Introduction

Preservatives are required to maintain the quality, extend shelf life and ensure safety of fresh and processed food products. Although chemical preservatives form an essential part in food preservation, legislation has restricted their use in different foods (Brul and Coote 1999). At the same time, consumer preferences are moving towards foods that contain lower levels of chemical preservatives, exhibit characteristics of fresh or natural products and are microbiologically safe as well.

Non-thermal methods such as the addition of naturally occurring compounds having antibacterial activity (Hayes et al. 2010), high pressure carbon dioxide (Garcia-Gonzalez et al. 2009), high intensity pulsed electric field (Mosqueda-Melgar et al. 2008), irradiation (Alighourchi et al. 2008) or ultrasound (Schenk et al. 2008; Salleh-Mack and Roberts 2007) are increasingly gaining attention for preservation of minimally processed foods. In recent years, the use of naturally occurring antimicrobial agents to inhibit pathogen growth and prevent food spoilage has received special attention (Hayes et al. 2010). Useful antimicrobial phytochemicals can be divided into several categories such as phenolics and polyphenols; quinines; flavones, flavonoids and flavonols; tannins; alkaloids and lectins; coumarins and polypeptides. Nowadays, minimal preservation processes based on the combined factors technology are also gaining importance for food preservation. Char et al. (2010) studied the response of Listeria innocua to combined treatments involving moderate temperatures and the addition of different levels of citral to obtain a minimally processed orange juice.

Seaweeds are considered a source of bioactive compounds as they are able to produce a great variety of secondary metabolites characterised by a broad spectrum of biological activities. Although seaweeds grow in a harsh environment, they seldom suffer any serious photodynamic damage during metabolism. This fact implies that seaweed cells have some protective compounds and mechanisms (Matsukawa et al. 1997). Since seaweeds are a good source of antimicrobial compounds, ω3 fatty acids, antioxidants and other bioactive compounds, there is an interest to utilize these products as nutraceuticals and in functional foods (Yuan 2008). Compounds, such as polyphenols, flavonoids and polysaccharides, having antioxidant and antimicrobial activities have been detected in brown, red and green algae (Cox et al. 2009; Zaragoza et al. 2008). Ara et al. (2002) reported brown algae to be active against a number of Gram-positive and Gram-negative organisms. Nagayama et al. (2002) reported that phlorotannins, brown-algal phenolic compounds, such as eckol and eckol-related compounds, from Ecklonia kurome, have strong bactericidal activity. A series of polyphenolic compounds such as catechins, flavonols and flavonol glycosides have been identified from methanol extracts of red and brown algae (Hosokawa et al. 2006) and found to have antioxidant and antimicrobial activity. Horie et al. (2008) isolated sargaquinoic acid derivatives from the brown alga Sargassum sagamianum having antibacterial properties.

Traditional microbial enumeration techniques are time-consuming, and therefore, mathematical microbial models are used to assess the potential for growth of microorganisms in foods during processing and storage (Bovil et al. 2001). Empirical sigmoidal type models such as the modified Gompertz and logistic models or the semi-mechanistic model of Baranyi–Roberts have been used for fitting bacterial growth (Xiong et al. 1999). However, data on the use of actual plant extract for inhibiting microbial growth and modelling the resulting kinetics are scarce. Most of the studies done till date use either thermal treatments or purified compounds having antimicrobial activity for studying growth inhibition.

This study was conducted to determine the effect of different concentrations of brown seaweed (Himanthalia elongata) extract against Listeria monocytogenes, Salmonella abony, Enterococcus faecalis and Pseudomonas aeruginosa. There are some reports available wherein the antimicrobial effect of seaweed extract has been studied on different organisms (Taskin et al. 2010; Cox et al. 2009; Ely et al. 2004; Nagayama et al. 2002), but no studies are available where the growth inhibition has been modelled. Hence, the present study investigates the utilization of methanolic extract from brown seaweed as a natural antimicrobial agent for food preservation by examining its effects on the growth kinetics of four common food spoilage and food pathogenic microorganisms. In order to describe growth inhibition in the presence of seaweed extract, performance of three commonly used primary models, namely the Baranyi–Roberts, modified Gompertz and logistic models, was evaluated.

Materials and Methods

Seaweed Material

Brown seaweed H. elongata (Pheophyta) was purchased from Quality Sea Veg., Co. Donegal, Ireland. Samples were received in September 2009 and washed thoroughly with freshwater to remove epiphytes and salt.

Preparation of Seaweed Extracts

The extraction of seaweed was carried out with 60% methanol under nitrogen atmosphere at 40 °C and 100 rpm in a shaker incubator (Innova 42, Mason Technology, Ireland). Samples were filtered and centrifuged at 10,000 rpm (8,720×g) for 15 min (Sigma 2K15, Mason Technology, Ireland). Resulting extracts were evaporated to dryness using vacuum polyevaporator (Buchi Syncore Polyvap, Mason Technology, Ireland) at 60 °C. A pressure gradient programme was designed for evaporation of the solvents with vacuum conditions of 33,700 and 7,200 Pa for methanol and water, respectively.

Antimicrobial Activity

Microbial Culture

Two species of common food pathogenic and food spoilage bacteria selected for this study were L. monocytogenes (ATCC 19115), S. abony (NCTC 6017), E. faecalis (ATCC 7080) and P. aeruginosa (ATCC 27853), respectively (Medical Supply Company, Dublin, Ireland). All cultures were maintained at −70 °C in 20% glycerol stocks and grown in Tryptic Soy Broth (TSB; Scharlau Chemie, Barcelona, Spain) at 37 °C, except for P. aeruginosa which was incubated at 30 °C, to obtain sub-cultures. A final cell concentration of 1 × 106 colony-forming units (CFU)/ml was used for the experiments.

Relation Between Turbidity and Viable Count

Before the kinetics study, a relationship between optical density at 600 nm and viable count was determined for all of the bacteria studied. A volume of 200 μl of bacterial suspension containing 6 log CFU/ml was dispensed into 50 wells of the 96-well microtitre plate (Sarstedt Ltd., UK). Every hour, the optical density (OD) of the microtitre plate was read. At the same time, an aliquot of 100 μl from one well was transferred into 900 μl of maximum recovery diluent (Scharlau Chemie, Barcelona, Spain) to determine the viable cell count. Spreading was carried out on Tryptic soy agar (Scharlau Chemie, Barcelona, Spain) plates by taking 100 μl of relevant dilution. Plates were incubated at 37 °C, with the exception of P. aeruginosa (30 °C), for 24 h before determining the CFU per millilitre. A standard curve (OD600nm vs. log CFU per millilitre) was drawn from the results obtained. This curve was later used for conversion of the OD values to log CFU per millilitre for respective bacteria in the presence of seaweed extract.

Antimicrobial Activity Assay

The influence of varying concentrations of extract on efficacy was assessed against the four organisms using 96-well microtitre plates. Extract (300 mg) obtained from 5 g fresh seaweed was dissolved in TSB (2.5 ml), and 200 μl was added to the first row of each plate. All other wells were filled with 100 μl of TSB, and 100 μl from the first well was serial diluted into 2-fold along each column. Finally, 100 μl of bacterial suspension containing 6 log CFU/ml was added to the wells. The last row was used for bacterium and media controls. Sample blanks were also prepared for all of the extracts. The plate was incubated in the microtitre reader for 24 h at respective temperature for each organism. Microbial growth was recorded every 2 h on a Powerwave microplate spectrophotometer (Powerwave, Biotek) driven by Gen5 reader control and data analysis software. Turbidity was measured as absorbance at 600 nm, with 20 s agitation before each OD measurement. The OD values were converted to log CFU per millilitre by the standard curve as described in “Relation Between Turbidity and Viable Count” section.

Growth Curve

To describe the inhibition of bacterial growth in the presence of seaweed extract, three primary growth models, namely modified Gompertz, logistic and Baranyi–Roberts model, were fitted to the data, and their performance was comparatively evaluated. Growth curves were plotted to evaluate the antibacterial activities of the seaweed extract.

Baranyi–Roberts Model

A programme implemented in Microsoft Excel (DM-Fit; Institute of Food Research, Norwich, UK) was used to fit the equation of Baranyi and Roberts (1994) to the growth data. To evaluate the effect of different extract concentrations main kinetic parameters such as exponential specific growth rate (μ), Log10 maximum population density (γ), lag time (λ) and the coefficient of determination (R 2) were calculated.

Modified Gompertz Model

The modified Gompertz model (Gibson et al. 1987) is given by Eq. 1,

$$ {N_t} = A + C\, \times \exp \,\left[ { - \exp \,\left\{ { - B \times \,\left( {t - M} \right)} \right\}} \right] $$
(1)

where N t is the cell number (log CFU per millilitre) at any time t, A is the lower asymptotic line of the growth curve as t decreases to zero (that is N 0: initial population level at time t = 0 (log CFU per millilitre)), C is the difference between the upper asymptotic line of the growth curve (maximum population level, N max) minus the lower asymptotic line (for example, N max − N 0 (log CFU per millilitre)), B is the relative maximum specific growth rate (per hour) at time M and M is the time at which the specific growth rate is maximum. Equations 2, 3 and 4 can then be used for the calculation of maximum specific growth rate (μ max (per hour)), lag phase duration (λ, hours) and maximum cell population (N max), respectively:

$$ {\mu_{{\max }}} = \frac{{B\, \times \,C}}{e} $$
(2)

where e = 2.7182

$$ \lambda = M - \frac{1}{B} $$
(3)
$$ N_{max} = A + C$$
(4)

Logistic Model

The logistic model used for defining bacterial growth as a function of time at constant environmental conditions, such as temperature, pH, water activity etc. is given by Eq. 5 (Gibson et al. 1987):

$$ {N_t} = A + \frac{C}{{1 + \exp \left[ { - B \times \left( {t - M} \right)} \right]}} $$
(5)

where N t , A, B, M and C have the same meaning as given for the modified Gompertz equation. The μ max and λ parameters can be calculated by Eqs. 6 and 7, respectively, as follows:

$$ {\mu_{{\max }}} = \frac{{B\; \times \;C}}{4} $$
(6)
$$ \lambda = M - \frac{2}{B} $$
(7)

Model Comparison

Root Mean Square Error

The smaller the root mean square error (RMSE) values, the better the fit of the model to the data.

$$ {\hbox{RMSE}} = \sqrt {{\frac{{\sum {{{\left( {{\text{predicted}} - {\text{observed}}} \right)}^2}} }}{{n - p}}}} $$
(8)

where n is the number of observations and p the number of parameters to be estimated.

Curve Fitting

A plot of microbial count versus time for each extract concentration was used to derive the starting values for the parameters, N 0 and N max, for all three models evaluated. The lag time was obtained from the raw data by noting the time when exponential growth started. The experimental data were fitted to equations described above by nonlinear regression with a Marquardt algorithm using the software Statgraphics Centurion XV (StatPoint Technologies, Inc., Warrenton, VA, USA). The aim of the fitting procedure was to find each model's parameters that best described the data by minimizing the sum of the squares of the differences between the model simulated and experimental values.

Statistical Analysis

All experiments were performed in duplicate and replicated at least three times. All statistical analyses were carried out using STATGRAPHICS Centurion XV. Statistical differences between extract activities were determined using ANOVA followed by least significant difference testing. Differences were considered statistically significant when p < 0.05.

Results and Discussion

Antimicrobial Effect of Different Concentrations of H. elongata Extracts

In our previous study (Cox et al. 2009), we had reported the antioxidant capacity of six species of Irish seaweeds and found the methanolic extracts from H. elongata to be the richest in terms of antioxidant properties. The selection of the pathogenic microbes (L. monocytogenes and S. abony) was made after discussions with the Food Safety Authority of Ireland as these were found to be the most challenging organisms for the safety of food products in Ireland. The other two (E. faecalis and P. aeruginosa) are the most widespread food spoilage microorganisms. Since the yield of the extract obtained was only 60 mg/g seaweed, growth inhibition was checked by measuring the OD by a microtitre plate-based assay rather than by the conventional spread plate method. As expected, control samples showed a rapid and prolific growth, as the populations were 9.6, 10, 14.1 and 17.2 log CFU/ml after 24 h for S. abony, L. monocytogenes, P. aeruginosa and E. faecalis, respectively (Fig. 1). The incorporation of seaweed extract resulted in variable levels of inhibition in the growth of the different organisms. Resistance to extract was not correlated with taxonomy, since E. faecalis (Gram positive) and P. aeruginosa (Gram negative) were the most sensitive to all of the different concentrations of the extract followed by S. abony and L. monocytogenes. Figure 1 shows the influence of the different concentrations of crude extracts obtained from H. elongata against the four studied organisms. The extract had a strong antagonizing effect on the food spoilage and pathogens studied, showing a remarkable dose–response relationship with an increase of the lag phase duration and decrease of the exponential growth rate. In addition, a reduction in the maximum number attained or a complete suppression of growth was observed. The addition of H. elongata extracts resulted in complete growth inhibition of all the studied organisms at the highest extract concentration (6%) used (Fig. 1). The bactericidal activity can be attributed due to the presence of phenolic compounds such as bromophenols and phlorotannins, produced by brown algae (Nagayama et al. 2002). Phenolic compounds from other plant sources have also been reported to inhibit various foodborne pathogens (Plaza et al. 2010; Kim et al. 2005; Prashanth et al. 2001). Polyphenols, such as tannins and flavonoids, are important antibacterial substances. Halogen-containing terpenoids, acetylenes and phenols have also been identified in several seaweed species as biologically active compounds having antibacterial and anti-tumoural activities (Cardozo et al. 2007; Vairappan et al. 2001; Carvalho and Roque 2000). Plaza et al. (2010) identified volatile compounds like fatty acids, alkanes, phenols and compounds such as phytol (2-hexadecen-1-ol, 3,7,11,15-tretramethyl) and neophytadiene in the ethanol extracts from Synechocystis sp. and H. elongata. These compounds have been already proposed to have antimicrobial activity (Alagić et al. 2006).

Fig. 1
figure 1

Growth kinetics of food spoilage and pathogenic bacteria in presence of different concentrations of seaweed extract: a L. monocytogenes, b P. aeruginosa, c E. faecalis and d S. abony. Different concentrations of extract used: diamond 6%, black triangle 3%, circle 1.5%, square 0.75%, asterisk 0.375% and white triangle 0%

Full size image

The cell density in the presence of 6% extracts upon completion of the assay (24 h) was lower for all four bacteria than the initial bacterial density. Similar effect was seen in earlier studies on the effect of seaweed extracts on growth of marine and fish pathogenic extracts (Dubber and Harder 2008) where it had been anticipated the reason for this frequently observed result could be associated with the complete disappearance of the bacterial DNA upon incubation with algal extract components. Therefore, the extracts presumably evoked not only a bacteriostatic but also bacteriolytic mode of action. Studies by Ceylan et al. (1998) revealed that addition of 1% spice (garlic, clove and cinnamon) to salami mixed with starter culture and Escherichia coli O157:H7 resulted in slight reduction of the pathogen; however, the addition of 7.5% garlic and clove killed 99% of the pathogen. Similar results were obtained in the present study as well wherein addition of 6% extracts resulted in growth inhibition and extract concentrations lesser than that caused a reduction in the cell numbers.

As the extract concentration was serially diluted, the inactivation effect was reduced. Although the addition of extracts at a concentration of 3% did not result in a complete inactivation of bacteria, the growth kinetics was highly altered. An increase in the cell number, after 24 h of incubation with 3% extract, for each of the four bacteria was in the range of 0.21–0.43 log CFU/ml. A lag phase much longer than the control was observed, and the specific growth rate was significantly reduced for all the organisms. At the same time, a reduction of 98% (E. faecalis), 97% (P. aeruginosa), 93% (S. abony) and 91% (L. monocytogenes) in the stationary level growth was observed as compared to the control. There was a significant difference (p < 0.05) in the stationary phase growth (24 h) of P. aeruginosa, E. faecalis, L. monocytogenes and S. abony upon the addition of 3%, 1.5% and 0.75% extract. The increase in the cell number upon the addition of 1.5% extract, after 24 h incubation, was in the range of 0.6–0.8 log CFU/ml. Thus, it can be said that addition of these concentrations of extracts resulted in an extended lag phase. Despite the fact that reducing the extract concentration to 0.75% resulted in a lag phase similar to that of the control, the specific growth rate was highly suppressed. Hence, an increase of 1.5-, 1.6-, 2.2- and 2.3-fold in the stationary phase growth of S. abony, L. monocytogenes, P. aeruginosa and E. faecalis, respectively, was seen in the control (0% extract) as compared to samples containing 0.75% extract. Thus, the bacteria started to grow at almost similar times, but the presence of the extract suppressed the maximum cell number attained. This was further evident from a slight increase of 1–1.9 log CFU/ml for each of the four bacteria after 24 h growth. Reducing the extract concentration further to 0.35% and 0.18% resulted in growth patterns very similar to the control in case of L. monocytogenes and S. abony.

The present study utilized methanol as a solvent for extraction of compounds responsible for the observed effect. Studies are also available wherein different solvents have been utilized for the extraction of biologically active compounds from seaweeds. Earlier reports on the effectiveness of extraction methods evidenced that methanol extraction yielded higher antimicrobial activity than other solvents such as n-hexane and ethyl acetate (Sastry and Rao 1994; Paul and Puglisi 2004). It is well documented that using organic solvents always provides a higher efficiency in extracting compounds for antimicrobial activities as compared to water-based methods (Masuda et al. 1997; Lima-Filho et al. 2002).

A significant finding of the present study was the potency of the extract against Gram-negative bacteria (P. aeruginosa). Gram-negative bacteria are more resistant pathogens compared to the Gram-positive bacteria. They have an additional lipopolysaccharide layer on the outer surface which prevents certain drugs and antibiotics from penetrating the cell thus accounting for the high resistance of these bacteria to antibiotics (Dowling 2004). Therefore, the present study brings out a new insight towards the development of antimicrobial agents against Gram-negative bacteria from seaweeds. In recent years, the use of non-thermal techniques for preservation of food has been gaining importance. The use of ozone, irradiation or ultrafiltration is better than thermal sterilization but can have a detrimental impact on the bioactive compounds present in foods if used at higher levels (Tiwari et al. 2009; Alighourchi et al. 2008; Zárate-Rodríguez et al. 2000). In this regard, the use of natural antimicrobials in foods to prevent spoilage might also provide the additional benefit of preserving the bioactive properties of foods. Thus, the addition of extracts from seaweeds can not only impart microbiological safety to food products as they are rich in bioactives (Cox et al. 2009) but can also provide foods with antioxidants in order to prevent oxidative spoilage.

Comparison of Kinetic Models

Generally, the models that can be used for describing the kinetics of survival curves are either empirical or based on biological assumptions. Three primary growth models (modified Gompertz (empirical), logistic (empirical) and Baranyi–Roberts model (semi-mechanistic model)) were used to analyse the delay or inhibition of growth against the four different organisms. In most of the cases, the R 2 values for all the models were greater than 0.9 (except when extract at a concentration of 1.5% was added to P. aeruginosa), indicating a good fit to the experimental data. Examples of the fit of the three models to the inactivation of the four organisms at an extract concentration of 0.75% and 0.375% are depicted in Fig. 2. All the parameters obtained for the three mathematical models were directly related to the extract concentration. Analyses of variance indicated that the maximum specific growth rate, μ max, was significantly reduced (p < 0.05) with increasing extract concentration suggesting that the cells became more sensitive (Table 1). The estimated values for the lag phase for all the three models tended to increase as the extract concentration increased. In individual model analysis, it was found that all the three models were capable of fitting the experimental data very reasonably and produced almost similar curves; however, no model could produce consistently best fit to all the growth curves analysed. RMSE (Table 2) was used as a statistical measure for comparison of the experimental and model simulated values. There was no significant difference between the RMSE (p > 0.05) for the three models. One way to discriminate the goodness of fit among different models is to compare them statistically Zwietering et al. (1990). The models were statistically validated with the use of F test. The calculated F values were lower than the F table values, indicating that there was no significant difference in the goodness of fit between the three models, except for 6% extract concentration against P. aeruginosa. In this case, fitting by Gompertz and logistic were found to be better than Baranyi–Roberts model (data not shown). Although the performance of the modified Gompertz model was better than the logistic and the Baranyi–Roberts model, the use of one primary model or the other in case of inactivation curves should be guided by specific requirements (Geeraerd et al. 1997). However, based on the RMSE and residual sum of squares (RSS) values, it can be said that all the three model equations were effective for describing sigmoidal curves as previously reported by Xiong et al. (1999), who modelled the thermal inactivation of L. monocytogenes. The model kinetic parameters such as the λ, μ and N max estimated using the experimental data with the modified Gompertz, logistic and Baranyi–Roberts model are summarized in Table 2. As there was no difference in the goodness of fit of the three models, the parameters of modified Gompertz were further analysed, as an example, to study the effect of extract concentration on the bacterial growth. The most prominent effect of the extract was an increase in the lag phase duration. The increase in lag phase due to the addition of extract at a concentration of 3% was more than 3-fold for L. monocytogenes and P. aeruginosa as compared to the control. A delay in, or inhibition of, microbial growth is particularly useful in terms of food safety. The extension of the lag phase is probably the most widely used parameter to describe the inhibitory effects of antimicrobial compounds, and a slight delay in the lag phase may have an important influence on the shelf life of food products. Thus, the prolonging of lag phase could be used as an appropriate index for evaluation of the activity of the antimicrobial compounds. In the present study, a linear positive correlation (R 2 ranging between 0.912 and 0.986) was obtained between the lag phase and the concentration of seaweed extract for the different organisms (Fig. 3). The seaweed extracts were potent even at an extract concentration as low as 0.75% resulting in 42% and 20% increase in the lag phase of L. monocytogenes and E. faecalis, respectively, as compared to the control. Extract concentration of 3% increased the lag phase in a range of 65% to 81% for all of the organisms. In an early report on microbial growth modelling of fresh filled pasta stored at different temperatures by Giannuzzi (1998), it was observed that the ratio of specific growth rate to generation time was nearly constant which suggests a linear relationship between lag phase and the reciprocal of the maximum specific growth rate. Similar observations were confirmed in the present study during the inactivation of the four organisms at different extract concentrations.

Fig. 2
figure 2

Fitting of the three models to the inactivation of the four organisms by extract at a concentration of 0.75% and 0.375%. a L. monocytogenes, b P. aeruginosa, c E. faecalis and d S. abony. Different concentrations of extract used: triangle 0.75% and diamond 0.375%. Different models: dotted line modified Gompertz model equation, dashed line Baranyi–Roberts model equation and dash-dotted line logistic model equation. Points represent experimental data

Full size image
Table 1 Estimations of the kinetic parameters using the logistic, modified Gompertz and the Baranyi-Roberts models
Full size table
Table 2 Values of the statistical indices, RMSE and RSS, for the three models against four different bacteria at six different extract concentrations
Full size table
Fig. 3
figure 3

Relation between lag phase and concentration of extract against the four different organisms (diamond—L. monocytogenes, circle—S. abony, square—P. aeruginosa and triangle—E. faecalis)

Full size image

The concentration of extract also had a strong effect on the maximum specific growth rate. A reduction of 99% and 96.8% was observed in the maximum specific growth rate, as compared to control at 3% extract concentration (E. faecalis and P. aeruginosa, respectively).

It has been reported in the literature that flavonoids, polysaccharides, sesquiterpenes and phlorotannins can be obtained from seaweeds. These active ingredients produce varied pharmacological effects such as anti-angiogenic, anti-inflammation, disinfection and anti-tumour. The application of the extracts of H. elongata in food industry may contribute to such pharmacological activities as food anti-oxidation, health care and in addition as food nutrient. Therefore, these extracts could be applied as natural additives with extensive market prospect.

Conclusion

H. elongata can be considered as a promising marine plant in the development of bioactive ingredients for functional foods, nutraceuticals and other applications. The extracts showed an evident antimicrobial effect against the microorganisms used in the present study in a dose-dependent manner. Complete growth inactivation of all of the studied organisms was observed at a concentration of 6%. Addition of extracts at a concentration less than that resulted in an extension of the lag phase and significantly reduced maximum specific growth rate. A reduction of 91–98% in the stationary level growth as compared to the control was also observed. The findings suggest that seaweed extracts have a good potential as natural antibacterial substances in food preservation. It might be possible that high concentrations of these extracts may adversely affect the organoleptic properties of food; however, lower concentrations may be sufficient for food safety in situations where bacterial load is low.