Investigation of clonal distribution and persistence of Salmonella Senftenberg in the marine environment and identification of potential sources of contamination

Authors

  • Jaime Martinez-Urtaza,

    Corresponding author
    1. Instituto de Acuicultura, Universidad de Santiago de Compostela, Campus Universitario Sur, 15782 Santiago de Compostela, Spain
    Search for more papers by this author
  • Ernesto Liebana

    1. Department of Food and Environmental Safety, Veterinary Laboratories Agency-Weybridge, Addlestone, Surrey KT15 3NB, UK
    Search for more papers by this author

*Corresponding author. Tel.: +34 981 528024/563100x16043; fax: +34 981 547165. ucmjmur@usc.es

Abstract

Salmonella Senftenberg was detected in the coastal areas of Galicia (NW Spain) in 1998, where it remained the predominant serovar for the next four years. Although the overall incidence of this serovar in the zone was lower than 1%, contamination by Salmonella serovar Senftenberg was located in very specific areas of the Ría de Arousa, where it persisted for more than five years. A total of 60 Salmonella serovar Senftenberg isolates, originating from surveillance activities in marine environments, was subjected to molecular characterization by pulsed-field gel electrophoresis (PFGE). PFGE analysis of the marine isolates allowed the differentiation of three main PFGE types, which contained the majority of the isolates, each type showing a specific spatial distribution in the coastal waters. The most prevalent pulse types persisted for more than four years, emphasizing their capacity to adapt and survive in marine environments. Using PFGE analysis, marine isolates were compared with Salmonella serovar Senftenberg isolates from neighbouring mussel-processing facilities and to other epidemiologically unrelated isolates from human, animal and feed sources. Comparison of the restriction patterns showed that indistinguishable PFGE types were present in the isolates from mussel-processing facilities and their surrounding marine areas, suggesting that the mussel processing is the main source for contamination with Salmonella Senftenberg in these marine environments. A molecular fingerprinting relationship was established between three shellfish isolates and a human isolate, which could be considered as preliminary evidence of infection caused by Salmonella Senftenberg associated with molluscan consumption.

1Introduction

Members of the genus Salmonella are gram-negative rods that spend a considerable part of their lifecycle as residents of animal hosts [1]. A great number of animal species function as natural reservoirs for these organisms, indicating the lack of host-specificity for most of the Salmonella serotypes [2]. This ubiquitous nature facilitates a lifecycle consisting of passage through a host to the environment and then to a new host [3]. Salmonella has been detected in a diverse range of environmental sources, such as water [4,5], sewage effluent [6] and soil, where it can survive for periods of at least one year [7]. These observations suggest that Salmonella possesses a high resistance to a large variety of stresses related to environmental changes, resulting in environmental persistence for a certain period of time [1]. This survival capacity plays a vital role in the Salmonella lifecycle and increases the probability of colonizing a new host [1].

In marine environments, Salmonella has been found associated to seawater, molluscs and other seafood products [4,5,8–12]. The presence of these enteric bacteria in coastal waters has been linked to heavy rain and storm-generated discharges, transporting the contamination from their sources to the sea via river waters [4,13–15]. In the recent years, Salmonella enterica serovar Senftenberg appeared to be one of the main serovars in the marine environment [11]. In contrast to other serovars, studies on the presence of Salmonella serovar Senftenberg in coastal waters have not shown any direct relationship to environmental factors [14]. This serovar has also been isolated in association with persistent contamination of fish feed factories [16] and mussel-processing facilities [17]. Nevertheless, there is very little information about this sudden protagonism of Salmonella serovar Senftenberg in coastal waters and related environments, and about the possible sources of contamination that have facilitated the introduction and spread of this serotype in these habitats. In the present study, the distribution of clonal types of Salmonella serovar Senftenberg in the marine environments was investigated, and, subsequently, the possible sources of contamination through comparison of different populations of these bacteria were identified.

2Materials and methods

2.1Bacterial strains

A total of 60 Salmonella serovar Senftenberg isolates from marine environments was characterized (Table 1). These isolates represented all the serovar Senftenberg isolates identified among a total of 133 Salmonella isolates obtained using standard bacteriological procedures from the analysis of 6317 samples of molluscs and seawater, taken during routine monitoring surveillance from 1998 to 2002 (1998, n= 933; 1999, n= 1134; 2000, n= 916; 2001, n= 1043; 2001, n= 2291) [11,14]. The samples were collected in the mollusc production areas, located in the four most important rias (estuaries similar to small fiords which extend from East to West) of Galicia in North-western Spain (Fig. 1).

Table 1.  Identification of the 60 isolates of Salmonella serovar Senftenberg from marine environments included in this study and the XbaI-types obtained by PFGE analysis
OriginaZoneaIsolate numberDate isolationSourceXbaI-type
  1. aRias of Galicia and molluscan harvesting zones.

ArousaIUCM-13729-Aug-01MusselX08
  UCM-13220-Aug-01MusselX36
 IIUCM-14903-Oct-01MusselX11
  UCM-01528-Sep-98MusselX19
 IIIUCM-25805-Nov-02SeawaterX08
  UCM-14424-Sep-01MusselX08
  UCM-15003-Oct-01MusselX08
  UCM-22610-Sep-02MusselX08
  UCM-22710-Sep-02SeawaterX08
  UCM-25204-Nov-02MusselX08
  UCM-10119-Oct-00MusselX10
  UCM-09822-Nov-00Waste waterX11
  UCM-15303-Oct-01Waste waterX11
  UCM-15203-Oct-01MusselX11
  UCM-15403-Oct-01Waste waterX11
  UCM-17312-Dec-01MusselX11
  UCM-26005-Nov-02MusselX11
  UCM-26406-Nov-02MusselX11
  UCM-27402-Dec-02MusselX11
  UCM-15103-Oct-01MusselX39
  UCM-16714-Nov-01MusselX39
  UCM-20713-Aug-02MusselX39
  UCM-10219-Oct-00MusselX42
 IVUCM-17410-Dec-01ClamX12
  UCM-16215-Oct-01MusselX19
 VUCM-20218-Jun-02MusselX04
  UCM-25505-Nov-02MusselX10
  UCM-05629-Sep-99ClamX31
  UCM-11527-Nov-00MusselX42
 VIUCM-27609-Dec-02MusselX11
  UCM-27709-Dec-02MusselX11
  UCM-05402-Sep-99MusselX31
  UCM-05212-Aug-99MusselX31
  UCM-07008-Nov-99MusselX41
 VIIUCM-08527-Mar-00ClamX19
  UCM-02830-Nov-98MusselX19
  UCM-03118-Jan-99MusselX19
  UCM-03719-Apr-99MusselX31
 VIIIUCM-16422-Oct-01OysterX10
  UCM-05022-Jul-99MusselX11
  UCM-09902-Oct-00ClamX19
  UCM-02630-Nov-98MusselX19
  UCM-02730-Nov-98MusselX19
  UCM-02930-Nov-98MusselX19
  UCM-03030-Nov-98MusselX19
  UCM-05725-Aug-99SeawaterX31
  UCM-08720-Jul-00MusselX31
  UCM-04915-Jul-99MusselX31
  UCM-05326-Aug-99MusselX31
  UCM-05509-Sep-99MusselX31
  UCM-12202-Apr-01OysterX35
  UCM-26206-Nov-02MusselX40
  UCM-11422-Nov-00OysterX43
  UCM-26306-Nov-02MusselX44
 IXUCM-20302-Jul-02MusselX04
      
Muros-NoiaIIUCM-23017-Sep-02MusselX11
  UCM-25028-Oct-02MusselX38
      
PontevedraIIIUCM-07129-Nov-99ClamX19
      
VigoIIIUCM-16315-Oct-01MusselX04
  UCM-16629-Oct-01MusselX37
Figure 1.

Location of the coastal areas included in the present study.

The XbaI-restriction patterns of marine environment isolates were compared to those obtained from isolates of mussel-processing facilities (n= 108) located in the same area, and to other epidemiologically unrelated isolates (human, n= 6; animal, n= 5; feed and mud, n= 11) of this serovar obtained from a previous study (J. Martinez-Urtaza and E. Liebana, unpublished). Using standard bacteriological procedures, all the isolates from mussel-processing facilities were obtained from processed mussel samples and items related to their manufacture (water, brine and grease), associated with persistent contaminant events between 1998 and 2002 [17]. Restriction patterns were compared using BioNumerics software (Applied Maths, Sint-Martens-Latem, Belgium) in order to establish a possible source of Salmonella serovar Senftenberg contamination in the coastal areas.

2.2PFGE

PFGE was preformed according to the “One-Day (24–28 h) Standardized Laboratory Protocol for Molecular Subtyping of Non-typhoidal Salmonella by PFGE” (Pulse-Net, CDC, Atlanta, USA) [18]. A single colony of each isolate was streaked on tryptic soy agar (TSA) and incubated overnight at 37 °C. Using a cotton swab, part of the cell culture on agar plate was transferred to 2 ml of Cell Suspension Buffer (100 mM Tris:100 mM EDTA, pH 8.0) and the concentration of cell suspensions was adjusted to 0.48–0.52 in a Dade Microscan Turbidity Meter (Dade Behring, USA). Immediately, 400 μl of adjusted cell suspension was transferred to 1.5-ml micro-centrifuge tubes with 20 μl of proteinase K (20 mg/ml), subsequently mixed with 400 μl of melted 1% SeaKem Gold (Cambrex, East Rutherford, NJ) :1% SDS agarose prepared with TE Buffer (10 mM Tris:1 mM EDTA, pH 8.0), and pipetted into disposable plug moulds. Three plugs were transferred to 50-ml polypropylene screw-tubes with 5 ml of Cell Lysis Buffer (50 mM Tris:50 mM EDTA, pH 8.0 + 1% sarcosyl) and 25 μl of proteinase K (20 mg/ml), and incubated at 54 °C in a shaker water bath for 2 h with agitation. Thereafter, the plugs were washed twice with 15 ml of sterile water and three more times with TE Buffer at 50 °C for 15 min. Chromosomal DNA was digested with 50 U of XbaI (Promega, Southampton, UK). PFGE was performed on a CHEF DRIII system (Bio-Rad, Hercules, CA) in 0.5× Tris-Borate-EDTA (TBE) extended range buffer (Bio-Rad) with recirculation at 14 °C. DNA macrorestriction fragments were resolved on 1% SeaKem Gold Agarose (Cambrex) in 0.5× TBE buffer. DNA from Salmonella Braenderup H9812, restricted with XbaI, was used as a size marker. Pulse times were ramped from 2.2 to 63.8 s during an 18-h run at 6.0 V/cm. Macrorestriction patterns were compared using BioNumerics software (Applied Maths). Different profiles were designated with the letters X (XbaI types) in accordance with the restriction patterns.

2.3Statistical analysis

The differences in the number of isolates belonging to each PFGE type by year and isolate source (marine environment and mussel-processing facilities) were determined using the χ2-test. All the statistical analysis was performed with the SPSS version 11.0.1 (SPSS Inc.) and the level of significance was set at P < 0.05.

3Results

PFGE typing of the 60 Salmonella serovar Senftenberg isolates from marine environments discriminated 16 different pulse types (Table 1 and Fig. 2). The four most prevalent pulse types were X11, X19, X31 and X08 with 13, 11, 9 and 8 isolates, respectively. Seven XbaI-types were comprised by a unique isolate. Types X08, X11 and X39 presented similar restriction patterns and isolates from these types were prevalently isolated in Zone III of Arosa (Fig. 3), while X31 and X19 types were mostly isolated from the same area of the Ria of Arosa (Zones VI, VII and VIII). The most persisting types (Table 2) were X11, detected from 1999 to 2002, and X19, detected from 1998 until 2001. Type X11 isolates were significantly more abundant (P < 0.05) in 2002 than during the rest of the years, while type X19 isolates were significantly more frequent (P < 0.05) in 1998 and then decreased significantly in 2001. Type X31 isolates were found from April 1999 to July 2000, although significantly more frequently in 1999. Type X08 isolates were detected from August 2001 to November 2002 and were significantly more abundant in 2002. The number of XbaI-types detected in the different years is shown in Table 2. The six isolates obtained in 1998 belonged to a single type (X19). The genetic diversity increased during the years, with 4 different pulse types identified in 1999, 6 in 2000, 10 in 2001 and 7 in 2002.

Figure 2.

Dendogram generated by Bionumerics software, showing representative restriction patterns (XbaI-types) of Salmonella serovar Senftenberg isolates obtained from the marine environments.

Figure 3.

Representation of the distribution of the XbaI-PFGE types identified during this study of isolates from the Ria of Arousa and facilities located in its surrounding coast. (a) Geographical location of isolates belonging to the different pulse types. (b) Area of distribution of the three most prevalent pulse types in the sea.

Table 2.  Annual distribution of the pulse types identified with PFGE analysis of the isolates from marine environments and facilities (Martinez-Urtaza and Liebana, unpublished), detailing the number of isolations of each pulse type per year according to the origin of the isolates
XbaI-types19981999200020012002TotalTotal
EaFbEFEFEFEFEF  
  1. aE: isolates from marine environment.

  2. bF: isolates from mussel-processing facilities.

  3. *Significant differences in the number of isolates from marine environment belonging to each PFGE type by year at P < 0.001.

  4. **Significant differences in the number of isolates belonging to each PFGE type by isolate source (marine environment and facilities) at P < 0.05.

X03   1 3 3 5 1212
X04 1    1 2 3**1**4
X05         2 22
X06       1 1 22
X07 1         11
X08     1355*881422
X09         2 22
X10    1 1 113**1**4
X11  181 5 6*14132235
X12      11  112
X13         1 11
X16       1 1 22
X17       1   11
X18 1 1       22
X196*929241*12 111**35**46
X20 1         11
X21         1 11
X22         1 11
X23 1         11
X30 2         22
X31 28*11     9**3**12
X35      1   1 1
X36      1   1 1
X37      1   1 1
X38        1 1 1
X39      2 1 3 3
X40  1       1 1
X41    2     2 2
X42    1     1 1
X43        1 1 1
              
Total6181220881724173860108168

The PFGE analysis, resulting from the integration of the restriction profiles of isolates obtained from the mussel-processing facilities (J. Martinez-Urtaza and E. Liebana, unpublished) and marine environments, detected indistinguishable PFGE types in isolates from both sources. Of the 16 XbaI-types identified in the marine environment, seven were coincident with pulse types identified among isolates from the eight mussel-processing plants studied (Table 2). Types X04, X10 and X31 were found significantly more frequent (P > 0.05) in the marine environment than in the facilities, while type X19 isolates were significantly more frequent in mussel-processing facilities. No significant statistical differences (P > 0.05) were detected for X08, X11 and X12 type-isolates obtained from both sources. A total of 14 pulse types originating from the processing facilities was not found in isolates from the marine environment. In 1998, only the X19 type was identified among isolates from the marine environment, while eight types were detected in facilities isolates, including the X19. The number of types detected in marine environment and facilities in 2002 was 7 and 12, respectively. The geographical distribution of the PFGE types found in isolates from mussel-processing facilities placed in the Ria of Arousa showed a clear spatial relationship (Fig. 3). Identical PFGE types were found in isolates from mussel-processing facilities and their surrounding marine areas. Type X11 was the most widespread, and it was present in the majority of the coastal areas and facilities in the north and south sides of the Ria of Arousa. Type X31 isolates appeared on coastal areas and factories located in the southern side, while isolates of the type X19 were exclusively detected in the marine environment of this ria.

When the restriction profiles from clinical and feed isolates (J. Martinez-Urtaza and E. Liebana, unpublished) were incorporated into the comparative analysis, the types of these isolates were clearly separated from isolates belonging to facilities and marine environments. Only the human isolate 391/01 obtained in 2001 presented an indistinguishable XbaI-type (X10) to two isolates from live mussels and one isolate from live oysters from years 2000, 2001 and 2002, respectively.

4Discussion

Recent studies evaluating the incidence of Salmonella spp. in the coastal waters of Galicia presented values that did not exceed 2.4%[14]. Salmonella serovar Senftenberg was the most frequently isolated serovar in these studies, as has been previously reported for other marine environments of temperate and tropical zones [8–11]. In spite of its dominance in the rias of Galicia from 1998, the general incidence of Salmonella serovar Senftenberg in the zone is lower than 1% and its presence is restricted to well-defined areas of the Ria of Arousa, where important activities of mussel production and processing take place. Data obtained in a previous study linked the presence of several Salmonella serovars in the Ría de Arousa to heavy and persistent rains [14]. This seasonal pattern in Salmonella spp. prevalence has been observed in other studies, where a relationship has been detected with storm-generated flows, torrential rains or the monsoon season [4,13,15]. However, the presence of Salmonella serovar Senftenberg in the Ria of Arousa could not be associated with rainfall or any other environmental factor [14]. Persistent contamination by Salmonella serovar Senftenberg in mussel-processing facilities and marine environments was detected during the same period of study [17]. Considering this persistent and localized contamination by Salmonella serovar Senftenberg and the lack of correlation with any environmental parameter, there appears to be a strong link between mollusc processing and seawater contamination. Interestingly, the relationship between mollusc processing and contamination in the marine environment is further supported by the coincident seasonality (June–December) of mussel processing and the presence of Salmonella serovar Senftenberg in the marine environment [14].

The present study aimed to provide further evidence by performing molecular typing of the isolates. The results obtained showed a coincidence, and a clear spatial relationship between the pulse types identified among isolates from processing facilities and among isolates form the surrounding marine areas. This suggests that mussel-processing industries were the main source of contamination. Seven PFGE types identified in the isolates from the Ria of Arousa had been previously detected in isolates from facilities located in the immediately surrounding coast. Identification of the facilities as the source of contamination is also supported by the fact that when a new type was detected in the marine environment, it had always been previously identified in isolates from processing facilities. Of special interest is the case of contamination of facility 3. This factory is located inland and lacks direct effluents to the sea. Although type X03 was the cause of persistent contamination of this facility for four consecutive years (J. Martinez-Urtaza and E. Liebana, unpublished), this type has never been detected in the marine environment.

Clustering analysis of PFGE types from Salmonella serovar Senftenberg isolates from other sources (human, animal and feed isolates) showed that these formed a clearly differentiated cluster compared to isolates from mussel processing and marine environment. This cluster separation reveals a genetic source-related difference between populations of this organism, and suggests that human, animal and feed isolates did not represent a significant source of contamination of Salmonella serovar Senftenberg in the marine environment during our study.

The discrepancy observed in the number of pulse types detected in isolates from facilities and marine environment was most likely due to the sampling strategy used in the studies. The sampling of the marine environment included almost the entire Galician coast, and included the analysis of a similar number of samples per year. The microbiological control of the mussel-processing plants, however, was limited to only eight participating facilities from dozens of mussel-processing plants and mussel-canning factories existing along the coast. This situation implies that only partial information was obtained of the Salmonella serovar Senftenberg clones present in the facilities, and may explain why not all the clones detected in the marine environment could be identified in the facilities. The absence in the marine environments of 14 of the 21 types identified in the facilities could be attributed to the prevalence of the different clones in the factories. The types detected in the coastal waters were always the most prevalent in facilities, while less frequent facility types were undetected in the marine environment.

The results from PFGE analysis of the isolates from marine environment revealed an increasing genetic heterogeneity from the first isolation. For the whole duration of the study, the number of samples analyzed and of zones studied per year was almost constant. However, when the contamination was first detected in 1998, only one single pulse type was identified in the marine environment. The number of types detected gradually increased in the following years until 2001. The lack of genetic diversity detected in the marine environments in 1998 contrasts with the degree of heterogeneity of the isolates present in this year in the facilities (Table 2). This observation indicates that the origin of contamination is unlikely to be associated with the live molluscs entering processing lines in the factories, and situates the origin of the contamination inside the mussel-processing plants. A thorough anamnestic investigation of the mussel-processing plants revealed that in the same period, massive importations of low-quality marine salt from foreign countries could have been the origin of contamination with Salmonella serovar Senftenberg [17] (J. Martinez-Urtaza and E. Liebana, unpublished). This serovar, well adapted to hypersaline environments, spread on the facilities and some of the clones reached the nearby marine environment, contaminating the seawater and the molluscs that live there. This exemplifies a contamination of the environment by Salmonella, derived from a non-faecal industrial source.

Previous investigations have described a long-last survival of Salmonella in non-host environments. They have been shown to exhibit higher survival ratios in the marine environment than other bacteria [19]. However, to date no life cycle of Salmonella without a host has been described. The distribution of the dominant clones in the Ría de Arousa demonstrates the great capacity of adaptation of Salmonella serovar Senftenberg to the marine conditions. The two most prevalent pulse types in this study (X11 and X19) were detected for four years in the same coastal area. Similar prevalence was identified in facilities, where some types were detected for five consecutive years (J. Martinez-Urtaza and E. Liebana, unpublished). The persistence of specific types of Salmonella serovar Senftenberg in marine environment and in facilities for years, without evidence of intervention of an intermediate host to guarantee their survival, gives proof of the survival capacity of this serovar in non-host environments. Likewise, this could be an early evidence of a long-term lifestyle of Salmonella in the environment without host intervention.

In Spain, Salmonella serovar Senftenberg is infrequently isolated from human infections. Only two cases of a total of 5480 isolations of human origin in 1998, four of the 5326 in 1999 and two of the 6043 in 2000 were reported [20–22]. None of these infections were associated with molluscan consumption. The detection of contamination of bivalve molluscs from 1998 did not reflect any increase in the human isolates of Salmonella serovar Senftenberg reported during this period. However, the PFGE analysis of the isolates included in this study showed a molecular fingerprinting relationship between three shellfish isolates and the human isolate 391/01, obtained from a stool sample in 2001. This could represent the first preliminary evidence of an infection caused by Salmonella serovar Senftenberg associated with molluscan consumption.

Acknowledgements

We are grateful to Lourdes García-Migura and Dr. Miranda Batchelor (VLA, Weybridge, UK) for their technical assistance. We also thank Jacobo de Novoa for assistance in statistical analysis of data.

Ancillary