The Ef fi ciency of UVC Radiation in the Inactivation of Listeria monocytogenes on Beef-Agar Food Models

Listeria monocytogenes has been identifi ed as a signifi cant public health risk in ready-to-eat (RTE) sliced meat products (1). This is due to the long shelf life of RTE meat products and the ability of L. monocytogenes to grow under refrigerated conditions. Contamination of RTE sliced meat products with Listeria usually occurs as a result of handling and slicing of the products following cooking and during packaging. Technologies that can eliminate L. monocytogenes from the surface of sliced meat before packaging are therefore required. One of the potential methods is ultraviolet light. Shortwave ultraviolet (UVC) radiation is a non-thermal microbial deactivation technique that does not leave any residues on the surface of treated products or produce any waste. It is an economical treatment and it can easily be installed on established production lines.


Introduction
Listeria monocytogenes has been identifi ed as a significant public health risk in ready-to-eat (RTE) sliced meat products (1).This is due to the long shelf life of RTE meat products and the ability of L. monocytogenes to grow under refrigerated conditions.Contamination of RTE sliced meat products with Listeria usually occurs as a result of handling and slicing of the products following cooking and during packaging.Technologies that can eliminate L. monocytogenes from the surface of sliced meat before packaging are therefore required.One of the potential methods is ultraviolet light.Shortwave ultraviolet (UVC) radiation is a non-thermal microbial deactivation technique that does not leave any residues on the surface of treated products or produce any waste.It is an economical treatment and it can easily be installed on established production lines.
The deactivation of foodborne pathogens to below the detection limit using UVC radiation has been achieved on agar plates (2)(3)(4).Meat products including ham slices (5), beef slice (6), raw salmon (7), and chicken breast fi llets (8) have also been treated by shortwave UV light.However, the results have not been as eff ective.One reason is believed to be that microorganisms are shielded by the surface irregularities on products such as beef, chicken or ham slices (2,5,(9)(10)(11).An alternative explanation is that the degree of survival is aff ected by the chemical composition of the food being treated.Bernbom et al. (12), in a study of the eff ect of UVC on L. monocytogenes in spent brine, used by the fi sh industry, concluded that the survival of the L. monocytogenes was enhanced by the presence of sodium chloride and organic food material.It was not clear whether this enhancement of survival was due to a physical protection of the cells in the food matrix or because of physiological changes in the att ached bacteria, or a combination of both.However, data on the ability of UVC light to deactivate microorganism(s) on foods with diff erent chemical or physical properties is lacking.
This study was carried out to determine whether the eff ectiveness of UVC against L. monocytogenes on meat surfaces is infl uenced by: (i) the chemical composition of the sample, (ii) surface irregularities of the sample being treated, or (iii) both.To achieve this goal, model products (beef-agar media) were prepared with a range of meat contents from 4.8 to 58.3 % (by mass per volume).Controls were produced of Listeria selective agar (LSA) and sliced cooked beef.The surface morphology of the beef--agar media and LSA plates was examined utilising a white light interferometer (WLI), which has not been used for food surface analysis purposes extensively, and the results were correlated with the UVC-deactivation of L. monocytogenes on these media.Unfortunately, the WLI could not measure the morphology of the sliced cooked beef sample surface.

Medium preparation
Sliced roast beef was purchased from a local supermarket and chopped into small pieces (approx.1-2 mm in all dimensions) using a hand held food blender (model 4169; Braun, Zamudio, Spain).An agar base solution was made by the addition of 1.5 g of agar powder (Scientifi c Laboratory Supplies Ltd, Manchester, UK) to 98.5 mL of distilled water.Nine bott les each containing the agar base solution were prepared.Diff ering amounts of chopped beef were added to the nine bott les to produce suspensions with diff erent beef contents (Table 1).A magnetic stirrer was added to each bott le.The contents of each bottle were mixed on a hot plate stirrer (model 375, VWR; Radnor, PA, USA) for 30 min followed by sterilisation at 121 °C for 15 min in an autoclave (Prestige Medical, Classic bench-top, Blackburn, UK).Stirring was continued for a further 30 min aft er sterilisation and before pouring the sterile beef-agar mix into sterile Petri dishes.Approximately 20 g of the hot suspension (approx.70 °C) was poured into each plate.Four replica plates of each meat mass per volume ratio and Listeria selective agar (LSA, CM0856; Oxoid, Hampshire, UK) plates with 0 % (by mass per volume) meat were produced.Care was taken during the pouring process to ensure a consistent volume in each plate with a homogenous distribution of particles and consistent content of the beef in all plates.
The poured plates were left in a refrigerator, operating at approx.6 °C, overnight to cool and set.Aft er sett ing the beef-agar plates were topped up to (40±0.1)g with sterile agar base solution and then returned to the refrigerator for another 3 h.LSA plates containing (40±0.1) g of agar were also poured.During the production of the plates the chopped beef particles sett led to the bott om of the plates.The set media were therefore inverted and aseptically inserted into sterile Petri dishes prior to UVC treatment.

Inoculation of beef-agar and LSA plates
An overnight inoculum of L. monocytogenes strain ATCC 19111 (purchased from National Centre for Type Cultures, Salisbury, UK) was prepared by adding few colonies from stock culture into 10 mL of tryptone soya broth (casein soya bean digest medium, Oxoid) bott les and incubating at 35 °C for (24±2) h.Beef-agar media and LSA plates were dried at 35 °C for 30 min before inoculation to remove any condensation from the surface.A volume of 0.1 mL of the overnight L. monocytogenes inoculum was deposited on top of each plate and spread using a sterile spreader.Four plates of each meat mass per volume ratio (40 plates in total) were inoculated, three of which were exposed to UVC radiation in the next stage and one was used as control sample.The plates were left (lid-on) at an ambient temperature of approx.22 °C for 60 min to dry and for L. monocytogenes to adhere.

Inoculation of beef slices
Using a Petri dish (r=5 cm) as a mould, beef slices were cut to have the same exposed surface area (78.5 cm 2 ) as the LSA and beef-agar medium plates.The same L. monocytogenes culture and the same inoculation method used for inoculating beef-agar medium and LSA plates were also used to inoculate four beef slices (three for treatment and one as control sample).Before UVC treatment, the inoculated slices were left at ambient temperature of 22 °C for 1 h to air-dry and for L. monocytogenes cells to adhere to the surface.

UVC treatment chamber
The UVC treatment chamber (Fig. 1) consisted of a stainless steel box (75 cm×93 cm×104 cm) with eight germicidal UVC bulbs (TUV PL-L 95 W; Philips, Pabianice, Poland).Four of the bulbs were placed in the top and other four at the base of the box.A welded wire-mesh sample holder with 4 cm 2 grid squares was installed between the emitt ers.The distance of the sample from the bulbs could be adjusted to achieve diff erent fi eld intensities.The UVC chamber was also equipped with two fans, operated via a thermostat, that removed the excess heat produced by UVC emitt ers, as well as an analogue timer (H3CR; Omron, Milton Keynes, UK), to control treatment time.

UVC treatment
Before each trial, the UVC emitt er bulbs were preheated for 10 min to stabilise the UVC emission.The sample support shelf was adjusted to be 5 cm below the upper UVC tubes.UVC peak intensity and dosage (fl ux) were measured using a digital UV metre (UVICURE ® Plus II, Eit, Sterling, VA, USA).Preliminary tests showed no signifi cant diff erence in the measured UVC radiation at diff erent points over the surface of the sample holder tray.
Three inoculated LSA plates were exposed to (195±11) mJ/cm 2 of UVC radiation in the treatment chamber.This experiment was repeated for beef slice and beef-agar media with diff erent beef contents (in %, by mass per volume), i.e. three samples from each type at a time.Preliminary tests with higher doses reduced the population of L. monocytogenes to below the detection limit on most of the samples.

Recovery of Listeria monocytogenes from control and treated samples
Aft er treatment each sample was removed from the Petri dish using a sterile scalpel and transferred aseptically to a fi lter stomacher bag (Stomacher Bags, 3500 mL; Seward Ltd, Worthing, West Sussex, UK).To each stomacher bag 40 mL of buff ered peptone water (BPW, CM0509; Oxoid) were added followed by stomaching (stomacher 400 circulator; Seward Ltd) for 60 s at 300 rpm.Diluting the samples was continued by conducting decimal dilutions in BPW followed by plating 0.1-mL aliquots on LSA plates.The plates were then incubated for 48 h at 35 °C, followed by enumeration.

Surface topography of the beef-agar media and LSA plates
To measure the surface roughness of the beef-agar media and LSA plates, a white light interferometer (New View 500; WLI; Zygo Ltd, Middlefi eld, CT, USA) and Me-troPro soft ware (Zygo Ltd) were utilised.A WLI is a machine which provides optical measurement for the depth of the irregularities on the sample surface.The WLI was set up using a ×2.5 Mirau lens with a zoom of ×1.00 and working distance of 10.3 mm.Sa and Ra roughness parameters for tested media were determined using the Me-troPro soft ware.Ra is the arithmetic average of the absolute values along a single specifi ed direction, and Sa is the arithmetic average of the absolute values over the whole measured area (13).
The surface roughness of each medium was measured in triplicates and 10 Ra values were obtained for the surface of each replicate (30 readings for each sample).Initially, it was not possible to get any reading for the surface of LSA plate using the WLI, as the medium was too transparent.To overcome this, 5 mL of ready-mixed white paint (300 mL; Wilko Colour Time, China) were added to 500 mL of the LSA solution, to make the mix opaque.The solution was then poured into Petri dishes aft er being mixed properly.The plates were left to set before being studied by WLI.Unfortunately, the WLI instrument was unable to provide any data regarding the surface topography of beef slices possibly due to the inability of the beef slice to refl ect light.
Colony forming units (CFU) of L. monocytogenes were calculated and then log transformed in Excel (Microsoft Excel, Microsoft Corporation, Redmond, WA, USA).To determine the error bars on the graphs a 95 % confi dence level was used.Further statistical analysis of the data to determine any probable signifi cant diff erence between calculated values was carried out in SPSS (IBM SPSS Statistics v. 19; IBM Company, Armonk, NY, USA) using one--way analysis of variance (ANOVA) followed by Scheff e's post-hoc test.

UVC-deactivation of Listeria monocytogenes on LSA, beef-agar and beef slices
Initial population of the L. monocytogenes on LSA plates, beef-agar medium plates and beef slices was 5.54, 5.89, and 6.19 log CFU/cm 2 , respectively.The detection limit was 1.01 log CFU/cm 2 .A UVC radiation dose of (195±11) mJ/cm 2 reduced the number of L. monocytogenes cells to below the detection limit in all replicates of both LSA and beef-agar media plates with 4.8, 9.1, and 16.7 % (by mass per volume) beef content (Fig. 2).At beef contents above 16.7 % (by mass per volume), the numbers of surviving L. monocytogenes tended to increase as the beef content increased (Fig. 2).Over 3 log CFU/cm 2 of L. monocytogenes cells survived on the beef-agar media with beef contents above 54.6 % (by mass per volume).
The application of a UVC radiation dose of (195±11) mJ/cm 2 produced reductions ranging from 2.95 to over 5.53 log CFU/cm 2 on beef-agar media with various meat contents.However, the same dosage on the beef slices only reduced the L. monocytogenes numbers from 6.19 to 5.32 log CFU/cm 2 , i.e. a 0.87 log reduction.

Surface topography of LSA and beef-agar plates
An example of the pictorial output of the WLI system is shown in Fig. 3.This picture shows the results of surface analysis of beef-agar with 55 % (by mass per volume) beef, provided by the MetroPro soft ware.The average of the Sa (μm) values with 95 % confi dence intervals are shown in Fig. 2. The lowest Sa (0.21 μm) value was obtained from the LSA plates, which confi rms that, as expected, this sample had the smoothest surface.
An analysis of variance (ANOVA) using IBM SPSS soft ware v. 19 showed that the LSA plates' surfaces were signifi cantly (p<0.05)smoother than the beef-agar media plates.The Sa values of the beef-agar media plates fl uctuated between 0.43 and 0.70 μm.Post-hoc test revealed an appreciable diff erence (p<0.05) between the surface irregularities of medium containing 50 % (by mass per volume) beef and the media with 4.8, 28.6, 37.5 and 44.4 % (by mass per volume) beef content.This increase in the roughness of medium with 50 % (by mass per volume) beef correlates with a decrease in the survival rate of L. monocytogenes cells shown in Fig. 2.
The survival rate of L. monocytogenes improved as the beef content in the media increased.However, although a signifi cant diff erence was seen between surface roughness of some of the beef-agar media at diff erent beef mass per volume ratios, the WLI analysis did not show a steady enhancement in the general trend of the surface roughness as the survival rate of L. monocytogenes increased (R 2 =0.24).The relationship between the survival of L. monocytogenes aft er treatment and the Sa value is shown in Fig. 4. No obvious trend was revealed and R 2 (0.09) and root mean square error (RMSE) (2.10) values calculated using Microsoft Excel of fi t of the trendline on the data show a very poor relationship.

Discussion
The results obtained in the current study have shown that the effi ciency of UVC in deactivating L. monocytogenes cells on the surface of agar plates decreases as the content of meat in the plates is increased.
Due to the impenetrability of UVC light, surface roughness has been suggested to be responsible for shielding microorganisms and hindering their deactivation (2,5,11,(14)(15)(16).It is known that the physical state of the food plays an important role in other decontamination techniques such as those based on thermal treatments (17).Studies using lux gene technologies showed substantial diff erences in the recovery of Salmonella typhimurium DT104 on the chicken surfaces with and without skin aft er the same heat treatment (18,19).However, no research has previously been conducted on the eff ect of a measure of surface topography on microbial inactivation by UVC radiation.In this trial beef agar medium plates with various meat contents but similar surface properties were prepared to challenge the impact of 'surface irregularity' factor on UVC treatment process.No correlation was found between the degree of deactivation and the surface roughness of the plates containing 4.8, 9.1, 16.7, 28.6, 37.5, 44.4 and 54.6 % (by mass per volume) beef, as measured by a WLI, thus indicating that, in this case, the degree of eff ectiveness was more related to the chemical composition.Product composition, solid content, colour, starch, and the overall chemistry of the food have been reported to have a key eff ect on the effi cacy of UVC treatments but the eff ects of these individual factors are not known (20).Organic materials have been known to absorb UVC radiation (21).It was speculated by Bernbom et al. (12) that the presence of sodium chloride and organic materials impacts the effi ciency of UVC radiation by absorption of the UVC rays, physical protection of the microorganism, and/ or possible induction of physiological changes in the microorganism, resulting in enhanced survivability.
The presence of organic materials has been found to reduce the effi ciency of other microbial decontamination techniques.Oteiza et al. (22) studied the heat resistance of two strains of Escherichia coli following addition of fat and starch to heating menstruum prepared from morcilla (a blood sausage).The authors concluded that when fat or starch contents are increased to obtain bett er yields, the risk to consumers increases since both components have a protective effect on the thermal inactivation of the microorganisms.In Bermúdez-Aguirre and Barbosa-Cánovas (23) study on the inactivation of L. innocua by thermo-sonication it was reported that inactivation of Listeria cells occurs fi rst in fat-free milk, and that the rate of inactivation decreases with increasing fat content.WLI is a powerful tool for optically measuring the depth of roughness on the sample surface.This device was, however, unable to analyse the surface topography of beef slice and un-manipulated LSA plates possibly owing to the incapability of such surfaces to refl ect light.The fl uctuation in the roughness measurement results implies the limitation of this machine in measuring surface imperfections of food products.In this study, no relationship was found between the surface roughness of seven out of nine types of produced beef-agar media and the degree of inactivation resulting from UVC radiation at the given dose.However, this is the fi rst time that the authors have read about this technique being used on a food surface.Further studies need to be carried out using the system on diff erent food surfaces that are known to be amenable to UVC decontamination technologies.
The results of this study show that the eff ects of the diff erent beef components and their contents on resistance of L. monocytogenes against UVC radiation need to be further investigated.In addition, more research needs to be done on the survival characteristics of spoilage or pathogenic microorganisms when exposed to UVC radiation on surfaces with diff erent chemical and physical characteristics.

Conclusions
Although surface irregularities of ready-to-eat meat slices have been reported to account for the limited success of UVC treatments in deactivating microorganisms on such surfaces, the fi ndings of the current study indicate that chemical composition may play a more important role in microbial survival/resistance than surface roughness.This could be through: (i) physical protection of the microorganisms, (ii) absorption of the UVC wavelengths by the chemical compounds present, or (iii) inducing physiological changes in the microorganisms making them resistant against UVC light.Further studies are required to identify the cause for microbial survival of microorganisms treated with UV on such food surfaces and how the eff ectiveness of such treatments can be improved.

Fig. 2 .
Fig. 2. Relationship between beef content in beef-agar media and levels of L. monocytogenes aft er UVC light treatment, and surface roughness (Sa).Beef content of zero represents Listeria selective agar (LSA) plates.Error bars indicate 95 % confi dence interval.Detection limit was 1 log CFU/cm 2

Fig. 4 .
Fig. 4. Relationship between the surface roughness (Sa) of beef--agar media plates and survival rate of L. monocytogenes.Root Mean Square Error (RMSE) and R 2 values are 2.1 and 0.09, respectively.Error bars indicate 95 % confi dence interval

Fig. 3 .
Fig. 3. Surface morphology of A120 beef-agar medium.Pictures a and b provide 3-dimensional and 2-dimensional drawing of the surface of A120, respectively.Picture c shows a graph which represents information regarding the depth of the surface irregularities over a certain distance (height×distance).Picture c also provides surface roughness (Ra, in μm) value of the surface of the sample along the line drawn in the middle of picture d.Surface roughness over whole measured area (Sa, in μm) values presented in this paper were calculated by averaging 10 Ra readings obtained from drawing 10 random lines over picture d

Table 1 .
Content of beef in diff erent beef-agar medium samples