Annual Report 1996 Wisconsin Center for Dairy Research

WisconsinCenter for Dairy Research

Annual Report 1996

Wisconsin

Center for Dairy Research

Annual Report 1996

1605 Linden Dr.

Madison, WI 53706

Phone: (608) 262-5970

Fax: (608) 262-1578 http://www.cdr.wisc.edu

CDR Annual Report 1996

ii

CDR Annual Report

Published September 1, 1996, by the Wisconsin Center for Dairy Research.

Our annual report is a technical overview of CDR funded research and other Centeractivities during fiscal year 1996. We prepared this report for organizations fundingCDR and for fellow dairy researchers. This document describes projects in progressand interpretations of data gathered to date. It is not a peer-reviewed publication.

Please seek the author's written consent before reprinting, referencing, orpublicizing any reports contained in this document.

For more information call Karen Paulus at (608) 262-8015.(E-mail: [email protected])

Overview

iii

Contents

Overview vii

Chapter 1Milkfat Research 1

Use Of Pregastric Lipases Immobilized In A Hollow Fiber Reactor To Produce Lipolyzed DairyProductsCharles G. Hill, Jr., Hugo S. Garcia, Louis Lessard, Souheil Ghannouchi ......................................................................................... 3

The Effect of pH on the Fatty Acid Specificity of Pregastric Esterase from Kid GoatHugo S. Garcia ................................................................................................................................................................................ 11

Evaluation of Monoacylglycerols Derived from Butteroil as Emulsifiers in FoodsKirk L. Parkin, Kathleen Wiederholt, Purwiyatno Hariyadi, Shu-Jung Kuo, Marit Reierstad ....................................................... 13

Application of Emulsifiers From Butteroil or Milkfat Fractions in Ice CreamR.W. Hartel, K.L. Parkin, B. Liang ................................................................................................................................................... 17

Preparation of Butteroil-in-Water Emulsions Using Blends of Milk Proteins and LipidsMarit Reierstad, Kirk L. Parkin ...................................................................................................................................................... 22

Kinetics of Milkfat CrystallizationR.W. Hartel, D.B. Patience, D. Illingworth ...................................................................................................................................... 23

Mechanisms for Formation of Milkfat CrystalsR.W. Hartel, Y. Shi ........................................................................................................................................................................... 25

Continued Studies of Surface Melt Crystallization Techniques for Fractionating MilkfatR.W. Hartel, J. Ulrich, M. Tiedtke .................................................................................................................................................... 27

Physical Chemistry of Lipid Mixtures: Dairy Based SpreadsR.W. Hartel, R.L. Lindsay, J. Knollenberg, B. Liang ........................................................................................................................ 29

Incorporation of Milkfat Fractions in Chocolates - Phase 2R. W. Hartel, J. Bricknell, S. Metin ................................................................................................................................................... 31

Investigation of Baked Milkfat Flavor Development in Milkfat Ingredients for theBakery and Food IndustriesRobert C. Lindsay, Ann Han (In collaboration with researchers at the New Zealand Dairy Research Institute) ......................... 33

Milkfat Applications Research ProgramKerry E. Kaylegian, Barbara H. Ingham, Orville H. Harris,Chris Kirk .......................................................................................... 35

Effects of Defined Milkfat Fractions on Postprandial Lipid Metabolism in the RatDenise Ney, Hui-Chuan Lai, Mike Grah ......................................................................................................................................... 37

Development of New Dairy Products and the Adoption Process by U.S. ConsumersBrian W. Gould, J.H. Park ............................................................................................................................................................... 39

Chapter 2Nonfat Solids 51

Fractionation of Whey Proteins Using Ion Exchange MembranesMark. R. Etzel, Clovis Ka Kui Chiu, Ida A. Adisaputro .................................................................................................................... 53


iv

Conversion of Whey Permeate to Propylene Glycol for Food and Non-Food UsesDouglas C. Cameron, Mark R. Etzel, Nedim E. Altaras, Roxanne M. Smith, Yi-Jui Wu .................................................................. 55

Chapter 3Cheese Research 57

Cheese Making Properties of Milk from Cows of Different Milk Protein GenotypeRobert Bremel, Josie Lewandowski, Joshua Ruff, Tricia Braaksma, Mark Johnson, John Jaeggi, Brian Gould ............................. 59

Studies of the Influence of Milkfat on the Formation of Flavor Compounds in Cheddar CheeseRobert C. Lindsay, Norman F. Olson, David Bogenrief, Qiaoling Zeng .......................................................................................... 60

Mechanisms for Production of Cheese Flavor CompoundsRobert C. Lindsay, Christine Nowakowski ..................................................................................................................................... 62

Identification of Microbial Enzymes and Metabolites Involved in the Development of LowfatCheddar Cheese Flavor (Phase I and Phase II)James L. Steele, Mark E. Johnson, Jeff Broadbent, Bart Weimer, Kristen Houck, Song Gao, Ed Dudley, Jeff Christensen ........... 64

Contribution of Endopeptidases from Lactobacillus helveticus CNRZ32 to Cheese FlavorDevelopmentJames L. Steele, Kurt M. Fenster, Yo–Shen Chen, Kirk L. Parkin, Mark E. Johnson ...................................................................... 68

Process Modification of Starter Cultures for Flavor Enhancement in Lowfat CheeseMark. R. Etzel, Brian Chi-Shung To ................................................................................................................................................ 70

Evaluating Microstructure of Reduced Fat Cheese by Computer Image ProcessingS. Gunasekaran, N.F. Olson, M. Johnson, K. Muthukumarappan, S. Y. Kim .................................................................................. 72

Machinability of Reduced Fat and Lowfat CheesesS. Gunasekaran, N. F. Olson, K. Muthukumarappan ...................................................................................................................... 76

Structure Function Relationships During Melting and Cooling of Lower Fat CheesesS. Gunasekaran, N. F. Olson, D.J. Klingenberg, R. Subramanian .................................................................................................... 78

Manufacture of a New Reduced Fat Cheese for Use on Pizza PiesCarol M. Chen, Mark E. Johnson, Amy L. Dikkeboom, John J. Jaeggi, William A. Tricomi .......................................................... 81

Minimizing the Watering Off of Unripened High Moisture Lower Fat and No Fat MozzarellaCheeseCarol Chen, Mark E. Johnson, Amy L. Dikkeboom, Kristen B. Houck, John J. Jaeggi, William A. Tricomi .................................... 84

Lower-fat Swiss cheese: Development of a Manufacturing Protocol and the Evaluation ofFlavor DevelopmentCarol M. Chen, Mark E. Johnson, Amy L. Dikkeboom, John J. Jaeggi,William A. Tricomi ............................................................. 86

CDR Specialty Cheese Applications ProgramJames Path, John Jaeggi .................................................................................................................................................................. 89

Improved Quality of Shredded Cheese - Antimycotics, Oxygen Scavengers and ModifiedAtmosphere PackagingRussell Bishop, Joseph E. Marcy, Tina Moler Grove ....................................................................................................................... 90

1996 Wisconsin Cheese Plant Management SurveyBrian W. Gould, Kurt A. Carlson ..................................................................................................................................................... 93

Overview

v

Developing a Graphical Paradigm for Organizing and Delivering Technical Information aboutCheeseJohn Norback, Candelaria Barcenas ............................................................................................................................................... 94

Chapter 4Safety and Quality 97

Biological Significance of Conjugated Dienoic Derivatives of Linoleic AcidM.W. Pariza, Wei Liu, Jayne Storkson, Karen Albright, Kisun Lee, Xiaoyun Yang ......................................................................... 99

Cleanability Assessment of Milking EquipmentDouglas J. Reinemann, Amy C. Lee Wong, Anton Muljadi, John Patoch ...................................................................................... 102

Verification of Dairy Product Safety System (HACCP) Incorporated into CheeseManufacturingEric Johnson, Steve Ingham, Marianne Smukowski, Ann Larson ............................................................................................... 104

Survival and Physiology of Listeria monocytogenes in Commercial BrinesJohn H. Nelson, Eric A. Johnson, Ann E. Larson .......................................................................................................................... 105

Microbiological Safety and Quality of Reduced Fat Cheddar CheeseEric A. Johnson, John B. Luchansky, Alvaro Quinones, Al Degnan, Greg Kulman ....................................................................... 106

Control of Clostridium botulinum and Related Sporeformers in Full Fat and Reduced FatCheddar CheeseEric A. Johnson, Ann E. Larson .................................................................................................................................................... 108

Application of Biopreservatives as Antilisterial Agents in Queso Fresco and Cheddar CheeseJohn B. Luchansky, Mark E. Johnson, Nana Y. Farkye, Alan J. Degnan ......................................................................................... 109

Chapter 5Communications 117

Outreach Events ............................................................................................................................................................................ 119Technical Seminars ...................................................................................................................................................................... 119Scientist exchanges ....................................................................................................................................................................... 120Industry Teams ............................................................................................................................................................................ 120Publications .................................................................................................................................................................................. 120Videos ........................................................................................................................................................................................... 121CDR On-Line ................................................................................................................................................................................ 122


vi

Dear Colleagues,

The key to success is planning and executing research programs withpartners. This has been an exciting year for CDR due to industry teams/consortia actively participating with us in the areas of milkfat fraction-ation, cheese as a food ingredient, and safety and quality. In addition,DMI’s planning of national dairy research included industry, WMMB, andthe Centers in an unprecedented effort to gain total buy-in from theresearchers and the users of the developed technologies.

Development of a totally coordinated national research program willbenefit all those involved, especially the major investors - the dairyproducers of the U.S. Thank you for your continued support.

Sincerely,

J. Russell Bishop, Ph.D.Director, Wisconsin Center for Dairy Research

Overview

vii

Wisconsin Center for Dairy Research Staff

Top Row: Bill Tricomi, Amy Dikkeboom, Tom Rowe, Kristen Houck, John Jaeggi, Tom SzalkuckiMiddle Row: Brian Gould, Carmen Huston, Sarah Quinones, Carol Chen, Karen Paulus, Rusty BishopBottom Row: Dave Bogenrief, Kerry Kaylegian, Linda Hewitt, Mark Johnson

Missing: Jim Path, Marianne Smukowski


viii

CDR was established in 1986 to:

• Provide technical expertise to strengthen the economy of the dairy industry

• Re-establish a focus on dairy research at University of Wisconsin-Madison

• Foster multidisciplinary research and transfer information and technology

• Integrate milk production, processing and marketing research

The Center for Dairy Research

CDR structureCDR is organized into three functional areas — administration, research, and communications. Threecommittees composed of both industry and academic representatives assist with administration andresearch planning.

CDR staffDirector — J. Russell Bishop, Ph.D.Assistant Director — Tom SzalkuckiCarmen Huston, program assistant supervisorLinda Hewitt, program assistant 2Sandra Sekel, program assistant 2

Research staffDavid Bogenrief, researcherKurt Carlson, research specialistCarol Chen, researcherAmy Dikkeboom, research specialistBrian Gould, senior scientistKristen Houck, research specialistBarbara Ingham, research specialistJohn Jaeggi, associate researcherMark Johnson, senior scientistKerry Kaylegian, researcherChristian Kirk, research specialistJim Path, outreach specialistMarianne Smukowski, research specialistWilliam Tricomi, assistant researcher

CommunicationsDawn Hyatt, program assistantKaren Paulus, editorSarah Quinones, outreach program managerTom Rowe, management information specialist

Overview

ix

The Research ProgramCDR sponsors a diverse range of research on dozens of dairy topics including disciplines from geneticengineering to economics. However, the CDR research program focuses on four areas: 1) demand and usefor milkfat, 2) developing new applications for nonfat milk components, 3) cheese technology, and 4) dairyfoods safety and quality.

CommunicationsInformation and technology transfer is an esssential component of CDR. CDR's communications programprovides publications, workshops, seminars, conferences, and scientist exchanges.

CommitteesAdministrative CommitteeThe Administrative Committee is responsible for policy formulation and appointment of the CDR Director.Its members (FY 1995-1996) are:

J. Russell Bishop, CDRJoe von Elbe, Dept. of Food ScienceJanet Greger, Graduate SchoolNeal Jorgensen, College of Agricultural and Life SciencesLeslie Lamb ,WMMBLinda Racicot, DMITom Szalkucki , CDR

Technical Advisory CommitteeThe Technical Advisory Committee (TAC) plans the CDR research program, and evaluates and approvesresearch projects for scientific merit. Members (FY 1995-1996) include:

Bishop, J. Russell, Wisconsin Center for Dairy ResearchBlaska, Gregory, DMI and WMMBBremel, Robert, Department of Dairy ScienceBumbalough, John, Land O’ LakesBurrington, David, Wisconsin Milk Marketing BoardDobson, William D., Dept. of Agricultural EconomicsEtzel, Mark, Dept. of Food ScienceGeyer,James, Foremost FarmsHartel, Richard, Department of Food ScienceHill, Charles, Dept. Chemical EngineeringJohnson, Eric, Dept. of Food Microbiology & ToxicologyJohnson, Mark, Center for Dairy ResearchJorgensen , Neal , College of Agricultural & Life SciencesKrug, David, DMI and WMMBLindsay, Robert, Department of Food ScienceMuck, George, Dean FoodsNey, Denise, Dept. of Nutritional SciencesOlson, Norman F., Department of Food ScienceRacicot, Linda, DMIRose, David, WMMBSellars, Robert, R. L. Sellars & Associates, Inc.Szalkucki, Thomas, Wisconsin Center for Dairy Research


x

Industry Advisory Committee

The Industry Advisory Committee determines the best methods for commercial investment in CDRprojects. Committee members bring an industry perspective to research planning, including a commercialview of the interaction between R&D, marketing, and economics. They include (FY 1995-1996):

Bishop, J. Russell, Wisconsin Center for Dairy ResearchBurrington, David, Wisconsin Milk Marketing BoardBush, Robert, Schreiber Foods, Inc.Byrne, Rob, National Cheese Inst/Amer Butter Inst.Cobian, Francis, Lake O' LakesCords, Bruce, Klenzade/Ecolab Research CenterCrawford, Robert, Borden Foods, Inc.Geyer, Jim, Foremost FarmsKasten, James A., AMPI-Morning Glory FarmsKozak, Jerry, Nat Cheese Inst/Amer Butter InstKrug, David, WMMBLamb, Leslie, Wisconsin Milk Marketing BoardLegreid, Bradley, Wisconsin Dairy Products Association, IncLemmenes, Larry, Alto Dairy CoopMathison, Matt, Systems Bio Industries, Inc.Muck, George, Dean FoodsNilsestuen, Rod, Wisconsin Federation of CoopsRacicot, Linda, DMIRank, Tom, Chr. Hansen’s LaboratoryRose, David, WMMBRice, Harold, DMISellars, Robert, R. L. Sellars & Associates, Inc.Sommer, Dean, AltoStorhoff, Donald, Foremost FarmsUmhoefer, John, Wisconsin Cheese Makers Assoc.Wagner, Dr. Richard, Weyauwega Milk ProductsWeiss, Ronald, FRIWuethrich, Dallas, Grassland Dairy Products, Inc.Yaghoubian, Dr. V. Edward, DMI

Program Area Coodinators

Cheese Technology— Robert Lindsay, Dept. of Food Science, University of WI-MadisonMikfat Utilization—Rich Hartel, Dept. of Food Science, University of WI-MadisonNonfat Solids Utilization—Mark Etzel, Dept. of Food Science, University of WI-MadisonQuality and Safety—Eric Johnson, Food Research Institute, University of WI-Madison

Chapter 1

Summary

Milkfat ResearchAlthough fractionation continues to be a primary feature of CDR’s milkfat program,modifying milkfat is also an important approach to produce value added milkfat products.Charlie Hill and his fellow researchers improved the purification and control of pregastricesterases to produce lipolyzed butteroil. Their immobilized reactor system uses a higherpurity enzyme system which can provide better control of the finished product at a lowercost when compared to the use of free enzyme. An immobilized enzyme can also preventundesirable continued enzymatic activity in the finished product.

In another example of value added milkfat modifications, Parkin and his group have usedenzymes to produce emulsifiers from butteroil. These emulsifiers have physical andfunctional properties similar to commercial monoacylglycerides (MAG). In an associatedproject, the MAG butteroil derived emulsifiers were compared to commercial MAG in icecream and they functioned well.

Despite the imminent commercial introduction of milkfat fractions in the US, CDRresearchers continue to investigate the underlying principles of suspension-type meltcrystallization, which is the process used to produce fractions. Currently, much of theequipment scaling and process parameters are a combination of “art” and science. Re-search conducted by Hartel and others will improve understanding of the process, whichbuilds the science base. Hartel and Ulrich are also investigating surface-layer melt crystal-lization as a method to improve the separation efficiency between solid and liquid phasesfor milkfat fraction production.

Milkfat fractions impart specific desirable characteristics to food. Determining thebenefits of using milkfat fractions continues to be an important part of CDR’s milkfatfractionation program. Research on dairy spreads, chocolates, baked goods and otherproducts is providing the technical information to use fractions to improve products. Toaid these investigations, milkfat fractions are now being produced in Babcock Hall withthe new fractionation pilot plant. This unit will allow us to produce milkfat fractionsamples for industry, as well as university researchers.

Technology transfer

Milkfat Technology Forum, April 23-24, 1996CDR Open House, March 27, 1996


2

Milkfat

3

FINAL REPORT

Use Of Pregastric Lipases Immobilized In A HollowFiber Reactor To Produce Lipolyzed Dairy Products

Personnel: Charles G. Hill, Jr., professor of Chemi-cal Engineering; Dr. Hugo S. Garcia, visitingscientist, associate professor, Department of FoodTechnology, Centro de Graduados, InstitutoTecnologico de Veracruz, Veracruz, Mexico; LouisLessard, research assistant, Department of Chemi-cal Engineering; Souheil Ghannouchi, researchassistant, Department of Chemical Engineering.

Dates: July 1994 – June 1996

Funding: Dairy Management Inc. HLL 95

Objectives

1. Generate experimental data to characterize therates of the lipolysis reactions constituting thereaction networks of interest. Determine the effectsof temperature and pH on both the overall rate ofhydrolysis and reaction specificity for each lipase ofinterest.

2. Use these kinetic data to develop bothuniresponse and multiresponse mathematicalmodels of the reaction network to use in processdesign and simulation, control, and optimizationanalyses. Both intrinsic kinetics (including inhibi-tion effects) and enzyme deactivation effects willbe quantified.

Summary

Our efforts on this project involved two aspects:

1. Developing a protocol for purifying and immobi-lizing pregastric esterases derived from the salivarytissue of calves, lambs and kid goats. These en-zymes are of particular interest because of theirhigh specificities for release of short chain fattyacids from fats and oils.

2. Use of the indicated immobilized enzymes tohydrolyze butteroil using a hollow fiber reactor

configuration. This involves generating kinetic dataand developing mathematical models to fit thesedata.

This project will continue using funds fromProfessor Hill’s NSF grant in order to obtaincomplementary information which will facilitatefurther development of this technology.

Purification Studies

The crude enzyme preparations obtained from acommercial source were purified prior to immobili-zation. This purification process gives a specificactivity which is an order of magnitude greaterthan the commercial precursor. It consists of thefollowing steps:

1. Refrigeration of a suspension of the crude tissuepreparation in a phosphate extraction buffer tofacilitate dissociation of the esterases from the largehydrophobic proteins with which they are normallyassociated in the salivary tissue.

2. Centrifugation of the suspension, followed bymicrofiltration through a 0.2 µm membrane toremove suspended solids.

3. Ultrafiltration of the permeate from step 2 usinga membrane characterized by a molecular weightcutoff value of 30,000 daltons.

The resultant concentrated solution of the enzymepossesses a much higher enzyme activity per unitof soluble protein than its precursors. In previouspublications we have described how to use thissolution to immobilize the esterase via physicaladsorption on the polypropylene hollow fiberscontained within the reactor. This approachdrastically improves the efficacy of fatty acidrelease. Figure 1 depicts the free fatty acid contentof the effluent butteroil stream as a function of thespace time (residence time) of the butteroil in thereactor. The closed circles represent data obtained


4

using an esterase purified by the protocol describedabove. The smooth curve represents the best fitcurve through data obtained using an earlierprotocol for purification of the esterase. Themarked improvement in the performance of thereactor implies that if this technology is used,significantly shorter contact times would berequired to achieve the concentrations of free shortchain fatty acids characteristic of commerciallipolyzed butteroil products. Using this approach,one could take advantage of the economic benefitsof employing immobilized enzyme technology toproduce a lipolyzed butteroil rather than the freeenzyme process traditionally employed by com-mercial manufacturers of such products.

Note that the stabilities of the immobilized en-zymes generated via our current protocol aresufficient for the proposed commercial application.In an experiment involving continuous operation at45°C for a week, there was no significant change inthe effluent composition during the course of theexperiment. There was no statistically significantchange in the activity of the immobilized calfpregastric esterase during the week.

0

50

100

150

200

0 1 2 3 4 5 6 7 8

µm

ol

ffa

/mL

Space Time (h)

a

FIGURE 1. Production of free fatty acids by the hollow fiber reactors loaded with calf PGE (a)and kid PGE (b). The solid lines correspond to the best fits of kinetic data obtained using animmobilized enzyme purified using only microfiltration while the closed circles correspond toimmobilized enzymes purified via both microfiltration and ultrafiltration.

Kinetics Studies

We have characterized the kinetics of the lipolysisof butteroil in the presence of immobilized forms ofthree pregastric esterases. In this work butteroiland McIlvane buffer flow in concurrent fashionthrough the shell and tube sides of a hollow fiberreactor. The enzyme is immobilized within thewalls of the porous propylene which constitutes thehollow fibers. The reactor configuration can beviewed as a miniature version of a shell-and-tubeheat exchanger. We have probed the effects ofreactor space time (fluid residence time), pH,temperature and type of enzyme on both the totalamount of free fatty acids released and the distri-bution of fatty acids. Some of these results aredescribed in manuscripts submitted for publicationin archival journals or presentation at professionalsociety meetings. Representative results for thesepregastric esterases are described below.

0

50

100

150

200

250

300

0 1 2 3 4 5 6 7 8

µm

ol

ffa

/mL

Space Time (h)

b

Milkfat

5

Effect of pH on the fatty acidspecificity of immobilized kid goatpregastric esterase.

For the kid goat esterase, the optimum operatingtemperature is 40°C. Data for the total concentra-tion of free fatty acids in the reactor effluent foroperation at this temperature are presented inFigure 2. These data indicate the effects of reactorspace time and pH on the overall extent of reaction.

Inspection of Figure 2 indicates that the effects ofpH become more evident at space times longerthan 5 hours. At space times of less than 4 hours thedifferences in conversion at different pH valueswere relatively small. A first order kinetic modeldeveloped by our research group was employed inthe regression analysis of the data to obtain thesmooth curves in Figure 2.

Figure 2. Effect of pH on the overall extent of hydrolysis of butteroil at various reactor spacetimes.

0

0.2

0.4

0.6

0.8

1

0 2 4 6 8 10 12 14

pH 5.0pH 6.0pH 7.0

FF

A

(mm

ol/

mL

)

Space Time (h)

Analyses for individual free fatty acids in thereactor effluent indicated that butyric acid isreleased faster at pH 6.0 than at pH 5.0 or 7.0. Forthe range of space times we investigated, mediumand long-chain fatty acids were released to only alimited extent. A good criterion for assessing flavorbalance involves calculating different ratios of freefatty acids. Ratios of particular interest (C4/C8, C4/C6, C6/C8, and C4/C12) are presented in Table 1.The first two ratios can be used as a measure ofdesirable flavor and flavor notes. Statistical analysis

by ANOVA and Fisher’s paired test indicated thatthe differences in the ratio C6/C8 at the three pHvalues were not statistically significant. High valuesof C4/C6 and C4/C8 correspond to intense, butdesirable, flavors. By contrast, low values of theratio C4/C12 can be used as an indication of soaplike (undesirable) flavors associated with anexcessive amount of C12. In our work, the ratio C4/C6 remained substantially constant as the spacetime increased whereas the ratio C4/C8 increased.However, the latter ratio approached an asymptotic


6

value for space times in excess of ca. 1 hour. Themain advantages of our immobilized configurationinclude facilitating continuous production oflipolyzed butteroil to give a relatively constantcomposition of the effluent and permitting users toadjust the reactor space time to allow strict controlof the extent of lipolysis.

The free fatty acid compositions of three commer-cial LBO samples are presented in Table 2. Thesesamples were probably prepared with either calf,kid or lamb PGE. This table also contains data fortwo different effluent streams from our reactor.These two samples were selected because theycontain concentrations of free butyric (C4)

Table 1. Average ratios of fatty acids for reactor space times from 1 to 17 hours at threedifferent pH values. Different superscripts indicate statistically different ratios as calculatedby ANOVA (p<0.05).

Ratios of fatty acids pH 5 pH 6 pH 7

C4/C6 1.79ab 1.91a 1.45b

C4/C8 4.06cd 4.77c 2.96d

C6/C8 2.34e 2.39e 2.07e

C4/C12 5.07f 6.49g 3.97f

Table 2. Free fatty acid compositions (µmol/mL) of three different commercial LBO’s andtwo samples obtained from the immobilized kid PGE that contain similar amounts of C4-C6 fatty acids.

Fatty Acid pH 5,5.2 h LBO-95 LBO-50 pH 6, 11.8 h LBO-110

C4 54.14 55.83 66.63 85.18 80.05

C6 24.88 21.83 47.09 37.68 43.30

C8 8.87 12.60 30.16 11.89 26.68

C10 11.94 19.02 43.92 14.86 40.40

C12 7.74 15.37 38.23 10.27 36.97

C14 8.26 31.83 130.6 11.83 114.5

C18:2 ** 18.61 51.6 ** 45.44

C16 8.91 34.0 154.6 12.58 138.7

C18:1 11.23 32.65 136.9 14.49 136.5

C18 4.71 10.23 36.35 5.24 31.19

C4/C6 2.17 2.56 1.41 2.26 1.85

C4/C8 6.10 4.43 2.21 7.16 3.00

C6/C8 2.80 1.72 1.56 3.17 1.62

C4/C12 6.99 3.63 1.74 8.29 2.16

SC4-C10 99.83 109.28 187.80 149.61 151.43

SC12-C18 40.85 142.69 548.28 54.41 503.3

** not detected

Milkfat

7

and caproic (C6) acids similar to those of the LBOsamples located to their right. By varying theoperating pH, you can produce effluents withdissimilar fatty acid profiles. Similarities betweenthe LBO’s and the immobilized PGE product shouldbe further validated by performing analyses for thevolatile free fatty acids (straight and branchedchain fatty acids) as well as nonacidic components.

It is worth noting that our samples containedsignificantly less C12 than the corresponding LBO.This suggests that our lipolyzates could havecleaner and less soapy flavor tones than the com-mercial products. Inspection of the tabular entriesfor short (C4-C10) and long (C12-C18) chain fattyacids, reveals very substantial differences in thelong chain fatty acid contents of our products andthe corresponding commercial LBOs, even thoughthe concentrations of short chain fatty acids arecomparable. The commercial LBO’s are character-ized by elevated concentrations of C12-C18 fattyacids. Although the same enzymes may have beenused to prepare both types of samples, there is amarked difference in the selectivities of theseenzymes for release of short chain fatty acids.

It is clear that the immobilized PGE system permitsyou to tailor the product composition by selectingoperating conditions; however, detailed sensoryassessment of the products for dairy type flavorsremains to be acomplished. Also, it will be neces-sary to analyze for the presence of other volatilecompounds contributing to specific flavor notescharacteristic of lipolyzed dairy products.

Kinetic behavior of the immobilizedlamb pregastric esterase

For the immobilized lamb pregastric esterase, theeffects of temperature and pH on the total quantityof free fatty acids released during hydrolysis ofanhydrous butterfat are shown in Figures 3 and 4,respectively. Even though the starting enzyme was acrude extract, fatty acids are released to the extentof 450 µmoles per mL of butterfat at a space time of15 hours. This release corresponds to approxi-mately 20% of the total number of fatty acidresidues present in the original butterfat. In orderto generate the data necessary for the reactor

modeling studies, experiments were conducted atseveral different values of pH and temperature. Inmost cases, the effects of temperature and pH onthe rate of reaction were readily distinguishable,even at short residence times (or holding times forthe batch reactor). We selected the ranges oftemperature and pH to investigate on the basis ofearlier work with immobilized lipases and theobservation that the activity of the lamb esterasedecreases markedly for pH values less than 5 andgreater than 6.5. Both hollow fiber and batchreactor experiments indicated that the optimumconditions (maximum conversion) correspond to apH of ca. 6.0 and a temperature of 45°C. At tem-peratures above 45°C, thermal deactivation of theenzyme adversely affects the rate of the hydrolysisreaction.

A mathematical model based on Michaelis-Mentenkinetics and a ping pong bi bi mechanism providesan appropriate fit of the data. When the reactor wasoperated continuously at 45°C for six consecutivedays, the immobilized enzyme did not suffer anyappreciable loss of activity. Hence the immobilizedlamb pregastric esterase appears to be sufficientlystable for industrial use in this application. Notethat we observed clogging of the hollow fibers afterlong space times (high conversions). This problemis attributed to precipitation of fatty acids (withhigh melting points) whose solubility limits wereexceeded. You can avoid this problem by periodi-cally operating at high flow rates to flush thecrystals from the reactor, or by operating at lowerconversions. We are currently in the process ofanalyzing the HPLC data from the multiresponseexperiments in order to determine the operatingconditions that optimize the release of those fattyacids which are primarily responsible for genera-tion of the desired flavor notes.

Kinetic behavior of the immobilizedcalf pregastric esterase

For the calf pregastric esterase, the method ofpurification and immobilization was optimized inorder to maximize the activity retained within thehollow fiber reactor. Optimum operating conditionsare a buffer pH of 6.0 and a temperature of 40°C.Under these conditions, the half-life of the enzyme


8

0

50

100

150

200

250

300

350

0 2 4 6 8 10 12 14 16

pH = 5.0pH = 5.5pH = 6.0pH = 6.5

Free

Fat

ty A

cid

Con

cent

ratio

n (m

mol

/L)

Space Time (hr)

Figure 4. Effects of pH and reactor space time on the overall extent of hydrolysis at 40°C

0

100

200

300

400

500

0 2 4 6 8 10 12 14 16

T = 40°CT = 45°CT = 50°C

Free

Fat

ty A

cid

Con

cent

ratio

n (m

mol

/L)

Space Time (hr)

Figure 3. Effects of temperature and reactor space timeon the overall extent of hydrolysis at pH 6.0

Milkfat

9

was 70 hours, and the maximum amount of fattyacids released was 120 µmoles per milliliter ofmilkfat at a reactor space time of 20 hours. Theseconditions correspond to release of approximately5% of the total number of fatty acid residues. Underthese conditions, the amount of butyric acidresidues released was 50 µmoles per milliliter ofmilkfat, a result which corresponds to release of ca.60% of the original butyric acid residues.

Because the pregastric esterase has a high specific-ity for releasing short chain fatty acids, the longchain fatty acids in the lipolyzed butterfat arepresent at very low concentrations (less than 10µmol/mL of fat). A new HPLC method has beendeveloped to analyze long chain fatty acids at theselow concentrations. Kinetic data for the lipolysisreaction at 40°C have also been collected at pHvalues of 4.0, 5.0, 6.0 and 7.0 for the calf pregastricesterase. The optimum pH for the hydrolysisreaction is near pH 6.0.

Significance to the dairy industry

The use of immobilized enzyme technology toproduce lipolyzed butter oils offers several advan-tages over the conventional batch process based onthe use of soluble enzymes. From an economicstandpoint, you can produce far more product perunit of enzyme. In addition, using immobilizedlipases offers the potential to develop products withunique sensory, functional, or nutraceuticalproperties by using lipases with different specifici-ties, either alone or in combination with oneanother.

Hydrolysis and interesterification reactions repre-sent techniques to manipulate the chemicalcomposition of milkfat to design foods withspecified physiological functions. There is agrowing awareness of the role of foods in humanhealth. These reactions can replace undesirablefatty acids in the triglycerides of milkfat withresidues which have positive physiological benefitsto produce nutraceuticals or pharmafoods. Forexample, investigators have attributedanticarcinogenic and antioxidant properties toconjugated linoleic acids. The modified milkfatproducts of interest thus have significant dietary

implications with respect to nutrition, flavorgeneration, and potential anti-cancer activity. Usingimmobilized lipases offers the intriguing possibilityof producing specially designed foods for selectedsegments of the population – particularly individu-als who are especially health conscious from adietary standpoint or are high risk candidates forcardiovascular, hypertensive, or other healthproblems. These products represent a very signifi-cant long term marketing opportunity for the dairyindustry.

Publications/Presentation

Manuscripts accepted for publication

“Immobilization of Pregastric Lipases in a HollowFiber Reactor for Continuous Production ofLipolyzed Butteroil,” by H.S. Garcia, A. Qureshi, L.Lessard, S. Ghannouchi, and C.G. Hill, Jr.,Lebensmittel-Wissenschaft und Technologie, 28, 253(1995).

“Improving the Continuous Production ofLipolyzed Butteroil with Pregastric EsterasesImmobilized in a Hollow Fiber Reactor,” by H.S.Garcia and C.G. Hill, Jr., Biotechnology Techniques, 9,467.

“Bioseparacíon e Immovilizacíon de EsterasPregástericas para la Produccíon de Grasa ButíricaLipolizada,” by H.S. Garcia and C.G. Hill, Jr., inFronteras en Biotecnología y Bioingeniería (E.Galindo, editor), Sociedad Mexicana deBiotecnología, A.C., Mexico City, 1996 (in press).

Manuscript Submitted for Publication

“Effect of pH on the Fatty Acid Specificity ofImmobilized Kid Goat Pregastric Esterase,” by H.S.Garcia and C.G. Hill, Jr., submitted for publicationin Biotechnology Techniques.

Presentations at Professional Society Meetings

“Production of Lipolyzed Butteroil by ImmobilizedCalf and Kid Goat Lipases,” by H.S. Garcia and C.G.Hill, Jr., presented at the 90th Annual Meeting of theAmerican Dairy Science Association, June 25-28,1995, Ithaca, New York.


10

“Production of Lipolyzed Butteroil by ImmobilizedLipases,” by H.S. Garcia and C.G. Hill, Jr., invitedpaper presented at the Symposium onBioseparations, Ixtapa, Mexico in September, 1995.

“Effect of pH on the Fatty Acid Specificity ofImmobilized Kid Goat Pregastric Esterase,” by H.S.Garcia and C.G. Hill, Jr., presented at the AnnualMeeting of the Institute of Food Technologists,June, 1996.

Papers Submitted for Presentation at ProfessionalSociety Meetings:

“Membrane Reactors: Ceramic Modules andImmobilized Enzyme Systems,” by C.G. Hill, Jr., F.Tiscareño-Lechuga, F.X. Malcata, K.E. Rice andM.A. Anderson, invited paper to be presented at theOctober, 1996 meeting of the CongresoMediterráneo de Ingeniería Química.

“Hydrolysis of Naturally Occurring Fats and Oils byEsterases Immobilized in a Hollow Fiber Reactor,”by C.G. Hill, Jr., H.S. Garcia, K.E. Rice, F.X. Malcata,L. Lessard, and S. Ghannouchi, invited paper to bepresented at the 6th Biochemical EngineeringConference, September 1996 at the Korea AdvancedInstitute of Science and Technology.

“Kinetics of the Lipolysis of Butteroil in a HollowFiber Reactor Containing an Immobilized Esterase,”by S. Ghannouchi and C.G. Hill, Jr., paper submittedfor presentation at the November 1996 AIChEmeeting.

Seminars

“Hydrolysis of Milkfat by Lipase Immobilized in aHollow-Fiber Membrane Reactor,” presented at TheUniversity of Illinois at Chicago, Department ofChemical Engineering, April 7, 1995.

“Recent Research Results Concerning the Produc-tion of Lipolyzed Butteroils Via ImmobilizedEnzyme Technology” Presented to a review panelconsisting of industry representatives and research-ers concerned with the Milkfat Fractionation andUtilization program of the University of WisconsinCenter for Dairy Research, the “Milkfat IndustryTeam” (May 26, 1995).

“Chemical Modification of Milkfat Via ImmobilizedEnzyme Technology”, Milkfat Fractionation andUtilization Conference sponsored by the Universityof Wisconsin Center for Dairy Research (April,1996).

Milkfat

11

VISITING SCIENTIST REPORT

The Effect of pH on the Fatty Acid Specificity ofPregastric Esterase from Kid Goat

Personnel: Hugo S. Garcia, visiting scientist,Instituto Tecnologico de Veracruz, Veracruz, Mexico,Charles Hill, professor, Dept. of Chemical Engineer-ing

Dates: June – July 1995

Objectives

1. Examine the influence of pH on the specificity ofone of the enzymes currently used by the flavorindustry, the kid goat pregastric esterase.

Summary

Work done during the 1993-1995 time frame hasprovided valuable information on procedures forpartial purification of pregastric esterases andimmobilization of these enzymes. In the summer of1995 our primary focus was an examination of theinfluence of pH on the specificity of one of theenzymes currently used by the flavor industry, thekid goat pregastric esterase. The relative rates ofrelease of various fatty acids during lipolysiscatalyzed by this enzyme at 40 ºC in a hollow fiberreactor were measured by HPLC using the methodpublished previously by our group. Total fatty acidsreleased were also monitored by alkali titration.

The reactor and procedures used in the indicatedresearch were the same as those described inprevious publications: We employed space times aslong as 13 hours and buffer pH values of 5, 6 and 7.Previous work indicated that this pH range corre-sponded to the highest activities of the immobi-lized enzyme. Measurement of the total amount offatty acids released indicated that the highestconversions were obtained at pH 6.0. The differ-ences were more evident at space times longer than5 hours. The differences in conversion were quitesmall at space times of less than 4 hours.

Analysis of individual free fatty acids in the reactoreffluent indicated that butyric acid is released faster

at pH 6 than at pH 5 or 7 and that medium andlong-chain fatty acid residues were released to onlyvery limited extents for the range of space timesinvestigated. A good criterion for assessing flavorbalance involves calculating different ratios of freefatty acids, for example C4/C8 and C4/C12. The firstratio provides a measure of a desirable flavor andflavor notes can be defined by the range of thisratio. High values of C4/C8 correspond to anintense, but desirable flavor. By contrast, low valuesof the ratio C4/C12 can be used as an indication ofsoapiness or undesirable flavors arising from thepresence of an excessive amount of C12.

The C4/C8 ratio increased with increasing spacetimes. However, the ratio approached an asymptoticvalue for space times in excess of ca. 1 hour.Comparison of the ratios obtained while operatingwith different buffer solutions indicated thathydrolysis at pH 6 gave C4/C8 and C4/C12 ratios(4.77 and 6.49, respectively) which differed fromthose obtained at pH 5 (4.06 and 5.06) and 7 (2.96and 3.97). These differences indicate that the sameenzyme/substrate system can be employed toproduce significantly different and distinct flavorprofiles under slightly different reaction conditions.This result requires further assessment by the usersof this technology to modify the flavor profiles oftheir products without having to make substantialchanges in their processing conditions. Comparisonof the free fatty acid contents and pertinent productratios for our effluent with those of commercialLBO samples provided by SANOFI indicated thatthe commercial LBO’s contained large amounts offree medium and long chain fatty acids. In somecases, e. g., stearic acid (C18), oleic acid (C18:1),palmitic acid (C16) and myristic acid (C14), theconcentrations of these fatty acids were an order ofmagnitude or more greater that the concentrationspresent in our product. Although the C4/C8 ratiosof two of the three LBO’s were similar to those ofour products, the C4/C12 ratios of our productswere always more favorable. This result may implythat our lipolysates may be characterized as


12

“cleaner” dairy flavors than the commercialsamples. However, this statement will have to bevalidated by strict sensory evaluation tests involv-ing both lipolysates.

In a different series of experiments, the secondretentate of the kid lipase was subjected to sizeexclusion chromatography. Most of the lipolyticactivity was associated with the fraction corre-sponding to proteins with molecular weightsgreater than 100,000. Separations of the samesecond retentate using ultrafiltration membraneswith nominal molecular weight cutoff values(equivalent to pore sizes) of 300,000 and 100,000.The permeate obtained by centrifugation did notcontain any measurable lipolytic activity. Thisresult implies that in the light of the known mo-lecular weight value for pregastric esterases (ca.43,000 as determined by SDS-PAGE), the kidsalivary lipases used in our studies are likely to bepresent in oligomeric forms. Thus, the use ofultrafiltration membranes with a molecular weightcutoff values greater than 30,000 is sufficient toeffect the concentration step. We suggest usingmembranes with a nominal cutoff of 100,000. Thisapproach provides higher transmembrane flowrates which would make the preparation processmore efficient.


Results obtained in the summer of 1995 indicatethat detailed work is needed to model the selectiv-ity of the kid pregastric esterase with respect to pH.Subtle changes in process conditions, such as a pHunit change of one, lead to different fatty acidprofiles which, in turn, may lead to differentproducts with different sensory properties.

The free fatty acid profiles obtained with theimmobilized lipase reactor contained a much lowerproportion of medium and long chain fatty acidsthan those of commercial LBO samples. This factimplies that the flavor profiles obtained with theimmobilized form of the enzyme are “cleaner” thanthose of products obtained using the same enzymein its free form with longer reaction times.

The observation that the kid goat pregastricesterase exists in oligomeric form should allow theuse of ultrafiltration membranes, with a molecularweight cut-off of 100,000 da., during the purifica-tion procedure. Such membranes will allow one toachieve higher values of both the transmembranefluxes and the specific activity of the concentratedenzyme.

Milkfat

13

FINAL REPORT

Evaluation of Monoacylglycerols Derived fromButteroil as Emulsifiers in Foods

Personnel: Kirk L. Parkin, associate professor,Kathleen Wiederholt, research specialist, (7/1/93-10/21/94) replaced by Purwiyatno Hariyadi andShu-Jung Kuo, research assistants, Department ofFood Science; Marit Reierstad, Honorary Fellow(graduate student), Agricultural University ofNorway

Dates: July 1993 – October 1995

Funding: Dairy Management Inc. PRK94

Objectives

1. Prepare monoacylglycerol (MAG)-rich fractionsfrom butteroil at levels suitable for functionalityand applications evaluation.

2. Develop procedures to partially purify and/orfractionate MAG-rich fractions prepared frombutteroil for functionality and applications testing.

3. Evaluate the physicochemical properties andrelated functionality of MAG-rich fractions pre-pared from butteroil as emulsifying agents.

4. Evaluate the effectiveness of selected MAG-richfractions prepared from butteroil in selected foodsystems.

Summary

Objective 1: Two basic reaction configurations wereevaluated for several lipase sources for their abilityto produce MAG (and DAG) blends derived frombutteroil fatty acids. Glycerolysis reactions usedbutteroil and glycerol as substrates, whereasesterification reactions used glycerol and fatty acids(FA) derived from butteroil (by saponification) assubstrates. In the latter case, the fatty acid composi-tion of the starting substrate was diminished inshort chain-length fatty acids because of partiallosses during saponification.

Of the enzymes evaluated, both soluble andimmobilized Rhizomucor miehei lipases (Palataseand Lipozyme, respectively) were capable ofyielding 85-90% [MAG + DAG] with >50% MAG inas little as 10 hr reaction time, in esterificationreactions. Pancreatic lipase, the other food-gradeenzyme evaluated in this manner, could yield only30-50% [MAG + DAG], primarily as DAG, in 48 hrreaction times. Non-food-grade enzymes fromPseudomonas sp (type AK and PS-30), and Candidarugosa yielded 60-70% [MAG + DAG] in as little as4-10 hr reaction time, but the principal product wasDAG.

In glycerolysis reactions, only the soluble R. mieheilipase was capable of yielding >70% [MAG + DAG],requiring 40-80 hr reaction time to do so, with thefungal preparations (Genus/species unknown)“Lipase Praparat” and “Lipolact K1” being lesseffective. Pre-gastric lipases in this reactionconfiguration were not effective. Previous work inthis laboratory established that the Pseudomonaslipases (Amano, type AK and PS-30) were the mosteffective in yielding [MAG + DAG] from butteroil inglycerolysis reactions. In a recent trial, Novozyme435 lipase (a Candida antarctica lipase cloned intoAspergillus oryzae, potentially a food-grade prepa-ration) showed promise in glycerolysis reactions,yielding up to 50% MAG after 24-48 hr reactiontime.

Objective 2: Starting with about 100 g of butteroil,an MAG-rich (>90%) fraction of about 40 g wasprepared using various lipases. The >90% pureMAG fraction was obtained from the final productmixture by repeated precipitation with 4-5 volumeshexane, followed by filtration. Amano type PS-30(Pseudomonas cepacia) and Novozym 435 (Candidaantarctica B lipase produced by a cloned Aspergillusoryzae strain) lipases were effective in this manner.Two food-grade lipases, Novo’s Lipozyme IM andPalatase M (both from Rhizomucor miehei) are alsoeffective, but only if butteroil was first hydrolyzedand the free fatty acids (FA) then combined with


14

glycerol and catalyst in an esterification reactionconfiguration.

Both PS-30- and Palatase M-derived MAG prepara-tions were evaluated fully by separating the productmixture into triacylglycerol (TAG), FA, DAG andMAG components by chromatography using silicicacid and aminopropyl-bonded solid phase car-tridges. Both enzyme-derived MAG preparationsbecame enriched in myristic, palmitic and stearicacids, and diminished in shorter chain-length andunsaturated fatty acids, compared to nativebutteroil (indicated by “product mixture”) (Figures1 and 2). Palatase-mediated processes tended toconcentrate C4-6 FA in the DAG/FA fractions, C8-12FA in the TAG/FA fractions, and C18:1-2 FA in theTAG/DAG fractions. The PS-30-mediated processestended to concentrate C4-6 FA in the TAG/DAGfractions, C8-12 in the FA fraction, and C18:1-2 inthe TAG/DAG/FA fractions. Thus, preparing MAGfrom butteroil in this fashion also resulted in aneffective fractionation of fatty acids into differentpools. High- mid- and low-melting butteroilfractions (HMF, MMF and LMF, respectively) werealso used as starting materials for preparing MAGfractions. Relative to MAG prepared from nativebutteroil, MAG from LMF was enriched in C8-12,18:1 FA and diminished in C14-16 FA; MAG fromMMF was enriched in C8-12, 18:1 FA and dimin-ished in C14-16 and C4-6 FA; MAG from HMF wasdiminished in C4-6 FA and enriched in C8-12 FA.

Objective 3: Hexane-precipitated MAG (66-74%purity) from the enzyme-modified butteroilpreparations were compared with a commercialMAG (>90%) preparation for physicochemicalanalyses. The commercial MAG preparation waseffective at reducing surface tension at an oil-waterinterface (at levels of MAG addition of 0.05-0.5% ofthe oil phase) to 15-5 mN m-1. (The commercialMAG product was 70, 17 and 11% in 18:1, 18:0 and16:0, respectively). MAG prepared from the high-and mid-melting butteroil fractions were asfunctionally capable as the commercial emulsifier,whereas MAG prepared from low-melting fractionand native butteroil were slightly less effective atreducing surface tension.

Solid fat content of the MAG-rich fractions indi-cated that they had behavior intermediate to the

commercial MAG preparation and native butteroil(Figure 3). MAG-rich fractions from HMF ofbutteroil had greater solidity than MAG preparedfrom LMF. However, the most important featurewas the relatively constant % solids content for thebutteroil-derived MAG preparations over the rangeof 0-35ºC, indicating a substantial plastic range.This behavior would be promising for the develop-ment of fat-based table spreads. At a 0.5% level ofaddition to native butteroil, MAG-rich fractionsprepared with Palatase tended to slightly increasesolidity over the range of 0-15ºC, whereas MAG-rich fractions prepared with PS-30 tended toslightly suppress solidity over the same tempera-ture range. Between 20-40ºC, there was little effectof MAG addition on melting behavior of butteroil.

Emulsion stabilizing ability was tested in 6, 12 and20% water-in-oil model systems using 0.5% MAG(based on oil phase). Stability was greatest in 6%water systems, and there was little difference instabilizing effect among the butteroil-derived andcommercial MAG preparations. All MAG prepara-tions were not very effective in stabilizing theseemulsions relative to the influence that a morehydrophilic emulsifier, tween 80, had on stabilizingthese emulsions.

Starch-complexing activity was evaluated using a0.19% amylose solution to which 0.012, 0.024 and0.048% different MAG preparations (84-90% purityin MAG) were added. Amylose-complexing abilityof the MAG preparations derived from the LMF andMMF of butteroil was greater than or equal to thatof the commercial MAG product at all levelsevaluated, whereas the MAG preparation derivedfrom the HMF of butteroil was least effective(Figure 4). These results indicate the ability of thebutteroil-derived MAG preparations to be effectiveanti-staling or crumb-softening agents in bakedgoods.

Objective 4: Preliminary studies indicated that thebutteroil-derived MAG was equally as effective as acommercial emulsifier blend when used in icecream (these results evolved into a project to focusspecifically on this application - see Final ProjectReport by Hartel and Parkin).

Milkfat

15

Figure 1. Palatase-mediated partitioning of fatty acyl groups

Figure 2. PS30-mediated partitioning of fatty acyl groups


16

The effect of the MAG-rich fractions incorporatedinto a simple cake product (48% each water andflour, 2.8% oil and 0.48% MAG preparation)indicated that staling (textural hardness as mea-sured with a TA.XT2 analyzer) took place to similarextents over a 7 day period at 4ºC for samplesformulated with MAG-rich fractions from butteroil,butteroil fractions or a commercial preparation.Preliminary studies with reduced fat (40, 60%)table spreads also indicated a similar degree ofemulsion stabilization by MAG-rich prepared frombutteroil compared to commercial MAG products.Samples of MAG-rich preparations derived frombutteroil were also supplied to Dr. W. James Harperat The Ohio State University for applications testingin baked goods.


These studies indicate that food-grade MAG-richpreparations can be enzymically derived frombutteroil and/or butteroil fractions, and that thesepreparations have physicochemical and functionalproperties similar to commercial MAG productsavailable from other sources. Thus, the opportunityexists to increase the utilization of butteroil-derivedresources in markets that are conventionallydominated by nondairy resources. Focus in futureproduct development efforts should be placed on 1)identifying appropriate products/markets in whichto utilize butteroil-derived MAG, and 2) exploitingany unique (or dual, including any flavoring impactof butteroil) functionality that butteroil-derivedMAG can offer.

Figure 3. Melting behavior

Figure 4. Amylose-complexing properties

Milkfat

17

FINAL REPORT

Application of Emulsifiers From Butteroil or MilkfatFractions in Ice Cream

Personnel: R.W. Hartel, associate professor, K.L.Parkin, associate professor, B. Liang, researchassociate, Dept of Food Science

Dates: September 1994 – August 1996

Funding: Dairy Management Inc. HRP 95

Objectives

1. To study the application of emulsifiers madefrom enzyme modified butteroil and milkfatfractions in ice cream.

2. Evaluate the cost of producing the butteroilemulsifiers and compare to commercial emulsifiersfor use in ice cream.

Summary

An enzyme process, developed at the University ofWisconsin, for producing mono- and di-glyceridesfrom butteroil has been used to produce emulsifiersfor ice cream. For this research, the procedureinvolved complete removal of fatty acids from thetriglyceride (TAG), followed by controlled enzy-matic reesterification, using Palatase M 1000 L(Novo Nordisk Bioindustrials, Franklinton, NC) toproduce mono- (MAG) and di-glycerides (DAG).MAG and DAG were then separated in hexane toprovide emulsifiers with high MAG or DAG content.Different blends of MAG/DAG were evaluated in icecream by mixing the two purified products atdifferent proportions. These blends were character-ized by their MAG content, and varied from 10 to87% in the final emulsifier. These emulsifierscontained only a few percent triglycerides and freefatty acids. In addition, emulsifiers were producedusing the same technique but with milkfat fractionsas the starting materials.

Milkfat fractions for this study were produced by alab-scale dry fractionation procedure. Anhydrousmilk fat (AMF) was melted at 60°C, cooled to an

initial crystallization temperature of 28°C, andallowed to crystallize for 24 hours. The crystalslurry was vacuum filtered to produce a high-melting fraction (solid) and liquid phase. Theliquid phase was cooled further to 20°C for asecond fractionation, and the second liquid fractioncooled further to 14°C for a third fractionation.This resulted in a total of 6 fractions (28S and 28L;20S and 20L; and 14S and 14L), as well as the initialmilk fat, available for modification into mono- anddiglycerides. Significant differences in fatty aciddistributions were observed among the milk fatfractions. Higher melting fractions contained morelong-chain saturated fatty acids, while lower-melting fractions contained shorter-chain andunsaturated fatty acids. Melting points of thesefractions varied between 17 and 45°C, dependingon fatty acid composition.

Ice creams were produced with these emulsifiers ina soft-serve freezer set for 50% overrun. Emulsifierswere added to the mix at 0.1%, and compared toseveral commercial emulsifiers. In addition, an overemulsified ice cream was made with 1% of abutteroil emulsifier containing 83% MAG. Icecreams were evaluated at draw, after hardening andafter storage at -15°C. Measurements includedrelative amount of fat destabilization (using aspectrophotometric technique), melt down rate(drip test), hardness (by penetrometer), ice crystalsize distribution, and air cell/bubble distribution.Ice crystals were measured using a refrigeratedglove box, optical microscopy technique, while aircells (or the bubbles arising from the air cells) weremeasured using a modified “squash” method,followed by image analysis from optical photomi-crographs.

Results

Significant differences were found in rate ofreesterification, and type and level of glycerideproduced when using different milk fat fractions asthe starting material. It was found that MAG and


18

DAG could not be produced by reesterification ofthe fatty acids from the lowest melting pointfractions (14L and 20L). Reesterification of thesecomponents resulted in formation of TAG at allexperimental conditions evaluated. Thus, emulsifi-ers were not produced from these two fractions.Typically, yields of MAG in the reesterificationprocess varied from 30 to 54%, while yields of DAGvaried from 5 to 19%. We found that reactiontemperature was the most critical parametercontrolling the reesterification reaction, with highertemperature leading to higher production of DAGand TAG and lower temperature leading to higher

production of MAG. For ice cream manufacture, 11different emulsifiers were prepared. Seven emulsifi-ers were prepared with different MAG/DAG ratio(from 10 to 87% MAG) by mixing the productsseparated from reesterification of the original AMF(labeled AMFE in Table 1). In addition, emulsifierswere produced from four milkfat fractions (28S,28L, 20S, and 14S), with varying amounts of MAGand DAG depending on their fatty acid composi-tion. The breakdown of lipid components for these11 emulsifiers, as well as for 5 commerciallyobtained emulsifiers, is shown in Table 1. Recognizethat this table does not provide details of the fattyacid composition of these emulsifiers.

Table 1 - Composition (mass%) of Emulsifiers

emulsifier ID* MAG DAG TAG FFA unidentified

AMFE87 86.8 11.9 0 0 1.3AMFE83 82.8 14.0 0.1 0 3.2AMFE70 69.9 22.4 2.1 4.3 1.3AMFE68 67.8 24.5 3.6 0.4 3.7AMFE49 48.8 44.7 3.4 0.2 2.9AMFE31 31.1 50.0 9.4 0.2 9.4AMFE10 10.0 75.8 10.6 0 3.6

28SE71 70.8 13.6 2.1 11.7 1.828LE70 70.3 25.6 1.5 2.2 0.420SE70 69.6 21.7 2.8 5.4 0.514SE71 70.5 16.0 3.9 9.2 0.4

700K37 36.9 53.4 8.4 0.7 0.6GMSE35 35.3 50.3 13.0 0.4 1.0POLMO44 44.1 41.9 13.4 0.3 0.31427E66 66.2 8.8 23.4 1.1 0.51438E65 65.1 14.7 18.2 1.6 0.4

*Group 1 - original anhydrous milkfat emulsifiers; Group 2 - milkfatfraction emulsifiers; Group 3 - commercial emulsifiers.

The last two-digit number in emulsifier ID represents the approximatepercentage content of MAG.

Milkfat

19

Ice creams made with emulsifiers produced fromthe original AMF and each of the four milkfatfractions were virtually identical in appearance andtexture to those made with the commercial emulsi-fiers. We noted that the ice cream made with 1.0%of emulsifier (over-emulsified) was the only onethat looked different. This ice cream was dull andpuffy in appearance, whereas all the other icecreams had a slightly glossy appearance. Also, theover emulsified ice cream did not have a smoothmouthfeel, but the other ice creams did. Thus,manufacture of ice creams with different emulsifi-ers based on milkfat or milkfat fractions had noeffect on overall texture and appearance. In addi-tion, overrun of all ice creams made with differentemulsifiers (whether from milkfat, milkfat frac-tions or the commercial emulsifiers) was the sameat approximately 40 to 45%, with the exception ofthe over emulsified product, which had an overrunof about 55%.

Hardness of ice creams increased after hardeningfor 24 hours at -40°C. Some differences wereobserved between ice creams made with differentemulsifiers, as seen in Figure 1. In general, ice

Figure 1. Penetrometry test results for sample ice creams after 1-day storage

creams made with emulsifiers from AMF wereslightly softer than those made with commercialemulsifiers. However, ice creams made withemulsifiers from milkfat fractions, particularly 28S,were harder than those made with emulsifiers fromAMF. A slight effect of emulsifier was seen on levelof fat destabilization, as measured by change inturbidity of melted ice cream when compared tothe original ice cream mix. Emulsifiers with lowerMAG content gave slightly greater level of destabili-zation. In general, hardness of ice cream correlatedwell with melt-down rate. That is, harder ice creamsdelayed the onset of melting when exposed to roomtemperature. Typically, the commercial emulsifierswere the slowest to melt, while ice creams madewith emulsifiers from either milkfat or milkfatfractions melted at about twice the rate. Theexceptions to this were the ice creams made withemulsifiers derived from the 28S and 14S milkfatfractions. These ice creams melted at the same rateas those made with the typical commercial emulsi-fiers. However, the ice cream containing the emulsi-fier made with the 20S milkfat fraction exhibitedthe faster rate of melt-down. The reason for thisbehavior is not clear.


20

We also measured ice crystal size and air celldistributions using optical microscopy and imageanalysis. Initial air cell distribution (after harden-ing) was compared with the distribution obtainedafter storage at -15°C for three weeks, as seen inFigure 2. For each of the emulsifiers, whether frommilkfat, milkfat fractions or commercial sources,the initial air cells had essentially similar distribu-tions with mean size about 13 µm and bubble sizevarying from about 1 µm to between 80-100 µm.There was no significant difference in these distri-butions. However, the ice cream made with 1%emulsifier had significantly smaller air bubbles,with a mean size of 7.7 µm and a range of sizesfrom about 1 to 70 µm. Overemulsification resultedin smaller air bubbles, as expected. After 3 weeks ofstorage, the air bubble distribution had increased inmean size, up to about 17-18 µm. Again, there wereno significant differences in air bubble distributionbetween ice creams made with any of the emulsifi-ers at addition level of 0.1%. Interestingly, the air

bubble distribution in the overemulsified ice creamdid not change at all during this storage period.

Ice crystal size distributions in the ice creams afterhardening were not different. Ice creams with eachof the emulsifiers had mean ice crystal size between32 to 37 µm with a range of sizes from about 6 µmto about 100 µm. The only ice cream with differentice crystal size distribution was the overemulsifiedice cream, with 1% emulsifier. Mean ice crystal sizefor this ice cream was about 29 µm. Storage studiesfor change in ice crystal size with time are stillunderway.

Emulsifiers can be produced from either milkfat ormilkfat fractions, and these emulsifiers (MAG/DAGblends) are effective in ice cream. Most propertiesof ice creams (appearance, hardness, overrun, icecrystal and air cell distributions) made with theseemulsifiers, with the exception of melt down rate,were identical to those obtained when commercial

Figure 2. Mean size of air cells in ice cream samples after 2 day and 3 week storage.

0

5

1 0

1 5

2 0

70

0K

37

GM

SE

35

14

27

E6

6

PO

LMO

44

AM

FE

87

AM

FE

70

AM

FE

68

AM

FE

49

AM

FE

31

AM

FE

10

28

SE

71

28

LE

70

20

SE

70

14

SE

71

AM

FE

83

air 2day air 3wk

me

an

siz

e (

µm

)

Emulsifier ID

Milkfat

21

emulsifiers were used. Ice creams made withemulsifiers produced from milkfat fatty acidsmelted down at slightly faster rate than those madewith commercial emulsifiers, with two exceptions.Ice creams made with emulsifiers produced fromthe hard fractions of milkfat (28S and 14S) exhib-ited the same melt down rates as those made withcommercial emulsifiers.


Ice cream requires a particular emulsifier content toprovide proper melt-down rates and appearance.However, emulsifiers (mono- and di-glycerides,polysorbates, etc.) listed on the ingredient label areoften viewed as negative. The one major commer-cial ice cream with no added emulsifiers, promotedas all natural, commands the largest market sharein the US. An emulsifier produced by a naturalprocess from butteroil or milkfat fractions wouldpotentially satisfy the requirements for an all-natural product, while still providing necessaryemulsification. This could increase the use of bothice cream and milkfat. The results of this projectclearly show the potential benefits of using emulsi-fiers made from milk fat in ice cream. Further workis needed to develop procedures for controllingemulsifier manufacture from milkfat and milkfatfractions.

Publications

Liang, B., R.W. Hartel, T. Hagiwara, K. Wiederholtand K.L. Parkin, Application of Emulsifiers FromButteroil in Ice Cream, paper presented at AOCSAnnual Meeting, San Antonio, TX (1995).


22

VISITING SCIENTIST REPORT

Preparation of Butteroil-in-Water Emulsions UsingBlends of Milk Proteins and Lipids

Personnel: Marit Reierstad, Agricultural Universityof Norway; Kirk L. Parkin, Dept. Of Food Science

Funding: Wisconsin Milk Marketing Board andAgricultural University of Norway

Dates: March 15 – October 15, 1995

Objectives

Evaluate the technical feasibility of preparingbutteroil-in-water emulsions (including lowfatspreads) emphasizing the incorporation of dairyingredients. Casein hydrolysates and mono- anddiacylglycerol blends derived from butteroil will beevaluated. Butteroil in water emulsion properties ofgreatest interest are stability, consistency,spreadability, and flavor.

Summary

A preliminary assessment of formulations forpreparing reduced-fat dairy spreads was based onfactorial designs using eight experimental param-eters: emulsifier type, emulsifier concentration,butteroil with and without low-melting butteroilfraction, fat level, whey protein isolate or hydroly-sate, protein concentration, salt concentration andtemperature. The emulsifiers that were comparedinclude a commercial distilled monoacylglycerol(MAG) product, and two MAG-rich productsderived from enzymic glycerolysis of butteroil.Emulsions were prepared with the use of a Polytron(combined homogenization and sonication). Weanalyzed the emulsion for viscosity, stability andappearance by light microscopy.

The most important findings in these preliminarystudies were that the MAG-rich fractions preparedfrom butteroil were as effective in yielding stableemulsions as the commercial emulsifier. Emulsionstability was greater at 60% added fat than at 40%added fat, but whether or not butteroil fractionswere used had little impact. Emulsion stability was

also enhanced when 10% protein was added,although protein type had little influence. Theaddition of salt (1.2% NaCl) tended to destabilizethe emulsions. Temperature (40 or 50oC) andemulsifier level (0.5 and 1.0%) appeared to havelittle impact on emulsion stability.


These preliminary results will provide a startingpoint for a more comprehensive effort to developformulations for preparing reduced-fat tablespreads with butteroil, butteroil fractions and milkcomponents as principal ingredients (projectplanned for FY 1996-1999). These efforts areintended to expand and develop new markets fordairy ingredients.

Milkfat

23

INTERIM REPORT

Kinetics of Milkfat Crystallization

Personnel: R.W. Hartel, associate professor, D.B.Patience, graduate research assistant, Dept of FoodScience, UW; D. Illingworth, New Zealand DairyResearch Institute

Funding: Wisconsin Milk Marketing Board 92-8/New Zealand Dairy Board

Dates: February 1994 – January 1998

Objectives

1. Determine the effect of crystallization tempera-ture on milkfat crystallization and crystal separa-tion.

2. Determine the effect of pretreatment and coolingrate on milkfat crystallization.

3. Determine the effect of milkfat source on milkfatcrystallization.

4. Determine the effect of various mixer andcrystallizer conditions and geometries on milkfatcrystallization.

5. Determine the effects of the processing variables(Objectives 1 to 4) on physical properties and yieldsof milkfat fractions.

6. Investigate the effects of scale of operation.

Summary

For isothermal crystallization of milkfat at 28°C, ineither 0.5 or 3.5 L vessels, a 3x3 randomized blockdesign of experiments was performed to investigatethe effects of mixing on crystallization kinetics andfiltration efficiency. Impeller diameter (1.5, 2 and3") and stirrer speed (50, 100, 150 RPM) wereevaluated, using triple, propeller blades. Analysesincluded change in slurry turbidity with time,image analysis of slurry samples for crystal sizedistribution and growth kinetics, and filtrationcharacteristics at 5 bar in a pressure filter. Acustom-built, lab-scale, pressure filtration unit was

used to evaluate filtration efficiency by measuringrate of filtrate with time for different milkfat crystalslurries. An additional mixing experiment wasperformed using the 3" stirrer operated at 12.5RPM to investigate the laminar region of mixing.

For fractionation at 0.5 L scale, an optimal (scaled)tip speed (RPM. impeller diameter/tank diameter)was found, where the fastest rate of filtrationoccurred. Filtration resistance was lowest at thispoint. This optimal filtration point correlated with acrystal size distribution that contained mediumsized crystals (150 µm) without too many small orlarge crystals in the distribution. At lower scaled tipspeeds, filtration rate was reduced due to agglom-eration of milkfat crystals. On the other hand, atlarger scaled tip speeds (higher RPM), many smallnuclei were evident which disrupted filtration. Nosuch optimal point in scaled tip speed was foundfor fractionation at 3.5 L scale. Instead, filtrationresistance increased directly as scaled tip speedincreased, although it could have been that we didnot study low enough values of this scaled tip speedfor the larger vessel.

Melting points and yields of solid fractions corre-late with this optimal filtration point, indicatingthat the minimum liquid entrainment occurred atthese conditions. In general, yield of solid fractionincreased as tip speed increased. However, thisincrease in yield was due to greater liquid entrain-ment, as measured using an absorbance technique.

A preliminary research project to evaluate theeffects of nontriglyceride (nonTAG) components ofanhydrous milkfat (AMF) on AMF crystallizationwas also undertaken. A chromatography columntechnique was used to separate classes of lipids(TAG from nonTAG). The nonTAG component wasadded back to the purified TAG component andcrystallization kinetics studied using an absor-bance/turbidity technique. Addition of thesenonTAG components significantly delayed onset ofmilkfat crystallization under these conditions.


24


Improving fractionation technologies for separat-ing milkfat into valuable food ingredients isnecessary to improve process economics and tounderstand means of producing specific fractionsfor targeted applications. Understanding thekinetics of crystallization of milkfat will allowbetter control of existing fractionation technolo-gies, and may lead to new and improved tech-niques. In addition, the relations between mixingconditions in the crystallizer and filtration effi-ciency will allow optimal design of fractionationprocesses.

Publications

Patience, D.B. and R.W. Hartel, CrystallizationKinetics and Pressure Filtration of AnhydrousMilkfat, paper presented at AOCS Meeting, India-napolis, IN (1996).

Patience, D.B. and R.W. Hartel, Crystallization andPressure Filtration of Anhydrous Milkfat, Proceed-ings of Industrial Crystallization, Toulouse, France(accepted).

Milkfat

25

INTERIM REPORT

Mechanisms for Formation of Milkfat Crystals

Personnel: R.W. Hartel, associate professor,Y. Shi,research associate, Dept of Food Science

Dates: June 1995 – June 1998

Funding: Wisconsin Milk Marketing BoardUW9503

Objectives

1. To determine the main parameters influencingnucleation of milkfat and relate these parameters totype (polymorph) and shape of crystal formed.

2. Determine the effects of milkfat composition,agitation rate and temperature on kinetics ofmilkfat nucleation in melt crystallization.

Summary

We have studied the effects of temperature andagitation rate on nucleation of milkfat crystals fromthe melt. Milkfat crystals were formed undercontrolled conditions of temperature, agitationintensity and duration of agitation. These crystalswere separated (by filtration) from the liquid atdifferent times during their formation and analyzedfor melting profile (DSC), and chemical composi-tion (GC). In addition, the filtered crystals werewashed with acetone to separate entrained liquidfrom true crystalline solids, and the crystallineportion analyzed in the same way.

A DSC technique for quantifying melting point wasdeveloped based on the peak of the melting curve,and correlated to clear point measurements. It wasshown that the melting point based on the peak inthe DSC melting curve gave a reliable indication ofclear point only for high-melting milkfat fractions.Since this project investigates characteristics ofnucleation of initial milkfat crystals, this is a usefultechnique for comparing crystals formed duringnucleation.

For spontaneously nucleated (without agitation)milkfat crystals, the melting point (based on DSCmelting curve) increased with higher crystalliza-tion temperature. For crystallization of a wintermilkfat at 27.5°C, the melting point was about 40°C,while crystallization at 30.5°C resulted in nucleiwith a melting point over 49°C. When the nuclei ateach condition were washed with acetone, and theremaining crystalline product analyzed, a meltingpoint of about 52°C was obtained regardless of theoriginal crystallization temperature. These resultssuggest that the base crystalline material is thesame regardless of the crystallization temperature,but higher level of entrained liquid causes themelting point to decrease with decreased crystalli-zation temperature.

When nuclei were formed during a burst of agita-tion for a brief period of time, followed by quies-cent conditions, an interesting phenomenon hasbeen observed. That is, the melting point of sepa-rated nuclei increases with time of crystallization,at least during the initial stages of crystallization,and then approaches a maximum value. Thissuggests that the initial milkfat nuclei that form aremore loosely packed molecular structures than thefinal crystals produced at a given temperature. Onepotential mechanism for this behavior is initialformation of the a polymorph, with lower meltingpoint, followed by rapid transformation to a morestable (perhaps b’) polymorph with higher meltingpoint. X-ray spectroscopy experiments shouldconfirm this behavior.

Composition differences (fatty acid and acyl carbonnumber) were also found between nuclei formed atdifferent conditions. As expected, higher levels oftrisaturated triglycerides were found to correlatewith higher melting point. However, some interest-ing relationships of groups of triglycerides arebeginning to arise as we analyze this data in greaterdetail. Further results will be discussed in subse-quent reports.


26


In order to control milkfat fractionation technolo-gies, it is critical to understand exactly how milkfatcrystals form, and the parameters that influencetheir formation process. By understanding forma-tion of milkfat crystals, we can improve fraction-ation technologies to produce higher qualitymilkfat fractions more efficiently.

Milkfat

27

INRERIM REPORT

Continued Studies of Surface Melt CrystallizationTechniques for Fractionating Milkfat

Personnel: R.W. Hartel, sssociate professor, Dept ofFood Science; J. Ulrich, professor, University ofBremen; M. Tiedtke, graduate research assistant,University of Bremen.


Funding: Dairy Management Inc. HRT 96

Objectives

1. To evaluate surface layer melt crystallization formulti-step fractionation of milkfat.

2. To make an economic comparison betweensurface layer and suspension techniques forfractionating milkfat.

3. To determine the effects of initial milkfat compo-sition on crystallization kinetics and overallefficiency for fractionating milkfat using surfacelayer techniques.

Summary

In the past several years, surface layer crystalliza-tion techniques have been used successfully forfractionating milkfat. Milkfat was cooled as itflowed across a cooling surface in a thin film. Bycontrolling the temperature of the cooling surface,crystallization of milkfat occurred directly on thesurface. As a solid layer formed on the coolingsurface, a partition between crystalline and liquidportions of milkfat occurred, depending on thetemperature gradients and milkfat composition.Temperature of the solid layer was kept constantduring the process by decreasing temperatureaccording to film thickness. To enhance separationof the crystalline layer, a sweating procedure and/ora washing step may be used, which may involve aslight heating of the layer. Finally, the layer wasremoved from the surface by raising the tempera-ture to melt the solid layer and the hard fractioncollected in the liquid form.

In this study, several multiple-step fractionationprotocols are being explored. In the first step,milkfat was fractionated at 25°C to give a solidfraction (yield of 54%, clear point of 39.4°C) and aliquid fraction (yield of 46%, clear point of 21.6°C).The liquid fraction was then cooled to 20°C for asecond fractionation step. A low-melting fraction(overall yield of 37%, clear point of 15.4°C) and amiddle-melting fraction (overall yield of 9%, clearpoint of 27.1°C) were produced. The solid fractionwas melted and recrystallized to produce a high-melting fraction. When the solid fraction from theinitial step was cooled rapidly, a high-meltingfraction (overall yield of 21%, clear point of 43.4°C)and a middle-melting fraction (overall yield of33%, clear point of 26.6°C) were produced. Whenthe solid fraction was cooled more slowly, a veryhigh-melting fraction (overall yield of 5%, clearpoint of 47.5°C) was produced, as well as a middle-melting fraction (overall yield of 49%, clear point of37.1°C). Thus, in our preliminary experiments onmulti-step fractionation, milkfat fractions withreasonable yields and clear points can be producedusing surface-layer crystallization.

An excellent correlation between yield of solid orliquid fraction and the clear point has been foundfor this technique. Higher yields during fraction-ation resulted in lower clear points. Further studiesto determine the effects of different milkfat startingmaterials on fractionation are underway.


Techniques that improve separation efficiencybetween solid and liquid phases to produce moredistinct milkfat fractions are needed. Layer-typemelt crystallization techniques, popular in theorganic chemical industry, have potential forproviding controlled crystallization with minimalseparation requirements to produce distinctfractions. In many instances, layer crystallization ismore economical than suspension-type crystalliza-tion. In this project, we have demonstrated that


28

surface-layer melt crystallization techniques canproduce high melting milkfat fractions that areindistinguishable from those produced by existingtechnologies and at similar yields. Multiple-stepfractionations are also feasible with surface-layercrystallization technology, and can producefractions of similar yield and composition assuspension fractionation. Further work is needed todetermine the economic standpoint of thistechnology.

Publications

Tiedtke, M., J. Ulrich and RW Hartel, Solid LayerMelt Crystallization - A Fractionation Process forMilkfats, paper presented at Third InternationalWorkshop on Crystal Growth of Organic Materials,Washington, DC (1995).

Tiedtke, M., Ulrich, J. and R.W. Hartel, Solid LayerMelt Crystallization - A Fractionation Process forMilkfat, ACS Conference Proceeding Series, CrystalGrowth of Organic Materials (Myerson, A.S., Green,D.A., Meenan, P., Eds.) pp. 137-144 (1996).

Milkfat

29

INTERIM REPORT

Physical Chemistry of Lipid Mixtures: Dairy BasedSpreads

Personnel: R.W. Hartel, associate professor, R.L.Lindsay, professor, J. Knollenberg, graduate assis-tant, B. Liang, research associate, Dept of FoodScience.


Funding: Dairy Management Inc. HRL 95

Objectives

1. To relate the specific triglyceride composition ofmilkfat fractions to phase and crystallizationbehavior of milkfat fractions with each other.

2. To relate the specific triglyceride composition ofmilkfat fractions to phase and crystallizationbehavior of mixtures of milkfat fractions andcanola oil.

Summary

A technique was developed for studying thesolubility of hard milkfat fractions in low-meltingfractions (liquid oils). This technique relies onmeasuring changes in turbidity (absorbance) afterincreasing the addition of hard fraction in theliquid fraction. In this technique, the liquid oil wasmaintained at constant temperature (25 - 40°C) in ajacketed beaker and a specific amount of hardfraction added as a liquid fat. The system wasallowed to equilibrate for 4 days before measuringabsorbance. Using this technique, a clear solutionwas obtained with essentially no change in absor-bance until the point where the hard fractioncrystallized separately from the liquid fraction. Atthis point, the absorbance increased dramaticallywith slightly higher addition level of the hardfraction. The point of increased absorbanceprovides a measure of the intersolubility of onelipid in another. While this is not a thermodynamicequilibrium, the intersolubility measured in thisway provides an estimate of how much hard

fraction can be accommodated in the liquidfraction before crystallization occurs. For thefractions studied (with melting points of 45º and15°C), the intersolubility was measured at less than1% at 25°C and increased exponentially to about35% at 40°C. A wider range of milkfat fractionsand canola oil is currently being studied.

The phase behavior of mixed lipid systems can beseen in isosolids diagrams, where solid fat contentis measured at different temperatures for the rangeof mixtures from 100% hard fraction to 100%liquid oil (either low-melting fraction or canola oil).These diagrams provide information about thecompatibility of two lipids. Using a BrukerMinispec PC-120 NMR, the solid fat curves formixtures of several hard and soft milkfat fractionswere measured. In general, mixtures of high-melting milkfat fractions with either canola oil orlow-melting milkfat fractions resulted in straightlines on the isosolids diagram, indicating completecompatibility between these fractions. DSC ther-mograms of mixtures of high-melting and low-melting milkfat fractions showed this compatibil-ity, where the peaks associated with each compo-nent changed uniformly with increasing addition ofone fraction to the other. However, mixtures ofhigh-melting milkfat fractions with canola oilexhibited more complex melting behavior. Analysisof these results is still ongoing.

Some preliminary experiments to measure induc-tion time for nucleation, using an isothermal DSCtechnique, have been completed for mixtures ofhigh-melting milkfat fractions in either lowmelting milkfat fractions or canola oil. Results todate show that the time required for formingcrystals decreased as crystallization temperaturedecreased for all mixtures, decreased as the concen-tration of high-melting milkfat fraction increased,and was shorter for crystallization of high-meltingmilkfat fractions dissolved in canola oil, as com-pared to low-melting milkfat fractions. Since


30

shorter induction time indicates more rapid crystal-lization, high-melting milkfat fractions crystallizedmore rapidly at lower temperatures, when itsconcentration in the liquid oil was higher and whenit crystallized out of canola oil (compared to low-melting milkfat fractions). Analysis and furtherexperimentation are still underway.


Increased utilization of milkfat and milkfat frac-tions in dairy-based spreads requires an under-standing of how chemical composition influencesphysical behavior. In this study, the focus is under-standing the effects of chemical composition onphase and crystallization behavior. Through thesestudies, improved applications of milkfat fractionsin dairy-based spreads will result.

Milkfat

31

INTERIM REPORT

Incorporation of Milkfat Fractions in Chocolates -Phase 2

Personnel: R. W. Hartel, associate professor, J.Bricknell, graduate research assistant, S. Metin,graduate research assistant, Dept. of Food Science


Funding: Wisconsin Milk Marketing Board 94-02

Objectives

1. To further study the fundamental aspects ofincorporating milkfat and milkfat fractions intochocolates.

2. To investigate the incorporation of milkfatfractions produced by the Tirtiaux pilot plant inboth milk and dark chocolate.

3. To study processing techniques to improve thecompatibility of milkfat fractions and cocoa butterin chocolates.

Summary

Our research in the past year focused on studyingeffects of milkfat fractions on crystallization ofcocoa butter, and developing improved techniquesfor studying the effects of milkfat fractions onbloom formation in chocolate.

Using Differential Scanning Calorimetry (DSC), wehave studied the effect of milkfat and milkfatfractions on cocoa butter crystallization. Two DSCtechniques were used to study crystallization ofcocoa butter in the presence of different levels andtypes of milkfat fractions. In the first method, themixed lipid was brought quickly to crystallizationtemperature in the DSC and allowed to crystallize atconstant temperature. This “isothermal DSC”technique allowed us to determine the inductiontime required for onset of cocoa butter crystalliza-tion, and estimate the rate of crystallization. Milkfatclearly crystallized more readily than cocoa butterover the temperature range of 10 to 25°C. Addition

of either milkfat or milkfat fractions to cocoa butterat 5 or 10% by weight generally increased theinduction time for cocoa butter crystallization. Ingeneral, this inhibition of crystallization wasgreater at the higher addition levels. No differenceamong a range of milkfat fractions was observed.These results were analyzed using the Avramiequation to obtain an exponent of one, indicatingmass transfer limitations. By applying differentscanning rates (from 5 to 20°C/minute) in the DSC,we developed information on crystallizationkinetics using a modified Avrami equation. Whiledifferent crystallization curves were obtained usingthis technique, analysis of these results were notsatisfactory, yielding information that contradictedthe isothermal technique.

In chocolate, one of the main theories of bloomformation is that cocoa butter undergoes a poly-morphic transformation (from ß-5 to ß-6) thatleads to rearrangement of the crystal structure andappearance of microcrystals at the surface. Thisrearrangement results in visual dulling and grayingof the surface, associated with bloomed chocolate.It has been hypothesized that milkfat inhibits thispolymorphic transformation and thus, delaysbloom formation. However, there is no conclusiveevidence of this in chocolate. Our work this pastyear has focused on developing a technique thatwill allow us to quantify both the polymorphictransformation (by X-ray spectroscopy) and theonset of visual bloom in chocolates, and to evaluatethe effects of different milkfat fractions. We havedeveloped a technique to manufacture chocolatewithout crystalline sucrose (which interferes withthe X-ray spectra of cocoa butter) in order toquantify the appearance of the ß-6 polymorph andcorrelate this with visual bloom formation. Study-ing the effects of milkfat fractions on this polymor-phic transformation in chocolates, and correlatingthese effects with visual observation of bloom is atask that remains.


32


One of the main applications for milkfat fractions isin the chocolate and confectionery industry. Inorder to optimize the use of milkfat fractions inchocolate applications, we need to understand theeffects of components of milkfat on bloom resis-tance, softening of chocolate and lipid compatibil-ity. In addition, we need to understand the kineticsof crystallization in the mixed system of cocoabutter and milkfat fractions to optimize processingconditions for producing (tempering) chocolates.

Publications

Hartel, R.W., Application of Milkfat Fractions inConfectionery, J. AOCS (in press).

Metin, S. and R.W. Hartel, Crystallization Kineticsof Blends of Cocoa Butter and Milkfat or MilkfatFractions, J. Thermal Analysis (accepted).

Milkfat

33

INTERIM REPORT

Investigation of Baked Milkfat Flavor Developmentin Milkfat Ingredients for the Bakery and FoodIndustries

Personnel: Robert C. Lindsay, professor, Dept. ofFood Science, Ann Han, graduate research assistant,Dept. of Food Science (In collaboration withresearchers at the New Zealand Dairy ResearchInstitute)

Dates: February 1993 – January 1995 (ExtendedJuly 1996)

Funding: Wisconsin Milk Marketing Board #92-7

Objectives

1. To establish a sensory panel trained in therecognition and descriptive sensory analysis ofbaked milkfat flavor in baked products (NDZRI).

2. To establish a model baking system for theassessment of baked milkfat flavor in baked goods(NZDRI).

3. Using the model baking system, undertake andcorrelate chemical and sensory analyses to identifykey compounds responsible for baked milkfatflavor in baked products (UW and NZDRI).

4. To define the specific chemical reactions involvedin the development of enhanced baked milkfatflavor (UW and NZDRI).

5. To formulate and prepare milkfat based ingredi-ents which provide enhanced baked milkfat flavor(UW and NZDRI).

Summary

Initial studies focused on the isolation of volatileflavor compounds in heated butter and buttercookies by headspace, vacuum steam distillation,and solvent extraction procedures. Many com-pounds were identified by gas chromatography andmass spectrometry, and included free fatty acids,

methyl ketones, gamma and delta-lactones, alde-hydes, esters, and Maillard browning products.However, all of the compounds that were analyzedin these traditional isolates were relatively abun-dant, well-documented compounds. While somecompounds were important to baked butter flavors,they did not greatly enhance the basic knowledgeabout or reveal the nature of the true baked butterflavor.

Earlier research at the UW had revealed that somenew volatile compounds were responsible for thethe milky flavor in skimmilk, and these were agroup of alkyl phenols, including p-cresol. Sincelittle was known about the flavor effects of thesecompounds, they were investigated in detail aspotential contributors to butter flavors. Althoughpresent in extremely low parts per billion levels, thealkyl phenols were found to be the key missingingredients to the baked butter flavor complex.

Basic information was obtained about the behaviorof alkyl phenols in the butter system, and it wasfound that when they were formed by heat degra-dation of flavorless bound or conjugate formsduring baking, they migrated to the milkfat phase.Most of the alkyl phenols were found to occur inmilk and butter as the bound or conjugate forms. Itwas found in this research that when excessiveconcentrations were formed through the action ofconjugase enzymes, strong cowy like flavors wereproduced.

Research studies on development of high-flavoredmilkfat and butter ingredients led to methods forthe enzymatic preparation of prototype productswith approximately 20X flavor strength. Whenadded to butter cookies, milk chocolate, and toffee,these intensely flavored ingredients greatly en-hanced baked butter flavors.

Recent studies have focused on investigations onsources of alkyl phenols in milk products, and


34

establishing basic information on the occurrenceand behavior of the bound forms which hold mostof the baked butter flavor potential in milk andbutter products. Additionally, studies have beencarried out on the interaction of alkyl phenolflavors with those produced by short chain freefatty acids and lactones when they are liberatedfrom milkfat triacylglycerols.


Many of the traditional baking industry uses ofbutter have been replaced by specialty vegetableoils. Bakers have stated that greater quantities ofbutter could be used if flavor could be distinctivelyprovided when reasonable amounts of butter areused as ingredients in baking recipes. This researchdirectly addresses the task of developing basicinformation to foster the development of newintensely flavored butter ingredients for the bakingindustry. Such ingredients will allow the dairyindustry to regain butter sales and competitivelymaintain butter as a functional sought-afteringredient for the baking industry.

Milkfat

35

APPLICATIONS PROGRAM REPORT

Milkfat Applications Research Program

Personnel: Kerry E. Kaylegian, researcher, BarbaraH. Ingham, research specialist, Orville H. Harris,research specialist, Chris Kirk, research specialist,Center for Dairy Research


Funding: Wisconsin Milk Marketing Board UWA9602, Dairy Management Inc.

Objectives

1. Provide technical support for butter and milkfatfractions to the dairy, bakery, and confectioneryindustries

2. Work toward increased expertise in establishedareas suited for milkfat fractions, and investigatepotential new applications suited for milkfatfractions

3. Conduct applied research in butter and milkfatfractions

4. Coordinate the Tirtiaux pilot plant operation andactivities

5. Coordinate technical transfer activities inconjunction with CDR’s Communication programfor UW projects relating to milkfat

Summary

This has been an exciting year for the MilkfatApplications Program. To start, we’ve grown from astaff of one person to three full-time people andone-part time person. We installed the Tirtiauxfractionation pilot plant, placed an order for aGerstenberg and Agger texturizer pilot plant, andimplemented the Milkfat Fractionation Consor-tium. The Milkfat Fractionation Consortium hasnine members, donating almost $80,000 per year,who are interested in producing and/or usingmilkfat fractions in a variety of food products.

The Tirtiaux fractionation pilot plant was installedin March 1996, just in time for our CDR OpenHouse. As expected with most equipment, we hadour challenges during start up but now things arerunning smoothly. We are currently producingsamples of milkfat fractions for our Consortiummembers, university researchers, and others. Themilkfat fractions available for sampling exhibit arange of melting behaviors that meet the needs of avariety of applications. These fractions havemelting behaviors similar to the fractions producedwith the leased Tirtiaux unit several years ago. TheGerstenberg and Agger unit is due in this summerand will give us the capability to turn the fractionsinto specialty butters, ranging from soft spreads tohard butters and ranging from 20% fat to 100% fat.

The additional staff and improvement of facilitiesallows the Milkfat Program to provide more in-depth technical support on butter, milkfat andmilkfat fractions to the dairy and food industries.We continue to receive and answer phone questionsabout butter, milkfat, and milkfat fractions. Ques-tions come from the butter makers, confectioners,bakers, general food industry, and media/marketersfor information on the manufacture, shelf life, anduse of butter, how milkfat is fractionated and why,and all sorts of questions ranging from basic tofairly technical. The milkfat group also provides on-site technical support, we travel to companies togive presentations, assist with laboratory set-upand analyses, as well as train personnel in ourlaboratories to perform analytical tests and to learnmilkfat fractionation at the pilot plant.

To improve our ability to understand the functionof milkfat in foods and to provide better technicalservice to our constituents, we constantly monitorthe literature, perform small-scale applicationsprojects, and participate in short courses whenappropriate. This year we completed an evaluationof experimental pastry butters made with milkfatfractions in puff pastries. The experimental buttersprovided the physical benefits of pastry margarine


36

to the pastries (i.e., improved height and volume)while contributing desirable butter flavors that themargarine samples lacked. Our results from thisstudy were presented at the annual meeting of theAmerican Association of Cereal Chemists inNovember 1995, and the paper is being submittedto the Journal of Cereal Chemistry. Members of themilkfat group attended a short course in “Applica-tions of Fats, Oils and Shortenings in BakeryProducts” sponsored by the American Oil Chemists’Society 11/95, and the “Ice Cream Makers ShortCourse” sponsored by the UW 1/96.

This year CDR helped to plan and then host the firstannual Milkfat Technology Forum, sponsored byDMI. This forum was held in April and broughttogether researchers and industry from around thecountry. Research updates were presented in keyareas of milkfat research, and the future directionof milkfat research was discussed by the group.Proceedings from this Forum are currently inpublication.


Fractionation provides an opportunity to tailormilkfat for specific applications that may benefitfrom the flavor of milkfat, but where the use ofmilkfat is hindered by its physical properties. Thefractionation of milkfat to produce specialtyingredients will increase the value of milkfat andexpand its use in the food industry. The CDRMilkfat Group provides much needed researchfacilities, data, and technical support to assist bothmanufacturers and users of milkfat fractions as theindustry begins to commercialize milkfat fraction-ation in the U.S.

Publications and Presentations

Kaylegian, K.E. Functional Characteristics andNontraditional Applications of Milk Lipid Compo-nents in Food and Nonfood Systems. J. Dairy Sci.1995 78:2524-2540.

Paeschke, T.M., and K.E. Kaylegian. Evaluation ofExperimental Pastry Butters in Puff Pastries.Presented at the annual meeting of the AmericanAssociation of Cereal Chemists, November 1995.

Kaylegian, K.E. Milkfat Fractions Innovative FoodIngredients. Governor’s Conference on AgricultureMarch 1996, Oshkosh, WI.

Kaylegian, K.E. Milkfat Fractions: A Global Per-spective of Research and Applications. UW FoodScience Club 10th Annual Symposium — FoodScience — A World of Opportunities, April 1996,Madison, WI.

Kaylegian, K.E. Milkfat Fractionation Technologiesand Applications. National Milkfat TechnologyForum, April 1996, Madison, WI.

Kaylegian, K.E. Applications of Milkfat Fractions inLowfat Foods. Presented at the annual meeting ofthe Institute of Food Technologists, June 1996, NewOrleans, LA.

Kaylegian, K.E. Milkfat Fractions Innovative FoodIngredients. Presented at the annual meeting of theInstitute of Food Technologists, June 1996, NewOrleans, LA.

Milkfat

37

FINAL REPORT

Effects of Defined Milkfat Fractions on PostprandialLipid Metabolism in the RatPersonnel: Denise Ney, associate professor, Hui-Chuan Lai, graduate student, Mike Grahn, researchspecialist, Department of Nutritional Sciences



Objectives

The overall objective of this research is to charac-terize the nutritional effects of defined milkfatfractions on lipid metabolism. Evidence suggeststhat cholesterol and triacylglycerol (TAG) metabo-lism will be altered in conjunction with changes inthe fatty acid and TAG profiles of milkfat due tofractionation. Our research focuses on postprandial(i.e., after ingestion of a meal) lipid metabolism inanimals fed diets containing either a low-meltingpoint (liquid) or high-melting point (solid) milkfatfraction. The specific objectives of this project were:

1. To determine the effects of ingesting definedmilkfat fractions on postprandial lipemia and theactivity of post-heparin lipoprotein lipase, an indexof plasma TAG catabolism.

2. To characterize the composition of lymphchylomicrons in animals fed liquid or solid milkfatfractions compared to palm oil or corn oil.

Summary

We have demonstrated that ingesting a liquidmilkfat fraction enriched in 18:1 and unsaturatedTAGs with <40 carbon atoms produces lowerconcentrations of cholesterol and triacylglycerol inserum of rats compared to a solid milkfat fractionor palm oil (Lai et al. 1995; Lai and Ney 1995).Interestingly, the CSIRO Division of HumanNutrition in Australia recently demonstrated lowerconcentrations of cholesterol in plasma fromhuman subjects fed dairy products with feed-induced modifications in the fatty acid profile

similar to those induced by milkfat fractionation inour studies (1996). This suggests that our studiesare relevant to humans and that changes in the TAGor fatty acid composition of milkfat by fraction-ation processes or feeding practices may improveits nutritional profile.

We investigated the time course of postprandiallipemia and lipolytic activity in rats trained to eatmeals containing milkfat fractions, palm oil or cornoil for 4 weeks. Rats fed saturated fats comparedwith corn oil showed a significantly greater peakincrease in postprandial TAG concentrations. Cornoil ingestion resulted in significantly less TAGaccumulation (millimoles per liter per 24 hr)compared with ingestion of saturated fats.Postheparin plasma lipoprotein lipase activity andplasma insulin concentration were generally greaterwith ingestion of corn oil compared with palm oilor butterfat. Palm oil ingestion resulted in abiphasic plasma TAG response curve and greaterpostheparin plasma lipoprotein lipase activitycompared with butterfat ingestion, suggestingdifferential effects of saturated fats on postprandiallipemia. In summary, greater postprandial lipemiawith ingestion of saturated fats compared with cornoil may be due in part to slower plasma TAGclearance.

Digestion of dietary fat begins in the mouth andstomach and is completed in the small intestinewhere final absorption occurs. The shortchain andmedium-chain length fatty acids in milkfat,primarily located at the sn-3 position, are subject togastric lipolysis. We studied how liquid and solidfractions of milkfat affect the composition of lymphchylomicrons in rats given gastric infusion of lipidemulsions containing liquid milkfat, solid milkfat,palm oil or corn oil. Our results show that ratschronically fed palm oil or solid milkfat show 1.4-2.5 fold higher mass output of cholesterol, TAG andphospholipid in baseline lymph samples comparedwith rats fed corn oil or liquid milkfat. The similar-ity in response when comparing liquid milkfat and


38

corn oil suggests that modification of the fatty acidand TAG profiles of milkfat improves the lipidresponse associated with ingestion of milkfat.

Determining the fatty acid and TAG profiles of thediets and lymph chylomicrons provided additionalinformation about the effects of milkfat on lipiddigestion. The profile of TAG species in lymphbased on total acyl carbon number, indicated thatTAG species with <40 carbons are present in themilkfat diets, but they do not appear in lymph.Similarly, lymph from animals infused with liquidor solid milkfat fractions contained few short-chainfatty acids but the profiles of C14:0, C16:0 andC18:1 were maintained compared to the emulsioninfused.

Taken together, our results suggest that a liquidmilkfat fraction enriched in oleic acid and short-chain fatty acids shows improved digestive effi-ciency (i.e., enhanced gastric lipolysis). Liquidmilkfat may be useful as an ingredient in humannutritional products. Nutritional products mayinclude infant formulas, supplements for healthyindividuals such as the elderly or athletes, andmedical foods for individuals with pancreaticdeficiency, which occurs in cystic fibrosis, pancre-atitis or premature birth.


The American public currently demonstrates greatinterest in issues related to diet, nutrition andhealth. In order to market dairy products toconsumers, the dairy industry needs valid informa-tion about the nutritional effects of milkfat. Ourresearch extends and compliments work conductedat CDR by providing new information regarding thein vivo nutritional effects of defined milkfatfractions on lipid metabolism. This will be keyknowledge for marketing future dairy products toconsumers.

Publications/presentations

Lai, H.C. Postprandial lipid metabolism withingestion of defined butterfat fractions in the rat.Ph.D. thesis. University of Wisconsin–Madison, 275pp, 1994.

Lai, H.-C., Lasekan, J.B., Monsma, C.C. and Ney,D.M. Alteration of plasma lipids in the rat byfractionation of modified milk fat (butterfat). J. ofDairy Science, 78:794–803, 1995.

Lai, H.-C. and Ney, D.M. Corn oil, palm oil andbutterfat fractions affect postprandial lipemia andlipoprotein lipase in meal-fed rats. J. Nutr.125:1536–1545, 1995.

CDR Technical Cheese Conference, March 22–23,1995, Green Bay, WI.

Lai, H.-C. and Ney, D.M. Lymph chylomicroncomposition and size with gastric infusion of cornoil and butterfat fractions in rats. In preparation, J.Nutr.

Milkfat

39

FINAL REPORT

Development of New Dairy Products and theAdoption Process by U.S. Consumers

Personnel: Brian W. Gould, senior scientist, J.H.Park, graduate research assistant,Center forDairy Research and Department of Agriculturaland Applied Economics


Funding: Wisconsin Milk Marketing BoardUWA9505

Objectives

1.Examine the role of reduced fat dairy products inmeeting US dairy product demand2. Develop statistical models of dairy productconsumption to examine the role of productpromotion on dairy production adoption and use

Summary

The U.S. dairy industry is restructuring both at thefarm and processing level. On the farm level, therehave been tremendous shifts in the production offarm milk away from traditional milk productionareas to the west and south central regions of theU.S. On the processing side, there has been signifi-cant consolidation of processing firms with theexpectation that such consolidation will continue tooccur over the near term. In terms of consumerdemand, with increased health concerns aboutdietary fat intake, the dairy processing industry hasincreased efforts to develop reduced fat dairyproducts. For example, in the cheese industry,reduced fat cheeses are in the marketplace. Reducedfat products currently account for less than 7% ofthe natural Cheddar cheese retail market, whereaslight (50% fat reduction) or nonfat products makeup 12% of the American (process) cheese market.The fluid milk market has had a longer history, andbeen more dramatic, with per capita reduced fatmilk consumption exceeding whole milk consump-tion since the mid-1980’s.

The increased development efforts by the dairy

processing industry to market new products raisesseveral questions regarding how consumers decideto buy these new forms of dairy products. Thequestions include: What is the role of promotion inthe adoption process? What is the role of price?What is the interactive effect of product character-istics, promotion and price? What is the impact ofadoption of new products on the purchases of thetraditional products?

Focusing on the demand for reduced fat dairyproducts, and using the Dairy Product ConsumerPurchase database system, we have the potential forfollowing the weekly dairy product purchase of thesame set of households over the 1991/94 period(e.g. 170 weekly observations for each household).We can examine the habits of consumers buyingproducts like reduced fat cheeses, reduced fat milks,butter and butter blends, yogurt, etc. Using thisdata, we have found that over the April 1991-March1992 period, more than 45% of cheese consuminghouseholds purchase some type of reduced fatcheese variety and these reduced fat varietiesaccount for more than 10% of cheese purchased.Most households purchased more than one type ofreduced fat dairy product (80%), reduced fat milkwas the most commonly purchased type. More than70% of households purchased 2% milk, 45%purchased 1% milk, 42% purchased skim milk and55% purchased whole milk. More than 45% ofhouseholds purchased butter, while more than one-third purchased butter/margarine blends (Gould,Carlson and Lin, 1993).

Reduced fat varieties of yogurt and fluid milk havelong possessed sensory characteristics that meetconsumer demands. As discussed in a recentWMMB sponsored reduced fat cheese conference,increased reduced fat cheese consumption hasoccurred despite the fact that, when compared tofull fat varieties, reduced fat cheese lacks desirableflavor, texture, functionality and shelf life character-istics (Center for Dairy Research, 1994). Shouldefforts conducted in both private industry and at


40

the regional dairy research centers be successful inimproving the quality of reduced fat cheeses,growth in consumer acceptance of these cheesescan be expected to exceed recently observed rates.

During FY’s 1994 and 1995, the Wisconsin Centerfor Dairy Research initiated a project to analyze thedeterminants of changing dairy product consump-tion patterns on the demand for farm milk. Wefocused on the impact of changing consumerhealth awareness, socio-economic status, and otherfactors that influence both household consumptionof dairy products and choices made within thehousehold regarding the purchase of reduced fatvarieties. Implications of such choices on farm milkdemand were also presented (Gould and Carlson,1993a; Gould and Carlson, 1993b; Gould, Carlsonand Lin, 1993; Carlson and Gould, 1994b).

With detailed dairy product purchase informationwe can identify whether consumers are willing topay a premium for reduced fat products comparedto similar full fat varieties. For cheese manufactur-ers this may be relatively important, since produc-tion costs are higher for reduced fat cheeses (Centerfor Dairy Research, 1994, p.29).

Modeling the demand for reducedfat dairy products

In order to investigate the differing demandstructure for full versus reduced fat dairy products,we developed an econometric model of demand.This model allows us to incorporate, within a singlemodel, the simultaneous purchase decision ofwhich type of dairy product to consume and theamount to be consumed (Gould, 1996a). Thismodel accounts for the “competing” nature of dairyproducts with differing fat contents with respect tothe consumers dairy foods budget, and it alsoincorporates statistical features which control forthe impact that a significant number of households(individuals) do not consume a particular com-modity. These commodities are referred to ashaving a “censored” demand and special economet-ric procedures need to be incorporated to avoid thegeneration of biased demand coefficient estimatesand resulting biased estimates of income, price, anddemographic elasticities.

As an example, we apply our econometric model ofdairy product demand to fluid milk using threecommodity definitions: whole, 2% and less than2%. We apply the model to one year of fluid milkpurchase obtained from the Dairy Product Con-sumer Purchase database over the April 1991-March 1992 period. Table 1 provides an overview ofthe annual purchase patterns of the sample house-holds according to what type of milk they pur-chased over the 1991/92 period. Table 2 provides alisting of the demographic variables used in theeconometric model to control for their effect onhousehold milk consumption.

Own- and cross-price and substitution elasticitieswere estimated. Table 3 provides estimated priceelasticities. As expected, all own-price elasticitiesare negative and significantly less than 1, indicatingthat fluid milk manufacturers face an inelasticdemand for their product. All cross-price elastici-ties were found to be positive and statisticallysignificant. These positive values support thehypothesis that milk of different fat contents aresubstitutes for one another.

Given the success of modeling consumer demandfor full and reduced fat milks, we are expanding itto analyze alternative full and reduced fat dairyproducts. We intend to apply this model to alterna-tive types of cheese, food fats and oils and yogurt.These results can be obtained from the author uponrequest.

Modeling the impact of promotion ondairy product adoption rates

The U.S. dairy industry has responded to decreas-ing per capita dairy product consumption byadopting industry-wide advertising and promotioncampaigns. Examples of these programs includethe 1995 $52 million Fluid Milk Processor Educa-tion Program and the cheese, butter and otherdairy product promotion efforts of the NationalDairy Promotion and Research Board. Recentevaluations of several dairy advertising campaignshave indicated a positive net effect of such pro-grams (Forker and Kinnucan, 1991; Cheese Re-porter, 1995; Sun, Blisard and Blaylock, 1995). For

Milkfat

41

8.7 6.916.932.6

19.725.526.536.7

Household Type No. of Consuming Amount Std.Households Consumed Dev.

(Gallons)

Consume all milks 1096 Whole 9.6 18.6

Skim 16.0 23.6 Lowfat 17.9 27.4 Total 43.6 43.6Only whole milk 420 27.3 31.2Only skim milk 411 26.3 35.4Only lowfat milk 480 29.4 34.2Whole and skim milk 291 Whole Skim Total

Whole and lowfat 621 Whole Lowfat Total

Skim and lowfat 984 Skim Lowfat Total

All households 4303 Whole Skim Lowfat Total

Source: Gould(1996)

17.115.132.2

23.825.030.9

14.022.136.0

25.430.436.6

15.916.732.6

31.524.937.9

Table 1. Milk Consumption Patterns by Consuming Households


42

Table 3. Estimated Milk Own- and Cross-Price Elasticities

Commodity Price Elasticity

Whole Milk Skim/1% 2%

Whole Milk -0.803 (0.096) 0.294 (0.050) 0.414 (0.057)

Skim/1% 0.242 (0.044) -0.593 (0.078) 0.253 (0.057)

2% 0.252 (0.039) 0.190 (0.043) -0.512 (0.057)

Note: Standard Errors are in parentheses.Source: Gould(1996a).

Table 2. Means of Exogenous Variables Used in Econometric Model

Variable VariableName Units Mean Std. Dev. Population Mean

Household Characteristics: Income as percent of poverty threshold PCTPOV % 352.3 227.7 ——— Percent of household members < 13 yrs. PERLT13 % 10.3 18.9 ——— Percent of household members > 65 yrs. PERGT65 % 23.3 40.0 ———

Meal Planner Characteristics: Non-White NONWHITE 0/1 11.7 ——— ——— Completed College COLLEGE 0/1 27.3 ——— 22.0

Region of Residence: Northeast NE-REG 0/1 5.0 ——— 5.4 South Atlantic SA-REG 0/1 17.0 ——— 16.2 Middle Atlantic MA-REG 0/1 17.1 ——— 17.6 East North Central ENC-REG 0/1 19.0 ——— 17.0 West North Central WNC-REG 0/1 9.2 ——— 7.3 East South Central ESC-REG 0/1 5.4 ——— 6.1 West South Central WSC-REG 0/1 9.6 ——— 10.5 Pacific/Mountain PAC-REG 0/1 17.7 ——— 19.9

Milkfat

43

example, Kaiser and Roberte(1995) evaluatedgeneric fluid milk advertising in New York City overthe January 1986 - December 1992 period andfound a significant impact on whole and low fatmilk demand. Larson(1992) argues for continuedexpansion of such generic agricultural commoditypromotion programs given the positive secondarybenefits of increased effectiveness of associatedbranded promotion efforts and protection ofmarket share from processed substitute products.

Previous evaluations of the effect of promotionefforts on U.S. dairy product demand have tendedto focus on generic advertising programs at theaggregate commodity or market level (Blaylock andBlisard, 1988, 1990; Kinnucan and Forker, 1991;Sun, Blisard and Blaylock, 1995; Ward and Dixon,1989; Ward and McDonald, 1986). Little work hasbeen done at the household level to analyze theimpact of dairy product promotion or to analyzethe impacts of non-advertising promotion efforts.With proposed changes in U.S. dairy policy to amore market oriented system, the use of bothadvertising and non-advertising promotion effortsto maintain industry revenues can be expected toincrease. Understanding household impact isimportant to understanding the potential benefitsof such efforts.

Household level analyses of the impact of commod-ity promotion programs typically hypothesizeimpact on both quantity and timing of productpurchases (Neslin, Hendersen, and Quelch, 1985;Gupta, 1988,1991). Figure 1 provides one represen-tation of the relationship between product promo-tion, quantity purchased and the dynamics of thesepurchases. With commodity promotion effort, theremay be a direct positive impact on quantitypurchased and a negative impact on the length oftime (shorter) between purchases. With less timebetween purchases, this implies that householdstocks may be larger. The larger household stocks,the lower the amount purchased. The net effect ofpromotion, therefore depends on the direct pur-chase impact and indirect stock effects. Thepossible conflict of the direct and indirect quantityimpact implies that unless shorter interpurchasetime is recognized, comparing quantity consumedwith and without product promotion may overesti-

mate the true amount by which a particularpromotion campaign increases demand (Neslin,Hendersen, and Quelch, 1985, p.150). In terms ofdairy food manufacturers, dairy product promo-tion efforts may result in increased sales during apromotion period but the net impact will dependon the extent to which this increase is due toconsumers switching from other products versusan acceleration of purchases by current consumersbut with no increase in total purchases (Blattberg,Eppen, and Lieberman, 1981).

The phenomenon of purchase acceleration may beviewed positively or negatively from the retailersview depending on marketing environment. Forexample, in a competitive marketing environment,purchase acceleration may be an explicit objectiveundertaken by retailers to counter-act anticipatedmarketing efforts of competitors. Alternatively,without such a predatory situation, purchaseacceleration may result in consumers stockpilingthe commodity that they would have purchasedregardless of the promotion (Neslin, Hendersen andQuelch, 1985).

For this project we made a first attempt to examinethe relationships shown in Figure 2, by focusing onthe dynamics of the consumer purchase process(Gould, 1996b). In particular, we examined theeffect of coupon-based price deals on the timebetween purchases (e.g., interpurchase times). Weapply econometric models of duration to a fre-quently purchased food commodity, cheese. Weconducted this analysis by using the 170 weekconsumer panel a data set which consists of ahousehold panel observed over 170 consecutiveweeks from March, 1991 to June 1994. Besidespurchase quantity and price, information withrespect to coupon use and household demographiccharacteristics are available.

We used event history analysis to analyze purchasesof four cheese types: all (excluding cottage),processed, natural Cheddar, and cottage cheese. Avariety of duration models were used in ouranalyses under alternative assumptions concerningthe shape of the frequency distribution of theamount of time between purchase occasions(interpurchase time), the role of household charac-teristics in the determination of such distributions,


44

Figure 1. Relationship between interpurchase time, quantity purchased andcommodity production.

Cheese Type Number ofHouseholds

All Cheese 658 52.5 18.8 3.2 3.3 2.9

Processed 526 35.0 18.7 4.8 4.7 5.0Cheese

Natural 324 28.5 13.3 5.8 5.7 6.4Cheddar

Cottage 529 34.9 3.4 4.8 4.7 6.1

Coupons Coupons

MeanPurchaseWeeks

Percent ofPurchaseWeeks WithCoupon Use

Cheese

All Weeks Without With

Interpurchase Timea

Table 4. Cheese Purchase Characteristics

aThese means are calculated over all purchase occasions and households.

Milkfat

45

Figure 2. Hazard rate profiles for alternative household types.The profiles are generated by using mean values of the demographic variables for the varioussubgroups. For example, Hispanic hazard rate profiles for all cheese are calculated based onmean values of the demographic variables for Hispanic households only. Non-minority house-holds for these figures are assumed to be households classified as neither Black nor Hispanic.


46

and the presence of unobserved heterogeneity(variation) across households in these distribu-tions.

In our duration models we hypothesized that thedistribution of cheese specific interpurchase timeswere influenced by the use of coupon-based pricedeals as well as household demographic character-istics. Along with two variables used to capture theeffect of coupon redemption on the timing ofcheese purchases, demographic variables were alsoincluded as distribution “shifters.” They werehousehold size, age composition of householdmembers, ethnicity of main meal planner, ratio ofhousehold income to poverty threshold incomes,lagged purchases, whether there was a shelf-pricechange since last purchase, dummy variablesidentifying the Thanksgiving/Christmas periodand summer months, and urbanization of resi-dence area.

Table 4 shows cheese purchase characteristics forour cheese purchase panel. For all cheese, we foundthat over 18 percent of the purchase weeks oc-curred with the use of some type of cents-offcoupon. This is similar to what we found forprocessed cheese. Only 3 percent of cottage cheesepurchase occasions involve the use of some type ofcoupon. From an examination of averageinterpurchase time there appeared to be someacceleration of purchases. For processed cheese therewas a 1.9 week difference in mean interpurchasetime for weeks associated with coupon use versusnon-coupon use. This compared with .6 weeks forCheddar and 1.1 weeks for cottage cheese.

Table 5 provides mean values of exogenous pur-chase and household characteristics used to explaincheese interpurchase time. Using the results fromthe duration models applied to the four cheeses, weconducted likelihood ratio tests on the null hypoth-esis that the distribution of interpurchase time isnot impacted by household demographic charac-teristics. This was rejected for all four cheese types.We also tested the null hypothesis that couponredemption does not influence the length of timebetween purchases. This hypothesis was alsorejected.

From the above results, we simulate “hazard” ratesprofiles for the above four cheeses. These hazardrate profiles show the instantaneous probability ofa household purchasing given the number of weekssince last purchase. Figure 1 shows hazard rates fornon-minority, Black, Hispanic, and single personhouseholds.3 From this figure we see that asinterpurchase time increases, Hispanic householdsexhibit the largest instantaneous probability(hazard rate) of purchase for all cheese, while singleperson households show the lowest. For all cheese,after 4 weeks, the hazard rate for the simulatedsingle-person household is slightly more than .48compared to more than .67 for Hispanic house-holds. Single-person households have the lowesthazard rates except for cottage cheese where Blackhouseholds exhibit the lowest profile.

Coupon-based incentive programs continue to bean important marketing tool. The present analysisinvestigated one facet of coupon usage, namely itsimpact on the timing of purchases for a frequentlypurchased non-durable commodity, cheese. Weestimated a series of duration models for fourcheese classifications: all, processed, naturalCheddar, and cottage cheese. We incorporated withinthese duration models household demographiccharacteristics that allow for the distribution ofinterpurchase times to vary across households. Theresults of likelihood ratio tests indicate that thehousehold characteristics included as distributionshifters are statistically significant factors.

Two variables are used to account for the impact ofcoupon redemption on interpurchase times.Likelihood ratio tests clearly reject the null hypoth-esis that coupon use has no impact on the timing ofcheese purchases. As hypothesized, the use ofcoupons results in reduced interpurchase times forall cheeses. This impact, however varies acrosscheese type, especially when considering the typeof household doing the purchasing.

Previous analyses have shown, coupon promotionhas a direct impact on purchase time and quantitypurchased and indirect impact on quantity pur-chased that may counteract the direct impact. Anarea of future research is one of developing a modelwhich takes into account the simultaneous

Milkfat

47

Variable

Definition

ExpectedType of C

heese

All

Processed Cheddar C

ottage Purchase C

haracteristicsPU

RC

H-LA

Ga

Am

ount purchased on last purchase occasion (lb.)_

1.21.2

1.01.5

CO

UP-V

ALU

Eb

Value of coupons used per occasion ($/lb.)

_0.80

0.741.02

0.47PO

S-CH

AN

GE

aR

atio of change in price to reference price when

--0.156

-0.112-0.064

-0.077there is a price increase (#)

NEG

-CH

AN

GE

aR

atio of change in price to reference price when

+0.116

0.0880.056

0.062there is a price decrease (#)

SUM

MER

aD

umm

y variable for June, July and August (0/1)

+0.279

0.2630.241

0.279H

OLID

AYa

Dum

my variable for N

ovember and D

ecember (0/1)

+0.127

0.1490.162

0.127

Household C

haracteristicsIN

V-H

HSIZ

Ec

Inverse of household size (1/# of Mem

bers)_

0.4770.466

0.4480.491

BLA

CK

cD

umm

y variable =1 if m

eal planner is Black (0/1)

?0.055

0.0490.062

0.011H

ISPAN

ICc

Dum

my variable =

1 if meal planner is H

ispanic (0/1)?

0.0330.061

0.0250.026

PER-5

cPercent of household m

embers less than 6 years (%

)_

0.0340.055

0.0500.044

PER6-13

cPercent of household m

embers betw

een 6 and 13 years (%)

_0.065

0.0750.081

0.059PER

14-18c

Percent of household mem

bers between 14 and 18 years (%

)_

0.0460.047

0.0420.036

POV-R

ATIOc

Ratio of household incom

e to poverty threshold income (#)

?3.43

3.353.45

3.31SU

BU

RB

cD

umm

y variable = 1 if household resides in suburb (0/1)

?0.120

0.1310.167

0.140R

UR

AL

cD

umm

y variable = 1 if household resides in rural area (0/1)

?0.064

0.0860.040

0.083

Note: The “Expected Sign” colum

n refers to hypothesized coefficient signs for variables in equations (11) and (13).a indicates m

ean taken over all purchase occassionsb indicates m

ean taken over all purchase occassions where there is a coupon used

Nam

eSign

Table 5. Definition and M

ean Values of D

emographic and C

heese Purchase Characteristics


48

decisions of coupon use, interpurchase time andquantity purchased. Previous analyses have notrecognized the simultaneous nature of theseconsumer decisions.


Our work provides the foundation for futureresearch to determine the net effect of couponpromotion on commodity demand and whether thecosts of such promotion are justified. We alsointend to expand the scope of this analysis byexamining purchase timing on other dairyproducts.

References

Blattberg, R.C., G.D. Eppen, and J. Lieberman, 1981.A Theoretical and Empirical Evaluation of PriceDeals in Consumer Nondurables, Journal ofMarketing, Vol.45:116-129.

Blaylock J.R. and W.N. Blisard, 1988. Effects ofAdvertising on the Demand for Cheese, TechnicalBulletin 1752, USDA, ERS, Washington D.C.

Blaylock J.R. and W.N. Blisard, 1990. Effects ofAdvertising on the Demand for Cheese, January1982-June 1989, ERS Staff Report No. AGES 9055,USDA, ERS, Washington D.C.

Carlson, K.A. and B.W. Gould, 1994a. The Role ofHealth Knowledge inDetermining Dietary FatIntake, Review of Agricultural Economics,Vol.16(3):373-386.

Carlson, K.A., and B.W.Gould, 1994b. At-HomeDairy Product Demand: Projections to the Year2010, unpublished, Phase III Report: The Implica-tions of Changing Dairy Product Consumption forthe Demand for Farm Milk, Center for DairyResearch, University of Wisconsin-Madison, April.

Center for Dairy Research, 1994. Lower Fat CheeseResearch Coordination Conference,unpublished,University of Wisconsin-Madison, January.

Cheese Reporter, 1995. “MilkPEP Having PositiveImpact on Consumer Milk Attitudes,”Vol.120(24):1.

Forker, O.D. and H.W. Kinnucan, 1991. EconometricMeasurement of Generic Advertising, Special IssueNo. 9202, International Dairy Federation, Brussels,Belgium.

Gould, B.W., 1996a. Factors Affecting U.S. Demandfor Reduced-Fat Fluid Milk, Journal of Agriculturaland Resource Economics, Vol.21(1):1-14.

Gould, B.W., 1996b. Consumer Promotion andPurchase Timing: The Case of Cheese, AppliedEconomics, forthcoming.

Gould, B.W. and K.A. Carlson, 1993a. Description ofNutrient Intake by Households in the 1991/92Nielsen Dairy Product Consumer Panel, unpub-lished manuscript, Center for Dairy Research,University of Wisconsin-Madison, September.

Gould, B.W. and K.A. Carlson, 1993b. The Role ofHealth Knowledge in Determining Nutrient Intake,unpublished, Phase II: The Implications of Chang-ing Dairy Product Consumption for the Demand forFarm Milk, Center for Dairy Research, University ofWisconsin-Madison, December.

Gould, B.W., K.A. Carlson and H.C. Lin, 1993.Consumption Characteristics for Full and ReducedFat Dairy Products, unpublished, Phase I Report:The Implications of Changing Dairy ProductConsumption for the Demand for Farm Milk,Center for Dairy Research, University of Wisconsin-Madison, December.

Gupta, S., 1991. Stochastic Models of InterpurchaseTime with Time-Dependent Covariates, Journal ofMarketing Research, Vol.28:1-15.

Gupta, S., 1988. Impact of Sales Promotion onWhen, What, and How Much to Buy, Journal ofMarketing Research, Vol.25:342-55.

Kaiser, H.M. and J.C. Roberte, 1995. Impact ofGeneric Fluid Milk Advertising on Whole, Lowfat,and Skim Milk Demand, NICPRE Report, 95-02,Department of Agricultural, Resource, and Mana-gerial Economics, Cornell University, Ithaca,August.

Milkfat

49

Kinnucan, H.W. and O.D. Forker, 1986. Seasonalityin the Consumer Response to Milk Advertisingwith Implications for Milk Promotion Policy,American Journal of Agricultural Economics,Vol.63(3):563-571.

Neslin, S.A, C. Hendersen and J. Quelch, 1985.Consumer Promotions and the Accelerationof Product Purchases, Marketing Science,Vol.4(2):147-165.

Sun, T.Y., N. Blisard and J.R. Blaylock, 1995. AnEvaluation of Fluid Milk and CheeseAdvertising, 1978-1993, Technical Bulletin 1839,USDA, ERS, Washington D.C.

Ward, R.W. and B.L. Dixon, 1989. Effectiveness ofFluid Milk Advertising Since the Dairy and To-bacco Adjustment Act of 1983, American Journalof Agricultural Economics, Vol.71(3):730-740,1989.

Ward R.W. and W.F. McDonald, 1986. Effectivenessof Generic Milk Advertising: A Ten-Region Study,Agribusiness, Vol.2(1):77-90.

Chapter 2

Nonfat Solids ResearchSummaryThe value added utilization of whey and whey components continues to be the focus forCDR’s nonfat solids program. Damodaran’s patented process, a modified version of theprocedure in CDR FY94 Annual Report, for lipid removal from whey is now used com-mercially to produce higher value whey products. CDR researchers are exploring othernew technologies to develop innovative products and processes which increase thedemand for whey products.

Another value added process involves purifying individual proteins from whey. MarkEtzel has been researching the feasibility of using ion exchange membranes to isolatethese valuable proteins from whey. Last year, he was able to purify glycomacropeptidefrom whey using an ion exchange membrane but further work was necessary to isolateother proteins, including lactoperoxidase and lactoferrin. Mark and his group developeda lab scale purification process for both the lactoperoxidase and lactoferrin, successfullyscaling this up to a pilot scale process. However, when they attempted the scale upprocess for the glycomacropeptide it developed excessive back pressure across themembrane. Mark is now working with a whey manufacturer, an equipment manufac-turer and an ion exchange bead manufacturer all interested in solving the problem andbringing the glycomacropeptide purification technology to a commercial scale.

Research focused on producing propylene glycol from whey or whey permeate continuesto progress. After a series of trials, the growth media has been optimized. Now, continu-ous removal of inhibitory products from a fermentation has been demonstrated using anion - exchange column.

Technology transfer

CDR Open House, March 27, 1996


52

Nonfat Solids

53

INTERIM REPORT

Fractionation of Whey Proteins Using Ion ExchangeMembranes

Personnel: Mark. R. Etzel, associate professor, Dept.of Food Science; Clovis Ka Kui Chiu, graduatestudent, Dept. of Food Science; Ida A. Adisaputro,student, Dept. of Chemical Engineering

Dates: November 1993 – November 1996


Objectives

The overall objective of this research is to developan economical large-scale technology for producingpure individual whey proteins so that existing wheyprotein concentrate manufacturers can convert overwithout having to invest in a new plant. This newtechnology is needed to exploit the unique nutri-tional and functional properties of dairy proteinsnot found in other proteins derived from soy beansand eggs. The specific objectives are to:

1. Show that glycomacropeptide, lactoferrin andlactoperoxidase can be fractionated from wheyusing ion-exchange membranes.

2. Show that beta-lactoglobulin, alpha-lactalbuminand immunoglobulin G can be fractionated intopure products using multicomponent adsorptionbehavior of the ion-exchange membrane.

3. Optimize the fractional properties by modifyingprocessing conditions such as whey pH and loadingvolume, and eluant pH.

4. Scale-up the process by collecting pilot-plantdata needed to design, build and operate a success-ful commercial-scale process.

Summary

In the past year, lactoferrin and lactoperoxidasewere fractionated from whey using a sulfopropylion exchange membrane cartridge (model S100X,Sartorius Corp., Edgewood, NY) having a mem-

brane volume of 2 ml. As opposed to prior workusing a large pore-size membrane (50-300 µm), nomass transfer limitations were observed using thissmall pore-size membrane (3-15 microns). Usingfiltered whey (0.7 µm filter paper, Micro FiltrationSystems, Dublin, CA), flow rates of 10 and 50 ml/min were used without significant pressure drop orloss of performance. Up to 250 ml of whey wasloaded before breakthrough of lactoperoxidase orlactoferrin. Loading more whey caused the lactop-eroxidase activity of the effluent to exceed thelactoperoxidase activity of the inlet whey becauselactoferrin, which binds more tightly than lactoper-oxidase, displaced lactoperoxidase from themembrane. Using a 250 ml loading volume of whey,essentially all the lactoferrin and lactoperoxidasewere recovered from the whey. Two elution peakswere collected. In the first peak, using 0.3 M NaCl,pure lactoperoxidase was obtained. In the secondpeak, using 0.9 M NaCl, pure lactoferrin wasobtained. Repeated loading and elution cycles wereperformed in sequence without cleaning betweencycles, and performance was stable. Based on thesesuccessful laboratory-scale experiments, a pilot-plant-scale membrane having a volume of 16 mlwas used to scale-up the process by eight-fold. Thecycle time was 9 min, the flow rate was 400 ml/min,and the loading volume was 2000 ml. Matchingperformance was observed, making the next stepthe commercial-scale evaluation of the newprocess.

Recent efforts centered on fractionatingglycomacropeptide from whey. It is the moietycleaved from kappa-casein by chymosin duringcheese making. It occurs at a concentration of 1.2 to1.5 g/L in sweet whey, comprising 15 to 20% of thetotal protein. Glycomacropeptide is an example of apurified whey protein that has unique medical orhealth benefits not found with other proteins.Developing a technology for fractionation ofglycomacropeptide from whey will allow produc-tion of a high-value nutraceutical product fromwhey.


54

Earlier in this research, it was shown thatglycomacropeptide could be fractionated fromwhey using a quaternary methyl amine ion ex-change membrane cartridge (model PSC10-QM,BPS Separations Ltd., Spennymoor, CountyDurham, U.K.) having a 10-mL membrane volumeand a large pore size. Optimal conditions were:whey pH of 5.5-6.0, loading volume of 100-150 ml,flow rate greater than 15 ml/min, and 0.3 M NaClfor elution. The process was scaled-up by a factor offour using the optimal conditions. Ten cycles weresuccessfully conducted, yielding 3.2 gglycomacropeptide from 4.8 L of whey, or ~ 50%recovery.

In recent research, the Sartorius membrane wasutilized for recovery of glycomacropeptide, butpressure build-up across the membrane wasexcessive. Subsequently, ion exchange beads(GibcoCel, Life Technologies, Gaithersburg, MD)packed into a column were used to successfullyrecover glycomacropeptide from whey. Thisresearch led to a joint project involving a Wisconsinwhey processor, a dairy equipment manufacturer,an ion exchange bead manufacturer, and theUniversity to evaluate the process on a commercialscale. The graduate student on this project per-formed an industry-sponsored internship in thesummer of 1996 to help operate and characterizethe prototype process.


Although the U.S. market for high value, purifiedwhey proteins is small at present, several factorsindicate it will grow. Other mayor dairying coun-tries are developing markets for these purifiedindividual whey protein products. In the future, thefood industry will demand proteins with highernutritional and functional properties because ofthe trend towards foods with enhanced healthbenefits, lower fat content and lower lactosecontent.

The proposed ion exchange membrane processoffers numerous advantages. Evaluating andoptimizing the process should provide the stimulusfor the U.S. dairy industry to expand into the newdomestic and international markets for whey

protein products with enhanced nutritional andfunctional properties.

Publications/Presentations

Adisaputro, I.A., Wu, Y.-J., and Etzel, M. R. Strongcation and anion exchange membranes and beadsfor protein isolation from whey. J. Liq. Chrom. & Rel.Technol., 19(9), 1437-1450 (1996)

Nonfat Solids

55

INTERIM REPORT

Conversion of Whey Permeate to Propylene Glycolfor Food and Non-Food Uses

Personnel: Douglas C. Cameron, associate professor,Dept. of Chemical Engineering, Mark R. Etzel,associate professor, Dept. of Food Science, Nedim E.Altaras, graduate student, Dept. of ChemicalEngineering, Roxanne M. Smith, Yi-Jui Wu, stu-dents, Dept. of Chemical Engineering


Funding: Dairy Management Inc. CME95

Objectives

The overall objective of this project is to develop afermentation process for converting lactose in wheypermeate to propylene glycol, a large volumecommodity chemical. The fermentation processyields an optically pure product, R-propylene glycol,with added value over synthetically producedpropylene glycol. The specific objectives are to:

1. Screen and select microorganisms capable offermenting lactose to propylene glycol.

2. Optimize the medium and environmentalconditions for propylene glycol production fromwhey permeate by the organism(s) identified inObjective 1. The effects of carbon to nitrogen ratio(C/N), trace nutrients, and medium pH will beinvestigated.

3. Develop technology to remove inhibitory acetateand lactate from whey permeate-based mediumduring fermentation to achieve increased propyleneglycol concentration and productivity

Summary

During the first year of this project we demon-strated that Clostridium thermosaccharolyticumHG-8 (ATCC 31960) produces some propyleneglycol on unhydrolyzed whey permeate supple-mented with yeast extract. We have now established

that C. thermosaccharolyticum also producespropylene glycol from hydrolyzed whey permeate(WPH) and WPH with added yeast extract, yieldingtiters of 4.5 g/l in 48 h in 300 ml anaerobic flasks.Both glucose and galactose were utilized in thefermentations. To further investigate this finding,eight different whey-based media fermentationswere carried out in triplicate for a total of twenty-four fermentations. The eight media were whey orwhey permeate, with or without prior hydrolysis oflactose, and with or without added yeast extract.The major products of the C.thermosaccharolyticumfermentation are propylene glycol, acetate, lactateand ethanol. Fermentations of WPH and hydro-lyzed whey gave greater propylene glycol selectivity(moles propylene glycol per moles of by-products)compared to the media with added yeast extract.However, media with yeast extract produced morepropylene glycol in absolute concentration.Unhydrolyzed whey and unhydrolyzed wheypermeate were converted to propylene glycol onlywith the addition of yeast extract, and selectivitywas lower than it was for hydrolyzed medium.

Although the above research utilizing anaerobicflasks demonstrated proof of concept, and allowedinitial optimization of media composition andenvironmental conditions, productivity was limitedby accumulation of inhibitory products. The focusof the work on Objective 3 was to increase theproductivity and final titer of the fermentation bydeveloping new technology for removal of inhibi-tory products from the broth with ion exchangeresins simultaneous to the fermentation. Whenacetate and lactate anions are not removed duringthe fermentation, low productivity results due toproduct inhibition. When acetate was added to alevel of 20 g/l during the fermentation, cell masswas significantly reduced, providing evidence ofstrong product inhibition. Our research establishedthat a column packed with Amberlite IRA-400effectively removes lactate from uninoculateddefined glucose medium, while controlling pH.


56

When one anion binds to the column, one hydrox-ide ion was displaced, which controlled the pHwithout the need for adding base. In subsequentresearch, we used a column to extract lactate andacetate from an actual fermentation using definedglucose medium. We set-up a fermentation systemwhere broth was continuously pumped from awater-jacketed (60ºC) one-liter fermentor into amicroporous hollow-fiber filter. Cells were returneddirectly to the fermentor, and cell-free broth passedthrough an anion-exchange column before it wasreturned to the fermentor. By establishing atechnology for the simultaneous removal of acetatefrom the fermentation broth, we met Objective 3 ofthe project.


This research will provide the basis for the indus-trial production of a new fermentation chemical,propylene glycol, from whey permeate. Propyleneglycol is a major organic chemical with extensiveapplications in the food industry, both as aningredient and as an antifreeze and heat transferfluid. Currently propylene glycol is producedentirely from propylene, a petrochemical. Thissynthetic propylene glycol is a racemic mixture, amixture of left and right handed forms of propyleneglycol. The fermentation process provides a uniqueroute to enantiomerically pure R-propylene glycol.Since the other major non-volatile products of thefermentation are acetic and lactic acid, it may alsobe possible to concentrate the fermentation brothand make a cultured whey product containingpropylene glycol and organic acids.

Chapter 3

Summary

Cheese ResearchWhat makes a Cheddar cheese taste really good? How does that remarkable flavor develop?These questions have been asked for many years and research continues to provide pieces ofthe flavor puzzle. Lindsay and Olson have been systematically looking at the role of milkfatand how it influences Cheddar flavor. They have determined that cell-associated lipases play akey role developing proper levels of free fatty acids and specific fatty acids are the key to goodCheddar flavor. In an associated project, Lindsay’s group is also looking into the mechanismof Cheddar flavor, and off flavor development, from the interaction of alpha-dicarbonylcompounds with other cheese constituents. The alpha-dicarbonyls are very reactive com-pounds. The reaction products can be involved in forming secondary browning flavorcompounds, which may be related to browned or brothy flavor defects. In a joint project,Wisconsin CDR researchers have been working with scientists at Western Center for DairyProtein Research and Technology, characterizing and evaluating selected starter and adjunctcultures used in cheese making. They have combined selected starter and adjunct culturesand then used them to make reduced fat Cheddar cheese. The cheeses were evaluated by bothsensory evaluation as well as instrument analyses. In addition, intracellular aminopeptidaseand esterase/lipase activities of the cultures have been analyzed.

Understanding the physical properties of cheeses, and how those properties develop, are bothessential elements needed to produce better reduced fat cheeses. This, in turn, will help toexpand the utilization of cheese as a food ingredient. Gunasekaran, Olson, and others havebeen modifying and developing procedures to understand the physical properties of cheese.Three-dimensional digital analysis, in conjunction with a confocal laser scanning micro-scope, is providing a picture of the structure of cheese as it ages. Objective machinability testsand refined rheological methods are providing more information about the physical charac-teristics of cheeses as a function of composition, manufacturing technique, and age.

Reduced fat and specialty cheeses continue to be an important part of CDR’s cheese technol-ogy program. CDR researchers Carol Chen and Mark Johnson have been granted a patent(Pat. # 5,554,398) for a make procedure that produces flavorful reduced fat Cheddar cheese.Reduced fat Cheddar made with this procedure was featured at the CDR booth at the Instituteof Food Technologists (IFT) annual meeting this year. The product was well received andcompanies have expressed interest in the technology. A no brine, no mixer molder pizzacheese has also been developed and was shown at the Dairy Management Inc. booth at theIFT meeting. Sold out specialty cheese seminars, one on Swiss cheeses and the other onEnglish cheeses, demonstrate the timeliness of these programs. These courses are valuablebecause they provide the “hands on” type of instruction that cheese makers appreciate. Gouldand Carlson report some interesting insight into the cost of production as they examined theeconomics of specialty cheese production in various regions of the country.


58

Technology Transfer

WCMA/CDR Cheese conference, April, 1996

CDR Open House, March 27, 1996

CDR booth at Institute of Food Technologists Annual Meeting, New Orleans,June 22-26, 1996

Cheese Research

59

INTERIM REPORT

Cheese Making Properties of Milk from Cows ofDifferent Milk Protein Genotype

Personnel: Robert Bremel, professor, JosieLewandowski, research specialist, Joshua Ruff,student, Tricia Braaksma, student, Animal Science,Mark Johnson, senior scientist, John Jaeggi, assis-tant researcher, Carol Chen, researcher, Brian Gould,senior scientist, Center for Dairy Research



Objectives

Year One1. Select animals from the herd based ongenotyping for kappa casein

2. On farm set-up to segregate milk based ongenotype

3. Manufacture cheese using segregated milk

4. Standard cheese analysis (sensory, composi-tional, functional)

5. Economic evaluation of milk segregation

Year Two1. Complete objectives 3, 4, and 5 from above

2. On-farm establishment of experimental herdsspecific for particular genotype combinations

Summary

We are working with a producer who has followed abreeding program that produces cows with a higherproportion of B allele kappa casein animals. Thisherd consists of approximately 200 animals that arecurrently milking or will be milking in the nextyear.

We delayed the start of this project because of theunusually hot weather during the summer of 1995.So far, we collected blood samples and extractedDNA to perform molecular genotyping. Results arelisted below.

Blood Sampling Results:

➞ ≅ 200 head

➞ 7 BB genotype (more kappa as % of total casein, less beta, less alpha)

➞ 30 AB genotype

➞ The rest AA genotype

We have now obtained small bulk tanks andestablished a procedure for collecting milk. We haveselected animals of each genotype, enough toprovide milk for a 500 lb. lot for cheese makingfrom each group. Some differences have alreadybeen noted in milk performance, but it is too earlyto determine any trends in the data.


The milk protein, kappa casein, can affect moistureand texture in different kinds of cheese. The Ballele, a variant of the gene for kappa casein, ispredominant in Brown Swiss and Jerseys. The Aallele is more common in American Holsteins. TheB allele produces kappa casein that enhancesclotting speed during cheesemaking and thus itmay be a more favorable genotype. If our study canconfirm an advantage, then breeding cows for the Ballele might be a productive way to influencecheesemaking through milk composition.


60

INTERIM REPORT

Studies of the Influence of Milkfat on the Formation ofFlavor Compounds in Cheddar Cheese

Personnel: Robert C. Lindsay, professor, Dept. ofFood Science; Norman F. Olson, professor, Dept. ofFood Science; David Bogenrief, associate re-searcher, CDR; Qiaoling Zeng, graduate researchassistant, Dept. of Food Science

Dates: October 1993 – December 1996

Funding: Wisconsin Milk Marketing Board 93-8,Kraft-General Foods

Objectives

1. To investigate the basic physical and chemicalinfluences of milkfat on the development of flavorcompounds in Cheddar-type cheeses.

2. To use the information to devise strategies formanufacturing low fat Cheddar cheeses with flavorssimilar to traditional full fat cheese.

Summary

In order to simulate more appropriately the condi-tions encountered in industry for the manufactureof Cheddar cheese, pilot-sized, mechanically-stirredcheese vats have been installed and used through-out the experiments. Trials composed of thefollowing lots of cheese are made when the fullrange of fats in Cheddar cheese is desired: full fat(32 % fat); 25% reduced fat (25% fat); 33% reducedfat (18% fat); 50% reduced fat (12% fat); 75%reduced fat (3% fat) no fat skim milk (defined as“0%” fat, but 1.2-2.5% fat).

Cheeses have been manufactured with severalculturing systems which were established withindustry collaboration on identifying desirabletraits. The cheeses were evaluated for flavor by theproject team (both UW and Kraft) according to thefollowing schedule based on achieved age of thecheese: 1-2, 3, 6, 9, and 12 months. When somecheese samples progressed to a point that the

characteristics were beyond commercial or scien-tific value they were removed from the experimen-tal design.

Flavor compounds in cheeses from the trials wereanalyzed according to the age-based samplingscheme, and included, where appropriate, thefollowing analyses: α-Dicarbonyls (glyoxal,methylglyoxal, diacetyl) by HPLC; neutral andsulfur headspace volatiles by GC; higher boilingflavor compounds by acetonitrile extraction; freefatty acids by alumina adsorption and GC; freeamino acids and selected non-volatile compoundsby HPLC.

Results have shown that the fat content greatlyinfluences the level of free fatty acids in cheese.Cheeses containing less fat than that obtained at a33% reduction do not develop adequate levels ofbutyric acid and other free fatty acids to providecheesy flavors. Addition of free lipase yieldsexcessive levels of free fatty acids. The key discov-ery is that cell-associated lipases of microbes innormal ripened cheese provide the required level offree fatty acids. Selected cultures of lipase-positivelactobacillus organisms used as adjunct culturesprovide proper levels of fatty acids.

Studies have revealed that several postulated aminoacid metabolites are present in the aging cheeses,and appear involved in the flavors of reduced fatcheeses. Included are p-hydroxybenzoic acid,phenyllactic acid, p-hydroxyphenyllactic acid, andmethyl p-hydroxyphenyllactic acid. These com-pounds are influenced by the conditions in full andreduced fat cheeses, and may be involved indistinguishing cheese flavors.

Cheeses that have 75%, or greater, fat reductionbuild-up much greater levels of methylglyoxal thanthose containing higher levels of milkfat. Addition-ally, the levels of methylglyoxal and otherdicarbonyls decrease during the later stages of

Cheese Research

61

ripening, and appear associated with the develop-ment of brothy, umami, and protein-like flavordefects. The compounds that cause these flavors arenot clearly indicated at this point, but work iscontinuing on their identification.


One of the major limitations in the development ofexpanded markets for lowfat Cheddar cheese is thedevelopment of off-flavors and the lack of develop-ment of typical desirable cheese flavors. Removingmilkfat and selecting cultures by trial and error hasonly produced partially acceptable reduced fatcheeses. This research is unraveling the complexassociations between the physical aspects and theflavor chemistry aspects of cheese flavor develop-ment, and will result in improved flavors in lowfatcheeses. Such cheeses will greatly increase thedemand for lowfat cheeses and will allow the dairyindustry to provide desirable contemporarycheeses that the consumer wants.


62

INTERIM REPORT

Mechanisms for Production of Cheese FlavorCompounds

Personnel: Robert C. Lindsay, professor, Dept. ofFood Science; Christine Nowakowski, graduateresearch assistant, Dept. of Food Science

Dates: July 1993 – December 1996

Funding: Dairy Management Inc. LD294

Objectives

1. To chemically define mechanisms for the forma-tion of Cheddar cheese flavor compounds frominteractions of alpha-dicarbonyl compounds(glyoxal, methylglyoxal, and diacetyl) with othercheese constituents (amino acids and peptides).

2. To investigate interactions between alpha-dicarbonyl and the aromatic amino acid flavorsystems in the development of Cheddar cheeseflavors.

3. To relate the findings from the research to theselection of lactic cultures used in the productionof Cheddar-type cheeses.

Summary

Quantitative analysis of cultures of lactic acidbacteria have documented the relative abilities ofthese organisms to produce glyoxal, methylglyoxal,and diacetyl. A specially developed method utiliz-ing bisulfite as a trapping agent has shown thatsubstantial amounts of dicarbonyls react veryrapidly with amino compounds after their produc-tion in culturing or cheese media. Dicarbonyls donot accumulate intracellularly, and are transportedout of cells. Using this procedure, cultures of lacticacid bacteria have been selected for further experi-ments based on their relative abilities (low, me-dium, and high) to produce the respectivedicarbonyls.

Survey studies have shown that the dicarbonyls,especially methylglyoxal and glyoxal, increase inyoung Cheddar cheeses (1 month), and thensteadily decline through 9 months aging. Thisperiod corresponds to that for aged cheese flavordevelopment in full fat cheese, but it also coincideswith the development of meaty-brothy off-flavorsin reduced fat cheeses.

We are continuing flavor chemistry studies on thesecondary browning flavor compounds that resultfrom initial amino acid and dicarbonyl reactions,and these particularly include the Strecker alde-hydes, pyrazines, and potentially the furanones.These compounds are produced in Cheddar cheese,including reduced fat cheeses, and particularly havebeen suggested to be related to the development ofbrowned or brothy flavor defects in low fat cheeses.

Investigations of rates of reactions of individualamino acids with the respective dicarbonyl com-pounds have shown that great differences betweenamino acids exist for this ability. Thus, likely cheeseflavor forming reactions have been identified basedon the resulting volatile compound properties andrelative rates of reactions. These processes areunder continued investigation in broth and cheese-based systems for their abilities to influence thedevelopment of cheese flavors. Included are thesulfur amino acids and aromatic amino acidswhich contribute to the low- and high-boilingflavor compound classes.

Cheese-based model slurry systems prepared byvarious methods of sterilization have been devel-oped. Slurries in barrier pouches that have beensterilized by selected ultra-high pressure treat-ments provide suitable models for cheese flavordevelopment studies. Studies using selected lacticacid bacteria, enzymes, and cheese flavor chemicalprecursors are in progress. Slurries are beinganalyzed for key indicator flavor compounds toassess their mechanisms of formation in cheese.

Cheese Research

63


Cheese flavor development is key to providingquality cheeses and ingredients to consumers andthe food industry. In order to develop and practiceculturing and manufacturing procedures that yieldconsistent and desirable flavors, it is essential tounderstand the basis of their formation. Thisresearch will lead to improved cheese flavors whichwill increase the demand for cheese and cheeseingredients.


64

INTERIM REPORT

Identification of Microbial Enzymes and MetabolitesInvolved in the Development of Lowfat CheddarCheese Flavor (Phase I and Phase II)

Personnel: James L. Steele, associate professor, UW-Madison Food Science, Mark E. Johnson, seniorscientist, Center for Dairy Research, Jeff Broadbent,assistant professor, Utah State Univ., Bart Weimer,assistant professor, Utah State Univ., Kristen Houck,research specialist, UW-Madison Food Science,Song Gao, research assistant, UW-Madison FoodScience, Ed Dudley, research assistant, UW-Madi-son Bacteriology, Jeff Christensen, research assis-tant, UW-Madison Bacteriology

Dates: June 1994 to June 1997


Objectives

Please note this report contains information from acollaborative project with funding from the WMMBand DMI (through the Western Center for DairyProtein Research and Technology). The principalinvestigators include Dr. Jeff Broadbent and Dr. BartWeimer from Utah State University and Dr. MarkJohnson and Dr. Jim Steele from The University ofWisconsin-Madison.

Stage 1:

1. A systematic characterization of metabolicproperties of starter cultures and flavor adjunctcultures.

2. Lowfat cheese (50% reduced) will be manufac-tured using various combinations of three startercultures and six flavor adjuncts.

3. Detailed sensory and chemical analysis of thelowfat cheese manufactured for objective 2.

4. Gene banks will be constructed of the selectedstarter and starter adjunct bacteria.

Stage 2: (The following are thought to be the mostlikely targets for the second stage; however, thesetargets will be altered in response to the resultsobtained in stage 1.)

5. To evaluate the role of primary proteolysis oncheese flavor development.

6. To characterize the influence of individualpeptidases from starter cultures and flavor adjunctson cheese flavor development.

7. To characterize amino acid degradation path-ways in starter cultures and flavor adjuncts andhow they influence cheese flavor development.

8. To characterize the influence of α -dicarbonylproduction by starter cultures and flavor adjunctson cheese flavor development.

Summary

The starter culture and flavor adjuncts used incheese trials were characterized with regard to theirintracellular general aminopeptidase activities andintracellular esterase/lipase activities. Significantvariation was observed among these strains forboth of these activities. The proteinase specificitywas determined for the three starter culturesemployed. Lactococcus lactis S1, SK11, and S3 weredetermined to have PI, PIII, and PI/PIII specificity,respectively. Previous investigators have suggestedthat cultures with PI specificity were more likely toproduce bitter cheese. Gene banks have beenconstructed of the selected starter and starteradjunct bacteria.

Forty-two vats of lowfat Cheddar cheese weremanufactured. A grid of three different startercultures and six different adjunct cultures (twostrains each of Lactobacillus helveticus, Lactobacil-lus casei, and Brevibacterium linens) was employed.

Cheese Research

65

Sensory analysis (both a small expert panel and alarge consumer panel) of these cheeses indicatedthat both the starter culture and the flavor adjuncthad a significant impact on the quality of the finalproduct. Both strains of Lb. helveticus and B. linensincrease the “Cheddar flavor intensity” and reducedbitterness and off-flavors. Cheeses made with Lb.casei were more bitter and had elevated levels ofoff-flavors. Cheeses made with Lc. lactis S3 wereconsistently very bitter.

Capillary electrophoresis, preparative HPLC,peptide sequencing, and mass spectroscopy havebeen utilized to identify seven distinct peptidesfrom the Cheddar cheese samples. The peptideswhich accumulated in specific cheeses are directlyrelated to the specificity of the proteinase of thestarter culture. A strong positive correlation (r2 =.90) between the presence of αs1-CN(1-9) and thelevel of bitterness in the cheese was observed. It isnot possible to determine at this point whether ornot this is a direct effect of αs1-CN(1-9) or simplyan indicator of proteinase specificity.

To determine which enzymes are most importantin Lb. helveticus CNRZ32’s demonstrated ability todebitter cheese when used as a starter adjunct,mutants lacking specific peptidases have beenconstructed. Specifically, mutants lacking the X-prolyl dipeptidyl aminopeptidase, a generalaminopeptidase designated PepC, or anothergeneral aminopeptidase designated PepN havebeen constructed. Additionally, all possible doubleand triple mutants of these three enzymes havebeen constructed. Next, these mutants will beexamined in a cheese slurry system to determinethe relative contribution of each enzyme toCNRZ32’s ability to debitter cheese.

The catabolism of aromatic amino acids is believedto play a significant role in the development ofunclean flavors in cheese. It will be important toexamine the interactions between the startercultures, non-starter lactic acid bacteria, andadjunct cultures in order to understand howaromatic amino acid catabolites are formed incheese. To initiate studies in this area, we havechosen to focus on characterization of the aromaticamino acid catabolism by lactococci, which areused as starter cultures in many ripened cheese

varieties. Screening of lactococcal strains forenzymes known to initiate catabolism of tryp-tophan was completed and a broad specificityaromatic amino acid transaminase was detected inall eight lactococcal strains examined. The varia-tion between strains for the level of this activitywas at least ten fold. Metabolites produced fromtryptophan by cell-free extracts of Lc. lactis S3 wereindolepyruvic acid, indoleacetic acid, and indole-3-aldehyde. Indoleacetic acid and indole-3-aldehydecan form spontaneously from indolepyruvic acidunder the conditions employed. A model systemwas developed to determine if theaminotransferase(s) was expressed and whichmetabolite(s) accumulate under conditions whichsimulate those of ripening Cheddar cheese. Theresults indicated that the aminotransferase(s) wasexpressed and stable in the model system. Thetryptophan metabolites which accumulated weredetermined to be strain specific. These resultsindicate that the aromatic amino acid catabolicpathways present in the starter culture may influ-ence whether or not unclean flavor compoundsaccumulate in cheese. Similar studies are currentlybeing conducted with various lactobacilli andbrevibacteria. The objective of this research will beto identify the metabolic pathways and interactionswhich lead to the accumulation of unclean flavorcompounds in cheese.


This project is a collaborative study betweenresearchers at the University of Wisconsin-Madisonand Utah State University. This collaboration bringstogether the expertise of the researchers in cheesemanufacture, physiology of lactic acid bacteria, andthe genetics of lactic acid bacteria in an attempt tosolve the problems of lack of flavor and off-flavorsin lowfat (50%) Cheddar-type cheeses. The ex-panded expertise and systematic approach is likelyto yield a significant advance in our understandingof lowfat cheese flavor development. The enhancedunderstanding of the basic biochemistry of cheeseflavor development will greatly facilitate thedevelopment of starter systems for the manufactureof high-quality, lowfat cheese. Lowfat cheese withthe organoleptic qualities of full-fat varieties willincrease consumer acceptance of lowfat dairy


66

products and expand the demand for these goodsto individuals that avoid cheese for reasons of dietand the absence of high quality lowfat alternatives.

Publications/PresentationsAbstracts

B. Weimer, C. Brennand, J. Broadbent, J. Jaegi, M.Johnson, F. Milani, B. Mistry, G. Reineccius, J. Steele,and M. Strickland. (1995). Chemical and sensoryattributes of 50% reduced-fat Cheddar made withvarious adjunct bacteria. International Dairy LacticAcid Bacteria Conference. Palmerston North, NewZealand. February 19-23, 1995. Abstract #S 3.4.

J. Steele, M. Johnson, J. Broadbent, and B. Weimer.(1995). Update: WCDR/Utah State lowfat cheeseresearch. Wisconsin Cheese Industry Conference.Green Bay, Wisconsin. March 22-23, 1995. Paper#12.

Weimer, B., Broadbent, J., Strickland, M., Steele, J.,Johnson, M., and Jaeggi, J. (1996). Differences incasein-derived peptides which accumulate in 50%reduced fat Cheddar cheese manufactured withvarious single-strain Lactococcus lactis starters.International Dairy Federation Symposium onRipening and Quality of Cheeses. 73.

Goa, S., Oh D-H., and J.L. Steele, J. (1996). Aromaticamino acid catabolism by lactococci. InternationalDairy Federation Symposium on Ripening andQuality of Cheeses. 22.

Johnson, M.E., J.L. Steele, B.C. Weimer, and J.R.Broadbent. (1996). Introduction to the session:Improvement in reduced-fat Cheddar cheese flavorthrough identification, isolation, and analysis ofenzymes and metabolites produced by starters andadjunct cultures. Amer. Dairy Sci. Abstr., 1996.

Gao, S., D-H. Oh, and J.L. Steele. (1996). Aromaticamino acid catabolism by lactococci. Amer. DairySci. Abstr., 1996.

Strickland, M., J. Broadbent, and B.C. Weimer.(1996). A combined HPLC and capillary electro-phoresis study of casein degradation during agingof Cheddar cheese. Amer. Dairy Sci. Abstr., 1996.

Brennand, C.P. (1996). Sensory evaluation of low fatCheddar cheese prepared with flavor adjuncts.Amer. Dairy Sci. Abstr., 1996.

Weimer, B.C., C. Brennand, M. Strickland, J.Broadbent, J. Jaegi, M. Johnson, and J. Steele. (1996).Flavor and chemical analysis of lower fat Cheddarcheese. Amer. Dairy Sci. Abstr., 1996.

Broadbent, J.R., B.C. Weimer, and M. Strickland.(1996). Influence of single-strain Lactococcus lactisstarters on the accumulation of casein-derivedpeptides in 50% reduced fat Cheddar cheese. Amer.Dairy Sci. Abstr., 1996.

Kong, Y., M. Strickland, and J.R. Broadbent. (1996).Tyrosine and phenylalanine catabolism by Lactoba-cillus casei flavor adjuncts: Biochemistry andimplications in cheese flavor. Amer. Dairy Sci.Abstr., 1996.

Gummalla, G., and J.R. Broadbent. (1996). Indoleproduction by Lactobacillus spp. in cheese: Apossible role for tryptophanase. Amer. Dairy Sci.Abstr., 1996.

Ummadi, M., and B.C. Weimer. (1996). Tryptophancatabolism in Brevibacterium linens and its influ-ence on lowfat Cheddar cheese flavor. Amer. DairySci. Abstr., 1996.

Dias, B., and B.C. Weimer. (1996). Degradation ofmethionine in low fat Cheddar cheese. Amer. DairySci. Abstr., 1996.

Christensen, J.E. and J.L. Steele. (1996). Character-ization of aminopeptidase-deficient Lactobacillushelveticus CNRZ32 derivatives. FEMS Fifth Sympo-sium on Lactic Acid Bacteria,

Publications

Gao, S., D-H. Oh, M.E. Johnson, J.R. Broadbent, B.C.Weimer, and J.L. Steele. 1996. Aromatic amino acidcatabolism by lactococci. submitted

Weimer, B., B. Diaz, U. Madhavi, J. Broadbent, C.Brennard, J. Jaegi, M. Johnson, F. Milani, J. Steele,and D.V. Sisson. 1996. Influence of NaCl and pH on

Cheese Research

67

intracellular enzymes which influence Cheddarcheese ripening. submitted

Presentations

“Influence of proteolysis and amino acid catabo-lism on lowfat Cheddar cheese flavor development.”Dairy Management Inc. Lowfat cheese researchplanning meeting. November 1995.

“Aromatic amino acid catabolism by lactococci”. Atthe International Dairy Federation’s symposium on“Ripening and Quality of Cheeses”. February 1996.

“Influence of proteolysis and amino acid catabo-lism cheese flavor development.” SoutheasternDairy Center. March 1996.

“Impact of lactic acid bacteria on cheese flavordevelopment: proteolysis.” Symposium on “Micro-bial Generation of Flavors and Pigments” at the1996 American Society of Microbiology AnnualMeeting. May 1996.

“Influence of proteolysis and amino acid catabo-lism cheese flavor development.” Systems Bio-Industries. May 1996.

“Lowfat cheese flavor development.” Kraft-GeneralFoods. June 1996

“Proteolytic enzymes of lactic acid bacteria andtheir importance in cheese flavor development.”Symposium on “Recent Advances in Lactic AcidBacteria” at the 1996 IFT Annual Meeting. June1996.

“How to reduce bitterness in lowfat cheese” at theTwelfth Biennial Cheese Conference. August 1996.


68

INTERIM REPORT

Contribution of Endopeptidases from Lactobacillushelveticus CNRZ32 to Cheese Flavor Development

Personnel: James L. Steele, associate professor, KurtM. Fenster, research assistant, Yo–Shen Chen,research assistant, Kirk L. Parkin, associate profes-sor, Dept. of Food Science, Mark E. Johnson, seniorscientist, CDR


Funding: Dairy Management Inc. SPJ95

Objectives

1. Characterization of the three endopeptidasegenes.

2. Construction of CNRZ32 derivatives with alteredendopeptidase activities.

3. Evaluation of the contribution of individualendopeptidases to degradation of casein–derivedpeptides.

4. Purification and characterization of selectedendopeptidase(s).

5. Determination of the role of selectedendopeptidase(s) in CNRZ32’s ability to reducebitterness and accelerate cheese flavor develop-ment.

Summary

An endopeptidase clone, designated EPII, waspreviously identified in a Lb. helveticus CNRZ32genomic library using the endopeptidase sub-strates, N–benzoyl–Phe–Val–Arg–pNA and N–benzoyl–Pro–Phe–Arg–pNA. Restriction analysisof this clone revealed that the clone contained a 2.5kbp insert. Tn1000 mutagenesis was used tolocalize the EPII gene to a 1.3 kbp region of theinsert. The insert was completely sequenced andthe open reading frame encoding EPII was found tobe 1,317 bp in length. A homology search with EPIIrevealed that EPII shared 40% identity with the

CNRZ32 aminopeptidase C and was a member ofthe cysteine proteinase family. Analysis of the EPIIgene strongly indicates that it is an intracellularendopeptidase due to the lack of a signal sequence.Analysis of the EPII gene also suggested that thisgene was transcribed monocistronically due to thepresence of a putative promoter region as well as astrong terminator. Northern analysis is currently inprogress to confirm that EPII is transcribedmonocistronically.

CNRZ32 mutants lacking EPII have been con-structed using a gene replacement technique.Preliminary characterization of these mutants inMRS broth, milk, and defined medium is inprogress to determine how EPII contributes togrowth of CNRZ32 in these media.

The purification of EPII is currently in progressusing a protein fusion purification kit supplied byNew England Biolabs, Inc. Preliminary steps, suchas the construction of the fusion protein necessaryfor the purification, have been completed.

To determine the role of EPII in casein hydrolysisby Lb. helveticus CNRZ32, isogenic strains differingin only their EPII activity will be evaluated, as wellas, the specificity of the purified EPII on casein–derived fragments.

A second endopeptidase clone, designated EPIII,was previously identified in a Lb. helveticusCNRZ32 genomic library using the endopeptidasesubstrates, N–benzoyl–Pro–Phe–Arg–pNA and N–benzoyl–Val–Gly–Arg–pNA. Restriction analysis ofthis clone revealed that the clone contained a 6.5kbp insert. This insert was subcloned and partiallysequenced. An open reading frame encoding EPIIIwas found to be approximately 2.0 kbp in length. Ahomology search with EPIII revealed that EPIIIshared 47% identity with the Lactococcus lactisPepO endopeptidase; therefore, EPIII was desig-nated PepO. The Lc. lactis PepO gene is part of anoperon containing the ATP–driven oligopeptide

Cheese Research

69

transport system. Identification of an open readingframe putatively encoding a protein associated withthe ATP binding cassette of the oligopeptidetransport system in CNRZ32 has been madeupstream from PepO. Research is currently inprogress to confirm the presence of theoligopeptide transport system and to characterizethe relationship between PepO and this transportsystem.

Construction of a PepO mutant in CNRZ32 is inprogress. The construct used to create a PepOmutant in CNRZ32 via gene replacement has beenmade. Once PepO mutants have been obtained,preliminary characterization of these mutants willbe made in MRS broth, milk, and defined medium.

To determine the role of PepO in casein hydrolysisby Lb. helveticus, isogenic strains differing in onlyPepO will be evaluated in cheese. PepO will alsohave to be purified to identify the specificity ofPepO on casein–derived fragments.


The construction of derivatives of Lb. helveticusCNRZ32 which differ only in the activity of anindividual endopeptidase, will allow the role of thatenzyme in flavor development to be demonstrated.If an endopeptidase is determined to be the rate-limiting enzyme for the degradation of bitterpeptides and/or the formation of free amino acidsor peptides with beneficial flavor attributes, thenstrains which overproduce this enzyme will beconstructed. These strains could then be used toproduce both traditional and lowfat cheese variet-ies; these cheeses should display less bitterness anddevelop cheese flavor more rapidly.


70

INTERIM REPORT

Process Modification of Starter Cultures for FlavorEnhancement in Lowfat Cheese

Personnel: Mark. R. Etzel, associate professor, BrianChi-Shung To, graduate student, Dept. of FoodScience


Funding: Dairy Management Inc. ETZ 95

Objectives

The overall objective of this research is to developmore flavorful lowfat Cheddar cheese by usingprocess-modified starter culture adjuncts thataccelerate the ripening process. Processing condi-tions are to be developed which reliably modify theculture characteristics to permit flavor enhance-ment in lowfat cheese. This allows the manufactureof more consistent and flavorful lowfat cheeses,thus increasing the demand for cheese and milk.The specific objectives are to:

1. Establish methods for small-scale production ofcell pastes of candidate cultures.

2. Freeze, freeze dry and spray dry the cell pastesolutions using processing conditions which rangefrom attenuating to preserving of metabolicactivity.

3. Analyze the cell pastes, solutions and cell pow-ders for metabolic activity, specifically cell survival,lactic acid production, and ß-galactosidase andaminopeptidase activity.

4. Select the cultures and processing conditionswhich are best for flavor enhancement in lowfatcheese, and then produce large amounts of adjunctneeded for cheesemaking.

Summary

The culture Brevibacterium linens ATCC 9174 wasselected as the first candidate for investigation. B.linens is responsible for the ripening of surface-

ripening cheeses such as Limburger and Brick.Incorporation of this obligate aerobe into Cheddarcheese, as an adjunct to the normal starter culture,has been shown to enhance the rate and extent offlavor development. Inexpensive methods forproducing, preserving, and distributing theseadjuncts cultures must be developed before wide-spread use in the dairy industry is economical. Forthis reason, the effects of processing (freezing,freeze drying and spray drying) on the characteris-tics of B. linens were measured in this research.

Survival was 100% for B. linens after freezing andfreeze drying. For spray drying, survival was halvedfor every 5 °C increase in the outlet air temperature.Heat was the sole mechanism decreasing cellsurvival during spray drying. Thermal resistancewas measured vs. moisture content and tempera-ture, resulting in D-values ranging from 120 min at25.3% solids and 44.0°C to 2.6 min at 46.2% solidsand 55°C. This is much more heat sensitive thanalkaline phosphatase, an indicator of properpasteurization of milk. However, thermal inactiva-tion of B. linens during spray drying can be elimi-nated by proper selection of the outlet-air tempera-ture. By extrapolation of the data, 100% survivalwould occur at an outlet air temperature of 57 °C.The freeze dried and spray dried cultures werestable during prolonged storage at 4 °C only ifoxygen and moisture scavengers were included inthe container.

In the second part of this investigation, five lacticacid bacteria, Lactococcus lactis ssp. cremoris D11,Lactobacillus casei ssp. pseudoplantarum UL137,Lactobacillus delbrueckii ssp. bulgaricus CH3B,Lactobacillus acidophilus NCFM and Streptococcussalivarius ssp. thermophilus CH3TH, were separatelyfrozen, freeze dried or spray dried, and tested forsurvival and lactic acid production before and afterprocessing. Greater than 86% of the cells survivedfreezing. Of these survivors, approximately twothirds survived the dehydration step of freezedrying, except for Lb. bulgaricus, for which one fifth

Cheese Research

71

survived. Survival after spray drying was greatestfor S. thermophilus and lowest for Lb. bulgaricus. Incontrast to freezing and freeze drying, spray dryingcaused a significant delay in lactic acid production.


The results of this research will help advance thecheese industry. The demand for lowfat cheese isgrowing. This new technology will help in develop-ing lowfat cheeses with a flavor and texture asdesirable as in full-fat cheeses. These lowfat cheeseswill be popular for producing spray-dried powdersfor use in lowfat convenience foods such as cheesenachos and frozen dinners, and will appeal toconsumers on lowfat diets who may have reducedtheir consumption of traditional full-fat cheeses.This new technology will help to accelerate thealready increasing demand for cheese, which willdirectly result in an increased demand for milk.

In addition, the results of this research will help toadvance the starter culture industry, most of whichis based in Wisconsin. According to Dr. DougWillrett, Technical Director at Marschall Products,“Much of the basic research on culture preservationis lacking and is not generally performed by thestarter culture industry. Most of our resources aredirected towards the direct servicing of our cus-tomers and assisting them with their applications.We depend on academic institutions such as theUW-Madison to conduct the type of basic studiesyou have proposed for our continued renewal.Without them, it would be difficult for our industryto advance.”


To, B.C.S. 1996. Properties of Brevibacterium linensand five different lactic acid bacteria attenuated byspray drying, freeze drying, or freezing. M.S. thesis,Univ. of Wisconsin, Madison.

To, B.C.S. and Etzel, M.R. 1996. Spray drying, freezedrying, or freezing of five different lactic acidbacteria. (submitted)

To, B.C.S. and Etzel, M.R. 1996. Survival ofBrevibacterium linens (ATCC 9174) after spraydrying, freeze drying, or freezing. (submitted)

To, B.C.S. and Etzel, M.R. 1996. Survival ofBrevibacterium linens ATCC 9174 after spraydrying, freeze drying, or freezing. Am. Inst. Chem.Eng. Conf. Food Eng., Nov. 2-3, 1995, poster 8.4.


72

INTERIM REPORT

Evaluating Microstructure of Reduced Fat Cheese byComputer Image Processing

Personnel: S. Gunasekaran, associate professor, Ag.Engineeering, N.F. Olson, professor, Dept. of FoodScience, M. Johnson, senior scientist, CDR, K.Muthukumarappan, associate researcher, Ag.Engineering, and S. Y. Kim, research assistant, Ag.Engineering


Funding: Dairy Management Inc. GK394

Objectives

1. Develop computer imaging system for acquiringdigital images of the microstructure of reduced fatcheeses.

2. Develop a computer algorithm and software toprocess the digitized images and extract variousdistinguishing features of fat globules such as size,shape, distribution, and surface characteristics.

3. Study the effect of composition, manufacturingprocess parameters, and age on the structuredevelopment in reduced fat cheeses.

4. Establish objective relationships among themicrostructural features and properties and qualityattributes of the cheese.

Summary

A number of manufacturing process parametersaffect the structure development in cheese. Theseinclude fat and moisture content of cheese, curd pHat draining of whey, curd handling, and proteolysis.Understanding how these variables affect function-ality and structure development in cheeses wouldallow better process control and use of a widervariety of raw materials. Objective evaluation ofthree-dimensional (3D) characteristics is impor-tant when studying the microstructure of materialswith non-uniform distribution of internal elements.Digital image analysis in conjunction with confocal

laser scanning microscope (CLSM) allows objectivecharacterization of 3-D microstructural features ofcheese.

Cheddar cheeses were manufactured at the UWdairy plant with different levels of fat (full-fat, 33%reduced & 50% reduced), moisture (53, 55 & 57%),curd pH at draining of whey (6.05, 6.2 & 6.35), curdhandling method (stirred- and milled-curd),chymosin level (0.33, 1.00 & 1.66) and trypsin level(normal, 3 times normal & 6 times normal).

Using the software we developed, the sequentiallayers of 2-D micrographs obtained from CLSMwere used to reconstruct its 3-D network. The 3-Dsize and shape characteristics of the fat globules incheese were evaluated. The data inTable 1 showaverage diameter (spherical), z-diameter (S Ni Di

3 /S Ni Di

2), number of fat globules, and sphericity ofthe fat globules. The fat globules were classified asboundary touching (BT) and boundary non-touching (BNT). The BT globules are those thatwere not totally contained within the sampleexamined in the CLSM. For analysis purposes, onlythose BNT globules of at least 1 µm3 in volume wereconsidered. They represented more than 99% of thefat. From 2-D micrographs of varying FDM ofCheddar cheeses, it appears that the average size ofthe fat globules decreased with decreased FDM. Incontrast to this appearance, the average sphericaldiameter and z-diameter of BNT fat globulesincreased with decreased FDM. Some very large fatglobules present in the low-fat cheeses, as evi-denced by high standard deviations, may explainthis. On the other hand, as cheese fat contentdecreased, the average spherical diameter and z-diameter of BT globules also decreased.

The size distribution of the BT globules variedhighly. Most of the BNT fat globules were smallerin size (< 4 µm) based on the total number of fatglobules present in the sample. However, based onthe total globular volume, most of the BNT fatglobules were greater than 4 µm. Number of BT

Cheese Research

73

globules in all size categories were higher than thenumber of BNT fat globules except for the sizecategory less than 2 µm. Based on the total globularvolume, the BT globules represented most of the fat(at least 90%). As cheese fat content decreased, thenumber of BT and BNT fat globules dispersedwithin the protein matrix increased. Similarcorrelations were observed during all stages ofmaturation and they were statistically significant.The BNT fat globules were more spherical than theBT fat globules. Average sphericity of BT and BNTfat globules of low-fat cheeses (26.4 and 46.9%FDM) was higher than that of full fat cheese (58.3%FDM).

The 2-D microstructure of 58.3% FDM Cheddarcheese samples during the ripening period areshown in Figure 1. It appears that there was anincreasing trend of confluence of smaller fatglobules into larger fat globules. Similar observa-tions have been previously reported in the litera-ture. The average fat globule size (spherical- and z-diameter) of three different fat Cheddar cheesesduring maturation is presented in Table 2. Both theaverage spherical- and z-diameter of all threecheeses increased during the first 8 wks of agingand then decreased during subsequent maturation.The differences were statistically significantbetween 0 and 8 wk and 8 and 24 wk age cheeses.

The fat globule size distribution of both stirred(58.3% FDM) and milled curd (55.9% FDM)cheeses at 0 wk age based on total globule numberis presented in Figure 2. All the fat globules in thestirred curd cheese sample were distributed within0-8 µm. However, the fat globules in the milled curdcheese sample were distributed between 0-12 µmrange and some of the globules were of the sizegreater than 12 µm. The fat globule size distributionchanges as the cheeses aged. At 0 wk, the stirredcurd cheese sample had a greater number ofsmaller (< 2 µm) fat globules than the milled curdcheese sample. However, at 24 wk the stirred curdcheese sample had a greater number of larger (> 2µm) fat globules than the milled curd cheese.


The digital image processing method we havedeveloped is a powerful tool for objectively evaluat-ing the microstructure of cheeses. In conjunctionwith the confocal laser scanning microscopy(CLSM), this method provides an efficient meansfor investigating structure (and thus texture) ofcheeses that is affected by a number of composi-tional and manufacturing process parameters. Thisinformation will provide useful feedback to cheesemanufacturers and may help them control thenecessary factors to manufacture cheeses ofprespecified textural characteristics.

Table 1. Fat globule size and shape of three different Cheddar cheeses at 1 day of age

% FDM Spherical diameter, µm z-diameter, µm Sphericity

BNT* BT BNT BT BNT BT BNT BT58.3 2.01± 0.60 9.24±16.13 2.36 49.05 14 15 0.534±0.07 0.499±0.1446.9 2.53±1.33 5.25±9.52 4.12 42.31 56 44 0.599±0.09 0.573±0.1126.4 2.77±1.42 5.67±6.97 4.41 25.66 104 79 0.581±0.06 0.544±0.12

* BNT-Boundary Non-touching; BT-Boundary Touching

No. of fatglobules


74

Table 2. Fat globule size and shape of three different Cheddar cheeses during maturation

Age, Average spherical diameter, µm Average z- diameter, µm

FF1 MF LF FF MF LF

0 2.13 ± 0.30 2.31 ±0.16 2.82 ±0.15 2.94 ±0.74 2.95 ±0.96 4.57 ±0.224 2.30 ± 0.46 2.39 ±0.26 2.89 ±0.30 3.03± 0.64 3.76 ±0.26 5.01 ±1.058 2.53±0.48 2.3 6± 0.31 3.19± 0.22 3.84± 0.80 4.00 ±0.87 5.53 ±0.4124 2.13± 0.07 1.79 ±1.27 2.78± 0.20 3.25± 0.75 2.63± 1.86 4.53 ±0.37

1 LF, MF & FF represent 26.4, 46.9 & 58.3 % FDM Cheddar cheeses, respectively

wk

Figure 1. CLSM Micrographs of 58.3% FDM Cheddar cheese at different aging.

Cheese Research

75

Figure 2. Distribution of fat globule size in stirred-curd, full-fat (SCFF) and milled curd full-fat(MCFF) cheese.


76

INTERIM REPORT

Machinability of Reduced Fat and Lowfat Cheeses

Personnel: S. Gunasekaran, associate professor, Ag.Engineeering, N. F. Olson, professor, Dept. of FoodScience and K. Muthukumarappan, associateresearcher, Ag. Engineering


Funding: Dairy Management Inc. GNO96

Objectives

1. Characterize rheological and fracture propertiesof cheese as a function of cheese type, composition,age, temperature (35º to 50ºF), and cutting speed.

2. Describe the reasons for poor machinability(cutting and shredding) based on the experimentaldata and physical chemistry of cheese.

3. Quantify effectiveness of the commercial shred-ding operation by characterizing the integrity anduniformity of the shreds and relating the shredquality information to cheese-machine factors.

Summary

With numerous new cheese-containing foodpreparations introduced in the market place, use ofmachined cheeses will continue to grow. Most ofthe cheese used as a food ingredient is in one of thefollowing machined forms – shredded, diced,grated, or sliced. Cheese manufactured in largeblocks is cut into smaller pieces for direct sales astable cheese or for other process operations such asshredding. In this project we are investigating anumber of variables and how they affect cheesefracture.

Cheddar cheese samples of different fat contents(8.0, 14.5 and 35.5%) were manufactured at theUW Dairy Plant. Mozzarella cheese samples of twofat contents (13.5 and 33%) were obtained from acommercial manufacturer. The cheese sampleswere stored at 4ºC until testing. All tests were

carried out at room temperature (20±2ºC). Thecheese samples were cut at different speeds (5 to 50cm/min) with an Instron model 1130 machine. Awire cutter was designed and fabricated usingstainless steel spring-tempered wire. Seven wires ofdifferent diameters (0.46 to 1.4 mm) were used forcutting cheese. Cheddar cheese samples were testedat 1, 3, 6, 9 and 12 wk after manufacture; and theMozzarella cheese samples were tested at 3 weeks.

Elastic-plastic fracture mechanics theory wasadopted to analyze the data. According to thistheory, the material is considered to flow only in alimited area around the crack tip. Therefore, thestored and flow energies are limited. During wire-cutting, only the material in the vicinity of the wireis disturbed. The force-time curves obtained duringwire-cutting had an initial linearly-increasing partcorresponding to the wire entry into the samplefollowed by a constant value. This constant forcewas recorded as the cutting force.

The cutting force vs. sample length plot (Fig. 1)shows that a linear increase in cutting force withthe sample length. The slope of this curve is termedas specific energy (SE). From the specific energy vs.wire diameter plot (Fig. 2), specific fracture energy(SFE) was calculated by determining the SE for zerowire diameter (by extrapolation). The SFE repre-sents the energy required only to fracture thecheese.

The SE increased with decreasing FDM (fat in thedry matter). For example, the SE of 57.6, 29.2 and16.7% FDM Cheddar cheese samples were 454.8,782.4 and 902.4 J/m2, respectively. Similarly, forMozzarella cheese, the SE of 33% and 13.5% FDMcheeses were 280.2 and 612.9 J/m2, respectively. SEincreased linearly with the speed of cutting for allwire diameters. At a given cutting speed, SFEcorrelated linearly with the wire diameter. We alsostudied the effect of age of cheese on SFE. Ingeneral, SFE of Cheddar cheeses was the lowest at 3weeks.

Cheese Research

77


The results of this project will help to elucidate thereasons for poor shreddability of lower fat cheeses.We will be able to help cheese makers and manu-facturers to selectively control such factors ascheese composition, age, temperature and cuttingspeed to obtain a product of expected quality. Theresults of quantitative analysis of cheese shred

characteristics will provide valuable feedbackregarding the effect of process parameters tomanufacture high quality shredded cheese. Endusers will be able to specify shred requirementsmore precisely in measurable parameters. Theability to meet their requirements will command abetter price and a higher market share for shreddedcheese.

Figure 2. Specific energy vs.Wire diameter for Mozzarellacheese (fat content: 33% FDM;age: 3 wk) at different cuttingspeeds.

Figure 1. Cutting force vs. Samplelength of one-week old Cheddarcheese of different fat levels (wirediameter: 1.3 mm; cutting speed:5.1 cm/min).


78

INTERIM REPORT

Structure Function Relationships During Melting andCooling of Lower Fat Cheeses

Personnel: S. Gunasekaran, associate professor, AgEngineering, N. F. Olson, professor, Food Science,D.J. Klingenberg, assistant professor, Engineering,and R. Subramanian, research associate, Ag Engi-neering



Objectives

1. Determine and evaluate the quantitative changesin the cheese microstructure using the confocallaser scanning microscopy and digital imageprocessing techniques.

2. Measure fundamental rheological parameters viatransient and dynamic viscoelastic experiments.

3. Determine melt and flow characteristics viaobjective rheological tests and empirical methods.

4. Develop hypotheses to explain changes infunctional properties in terms of microstructuraland rheological properties, compositional factorsand chemical changes in the cheese.

Summary

To better understand the relationship betweenstructural and rheological characteristics smallamplitude or linear deformation is imposed inorder to prevent any structural damage. Within thislinear range of deformation, viscoelastic rheologi-cal properties are independent of the magnitudeand rate of deformation or force. These measure-ments are normally made by applying a sinusoi-dally varying strain (or stress) at a certain fre-quency (ω) and measuring the resulting stress (orstrain) in the sample. Data obtained include twocomponents of complex shear modulus (G*): 1.Storage modulus (G’) which represents solid-like or

elastic character of a viscoelastic material, and it isa measure of energy stored and subsequentlyreleased. 2. Loss modulus (G”) which represents theliquid-like or viscous character of a viscoelasticmaterial, and is the energy dissipated per cycle ofdeformation. They are related as: |G*|2 = (G’)2 +(G”)2 . Also, G”/G’ is the loss tangent (tan d). Foran elastic (Hookean) solid δ = 0%; For a viscous(Newtonian) fluid, δ = 90%; and for a viscoelasticmaterial, 0 <δ< 90%.

A Bohlin VOR rheometer (with a 89.526 g-cmtorsion bar and 30 mm diameter parallel platesmeasuring system) was used to study the effect ofheating and changes in texture of cheese bymeasuring the dynamic mechanical spectra. Inorder to prevent slippage a coarse sand paper wasglued to the upper plate. Low-moisture, and lowfatpart-skim Mozzarella cheeses were made in theUW Dairy Plant. Cheese blocks were vacuumpackaged in plastic bags and ripened at 6-8ºC.Disk-shaped samples (mean thickness of 3.5 mmand diameter of 30 mm) were cut from refrigeratedcheese blocks after 1, 4 and 12 weeks of aging. Thetemperature of the lower plate of the measuringsystem was maintained by circulating water from awater bath. The sample was placed on the lowerplate and then the upper plate was brought incontact with the sample. The sample was held for 2-3 min to attain temperature equilibrium.

Strain sweep measurements were made at frequen-cies of 9.43 and 62.83 rad/s in order to obtain thelimits of linear viscoelasticity. The linear viscoelas-ticity region decreased with increasing temperature(Fig. 1) and age. The lowest linear range value ofabout 0.05% shear strain was obtained at a tem-perature of 70ºC after 12 wk of ripening. Therefore,for a wide temperature range study of linearviscoelastic properties, the strain should be limitedto 0.05%.

Frequency sweep measurements (frequency rangefrom 0.63 to 125.66 rad/s) were performed at a

Cheese Research

79

shear strain of 0.05% for different temperatures(10-70ºC) and age. Proteolysis during ripening ledto softening of the cheeses and thus decrease indynamic viscoelastic properties. Variation of G’ vs.ω, as a function of temperature, for one week-oldcheese is shown in Fig. 2. The storage modulus oflow-fat, part-skim Mozzarella cheese was greaterthan that of low-moisture, part-skim Mozzarellacheese at the same conditions. Master curves (at areference temperature of 40ºC) were obtained byshifting the temperature-dependent frequencydispersion of storage modulus after 1, 4 and 12weeks of aging (Fig. 3). We found no significantchange in the storage modulus after 4 weeks of

aging. The master curve significantly extends thefrequency range which is otherwise very difficult, ifnot impossible, to cover by direct methods ofmeasuring the moduli at a single temperature.


The results of this investigation will help us betterunderstand the observed loss of textural qualitiesin low-fat cheeses in terms of changes in structural,rheological, chemical, and sensory properties. Suchan understanding is crucial for systematic modifi-cation of cheeses to attain desired functionalities.

Figure 1. Storage modulus vs. Shear strain as a function of temperature (lowfat part-skimMozzarella cheese; age = 1 wk; frequency = 9.43 rad/s).


80

Figure 2. Frequency sweep test showing changes in G’ as a function of temperature (lowfat,part-skim Mozzarella cheese; age = 1 wk; strain = 0.05%).

Figure 3. Master curve constructed from several frequency sweep tests of data for differentaging (lowfat part-skim Mozzarella cheese; reference temperature = 40ºC).

Cheese Research

81

FINAL REPORT

Manufacture of a New Reduced Fat Cheese for Useon Pizza Pies

Personnel: Carol M. Chen, researcher, Mark E.Johnson, senior scientist, Amy L. Dikkeboom,research specialist, John J. Jaeggi, associateresearcher,William A. Tricomi, assistant researcher,CDR

Funding: Dairy Management Inc. CH 295


Objectives

1. To determine a non-pasta filata manufacturingprocedure for a 25% and 75% reduced-fat pizza cheese (correlates 42% and 19% FDB,respectively). These cheeses should have similarmelt and stretch characteristics as Mozzarellacheese.

2. To evaluate the physical and sensory characteris-tics of a 25% reduced-fat pizza cheese compared tolow moisture, part-skim Mozzarella and a 75%reduced-fat pizza cheese to 50% reduced-fatMozzarella cheese.

Summary

In the first year of the project two different manu-facturing approaches were taken to produce areduced-fat pizza cheese. A manufacturing protocolsimilar to that of lower fat Mozzarella (2) or 50%reduced-fat Cheddar cheese (1) was completed.These protocols included removing a portion ofwhey after cutting and adding water back. In bothcases the moisture content of the cheese was toolow (40-43% for 25% RF Pizza cheese, 46-51% for75% RF Pizza cheese.) Overall, these cheeses weremore meltable and less stretchable than desired.

In the second year of the study, we focused on twoareas: increasing cheese moisture and limitingproteolysis during aging. To achieve a relativelyhigh moisture and reduced-fat content, we beganwith part-skim milk, then slowed the rate of acid

production and shortened the total manufacturingtime (less than 2 h from addition of milk coagulantto hooping). A wash treatment was incorporated toinsure a final cheese pH of 5.25-5.35. Cheesecomposition and pH results can be found in Table 1.Mozzarella’s unique characteristics of both meltand stretch are related to pH and the heat treatmentit receives when going through the mixer. The heattreatment inactivates residual milk coagulant,reduces starter populations and decreases thepotential for casein hydrolysis in the cheese duringrefrigerated storage. (Good stretch is correlatedwith intact casein.) The pizza cheese manufactur-ing protocol does not use a mixer. Less milkcoagulant was used to decrease the amount theresidual milk coagulant activity.

The final product

Laboratory tests demonstrated that pizza cheesehas similar physical characteristics to Mozzarella(see Table 2). Cheese meltability (flow) was as-sessed by thermal and microwave melt assays. Inbaking applications, the pizza cheese did not flowoff of the crust, whereas Mozzarella cheese did. Thepizza cheese maintained a 10" and 5" stretchthrough 3 months of aging, for the 25% and 75%reduced-fat cheeses, respectively. Shredding thepizza cheese posed no problems. A consumer tastepanel conducted by the UW Sensory AnalysisLaboratory found no significant differences in thepreference of the 25% reduced-fat pizza cheese andlow moisture, part-skim Mozzarella. Of the 152unscreened panelist, 127 scored the pizza cheese ina “like” category as compared to 116 for the lowmoisture, part-skim Mozzarella.

There are differences between our pizza cheese andMozzarella that may be considered advantageous.They include:• No browning. Due to the starter culture used andaltered manufacturing protocol Wisconsin CDR’spizza cheese has no residual sugar and will notbrown during baking.


82

• 50% less oiling off. During the mixing process inMozzarella manufacture, the curd is heated,stretched and molded under hot water. The liquidfat and water pool around the protein strands. InWisconsin CDR’s pizza cheese process, hightemperatures are not used and thus the fat globulesremain smaller and more dispersed within thecheese matrix. Pools of oil are less likely to formfrom the cheese during the heating and bakingprocess of the pizza.• 50% fewer blisters. With a more homogenousproduct, fewer pockets of water develop which canturn into steam upon cooking. These pockets ofsteam cause the cheese to bubble, forming blisters.• White, more opaque color (both unmelted andmelted). Smaller and more numerous fat globulesreflect light more effectively. Due to the whiterappearance, tasters commented it looked like therewas more cheese on the pizza. Many commercialpizza makers request an extremely white cheese ontheir pizza pies.

Table 1. Cheese composition1.

Moisture Protein Fat Salt pH

25% Reduced-Fat Pizza cheese 47.04 27.33 22.30 1.63 5.18Low moisture, part-skim Mozzarella 46.46 27.54 21.70 1.53 5.2375% Reduced-Fat Pizza cheese 54.06 33.16 8.45 1.65 5.21Lower Fat Mozzarella 54.09 33.88 8.30 1.61 5.21

1 Results are the means of 3 trials.


The results of this study benefit cheesemakers inseveral ways. Because no mixer molder or brinesystem is needed, manufacturers of stirred curdcheeses (i.e. Cheddar, Colby, Muenster, or Brick) canproduce a cheese appropriate for pizza. The result-ing cheese is more homogeneous (brining canresult in salt gradients within the cheese whichinfluences cheese properties.) Bypassing the mixerstep results in less fat loss. In our trials, fat recoveryincreases from about 86 to 91%, giving highercheese yields. See Table 3.

(%) (%) (%) (%) at 1 month

Table 2. Summary of Physical characteristics at 30 days1.

Thermal melt Microwave Stretchat 8 min melt(mm) (mm) (cm)

25% Reduced-Fat Pizza cheese 53 84 31Low moisture, part-skim Mozzarella 55 81 5175% Reduced-Fat Pizza cheese 35 71 20Lower fat Mozzarella 35 70 18

1 Results are the means of 3 trials.

Cheese Research

83

References

Chen, C. M., J. J. Jaeggi, M. E. Johnson. 1994. Effect ofcoagulum firmness at cutting on quality and yieldof 50% reduced-fat Cheddar cheese. J. Dairy Sci.77:Suppl 1.

Johnson, M. E. March 1995. Update: CDR Mozza-rella research. Wisconsin Cheese Industry Confer-ence. Green Bay, WI.

Presentations/Publications

Chen, C. M. and A. L. Dikkeboom. March 1996. Anew cheese for Pizza pies. Wisconsin Center forDairy Research Open House, Madison, WI.

Chen, C. M., J. J. Jaeggi, M. E. Johnson. 1996. Com-parative study of pizza cheese made by stretchedand non-stretched methods. J. Dairy Sci. 79:Suppl 1.

Chen, C. M. June 1996. Pizza cheese made without aMixer. DMI featured technology at Annual IFTMeeting, New Orleans, LA.

Table 3. Estimated fat recovery and cheese yield for 25% reduced-fat pizza cheese and lowmoisture, part-skim Mozzarella.

% Fat Recovery % Cheese yield Increase inlbs cheese per50,000 lb milk

Mozzarella 86.5 9.26 NAPizza Cheese 90.9 9.47 109 lb


84

INTERIM REPORT

Minimizing the Watering Off of Unripened HighMoisture Lower Fat and No Fat Mozzarella Cheese

Personnel: Carol Chen, researcher, Mark E. Johnson,senior scientist, Amy L. Dikkeboom, researchspecialist, Kristen B. Houck, research specialist,John J. Jaeggi, associate researcher, William A.Tricomi, assistant researcher, CDR

Funding: Dairy Management Inc. CJH96

Dates: January 1996 – June 1997

Objectives

To evaluate the effect of different manufacturingprotocols on minimizing the watering off in highmoisture lower and no fat Mozzarella cheese.Specifically, manipulations will focus on water-to-protein interactions to maximize the water absorp-tion capability of the cheese proteins.

Summary

A newly manufactured Mozzarella melts to a tough,granular consistency, has low free oil and readilylosses water. About 2 weeks after manufacture, thecheeses are more homogeneous with improvedwater holding capacity. The very high moisturelevels of lower and no fat Mozzarella cheeses resultin a spongy-watery texture of the unripened cheese.Thus, while the higher moisture contents improvethe body and meltability of lower and no fatMozzarella cheese, it magnifies the physical defectof watering off. This can lead to problems becausethese high moisture cheeses are often shredded andfrozen for use as an ingredient on pizza pies orother prepared entrees within 1 to 4 days aftermanufacturing. Upon melting, these cheeses exhibitsevere watering off.

Our cheese making experiments to date haveexplored two aspects of protein-water interactionsand the effect on the water holding capacity oflower and skim milk Mozzarella cheese. We’ve

added sodium chloride to the whey during cookingand we have used higher than normal pasteuriza-tion temperatures. We observed an increase incheese moisture of 1.0% (the addition of salt to thewhey) and 3.5% (higher than normal pasteuriza-tion temperatures.) However, this moisture is not‘bound’ and can be expressed from the cheeseduring the first 14 days post-manufacture. Asecond observation is that, initially, more serumcan be expressed from lower fat Mozzarella cheeses,but these cheeses are able to ‘absorb’ moisture morequickly than the skim milk Mozzarella.


Consumption of Mozzarella cheese has increased ata constant rate of 8 to 12% for the past 12 years,with approximately 70% of the Mozzarella manu-factured produced for pizza pies. Unripened, highmoisture, lower, or no fat Mozzarella cheese whichseparates during melting is physically unsuitablefor pizza pies. Providing solutions that improve thefunctional properties of this economically impor-tant cheese will help ensure use of this cheese as afood ingredient.

Cheese Research

85

Table 1. Levels of expressible serum for experimental lower fat and skim milk Mozzarellacheeses.

Lower Fat Mozzarella Skim milk Mozzarella% Cheese Fat 9% 2%Age of the cheese (days) 1 4 7 1 4 7

% Expressible Serum (wt serum/wt cheese)

Pasteurization 164ºF/16 s - 7.0 .6 -0- 4.5 3.8 1.4no salt addedPasteurization 164ºF/16 s - 11.8 2.8 -0- 4.9 3.9 3.4salt addedPasteurization 184ºF/16 s - 11.3 1.5 -0- 9.9 8.0 5.4no salt addedPasteurization 184ºF/16 s - 11.7 1.6 -0- 10.3 6.4 6.4salt added


86

FINAL REPORT

Lower-fat Swiss cheese: Development of aManufacturing Protocol and the Evaluation of FlavorDevelopment

Personnel: Carol M. Chen, researcher, Mark E.Johnson, senior scientist, Amy L. Dikkeboom,research specialist, John J. Jaeggi, associateresearcher,William A. Tricomi, assistant researcher,CDR

Funding: Dairy Management Inc. JC296


Objectives

To develop a manufacturing protocol to producelower fat Swiss cheese (25% and 50% reduced-fat),that mimics the body and eye development of fullfat Swiss cheese. Once we produce an acceptablecheese, in terms of body and eye development,another project will be conducted to study methodsof flavor enhancement in reduced-fat Swiss.

Summary

An acceptable 25% reduced-fat Swiss cheese wasmade that was comparable to full-fat Swiss cheese(Table 1). The manufacturing protocol employed amesophilic starter culture (in addition to typicallevels of propionic acid bacteria), low pH at coagu-lant addition, firm milk coagulum at cutting, lowcooking temperatures, whey dilution, and minimalstirout. The same manufacturing protocol was usedto make a 50% reduced-fat Swiss (Table 2). How-ever, the cheese was too firm. We will be evaluatingall cheeses as they age, so it is possible that thebody of the 50% reduced-fat Swiss cheese willimprove.

One of the variables tested in this project wasadding salt (sodium chloride) to the whey duringcooking. Adding salt increased the final moisturecontent of the cheese by about 1% (Tables 1 & 2).These experimental trials yielded some interestingresults. We varied the level of salt addition from0.5% to 1.5%, but the increase in cheese moisture

over the control (no salt added) remained relativelyconstant. In addition, the salt addition with itscorresponding increased moisture did not appearto have an effect on the body and texture of thecheeses.

The eye development (number and size) in bothreduced-fat cheeses was comparable to full-fatSwiss i.e. round, shiny, and smooth. As the percent-age of cheese fat decreased, the length of time inthe warm room to develop eyes decreased from 33to 25 to 20 days, respectively for full fat, 25% and50% reduced-fat (Table 3). However, the datagenerated from these experiments did not show alinear relationship between the length of time inthe warm room and either pH or percentage oflactic acid of the cheeses going into the warm room.

The flavor of the 25% reduced-fat Swiss alsomimicked that of the full-fat Swiss cheese while the50% reduced-fat cheese lacked Swiss cheese flavor.


Developing a cheese making protocol for themanufacture of lower fat Swiss cheese will beessential to assist the cheese industry in manufac-turing a quality lower fat Swiss cheese. Over thepast several years, individual cheese manufacturershave contacted researchers at the Center for DairyResearch for expertise on lower fat Swiss cheese.This project allows researchers to effectivelycommunicate and meet the needs of the cheeseindustry. In addition, monitoring volatile acids andfree fatty acids in lower fat Swiss cheese will help usunderstand why lower fat cheese often lacks theflavor development of the full fat counterpart.

Cheese Research

87

Table 1. Composition, texture profile analysis and 2 month sensory analysis of 25% Reduced-fatSwiss cheese.

Target make schedule make schedule make scheduleA B C

% Moisture 42-44% 43.66 44.85 45.44% Fat 18-19% 18.2 19.8 18.4% Salt 0.8 - 1.0% 1.32 1.34 0.79% MNFS 52 - 54% 53.4 55.9 55.7pH (going into warm room) 5.2-5.4 5.31 5.18 5.31TPA Hardness 696 1,055 891 657(80% Compression - Ns) (full fat)TPA Cohesivenss 0.752 0.779 0.811 0.755

(full fat)Swiss flavor intensity 3.5 2.3 3.9Body 3.1 3.2 3.9Body Breakdown 2.8 3.1 4.1Body & texture preference 4.3 4.2 4.6

Make Schedules:A: pH at renneting = 6.63, no salt added during cookingB: pH at renneting = 6.63, 0.5 - 1.5% salt added during cookingC: pH at renneting = 6.15, no salt added during cooking

Sensory Scores:Swiss flavor intensity (0-7): 2 - very mild, 3- mild, mild to mediumBody (0-7): 2 - firm, 3 - slightly firmBody breakdown (0-7): 2 - curdy, 3 - slightly curdy, 4 - slightly firmBody and texture preference (0-7): 3 - dislike slightly, 4 - like slightly, 5 - like moderately


88

Table 2. Composition, texture profile analysis and 2 month sensory analysis of 50%Reduced-fat Swiss cheese.

Target make schedule make schedule make scheduleA B C

% Moisture 47-49% 45.83 46.75 47.11% Fat 11-12% 12.8 13.6 12.8% Salt 0.8 - 1.0% 1.32 1.36 0.75% MNFS 53 - 55% 52.6 54.1 54.1pH 5.2 - 5.4 5.42 5.34 5.55(going into warm room)TPA Hardness 696 1,400 1,793 985(80% Compression-Ns)(full fat)TPA Cohesivenss 0.752 0.795 0.819 0.786(full fat)Swiss flavor intensity 3.1 2.7 3.4Body 2.1 1.7 2.0Body Breakdown 1.9 1.4 2.6Body & texture preference 3.3 3.4 3.4

Make Schedules:A: pH at renneting = 6.63, no salt added during cookingB: pH at renneting = 6.63, 0.5 - 1.5% salt added during cookingC: pH at renneting = 6.15, no salt added during cooking

Sensory Scores:Swiss flavor intensity (0-7): 2 - very mild, 3- mild, 4 - mild to mediumBody (0-7): 2 - firm, 3 - slightly firmBody breakdown (0-7): 2 - curdy, 3 - slightly curdy, 4 - slightly firmBody and texture preference (0-7): 3 - dislike slightly, 4 - like slightly, 5 - like moderately

Table 3. Length of Time in the warm room. Type of Swiss cheese Length of time in pH going into

warm room (days) warm room % Lactic acid

Full Fat Swiss (n=10) 33 5.38 1.03%25% Reduced-fat Swiss (n=13) 25 5.31 1.40%50% Reduced-fat Swiss (n=13) 20 5.46 1.25%

Cheese Research

89

APPLICATIONS PROGRAM REPORT

CDR Specialty Cheese Applications Program

An English Cheese Seminar was held on September26-28 1996 in Madison. Primary instructors werefrom Reaseheath College in Nantwich, England andthe Center for Dairy Research in Madison. Topicscovered included “Overviews of the English DairyIndustry” and “English Cheeses.” As a result of thisseminar over five Wisconsin Cheese manufacturerswere contacted by an English cheese company toproduce English type cheeses in Wisconsin.

On Feb. 27-28, 1996 the second Wisconsin ProcessCheese Course was held as part of the WisconsinMaster Cheese Makers Curriculum. This course, theonly course of its type in the US, had a full enroll-ment of over 40 people.

On May 14-15, 1996 the first Wisconsin Dairy PlantWater & Waste Management Short course was heldat Babcock Hall in Madison. It was a joint effortbetween CDR and Agricultural Extension.

We are continuing our work on the computerlibrary of cheese names, descriptions and manufac-turing procedures. The computer program portionof the database is complete, but we are still enteringdescriptions and manufacturing procedures tomatch these names.

The Wisconsin Master Cheese Maker program willcontinue with plant visits and cheese samplingduring summer and fall of 1996. The first class ofWisconsin Master Cheese Makers should becertified in the spring of 1997. We are developingmore courses for the Wisconsin Master CheeseMaker program.


The specialty cheese program has helped topromote the image of Wisconsin Specialty CheeseInstitute (WSCI), improve the educational processfor industry personnel (Master Cheese Maker), andincrease practical cheese making knowledge(Cheese Artisan Seminar Series). The program hasalso worked directly with cheese makers to developspecialty types of cheese.

Personnel: James Path, outreach specialist, JohnJaeggi, assistant researcher, Center for DairyResearch



Objectives

1. Work on Wisconsin Master Cheese Makerprogram and Seminar Series

2. Laboratory and field research consisting ofdevelopment of traditional and new specialtycheeses

3. Transfer of CDR cheese research results toindustry

4. Liaison between CDR, WMMB, and Wisconsincheese manufacturer’s by assisting in solvingtechnical problems with the manufacture ofspecialty cheese.

Summary

Wisconsin Master Cheese Maker program

We have now completed the plant visits, oralexaminations, and cheese sample testing of 9 newpersons for the apprenticeship phase of the pro-gram. The board received 15 new applications forthe 1996 summer test period and approved 10 fortesting.

A Swiss Cheese Seminar was held on April 23-25,1996. Instructors from the BernisheMolkereischule in Rutti-Zolilofen, Switzerlandtaught at this seminar. It is one of two GermanLanguage schools in Switzerland which is able toqualify Swiss Master Cheese Makers. The manufac-ture of Swiss specialty cheeses and a discussion ofSwiss cheese defects was presented.


90

INTERIM REPORT

Improved Quality of Shredded Cheese -Antimycotics, Oxygen Scavengers and ModifiedAtmosphere Packaging

Personnel: Russell Bishop, associate professor,Center for Dairy Research, University of Wisconsin;Joseph E. Marcy, associate professor and Tina MolerGrove, graduate student, Dept. of Food Science &Technology, Virginia Polytechnic Institute & StateUniversity.

Date: July 1995 – June 1997


Objectives

The main objective is to determine the combinedeffectiveness of antimycotics, oxygen scavengersand modified atmosphere packaging in preventingthe growth of Penicillium roqueforti on shreddedCheddar and Mozzarella cheeses. This research willprovide information concerning the ability of theseapplications to extend the storage and shelf life ofshredded cheese, which in turn should decreasecheese losses due to mold growth.

In addition to the above objective, Virginia Tech willexamine the oxygen requirement for mold growthon shredded cheese packaged under MAP condi-tions.

Summary

Project objectives, as proposed by the cheeseindustry team, were developed into a research planby the Center for Dairy Research, University ofWisconsin and the Dept. of Food Science, VirginiaTech. The experimental design and testing schemewas created by the Dept. of Food Science andexamined by the VT Statistics Dept. to insurestatistically valid results.

The following pilot plant equipment and supplieshave been acquired, assembled and tested by theDept. of Food Science, Virginia Tech.

A cheese tumbler was designed using a 30 galcylindrical polypropylene tank with a cover. A sixinch hole was cut into the cover to allow for entry ofthe spray apparatus. The tank resides horizontallyon rollers supported by steel rods. An electric motorwith a belt drive spins the rollers thus rotating thetank at a rate of 6.5 rotations per min.

Natamycin was applied to the cellulose and cheeseby spraying. A spray apparatus was constructedfrom 1/8" stainless steel piping. In-line ball valveswere installed for easy operation. The spray nozzleswere obtained from Spraying Systems Co.Charolette, N.C and stainless steel fogger tips wereprovided by Gist-brocades, Menomonee Falls, WI.Each nozzle had a 15 inch diameter spray pattern,thus two nozzles sufficiently covered the length ofthe 30" tank. The apparatus delivers a total of 6.25mls/sec with the aid of a light volume transferpump and 80 psi of air.

A rapid package integrity tester was designed andbuilt from 1/2" Lexan™ (polycarbonate) to insurethe integrity of the package seal, thus allowingleakers to be rejected from the study. A single 8 oz.package of cheese is placed on the bottom of a 7 x10 x 10 vacuum tight polycarbonate box. Theunderside of the lid supports a depth gauge whichis mounted perpendicular to the lid surface.Vacuum is applied causing the bag to expand. Thedepth gauge caliper is pushed up by the bag and theneedle registers the expansion in the package. Theneedle on the depth gauge should remain station-ary, indicating a good seal. However, if the gaugeshows that the package is losing volume, due to thepackage deflating, then the package is regarded as aleaker.

Natamycin was provided by Gist-Brocades,Menomonee Falls, WI and Pfizer Inc. Milwaukee,WI. The objective of this study is to determine theeffectiveness of natamycin on preventing mold. Tengrams of Delvocid (Gist-brocades) and 10 grams of

Cheese Research

91

Natamax (Pfizer Inc.) were throughly mixedtogether in order to eliminated any differencesbetween antimycotic supplier. Both Delvocid andNatamax contain 50% natamycin (pimaricin) and50% lactose. The 1250 ppm natamycin solution wasmade by placing 2.5 g natamycin in a 1 L graduatedcylinder and filling to volume with ddH2O. The 300ppm solution was made by bringing 2.4 gnatamycin up to 4 L with ddH2O.

The Code of Federal Regulations (CFR) places arestriction on the concentration of natamycin thatmay be applied to cheese. Code21CFR172.155(c)(1), referring to Natamycin(pimaricin), states the additive may be applied bydipping or spraying, using an aqueous solutioncontaining 200 to 300 parts per million of additive.Due to this restriction, a 300 ppm natamycinsolution was applied to the cheese in this experi-ment.

Code 21CFR133.146(a) states that moisture may beremoved from the cheese ingredients in the manu-facture of the finished food, but no moisture isadded. The code also states that antimycotics andanticaking agent are considered optional ingredi-ents which renders the moisture limits of this codenon-applicable to their use.

Natamycin concentration on the cheese is beingdetermined using Gist-brocades modified versionof the International Dairy Federation’s protocol forthe Determination of Natamycin Content, Methodof Spectrometry. 140:1987. The procedure was usedto confirm the concentration of the 300 and 1250ppm stock solutions as well as the 4-6 ppmnatamycin concentration on the cheese.

Qual Flo™ (powdered cellulose) and Flow Am® 200(powdered cellulose, dextrose and enzymes) weresupplied by Qualcepts, Minneapolis, MN. Thecelluloses were preweighed in 1250 g amounts forapplication to identical treatments groups in theCheddar and Mozzarella cheese experiments.Cellulose was applied at 2% the product weight.

Two of the treatments required the natamycinsolution be sprayed onto the cellulose and allowedto dry, before being applied to the cheese. Since theCFR does not regulate the concentration of

natamycin that can be applied to cellulose, a 1250ppm solution was used. This solution was appliedto the cellulose at a rate equivalent to the 300 ppmsolution being applied to the cheese. Applyingnatamycin at this concentration decreased theoverall drying time. Natamycin treated cellulosewas covered with slotted foil and allowed to air dryin a laminar flow hood for two days. Cellulose wasstirred periodically to facilitate drying.

Penicillium roqueforti was selected as the inoculumfor this study because of its resistance to manytreatments. Enumeration of the Midwest® bluemold powder stock, supplied by Systems Bio-Industries Inc., Waukesha, WI, resulted in a concen-tration of 2.5 x 108 spores/g. To insure the bestpossible distribution of P. roqueforti, the inoculumwas mixed into preweighed cellulose with the aid ofa planetary type pan mixer. The mixing bowl andtool, both of stainless steel construction, weresanitized with a 200 ppm chlorine solution. Al-though this type of mixer visits all parts of thestationary pan; mixing was stopped every fiveminutes to scrape the edges of the bowl into thecenter. Total mixing period was twenty minutes.The initial “spore stock” was diluted by placing 2 g2.5 x 108 spores/g into 998 g of Qual-Flo™ andmixing as described above. The final concentrationof the “P. roqueforti stock cellulose” was 50 x 104

spores/g cellulose. Penicillium roqueforti stockcellulose (3.1g) was used to inoculate 1250 g ofpreweighed cellulose to obtain a concentration of1240 spores/g cellulose. Enumeration of inoculatedcellulose indicated a concentration of 3000spores/g.

The Cheddar and Mozzarella cheeses, provided byAlto Dairy, were processed in 9080 g (20 lb) and4540 g (10 lb) lots, respectively. Cellulose containing1240-3000 spores/g, prepackaged in 182 g and 91 glots, was applied to 9080 g and 4540 g cheese,respectively. The final inoculum level was 10-100spores/g cheese.

Cheeses were packaged in BDF Bags provide byCryovac, Duncan, SC. The bags were made from amulti-layered, coextruded polyolefin formulationwith an oxygen transmission rate of 4.0 cc/m2/24hrs. A gas mixer was used to deliver the proper ratioof nitrogen:carbon dioxide (75:25). Samples of this


92

mixture were analyzed on a gas partitionerthroughout the process to insure the gas wouldproperly modify the package atmosphere.

It is difficult to completely evacuate all the air frominside the bag resulting in low levels of residualoxygen in the package headspace. Headspace gasanalysis was performed on 2 packages per treat-ment immediately after packaging to determineinitial headspace gas composition. Gas drawn froman individual package was analyzed using a gaspartitioner. Changes in gas composition will bemonitored throughout the storage period(0,60,120,180 d) on 2 packages treatment.

Enumeration of yeast and mold will be performedby placing an 11g sample of cheese and a 99 mlphosphate/magnesium chloride dilution blank(Marshal, 1992) into a Seward 400 Stomacher bag(Scientific Products S8254-1), stomaching for 2min, plating 1 ml of the dilution onto a 3 M Yeast &Mold Petrifilm (Beuchat, 1991), incubating at 25º Cand counting at both 3 and 5 days.

A spreadsheet format has been established tohandle raw data. The data will be entered as thestudy progresses and analyzed for significanceusing SAS®, SAS Institute Inc., Cary, NC., with theaid of the Statistics Dept. at Virginia Tech.

The In and Out method, as described in SensoryEvaluation in Quality Control (Munoz,1992) isbeing used to determine if the cheese is in or out of“specs” for a saleable product. The panel consist of4 to 5 people with experience with dairy foods. Thepanel evaluates 16 samples per session including afreshly shredded sample of both cheddar andMozzarella cheeses, which serves as a standard.Each product is discussed to determine if it is “in”or “out” of specifications. There are no definedspecifications or guidelines and there is no trainingor product orientation. As a result, each panelistmakes a decision based solely on their individualexperience and familiarity with the product. Avisual score is also recorded according to thefollowing guidelines: 0 = no visible growth, 1 = 1-3spots, 2 = slight surface growth, 3 = mediumsurface growth, and 4 = pronounced surfacegrowth.

Yeast and mold air samples were taken in theprocessing facility and the cooler to determine airquality at the time of packaging and initial storage.Counts in the processing lab averaged 146 CFU/m3 ,while the 50ºF cooler had no visible growth.

Processing equipment and cheese contact surfaceswere sanitized prior to use with 200 ppm XY-12liquid sanitizer provided by Klenzade, Division ofEcoLab Inc., St. Paul, Minnesota.

Eleven hundred twenty pounds (560 lbs each) of 10day old Mozzarella and 30 day old Cheddar cheesewas provided by Alto Dairy, Waupun, WI. Thecheeses were processed with the aid of the under-graduates, graduates, faculty and staff of theVirginia Tech’s Food Science Dept.

Significance to dairy industry

Currently, the cheese industry absorbs the financialloss when mold growth on shredded cheese rendersthe product unsaleable. Consumers may be con-fronted with mold within days of opening the pre-shredded cheese. Limited research is available onthe effectiveness of antimycotics and oxygenscavengers in preventing mold growth on shreddedcheeses. The purpose of this study is to increase thestorage and shelf- life of shredded Mozzarella andCheddar cheese by preventing mold growth.

References

Beuchat, L.R., Nail, B.V., Brackett, R. E., Fox, T.L.(1991) Comparison of Petri films TM Yeast & MoldCulture Film Method to Conventional Methods forenumerating Yeast & Molds in Food. Journal ofFood Protection. Vol 54, N#6. Pg.443

Marshal R.T. (1992) Standard Methods for theExamination of Dairy Products. 16th edition.American Public Health Association, Washington,DC 20005. pg. 87-88

Munoz, A. M., G. V. Civelle, and B. T. Carr. (1992)Sensory Evaluation in Quality Control.VanNostrand Reinhold, NY, New York. 10003. Pg141-167.

Cheese Research

93

INTERIM REPORT

1996 Wisconsin Cheese Plant Management Survey

Personnel: Brian W. Gould, senior scientist, Kurt A.Carlson, visiting scientist, Center for Dairy Re-search and Department of Agricultural and AppliedEconomics

Dates: February 1996 – October 1996


Objectives

The primary objective of this study is to profile theorganization, management, and marketing charac-teristics of a representative sub-sample of small/medium sized cheese manufacturers. This infor-mation will be obtained by an in-plant survey of60-70 Wisconsin plant managers. There are threegeneral sections to the survey: market and market-ing characteristics, firm structure and organization,and management goals and objectives. Thefollowing questions will be addressed in thisresearch:1. What competitive strategies do cheese manufac-turers adopt? Why?

2. How do these strategies impact company perfor-mance?

3. Is there consistency in the strategies chosen withrespect to production, sales and marketing channelrelationships?

Summary

We have completed the development of the surveyinstrument. Potential survey participants havebeen identified. By July 1, forty plant managers hadagreed to participate in the survey. Another 30managers will be contacted. By mid-July, in plantvisits will be started. It is anticipated that by theend of August, the collection of surveys will becompleted. A first draft of a report to WMMB willbe generated by the middle of September and finalreport generated by October 15.


This study is timely and relevant for Wisconsin’scheese manufacturers. Currently, WMMB is facedwith making decisions about how to package theservices it offers. Ideally, WMMB would offer aunique package of services to enhance the eco-nomic position of each of Wisconsin’s cheese plants.However, due to resource constraints, this is notpossible. As an alternative, WMMB would like tooffer packages of services to plants with similarneeds. In order to do so, WMMB will requiredetailed information about Wisconsin cheese plantsbusiness strategies.

While the 1995 Wisconsin Small Cheese PlantSurvey identified the existence of different produc-tion strategies among Wisconsin’s small cheeseplants, it did not collect information on organiza-tional structure, management goals and objectives,marketing activities and distribution/sales systems.Thus, WMMB needs information to segmentWisconsin’s cheese manufacturers based uponneeds. Successful completion of this project willprovide WMMB with the information it requires totarget the services it offers to meet the needs ofWisconsin’s cheese plants.


94

INTERIM REPORT

John Norback, professor, Candelaria Barcenas,research assistant, Dept. of Food Science,R. L.Bishop, professor, Center for Dairy Research

Funding: Wisconsin Milk Marketing Board UWA9506

Dates: February 1996 – June 1996

Objectives

To build a software which uses the flow of materialsdiagram to enable convenient and friendly accessto technical information about Cheddar cheese. Itwill also provide the user with computationalinformation about intermediate and finishedproduct costs, and the coefficients constraints foroptimization models. The system will allow users toidentify and justify cheese research needs when noinformation exist about connecting a problemobserved or a desired quality with the manufactur-ing process.

The first part of this project has been focused todefine the type of information the software willoffer, and the development of organizing theinformation and when it will be displayed to meetuser needs.

Summary

The initial work of this project has been to defineinformation delivery to the user. The users fromindustry will obtain all types of informationavailable to improve quality of the finished product.Research Centers will obtain possible researchsubjects to provide the industry with missingtechnical information to satisfy consumers ofCheddar cheese.

Sensory methods are used to evaluate quality offinished product. The software will provide theinformation connecting each quality and function-ality parameter (melt, shred, bitterness, firmness)

of Cheddar cheese with the operations during thecheese making process. In other words, by clickingon the final product, a window with a menu of thequality parameters will appear. The user will clickon the desired parameters and the operationswhere product may be modify or affected will behighlighted in a different color. The user will clickon the operation and a window will open withchoices such as references (articles or book),process control parameters and decisions. Bychoosing one of these options information will bedisplayed, and printed if desired.

In the case that users need information about anintermediate product, they only have to click onthat point of the manufacturing process and amenu will be displayed. The first two optionsdisplayed at this point are controls (references withexpected value of parameters to be controlled atthat point such as pH, moisture, etc.), and problems(references about possible problems observed atthis point in the process with causes and alternativesolutions).

Another area included in this initial stage ofbuilding the software covers raw materials. A clickon an ingredient will offer information such ascomposition (suggested for the specific type ofcheese), microbiological conditions (desiredmicrobial counts), and tests (type and desiredresult). All this information will be gathered toinform the user about the quality of the rawmaterial that will allow him to obtain a “qualitycheese” when processed under the optimal condi-tions.

The options menu may vary through the develop-ment of the software.

Next Steps

The coding of the software and building the systemwill start on July 1996. Simultaneously moreinformation will be gathered and organized and

Developing a Graphical Paradigm for Organizing andDelivering Technical Information about Cheese

Cheese Research

95

then incorporated into the software. At the end ofFall semester 1996 there will be a test prototypesoftware to be used by researchers and industries toobtain feedback about the utility and ease of use ofthe software, and what else (type of information,menu items, organization, etc.) should be consid-ered to improve the final outcome of this system.

The literature review does not offer enough infor-mation about the connection between some of thequality parameters and the manufacturing process.In this cases, expert information is essential.


Communicating research results to people whomight use it has long been problematic. Numbers,charts and graphs developed in the laboratory arealways more valuable when linked to some back-ground information. Our approach – the flow ofmaterials diagram – provides some context toresearch information. Cheese manufacturers cansee how the research applies to practical problemsand why the research is important.


96

Safety and Quality

97

Chapter 4

Summary

Safety and QualityRecently, the reported health benefits of conjugated linoleic acids (CLA) have received muchpublicity. CLA may play a role in cancer prevention, and Pariza, et al. have discovered that CLAreduces atherosclerosis in rabbits. Researchers hope this effect will carry over to humans.

Obtaining safe good quality milk is an important starting point for all dairy plants. To achievethis goal, farm milking equipment must be cleaned effectively. Reinemann et al. have investi-gated the dynamics of air injected, clean-in-place (CIP) systems used on farm milkingequipment. (These cleaning systems are quite different from the “flooded” CIP systemstypically used in dairy plants) This analysis was important because cleaning procedures forair-injected CIP systems were developed primarily by trial and error. Several milking machinecompanies have already implemented the recommendations for improved CIP proceduresdeveloped from this research. As demonstrated in the Bishop et al. project, good manufactur-ing practices (GMPs) and a HACCP (hazard analysis critical control point) program can playan important part in reducing dairy plant microflora, which then produces a safer product.

Luchansky, et al. demonstrated that biopreservatives can appreciably reduce the numbers of L.monocytogenes in Hispanic style cheeses. This may be an important hurdle for potentiallyhigh risk cheeses.Although many cheese makers might view reduced fat cheeses as a betterenvironment for pathogen growth, Luchansky and Johnson report data that show just theopposite. They compared the survival of several pathogens in a reduced fat Cheddar madewith a carbohydrate based fat substitute to standard Cheddar cheese and found the reducedfat product less hospitable to the organisms tested.

Technology Transfer

Producing Safe Dairy Foods, August 1995CDR Open House, March 1996


98

Safety and Quality

99

FINAL REPORT

Biological Significance of Conjugated DienoicDerivatives of Linoleic Acid

Personnel: M.W. Pariza, professor, Wei Liu, researchspecialist, Jayne Storkson, senior research special-ist, Karen Albright, senior research specialist, KisunLee, graduate student, Xiaoyun Yang, graduatestudent, Food Research Institute

Dates: June 1990 – November 1995

Funding: Wisconsin Milk Marketing Board 8927

Objectives

1. Determine the function of CLA in protecting cellmembranes from oxidation

2. Determine the effects of CLA in regulating/modulating various key membrane enzymes andenzyme systems

3. Determine practical ways to synthesize CLA forcommercial application

Summary

1. Determine practical ways to synthesize CLA forcommercial application

Generated a data base on CLA content of over 100foods, to provide a “starting point” for possiblesupplementation of foods and feeds with CLA

Isolated and characterized a Lactobacillus from ratcolon that produces CLA from linoleic acid

In collaborative study with the U.W. Dairy ForageResearch Laboratory, discovered ways of feedingcows to substantially increase CLA content of milk.

2. Determine the effects of CLA in regulating/modulating various key cell membrane enzymesand enzyme systems

Discovered that CLA reduces atherosclerosis inrabbits by influencing lipid metabolism and

apparently also the expression of key cell mem-brane proteins;

Obtained evidence that CLA affects in positive waysprotein kinase C activity and arachidonic acidmetabolism

3. Determine the function of CLA as an antioxidant

Obtained evidence that CLA effectively chelatesiron, which possibly explains this activity

4. Investigate the inhibition of food spoilagemicroorganisms by CLA

Found that the effectiveness of CLA in a foodproduct is inversely proportional to lipid content.

Discussion

We began this project by generating a data base onCLA levels in over 100 foods. We established thatdairy products and meat from ruminant animalsare the primary dietary sources of CLA for bothchildren and adults. The findings are presented inthe report of Chin et al., 1992. This research wasimportant in that it demonstrated the importanceof dairy products as dietary sources of CLA, andalso to provide a basis for supplementing foods andfeeds with appropriate levels of CLA.

When this project began the only known “natural”means of producing CLA in the laboratory was touse a crude enzyme preparation isolated from astrict anaerobic rumen bacterium, Butyrivibriofibrosolvens. However, we observed CLA accumu-lated in the tissues of rats fed high levels of freelinoleic acid, indicating that microorganisms in therat colon might contain the necessary isomerase toconvert linoleic acid to CLA. The findings werepublished in the report by Chin et al., 1994.Xiaoyun Yang, a graduate student, then isolated alactobacillus from rat colon that has this property,and is characterizing the organism for possible


100

commercial use. This organism is the central focusof her Ph.D. research which should be completedlater this year. Our discovery may directly benefitdairy farmers through commercial production ofCLA for cattle and calf feeds, and for direct additionto dairy products to increase value.

In an ongoing collaborative project with the UW’sDairy Forage Research Laboratory, we discoveredways of feeding cows to substantially increase CLAlevels. Specifically, the amount of grass in a cow’sdiet appears to be directly correlated with theamount of CLA in her milk. This work will soon bepublished. Using these new findings it is possible toincrease the CLA level in dairy products at least 6-fold, a finding of obvious importance to dairyfarmers.

We conducted a study in rabbits fed an atherogenicdiet with or without CLA, and discovered that CLAsignificantly reduced serum LDL cholesterol andtriglyceride levels, and the symptoms of atheroscle-rosis (Lee at al., 1994). This work has been repro-duced by others using hamsters. We also haveobtained strong indications that CLA modifies cellmembrane proteins involved in the initiation andprogression of atherosclerosis. We found that CLAaffects the regulation of protein kinase C andarachidonic acid metabolism in positive ways,biochemical activities that are related to cancerdevelopment. These results show that CLA actuallyprotects against heart disease; the finding of such asubstance in dairy products is indeed a newconcept that should be of considerable interest toWisconsin dairy farmers. The data also provide abiochemical basis for the inhibition of both heartdisease and cancer by CLA.

There is commercial interest in using CLA as a foodpreservative, both as an antioxidant and as a moldinhibitor. During the course of this project westudied both of these areas. We obtained evidencethat CLA chelates iron, which may specifically relateto the antioxidant activity that we and others havereported to occur under some circumstances. Thiswork is being prepared for publication. In additionwe studied the ability of CLA to inhibit the growthof spoilage microorganisms, especially molds andcertain lactobacilli. We established that the effect is

inversely proportional to fat content, i.e., CLA ismore effective as fat is decreased. These data mayexpedite FDA approval of CLA in that they providea functional (as opposed to health-enhancing)basis for adding CLA to foods and feeds.


Dairy products are a principal dietary source ofCLA. Hence, work aimed at possible health benefitsof CLA is expected to enhance that already fineimage of dairy products as important for soundhealth. The fact that CLA is a component of dairyfat is particularly intriguing in this regard. Addi-tionally, CLA is potentially of great importance as anatural antioxidant and mold inhibitor for use infood systems.

Publications

Chin, S. F., Liu, W., Storkson, J. M., Ha, Y. L., andPariza, M. W. Dietary sources of conjugated dienoicisomers of linoleic acid, a newly recognized class ofanticarcinogens. J. Food of Composition andAnalysis 5: 185-197, 1992.

Pariza, M. W. Diet and cancer: where do mattersstand? Commissioned by the Council on ScientificAffairs, American Medical Association, Chicago, IL.Arch. Intern. Med. 153:50-56, 1993.

Chin, S. F., Storkson, J. M., Liu, W., Albright, K. J., andPariza, M. W. Conjugated linoleic acid (9,11- and10,12-octadecadienoic acid) is produced in conven-tional but not germ-free rats fed linoleic acid. J.Nutrition. 124: 694-701, 1994.

Lee, K. N., Kritchevsky, D., and Pariza, M. W.Conjugated linoleic acid and atherosclerosis inrabbits. Atherosclerosis 108:19-25, 1994.

Abstracts

Chin, S. F., Liu, W., Albright, K., and Pariza, M. W.Tissue levels of cis-9, trans-11 conjugated dienoicisomer of linoleic acid (CLA) in rats fed linoleicacid (LA). FASEB J. 6, abstract #2665 (1992).

Safety and Quality

101

Benjamin, H., Storkson, J. M., Liu, W., and Pariza, M.W. The effect of conjugated dienoic derivatives oflinoleic acid (CLA) on mouse forestomach proteinkinase C (PKC)-like activity. FASEB J. 6, abstract#2666 (1992).

Bonorden, W., Storkson, J., Liu, W., Albright, K., andPariza, M. Fatty acid inhibition of 12-0-tetradecanoylphorbol 13-acetate (TPA)-inducedphospholipase C activity. FASEB J. 7 (4), abstract#3580 (1993).


102

FINAL REPORT

Cleanability Assessment of Milking Equipment

Personnel: Douglas J. Reinemann, associate profes-sor, Department of Biological Systems Engineering,associate professor, Amy C. Lee Wong, Food Re-search Institute, Anton Muljadi, graduate researchassistant, John Patoch, research specialist


Funding: Dairy Management Inc. RNM 94UW,Babson Brothers Company

Objectives

1. Characterize the interaction of mechanical,chemical and thermal processes on air injectedClean-In-Place (CIP) of milking systems.

2. Use information from (1) to develop designcriteria and recommendationsfor installation and operation of CIP systems toassure effective cleaning of milking systems.

Summary

Ineffective cleaning and sanitizing procedures ofmilking systems may cause milk-soil and microbialdeposition which eventually will decrease milkquality. The dynamics of air injected CIP systemsare not well understood. Many different materialsand several distinctly different flow regimes areencountered in milking, (CIP) systems. Materialsinclude: stainless steel, synthetic rubber hoses andgaskets, and various types ofplastic used in hoses and milking system compo-nents. Cleaning flow dynamics and the resultingmechanical cleaning action vary considerablydepending on the location in the system and theway in which the milking CIP system is installedand operated. Developments of CIP systems formilking machines have been by trial and errorbasis. The need for research to develop rationaldesign criteria is recognized by the industry.

We have performed extensive studies of the flowdynamics of milking CIP systems on full scale

systems at the University of Wisconsin MilkingResearch and Instruction Laboratory (UW-MRIL)and in the field. We also developed and testedseveral methods of assessing mechanical cleaningaction at the UW-MRIL. These studies were carriedout in collaboration with Dr. Albrecht Grasshoff ofthe Federal Center for Dairy Research inKiel,Germany. The roles of mechanical, thermaland chemical factors were characterized. In addi-tion, the effectiveness of the ATP bioluminescenceassay as a rapid detection method in assessing thesanitary conditions of milking systems was alsostudied.

Methods have been developed to apply a standard-ized milk soil and bacterial residue to test chips andassess the effectiveness of various cleaning regimeson their removal. Four types of milk-soil were usedto test the effectiveness of various permutationsinvolving different chemical agents, shear forcesand cleaning solution temperatures. The interac-tion of bacteria and milk-soil can enhance theattachment to surfaces. It was demonstrated thateach CIP cycle has a certain degree of effectivenessin either reducing the amount of milk-soil orinactivating bacterial cells. More effective cleaningwas achieved by applying a high shear level (114 N/m2) and a high temperature (60ºC).

The ability to identify and compare the efficaciesamong the cleaning regimes indicates that thestudy method is reliable to evaluate the perfor-mance of the CIP systems. The overall results ofATP bioluminescence and plate count methodsgave good correlation (r=0.904). However, somelimitations of the ATP assay were encountered inthis study.

The results of these studies have already beenimplemented by several milking machine compa-nies. Recommendations for CIP system design andcontrol developed as part of this project have beenincorporated into company recommendations andtraining programs. The UW milking lab has alsooffered a series of courses to train technical andnon-technical audiences in the application of these

Safety and Quality

103

methods and recommendations. Several companieshave also requested that new cleaning products beevaluated using the methods developed in thisproject. The methods developed to evaluate clean-ing chemicals under worst-case field conditionshave not been available previously. As a result ofthese activities, numerous, longstanding cleaningand sanitation problems in the field have beenresolved. Considerable energy cost savings for dairyoperators have also been achieved.


The dynamics of air injected CIP systems are notwell understood. Most developments have occurredas a result of trial and error, although industryrecognizes the need for research to develop arational design criteria. Improved methods ofcleaning and sanitizing milking equipment willlead to safer and better milk quality, maintainingconsumer confidence in dairy products.

Publications

Muljadi, A., D.J. Reinemann, and A.C.L. Wong, 1996.Air injected Clean-In-Place for Milking systems:Development of a Study Method and Characteriza-tion of Chemical, Mechanical and Thermal Factors.ASAE paper No. 963019. Written for Presentation atthe 1996 International Meeting Sponsored by theAmerican Society of Agricultural Engineers, July14-18, 1996, Phoenix, AZ USA

Reinemann, D.J., 1996, Testing Cleaning Perfor-mance of Milking systems witha Vacuum Recorder. University of Wisconsin -Extension, Bulletin number A3649

Reinemann, D.J., 1995. System Design and Perfor-mance Testing for CleaningMilking Systems. Proc. Designing a ModernMilking Center, Northeast Regional AgriculturalEngineering Service National Conference, Roches-ter New York, Nov 29 - Dec. 1, 1995.

Reinemann, D.J., and A. Grasshoff, 1994. Two phasecleaning flow dynamicsin air injected milklines. Transactions of the ASAE,Vol. 37, No. 5, pp 1531-1536.

Reinemann, D. J., A.C.L. Wong and E. Rabotski,1993. Interaction of chemical, thermal and physicalactions on the removal of bacteria from milkcontact surfaces. ASAE paper No. 933536, Presentedat the 1993 Winter meeting of the American Societyof Agricultural Engineers, Chicago, Illinois, USA.

Grasshoff, A. and D.J. Reinemann, 1993. ZurReinigung von Milchsammeleitungen mit Hilfeeiner 2-Phasen Stroemung. KielerMischwirtshaftliche Forschungsberichte, 45, 205-234 (1993).

Reinemann, D.J., and A. Grasshoff, 1993. Milklinecleaning dynamics:design guidelines and troubleshooting. Dairy, Foodand Environmental Sanitation, Vol 13, No 8. p 462-467.


104

INTERIM REPORT

Verification of Dairy Product Safety System(HACCP) Incorporated into Cheese Manufacturing

Personnel: J. Russell Bishop, assoc. professor, CDR,Mark Johnson, senior scientist, CDREric Johnson, assoc. professor, FRI, Steve Ingham,assist. professor, Dept. of Food Science, MarianneSmukowski, research specialist, CDR, Ann Larson,research specialist, FRI

Dates: July 1994 – December 1996


Objectives

1. Develop a verification/sampling procedure foruse in HACCP implementation

2. Deliver verification/sampling procedure toWisconsin’s dairy industry by: (a) general information shared during a

seminar setting (b) on-site establishment of procedure

3. Assist other Wisconsin cheese manufacturers inHACCP implementation

Summary

The International Dairy Foods Association (IDFA)has provided leadership to the dairy industry byemphasizing the importance of safe dairy products.IDFA ‘s food safety system, based on HACCP, is amanagement tool that stresses prevention byidentifying and controlling potential hazards. Weare applying IDFA’s system in cheese manufactur-ing.

Three cheese plants, varying in size and manufac-tured products, were sampled quarterly for micro-biological analyses. Preliminary data shows asteady decline in microbial counts after implement-ing and utilizing GMP’s and a HACCP program.Despite the overall decline in counts, we have noted

a diurnal increase in microorganism counts,depending on milk source and product manufac-tured. We will continue to investigate those areas inthe make procedures.

We are unable to test for pathogens in the finishedproduct due to Wisconsin Ag 80.56 regulations.This rule requires dairy facilities to report results ofmicrobiological tests conducted on pasteurized, orready-to-eat dairy products, that confirm thepresence of pathogenic organisms in the product.We are continuing to work with the WI Dept. ofAgriculture to waive this reporting of pathogenicorganisms for research purposes.

Significance to dairy industry

HACCP systems, as part of a total safety system,represent a method for assuring the production ofsafe dairy foods. There is a great need to verifywhether an implemented HACCP system is effec-tive, especially from a microbiological standpoint.This study will establish a microbiological sam-pling regime that cheese manufacturers can use toverify the effectiveness of their HACCP programs.

Safety and Quality

105

FINAL REPORT

Survival and Physiology of Listeria monocytogenesin Commercial Brines

Personnel: John H. Nelson, Emeritus researchprogram manager, Eric A. Johnson, associateprofessor, Ann E. Larson, research specialist

Dates: July 1993 – August 1994

Funding: Wisconsin Milk Marketing Board 9310

Objectives

1. Determine the behavior of L. monocytogenes inconcentrated NaCl solutions and commercialbrines.

2. Determine nutritional and environmental factorsthat affect survival of L. monocytogenes in concen-trated NaCl solutions and commercial brines.

3. Determine the effect of the replacement ofsodium with other cations in salt solutions on thesurvival of L. monocytogenes.

4. Evaluate antimicrobial substances for inactiva-tion of L. monocytogenes in salt brine systems.

5. Determine the effect of filtration of commercialbrines on the survival of L. monocytogenes.

6. Provide practical guidelines to the cheeseindustry for treatment of brines.

Summary

Thirty eight commercial brine samples wereobtained from fourteen cheese plants in Wisconsinand northern Illinois. These brines represent avariety of cheese and brining systems. Each brinewas analyzed for levels of total nitrogen, variousminerals, hypochlorite, and percent NaCl. Standardplate counts, psychrotrophic plate counts, yeast andmold plate counts, and pH were also determined. L.monocytogenes was then added to the commercialbrines, and to laboratory prepared NaCl solutions,under a variety of conditions. Survival was deter-mined over time by plating on selective media.

L. monocytogenes survived in certain brines.Survival of L. monocytogenes was not stronglycorrelated with total nitrogen content, mineralcontent, level of filtration before inoculation, pH,bacterial counts, or yeast and mold counts of thecommercial brines. L. monocytogenes survivedslightly longer in laboratory prepared brines madewith KCl than those containing NaCl. Lowertemperatures promoted L. monocytogenes survival.The presence of sodium hypochlorite at low levelsin brines resulted in rapid inactivation of L.monocytogenes. Other antimicrobials, includinghydrogen peroxide, sodium benzoate, potassiumsorbate, acetic acid, and lactic acid, had a negativeimpact on the survival of L. monocytogenes.


The overall goal of this project was to quantitate thebehavior of L. monocytogenes in brining systemsincluding commercial brines and to elucidate thenutritional and environmental factors determiningsurvival. A major goal of this study was to providepractical guidelines for the cheese industry forrotation, cleaning, maintenance and treatment ofbrines. The recommendations will be formulatedafter further review and discussions by the FoodSafety Task Force of WMMB and through othercommunications with the cheese industry. Theresearch should help avoid contamination of brinedcheeses by L. monocytogenes.


Presentation to the Dairy Safety Task Force, Wis-consin Milk Marketing Board, July 6, 1994.

Larson, A., E. A. Johnson, and J. H. Nelson. Behaviorof Listeria monocytogenes in commercial cheesebrines. In preparation.

Larson, A., E. A. Johnson, and J. H. Nelson. Factorsaffecting survival of Listeria monocytogenes inlaboratory brines. In preparation.


106

FINAL REPORT

Microbiological Safety and Quality of Reduced FatCheddar Cheese

Personnel: Eric A. Johnson, professor, John B.Luchansky, professor, Alvaro Quinones, researchspecialist, Al Degnan, research specialist, GregKulman, student, Food Research Institute

Dates: July1993 – December 1995

Funding: Dairy Management Inc. JNLC94

Objectives

1. Determine viability of Listeria monocytogenesand Salmonella and toxin formation by Clostridiumbotulinum in reduced fat Cheddar cheese. Comparethe behavior of these pathogens in reduced fatCheddar to conventional full fat Cheddar.

2. Evaluate the efficacy of natural preservativesincluding bacteriocins (e.g., pediocins, sakacins,enterocins, nisin), monoglyerides, antimicrobialpeptides from lactoferrin or conalbumin, andlysozyme for control of Listeria monocytogenes,Salmonella spp., Clostridium botulinum, andspoilage bacteria in reduced fat cheese.

Summary

The production of reduced fat cheeses holdsconsiderable promise for the Wisconsin cheeseindustry and, thus, considerable research has beensupported for this category of cheese. Research hasbeen directed at producing products with satisfac-tory taste and texture and manufacturing proce-dures for reduced fat cheeses have been established.Reduced fat cheeses are higher in moisture andusually lower in salt content and acidity than full fatcheeses. The moisture increase is due in part tocurd pH at drain and mill, curd washing or wheydilution, and the binding of water by fat substitutes.Reducing the fat content of cheese also affectsphysical properties including firmness, elasticity,and adhesiveness. As a result certain physical andchemical properties of reduced fat cheese areunique in comparison to their full fat counterparts.

Due to the higher moisture and unique physicalproperties, concerns have been raised regarding themicrobiological safety of reduced fat cheeses. Theoverall goal of this project was to evaluate themicrobiological safety and quality of reduced fatCheddar cheese.

Organisms and growth conditions

We used four strains of L. monocytogenes (Scott A,V7, Ohio, and California), four serovars of Salmo-nella (S. typhimurium, S. heidelberg, S. javiana, andS. dublin), and six strains of C. botulinum (56A,62A, 69A, Okra B, 113B, 17B). Listeriamonocytogenes was enumerated on MOX agaraccording to the FDA procedure (Lovett andHutchins, 1991). Salmonella was enumerated onHektoen agar. Botulinal toxin was determined bymouse assay as previously described (Malizio andJohnson, 1991).

Preparation of cheese

Reduced fat cheese was prepared using 45.4 kg(100 lb) low fat (1.3% butterfat) milk with acarbohydrate-based fat substitute (4% w/w Stellar,from A.E. Staley Mfg. Co., Decatur, IL). Milk waspasteurized, tempered to 31°C, and inoculated withstrains of each pathogen. After 5 min, antimicrobi-als were separately added to the milk, and inocu-lated with 0.015% (vol/vol) DVS (Direct Vat Set)starter culture (Chr. Hansen’s Laboratory, Inc.Milwaukee, WI). Milk was incubated for 15 min,and 0.01% chymosin (Chymax-Double strength,Pfizer, Milwaukee, WI) was added to the milk. Astandard Cheddar cheese-making schedule wascarried out. Control vats were inoculated withpathogens, but antimicrobials were not added. Forwash treatments, most of the whey was withdrawn,and the vat was filled with water at 38°C for 10 minwith gentle stirring. After pressing for 16 hr, 100 gportions were vacuum packaged, stored at either 12or 4°C, and sampled for botulinal toxin, Listeria

Safety and Quality

107

monocytogenes and Salmonella sp. at 0, 1, 7, 14, 21,30, 60, 90, and 180 d.

Results and Conclusions

Growth and survival of Listeria monocytogenes, andSalmonella and toxin production by Clostridiumbotulinum was determined in reduced fat and fullfat cheeses prepared with or without antimicrobials.In the absence of antimicrobials, counts of Salmo-nella (1-2 log10 cfu/g) and Listeria monocytogenes(2-3 log10 cfu/g) were lower in low-fat compared tofull-fat cheese at both 4° and 12°C. Botulinal toxinformation was delayed in reduced fat Cheddar at 4ºand 12°C compared to the full fat counterpart.

Antimicrobials affected the survival of pathogensand botulinal toxin production in reduced fat andfull fat cheese at 4° and 12°C. Nisin, enterocin, andmonolaurin had the strongest impact on pathogensurvival and toxin production. In samples treatedwith nisin, there was an initial large drop in viabil-ity of L. monocytogenes, and the drop was greatestin full fat cheese. In full fat but not in reduced fatcheeses, counts of L. monocytogenes eventuallyincreased in number. Nisin did not significantlynegate Salmonella survival in full-fat or reduced fatcheese. Enterocin promoted inactivation of L.monocytogenes and Salmonella in reduced fat butnot in full-fat cheese. Monolaurin also promotedinactivation of Salmonella and L. monocytogenes inreduced fat and full-fat cheeses. Monolaurin alsosignificantly delayed production of botulinal toxinin reduced fat cheese and to a lesser extent in full-fat cheese. None of the other antimicrobials testedsignificantly delayed production of botulinal toxin.

In conclusion, reduced fat cheese was less permis-sive to pathogen survival and toxin production. Theantimicrobials showed the same general pattern inwashed-curd cheeses. The results of this study showthat reduced fat cheeses are less permissive tosurvival of Salmonella and L. monocytogenes, anddelayed toxin production by C. botulinum. Themechanisms of pathogen inhibition are not clear atthis time, but are currently being investigated.

The detailed results including graphs of survival ofSalmonella sp., Listeria monocytogenes and toxin

production by C. botulinum are currently beingprepared for publication.


The results of this study show that reduced fatCheddar is less permissive to survival of Salmonellaand L. monocytogenes, and delays toxin productionby C. botulinum. These results support the conclu-sion that reduced fat Cheddar holds potential as asafe dairy product. The results of this projectshould be valuable in preventing foodborneoutbreaks in reduced fat Cheddar cheese andshould also be helpful in extending the shelf-life ofthis promising class of dairy products and main-taining the excellent safety record and reputation ofcheeses.

Publications

Luchansky, J. B. Knowing and controlling cheesepathogens. Presentation at the International CheeseTechnology Exposition, April 5-7, 1994, LaCrosse,Wisconsin.

Luchansky, J. B. Processing factors in cheese andlowfat cheese that control growth and survival ofListeria. Presentation at the symposium “Conquer-ing Listeria,” Dairy Research Foundation,Rosemont, Illinois.

Quinones, A, A. J. Degnan, J. B. Luchansky, and E. A.Johnson. Pathogen survival in reduced fat and full-fat Cheddar cheese containing antimicrobials.ASDA Abstract, J. Anim. Sci. 72:, Suppl 1, J. DairySci.: 77, Suppl 1, 1994.

Quinones, A, A. J. Degnan, J. B. Luchansky, and E. A.Johnson. Pathogen survival in reduced fat and fullfat Cheddar cheese containing antimicrobials.Poster presentation at the Annual Meeting of theFood Research Institute, May 25, 1994.

Gorski, D. Food Safety. Technological advancements.Dairy Foods95 (May 1994): 34-37.

Johnson, E. A. Microbiological safety of reduced fatCheddar cheese. Presentation at the AnnualIAMFES Meeting, Seattle, Washington, July 1996.


108

INTERIM REPORT

Control of Clostridium botulinum and RelatedSporeformers in Full Fat and Reduced Fat CheddarCheese

Personnel: Eric A. Johnson, professor, Ann E.Larson, senior research specialist, Food Microbiol-ogy and Toxicology


Funding: Dairy Management Inc. JHN95

Objectives

1. Identify clostridia responsible for gaseousspoilage of commercial Cheddar cheese obtainedfrom manufacturers in Wisconsin. Determinegrowth and spoilage by these isolates in full fat andreduced fat Cheddar cheeses.

2. Determine growth and toxin production by C.botulinum and toxigenic C. butyricum in Cheddar,reduced fat Cheddar (25% and 50% fat reduction),and in commercial reduced fat cheese and cheesesauce.

3. Determine the limiting nutrients and physicalconditions that govern growth of C. botulinum andC. sporogenes. Examine the location of sporeswithin the cheeses by scanning electron micros-copy (SEM) and by confocal microscopy.

4. Prepare peptide (cationic) and lipid fractionsfrom the cheese and examine these for inhibitoryactivity against pathogens: gram-positive(Clostridium and Listeria) and gram-negative(Salmonella, Campylobacter, E. coli).

Summary

Letters were sent by the Wisconsin CheesemakersAssociation to many commercial cheese factories inorder to obtain samples of gassy/spoiled cheese.About 4 to 5 samples of gaseous cheese wereobtained from the cheese industry. Efforts to isolateanaerobic sporeformers from these cheeses wereunsuccessful. Our data supports the conclusion that

anaerobic sporeformers are not a major cause ofgaseous spoilage in cheeses in Wisconsin. However,we continue our efforts to obtain natural isolates ofclostridia species from gassy cheeses. We welcomemanufacturers to send spoiled cheeses to ourlaboratory.

Our laboratory has a wide variety of strains ofClostidium botulinum, C. sporogenes, and otherclostridial species from various food and clinicalsources that are available for subsequent parts ofthis study. Botulism outbreaks and recalls haveoccurred in recent years in commercial cheesesauces due to temperature abuse. We obtainedsamples of product involved in one outbreak, aswell as many similar products. We have alsoobtained isolates of the causitive C. botulinum agentfrom the FDA. We have used this isolate and otherstrains of C. botulinum to inoculate various prod-ucts with added nutrients at various temperaturesto define the limiting nutrients and conditions forgrowth of anaerobic sporeformers. Work is inprogress to determine the effect of fat reduction onvarious anaerobic sporeformers, as well as definingspecific anticlostridial properties of Cheddar cheeseand cheese sauces.


Safety is the highest priority in producing dairyfoods. Recent outbreaks of botulism in commercialcheese products and in foods containing cheeseshave indicated that research is necessary to definethe condtions that enable growth and toxin produc-tion by C. botulinum and C. butyricum that pro-duces botulinal toxin. Additonally, this research willenhance the safety of dairy products by the identifi-cation and implementation of antimicrobialspresent in cheese including lipid compounds andpeptides.

Safety and Quality

109

INTERIM REPORT

Application of Biopreservatives as AntilisterialAgents in Queso Fresco and Cheddar Cheese

Personnel: John B. Luchansky, associate professor,Food Research Institute, Mark E. Johnson, seniorscientist, Center for Dairy Research, Nana Y. Farkye,research scientist, California Polytechnic StateUniversity Alan J. Degnan, senior research special-ist, Food Research Institute

Funding: Dairy Management Inc. LCH 95


Objectives

1. Identify lactic acid bacteria (LAB), for use inQueso Fresco, that produce bacteriocins effectiveagainst Listeria monocytogenes without producingsignificant amounts of organic acid(s). Identifybacteriocinogenic starters or adjuncts for use inCheddar cheese.

2. Validate bacteriocinogenic LAB and fermentatesof LAB as antilisterial agents in Queso Fresco andCheddar cheese at different steps during manufac-ture.

3. Evaluate the effect of biopreservatives on sensoryor biochemical qualities of Queso Fresco andCheddar cheese during storage at refrigeration andabuse temperatures.

Summary

Queso Fresco (QF) is a Hispanic, fresh-style softcheese that relies on enzymes rather than acid tocoagulate the cheese milk. Thus, the final product isa low acid (pH 6.2-6.4), high moisture (ca. 55%),buttery-tasting cheese, whose popularity is growingrapidly, especially in the southern and westernregions of the United States. The absence of an acid“hurdle” provides favorable conditions for bacterialproliferation and, presumably, a greater potentialfor foodborne hazard. Listeria monocytogenes is afoodborne pathogen which persists in a wide rangeof raw and processed foods. The proximate compo-sition and manufacture/storage conditions for QFprovide a suitable growth environment for this acidand salt tolerant, cold-loving pathogen. Therefore,we investigated the efficacy of biopreservativesdelivered during production of QF for the control ofL. monocytogenes.

In preliminary experiments, bacteriocinogeniclactic acid bacteria (LAB) from our culture reposi-tory with antilisterial activity (Table 1) wereevaluated to identify strains which grew adequatelyand produced sufficient bacteriocin but minimallactic acid during growth in milk compared tosynthetic media. Each strain was separately grownat 30°C for 20 hours in whole or skim milk, as well

Table 1. Bacteriocinogenic lactic acid bacteria with antilisterial activity.

Strain Bacteriocin

Lactococcus lactis subsp. lactis ATCC 11454 nisinPediococcus pentosaceus FBB611 pediocinLactobacillus sake LB706 sakacin AEnterococcus faecium 1083 enterocin 1083Pediococcus acidilactici JBL1095 pediocin AcH


110

as a lactose-based or glucose-based syntheticmedium. Strains that did not readily utilize lactoseshowed low growth rates (Figure 1a) and little or nobacteriocin production (Figure 1b) in either milkor lactose-based media compared to the glucose-containing synthetic medium. Only Lactococcuslactis subsp. lactis ATCC 11454 produced sufficientantilisterial activity [ca. 25,000 Arbitrary Units(AU) per ml of skim milk] for use incheese-making trials. As a disadvantage, L. lactisalso produced more acid in milk (pH 5.7) than wasconsidered ideal (data not shown). However, it wasassumed that adding this culture, or milkprefermented with this culture, to a relatively largevolume of cheese milk would have little effect onthe pH of QF.

Queso Fresco was prepared using a traditionalprocess that resulted in a final product with about52% moisture, 38% fat, 1.9% NaCl, and pH 6.4. Toinoculate QF with L. monocytogenes, pasteurized,

whole milk was challenged with a cocktail contain-ing strains V7, Ohio, and Scott A at about 104 cfu/ml. The final product was stored at 4 or 12°C andtested for viable L. monocytogenes at 0, 1, 3, 7, 14,and 21 days. For experimental treatments, thebacteriocin, nisin was added to QF by 3 differentstrategies. First, a prefermented milk (PFM) wasprepared by overnight growth of L. lactis ATCC11454 in 2.5 L of whole milk at 30°C and theresulting nisin-containing PFM (ca. 25,000 AU/ml;pH 5.5) was added to the cheese milk. The PFM wasmixed at a level of 5% with the cheese milk contain-ing the L. monocytogenes cocktail just before addingthe starter culture, delivering about 500 AU of nisinper ml of cheese milk. Second, nisin activity (1,000AU/ml) was added directly to cheese milk byblending 50 g of NisaplinTM, a commercial prepara-tion of nisin, with 50 kg milk. Third, nisin activitywas delivered by “salting” 50 g Nisaplin simulta-neously with cheese salt directly onto 5 kg of curd(produced from 50 kg of milk). The salt/nisin

Figure 1a. Population increases (log10 cfu/ml) of bacteriocinogenic lactic acid bacteria in milkand lactose- or glucose-based synthetic media.

Safety and Quality

111

Figure 1b. Antilisterial activity (AU/ml) produced by bacteriocinogenic lactic acid bacteria inmilk and lactose- or glucose-based synthetic media.

Figure 2. Fate of L. monocytogenes in Queso Fresco prepared with nisin (500 AU/ml) addedvia prefermented milk (PFM; 5%).


112

combination was thoroughly distributed through-out the cheese curd by hand mixing just prior tohooping and pressing.

In control batches (nisin-free) of QF stored at 12°C,counts of L. monocytogenes increased over 3 log10

cfu/g cheese within 3 days and remained at maxi-mum levels over the 21 day sampling period(Figure 2). At 4°C, pathogen counts in the controlbatches increased more slowly, reaching about 1log10 cfu/g over initial inoculum levels over 21 days(Figure 2). For QF prepared with PFM and stored at12°C, populations of L. monocytogenes decreasedabout 2 log10 units within 1 day. However, countsrecovered similar to those in control batches within21 days. When stored at 4°C, counts of the pathogenalso declined about 2 log10 units within 1 day, andcounts remained about 2 log10 units below countsof the control over the 21 day sampling period.Compared to storage at 12°C, after the initial 2 log10

unit decrease the additional hurdle of refrigerationsubstantially suppressed the pathogen duringstorage at 4°C. The initial decrease in pathogennumbers observed in QF prepared with PFM wasattributed to the presence of nisin and perhaps theslight acidity introduced via the PFM.

In QF prepared with Nisaplin added to the cheesemilk, and then stored at 12°C, counts of L.monocytogenes decreased 1-2 log10 units within 1day and remained 2-3 log10 units below those incontrol batches over the 21 day storage period(Figure 3). Similarly, in batches of QF stored at 4°C,populations of L. monocytogenes decreased about1-2 log10 units within 1 day and remained staticthereafter at 2-3 log10 units below control popula-tions over 21 days at 4°C. However, the combinationof Nisaplin and refrigeration displayed greaterantilisterial activity than using Nisaplin and storingat 12°C. The greater reduction of L. monocytogenesachieved by adding Nisaplin to the cheese milkrather than adding nisin to the cheese milk viaPFM was attributed to the higher levels of nisinactivity (1,000 vs 500 AU/ml) in the former. Also,the use of a nisin preparation compared to anisin-producing strain in PFM enabled precisecontrol over the degree of antilisterial activity.Sufficient antilisterial activity was not produced byL. lactis ATCC 11454 for efficacious use of PFM at alevel of 5%.

In QF prepared with Nisaplin “salted” into cheesecurd with subsequent storage at 12°C, counts of L.monocytogenes initially dropped 3 log10 units

Figure 3. Fate of L. monocytogenes in Queso Fresco prepared with 50 g Nisaplin added to 50kg cheese milk.

Safety and Quality

113

within day 1 (Figure 4). Despite a 2-3 log10 resur-gence in pathogen levels, counts of L.monocytogenes remained 2-3 log10 units belowlevels observed in control batches. Similarly, for QFstored at 4°C, pathogen numbers dropped belowdetection (5 cfu/g cheese) and remained static forup to 14 days, after which low levels of L.monocytogenes were detected. The differencebetween populations of L. monocytogenes incontrol (nisin-free) cheese compared to cheeseprepared with Nisaplin added at the salting stage at4°C was about 4 log10 units.

Batches of QF treated with Nisaplin receivedequivalent amounts, whether added to the cheesemilk (50 g Nisaplin/50 kg milk), or “salted” onto thecurd (50 g Nisaplin/5 kg curd). However, activitylevels were increased by adding nisin activity to 5kg curd (10,000 AU/g) compared to 50 kg milk(1,000 AU/g). This may explain, in part, whypopulations of L. monocytogenes were lower in QFprepared by adding Nisaplin to the curd ratherthan the cheese milk. Also, delivering nisin activity

to the curd precludes the loss of activitythat presumably occurs when the whey isdrained. Thus, the “salting” method ofdelivering antilisterial activity to QF at thecurd phase was more efficient than deliver-ing an equivalent amount to the cheesemilk. Regardless of when or howantilisterial activity was delivered to QF, thepresence of the bacteriocin was sufficient toreduce pathogen numbers appreciably forup to 5 days, which is the approximatetimeframe for consumption of QF. Lastly,the addition of nisin-containing PFM orNisaplin to cheese milk or curd did notappreciably alter pH. Generally, pH values inall treatments discussed in this report werenot significantly different, with batchesstored at 12°C ranging from an initial pH ofpH 6.4 to a final pH of pH 5.7 at day 21. ThepH of the cheese stored at 4°C ranged froman initial pH of pH 6.4 to a final pH of pH6.2 at day 21.

Figure 4. Fate of L. monocytogenes in Queso Fresco prepared with 50 g Nisaplin “salted”onto 5 kg curd prepared from 50 kg cheese milk.


114


Queso fresco (QF) is a high pH, high moisture softcheese prepared with minimal starter activity,considerable hand manipulation, and is stored atelevated temperatures. Thus, QF is particularlyprone to microbial hazard, notably from thefoodborne pathogen L. monocytogenes. Althoughgood manufacturing practices and proper sanita-tion can reduce the likelihood of hazard, additionalstrategies are needed to manage L. monocytogenesin Hispanic-style cheeses. Our results demonstratethat biopreservatives can appreciably reducenumbers of L. monocytogenes in QF, particularlywhen added directly to the cheese milk or curd. Assuch, bacteriocins offer great potential for improv-ing the safety of QF and dairy products in general.

Publications

Glass, K. A., B. Prasad, J. H. Schlyter, H. E. Uljas, N. Y.Farkye, and J. B. Luchansky. 1995. Effects of acidtype and ALTATM2341 on Listeria monocytogenes ina Queso Blanco cheese. J. Food Prot. 58:737-741.

Luchansky, J. B., and N. Y. Farkye. 1996. Controllingpathogenic and spoilage bacteria in cheese withbiopreservatives. California Dairy Beat 3(3):6.

Degnan, A. J., A. Piva, and J. B. Luchansky. 1996.Preservation of food using lactic acid bacteria andbacteriocins. Food Testing & Analysis 2(2):17-19.

Abstracts

Degnan, A. J., N. Y. Farkye, M. E. Johnson, and J. B.Luchansky. 1996. Use of nisin to control Listeriamonocytogenes in a Queso Fresco cheese. Abstractsof the Annual Meeting of the International Associa-tion of Milk, Food and Environmental Sanitarians(#154), pg 70.

Presentations

“Bacteriocin use in foods.” Invited speaker at theFDA Science Forum on Regulatory Sciences.Washington, DC, September 30, 1994. (J.B.Luchansky)

“Managing pathogens in processed food.” Presen-tation at the Vision for Food Safety Symposium.Madison, WI, January 18, 1995. (J.B. Luchansky)

“Update on biocontrol and subtyping of foodbornepathogens.” Scientific Lectureship, Institute ofFood Technologists. Honolulu, HI, January 25, 1995.(J.B. Luchansky)

“Applications of bacteriocins and bacteriocin-producing lactic acid bacteria in foods.” Presenta-tion at the International Dairy Lactic Acid BacteriaConference. Palmerston North, New Zealand,February 20, 1995. (J.B. Luchansky)

“Applications of pulsed-field gel electrophoresisand lactic acid bacteria for subtyping andbiocontrol of foodborne pathogens.” DistinguishedScience in Microbiology and Cell Science Lecturer,University of Florida. Gainesville, FL, March 24,1995. (J.B. Luchansky)

“Microbial safety of reduced fat Cheddar cheese.”Invited speaker at the International BusinessCommunications Conference of Fat and Choles-terol-reduced Foods. New Orleans, LA, March 30,1995. (J.B. Luchansky)

“Applications of lactic acid bacteria in food preser-vation.” Keynote speaker at the Advanced Work-shop on Bacteriocins of Lactic Acid Bacteria. Banff,Alberta, Canada, April 18, 1995. (J.B. Luchansky)

“Applications of genomic fingerprinting andbacteriocins for food microbiology.” AnnualMeeting of the Food Research Institute. Madison,WI, May 16, 1995. (J.B. Luchansky)

“Molecular typing and biocontrol of L.monocytogenes in foods.” Distinguished Lecture-ship, Czechoslovakia Society of Microbiology. Brno,Czech Republic, October 25, 1995. (J.B. Luchansky)

“Dairy applications for biopreservatives andmolecular subtyping.” The 63rd Annual Dairy andFood Industry Conference, The Ohio State Univer-sity. Columbus, OH, February 14, 1996. (J.B.Luchansky)

Safety and Quality

115

“Applications of biopreservatives and pulsed-fieldfingerprinting for the dairy and foods industries.”Invited speaker at the California Polytechnic StateUniversity, Dairy Products Technology Center. SanLuis Obispo, CA, March 11, 1996. (J.B. Luchansky)


116

Chapter 5

Communications


118

Communications

119

Center for Dairy Research Communications

Personnel: Sarah Quinones, outreach program manager, Karen Paulus, editor, Dawn Hyatt, programassistant

Outreach Events

CDR hosted a technical session on cheese yield. The proceedingsfor this session, distributed at the event, were very popular.

This annual workshop provides participants with an understand-ing of the nature and the control of dairy foods pathogens. Acomprehensive set of course materials is developed by the speakersand distributed for this event.

The CDR Open House was a tremendous success with 175 attendeesrepresenting producers and processors. This event showcased CDRand CDR produced technologies. Over 16 press organizationscovered the CDR event.

Over 50 researchers, industry representatives, board members, andproducers met in Madison to participate in the National MilkfatTechnology Forum, co-sponsored by Dairy Management Inc., theWisconsin Milk Marketing Board and the California Dairy ResearchFoundation. Scientists presented updates on all aspects of milkfattechnology and applications, including pre-harvest technologies,butter manufacturing, milkfat fractionation, enzymatic modifica-tion, butter flavor, and nutrition. The presentations were followed bya lengthy discussion of the current program and issues, and needsfor future milkfat research. DMI, WMMB and CDR are using theinformation to develop the National Milkfat Plan.

CDR staffed a booth at this annual meeting to present informationabout our research and applications programs. We offered samplesof reduced fat cheese, along with a technical fact sheet describingthe patented make procedure.

Technical Seminars

“Rheology and Microstructure of Dairy Products,” Dr. Karsten B. Qvist, Director, Institute for DairyResearch, Copenhagen, Denmark. August, 1995.

“Recent Research on Proteolysis During Cheese Ripening,” Dr. Pat Fox, Professor, Dept. of Food Chemistry,Cork, Ireland. August 21, 1995.

WCMA/CDR Cheese conferenceApril, 1996

Producing Safe Dairy FoodsAugust 29-30, 1995

CDR Open HouseMarch 27, 1996

Milkfat Technology Forum,April 23-24, 1996

CDR Booth, Institute of FoodTechnologists Annual Meeting,June, 1996


120

Scientist exchanges

Hosting professor: Kirk Parkin, Dept. of Food Science. March, 1995to October, 1995. Project: Preparation of butteroil-in-wateremulsions using blends of milk proteins and lipids. (See report inmilkfat research chapter.)

Hosting professor: S. Gunasekaran, Dept. of AgriculturalEngineering. July - September, 1995. Project: Objective method formeasuring flow and cheese melt characteristics. (Report in Cheeseresearch chapter.)

Hosting professor: Charles Hill Jr., Dept. of Chemical Engineering.July - August, 1995. Project: Use of immobilized lipases for themodification of milkfat. ( Report published in milkfat researchchapter.)

Visiting Mentor. Dr. Qvist is an internationally known expert inrheology of dairy foods. While in Madison, he met with over fifteenresearchers at the University to discuss cheese structure, function-ality, and manufacturing technology. Dr. Qvist also visited LandO’Lakes to make a presentation.

Industry Teams

The thirty members of the cheese industry team met twice in thepast year to develop the Center’s research program and to hearabout the progress of research projects. It was decided that theteam membership would be better defined in the future to includeclear objectives, benefits, and an annual membership fee.

The safety industry team consists of 20 members who met oncethis past year to hear about research progress, to discuss currentsafety concerns, and to identify future research topics.

Publications

Although we have considered different options to distribute thisinformation, CDR continues to publish and mail the annual report.(Reports will also be available through our web site.) This technicalannual report summarizes all current research projects and isdistributed to funding agencies and scientists in academia, indus-try and at other dairy research centers around the world.

Marit Reierstad. AgriculturalUniversity of Norway

Muhammet Ak, IstanbulTechnical University

Hugo Garcia,Centro deGraduados, Mexico

Karsten Qvist, Director,Institute for Dairy Research,Copenhagen, Denmark

Cheese

Safety

Annual report

“Physico-Chemical Changes in the Water Phase of Cheese During Aging: Implications for ControllingFunctionality,” Dr. Paul Kindstedt, Associate Professor, Dept. of Animal and Food Sciences, University ofVermont, Burlington. Sept. 19, 1995. Kindstedt also met with industry and the Wisconsin Dept. of Ag inindividual consultations.

Communications

121

Cutting Edge We also publish a shorter, general report which we now call theCutting Edge. We use it to describe CDR and highlight severalresearch projects in an easy-to-read style. We have taken the CuttingEdge to the Dairy Expo, IFT, and the joint meeting with WisconsinCheesemakers.

The Dairy Pipeline continues to be a practical and dependabletraining publication for the dairy industry. This year, in honor of our10th Anniversary, we published a collection of all the Curd Clinicspublished in the Pipeline and distributed it at the LaCrosse, Wiscon-sin Cheesemakers Conference. The Pipeline, including back issues,will soon be available through our web site.

Technical fact sheets are a simple, efficient way to describe andexplain recently completed research projects. For example, weprinted a technical fact sheet about the make schedule for reducedfat Cheddar and had it available at our IFT booth, next to samples ofreduced fat Cheddar.

An information packet was developed which provides an overviewof CDR’s vision, research, applications, and projects. This packet ofinformation will be given to visitors and taken on calls to industrysites. A series of visual icons were designed for this packet whichcan be used throughout CDR communications pieces includingCDR’s homepage and promotional videos.

A handout was prepared for each of the four applications programareas, Milkfat as an Ingredient, Cheese as an Ingredient, Safety andQuality, and Wisconsin Specialty Cheese. These handouts provideprospective industry clients with an understanding of the vision foreach area, their services and research facilities.

Videos

Several segments of video footage had been done by news stationsand promotional agencies. This promotional video put togetherclips from various sources to present a 6 minute overview of the 5key aspects of CDR — the vision, research program, applicationsprograms, transfer of information and technology, and researchprojects. This video was designed for the director to use in hispromotional travel and exhibits at industry events.

A TV news feature segment was done on the cheese functionality

Dairy Pipeline

Technical Fact Sheets

CDR Information Packet

CDR Applications Fact sheets

CDR promotional video

ABC News Milwaukee


122

research of S. Gunasekaran and Norm Olson. This segment was anexcellent means of conveying CDR’s partnership with the foodindustry and role in improving the quality of cheese as aningredient.

CDR On-Line

Established CDR On-line Web site.Established databases based on industry needs of informationaccessBegan porting databases to web siteBegan setup of Local Area Network

CDR’s web site went on-line in September 1995 and made it’s publicdebut at the World Dairy Expo in October of ’95. Our web site islisted in Yahoo, Webcrawler and other Web search databases and weare receiving e-mail via the web site on dairy research related topics.It has been well regarded and received a “must see” in a review ofdairy related web sites.

Key Accomplishments:

Annual Report 1996 Wisconsin Center for Dairy Research

Documents

Transcript of Annual Report 1996 Wisconsin Center for Dairy Research