Development of a toxR-based real-time PCR assay for V. parahaemolyticusA. Powell1, R. Griffin2, C. Baker-Austin1 and R. Hartnell1

1Centre for Environment, Fisheries and Aquaculture Science, Weymouth Laboratory, Weymouth, Dorset, England (andy.powell@cefas.co.uk), 2University of Surrey, Guildford, Surrey, England

IntroductionVibrio parahaemolyticus is a Gram-negative bacterium found naturally in marine and estuarine waters. V. parahaemolyticus is halophillic, oxidase positive, and motile with a single polar flagellum. Globally, V. parahaemolyticus is a significant cause of foodborne disease, and is considered the leading agent of bacterial illness associated with seafood consumption (Joseph et al. 1983). An estimated 40,000 people contract V. parahaemolyticus infections each year in the USA (Scallan et al. 2011), highlighting the clinical burden associated with this pathogen. Several large outbreaks of V. parahaemolyticus have also emerged in Europe in the last decade (Baker-Austin et al. 2010), highlighting the need for improved global surveillance systems. Clinical characteristics of V. parahaemolyticus infections include abdominal cramps, diarrhoea, nausea, headaches, fever, and chills (Honda and Iida 1994, Hardy and Klontz 1996). V. parahaemolyticus symptoms typically resolve in less than 72 hours. A small number of cases have been reported that persist for up to 10 days, but the majority of these cases tend to involve immune-compromised individuals. Several studies have demonstrated that toxR represents the most reliable species-specific molecular target for the identification of V. parahaemolyticus (Croci et al. 2007). Although several conventional PCR assays have been developed and published that target toxR no real time PCR approaches have been published to date. We describe here a real-time PCR assay for toxR that is both specific, reliable, fully quantitative and rapid as well as easy to use.

Materials and methods

Primer and probe design and RT-PCR assayThe complete nucleotide sequences (open reading frame regions only) for all full length toxR nucleotide sequences were aligned using clustalW (Thompson et al. 1994). A segment of the toxR genes was analysed using primer express software from DNASTAR (Madison, WI, USA). The probe and primers were subsequently assessed for species as well as strain specificity using a BLAST search against publically available databases. The primer and probe sequences are listed in Table 1. Our PCR primers and TaqMan probe were designed to specifically target the toxR gene (Fig 1, Table 1). For real-time PCR experiments, the assay comprised of 25 µl, consisting of 12.5 µl TaqMan Universal PCR Master Mix (Applied Biosystems), 0.45 µl each of forward and reverse V. parahaemolyticus primers (100 nM), 5.6 µl nuclease-free water, and 1 µl of probe (500 nM) (Table 1). Five microliters of template (either extracted DNA, plasmids or boiled cell lysate) was then added, and each reaction was performed in triplicate. An ABI PRISM® 7500 Sequence Detection System (CGRB Core Laboratories), was used, and a thermal cycle performed with a two-step PCR protocol: 1 cycle at 95 °C for 10 min followed by 40 cycles at 95 °C for 15 s and 60 °C for 90 s. To test the efficacy of our RT-PCR assay, V. parahemolyticus strains, encompassing geographically diverse isolates, were grown at 28oC for 24 h in tryptone soy broth or on solid agar media, supplemented with 5 g/liter NaCl. Strains were cryogenically stored at -80oC prior to use, supplemented with 20% (vol/vol) glyerol. A total of 77 bacterial strains, including V. parahaemolyticus (n = 60), other Vibrio species (n = 11) and distantly related reference strains (n = 6) was also used to assess the specificity of the oligonucleotide probe and primer sets used (Table 2).

Shellfish bioaccumulation experimentsShellfish bioaccumulation experiments. In this study, two 50 litre-capacity marine tanks were filled with sterile seawater, and maintained at 15 oC (± 1 oC) with constant air sparging. Un-depurated live C. gigas (60) were obtained via a wholesaler, and were evenly distributed into tanks for 24 hours to acclimatize, prior to the addition of bacterial suspensions. For bioaccumulation experiments, a previously identified pathogenic strain of V. parahaemolyticus was grown overnight to late logarithmic stage and subsequently used in all experiments. For each sample, the digestive glands (stomach and digestive diverticula) were removed from each animal and weighed. The digestive glands were subsequently pooled together, and then finely chopped using a sterile razor blade. Homogenates were then prepared by treating the chopped digestive glands with 100 μg/ml Proteinase K solution (30U/mg; Promega) and DNA extracted, as previously described (Baker-Austin et al. 2010). Homogenates were stored at 4 ºC prior to testing using RT-PCR (Figure 1).

ResultsTable 2 summarizes the results of the analyses of the 77 strains examined by real-time PCR. The results demonstrated that the toxR RT-PCR was capable of correctly identifying V. parahaemolyticus on all tested strains (100% accuracy). No amplification of closely related Vibrio or non-Vibrio strains were observed, indicating specificity of the assay (Table 2). Real-time PCR assays were performed on DNA extracted from bioaccumulated shellfish samples. We were able to identify toxR directly from shellfish matrices (data not shown), whereas no signal was evident from control samples, or from extracted water. Using cloned toxR material we demonstrated that the assay could detect down to less than 10 genome copies of target per reaction volume. A serial dilution range of cloned toxR material demonstrated excellent expected dynamic range, with amplification curves approximately 3.3. ct values apart (Figure 2).

ConclusionsIn summary, this real-time PCR assay represents a rapid means of identifying V. parahaemolyticus strains. The assay demonstrated excellent discriminatory powers, as it was confirmed to be 100% accurate in identifying V. parahaemolyticus isolates in our strain library. The assay can be applied directly on isolated strains and bacterial DNA, but also on extracted shellfish matrices. Future work should include the use of this assay alongside pathogenicity markers in V. parahaemolyticus such as trh and tdh (Nordstrom et al. 2007) in a multiplex format. The use of this assay on naturally contaminated shellfish samples should also be a key focus of future work. This tool will enable early detection capability in a range of different applications, such as food processing, regulatory and clinical settings.

Figure 2. Serial dilution of cloned toxR target showing expected amplification curves across a series of 5 orders of magnitude dilution.

Table 1. Primers and probes used in this study

Name Sequence (5’-3’)toxR - F GAACCAGAAGCGCCAGTAGTtoxR - R AAACAGCAGTACGCAAATCGProbe (FAM)- TCACAGCAGAAGCCACAGGTGC-(TAMRA)

Table 2. Percentage accuracy of assay

V. parahaemolyticus strains 60/60 (100% accurate)

Other Vibrio spp.V. alginolyticus ND (8 strains, 0 positive)V. cholerae ND (2 strains, 0 positive)V. vulnificus ND (1 strain, 0 positive)

Other bacteriaS. paucimobilis (NCTC 11030) NDP. aeruginosa (NCTC 10332) NDP. mirabilis (NCTC 10975) NDE .coli (NCTC 12241) NDK. aerogenes (NCTC 9528) NDE. faecalis (NCTC 775) ND

Figure 1. Schematic of major steps used to extract nucleic acid from shellfish matrices

2. Glanding. Diges�ve glands from each shellfish (C. gigas) are cut away from shellfish meat and trimmed with a sterile scalpel.

3. Physical disrup�on. Diges�ve glands from all animals are combined in a sterile petri dish and weighed. Glands are mashed in the presence of equal (w/v)100 mg ml-1 Proteinase K solu�on (30 U mg-1.

5. Nucleic acid extrac�on. Each 500 µl aliquot is processed using a Biomerieux magne�c silica extrac�on procedure, using proprietary alkaline lysis coupled to magne�c silica purifica�on.

1. Shucking. 10 shellfish (C. gigas) are shucked, and the meat from each animal is removed.

4. Incuba�on. Samples are added to a sterile 50 ml falcon tube, and shaken gently at 37 ºC for 1 hour. Samples are then placed in a 65 ºC water bath for 15 min. Samples are subsequently centrifuged (5 min, 3K) and these homegenates are collected in 500 µl aliquots.

2-2.

5 ho

urs

Persistence  of  human  pathogenic  V.  cholerae  in  cold  water  condi5ons  and  a6er  temperature  up-­‐shi6  

Aims    

•  Is  a  human  pathogenic  strain  persistent  in  aqua5c  environment?  

•  Is  the  culturability  of  a  clinical  and  a  marine  strain  in  cold  water  comparable?  

•  How  will  the  strains  react  on  a  temperature  up-­‐shi6  to  20°C  a6er  three  weeks  at  4°C?  

•  Is  natural  sediment  a  favourable  reservoir  for  human  pathogenic  bacteria?  

•  May  the  ability  to  regain  culturability  be  a  more  important  pathogenic  trait  than  the  harbouring  of  virulence  genes?  

•  The  clinical  strain  lost  culturability  directly  a6er  inocula5on,   but   regained   and   peaked   in  number  one  week  a6er  temperature  up-­‐shi6.    

•  The   marine   strain   lost   were   culturable   un5l  temperature  shi6,  when  it  lost  the  culturability  completely.  

•  The  total  cell  count,  shown  by  real-­‐5me  qPCR,  was  quite  stable  throughout  the  experiment  for  both  strains.  

•  The  clinical  strain  was  culturable  for  one  more  week  at  4°C  in  the  sediment  than  in  the  water  microcosm.   A6er   temperature   up-­‐shi6,   the  strain  regained  culturability.    

•  The   total   cell   count   was   lower   than   in   the  water   microcosm,   which   may   be   due   to  compe55on  from  other  microorganisms.    

•  The   natural   boLom   sediment   was   a   suitable  reservoir   for   the   clinical   strain,   despite  compe55on.  

Set-­‐up:   the  experiments  were  performed   in  50  ml  microcosms.  Filtered   seawater  microcosms  and  natural   sediment  microcosms  collected   from   the   Sound   (Öresund)   was   added   to   the   microcosms   and   then   bacteria   (OD   1.0)   was   added   to   the   different  microcosms.  The  flasks  were  incubated  at  4°C  for  three  weeks  and  then  the  temperature  was  increased  to  20°C  and  kept  at  that  temperature  for  another  week.  At  8  sampling  points,  the  CFU/ml  was  determined  on  TCBS  agar  plates  and  Vibrio/ml  by  qPCR.  

Conclusions    •  The  ability  of   the  clinical  Vibrio   strain  to  quickly  adapt  to  temperature  

changes   and   thereby   regain   culturability   a6er   a   temperature   increase  may  be  more   important   for  determining   the  pathogenicity   to  humans  than  the  presence  of  the  screened  virulence  genes.    

•  To   inhabit   sea   sediment   seems   to   be   a   successful   strategy   of   clinical            V.  cholerae  strains  to  survive  low  water  temperatures.  

Fig.  1.  The  mean  cell  abundance  (±SD)  of  the  clincal  and  marine  V.  cholerae  strains  cultured  in  sterile  seawater,  presented  as  the  CFU  ml-­‐1  and  total  Vibrio  ml-­‐1  based  on  real-­‐5me  qPCR  analysis.  The  temperature  was  increased  from  4  to  20°C  a6er  21  days.  Clinical  strain  CFU▲,  marine  strain  CFU●,  clinical  strain  qPCR  Δ,  marine  strain  qPCR○.    

Fig.  2.  The  mean  (±SD)  cell  abundance  of  the  clinical  V.  cholerae  strain  in  natural  sediment.  The  temperature  was  increased  from  4  to  20°C  a6er  21  days.  CFU  ml-­‐1  ▲;  total  Vibrio  ml-­‐1  calculated  from  qPCR  data  Δ  .    

!

!

BeLy  Collin,  Bodil  Hernroth  and  Ann-­‐Sofi  Rehnstam-­‐Holm  Kris5anstad  University,  University  of  Gothenburg  and  Sven  Lovén  Center  for  Marine  Sciences  

BeLy  Collin  

Kris5anstad  University  

Leibniz Institute for Baltic Sea Research Warnemünde

Bianca Schippmann | Gerald Schernewski Leibniz Institute for Baltic Sea Research Warnemuende, Germany

Introduction

Modelling Vibrio transport in a Baltic Sea Bay - A promising tool for coastal management?

Water quality measurements are often taken only on fixed stations and are hence spatially restricted. In a best case situation (no gradients) a good temporal distribution of the organisms under study can be achieved at one single spot. However, a conclusion of the situation in a larger area (several kilometers) is not possible since there is a high spatial variance of the organism distribution. Modelling particle pathways can fill this gap and has been shown to be a promising tool for coastal management (Schernewski et al. 20121, Schippmann et al. submitted2). Vibrio vulnificus has often occurred in the Bay of Greifswald and at the coast of Usedom during the last decade (3). It caused several wound infections and deaths. With the simulation of transport and spatial behaviour of the organisms, analyses of the risk and potential sources are possible which allow forecasts for the along coast bathing areas.

References | 1Schernewski, G., Fischer, E., Huttula, T., Jost, G., Ras, M., 2012. Model tools to support bathing water quality management: Escherichia coli bacteria in a Baltic lagoon. Journal of Costal conservation 2Schippmann, B., Schernewski, G and Graewe, U., submitted, Escherichia coli pollution in a Batic Sea lagoon: A model-based source and spatial risk assessment 3Landesamt fuer Gesundheit und Soziales Mecklenburg-Vorpommern

Acknowlegdements

Hans Burchard, Ulf Graewe

North-German Supercomputing

Alliance, Knut Klingbeil,

Richard Hofmeister

Methods

Tracks:

Figure 3: Tracks of 5000 salmonella bacteria emitted off the Wolin coast in Misdroy (Poland). T90 survival time of salmonella 142 h. Different colours indicate different particle tracks.

Concentrations:

Figure 4: Salmonella density according to the computed tracks in order of magnitude per m3 .

General Estuarin Transport Model (GETM): 3D free-surface primitive equation model based on Navier-Stokes equations computes physical state variables (temperature, salt, currents, diffusivity,..) and provides flow fields for the particle tracking model (Figure 2) model setup with a resolution of 100x100m (Figure 1)

information about Vibrio behaviour and properties at e.g. changing light, salinity and temperature conditions

infection dose to quantify the measurements and model

results (risk assessment)

Challenges: model vibrio :

• dependence on salt and temperature (model extension) • mortality and growth rate

find the correct number of start concentration

Possibilities

General Individuals Tracking Model (GITM): 3D Lagrangian particle tracking model computes particle movement according to dvection/diffusion forced with 3D flow fields contains a bio-module for particle properties (sinking speed, growth, …) for modelling algae, mussels, etc. following results from a salmonella simulation are shown exemplarily (Figure 3 & 4) Figure 2: Flow field in the Pomeranian Bight computed with GETM for salmonella particle tracking simulations.

Needs & Challenges

Needs:

small -spatial monitoring program with high resolution (not

only fixed stations where temporal and spatial uncertainties

too high)

monitoring of boundary conditions:

• temperature and salinity

• wind conditions

• exact measure position and water depth

Figure 1: Model domain and bathymetry of the future vibrio setup in Bay of Greifswald. The background shows every 20-th grid.

steady state simulations to determine risk at region under different wind, river, etc. conditions realistic simulations to show spatial distribution at specific dates

DYNAMICS,  DIVERSITY  AND  PATHOGENIC  POTENTIAL  OF  VIBRIOS  IN  THE  HEMOLYMPH  OF  PACIFIC  OYSTERS

Wendling CC, Wegner KMHelmhotz Centre for Ocean Research Kiel, Düsternbrooker Weg 20, 24105 Kiel, Germany

Alfred Wegener Institute for Polar und Marine Research Sylt, Hafenstraße 43, 25992 List/ Sylt, Germany

         Mass  mortality  events  of  Pacific  oysters  Crassostrea  gigas  triggered  by  microbs,  such  opportunis9c  Vibrio  in  conjunc9on  with  high  

temperatures   and   physical   weakness   a=er   matura9on   have   been   observed   worldwide.   However,   studies   addressing   the   naïve  

community  structure  of  oyster  associated  Vibrio,  thereby  evalua9ng  the  risk  of  mortality  events,  are  rare.  We  aimed  at  characterizing  

the  dynamics,  diversity  and  pathogenic  poten9al  of  the  Vibrio  community  structure  isolated  from  the  Pacific  oyster  around  the  island  

of  Sylt,  North  Sea,  Germany,   a  coastal  area  where  oyster  mortality  has  not  been  observed,  yet.  A   total  of  78  Vibrio   isolates  were  

obtained  during   a   six  months   in   situ   study  by   repeated  hemolymph   sampling   and  the  genomic  diversity  based  on  16s  rRNA  was  

analyzed.  Dynamics  and  diversity  of  Vibrio  varied  significantly  over  9me  and  were  highest  a=er  matura9on.  By  means  of  controlled  

infec9on  experiments  we  provide  first  evidence  that  pathogenic  Vibrio  are  indeed  present  around  the  island  of  Sylt  and  discuss  the  

risk  of  future  oyster  summer  mortality  in  the  context  of  predicted  global  warming.  

SUMMARY

Hemolymph  sampling  from  wild  oysters  

Isola9on  of  Vibrio  spp. Iden9fica9on  of  Vibrio  spp.  based  on  16s  rRNA

Construc9on  of  a  phylogene9c  tree  based  on  16s  rRNA  including  known  virulent  strains

Controlled  infec9on  experiments  to  estaimate  pathogenic  poten9al

Determine  pathogenic  poten9al  by  induced  virulence

1 2 3 4 5 6

METHODS

Contact:Carolin  C  Wendling

Alfred  Wegener  Ins9tute  for  Polar  and  Marine  ResearchHafenstraße  43  -­‐  25992  List

mail:  carolin.wendling@awi.dephone:  +49  4651  952  4206

๏Characterize the dynamics and diversity of the naïve   Vibrio   community   structure   of   healthy  

Pacific  Oysters  

๏Estimate the pathogenic potential of oyster associated Vibrios

AIM Occurence of oyster mass m

ortalities

Sampling Site: Sylt

No observed oyster mass m

ortalities

This  is  a  first  survey  aiming  at  the  characteriza9on  of  the  dynamics,  diversity  and  pathogenic  poten9al  of  oyster  associated  Vibrio   in  a  coastal  area  that  has  not  experienced  oyster  summer  mortality,  yet.  High  Vibrio  abundance  

a=er  matura9on  and  the  evidence,  that  pathogenic  Vibrio  strains  naturally  exist   in  apparently  healthy  oysters  point  to  the  risk  of  poten9al  oyster  mortality  outbreaks  when  average  summer  water   temperatures  increase  with  

global  climate  change.  As  summer  mortali9es  of  pacific  oysters  have  been  observed  solely  200  km  southward,  ongoing  research  focusing  on  the  study  of  the  Vibrio  ecology  including  methods  to  predict  the  pathogenic  poten9al  of  

a  strain  are  pivotal   in  this  area,  where  oyster  aquaculture  is  prominent.  We   intend  to  con9nuously  monitor   the  dynamics  and  diversity  of  Vibrio   in  oyster  hemolymph,   thereby  detect  first  signs  of  summer  mortality  indicated  

through  a  change  in  the  natural  community  structure.  The  ability  to  iden9fy  epidemiological  paaerns  linked  to  bacterial  community  structure  and  specific  dynamics  of  candidate  pathogens  is  necessary,  especially  when  infec9ous  

diseases  are  predicted  to  increases  with  global  warming.

CONCLUSION

0.1

Legend:

season

location

Color ranges:

V. splendidus

V. cyclitrophicus

V. crassostreae

Allivibrio fisherii

V. tubiashii

V. alginolyticus

V. bregonii

V. cholerae

Shewanella

Tx5.2

enuJ

b81

O

O46

bOct

D01

wJu

ne

O33

bJun

e

O33

wM

ay

D35

bJul

y

O02

wOct

D28b

July

D41b

June

O12

bJun

e

D12w

Oct

D07b

Sept

O33wS

ept

D06wSe

pt

O43bA

ug

O05bJ

une

O22wOctTx6.2D35wJuneD12bSeptO47bOctO33bMayO18bMayO18wMay

D35wMayD45bMay

O03wOct

Tx4wD27bMay

D10wSept

D19wOct

D35bJune

D10w

Oct

O07wJuly

Tx03w

Tx3O07wSept

O33wJuly

O02wSept

D15wJuly

D29w

Sept

D14w

June

O45wJulyD28wJuly

O44wSept

O45wMay

Tx5.1

D35wJuly

D30wJuly

D12wJuly

O07bJuly

O32wSept

O32bSept

D07wSept

D12bJulyO33wOct

D03wJulyO03wAug

D03bJulyD29bAugO20wAugD05bAugD46wAugD20wAugD17wOct

Vibrio splendidus biovar II AB038030

D10bAug

Tx3b

O05bJuly

O24bSept

D29wAug

Vibrio tubiashii X74725O44wAugD47wSept

Vibrio anguilla

rum X16895

Vibrio

aestu

arian

us X74

689

Vibrio

algino

lyticu

s X74

690

Vibrio

parah

aemoly

ticus

X7472

0

O46bJ

ulyTx2b

O05wJ

uly

O24

wSe

ptO

03bS

ept

D47b

Aug

D41w

June

O45

bJun

eO3

3wJu

neVibr

io c

hole

rae

X763

37

D15

bJul

y

O20

bAug

D28

bSep

t

D17

bOct

V.  splendidus

V.  cyclitrophicus

V.  crassostreae

V.  tubiashii

V.  alginoly9cus

V.  bregonii

V.  cholerae

Shewanella

Allivibrio  fisherii

Figure  3  Phylogene9c  rela9onship  based  on  Maximum  likelihood  method  (TN  93  +  G+  I  subs9tu9on  model)  using  par9al  16s  rRNA  sequences  (1095  bp)  of  Vibrio  isolates  sampled  from  oyster  hemolymph.  Bootstrap  percentages  above  80%  are  represented   by   grey  dots   at  the  parent  nodes.  Sampling  site  is   depicted  by  grey  colours   on  the  first  outer  ring  for  two  different  sampling  sites   on   Sylt,  i.e.  Dietrichsenbank   (DB)  and  Oddewaa   (OW)  and   for  a  third   sampling  site  on  the  island  of   Texel,  Netherlands   (Txl).  Sampling  month  is  depicted  by  blue  and   green   colours  on   the  second  outer  ring.  Assignment  to   the  different  clades   (see  legend)  based  on  99%  sequence  similyrity  by  BLASTN  analysis  in  depicted  in  colour  on  the  inner  ring.

Survival  (%

)

0 2 4 6 8 10

0.0

0.2

0.4

0.6

0.8

1.0

Control

D14w_june

O7w_julyD29w_augTx3w

D45w_may

D47w_sept

Tx5.1D29w_sept

Tx2b

O43b_aug

Time  post  Infec9on  (days)

Figure  1  Kaplan-­‐Meier  death  censored  survival  by  selected  Vibrio  strains  B-­‐L  and  control  A  (n=15).  Colour  codes  correspond  to  those  in  Figure  3.  

RESULTS

๏  Vibrio  bacteria  are  constantly  present  in  the  hemolyph  of  Pacific  Oysters  in  the  North  Sea  

๏  Vibrio  load  is  highest  a=er  matura9on

๏  V.  splendidus  is  dominant  when  water  temperatures  are  below  18°C  -­‐  V.  crassostreae  is  dominant  

when  water  temperatures  are  above  18°C

๏  Vibrio  diversity  is  posi9vely  correlated  with  water  temperature  

๏  Highly  pathogenic  strains  are  present  in  North  Sea  oysters

0

2

4

6

8

Index

Cel

ls/..

l

Month

tem

p$Te

mp

May Jun Jul Aug Sep Oct Nov

0

5

10

15

20

Colony  fo

rming  Units  (cells/µl  hem

olym

ph)

Temp  (°C)

Figure   2   Seasonal   varia9on   in   amount   of   Vibrio   bacteria   isolated   from   oyster   hemolymph   during   monthly   sampling   trips.   (A)  Quan9fica9on  of  total  Vibrio  spp.  (open  squares  and  rectangles)   isolated   from  oyster  hemolymph,  water  temperature  (full  circle)  and  occurrence  of  spawning  period  (black  arrows).  (B)  Rela9ve  propor9on  of  Vibrio  isolates  by  phylogene9c  associa9on  based  on  99%  similarity  by  BLASTN  analysis.  Colour  codes  correspond  to  those  in  Figure  3.

(A)

(B)

CONCLUSION

1,00E+00

1,00E+02

1,00E+04

1,00E+06

1,00E+08

1,00E+10

0 1.5 3 4.5 6 24

CFU

/g

Time post inoculation [h]

MDD MR IVW CV

spiking

0 1 2 3 4 5 6 7 8 9 10 110 1 2 3 4 5 6 7 8 9 10 11

Time post inoculation [d]

MDD vacuum 3°C MDD gunnysack 3°C

MDD vacuum 15°C MDD gunnysack 15°C

1,00E+00

1,00E+02

1,00E+04

1,00E+06

1,00E+08

0 1 2 3 4 5 6 7 8 9 10 11

In vivo survey of Vibrio spp. in artificially contaminated Mytilus edulis

Doreen Herrfurth1, Keike Schwartz2, Ralph Pund2, Eckhard Strauch2, Thomas Alter1, Stephan Huehn1

1 Institute of Food Hygiene, Free University Berlin, Germany 2 Federal Institute for Risk Assessment, Berlin, Germany

Background All studies performed on the appearance of Vibrio spp. in mussels on pre-harvest and retail level in Germany so far showed high loads with these microorganisms. As far as food safety is concerned Vibrio (V.) parahaemolyticus, V. vulnificus and V. cholerae are the most important species within this genus. In a previous study, 106 retail mussel samples of different species and origin were investigated for the presence of Vibrio. 54.8 % of the samples were positive for Vibrio spp. (approx. 59 % V. alginolyticus, 12 % V. cholerae (non-O1, non-O139), 6 % V. parahaemolyticus). In addition, single V. metschnikovii, Photobacterium damselae and Listonella anguillarum were found. One of the most frequently consumed mussel species in Germany is the blue mussel, Mytilus edulis. This mussel species filtrates approx. 1.5 l of water per hour, thus the surrounding water is clarified effectively (Riisgard, 1988). Along with the water, pathogenic agents, e. g. bacteria (including Vibrio

spp.), viruses or toxic algae are affiliated. This potential accumulation of Vibrio spp. can cause food borne infections if mussels are not properly heated before consumption. The aim of this study was to determine the clearance kinetics of Mytilus edulis after artificial incubation with V. cholerae, V. parahaemolyticus or V. vulnificus. Additionally, the change of Vibrio loads in mussels during storage on ice and at 15 ºC was investigated.

Material and Methods

Clearance In clearance assays approx. 100 mussels were transferred after 24 h incubation with approx. 1011 CFU (5 x 106 CFU/ml) for each Vibrio spp. into the husbandry unit to cleanse the mussels from Vibrio. The accumulation of Vibrio spp. and clearance of mussels was observed over 7 days.

Storage For the storage test approx. 200 mussels were contaminated with 1010 CFU of Vibrio spp. which corresponds to 5 x 105 CFU/ml in the contamination unit. After 24 hours of incubation mussels were packed in vacuum bags and gunnysacks. A vacuum bag contained five mussels with artificial seawater, whereas each gunnysack contained five mussels only. Samples of both packing strategies were stored for ten days at 3 ºC on ice and at 15 ºC. Vibrio spp. load was investigated every day for storage at 3 ºC and subsequently every two days for storage at 15 ºC.

Results

Discussion Semi-quantitative methods were used to describe Vibrio concentrations within contamination ponds and clearance ponds as well as in mussel lumen water, digestive glands and the remaining mussel meats. The bioaccumulation assays in Mytilus edulis showed an enrichment of Vibrio especially within digestive glands. After 1.5 h the concentration in the digestive glands was approx. 2 log higher than the corresponding concentration in the tank water. This confirms the 200-fold enrichment described by DePaola et al. (2000). At the end of incubation the loads of Vibrio spp. were reduced by 1 - 2 log to approx. 104 – 106 CFU/g in all sample segments. During clearance a slow but continuous reduction of Vibrio concentrations was determined within a seven day period in all mussel segments. Vibrio loads started in digestive glands at 106 CFU/g and were reduced by 4 – 5 log to approx. 102 – 103 CFU/g in the husbandry unit for seven days. Based on the data, a general clearance rate of 0.023 log-CFU/g per h was calculated (Morisson et al. 2011). Once a mussel is contaminated with Vibrio, which can occur very likely, even a clearance over a seven day period seems to be insufficient to clear the mussel completely. Within ten days of storage at 3 ºC on ice or at 15 ºC almost no reduction of the Vibrio load occurred in Mytilus edulis for V. parahaemolyticus and V. cholerae, only variations in a 2 log interval. In contrast, in V. vulnificus a cell reduction can be seen, probably due to the fact that in the assay a human isolate and no mussel isolate was used. Therefore all Mytilus edulis should be thoroughly heated before consumption to avoid potential infection.

Clearance assay Storage assay

Fig. 4: Cell count of V. cholerae,

V. parahaemolyticus and V. vulnificus per g Mytilus edulis digestive glands (MDD) in husbandry unit at 15 ºC for seven days.

After 24 h of incubation the loads of V. cholerae and V. vulnificus were at 106 CFU/g, V. parahaemolyticus loads at 107 CFU/g. Vibrio cell counts were reduced by 4 - 5 log at 15 ºC to approx. 102 – 103 CFU/g after 7 days (Fig. 4). To mimic a natural water exchange, 50 % of the water was exchanged with fresh artificial sea water after 48, 96 and 144 h. However, at the end of the clearance period Vibrio was still detectable.

To prove that spiking strains and strains isolated from mussels after contamination are identical, fAFLP was performed with selected isolates. fAFLP results before and after storage showed no differences among V. parahaemolyticus clones (Fig. 6A) as well as among V. cholerae clones (Fig. 6B).

Fig. 5: Cell count of V. parahaemolyticus (A), V. cholerae (B) and V. vulnificus (C) per g Mytilus edulis digestive glands during storage for ten days.

A

Institute of Food Hygiene • FU Berlin • Königsweg 69 • 14163 Berlin • Tel. 0 30 - 838 - 62538 • Fax 0 30 - 838 - 62552 • stephan.huehn@fu-berlin.de • www.vetmed.fu-berlin.de

Fig. 3: Cell count of V. parahaemolyticus per g Mytilus edulis during contamination at 15 ºC for 24 hours.

B

Starting loads of V. cholerae and V. parahaemolyticus were approx. 105 CFU/g at the beginning of storage (Fig. 5). The concentration was constant at around 106 CFU/g with a variance of 2 log during storage for ten days. The V. vulnificus load started at 104 CFU/g after 24 h of contamination and was reduced gradually from the first day of storage to approx. 102 – 104 CFU/g . Among the different storage techniques (vacuum packaging and gunnysacks at 3 ºC or 15 ºC) no differences in the concentrations of Vibrio spp. were observed. At the 4th day mussels in gunnysacks at 15 ºC started dying.

Bioaccumulation assay

Bacterial strains and mussels The spiking strains V. cholerae non-O1/non-O139 (ctx-) and V. parahaemolyticus (tdh-/trh-) were isolated from wild Mytilus edulis while the V. vulnificus strain was a human isolate, Biotyp 1. For contamination Vibrio strains were cultured in alkaline peptone water. Mytilus edulis mussels used for the experiments originated from the Baltic sea (Kiel Fjord) and were held in artificial seawater at all times. For all assays the mussels were incubated with Vibrio spp. in the contamination unit of the experimental unit (Fig. 1).

Vibrio spp. accumulated in Mytilus edulis

shown by the example of V. parahaemolyticus (Fig. 3). Highest enrichment was found after 1.5 h in digestive glands (MDD). Therefore, in the following figures only Vibrio levels in digestive glands are shown. Assays with V. cholerae and V. vulnificus showed similar results (data not shown).

Fig. 1: Experimental unit

Sample preparation To quantify Vibrio spp. the mussels were prepared and subdivided into different sample segments: digestive gland (MDD), lumen water of mussel (IVW), mussel remnants (MR), water of the contamination unit (CV) and husbandry unit (HU) (Fig. 2). Sample processing to quantify Vibrio was performed according to ISO/TS 21872.

Fig. 2: Mytilus edulis segments

Musculus retractor bysso-pedalis

MR

MDD

Partitioning of samples: MDD digestive gland IVW lumen water of mussel MR mussel remnants HU husbandry unit CV contamination unit

MLW

fAFLP-Analysis The fluorescent Amplified Fragment Length Polymorphism (fAFLP) analysis was performed according to Duim et al. (1999) except using Taq1 instead of Hha1. This genotyping method was applied to analyse whether the isolated strains at the end of the assay correspond to the spiking strains.

Fig. 6: fAFLP of V. parahaemolyticus (A) and V. cholerae (B)

A B

Isolates of spiking strain

Isolates of the end of assay

C

1,00E+00

1,00E+02

1,00E+04

1,00E+06

1,00E+08

1,00E+10

0 1.5 3 4.5 6 12 24 48 72 96 120 144 168

CFU

/g

Time post inoculation [h]

MDD V. cholerae

Contamination / Bioaccumulation Overnight cultures of bacterial strains were pooled to 1012 CFU for V. parahaemolyticus and V. cholerae and 1010 CFU for V. vulnificus which corresponds to a final concentration in the contamination unit of 5 x 107 CFU/ml and 5 x 105 CFU/ml, respectively. For each experiment approx. 40 mussels were contaminated with Vibrio spp. and tested for 24 hours.

Genotyping assay

References: Duim et al., 1999, High-resolution genotyping of Campylobacter strains isolated from poultry and humans with Amplified Fragment Length Polymorphism fingerprinting DePaola et al., 2000, Environmental investigations of Vibrio parahaemolyticus in oysters after outbreaks in Washington, Texas, and New York (1997 and 1998) Riisgard, H. U., 1988, Efficiency of particle retention and filtration-rate in 6 species of northeast american bivalves Morrison et al., 2011, Survival of Salmonella Newport in oysters

CFU

/g

1,00E+00

1,00E+02

1,00E+04

1,00E+06

1,00E+08

1,00E+10

0 24 25.5 27 28.5 30 36 48 72 96 120 144 168 192

CFU

/g

Time post inoculation [h]

MDD V. cholerae MDD V. parahaemolyticus MDD V. vulnificus

spiking spiking spiking spiking

packaging packaging packaging transfer Water change

spiking dose: 5 x 107 CFU/ml

spiking dose: 5 x 105 CFU/ml

spiking dose: 5 x 106 CFU/ml

Integrated evaluation of environmental parameters influencing Vibrio occurrence in the coastal Northern Adriatic Sea (Italy)

Greta Caburlotto, Franco Bianchi, Giorgio Socal, Ilaria Trento* and Maria M. Lleo

Aim of the study Although temperature has been identified as a key factor, there is no general agreement on which parameters are truly determinant in influencing the bacterial concentration and persistence in the marine environment; different oceanic parameters could have determinant roles in the specific geographic areas. Most of the studies regarding the influence of environmental parameters on Vibrio occurrence report the effect of individual oceanographic variables not always taking into account the inter-relations existing among these parameters: an integrated approach was used to identify key parameters determining the abundance of Vibrio spp in marine samples from the Venetian Lagoon in Italy, which is an important area for fish farming and tourism. Permanent monitoring of environmental parameters is mandatory to predict and prevent climate-associated diseases and threats

Materials and Methods Samples: water, plankton and sediment. Sampling frequency: twice a month during the warm season (June to September) and once each month from October to May. Monitoring period: from June 2006 to November 2007 for six stations while for Foce, from May 2007 to November 2007. Data on temperature ( C), salinity (PSU), dissolved oxygen (% of saturation), pH, and turbidity (NTU) were acquired with a multiparametric probe. Vibrio spp concentration: aliquots of marine samples were inoculated in modified alkaline peptonate water (pH 8.5, added with 3.5% NaCl, m-APW) and incubated at 37 C for 7 h; growth was assessed by visual inspection and the number of Vibrio cells/ml calculated by the MPN method. Vibrio ID confirmation by culture and PCR: a loop from each of the tubes with visible growth was also inoculated on the selective medium TCBS to confirm Vibrio colonies. DNA was prepared from cell pellets using a standard protocol and PCR assays were carried out using primer pair designed on the 16S rDNA Vibrio sequence. Statistical analysis: A non-parametric multivariate multiple regression, done using the computer programs DISTLM® and STATISTICA®, was set up, with a previous log10 transformation of biological data because of their non-normal distribution. Individual variables were analyzed separately for their relation with the Vibrio concentration.

Conclusions Multivariate techniques have been adopted to analyze the dataset: using PCA, it was shown that a relatively high proportion of the total variance in this area was mainly due to two independent variables, namely salinity and temperature. by cluster analysis it was possible to categorize different groups with homogeneous features as regards space (“stations”) and time (“seasons”) distribution, as well as to quantify the values of environmental variables and the Vibrio abundances in each category. Integrating key environmental factors and bacterial concentration values, we identified levels of salinity and sea surface temperature which were optimal for Vibrio concentration in water, plankton, and sediment samples.

The identification of key environmental variables conditioning Vibrio occurrence should facilitate ocean monitoring making it possible to predict unexpected variations in marine microflora which could determine public health risks in coastal areas.

avg std avg std avg std

Temperature °C 21.2 3.5 19.7 5.8 20.4 5.9

Salinity PSU 23.7 9.0 32.2 3.3 35.3 1.5

Dissolved Oxygen % 98 11 107 16 101 11

pH 8.15 0.12 8.20 0.07 8.19 0.06

Turbidity NTU 10.4 8.1 3.6 2.7 2.0 1.2

CPE μg/dm³ 3.6 2.4 3.0 2.1 2.1 1.5

Vibrio in water cells/cm³ 4170 8744 8648 19882 3769 16049

Vibrio in plankton cells/cm³ 419276 1123834 2455162 9938648 551750 1357550

Vibrio in sediment cells/cm³ 9565 26339 8011 27245 2430 5496

GROUP 1 (n = 8)

GROUP 2 (n = 84)

GROUP 3 (n = 46)Cluster among stations

avg std avg std avg std

Temperature °C 10.1 0.8 25.1 2.1 17.9 4.3

Salinity PSU 33.6 2.5 32.1 4.8 33.1 4.2

Dissolved Oxygen % 96 5 108 14 103 15

pH 8.24 0.02 8.19 0.08 8.19 0.07

Turbidity NTU 1.9 1.1 3.7 3.9 3.6 3.2

CPE μg/dm³ 1.1 0.4 3.1 2.2 2.7 1.8

Vibrio in water cells/cm³ 250 272 14657 26552 1926 6253

Vibrio in plankton cells/cm³ 101757 215722 3931604 12259455 312042 941213

Vibrio in sediment cells/cm³ 224 333 2229 4657 10705 30267

Cluster among seasonsGROUP A

(n = 12)GROUP B

(n = 54)GROUP C

(n = 72)

GROUP A: winter GROUP B: summer GROUP C: spring + autumn

GROUP 1: Foce GROUP 2: Caleri1, Caleri2, Fosson1, and Fosson2 GROUP 3: S.Pietro1 and S.Pietro2

Cluster Analysis, a method for grouping objects of similar kind into respective categories, was applied to the hydrological matrix in order to verify the way in which data grouped into two categories, respectively “homogeneous stations” and “homogeneous seasons”

Principal component analysis PCA analysis is a useful tool to quantify the contribution of each variable to describe the environmental variability of a region and to visualize them in the space of two or more dimensions, called factors. The results of this analysis, applied to our environmental dataset (SST, SSS, pH, DO%, turbidity and total pigments), showed that a high fraction of the total variance (> 54%) was explained by the first two factors: salinity and temperature are the most significant parameters, with a respective contribution of 19 and 15% to the variability of the whole system. Other variables contribute as well, but their influence acts mainly in a indirect way, i.e. by means of salinity or temperature. As shown in Figure 1, salinities around 30–35 PSU and maximal temperatures (27–29 C) seem to be the optimal range for vibrios concentration in plankton (Figure 1B), while the same salinity range but associated with lower temperatures (18–24 C) favored the presence of Vibrio in sediment (Figure 1C). Free vibrios in water are favored at lower salinity ranges (20–28 PSU) both at intermediate (17–23 C) and high temperatures (27–29 C) (Figure 1A).

Cluster analysis In the present study, we have considered globally a number of parameters to accurately characterize sampling sites in the study area and, in this context, be able to analyze the presence and abundance of the Vibrio population.

The study area The area extends from the Venetian Lagoon to the estuary of the Po river. This stretch of coast was chosen because of the presence of riverine inputs rich in pollutants, industrial wastes, and nutrients and because it is subject to anthropogenic disturbance. The area is an important recreational zone and is dedicated to fish farming. Three stations were chosen in the area: • San Pietro, a control site with “low-risk” environmental conditions • Fosson and Caleri, situated in areas considered to be at “high risk” on the basis of conditions that favour Vibrio persistence. For each station, two sites were chosen: one 500 m from the coast and the other 3000 m offshore. A seventh site (Foce, waters of low salinity) was included in the study since May 2007.

FIGURE 1: Combined effect of temperature and salinity on Vibrio abundance in water, plankton, and sediment

Federal Institute for Risk Assessment • Dept. Biological Safety • Max-Dohrn-Str. 8-10 • D-10589 Berlin • Tel. + 49 (0) 30 - 184 12 - 0 • Fax + 49 (0) 30 - 184 12 - 47 41 • www.bfr.bund.de

Phenotypic and genotypic

characterisation of environmental and

clinical Vibrio vulnificus isolates from

the Baltic sea

Background

The number of infections caused by pathogenic Vibrio spp. has increased in recent years.

Vibrio spp. are ubiquitous Gram-negative bacteria found naturally in marine and estuarine

waters, including aquaculture settings. They are a leading cause of seafood-borne bacterial

illness and can cause serious wound infections. Increasing incidence of vibrioses in marine

animals and humans has been linked to rising seawater temperature due to global warming.

V. vulnificus can be taken up by ingestion of contaminated seafood such as raw

oysters/mussels or through an open wound exposed to contaminated seawater. Both routes of

infections can lead to serious diseases, sometimes with lethal outcome. In the Baltic Sea

region a number of serious cases of wound infections have been reported, especially in years

with hot summers and periods of elevated sea water temperatures. Although V. vulnificus can

be frequently isolated from the sea water, the number of human infections is comparatvely low.

The aim of this study is to characterize V. vulnificus strains from the Baltic Sea region for

pathogenicity factors and to discriminate between potential pathogenic strains and

environmental strains without clinical relevance.

Nadja Bier1, Silke Bechlars1, Susanne Diescher1, Florian Klein1, Gerhard Hauk2, Oliver Duty2, Eckhard Strauch1, Ralf Dieckmann1

Determination of 16S rRNA typeBased on polymorphisms in the 16S rRNA-gene Vibrio vulnificus isolates can be assigned

to two types A & B (Aznar et al., 1994), which are described to correlate with the strain

origin (Nilsson et al., 2003). To assess the 16S rRNA type Real Time PCR assays were

performed according to Vickery et al., 2007. In regard to our Baltic sea isolates, we couldn`t

observe a clear correlation. The majority of isolates belonged to type A regardless of their

origin (Fig. 3).

Conclusions

• Phenotypic examination along with in-vitro and in-vivo assays are important to assess the pathogenic potential of Vibrio strains. Nearly all strains show

cytotoxicity to Caco-2 cells. Resistance to human serum was observed in most clinical Baltic sea strains, but was also found in some environmental strains.

• So far genotypic techniques do not suggest a clear discrimination between clinical strains and environmental strains as revealed by the analyses of the 16S rRNAtype, vcg type and pathogenicity region XII.

• Most environmental strains may be potentially dangerous to public health and

further studies for hazard identification and risk assessment are needed.

References

Aznar et al., 1994. Int. J. Syst. Bacteriol. 44: 330-337.Cohen et al., 2007. Appl. Environ. Microbiol. 73:5553-65.Nilsson et al., 2003. J. Clin. Microbiol. 41:442-46. Roig et al., 2010. Appl. Environ. Microbiol. 76:1328-33.

Rosche et al., 2005. Microbiol. Immunol. 49:381-89.

Vickery et al., 2007. J. Microbiol. Methods. 68: 376-384.

Acknowledgement: VibrioNet is funded by the Federal Ministry of Education and Research (Grant 01KI1015A)

Cytotoxicity assaysThe cytotoxicity to Caco-2 cells induced by supernatants of V. vulnificus liquid cultures was

measured using a CytoTox96 Non-Radioactive cytotoxicity assay kit (Promega, Madison,

WI). Except for VN-206, supernatants of all strains showed cytotoxic activity. Most strains

induce complete lysis of the host cells in a short time. However, few strains that were

negative for cell lysis in this assay showed cytopathic effects by inducing rounding of host

cells (Figure 1).

Phenotypic and genotypic assays associated to virulenceTo assess the virulence potential of Vibrio vulnificus isolates, phenotypic ad genotypic

assays were applied. Phenotypic assays may reveal valuable information on known and

putative virulence–associated traits such as expression of extracellular proteins (hemolysins,

proteases, chitinases), and on interaction with eukaryotic cells (serum resistance properties,

cytotoxicity). Genotypic methods combine high discriminatory power and high reproducibility.

These included vcgC/E typing (Rosche et al., 2005), 16S rRNA A/B typing (Nilsson et al.,

2003) and presence/absence of 33-kb genomic region XII (Cohen et al., 2007) by

established PCR assays.

Fig. 1 Change in morphology of Caco-2 cells after treatment with supernatants of V. vulnificus overnight cultures.

A: complete lysis (+ LDH), B: cell rounding (- LDH), C: control (- LDH).

A B C

1) Federal Institute for Risk Assessment, Berlin, 2) Landesamt für Gesundheit und Soziales Mecklenburg-Vorpommern, Rostock

Fig. 5 A) Pathogenicity Region XII of V. vulnificus (Cohen et al., 2007). B) Distribution of pathogenicity

region XII among clinical and environmental strains.

Clinical Environmental

Pathogenicity Region XII of V. vulnificus:

4 PCRs described by Cohen et al., 2007 were performed to investigate the presence of a

33 kb genomic island which could be associated with virulence (Fig. 5a). We could prove

the presence of this region XII in the majority of clinical isolates (Fig. 5b) which supports its

relevance for pathogenicity.

Nilsson et al., 2003 :

Type A ���� environmental origin

Type B ���� clinical origin

Clinical Environmental

Fig. 3 : Distribution of different 16S rRNA-types among clinical and environmental strains

type A

type B

type AB

Rosche et al., 2005 :

Type E ���� environmental origin

Type C ���� clinical origin

EnvironmentalClinical

Fig. 4: Distribution of different vcg types among clinical and environmental strains

vcg type C

vcg type E

Determination of vcg typeAnother method to predict virulence of V. vulnificus isolates takes advantage of a

polymorphism in the so-called virulence correlated gene (vcg) which discriminates two

types (C=clinical type, E=environmental type). To assess the vcg type we performed PCRs

as described by Rosche et al., 2005. Only 19 % of the clinical strains possessed the vcg

type C, which should correlate with clinical origin (Fig. 4).

Serum resistanceSurvival and growth in blood are characteristic traits of pathogenic V. vulnificus.

Examination of serum resistance as an important virulence factor could therefore serve to

predict the pathogenic potential of V. vulnificus isolates. 84 % of the clinical isolates

showed growth in the presence of 60-80% human serum and could be classified as serum

resistant (Fig. 2).

Fig. 2 Flow chart for serum resistance assay. Bacteria are grown in 96 well plates for 5 hours and transferred to new plates containing media with different concentrations of human

serum (0 – 80% serum). Growth is visible by colour change of bromothymolblue from blue to yellow.

24h, 37°Csensitive

resistant

control

80% 60% 40% 20% 10% 0% serum

80% 60% 40% 20% 10% 0% serum

A

B

Analyzing Vibrio populations by MALDI-TOF

Compared to DNA-based techniques, Matrix-Assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) used as a mass spectrometric profiling method provides rapid and cost-effective identification results. Creating a MALDI-TOF Vibrio reference database leads to a clear discrimination between vibrios. Hence, whole environmental Vibrio populations can easily be screened for potentially human pathogenic species.

MALDI-TOF Principle Colonies of Overnight-Cultures are getting picked. Transferring to target plates occurs either directly or by an ethanol / formic acid extraction. Covering with Matrix (cinnamic acid) yields to crystal structures. Therein embedded proteins getting desorpted by Laser shots. Simultaneously these molecules become ionized due to proton transfers. After electric field acceleration, protein ions passing a vaccum flight tube. The flight duration differs depending on masses and charges of these molecules (m/z). Thus, there is a time-dependent detection of ion impact intensities. The outcome are spectra with the m/z ratio (molecular masses) on the abscissa and measured intensities on the ordinate (Fig. 1). Fig. 1: Sample preparation approaches and measurement principle of MALDI-TOF.

Alfred Wegener Institute for Polar and Marine Research, Shelf Sea System Ecology, Helgoland René Erler

Vibrio Population Analyses

Species Identification

MALDI-TOF can be a powerful tool in Vibrio population analyses. The approach is currently tested on monitoring cruises to observe the relative proportion of Vibrio species in the German Bight. In addition to this quantitative application of MALDI-TOF, Cluster Analyses can be performed to detect interspecific, maybe pathogenic or sampling site associated groups. Our goal is to establish this mass spectrometric method in Vibrio research. This offers the opportunity for Vibrio population analyses on a larger scale.

In a pilot study, mass spectra of 61 Vibrio isolates from different sampling sites were acquired. They were all included in one correlation matrix (Fig.2). Matching scores over 0.66 are coloured red, partly matches (0,33 to 0,66) are yellow and matching scores lower than 0,33 are green coloured. The average correlation score between vibrios of the same species varies from 0,69 to 0,89 (CV: 0,14 to 0,21). Clustering results in a clear separation of single species (Fig.3). Mass spectra generated from one isolate correlate with a score of 0,97 and a standard deviation of 2 %.

Num

ber o

f iso

late

s

Vib

rio a

lgin

oly

ticus

Vib

rio c

hole

rae

Vib

rio p

ara

ha

em

oly

ticu

s

Vib

rio v

uln

ific

us

Vibrio alginolyticus 15 0,69 0,19 0,57 0,37

Vibrio cholerae 11 0,19 0,80 0,21 0,23

Vibrio parahaemolyticus 15 0,57 0,21 0,85 0,45

Vibrio vulnificus 15 0,37 0,23 0,45 0,89

Vibrio diazotrophicus 1 0,33 0,23 0,32 0,29

Vibrio fluvialis 1 0,32 0,20 0,23 0,23

Vibrio harveyi 1 0,55 0,23 0,44 0,34

Listonella anguillarum 1 0,22 0,16 0,21 0,20

Shewanella putrefaciens 1 0,11 0,08 0,12 0,07 Fig. 3: Cluster Analysis Dendogram according to data from Fig.2. Fig. 3: Average correlation scores of Vibrio isolates .

Summary & Outlook

V. parahaemolyticus

V. alginolyticus

V. vulnificus

V. navarrensis

V. diazotrophicus

Two dimensional principal component analysis

In August 2011, vibrios of the North and Baltic Sea were isolated during a cruise with the research vessel Heincke. Comparison of single spectra shows distinct species-specific Cluster, even in the case of closely related species like V. parahaemolyticus / V.

alginolyticus and V. vulnificus / V. navarrensis

(Fig.4). An even clearer differentiation of species cluster can be reached by multi-dimensional scaling of correlation data (Fig.5).

Fig. 4: Two dimensional component analysis of single spectra, acquired from environmental isolates.

V. parahaemolyticus

V. alginolyticus

V. diazotrophicus

V. vulnificus

V. navarrensis

Fig. 5: MDS Plot of selected Heincke isolates.

Bacterial-Strains In total 78 V. parahaemolyticus-isolates were analysed. Thereof, 30 strains derived from seawater, bivalves and shrimp of different geographic origin. Additionally, 46 strains were isolated from shrimp of farms located in three Sri Lankan regions (Fig. 1). One isolate per farm and per shrimp-pond was applied to MLST analysis, respectively. The Japanese clinical strains ATCC17802 and RIMD2210633CM29 served as reference strains. Multilocus Sequence Typing (MLST) In MLST analysis the nucleotide sequences of multiple reference genes are compared. Therefore, internal fragments of the genes dnaE, gyrB, recA, dtdS, pntA, pyrC and tnaA were amplified via PCR and sequenced (Fig. 2A; González-Escalona et al. 2008). The received sequence data were analysed with Bionumerics (Applied Maths, Version 6.01) and compared to already published sequences on the PubMLST web page (http://pubmlst.org/vparahaemolyticus; Jolley et al. 2004). Based upon the concatenated nucleotide sequences a dendrogram was created via UPGMA-Analysis. The resulting similarity matrix served as basis for the generation of Minimum Spanning Trees.

Literature: Chao 2011, Origin of Vibrio parahaemolyticus O3:K6 pandemic clone DePaola et al. 2000, Environmental investigations of Vibrio parahaemolyticus in Oysters after outbreakes in Washington, Texas and New York (1997 and 1998) Feil 2004, eBURST: inferring patterns of evolutionary descent among clusters of related bacterial genotypes from González-Escalona et al. 2008, Determination of Molecular Phylogenetics of Vibrio parahaemolyticus Strains by Multilocus Sequence Typing Jolley et al. 2004, mlstdbNet – distributed multi-locus sequence typing (MLST) databases; Vibrio parahaemolyticus MLST web page (http://pubmlst.org/ vparahaemolyticus) developed by Keith Jolley and sited at the University of Oxford

MULTILOCUS SEQUENCE TYPING FOR GENOTYPING OF VIBRIO PARAHAEMOLYTICUS

Background Vibrio (V.) parahaemolyticus, a naturally present potentially pathogenic bacteria, is isolated frequently from seawater, sediment and raw or insufficiently cooked seafood (e.g. shellfish and bivalves). Owing to the high filtrating activities of bivalves the level of V. parahaemolyticus was found to be 200-fold higher than in the surrounding seawater (DePaola et al. 2000). Consumption of or contact to raw or undercooked seafood, that contains Vibrio in appropriate numbers, can lead to infections (gastroenteritis, wound infections and septicaemia). Isolates with (A) different geographic origin were analysed via MLST concerning their genetic relatedness and possible distribution patterns. One focus was the genomic analysis of (B) isolates gained from shrimp samples of three Sri Lankan regions.

Material and Methods

Institute of Food Hygiene • FU Berlin • Königsweg 69 • 14163 Berlin • Germany • Tel. 030 - 838 - 62538 • Fax 030 - 838 - 62552 • sara.urmersbach@fu-berlin.de • www.vetmed.fu-berlin.de

Fig. 1: Map of Sri Lankan Sampling Areas (1 Chillaw, 3 Madurankuliya, 5 Puttalam)

Sara Urmersbach1, Madura Sanjeevani Gonsal Koralage2, Thomas Alter1, Ute Messelhäusser3, Duangporn Pichpol4, Stephan Huehn1 1Institute of Food Hygiene, Free University Berlin, Germany 2Government Veterinary Office, Walikanda, Polonnaruwa, Sri Lanka 3Bavarian Health and Food Safety Authority, Oberschleißheim, Germany 4Faculty of Veterinary Medicine, Chiang Mai University, Thailand

Often strains are not typable owing to missing sequences of recA and gyrB fragments (Chao 2011). With the application of newly designed primer gyrB and recA fragments of former non-typable isolates could be analysed via MLST. The distribution of sequence types is independent from the geographic origin of analysed isolates. Strains from Asia as well as from Europe and South-America were genetically diverse and no dominating ST was identified. Owing to the high fraction of Asian strains (88%) the distribution of European and South-American strains will be restudied after analysis of further non-Asian isolates. Sri Lankan strains from shrimp farms of three nearby localized regions possessed a high genetic

diversity and no local dominating ST was found. However, strains with the same ST were isolated in different regions. Owing to the lack of more published data of Sri Lankan isolates and the naturally high genetic diversity of environmental isolates, most of the strains isolated possessed new STs (94%). Close relatedness of environmental and clinical isolates was shown in Minimum Spanning Trees and also discovered by eBURST-analysis, thus emphasising the high pathogenic potential of environmental and food isolates of V. parahaemolyticus.

Discussion

Most of the analysed isolates originated from Asia (88%), especially West-Asia. Sequencing of 78 isolates led to identification of 65 new alleles: 7 for dnaE, 15 for gyrB, 9 for recA, 9 for dtdS, 6 for pntA, 10 for pyrC und 9 for tnaA. A complete allelic profile and thus an assignment of a ST was revealed for 69 isolates. Most of them (n= 52) belonged to new STs, consisting of either recombination of already published or combinations with new STs. The Minimum Spanning Tree reveals no clear clustering based on the geographic origin of the isolates (Fig. 3A). Strains from Sri Lanka formed two clusters, which are connected by strains of diverse origin (Japanese reference strains and food isolates).

Isolates from food samples could be assigned to 21 STs (8 already published, 13 new) and were distributed over the whole Minimum Spanning Tree. Isolates from South-America where located in both clusters, the two STs of European strains were located in one cluster but not nearby. The majority of West-Asian isolates (n= 43) originated from shrimp-samples out of three Sri Lankan regions (Fig. 3B). They belonged to 15 STs. Ten of these derived from isolates out of one region, two appeared in two, and three in all regions. Isolates with the same ST and region originated mostly from different farms but at least from different ponds.

Results

B – Distribution of MLST-Sequence Types within three regions of Sri Lanka

21%

35%

44%

Sri Lanka 5 Sri Lanka 3 Sri Lanka 1

Fig. 3B: Minimum Spanning Tree and pie chart of isolates from Sri Lanka; Colouring according to sampling area.

recA and gyrB Dividing the original fragments of gyrB and recA (Fig. 2B) into two fragments led to complete allelic profiles. Therefor two inner primer were designed, allowing PCR amplification (Fig. 2C) and sequencing with the conditions published by González-Escalona (2008). eBURST-Analysis The eBURST algorithm identifies groups of related genotypes on the basis of their allelic profiles, called sequence type (ST) and predicts the descent from the presumably founder to other genotypes in this group (http://eburst.mlst.net/).

www.weltkarte.com

C – eBURST-Analysis eBURST-Analysis of all published V. parahaemolyticus STs led to the detection of a new clonal complex (CC; Fig. 4). The founder, ST412 (trh+tdh-), was isolated from an Ecuadoran shrimp sample. All other strains of the CC were isolated in the USA. Single Locus Variants (SLV) were from environmental samples: ST313 (trh+tdh+), ST314 (trh+tdh+) and ST315 (trh-

tdh+).The Double Locus Variant (DLV) ST43 (trh+tdh+) was found in environmental and clinical samples. The Satellite ST44 (trh+tdh+) was only found in clinical samples.

Fig. 4: Graphical output of eBURST-analysis (http://pubmlst.org/)

A – Influence of geographic origin on the MLST-Sequence Types

Sri Lanka Bangladesh Ecuador Thailand Vietnam Pacific Ocean Baltic Sea Japan China Honduras India

66%

12% 10%

7%

5%

88%

Fig. 3A: Minimum Spanning Tree depending on sequence differences of STs. Pie chart of geographic origin. Colouring according to the region of origin.

West-Asia

South-America

Central-Asia

Europe

East-Asia

Asia

Fig. 2: Gelelectrophoretic illustration of amplified fragments before (A) and after (C) usage of additional primers (B).

A

B

C

Acknowledgement: The project was founded by BMBF (VibrioNet-Project).