influence of sampling depth and post-sampling …

INFLUENCE OF SAMPLING DEPTH AND POST-SAMPLING ANALYSISTIME ON THE OCCURRENCE OF COLIFORMS AND VIBRIO IN

WATER AND SHELLFISH

RE Sallema1 and GYS Mtui2

1National Environment Management Council, Vice President’s Office,P. O .Box 63154, Dar es Salaam, Tanzania

2Applied Microbiology Unit, Department of Botany, University of Dar es Salaam,P.O. Box 35060, Dar es Salaam, Tanzania

ABSTRACTThe bacteriological quality was examined at the water surface, 3 m depth and in the shellfishflesh, and the results were compared to other potential pathogenic indicator organisms. Thestudy was conducted at Long Harbour (mussel farm), St. John’s and Outer Cove sites ofNewfoundland, Canada. Bacteriological analysis was carried out for samples taken at waterdepth and at 1, 6, 12 and 24 hours post-sampling. It was observed that the total and faecalcoliform bacteria were significantly higher in the 3 m water depth samples than in the surfacewater samples (ANOVA, F = 59.41, 26.751, 9.82 (T.C); 46.41, 26.81, 10.72 (F.C); P <0.001). Inaddition, shellfish tissues had substantial amounts of coliform bacteria levels, which variedsignificantly with station and date of sampling (F = 128.21,37.42 (T.C); 1281, 37.42 (F.C); P<0.05). The higher levels reflect bioaccumulation. There were no correlations between estimatesof total or faecal coliforms with potential pathogenic Vibrio groups. Furthermore, there were nosignificant differences in total and faecal coliforms among the post-sampling time intervals. Theresults suggest that sampling and bacteriological analysis of water and shellfish for qualitycontrol should consider both the water surface and depths proximal to the shellfish. Moreover,adoption of extended post-sampling time may lead to a more convenient and less costly approachto monitoring of bacteriological impact on the coastal marine environments.

INTRODUCTIONBivalve molluscs are commerciallyharvested marine organisms and are amongthe few shellfish consumed raw in manycountries. As the world populationincreases, food security becomes a majorconcern. Although shellfish farming hasattracted a great deal of attention inaquaculture as a source protein-rich food, itis being developed in shallow, shelteredcoastal areas which are also used asreceptacles for domestic sewage discharge,industrial effluents and freshwater surfacerunoff. This being the case, sedentary filter-feeders such as shellfish can potentiallyaccumulate chemicals, bacteriologicalpollutants and naturally occurring toxins tolevels far in excess of the surrounding water(Burkhardt and Calci 2000; Cavallo andStabili 2002).

Total and faecal coliform bacteria groups areindicator organisms currently used to assesswater quality in shellfish growing andharvesting areas. However, these organismsdo not provide indications of potentialhuman pathogens and are generally notpathogens themselves. Consequently,disease outbreaks associated with shellfishharvested from waters meeting faecalcoliform standards have been increasing inrecent years (Levitan 1991, Hackney andPierson 1994). In several cases, variousVibrio bacteria are the causes of the illnessbut they go undetected in standardprotocols, since they are not easily culturedby these methods. The use of thiosulphatetitrate bile salt (TCBS) agar is a relativelymore reliable culture method as it allowsmost Vibrio cells to grow (Kaysner et al.1994). In this study, bacterial counts(coliform levels) from standard analytical

Sallema & Mtui. – Influence of sampling depth and post-sampling analysis time …

56

methods were compared with TCBS-positive cultures in order to evaluatepossible relationships.

Bacteriological water sampling withconventional methods are based on samplingwater from the surface layers, normally thefirst 20 cm, while shellfish are suspended atdeeper levels, usually one meter depth orgreater. The surface waters for whichsamples are taken do not necessarily reflectwaters present next to the shellfish (Nix etal. 1993; Beliaeff and Cochard 1995).

Sampling and analysis intervals areimportant aspects for consideration inbacteriological examination of seawaterand/or shellfish samples. According to thecurrent method guidelines, bacteriologicalsample analysis needs to be done strictlywithin six hours after sample collection.This is a major constraint because shellfishsites are located long distances away fromanalytical laboratories.

In this study, indicator organisms togetherwith other potential pathogenic groups wereevaluated from water sampled from both thesurface and 3 m depth; after several analysistime intervals within 24 hours aftercollection, in order to elucidate the effect ofextending the time for post-collectionbacteriological sample analyses for watersurface.

MATERIALS AND METHODSStudy sitesTwo stations along the Long Harbourcommercial shellfish production area andone station in each of the St. John’s andOuter Cove in Newfoundland, Canada, werestudied (Fig. 1). Long Harbour musselfarms are situated near the Head of PlacentiaBay. Samples were collected from thesefarms, one classified as “Closed” or“Polluted” (Station 1) and another as“Open” or “Unpolluted” (Station 2). St.John’s Harbour site (polluted) is located onthe East end of Avalon Peninsula. The site

is polluted as the whole sewage (untreated)from St. Johns city Mount Pearl andParadise is released directly to the ocean atthis point. The Outer Cove site lies NorthEast of St. John’s and is unpolluted. Thethree sites have several freshwater streamsand are characterized by sheltered bays andtherefore protected against strong current andwaves.

Sampling procedureTriplicate samples of water and shellfishwere taken at three arbitrary points at eachsite/station. Water was sampled on thesurface and at 3 m depths while musselsampling was taken at 3 m depth. Surfacewater samples were collected using a sterilepolypropylene bottles attached to apolyvinyl-chloride pipe at a 20 cm mark.Sampling at 3 m depth was done by asterilised tubing attached to polyvinylchloride pipes. Each sample wasimmediately emptied into sterilised ice-cooled (0ºC - 0.2ºC) plastic bottle, sealedand labelled in the field. During musselsampling, 10 -12 animals (sufficient to yield150-250 g of soft tissue), were taken fromeach shellfish stock sample, sealed in coolsterile plastic bags and kept in plasticcontainers. Salinity, temperature and weatherconditions were recorded at the time ofsampling. Sources of freshwater input to thesites, such as streams and rivers, were alsorecorded.

Culture mediaSingle and double strength (in g sample/Lof distilled water) media for Lauryl SulfateTryptose Broth (LTB) were 35.6 g and 71.2g sample/L; Brilliant Green Lactose 2% BileBroth (BGLB) were 40 g and 80 gsample/L; and Escherichia coli Broth (EC)were 37 g and 74 g sample/L, respectively.Bacto Alkaline Peptone Water (APW) mediawas 1 g/L, Bacto Tripticase Citric BileSulfate Agar (TCBS) was 89 g/L andTripticase Soy Agar (TSA) was 30 gsample/L. Culture media were sterilised inan autoclave at 121ºC for 15 minutes.

Tanz. J. Sci. Vol 31(1) 2005

57

Figure 1: Map of Newfoundland showing two stations along the Long Harbour commercialshellfish production area in St. John’s and Outer Cove sites.

Bacteriological analysisAbout 200 g of mussel tissue was weighed,mixed with 200 ml of 1% APW, thenblended for 90 seconds. Decimal dilutionsof the tissue and water samples wereintroduced into replicate tubes of a mediumdesigned to select for growth of totalcoliform and faecal coliform bacteria. Fivetubes were used for each of the three decimaldilutions (10mL, 1mL and 0.1mL). Alltubes were incubated in a water bath at 48ºCfor 48 h and any gas produced was noted as

a positive test for coliforms. The numbersof positive tubes were correlated to thenumber of bacteria using the Most ProbableNumber (MPN) tables based on probabilityformulae (APHA 1995). TCBS was used asa selective plating medium for Vibrioculture following the method by Kaysner etal. (1994). For total plate counts (TPC),plates of TCBS media were inoculated andincubated overnight at 35ºC. Colonyforming units were counted per plate( C F U ) / 1 0 0 m L a n d recorded.


58

Statistical analysisMeans and Standard Error of Means (SEM)were calculated for counts of coliforms andVibrios.

To determine the treatment effects betweenbacteria (coliforms and vibrios) and day,depth, analysis time and station ofsampling, statistical analysis of data wasperformed using 2-way and 3-way Analysis

of Variance (ANOVA) for all variables(<0.05). Tukey Post hoc comparisontest was carried out following significantANOVA results. A correlation analysis (9 x9 matrix) was done to determine ifrelationships existed among the variables.All data were analysed by the Microsoft©

Excel statistical package.

S

3 m

S 3 m S 3 m

0.0

1.0

2.0

3.0

4.0

5.0

Fig. 2a: Faecal Coliform at Station Two

To

tal C

olif

orm

M

PN

S 3 mS

3 mS

3 m

0.0

1.0

2.0

3.0

4.0

5.0

Fig. 2b: Total Coliform at Station TwoTo

tal C

olif

orm

M

PN

S3 m

S

3 m S 3 m

0.0

1.0

2.0

3.0

4.0

5.0

Fig. 2c: Faecal Coliform at Station OneFa

eca

l C

olif

orm

M

PN

Fig. 2d: Faecal Coliform at Station Two

S3 m

S3 m

S

3 m

0.0

1.0

2.0

3.0

4.0

5.0

Fa

eca

l C

olif

orm

M

PN

Dec./04/96 Dec./13/96 Dec./17/96

Figure 2 (a-d): Total and faecal coliform levels in water samples from the surface (S) and 3 mdepth (3m) at Stations 1 and 2, Long Harbour Mussel farm.

Tanz. J. Sci. Vol 31(1) 2005

59

RESULTSTotal and faecal coliforms were significantlyhigher in water samples taken at a depth of 3meters than in surface water samples(ANOVA, F = 59.41, 26.751, 9.82 (T.C);46.41, 26.81, 10.72 (F.C); P < 0.001) (Fig.2 (a – d)). There were significant differences

in coliform levels between Dec. 4, 96 andboth Dec. 13, 96 (post hoc comparison: p =0.001; 0.001) and Dec. 17, 96 (p = 0.007;0.004) but no significant difference betweenthe later two dates (p = 0.684; 0.719).

0.0

100.0

200.0

300.0

400.0

500.0

Fig. 3a: Total Coliform Mussels at Station OneTo

tal C

olif

orm

M

PN

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

Fig. 3b: Total Coliform Mussels at Station Two

To

tal C

olif

orm

MP

N

0.0

100.0

200.0

300.0

400.0

Fig. 3c: Faecal Coliform Mussels at Station OneFa

eca

l C

olif

orm

M

PN

0.0

2.0

4.0

6.0

8.0

10.0

Fig. 3d: Faecal Coliform Mussels at Station Two

Fa

eca

l C

olif

orm

M

PN

Dec./04/96 Dec./13/96 Dec./17/96

Figure 3 (a-d): Total and faecal coliform levels in mussel tissue samples from Stations 1 and 2at Long Harbour mussel farm. The sampling depth was 3 m.


60

Mussel tissues accumulated higher levels offaecal and total coliforms than thesurrounding waters, particularly in Station 1(Fig. 3 (a-d)). The levels also variedsignificantly with stations and dates ofsampling (ANOVA, F = 128.21, 37.42

(T.C); 1281, 37.42 (F.C); P <0.05). Posthoc comparison test shows that there weresignificant differences between Dec. 4, 96and both Dec. 13, 96 (p = 0.000; 0.000) andDec. 17, 96 (p = 0.000; 0.000) but no

significant differences between the later twodates (p = 0.798; 0.980). No significantdifferences were detected in Vibrio coloniesbetween depths, sampling dates or water andtissues (ANOVA, F = 0.721, 1.231, 0.692,0.082; P > 0.05). However, a significantdifference was found between stations(ANOVA, F = 5.51; P < 0.05) - (Fig. 4 (a-d)).

S

3 m

S

3 m S3 m

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

Fig. 4a: Vibrio Water at Station One

cf u

/10

0m

l V

ibrio

Co

lon

ies

S

3 m S 3 m

S 3 m

-1.0

1.0

3.0

5.0

7.0

Fig. 4b: Vibrio Water at Station Two

cf u

/10

0m

l V

ibrio

Co

lon

ies

0.0

5.0

10.0

15.0

Fig. 4c: Vibrio Mussels at Station One

cf u

/10

0m

l V

ibrio

Co

lon

ies

0.0

10.0

20.0

30.0

Fig. 4d: Vibrio Mussels at Station Two

cf u

/10

0m

l V

ibrio

Co

lon

ies

Dec./04/96 Dec./13/96 Dec./17/96

Figure 4 (a-d): Vibrio colony levels in water and in mussel samples from Stations 1 & 2, LongHarbour Mussel farm. (Water sampling at the surface, (S): and 3 m depth, (3m);Mussel sampling at 3m depth).

Tanz. J. Sci. Vol 31(1) 2005

61

The presence of Vibrio was presumptive,based on the TCBS-positive patterns(Kaysner et al., 1994). The Vibrio coloniesvaried from 2 to 4 mm in size while thecolour varied from green, cream, yellow andsometimes translucent. Pearson Correlation

Analysis indicated a significant relationshipamong all variables (water and tissue MPNcoliforms); values ranged from R = 0.92 to0.99 (n = 18, P = 0.01), but not MPN andVibrio data (R = 0.78 to - 0.43, n = 18, P >0.05) (Table 1).

Table 1: Matrix correlation analysis showing possible combinations of Total coliform(T.C.), Faecal coliform (F.C.), Water (W), Mussel meat (M) and Vibrios (V)for water and mussels samples.

FC (W) FC (M) TC (W) TC (M) V (W) V (M)Pearson

correlationF.C. (W)F.C. (M)T.C. (W)T.C. ( M)V. ( W)V. (M)

1.000.933**.987**.924**

.180-.383

.933**1.000.933**.973**

.104-.425

.987**

.933**1.000.924**

.163-.383

.924**

.973**

.924**1.000.070-.415

.180

.104

.163

.0701.000-.037

-.383-.425-.383-.415-.0371.000

N F.C. (W)F.C. (M)T.C. (W)T.C. (M)V. (W)V. (M)

361836183618

181818181818

361836183618

181818181818

361836183618

181818181818

The analyses done at 1, 6, 12 and 24 hoursof post-sampling time showed no significantdifferences in total and faecal coliforms(ANOVA, F = 0.02, 0.023; P > 0.05) at anyof the post-sampling time intervals (Fig. 5(a-d)). There was no significant difference intotal and faecal coliform levels with regardto date of sampling within stations. Posthoc comparison indicates that, there were

significant differences between Jan. 31, 97and both Feb. 7, 97 (p = 0.000; 0.000) andFeb. 14, 97 (p= 0.000; 0.000) but nosignificant differences between the later twodates (p = 0.999; 0.931). However, asignificant difference was noted betweensites (ANOVA, F = 17366261 (T.C); 467811

(F.C.); P < 0.001).


62

241261241261241261

0.0

500.0

1000.0

1500.0

2000.0

Fig. 5a: Variation of Total Coliforms with Analysis Hour - St. John's Harbour

Jan./31/97 Feb./07/97 Feb./14/97

To

tal C

olif

orm

MP

N

24

12

61

2412

6

1

0.0

5.0

10.0

15.0

Fig. 5b: Variation of Total Coliform with Analysis Hour - Outer Cove Site

Feb./07/97 Feb./14/97

To

tal C

olif

orm

M

PN

24126

1

24

12

612412

6

1

1610.0

1620.0

1630.0

1640.0

1650.0

Fig. 5c: Variation of Faecal Coliform with Analysis Hour - St. John's HarbourF

ae

ca

l C

olif

orm

MP

N

Jan./31/97 Feb./07/97 Feb./14/97

24

126

1

2412

6

1

0.0

5.0

10.0

15.0

Fig. 5d: Variation of Faecal Coliform with Analysis Hour - Outer Cove Site

Fa

eca

l C

olif

orm

MP

N

Feb./07/97 Feb./14/97DATE

Fig. 5 (a-d): Variations of Total and Faecal Coliform levels with post-sampling analysis time ofsurface sample from St. John’s and Outer Cove. The number of hours are indicatedabove each bar.

DISCUSSIONThe indication of higher numbers of

coliform bacteria at water depths andvariation of coliform levels with day and

Tanz. J. Sci. Vol 31(1) 2005

63

station of sampling in this study can beinfluenced by the time of sampling. Sincesampling was carried out during Fall andWinter season, the declining effect oftemperatures, coupled with wind, stationlocation and seasonal thermal turnover of thewater, might have been the core factorsinfluencing the higher levels of coliformbacteria at water depths. The suggestion is,the unique weather patterns ofNewfoundland as discussed by Hewson(1993) and Banfield (1993) have to bestudied thoroughly to indicate their effect onthe distribution of coliform organisms in thewater column. On the other hand, bothexperimental sites were located in an areawith runoff events and freshwater streamshence contributing to increased coliformlevels (Ryan, 2003). The release of sewageand other wastes into the ocean might havealso facilitated accumulation of the coliformbacteria in the sediment (unpublished data).Re-suspension of coliforms from thesediment, caused by shear forces of wind,current and tides, might have contributed tothe increased levels of coliforms in the watercolumn (Hewson, 1993 and Banfield, 1993).The occurrence of higher content ofcoliforms in mussel tissues than watercoliforms has also been reported by Levitan(1991); Valiela et al. (1991); and Hackneyand Pierson (1994). The higher levels maybe attributed to the bioaccumulation capacityof shellfish (Bonnadona et al. 1990, Valielaet al. 1991, Burkhardt and Calci 2000).

The lack of significant difference in coliformbacteria within 24 h period suggests thatwith careful sample handling, analysis timecan be taken at longer time intervals up to24 h post-sampling. The possibility ofanalysing samples at a longer post-samplingtimes permits access to laboratory testing,especially for samples from remote areas.Although coliform positive test is no longerstrictly being used as an indicator of waterquality, the total and faecal indicator groupscan be used to provide useful information tobe used in tracking the dispersion and

dilution of certain sewage effluents inshellfish harvesting areas.

The higher levels of Vibrio colony countsnoted in both mussel tissues and in watersamples even for the site with very lowcoliform levels can be explained by the free-living nature of the Vibrio species and theability to multiply in the naturalenvironment – a trait which could impact amajor effect to this study. This thereforeconfirms that coliform organisms cannotalways represent the presence or absence ofpotential pathogens like Vibrios, and similarconclusions have also been made by Levitan(1991) and Gosling (1992).

This study demonstrated that sampling atboth the surface and 3 m depth forbacteriological monitoring of shellfish andwater quality is necessary in order tointegrate both well-mixed and stratifiedwaters. Sample analysis at various timeintervals indicated that up to 24 h post-sampling period is feasible. Adoption ofthis extended post-sampling time may leadto a more convenient and less costlyapproach to monitoring of bacteriologicalimpact on the coastal marine environments,especially in remote shellfish growing andharvesting sites. Future studies shouldfocus on sampling optimisation byincreasing sampling frequency and takinginto account the water stratification of agiven place.

ACKNOWLEDGEMENTWe are grateful to the Canadian InternationalDevelopment Agency (CIDA) for thefinancial support. Dr. C. Couturier of theMarine Institute of Memorial University,Newfoundland, Canada, is acknowledged forguiding this study. We appreciate thetechnical assistance from Mr Keith Rideout,Mr. Keith Mercer and the staff of LongHarbour Mussel Farm. Special thanks aredue to Dr. A. B. Dickinson, Ms. ColleenClarke, Dr. M.A.K. Ngoile and Dr. J.Francis for their constant moral support.Prof. A. K. Kivaisi is cordially thanked for


64

her constructive comments on themanuscript.

REFERENCESAPHA 1995. Standard methods for the

examination of water and wastewater,19th edition, American Public HealthAssociation, Washington DC.

Banfield CE 1993. The climate ofNewfoundland. In: Climate andWeather of Newfoundland andLabrador. (eds.) Robertson A., PorterS., and Brodie G. CreativePublishers, St. John's, p 13 - 31.

Beliaeff B and Cochard M 1995. Applyinggeostatistics to identification ofspat ia l pat terns of faecalcontamination in a mussel farmingarea. Wat. Res. 29(6): 1541-1548.

Bonadonna LL, Volteria L, Auliano FA andMancini L 1990. Accumulationpower of some bivalve molluscs.Mar. Poll. Bull. 21(2): 81 - 84.

Burkhardt W and Calci KR 2000. Selectiveaccumulative may account forshellfish-associated viral illness.Applied and EnvironmentalMicrobiology. 66(4): 1375 –1378

Cavallo RA and Stabili L 2002. Presence ofVibrios in seawater and Mytilusgalloprovincialis (Lam.) from theMar Piccolo of Taranto (Ionian Sea).Water Res. 36(15): 3719-26.

Gosling E (ed.) 1992. The Mussel Mytilus:Ecology, Physiology, Genetics andCulture. Elsevier Amsterdam, 589pp.

Hackney CR and Pierson MD (eds.) 1994.Historical Overview: Genesis of the

NSSP. In: Environmental Indicatorsand Shellfish Safety, Chapman andHall, Great Britain, p 1- 17.

Hewson MD 1993. Local effects: inland andmarine. In: Climate and Weather ofNewfoundland. (eds.) Robertson A.,Porter S., and Brodie G. CreativePublishers, St. John's p 38 - 47.

Kaysner CA, Tamplin ML and Twedt RM1994. Vibrio: In: Compendium ofMethods for the MicrobiologicalExamination of Foods. (eds.)Vanderzan t C . and D.F .Splittstoesser. American PublicHealth Association, Washington, DCp 450 - 463.

Levitan CW 1991. Faecal coliformcontamination. A literature reviewand case history analysis in B.C.Bamfield Marine Station. Bamfield,B.C. VOR 1BO. 48 pp.

Nix PG, Daykin MM and Vilkas KL 1993:Sediment bags as an integrator offaecal contamination in aquaticsystems. Water Res. 27(10): 1569-1576.

Ryan AC 2003. Cardigan WatershedBacteria Study. For SoutheastEnvironmental Association TechnicalAdvisory Committee FundingPartners. Southeast EnvironmentalAssociation.

Valiela I, Alber M and Lamontagne M1991. Faecal coliform loadings andstocks in Buttermilk Bay,Massachuse t t s , USA, andmanagement implications. Envir.Man. 15 (5): 659 - 674.

influence of sampling depth and post-sampling …

Documents