valid species level identification. Values between 2.0 and 1.7 represent reliable genus level identifications. Furthermore, the software displays additional results next to the best match for plausibility checks (Fig. 4). Current algorithms allow the entire computational analysis to be performed in near real time (Jarman et al. 2000; Sauer et al. 2008). Therefore, if only one sample is to be measured, it can be processed in 5–7 min and provides a species level identification. If a target plate containing 96 isolates is used, results can be obtained in about 1 h starting from the time-point the first sample is loaded on the plate.
Fig. 3

MALDI-TOF mass spectrum of Enterococcus faecium. The measured range of 3,000 to 11,000 Da is displayed. The characteristic mass peaks are predominantly ribosomal proteins. Subsequently, the integrated MALDI software matches the pattern with entries of a database
Fig. 4

Computer display of identification results after automatic comparison of the generated spectrum with the MALDI-TOF database. The ten best matching entries are shown in a tabular form. The degree of similarity to the reference spectrum is represented by a score value. Identification results with score values above 2.0 are considered to be correct for determination of the respective species

After the analysis process in the MALDI-TOF MS, the used target plate is removed from the machine. Disposable target plates can be discarded in the regular laboratory waste while reusable versions are cleansed for further use. Most recommended cleaning procedures start with treating the plate with ethanol and TCA solutions and include mechanical cleansing steps. In our hands, a quick cleaning protocol with 5 min of 70% ethanol and subsequent mechanical cleaning with detergent and cloth is sufficient for regular workup and produces significantly less chemical waste (own validated protocol). Further, the more aggressive cleaning protocols can be performed on a weekly basis (own validated protocol).

Use of MALDI-TOF MS in microbiological diagnostics—identification and differentiation of bacterial and fungal isolates grown on agar plates

The main goal of microbiological diagnostics is to identify the causative agents of infectious diseases such as bacteria, fungi, and parasites. Classically, microbiological diagnostics is based on microscopy of specimens and culturing on media, which may also include antimicrobial susceptibility testing. For more than 50 years, differentiation of bacteria and fungi has mainly been accomplished through microscopy and analyzing metabolic traits using biochemical reaction profiles. Studies have evaluated the potential of MALDI-TOF MS to perform these identification procedures. It could be demonstrated that MALDI-TOF MS and conventional biochemical differentiation methods are both highly accurate for the identification of isolated subcultured bacteria and yeast.

A retrospective study of 1,116 clinical isolates comparing MALDI-TOF MS with conventional biochemical testing systems showed correct species identification of bacteria in 95.2% of the cases (Eigner et al. 2009). In a similar prospective study including 1,660 bacterial isolates from 109 different species, 84.1% were correctly identified by MALDI-TOF MS at the species level (Seng et al. 2009). At the time, missing database entries were at fault for the majority of unsuccessfully identified samples. Our validation studies with the Bruker Daltonik (Bremen, Germany) MALDI-TOF MS system have also confirmed the validity of the system. The mass spectrometry results have been compared with those achieved by biochemical identification systems (Phoenix, Becton Dickinson, Heidelberg, Germany; and API, bioMérieux, Nürtingen, Germany). In the case of discrepancies, 16S-rDNA/28S-rDNA sequencing was performed (unpublished data). Out of 1,200 prospectively analyzed isolates, 93.5% of the identifications were in accordance with the results achieved through biochemical methods. The isolates included 370 Enterobacteriaceae, 300 nonfermenters, 110 other Gram-negatives, 320 Gram-positives, and 100 yeast. When 16S-rDNA sequences were used as the gold standard, biochemical differentiation was superior in 3% of the cases whereas MALDI-TOF MS provided a more accurate identification result in 3.5% (unpublished data). In many cases where the MALDI-TOF MS yielded no result, the relevant species information was absent from the database. By now, most of the clinically relevant organisms are included in the spectra database and have nearly closed the diagnostic gap.

With the corresponding additions to the reference database, the identification rate was almost 100% for Neisseria (Ilina et al. 2009; Ilina et al. 2008), Clostridia (Grosse-Herrenthey et al. 2008), Mycobacteria (Pignone et al. 2006), Salmonella (Dieckmann et al. 2008), viridans group streptococci (Friedrichs et al. 2007), Helicobacter pylori (Ilina et al. 2010), and Campylobacter (Martiny et al. 2011). The results were very similar when comparing the MALDI-TOF MS system with biochemical methods to identify Enterobacteriaceae, staphylococci, and streptococci (Holler et al. 2011). The MALDI-TOF MS showed advantages in the identification of Gram-positive rods (Barbuddhe et al. 2008), anaerobes, and some nonfermenters (Mellmann et al. 2008; Vanlaere et al. 2008).

MALDI-TOF MS has also been used successfully by several groups to differentiate yeast and fungi. Marklein et al. proved that 96% out of 250 clinical Candida isolates from 15 different species could be correctly identified by MALDI-TOF MS (Marklein et al. 2009). Two prospective studies looking at the identification of yeast showed comparably high results of identification (Bizzini et al. 2010; van Veen et al. 2010). Far less data are available for the differentiation of molds like Aspergillus sp., Penicillium sp., Fusarium sp., and dermatophytes (Amiri-Eliasi and Fenselau 2001;Fenselau and Demirev 2001; Erhard et al. 2008; Hettick et al. 2008b; Hettick et al. 2008a; Marinach-Patrice et al. 2009; Putignani et al. 2011). As MALDI-TOF MS identification can only be performed from cultured fungi, the various growth forms of molds, such as mycelium and conidia, complicate the analysis due to differences in protein composition. Adjustments and optimizations are needed to enhance the performance of MALDI-TOF MS-based identification systems for routine diagnostics of molds. Furthermore, the cultivation of various types of fungi, such as dermatophytes is relatively time consuming and no data exist regarding the potential benefits of MALDI-TOF MS in this setting. In many cases, there is no clinical relevance for exact identification (e.g., athlete’s foot) and therefore many routine diagnostic laboratories do not even perform dermatophyte differentiation.

Currently used MALDI-TOF MS techniques are nearly independent of culture conditions. Selective media such as McConkey and XLD Agar can also be used in addition to standard media formulations such as Columbia and Chocolate Agar. Since small amounts of culture material are enough for successful analysis, it is often possible to use a colony from a culture or subculture only a few hours after inoculation. This is why it is possible to obtain a pure culture identification from a mixed culture within the same day. Besides culturing the organism, sample preparation is an important factor contributing to analysis quality. In most cases, the MALDI-TOF MS analysis can be performed with little sample preparation other than streaking a colony on the target plate. Directly streaked colonies used for analysis should be as fresh as possible (not more than 48 h) because with increasing cultivation time, weaker and less distinguished peaks will appear in the spectra (unpublished observation). This effect is probably due to ribosomal protein degradation and leads to less efficient species identification.

In some instances, a sample may have a strong cell wall (e.g., yeast) and may require a short extraction procedure to render the ribosomal proteins available for analysis and to provide a more defined profile for database matching and differentiation with higher confidence. For this, the colony to be analyzed is suspended in an 80% ethanol solution, centrifuged and resuspended in an acetonitrile and formic acid solution. The resulting supernatant is again centrifuged and placed on the target plate for analysis.

An efficient “quick extraction protocol” can be performed directly on the target plate by adding 1 μl of neat formic acid on the dried crude sample spot (Haigh et al. 2011).

Direct identification of bacteria and yeast from positive blood cultures using MALDI-TOF MS

Early on, attempts were made to use the MALDI-TOF MS for differentiation of pathogens directly from positive blood cultures (La Scola and Raoult 2009). The problem with this application is that a high concentration of host proteins from the patient’s blood sample interferes with the detection of specific bacterial and fungal proteins. As a result, several research groups have developed and studied preparation protocols that separate bacterial and host proteins through cell lysis or differential centrifugation steps (Christner et al. 2010; Ferroni et al. 2010; Stevenson et al. 2010; Schubert et al. 2011). In 75–95% of tests performed, correct species level identification could be achieved (Prod’hom et al. 2010; Ferreira et al. 2011).

For use in a routine diagnostics laboratory, protocols need to be accurate and concise. Many of the published procedures are time consuming and difficult to incorporate into the laboratory workflow. Therefore, cell lysis methods are generally favored due to their simplicity and effectivity (Fig. 5). Such protocols allow about 86% of positive blood cultures to be successfully analyzed within 30 min. To obtain such high levels of identification, modifications to the software thresholds have been implemented (Stevenson et al. 2010; Schubert et al. 2011; Moussaoui et al. 2010). For example, score values lower than 2.0 are considered valid if the first three results in the list presented for each sample analysis refer to the same species. Secondly, the protein mass range used for homology search in the database is changed to remove unstable spectra caused by contaminations introduced by the patient’s blood and the blood culture system. These modifications do not abrogate the inherent limitations of the system, such as difficulties distinguishing between Streptococcus pneumoniae and Streptococcus mitis/oralis.
Fig. 5

Workflow of sample preparation for identification of bacteria and yeasts directly from positive blood cultures. 1 ml of blood culture liquid is filled into an Eppendorf tube (a), lysis buffer is added (b), sample is centrifuged for 1 min at 17,900g and subsequently supernatant is discarded (c), washing buffer is added to the pellet and mixed carefully (d), centrifugation step is repeated (e)

MALDI-TOF MS enables the identification of pathogenic bacteria and yeast directly from patients’ urine samples

Analysis of pure isolated cultures from agarose plates is routinely performed with simple protocols. Identification of germs from a blood culture system requires extraction and enrichment protocols. The most challenging task, however, is to directly analyze patient samples without a cultural enrichment step. Such sample material is most often rich in host proteins and normal flora, which both overlay a pathogen’s mass spectrum. Urine is a good candidate as it does not contain normal flora and there are almost no host proteins in the sample. At the same time, urinary tract infections are normally monomicrobial and exhibit high concentrations of pathogens during the course of infection (Ferreira et al. 2010; Ferreira et al. 2011).

To prepare the samples for analysis, they are slowly centrifuged to pellet out the leukocytes. Afterwards, a fast centrifugation step will pellet the pathogen. It is then washed with distilled water to remove residual contaminants. These refined samples can be directly analyzed with a MALDI-TOF MS. In a large study using this protocol, it was possible to correctly identify pathogens at the species level directly from urine samples at rates of 91.8% (significant bacterial load, >105 CFU/mL) (Ferreira et al. 2010). Thus, MALDI-TOF MS presents itself as a true alternative to more lengthy and expensive culture-based microbial identification systems, especially in the outpatient setting where resistance testing is not required.

MALDI-TOF MS in microbiological diagnostics—advantages, opportunities, and limitations

MALDI-TOF MS is increasingly used for microbiological diagnostics and has already replaced conventional biochemical differentiation methods in some laboratories. Minor discrepancies between biochemical, molecular, and MALDI-TOF MS-based differentiation results have been observed. Differences are seen because MALDI-TOF MS pathogen identification is based on the analysis of ribosomal protein spectra. Such differentiation results therefore are closely related to the results of the 16 s-rDNA sequence database comparisons. Consequently, species which do not differ sufficiently in their ribosomal protein sequences, such as Shigella spp. and Escherichia coli or S. pneumoniae (pneumococcus) and members of the S. oralis/mitis group, cannot be distinguished by MALDI-TOF MS. This is where classical biochemical tests, antigen detection, or molecular methods are required. On the other hand, MALDI-TOF MS reveals the shortcomings of the conventional metabolic testing methods. The latter are very vulnerable to the gain and loss of function in metabolic pathways and can therefore misclassify isolates. In contrast, MALDI-TOF MS does not depend on metabolic reactions. Further, MALDI-TOF MS can perform species level identification without the need for predifferentiation. A single MALDI-TOF MS system can be used for gram-positive bacteria, gram-negative bacteria, and yeast, which is not the case with biochemical differentiation methods.

Fast pathogen identification is critical in clinical diagnostics. With MALDI-TOF MS, species identifications can be obtained from a colony within a few minutes, so valuable differentiation results can be communicated to the responsible physician right after the cultures have been reviewed for the first time, 12–24 h after receiving the sample. Further, sample preparation can be automated allowing high sample throughput analysis of clinical samples (Kiestra™, Wasp™).

Another advantage of MALDI-TOF-MS is that the reference spectra database can be amended and edited by either commercially available software updates or by internal laboratory personnel. An open source platform can be established as a way for users to exchange spectra of isolates to increase their own reference databases. However, the quality of the entries must be controlled to ensure that no wrong data are distributed. The flexibility of the system is attractive to users in various fields because it provides a high level of customized features, pertinent to their respective interests.

As for any diagnostic system, there are limitations to MALDI-TOF MS. To date, in nearly all cases a preculture is required for successful analysis of most patient samples. Further, MALDI-TOF MS has only limited ability to detect bacterial resistance mechanisms.

Costs are an important factor when considering integrating new systems into a laboratory. A typical MALDI-TOF MS system used in routine diagnostics currently costs in the range of US $180,000–200,000 including analysis equipment, computers, relevant software, and integrated databases. These initial investments are offset by low overall operating costs due to high sample throughput analysis and the low amount of consumables required. Expenditures per analysis can be as low or even lower than 50 cents. In the near future, it is expected that technological improvements in semiconductor lasers will allow for longer lifespan and faster pulse frequencies, further reducing running costs while improving speed and efficiency. The introduction of automatic heat cleaning systems for the MALDI-TOF MS can extend the intervals between routine cleaning, decreasing maintenance costs and downtime. In the study by Seng et al. (Seng et al. 2009), biochemical testing was found to be four times more expensive than MALDI-TOF MS analysis. In this calculation, the authors included expenses for consumables, salaries of employees, and equipment depreciation over a 5-year period. Similar savings were identified in a study by Cherkaoui et al. (Cherkaoui et al. 2010). However, each laboratory must carefully calculate the cost reduction for their specific situation, as not all study results can easily be transferred.

Future applications of MALDI-TOF for microbiological diagnostics

Currently, further applications of MALDI-TOF MS are being developed for clinical microbiological diagnostics. One major area of interest is detecting antimicrobial resistance mechanisms. One of the organisms in focus is methicillin-resistant Staphylococcus aureus (MRSA). Studies were performed to detect the penicillin-binding protein PBP2a in MRSA using MALDI-TOF MS to rapidly confirm the methicillin-resistant phenotype. Some trials found differences in protein patterns between methicillin-susceptible (MSSA) and resistant (MRSA) S. aureus strains. However, the detected differences were not due to the resistance determinant PBP2a, but due to the clonality of the S. aureus isolates (Wolters et al. 2011; Bernardo et al. 2002). This means typical MRSA clones were separated from typical MSSA clones. As a consequence, isolates of a “typical” MRSA clone can be misidentified as resistant to methicillin although they have lost a functional resistance determinant. Conversely, susceptible clones which have newly acquired the resistance genes can also be misidentified as susceptible (Wolters et al. 2011; Jackson et al. 2005; Friedrichs et al. 2007; Majcherczyk et al. 2006). Given this, it is not surprising that conflicting data have been published (Bernardo et al. 2002; Du et al. 2002).

Another important diagnostic challenge is the detection of β-lactamase production in Enterobacteriaceae, the main mechanism for bacterial resistance to carbapenem and β-lactam antibiotics. A promising approach for identification of β-lactam-resistant strains using MALDI-TOF MS is the detection of cleavage products of the respective compounds (Fig. 6). These products are produced by hydrolysis of β-lactam rings by the bacterial enzymes. In case of the β-lactam antibiotic ampicillin, the molecule will be decarboxylated after the addition of a H2O. These chemical modifications change the mass of the antibiotic compound which can be detected in the mass spectrometer and prove the presence of a β-lactamase. To produce enough cleavage products for successful detection, the bacterial strain needs to be co-incubated with the antibiotic for a period of 1–3 h. Using this method, resistance to various penicillins, cephalosporins, and carbapenems could be successfully detected (Burckhardt and Zimmermann 2011; Hrabak et al. 2011; Sparbier et al. 2012).
Fig. 6

Use of MALDI-TOF MS to detect β-lactamase activity in E. coli isolates after coincubation with ampicillin for 2.5 h. The mass peak of ampicillin is detectable only in the supernatant of the β-lactamase negative control strain DH5a (a). In contrast, the supernatant of the ESBL-positive E. coli strain reveals the mass peaks of the hydrolyzed ampicillin (b) with a mass of plus 18 Da according to the addition of H2O. The further decarboxylated product is visible as a 44-Da lower mass peak (c). Modified according to Sparbier et al., Poster no. 65 at the MSACL conference, San Diego, 2011

A further application of MALDI-TOF MS is the differentiation of isolates below the species level. This allows epidemiological analysis of patient and environmental samples. The idea is to use discrete differences between the protein spectra for individual strain typing and to save the time and costs for molecular analysis such as PCR and restriction fragment length polymorphism analysis. Recent studies on Salmonellae, Francisella tularensis, Bacteroides fragilis, and Streptococcus agalactiae demonstrated the potential of MALDI-TOF MS as a tool to differentiate bacteria on a subspecies level (Dieckmann et al. 2008; Dieckmann and Malorny 2011; Lartigue et al. 2009; Seibold et al. 2010; Nagy et al. 2011). Other examples for subspecies typing include the accurate identification of genomic species from the Acinetobacter baumannii group and the rapid subtyping of Yersinia enterocolitica isolates (Stephan et al. 2011; Espinal et al. 2011).

One of the current limitations of the MALDI-TOF system is the inability to detect pathogens directly from patient material with the exception of urine (see above). Various attempts are made to directly analyze patient material from other sources, such as cerebrospinal fluid and blood. However, no working protocols have been published to date.

An increasing number of laboratories with high sample throughput are pushing towards full laboratory automation. This concept includes automatic inoculation of patient material on growth media as well as automated cultivation, growth detection, and digital photography of the sample cultures. The next goal is to have colonies automatically picked and differentiated to the species level. Systems incorporating an automatically loaded MALDI-TOF MS for microbial identification are in the promising stages of development and are expected to be fully operational in the near future (WASP™, Kiestra™). MALDI-TOF MS has only recently been introduced into the field of microbiological diagnostics and many innovations and new developments point to a promising future for its use in the coming years.


The work to refine the diagnostic MALDI-TOF MS technology in our lab was supported by Bayerische Forschungsstiftung (FORPROTECT—Protection against infectious diseases, new diagnostic, and therapeutic approaches).

Copyright information

© Springer-Verlag 2011