- Email: [email protected]
Differentiation of Flavobacterium psychrophilum from Flavobacterium psychrophilum-like species by MALDI-TOF mass spectrometry
Research in Veterinary Science 115 (2017) 345–352
Contents lists available at ScienceDirect
Research in Veterinary Science journal homepage: www.elsevier.com/locate/rvsc
Differentiation of Flavobacterium psychrophilum from Flavobacterium psychrophilum-like species by MALDI-TOF mass spectrometry
MARK
Marta Pérez-Sanchoa, Ana Isabel Velaa,b, Tom Wiklundc, Markus Kostrzewad, Lucas Domíngueza,b, José Francisco Fernández-Garayzábala,b,⁎ a
Centro de Vigilancia Sanitaria Veterinaria (VISAVET), Universidad Complutense de Madrid, Avenida Puerta de Hierro s/n, 28040 Madrid, Spain Departamento de Sanidad Animal, Facultad de Veterinaria, Universidad Complutense, Avenida Puerta de Hierro s/n, 28040 Madrid, Spain c Laboratory of Aquatic Pathobiology, Environmental and Marine Biology, Åbo Akademi University, BioCity, Artillerigatan 6, 20520 Åbo, Finland d Bruker Daltonik GmbH, Bremen, Germany b
A R T I C L E I N F O
A B S T R A C T
Keywords: Flavobacterium psychrophilum MALDI-TOF MS Identification Differentiation Diagnosis
Rainbow trout fry syndrome (RTFS) is an important infectious disease caused by Flavobacterium psychrophilum affecting farmed salmonids worldwide. Other Flavobacterium psychrophilum-like species (F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis) have been isolated from diseased rainbow trout fry suspected of RTFS although the epidemiological and clinical relevance of these pathogens are unknown. The objective of this study was to evaluate the potential use of MALDI-TOF (Matrix-Assisted Laser Desorption/Ionization-Time of Flight) Mass Spectrometry as method for specific identification of F. psychrophilum and its differentiation from other F. psychrophilum-like species isolated from diseased fish. Fifty-three isolates were analyzed after the creation of the Main Spectrum Profile (MSP) of reference strains of each of abovementioned species. F. psychrophilum exhibited a mass spectra very different from those of F. psychrophilum-like species, with five peaks (m/z 3654, 4585, 5388, 6730 and 7310) present only in F. psychrophilum isolates, and three peaks (m/z 6170, 7098 and 9241) absent in F. psychrophilum but present in all F. psychrophilum-like species. All F. psychrophilum isolates were correctly identified and differentiated from the F. psychrophilum-like species by MALDI-TOF. Although this approach showed a limited ability to differentiate among F. psychrophilum-like species, its complementation with a few simple biochemical tests may represent an alternative approach for the routine identification of the Flavobacterium psychrophilum-like species.
1. Introduction Flavobacterium psychrophilum is an important bacterial pathogen of wild and farmed salmonids worldwide and the etiological agent of rainbow trout (Oncorhynchus mykiss) fry syndrome (RTFS; Austin and Stobie, 1991) and bacterial cold water disease (BCWD; Wood and Yasutake, 1956). In RTFS, F. psychrophilum infections cause high mortalities and significant losses in the affected fish populations (Nematollahi et al., 2003). In recent years, several new described Flavobacterium species (F. plurextorum, F. tructae, F. collinsii, F. oncorhynchi and F. piscis) have been isolated from diseased rainbow trout fry that present clinical symptoms consistent with F. psychrophilum infection (Zamora et al., 2012; Zamora et al., 2013a; Zamora et al., 2013b; Zamora et al., 2014). For this reason and the resemblance in the macroscopic characteristics of their colonies on Anacker and Ordal agar with those of F. psychrophilum and similar biochemical characteristics (Zamora, 2015), these species are considered under the generic name of ⁎
F. psychrophilum-like species in the present work. The immunological and molecular methods available for the detection and identification of F. psychrophilum are useful for an accurate identification of RTFS but not for identification of the etiological agent of those episodes in which F. psychrophilum-like species are involved. These F. psychrophilum-like species can be differentiated from F. psychrophilum by several biochemical characteristics (e.g. growth at 25 °C, degradation of aesculin and carbohydrates and growth on trypticase soy agar) or sequencing of their 16S rRNA gene (Zamora, 2015). However, conventional phenotypic identification methods are labor-intensive and time-consuming and costs of DNA sequencing generally are too high for its routine use in veterinary diagnostic laboratories. These limitations can be overcome by MALDI-TOF MS approach, a useful tool, which has become a rapid, accurate, and cost-effective identification technique in clinical microbiology laboratories (Patel, 2015; Singhal et al., 2015). In the veterinary field, the method has been successfully applied for the identification of a wide range of bacterial pathogens (Randall et al.,
Corresponding author at: Departamento de Sanidad Animal, Facultad de Veterinaria, Universidad Complutense, Avenida Puerta de Hierro s/n, 28040 Madrid, Spain. E-mail address: [email protected] (J.F. Fernández-Garayzábal).
http://dx.doi.org/10.1016/j.rvsc.2017.06.022 Received 22 December 2016; Received in revised form 9 June 2017; Accepted 28 June 2017 0034-5288/ © 2017 Elsevier Ltd. All rights reserved.
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
2015) including several bacterial fish pathogens (Beaz-Hidalgo et al., 2009; Böhme et al., 2010; Erler et al., 2015; Kurokawa et al., 2013; Regecová et al., 2014; Strepparava, 2012). However, to our knowledge, there is only one study evaluating the suitability of MALDI-TOF for identification of F. psychrophilum but other Flavobacterium species associated with infection in fish were not included (Strepparava, 2012). The aim of this work is to evaluate the suitability of MALDI-TOF MS for the identification of F. psychrophilum and its differentiation from other F. psychrophilum-like species (F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis) that can be isolated from outbreaks with presumptive diagnosis of F. psychrophilum infections based on clinical symptoms.
Table 1 Potential species-specific MALDI-TOF peaks (mass/charge, m/z) used for differentiation of Flavobacterium psychrophilum and Flavobacterium psychrophilum-like species included in the present study. m/z
3654 4585 5388 6730 7310 6170
2. Materials and methods 2.1. Bacterial isolates and culture conditions 7098
A total of 64 Flavobacterium isolates were included in the present study. A first set of isolates (n = 11) was used to create Main Spectra (MSPs) for supplementation of the Bruker MALDI-TOF MS database (version 3.4; 5989 entries): F. psychrophilum (JIP02/86 and H7/00), F. plurextorum (1126-1H-08T, 424-08), F. oncorhynchi (631-08T, 662-09), F. tructae (435-08T) F. collinsii (983-08T and 978B-08) and F. piscis (41209T and 60B-3-09). A second set of isolates (n = 53; Table 2) was used for the external validation of the new MSPs confirming the speciesdifferential ions detected and to assess the suitability of MALDI-TOF technique for the differentiation of F. psychrophilum from the other F. psychrophilum-like species. All isolates, except F. psychrophilum (n = 31), had been part of taxonomic polyphasic studies; therefore, all isolates had been previously characterized by biochemical, chemotaxonomic and genetic methods (Zamora et al., 2012; Zamora et al., 2013a; Zamora et al., 2013b; Zamora et al., 2014). Identification of F. psychrophilum was achieved by biochemical characterization, sequencing of 16S rRNA gene or species-specific PCR (Nilsen et al., 2014). The reference strain JIP02/86 of F. psychrophilum (DSM 21280) was acquired from Leibniz Institute DSMZ (German Collection of Microorganisms and Cell cultures). Clinical isolates were mostly recovered from rainbow trout (n = 50), although a subset of F. psychrophilum isolates (n = 14) was isolated from European perch (Perca fluviatilis), Atlantic salmon (Salmo salar), brook trout (Salvelinus fontinalis), brown trout (Salmo trutta) and water. All isolates used in this study were grown on R2A agar (Reasoner's 2A agar, Reasoner et al., 1979) and incubated for 72 h at 15–22 °C under aerobic conditions.
9241
Flavobacterium sp.a,b
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum collinsii oncorhynchi piscis plurextorum tructae collinsii oncorhynchi piscis plurextorum tructae collinsii oncorhynchi piscis plurextorum tructae
Total isolatesc
31 31 31 31 31 5 18 3 5 2 5 18 3 5 2 5 18 3 5 2
Species-specific peaks Intensity (a.u.)
Signal-tonoise (s/n)
Frequencyd
1727–9916 6510–34,759 3988–16,944 1462–9716 1755–14,152 3202–5121 1617–9060 1744–4581 2795–8694 1426–9515 682–3271 941–4837 2667–5074 1557–3746 1054–5100 848–1731 1406–5413 1435–2112 1574–2888 2604–2708
19–55 67–258 39–109 28–81 36–125 24–72 12–66 18–27 26–52 11–75 6–28 9–35 29–42 16–38 9–45 18–24 17–41 24–28 25–32 33–34
100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 80% 100% 100% 100% 100% 80% 100% 100% 100% 100%
a Flavobacterium species identification determined by biochemical, chemotaxonomic and genetic methods. b F. oncorhynchi and F. plurextorum showed a very similar protein profile. Some additional peaks (m/z 4820, 5055 and 10,110) have been proposed in this study for their differentiation. c Number of isolates per species. d Number of isolates displaying each specific peak.
was overlaid with 1 μL of HCCA matrix solution (saturated solution of α-cyno-4-hydroxycinnamic acid in 50% acetonitrile 2.5% trifluoroacetic acid) and air dried at room temperature. Secondly, the extraction protocol was repeated with a subset of 11 isolates, which were used for the creation of reference spectra (as explained below): protein extracts were deposited onto 8 spots and three subsequent spectra acquisitions were performed to obtain 24 spectra/isolate. Finally, in the case of F. psychrophilum-like isolates, the extraction protocol was repeated and one microliter of the protein extract was spotted onto each of 8 spots. After adding 1 μL of HCCA matrix solution to each spot and let dry, three subsequent spectra acquisitions were carried out from each spot to obtain 24 spectra per sample for cluster analysis.
2.2. Preparation of bacterial cell lysates for MALDI-TOF MS
2.3. MALDI-TOF MS: acquisition and processing of mass spectra
First, cells from bacterial colonies of each isolate were transferred from the agar plate to a 2 mL tube (Eppendorf, Germany) containing 300 μL of HPLC grade-water (Sigma Aldrich, Germany). A total of 900 μL of HPLC-grade (High Performance Liquid Chromatography) absolute ethanol was added to the bacterial suspension. A protein extraction protocol based on formic acid/acetonitrile was subsequently performed in all isolates following the manufacturer's instructions (Bruker Daltonik GmbH, Germany) and previously described elsewhere (Pérez-Sancho et al., 2015). The extraction protocol was repeated twice or three times for different purposes. Firstly, each bacterial suspension was subjected to protein extraction and the extracts were used for bacterial identification using the Bruker BDAL MALDI Biotyper database (5989 entries). Additionally, these spectra were also used to assess the robustness of proposed species-specific biomarkers detected on MSP (Main Spectra) spectra. In this case, 1 μL of the protein extracts containing the bacterial proteins and peptides was spotted onto one spot of a polished steel target plate and let dry at room temperature. Subsequently, each spot
MALDI-TOF spectra were acquired using two different devices: an UltrafleXtreme machine (Bruker Daltonik; used for spectra acquisition for bacterial identification and for evaluation of the robustness of proposed species-specific biomarkers) (Table 1) and a Microflex LT equipment (Bruker Daltonik; used for spectra acquisition for creation of MSPs, evaluation of the presence of species-specific peaks on these reference spectra and cluster analysis). Regardless of the device type, spectra were acquired in the linear and positive mode within a mass range between 2 and 20 kDa. Each spectrum was acquired using the FlexControl software (version 3.4) in an automatic mode and externally calibrated using the Bacterial Test Standard (Bruker Daltonik GmbH, Bremen, Germany). 2.4. Identification based on MALDI Biotyper For the bacterial identification study, one spectra from each clinical Flavobacterium isolate (n = 64) acquired by UltrafleXtreme device was 346
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
Table 2 Identification results using MALDI TOF approach in 53 Flavobacterium isolates recovered from diseased fish after the inclusion of F. psychrophilum and F. psychrophilum-like MSP to construct the fish Flavobacterium MALDI database (FDB). Speciesa
Reference
1 ID - FDB (Score value)b
F. psychrophilum
P4-4/06 1.12.99-47 P11-1B/97 P16-8/97 P4-3/06 K2/99 T1-1 L4/00 P7-9/2R/10 P29-1A/10 P18-2/11 OG-83/28 M2/99 255-ii/93 V43 3474/94 21.10.99-3 H2/00 P11-1A/97 ST9/00 P7-9/10 P44-9/11 P4-2B/06 P56-2A/06 A6/00 P5-1/12 P69-5/97 15,527/2/96 H1/00
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
53B-3-09 991-08 977H-09
F. collinsii (2.518) F. collinsii (2.164) F. collinsii (2.049)
F. collinsii (2.080) F. collinsii (2.031) F. collinsii (2.045)
A A A
F. oncorhynchi
426B-08 628-1-08 646-08 437B-08 441B-08 695B-08 947B-08 646B-08 950B-08 650-08 688B-08 22B-09 425B-08 433B-08 47B-2-09 666-09
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
plurextorum (1.948) oncorhynchi (1.861) plurextorum (1.927) oncorhynchi (1.953) plurextorum (1.907) oncorhynchi (1.865) johnsoniae (1.748) plurextorum (1.921) oncorhynchi (1.867) oncorhynchi (1.677) plurextorum (1.931) oncorhynchi (1.885) oncorhynchi (1.985) oncorhynchi (1.963) plurextorum (2.021) plurextorum (2.024)
A A A A A A A A A A A A D D B B
F. piscis
412R-09
F. piscis (2.281)
F. saccharophilum (2.208)
B
F. plurextorum
1084B-08 51B-09 986-08
F. plurextorum (2.206) F. plurextorum (2.175) F. plurextorum (2.257)
F. plurextorum (2.136) F. plurextorum (2.083) F. plurextorum (2.051)
A A A
F. tructae
47B-3-09
F. hibernum (1.840)
F. tructae (1.824)
D
F. psychrophilum-like F. collinsii
psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum
oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi
Consistencyd
2 ID- FDB (Score value)c (2.644) (2.563) (2.537) (2.588) (2.685) (2.619) (2.514) (2.653) (2.686) (2.648) (2.592) (2.718) (2.521) (2.569) (2.680) (2.611) (2.558) (2.621) (2.698) (2.551) (2.502) (2.588) (2.523) (2.518) (2.523) (2.494) (2.570) (2.532) (2.557)
(2.220) (2.210) (2.015) (2.210) (2.240) (2.217) (2.156) (2.113) (2.180) (2.125) (2.294) (2.200) (1.993) (1.988) (2.120) (2.192)
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum
(2.096) (2.072) (2.209) (1.885) (2.094) (2.057) (2.030) (2.197) (2.026) (2.083) (2.151) (2.049) (2.139) (1.965) (2.127) (2.212) (1.692) (2.281) (1.832) (1.760) (2.013) (2.045) (2.026) (2.002) (1.928) (1.940) (2.007) (2.141) (2.142)
A A A A A A A A A A A A A A A A A A A A A A A A A A A A A
a
Flavobacterium species determined by biochemical, chemotaxonomic and genetic methods. Best match in the ranking list of MALDI TOF identification results and corresponding score value provided by MALDI Biotyper (Bruker Daltonik, Germany). c Second match in the ranking list of MALDI TOF identification results and corresponding score value provided by MALDI Biotyper (Bruker Daltonik, Germany). d Consistency ranking list of the first 5 best matches = A, the correct species is the unique species with score value ≥ 2.000; B, the correct species is first ranked but a different species in the second rank also with score value ≥ 2.000; D, first and second matches have score values < 2.000 (correct genus identification). b
Flavobacterium saccharophilum). After MSP creation using the 11 Flavobacterium isolates as described above, the remaining isolates (n = 53) were identified using the fish Flavobacterium databases (FDB, 5989 entries of BDAL Biotyper Database plus eleven new Flavobacterium MSPs, Table 2). The Biotyper software showed 10 top matches (as defined by the user) in an identification ranking list ordered according to
processed using the MALDI Biotyper RTC software (version 3.1) with default settings in the automatic mode. All spectra were identified using the Bruker BDAL MALDI Biotyper (5989 entries; which includes Flavobacterium flevense, Flavobacterium gelidilacus, Flavobacterium hibernum, Flavobacterium hydatis, Flavobacterium johnsoniae, Flavobacterium lindanitolerans, Flavobacterium pectinovorum and 347
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
To incorporate the MSP of F. psychrophilum and F. psychrophilum-like species in the Bruker database (version 3.4; 5989 entries), two wellcharacterized isolates of each species (except one isolate in F. tructae) were analyzed. Analysis of the spectra of these species was performed by selecting a range of m/z between 3000 and 10,000 Da that included the major differences between their MSP. The visual inspection of these spectra using ClinProTools revealed important differences between the mass spectra of F. psychrophilum and those of the F. psychrophilum-like species (Table 1). Thus, peaks 3654, 4585, 5388, 6730 and 7310 m/z were present in F. psychrophilum strains, while peaks 6170, 7098 and 9241 m/z, present in all F. psychrophilum-like species, were absent in F. psychrophilum. These distinctive peaks were also observed by visual inspection using FlexAnalysis software (data not shown). Data of the intensity, signal-to-noise and frequency of detection of these ions in the 64 isolates of the different Flavobacterium species is shown in Table 1. Most peaks were identified in 100% of the isolates examined with average intensity and signal-to-noise values of > 600 and > 6, respectively. A cluster analysis of MSP obtained from the 11 initial Flavobacterium isolates based on the construction of a dendrogram using MALDI Biotyper OffLine Classification software (version 3.1) demonstrated that F. psychrophilum spectra form a well-defined clade separated from those formed by the F. psychrophilum-like species and the other Flavobacterium species included currently in the MALDI Biotyper database (Fig. 1). Visual inspection showed that the spectra of the F. psychrophilumlike species were in general quite similar except for F. piscis that showed one indicative peak at 5418 m/z (Table 1). Visual inspection using ClinProTools and FlexAnalysis of the mass patterns of F. oncorhynchi and F. plurextorum revealed almost no differences in their spectra. Similarly, F. tructae and F. collinsii presented very similar mass patterns (Fig. 1). Fig. 3A is a strain distribution map base on the spectra of all 33 F. psychrophilum-like isolates showing the close relationship between F. plurextorum and F. oncorhynchi and between F. tructae and F. collinsii. These relationships reflect the similarity in their MALDI profiles observed by visual inspection and may suggest potential problems of misidentification among these species. Results of PCA score plot confirmed these close relatedness (Fig. 3B). After the MSPs of F. psychrophilum and F. psychrophilum-like species were incorporated in the Bruker database, the second set of isolates (n = 53) of these species recovered from clinical specimens of diseased fish was again identified by MALDI-TOF MS (Table 2). All F. psychrophilum isolates were correctly identified with a mean score value of 2.588 and a consistency identification category A that indicate the absence of misidentifications with any of the F. psychrophilum-like species. Regarding isolates of the F. psychrophylum-like species, the top hit of the ranking identification list were in accordance with the biochemical, chemotaxonomic and/or genetic results in all isolates (except DICM09/0047B-3) (Table 2) (n = 23; 95.83%). Most of correctly identified isolates (91.3%; n = 21) gave score values > 2.0 in the first option in the identification ranking list provided by the MALDI Biotyper. The single strain of F. tructae was identified as F. hibernum or F. tructae in the first and second options in the identification-ranking list (Table 2) with score values of 1.84 and 1.824, respectively, which indicate a reliable identification at genus level but only a low confidence species identification. A closer analysis of the mass peak profiles of both strains of F. tructae (one used for MSP creation and the other used in the blind panel) revealed important differences between them (e.g. m/z 3084, 3728 and 6169 in strain 435-09T and m/z 3716, 7433 and 7608 in strain 47B-3-09) suggesting a potentially high intraspecies variability. Analysis of further isolates of this species in the future will allow confirmation whether these differences are common in F. tructae. The single strain of F. piscis was correctly identified by MALDI-TOF with score value of 2.281 in the first match of the identification-ranking list. However, F. saccharophilum was the identification result on the second position of the raking list with a very close score value of 2.208
the score value of each identification. For interpretation of the identification results, two variables were considered: (i) score values following the MALDI Biotyper Compass (≥ 2.000: identification at species level with high confidence; 1.999–1.800: identification with low confidence and ≤ 1799: non-reliable identification) and (ii) consistency ranking list of the first 2 best matches [A, the correct species is the unique species with score value ≥ 2.000; B, the correct species is first ranked but a different species in the second rank also with score value ≥ 2.000; C, first and second matches have score values ≥2.000 but the correct species is second ranked; D, first and second matches have score values < 2.000 (correct genus identification)]. 2.5. Evaluation of the presence of species-identifying biomarkers and creation of MSP for F. psychrophilum, F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis Visual inspection of spectra of all Flavobacterium isolates was performed using the pseudogel view option of ClinProTools program (version 3.0). Those potential species-identifying ions were visually confirmed using FlexAnalysis software (Bruker Daltonik). Analysis of the presence of candidate differential biomarkers for each Flavobacterium species was confirmed by ClinProTools (version 3.0). This software allows the data processing and spectra preparation realignment to reduce measurements variations. The pseudogel view option of ClinProTools allows evaluating the presence/absence of potential species-specific ions. Finally, a principal component analysis (PCA) was created using 24 spectra/isolate of all F. psychrophilum-like species. PCA is based on the Euclidian distance method and was constructed using ClinProTools software. 2.6. Construction of fish Flavobacterium MALDI database (FDB) The construction of the FDB was carried out according to manufacture's instructions. A total of 11 new MSP were incorporated in the Bruker BDAL MALDI Biotyper database available that time (5989 entries): two new entries per species were included except in the case of F. tructae in which only one entry was included due to the limitation in the number of isolates available. MSP construction was performed using 20–24 spectra for each Flavobacterium isolate (three measurements at 8 different spots). The quality and reproducibility of all spectra included in each MSP were assessed using FlexAnalysis software (version 3.4, Bruker Daltonik) after smoothing, normalization, baseline subtraction and peak picking. All MSP were created by MALDI Biotyper Compass Explorer software (version 4.1.16, Bruker Daltonik) using the MSP series creation option. To determine if MALDI profiles of Flavobacterium species were distributed in different clades, we combined MSP of F. psychrophilum, F. psychrophilum-like species, and those Flavobacterium species currently present in the Biotyper database for cluster analysis (MSP based dendrogram construction using default parameter set) using MALDI Biotyper software. 3. Results Before the incorporation of the MSP in the Bruker database, the 64 Flavobacterium isolates were analyzed as a preliminary test of the suitability of MALDI-TOF MS for the identification of Flavobacterium species. Most isolates of the Flavobacterium species included in the study (n = 49; 79%) showed a mean score value of < 1.699, indicating an unreliable identification and consequently their unlikely misidentification with any of those species currently included in this database (F. flevense, F. gelidilacus, F. hibernum, F. hydatis, F. johnsoniae, F. lindanitolerans, F. pectinovorum and F. saccharophilum). Twelve isolates gave score values between 1.999 and 1.800 indicative of a correct identification at low confidence level. The three isolates of F. piscis were identified as F. saccharophilum with scores values > 2.000, which suggests a misidentification at the high confidence species level. 348
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
Fig. 1. Dendrogram derived from MSP of Flavobacterium species analyzed in this study and those currently included in the MALDI Biotyper database (5989 entries) demonstrates that F. psychrophilum and F. psychrophilum-like species grouped in different clades (relative distance values are normalized to a maximum value = 1000).
(Seng et al., 2009; Cherkaoui et al., 2010). In the same way, the speed of bacterial identification by MALDI-TOF is estimated to be few minutes compared with hours or days required for conventional methods (Seng et al., 2009; Cherkaoui et al., 2010). In this study, we have evaluated the usefulness of MALDI-TOF MS for identifying F. psychrophilum and discriminating it from other F. psychrophilum-like species, which can be relevant from a diagnosis and treatment point of view. Neither F. psychrophilum nor F. psychrophilum-like species analyzed in this study are included among the Flavobacterium species present in the MALDI-TOF database (5989 entries). An initial attempt to identify these species by MALDI-TOF was unsuccessful as score values for most species were < 1.9. Only the isolates of F. piscis gave scores > 2.000 with F. saccharophilum indicating a probable misidentification with this species. These data confirm that, with the exception of F. piscis, all other Flavobacterium species were not misidentified with any of the species of Flavobacterium included in the Bruker database. This highlights the need to supplement specific MSP of these species to assess the suitability of MALDI approach for these fish pathogens as previously observed for other microorganisms (Calderaro et al., 2014; Christensen et al., 2012; De Carolis et al., 2014). Therefore, the first objective of this study was to construct the MSP of the reference strains of F. psychrophilum and F. psychrophilum-like species. The new MSP of Flavobacterium here created are at the disposal of Bruker Daltoniks to be included in further versions of Bruker Biotyper database. Visual inspection of generated MSP revealed very different mass spectra exhibited by F. psychrophilum isolates compared to those of the F. psychrophilum-like species, with six conserved and stable peaks (m/z 3654, 4585, 5388, 6730 and 7310) present in all F. psychrophilum isolates, and three peaks (m/z 6170, 7098 and 9241) absent in F. psychrophilum but present in all isolates of F. psychrophilum-like species (Table 1, Fig. 2). Most of these distinctive peaks exhibited significant intensity and signal-to-noise values (> 600 and > 6, respectively) and they were identified in all isolates. These results confirmed their robustness and specificity and, therefore, their potential usefulness as species-specific biomarkers for the differentiation of F. psychrophilum
indicating an inconclusive identification (B category consistency identification). MALDI identification results were in accordance with those observed for genetic and biochemical methods for F. collinsii showing category A consistency identification. The three isolates of F. plurextorum were also correctly identified with a consistency identification category A and average score values of 2.216 (Table 2). Most isolates of F. oncorhynchi were correctly identified (category A; Table 2) but in two strains the second identification option was F. plurextorum also with scores values > 2.0 (category B) and another two isolates were identified only to low confidence level (category D). 4. Discussion The existence of different Flavobacterium species (F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis), which may cause RTFSlike disease on rainbow trout, may complicate the species-specific identification of F. psychrophilum invalidating those immunologic and molecular methods available for this purpose (Zamora, 2015). These species have been also isolated from gills and several internal organs (brain and kidney) of different fish species (Zamora et al., 2012; Zamora et al., 2013a; Zamora et al., 2013b; Zamora et al., 2014; Loch and Faisal, 2015), and their identification is based upon near-to-complete sequencing of their 16S rRNA gene (Loch et al., 2013; Loch and Faisal, 2014); which is a technique that is laborious, expensive and relatively time consuming. Consequently, it is necessary to develope diagnostic tools not only for the rapid differentiation of F. psychrophilum from F. psychrophilum-like species, but also for the species-specific identification of the latter. In this context, MALDI-TOF MS can be a valuable alternative. The method has been recently become routine in some diagnostic laboratories for the identification of many bacterial pathogens. Different studies have assessed the improvements of implementing MALDI-TOF technique on the routine laboratory diagnosis (Seng et al., 2009; Cherkaoui et al., 2010; Bizzini and Greub, 2010). In this sense, it has been calculated that the cost of this proteomic approach to be 17–32% of the cost of automated identification system 349
Fig. 2. Differential species-specific peaks on the MALDI-TOF MS spectra of whole-cell extracts of F. psychrophilum and F. oncorhynchi. F. psychrophilum-like species exhibited very similar mass patterns and the spectra of F. oncorhynchi is given as an example of the other F. psychrophilum-like species (A). In inserts, an enlarged spectrum of 3500–5550 m/z (B), 6000–7500 m/z (C) and 9000–10,500 m/z (D) to highlight the differential peaks between F. psychrophilum and F. oncorhynchi spectra.
M. Pérez-Sancho et al.
Research in Veterinary Science 115 (2017) 345–352
350
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
Fig. 3. Cluster analysis of F. psychrophilum-like species [F. oncorhynchi (○), F. plurextorum (◊), F. collinsii (×), F. tructae (Δ) and F. piscis (□)]. 2D peak distribution diagram (A) and PCA plot (B) illustrating the close relation between F. oncorhynchi/F. plurextorum and F. collinsii/F. tructae.
The evaluation of MALDI profiles of F. oncorhynchi and F. plurextorum was only performed to assess the discriminatory power of species differentiating-ions because the number of isolates of these species was not high enough for further statistical analysis. However, the rigorous visual inspection of their spectra demonstrated the presence of some mass-peaks (m/z 4822, 5057, 6902, 9640 and 10,112 for F. oncorhynchi and m/z 4828, 5066, 9654 and 10,127 for F. plurextorum; data not shown) potentially useful for their differentiation. Future isolates belonging to these species may allow the validation of the proposed differentiating-peaks to confirm their suitability for the rapid and reliable differentiation of F. oncorhynchi and F. plurextorum allowing the construction of classification models for their automatic differentiation complementing the standard identification algorithm. Meanwhile, for isolates of species for which the two highest Biotyper match scores are > 2.0 MALDI-TOF can be used in combination with simple complementary tests to achieve an unequivocal identification. Thus, isolates of F. oncorhynchi can be easily differentiated from F. plurextorum by the ability of the former species to produce the enzyme β-galactosidase (Zamora et al., 2012; Zamora et al., 2013a). Similarly, F. piscis can be differentiated from F. saccharophilum by the ability of the latter species to produce the enzymes β-galactosidase and DNAse (Bernardet et al., 1996; Zamora et al., 2014).
from the F. psychrophilum-like species. The clear differentiation of F. psychrophilum and F. psychrophilum-like species after cluster analysis of their MSP (Fig. 1) indicates that MALDI-TOF approach represents an adequate alternative for their differentiation. In agreement with these data, MALDI-TOF MS correctly identified all F. psychrophilum isolates with score values always higher than 2.3 for the best match in Biotyper identification list (average 2.588; Table 2). F. psychrophilum isolates reached a category A consistency identification, further indicating it was clearly differentiated from the F. psychrophilum-like species that exhibited always score values < 2.0. Therefore, MALDI-TOF MS represents an easy and accurate alternative to the immunological and genetic methods currently available for the identification of F. psychrophilum. On the other hand, visual inspection of MSP generated by F. psychrophilum-like species revealed they were very similar, in particular the mass spectra of F. plurextorum and F. oncorhynchi and the spectra of F. collinsii and F. tructae after 2D peak distribution diagram and PCA plot analysis (Fig. 3). The close relatedness between these species is congruent with the high similarity of their 16S rRNA gene sequences (Zamora et al., 2012; Zamora et al., 2013a; Zamora et al., 2013b; Zamora et al., 2014). The utility of MALDI-TOF MS for the identification of the F. psychrophilum-like species was variable. Isolates of F. plurextorum and F. collinsii were correctly identified with category A consistency identification and score values > 2.0 (average scores of 2.212 and 2.242, respectively, Table 2). Similarly, most isolates (n = 12) of F. oncorhynchi were correctly identified as at least the two best matches generated by Biotyper indicated this species or second best matches indicating a different species was lower than 2.0. The isolate of F. tructae and two isolates of F. oncorhynchi were identified correctly only to genus level with scores values < 2.0 (low confidence, D category consistency identification; Table 2). The resolution of MALDI-TOF MS was not high enough to discriminate between some species (Table 2) even though the best matching score was the correct species, as it can be observed by comparing the corresponding score values. Thus, the discrimination was not clear cut for the isolate of F. piscis that was identified in the first best match in Biotyper, but the second best match was F. saccharophilum also with a score value > 2.0 (B category consistency identification; Table 2). Similar observation was made with two isolates of F. oncorhynchi in which the second match with scores values > 2.0 pointed to F. plurextorum (Table 2). These four species have been isolated from diseased fish and the limited discrimination of MALDI-TOF MS for their differentiation could represent a challenge to diagnostic laboratories.
5. Conclusion MALDI-TOF MS seems to be a reliable and alternative approach for the accurate identification of F. psychrophilum. In addition, MALDI-TOF MS, used alone or complemented with a few simple biochemical tests, can be used for the routine identification of different Flavobacterium species (F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis) that can be isolated from diseased fish suspected of F. psychrophilum infections, facilitating the differential diagnosis. The use of MALDI-TOF MS in veterinary diagnostic laboratories would advance knowledge of the epidemiological and clinical significance of these flavobacteria. Role of the funding source This study was supported by the project Advance Technologies in Health Surveillance [S2013/ABI-2747 (TAVS-CM)] funded by the Regional Government of Madrid and the European Union. Marta PérezSancho is recipient of a technical support staff for scientific infrastructures of the Moncloa Campus of International Excellence (Programa CEI09-0019) and Ministerio de Educación, Cultura y 351
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
phylogenetic analyses and MALDI Biotyper identification system. Mar. Biotechnol. (NY) 15, 340–348. Loch, T.P., Faisal, M., 2014. Deciphering the biodiversity of fish-pathogenic Flavobacterium spp. recovered from the Great Lakes basin. Dis. Aquat. Org. 112, 45–57. Loch, T.P., Faisal, M., 2015. Emerging flavobacterial infections in fish: a review. J. Adv. Res. 6, 283–300. Loch, T.P., Fujimoto, M., Woodiga, S.A., Walker, E.D., Marsh, T.L., Faisal, M., 2013. Diversity of fish-associated flavobacteria of Michigan. J. Aquat. Anim. Health 25, 149–164. Nematollahi, A., Decostere, A., Pasmans, F., Haesebrouck, F., 2003. Flavobacterium psychrophilum infections in salmonid fish. J. Fish Dis. 26, 563–574. Nilsen, H., Sundell, K., Duchaud, E., Nicolas, P., Dalsgaard, I., Madsen, L., Aspán, A., Jansson, E., Colquhoun, D.J., Wiklund, T., 2014. Multilocus sequence typing identifies epidemic clones of Flavobacterium psychrophilum in Nordic countries. Appl. Environ. Microbiol. 80, 2728–2736. Patel, R., 2015. MALDI-TOF MS for the diagnosis of infectious diseases. Clin. Chem. 61, 100–111. Pérez-Sancho, M., Vela, A.I., García-Seco, T., Gottschalk, M., Domínguez, L., FernándezGarayzábal, J.F., 2015. Assessment of MALDI-TOF MS as alternative tool for Streptococcus suis identification. Front.Public Health 3, 202. Randall, L.P., Lemma, F., Koylass, M., Rogers, J., Ayling, R.D., Worth, D., Klita, M., Steventon, A., Line, K., Wragg, P., Muchowski, J., Kostrzewa, M., Whatmore, A.M., 2015. Evaluation of MALDI-TOF as a method for the identification of bacteria in the veterinary diagnostic laboratory. Res. Vet. Sci. 101, 42–49. Reasoner, D.J., Blannon, J.C., Geldreich, E.E., 1979. Rapid seven-hour fecal coliform test. Appl. Environ. Microbiol. 38, 229–236. Regecová, I., Pipová, M., Jevinová, P., Marušková, K., Kmeť, V., Popelka, P., 2014. Species identification and antimicrobial resistance of coagulase-negative staphylococci isolated from the meat of sea fish. J. Food Sci. 79, 898–902. Seng, P., Drancourt, M., Gouriet, F., La Scola, B., Fournier, P.E., Rolain, J.M., Raoult, D., 2009. Ongoing revolution in bacteriology: routine identification of bacteria by matrix-assisted laser desorption ionization time-of-flight mass spectrometry. Clin. Infect. Dis. 49, 543–551. Singhal, N., Kumar, M., Kanaujia, P.K., Virdi, J.S., 2015. MALDI-TOF mass spectrometry: an emerging technology for microbial identification and diagnosis. Front. Microbiol. 6, 791. Strepparava, N., 2012. Novel Approaches to the Diagnosis of Flavobacteriosis: Identification of Pathogenic Flavobacteriaceae and Epidemiology of Flavobacterium psychrophilum. Universität Bern, PhD dissertation. Wood, E.M., Yasutake, W.T., 1956. Histopathology of fish. III. Peduncle (‘coldwater’) disease. Progress. Fish Cult. 18, 58–61. Zamora, L., 2015. Estudio Taxonómico de Bacterias Relacionadas con el Síndrome del Alevín de la Trucha (Taxonomic study of bacteria associated with rainbow trout fry syndrome). (PhD dissertation) Universidad Complutense de Madrid. Zamora, L., Fernández-Garayzábal, J.F., Svensson-Stadler, L.A., Palacios, M.A., Domínguez, L., Moore, E.R., Vela, A.I., 2012. Flavobacterium oncorhynchi sp. nov., a new species isolated from rainbow trout (Oncorhynchus mykiss). Syst. Appl. Microbiol. 35, 86–91. Zamora, L., Fernández-Garayzábal, J.F., Sánchez-Porro, C., Palacios, M.A., Moore, E.R., Domínguez, L., Ventosa, A., Vela, A.I., 2013a. Flavobacterium plurextorum sp. nov. isolated from farmed rainbow trout (Oncorhynchus mykiss). PLoS One 8, e67741. Zamora, L., Vela, A.I., Sánchez-Porro, C., Palacios, M.A., Domínguez, L., Moore, E.R., Ventosa, A., Fernández-Garayzábal, J.F., 2013b. Characterization of flavobacteria possibly associated with fish and fish farm environment. Description of three novel Flavobacterium species: Flavobacterium collinsii sp. nov., Flavobacterium branchiarum sp. nov., and Flavobacterium branchiicola sp. nov. Aquaculture 416–417, 346–353. Zamora, L., Vela, A.I., Sánchez-Porro, C., Palacios, M.A., Moore, E.R., Domínguez, L., Ventosa, A., Fernández-Garayzábal, J.F., 2014. Flavobacterium tructae sp. nov. and Flavobacterium piscis sp. nov., isolated from farmed rainbow trout (Oncorhynchus mykiss). Syst. Appl. Microbiol. 64, 392–399.
Deporte (Orden PRE/1996/2009 de 20 de Julio). Declaration of interest Markus Kostrzewa is an employee of Bruker Daltonik GmbH (Bremen, Germany). Acknowledgments The authors wish to thank Almudena Casamayor and Elisa Pulido for their invaluable technical assistance and Sergio González for his kind help with the design of figures. We are grateful to Dr. V. HirveläKoski, Dr. P. Rintamäki and Dr. S. Viljamaa-Dirks for kindly providing some of the isolates used in this study. References Austin, B., Stobie, M., 1991. Recovery of yellow pigmented bacteria from dead and moribund fish during outbreaks of rainbow trout, Oncorhynchus mykiss (Walbaum), fry syndrome in England. J. Fish Dis. 14, 677–682. Beaz-Hidalgo, R., Alperi, A., Figueras, M.J., Romalde, J.L., 2009. Aeromonas piscicola sp. nov., isolated from diseased fish. Syst. Appl. Microbiol. 32, 471–479. Bernardet, J.F., Segers, P., Vancanneyt, M., Berthe, F., Kersters, K., Vandamme, P., 1996. Cutting a Gordian knot: emended classification and description of the genus Flavobacterium, emended description of the family Flavobacteriaceae, and proposal of Flavobacterium hydatis nom. nov. (basonym, Cytophaga aquatilis Strohl and Tait 1978). Int. J. Syst. Evol. Microbiol. 46, 128–148. Bizzini, A., Greub, G., 2010. Matrix-assisted laser desorption ionization time-of-flight mass spectrometry, a revolution in clinical microbial identification. J. Clin. Microbiol. 48, 1549–1554. Böhme, K., Fernández-No, I.C., Barros-Velázquez, J., Gallardo, J.M., Calo-Mata, P., Cañas, B., 2010. Species differentiation of seafood spoilage and pathogenic gram-negative bacteria by MALDI-TOF mass fingerprinting. J. Proteome Res. 9, 3169–3183. Calderaro, A., Gorrini, C., Piccolo, G., Montecchini, S., Buttrini, M., Rossi, S., Piergianni, M., Arcangeletti, M.C., De Conto, F., Chezzi, C., Medici, M.C., 2014. Identification of Borrelia species after creation of an in-houseMALDI-TOF MS database. PLoS One 9, e88895. Cherkaoui, A., Hibbs, J., Emonet, S., Tangomo, M., Girard, M., Francois, 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. Christensen, J.J., Dargis, R., Hammer, M., Justesen, U.S., Nielsen, X.C., Kemp, M., Danish MALDI-TOF MS Study Group, 2012. Matrix-assisted laser desorption ionization-time of flight mass spectrometry analysis of gram-positive, catalase-negative cocci not belonging to the Streptococcus or Enterococcus genus and benefits of database extension. J. Clin. Microbiol. 50, 1787–1791. De Carolis, E., Vella, A., Vaccaro, L., Torelli, R., Posteraro, P., Ricciardi, W., Sanguinetti, M., Posteraro, B., 2014. Development and validation of an in-house database for matrix-assisted laser desorption ionization-time of flight mass spectrometry-based yeast identification using a fast protein extraction procedure. J. Clin. Microbiol. 52, 1453–1458. Erler, R., Wichels, A., Heinemeyer, E.A., Hauk, G., Hippelein, M., Reyes, N.T., Gerdts, G., 2015. VibrioBase: A MALDI-TOF MS database for fast identification of Vibrio spp. that are potentially pathogenic in humans. Syst. Appl. Microbiol. 38, 16–25. Kurokawa, S., Kabayama, J., Fukuyasu, T., Hwang, S.D., Park, C.I., Park, S.B., del Castillo, C.S., Hikima, J., Jung, T.S., Kondo, H., Hirono, I., Takeyama, H., Aoki, T., 2013. Bacterial classification of fish-pathogenic Mycobacterium species by multigene
352
Contents lists available at ScienceDirect
Research in Veterinary Science journal homepage: www.elsevier.com/locate/rvsc
Differentiation of Flavobacterium psychrophilum from Flavobacterium psychrophilum-like species by MALDI-TOF mass spectrometry
MARK
Marta Pérez-Sanchoa, Ana Isabel Velaa,b, Tom Wiklundc, Markus Kostrzewad, Lucas Domíngueza,b, José Francisco Fernández-Garayzábala,b,⁎ a
Centro de Vigilancia Sanitaria Veterinaria (VISAVET), Universidad Complutense de Madrid, Avenida Puerta de Hierro s/n, 28040 Madrid, Spain Departamento de Sanidad Animal, Facultad de Veterinaria, Universidad Complutense, Avenida Puerta de Hierro s/n, 28040 Madrid, Spain c Laboratory of Aquatic Pathobiology, Environmental and Marine Biology, Åbo Akademi University, BioCity, Artillerigatan 6, 20520 Åbo, Finland d Bruker Daltonik GmbH, Bremen, Germany b
A R T I C L E I N F O
A B S T R A C T
Keywords: Flavobacterium psychrophilum MALDI-TOF MS Identification Differentiation Diagnosis
Rainbow trout fry syndrome (RTFS) is an important infectious disease caused by Flavobacterium psychrophilum affecting farmed salmonids worldwide. Other Flavobacterium psychrophilum-like species (F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis) have been isolated from diseased rainbow trout fry suspected of RTFS although the epidemiological and clinical relevance of these pathogens are unknown. The objective of this study was to evaluate the potential use of MALDI-TOF (Matrix-Assisted Laser Desorption/Ionization-Time of Flight) Mass Spectrometry as method for specific identification of F. psychrophilum and its differentiation from other F. psychrophilum-like species isolated from diseased fish. Fifty-three isolates were analyzed after the creation of the Main Spectrum Profile (MSP) of reference strains of each of abovementioned species. F. psychrophilum exhibited a mass spectra very different from those of F. psychrophilum-like species, with five peaks (m/z 3654, 4585, 5388, 6730 and 7310) present only in F. psychrophilum isolates, and three peaks (m/z 6170, 7098 and 9241) absent in F. psychrophilum but present in all F. psychrophilum-like species. All F. psychrophilum isolates were correctly identified and differentiated from the F. psychrophilum-like species by MALDI-TOF. Although this approach showed a limited ability to differentiate among F. psychrophilum-like species, its complementation with a few simple biochemical tests may represent an alternative approach for the routine identification of the Flavobacterium psychrophilum-like species.
1. Introduction Flavobacterium psychrophilum is an important bacterial pathogen of wild and farmed salmonids worldwide and the etiological agent of rainbow trout (Oncorhynchus mykiss) fry syndrome (RTFS; Austin and Stobie, 1991) and bacterial cold water disease (BCWD; Wood and Yasutake, 1956). In RTFS, F. psychrophilum infections cause high mortalities and significant losses in the affected fish populations (Nematollahi et al., 2003). In recent years, several new described Flavobacterium species (F. plurextorum, F. tructae, F. collinsii, F. oncorhynchi and F. piscis) have been isolated from diseased rainbow trout fry that present clinical symptoms consistent with F. psychrophilum infection (Zamora et al., 2012; Zamora et al., 2013a; Zamora et al., 2013b; Zamora et al., 2014). For this reason and the resemblance in the macroscopic characteristics of their colonies on Anacker and Ordal agar with those of F. psychrophilum and similar biochemical characteristics (Zamora, 2015), these species are considered under the generic name of ⁎
F. psychrophilum-like species in the present work. The immunological and molecular methods available for the detection and identification of F. psychrophilum are useful for an accurate identification of RTFS but not for identification of the etiological agent of those episodes in which F. psychrophilum-like species are involved. These F. psychrophilum-like species can be differentiated from F. psychrophilum by several biochemical characteristics (e.g. growth at 25 °C, degradation of aesculin and carbohydrates and growth on trypticase soy agar) or sequencing of their 16S rRNA gene (Zamora, 2015). However, conventional phenotypic identification methods are labor-intensive and time-consuming and costs of DNA sequencing generally are too high for its routine use in veterinary diagnostic laboratories. These limitations can be overcome by MALDI-TOF MS approach, a useful tool, which has become a rapid, accurate, and cost-effective identification technique in clinical microbiology laboratories (Patel, 2015; Singhal et al., 2015). In the veterinary field, the method has been successfully applied for the identification of a wide range of bacterial pathogens (Randall et al.,
Corresponding author at: Departamento de Sanidad Animal, Facultad de Veterinaria, Universidad Complutense, Avenida Puerta de Hierro s/n, 28040 Madrid, Spain. E-mail address: [email protected] (J.F. Fernández-Garayzábal).
http://dx.doi.org/10.1016/j.rvsc.2017.06.022 Received 22 December 2016; Received in revised form 9 June 2017; Accepted 28 June 2017 0034-5288/ © 2017 Elsevier Ltd. All rights reserved.
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
2015) including several bacterial fish pathogens (Beaz-Hidalgo et al., 2009; Böhme et al., 2010; Erler et al., 2015; Kurokawa et al., 2013; Regecová et al., 2014; Strepparava, 2012). However, to our knowledge, there is only one study evaluating the suitability of MALDI-TOF for identification of F. psychrophilum but other Flavobacterium species associated with infection in fish were not included (Strepparava, 2012). The aim of this work is to evaluate the suitability of MALDI-TOF MS for the identification of F. psychrophilum and its differentiation from other F. psychrophilum-like species (F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis) that can be isolated from outbreaks with presumptive diagnosis of F. psychrophilum infections based on clinical symptoms.
Table 1 Potential species-specific MALDI-TOF peaks (mass/charge, m/z) used for differentiation of Flavobacterium psychrophilum and Flavobacterium psychrophilum-like species included in the present study. m/z
3654 4585 5388 6730 7310 6170
2. Materials and methods 2.1. Bacterial isolates and culture conditions 7098
A total of 64 Flavobacterium isolates were included in the present study. A first set of isolates (n = 11) was used to create Main Spectra (MSPs) for supplementation of the Bruker MALDI-TOF MS database (version 3.4; 5989 entries): F. psychrophilum (JIP02/86 and H7/00), F. plurextorum (1126-1H-08T, 424-08), F. oncorhynchi (631-08T, 662-09), F. tructae (435-08T) F. collinsii (983-08T and 978B-08) and F. piscis (41209T and 60B-3-09). A second set of isolates (n = 53; Table 2) was used for the external validation of the new MSPs confirming the speciesdifferential ions detected and to assess the suitability of MALDI-TOF technique for the differentiation of F. psychrophilum from the other F. psychrophilum-like species. All isolates, except F. psychrophilum (n = 31), had been part of taxonomic polyphasic studies; therefore, all isolates had been previously characterized by biochemical, chemotaxonomic and genetic methods (Zamora et al., 2012; Zamora et al., 2013a; Zamora et al., 2013b; Zamora et al., 2014). Identification of F. psychrophilum was achieved by biochemical characterization, sequencing of 16S rRNA gene or species-specific PCR (Nilsen et al., 2014). The reference strain JIP02/86 of F. psychrophilum (DSM 21280) was acquired from Leibniz Institute DSMZ (German Collection of Microorganisms and Cell cultures). Clinical isolates were mostly recovered from rainbow trout (n = 50), although a subset of F. psychrophilum isolates (n = 14) was isolated from European perch (Perca fluviatilis), Atlantic salmon (Salmo salar), brook trout (Salvelinus fontinalis), brown trout (Salmo trutta) and water. All isolates used in this study were grown on R2A agar (Reasoner's 2A agar, Reasoner et al., 1979) and incubated for 72 h at 15–22 °C under aerobic conditions.
9241
Flavobacterium sp.a,b
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum collinsii oncorhynchi piscis plurextorum tructae collinsii oncorhynchi piscis plurextorum tructae collinsii oncorhynchi piscis plurextorum tructae
Total isolatesc
31 31 31 31 31 5 18 3 5 2 5 18 3 5 2 5 18 3 5 2
Species-specific peaks Intensity (a.u.)
Signal-tonoise (s/n)
Frequencyd
1727–9916 6510–34,759 3988–16,944 1462–9716 1755–14,152 3202–5121 1617–9060 1744–4581 2795–8694 1426–9515 682–3271 941–4837 2667–5074 1557–3746 1054–5100 848–1731 1406–5413 1435–2112 1574–2888 2604–2708
19–55 67–258 39–109 28–81 36–125 24–72 12–66 18–27 26–52 11–75 6–28 9–35 29–42 16–38 9–45 18–24 17–41 24–28 25–32 33–34
100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 80% 100% 100% 100% 100% 80% 100% 100% 100% 100%
a Flavobacterium species identification determined by biochemical, chemotaxonomic and genetic methods. b F. oncorhynchi and F. plurextorum showed a very similar protein profile. Some additional peaks (m/z 4820, 5055 and 10,110) have been proposed in this study for their differentiation. c Number of isolates per species. d Number of isolates displaying each specific peak.
was overlaid with 1 μL of HCCA matrix solution (saturated solution of α-cyno-4-hydroxycinnamic acid in 50% acetonitrile 2.5% trifluoroacetic acid) and air dried at room temperature. Secondly, the extraction protocol was repeated with a subset of 11 isolates, which were used for the creation of reference spectra (as explained below): protein extracts were deposited onto 8 spots and three subsequent spectra acquisitions were performed to obtain 24 spectra/isolate. Finally, in the case of F. psychrophilum-like isolates, the extraction protocol was repeated and one microliter of the protein extract was spotted onto each of 8 spots. After adding 1 μL of HCCA matrix solution to each spot and let dry, three subsequent spectra acquisitions were carried out from each spot to obtain 24 spectra per sample for cluster analysis.
2.2. Preparation of bacterial cell lysates for MALDI-TOF MS
2.3. MALDI-TOF MS: acquisition and processing of mass spectra
First, cells from bacterial colonies of each isolate were transferred from the agar plate to a 2 mL tube (Eppendorf, Germany) containing 300 μL of HPLC grade-water (Sigma Aldrich, Germany). A total of 900 μL of HPLC-grade (High Performance Liquid Chromatography) absolute ethanol was added to the bacterial suspension. A protein extraction protocol based on formic acid/acetonitrile was subsequently performed in all isolates following the manufacturer's instructions (Bruker Daltonik GmbH, Germany) and previously described elsewhere (Pérez-Sancho et al., 2015). The extraction protocol was repeated twice or three times for different purposes. Firstly, each bacterial suspension was subjected to protein extraction and the extracts were used for bacterial identification using the Bruker BDAL MALDI Biotyper database (5989 entries). Additionally, these spectra were also used to assess the robustness of proposed species-specific biomarkers detected on MSP (Main Spectra) spectra. In this case, 1 μL of the protein extracts containing the bacterial proteins and peptides was spotted onto one spot of a polished steel target plate and let dry at room temperature. Subsequently, each spot
MALDI-TOF spectra were acquired using two different devices: an UltrafleXtreme machine (Bruker Daltonik; used for spectra acquisition for bacterial identification and for evaluation of the robustness of proposed species-specific biomarkers) (Table 1) and a Microflex LT equipment (Bruker Daltonik; used for spectra acquisition for creation of MSPs, evaluation of the presence of species-specific peaks on these reference spectra and cluster analysis). Regardless of the device type, spectra were acquired in the linear and positive mode within a mass range between 2 and 20 kDa. Each spectrum was acquired using the FlexControl software (version 3.4) in an automatic mode and externally calibrated using the Bacterial Test Standard (Bruker Daltonik GmbH, Bremen, Germany). 2.4. Identification based on MALDI Biotyper For the bacterial identification study, one spectra from each clinical Flavobacterium isolate (n = 64) acquired by UltrafleXtreme device was 346
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
Table 2 Identification results using MALDI TOF approach in 53 Flavobacterium isolates recovered from diseased fish after the inclusion of F. psychrophilum and F. psychrophilum-like MSP to construct the fish Flavobacterium MALDI database (FDB). Speciesa
Reference
1 ID - FDB (Score value)b
F. psychrophilum
P4-4/06 1.12.99-47 P11-1B/97 P16-8/97 P4-3/06 K2/99 T1-1 L4/00 P7-9/2R/10 P29-1A/10 P18-2/11 OG-83/28 M2/99 255-ii/93 V43 3474/94 21.10.99-3 H2/00 P11-1A/97 ST9/00 P7-9/10 P44-9/11 P4-2B/06 P56-2A/06 A6/00 P5-1/12 P69-5/97 15,527/2/96 H1/00
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
53B-3-09 991-08 977H-09
F. collinsii (2.518) F. collinsii (2.164) F. collinsii (2.049)
F. collinsii (2.080) F. collinsii (2.031) F. collinsii (2.045)
A A A
F. oncorhynchi
426B-08 628-1-08 646-08 437B-08 441B-08 695B-08 947B-08 646B-08 950B-08 650-08 688B-08 22B-09 425B-08 433B-08 47B-2-09 666-09
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
plurextorum (1.948) oncorhynchi (1.861) plurextorum (1.927) oncorhynchi (1.953) plurextorum (1.907) oncorhynchi (1.865) johnsoniae (1.748) plurextorum (1.921) oncorhynchi (1.867) oncorhynchi (1.677) plurextorum (1.931) oncorhynchi (1.885) oncorhynchi (1.985) oncorhynchi (1.963) plurextorum (2.021) plurextorum (2.024)
A A A A A A A A A A A A D D B B
F. piscis
412R-09
F. piscis (2.281)
F. saccharophilum (2.208)
B
F. plurextorum
1084B-08 51B-09 986-08
F. plurextorum (2.206) F. plurextorum (2.175) F. plurextorum (2.257)
F. plurextorum (2.136) F. plurextorum (2.083) F. plurextorum (2.051)
A A A
F. tructae
47B-3-09
F. hibernum (1.840)
F. tructae (1.824)
D
F. psychrophilum-like F. collinsii
psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum
oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi oncorhynchi
Consistencyd
2 ID- FDB (Score value)c (2.644) (2.563) (2.537) (2.588) (2.685) (2.619) (2.514) (2.653) (2.686) (2.648) (2.592) (2.718) (2.521) (2.569) (2.680) (2.611) (2.558) (2.621) (2.698) (2.551) (2.502) (2.588) (2.523) (2.518) (2.523) (2.494) (2.570) (2.532) (2.557)
(2.220) (2.210) (2.015) (2.210) (2.240) (2.217) (2.156) (2.113) (2.180) (2.125) (2.294) (2.200) (1.993) (1.988) (2.120) (2.192)
F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F.
psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum psychrophilum
(2.096) (2.072) (2.209) (1.885) (2.094) (2.057) (2.030) (2.197) (2.026) (2.083) (2.151) (2.049) (2.139) (1.965) (2.127) (2.212) (1.692) (2.281) (1.832) (1.760) (2.013) (2.045) (2.026) (2.002) (1.928) (1.940) (2.007) (2.141) (2.142)
A A A A A A A A A A A A A A A A A A A A A A A A A A A A A
a
Flavobacterium species determined by biochemical, chemotaxonomic and genetic methods. Best match in the ranking list of MALDI TOF identification results and corresponding score value provided by MALDI Biotyper (Bruker Daltonik, Germany). c Second match in the ranking list of MALDI TOF identification results and corresponding score value provided by MALDI Biotyper (Bruker Daltonik, Germany). d Consistency ranking list of the first 5 best matches = A, the correct species is the unique species with score value ≥ 2.000; B, the correct species is first ranked but a different species in the second rank also with score value ≥ 2.000; D, first and second matches have score values < 2.000 (correct genus identification). b
Flavobacterium saccharophilum). After MSP creation using the 11 Flavobacterium isolates as described above, the remaining isolates (n = 53) were identified using the fish Flavobacterium databases (FDB, 5989 entries of BDAL Biotyper Database plus eleven new Flavobacterium MSPs, Table 2). The Biotyper software showed 10 top matches (as defined by the user) in an identification ranking list ordered according to
processed using the MALDI Biotyper RTC software (version 3.1) with default settings in the automatic mode. All spectra were identified using the Bruker BDAL MALDI Biotyper (5989 entries; which includes Flavobacterium flevense, Flavobacterium gelidilacus, Flavobacterium hibernum, Flavobacterium hydatis, Flavobacterium johnsoniae, Flavobacterium lindanitolerans, Flavobacterium pectinovorum and 347
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
To incorporate the MSP of F. psychrophilum and F. psychrophilum-like species in the Bruker database (version 3.4; 5989 entries), two wellcharacterized isolates of each species (except one isolate in F. tructae) were analyzed. Analysis of the spectra of these species was performed by selecting a range of m/z between 3000 and 10,000 Da that included the major differences between their MSP. The visual inspection of these spectra using ClinProTools revealed important differences between the mass spectra of F. psychrophilum and those of the F. psychrophilum-like species (Table 1). Thus, peaks 3654, 4585, 5388, 6730 and 7310 m/z were present in F. psychrophilum strains, while peaks 6170, 7098 and 9241 m/z, present in all F. psychrophilum-like species, were absent in F. psychrophilum. These distinctive peaks were also observed by visual inspection using FlexAnalysis software (data not shown). Data of the intensity, signal-to-noise and frequency of detection of these ions in the 64 isolates of the different Flavobacterium species is shown in Table 1. Most peaks were identified in 100% of the isolates examined with average intensity and signal-to-noise values of > 600 and > 6, respectively. A cluster analysis of MSP obtained from the 11 initial Flavobacterium isolates based on the construction of a dendrogram using MALDI Biotyper OffLine Classification software (version 3.1) demonstrated that F. psychrophilum spectra form a well-defined clade separated from those formed by the F. psychrophilum-like species and the other Flavobacterium species included currently in the MALDI Biotyper database (Fig. 1). Visual inspection showed that the spectra of the F. psychrophilumlike species were in general quite similar except for F. piscis that showed one indicative peak at 5418 m/z (Table 1). Visual inspection using ClinProTools and FlexAnalysis of the mass patterns of F. oncorhynchi and F. plurextorum revealed almost no differences in their spectra. Similarly, F. tructae and F. collinsii presented very similar mass patterns (Fig. 1). Fig. 3A is a strain distribution map base on the spectra of all 33 F. psychrophilum-like isolates showing the close relationship between F. plurextorum and F. oncorhynchi and between F. tructae and F. collinsii. These relationships reflect the similarity in their MALDI profiles observed by visual inspection and may suggest potential problems of misidentification among these species. Results of PCA score plot confirmed these close relatedness (Fig. 3B). After the MSPs of F. psychrophilum and F. psychrophilum-like species were incorporated in the Bruker database, the second set of isolates (n = 53) of these species recovered from clinical specimens of diseased fish was again identified by MALDI-TOF MS (Table 2). All F. psychrophilum isolates were correctly identified with a mean score value of 2.588 and a consistency identification category A that indicate the absence of misidentifications with any of the F. psychrophilum-like species. Regarding isolates of the F. psychrophylum-like species, the top hit of the ranking identification list were in accordance with the biochemical, chemotaxonomic and/or genetic results in all isolates (except DICM09/0047B-3) (Table 2) (n = 23; 95.83%). Most of correctly identified isolates (91.3%; n = 21) gave score values > 2.0 in the first option in the identification ranking list provided by the MALDI Biotyper. The single strain of F. tructae was identified as F. hibernum or F. tructae in the first and second options in the identification-ranking list (Table 2) with score values of 1.84 and 1.824, respectively, which indicate a reliable identification at genus level but only a low confidence species identification. A closer analysis of the mass peak profiles of both strains of F. tructae (one used for MSP creation and the other used in the blind panel) revealed important differences between them (e.g. m/z 3084, 3728 and 6169 in strain 435-09T and m/z 3716, 7433 and 7608 in strain 47B-3-09) suggesting a potentially high intraspecies variability. Analysis of further isolates of this species in the future will allow confirmation whether these differences are common in F. tructae. The single strain of F. piscis was correctly identified by MALDI-TOF with score value of 2.281 in the first match of the identification-ranking list. However, F. saccharophilum was the identification result on the second position of the raking list with a very close score value of 2.208
the score value of each identification. For interpretation of the identification results, two variables were considered: (i) score values following the MALDI Biotyper Compass (≥ 2.000: identification at species level with high confidence; 1.999–1.800: identification with low confidence and ≤ 1799: non-reliable identification) and (ii) consistency ranking list of the first 2 best matches [A, the correct species is the unique species with score value ≥ 2.000; B, the correct species is first ranked but a different species in the second rank also with score value ≥ 2.000; C, first and second matches have score values ≥2.000 but the correct species is second ranked; D, first and second matches have score values < 2.000 (correct genus identification)]. 2.5. Evaluation of the presence of species-identifying biomarkers and creation of MSP for F. psychrophilum, F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis Visual inspection of spectra of all Flavobacterium isolates was performed using the pseudogel view option of ClinProTools program (version 3.0). Those potential species-identifying ions were visually confirmed using FlexAnalysis software (Bruker Daltonik). Analysis of the presence of candidate differential biomarkers for each Flavobacterium species was confirmed by ClinProTools (version 3.0). This software allows the data processing and spectra preparation realignment to reduce measurements variations. The pseudogel view option of ClinProTools allows evaluating the presence/absence of potential species-specific ions. Finally, a principal component analysis (PCA) was created using 24 spectra/isolate of all F. psychrophilum-like species. PCA is based on the Euclidian distance method and was constructed using ClinProTools software. 2.6. Construction of fish Flavobacterium MALDI database (FDB) The construction of the FDB was carried out according to manufacture's instructions. A total of 11 new MSP were incorporated in the Bruker BDAL MALDI Biotyper database available that time (5989 entries): two new entries per species were included except in the case of F. tructae in which only one entry was included due to the limitation in the number of isolates available. MSP construction was performed using 20–24 spectra for each Flavobacterium isolate (three measurements at 8 different spots). The quality and reproducibility of all spectra included in each MSP were assessed using FlexAnalysis software (version 3.4, Bruker Daltonik) after smoothing, normalization, baseline subtraction and peak picking. All MSP were created by MALDI Biotyper Compass Explorer software (version 4.1.16, Bruker Daltonik) using the MSP series creation option. To determine if MALDI profiles of Flavobacterium species were distributed in different clades, we combined MSP of F. psychrophilum, F. psychrophilum-like species, and those Flavobacterium species currently present in the Biotyper database for cluster analysis (MSP based dendrogram construction using default parameter set) using MALDI Biotyper software. 3. Results Before the incorporation of the MSP in the Bruker database, the 64 Flavobacterium isolates were analyzed as a preliminary test of the suitability of MALDI-TOF MS for the identification of Flavobacterium species. Most isolates of the Flavobacterium species included in the study (n = 49; 79%) showed a mean score value of < 1.699, indicating an unreliable identification and consequently their unlikely misidentification with any of those species currently included in this database (F. flevense, F. gelidilacus, F. hibernum, F. hydatis, F. johnsoniae, F. lindanitolerans, F. pectinovorum and F. saccharophilum). Twelve isolates gave score values between 1.999 and 1.800 indicative of a correct identification at low confidence level. The three isolates of F. piscis were identified as F. saccharophilum with scores values > 2.000, which suggests a misidentification at the high confidence species level. 348
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
Fig. 1. Dendrogram derived from MSP of Flavobacterium species analyzed in this study and those currently included in the MALDI Biotyper database (5989 entries) demonstrates that F. psychrophilum and F. psychrophilum-like species grouped in different clades (relative distance values are normalized to a maximum value = 1000).
(Seng et al., 2009; Cherkaoui et al., 2010). In the same way, the speed of bacterial identification by MALDI-TOF is estimated to be few minutes compared with hours or days required for conventional methods (Seng et al., 2009; Cherkaoui et al., 2010). In this study, we have evaluated the usefulness of MALDI-TOF MS for identifying F. psychrophilum and discriminating it from other F. psychrophilum-like species, which can be relevant from a diagnosis and treatment point of view. Neither F. psychrophilum nor F. psychrophilum-like species analyzed in this study are included among the Flavobacterium species present in the MALDI-TOF database (5989 entries). An initial attempt to identify these species by MALDI-TOF was unsuccessful as score values for most species were < 1.9. Only the isolates of F. piscis gave scores > 2.000 with F. saccharophilum indicating a probable misidentification with this species. These data confirm that, with the exception of F. piscis, all other Flavobacterium species were not misidentified with any of the species of Flavobacterium included in the Bruker database. This highlights the need to supplement specific MSP of these species to assess the suitability of MALDI approach for these fish pathogens as previously observed for other microorganisms (Calderaro et al., 2014; Christensen et al., 2012; De Carolis et al., 2014). Therefore, the first objective of this study was to construct the MSP of the reference strains of F. psychrophilum and F. psychrophilum-like species. The new MSP of Flavobacterium here created are at the disposal of Bruker Daltoniks to be included in further versions of Bruker Biotyper database. Visual inspection of generated MSP revealed very different mass spectra exhibited by F. psychrophilum isolates compared to those of the F. psychrophilum-like species, with six conserved and stable peaks (m/z 3654, 4585, 5388, 6730 and 7310) present in all F. psychrophilum isolates, and three peaks (m/z 6170, 7098 and 9241) absent in F. psychrophilum but present in all isolates of F. psychrophilum-like species (Table 1, Fig. 2). Most of these distinctive peaks exhibited significant intensity and signal-to-noise values (> 600 and > 6, respectively) and they were identified in all isolates. These results confirmed their robustness and specificity and, therefore, their potential usefulness as species-specific biomarkers for the differentiation of F. psychrophilum
indicating an inconclusive identification (B category consistency identification). MALDI identification results were in accordance with those observed for genetic and biochemical methods for F. collinsii showing category A consistency identification. The three isolates of F. plurextorum were also correctly identified with a consistency identification category A and average score values of 2.216 (Table 2). Most isolates of F. oncorhynchi were correctly identified (category A; Table 2) but in two strains the second identification option was F. plurextorum also with scores values > 2.0 (category B) and another two isolates were identified only to low confidence level (category D). 4. Discussion The existence of different Flavobacterium species (F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis), which may cause RTFSlike disease on rainbow trout, may complicate the species-specific identification of F. psychrophilum invalidating those immunologic and molecular methods available for this purpose (Zamora, 2015). These species have been also isolated from gills and several internal organs (brain and kidney) of different fish species (Zamora et al., 2012; Zamora et al., 2013a; Zamora et al., 2013b; Zamora et al., 2014; Loch and Faisal, 2015), and their identification is based upon near-to-complete sequencing of their 16S rRNA gene (Loch et al., 2013; Loch and Faisal, 2014); which is a technique that is laborious, expensive and relatively time consuming. Consequently, it is necessary to develope diagnostic tools not only for the rapid differentiation of F. psychrophilum from F. psychrophilum-like species, but also for the species-specific identification of the latter. In this context, MALDI-TOF MS can be a valuable alternative. The method has been recently become routine in some diagnostic laboratories for the identification of many bacterial pathogens. Different studies have assessed the improvements of implementing MALDI-TOF technique on the routine laboratory diagnosis (Seng et al., 2009; Cherkaoui et al., 2010; Bizzini and Greub, 2010). In this sense, it has been calculated that the cost of this proteomic approach to be 17–32% of the cost of automated identification system 349
Fig. 2. Differential species-specific peaks on the MALDI-TOF MS spectra of whole-cell extracts of F. psychrophilum and F. oncorhynchi. F. psychrophilum-like species exhibited very similar mass patterns and the spectra of F. oncorhynchi is given as an example of the other F. psychrophilum-like species (A). In inserts, an enlarged spectrum of 3500–5550 m/z (B), 6000–7500 m/z (C) and 9000–10,500 m/z (D) to highlight the differential peaks between F. psychrophilum and F. oncorhynchi spectra.
M. Pérez-Sancho et al.
Research in Veterinary Science 115 (2017) 345–352
350
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
Fig. 3. Cluster analysis of F. psychrophilum-like species [F. oncorhynchi (○), F. plurextorum (◊), F. collinsii (×), F. tructae (Δ) and F. piscis (□)]. 2D peak distribution diagram (A) and PCA plot (B) illustrating the close relation between F. oncorhynchi/F. plurextorum and F. collinsii/F. tructae.
The evaluation of MALDI profiles of F. oncorhynchi and F. plurextorum was only performed to assess the discriminatory power of species differentiating-ions because the number of isolates of these species was not high enough for further statistical analysis. However, the rigorous visual inspection of their spectra demonstrated the presence of some mass-peaks (m/z 4822, 5057, 6902, 9640 and 10,112 for F. oncorhynchi and m/z 4828, 5066, 9654 and 10,127 for F. plurextorum; data not shown) potentially useful for their differentiation. Future isolates belonging to these species may allow the validation of the proposed differentiating-peaks to confirm their suitability for the rapid and reliable differentiation of F. oncorhynchi and F. plurextorum allowing the construction of classification models for their automatic differentiation complementing the standard identification algorithm. Meanwhile, for isolates of species for which the two highest Biotyper match scores are > 2.0 MALDI-TOF can be used in combination with simple complementary tests to achieve an unequivocal identification. Thus, isolates of F. oncorhynchi can be easily differentiated from F. plurextorum by the ability of the former species to produce the enzyme β-galactosidase (Zamora et al., 2012; Zamora et al., 2013a). Similarly, F. piscis can be differentiated from F. saccharophilum by the ability of the latter species to produce the enzymes β-galactosidase and DNAse (Bernardet et al., 1996; Zamora et al., 2014).
from the F. psychrophilum-like species. The clear differentiation of F. psychrophilum and F. psychrophilum-like species after cluster analysis of their MSP (Fig. 1) indicates that MALDI-TOF approach represents an adequate alternative for their differentiation. In agreement with these data, MALDI-TOF MS correctly identified all F. psychrophilum isolates with score values always higher than 2.3 for the best match in Biotyper identification list (average 2.588; Table 2). F. psychrophilum isolates reached a category A consistency identification, further indicating it was clearly differentiated from the F. psychrophilum-like species that exhibited always score values < 2.0. Therefore, MALDI-TOF MS represents an easy and accurate alternative to the immunological and genetic methods currently available for the identification of F. psychrophilum. On the other hand, visual inspection of MSP generated by F. psychrophilum-like species revealed they were very similar, in particular the mass spectra of F. plurextorum and F. oncorhynchi and the spectra of F. collinsii and F. tructae after 2D peak distribution diagram and PCA plot analysis (Fig. 3). The close relatedness between these species is congruent with the high similarity of their 16S rRNA gene sequences (Zamora et al., 2012; Zamora et al., 2013a; Zamora et al., 2013b; Zamora et al., 2014). The utility of MALDI-TOF MS for the identification of the F. psychrophilum-like species was variable. Isolates of F. plurextorum and F. collinsii were correctly identified with category A consistency identification and score values > 2.0 (average scores of 2.212 and 2.242, respectively, Table 2). Similarly, most isolates (n = 12) of F. oncorhynchi were correctly identified as at least the two best matches generated by Biotyper indicated this species or second best matches indicating a different species was lower than 2.0. The isolate of F. tructae and two isolates of F. oncorhynchi were identified correctly only to genus level with scores values < 2.0 (low confidence, D category consistency identification; Table 2). The resolution of MALDI-TOF MS was not high enough to discriminate between some species (Table 2) even though the best matching score was the correct species, as it can be observed by comparing the corresponding score values. Thus, the discrimination was not clear cut for the isolate of F. piscis that was identified in the first best match in Biotyper, but the second best match was F. saccharophilum also with a score value > 2.0 (B category consistency identification; Table 2). Similar observation was made with two isolates of F. oncorhynchi in which the second match with scores values > 2.0 pointed to F. plurextorum (Table 2). These four species have been isolated from diseased fish and the limited discrimination of MALDI-TOF MS for their differentiation could represent a challenge to diagnostic laboratories.
5. Conclusion MALDI-TOF MS seems to be a reliable and alternative approach for the accurate identification of F. psychrophilum. In addition, MALDI-TOF MS, used alone or complemented with a few simple biochemical tests, can be used for the routine identification of different Flavobacterium species (F. plurextorum, F. oncorhynchi, F. tructae, F. collinsii and F. piscis) that can be isolated from diseased fish suspected of F. psychrophilum infections, facilitating the differential diagnosis. The use of MALDI-TOF MS in veterinary diagnostic laboratories would advance knowledge of the epidemiological and clinical significance of these flavobacteria. Role of the funding source This study was supported by the project Advance Technologies in Health Surveillance [S2013/ABI-2747 (TAVS-CM)] funded by the Regional Government of Madrid and the European Union. Marta PérezSancho is recipient of a technical support staff for scientific infrastructures of the Moncloa Campus of International Excellence (Programa CEI09-0019) and Ministerio de Educación, Cultura y 351
Research in Veterinary Science 115 (2017) 345–352
M. Pérez-Sancho et al.
phylogenetic analyses and MALDI Biotyper identification system. Mar. Biotechnol. (NY) 15, 340–348. Loch, T.P., Faisal, M., 2014. Deciphering the biodiversity of fish-pathogenic Flavobacterium spp. recovered from the Great Lakes basin. Dis. Aquat. Org. 112, 45–57. Loch, T.P., Faisal, M., 2015. Emerging flavobacterial infections in fish: a review. J. Adv. Res. 6, 283–300. Loch, T.P., Fujimoto, M., Woodiga, S.A., Walker, E.D., Marsh, T.L., Faisal, M., 2013. Diversity of fish-associated flavobacteria of Michigan. J. Aquat. Anim. Health 25, 149–164. Nematollahi, A., Decostere, A., Pasmans, F., Haesebrouck, F., 2003. Flavobacterium psychrophilum infections in salmonid fish. J. Fish Dis. 26, 563–574. Nilsen, H., Sundell, K., Duchaud, E., Nicolas, P., Dalsgaard, I., Madsen, L., Aspán, A., Jansson, E., Colquhoun, D.J., Wiklund, T., 2014. Multilocus sequence typing identifies epidemic clones of Flavobacterium psychrophilum in Nordic countries. Appl. Environ. Microbiol. 80, 2728–2736. Patel, R., 2015. MALDI-TOF MS for the diagnosis of infectious diseases. Clin. Chem. 61, 100–111. Pérez-Sancho, M., Vela, A.I., García-Seco, T., Gottschalk, M., Domínguez, L., FernándezGarayzábal, J.F., 2015. Assessment of MALDI-TOF MS as alternative tool for Streptococcus suis identification. Front.Public Health 3, 202. Randall, L.P., Lemma, F., Koylass, M., Rogers, J., Ayling, R.D., Worth, D., Klita, M., Steventon, A., Line, K., Wragg, P., Muchowski, J., Kostrzewa, M., Whatmore, A.M., 2015. Evaluation of MALDI-TOF as a method for the identification of bacteria in the veterinary diagnostic laboratory. Res. Vet. Sci. 101, 42–49. Reasoner, D.J., Blannon, J.C., Geldreich, E.E., 1979. Rapid seven-hour fecal coliform test. Appl. Environ. Microbiol. 38, 229–236. Regecová, I., Pipová, M., Jevinová, P., Marušková, K., Kmeť, V., Popelka, P., 2014. Species identification and antimicrobial resistance of coagulase-negative staphylococci isolated from the meat of sea fish. J. Food Sci. 79, 898–902. Seng, P., Drancourt, M., Gouriet, F., La Scola, B., Fournier, P.E., Rolain, J.M., Raoult, D., 2009. Ongoing revolution in bacteriology: routine identification of bacteria by matrix-assisted laser desorption ionization time-of-flight mass spectrometry. Clin. Infect. Dis. 49, 543–551. Singhal, N., Kumar, M., Kanaujia, P.K., Virdi, J.S., 2015. MALDI-TOF mass spectrometry: an emerging technology for microbial identification and diagnosis. Front. Microbiol. 6, 791. Strepparava, N., 2012. Novel Approaches to the Diagnosis of Flavobacteriosis: Identification of Pathogenic Flavobacteriaceae and Epidemiology of Flavobacterium psychrophilum. Universität Bern, PhD dissertation. Wood, E.M., Yasutake, W.T., 1956. Histopathology of fish. III. Peduncle (‘coldwater’) disease. Progress. Fish Cult. 18, 58–61. Zamora, L., 2015. Estudio Taxonómico de Bacterias Relacionadas con el Síndrome del Alevín de la Trucha (Taxonomic study of bacteria associated with rainbow trout fry syndrome). (PhD dissertation) Universidad Complutense de Madrid. Zamora, L., Fernández-Garayzábal, J.F., Svensson-Stadler, L.A., Palacios, M.A., Domínguez, L., Moore, E.R., Vela, A.I., 2012. Flavobacterium oncorhynchi sp. nov., a new species isolated from rainbow trout (Oncorhynchus mykiss). Syst. Appl. Microbiol. 35, 86–91. Zamora, L., Fernández-Garayzábal, J.F., Sánchez-Porro, C., Palacios, M.A., Moore, E.R., Domínguez, L., Ventosa, A., Vela, A.I., 2013a. Flavobacterium plurextorum sp. nov. isolated from farmed rainbow trout (Oncorhynchus mykiss). PLoS One 8, e67741. Zamora, L., Vela, A.I., Sánchez-Porro, C., Palacios, M.A., Domínguez, L., Moore, E.R., Ventosa, A., Fernández-Garayzábal, J.F., 2013b. Characterization of flavobacteria possibly associated with fish and fish farm environment. Description of three novel Flavobacterium species: Flavobacterium collinsii sp. nov., Flavobacterium branchiarum sp. nov., and Flavobacterium branchiicola sp. nov. Aquaculture 416–417, 346–353. Zamora, L., Vela, A.I., Sánchez-Porro, C., Palacios, M.A., Moore, E.R., Domínguez, L., Ventosa, A., Fernández-Garayzábal, J.F., 2014. Flavobacterium tructae sp. nov. and Flavobacterium piscis sp. nov., isolated from farmed rainbow trout (Oncorhynchus mykiss). Syst. Appl. Microbiol. 64, 392–399.
Deporte (Orden PRE/1996/2009 de 20 de Julio). Declaration of interest Markus Kostrzewa is an employee of Bruker Daltonik GmbH (Bremen, Germany). Acknowledgments The authors wish to thank Almudena Casamayor and Elisa Pulido for their invaluable technical assistance and Sergio González for his kind help with the design of figures. We are grateful to Dr. V. HirveläKoski, Dr. P. Rintamäki and Dr. S. Viljamaa-Dirks for kindly providing some of the isolates used in this study. References Austin, B., Stobie, M., 1991. Recovery of yellow pigmented bacteria from dead and moribund fish during outbreaks of rainbow trout, Oncorhynchus mykiss (Walbaum), fry syndrome in England. J. Fish Dis. 14, 677–682. Beaz-Hidalgo, R., Alperi, A., Figueras, M.J., Romalde, J.L., 2009. Aeromonas piscicola sp. nov., isolated from diseased fish. Syst. Appl. Microbiol. 32, 471–479. Bernardet, J.F., Segers, P., Vancanneyt, M., Berthe, F., Kersters, K., Vandamme, P., 1996. Cutting a Gordian knot: emended classification and description of the genus Flavobacterium, emended description of the family Flavobacteriaceae, and proposal of Flavobacterium hydatis nom. nov. (basonym, Cytophaga aquatilis Strohl and Tait 1978). Int. J. Syst. Evol. Microbiol. 46, 128–148. Bizzini, A., Greub, G., 2010. Matrix-assisted laser desorption ionization time-of-flight mass spectrometry, a revolution in clinical microbial identification. J. Clin. Microbiol. 48, 1549–1554. Böhme, K., Fernández-No, I.C., Barros-Velázquez, J., Gallardo, J.M., Calo-Mata, P., Cañas, B., 2010. Species differentiation of seafood spoilage and pathogenic gram-negative bacteria by MALDI-TOF mass fingerprinting. J. Proteome Res. 9, 3169–3183. Calderaro, A., Gorrini, C., Piccolo, G., Montecchini, S., Buttrini, M., Rossi, S., Piergianni, M., Arcangeletti, M.C., De Conto, F., Chezzi, C., Medici, M.C., 2014. Identification of Borrelia species after creation of an in-houseMALDI-TOF MS database. PLoS One 9, e88895. Cherkaoui, A., Hibbs, J., Emonet, S., Tangomo, M., Girard, M., Francois, 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. Christensen, J.J., Dargis, R., Hammer, M., Justesen, U.S., Nielsen, X.C., Kemp, M., Danish MALDI-TOF MS Study Group, 2012. Matrix-assisted laser desorption ionization-time of flight mass spectrometry analysis of gram-positive, catalase-negative cocci not belonging to the Streptococcus or Enterococcus genus and benefits of database extension. J. Clin. Microbiol. 50, 1787–1791. De Carolis, E., Vella, A., Vaccaro, L., Torelli, R., Posteraro, P., Ricciardi, W., Sanguinetti, M., Posteraro, B., 2014. Development and validation of an in-house database for matrix-assisted laser desorption ionization-time of flight mass spectrometry-based yeast identification using a fast protein extraction procedure. J. Clin. Microbiol. 52, 1453–1458. Erler, R., Wichels, A., Heinemeyer, E.A., Hauk, G., Hippelein, M., Reyes, N.T., Gerdts, G., 2015. VibrioBase: A MALDI-TOF MS database for fast identification of Vibrio spp. that are potentially pathogenic in humans. Syst. Appl. Microbiol. 38, 16–25. Kurokawa, S., Kabayama, J., Fukuyasu, T., Hwang, S.D., Park, C.I., Park, S.B., del Castillo, C.S., Hikima, J., Jung, T.S., Kondo, H., Hirono, I., Takeyama, H., Aoki, T., 2013. Bacterial classification of fish-pathogenic Mycobacterium species by multigene
352