foods Article Evaluation of Continuous UVC Treatments and its Combination with UHPH on Spores of Bacillus subtilis in Whole and Skim Milk María Martinez-Garcia 1 , Jezer N. Sauceda-Gálvez 1 , Idoia Codina-Torrella 1 , Mª Manuela Hernández-Herrero 1 , Ramón Gervilla 2 and Artur X. Roig-Sagués 1, * 1 Centre d’Innovació, Recerca i Transfèrencia en Tecnologia dels Aliments (CIRTTA), XaRTA, TECNIO-CERTA, MALTA-Consolider Team, Departament de Ciència Animal i dels Aliments, Facultat de Veterinària, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; maria.martinez.garcia@uab.cat (M.M.-G.); jezernoe.sauceda@e-campus.uab.cat (J.N.S.-G.); idoia.codina@uab.cat (I.C.-T.); manuela.hernandez@uab.cat (M.M.H.-H.) 2 SPTA-Servei Planta Tecnologia Aliments, Universitat Autònoma de Barcelona, c/ de l‘Hospital S/N, 08193 Bellaterra (Barcelona), Spain; ramon.gervilla@uab.cat * Correspondence: arturxavier.roig@uab.cat; Tel.: +34-935811960  Received: 14 September 2019; Accepted: 24 October 2019; Published: 2 November 2019  Abstract: The aim of this study was to evaluate the effectiveness of different UVC treatments, alone or in combination with ultra-high pressure homogenization (UHPH) on Bacillus subtilis spores in milk. Spores of B. subtilis (CECT4002) were inoculated in whole and skim milk to an initial concentration about 6 log CFU/mL. Milk was subjected to different ultraviolet radiation treatments at 254 nm (UVC) using a concentric tubular reactor in a dose ranging from 10 to 160 J/mL. Different number of passes were used to adjust the final dose received by the matrix. In general, increasing the number of passes (defined as number of entries to the tunnel-NET) increased the inactivation of spores of B. subtilis. The best lethality results (above 4 Log CFU/mL) were obtained by applying doses from 100 J/mL with several NET. When the same doses were achieved with a single pass lethality in most cases did not exceed 1 log CFU/mL. Increasing the NET also increased the likelihood for the spores to remain longer in the effective distance from the UVC source, estimated as 0.02 mm for whole milk and 0.06 mm for skim milk. Combination of UHPH and UVC did not clearly increase the efficiency of a single UVC treatment, and a lower lethality was even observed in some cases. UHPH treatments increased the turbidity and absorption coefficient (254 nm) of both whole and skim milk. Keywords: UVC; UHPH; milk; Bacillus subtilis 1. Introduction Short-wave ultraviolet radiation (UVC) is a non-thermal technology proposed as an alternative to heat pasteurization that is being investigated nowadays for the reduction of microorganisms in liquid foods that may lead to spoilage and potential health risks for the consumers [1]. The application of UVC using reactors allows processing liquid foods in a continuous regime which increased the interest on this technology because it is easy to use, has a lower cost and offers a better preservation of the nutritional content and quality aspects of foods [2–5]. UVC is lethal for most of microorganisms through the formation of pyrimidine dimers which blocks DNA duplication leading to cell death [6,7]. Moreover, unlike other non-thermal emerging technologies, UVC is capable of inactivating bacterial spores [8,9]. The microbiocidal effect of UVC has been tested in different foodstuffs, such as milk, demonstrating an effectiveness similar to thermal pasteurization [2,10–22]. However, milk presents the limiting factor Foods 2019, 8, 539; doi:10.3390/foods8110539 www.mdpi.com/journal/foods Foods 2019, 8, 539 2 of 17 of its high absorption coefficient at 254 nm (α(254)) which reduces the possibility for UVC radiation to penetrate through the matrix and reach those areas farther from the UVC source with the same intensity, underexposing some of the present microorganisms, and consequently, reducing the effectiveness of the UVC treatment [23–26]. In order to increase the microbiocidal efficiency of UVC when a continuous flow reactor is used, both the UVC radiation dose and the flow pattern (laminar flow or turbulent flow) must be correctly adjusted to reduce the distance from the UVC source and the microorganisms to make them access the area were the UVC radiation is more efficient inside the reactor [4]. Two basic designs of continuous flow UVC reactors are the most frequently described in the literature to overcome these inconveniences. Some are based on increasing the turbulence generated inside the system and consequently the probability that all parts of the matrix could get into close contact with the UVC source. The conditions for a turbulent flow are described mathematically by the Reynolds number (Re), a minimum Re of 2100 should be achieved to ensure a sufficient turbulence to guarantee a more homogeneous distribution of the residence times inside the reactor [24,27]. The other proposed strategy is to reduce as much as possible the gap between the UVC source (lamp) and the farthest particle of the matrix, reducing the path length through creating a thin film around the UVC source. However, in this system, the flow of the liquid is laminar, and that represents an inconvenience to treat matrices with high α(254). One option to overcome this inconvenience is to pass the matrix several times throughout the reactor or several connected reactors, forcing a mixing effect between each pass that increases the efficiency. A different strategy to increase the UVC treatments efficiency would be to combine them with another emerging non-thermal technology, such as ultra-high pressure homogenization (UHPH) seeking for a synergic o additive effect following the “hurdle” principle. UHPH has been proposed as a good alternative to thermal pasteurization as it reduces the load of vegetative cells of pathogen and spoilage microorganisms significantly while sparing most of the organoleptic and physicochemical characteristics of raw milk [28–30]. However, its effect on bacterial spores is much more limited [31,32]. Both strategies have been recently tested on Bacillus subtilis inoculated into phosphate-buffer saline solution (PBS), increasing the efficiency compared with both UVC and UHPH single treatments [33]. A complementary effect between these technologies was also reported, as UVC was able to inactivate microorganisms that were resistant to UHPH and vice versa. The aim of this study was to evaluate the effectiveness of UVC treatments applied in different ways, alone or in combination with UHPH on Bacillus subtilis inoculated in whole and skim milk, and exploring the potential of these technologies to obtain commercial sterile products. Mathematical models were used to determine the importance of the treatment variables on their efficacy. 2. Materials and Methods 2.1. Preparation of the Spore Suspension of B. Subtilis The strain used in this study was Bacillus subtilis CECT4002, supplied by the Spanish Type Culture Collection (CECT, University of Valencia, Valencia, Spain). A modification of the UNE EN ISO 13704:2002 procedure [34] was used to obtain the spores. Briefly, the lyophilized culture was rehydrated in 10 mL of glucose and tryptone broth (TGB: 2.5 g of yeast extract (Oxoid, Basingstoke, UK), 5 g of tryptone (Oxoid), 1 g of glucose (Sigma-Aldrich, St. Louis, USA), in 1 L of distilled water, pH adjusted to 7.2), and incubated for 24 h at 30 ◦ C. Two mL of this culture was transferred to Roux bottles containing yeast extract agar (MYA: 10 g of meat extract (Oxoid), 2 g of yeast extract (Oxoid), 15 g of agar (Oxoid) and 0.04 g of MnSO4 ·H2 O (Merck, Darmstadt, Germany) in 1 L of distilled water), which were incubated at 30 ◦ C for up to 30 days. The formed spores were collected by adding 20 mL of sterile distilled water to the Roux bottles and scraping the surface with a Digralsky stick. Spore suspensions were pooled and washed four times in 15 mL cold sterile water by centrifugation at 10,000× g for 20 min at 4 ◦ C using a Sigma 4K15 centrifuge (Sigma Laborzentrifugen GmbH, Osterode am Harz, Germany). The resulting sediment was then suspended in 30 mL of sterile distilled Foods 2019, 8, 539 3 of 17 Foods 2019, 9, x FOR PEER REVIEW 3 of 17 water and subjected to a heat treatment at 75 ◦ C for 10 min to ensure the inactivation of vegetative cells. Theam resulting spore suspension Harz, Germany). was The resulting stored at sediment 4 ◦then was C until use. One suspended in 30mL mLof ofthe spore sterile suspension distilled water was and subjected to a heat treatment at 75 °C for 10 min to ensure the inactivation used to inoculate 1 L of milk to guarantee a minimum initial load close to 10 CFU/mL. 6 of vegetative cells. The resulting spore suspension was stored at 4 °C until use. One mL of the spore suspension was used to 2.2. Typesinoculate 1 L of milk to guarantee a minimum initial load close to 106 CFU/mL. of Milk Used Two2.2. types Typesofoflocally Milk Usedpurchased UHT cow milk were used to carry out this study: whole milk (3.5% Fat) andTwo skim milk types of (<0.2% Fat). TheUHT locally purchased characteristics of the cow milk were matrices used to carrythat maystudy: out this influence wholethe efficacy milk of the treatments (3.5% Fat)were determined and skim withFat). milk (<0.2% theThe following methods. characteristics of the matrices that may influence the efficacy of the treatments were determined with the following methods. 2.2.1. Absorption Coefficient at 254 nm (α(254)) 2.2.1. Absorption Coefficient at 254 nm (α(254)) It was determined using a Nanophotometer Pearl model spectrophotometer (IMPLEN GmbH, It was determined using a Nanophotometer Pearl model spectrophotometer (IMPLEN GmbH, München, Germany) with 1 cm quartz cuvettes (Fisher Scientific, Hanover, IL, USA). Sample was diluted München, Germany) with 1 cm quartz cuvettes (Fisher Scientific, Hanover, IL, USA). Sample was 1000 folddiluted with distilled water 1000 fold withindistilled order to fit into water the reliable in order reading to fit into range ofreading the reliable the spectrophotometer. range of the spectrophotometer. 2.2.2. Turbidity 2.2.2. Turbidity It was measured with a portable EPA 2100Q turbidimeter (HACH, Hospitalet de Llobregat, Spain). It was measured Sample was diluted 1000 foldwith witha portable distilledEPA 2100Q water turbidimeter in order (HACH, to fit into Hospitalet the reliable de Llobregat, reading range of the Spain). Sample was diluted 1000 fold with distilled water in order to fit into the reliable reading range turbidimeter. Results were expressed as Nephelometric Turbidity Units (NTU). of the turbidimeter. Results were expressed as Nephelometric Turbidity Units (NTU). 2.3. Application of UVC Radiation Treatments 2.3. Application of UVC Radiation Treatments 2.3.1. UVC Reactor 2.3.1. Features UVC Reactor Features A UVC reactor (UV-Therm, A UVC reactor Ypsicon (UV-Therm, S.L., Ypsicon Barcelona, S.L., Barcelona,Spain; European Spain; European Patent Patent EP 2965766-A1) EP 2965766-A1) was was used in this usedsurvey. The reactor in this survey. waswas The reactor fed fed by bya Flowmaster a Flowmaster FMT300 peristaltic FMT300 peristaltic pumppump (ISMATEC (ISMATEC Lab. Lab. GmbH, Wertheim-Mondfeld, GmbH, Wertheim-Mondfeld, Germany) Germany) througha asilicone through silicone pipe. pipe.ItIthas hasa capacity of 673 a capacity of mL 673ofmLtotal of total volume and consists of two low pressure mercury UVC lamps (UV-Consulting Peschl Spain, volume and consists of two low pressure mercury UVC lamps (UV-Consulting Peschl2 Spain, Castellón, Castellón, Spain), with a total electrical power of 55 W and an irradiance of 41 mW/cm . Each of the Spain), with a total electrical power of 55 W and an irradiance of 41 mW/cm2 . Each of the lamps is lamps is protected by a 2 mm thick quartz tube (UV-Consulting Peschl, Geldo, Spain), leaving a radial protectedspace by a to2 mm thick circulate forquartz the foodtube of 1(UV-Consulting Peschl,ofGeldo, mm. The internal section the UVC Spain), reactorleaving consists aofradial space to different circulate concentric for the food of 1 mm. cylinders The (Figure 1): internal sectionAofand between sections the UVC B air reactor circulates to consists of different cool the UVC concentric lamp; between cylinderssections (FigureB1):andbetween C the food matrix circulates sections A and Binaira layer of 1 mm; circulates externally to cool the UVC(between sections lamp; betweenC and D) sections B water flows to control the temperature of the treatments. In this survey, the temperature and C the food matrix circulates in a layer of 1 mm; externally (between sections C and D) water flows was adjusted to 20 °C. to control the temperature of the treatments. In this survey, the temperature was adjusted to 20 ◦ C. Figure 1. Structure of the UVC reactor used in this survey. A: UVC lamp; B: quartz protection glass; Figure 1. Structure of the UVC reactor used in this survey. A: UVC lamp; B: quartz protection glass; C and D stainless steel tubes. Section A–B: air circulation; section B–C: food matrix circulation area; C and D stainless steel tubes. Section A–B: air circulation; section B–C: food matrix circulation area; section C–D: section C–D:cooling water system. water cooling system. 2.3.2. UVC Treatments 2.3.2. UVC Treatments The parameters that defined each UVC treatment were: the UVC energy emitted by the lamp per volume unit (dosimetry); the flow rate at which the sample circulated (expressed as mL/s) and correlated with the revolutions per minute or rpm of the peristaltic pump) that determines the retention time inside the UVC lamp during each pass; and the number of passes through the reactor defined as the number of entries into the reactor or NET. Table 1 shows the combination of these parameters that Foods 2019, 8, 539 4 of 17 defined each treatment. The different combinations were divided into three groups depending on the flow rate used (T1, T2 and T3). In each group, the final doses achieved ranged from 20 to 160 J/mL. Table 1. Parameters of the UVC Treatments Tested. Treatment Flow Rate (mL/s) NET RPM Retention Time (s) UVC Dose (J/mL) Reynolds Number 21 24 20 52 12 48 40 26 6 96 80 13 T1 ≤2.9 2 5 120 100 10 3 144 120 9 1.5 192 160 7 30 360 20 60 720 40 121 1440 80 T2 41.9 300 658 151 1800 100 181 2160 120 241 2880 160 47 360 20 93 720 40 186 1440 80 T3 64.6 500 1015 233 1800 100 279 2160 120 372 2880 160 NET: number of entries to the tunnel; RPM: revolutions per minute (RPM). 2.3.3. Dosimetry Using a Potassium Iodide/Iodate Actinometer In this study, the UVC energy emitted by the lamp per volume unit at 254 nm was measured by the chemical iodide/iodate actinometer developed and described by Rahn [35] using a solution of potassium iodide that was passed through the UVC reactor. Its absorbance was then measured at 352 nm using a Nanophotometer Pearl spectrophotometer (IMPLEN) after diluting ten times the sample with distilled water. The formula proposed by Linden and Mofidi [36] was used to calculate the power of the UVC reactor lamps: (∆OD352 )(V)(23, 786.4) HMP = (1) (I)(1 + 0.02(T − 20.7))A where HMP is the applied UVC dose of the UVC lamp in mJ/cm2 ; ∆OD352 ∆OD352 is the difference of the absorbance of the irradiated sample and absorbance of the untreated (white) sample measured at 352 nm; V is the volume of the irradiated sample in L; I is the path length of light passing through the solution in cm (in the reactor used it was 0.1 cm); T is the temperature of the treatment, expressed in ◦ C; A is the area of sample facing the light source (700 cm2 ) and 23,786.4 is the constant (mJ/cm2 ) specific for LMP-UVC lamps. 2.3.4. Determination of the Type of Flow into the Reactor To determine the type of flow at which the matrix was circulating through the reactor the Reynolds number (Re) was determined according to the formula described by Müller et al. [37]: d·v·ρ Re = (2) η where d is the diameter of the tube (m), v is the velocity (m/s), ρ is the fluid density (kg/m3 ), and η is dynamic viscosity of the fluid (Pa·s). The relative viscosity was obtained using an Ostwald 1293 model Foods 2019, 8, 539 5 of 17 viscometer (CIVEQ, Mexico City, Mexico) and density with a hydrometer–aerometer densitometer HYDR-100-001 (Labbox, Vilassar de Dalt, Spain). According to that, the flow is considered laminar when the value of Re is less than 2100 and turbulent when it is above 4000. Table 1 shows the Re values obtained in the treated samples. 2.3.5. Determination of the Effective Depth of UVC In liquid matrices of high opacity or turbidity the UVC photons cannot penetrate deeply into the liquid, for this reason, it is necessary to know the depth at which the action of the UVC254 nm radiation is effective. This can be determined by applying the Lamber-Beer law by the equation: I = I0 ·e−k·c·d (3) where I and I0 correspond to the intensity of the UVC radiation expressed in mW/cm2 ; I corresponds to the intensity that the matrix receives at a given point, and I0 corresponds to the intensity emitted by the reactor. The constant k corresponds to the α(254) of the matrix expressed in cm−1 ; c is the concentration of solutes capable of absorbing UVC of the sample expressed in mol/L, and finally, d is the depth or distance at that must UVC light pass through expressed in cm. 2.4. UHPH Treatments For these experiments, an ultra-high pressure homogenizer Stansted FPG12500 (Stansted Fluid Power Ltd., Essex, United Kingdom), with a flow rate of 15 L/h, was used. The equipment consisted in two intensifiers driven by a hydraulic pump and a pressure valve consisting of a combination of a needle and a zirconium seat, with a cutting angle of 60 ± 0.5◦ at the needle and 45 ± 0.5◦ in the seat. Two heat exchangers (Garvía, Barcelona, Spain) were placed before and after the UHPH equipment to control the inlet and outlet temperatures. Whole and skim milk inoculated with spores of B. subtilis were treated at a pressure of 200 MPa in a single stage at an inlet temperature of 60 ◦ C. Then, the samples were subjected to the UVC treatment conditions T1 and T3 (Table 1). 2.5. Microbiological Analysis of the Samples and Calculation of the Lethality Achieved Once treated, samples were collected aseptically in a Telstar PCR Mini-V biosafety cabinet (Telstar, Terrassa, Spain), in sterile containers, and kept refrigerated (4 ◦ C) until analysis. Ten-fold dilutions were made from each sample in phosphate buffered saline solution consisting of 0.24 g of KH2 PO4 in 1 L of distilled water and pH adjusted to 7.4 (PBS, Panreac). Aliquots of each dilution were plated in Petri plates with trypticase soy agar medium enriched with 0.6% yeast extract (TSA-YE, Oxoid). The plates were incubated for 24 h at 37 ◦ C. The lethality caused by the UVC and/or UHPH treatments on B. subtilis was estimated with the formula: N0   Lethality = Log10 (4) N+1 where N0 is the initial amount of spores of B. subtilis present in the samples before the treatments and N is the number of remaining viable spores after the treatments, both expressed as in CFU/mL. 2.6. Determination of the Inactivation Kinetics and Effect of Treatment Variables Different statistical models obtained from the results obtained with the different matrices after the UVC treatments (T1, T2 and T3) were elaborated based on the lethality as a response variable, and different explanatory or independent variables: (1) Log C, corresponding to the quantity of inoculated microorganisms expressed as log CFU/mL; (2) UVC, the dose of UVC radiation, expressed in J/mL; (3) the treatment flow rate, expressed in mL/s; and (4) the NET, expressed as its absolute value. Five models were elaborated: model 1 presents the effect of the combination of variables of dose and Foods 2019, 8, 539 6 of 17 flow rate with a linear adjustment; model 2 presents the influence of the NET value; model 3 also presents the influence of NET but from a quadratic approach; model 4 included the variables of UVC Foods 2019, 9, x FOR PEER REVIEW 6 of 17 dose, flow rate and NET with a linear adjustment, and finally model 5 is a quadratic adjust using the variables dose, flow rate and NET. value. Five models were elaborated: model 1 presents the effect of the combination of variables of dose and flow rate with a linear adjustment; model 2 presents the influence of the NET value; model 2.7. Determination of the 3 also presents UVC Four the influence of Decimal Reduction NET but from Valueapproach; a quadratic (4Duvc) model 4 included the variables of UVC dose, flow rate and NET with a linear adjustment, and finally model 5 is a quadratic adjust Survival data of B. subtilis spores after the different UVC treatments (T1, T2 and T3) were adjusted using the variables dose, flow rate and NET. to different linear and nonlinear models of survival curves (biphasic, Weibull, Weibull with tail, Double 2.7. Weibull and biphasic Determination models of the UVC with shoulder) Four Decimal using(4Duvc) Reduction Value the GInaFiT software a freeware add-in for Microsoft® Survival Excel [38]. The suitability of the adjustment was data of B. subtilis spores after the different UVC determined by(T1, treatments determining T2 and T3)the root mean were square error (RMSE) value and the coefficient (R 2 ). The 4D value that is the UVC dose necessary to adjusted to different linear and nonlinear models of survivaluvc curves (biphasic, Weibull, Weibull with reduce 4tail, logDouble CFU/mL of the Weibull andinitial load biphasic of the models microorganisms with tested shoulder) using the wassoftware GInaFiT determined fromadd- a freeware the models in for that show Microsoft the ® Excel [38]. The suitability of the adjustment was determined by determining the root best adjustment. mean square error (RMSE) value and the coefficient (R2). The 4Duvc value that is the UVC dose necessary to reduce 2.8. Statistical Analysis 4 log of the CFU/mL of the initial load of the microorganisms tested was determined Results from the models that show the best adjustment. Three independent trials were performed for each experiment and two different samples, from each = 6). trial (n 2.8. Statistical AnalysisAnalysis of the Results of variance (ANOVA) and Tukey test was used to compare results between treatments Three and/or matrices. Differences were considered independent trials were performed to be significant for each experiment at p