Abstract
The present study investigates the suitability of direct bacterial profiling as a tool for the identification and subtyping of pathogenic Neisseria. The genus Neisseria includes two human pathogens, Neisseria meningitidis and Neisseria gonorrhoeae, as well as several nonpathogenic Neisseria species. Here, a matrix-assisted laser desorption/ionization time-of-flight mass spectrometry profiling protocol was optimized using a laboratory strain of E. coli DH5α to guarantee high quality and reproducible results. Subsequently, mass spectra for both laboratory and clinical strains of N. gonorrhoeae, N. meningitidis, and several nonpathogenic Neisseria species were collected. Significant interspecies differences but little intraspecies diversity were revealed by means of a visual inspection and bioinformatics examination using the MALDI BioTyper software. Cluster analysis successfully separated mass spectra collected from three groups that corresponded to N. gonorrhoeae, N. meningitidis, and nonpathogenic Neisseria isolates. Requiring only one bacterial colony for testing and using a fast and easy measuring protocol, this approach represents a powerful tool for the rapid identification of pathogenic Neisseria and can be adopted for other microorganisms.
The prevalence of related diseases makes accurate identification of pathogenic Neisseria difficult. Gonococcal infection, caused by Neisseria gonorrhoeae, is still one of the most common bacterial sexually transmitted diseases in the world, estimated to infect more than 60 million people each year. The incidence of gonorrhea is lowest in Western Europe, but the disease remains epidemic in Eastern Europe, much of Africa, the Indian subcontinent, and parts of Asia, South America, and the United States. In Russia, more than 89,000 cases of gonorrhea were diagnosed in 2005.1 Neisseria meningitidis is a common inhabitant of the mucosal membranes of the nose and throat; infections occur sporadically throughout the world with seasonal variations and account for a proportion of endemic bacterial meningitis. Although the highest burden of disease currently exists in Africa, epidemics can occur in any part of the world. Asia has suffered from some major epidemics of meningococcal disease in the last 30 years (China 1979 and 1980, Vietnam 1977, Mongolia 1973–1974 and 1994–1995, Saudi Arabia 1987, Yemen 1988). There have also been epidemics in Europe and in America during the last 30 years, but they have not reached the very high incidence levels of epidemics in other parts of the world. In Russia, the frequency of bacterial meningitis and sepsis caused by N. meningitidis reaches 5000 per year with mortality rates from 5% to 20%.2
Conventional diagnostic methods for Neisseria species identification include series of enzyme reaction tests, which have a certain number of false negative results.3 As taxonomic differences between members of the Neisseria genus are small, the identification of these pathogens remains problematic and as such, no single method is currently recommended for a definitive identification. Prolyliminopeptidase was previously considered to be almost universally present in N. gonorrhoeae, but the widespread occurrence of Prolyliminopeptidase-negative N. gonorrhoeae strains4 potentially results in incorrect, doubtful, or delayed species identification. Additionally, N. meningitidis strains producing Prolyliminopeptidase like N. gonorrhoeae have been reported.
Traditional microbiological methods based on morphological and biochemical characteristics are time-consuming, error-prone, and laborious. Therefore, the development of new rapid and effective methods for Neisseria species identification is required. Mass spectrometry is a powerful routine tool in biological studies and presents an attractive alternative to traditional methods, in particular for bacterial identification and differentiation.
Direct bacterial profiling (DBP) by means of matrix-assisted laser desorption−ionization time-of-flight (MALDI-TOF) mass spectrometry has been suggested as an alternative approach for rapid and accurate classification and species identification of microbes,5,6,7,8,9 overcoming the existing limitations of classical biochemical methods. The most important advantages are: 1) requirement of only a small amount of biological material, 2) a fast and easy measurement protocol, and 3) high specificity in species differentiation.
Although the method has been described several years ago there is no common protocol of sample preparation for mass spectrometry measurement. Suggested procedures include “intact” cells technique (native cells are deposited directly onto the sample holder and mixed with matrix),10 different ways of protein extraction, and even the utilization of non-ionic surfactants to expand the upper mass range for bacterial proteins obtained from whole cell MALDI-TOF mass spectrometry analysis.11 The influence of cultivation condition, matrix, solvent, and spotting technique on the reproducibility and quality of mass spectra has been reported.12 Therefore, today the majority of studies in this field are intended to develop effective and universal sample preparation and measurement protocols and the optimization of analytical algorithms. Whether it will be a single protocol for all samples or several techniques for different microbial groups, the future will show. Further studies will have to prove the applicability of mass spectrometry profiling for the reliable identification of clinically relevant microorganisms in routine and diagnostic settings.
Here, we present an evaluation of DBP for the identification of pathogenic Neisseria and of its ability to distinguish between closely related species from the genus Neisseria. For this purpose, a dedicated database of mass spectra has been created using the recently presented MALDI Biotyper bioinformatics package.13 Furthermore, the intrageneric and intraspecies variability of DBP mass spectra was analyzed.
Materials and Methods
Bacterial Strains
Clinical strains of N. gonorrhoeae, N. meningitides, and non-pathogenic Neisseria (n = 29, n = 13, and n = 15, respectively) were collected. Additionally, ATCC strain 49226 of N. gonorrhoeae was included for validation of assay performance. A laboratory strain of Escherichia coli DH5α (Life Technologies, UK) was used for mass spectrometry measurement optimization. Detailed information about the strains is noted in Table 1.
Table 1.
Bacterial Strains
| Strain no. | Description | Isolation region/additional information | |
|---|---|---|---|
| E. coli | |||
| 1 | Ec DH5alpha | Laboratory strain DH5α | |
| Neisseria gonorrhoeae | |||
| Por type | |||
| 1 | Ng ATCC 49226 | Reference strain ATCC 49226 | |
| 2 | Ng a04 | PIB 26 | Archangelsk |
| 3 | Ng a032 | PIA 6 | Archangelsk |
| 4 | Ng a040 | PIA 6 | Archangelsk |
| 5 | Ng a041 | PIB 3 | Archangelsk |
| 6 | Ng ch012 | PIB 1 | Cheboksary |
| 7 | Ng e004 | PIB 2 | Yekaterinburg |
| 8 | Ng i6001 | PIB 2 | Irkutsk |
| 9 | Ng i6002 | PIB3 | Irkutsk |
| 10 | Ng i6003 | PIB 22 | Irkutsk |
| 11 | Ng i6004 | PIB22 | Irkutsk |
| 12 | Ng i6005 | PIA 2 | Irkutsk |
| 13 | Ng i6007 | PIB 3 | Irkutsk |
| 14 | Ng i6008 | PIA 6 | Irkutsk |
| 15 | Ng i6010 | PIB 3 | Irkutsk |
| 16 | Ng i6011 | PIB 22 | Irkutsk |
| 17 | Ng i6012 | PIB 22 | Irkutsk |
| 18 | Ng i6013 | PIA 10 | Irkutsk |
| 19 | Ng ka003 | PIB 2 | Kaluga |
| 20 | Ng ka005 | PIB 2 | Kaluga |
| 21 | Ng m024 | PIB 2 | Murmansk |
| 22 | Ng mr012 | PIB 3/6 | Moscow region |
| 23 | Ng n012 | PIB3 | Nizhniy Novgorod |
| 24 | Ng n020 | PIB 2 | Nizhniy Novgorod |
| 25 | Ng ps601 | PIB 3 | Pskov |
| 26 | Ng sp030 | PIB 8 | Saint-Petersburg |
| 27 | Ng sp031 | PIB 3/6 | Saint-Petersburg |
| 28 | Ng sp032 | PIB 2 | Saint-Petersburg |
| 29 | Ng sp033 | PIB 2 | Saint-Petersburg |
| 30 | Ng sp034 | PIB 3/6 | Saint-Petersburg |
| Neisseria meningitidis | |||
| Serogroup | |||
| 1 | Nm 229 | A | Moscow region |
| 2 | Nm 537 | B | Moscow region |
| 3 | Nm 549 | A | Moscow region |
| 4 | Nm 552 | A | Moscow region |
| 5 | Nm 592 | A | Moscow region |
| 6 | Nm 605 | A | Moscow region |
| 7 | Nm 266 | A | Moscow region |
| 8 | Nm 441 | B | Moscow region |
| 9 | Nm 406 | A | Moscow region |
| 10 | Nm 634 | B | Moscow region |
| 11 | Nm 639 | A | Moscow region |
| 12 | Nm 1522 | B | Moscow region |
| 13 | Nm 1539 | A | Moscow region |
| 14 | Nm 1541 | C | Moscow region |
| 15 | Nm 070309 | A | Moscow region |
| Nonpathogenic Neisseria | |||
| 1 | N. elongata 605 | Moscow region | |
| 2 | N. flavescens 533 | Moscow region | |
| 3 | N. flavescens 8552 | Moscow region | |
| 4 | N. lactamica 79 | Moscow region | |
| 5 | N. mucosa 839 | Moscow region | |
| 6 | N. mucosa 1591 | Moscow region | |
| 7 | N. perflava 1621 | Moscow region | |
| 8 | N. polysaccharea 547 | Moscow region | |
| 9 | N. sicca 70 | Moscow region | |
| 10 | N. flava 110 | Moscow region | |
| 11 | N. subflava 239 | Moscow region | |
| 12 | Neisseria spp 8503 | Moscow region | |
| 13 | Neisseria spp 8553 | Moscow region | |
| 14 | N. subflava 227 | Moscow region | |
| 15 | N. subflava 1672 | Moscow region | |
Before analysis, E. coli strain DH5α was grown on lysogeny broth (LB) agar plates at 37°C for 18 hours. N. gonorrhoeae strains were grown on the medium prepared by supplementing GC agar base (BBL-DIFCO) with 1% of defined supplement (IsoVitaleX; BBL-DIFCO) at 37°C in 5% CO2 atmosphere for 18 hours. Crystal Haemophilus/Neisseria (BBL-DIFCO) test was used for species identification. N. meningitidis strains as well as non-pathogenic Neisseria were cultured using special media “Chocolate Agar” or “Columbia III” (BBL-DIFCO) with 5% sheep blood at 37°C in 5% CO2 atmosphere for 19 to 24 hours. Biochemical species identification was performed by the automated identification system WalkAway40 (MicroScan Dade Behring) using Hemophilus and Neissena Identification (HNID) kit-panels. Serotyping of N. meningitidis strain was done by agglutination test.
Por Typing of N. gonorrhoeae Isolates
The N. gonorrhoeae typing based on analysis of por gene variability was done as described previously.14,15,16,17 The por-type of a particular isolate was determined by identification of the obtained nucleotide sequences of the por genes in the EMBL (European Molecular Biology Laboratory) and NCBI (National Center for Biotechnology Information) databases.
Bacterial Preparation for MALDI Mass Spectrometry Analysis
For MALDI-TOF mass spectrometry analysis single colonies of fresh bacterial cultures have been used. Cells were transferred from the plate into an extraction 0.5 ml tube (Eppendorf, Germany) with a 1.0 μl plastic loop (FL medical, Italy) corresponding to 5 to 10 mg cell material, approximately. Then, 50 μl of extraction solution consisting of 50% acetonitrile (Sigma-Aldrich, Germany), 2.5% trifluoroacetic acid (Sigma-Aldrich, Germany), and 47.5% water (Fluka, Switzerland) was added. The mixture was resuspended well and centrifuged (12000 rpm for 1 minute). Supernatant was used for further mass spectrometry analysis.
Saturated solutions of α-cyano-4-hydroxy cinnamic acid, sinapinic acid, or ferulic acid (all substances from Bruker Daltonics, Germany) were prepared in 1.0 ml of 50% acetonitrile, 2.5% trifluoroacetic acid, and 47.5% water.
Cocrystallization method depended on the matrix used: a) 1 μl of the bacterial cell lysate was deposited onto a sample spot of a MALDI target plate (MSP 96 target, ground steel; Bruker Daltonics, Germany) and was allowed to dry on air at room temperature. Finally, 2 μl of α-cyano-4-hydroxy cinnamic acid or ferulic acid matrix solution were deposited onto the dried sample and allowed to dry at room temperature; or b) 0.3 μl of sinapinic acid solution saturated in absolute ethanol was spotted onto a sample spot of a MALDI target plate (MSP 96 target ground steel; Bruker Daltonics, Germany) and allowed to dry at room temperature. Then, 1 μl of the bacterial cell lysate was added. Finally, 2 μl of sinapinic acid matrix were deposited onto the dried crystals and allowed to dry at room temperature. Each bacterial lysate was spotted onto the MALDI target plate in five replicas.
Preparation of Ribosomal Protein Fraction for MALDI Mass Spectrometry Analysis
Isolation and purification of ribosomal proteins were performed according to the protocol described earlier.18
Then five cycles of sonication (20 seconds at 1-minute intervals) followed by centrifugation to remove cell debris for 20 minutes were done. Briefly, ribosomes were collected by ultracentrifugation through a 10% to 30% (w/v) sucrose gradient in buffer containing 1 mmol/L HEPES (pH 7.8), 1 mmol/L MgCl2, and 200 mmol/L NH4Cl, over 18 hours at 23,000 rpm. The 30S and 50S peak fractions were collected separately pelleted by ultracentrifugation (20 hours at 45,000 rpm) and resuspended in buffer 1 mmol/L HEPES (pH 7.8), 1 mmol/L MgCl2, and 200 mmol/L NH4Cl. All procedures were conducted at 4°C. During the preparation extreme care was taken to prevent ribonuclease contamination. Therefore, only glassware pre-baked at 180°C for at least 2 hours, or sterile plastic tubes and pipette tips were used.
Purified protein solution (2 μl) was spotted onto a MALDI target plate (MSP 96 target, ground steel; Bruker Daltonics, Germany) and was allowed to air dry at room temperature. Finally, 2 μl of α-cyano-4-hydroxy cinnamic acid matrix solution were deposited onto the dried sample and allowed to dry at room temperature.
Mass Spectrometry Analysis
Mass spectra profiles were acquired using a Microflex LT MALDI-ToF mass spectrometer (Bruker Daltonics, Germany) equipped with a N2 337 nm laser. The molecular ions were measured in the linear positive ion mode with instrument parameters optimized for the range of 2000 to 20,000 m/z. The laser intensity was set slightly above the desorption and ionization threshold of the sample.
E. coli DH5α was used for calibration and instrument parameter optimization. Precisely, there were masses matched to RL36, RS22, RS34, RL32, RL30, RL35, RL29, RL31, RS21, and RS19 (see Table 2) included in the calibration list. A three-step calibration procedure was performed. First, the x-axis was calibrated using mentioned above masses of reproducible and high resolution proteins included in mass control list (internal calibration). The calibration constant of E. coli DH5α was applied to the measurement of experimental strains. Further spectra examination was performed using the FlexAnalysis 2.4 software (Bruker Daltonics, Germany).
Table 2.
Analysis of the most reproducible peaks (registration frequency equal or more than 0.7) for E. coli strain DH5α
| Peak number | M (m/z) | σ (M) | I | σ (I) | f | M (match) | Ion type | Description |
|---|---|---|---|---|---|---|---|---|
| 1 | 4365 | 1.0 | 0.50 | 0.2 | 1 | 4364* | M+H+ | RL36 |
| 4364 | M+H+ | UreG (fragment) | ||||||
| 2 | 4777 | 1.61 | 0.13 | 0.05 | 1 | 4778 | M+H+ | Q 8FKS1 |
| 4779 | M+H+ | Q 8FKD1 | ||||||
| 4779 | M+H+ | Q 1RFE4 | ||||||
| 9553 | M+2H+ | RS20 | ||||||
| 9553 | M+2H+ | RL26 | ||||||
| 3 | 4870 | 1.71 | 0.05 | 0.04 | 0.9 | 9737 | M+2H+ | AFAF |
| 9737 | M+2H+ | Q 27TL7 | ||||||
| 9736 | M+2H+ | Q 1R2B8 | ||||||
| 4 | 5097 | 1.34 | 0.82 | 0.15 | 1 | 5096* | M+H+ | RS22 |
| 5098 | M+H+ | LPPY | ||||||
| 5 | 5150 | 1.56 | 0.08 | 0.03 | 0.9 | 10300 | M+2H+ | ACYP |
| 10299 | M+2H+ | RS19 | ||||||
| 10303 | M+2H+ | YJCB | ||||||
| 6 | 5340 | 1.61 | 0.03 | 0.01 | 0.7 | |||
| 7 | 5381 | 1.38 | 0.64 | 0.18 | 1 | 5380* | M+H+ | RL34 |
| 8 | 6255 | 1.47 | 0.82 | 0.29 | 1 | 6254 | M+H+ | RL33 (methylated) |
| 9 | 6316 | 1.56 | 0.38 | 0.13 | 1 | 6315* | MH+ | RL32 |
| 10 | 6412 | 1.57 | 0.04 | 0.02 | 0.8 | 6411* | M+H+ | RL30 |
| 11 | 6509 | 1.61 | 0.20 | 0.05 | 0.7 | 6507 | M+H+ | RMF |
| 12 | 7158 | 1.81 | 0.27 | 0.12 | 1 | 7158* | M+H+ | RL35 |
| 13 | 7275 | 1.69 | 0.39 | 0.17 | 1 | 7273* | M+H+ | RL29 |
| 14 | 7872 | 1.98 | 0.07 | 0.01 | 0.9 | 7871* | M+H+ | RL31 |
| 15 | 8370 | 1.93 | 0.14 | 0.06 | 1 | 8369* | M+H+ | RS21 |
| 8371 | M+H+ | FEOA | ||||||
| 16 | 8877 | 2.03 | 0.04 | 0.02 | 0.8 | 8875 | M+H+ | RL28 |
| 17 | 8995 | 1.92 | 0.05 | 0.02 | 0.8 | 8993 | M+H+ | RL27 |
| 18 | 9064 | 1.98 | 0.17 | 0.07 | 1 | 9064 | M+H+ | KOC1 (loss of Met) |
| 19 | 9227 | 2.11 | 0.04 | 0.01 | 0.8 | 9227 | M+H+ | KEC2 |
| 20 | 9553 | 2.09 | 0.27 | 0.06 | 1 | 9553 | M+H+ | RS20 |
| 21 | 9740 | 2.31 | 0.38 | 0.21 | 1 | 9737 | M+H+ | AFAF |
| 9737 | M+H+ | Q 27TL7 | ||||||
| 9736 | M+H+ | Q 1R2B8 | ||||||
| 22 | 10300 | 2.25 | 0.08 | 0.03 | 1 | 10300 | M+H+ | ACYP |
| 10299* | M+H+ | RS19 | ||||||
| 10303 | M+H+ | YJCB |
Peaks already reported for E. colistrain K-12 [20] are marked in italics. Ribosomal proteins are indicated by bold type. M − average value of experimental m/z for each peak (mass spectra collected from 10 independent experiments have been used); σ (M) − standard deviation of M; I–relative intensity of peak; σ (I) − standard deviation of I; f − frequency of peak registration; M (match) − value of protein mass from SwissProt/TrEMBL database (Da);
proteins' masses included in calibration list; Ion type − for matched peaks only; Description − description of matched protein according to SwissProt/TrEMBL database.
Four different mass spectra from each replicate were acquired resulting in 20 spectra per sample (four spectra for each of five replicas). Five hundred individual laser shots were accumulated for each mass spectrum. Spectra quality criteria were: high spectra reproducibility, little noise and a flat baseline (signal-to-noise ratio above 15), presenting as much peaks as possible and an average resolution of 600 within the range of m/z selected for analysis. Both manual and automatic measurements were applied. The mass accuracy after external calibration using known E. coli protein signals was about 200 ppm.
Spectra processing for visual inspection was performed using the FlexAnalysis 2.4 software: peak lists were calculated using the centroid peak algorithm, with signal-to-noise ratio set at 15 and relative intensity threshold set at 10%.
For matching the peaks of interest, their m/z values were compared against the SwissProt/TrEMBL database (Expasy, Swiss Bioinformatics Institute) by manual search using the Sequence Retrieval System module at http://www.expasy.ch/srs5/ of the definite bacteria. The only parameters used in the search were ≪Organism≫ and ≪MolWeight≫ (experimental mass allowing for a mass window) corresponding to an experimental mass accuracy of 300 ppm. When no database protein was found in the mass window, the search was repeated with a mass value corresponding to N-terminal methionine loss, which is the most common protein modification in bacteria.
To construct and compare the characteristic reference spectra (main spectra) of certain strains, the mass spectra acquired (20 raw spectra per sample) were processed with the MALDI Biotyper 1.1 software (Bruker Daltonics, Germany). The main spectra were constructed based on processed raw spectra after smoothing, baseline subtraction, normalization, and peak picking. The resulting twenty individual peak lists per strain were used to construct a main spectrum, which contained the average peak position and peak intensities and their frequency in the set of spectra for the most prevalent peaks. All processing steps were performed automatically using the default settings of the software. Thereby, a main spectrum displayed the most reproducible peaks typical for a certain bacterial strain. Cluster analysis was performed based on comparison of strain specific main spectra created as described above. In a first step, each main spectrum of the dataset was compared with each other main spectrum resulting in a matrix of cross-wise identification scores. This matrix was used to calculate distance values for each pair of main spectra. Based on these values a dendrogram was generated using the according function of the statistical toolbox of Matlab 7.1 (The MathWorks Inc.) integrated in the MALDI Biotyper software. The parameter settings were: “Distance Measure = Euclidian” and “Linkage = complete.” The linkage function is normalized according to the distance between 0 (perfect match) and 1000 (no match).
Extraction of Genomic DNA, Amplification, and Sequencing of Neisseria Genes
Bacterial DNA was isolated according to Boom et al.19 N. gonorrhoeae genes rpmE, rpmH, rpmJ and N. meningitidis genes rpsT, rpmA were amplified with followed primer sets: rpmE-f 5′- TTTTATCGAATATCCTCTTGACG-3′, and rpmE-r 5′-GGCTTCGTCGCCTTGTCC-3′; rpmH-f 5′-CCGATTATACCGACTCGCCG-3′, and rpmH-r 5′-CGCGTTTCATATAGTTCCGCTC-3′; rpmJ-f 5′-ATATTTCTGGGAAGATGCGG-3′, and rpmJ-r 5′-AATTTAGCACGAGTAGCACC-3′; rpsT-f 5′-GCACCCTTTCCGCCATTTC-3′, and rpsT-r 5′-CCGTTTTTTTGCGCTGTCAG-3′; rpmA-f 5′-GGCTAAAGTAGTGGCACACG-3′, and rpmA-r 5′-TGAAAAGATGCGGCTTCAGG-3′. All primers were designed using the Oligo_6.37 software (Molecular Biology Insights Inc.) and sequences of N. gonorrhoeae strain FA 1090 (NC_002946), N. meningitidis strains FAM18, serogroup C (NC_008767), MC58, serogroup B (NC_008767), Z2491, serogroup A (NC_003116).
PCR amplification was performed in a reaction mixture containing 66 mM Tris-HCl (pH 9.0), 16.6 mmol/L (NH4)2SO4, 2.5 mM MgCl2, 250 μM of each dNTP, 1U of TaqDNA polymerase (Promega), and 10 pmol of each primer in final volume 10 μl using the thermocycler DNA Engine Tetrad 2 (MJ Research). The products of amplification were analyzed in 2% agarose gel.
Then, 0.5 U of shrimp alkaline phosphatase (Promega), and 5 U Exonuclease I (Fermentas) were added to each PCR product and incubated at 37 °C for 30 minutes for dephosphorylation of the 5′-end phosphate groups of deoxynucleoside triphosphates and primer elimination. Subsequently, the enzymes were inactivated by heating at 85 °C for 10 minutes.
Sequencing was performed using the ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kit and the ABI Prism 3100 Genetic Analyser (Applied Biosystems; Hitachi, Japan).
The amino acid sequences of N. gonorrhoeae ribosomal proteins RL31, RL34, RL36, and N. meningitidis ribosomal proteins RS20, RL27 were deduced from the nucleotide sequences of corresponding genes rpmE, rpmH, rpmJ, and rpsT, rpmA, respectively, using the Vector NTI Suite v.9.0 software (InforMax Inc., USA).
Results
In our investigation, the initial DBP experiments were started with already well-characterized bacteria such as E. coli. In the well-studied case of E. coli MALDI-TOF profiling, about 50% of the observed peaks could be assigned to ribosomal proteins, which are known to account for more than 20% of total cell protein content.20 Because ribosomal proteins are relatively basic, they can be well extracted by acidic solutions. In addition, these positively charged proteins can be measured with high sensitivity in the positive MALDI-TOF measurement mode.
Different acetonitrile and trifluoroacetic acid concentrations for preparation of whole-cell lysates and protein extraction were tested. The best results were obtained using 50% acetonitrile/2.5% trifluoroacetic acid/47.5% water solution. Three different MALDI matrices have been examined (for more details see “Material and Methods”). For further work, α-cyano-4-hydroxy cinnamic acid matrix was selected because of the best mass spectra quality (ie, resolution, signal-to-noise ratio, number of peaks, intensity, and mass range of detected proteins) and the matrix crystals homogeneity allowing fast automated measurements of high quality (data not shown).
A MALDI-TOF mass spectrum of E. coli strain DH5α whole-cell lysate is shown in Figure 1. Visual review showed that mainly the mass range up to 12000 m/z is informative and the majority of peaks within 2000 to 4000 m/z range correspond to doubly charged ions. In addition, some doubly charged ions were located within the 4000 to 5000 m/z region. However, the majority of these peaks were reproducible and of good resolution (higher than 500). Thus, further bioinformatic mass spectra analyses were performed within the 2000 to 12000 m/z range.
Figure 1.
MALDI-TOF mass spectrum of a whole-cell extract of E. coli strain DH5α. The mass-spectrum was recorded within the range of m/z 2000 to 20,000, zoomed region (m/z 2000 to 4000) illustrates peaks of doubly charged ions. The roman numerals indicate signal pairs of singly and doubly charged ions of the same molecules. The relative intensities of the ions are shown on the y axis, the mass to charge ratios are shown on the x axis.
Peak lists of 10 mass spectra collected from three different cultures of E. coli strain DH5α were generated. The most reproducible peaks (registration frequency is equal or more than 0.7) with SwissProt/TrEMBL database searching results are listed in Table 2. Maximum mean-square deviation found was 2.6 Da for m/z value 10,300 Da and 0.29 for relative intensity. The peak list obtained for E. coli DH5α was similar to the one for E. coli K-12 described earlier.21 Highly reproducible peaks (registration frequency is 1.0) were used as calibrators in further DBP tests.
The technique optimized for DBP analysis of E. coli was applied to study bacteria of the genus Neisseria, starting with N. gonorrhoeae laboratory strain ATCC 49226 and N. meningitidis clinical strain 537, whose mass spectra are shown on Figure 2A, B).
Figure 2.
MALDI-TOF mass spectra of whole-cell extracts of pathogenic Neisseria strains: N. gonorrhoeae strain ATCC 49226 (A) and N. meningitidis clinical strain 537 (B). The relative intensities of the ions are shown on the ϒ axis, the mass to charge ratios on the x axis. Changes of MALDI-TOF mass spectra observed for clinical strains of N. gonorrhoeae (C) and N. meningitidis (D). Variants of peaks corresponding to ribosomal proteins RL 36, RL 34, and RL 31 of N. gonorrhoeae and of those corresponding to ribosomal proteins RS 20 and RL27 of N. meningitidis are shown.
Spectra quality and reproducibility for N. gonorrhoeae cells were similar to E. coli. The described protocol allowed us to obtain good quality MALDI mass spectra for N. gonorrhoeae laboratory strain ATCC 49226, which are much richer in peaks, and thereby more informative, than analogous data reported previously.22 Maximum mean-square deviation was shown as 2.8 Da for m/z value 9570 and 0.24 for relative intensity. Results of accuracy and reproducibility analysis for 10 different spectra (corresponding to two different cultivations of N. gonorrhoeae strain ATCC 49226) are summarized in Table 3. Generally, the mass spectra contained more than 70 peaks, with 20 of them being the most reproducible (registration frequency equal or more than 0.7). As expected most of these peaks could be matched to ribosomal proteins, a finding that was confirmed by analysis of protein signals from a purified ribosomal fraction of the cell lysate (Figure 3).
Table 3.
Comparative Analysis of the Most Reproducible Peaks (Registration Frequency Equal or More than 0.7) for N. gonorrhoeae and N. meningitidisStrains
|
Neisseria gonorrhoeae ATCC 49226 |
Neisseria meningitidis 537 (serogroup B) |
||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Peak number | M | s (M) | I | σ (I) | f | M (match) | Ion type | D | Peak number | M | σ (M) | I | σ (I) | f | M (match) | Ion type | D |
| 1 | 4474 | 1.04 | 0.66 | 0.07 | 1 | 4473 | M+H+ | RL36 | |||||||||
| 1 | 4489 | 0.52 | 0.70 | 0.25 | 0.9 | 4489 | M+H+ | RL36 | |||||||||
| 2 | 4511 | 1.71 | 0.04 | 0.01 | 0.8 | ||||||||||||
| 3 | 4689 | 1.17 | 0.37 | 0.24 | 1 | 9377* | M+2H+ | RS20 | 2 | 4688 | 0.26 | 0.35 | 0.21 | 0.7 | 9377* | M+2H+ | RS20 |
| 4 | 4784 | 1.22 | 0.15 | 0.08 | 1 | 9570 | M+2H+ | RL27 | |||||||||
| 5 | 5010 | 1.09 | 0.08 | 0.03 | 1 | 5010 | |||||||||||
| 6 | 5052 | 1.2 | 1 | 0 | 1 | 5051 | M+H+ | RL34 | 3 | 5051 | 0.03 | 0.77 | 0.24 | 1 | 5051 | M+H+ | RL34 |
| 4 | 5115 | 0.36 | 0.25 | 0.11 | 0.7 | 10230* | M+2H+ | RS19 | |||||||||
| 7 | 5130 | 1.28 | 0.21 | 0.13 | 0.9 | 10259* | M+2H+ | RS19 | 5 | 5129 | 0.53 | 0.31 | 0.17 | 0.8 | |||
| 8 | 5484 | 1.33 | 0.03 | 0.01 | 0.7 | ||||||||||||
| 6 | 5541 | 0.35 | 0.57 | 0.05 | 0.9 | ||||||||||||
| 7 | 5624 | 0.18 | 0.38 | 0.14 | 0.7 | ||||||||||||
| 9 | 5908 | 1.4 | 0.72 | 0.06 | 1 | 5907 | M+H+ | RL33 | |||||||||
| 8 | 5937 | 0.46 | 0.47 | 0.18 | 1 | 5937 | M+H+ | RL33 | |||||||||
| 10 | 5946 | 1.31 | 0.10 | 0.03 | 1 | ||||||||||||
| 11 | 6053 | 1.54 | 0.03 | 0.01 | 0.9 | ||||||||||||
| 9 | 6342 | 0.70 | 0.32 | 0.15 | 1 | 6342 | M+H+ | RL32 | |||||||||
| 12 | 6404 | 1.49 | 0.55 | 0.04 | 1 | 6402* | M+H+ | RL32 | |||||||||
| 10 | 6431 | 0.50 | 0.47 | 0.29 | 0.9 | ||||||||||||
| 13 | 7080 | 1.59 | 0.66 | 0.15 | 1 | 7078 | M+H+ | RL29 | 11 | 7078 | 0.52 | 0.55 | 0.24 | 1 | 7078 | M+H+ | RL29 |
| 14 | 7227 | 1.65 | 0.32 | 0.05 | 1 | 7226* | M+H+ | RL35 | 12 | 7226 | 0.51 | 0.25 | 0.07 | 0.7 | 7226* | M+H+ | RL35 |
| 15 | 8068 | 1.81 | 0.28 | 0.06 | 1 | ||||||||||||
| 16 | 8167 | 1.95 | 0.06 | 0.03 | 1 | 8165 | M+H+ | RL31 | |||||||||
| 17 | 8225 | 1.81 | 0.49 | 0.25 | 1 | 8224* | M+H+ | RS21 | |||||||||
| 13 | 9350 | 1.15 | 0.24 | 0.08 | 0.7 | ||||||||||||
| 18 | 9379 | 1.97 | 0.28 | 0.13 | 1 | 9377* | M+H+ | RS20 | 14 | 9377 | 2.35 | 0.19 | 0.09 | 0.9 | 9377* | M+H+ | RS20 |
| 15 | 9557 | 0.53 | 0.31 | 0.17 | 0.8 | 9554* | M+H+ | RL27 | |||||||||
| 19 | 9570 | 2.17 | 0.1 | 0.04 | 1 | 9568* | M+H+ | RL27 | |||||||||
| 16 | 10237 | 1.78 | 0.06 | 0.04 | 0.7 | 10237* | M+H+ | RS15 | |||||||||
| 20 | 10260 | 2.13 | 0.09 | 0.05 | 1 | 10259* | M+H+ | RS19 | |||||||||
| 17 | 11248 | 0.50 | 0.47 | 0.29 | 0.9 | ||||||||||||
Identical peaks are marked by italics. Ribosomal proteins are indicated by bold type. M – average value of experimental m/z for each peak (mass spectra collected from 10 independent experiment have been used); σ (M) – standard deviation of M; I – relative intensity of peak; σ (I) – standard deviation of I; f – frequency of peak registration; M (match) – value of protein masses from SwissProt/TrEMBL database (Da);
with N-terminal methionine lost; Ion type – for matched peaks only; D – description of matched protein according to SwissProt/TrEMBL database.
Figure 3.
MALDI-TOF mass spectra of the purified ribosomal fraction and cell lysate of N. gonorrhoeae ATCC 49226 strain. The relative intensities of the ions are shown on the y axis, and the mass to charge ratios are shown on the x axis.
In contrast to N. gonorrhoeae, mass spectra of the closely related species N. meningitidis (clinical strain 547) exhibited only 47 significant peaks; 17 were found to be the most reproducible ones. Despite taxonomical (phylogenetic) relationship between N. gonorrhoeae and N. meningitidis, only six peaks were found identical for these species (selected by gray in Table 3).
To reveal the heterogeneity of strains under investigation Por-typing of N. gonorrhoeae and serotyping of N. meningitidis was performed (Table 1). Commonly, N. gonorrhoeae strains were distributed among 10 different por-types. Five of the investigated isolates were found to belong to the PorB1a serovar (PIA); the remaining 24 strains were assigned to serovar PorB1b. For N. meningitidis strains, all three serogroups were revealed, but serogroup A was predominant.
All together, MALDI-TOF mass spectra were collected for 29 clinical strains of N. gonorrhoeae, 13 clinical strains of N. meningitidis and 15 strains of non-pathogenic Neisseria (listed in Table 1). Comparative analysis of peak lists showed good reproducibility within each species with slight differences between strains.
Thus, for N. gonorrhoeae the majority of significant peaks were constant and only three peaks m/z 4473, 5051, and 8165 changed their m/z value to 4487, 5081, and 8146 respectively for several strains (Figure 2C). For N. meningitidis strains similar changes of two peaks with m/z 9377, and 9557 (transformed to 9393 and 9587, respectively) have been observed (Figure 2D).
For N. gonorrhoeae strains, it was possible to unveil the molecular basis of the observed changes. The variable peaks were matched to ribosomal proteins RL36, RL34, and RL31, respectively. Full lengths of the corresponding rpmJ, rpmH, and rpmE genes were sequenced for strains with the original and modified mass profile. A single nucleotide alteration leading to an amino acid substitution was found for each gene. The expected mass differences calculated from the amino acid substitutions were coincident with the observed mass changes (Table 4). These findings confirmed the results of SwissProt/TrEMBL database search and supported previous reports that mainly ribosomal proteins contribute to the MALDI-TOF profiles of bacteria in the mass range that we have investigated. Based on m/z values of these peaks the N. gonorrhoeae strains were divided into four different groups with m/z peaks' combinations 4473/5051/8165; 4487/5051/8165; 4487/5051/8146; and 4473/5081/8165. It should be mentioned that most clinical strains as well as strain ATCC 49226 belonged to the first group corresponding to the wild-type sequence.
Table 4.
Molecular Changes of Ribosomal Proteins Corresponding to Observed Differences within N. gonorrhoeae and N. meningitidis MALDI-TOF Mass Spectra Groups
| m/z | Matched protein (SwissProt/TrEMBL) | Gene | Nucleotide changes | Amino acid changes | Calculated mass changes, Da |
|---|---|---|---|---|---|
| N. gonorrhoeae | |||||
| 4473 | RL36 | rpmJ | *wt | wt | 4473 |
| G19→A | V7→I | 4487 | |||
| 5051 | RL34 | rpmH | Wt | wt | 5051 |
| G109→A | A37→T | 5081 | |||
| 8166 | RL31 | rpmE | Wt | wt | 8165 |
| G29→A | R10→H | 8146 | |||
| N. meningitides | |||||
| 9378 | RS20 | rpsT | Wt | wt | 9378 |
| G169→A | V57→I | 9393 | |||
| 9557 | RL27 | rpmA | Wt | wt | 9554 |
| G79→A | G27→S | 9584 |
wt – nucleotide and amino acid sequences are corresponding to those from N. gonorrhoeae strain FA 1090 (NC 002946) and N. meningitidis strains FAM18, serogroup C (NC 008767), MC58, serogroup B (NC 008767), Z2491, serogroup A (NC 003116), respectively.
For N. meningitidis strains, the variable peaks were matched to ribosomal proteins RS20 and RL27. Full-length sequencing of the corresponding genes rpsT, and rpmA was performed. As for N. gonorrhoeae strains, the nucleotide alterations leading to amino acid substitutions were found, and the theoretically calculated mass differences for these proteins were coincident with the mass changes observed (Table 4). In the case of N. meningitidis, only two m/z combinations of variable peaks were found—9377/9557 and 9393/9587.
While the profiles obtained for N. gonorrhoeae and N. meningitidis showed only limited intraspecies variability comparison of MALDI-TOF mass spectra collected for N. gonorrhoeae, N. meningitidis and non-pathogenic Neisseria species had numerous significant differences in peak positions (Figure 4). For further cluster analysis, a database containing the main spectra of the 57 Neisseria strains was created using the MALDI Biotyper 1.1 software. A main spectrum based on the comparative analysis of twenty raw spectra was generated for each strain using the dedicated data processing algorithms of the software.
Figure 4.
MALDI-TOF mass spectra comparison of closely related species of N. gonorrhoeae, N. meningitidis, N. sicca, N. preflava, and non-determined Neisseria spp 8503. The relative intensities of the ions are shown on the ϒ axis, and the mass to charge ratios are shown on the x axis. The patterns display the individual distinguishes among isolates that lead to Neisseria species separation.
The established database was used to examine specificity of the technique for the identification of the different Neisseria species. Alignment of all 57 main spectra divided them into three separate groups corresponding to N. gonorrhoeae, N. meningitidis and non-pathogenic Neisseria (Figure 5) that demonstrated the possibility of isolates identification at the species level.
Figure 5.
A score oriented dendrogram of MALDI-TOF mass spectra profiles was generated using the MALDI Biotyper 1.1 software with the following settings for peak picking: lower bound 3000, upper bound 11,000, maximum peaks 100 and threshold 0.001. Neisseria species are listed in Table 1. The dendrogram was generated with the following settings: the distance measure was set at Euclidian and linkage set at completion.
To investigate if different MALDI-TOF mass spectrometers can be used for bacteria species identification, two different sets of mass spectra collected by means of a benchtop microflex MALDI-TOF mass spectrometer and an ultraflex II MALDI-TOF/TOF mass spectrometer were compared (Table 5). We found a good correlation (more than 0.86) between the mass spectra collected by the different devices for E. coli, N. gonorrhoeae, N. meningitides, and non-pathogenic Neisseria strains.
Table 5.
Comparison of MALDI Mass Spectra
| Np (U) 14 | 0.00 | 0.00 | 0.09 | 0.09 | 0.05 | 0.05 | 0.18 | 0.18 | 0.15 | 0.15 | 0.93 | 0.98 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Np (M) 13 | 0.00 | 0.01 | 0.23 | 0.23 | 0.16 | 0.16 | 0.36 | 0.36 | 0.30 | 0.30 | 0.97 | |
| Nm2 (U) 10 | 0.00 | 0.02 | 0.73 | 0.81 | 0.70 | 0.74 | 0.87 | 0.87 | 0.93 | 0.93 | ||
| Nm2 (M) 9 | 0.00 | 0.00 | 0.73 | 0.81 | 0.70 | 0.74 | 0.87 | 0.87 | 0.93 | |||
| Nm1 (U) 8 | 0.00 | 0.00 | 0.64 | 0.73 | 0.60 | 0.64 | 0.97 | 0.97 | ||||
| Nm1 (M) 7 | 0.00 | 0.00 | 0.67 | 0.73 | 0.60 | 0.64 | 0.97 | |||||
| Ng2 (U) 6 | 0.00 | 0.00 | 0.80 | 0.92 | 0.89 | 0.99 | ||||||
| Ng2 (M) 5 | 0.00 | 0.00 | 0.88 | 0.83 | 0.98 | |||||||
| Ng1 (U) 4 | 0.00 | 0.00 | 0.86 | 0.99 | ||||||||
| Ng1 (M) 3 | 0.00 | 0.00 | 0.95 | |||||||||
| Ec (U) 2 | 0.88 | 0.91 | ||||||||||
| Ec (M) 1 | 0.99 | |||||||||||
| EC (M) 1 | EC (U) 2 | Ng1 (M) 3 | Ng1 (U) 4 | Ng2 (M) 5 | Ng2 (U) 6 | Nm1 (M) 7 | Nm1 (U) 8 | Nm2 (M) 9 | Nm2 (U) 10 | Np (M) 13 | Np (U) 14 |
E. coli (EC) strain DH5α, N. gonorrhoeae strains ATCC 49226 (Ng1) and ps601 (Ng2), N. meningitidisstrains 537(Nm1) and 229 (Nm2), and non-pathogenic N. subflava 227 (Np) strain. We used a microflex LT MALDI-TOF mass spectrometer (M) and an ultraflex II MALDI-TOF/TOF mass spectrometer (U) from Bruker Daltonics, Germany. The composite correlation index (CCI) matrix was calculated using the MALDI Biotyper 1.1 software with the following settings: lower bound 3000, upper bound 12000, resolution of the mass range 4, number of intervals for CCI 4. CCIs for comparison mass spectra are indicated by bold.
In addition, a separate MALDI Biotyper database containing 1671 main spectra generated from different environmental and clinically relevant microorganisms was used to show the specificity of pathogenic Neisseria species identification. N. gonorrhoeae and N. meningitidis isolates could be identified unambiguously on the species level regardless of the instrument used for the profile spectra acquisition.
Discussion
Advantages of MALDI-TOF mass spectrometry make this technique very attractive for the rapid identification of bacteria based on the simultaneous detection of a mixture of cellular components, mainly proteins and peptides, directly from “whole” cells without preliminary extraction and separation. The current trend of this approach to introduce it into routine is finding the balance between simplicity, speed of analytical procedure and optimal characteristics of the obtained mass spectra required for particular applications.
The study presented is focused on features of DBP by MALDI TOF mass spectrometry for investigation of clinically relevant microorganisms such as Neisseria. Interest in N. meningitidis and N. gonorrhoeae remains permanent because of their well-known contribution in the structure of human infection diseases. The annual local outbreaks of meningitis and the spreading of multidrug resistant gonococci require the development the new effective molecular systems for species identification and typing. Mass spectrometry based methods seem to be most promising in this field. Recently, Oliver Schmid et al, 2005, used surface enhanced laser desorption/ionization-time of flight mass spectrometry in combination with artificial neural networks for the identification of N. gonorrhoeae.23 In addition the automated comparative sequence analysis by base-specific cleavage and mass spectrometry was used for N. meningitidis typing.24
The challenge of our investigation was to apply a MALDI-TOF mass spectrometry technique for species identification of closely related bacteria from the genus Neisseria. This approach based on mass spectral profile analysis, which does not require any sophisticated purification of bacterial proteins, enables species identification with significantly lower costs than surface enhanced laser desorption/ionization-time of flight mass spectrometry without any loss of efficiency. Some years ago, a similar approach was shown to be suitable for investigation of different Haemophilus spp including N. gonorrhoeae as a microorganism that exhibits a similar biochemical and bacteriological profile.25 In our case, the DBP technique was extended by using correlation analysis to compare mass spectra acquired with different instruments. The high mass spectra reproducibility together with low intraspecies variability and evident interspecies heterogeneity indicated the applicability of this approach for Neisseria species identification. Moreover, analysis of N. gonorrhoeae ribosomal fraction confirmed that most of the peaks registered using the described protocol belong to ribosomal proteins, conserved for certain bacterial species.
Furthermore, we evaluated the DBP approach as a rapid tool useful for identification of bacterial isolates. The following steps of the DBP protocol investigation and optimization were performed: theoretical analysis of earlier accumulated experience, adjustment of the sample preparation protocol using E. coli to compare our results with data of other studies (matrix, extraction solution composition, etc), adjustment of the sample preparation protocol and MALDI-TOF mass spectrometry procedure using first references and then clinical strains of the investigated bacterial species. An important point is that the clinical strains were epidemiologically different strains covering interspecies variability in the investigated Neisseria species.
The problem of the Neisseria genus study is complex. Although the clinically important microbes—N. gonorrhoeae and N. meningitidis—are relatively straightforward to identify, the differences between many of the non-pathogenic strains are small and the speciation of these strains within a diagnostic setting is not always possible. While the identification of non-pathogenic Neisseria to species level is generally not required, the misidentification of these strains as N. gonorrhoeae or N. meningitidis can have serious health, legal and social consequences.
Based on the described experiments we can conclude that the suggested DBP technique works very well for species identification of N. gonorrhoeae and N. meningitidis. Demonstrated performance of the MALDI profiling approach is similar to recently published data describing the capability of analogue techniques for species identification of closely related mycobacteria26,27 and for Gram-negative non-fermenting bacteria.8
As far as non-pathogenic Neisseria are concerned, their mass spectral profiles are completely different from N. gonorrhoeae and N. meningitidis (distance level is 1000) and form a rather heterogenic group. Probably, further enlargement of this group by introduction of more commensally Neisseria allows obvious species discrimination as well.
It should be mentioned that despite investigated N. gonorrhoeae and N. meningitidis strains belonged to different genotypes or serotypes, little intraspecies heterogeneity has been found looking at the prominent peaks of the spectra. The four main proteotypes described forN. gonorrhoeae and two for N. meningitidis did not reflect the true variability of these microorganisms. Therefore, for Neisseria genus the described DBP approach seems more suitable for robust species identification in contrast to resequencing techniques, eg, MALDI-TOF mass spectrometry based multilocus sequence typing,24 used for strain differentiation. But the capability of subtyping using low-intensity peaks, eg, in a hierarchical approach after speciation by standard pattern matching, has to be evaluated.
On the other hand, species identification by means of DBP is much easier, cheaper and more rapid than by using DNA-analysis based technologies. The optimized measuring protocol based on the strong criteria of mass spectra collection (high spectra reproducibility, little noise and a flat baseline, presenting as many peaks as possible and medium resolution of 600 within the range of m/z selected for analysis) enabled fast acquisition of high-quality spectra in automatic measurements. In general, the MALDI mass spectra produced by our methodology were conserved for each specimen and slightly different within the same species. Moreover, most differences observed comparing individual measurements of one strain were in intensities, and only in few cases in absence or presence of peaks. These measurement variations, which are intrinsic to the MALDI-TOF technology, are reduced significantly by merging 20 raw spectra for each sample into one main spectrum, which displays the most reproducible peaks. Nevertheless, it is important to keep experimental conditions constant to guarantee a secure species identification.
At the same time, the approach has great advantages such as the requirement of a small amount of biological material, very easy sample preparation, highly accurate measurement, and low cost per analysis. As a whole-cell fingerprinting technique and due to the low amount of material required, DBP allows the rapid identification of large numbers of microbial isolates at a very early stage in the screening process, eg, from bacterial colonies growing on pre-isolation plates.
Thus, we conclude that the DBP approach using the described rapid, simple and effective protocol coupled with special software, eg, the MALDI Biotyper, equipped with optimized algorithms for identification and classification is highly suitable for the secure Neisseria species identification in routine settings. Further studies will be dedicated to investigations of other bacterial genera and to the utilization of the minor spectra variability between different subtypes for subspecies classification.
Acknowledgements
We sincerely thank Vladimir A. Karpov for oligonucleotide primers synthesis and Marina V. Serebraykova for useful discussions.
Footnotes
Supported by Development Contract No 24-05 from Russian Agency of Health; Sächsische Aufbaubank (SAB10634).
The results were partly presented at the 15th International Pathogenic Neisseria Conference, September 2006, Australia (Abstract P4.3.08).
References
- 1.Kubanova AA, Frigo NV, Priputnevich TV, Sidorenko SV. Antimicrobial resistance of Neisseria gonorrhoeae in the Russian Federation in 2006. Clin Microbiol Infect. 2007;13:S8. Abstract 031. [Google Scholar]
- 2.Achtman M, van der Ende A, Zhu P, Koroleva IS, Kusecek B, Morelli G, Schuurman IG, Brieske N, Zurth K, Kostyukova NN, Platonov AE. Molecular epidemiology of serogroup a meningitis in Moscow, 1969 to 1997. Emerg Infect Dis. 2001;7:420–427. doi: 10.3201/eid0703.010309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Alexander S, Ison C. Evaluation of commercial kits for the identification of Neisseria gonorrhoeae. J Med Microbiol. 2005;54:827–831. doi: 10.1099/jmm.0.46108-0. [DOI] [PubMed] [Google Scholar]
- 4.Unemo M, Palmer HM, Blackmore T, Herrera G, Fredlund H, Limnios A, Nguyen N, Tapsall J. Global transmission of prolyliminopeptidase-negative Neisseria gonorrhoeae strains: implications for changes in diagnostic strategies. Sex Trans Infect. 2007;83:47–51. doi: 10.1136/sti.2006.021733. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lynn EC, Chung MC, Tsai WC, Han CC. Identification of enterobacteriaceae bacteria by direct matrix-assisted laser desorption/ ionization mass spectrometric analysis of whole cells. Rapid Comm Mass Spectrom. 1999;13:2022–2027. doi: 10.1002/(SICI)1097-0231(19991030)13:20<2022::AID-RCM750>3.0.CO;2-3. [DOI] [PubMed] [Google Scholar]
- 6.Leuschner RGK, Beresford-Jones N, Robinson C. Difference and consensus of whole cell Salmonella enterica subsp enterica serovars matrix-assisted laser desorption/ionization time-of-flight mass spectrometry spectra. App Microbiol. 2003;38:24–31. doi: 10.1046/j.1472-765x.2003.01436.x. [DOI] [PubMed] [Google Scholar]
- 7.Mandrell RE, Harden LA, Bates A, Miller WG, Haddon WF, Fagerquist CK. Speciation of campylobacter coli, c. jejuni, c. helveticus, c. lari, c. sputorum, and c. upsaliensis by matrix-assisted laser desorption ionization-time of flight mass spectrometry. App Environ Microbiol. 2005;71:6292–6307. doi: 10.1128/AEM.71.10.6292-6307.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Mellmann A, Cloud J, Maier T, Keckevoet U, Ramminger I, Iwen P, Dunn J, Hall G, Wilson D, LaSala P, Kostrzewa M, Harmsen D. Evaluation of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) in comparison to 16S rRNA gene sequencing for species identification of nonfermenting bacteria. J Clin Microbiol. 2008;46:1946–1954. doi: 10.1128/JCM.00157-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fenselau C, Demirev PA. Characterization of intact microorganisms by MALDI mass spectrometry. Mass Spectrom Rev. 2001;20:157–171. doi: 10.1002/mas.10004. [DOI] [PubMed] [Google Scholar]
- 10.Holland RD, Wilkes JR, Rafii F, Sutherland JB, Persons CC, Voorhees KJ, Lay JO., Jr 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. 1996;10:1227–1232. doi: 10.1002/(SICI)1097-0231(19960731)10:10<1227::AID-RCM659>3.0.CO;2-6. [DOI] [PubMed] [Google Scholar]
- 11.Meetani MA, Voorhees KJ. MALDI mass spectrometry analysis of high molecular weight proteins from whole bacterial cells: pretreatment of samples with surfactants. J Am Soc Mass Spectrom. 2005;16:1422–1426. doi: 10.1016/j.jasms.2005.04.004. [DOI] [PubMed] [Google Scholar]
- 12.Williams TL, Andrzejewski D, Lay JO, Jr, Musser SM. Experimental factors affecting the quality and reproducibility of MALDI-TOF mass spectra obtained from whole bacteria cells. J Am Soc Mass Spectrom. 2003;14:342–351. doi: 10.1016/S1044-0305(03)00065-5. [DOI] [PubMed] [Google Scholar]
- 13.Maier T, Klepel S, Renner U, Kostrzewa, M: Fast and reliable MALDI-TOF MS-based microorganism identification. Nature Methods 2006, Nature Publishing Group. Application note
- 14.Fudyk TC, Maclean IW, Simonsen NJ, Njagi EN, Kimani J, Brunham RC, Plummer FA. Genetic diversity and mosaicism at the por locus of Neisseria gonorrhoeae. J Bacteriol. 1999;181:5591–5599. doi: 10.1128/jb.181.18.5591-5599.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ilina EN, Malakhova MV, Vereshchagin VA, Govorun VM, Sergienko VI, Zubkov MM, Vasilev MM, Kubanova AA. Molecular typing of N. gonorrhoeae strains prevalent in the Russian Federation. Bull Exp Biol Med. 2003;136:179–182. doi: 10.1023/a:1026379209810. [DOI] [PubMed] [Google Scholar]
- 16.Posada D, Crandall KA, Nguyen M, Demma JC, Viscidi RP. Population genetics of the porB gene of Neisseria gonorrhoeae: different dynamics in different homology groups. Mol Biol Evol. 2000;17:423–436. doi: 10.1093/oxfordjournals.molbev.a026322. [DOI] [PubMed] [Google Scholar]
- 17.Viscidi RP, Demma JC, Gu J, Zenilman J. Comparison of sequencing of the por gene and typing of the opa gene for discrimination of Neisseria gonorrhoeae strains from sexual contacts. J Clin Microbiol. 2000;38:4430–4438. doi: 10.1128/jcm.38.12.4430-4438.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Blaha G, Stelzl U, Spahn CMT, Agrawal RK, Frank J, Nierhaus KH. Preparation of functional ribosomal complexes and the effect of buffer conditions on tRNA positions observed by cryoelectron microscopy. Methods Enzymol. 2000;317:292–309. doi: 10.1016/s0076-6879(00)17021-1. [DOI] [PubMed] [Google Scholar]
- 19.Boom R, Sol CJ, Salimans MM, Jansen CL, Wertheim-van Dillen PM, van der Noordaa J. Rapid and simple method for purification of nucleic acids. J Clin Microbiol. 1990;28:495–503. doi: 10.1128/jcm.28.3.495-503.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Arnold RJ, Reilly JP. Fingerprint matching of E. coli strains with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry of whole cells using a modified correlation approach. Rapid Commun Mass Spectrom. 1998;12:630–636. doi: 10.1002/(SICI)1097-0231(19980529)12:10<630::AID-RCM206>3.0.CO;2-0. [DOI] [PubMed] [Google Scholar]
- 21.Ryzhov V, Fenselau C. Characterization of the protein subset desorbed by MALDI from whole bacterial cells. Anal Chem. 2001;73:746–750. doi: 10.1021/ac0008791. [DOI] [PubMed] [Google Scholar]
- 22.Keys CJ, Dare DJ, Sutton H, Wells G, Lunt M, McKenna T, McDowall M, Shah HN. Compilation of a MALDI-TOF mass spectral database for the rapid screening and characterization of bacteria implicated in human infectious diseases. InfectGenet and Evol. 2004;4:221–242. doi: 10.1016/j.meegid.2004.02.004. [DOI] [PubMed] [Google Scholar]
- 23.Schmid O, Ball G, Lancashire L, Culak R, Shah H. New approaches to identification of bacterial pathogens by surface enhanced laser desorption/ionization time of flight mass spectrometry in concert with artificial neural networks, with special reference to Neisseria gonorrhoeae. J Med Microbiol. 2005;54:1205–1211. doi: 10.1099/jmm.0.46223-0. [DOI] [PubMed] [Google Scholar]
- 24.Honisch C, Chen Y, Mortimer C, Arnold C, Schmidt O, van den Boom D, Cantor CR, Shah HN, Gharbia SE. Automated comparative sequence analysis by base-specific cleavage and mass spectrometry for nucleic acid-based microbial typing. Proc Natl Acad Sci USA. 2007;104:10649–10654. doi: 10.1073/pnas.0704152104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Haag AM, Taylor SN, Johnston KH, Cole RB. Rapid identification and speciation of Haemophilus bacteria by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. J Mass Spectrom. 1998;33:750–756. doi: 10.1002/(SICI)1096-9888(199808)33:8<750::AID-JMS680>3.0.CO;2-1. [DOI] [PubMed] [Google Scholar]
- 26.Pignone M, Greth KM, Cooper J, Emerson D, Tang J. Identification of mycobacteria by matrix-assisted laser desorption ionization–time-of-flight mass spectrometry. J Clin Microbiol. 2006;44:1963–1970. doi: 10.1128/JCM.01959-05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Hettick JM, Kashon ML, Simpson JP, Siegel PD, Mazurek GH, Weissman DN. Proteomic profiling of intact mycobacteria by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Anal Chem. 2004;76:5769–5776. doi: 10.1021/ac049410m. [DOI] [PubMed] [Google Scholar]