Microbial inactivation by new technologies of food preservation

Abstract

The increasing consumer demand for ‘fresh-like’ foods has led to much research effort in the last 20 years to develop new mild methods for food preservation. Nonthermal methods allow micro-organisms to be inactivated at sublethal temperatures thus better preserving the sensory, nutritional and functional properties of foods. The aim of this review is to provide an overview of the microbiological aspects of the most relevant nonthermal technologies for microbial inactivation currently under study, including irradiation, high hydrostatic pressure, pulsed electric field and ultrasound under pressure. Topics covered are the mechanisms of inactivation, sensitivity of different microbial groups and factors affecting it and kinetics of inactivation. Micro-organisms are the main agents responsible for food spoilage and food poisoning and therefore food preservation procedures are targeted towards them. Food preservation methods currently used by the industry rely either on the inhibition of microbial growth or on microbial inactivation. Methods which prevent or slow down microbial growth cannot completely assure food safety, as their efficacy depends on the environmental conditions such as, for instance, the maintenance of the chill chain. Thermal treatment is the most widely used procedure for microbial inactivation in foods. However, heat causes unwanted side-effects in the sensory, nutritional and functional properties of food. This limitation together with increasing consumer demand for fresh-like foods has promoted the development of alternative methods for microbial inactivation, among which ionizing irradiation, ultrasound under pressure, high hydrostatic pressure (HHP) and pulsed electric field (PEF) are attracting much interest. The irradiation process involves the application of electromagnetic waves or electrons to foods. Radiation sources are either gamma rays from cobalt-60, electron beams or X-rays, and the amount of irradiation absorbed by a food is measured in kGy (1 Gy = 1 J kg−1). Commercial application of ionizing radiation (IR) treatment on foods was started at the beginning of the 1980s, but its success has been prevented by consumer concerns. Nowadays, social perception of IR is changing and this technology is being re-examined. Ultrasound is defined as sound waves with frequencies above the threshold for human hearing (>16 kHz). Although ultrasound was initially discarded for food preservation because of its weak lethal action, the application of an external hydrostatic pressure of up to 600 kPa [manosonication (MS)] increases substantially the lethality of the treatment. In addition, a combination of MS with temperature [manothermosonication (MTS)] has been proposed (Raso et al. 1998a). The HHP involved the application of pressures from 100 to 1000 MPa. The first studies on the lethal effect of HHP were conducted at the end of the 19th century, but it has been in recent years when commercial applications of this procedure have started. Finally, PEF technology consists in the application of short duration (1–100 μs) high electric field pulses (10–50 kV cm−1) to a food placed between two electrodes. Like ultrasound under pressure, PEF technology is not yet being used to preserve food commercially. This review discusses the state of the art of the research of the microbial aspects of the four technologies described above. The successful implementation of a novel technology for food preservation relies on the progress in the field of mechanisms of inactivation. An adequate knowledge of the physiological behaviour of micro-organisms towards inactivation agents is essential for the development of safe foods. It is necessary for understanding the effect of environmental factors on resistance and also for identifying critical factors. It would help to interpret kinetics of inactivation and to develop mathematical models based on parameters with a biological meaning and therefore more useful and able to predict microbial inactivation in a wider range of conditions. A better knowledge of the effect of the preservation agents on the micro-organisms would lead to a more rational design of processes. Cell death has been associated with either structural damage or physiological dysfunctions. Among structural damage, disruption of the envelopes, DNA conformational changes, ribosome alterations or protein aggregation are the most frequently described. Also physiological disorders, such as membrane selective permeability alterations or loss of function of key enzymes have been proposed as events leading to cell death (Gould 1989). Perhaps the most important difficulties that researchers encounter in this area is that several of these lesions may occur simultaneously when the cells are subjected to an agent and it is therefore difficult to attribute the loss of viability of the cell to a single event. Nevertheless, it must be kept in mind that also multitarget inactivation is feasible, this being the addition of several lesions that together cause death. It is also possible that the key target is only affected when a secondary structure is previously damaged. For instance, heat causes membrane damage, loss of nutrients and ions, ribosome aggregation, DNA strand breaks, inactivation of essential enzymes, protein coagulation, etc. (Gould 1989). In other words, almost every cellular structure is somehow affected by elevated temperatures, and it is very difficult to discern which events are leading to cell death. In fact, this is only possible if a direct relationship between the degree of inactivation and the degree of modification of a given target under different environmental conditions is found. Regarding novel inactivation technologies, progress on inactivation mechanisms research is heterogeneous. In this way, sound hypotheses have been put forward to describe the effect of IR and ultrasound on micro-organisms, but much research is still needed to completely explain the way PEF and HHP can inactivate bacterial cells. Table 1 summarizes the relevant events related to bacterial inactivation by the four technologies. It is well established that the critical target for irradiation is the chromosome (Moseley 1989). The effects of ionizing irradiation on bacterial cells are classified as direct and indirect. Direct actions comprise the events caused by the absorption of radiation energy by the target molecules, whereas indirect actions are those derived from the interaction between the reactive species formed by the radiolysis of water, such as the hydroxyl radical, and the target molecules. The hydroxyl radical OH• is able to react with the sugar-phosphate backbone of the DNA chain giving rise to the elimination of hydrogen atoms from the sugar. This causes the scission of the phosphate ester bonds and subsequent appearance of single strand breaks. Double strand breaks occur when two single strand breaks take place in each chain of the double helix at a close distance. Bases are also attacked by the free radicals generated by radiolysis, but it is not clear whether this is relevant to cell death (Moseley 1989). The mechanism of inactivation of bacterial cells by ultrasound under pressure has also been described. Most authors agree that the cavitation phenomenon is responsible for the lethal effects of ultrasound (Kinsloe et al. 1954; Raso et al. 1998a). When bubbles implode under an intense ultrasonic field, very high pressures and temperatures are generated, and consequently strong mechanical forces and free radicals are formed (Suslick 1990). Free radicals could therefore inactivate bacterial cells in a similar mode as that described for IR. However, experimental data using free radical scavengers have lead to the conclusion that the possible effect of free radicals is negligible in comparison with that of the strong mechanical effects generated by cavitation (Allison et al. 1996; Raso et al. 1998a). Raso et al. (1998a) studied the effect of equivalent heat, MS and MTS treatments (99% of inactivation) on the degree of cell disruption evaluated through phase contrast microscopy and they observed that whereas heat-treated cells maintained full cellular integrity, MS treated cells were completely broken. MTS treated cells showed a medium degree of disruption. These results confirmed that ultrasound inactivate microbial cells through envelope breakdown in an ‘all or nothing’ type phenomenon. There are however some unclear aspects still not solved regarding the mechanism of action of ultrasound. In some experimental conditions, and with some bacterial species, a synergistic effect of MS and heat (MTS) has been observed. This is the case of Enterococcus faecium (Pagán et al. 1999c), heat-shocked cells of Listeria monocytogenes (Pagán et al. 1999b) and cells suspended in a low water activity media (Álvarez et al. 2003b). The reasons for the increased sensitivity of these cells to a combined MS-heat treatment are still not known. It has been suggested that moderately elevated temperatures (55–60°C) would cause a weakening effect on cell envelopes, facilitating the mechanical disruption of the cell by ultrasonic waves (Álvarez et al. 2003b). This weakening effect would have no relevance for thermosensitive cells, as they are killed by heat at relatively low temperatures. In addition, a synergistic effect of MS and heat has been described for bacterial spores (Raso et al. 1998b). Ultrasonic treatments cause the release of some low molecular weight polypeptides and dipicolinic acid from the spore (Palacios et al. 1991). It has also been found that MS treatments sensitize spores of Bacillus subtilis to lysozyme (Raso et al. 1998b). Therefore, it has been suggested that ultrasonic waves could damage the external layers of the spore, facilitating its rehydration and consequently reducing its extreme heat resistance. In contrast to the clear mechanisms of inactivation proposed for irradiation and ultrasound, a much more complicated picture emerges for high hydrostatic pressure inactivation. Much research has been carried out in this matter, but there is still controversy about the key target leading to cell death by high pressure. Initial investigations on mechanism of inactivation of HHP suggested the cytoplasmic membrane as the key target (Cheftel 1995; Smelt 1998; Patterson 1999), and most researchers still agree with that hypothesis despite some apparent inconsistencies having been reported. Evidences of bacterial membranes being a target for HHP inactivation are clear. High pressure causes tighter packing of the acyl chains within the phospholipid bilayer of membranes and promotes membrane transition from liquid crystalline to gel phase, in a similar way as a temperature downshift (MacDonald 1993). Although phase transition of membrane lipids is not necessarily lethal to bacteria, it has been demonstrated that the composition and state of the bacterial cell membrane prior to pressure treatment affect bacterial resistance to HHP (Casadei et al. 2002). Cells with a more fluid membrane, i.e. with a higher degree of unsaturation are more barotolerant (Casadei et al. 2002). It is not clear how a more fluid membrane renders a more resistant cell to HHP. The pressure at which phase transition occurs would be higher in cells with a more fluid membrane, but it is not known in which circumstances, if any, cell damage is linked to phase transition. Damage to the cytoplasmic membrane after pressurization has also been repeatedly reported, through loss of osmotic responsiveness (Pagán and Mackey 2000; Mañas and Mackey 2004), uptake of vital dyes (Shigehisa et al. 1991; Benito et al. 1999; Pagán and Mackey 2000; Mañas and Mackey 2004), loss of intracellular material, and formations of buds and vesicles of lipidic origin (Perrier-Cornet et al. 1999; Ritz et al. 2000; Mañas and Mackey 2004). The loss of function of some proteins including the F1–F0 ATPase or multidrug efflux pumps has also been described (Smelt et al. 1994; Wouters et al. 1998; Molina-Hoppner et al. 2004). Even a direct relationship between loss of membrane integrity and loss of viability has been found for pressure-treated exponentially growing cells (Pagán and Mackey 2000; Mañas and Mackey 2004). However, it has also been demonstrated that both outer and cytoplasmic membrane permeabilization is transient to some extent (Hauben et al. 1996; Pagán and Mackey 2000; Ganzle and Vogel 2001; Mañas and Mackey 2004) and that pressure-treated stationary-phase cells of Escherichia coli may maintain a physically intact cytoplasmic membrane upon decompression even in dead cells (Pagán and Mackey 2000; Mañas and Mackey 2004). Therefore, other structures inside the cell have also been proposed as potential key targets for inactivation by HHP. Some authors have reported similarities between cell inactivation and protein denaturation kinetics by HHP (Sonoike et al. 1992), and changes in the conformation of the nucleoid, ribosomes and cytoplasmic protein have been described (Mackey et al. 1994; Niven et al. 1999; Mañas and Mackey 2004). Niven et al. (1999) found a direct relationship between loss of viability in E. coli and ribosome damage, evaluated by differential scanning calorimetry. Further incubation of treated cells in a magnesium-rich medium, which is known to have a stabilizing effect on ribosome structure, aided the ribosomes to recover the initial conformation. The authors concluded that other factors together with ribosome initial destabilization accounted for cell death and suggested that the loss of essential ions like magnesium through a damaged membrane could be the event triggering ribosome destabilization. Moreover, Perrier-Cornet et al. (1999) correlated loss of viable cells of yeasts of the genus Saccharomyces with the loss of internal solutes caused by the pressure-induced cell permeabilization during HHP treatment. It seems clear that some of these cellular lesions like DNA and protein condensation are not necessarily lethal (Mañas and Mackey 2004) and are repairable if the cell keeps a functional membrane and the environmental conditions are suitable. In conclusion, HHP inactivation seems to be multitarget in nature. Membrane is a key target, but in some cases additional damaging events such as extensive solute loss during pressurization, protein coagulation, key enzyme inactivation and ribosome conformational changes, together with impaired recovery mechanisms, seem also needed to kill bacteria. Bacterial spores are extremely resistant to HHP, being able to withstand up to 1000 MPa for long treatment times, unless they are in the germinated state (Cheftel 1995). Pressure itself at a moderate level induces spore germination (Gould and Sale 1970). This has been the basis for the design of a cyclic combined treatment in which spores are induced to germinate in a first step and inactivated in a second step, generally by a combination of mild heat and pressure (Raso and Barbosa-Cánovas 2003). Membrane structural or functional alteration is generally accepted as the cause of cell death by PEF. Sale and Hamilton (1967) demonstrated that the inactivation was the result of the direct effect of PEF on the membrane (electroporation) rather than because of the temperature increase or the electrolysis products. Electroporation can be defined as the formation of pores in cells and organelles, and the most accepted theory to explain it is that proposed by Zimmermann et al. (1974). Zimmermann compares the cell membrane to a capacitor. Free charges tend to accumulate in the inner and outer surface of the membrane generating a transmembrane potential of about 10 mV. When an external electric field is applied, as in PEF treatment, a higher amount of free charges of opposite charge accumulate at both membrane surfaces, resulting in compression of the membrane. When the external electric field exceeds a critical value or threshold, the membrane is unable to withstand the electrocompression and pores are formed. The size and amount of pores depend on the electric field strength and the duration of the treatment. Alternative theories propose that the permeabilization is the consequence of dipolar reorientation of the membrane phospholipids under an electric field (Tsong 1991). Tsong (1991) has also suggested that the formation of hydrophilic pores would lead to a localized Joule heating phenomenon that could be responsible for the denaturation of proteins and phase changes in the membrane. Studies on electron microscopy of several bacteria and yeasts have shown morphological alterations like surface roughness, disruption of organelles, ruptures in the membrane, etc. (Jayaram and Castle 1992; Pothakamury et al. 1997; Dutreux et al. 2000). However, no correspondence between the frequencies of appearance of morphological alterations with loss of viability has been proven (Aronsson and Rönner 2001). Many attempts have been made to establish whether there is a relationship between membrane damage and microbial inactivation by PEF. Using vital staining and detection of UV-absorbing material leakage various researchers have shown the occurrence of membrane permeabilization in relation to cell death (Hamilton and Sale 1967; Simpson et al. 1999; Wouters et al. 2001b). Wouters et al. (2001b) have described a linear relationship between the percentage of permeabilized cells and the intensity of the PEF treatment, supporting the hypothesis of membrane permeabilization being the cause of cell inactivation. An interesting and almost nonexplored aspect of PEF treatments is the occurrence of reversible pores. The same way as it happens with high pressure (Pagán and Mackey 2000), a proportion of cell membranes could become leaky during PEF treatment but reseal to a certain extent after it. Experiments carried out in our laboratory using the addition of propidium iodide to the treatment medium as a marker of nonpermanent permeabilization of the cytoplasmic membrane, have shown that the degree of staining of Salmonella serotype Senftenberg 775 W cells was approximately twice as that observed when the propidium iodide was added after PEF treatment. These results would indicate that a percentage of cells were able to reseal their pores just after PEF treatment. Micro-organisms surviving the lethal action of preservation agents may be sublethally injured: able to repair the damage and outgrow only if the environmental conditions are suitable (Mackey 2000). The occurrence of sublethal injury has two main consequences. First, injured cells might not be detected when selective conditions are used for enumeration of survivors. This can lead to an overestimation of the lethality of the treatment. Secondly, the occurrence of sublethal injury means in practice that a delay time between treatment and outgrowth of survivors takes place. Alternatively, if repair is adequately prevented by the combination of additional preservation agents (hurdles) that interfere with cellular homeostasis maintenance, the cell might not be able to outgrow, and the inactivation level attained might be higher (Mackey 2000). Thus, detection and characterization of sublethal injury by novel preservation technologies is essential for the optimization of mild combined methods with a higher lethal effect on microbes. Among the four methods for microbial inactivation reviewed here, only in the case of ultrasound under pressure have no sublethally injured cells been detected. Survival curves of MS treatments of Gram-positive and Gram-negative cells recovered in media with sodium chloride added are virtually identical to those recovered in a nonselective medium (Pagán et al. 1999a). This indicates the total absence of repairable membrane damage and is in concordance with the hypothesis of the all or nothing lysis of the cell as the mechanism of inactivation of ultrasonic waves on bacterial cells. Sublethal or repairable injury has been detected for irradiated, pressurized and PEF-treated cells. For each agent the mechanism of inactivation is different, and so is the nature and the magnitude of the sublethal injury. Studies on sublethal cell injury and repair of irradiated cells have focused on DNA. It seems clear that the relative sensitivity of the different microbial groups depends not only on their susceptibility to the direct and indirect action of irradiation itself, but also on their capability of repairing the single and double strand breaks through several enzymatic actions (Moseley 1989). Irradiation, however, does not discriminate among molecules in a sample, and virtually all the molecules in an irradiated cell may be affected. Irradiation damage to other structures has been scarcely studied. Some authors have reported the sensitization of irradiated cells to selective media (Patterson 1989; Tartéet al. 1996; Buchanan et al. 1999) but Kim and Thayer (1996) showed that irradiation did not induce membrane damage as detected by vital dye staining. It is not clear whether the secondary hurdle added (salt, acid, CO2) would exert its inhibitory action either towards the damaged DNA as suggested by Kim and Thayer (1996) with heat, interfering with DNA repair systems or at any other level. Bacterial membranes are the main target for sublethal injury in HHP treated cells. Permeabilization of the outer and cytoplasmic membrane has been described (Pagán and Mackey 2000; Ganzle and Vogel 2001) and combined treatments of HHP and antimicrobial peptides in the pressurizing medium, such as lysozyme, nisin, pediocin AcH, lacticin, lactoferrin and lactoferricin with increased efficacy for both Gram-positive and Gram-negative micro-organisms have been proposed (Hauben et al. 1996; Kalchayanand et al. 1998). However the permeabilization of the outer membrane of Gram-negative cells is transient and a fully functional membrane is recovered automatically shortly after decompression (Chilton et al. 2001). On the contrary, cytoplasmic membrane damage repair is highly demanding, and requires energy, RNA and protein synthesis (Chilton et al. 2001). An intact cytoplasmic membrane is essential for the maintenance of the homeostasis under unfavourable environmental conditions, and thus, from a practical point of view, combinations of HHP with alkali, salt or acidic environment also attain a higher lethal effect on micro-organisms (García-Graells et al. 1998; Pagán et al. 2001; Wuytack et al. 2003; Sherry et al. 2004). The occurrence of sublethal cell damage in PEF-treated cells is a matter of controversy. Most authors have not observed the occurrence of sublethal injury using the selective medium plating technique (Simpson et al. 1999; Dutreux et al. 2000; Ulmer et al. 2002; Wuytack et al. 2003) so it was generally accepted that bacterial inactivation by PEF was an all-or-nothing effect. However, recent results obtained in our laboratory (García et al. 2003) have shown the occurrence of sublethal cytoplasmic membrane damage in a large proportion of the population of Gram-negative cells (≥99·9% of survivors) estimated by differential plating. Unpublished results of our research group indicate that the discrepancies in published data may arise from the fact that the occurrence of sublethal membrane damage by PEF treatments depends on the pH of the suspending medium and on the bacterial species. In this way, several Gram-negative bacteria showed a higher resistance to PEF at acidic pH (when compared with neutral pH) that was correlated to the capability to repair the cytoplasmic membrane, which was extensively damaged. For Gram-positive cells the opposite was true, the greater resistance to PEF, the occurrence of membrane damage and the subsequent repair ability was detected at neutral pH. At acidic pH the cells became more sensitive and irreversibly damaged. It is worthy of note that PEF-sublethally injured cells stored under acidic conditions lost viability during storage time. This means, from a practical point of view, that if adequate post-treatment holding conditions are selected, the intensity of treatments, both for HHP and PEF, could be diminished without affecting the microbial quality of the product. Alternatively, a higher degree of safety could be attained. We have also detected the occurrence of sublethal damage to the outer membrane in PEF-treated Psychrobacter immobilis by differential plating techniques (P. Mañas, unpublished data). The role of the outer membrane in PEF inactivation needs further attention. These observations provide new useful data that contribute to the understanding of mechanisms of membrane electroporation in bacterial cells. They also contribute to clarify the environmental circumstances under which PEF might act synergistically with other hurdles for food preservation. Further work is in progress. It is well known that micro-organisms can develop adaptive responses and resistances when exposed to sublethal stresses (Abee and Wouters 1999), which may have serious implications for food safety. In the last 15 years much research interest has been directed towards the elucidation of bacterial stress adaptation mechanisms and gene regulation behind it. The modification of sigma factors (σ) bound to core RNA polymerase, conferring promoter specificity, is possibly the most important regulatory mechanism in bacterial cells (Abee and Wouters 1999). Sigma factor σs regulates, in Gram-negative bacteria, the transcription of more than 50 genes involved in resistance to osmotic, heat, oxidative and acid stress, among others (Huisman et al. 1996). The induction of this sigma factor occurs in response to starvation, generally when cells enter the stationary phase of growth, and also when exponentially growing cells are subjected to stresses other than starvation (Dodd and Aldsworth 2002). In Gram-positive bacteria (B. subtilis, L. monocytogenes and Staphylococcus aureus) an alternative sigma factor with equivalent physiological functions has been described (sigB) (Abee and Wouters 1999; Hill et al. 2002). Therefore, it seems that a parallel mechanism for the acquisition of multiple stress resistance exists in Gram-positive and Gram-negative cells. The influence of the activation of these regulatory networks on bacterial resistance to novel preservation processes has not yet been sufficiently studied. Nevertheless, it is known that the higher resistance to HHP of stationary phase cells is partly the result of the presence of the RpoS protein in E. coli and sigB in L. monocytogenes (Robey et al. 2001; Wemekamp-Kamphuis et al. 2004). A range of morphological and physiological changes dependent on RpoS expression has been described for Salmonella and Escherichia cells that might possibly account for the increase in pressure resistance (Huisman et al. 1996). It can be foreseen that these changes might also have an influence on bacterial resistance to PEF, ultrasound and irradiation, but up to now no data are available on the influence of the RpoS/sigB regulon on the resistance to these technologies. Sigma factor σ32, encoded by the rpoH gene, controls the heat shock response, which consists of a rapid and transient overexpression of chaperons and proteases. The application of sublethal HHP treatments promotes the expression of several heat shock proteins (Welch et al. 1993), suggesting a direct role of the heat shock response in HHP resistance. In fact, it has been reported that a previous heat shock may protect cells against HHP (Pagán and Mackey 2000; Aersten et al. 2004). In addition, Aersten et al. (2004) have described that the basal expression of several heat shock proteins like dnaK, lon or clp is increased in pressure-resistant mutants (Hauben et al. 1997). The influence of a sublethal heat shock on PEF resistance has been scarcely investigated but Evrendilek and Zhang (2003) have shown that survival seems to be increased. Preliminary results obtained in our laboratory suggest that the ability to recover from PEF damage is also increased. More research is needed in this field. The application of a heat shock does not protect bacteria to a subsequent MS treatment (Pagán et al. 1999b). This indicates that the possible changes that heat shock may induce in cell envelopes, if any, are not relevant to ultrasound resistance. Although scattered information is available about the influence of other types of stresses on bacterial resistance to novel preservation methods, from published data it can be deduced that in most cases cross resistance is induced. Palhano et al. (2004) have described how the exposure to hydrogen peroxide, ethanol and cold-shock induces baroresistance in Saccharomyces cerevisie. Buchanan et al. (2004) have reported the increase of irradiation resistance of E. coli O157:H7 when cells were previously acid adapted. Likewise, acid, cold and heat adaptation also seems to protect this micro-organism to PEF treatments (Evrendilek and Zhang 2003). In summary, adaptation of micro-organisms to adverse environmental conditions during processing poses a risk that should not be underestimated. Extensive research is still needed to characterize the physiology and genetics of microbial stress responses involved in survival to food preservation processes. Microbial inactivation by irradiation, ultrasound under pressure, HHP and PEF has been found to depend on many factors. Effective comparison of data published in literature is hampered by the diversity of equipments and experimental conditions employed by the different authors. Nevertheless, this section tries to give an overview on the most relevant factors affecting resistance to novel technologies. The factors are classified into three groups: process parameters, microbial characteristics and product parameters. A summary of the most important factors for each technology is given in Table 2. Some process parameters are intrinsic to each technology and no general conclusions can be drawn. For