MALDI-TOF MS for the Identification of Anaerobic Bacteria

University Medical Center Groningen Department Medical Microbiology and Infection Prevention Linda Veloo Expert Center Anaerobic Infections The Netherlands www.mmb-umcg.eu © by author ESCMID Online Lecture Library Matrix Assisted Laser Desorption/Ionization time-of-flight Mass Spectrometry (MALDI-TOF MS)

Time of flight tube Target at 15-25 kV Detector

Ion source © by author

ESCMID- “Time of flight” Online of individual Lecture proteins is converted Library into mass information. - Spectrum is produced - Database is built Veloo et al. Anaerobe 2011; 17:211-212

Workflow

Direct spotting of bacteria on target using a toothpick

Add HCCA matrix

Data acquisition

© Databy analyses author ESCMID Online Lecture Library Log score: <1.7 no reliable identification 1.7 – 2.0 reliable genus identification ≥ 2.0 reliable species identification Obtained spectrum is unique for bacterial species

Intensity © by author ESCMID Online Lecture Library Anaerobic culture

Phenotypic pure culture primary incubation 2 days 2-7 days

aerotolerance identification MALDI-TOF MS 1 day 2-14 days

primary incubation 2-7 days © by author MALDI-TOF MS testing minutes ESCMID Online Lecture Library How many anaerobic bacteria can be identified using MALDI-TOF MS?

UMCG 2011/2012

Total no. of strains 1000 Species ID 650 65% Genus ID 149 15% No ID © 201by author20% ESCMID Online Lecture Library Performance differs per genus

Genus* % species ID % genus ID Clostridium sp. (n=149) 97 3 B. fragilis sp. (n=179) 97 3 Parabacteroides sp. (n=14) 93 7 GPAC (n=133) 85 15 Prevotella sp.(n=83) 78 12 Propionibacterium sp. (n=129) 64 36 Actinomyces sp. (n=28) 57 43 Fusobacterium sp. (n=16)© by author50 50 Campylobacter sp. (n=17) 41 59 Bilophila sp. (n=14) 0 100 ESCMIDDivers (n=26) Online Lecture77 Library13

* Only isolates which could be identified with MALDI- TOF MS (approx. 80% of all isolates) Example from literature

Species identification of clinical Prevotella isolates by Matrix-Assisted Laser Desorption Ionization-Time of Flight mass spectrometry Wybo et al. J. Clin. Microbiol. 2012; 50:1415-1418

Strains: 102 sequenced clinical isolates

Identification MALDI-TOF MS: 1. using the commercial reference database 2. using the commercial reference database + the addition of main spectral profiles (MSP)© ofby 23 referenceauthor strains and 7 sequenced clinical isolates ESCMID Online Lecture Library How to make a Main Spectral Profile (MSP)

x104 Intens. [a.u.] Intens.

1.5 Intens. [a.u.] Intens.

4000

1.0

3000 0.5 © by author

2000

0.0 2000 4000 6000 8000 10000 12000 14000 16000 m/z

ESCMID Online1000 Lecture Library Average of at least 20 spectra 0 5920 5930 5940 5950 5960 5970 5980 m/z Example from literature

Commercial reference database: Commercial database + additions: - 63 % correct species identification - 83 % correct species identification - 11 % correct genus identification - 6 % correct genus identification - 26 % no identification - 11 % no identification

14 % due to the addition of© main by spectra author of species not yet present in database

6 % due to the addition of main spectra of species already present in ESCMIDdatabase Online Lecture Library Summary (I)

1. ± 70 % of the anaerobes can be identified using MALDI-TOF MS

2. Performance differs per species/genus

3. MALDI-TOF MS database needs optimizing for the identification of anaerobic bacteria

- addition of spectra to existing database increases performance - reference strains - clinical© isolates by author ESCMID Online Lecture Library ENRIA

European Network for the Rapid Identification of Anaerobes

Lead: Prof. A.W. Friedrich (The Netherlands) Prof. E. Nagy (Hungary)

Coördination: © by author Dr. A.C.M. Veloo (The Netherlands) Prof. A.J. van Winkelhoff (The Netherlands) ESCMID Online Lecture Library In collaboration with: Dr. W. Pusch (Bruker, Germany) Dr. M. Kostrzewa (Bruker Germany) EU INTERREG ENRIA project

Expertise laboratories:

- UMCG, Groningen, The Netherlands - University of Szeged, Szeged, Hungary - Odense University Hospital, Odense, Denmark - University Hospital Brussels, Brussels, Belgium - Health Protection Agency, London, England - Praticien Hospitalier, Montpellier, France - Public Health Wales, Cardiff, England © by author ESCMID Online Lecture Library ENRIA project Goal:

To optimize the MALDI-TOF MS database for the most common encountered anaerobic bacteria

At least 5 MSP’s present of each species

Strains are sent to the UMCG by the expertise laboratories © by author UMCG: - Culture - Storage - Identification by sequencing ESCMID Online- Ethanol suspension Lecture Library

Validation of the optimized database How does this look like in practice:

Which preanalytical factors influence the quality of the spectrum of anaerobic bacteria???

- Reproducibility of spotting

- Incubation time

- Pretreatment method -© Direct by spotting author - On target extraction - Full extraction

ESCMID -Online Exposure to oxygenLecture Library

Conditions Reproducibility spotting

Amount of bacteria is important:

Too little no peaks with sufficient intensity

Too much saturation of detector and only prominent peaks are measured

ESCMID Online Lecture Library Two examiners: an experienced one and a less experienced one Reproducibility spotting results

24 hours 48 hours 72 hours 96 hours

Pre-treatment method: © by author - direct spotting - on target extraction with 70% formic acid ESCMID Online - full extraction Lecture Library Incubation time results

Incubation time (hours)

Species 24 48 72 96

Gram-negative bacteria Bacteroides thetaiotaomicron 2.23 2.21 2.17 2.19 Prevotella intermedia 1.95a 2.03 2.27 2.04 Fusobacterium necrophorum 2.27 2.18 2.31 2.29 Fusobacterium nucleatumb 2.33 2.19 2.22 2.17 Campylobacter ureolyticus 1.90 2.07 2.01 1.92a Veillonella parvula 2.29 2.45 2.38 2.37

% reliable species ID direct spotting 67 100 100 83 % reliable species ID total 67 100 100 83

Gram-positive bacteria Parvimonas micra 2.33 2.29 2.30 2.27 Finegoldia magna 2.51c 2.43c 2.23c 2.21 Peptoniphilus harei 2.02c 2.23 2.19c 2.20 Peptoniphilus ivorii 1.78c 1.86d 2.02d 1.88d Atopobium minutum 1.96a 2.26 2.29 2.25 Clostridium butyricum 2.29 2.13 2.17 2.08 Clostridium hathewayii© by author2.20d 2.36 2.26 2.36 Actinomyces israellii <1.7e <1.7e <1.7e 2.03 Actinomyces graevenitzii 2.09c 2.36 2.00d 2.11 Actinomyces meyeri 2.27 2.15 2.12 2.20 Bifidobacterium dentium 2.30 2.12 2.37 2.28 ESCMID Bifidobacterium Online longum Lecture 2.18c 2.08c 2.07 Library2.01 Propionibacterium acnes 2.14c 2.24 2.10 2.16 Eggerthella lenta 2.21c 2.20 2.20d 2.10d

% reliable species ID direct spotting 29 71 57 86 % reliable species ID total 79 86 93 93

Conditions

Exposure to oxygen

0 hour 6 hours 24 hours 48 hours

- direct spotting ESCMID Online - on target Lecture extraction with Library 70% formic acid - full extraction Exposure to oxygen results

Exposure to oxygen (hours)

Species 0 1 6 24 48

Gram-negative bacteria Bacteroides thetaiotaomicon 2.14 2.24 2.10 2.17 2.35 Bacteroides stercoris 2.19 2.14 2.18 2.25 2.27 Parabacteroides johnsonii 2.23 2.32 2.29 2.31 2.28 Fusobacterium necrophorum 2.08 2.30 2.19 <1.7a <1.7a Fusobacterium nucleatumb 2.08 2.27 2.32 2.19 2.08 Prevotella intermedia 1.92 1.93 1.86 1.95c <1.7a Prevotella oris 2.08 2.17 2.01 2.35 2.29 Alistipes onderdonkii 2.21 2.24 2.29 2.29 2.21 Veillonella parvula 2.25 2.13 2.2 2.20 2.16

% reliable species ID direct spotting 89 89 89 78 78 % reliable species ID total 89 89 89 78 78

Gram-positive bacteria Finegoldia magna 2.24 2.05c 2.34 2.52 2.50 Peptoniphilus harei 2.15 2.00 2.13 2.23b 2.19 Peptoniphilus ivorii © by<1.7 d author1.81e <1.7d 1.76e 1.80e Clostridium hathewayii 2.36 2.23 2.29 2.07 2.19 Clostridium ramosum 2.03 2.19 2.27 2.06 2.17 Actinomyces graevenitzii 2.05c 2.03e 2.11 2.06e <1.7d Actinomyces meyeri 2.03 2.27 2.35 2.19 2.25 Bifidobacterium longum 2.00c 2.12c 1.99c 2.15 2.17 ESCMID Propionibacterium Online acnes 2.18 Lecture 2.14c 2.12 2.13 2.13 Library Eggerthella lenta 2.11 2.30 2.27 2.16 2.02e Collinsella aerofaciens 2.26c 2.16 2.23c 2.32c 2.28c

% reliable species ID direct spotting 64 55 73 64 64 % reliable species ID total 91 82 82 91 82

Summary (II)

The fact whether an unknown anaerobic bacterium can be identified using MALDI-TOF MS mainly depends on:

- The person who is spotting

- The type of colony

- Amount of bacteria spotted © by author - If sufficient MSP’s are present in the MALDI-TOF MS database ESCMID Online Lecture Library Recommendations in order to obtain a good quality spectrum

- perform the MALDI-TOF MS measurement after 48 hours of incubation in an anaerobic environment

- spot the bacteria on the same day

- perform an on target extraction with 70% formic acid for gram-positive anaerobic© bybacteria author

- have the spotting performed by an experienced examiner ESCMID Online Lecture Library Accomplished so far:

Collected: ± 650 strains © by author ESCMID Online Lecture Library Preliminary results ENRIA

Misidentifications:

Sequence ID MALDI-TOF MS ID

Anaerococcus vaginalis Anaerococcus hydrogenalis

Bacteroides dorei Bacteroides vulgatus

Bacteroides cellulosilyticus Bacteroides intestinalis

Parabacteroides merdae Parabacteroides johnsonii © by author Clostridium rectum/ Fusobacterium mortiferum Fusobacterium mortiferum ESCMID Online Lecture Library Moryella indoligenes Fusobacterium naviforme

No MSPs present in the database Dendrogram Anaerococcus vaginalis/hydrogenalis

Bacteroides massiliensis ENR0260 Bacteroides massiliensis ENR0230 Bacteroides massiliensis ENR0154 Bacteroides massiliensis DSM 17679T DSM Bacteroides vulgatus PNU 71838 PNU Bacteroides vulgatus HU40347_2 PNU Bacteroides vulgatus PNU 1536_5 PNU Bacteroides vulgatus DSM 3289 DSM Bacteroides vulgatus DSM 1447T DSM Bacteroides dorei ENR0141 Bacteroides dorei ENR0140 Bacteroides dorei ENR0139 Bacteroides dorei ENR0137 Bacteroides stercoris ENR0247 Bacteroides eggerthii ENR0264 Bacteroides eggerthii ENR0220 Bacteroides ovatus DSM 1896T DSM Bacteroides ovatus 53152 PNU Bacteroides xylanisolvens ENR0364 Bacteroides ovatus HU36898_1 PNU Bacteroides ovatus IBS_MS_21 IBS Bacteroides ovatus 32456_1 PNU © by author Bacteroides ovatus 11483_2 PNU Bacteroides finegoldii ENR0155 Bacteroides finegoldii DSM 17565T DSM Bacteroides caccae 103 PIM Bacteroides caccae DSM 19024T DSM Bacteroides caccae ENR0145 ESCMID Online Lecture BacteroidesLibrary caccae ENR0144 1000 900 800 700 600 500 400 300 200 100 0 Distance Level Identification of a B. dorei strain after addition of B. dorei MSP’s

B. dorei, log score 2,565

MSP Dendrogram

Parabacteroides merdae ENR0277 Parabacteroides merdae ENR0273 Parabacteroides johnsonii DSM 18315T DSM Parabacteroides merdae ENR0283 Parabacteroides goldsteinii DSM 19448T DSM Parabacteroides goldsteinii BB28 PNU Parabacteroides goldsteinii BB02 PNU Parabacteroides distasonis RV_00005_09 ERL Parabacteroides distasonis HU35718_3 PNU Parabacteroides distasonis SW26 PNU Parabacteroides distasonis HU41816_4 PNU Parabacteroides distasonis HU36220 PNU Parabacteroides distasonis RV 412_0209_5 LBK Parabacteroides distasonis DSM 20701T DSM Parabacteroides distasonis 53181 PNU Bacteroides massiliensis ENR0230 © by authorBacteroides massiliensis ENR0260 Bacteroides massiliensis ENR0154 Bacteroides massiliensis DSM 17679T DSM Bacteroides intestinalis DSM 17393T DSM Bacteroides intestinalis 110706_E9 LUMC Bacteroides intestinalis SW20 PNU ESCMID Online LectureBacteroides cellulosilyticusLibrary ENR0168 Bacteroides cellulosilyticus ENR0149 Bacteroides cellulosilyticus ENR0228 Bacteroides cellulosilyticus ENR0237 Bacteroides cellulosilyticus ENR0236

1000 800 600 400 200 0 Distance Level Phylogenetic tree of fusobacteria

MSP Dendrogram

Fusobacterium nucleatum BK635_11 ERL Fusobacterium nucleatum 100617_02 PNU Fusobacterium nucleatum 110706_B8 LUMC Fusobacterium naviforme DSM 20699 DSM Fusobacterium naviforme DSM 20699 BRB Fusobacterium canifelinum DSM 15543 BRB Fusobacterium canifelinum DSM 15542T BRB Fusobacterium necrophorum ssp necrophorum DSM 20698 DSM Fusobacterium necrophorum ssp necrophorum DSM 21784T DSM Fusobacterium gonidiaformans DSM 19810T DSM Fusobacterium gonidiaformans 110706_B5 LUMC Fusobacterium varium DSM 19868T DSM Fusobacterium varium DSM 19868T BRB Fusobacterium mortiferum DSM 19809T BRB Clostridium rectum ENR0178 Clostridium rectum ENR0317 Clostridium rectum ENR0170 Clostridium perfringens RV_BA_03_D LBK © by authorClostridium perfringens HU51221 PNU Clostridium perfringens DSM 798 VML Clostridium difficile DSM 1296T DSM Clostridium difficile DSM 12057 DSM Clostridium clostridioforme ENR0380 Clostridium clostridioforme 110706_G8 LUMC Clostridium clostridioforme DSM 933T DSM ESCMID Online LectureClostridium clostridioforme Library 1021_NCTC 11224T BOG Bacteroides pyogenes DSM 20612 DSM Bacteroides pyogenes DSM 20611T DSM Bacteroides fragilis DSM 9669 DSM Bacteroides fragilis DSM 2151T DSM 1000 900 800 700 600 500 400 300 200 100 0 Distance Level Dendrogram Campylobacter gracilis: unexpected finding

MSP Dendrogram

Campylobacter gracilis DSM 19528T DSM

Campylobacter gracilis CCUG 27720T NVU

Campylobacter gracilis ENR0074

Campylobacter gracilis 0057

Campylobacter gracilis 0027

Campylobacter gracilis ENR0036

Campylobacter gracilis 0058

Camypylobacter gracilis ENR0078 ESCMID Online LectureCampylobacter gracilisLibrary ENR0073 1000 900 800 700 600 500 400 300 200 100 0 Distance Level % sequence similarity of ENR strains with C. gracilis and each other: >99 % Achieved so far

Number of species

Species originally available with >5 MSP’s 35 Species available with >5 MSP’s after 113 optimization Species improved (available now with 2 to 5 © by author 94 MSP’s) SpeciesESCMID that are new Online in the database Lecture Library33 Estimated end result of strains

Total no. of strains: ± 700

Number of species

Species originally available with >5 MSP’s 35 Species available with >5 MSP’s after 175 optimization © by author Species improved (available now with 2 to 5 198 MSP’s)ESCMID Online Lecture Library Species that are new in the database 50 Future plans

Validation of the optimized database:

- proficiency testing validating laboratories

- performed by several laboratories throughout Europe

- validation of optimized database with known clinical isolates

Initiated by “Neglected Infections”of the EurSafety Health-net, which is financially supported by the EU, Niedersachsen, ESCMID OnlineNordrheinWestfalen Lecture and the Dutch border Library Provinces. Ringtest

Validating laboraties

7 core laboratories © by author ESCMID Online Lecture Library © by author ESCMID Online Lecture Library © by author Thank you for your attention ESCMID Online Lecture Library