Abstract
Matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) has demonstrated its ability to promptly identify nontuberculous mycobacteria using the Mycobacteria Library v2.0. However, some species are particularly difficult to identify reliably using this database, providing a low log(score). In this study, the identification power of an updated Mycobacteria Library (v3.0) has been evaluated. Overall, 109 NTM isolates were analyzed with both databases. The v3.0 database allowed a high-level confidence in the identification [log(score) value, ≥1.8] of 91.7% of the isolates versus 83.5% with the v2.0 version (P < 0.01).
TEXT
Rapid identification of nontuberculous mycobacteria (NTM) with matrix-assisted laser desorption–ionization time of flight mass spectrometry (MALDI-TOF MS) has outperformed molecular techniques, such as GenoType (Hain Lifescience GmbH, Nehren, Germany), and provides accurate identification that correlates well with 16S rRNA gene sequencing when applied to the most common species of NTM (1–3). The Mycobacteria Library database (Bruker Daltonik GmbH, Bremen, Germany) available so far (v2.0) provided low scores, particularly for NTM belonging to the slow-growing groups (3–5). In the present study, we assessed the power of a new database to identify NTM (i.e., Mycobacteria Library v3.0) using 109 isolates from 26 NTM species (Table 1) and compared the identification scores of version v2.0 and version v3.0.
TABLE 1.
MALDI-TOF MS identification of NTM isolates grouped according the Runyon classification
| Mycobacterium group | Identification score with Mycobacteria Library v2.0: |
Identification score with Mycobacteria Library v3.0: |
No. of isolates | ||||
|---|---|---|---|---|---|---|---|
| 3.000–1.800 | 1.799–1.600 | ≤1.599 | 3.000–1.800 | 1.799–1.600 | ≤1.599 | ||
| Rapid growers | |||||||
| M. abscessus | 13 | 13 | 13 | ||||
| M. arupense | 1 | 1 | 1 | ||||
| M. chelonae | 6 | 6 | 6 | ||||
| M. fortuitum | 16 | 2 | 18 | 18 | |||
| M. insubricum | 1 | 1 | 1 | ||||
| M. mageritense | 3 | 3 | 3 | ||||
| M. mucogenicum | 2 | 2 | 2 | ||||
| M. peregrinum | 2 | 2 | 2 | ||||
| M. porcinum | 1 | 1 | 1 | ||||
| M. smegmatis | 1 | 1 | 1 | ||||
| No. of rapid growers (% of total organisms) | 46 (95.8) | 2 (4.2) | 0 (0.0) | 48 (100) | 0 (0.0) | 0 (0.0) | 48 |
| Slow-growing nonchromogens | |||||||
| M. avium | 11 | 5 | 2 | 14 | 4 | 18 | |
| M. haemophilum | 1 | 1 | 1 | ||||
| M. intracellulare | 5 | 1 | 6 | 6 | |||
| M. malmoense | 1 | 1 | 1 | ||||
| M. palustre | 1 | 1 | 1 | ||||
| M. shimoidei | 1 | 1 | 1 | ||||
| M. triplex | 1 | 1 | 1 | ||||
| No. of slow-growing nonchromogens (% of total organisms) | 21 (72.4) | 6 (20.7) | 2 (6.9) | 25 (86.2) | 4 (13.8) | 0 (0.0) | 29 |
| Slow-growing scotochromogens | |||||||
| M. bohemicum | 1 | 1 | 1 | ||||
| M. europaeum | 1 | 1 | 1 | ||||
| M. gordonae | 5 | 3 | 1 | 6 | 3 | 9 | |
| M. lentiflavum | 5 | 5 | 5 | ||||
| M. szulgai | 1 | 1 | 1 | ||||
| M. xenopi | 4 | 1 | 5 | 5 | |||
| No. of slow-growing scotochromogens (% of total) | 17 (77.3) | 3 (13.6) | 2 (9.1) | 19 (86.4) | 3 (13.6) | 0 (0.0) | 22 |
| Slow-growing photochromogens | |||||||
| M. kansasii | 5 | 3 | 6 | 2 | 8 | ||
| M. marinum | 1 | 1 | 1 | ||||
| M. simiae | 1 | 1 | 1 | ||||
| No. of slow-growing photochromogens (% of total) | 7 (70.0) | 3 (30) | 0 (0.0) | 8 (80.0) | 2 (20) | 0 (0.0) | 10 |
| Total no. (%) | 91 (83.5) | 14 (12.8) | 4 (3.7) | 100 (91.7) | 9 (8.3) | 0 (0.0) | 109 |
Ninety-nine nonselected NTM isolates from clinical samples and 10 reference strains (Table 1) were collected in the clinical microbiology laboratory from the Hospital Gregorio Marañón (Madrid, Spain) between January 2011 and May 2015. These isolates were routinely identified by 16S rRNA hsp65 sequencing and, in parallel, by MALDI-TOF MS using a Microflex LT benchtop mass spectrometer (Bruker Daltonik) and the Mycobacteria Library v2.0, containing 313 Mycobacterium isolates from 131 species (Bruker Daltonik) (4). Sample preparation was described elsewhere (6). Briefly, colonies of NTM isolates grown on Lowenstein-Jensen medium were harvested into a 1.5-ml Eppendorf tube with 300 μl of deionized water and inactivated for 30 min at 95°C under biosafety level 3 conditions. Then, they were centrifuged at maximal speed and subsequently resuspended in 300 μl of water and 900 μl of absolute ethanol and centrifuged again at 13,000 rpm. The supernatant was discarded, and the pellet was taken to biosafety level 2 conditions in order to disrupt the mycobacteria cell aggregates with silica bead vortexing and extract the bacterial proteins using formic acid and acetonitrile. In the end, 1 μl of supernatant was placed onto a steel plate for MALDI-TOF MS analysis. Samples were analyzed in duplicates; the species identification, using the Mycobacteria Library v2.0, and the higher log(score) value result were recorded. For comparison reasons, all of the protein spectra from the 109 NTM isolates were reanalyzed using the new Mycobacteria Library v3.0, containing 853 references from 149 Mycobacterium species. Log(score) value differences of 0.1 or greater between the databases were considered significant (Table 2).
TABLE 2.
Isolates with an improved score when identified with the Mycobacteria Library v3.0
| Isolate number | Mycobacteria Library v2.0 |
Mycobacteria Library v3.0 |
Consistency Category | ||
|---|---|---|---|---|---|
| Identification | Score | Identification | Score | ||
| 4 | M. abscessus subsp. abscessus DSM 44567 DSM b | 1,815 | M. abscessus 545_11 FZB b | 2,012 | A |
| 14 | M. arupense DSM 44942T DSM b | 1,918 | Mycobacterium arupense CCUG 52359 CCUG b | 2,067 | A |
| 17 | M. avium subsp. avium DSM 44157 DSM b | 1,889 | M. avium subsp. hominissuis 276_04 FZB b M | 2,026 | A |
| 18 | M. avium subsp. avium CCUG 28067 CCUG b | 1,636 | M. avium subsp. hominissuis 252_11 HLG b | 1,742 | B |
| 20 | M. avium subsp. avium DSM 44158 DSM b | 1,711 | M. avium subsp. hominissuis 1244_09 FZB b | 1,917a | B |
| 22 | M. avium subsp. avium DSM 44158 DSM b | 1,589 | M. avium subsp. hominissuis 1244_09 FZB b | 1,875 | B |
| 23 | M. avium subsp. avium DSM 44156T DSM b | 1,587 | M. avium subsp. hominissuis 9469_10 FZB b | 1,745 | B |
| 24 | M. avium subsp. avium CCUG 28067 CCUG b | 1,608 | M. avium 22_027242 MML b | 1,788 | B |
| 25 | M. avium subsp. avium DSM 44157 DSM b | 1,924 | M. avium subsp. hominissuis 1840_09 FZB b | 2,117 | A |
| 26 | M. avium subsp. silvaticum DSM 44175T DSM b | 1,705 | M. avium subsp. hominissuis 10756_10 FZB b M | 1,820 | B |
| 32 | M. bohemicum DSM 44408 DSM b | 1,903 | M. bohemicum DSM 44277T DSM b | 2,048 | A |
| 40 | M. fortuitum subsp. fortuitum DSM 43477 DSM b | 1,985 | M. fortuitum subsp. fortuitum CCUG 46694 CCUG b | 2,092 | A |
| 41 | M. fortuitum subsp. fortuitum DSM 46622 DSM b | 1,953 | M. fortuitum subsp. fortuitum CCUG 46694 CCUG b | 2,193 | A |
| 43 | M. fortuitum subsp. fortuitum DSM 43477 DSM b | 1,963 | M. fortuitum subsp. fortuitum CCUG 46694 CCUG b | 2,126 | A |
| 45 | M. fortuitum subsp. fortuitum DSM 46622 DSM b b | 2,056 | M. fortuitum subsp. fortuitum CCUG 46694 CCUG b | 2,161 | A |
| 46 | M. fortuitum subsp. fortuitum DSM 46622 DSM b b | 2,109 | M. fortuitum subsp. fortuitum CCUG 46694 CCUG b | 2,265 | A |
| 47 | M. fortuitum subsp. fortuitum DSM 43477 DSM b | 1,797 | M. fortuitum 121126_11 UOK b | 1,950 | B |
| 48 | M. fortuitum subsp. fortuitum DSM 46622 DSM b | 1,888 | M. fortuitum subsp. fortuitum CCUG 46694 CCUG b | 2,032 | A |
| 53 | M. fortuitum subsp. fortuitum DSM 43477 DSM b | 1,950 | M. fortuitum subsp. fortuitum CCUG 46694 CCUG b | 2,056 | A |
| 54 | M. fortuitum subsp. fortuitum DSM 43477 DSM b | 1,648 | M. fortuitum 7655_05 FZB b M | 1,806 | B |
| 56 | M. fortuitum subsp. fortuitum DSM 43477 DSM b | 2,220 | M. fortuitum subsp. fortuitum CCUG 46694 CCUG b | 2,390 | A |
| NC10394 | M. fortuitum subsp. acetamidolyticum DSM 44220T DSM b | 1,912 | M. fortuitum 5879_96 FZB b | 2,097 | A |
| 58 | M. gordonae NO1493 LIG b | 2,123 | M. gordonae CCUG 21812 CCUG b L | 2,226 | A |
| 59 | M. gordonae NO1493 LIG b | 1,909 | M. gordonae 21_006209 MML b | 2,012 | A |
| 61 | M. gordonae NO1493 LIG b | 1,543 | M. gordonae 177_12 FZB b M | 1,737 | B |
| 65 | M. gordonae NO1493 LIG b | 1,843 | M. gordonae CCUG 21812 CCUG b L | 2,098 | A |
| DSM 44634 | M. haemophilum DSM 44634T DSM b | 1,891 | M. haemophilum7958_11 FZB b | 2,058 | A |
| 67 | M. chimaera-M. intracellulare group (M. intracellulare DSM 44365 DSM b) | 1,705 | M. chimaera-M. intracellulare group (M. intracellulare 5878_96 FZB b) | 1,939 | B |
| 68 | M. chimaera-M. intracellulare group (M. chimaera 12030617 MVD b) | 1,974 | M. chimaera-M. intracellulare group (M. chimaera 3720_10 FZB b) | 2,170 | A |
| 69 | M. chimaera-M. intracellulare group (M. intracellulare DSM 44365 DSM b) | 1,896 | M. chimaera-M. intracellulare group (M. intracellulare 9643_11 FZB b) | 2,167 | A |
| 72 | M. chimaera-M. intracellulare group (M. chimaera 12030617 MVD b) | 2,061 | M. chimaera-M. intracellulare group (M. chimaera 3720_10 FZB b) | 2,165 | A |
| 80 | M. kansasii DSM 44162T DSM b | 1,857 | M. kansasii NLA001001128 UMR b | 2,207 | A |
| 82 | M. lentiflavum DSM 44421 DSM b | 2,071 | M. lentiflavum DSM 44420 DSM b | 2,308 | A |
| 83 | M. lentiflavum DSM 44421 DSM b | 2,074 | M. lentiflavum DSM 44420 DSM b | 2,186 | A |
| 85 | M. lentiflavum DSM 44419 DSM b | 2,014 | M. lentiflavum DSM 44420 DSM b | 2,115 | A |
| 87 | M. mageritense DSM 44476T DSM b | 2,140 | M. mageritense 22_043151 MML b | 2,310 | A |
| 88 | M. mageritense DSM 44476T DSM b | 2,091 | M. mageritense 2105743 IIUB b | 2,222 | A |
| 90 | M. mucogenicum-M. phocaicum group (M. mucogenicum CCUG 47451T CCUG b) | 2,218 | M. mucogenicum-M. phocaicum group (M. mucogenicum 1323_12 FZB b) | 2,362 | A |
| 93 | M. peregrinum DSM 43271T DSM b | 2,155 | M. peregrinum 130506271201 HPL b | 2,299 | A |
| 94 | M. porcinum DSM 44242T DSM b | 2,150 | M. porcinum WC14_0103 NYDH b | 2,361 | A |
| DSM 44626 | M. triplex DSM 44626T DSM b | 1,933 | M. triplex 19_017398 MML b | 2,049 | A |
| 96 | M. xenopi DSM 43995T DSM b | 1,943 | M. xenopi 3238_11 FZB b | 2,237 | A |
| 97 | M. xenopi DSM 44169 DSM b | 1,485 | M. xenopi 11720_11 FZB b | 2,033 | A |
| 99 | M. xenopi DSM 43995T DSM b | 1,827 | M. xenopi 11163_11 FZB b | 1,947 | B |
| DSM 43995 | M. xenopi DSM 43995T DSM b | 1,905 | M. xenopi AFB047 MCW b | 2,125 | A |
When the score obtained with the v3.0 database allowed an upgrade within the low- and high-confidence level categories established in this study, the score is shown in bold.
Log(score) values of 1.6 and 1.8 were proposed as new mycobacteria thresholds for low- and high-confidence identification levels, respectively (7). Sensitivity values obtained with both Mycobacteria Library databases were compared using the McNemar test for paired samples, with two tails. SPSS software package 18.0 (IBM, Chicago, IL, USA) was used for the data analysis.
All of the NTM isolates included in this study obtained a correct MALDI-TOF MS identification using the Mycobacteria Library v2.0, although 18 isolates were identified with scores below 1.8 and 4 of these had scores below 1.6 (Table 1). The isolates with low scores belonged mainly to Mycobacterium avium, Mycobacterium gordonae, and Mycobacterium kansasii species. However, the Mycobacteria Library v3.0 improved the log(score) value of 45 (41.3%) NTM isolates: mass spectra exceeded 1.6 for 4 of the isolates (2 M. avium, 1 M. gordonae, and 1 Mycobacterium xenopi) and reached or exceeded 1.8 for 5 isolates (2 M. avium, 2 Mycobacterium fortuitum, and 1 Mycobacterium intracellulare); the log(score) for the other 36 isolates was ≥1.8 with both database versions, although these score increased when they were identified with the Mycobacteria Library v3.0 (Table 2). Also, the log(score) values of 4 M. avium, 3 M. gordonae, and 2 M. kansasii isolates remained below 1.8 despite the increase. On the other hand, the log(score) value for 1 additional isolate of M. gordonae reached ≥1.8 with the updated library, although it did not meet the criteria established to be considered an increased log(score). In all the cases, the improvement in the identification quality was due to higher homology with mycobacteria strains added to the updated library. The impact of the update on the identification of different NTM species is detailed in Table 3.
TABLE 3.
Impact of database update on identification of clinical isolates
| Mycobacterium species | No. of isolates | No. of references in: |
Mean (range) log(score) value(s) with: |
||
|---|---|---|---|---|---|
| v2.0 | v3.0 | v2.0 | v3.0 | ||
| M. abscessus | 13 | 11 | 24 | 2.036 (1.815–2.183) | 2.059 (1.830–2.209) |
| M. arupense | 1 | 1 | 8 | 1.918 | 2.067 |
| M. avium | 17 | 10 | 39 | 1.872 (1.587–2.222) | 1.963 (1.738–2.298) |
| M. bohemicum | 1 | 3 | 10 | 1.903 | 2.048 |
| M. chelonae | 6 | 10 | 19 | 2.058 (1.989–2.208) | 2.084 (2.017–2.208) |
| M. europaeum | 1 | 1 | 2 | 1.847 | 1.847 |
| M. fortuitum | 17 | 9 | 20 | 2.052 (1.648–2.393) | 2.160 (1.806–2.393) |
| M. gordonae | 8 | 10 | 23 | 1.824 (1.543–2.123) | 1.904 (1.737–2.226) |
| M. haemophilum | 1 | 1 | 6 | 1.891 | 2.058 |
| M. insubricum | 1 | 3 | 3 | 2.018 | 2.018 |
| M. chimaera-M. intracellulare group | 6 | 9 | 32 | 1.971 (1.705–2.137) | 2.105 (1.939–2.170) |
| M. kansasii | 8 | 14 | 26 | 1.927 (1.636–2.243) | 1.992 (1.636–2.243) |
| M. lentiflavum | 5 | 3 | 13 | 2.063 (2.014–2.132) | 2.175 (2.043–2.308) |
| M. mageritense | 2 | 2 | 4 | 2.146 (2.091–2.207) | 2.249 (2.214–2.310) |
| M. malmoense | 1 | 5 | 14 | 2.233 | 2.233 |
| M. marinum | 1 | 8 | 18 | 2.345 | 2.345 |
| M. mucogenicum-M. phocaicum group | 2 | 3 | 21 | 2.218/2.101a | 2.362/2.127a |
| M. palustre | 1 | 2 | 2 | 2.086 | 2.086 |
| M. peregrinum | 2 | 1 | 15 | 2.349/2.155b | 2.349/2.299b |
| M. porcinum | 1 | 2 | 8 | 2.150 | 2.361 |
| M. shimoidei | 1 | 1 | 6 | 2.225 | 2.225 |
| M. simiae | 1 | 3 | 13 | 2.166 | 2.166 |
| M. smegmatis | 1 | 5 | 9 | 2.034 | 2.034 |
| M. szulgai | 1 | 5 | 15 | 2.125 | 2.199 |
| M. triplex | 1 | 1 | 2 | 1.933 | 2.019 |
| M. xenopi | 4 | 5 | 18 | 1.852 (1.485–2.099) | 2.093 (1.947–2.237) |
| Total | 104 | 128 | 370 | ||
Score values for the two representative M. mucogenicum-M. phocaicum group strains.
Score values for the two representative M. peregrinum strains.
None of the log(score) values decreased with the Mycobacteria Library v3.0 according to the criteria established in this study. The identification obtained with both databases matched in all cases.
Using the older version of the database, only 91 isolates (83.5%) met the criteria for the high-confidence identification level [log(score), ≥1.8] (Table 1). However, the new database allowed the identification of 100 isolates (91.7%; P < 0.01) with a log(score) value of ≥1.8 and also improved the rate of NTM isolates identified at a low-confidence level [log(score), ≥1.6], since none of the analyzed isolates was identified with a log(score) below 1.636.
NTM isolates were grouped following the Runyon classification, and the identification results obtained with each database were compared. Using the Mycobacteria Library v3.0, a higher number of isolates belonging to all of the mycobacteria groups was correctly identified with a log(score) value of ≥1.8. Although the number of rapid growing isolates reaching high scores was similar with both databases (46 [95.8%] in v2.0 versus 48 [100%] in v3.0), the number of slow-growing mycobacteria isolates identified at a high-confidence level was higher when the Mycobacteria Library v3.0 was used (Table 1). Differences were mainly due to the increased number of M. avium, M. gordonae, M. kansasii, and M. xenopi isolates correctly identified at a high-confidence level. Previous studies have already shown that the identification of M. avium and M. gordonae with a log(score) value of ≥1.8 was sometimes difficult to reach using older databases (3–5). However, the updated Mycobacteria Library v3.0 overcame this drawback and allowed a high-confidence identification of 14 of 18 M. avium and 6 of 9 M. gordonae isolates, while the remaining 4 M. avium and 3 M. gordonae isolates were reliably and unequivocally identified with log(score) values of ≥1.7.
In conclusion, the use of the improved Mycobacteria Library v3.0 has increased the number of NTM isolates that attain a high-confidence identification level. The NTM species reportedly identified with low-confidence log(score) values using the previous, v2.0, database (M. avium and M. gordonae isolates in particular) (1, 3) are now reliably identified with log(score) values of ≥1.7. This is very important for clinical microbiology laboratories, where the achievement of a rapid identification is as crucial as the robustness of the identification itself. Finally, the high rate of high-confidence NTM identifications obtained with the Mycobacteria Library v3.0 underlines the importance of implementing MALDI-TOF MS technology, coupled with updated databases containing several strains per NTM species, for the rapid and accurate identification of these microorganisms.
ACKNOWLEDGMENTS
This study was supported by the Miguel Servet Program CP14/00220, part of the Plan Estatal I+D+I 2013-2016, and cofinanced by ISCII-Subdirección General de Evaluación y Fomento de la Investigación and the European Regional Development Fund.
B.R.-S. is supported by Miguel Servet contract MS14/00220. M.T. and M.K. are employees at the mass spectrometry company Bruker Daltonik GmbH. Other authors declare no conflicts of interest.
Funding Statement
The Miguel Servet Program is a governmental program to promote emerging scientists. This study was supported by the Miguel Servet Program CP14/00220, part of the Plan Estatal I+D+I 2013-2016, and cofinanced by ISCII-Subdirección General de Evaluación y Fomento de la Investigación and the European Regional Development Fund.
REFERENCES
- 1.Buchan BW, Riebe KM, Timke M, Kostrzewa M, Ledeboer NA. 2014. Comparison of MALDI-TOF MS with HPLC and nucleic acid sequencing for the identification of Mycobacterium species in cultures using solid medium and broth. Am J Clin Pathol 141:25–34. doi: 10.1309/AJCPBPUBUDEW2OAG. [DOI] [PubMed] [Google Scholar]
- 2.Mather CA, Rivera SF, Butler-Wu SM. 2014. Comparison of the Bruker Biotyper and Vitek MS matrix-assisted laser desorption ionization–time of flight mass spectrometry systems for identification of mycobacteria using simplified protein extraction protocols. J Clin Microbiol 52:130–138. doi: 10.1128/JCM.01996-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Balada-Llasat JM, Kamboj K, Pancholi P. 2013. Identification of mycobacteria from solid and liquid media by matrix-assisted laser desorption ionization–time of flight mass spectrometry in the clinical laboratory. J Clin Microbiol 51:2875–2879. doi: 10.1128/JCM.00819-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Rodríguez-Sánchez B, Ruiz-Serrano MJ, Marín M, López Roa P, Rodríguez-Créixems M, Bouza E. 2015. Evaluation of MALDI-TOF MS for the identification of nontuberculous mycobacteria (NTM) from clinical isolates. J Clin Microbiol 53:2737–2740. doi: 10.1128/JCM.01380-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Machen A, Kobayashi M, Connelly MR, Wang YF. 2013. Comparison of heat inactivation and cell disruption protocols for identification of mycobacteria from solid culture media by use of Vitek matrix-assisted laser desorption ionization–time of flight mass spectrometry. J Clin Microbiol 51:4226–4229. doi: 10.1128/JCM.02612-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mediavilla-Gradolph MC, De Toro-Peinado I, Bermúdez-Ruiz MP, García-Martínez MDLA, Ortega-Torres M, Quezel-Guerraz NM, Palop-Borrás B. 2015. Use of MALDI-TOF MS for identification of nontuberculous Mycobacterium species isolated from clinical specimens. BioMed Res Int 2015:854078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Pranada AB, Timke M, Witt E, Kostrzewa M. 2015. Mycobacteria identification by MALDI Biotyper system: evolution of database content and evaluation criteria. Abstr 25th Eur Congr Clin Microbiol Infect Dis, poster 1190. [Google Scholar]