Microbiological evaluation and development of quality index method (QIM) scheme for farmed pintado fish (Pseudoplatystoma corruscans)

Experiments were conducted to evaluate the microbiological quality and develop a quality index method (QIM) scheme for farmed pintado fish (Pseudoplatystoma corruscans), with estimation of shelf life. Microbiological analyses showed that two samples of pintado fish were unacceptable for human consumption due to the presence of Salmonella spp. The presence of Staphylococcus aureus in all six samples of pintado can be related with poor hygiene during handling, since S. aureus is a bacterium known to be carried by food handlers. A QIM scheme based on a total of 16 demerit points was developed. The QIM scheme proved to give a good description of the changes in whole fresh pintado during ice storage. The quality index (QI) evolved linearly with storage time in ice (R 2 = 0.97). Results indicated that the shelf life of pintado fish (whole and gutted) stored in ice is about 13 days.


INTRODUCTION
The native freshwater fish species of Brazil are numerous and many of these species are important for commercial production in South America, with good potential for aquaculture.The pintado (Pseudoplatystoma corruscans), also known as surubim, is a carnivorous species native of the Paraná River and São Francisco River basins.Pintado has a great commercial potential due to its size (it may reach 50 kg) and also due to its clear and delicate flesh, which presents few inter-muscle bones (Campos et al., 2006;Tanamati et al., 2009).
Due to over exploitation of the stocks, wild populations of pintado have been decreased in the last decade (Martino et al., 2002).Nowadays, some commercial farms are achieving good growth rates in captivity.The farming of pintado is growing in Brazil; however the production is still modest and concentrated in the Centro-Oeste region (Campos et al., 2006;Martino et al., 2002).According to IBGE (2013), the Brazilian production of pintado in 2013 was 15.715 tons and the Centro-Oeste region (Mato Grosso, Mato Grosso do Sul and Goiás) was responsible for 82.9% of the production.
Fresh fish are susceptible to rapid spoilage.The high water and protein content and neutral pH of fresh fish renders them sensitive to microbial attack (Boziaris et al., 2013;Gram and Dalgaard, 2002).Considering that fresh fish spoil mainly owing to microbial action, the microbiological evaluation is essential in determining the quality and the safety of fish (Gram and Dalgaard, 2002;Teklemariam et al., 2015).
The main quality attributes of fish are safety, nutritional characteristics and freshness.The most common methods to evaluate freshness of fish are the sensory methods.Evaluation procedures for raw fish should be simple to apply and specific for particular fish species.The quality index method (QIM) is a grading system for estimating the freshness of fish, based on objective evaluation of the main sensory attributes of each fish species using a demerit points scoring system.The QIM scheme is designed in a way that the quality parameters of very fresh fish receive zero points.Deterioration progresses of fish with storage time in ice receive the maximum of 3 demerit points.Therefore, higher scores are given as storage time progresses.A quality index (QI) is obtained by the sum of demerit points, which allows evaluating fish quality and predicting the shelf life of the species (Bernardi et al., 2013;Lanzarin et al., 2016;Sveinsdottir et al., 2003).
QIM schemes have been developed for a number of fish species including: farmed Atlantic salmon (Sveinsdottir et al., 2003), fresh cod fillets (Bonilla et al., 2007), Brazilian corvine (Teixeira et al., 2009), Brazilian sardines (Andrade et al., 2012), Brazilian Amazonian Pintado (Lanzarin et al., 2016) and others.The aim of this study was to determine the microbiological quality and to develop a quality index method (QIM) scheme for farmed pintado fish (P.corruscans), with estimation of shelf life.

Samples used for microbiological analyses
Samples were collected from six different supermarkets in the city of Brasilia, DF, Brazil, from September to December 2015.The samples of pintado stored in ice were represented by six whole fish gutted and cut in pieces, prepared at the time of purchase in the supermarkets.Fresh pintado was from aquaculture and the samples had an average weight between 2.3 and 2.7 kg.The time needed to transport samples from supermarkets to the laboratory was 20 to 40 min.The samples were transported in a cool box and microbiologically examined within maximum 1 h of sampling.The samples were analyzed in triplicate and results were expressed as de Oliveira et al. 427 mean.

Microbiological analyses
All the samples were analyzed for each of the following microorganisms or microbial groups: total aerobic mesophilic and psychrotrophic bacteria, total and thermotolerant coliforms, Staphylococcus spp.and Salmonella spp.An amount of 25 g of each sample was diluted on 225 mL of 0.1% peptone water.Samples were homogenized and an initial 10 -1 dilution was obtained.Then, serial dilutions of the homogenates were prepared in 0.1% peptone water (up to 10 -5 ).
For total aerobic mesophilic and psychrotrophic bacteria counts, serial dilutions of the samples were surface plated in plate count agar (PCA) (HiMedia, USA), following incubation at 37°C for 24 h for mesophilic bacteria and 8-10°C for 7 days for psychrotrophic bacteria.The results were expressed by colony forming unit per gram (CFU g -1 ).
For detection of total and thermotolerant coliforms, 1 mL of each dilution was transferred to three-tube series containing lauryl sulfate tryptose (LST) (HiMedia, USA) with Durham tubes in its interior.Total coliforms were enumerated in Brilliant Green Bile Broth 2% (HiMedia, USA), incubated at 37°C for 24 h and thermotolerant coliforms were determined in E. coli broth (EC) (Acumedia, USA) incubated at 45°C for 24 h.The results were expressed by most probable number per gram (MPN g -1 ).
For Staphylococcus aureus counts, serial dilutions of the samples were surface plated in Mannitol Salt Agar (HiMedia, USA), following incubation at 37°C for 48 h.The colonies were counted and sub-cultured in Mannitol Salt Agar tubes.The characteristic colonies of S. aureus (yellow colonies with yellow zones, mannitolfermenting) were stained by Gram's method to confirm Grampositive cocci.The colonies of S. aureus were further confirmed through molecular analyses.
The detection of Salmonella spp. was carried out with preenrichment in lactose broth and incubation at 37 -1 C for 24 h.Subsequently, 1 mL of this medium was inoculated on Selenite cystine broth that was incubated at 37 -1 C for 24 h.Further, Salmonella Shigella Agar (SS) (HiMedia, USA) was streaked with sterile loops carrying inoculums taken from Selenite cystine broth.The plates were incubated at 37 -1 C for 24 h in order to isolate characteristic colonies of Salmonella spp.Triple sugar iron (TSI) (HiMedia, USA) was used for presumptive confirmation of colonies that were further confirmed through molecular analyses.

Molecular analyses
The potential pathogenic bacteria S. aureus and Salmonella spp.were identified using the technique of polymerase chain reaction (PCR).For identification of S. aureus the primers CoA-F and CoA-R (Table 1) (Life Technologies, Brazil) specific for the coagulase gene (coa gene) of S. aureus were used.For identification of Salmonella spp. the primers InvA-F and InvA-R (Life Technologies, Brazil) (Table 1) specific for the invasion A gene (invA gene) of Salmonella spp.were used.For DNA extraction, the NucleoSpin Plasmid ® kit (Macherey-Nagel Inc.; Germany) was used, following the manufacturer's instructions.Extracted DNA was stored at -20°C.PCR was performed in a reaction mixture of 25 µl final volume containing 2.5 µl of PCR buffer; 0.7 µl of MgCl 2 ; 1.5 μL of dNTP (2,5 mM); 0.5 µl of Taq DNA polymerase; 1.5 μL of each primer forward and reverse and 18.3 µl of Milli-Q water.
PCR amplification was performed with an initial denaturing step at 95°C for 1 min, followed by a 35-cycle reaction (95°C for 1 min  and 60°C for 1 min).A final extension step was undertaken at 72°C for 1 min.All thermal cycling reactions were performed with Techne TC-512 thermal cycler (Bibby Scientific Inc., USA).Both negative and reagent controls were included in each PCR run.The reagent control consisted of all PCR components except the template DNA.The entire amplified DNA was separated by electrophoresis at 100 V for 50 min in 1.5% (w/v) agarose gel and stained with ethidium bromide.Gels were visualized under UV light.A 100 bp DNA ladder was used as a molecular weight marker.

Samples used for quality index method
The experiments for quality index method were performed between November, 2015 and April, 2016.Fresh pintado was from an aquaculture farm located in the state of Tocantins, Brazil.Three experiments for quality index method were performed and then 18 fish were used in total.For each experiment, 6 whole and gutted fish were kept iced and boxed in a refrigerator set at 8°C.These boxes had perforated bottoms, which allowed drainage of melted ice.Thus, ice was added on a daily basis to replace the melted fraction.Direct contact of ice with fish skin was the option chosen.The samples had an average weight of 2.3 and 2.7 kg.

Quality index method (QIM)
To develop the quality index method, the first experiment included the preparation of the preliminary QIM scheme for sensory evaluation of fresh pintado fish.Two experts observed the 6 samples of pintado after 1, 3, 5, 7, 9, 11, 13, 15 and 17 days of storage in ice.All observations were conducted under standardized conditions at room temperature.The first experiment was developed to find the characteristics that change clearly with time, necessary for the first draft of the QIM table.Major sensory changes were selected among the listed attributes and described in detail in preliminary scheme.A score of 0 to 3 demerit points was given for every change of evaluated parameter.
The second experiment of the QIM development included training of the panelists and testing of the preliminary scheme.The panel consisted of 6 panelists and the preliminary QIM table and instructions on how to perform the correct evaluation was used to train the panel.For testing the preliminary scheme, a total of 6 fish were examined after 1, 3, 5, 7, 9, 11, 13, 15, 17 and 19 days of storage in ice.All suggestions of improvements by the panel members, in the evaluation, were included in the final scheme.During this period, microbiological analyses were performed with count of total mesophilic bacteria.
The third experiment was used to validate the QIM scheme developed in the previous evaluations.The trained panel members evaluated a total of 6 fish after 1, 3, 5, 7, 9, 11, 13, 15, 17 and 19 days of storage in ice.The QIM scheme for whole fresh and gutted pintado listed quality attributes for appearance, eyes, gills and abdomen area and described how they change with storage time.Scores were given for each quality attribute ranging from 0 to 3 demerit points.The overall sum of demerit points was designated as quality index (QI) and could reach a total of 16 demerit points.During this period, microbiological analyses were performed with count of total mesophilic bacteria.

Microbiological and molecular analyses
The results of microbiological analyses are shown in Table 2.The count of mesophilic bacteria ranged from 3.0 to 5.0 log CFU/g and the count of psychrotrophic bacteria ranged from 4.2 to 6.8 log CFU/g in most samples.Only one sample of pintado fish (sample 2) had a higher count of 6.8 log CFU/g for mesophilic bacteria and 8.0 log CFU/g for psychrotrophic bacteria.
The International Commission on Microbiological Specifications for Foods (ICMSF, 1986) recommends the maximum acceptable limit of 7 log CFU/g for total count of bacteria in fresh fish.An increase in total count to levels in excess of 6 log CFU/g is usually indicative of long storage at chilling temperatures or temperature abuse prior to freezing.Thus, total count of bacteria is indicative of general quality and to a lesser extent of handling and storage procedures (ICMSF, 1986).
In this study, most samples (excluding sample 2) showed acceptable counts of total mesophilic and psychrotrophic bacteria.Delbem et al. (2012) found similar results and reported average populations of 3.3 to 4.4 log CFU/g of mesophilic bacteria in samples of pintado fish acquired in the city of Corumbá, Mato Grosso do Sul, Brazil.
All samples of pintado presented low numbers of total and thermotolerant coliforms.The low numbers of total and thermotolerant coliforms in samples indicates a good quality of the water where fish inhabit.If the fish were caught in polluted waters, Enterobacteriaceae (total coliforms) could be the mainly spoilers.Fish do not usually carry these organisms generally considered to be typical of the mammalian microflora, including Escherichia coli (thermotolerant coliforms).The presence of human enteric organisms in flesh of fish is clear evidence of faecal contamination in waters (Ampofo and Clerk, 2003;Guzmán et al., 2004).
Despite the inexistence in Brazilian legislation (Brasil, 2001) of microbiological limits for thermotolerant coliforms or E. coli in fresh fish, the maximum value suggested by ICMSF (1986) for E. coli is 2.7 CFU/g.Delbem et al. (2012) found similar results and reported average enumeration of 1.3 log MPN/g of thermotolerant coliforms in samples of pintado acquired in the city of Corumbá, Mato Grosso do Sul, Brazil.
On the other hand, Salmonella spp. was found in two samples of pintado fish (samples 1 and 2).Salmonella is not a component of the normal flora of fish, thus its contamination is consequence of fecal contamination through polluted water, infected food handlers or crosscontamination during production or transport (Carrasco et al., 2012).According to Brazilian legislation (Brasil, 2001), the presence of Salmonella in fresh fish is unacceptable (absence/25 g) and may represent a risk for consumers.Therefore, samples 1 and 2 were unacceptable for human consumption.Onmaz et al. (2015) reported 5 samples of a total of 100 fish collected in Turkey contaminated with Salmonella spp.It was concluded that the prevalence of Salmonella spp. is frequently attributed to poor hygienic practices during handling and transportation from landing de Oliveira et al. 429 centers to fish markets.After molecular analyses, there was specific amplification of gene region invA in samples 1 and 2. According to Kumar et al. (2006), the invA gene (responsible for invasion protein A), recognizes the Salmonella genus.The invA gene is an important virulence factor of Salmonella spp.It is responsible for invasion of epithelial host cells and has been reported to be present in all salmonellas.All samples of pintado were contaminated with S. aureus (confirmed by PCR).The maximum value allowed by the Brazilian legislation (Brasil, 2001) for S. aureus in fresh fish is 3 log CFU/g.Although, all samples were in accordance with the Brazilian legislation, it is important to underline that S. aureus may occur on fish usually only in low numbers and its contamination probably is consequence of food handlers (ICMSF, 1986).Therefore, good hygiene practices must be implemented by producers in order to prevent S. aureus contamination.Gatti Junior et al. (2014) reported counts of Staphylococcus spp. in tilapia's fillet ranging from 2.0 to 4.0 log CFU/g and concluded that this contamination was a consequence of improper and unsanitary handling.
The coagulase gene (coa) has been reported to be specific for confirmation of S. aureus and the coagulase protein is an important virulence factor of S. aureus.Others studies (Gandra et al., 2005;Motta et al., 2001;Shopsin et al., 2000) used the coa gene to confirm S. aureus in samples.Gandra et al. (2005) used the coa gene for the differentiation between S. aureus and Staphylococcus intermedius and Staphylococcus hyicus.Motta et al. (2001) reported that all S. aureus isolates from fresh milk were positive for the presence of the coa gene determined by PCR.

Quality index method (QIM)
The QIM scheme presented (Table 3) proved to give a good description of the changes in whole fresh pintado during ice storage.The final QIM scheme developed for the farmed pintado fish included 8 parameters, grouped in 5 main attributes and had a total of 16 demerit points.Attributes were defined as appearance, eyes, gills, flesh and abdomen area.During the development of the scheme, the rotten odour was changed to very sour and for the attribute eyes, scores were added to 3 points because sunken shape eyes were observed.
At the beginning of the storage time (0-7 days), the pintado was very fresh (QI = 0-1.2).From the seventh day of storage in ice, QI average was 3.5 and it has been possible to observe the gills with dark red colour, the skin with darker colour and the eyes with flat shape, black pupils and slightly opaque cornea, but the odour still remained neutral.
The maximum storage life of pintado was 13 days in ice.During this period, the average QI was between 4.5 and 7.5 demerit points.Various sensory changes were observed as the beginning of sour odour upon opening the box (slightly sour) which dissipated completely when handling the fish, the gills with dark red colour, the eyes with little sunken shape, pupils and cornea opaque and the abdomen starting to get old aspect (opaque, rusty pink colour).
After 15 days of storage in ice, the odour was increasingly characterized by sour and this indicated that pintado was in the end of shelf life and acceptability.After 17 days of storage in ice, pintado had already been rejected, with sour odour and flesh with old aspect (opaque, rusty pink colour).Between 15 and 19 days of storage in ice, typical sensory changes of the deterioration progressed rapidly, with very sour odour quite noticeable in the abdomen area, the flesh was with softened texture and fat crumbling and the dorsal area was with a yellowish colour in the area between the mouth and gills.
Similar results for the proposition of the QIM scheme were found by Lanzarin et al. (2016) with maximum score of 18 points for gutted Amazonian Pintado (Pseudoplatystoma fasciatum x Leiarius marmoratus).In other studies, Sveinsdottir et al. (2003) and Teixeira et al. (2009) developed a QIM scheme based on a total of 22 demerit points for Atlantic salmon and Brazilian corvine, respectively.The differences between the QIM schemes show that this is a specific method that considers the differences observed in the deterioration process for the distinct species.Lanzarin et al. (2016) described that the quality attributes that most influenced the QIM protocol of fresh gutted Amazonian Pintado were abdominal odour and abdomen colour (yellowish flesh).In this study, these characteristics were observed when pintado was rejected, between 15 and 19 days, with very sour odour in the abdomen area and the dorsal area with a yellowish colour.In the final developed QIM scheme, there was a linear relationship with high correlation (R 2 =0.97) between the average QI with storage time in ice (Figure 1).
Table 4 presents the counts of mesophilic bacteria in  the samples of pintado fish stored in ice for 19 days.An initial bacterial flora of around 4 log CFU/g remained constant along the first 7 days in ice, while fish was still fresh, with good quality.Between 9 and 11 days of storage in ice, when the fish was less fresh, the mesophilic count increased to 5 log CFU/g.The maximum storage life of pintado was 13 days in ice, when the mesophilic bacteria counts were 6.0 log CFU/g.In 17 days of storage in ice, when the fish was rejected, the mesophilic bacterial count reached 6.5 log CFU/g.According to Barbosa et al. (2002), the total bacterial count just after catch fish or at the start of storage is variable between 1 and 4 log CFU/g, but at the time of sensorial rejection, it was between 6 and 8 log CFU/g, levels which are in accordance with the counts of microorganisms obtained in the present work.Lanzarin et al. (2016) estimated that the shelf life for ice stored gutted Amazonian Pintado was 12 days, when the mesophilic bacteria counts were 6.5 log CFU/g.At 14 days of storage, the population of these bacteria increased to 7.1 log CFU/g and Amazonian Pintado was considered unacceptable for consumption In conclusion, microbiological analyses showed that two samples of pintado were unacceptable for human consumption due to the presence of Salmonella spp.The other four samples were acceptable for consumption, although the presence of S. aureus in all samples can be related with poor hygiene during handling.The QIM scheme has proved to be effective for assessing the quality of fresh water pintado fish (whole and gutted) stored in ice, indicating that its shelf life is about 13 days.

Figure 1 .
Figure 1.QI linear correlation with ice storage days of the 16 demerit points for QIM scheme proposed for farmed pintado fish (Pseudoplatystoma corruscans)

Table 1 .
Primers sequence and size of amplified products of PCR.

Table 2 .
Counts of mesophilic and psychrotrophic bacteria, enumeration of total and thermotolerant coliforms, counts of S. aureus and detection of Salmonella spp. in the samples of pintado fish.
Results are reported as means ± standard deviation of triplicate measurements; ND: not detected.

Table 4 .
Count of mesophilic bacteria in pintado fish stored in ice.