Abstract
Matrix-assisted laser desorption ionization−time of flight mass spectrometry (MALDI-TOF MS) has entered clinical laboratories, facilitating identification of bacteria. Here, we evaluated the MALDI Biotyper (Bruker Daltonics) for the identification of fastidious Gram-negative rods (GNR). Three sample preparation methods, direct colony transfer, direct transfer plus on-target formic acid preparation, and ethanol-formic acid extraction, were analyzed for 151 clinical isolates. Direct colony transfer applied with the manufacturer's interpretation criteria resulted in overall species and genus identification rates of 43.0% and 32.5%, respectively; 23.2% of the isolates were not identified, and two misidentifications (1.3%) were observed. The species identification rates increased to 46.4% and 53.7% for direct transfer plus formic acid preparation and ethanol-formic acid extraction, respectively. In addition, we evaluated score value cutoff alterations. The identification rates hardly increased by reducing the genus cutoff, while reducing the 2.0 species cutoff to 1.9 and to 1.8 increased the identification rates to up to 66.2% without increasing the rate of misidentifications. This study shows that fastidious GNR can reliably be identified using the MALDI Biotyper. However, the identification rates do not reach those of nonfastidious GNR such as the Enterobacteriaceae. In addition, two approaches optimizing the identification of fastidious GNR by the MALDI Biotyper were demonstrated: formic acid-based on-target sample treatment and reductions in cutoff scores to increase the species identification rates.
INTRODUCTION
Fastidious Gram-negative rods (GNR) are slow-growing bacteria that belong to different genera, such as Neisseria, Pasteurella, and the HACEK organisms (Haemophilus, Aggregatibacter, Cardiobacterium, Eikenella, and Kingella). Many of these species are commensal organisms of the human or animal oral cavity. They are isolated from wound infections, e.g., after a human or animal bite, but are also associated with endocarditis, septicemia, and abscesses, particularly in immunocompromised and elderly patients (1–4).
Traditionally, identification of fastidious GNR in diagnostic laboratories is based on morphological and biochemical criteria. Identification is particularly challenging and time-consuming as these bacteria usually do not grow on standard media and require supplemental nutrients and/or a CO2-enriched atmosphere for growth (1–3). 16S rRNA gene sequencing is widely used in addition to conventional identification methods (5–7). Matrix-assisted laser desorption ionization−time of flight mass spectrometry (MALDI-TOF MS) has been suggested as an alternative to molecular methods for the identification of fastidious GNR and other bacteria that are difficult to culture (8–10).
Numerous studies have compared MALDI-TOF MS-based bacterial identification with molecular or biochemical identification methods (11–22). Most of these studies included only limited numbers of fastidious GNR or analyzed only a single genus, such as Haemophilus spp. (23–26), Neisseria spp. (1, 27), or Pasteurella spp. (28). Reported identification rates varied between different studies, mainly because different sets of fastidious GNR were investigated, different sample preparation methods were applied, and different MALDI-TOF MS systems with different databases were used.
We evaluated the capability of the MALDI Biotyper (Bruker Daltonics, Bremen, Germany) to identify 151 clinical isolates of fastidious GNR, including 40 species from 15 genera. We used three sample preparation methods (direct colony transfer, direct transfer plus on-target formic acid preparation, and tube-based extraction with ethanol and formic acid) and applied the manufacturer's interpretation criteria. Furthermore, we evaluated the feasibility of altered identification cutoffs for the species and genus level. Based on our data, we propose an optimized combination of sample preparation and cutoff scores to increase the identification rates for fastidious GNR.
MATERIALS AND METHODS
Bacterial isolates and culture conditions.
In this study, 151 clinical isolates of fastidious GNR collected at the Institute of Medical Microbiology, University of Zurich, Switzerland, from 1993 to 2012 were included (Table 1). The isolates were characterized using biochemical assays and 16S rRNA gene sequencing, the latter being used as the identification gold standard. When the discriminatory power of the 16S rRNA gene sequence analysis was insufficient, supplemental biochemical tests were used for the final identification (6). Bacteria were cultivated on Columbia sheep blood agar and on chocolate agar containing PolyViteX (bioMérieux, Marcy l'Etoile, France) at 37°C with 7.5% CO2 for 24 to 48 h.
TABLE 1.
Identification of 151 fastidious GNR by MALDI-TOF MS by applying direct colony transfer and standard Bruker interpretation criteria (species cutoff of 2.0; genus cutoff of 1.7)
| Organisma | No. (%) of isolates | No. (%) of MALDI-TOF MS resultsb: |
No. of reference spectrac | |||
|---|---|---|---|---|---|---|
| Species identification | Genus identification | No identification | Misidentification | |||
| Actinobacillus hominis | 2 | 2 | 0 | |||
| Aggregatibacter actinomycetemcomitans | 2 | 1 | 1 | 5 | ||
| Aggregatibacter aphrophilus | 10 | 5 | 3 | 2 | 3 | |
| Aggregatibacter sp. | 3 | 3 | 9 | |||
| Bordetella bronchiseptica | 1 | 1 | 9 | |||
| Bordetella petrii | 2 | 2 | 4 | |||
| Bordetella trematum | 1 | 1 | 3 | |||
| Capnocytophaga canimorsus | 3 | 1 | 2 | 1 | ||
| Capnocytophaga gingivalis | 1 | 1 | 1 | |||
| Capnocytophaga sputigena | 6 | 6 | 2 | |||
| Capnocytophaga sp. | 1 | 1 | 10 | |||
| Cardiobacterium hominis | 7 | 3 | 4 | 1 | ||
| Dysgonomonas capnocytophagoides | 1 | 1 | 0 | |||
| Dysgonomonas mossii | 1 | 1 | 0 | |||
| Eikenella corrodens | 13 | 10 | 3 | 2 | ||
| Haemophilus influenzae | 5 | 4 | 1 | 10 | ||
| Haemophilus parainfluenzae | 5 | 3 | 2 | 5 | ||
| Haemophilus sp. | 1 | 1 | 23 | |||
| Kingella denitrificans | 2 | 1 | 1 | 2 | ||
| Kingella kingae | 1 | 1 | 4 | |||
| Leptotrichia trevisanii | 1 | 1 | 0 | |||
| Moraxella atlantae | 1 | 1 | 1 | |||
| Moraxella canis | 1 | 1 | 2 | |||
| Moraxella lacunata | 2 | 1 | 1 | 1 | ||
| Moraxella nonliquefaciens | 5 | 5 | 3 | |||
| Moraxella osloensis | 10 | 1 | 5 | 4 | 7 | |
| Neisseria animaloris | 1 | 1 | 0 | |||
| Neisseria bacilliformis | 4 | 4 | 1 | |||
| Neisseria cinerea | 2 | 2 | 1 | |||
| Neisseria elongata | 9 | 1 | 7 | 1 | 2 | |
| Neisseria oralis | 1 | 1 | 0 | |||
| Neisseria shayeganii | 2 | 2 | 0 | |||
| Neisseria sicca | 1 | 1 | 3 | |||
| Neisseria subflava | 2 | 2 | 2 | |||
| Neisseria weaveri | 5 | 5 | 1 | |||
| Neisseria zoodegmatis | 5 | 5 | 1 | |||
| Neisseria sp. | 2 | 2 | 42 | |||
| Oligella urethralis | 6 | 3 | 3 | 2 | ||
| Pasteurella bettyae | 4 | 4 | 0 | |||
| Pasteurella canis | 1 | 1 | 3 | |||
| Pasteurella dagmatis | 1 | 1 | 1 | |||
| Pasteurella multocida | 11 | 11 | 10 | |||
| Pasteurella stomatis | 1 | 1 | 0 | |||
| Pasteurella sp. | 3 | 1 | 2 | 17 | ||
| Psychrobacter pulmonis | 1 | 1 | 0 | |||
| Psychrobacter sp. | 1 | 1 | 1 | |||
| Total | 151 (100) | 65 (43.0) | 49 (32.5) | 35 (23.2) | 2 (1.3) | |
Identification by 16S rRNA gene sequence analysis and biochemical methods according to de Melo Oliveira et al. (6).
MALDI-TOF MS identification applying direct transfer plus formic acid preparation, a genus cutoff value of 1.7, and a species cutoff value of 2.0 with Bruker database V.3.1.2.0 (3,995 entries).
Number of reference entries for the corresponding species and genus, respectively, included in Bruker database V.3.1.2.0 (3,995 entries).
Biochemical identification.
Fastidious GNR isolates were identified by means of biochemical reactions as specified by de Melo Oliveira et al. (6).
16S rRNA gene sequence analysis.
The identification of fastidious GNR by partial 16S rRNA gene sequencing was performed according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI) (29), and as described by de Melo Oliveira et al. (6). A 16S rRNA gene fragment, corresponding to Escherichia coli positions 10 to 806, was amplified using primers BAK11w (5′-AGTTTGATC[A/C]TGGCTCAG) and BAK2 (5′-GGACTAC[C/T/A]AGGGTATCTAAT) and sequenced with forward primer BAK11w (6). Sequences were analyzed for homology using a GenBank BLAST search and SmartGene IDNS software (SmartGene GmbH, Zug, Switzerland). The identification was done according to the following criteria: (i) species identification when the sequence determined had a similarity score of ≥99% with that of a reference sequence of a classified species and when the sequence divergence between different species was ≥0.5%; (ii) a genus was assigned for similarity scores of <99% to ≥95%; and (iii) a family was assigned for similarity scores of less than 95% (6).
Sample preparation for MALDI-TOF MS.
Bacteria were prepared for MALDI-TOF MS as described previously (30). In brief, for the direct colony transfer procedure, colony material from overnight cultures was deposited on a polished steel MSP 96 target (Bruker Daltonics) and overlaid with 1 μl of a saturated α-cyano-4-hydroxycinnamic acid (HCCA) matrix solution (Bruker Daltonics). Direct transfer plus formic acid preparation was done by the addition of 1 μl of a 70% formic acid solution to the bacterial spot prior to overlaying of the sample with the matrix solution. The tube-based extraction was performed by suspending a loopful of bacterial colony material in 300 μl of distilled water and 900 μl of 100% ethanol. The cell suspension was centrifuged at 17,000 × g for 2 min, and the pellet was resuspended in 5 to 50 μl of 70% formic acid according to the pellet size. Subsequently, acetonitrile was added in an equal volume. The suspension was centrifuged at 17,000 × g for 2 min, and 1 μl of the supernatant was spotted on the MALDI target. After being dried, the spot was overlaid with 1 μl of matrix solution.
MALDI-TOF MS analysis.
Mass spectra were acquired and analyzed using a microflex LT mass spectrometer (Bruker Daltonics) in combination with research-use-only (RUO) versions of the MALDI Biotyper software package (version 3.0) and the reference database V.3.1.2.0 (3,995 entries) as previously published (30). Calibration was done by following the manufacturer's instructions and using the manufacturer's recommended bacterial test standard (Bruker Daltonics). Analyses of all strains were done in duplicate from independent preparations.
MALDI-TOF MS data interpretation.
Data were interpreted by application of the manufacturer's standard criteria (30). In short, species were assigned for scores of ≥2.0, and genera were assigned for scores of ≥1.7 but <2.0. If scores were lower than 1.7, no identification was assigned. The cutoff scores were altered by reducing the standard 2.0 species cutoff to 1.9, 1.8, and 1.7 and the standard 1.7 genus cutoff to 1.6 and 1.5.
Statistical analysis.
Statistical calculations were performed with IBM SPSS, version 20 (SPSS Inc., Chicago, IL). Overall differences between the three MALDI-TOF MS preparation methods and the altered cutoff scores were tested using the Friedman test. In addition, follow-up tests were conducted using the Wilcoxon signed-rank test for pairwise comparison of the three preparation methods and the altered cutoffs. A P value of <0.05 was considered statistically significant. A Bonferroni correction was applied at the 0.0166 (0.05/3) and 0.0083 (0.05/6) levels of significance, respectively, across pairwise comparisons.
Nucleotide sequence accession numbers.
Partial 16S rRNA gene sequences of all 151 isolates have been deposited in NCBI GenBank by de Melo Oliveira et al. (6) (GU797849, KC866143 to KC866155, KC866157, KC866158, KC866160 to KC866167, KC866170 to KC866178, KC866182 to KC866197, KC866199 to KC866208, KC866210 to KC866224, KC866226, KC866229, KC866232 to KC866237, KC866239 to KC866242, KC866244, KC866245, KC866247 to KC866249, KC866251 to KC866256, KC866259 to KC866268, KC866270, KC866271, KC866274, KC866275, KC866277, KC866281 to KC866283, KC866285 to KC866287, KC866290 to KC866292, KC866294, and KC866299) or were deposited in the course of the present study (KJ557352 to KJ557379).
RESULTS
Identification of fastidious GNR by the MALDI Biotyper.
We analyzed 151 clinical fastidious GNR, including 40 species from 15 genera, by the Bruker MALDI Biotyper (Table 1). Using direct colony transfer and the manufacturer's recommended criteria for the identification of species (cutoff of 2.0) and genus (cutoff of 1.7), identification numbers were 65 out of 151 strains (43.0%) at the species level and 49 out of 151 strains (32.5%) at the genus level (Table 1). For 35 of 151 strains (23.2%) no identification was achieved. Incorrect species assignment was seen with two isolates: one Dysgonomonas capnocytophagoides isolate was misidentified as Dysgonomonas gadei, and one isolate of Haemophilus sp. was misidentified as Haemophilus influenzae (Table 2). In 5 out of the 49 isolates which were identified only to genus level, more than one species was listed with a score of ≥2.0 in the ranked MALDI-TOF MS results. According to the manufacturer's criteria, this setting indicates inconsistency in species assignment and results in genus assignment only. Such species inconsistency was observed for Actinobacillus hominis (scores of ≥2.0 for Actinobacillus suis/Actinobacillus equuli/Actinobacillus ureae), Bordetella bronchiseptica (scores of ≥2.0 for Bordetella bronchiseptica/Bordetella parapertussis), Neisseria sicca (scores of ≥2.0 for Neisseria mucosa/Neisseria macacae), and Neisseria subflava (scores of ≥2.0 for Neisseria perflava/Neisseria flavescens).
TABLE 2.
Overview of misidentifications by MALDI-TOF MS
| Organisma | No. (%) of isolates | MALDI-TOF MS resultsb |
Homology to 16S rRNA reference sequencec | ||
|---|---|---|---|---|---|
| Misidentification as: | Misidentification at cutoff 2.0 | Misidentification at cutoff 1.7 | |||
| Aggregatibacter sp. | 3 | A. aphrophilus | No | Yes | 97.7%–98.3% A. aphrophilus |
| Capnocytophaga sp. | 1 | C. sputigena | No | Yes | 98.5% C. ochracea |
| Dysgonomonas capnocytophagoides | 1 | D. gadei | Yes | Yes | 100% D. capnocytophagoides |
| Haemophilus sp. | 1 | H. influenzae | Yes | Yes | 96.8% H. haemolyticus |
| Neisseria sp. | 2 | N. zoodegmatis | No | Yes | 97.8%–97.9% N. zoodegmatis |
| Pasteurella stomatis | 1 | P. canis | No | Yes | 99.6% P. stomatis |
| Pasteurella sp. | 1 | P. canis | No | Yes | 98.4% P. stomatis |
Identification by 16S rRNA gene sequence analysis and biochemical methods according to de Melo Oliveira et al. (6).
MALDI-TOF MS identification applying direct transfer plus formic acid preparation, a genus cutoff value of 1.7, and a species cutoff value of 2.0 with Bruker database V.3.1.2.0 (3,995 entries).
Species identification when sequence homology is ≥99% and a sequence divergence with the next homologous species is ≥0.5%; genus assignment when sequence homology is <99% and ≥95%.
Species coverage of the Bruker database.
Ten species, accounting for 15 isolates in this study (i.e., Actinobacillus hominis, Dysgonomonas capnocytophagoides, Dysgonomonas mossii, Leptotrichia trevisanii, Neisseria animaloris, Neisseria oralis, Neisseria shayeganii, Pasteurella bettyae, Pasteurella stomatis, and Psychrobacter pulmonis), were not included in the Bruker database version (3,995 entries) we used (Table 1). While 11 of these 15 isolates were not identified by MALDI-TOF MS, both A. hominis isolates and the P. stomatis isolate were assigned to the genus Actinobacillus and Pasteurella, respectively. The D. capnocytophagoides isolate was misidentified as Dysgonomonas gadei. The latter discrepancy was also reported by Bizzini et al. (19).
Comparison of different MALDI-TOF MS sample preparation methods.
The rate of species identification increased significantly from 43.0% to 46.4% (Z = −2.646, P = 0.008) and to 53.7% (Z = −4.491, P < 0.0001), respectively, when direct transfer plus formic acid sample preparation and ethanol-formic acid extraction were used. In parallel, the number of isolates identified only to the genus level decreased from 32.5% to 30.5% and 25.8%, and the rate of no identifications decreased from 23.2% to 21.9% and 19.2%, respectively (Table 3). The rate of misidentification was 1.3% for all three methods. The average MALDI score values for direct colony transfer (1.98 ± 0.29), for direct transfer plus formic acid preparation (1.99 ± 0.28), and for ethanol-formic acid extraction (2.02 ± 0.31) were comparable.
TABLE 3.
Identification of 151 fastidious GNR by MALDI-TOF MS: comparison of three sample preparation methods and different score value cutoffs
| MALDI-TOF MS method | No. (%) of isolates identified using a species score value cutoff of: |
No. (%) of isolates identified using a genus score value cutoff of: |
|||||
|---|---|---|---|---|---|---|---|
| 2.0 | 1.9 | 1.8 | 1.7 | 1.7 | 1.6 | 1.5 | |
| Direct colony transfer | |||||||
| Species identification | 65 (43.0) | 85 (56.3) | 92 (60.9) | 98 (64.9) | |||
| Genus identification | 49 (32.5) | 29 (19.2) | 22 (14.6) | 8 (5.3) | 116 (76.8) | 125 (82.8) | 115 (76.2) |
| No identification | 35 (23.2) | 35 (23.2) | 35 (23.2) | 35 (23.2) | 35 (23.3) | 26 (17.2) | 36 (23.8) |
| Misidentification | 2 (1.3) | 2 (1.3) | 2 (1.3) | 10 (6.6) | 0 (0) | 0 (0) | 0 (0) |
| Direct transfer plus formic acid preparation | |||||||
| Species identification | 70 (46.4) | 86 (57.0) | 95 (62.9) | 101 (66.9) | |||
| Genus identification | 46 (30.5) | 29 (19.2) | 17 (11.3) | 8 (5.3) | 118 (78.1) | 128 (84.8) | 119 (78.8) |
| No identification | 33 (21.9) | 33 (21.9) | 33 (21.9) | 33 (21.9) | 33 (21.9) | 23 (15.2) | 32 (21.2) |
| Misidentification | 2 (1.3) | 3 (2.0) | 6 (4.0) | 9 (6.0) | 0 (0) | 0 (0) | 0 (0) |
| Ethanol-formic acid extraction | |||||||
| Species identification | 81 (53.7) | 89 (58.9) | 100 (66.2) | 105 (69.5) | |||
| Genus identification | 39 (25.8) | 27 (17.9) | 16 (10.6) | 9 (6.0) | 122 (80.8) | 124 (82.1) | 128 (84.8) |
| No identification | 29 (19.2) | 29 (19.2) | 29 (19.2) | 29 (19.2) | 29 (19.2) | 27 (17.9) | 23 (15.2) |
| Misidentification | 2 (1.3) | 6 (4.0) | 6 (4.0) | 8 (5.3) | 0 (0) | 0 (0) | 0 (0) |
Individual cutoff scores for species and genus identification.
The manufacturer recommends cutoff scores for an identification at the species and genus level of 2.0 and 1.7, respectively. In this study, we analyzed the effect of altering the cutoff for species assignment from 2.0 to 1.9, 1.8, and 1.7 and for genus assignment from 1.7 to 1.6 and 1.5 (Table 3).
The reduction in the genus cutoff value did not significantly increase the genus identification rate for any of the three methods [χ2(2,|151) = 3.957, P = 0.138 for direct colony transfer; χ2(2.|151) = 4.333, P = 0.115 for direct transfer plus formic acid preparation; and χ2(2,|151) = 1.867, P = 0.393 for ethanol-formic acid extraction]. However, decreasing the genus cutoff resulted in increasing numbers of genus inconsistencies, i.e., scores of ≥1.6 or ≥1.5 for more than one genus. A genus inconsistency rate of 2% was observed for a genus cutoff of 1.7; however, genus inconsistency increased to up to 15% when a genus cutoff of 1.5 was applied. This increase was mainly caused by Aggregatibacter spp. that showed high homology (scores of ≥1.5) to H. influenzae reference entries.
The reduced species cutoffs led to significantly higher species identification rates (Table 3). Decreasing the species cutoff from 2.0 to 1.7 resulted in a species identification rate of 64.9% compared to 43.0% for direct colony transfer (Z = −5.972, P < 0.0001), of 66.9% compared to 46.4% for direct transfer plus formic acid preparation (Z = −5.684, P < 0.0001), and of 69.5% compared to 53.7% for ethanol-formic acid extraction (Z = −4.992, P < 0.0001). Concomitantly, the rate of inconsistencies at the species level increased from 4.0% to 6.0%. Species inconsistency was observed in particular for Neisseria spp. Misidentifications at the species level increased from 1.3% applying a cutoff of 2.0 to ∼6.0% applying a cutoff of 1.7. An increased misidentification rate was observed mainly for those isolates that could be assigned only to the genus level by 16S rRNA gene sequence analysis due to the low sequence homology with reference sequences (Table 2). These isolates yielded erroneous species assignments with MALDI-TOF MS.
DISCUSSION
Recent studies have evaluated the identification of fastidious GNR using the Vitek MS system (bioMérieux, Marcy l'Etoile, France) and show reliable identification of HACEK organisms and other fastidious Gram-negative bacteria (31, 32). The aim of this study was to evaluate the capability of the MALDI Biotyper system (Bruker Daltonics) to correctly identify fastidious GNR. We analyzed the feasibility of various sample preparation methods and of altered MALDI-TOF MS interpretation criteria for improving reliable identification rates, in particular the alteration of score value cutoffs.
Sample preparation.
If fastidious GNR bacteria were pretreated with formic acid either on-target or using the tube-based extraction method, the rate of species identification increased compared to the identification by direct colony transfer without formic acid. Our observations parallel those of previous reports that demonstrated improved identification rates when formic acid was used in the process of sample preparation for MALDI-TOF MS analysis of bacteria (30, 33, 34). The highest identification rates were yielded by applying ethanol-formic acid extraction, which is the manufacturer-recommended gold standard and which is the standard method for generating reference spectra. Ethanol-formic acid extraction, however, comprises a considerable number of manual preparation steps and is, thus, time-consuming and laborious (35). Use of the direct transfer plus formic acid preparation, on the other hand, hardly increases the total time to the result compared to that for direct colony transfer (30). Therefore, the direct transfer plus formic acid preparation better meets the requirements of high-throughput clinical laboratories.
Altered MALDI Biotyper data interpretation criteria.
No significant increase in overall genus assignments was observed when the genus cutoff was reduced. In contrast, the rate of species identification was significantly improved for all sample preparation methods tested by a reduction of the species cutoff value from 2.0 to 1.7. However, reducing the cutoff to 1.7 led to a significantly higher number of misidentifications (up to 6.6% instead of 1.3%). Similar observations were made previously by Bizzini et al., who reported that a species cutoff of 1.7 led to a higher identification rate for difficult-to-identify bacteria but that this increase was accompanied by additional misidentifications (19). Species cutoffs of 1.9 and 1.8 increased the identification rates less than a cutoff of 1.7. However, the misidentification rates were significantly lower. Thus, species cutoffs of 1.9 or 1.8 may offer an optimal balance between increased correct identifications and limited misidentifications. This finding is supported by other studies that suggest lower species cutoffs (mostly 1.9) for the identification of bacteria in general (34).
Coverage of the reference database.
While this study was ongoing, a new version of the Bruker database was released (v.4.0.0.1, 5,627 entries). The rate of species identification increased from 53.7% to 64.2% when the new database was used with the standard interpretation criteria and ethanol-formic acid extraction (data not shown). Applying a species identification cutoff of 1.8 instead of 2.0 led to an additional increase in the identification rate to 70.2%. The increase was due to the addition of new species to the database, such as Pasteurella bettyae and Leptotrichia trevisanii, and additional reference entries for existing species, such as Capnocytophaga gingivalis, Moraxella osloensis, Neisseria bacilliformis, and Oligella urethralis. These data indicate that the low identification rates for fastidious GNR do not reflect a technical issue, but are rather the result of poor database coverage and that further expansion of the database is needed.
Comparison of MALDI Biotyper-based identification and conventional identification algorithms.
Traditionally the workflow of our diagnostic laboratory for the identification of fastidious GNR comprises both phenotypic and molecular methods (6). Initially, the identification of putative fastidious GNR isolates is attempted by investigating phenotypic and biochemical characteristics. For isolates which cannot be identified by phenotypic and biochemical methods, 16S rRNA gene sequencing is performed. In a previous study, we reported rates of 40%, 13%, and 47% for species identification, genus identification, and no identification, respectively, of fastidious GNR by phenotypic methods (6). The use of MALDI-TOF MS identification in the clinical laboratory will, therefore, reduce the number of additional 16S rRNA gene sequence analyses and lead to less cost, time, and effort for identification of fastidious GNR. We suggest an integrated approach combining MALDI-TOF MS and 16S rRNA gene sequencing: MALDI-TOF MS is used as the primary identification method. If no identification is achieved, 16S rRNA gene analysis is done as the secondary identification method. To optimize the species identification rate, direct transfer plus on-target formic acid sample preparation in combination with a species cutoff of 1.8 is proposed.
In conclusion, we demonstrated that the Bruker MALDI Biotyper is a reliable tool for the identification of fastidious GNR, although the identification rates were not as high as those reported for Gram-positive cocci or nonfastidious Gram-negative bacteria. The number of reliable species identifications can be increased by the following three measures: direct transfer plus on-target formic acid sample pretreatment, alteration of the species cutoff score value from 2.0 to 1.8, and amelioration of the database with additional spectra.
ACKNOWLEDGMENTS
We thank the laboratory technicians of the Institute of Medical Microbiology for their assistance.
This study was supported by the University of Zurich.
REFERENCES
- 1.Zbinden R, Von Graevenitz A. 2011. Actinobacillus, Capnocytophaga, Eikenella, Kingella, Pasteurella, and other fastidious or rarely encountered Gram-negative rods, p 574–588. In Versalovic J, Carroll K, Funke G, Jorgensen J, Landry M, Warnock D (ed), Manual of clinical microbiology, 10th ed, vol 1 ASM Press, Washington, DC. [Google Scholar]
- 2.Elias J, Frosch M, Vogel U. 2011. Neisseria, p 559–573. In Versalovic J, Carroll K, Funke G, Jorgensen J, Landry M, Warnock D (ed), Manual of clinical microbiology, 10th ed, vol 1 ASM Press, Washington, DC. [Google Scholar]
- 3.Ledeboer N, Doern G. 2011. Haemophilus, p 588–602. In Versalovic J, Carroll K, Funke G, Jorgensen J, Landry M, Warnock D (ed), Manual of clinical microbiology, 10th ed, vol 1 ASM Press, Washington, DC. [Google Scholar]
- 4.Brouqui P, Raoult D. 2001. Endocarditis due to rare and fastidious bacteria. Clin Microbiol Rev 14:177–207. doi: 10.1128/CMR.14.1.177-207.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sönksen UW, Christensen JJ, Nielsen L, Hesselbjerg A, Hansen DS, Bruun B. 2010. Fastidious Gram-negatives: identification by the Vitek 2 Neisseria-Haemophilus card and by partial 16S rRNA gene sequencing analysis. Open Microbiol J 4:123–131. doi: 10.2174/1874285801004010123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.de Melo Oliveira MG, Abels S, Zbinden R, Bloemberg GV, Zbinden A. 2013. Accurate identification of fastidious Gram-negative rods: integration of both conventional phenotypic methods and 16S rRNA gene analysis. BMC Microbiol 13:162. doi: 10.1186/1471-2180-13-162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tang YW, Ellis NM, Hopkins MK, Smith DH, Dodge DE, Persing DH. 1998. Comparison of phenotypic and genotypic techniques for identification of unusual aerobic pathogenic gram-negative bacilli. J Clin Microbiol 36:3674–3679. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Biswas S, Rolain JM. 2013. Use of MALDI-TOF mass spectrometry for identification of bacteria that are difficult to culture. J Microbiol Methods 92:14–24. doi: 10.1016/j.mimet.2012.10.014. [DOI] [PubMed] [Google Scholar]
- 9.Seng P, Abat C, Rolain JM, Colson P, Lagier JC, Gouriet F, Fournier PE, Drancourt M, La Scola B, Raoult D. 2013. Identification of rare pathogenic bacteria in a clinical microbiology laboratory: impact of matrix-assisted laser desorption ionization-time of flight mass spectrometry. J Clin Microbiol 51:2182–2194. doi: 10.1128/JCM.00492-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.McElvania TeKippe E, Burnham CA. 2014. Evaluation of the Bruker Biotyper and Vitek MS MALDI-TOF MS systems for the identification of unusual and/or difficult-to-identify microorganisms isolated from clinical specimens. Eur J Clin Microbiol Infect Dis 33:2163–2171. doi: 10.1007/s10096-014-2183-y. [DOI] [PubMed] [Google Scholar]
- 11.Neville SA, Lecordier A, Ziochos H, Chater MJ, Gosbell IB, Maley MW, van Hal SJ. 2011. Utility of matrix-assisted laser desorption ionization−time of flight mass spectrometry following introduction for routine laboratory bacterial identification. J Clin Microbiol 49:2980–2984. doi: 10.1128/JCM.00431-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dubois D, Grare M, Prere MF, Segonds C, Marty N, Oswald E. 2012. Performances of the MALDI-TOF mass spectrometry system Vitek MS for the rapid identification of bacteria in routine clinical microbiology. J Clin Microbiol 50:2568–2576. doi: 10.1128/JCM.00343-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Martiny D, Busson L, Wybo I, El Haj RA, Dediste A, Vandenberg O. 2012. Comparison of the Microflex LT and Vitek MS systems for routine identification of bacteria by matrix-assisted laser desorption ionization−time of flight mass spectrometry. J Clin Microbiol 50:1313–1325. doi: 10.1128/JCM.05971-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Carbonnelle E, Grohs P, Jacquier H, Day N, Tenza S, Dewailly A, Vissouarn O, Rottman M, Herrmann JL, Podglajen I, Raskine L. 2012. Robustness of two MALDI-TOF mass spectrometry systems for bacterial identification. J Microbiol Methods 89:133–136. doi: 10.1016/j.mimet.2012.03.003. [DOI] [PubMed] [Google Scholar]
- 15.Bizzini A, Durussel C, Bille J, Greub G, Prod'hom G. 2010. Performance of matrix-assisted laser desorption ionization−time of flight mass spectrometry for identification of bacterial strains routinely isolated in a clinical microbiology laboratory. J Clin Microbiol 48:1549–1554. doi: 10.1128/JCM.01794-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tan EK, Ellis BC, Lee R, Stamper PD, Zhang SX, Carroll KC. 2012. Prospective evaluation of a MALDI-TOF MS system in a hospital clinical microbiology laboratory for the identification of bacteria and yeasts: a bench-by-bench study to assess the impact on time-to-identification (TTI) and cost-effectiveness. J Clin Microbiol 50:3301–3308. doi: 10.1128/JCM.01405-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Cherkaoui A, Hibbs J, Emonet S, Tangomo M, Girard M, François P, Schrenzel J. 2010. Comparison of two matrix-assisted laser desorption ionization−time of flight mass spectrometry methods with conventional phenotypic identification for routine identification of bacteria to the species level. J Clin Microbiol 48:1169–1175. doi: 10.1128/JCM.01881-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.van Veen SQ, Claas EC, Kuijper EJ. 2010. High-throughput identification of bacteria and yeast by matrix-assisted laser desorption ionization-time of flight mass spectrometry in conventional medical microbiology laboratories. J Clin Microbiol 48:900–907. doi: 10.1128/JCM.02071-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Bizzini A, Jaton K, Romo D, Bille J, Prod'hom G, Greub G. 2011. Matrix-assisted laser desorption ionization-time of flight mass spectrometry as an alternative to 16S rRNA gene sequencing for identification of difficult-to-identify bacterial strains. J Clin Microbiol 49:693–696. doi: 10.1128/JCM.01463-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Marko DC, Saffert RT, Cunningham SA, Hyman J, Walsh J, Arbefeville S, Howard W, Pruessner J, Safwat N, Cockerill FR, Bossler AD, Patel R, Richter SS. 2012. Evaluation of the Bruker Biotyper and Vitek MS matrix-assisted laser desorption ionization−time of flight mass spectrometry systems for identification of nonfermenting Gram-negative bacilli isolated from cultures from cystic fibrosis patients. J Clin Microbiol 50:2034–2039. doi: 10.1128/JCM.00330-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Alby K, Gilligan PH, Miller MB. 2013. Comparison of matrix-assisted laser desorption ionization−time of flight (MALDI-TOF) mass spectrometry platforms for the identification of Gram-negative rods from patients with cystic fibrosis. J Clin Microbiol 51:3852–3854. doi: 10.1128/JCM.01618-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Saffert RT, Cunningham SA, Ihde SM, Jobe KE, Mandrekar J, Patel R. 2011. Comparison of Bruker Biotyper matrix-assisted laser desorption ionization−time of flight mass spectrometer to BD Phoenix automated microbiology system for identification of gram-negative bacilli. J Clin Microbiol 49:887–892. doi: 10.1128/JCM.01890-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Haag AM, Taylor SN, Johnston KH, Cole RB. 1998. Rapid identification and speciation of Haemophilus bacteria by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. J Mass Spectrom 33:750–756. doi:. [DOI] [PubMed] [Google Scholar]
- 24.Frickmann H, Christner M, Donat M, Berger A, Essig A, Podbielski A, Hagen RM, Poppert S. 2013. Rapid discrimination of Haemophilus influenzae, H. parainfluenzae, and H. haemolyticus by fluorescence in situ hybridization (FISH) and two matrix-assisted laser-desorption-ionization time-of-flight mass spectrometry (MALDI-TOF-MS) platforms. PLoS One 8:e63222. doi: 10.1371/journal.pone.0063222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Couturier MR, Mehinovic E, Croft AC, Fisher MA. 2011. Identification of HACEK clinical isolates by matrix-assisted laser desorption ionization-time of flight mass spectrometry. J Clin Microbiol 49:1104–1106. doi: 10.1128/JCM.01777-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Bruin JP, Kostrzewa M, van der Ende A, Badoux P, Jansen R, Boers SA, Diederen BM. 2014. Identification of Haemophilus influenzae and Haemophilus haemolyticus by matrix-assisted laser desorption ionization-time of flight mass spectrometry. Eur J Clin Microbiol Infect Dis 33:279–284. doi: 10.1007/s10096-013-1958-x. [DOI] [PubMed] [Google Scholar]
- 27.Ilina EN, Borovskaya AD, Malakhova MM, Vereshchagin VA, Kubanova AA, Kruglov AN, Svistunova TS, Gazarian AO, Maier T, Kostrzewa M, Govorun VM. 2009. Direct bacterial profiling by matrix-assisted laser desorption-ionization time-of-flight mass spectrometry for identification of pathogenic Neisseria. J Mol Diagn 11:75–86. doi: 10.2353/jmoldx.2009.080079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Zangenah S, Guleryuz G, Borang S, Ullberg M, Bergman P, Ozenci V. 2013. Identification of clinical Pasteurella isolates by MALDI-TOF—a comparison with Vitek 2 and conventional microbiological methods. Diagn Microbiol Infect Dis 77:96–98. doi: 10.1016/j.diagmicrobio.2013.06.024. [DOI] [PubMed] [Google Scholar]
- 29.Clinical and Laboratory Standards Institute. 2008. Interpretative criteria for identification of bacteria and fungi by DNA target sequencing; approved guideline. CLSI document MM18-A. Clinical and Laboratory Standards Institute, Wayne, PA. [Google Scholar]
- 30.Schulthess B, Brodner K, Bloemberg GV, Zbinden R, Böttger EC, Hombach M. 2013. Identification of Gram-positive cocci by use of matrix-assisted laser desorption−time of flight mass spectrometry: comparison of different preparation methods and implementation of a practical algorithm for routine diagnostics. J Clin Microbiol 51:1834–1840. doi: 10.1128/JCM.02654-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Branda JA, Rychert J, Burnham CA, Bythrow M, Garner OB, Ginocchio CC, Jennemann R, Lewinski MA, Manji R, Mochon AB, Procop GW, Richter SS, Sercia LF, Westblade LF, Ferraro MJ. 2014. Multicenter validation of the Vitek MS v2.0 MALDI-TOF mass spectrometry system for the identification of fastidious Gram-negative bacteria. Diagn Microbiol Infect Dis 78:129–131. doi: 10.1016/j.diagmicrobio.2013.08.013. [DOI] [PubMed] [Google Scholar]
- 32.Powell EA, Blecker-Shelly D, Montgomery S, Mortensen JE. 2013. Application of matrix-assisted laser desorption ionization−time of flight mass spectrometry for identification of the fastidious pediatric pathogens Aggregatibacter, Eikenella, Haemophilus, and Kingella. J Clin Microbiol 51:3862–3864. doi: 10.1128/JCM.02233-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Schulthess B, Ledermann R, Mouttet F, Zbinden A, Bloemberg GV, Böttger EC, Hombach M. 2014. Use of the Bruker MALDI Biotyper for identification of molds in the clinical mycology laboratory. J Clin Microbiol 52:2797–2803. doi: 10.1128/JCM.00049-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Khot PD, Couturier MR, Wilson A, Croft A, Fisher MA. 2012. Optimization of matrix-assisted laser desorption ionization−time of flight mass spectrometry analysis for bacterial identification. J Clin Microbiol 50:3845–3852. doi: 10.1128/JCM.00626-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Freiwald A, Sauer S. 2009. Phylogenetic classification and identification of bacteria by mass spectrometry. Nat Protoc 4:732–742. doi: 10.1038/nprot.2009.37. [DOI] [PubMed] [Google Scholar]