Food Chemistry 166 (2015) 206–214

Contents lists available at ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Agent selection and protective effects during single droplet dryingof bacteria

http://dx.doi.org/10.1016/j.foodchem.2014.06.0100308-8146/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: bee.may@rmit.edu.au (B.K. May).

Sarim Khem a, Meng Wai Woo b, Darryl M. Small a, Xiao Dong Chen b,c, Bee K. May a,⇑a School of Applied Sciences, RMIT University, 124 La Trobe St, Melbourne, Victoria 3001, Australiab Department of Chemical Engineering, Monash University, Clayton, Victoria 3800, Australiac Chemical Engineering Innovation Laboratory, College of Chemistry, Chemical Engineering and Material Science, Soochow University, Suzhou, Jiangsu Province, China

a r t i c l e i n f o

Article history:Received 28 January 2014Received in revised form 22 April 2014Accepted 3 June 2014Available online 12 June 2014

Keywords:Lactobacillus plantarumLactoseProtective mechanismSingle droplet dryingTrehaloseWhey protein isolate

a b s t r a c t

The protective mechanisms of whey protein isolate (WPI), trehalose, lactose, and skim milk onLactobacillus plantarum A17 during convective droplet drying has been explored. A single droplet dryingtechnique was used to monitor cell survival, droplet temperature and corresponding changes in mass.WPI and skim milk provided the highest protection amongst the materials tested. In situ analysis ofthe intermediate stage of drying revealed that for WPI and skim milk, crust formation reduces the rateof sudden temperature increase thereby imparting less stress on the cells. Irreversible denaturation ofthe WPI components might have also contributed to the protection of the cells. Skim milk, however,‘loses’ the protective behaviour towards the latter stages of drying. This indicates that the concentrationof the WPI components could be another possible factor determining the sustained protective behaviourduring the later stages of drying when the moisture content is low.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Lactic acid bacteria (LAB) are a group of ubiquitous Gram posi-tive bacteria which are generally regarded as safe (GRAS) in foodpreservation. LAB, especially those in the genus Lactobacillus, aremost widely utilised in food and theopathic applications(Bourdichon et al., 2012; Vogel et al., 2011). In addition, LAB canenhance food safety and consumer health by preventing or reduc-ing the incidence of pathogens (Gaggìa, Gioia, Baffoni, & Biavati,2011; Gaggìa, Mattarelli, & Biavati, 2010).

The success in the utilisation of the LAB cells for such purposesoften requires high cell density and retention of activity for a rea-sonable amount of time before incorporation into the food formu-lation, in order to ensure the desired effect. A variety of differentmethods have been reported in the literature, including freezedrying, spray drying, vacuum drying, air drying and fluidised beddrying (Meng, Stanton, Fitzgerald, Daly, & Ross, 2008) and thesehave resulted in varying degrees of cell survival.

Amongst the processing techniques for producing dried micro-bial cells, optimising spray drying and developing product formu-lations to minimise activity losses have attracted increasingresearch interest as it is much more cost effective than freeze

drying and larger volumes can be processed (Knorr, 1998). Spraydrying involves the fine atomisation of a liquid feed material andthe subsequent rapid evaporation of water makes it especially suit-able for heat sensitive products. Yet, the success in utilising thisapproach has been based on trial and error, while differentresearch groups have reported very variable rates of cell survivalranging from less than 1–100% (Fu & Chen, 2011). Therefore, theobjective of enhancement in cell viability warrants further investi-gation and an insight into the inactivation mechanisms of cells is aprerequisite (Santivarangkna, Kulozik, & Foerst, 2008).

Several factors have been reported to affect cell survival duringspray drying, including the tolerance of the cells as well as process-ing parameters applied during the drying stages (Meng et al.,2008). It has been reported that the inactivation of microorganismsmay occur due to high temperature and dehydration stress (Fu &Chen, 2011; Meng et al., 2008; Peighambardoust, Golshan, &Hesari, 2011). Damage to macromolecules particularly includingDNA and RNA, as well as cell membranes and ribosomes resultsfrom the exposure of the cell to high temperature. In addition, dur-ing dehydration, damage to cytoplasmic membranes can occur,leading to the loss of some intracellular components (Ananta,Volkert, & Knorr, 2005).

In order to overcome the stress encountered during spray dry-ing, microbial cells are often formulated with various protectants,and research efforts have been devoted to the choice of agents pro-viding optimal survival (Leslie, Israeli, Lighthart, Crowe, & Crowe,

S. Khem et al. / Food Chemistry 166 (2015) 206–214 207

1995). Skim milk and carbohydrates in the form of disaccharidesused either alone, or in combination, have been the most com-monly used protective substances in dehydration of microorgan-isms in the last decade with varying effects upon cell survival.Various hypotheses have been proposed to explain the beneficialeffects of sugars including the involvement of water replacementand vitrification (Santivarangkna, Higl, & Foerst, 2008). Althoughthe protective mechanism of skim milk has not been fully explored,it has been suggested that lactose in skim milk interacts with thecell membrane and helps to maintain membrane integrity in amanner similar to the protection by other sugars including treha-lose (Corcoran, Ross, Fitzgerald, & Stanton, 2004). Another majorconstituent of skim milk is protein and whether or not this exertsa significant protective effect remains to be elucidated (Fu & Chen,2011).

There has been an increasing number of reports on the applica-tion of milk protein in the protection of probiotic cells in recentyears, relating to roles in protection both whilst drying is occurringas well as during exposure to gastrointestinal or bile fluids. Thisprobably reflects the desirable gelation properties of the proteinswhich have been shown to be useful for the microencapsulationof probiotics (Rathore, Desai, Liew, Chan, & Heng, 2013). Therehave been a few reports specifically on the application of wheyfor the protection of probiotics during spray drying. These havedemonstrated good survival; however, these studies utilised eitherwhey protein isolates (WPI) in combination with carbohydrate orliquid whey which also contains lactose (De Castro-Cislaghi,Silva, Fritzen-Freire, Lorenz, & Sant’Anna, 2012; Ying et al., 2011,2013). Furthermore, none of these studies has addressed the mech-anism(s) whereby the probiotics were protected.

During pilot scale spray dying, billions of droplets are sprayedinto a relatively large tower. As a result of the varied trajectoriesthe droplets could have quite different drying histories, posingchallenges to understand bacterial inactivation and the mecha-nisms of protection afforded by various drying media. To circum-vent these difficulties, a single droplet drying device mimickingthe spray drying environment was employed here to investigatedifferent inactivation histories of bacterial cells as the dropletwas being dried under controlled conditions. This single dropletdrying device allows accurate measurements of changes in thedroplet temperature, mass and diameter as drying progresses(Fu, Woo, Lin, Zhou, & Chen, 2011). Therefore, the objective of thisstudy was to examine the protection mechanism of Lactobacillusplantarum A17 during convective drying using WPI, lactose, skimmilk, mixtures of lactose and WPI and the well-known non-reducing disaccharide trehalose.

2. Materials and methods

2.1. Materials

2.1.1. ProtectantsFive forms of protectant were used: (1) WPI 894 (Fonterra,

Australia) in both pasteurised and native forms; (2) trehalose dihy-drate (T9531, Sigma Aldrich, Australia); (3) lactose monohydrate(L3625, Sigma Aldrich, Australia); (4) long life skim milk (pur-chased from a local supermarket and used without further treat-ment) and (5) a mixture of lactose and WPI in a ratio of9.94:0.06. Deionised water (Milli-Q system QGARD00R1, Millipore,Australia) was used in all experiments.

2.1.2. Lactobacillus plantarum A17The test strain of L. plantarum A17 was obtained from the col-

lection of the Laboratory of Food Microbiology, School of AppliedSciences, RMIT University, Australia. The strain was frozen at

�80 �C in MRS Broth (Oxoid, Australia) with 40% (v/v) glycerol.Bacteria cells were sub-cultured and grown to the stage of peaklog-phase in deMann, Rogosa and Sharpe (MRS) broth.

2.1.3. deMann Rogosa Sharpe (MRS) agarThis was obtained from Oxoid, Australia and contained peptone

(1%), meat extract (0.8%), yeast extract (0.4%), glucose (2%), sodiumacetate trihydrate (0.5%), polysorbate 80 (0.1%), dipotassiumhydrogen phosphate (0.2%), triammonium citrate (0.2%), magne-sium sulphate heptahydrate (0.02%), manganese sulphate tetrahy-drate (0.005%) and agar (1.0%). MRS agar was used for furthergrowth and identification of LAB colonies.

2.2. Sample preparation and measurements

2.2.1. Preparation of protectantsProtectant solutions were prepared by mixing the selected com-

ponents at a concentration of 10% (w/w) in deionised water with-out taking into consideration the inherent moisture content in theoriginal powder. WPI solution was dispersed by mixing with mag-netic stirring for at least 30 min, stored at 4 �C overnight to ensurefull hydration and the pH of the resulting solution was 6.6. Lactoseand trehalose solutions were autoclaved at 121 �C for 15 minbefore use. For the WPI solution, two solutions were used: a solu-tion pasteurised to 70 �C for 1 min and a non-pasteurised solution(native). As WPI tends to denature at temperatures beyond 65 �C,the latter solution was used as a control having minimal initialdenaturation. Sterile deionised water (without protectant) wasalso used as a control.

2.2.2. Micro differential scanning calorimetryTo determine the extent of whey protein denaturation during

pasteurisation, micro differential scanning calorimetry, SetaramMicro DSC VII (Setu-rau, Caluire, France) was used. Approximately800 mg sample was filled into a DSC pan and sealed hermetically; areference pan was also taken with deionised water of equal weight.Both pans were placed into the instrument chamber and equili-brated for 1 h at 20 �C to eliminate the effect of thermal historyprior to heating to 95 �C at a programmed heating rate of 1 �C/min. Duplicate measurements were taken. During the micro differ-ential scanning calorimetry analysis, the remaining undenaturedprotein in the pasteurised sample undergoes denaturation, andthus the enthalpy of denaturation is a measure of the amount ofundenatured protein present in the sample after pasteurisation.The enthalpy values were calculated from the area under a lineextending from 60.2 to 87.5 �C and the values were normalisedto give enthalpies per gram of dry protein.

2.2.3. Preparation of cells for drying experimentsL. plantarum cells were grown in MRS broth for 17 h to reach the

late growth phase. The active growing cells were homogeneouslymixed and harvested by pipetting 1 ml into each of a number ofEppendorf tubes before centrifugation at 4000g for 10 min. Afterdecanting the supernatant, the pellet of cells in each tube waswashed with sterile saline water (0.85%) and then resuspendedin one of a series of different protectant dispersions (10% w/w)with a tube containing deionised water as a control. The tubes withthe cell suspensions in carrier media and water were all stored atroom temperature (�24 �C) for approximately 30 min before com-mencement of the drying experiments. All tubes were placed in anice bath throughout the drying experiments to prevent undesiredcell proliferation. The viable cell concentration in the sampleswas checked immediately after preparation as well as after 4 h toensure that there was no cell proliferation. All suspensions weremixed by gently vortexing before the withdrawal of sub-samplesfor drying. Initial cell concentrations were 2–3 � 109 CFU ml�1.

208 S. Khem et al. / Food Chemistry 166 (2015) 206–214

2.2.4. Single droplet drying experimentA single droplet drying system in Monash University was uti-

lised in this experiment. The principle for this unit and the proce-dure of droplet drying experiments were adapted from Fu et al.(2011). A schematic figure of the glass filament drier set up wasadapted from Fu, Woo, Moo, and Chen (2012) (see SupplementaryFig. 1). Briefly, an air stream with controlled temperature, velocity,and humidity was used for drying a suspended single droplet in aconfined drying chamber. The single droplet was suspended at thetip of a specially made fine glass filament.

Droplets were dried for up to 240 s and cell survival was mon-itored every 30 s. A droplet of the cell suspension for each carrierwas generated using a 5 ll gas chromatograph micro-syringe(5FX, Part# 001100, SGE Analytical Science Pty Ltd., Australia).Between drying runs, the syringe was washed with deionised,70% ethanol, followed by sterilised deionised water to avoidcross-contamination. The initial droplet size for each run was2 ll and the two drying temperatures (90 �C and 110 �C) with anair velocity of 0.75 m/s and moisture content of 0.0001 kg/kg wereused in this study. The generated droplet was transferred and hungon the suspending glass filament inside the drying chamber with aseparate transferring glass filament (see Supplementary Fig. 2a).The temperature and mass data of the droplet during drying wereobtained in separate runs with identical drying conditions follow-ing the methods described by Fu et al. (2012).

For temperature measurement, the droplet was suspended by astatic glass filament (see Supplementary Fig. 2b). In contrast, formass measurement a flexible long glass filament with a bend hor-izontal section was used (see Supplementary Fig. 2c). Temperaturedata was obtained by inserting a fine-wire thermocouple (Type K,Part# CHAL-001, Omega Engineering Inc., USA) in the centre of thedroplet and collecting the temperature data from the thermocou-ple connected to a computer via a Picometer TC-08 (Pico Technol-ogy, UK). Mass data was obtained from the changes ofdisplacement of the mass measuring filament during drying. Whenthere is a mass hanging at the tip, the glass filament will deflect,leading to a displacement from its original location. Throughoutthe drying process, this deflection will continuously reduce asmoisture evaporates. The displacement history was then capturedby the camcorder. From calibration, such displacement is in a lin-ear relationship to the mass of the weight, and hence can be con-verted to the actual mass value by reference to a standard (seeSupplementary Fig. 3). Weight standards were prepared by sus-pending known amounts of Vaseline-agglomerated fine glassbeads (0.1–0.2 mg each) under the same drying conditions andrecording the corresponding displacement of the glass filament.The actual mass of the standards was obtained using a five decimalanalytical balance (Sartorius CP225D, Sartorius AG, Germany).Moisture content results were calculated from the weight mea-surement considering only the dissolved protectant material asthe solid mass within the droplet. Values are expressed in unitsof kg water/kg solid matter, that is, on a dry basis (db).

Fig. 1. Micro differential scanning calorimetry thermogram of native and pasteur-ised whey protein isolate (WPI) at a concentration of 10%w/w at natural unbufferedpH of 6.6 subjected to a heating scan of 1 �C/min.

2.2.5. Isothermal heat treatmentTo monitor the kinetics of cell survival during isothermal heat-

ing, cell suspensions were prepared in two separate solutions (oneof WPI and the other of lactose) with the same conditions as thecell suspensions used for single droplet drying. Aliquots (1.0 ml)of the cell suspensions in either lactose or WPI (10% (w/w)) weretransferred into 4 ml sterile test tubes. These tubes were thenheated at 50, 60 and 90 �C for up to 180 s and the cell survivalwas monitored at the same time regime as used in single dropletdrying. Micro-thermocouples were inserted into two additionalvials containing the same amount of cell suspension in order torecord the temperature history during these tests.

2.2.6. Enumeration of bacterial cellsFor each cell suspensions, initial cell counts were taken after

30 min of mixing with carrier media and subsequently after fourhours of storage in the ice bath, in order to establish whether cell pro-liferation was occurring. During single droplet drying experiments,drying was stopped at different time interval and the semi-drieddroplet was diluted into peptone water (2 ml) without removing itfrom the glass filament. After appropriate dilution in peptone water,aliquots of cell suspension were plated on MRS agar followed byincubation at 30 �C for 48 h. Likewise, cell suspensions from isother-mal heating, after heat treatment, the cell suspension was seriallydiluted and plated on MRS agar and incubated at 30 �C for 48 h.

2.3. Statistical analysis

All experiments were completed in three separate runs.Statistical significance testing for the difference between treat-ments were analysed using an ANOVA test with Tukey Post-Hocanalysis (Minitab 16 Statistical Software). As per convention, a pvalue of 0.05 or below was considered statistically significant.

3. Results and discussion

3.1. Protein denaturation during pasteurisation

The extent of WPI denaturation during pasteurisation wasdetermined by micro differential scanning calorimetry with thethermogram of native and pasteurised WPI shown in Fig. 1. Thedenaturation peak of both native and pasteurised WPI (10% w/w)is very similar, corresponding to a temperature of approximately77 �C at pH 6.6. These values are in close agreement with thosein a previous study of WPI (20% w/w) for which the denaturationtemperature was reported as 75 �C at pH of 6 (Duongthingoc,George, Katopo, Gorczyca, & Kasapis, 2013).

In the current study, enthalpy changes calculated from the ther-mal denaturation of native and pasteurised WPI were 0.874 and0.748 J/g respectively. This indicates protein has been partly dena-tured during the pasteurisation treatment, corresponding to lessthan 15% of the protein being denatured, consistent with the pre-vious study showing that WPI only started to denature at temper-ature higher than 65 �C during isothermal heating (AmdadulHaque, Aldred, Chen, Barrow, & Adhikari, 2013).

3.2. Protective mechanism during isothermal heat treatment

The survival kinetics of cell suspension in lactose and WPI sub-jected to isothermal heating at 50, 60 and 90 �C are presented in

Fig. 2. Kinetics of cell survival during isothermal heating at 50, 60 and 90 �C of cellsuspension in whey protein isolate (WPI) and lactose (Lac).

S. Khem et al. / Food Chemistry 166 (2015) 206–214 209

Fig. 2. There was no significant difference (p > 0.05) in survival ofcells in either lactose or WPI when treated at either 50 or 60 �Calthough cells subjected to heat treatment at 50 �C for 180 sresulted in only a slight reduction in survival. In contrast, a reduc-tion of almost 5 logCFU ml�1 in viability was observed to the end ofheating time when the temperature was set at 60 �C. Interestingly,there was a significant difference (p < 0.05) in cell survival betweencell suspension in lactose and WPI when subjected to water bath

(a)

(b)

Fig. 3. Droplet temperature during single droplet drying of pasteurised whey protein ilactose and whey protein isolate (LacWPI) and water: (a) Drying at 90 �C, (b) Drying at

heating at 90 �C where almost no cells survived the first 60 s ofheating for cells suspended in lactose; whereas for cells suspendedin WPI, survival was slightly above 3 logCFU ml�1 at the end of aheating period of 180 s.

There have been numerous studies on cell death during heatingas it is a commonly utilised method to eliminate or to reduce thenumber of microorganisms in food to ensure microbiological safetyor lengthen shelf life. Apart from affecting a critical component ofthe cells, heat has also been assumed to influence or cause the lossof many other components which are present in larger numberswithin the cell. The loss of these less critical components doesnot cause cell death until the numbers are reduced to a very lownumber or the cell is subjected to additional stress (Gould, 1989).In addition, heat is also believed to damage some macromolecularcomponents including DNA, RNA and proteins as well as cell mem-branes and ribosomes (Abee & Wouters, 1999).

Lactose at 10% (w/w) did not provide protection to strain A17 at90 �C in comparison with WPI at the same concentration, particu-larly at extended heating times of 120–180 s. On this basis, wehypothesise that the aggregation of denatured WPI is responsiblefor the protection of cells during prolonged heating as it was pre-viously reported that WPI only starts to denature at temperatureshigher than 65 �C during isothermal heating (Amdadul Haqueet al., 2013). In the current studies, although the temperature ofthe cell suspension reached over 65 �C only during the first 20 sof heating; we also found that less than 15% of protein becomesdenatured during isothermal heating. This finding is also

solate (WPI), native WPI, lactose (Lac), trehalose (Tre), skim milk (SM), mixture of110 �C. Note that different vertical scale are used in the two graphs.

(a)

(b)

Fig. 4. Kinetics of cell survival during single droplet drying of cell suspension inpasteurised whey protein isolate (WPI), native WPI, lactose (Lac), trehalose (Tre),skim milk (SM), mixture of lactose and whey protein isolate (LacWPI) and water: (a)Drying at 90 �C, (b) Drying at 110 �C.

210 S. Khem et al. / Food Chemistry 166 (2015) 206–214

consistent with a previous report in which only about 30% of WPIbecame denatured during isothermal heating of a 10% solution at80 �C for 60 s (Amdadul Haque et al., 2013). As a consequence,we observed a statistically significant (p < 0.05) decrease in cellsurvival during the initial heating period. The pH of the suspensionof bacterial cells in WPI used here was about 6.6 and the majorconstituent of WPI is of b-lactoglobulin which is a globular protein.It has been shown that upon prolong heating at either 70 or 85 �C,particles agglomerate into large clusters and slowly precipitate(Donato, Schmitt, Bovetto, & Rouvet, 2009). King and Su (1993) alsosuggested that milk proteins may form a protective coating overcell wall protein, thereby protecting the cell during heating.

Another interesting trend that can be observed in Fig. 2, per-tains to the unique strain of bacteria utilised in this research. At50 �C, the cell survival count reduced by approximately 1 magni-tude. However, a statistically significant (p < 0.05) reductionoccurred, by several magnitudes, at 60 and 90 �C. Therefore, itcan be deduced that the apparent ‘massive destructive’ tempera-ture threshold able to be tolerated by the bacterium is in the rangeof 50 < Tlethal < 60 �C.

The observations here indicate that whey protein reduces ther-mal stress as it has been shown in the isothermal heating due todenaturation of the protein, possibly forming a layer around thecell. This is consistent with the suggestion by Duongthingoc et al.(2013) that the denaturation mechanism helps protect the yeastcell by adjustment of the initial pH of the suspension close to theisoelectric point. The results in Fig. 2 clearly demonstrate thatWPI offers a protective mechanism to the bacterium L. plantarumA17 strain.

3.3. Skin forming protective mechanism during single droplet drying

The temperature–time curves for the droplets containing thedissolved solids are quite distinct from those for water (Fig. 3)which with the latter being characterised by a long period ofalmost constant temperature followed by a dramatic increase tojust below the drying temperature once almost all the water hadevaporated. This extended period of constant temperature can beexplained by evaporative cooling caused by a constant rate ofmoisture loss. The temperature for the droplets containing the dis-solved protectants starts to increase much earlier than observedfor the samples of pure water. This early initiation of a temperaturerise within the droplets may be caused by the polar water mole-cules becoming bound to the polar protectant molecules, therebyreducing the drying rate and hence the magnitude of evaporativecooling. The progressive formation of a solid outer crust will alsoimpede the removal of moisture during the drying process. Thetemperature-time curve clearly shows that the temperature forthe droplets containing either WPI or skim milk rose even soonerthan for the droplets containing other protectants. This reflectsthe skin-forming surface-active behaviour of WPI enhancing thesurface crust formation during convective drying. For the dryingat a temperature of 110 �C, the pasteurisation of WPI resulted inenhanced skin-forming behaviour. One possible explanation forthis could be that as a result of pasteurising the WPI componentsclose to the denaturation temperature of b-lactoglobulin (78 �C)(Gulzar, Bouhallab, Jeantet, Schuck, & Croguennec, 2011), the wheyproteins agglomerate and can be considered to be in a globularmolten state.

Corresponding to the temperature–time trend, Fig. 4 shows thekinetics of cell survival with different protective substances, alongwith the cell suspension in water as a control, during single dropletdrying at 90 and 110 �C. During the first 90 s of drying for 90 �C,there was no significant cell reduction (p > 0.05) as the survivalrate was approximately 109 CFU ml�1 for each of the protectantstested. The same trend was also observed for up to 60 s of drying

at 110 �C. Whilst the low reduction in survival rates of the otherprotectants during this period can be attributed to the relativelylow temperature of about 30–40 �C, it was surprising that thehigher temperatures, in the range of 47–60 �C for the WPI (bothpasteurised and native), did not result in a significant reduction(p > 0.05) in survival. This could be due to the temperature,although higher, but still being lower than the ‘massive destruc-tion’ temperature in the range of 50–60 �C, established in the stud-ies of isothermal heating.

Beyond the initial period characterised by relatively high cellsurvival, a sudden reduction in cell counts of several orders of mag-nitude was observed. For both drying temperatures used in thisstudy, the sudden reduction in cell survival appeared to correspondto the sudden increase in temperature of the droplet for purewater, as well as those containing lactose, trehalose and mixtureof lactose and WPI. For the water there was a slight delay of cellsurvival reduction due to the delay in sudden temperature increasein the absence of crust formation. This evidence supports the viewthat the rate of change of temperature might be the primary factorwhich contributes to the death of the cell. The maximum rates oftemperature increase for these droplets are approximately 1.2 �C/s (90 �C) and 1.9 �C/s (110 �C).

Throughout the time when there is a sudden reduction in cellsurvival (Fig. 4), corresponding to approximately 90–120 s (90 �C

S. Khem et al. / Food Chemistry 166 (2015) 206–214 211

drying condition) and 60–90 s (at 110 �C), the moisture contentremains relatively high, at about 0.70–2.05 kg/kg db (Fig. 5). Thisis above the critical moisture content at which WPI is reported torapidly denature during convective drying, with values of0.35 kg/kg db and 0.65 kg/kg db during drying at 65 and 80 �Crespectively (Amdadul Haque et al., 2013). We suspect that thissudden reduction in cell survival may have been due to the effectof the sudden temperature rise.

On the other hand, it could also be argued that the dropletwould have reached the maximum destruction temperature ofapproximately 50–60 �C; it is also possible that the destructiontemperature may even be lower when the droplet is dehydratedto a moisture content of around 50% (w/w). However, due to therapid increase in temperature during that period, it is difficult todiscern which is the dominant factor contributing to the reductionin cell viability.

Trehalose is widely reported to be effective as a protectant dur-ing spray drying of bacteria and this is particularly attributed to thereplacement mechanism (Santivarangkna et al., 2008).Surprisingly, in the current study, this disaccharide did not appearto offer any protection to our strain of bacteria. Lactose in skimmilk has been suggested to interact with cell membrane and helpsto maintain membrane integrity in a similar way to the non-reducing disaccharide trehalose (Corcoran et al., 2004). The latteris one of the most widely recognised osmotic protectants againstboth osmotic and thermal stresses (Obuchi, Iwahashi, Lepock, &Komatsu, 2000) as it exhibits a universal protective effect onanhydrobiotic cells including those of microorganisms. Waterreplacement, high glass transition temperature and thermo-protectant against protein denaturation have been proposed forthe mechanism of protection by trehalose during dehydration(Santivarangkna et al., 2008). For example, adding 0.25 M trehalosereduces the loss of viability of Lactobacillus bulgaricus during bothheat and osmotic dehydration at 70 and 20 �C respectively(Zavaglia, Tymczyszyn, Antoni, & Anibal, 2003). Similarly, suspen-sion of cells in 20% trehalose resulted in up to 68% of survival dur-ing spray drying for outlet temperatures in the range of 65–70 �C(Sunny-Roberts & Knorr, 2009). Therefore the current results indi-cate that, the water replacement protective mechanism might notbe suitable for stresses existing under convective droplet dryingconditions. This may explain why sometimes such protectants donot work for certain drying processes or strains of bacteria(Santivarangkna, Kulozik, & Foerst, 2006).

Skim milk appeared to have resulted in higher cell survival dur-ing drying than native WPI (Fig. 4a (120 s) and Fig. 4b (90 s).However, the difference was found to be not significant(p > 0.05). In contrast to the sugars, the sudden reduction of the cellsurvival count for the skim milk and the WPI droplets did not cor-respond to the period of sudden increase in temperature of thedroplet. The sudden increase in temperature occurred relativelyearly at approximately 45–60 s (90 �C) and 32–52 s (110 �C),whereas the sudden drop in cell survival count only began atapproximately 90 s and 60 s, respectively. This could be due tothe slower rate of temperature increase, approximately 0.45 �C/s(90 �C) and 0.66 �C/s (110 �C), which might have imparted lessstress for the bacterial cells. This may reflect the lower rate of tem-perature change due to the inherent skin forming behaviour of theproteins in both the WPI and the skim milk droplets. If the slowerrate of temperature increase imparts less stress upon the cells,then what caused the sudden magnitude decrease in cell viabilitycount at approximately the same time with the other droplet?From an examination of the temperature curves, we speculate thatthe subsequent decrease in cell survival count at around 120 s(90 �C) and 90 s (110 �C) was primarily due to the dropletapproaching the ‘maximum destruction’ temperature. The signifi-cance of the skin forming behaviour could provide one explanation

as to why the lactose-WPI mixture, which only contains a veryminimal percentage of WPI, and resulting in no early crust forma-tion behaviour, did not offer similar protection.

Although solutions used as protectants were prepared at 10%(w/w), inherent powder moisture content was not taken into con-sideration; thus, initial solid contents may have differed slightlyfrom the reported values. From the moisture loss curves in Fig. 5,during the intermediate stage of drying, the drying rates for eachof the droplets were very similar. Therefore, it is unlikely that dif-ferences in dehydration stresses will be the principal factor in theenhanced protective mechanism of the WPI and skim milk drop-lets. One can argue that the differences in the moisture contentduring the intermediate drying period might affect the cell viabil-ity count. The values of moisture content obtained at the start andend of the period of rapid temperature increase are summarised inTable 1. An average value is given for droplets containing protec-tants other than WPI and skim milk as their temperature-timecurves are quite closely superimposed. Droplets containing WPIhad moisture contents that were 34–41% higher than thosecontaining other protectants at the start of the period of rapid tem-perature increase, whilst droplets containing WPI had contents48–49% higher at the completion of the period of rapid increase.It is unclear as to whether higher moisture contents help to pre-serve the cell viability. Nevertheless, even if it contributes to thecell counts, the higher moisture content manifests from the sur-face-active development of crust for WPI and skim milk; furtherhighlighting the importance of the earlier skin formation.

Examining the survival count for extended drying periods of upto 240 s, for which the moisture content in all particles are similar,within the range of 0.1–0.2 kg/kg db, the survival count for WPI isstill higher than that for the other protectants. Surprisingly, theextended drying conditions revealed that skim milk ‘loses’ the pro-tective behaviour within the extended drying region, approachingthe cell counts found for the other protectants, in contrast to WPI(p < 0.05). This was despite both skim milk and WPI displaying theskin formation behaviour described earlier. This indicates that adifferent mechanism might be responsible for the protectivebehaviour towards the later stages of drying. At the moment, morework is required to determine the exact protective mechanism forthe latter stages of drying. One possible mechanism is the denatur-ation and aggregation of protein during drying (Duongthingocet al., 2013). Previous work on various types of milk proteins(including casein) provided evidence that the aggregation of wheyprotein resulted in the release of sulphur-containing amino acidsthereby inhibiting lipid oxidation of the bacterial membrane dur-ing spray drying (Soukoulis, Behboudi-Jobbehdar, Yonekura,Parmenter, & Fisk, 2013). This aspect will be discussed in the nextsection.

3.4. Possible influence of protein denaturation and aggregation on theprotective mechanism

The denaturation and aggregation of WPI was previously sug-gested to be a dominant mechanism contributing to the protectivemechanism on cell viability (Duongthingoc et al., 2013). This wasdeduced primarily from the conditions of the final dried productwithout examining in detail how this protective behaviour isdeveloped during in situ drying. In the current study we haveattempted to unveil more on this mechanism by comparing theresults for native WPI with those of the pasteurised WPI solution.Although the difference in the degree of denaturation betweenthe two solutions was less than 15%, determined calorimetrically,the increase in temperature of the pasteurised WPI droplet wasmore rapid and occurred earlier than the native one. This mightindicate a higher degree of protein aggregation leading to crust for-mation during convective drying. However, comparison of the cell

(a)

(b)

Fig. 5. Moisture content on dry basis during single droplet drying of cell suspension in pasteurised whey protein isolate (WPI), native WPI, lactose (Lac), trehalose (Tre), skimmilk (SM), mixture lactose and whey protein isolate (LacWPI) and water: (a) Drying at 90 �C, (b) Drying at 110 �C.

Table 1Comparison of moisture content of droplets containing WPI with droplets containing other protectants at the start and finish of the period of rapid temperature increase.

Drying temperature (�C)

90 110

Start/finish of rapid increase (s) 90 120 60 90Moisture content (kg/kg db) of droplets containing

WPI 1.36 0.51 1.96 0.58Average for lactose and trehalose 0.89 0.31 1.39 0.35

Difference 41% 48% 34% 49%

212 S. Khem et al. / Food Chemistry 166 (2015) 206–214

survival count between both droplets for the two drying tempera-tures revealed no significant differences (p > 0.05). Therefore, thedegree of denaturation of less than 15% induced by the pasteurisa-tion step did not influence the protective mechanism observed forthe WPI experiments.

We then further explored the possible influence of in situ devel-opment of denaturation affecting the protective mechanism on thecell viability count for the WPI and skim milk droplets. For convec-tive drying conditions, Amdadul Haque et al. (2013) reported anirreversible denaturation temperature for whey of approximately54 �C at drying temperatures of 65 �C and 80 �C. The report alsonoted critical moisture of 0.35–0.56 kg/kg db below which, rapiddenaturation can potentially occur due to thermal stresses. Basedon this information, rapid development of denaturation could havepossibly occurred between 90–120 s (for the 90 �C case) and50–95 s (for the 110 �C case taking WPI droplet for the lower time

limit). These two time durations correspond to the period of drasticreduction in cell viability for the two drying conditions respec-tively. Therefore, there is a chance of in situ irreversible proteindenaturation contributing to the enhanced protective characteris-tics when compared to those of the sugar protectants studied.

If the denaturation of the WPI components contributes to theprotective mechanism, then why did the mixture of lactose andWPI not display such protective behaviour and why did the skimmilk ‘lose’ the protective mechanism towards the latter stages ofdrying? In comparison with the droplet prepared with WPI, themixture of lactose and WPI, as well as skim milk, contained onlyaround 0.6% of whey components. Drawing an analogy to previousresearch, although not a direct comparison, lactose-sodium casei-nate protectant mixture was shown to improve the survival rateof Lactococcus lactis when compared to pure lactose (Ghandi,Powell, Chen, & Adhikari, 2012). However, the proportion of

S. Khem et al. / Food Chemistry 166 (2015) 206–214 213

sodium-caseinate to lactose used was 1:3, which was significantlymuch higher than for the skim milk protein and lactose used here.Study on the usage of whey permeate consisted mainly of lactosewith low whey protein concentration used as a protectant in spraydrying of Lactobacillus acidophilus resulted in only half of the sur-vival rate when compared to reconstituted skim milk (Riveros,Ferrer, & Borquez, 2009). Therefore, the current results showstrong evidence that the high proportion of lactose with a low pro-portion of whey protein is not sufficient to confer protectionagainst such thermal and moisture stress, indicating that a mini-mal threshold amount of WPI is required to offer effective protec-tive behaviour.

3.5. Possible dual protective mechanism and future work

From the discussions presented here, there appears to be twopossible mechanisms in which WPI provides protection to cellsduring convective drying of droplets. In preventing a sudden tem-perature increase, the skin formation characteristic of WPI reducesthe thermal stress and possibly preserves higher moisture contentsfor the microorganisms during the intermediate stage of drying.There are potentially many ways that can be adopted to controlthe process of skin formation. An increased rate of formation mightbe induced by adjusting the initial pH of the cell suspension, or byusing either a higher drying temperature or increased initial con-centrations of protectant. The first of these, however, has to be bal-anced to ensure that the outlet temperature of the spray dryer isnot excessively high, which would induce a different type of ther-mal stress due to prolonged exposure to high temperature condi-tions (Fu & Chen, 2011). In addition, variations in initial pH couldalso induce early skin formation which could bring about cell pro-tection and research on this is in progress. Concurrent to the skinformation mechanism, the irreversible denaturation of the WPImight also contribute to the protection. More work is required toindependently assess both factors to determine which mechanismpre-dominates in contributing to this protective phenomenon.

4. Conclusions

A potential probiotic strain of L plantarum A17 suspended in aseries of different carriers (WPI, lactose, skim milk, trehalose, aswell as a mixture of lactose and WPI), were convectively driedusing the single droplet drying approach to study the cell protec-tion mechanism afforded by different substances. WPI providedthe highest protection during the intermediate and latter stagesof drying. During the intermediate drying stage, a reduction inthe rate of temperature increase and the resultant preservationof higher moisture content, due to the skin forming behaviour,could be a major factor contributing to the protective behaviour.Concurrently, the development of irreversible denaturation mightalso be a further factor influencing the survival of the bacteria. Itwas deduced that WPI concentration might be another importantfactor in providing sustained protection towards the latter stageof drying.

Acknowledgement

The first author would like to thank the Australian governmentfor providing him with Endeavour postgraduate award.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.foodchem.2014.06.010.

References

Abee, T., & Wouters, J. A. (1999). Microbiasl stress response in minimal processing.International Journal of Food Microbiology, 50(1–2), 65–91.

Amdadul Haque, M., Aldred, P., Chen, J., Barrow, C. J., & Adhikari, B. (2013).Comparative study of denaturation of whey protein isolate (WPI) in convectiveair drying and isothermal heat treatment processes. Food Chemistry, 141(2),702–711. <http://dx.doi.org/10.1016/j.foodchem.2013.03.035>.

Ananta, E., Volkert, M., & Knorr, D. (2005). Cellular injuries and storage stability ofspray dried Lactobacillus rhamnosus GG. International Dairy Journal, 15(4),399–409.

Bourdichon, F., Casaregola, S., Farrokh, C., Frisvad, J. C., Gerds, M. L., Hammes, W. P.,et al. (2012). Food fermentations: Microorganisms with technological beneficialuse. International Journal of Food Microbiology, 154, 87–97.

Corcoran, B. M., Ross, R. P., Fitzgerald, G. F., & Stanton, C. (2004). Comparativesurvival of probiotic lactobacilli spray-dried in the presence of prebioticsubstances. Journal of Applied Microbiology, 96(5), 1024–1039.

De Castro-Cislaghi, F. P., Silva, C. D. R. E., Fritzen-Freire, C. B., Lorenz, J. G., &Sant’Anna, E. S. (2012). Bifidobacterium Bb-12 microencapsulated by spraydrying with whey: Survival under simulated gastrointestinal conditions,tolerance to NaCl, and viability during storage. Journal of Food Engineering,113, 186–193.

Donato, L., Schmitt, C., Bovetto, L., & Rouvet, M. (2009). Mechanism of formation ofstable heat induced beta lactoglobulin mirogels. International Dairy Journal,19(5), 295–306.

Duongthingoc, D., George, P., Katopo, L., Gorczyca, E., & Kasapis, S. (2013). Effect ofwhey protein agglomeration on spray dried microcapsules containingSaccharomyces boulardii. Food Chemistry, 141(3), 1782–1788.

Fu, N., & Chen, S. D. (2011). Towards a maximal cell survival in convective thermaldrying processes. Food Research International, 44, 1127–1149.

Fu, N., Woo, M. W., Lin, S. X. Q., Zhou, Z., & Chen, X. D. (2011). ReactionEngineering Approach (REA) to model the drying kinetics of droplets withdifferent initial sizes-experiments and analysis. Chemical Engineering Science,66, 1738–1747.

Fu, N., Woo, M. W., Moo, F. T., & Chen, X. D. (2012). Microcrystallization of lactoseduring droplet drying and its effect on the property of the dried particle.Chemical Engineering Research and Design, 90, 138–149.

Gaggìa, F., Gioia, D. D., Baffoni, L., & Biavati, B. (2011). The role of protective andprobiotic cultures in food and feed and their impact in food safety. Trends inFood Science & Technology, 22, S58–S66.

Gaggìa, F., Mattarelli, P., & Biavati, B. (2010). Probiotics and prebiotics in animalfeeding for safe food production. International Journal of Food Microbiology,141(suppl. 1), S15–S28.

Ghandi, A., Powell, I. B., Chen, X. D., & Adhikari, B. (2012). The effect of dryer inletand outlet air temperatures and protectant solids on the survival of Lactococcuslactis during spray drying. Drying Technology, 30(14).

Gould, G. W. (Ed.). (1989). Mechanisms of action of food preservation procedures.Barking: Elservier.

Gulzar, M., Bouhallab, S., Jeantet, R., Schuck, P., & Croguennec, T. (2011). Influence ofpH on the dry heat-induced denaturation/aggregation of whey proteins. FoodChemistry, 129, 110–116.

King, V. A.-E., & Su, J. T. (1993). Dehydration of Lactobacillus acidophilus. ProcessBiochemistry, 28, 47–52.

Knorr, D. (1998). Technology aspects related to microorganims in functional foods.Trends in Food Science & Technology, 9, 295–306.

Leslie, S. B., Israeli, E., Lighthart, B., Crowe, J. H., & Crowe, L. M. (1995). Trehalose andsucrose protect both membranes and proteins in intact bacteria during drying.Applied and Environmental Microbiology, 61(10), 3592–3597.

Meng, X. C., Stanton, C., Fitzgerald, G. F., Daly, C., & Ross, R. P. (2008).Anhydrobiotics: The challenges of drying probiotic cultures. Food Chemistry,106, 1406–1416.

Obuchi, K., Iwahashi, H., Lepock, J. R., & Komatsu, Y. (2000). Calorimetriccharacterization of critical targets for killing and acquired thermotolerance inyeast. Yeast, 16, 111–119.

Peighambardoust, S. H., Golshan, T. A., & Hesari, J. (2011). Application of spraydrying for preservation of lactic acid starter culutres: A review. Trends in FoodScience & Technology (22), 215–224.

Rathore, S., Desai, P. M., Liew, C. V., Chan, L. W., & Heng, P. W. S. (2013).Microencapsulation of microbial cells. Journal of Food Engineering, 116, 369–381.

Riveros, B., Ferrer, J., & Borquez, R. (2009). Spray drying of a vaginal probiotic strainof Lactobacillus acidophilus. Drying Technology, 27(1), 123–132.

Santivarangkna, C., Higl, B., & Foerst, P. (2008). Protection mechanisms of sugarsduring different stages of preparation process of dried lactic acid startercultures. Food Microbiology, 25, 429–441.

Santivarangkna, C., Kulozik, U., & Foerst, P. (2006). Effect of carbohydrates on thesurvival of Lactobacillus helveticus during vacuum drying. Letters in AppliedMicrobiology, 42(3), 271–276.

Santivarangkna, C., Kulozik, U., & Foerst, P. (2008). Inactivation mechanisms of lacticacid starter cultures preserved by drying processes. Journal of AppliedMicrobiology, 105(1), 1–13.

Soukoulis, C., Behboudi-Jobbehdar, S., Yonekura, L., Parmenter, C., & Fisk, I. (2013).Impact of milk protein type on the viability and storage stability ofmicroencapsulated Lactobacillus acidophilus NCIMB 701748 using spraydrying. Food Bioprocessing Technology, 1120-x. doi: 10.1007/s11947 -013-1120-x.

214 S. Khem et al. / Food Chemistry 166 (2015) 206–214

Sunny-Roberts, E. O., & Knorr, D. (2009). The protective effect of monosodiumglutamate on survival of Lactobacillus rhamnosus GG and Lactobacillusrhamnosus E-97800 (E800) strains during spray drying and storage intrehalose-containing powders. International Dairy Journal, 19, 209–214.

Vogel, R. F., Hammes, W. P., Habermeyer, M., Engel, K.-H., Knorr, D., & Eisenbrand, G.(2011). Microbial food cultures – Opinion of the senate commission on foodsafety (SKLM) of the German research foundation (DFG). Molecular Nutrition andFood Research, 55, 654–662.

Ying, D., Sanguansri, L., Weerakkody, R., Singh, T. K., Leischtleld, S. F., Gantenbein-Demarchi, C., et al. (2011). Tocopherol and ascorbate have contrasting effects on

the viability of microencapsulated Lactobacillus rhamnosus GG. Journal ofAgricultural and Food Chemistry, 59, 10556–10563.

Ying, D., Schwander, S., Weerakkody, R., Sanguansri, L., Gantenbein-Demarchi, C., &Augustin, M. A. (2013). Microencapsulated Lactobacillus rhamnosus GG in wheyprotein and resistant starch matrices: Probiotic survival in fruit juice. Journal ofFunctional Foods, 5, 98–105.

Zavaglia, A. G., Tymczyszyn, T., Antoni, G. D., & Anibal, D. E. (2003). Action oftrehalose on the preservation of Lactobacillus delbrueckii ssp. bulgaricus by heatand osmotic dehydration. Journal of Applied Microbiology, 95, 1315–1320.