Bruker Daltonics

MALDI-TOF MS is a high throughput and reliable method for the classification and identification of microorganisms (Holland et al. 1996; van Baar 2000; Fenselau and Demirev 2001). Applications are feasible in clinical routine moni-toring, environmental and taxonomical research, or food processing quality control. The robust method requires minimal sample preparation efforts and life cycle costs. The Ready-to-Go Database together with easy walk-up analysis software allows unambiguous identification of thousands of bacteria, yeast and fungi. Moreover, the reference library is open, expandable and continuously growing.

Summary

The general workflow of microorganism profiling is a straightforward approach (Fig.1). Starting from a single colony or other biological material, sample deposition and preparation with MALDI matrix is performed within a few minutes. After sample drying and loading the instrument spectra acquisition is completed rapidly. Data transfer to dedicated software package for identification or classifica-tion is seamless and quick. In contrast to PCR or metabolic pathway analysis MALDI-TOF MS needs no initial assess-ment like gram staining, oxidase test of unknown samples or choice of PCR primers. Each class of sample material is treated in the same manner. The technology is an excellent alternative to classical microbiological identification and classification techniques.

Application Note # MT-80

Microorganism Identification and Classification Based on MALDI-TOF MS Fingerprinting with MALDI Biotyper

Sample preparation

The most straightforward analysis starts by apply-ing a small amount of biological material directly on a MALDI target plate in a thin film (also referred as “smear preparation”). The starting material can be a single colony or a centrifuged portion of a liquid cul-ture. Normally, “visible” amounts of biological mate-rial are prepared as smear, representing millions of cells roughly, although a minimum of 105 cells can be measured accurately on a standard MALDI steel target plate. The thin microbial film is overlaid with 2 μl of a saturated alpha-cyano-4-hydroxycinnamic acid matrix solution (50% acetonitrile, 2.5% triflouroacetic acid) and air dried.

For critical samples a short extraction protocol is recommended as cell wall structures or cell wall components can hamper unambiguous identification of smear preparations. Again visible amounts of cells are suspended in 70 % ethanol and centrifuged briefly. After discharging of supernatant the cell pellet is re-suspended in 30 μl of 70% formic acid by vigorous mixing followed by adding 30 μl of 100% acetonitrile and additional vigorous mixing. The amounts of formic acid and acetonitrile added may be lowered if less material is available. In a second centrifugation step (approx. 2 minutes at maximum speed, standard cen-

Unknown“single colony”

Acquisition of MALDI- TOF MS spectra

Data evaluation

Identification/Classification

Fig. 1: General workflow for microorganism identi-fication/classification

Ps. mendocina DSM50017 |Ps. oleovorans B396 |Ps. fluoreszens B340 |Ps. veronii DSM11331 |Ps. putida DSM291 |Ps. putida B401 |

Bru

ker

Dal

toni

cs is

con

tinua

lly im

prov

ing

its p

rodu

cts

and

rese

rves

the

rig

ht

to c

hang

e sp

ecifi

catio

ns w

ithou

t no

tice.

© B

ruke

r D

alto

nics

08

-20

08,

MT-

80/

3

trifuge) 1 - 2 microliters of the almost clear supernatant or extract is placed on a MALDI standard target, air dried and overlaid with the same matrix solution mentioned before. Mass spectrometry compatible tubes and solvents (gradi-ent grade or comparable) are recommended to obtain best results.

In addition, the extraction protocol inactivates a broad range of microorganisms, however not spore forming and/or some highly pathogenic bacteria, although different protocols for sample handling are available (like 80% TFA treatment or heat inactivation). All usually relevant safety rules for han-dling of biological material are still valid and shall be applied to the best of ones knowledge.

MALDI-TOF measurement

Mass spectra are acquired between 2000 and 20 000 Da automatically using a FLEX series MALDI-TOF mass spectrometer working in linear positive mode and controlled via dedicated, custom-made software for ease of use. The automation workflow of the MALDI Biotyper enables opti-mal sample acquisition (accumulation of typically 300 to 500 shots of high quality from different optimal spot positions), raw data processing and final identification in a few simple steps.

Reproducibility

Method robustness was monitored over a broad range of conditions. HCCA matrix was decided as the optimum choice for this application. Alternative matrices like DHB (2,5-Dihydroxybenzoic acid), FA (ferulic acid), SA (sinapinic acid), and DHAP (2,6-dihydroxyacetophenone) resulted in

similar peak pattern but exhibited disadvantages in regard to homogeneity of preparation, automation of measurement, reproducibility, and peak number.

Different compositions of growth media have little effect in the peak pattern distribution; especially in the range of 4000 to 12000 Da nearly no influence of culture medium is observed. The presence of culture medium adhering to the colonies has no effect and no additional signals appear. The growth state of the cells has little effect to the peak pattern as well. Cells in the lag phase of growth show a very similar pattern as cells in log, stationary or early death phase. Furthermore, given that sample preparation and measure-ment are performed under standardized conditions, the acquired profile spectra are highly comparable between different MALDI-TOF instruments. Spectra from the same sample target measured in three different instruments (Bruker Daltonics microflexTM, autoflexTM and ultraflexTM) are virtually identical. Therefore spectra from different MALDI-TOF MS instruments can be used to build a substantial and reliable database.

The remarkable reproducibility of the methodology is based on the measurement of constantly expressed high-abun-dant proteins, i.e. ribosomal proteins (Ryzhov and Fenselau 2001). The observed mass range of spectra is between 2000 and 20000 Da. Analysis in this range contains very few measurable metabolites as Fig. 2 illustrates. The ribosomal proteins, were identified by searching the expasy database (http://au.expasy.org/srs5bin/cgi-bin/wgetz). Also for other microorganisms, where genomic sequence information is available, profile peaks can be assigned to ribosomal proteins.

E. coli proteomics fingerprint

Fig. 2: Spectrum of E. coli and annotated masses of ribosomal proteins.

6254

The MALDI Biotyper software

The analysis software incorporates functionalities for the processing of mass spectra as well as for identification and classification. The processing parameters for smoothing, data reduction or baseline correction are user definable and the result is a dedicated peak list.

Pattern matching, for identification of unknown microorgan-isms, is accomplished through comparison of the generated peak lists with a library of stored spectra which contain the characteristic spectra information of species and subspe-cies. The library spectra are generated by measurement of known bacterial species and strains at different days under slightly different conditions and extracting the specific peak information. A good average is achieved by the measure-ment of 20 spectra. The software automatically generates the peak lists from the whole set of spectra and extracts the typical peaks which are present in a certain number of spectra from one species.

Additionally; the fully automated main spectra generation which starts with raw spectra allows the user to manually input information before calculation of main spectra. This enables careful spectra processing and peak list editing before the main spectra are generated for a database. Identification of unknown microorganisms is performed by comparing their individual peak lists to the database. A matching score based on identified masses and their intensity correlation is generated and used for ranking of the results (Fig. 3).

To increase the confidence of database searches the MALDI Biotyper is able to correct peak mass deviations

through a sophisticated recalibration algorithm. After peak picking it is possible to set an initially accepted error window and a desired adjustment result. The software can adapt the calibration from a new peak list to the known peak list in adjustable borders. This functionality makes the MALDI Biotyper identification exceptionally robust and accurate.

For dereplication, clustering and generation of family trees the MALDI Biotyper offers a variety of functionalities. First, it is possible to calculate the similarity of all main spectra in a library with each of the other spectra in this set by pat-tern matching. Based on the similarity scores dendrograms can be constructed. Second, an unsupervised multivariate analysis based on principle component analysis is possible for a set of spectra. Based on the calculated principle com-ponents a variety of clustering algorithms and visualizations are available. Fig. 4 and 5 show examples of dendrogram building and clustering in the genus Pseudomonas.

Identification by peak-matching

Fig. 3: Typical identification result. Green marks the perfect matching peaks, yellow marks the peaks matching in broader boar-ders, red peaks did not match within the reference spectrum. Lower table gives an impression of the output of the evaluated and ranked results.

Family tree generation

Fig. 4: Dendrogram based on identification score values.

Applications

MALDI-TOF MS bacteria identification offers an excellent alternative to traditional laboratory methods in a variety of domains, e.g. environmental research, food and water con-trol, quality control of microorganism collections, microbial identification in veterinary medicine and clinical microbiol-ogy (for research use only). Prerequisite is the establish-ment of high quality spectra libraries for the particular area. The speed and minimal costs of sample preparation and measurement for this method makes it exceptionally well suited for routine and high-throughput use. A further field of application is the analysis of taxonomic relationships. MALDI-TOF MS profile analysis results in similar family trees compared to classical methods like morphological and biochemical comparison or 16S ribosomal DNA sequenc-ing. Because ribosomal proteins are highly abundant and appear very stable the observed protein pattern enables a direct view to the translated DNA sequence. Therefore the method comprising the determination of the molecular mass can be assumed to have a value as a rough multi-locus sequencing. Dereplication of complex microorgan-ism communities based on their mass pattern offers new scientific capabilities, e.g. in environmental research and

Keywords:

Microorganism ID

Taxonomy

Profiling

Fingerprinting

Instrumentation & Software

microflex

autoflex

ultraflex

MALDI Biotyper

Multivariate analysis of pseudomonas species

Fig. 5: 3D scatter plot of the first three principle component scores for six individual species of the genus Pseudomonas, Pseudomo-nas veronii, P. fluorescens, P. mendocina, P. oleovorans, P. putida strain B401, P. putida strain DSM291

-4 -2

0 2

-4 -2

0 2

4

-4

-3

-2

-1

0

1

P C 1

3D scatter plot

P C 2

P C 3

investigations of biodiversity. Thousands of microorganisms can easily be analyzed and serve as the basis for further analysis. Several spectra measured from each unknown isolate can be used for construction of main spectra. An initial set of e.g. 50 microorganism’s main spectra may be used for comparing with each other. Spectra of several fold existing isolates show high scores, thereby redundancies can be removed. Further isolates then can be compared with the initial reference data set and either discarded or added to the library creating an overview about the diversity of microorganisms.

Conclusion

MALDI Biotyper integrates a ready-to-use reference library comprising thousands of individual strains and is grow-ing continuously. The system is open and can be easily expanded by specific requests of the user including data for ID of bacteria, yeasts and fungi.

Literature

Fenselau, C. and P. A. Demirev (2001). “Characterization of intact microorganisms by MALDI mass spectrometry.” Mass Spectrom Rev 20(4): 157-71.

Holland RD et al. (1996). “Rapid identification of intact whole bacteria based on spectral patterns using matrix-assisted laser desorption/ionization with time-of-flight mass spectrometry.” Rapid Commun Mass Spectrom 10(10): 1227-32.

Naumann, D., P. Lasch, et al. (2005). MALDI-TOF Mass Spectros-copy - a Versatile Tool in Microbiology. talk at Statens Serum Institut.

Ryzhov, V. and C. Fenselau (2001). “Characterization of the protein subset desorbed by MALDI from whole bacterial cells.” Anal Chem 73(4): 746-50.

Van Baar, B. L. (2000). “Characterisation of bacteria by matrix-assisted laser desorption/ionisation and electrospray mass spec-trometry.” FEMS Microbiol Rev 24(2): 193-219.

Authors

Thomas Maier, Gerold Schwarz & Markus Kostrzewa Bruker Daltonik GmbH

Bruker Daltonik GmbH

Bremen · GermanyPhone +49 (421) 2205-0 Fax +49 (421) 2205-103 sales@bdal.de

Bruker Daltonics Inc.

Billerica, MA · USAPhone +1 (978) 663-3660 Fax +1 (978) 667-5993 ms-sales@bdal.com

www.bdal.com

Bru

ker

Dal

toni

cs is

con

tinua

lly im

prov

ing

its p

rodu

cts

and

rese

rves

the

rig

ht

to c

hang

e sp

ecifi

catio

ns w

ithou

t no

tice.

© B

ruke

r D

alto

nics

08

-20

08,

MT-

80/

3

For research use only. Not for use in diagnostic procedures.