Table

Microbiological applications of pyrolysis mass spectrometry

Purpose

Problem

Microorganism

References

CIassification

Identification intra-species heterogeneity epidemiology (hospital infections)

pathogenici ty cariogenicity

Mycobacterium

Bacillus

Listeria

Neisseria gonorrhoea Pseudomonas Escherichia coli Arthrobacter Staphylococous Proteus Streptococcus

Klebsiella

Escherichia coli

Mycobacterium

Streptococcus

Wieten et al. (123-125) Meuzelaar et al. (43,151 Böhm and Meuzelaar (152) Risby and Yergey (19) Eshuis et al. (45), Meuzelaar et al. (43)

Risby and Yergey (19) Risby and Yergey (19) Risby and Yergey (19) Anhalt and Fenselau (18) Anhalt and Fenselau (18) Kistemaker et al. (66.), Meuzelaar et al. (48)

Haverkamp et al. (154)

Haverkamp et al. (65) Wieten et al. (123-125) Kistemaker et al. (66), Meuzelaar et al. (48)

Figure 36. Non-linear map (stress 12.2%) of the pyrolysis mass spectra of 13 Neisseria gonorrhoea strains, showing the presence of two distinct pyrotypes. Mote correlation between pyrotype II and the presence of delayed glucose utilization (- sign). Glucose utilization tests were repeated in different laboratories (see double signs).

Figure 36. Non-linear map (stress 12.2%) of the pyrolysis mass spectra of 13 Neisseria gonorrhoea strains, showing the presence of two distinct pyrotypes. Mote correlation between pyrotype II and the presence of delayed glucose utilization (- sign). Glucose utilization tests were repeated in different laboratories (see double signs).

Figure 37. Examples of pyrolysis mass spectra of two different pyrotypes of neisseria gonorrhoea. Arrows indicate significantly higher peak intensities. Conditions: sample 20 yg; Tc 510°C; Eel 14 eV.

Figure 37. Examples of pyrolysis mass spectra of two different pyrotypes of neisseria gonorrhoea. Arrows indicate significantly higher peak intensities. Conditions: sample 20 yg; Tc 510°C; Eel 14 eV.

Figure 36 gives an example of fingerprinting results obtained by Curie-point Py-MS. Analysis of N. gonorrhoea strains, harvested directly from chocolate agar culture plates, showed the presence of two hitherto unknown pyrotypes. Careful study of the conventional biochemical typing results revealed a possible correlation of the two pyrotypes with the rate of glucose metabolism; pyrotype II appeared to correlate with delayed metabolic conversion of glucose. Inspection of the spectra of both pyrotypes (Figure 37) shows a complex difference, with peak intensities at m/z 48, 67, 92, 114 and 117 being higher in type I and peak intensities at m/z 30, 59, 73, 95, 98, 109 and 125 being higher in type II. Tentative biochemical interpretation

Figure 38. Non-linear map (stress 12.7%) of two sets of pyrolysis mass spectra of Neisseria gonorrhoea strains. One set of spectra (•) corresponds to the set shown in Figure 36 (only centroids of the replicate analyses shown here). The second set (o) was analysed 3 months later and four of these strains were identical with strains of the first set (see broken lines). Note that both pyrotypes are still clearly distinguishable after 3 months, in spite of the obvious drift in the patterns caused by changes in instrumental and/or biological conditions.

points to an increase in some N-acetylamino sugar component in type II, accompanied by a decrease in a protein component rich in methionine (m/z 48), proline (m/z 67), phenylalanine (m/z 92) and tryptophan (m/z 117). Although later analyses showed the extent of the difference to be variable, the same two pyrotypes were clearly distinguishable three months later (see Figure 38). The reproducibility of the spectra was good enough to enable the older spectra to be used as reference patterns for identifying the new spectra, even though the latter included several strains not analysed before.

Probably the most detailed studies on fingerprinting bacteria by Py-MS are those carried out by Wieten et al. on mycobacteria (ref. 123 - 125). The main aim of this work is the identification of clinical Mycobacterium strains as belonging to either M. tuberculosis, M. bovis or M. bovis BCG, or to other mycobacteria. The three mentioned species, together referred to as the Tuberculosis complex, are most important to man; M. tuberculosis and M. bovis are pathogenic, whereas M. bovis BCG is used for vaccination purposes and may also cause disease. Special computerized data processing procedures for matching unknown strains with reference strains of the Tuberculosis complex were developed so as to confirm the existing microbiological

Figure 39. Non-linear map (stress 13.8%) of 18 clinical isolates of Myaobaatevium strains and 9 reference strains of the Tuberculosis complex (shaded), calculated on the basis of the key masses m/z 31, 50, 58, 59, 71, 86, and 98. Identification of the clinical isolates as whether or not belonging to the Tuberculosis complex was made on the basis of their distance (see Section 6.4) with respect to the reference strains. In this series, strains 1-6 belong to the Tuberculosis complex, whereas strains 8-18 were identified as other, non-pathogenic mycobacteria. Strain 7, micro-biologically typed as a non-Tuberculosis complex strain (M. xenopi), was misidentified as belonging to the Tuberculosis complex by Py-MS, using this key mass identification procedure. Note the clustering of the two strains 17 and 18, bacteriologically typed as M. kansasii, and also the relative closeness of strains 13 and 14 (M. fovtuitum) and of strains 8 and 9 (M. terrae).

Figure 39. Non-linear map (stress 13.8%) of 18 clinical isolates of Myaobaatevium strains and 9 reference strains of the Tuberculosis complex (shaded), calculated on the basis of the key masses m/z 31, 50, 58, 59, 71, 86, and 98. Identification of the clinical isolates as whether or not belonging to the Tuberculosis complex was made on the basis of their distance (see Section 6.4) with respect to the reference strains. In this series, strains 1-6 belong to the Tuberculosis complex, whereas strains 8-18 were identified as other, non-pathogenic mycobacteria. Strain 7, micro-biologically typed as a non-Tuberculosis complex strain (M. xenopi), was misidentified as belonging to the Tuberculosis complex by Py-MS, using this key mass identification procedure. Note the clustering of the two strains 17 and 18, bacteriologically typed as M. kansasii, and also the relative closeness of strains 13 and 14 (M. fovtuitum) and of strains 8 and 9 (M. terrae).

classification scheme. Identification of Tuberculosis complex strains can be made on the basis of the relative peak intensities of just seven "key masses" in the pyrolysis mass spectra (see Figure 39; ref. 124). Py-MS fingerprinting has also been used

Figure 40. Non-linear map of the pyrolysis mass spectra (only centroids shown) of 20 Klebsiella isolates from 13 different patients (arabic numerals) in the same hospital. Encircled spectra cannot be differentiated in the distance matrix (not shown). Note that only the isolates from patients 5 and 7 are possibly identical. However, these isolates could be differentiated by conventional biotyping. Thus, no cross-infection was found to exist between patients.

Figure 40. Non-linear map of the pyrolysis mass spectra (only centroids shown) of 20 Klebsiella isolates from 13 different patients (arabic numerals) in the same hospital. Encircled spectra cannot be differentiated in the distance matrix (not shown). Note that only the isolates from patients 5 and 7 are possibly identical. However, these isolates could be differentiated by conventional biotyping. Thus, no cross-infection was found to exist between patients.

for taxonomic problems within the genus Mycobacterium for differentiation of M. kansasii and M. gastri (ref. 125).

Apart from the above applications of fingerprinting for classification and identification, Curie-point Py-MS techniques can also be used for epidemiological purposes in microbiology, as shown in Figures 40 and 41. Klebsiella infections present a difficult problem in hospital epidemiology because conventional biochemical typing techniques often do not distinguish adequately between different Klebsiella strains and serological techniques are hampered by limited availability of reliable sera (ref. 155). The other major source of hospital infections, namely Staphylococcus, has so far eluded reliable classification by biochemical and serological (ref. 156) techniques, and so differentiation of Staphylococcus strains relies almost entirely on phage typing. However, about 10-15% of all Staphylococcus strains isolated in hospitals do not respond to any of the more generally available phages. In these cases, promising results have been obtained with subtyping of "phage-negative" Staphylococcus strains by Curie-point Py-MS (ref. 157).

0 0

Post a comment