CHAPTER 10

Biotechnology and Food Safety:Benefits of Genetic Modifications

T. Verrips

INTRODUCTION

A number of consumer research studies haveshown that fermented food products are re-garded by consumers as healthy and natural.9

Factually and from a historic point of view, thisis correct. During the approximately 7,000 yearsthat fermented foods have been produced, anenormous selection process has taken place, andthe tasty products that are presently on the mar-ket are the result of this selection process. It iswell known that fermentation of foods anddrinks prolongs the shelf life of these products.The historic claims that fermented foods, par-ticularly those derived from milk, prevent dis-eases are supported by many studies. Thoseclaims related to mucosal health are excellentlysummarized by Salminen et 0/.23 Nevertheless,the title of this chapter suggests that at present,there is concern about the safety of fermentedfood products. This is due to the rapidly growingapplication of genetically modified food prod-ucts, including fermented food products. In par-ticular, the penetration on the market of geneti-cally modified plants is remarkable (Figure 10-1and Table 10-1). In addition, more than 20 foodenzymes that are on the market at present areproduced using recombinant DNA technology(Table 10-2).

The perception of genetic modification offood products by European consumers is notvery positive and varies from country to country.Generally, the attitude of Western consumers isneutral to positive (Figure 10- 2), whereas in

most of the developing countries, this technol-ogy is seen as an opportunity.20 However, inspite of the present problems in a number ofWest European countries, and excluding smallgroups of dogmatic opponents of genetic engi-neering, the perception of the consumer will ulti-mately be determined by the benefits versusrisks ratio.

In spite of many efforts, neither scientists noropinion makers have been able to properly ex-plain the potential risks of genetically modifiedfood products. Therefore, many consumers per-ceive recombinant DNA technology as an intrin-sically dangerous technology, although in the 25years that this technology has been applied, nounintended dangerous materials have been pro-duced. The main reason for the present percep-tion of this technology by West European con-sumers is that in many discussions, no cleardefinitions of the nature of the genetically modi-fied food are used. In an attempt to rationalizethe discussions and to avoid inappropriate ethi-cal discussions, let us examine simple decisionschemes based on rules of the U.S. Food andDrug Administration (FDA), the United King-dom, and the Netherlands,29 and a model forvarious genetically modified products (Figure10-3).30 It should be emphasized that this figurereflects the author's personal view concerningthe safety of genetically modified foods that donot contain (antibiotic) resistance markers("clean"). Genetically modified plants or micro-organisms containing antibiotic resistance mark-ers present a very difficult to quantify but realis-

tic risk of spreading genes encoding antibioticresistance in the environment.6'10 The chancethat these distributed antibiotic-resistant genesare taken up is very low, but not zero, and there-

Table 10-1 Rapid Increase of TransgenicPlants

Crop 1996 1997

Maize 525,000 4,400,000Soybean 400,000 5,250,000*Potato/tomato 40,000 500,000+Oilseed 200,000 1,600,000Cotton 810,000 1,200,000

"Includes the South Amerian 1996/1997 harvest.+Includes estimates for China.Note: In 1994, no transgenic crops were cultivated for

usage in the agrofoods industry.

fore undermines the argument that geneticallymodified plants or microorganisms containing agene that encodes an intrinsically safe proteinare safe.

In addition to the lack of proper explanationof the potential risk of genetically modifiedfoods, the benefits for the consumer are eithernonexistent or not communicated properly toconsumers. This is a major weakness, and onlywhen the benefits to consumers become clearwill genetically modified foods be accepted bythe majority of consumers. It is obvious that ben-efits are not absolute values but relative values,strongly influenced by demographic, geo-graphic, and socioeconomic factors.

In this chapter, the supply chains of fermentedfoods are taken as guides to point out the poten-tial benefits and risks of every step of thesechains. The emphasis will be on fermented foods

first GMO first GMObacterium plant

Note: X-axis-penetration as a function of time: The expression of somatostatin in E. coll is taken asthe starting point and insulin as the first large-scale pharmaceutical product. Y-axis-estimatedpercent of rDNA product in certain area (Detergent enzymes are nearly 100% rDNA products, about25% of the food enzymes (e.g., chymosin) are made via rDNA technology).

Figure 10-1 Penetration of rDNA technology in agriculture, Pharmaceuticals, and consumer products.

SomatostatinInsulin

Foodenzymes

7 Qualitytrait crops

GMOMicroorganisms

Agronomictrait crops

Agronomictrait cropsDetergent

enzymes

Estimatedsales GMO/sales total

Table 10-2 Commercial Food Enzymes Made by rDNA Technology (as on the Market at 1.5.1998)

Enzyme Commercial Name Producer Application

oc-Galactosidase cc-galactosidase NOVO N. Animal feedXylanase Bio-feed wheat id idLipase Lipozym id lnteresterificationALDC Maturex id BeerAmylase Novamyl id Bread1,3 Lipase Novozym 677 id BreadXylanase Peptopan id BreadAmylase Maltogene id StarchChymosin Novoren id CheeseLipase Novozym 398 id lnteresterificationLipase Novozym 435 id idFytase Pytan id Animal feedAmylase Termamyl id StarchTransferase Toryzym id StarchChymosine Maxiran DSM/Gb CheesePhytase Natuphos id Animal FeedChymosin Chymogen C. Hansen CheeseChymosin Chymax Pfizer CheeseXylanase Biobake710 Quest BreadInvertase id ChocolateGlucanase id Beeroc-Galactosidase id Guar modification

starting from traditionally bred or geneticallymodified plants and milk from traditionally bredcattle. The approach of analyzing the safety of fer-mented products as a function of the supply chainhas been adopted, because only with such an ap-proach will all potential risks be included, which isessential to gain the confidence of consumers.

By analogy with microbial risk assessments,28

risk is defined as the probability of a hazard oc-curring and hazard as a harmful event. Thismeans that when the newly introduced gene en-codes a protein that is intrinsically safe and doesnot provide a selection benefit for any receivingorganism, the hazard is considered as zero. Con-sequently, the risk is zero and the safety of thefood product is 100%. If, on the other hand, thenewly introduced gene encodes for a protein thatis intrinsically safe but provides the receivingorganisms with a selection benefit in the openenvironment, then the health hazard is zero, butthe environmental hazard is not zero. Conse-

quently, the probability of transfer of the gene tothe receiver has to be estimated. Genes encodingintrinsically unsafe proteins (i.e., health hazard> O) are not considered because authorities willnot give permission for cloning of such genes inplants and/or microorganisms that enter the foodchain; neither will the food industry ever usesuch organisms.

SUPPLY CHAIN OF FERMENTEDFOODS AND ITS IMPORTANCE TOENSURE SAFE PRODUCTS

Table 10-3 provides a selection of fermentedfoods based on the raw material(s) that is (ormay become in the near future) derived from ge-netically modified plants.

The steps in the supply chains (including po-tential beneficial and risk aspects) of fermentedfoods using traditionally bred plant materials ormilk from traditionally bred cows with fermen-

Figure 10-2 Acceptance of rDNA technology for consumer products derived from plants and microorganismsin various countries. Acceptance varies from approximately 90% in Canada, United States, and Portugal to 30and 45% in Austria and Germany, respectively.

tation processes using genetically modified mi-croorganisms are provided in Figure ICM. How-ever, Tables 10-1 and 10-3 illustrate that geneti-cally modified plants will become a majorsource of plant raw material for fermented foodsin the near future, and therefore a separate sup-ply chain risk/benefits analysis for fermentedproducts derived from genetically modifiedplants is provided in Figure 10-5.

It is impossible to deal in this chapter with all ofthe products in Table 10-3 for both traditional andgenetically modified raw materials and/or for tra-ditional and genetically modified microorganismsused in the fermentations. Therefore, only two ex-amples are worked out in more detail.

1 . soy sauce — a fermented product using ge-netically modified wheat and/or soybeans,traditional or genetically modified micro-

organisms, and traditional or recombinantDNA enzymes

2. cheese — a fermented product using nor-mal cow's milk, genetically modifiedfunctional microorganisms, and/or en-zymes made by genetically modified mi-croorganisms

These two examples cover quite well thewhole spectrum of fermented food products inwhich recombinant technology is used at somestep in the overall process.

As stated in the introduction to this chapter, fer-mented foods derived from genetically modifiedplant material or produced by genetically modi-fied microorganisms will only be accepted if thereare clear direct (e.g., significantly better quality orshelf life, more healthy or more convenient) or in-direct (e.g., availability, environment) benefits for

SpainItaly

PortugalFrance

BelgiumIreland

UK

CanadaNorway

NetherlandSwedenFinlandJapan

DenmarkAustria

GermanyUSA

Known Unknown Knownnonintegrated integrated integrated

Note: X-axis-physical state of foreign gene(s): "known" means that the location of foreign gene(s) is/areknown exactly, "integrated" means stable integration of foreign gene(s) on the chromosome of the host;Y-axis-host organism; Z-axis-whether the consumer product is free of DNA (e.g., cheese made withchymosin) or product contains inactivated cells that produced the rDNA product (i.e., tomato paste) orthe rDNA producing cells are "alive" present in the product (i.e., fresh tomatoes). Vertical spheres-norisk; white spheres-risk assessment inconclusive; dotted spheres-either risk assessment inconclusiveor ethical objections against such consumer products.

/Vote: This scheme is based on the assumption that the used constructs do not contain an antibioticresistance marker. Otherwise, the risk to the environment will increase for all cases except containedfermentation of microorganisms.29

Figure 10-3 Summary of risk assessment of rDNA products on basis of three parameters.

the consumer. Some potential benefits will be dis-cussed in the following sections.

Some Consumer Benefits of GeneticallyModified Plants

Benefits are relative, and for each group ofconsumers, benefits will be perceived differ-ently. Consumer studies in Western Europeshow that the order of consumer criteria to buyproducts is quality, health aspects, convenience,environment, and price. It is likely that price to-gether with availability will be on top of such alist in developing countries. In the Westernworld, there is a surplus of food raw materials

and improvement of crop yield is not perceived asa benefit by the consumer; neither is the loss ofmaterial during transport seen as a problem. How-ever, the maintenance of quality during transportand at home is considered as a benefit by consum-ers, as was demonstrated by the initial successfulintroduction of the Flav Sav tomato (with anantisense polygalacturonase gene that ensured de-lay of softening of the tomato) by Calgene on theUnited States market in the mid-1990s.

In many developing countries, however, thereis at present a shortage of food raw materials.Therefore, an increase in the yield of crops bygenetic modification is seen as a major advan-tage. There is evidence that shows that in the

Host

Microorganism

Plant

Animal

Type of product

DNA in intact cell

DNA in dead cell

Free of DNA

Table 10-3 Overview of Most Important Fermented Food Products Derived from Plants of which Ge-netically Modified Varieties Are on the Market and the Main Functional Microorganisms

Raw Materials

Cassava

/ sorghum

Maize

/ cassavaRice

/ Black gram

/ Carrots/ Soybeans

or cassavaSoybean

/ rice

/ wheat

/ wheat

/ wheatWheat

/milk

Functional Microorganisms

BacteriaCorynebacterium,GeotrichumLactic acid bacteria,Saccharomyces, CandidaAspergillus, lactic acidbacteria, yeastYeasts, bacteriaLactic acid bacteria,yeasts, moldsLactic acid bacteria, yeastsMonascusRhizopusLactic acid bacteriaSaccharomycesAspergillus, BacillusHansenula, Candida,GeotrichumLactic acid bacteriaTorulopsisHansenulaAspergillus, Lactic acidbacteria, yeastsYeasts, moldsMucor, AspergillusAspergillus, RhizopusActinomucorBacillus nattoActinomucor, MucorRhizopusAspergillus, yeastsLactic acid bacteriaAspergillus, Lactic acidbacteriaAspergillus, yeastsLactic acid bacteriaAspergillusSaccharomycesMoldsLactic acid bacteria

Product

Chickwangne (Congo)

Gari (Nigeria)

Burukutu (Nigeria)

Chicha (Peru)Jamin-bang (Brazil)

Ogi (Nigeria)Banku (Ghana)Ang-kak (China)Lao-chao (China)

Puto (Philippines)Sierra rice (Ecuador)

Torani (India)

IdIi (India)Kanji (India)

Miso (China, Japan)Tape (Indonesia)Chee-fan (China)Meju (Korea)Meitauza (China)Natto (Japan)Sufu (China)Tempeh (Indonesia)

Miso (China, Japan)

Hamanatto (Japan)

Soy sauce (southeast Asia)Tao-si (Philippines)Jalebies (India, Pakistan)Minhin (China)Kishk (Egypt)

next 10 years, the increase in the supply of agri-cultural products will be less than the expectedincrease in world population and food shortage,especially in developing countries, will in-crease16 (Figure 10-6). It has also been sug-

gested that only when modern biotechnology re-sults in a second green revolution may this glo-bal problem be prevented.17

Also, the prevention of the substantial lossesduring transportation of plant raw materials (up

Figure 10-4 Schematic benefit versus risk ratio of rDNA products made by microorganism.

to 30% of harvested plant materials is destroyeddue to microorganisms, insects, or uncontrolledendogenous processes) is seen as a major benefitfor consumers in that part of the world.

There is evidence that modern biotechnologycan significantly contribute to increasing yieldsor reducing losses of plants and plant products inthe first phases of the supply chain. Monsanto 'sRoundup herbicide-resistant crops need less her-bicides (approximately 10% less) and show in-creased yields (5-15%) (Farmers OrganizationArgentina, personal communication, November1997).

Plants producing Bacillus thuringiensis BT-protein (preferably more than one variety of this

protein so as to minimize the probability of ad-aptation) are quite well protected against insects,thereby increasing the yield. Modern biotech-nology can also contribute to the reduction oflosses due to microbial spoilage during distribu-tion. Plants have a rich arsenal to defend them-selves against invading microorganisms.5 Theycan change the structure of plant tissue by exten-sive cross linking catalyzed by redox enzymes,produce low molecular mass antimicrobialagents, and degrade cell walls of invading mi-croorganisms using hydrolases they produce. Inparticular, plant pathogenic fungi can be effec-tively destroyed by these enzymes. Some com-panies like Novartis and Zeneca are very active

Environment

Consumption

Distribution

Raw materials

Processing, includingfermentation

Faster process,less salt

Price reduction,better quality, healthier

See 2.2 a-d and3.1, step 1a, 1b

ZEROSee 3.1, step

2a-c

Benefits Risks

Figure 10-5 Schematic benefit versus risk ratio of rDNA product made by plants.

in this field, as shown by the five and seven pat-ents they filed respectively on this technologybetween 1980 and 1997. This approach is lim-ited to improving existing plant protection sys-tems. However, recombinant DNA technologycan do more. Plant Genetic Systems had aproject in the mid-1980s to express potent anti-microbial peptides (e.g., apidaecin)3 in plants,although with the intention to isolate these anti-microbials from the plant and use them in alltype of products. At present, a large number ofantimicrobial peptides are known, some ofamazing effectivity.1

A consumer benefit of currently available ge-netically modified plants can be a lower price dueto higher yield, reduced cost for chemicals, andreduced losses during harvesting and transport.

However, an important benefit can be delivered byplants with the right balance of micronutrients thatcontribute to consumer health. Plants are a well-known and extremely rich source of all kinds ofcomponents that may contribute to consumerhealth, such as antioxidants, organic metal com-plexes, phytoestrogens, phytosterols, and vita-mins. At present, the diet of approximately 2.5 bil-lion people contains insufficient amounts ofminerals and vitamins. The consequences of theseshortages are dramatic. For example, each year, 2million young children die and approximately300,000 children go blind due to a shortage of vi-tamin A. With the rapid increase in our knowledgeof cell and molecular biology, and much better andfaster analytical techniques, the number and kindof plant and microbial components with sustain-

Environment

Consumption

Distribution

Processing, includingfermentation

Raw materials

See 3.2,step 2a-2b

See 3.2, step 3

See 3.2, step 4

See 3.2, step 5

See Table 10-2

Benefits Risks

Year

Figure 10-6 World grain harvest during the last 35 years. A steady increase in total production from 1.2 to 2.2billion of metric tons; a rapid increase in amount per capita from 1966 to 1987, followed by a decrease from 1988to present.16 Source: Reprinted with permission from FAOSTAT, C.C. Mann, Crop Scientists Seek a New Revo-lution, Science, Vol. 283, pp. 310-314. Copyright 1999 American Association for the Advancement of Science.

able health claims will increase significantly in thenear future.

What happens to these components with po-tential healthy properties during fermentation isnot very well known, but it is likely that chemi-cal nature and bioavailability of these compo-nents will be influenced (for better or worse) bythe various steps in a fermentation process.Some literature is available on the conversion ofestrogens and phytoestrogens by the microfloraof the gastrointestinal system and show thatphytoestrogens can be (extensively) convertedinto more effective compounds, such as estroneinto oestradiol.21 However, knowledge in thisarea is too fragmented and limited, and this im-portant aspect has to be researched.

Some Potential Hazards and Risks ofGenetically Modified Plants

It is outside the scope of this chapter to go intodetail on the potential hazards of genetically

modified plants. However, the integral supplychain approach followed requires that some as-pects of the potential risk are discussed. Atpresent, there are at least four issues in relationto genetically modified plants.

1. Most of the genetically modified plantscontain an antibiotic resistance gene asmarker and, although the marker gene assuch is not expressed, the potential transferof this marker gene to other crops in thesurrounding area has been studied.11 Al-though the frequency of such an event islow, it is measurable; therefore, this ap-proach may contribute to a further spread ofantibiotic resistance genes in the environ-ment, thereby contributing to an extremelyundesirable further increase of antibioticresistance of human and animal pathogens.

2. The gene providing the plant with a new de-sired property, such as herbicide resistance,can also be spread in the environment. This

Kilo

gram

s/ca

pita

World grain harvest per capitaP

roduction (billions of metric tons)

means that there is a small probability thatweeds will pick up this property and be-come resistant to herbicides. This spread ofgenes providing herbicide resistance is anissue that has to be addressed.

The spread of genes providing plantswith new desired properties related tohealth benefits of consumers will also oc-cur, but in this case, the gene will mostprobably not provide the recipient with anecological benefit; therefore, this risk is solow that it can be neglected.

3. It is still not possible to integrate the newgenetic information in a predeterminedplace in the genome of the plant. The ran-dom integration may result in the destruc-tion or enhancement of the expression ofgenes at the locus of integration and there-fore may change the metabolism of theplant. To exclude the introduction of a haz-ard, one should precisely determine the po-sition of integration and use techniques suchas DNA microarrays,4 proteomics,13 and gaschromatography coupled to mass spectrom-etry (GC/MS) analysis of metabolites to de-termine the effects of this random integra-tion on the metabolism of the plant.

4. It is possible that pieces of plant DNA aretaken up by the epithelial cells of the gas-trointestinal tract (GIT) of consumers.Hard data on this transfer are scarce, butthis is a general phenomenon of digestionand uptake of foods and not just an issuerelated to genetically modified plant mate-rial. However, particularly when antibi-otic-resistant markers are used in geneti-cally modified organisms (GMOs), it isnecessary to determine the probability ofthis transfer in order to make a good riskassessment. Of course, it would be muchbetter if antibiotic resistance markers werenot present at all.

Some Consumer Benefits of GeneticallyModified Microorganisms

As stated in the introduction to this chapter,the popularity of fermented food is based on sev-

eral aspects, of which the enormous variation inappearance, flavor, and texture is the most im-portant for Western consumers. This enables theconsumer to select a product out of this enor-mous range that is close to personal preferenceand not an "average" product. When fermentedfoods are made from plant materials that do nothave an optimum composition from a nutritionalpoint of view, such as cassava, fermentationcontributes to (partial) supplementation of thesecomponents (e.g., amino acids such as methion-ine, lysine, and tryptophan; vitamins, in particu-lar A and B vitamins; minerals). Fermentationand the accompanying physical processes alsocontribute to the elimination of undesirablecomponents that are present in plants2 (seeChapter 4). In general, fermented products alsohave a better microbiological stability, which isa considerable consumer benefit, especially indeveloping countries with less well-organized(chilled) distribution chains and often no refrig-erator in the homes (see Chapter 2).

On top of the general benefits for the consumerdescribed above, certain microorganisms used infermentation processes, particularly lactic acidbacteria (LAB), may provide additional healthbenefits (Exhibit 10-1). Recombinant DNA tech-nology may be able to improve these positive as-pects of fermented food products. It is even pos-sible to extend this range of potential healthbenefits to consumers, for example, by oral immu-nization against pathogens and/or toxins.31

Some Potential Hazards and Risks ofGenetically Modified Microorganisms orTheir Enzymes Used in theManufacturing of Fermented Foods

There are several aspects related to the safetyof fermented foods. Assuming that extensive re-search and risk assessment have proved that thenewly introduced gene produces an intrinsicallysafe protein and that this protein does not con-tribute to the biogenesis of any harmful compo-nent, either during fermentation or during diges-tion of the fermented product, the remainingsafety aspects are:

Exhibit 10-1 Potential Consumer Benefits ofFermented Foods

• Improved appearance, flavor, and texture• Conversion of antinutritional and toxic com-

pounds (e.g., removal of cyanogenic glyco-sides in cassave)

• Increased microbiological keepability• Enrichment with amino acids, minerals, and

vitamins• Agonistic action against enteric pathogens*• Improved lactose utilization (important for

developing countries)*• Conversion of potential precarcinogens into

less harmful compounds*• Stimulation of the mucosal immune systemf

*Mainly related to foods fermented with lacticacid bacteria.

fOnly a small number of lactic acid bacteria mayhave this property.

• Can the transfer of newly introducedgene(s) to recipient microorganisms occurin the fermentation process or after con-sumption of the product in the GIT of con-sumers?

• Can the transfer of newly introducedgene(s) to recipient cells of the GIT of con-sumers occur?

• Can the transfer of newly introducedgene(s) to recipient microorganisms in theenvironment occur?

The risk assessment of the aspects mentionedunder the first and third bullet have been de-scribed in some detail earlier.30 As the potentialfor transfer of genetic information from donormicroorganisms to recipient microorganisms isa matter of concern, and new data are available,this will be discussed in some detail using thedecision tree provided in Figure 10-7.

Provided that the microorganism used in a fer-mentation process is intrinsically safe and therecombinant version of this microorganism issubstantially equivalent and free of any DNA ofthe marker free (clean) host organisms, the risk

related to the use of these recombinant DNA mi-croorganisms during fermentation processes iszero for both consumer and environment.30

When, during the fermentation process, achemical or physical treatment is applied that Iy-ses the genetically modified microorganism anddestroys the functionalities of the genome, it isnot necessary to take into account the aspect ofpotential transfer of genetic material from thesemicroorganisms to recipient cells in the GIT ofconsumers or safety aspect (third bullet above).

TWO SETS OF EXAMPLES OF SAFETYEVALUATION OF FERMENTEDFOODS FOR WHICH RECOMBINANTDNA IS USED

Supply Chain for Soy Sauce Produced fromGenetically Modified Soybeans and/orWheat and/or Using Genetically ModifiedMicroorganisms and/or EnzymesProduced by Genetically ModifiedMicroorganisms

There are a large number of fermented prod-ucts that are derived from wheat and/or soybean(see Chapter 1). For the supply chain risk/ben-efits analysis, the well-known soy sauce wasused as an example. A general outline of thebenefits and risks of genetically modified soy-beans was provided on pages 223 and 227, andfor microorganisms and enzymes used in the fer-mentation process, on pages 228. This outlinewill now be worked out in detail for soy sauce. Asimplified scheme of the soy sauce process isdepicted in Figure 10-8.

The safety of soy sauce that is derived from ge-netically modified soybean will be analyzed as afunction of the supply chain. As pointed out onpage 227, a genetically modified soybean shouldnot have an antibiotic resistance marker becausethis will increase the risk and is of no benefit to theconsumer. For this example, it is assumed that thegenetic modification of the soybean results in afaster fermentation process due to the incorpora-tion of enzymes in the soybean and/or wheat thatcontribute in the first steps of the process. Thefaster process is assumed to also contribute to thesensory quality of the soy sauce. Therefore, the

benefits for the consumer will be better quality anda price reduction.

• Step Ia (Figure 1(M). The risk related tothe first step of the supply chain (cultivationof the soybeans): Because the newly intro-duced gene encodes an intrinsically safe pro-tein that does not provide any receiving or-ganism with an ecological advantage, boththe environmental and consumer health risksare zero. It should be noted, however, that theenvironmental hazard is not zero. Afterstudying the frequencies of spontaneous hy-bridization between oilseed rape (Brassicanapus) and weedy rape (B. campestris)Jorgensen & Anderson11 concluded thattransgenes from oilseed rape may be pre-

served for many years, and that weedy B.campestris with transgenes may present eco-nomic (ecological) risks to farmers andplant-breeding companies or biochemical in-dustries. Although their findings of sponta-neous hybridization have been confirmed inother studies, unfortunately, they mixed uprisk and hazard. Only when the transgene re-sults in a clear ecological advantage for therecipient is there a hazard, and, based on theirstudies, a realistic probability (> 1 0~8) that thehazard occurs (= risk). Moreover, the transferof genetic material has happened for millionsof years and has not resulted in ecologicalproblems because the transgene did not givethe recipient an advantage under a large num-ber of ecological conditions.

1(E). Generally recognized as safe ge-netically modified microorganismused in fermentation and still livingin the final product

2(Q). Encode the newly introduced genefor an intrinsically safe product?

3(Q). Is the new gene stable integrated inthe genome of the host cell?

4(Q). Is an antibiotic resistance markerused during construction of theGMO?

5(Q). Does the newly introduced gene en-code for a protein that may resultunder certain conditions in an eco-logical advantage for the host cell?

6(A). Carry out full risk assessment.7(Q). Does the newly introduced gene en-

code for a protein that after transfermay result under certain conditionsin an ecological advantage to recipi-ent mammalian cells?

8(Q). Does the newly introduced gene en-code for a protein that after transfermay result under certain conditionsin an ecological advantage to recipi-ent microbial cells?

9(A). Integrate new gene, preferably onpredetermined locus on the chromo-some of the host.

10(Q). Is the antibiotic resistant markergene eliminated?

(E). Entry or end; Diamonds(Q = questions); Y = yes, N = No

Figure 10-7 General scheme to determine safety to consumers and some environmental consequences of geneti-cally modified microorganisms used in fermentation processes. Note: This is a modification of the schemespublished in Verrips and Vandenberg.30

STOP

STOP

STOP

No safety orecological hazard

It is also important to consider the stabil-ity of plant DNA in soil because that DNAcan be picked up by microorganisms thatare present in the soil. A study showed that"survival" of a neomycin/kanamycin resis-tance gene from transgenic tobacco was0.14% after 120 days in soil. Most prob-ably, the persistence is so high because ofstabilization of the DNA by soil.26 Becausethe stability of DNA in soil is higher thanwas assumed in the past, the probabilitythat DNA is taken up by soil organisms isalso higher. Indeed, various studies haveshown that horizontal transfer of plantDNA to microorganisms can occur, but atlow frequencies. For instance, based on anumber of well-controlled experiments, thetransfer of DNA from transgenic potato tothe plant pathogenic bacterium Erwiniachrysanthemi under natural conditions iscalculated to be 10~17, which is much lowerthan the detection limit of approximately10-12 24 Eariier experiments showed that thetransfer of hygromycin resistance fromvarious transgenic plants (B. napus, B. ni-

gra, Datura innoxia, and Vicia narbonen-sis) to the fungus Aspergillus niger aftercocultivation occurred with an unexpect-edly high frequency.10

• Step Ib (Figure 10-4). In cases where thenewly introduced gene encodes a proteinthat can provide an ecological advantagefor any recipient organism, the risk is notzero and the risk assessment (Figure 10-7)should be performed before proceeding tostep 2.

• Step 2a (Figure 10-4). The risk related tothe second step of the supply chain (fer-mentation process): During this process,the soybean is physically denatured; more-over, the chemical composition in certainparts of the process is very hostile for or-ganisms, with the exception of some func-tional organisms.

• Step 2b (Figure 10-4). In the cases that thesoy sauce process contains geneticallymodified microorganisms, it is very likelythat the DNA of these microorganisms willalso be destroyed during the final steps ofthe process when the conditions are quite

Figure 10-8 Schematic outline of soy sauce process.

bottling soy sauce

pasteurization

infusion/flotation

fermentation (30-40 0C, NaC113%, 10 days)

yeast/LAB seedcultures

hydrolyses (40-50 0C, NaCI 8%, 10 days)

mold seed culturekoii

autoclaving

sugar/salinesolution

liquefactionsaccharification

syrup

wheat flour/rice brinedef. soy bean

soaking in water

wheat bran

brine

harsh (but hard evidence is missing). It isagain highly recommended to avoid thepresence of any nonfunctional foreigngenes or genes encoding antibiotic resis-tance properties in these microorganisms.The hazard of the newly introduced geneproduct should be analyzed in the sameway as described in step Ia for new genesintroduced in soybeans and/or wheat.

• Step 2c (Figure 10-4). In the case that en-zymes produced by genetically modifiedmicroorganisms are used in the soy sauceprocess, the regulations require that the en-zyme should be intrinsically safe and freeof DNA of the microorganism involved.29

Practice proves that this requirement can befulfilled without much difficulty in indus-try. Even though there is evidence that en-zymes can be made free of any DNA, as ad-vocated earlier, it is highly recommendedto produce enzymes only with food grademicroorganisms that are free from non-functional foreign genes, in particular, freefrom genes encoding antibiotic resistanceproperties. The enzyme as such should be"substantially equivalent" to the wild typeenzyme, another requirement that can bemet by industry. Consequently, using re-combinant DNA enzymes in the soy sauceprocess does not pose any hazard and there-fore no risk to consumer and environment.

• Steps 3-5 (Figure 10-5). In all of the sub-sequent steps of the supply chain, includingthe step in which consumers introduced the(digested) soy sauce in the environment,the risk related to the use of geneticallymodified soybean and/or wheat or the useof genetically modified microorganisms orenzymes produced by recombinant DNAtechnology remains zero.

Supply Chain for Cheeses Produced withGenetically Modified LAB and/orEnzymes Produced by GeneticallyModified Microorganisms

As stated in the introduction, fermented prod-ucts, particularly fermented milk products, have

a healthy image. Although a number of studiesstill have to be carried out to substantiate the po-tential health benefits of LAB, in particular, theeffects of LAB and their products on mucosalhealth, there are already a large number of publi-cations that support these benefits for consum-ers.23 The most important potential health andother benefits are summarized in Exhibit 10-1 .

Some of the main aspects of cheese processesare provided in Figure 10-9, and the risk assess-ment of that process and the other steps in thesupply chain will be performed similarly to theassessment for soy sauce.

• Step 1 (Fig 10-4). Traditional milk will beused for cheese production; therefore, a riskassessment of GMO-related issues in thisstep is not necessary.

• Step 2a (Fig 10-4). In this step, geneticallymodified LAB are introduced in the cheeseprocess. As the safety of the end product(cheese) for consumers will be discussed instep 4, only the environmental safety aspectsare discussed in this step. It is unrealistic toconsider the cheese fermentation process as acompletely closed process; therefore, there isa probability that genetically modified LABwill enter the environment. Assuming theseLAB do not contain any foreign, nonfunc-tional genes, and that the newly introducedgene encodes an intrinsically safe protein, thehazard will be zero and consequently the riskwill also be zero.

If, however, the LAB contain a gene-en-coding antibiotic resistance or a propertyproviding an ecological advantage to therecipient, the environmental hazard is notzero and a formal risk assessment as dis-cussed in detail in an earlier publication30

should be performed. More recent data15'18

support the view that horizontal transfer ofgenetic material from one microorganismto another, either directly by conjugation ortransformation or indirectly via transduc-tion, occurs at a much higher frequencythan was assumed in the mid-1980s.

• Step 2b (Figure 10-4). In this step, en-zymes made by recombinant DNA technol-

ogy are introduced.27 As described in thesoy sauce example, an intrinsically safe en-zyme produced under proper conditionsdoes not pose a risk to consumer or envi-ronment.

• Step 3 (Figure 10-5). During distribution,the probability of the escape of LAB intothe environment is extremely low, and ashorizontal transfer also occurs with low fre-quencies, this probability is considered tobe effectively zero.

• Step 4 (Figure 10-5). In this step, the con-sumer will be in direct contact with the ge-netically modified LAB and/or the enzyme.For the enzyme, a relatively straightfor-ward procedure has to be followed to provethat the enzyme is intrinsically safe, origi-nates from a well-known source, and issubstantially equivalent to the wild type en-zyme.29 If that is the case, the enzyme willnot pose a hazard and consequently no risk.For the genetically modified LAB, a formaldecision scheme has been developed.30

Two aspects should be analyzed: the gen-eral safety of the fermented food productmade with genetically modified LAB andthe probability of transfer of genetic infor-mation from the LAB to other bacteria inthe GIT of consumers (Figure 10-10). Inthis figure, the following aspects related toa quantitative risk assessment are consid-ered: (1) probability of transfer of geneticmaterial from the GMO to other microor-ganisms in the GIT as a function of the resi-dence time of the GMO in the GIT, (2)whether the GMO remains alive, (3) whetherDNA from lysed GMOs are taken up by nor-mal inhabitants of the GIT, (4) the probabil-ity that transfer of genetic information of theGMO to normal inhabitants results in an eco-logical advantage of the transformed inhabit-ants and, if so, (5) whether that may pose ahealth risk to the consumer, or environmentalrisks. Unfortunately, insufficient data areavailable to carry out such risk assessment inthe proper way.

Brie StiltonCamembert RoquefortCoulommier Blue

Gorgonzola

Figure 10-9 Schematic outline of milk fermentation processes and their cheese products.

Internal mold-ripenedSurface ripened

CheddarEmmenthalParmesanPort du salut

BrickEdamGouda

LimburgNeufchatel

Pressed Not pressedPressed

Curd scalded Curd scalded Curd not scalded

Part. ferm. milk, treated with rennetFermented and treated with rennet

Milk

I (E). Genetically modified lactic acid bacterium (GMO)2(A). Determine the distribution of the residence time of GMO in the gastrointestinal tract (GIT) of consumers. Take the time

corresponding with 95% of this distribution curve as t(r).3(Q). Will the GMO lyse with P(b)> b in the GIT within t(r)?4(A). Determine the probability P(c) that intact cells of the GMO transfer genetic information to normal inhabitants of the GIT.

In these studies, use t(r) as contact time and the conditions of the GIT.5(Q). Is P(C)XX6(A). Determine the probability P(f) that DNA originating from lysed GMO transform normal inhabitants of the GIT (resulting

in transformed inhabitants).7(Q). Is P(f)>f?8(A). Determine whether the transformed inhabitants obtain an advantage over untransformed inhabitants in the GIT: A(i).

Define A(i) in either faster growth rates t'(g), better adhesion to epithelial cells h', or higher production of certain metabo-lites {p(x}', x= 1,....}.

9(A) Determine the probability P(e) that any of the events described under action 8 will result in the formation of a hazardous(transformed inhabitant produce toxin or that will replace beneficial microorganisms in the GIT) microorganism.

10(Q). Is {P(c) + P(f)}*P(e) > e?I1 (A). Determine also the probability P(d) that the GMO will be transformed by genetic material originating from the common

GIT microorganisms (=modified GMO).12(A). Determine whether the modified GMO gains an advantage over untransformed GMO: B(i). Define B(I) in either faster

growth rates t"(g), better adhesion to epithelial cells h", or higher production of certain metabolites {p(x)", x=1,....}.13(Q). \sP(d)>d?14(A). Determine the probability P(g) that any of the events described under action 12 will result in the formation of a hazard-

ous GMO.15(Q). \sP(d)*P(g)>g?16(E). This risk is unacceptable and the GMO should not be released.17(E). This risk is acceptable and the GMO can be released.A = action; E = entry or end; Q = question

Figure 10-10 A proposal for a structured assessment of the risk related to the introduction of genetically modi-fied lactic acid bacteria in (or as) food products.29

In fact, only fermented products ofwhich the safety has been demonstrated,ecological hazards have been proven to bezero, and consequently, the risk is zero,should be approved. However, assumingthat the ecological hazard is not zero, whichmeans that the newly introduced gene con-fers an ecological benefit on recipient cells,both risk assessments should result in anacceptably low risk. This can only beachieved if the probability of transfer of ge-netic information from the donor to recipi-ent cells is very low. For horizontal transferof genetic material between microorgan-isms, conjugation, transformation, andtransduction show a higher probability forplasmid DNA than for chromosomal DNA.Therefore, it is recommended that the for-eign DNA be integrated (preferably on apreknown locus) into the chromosome ofthe microorganism. Integration atpreknown loci therefore forms an impor-tant component of the risk cube presentedin Figure 10-3.

A parameter that is essential for properrisk assessment of events in the human GITis the residence time of genetically modi-fied microorganisms in the GIT, whetherthey lyse, whether they conjugate, orwhether the DNA liberated during lysis cantransform other cells, including humancells. Although the number of reliable datasets have increased during the last decade,they are still not sufficient to solve all of thequestions of the decision scheme given inFigure 10-10. Nevertheless, the data avail-able indicate that LAB can survive passagethrough the GIT.19 The functional microor-ganism in many cheese fermentations,Lactococcus lactis, survives up to threedays, although the survival rate is only 1-2%. Another noteworthy observation inthis study was that the PCR method coulddetect special DNA stretches of this bacte-rium for up to four days after ingestion,12

indicating that either the feces containednonviable L. lactis or that these stretches ofDNA were liberated during lysis of the bac-

terium. Experiments with phage M 13 DNAingested by mice showed that a small butmeasurable percentage of this DNAreached peripheral leukocytes, spleen, andliver via the intestinal wall mucosa andcould be covalently linked to mouseDNA.25 Even taking into account that theseexperiments are rather artificial, for aproper risk assessment, they have to beincluded.

From these studies, it can be concludedthat there is a probability that geneticallymodified microorganisms, including LAB,can transfer genetic information to other mi-croorganisms in the GIT. This information isanother strong argument to ban geneticallymodified microorganisms that contain an an-tibiotic resistance gene. In cases where anongenetically modified microorganism hassuch a gene, it should be deleted because fur-ther unnecessary spread of antibiotic-resis-tant genes creates very serious health and en-vironmental problems.

• Step 5 (Figure 10-5). The release of fecesof consumers into the environment is the fi-nal step to be assessed. If the newly intro-duced gene does not encode for a proteinthat provides an ecological advantage forthe recipient cells, then a risk assessment ofthis step is not necessary. If this is not thecase, then the decision scheme outlined inan earlier publication30 can be applied forthis step. However, this step may becomemore complex because it should take intoaccount whether the transfer of genetic in-formation from the LAB to recipient mi-croorganisms has occurred in the GIT ofconsumers and, if so, these modified recipi-ents should be evaluated as well.

In this example, LAB are the functional mi-croorganisms in the fermented product, but forother organisms that are generally recognized assafe (GRAS), such as Saccharomyces cerevi-siae, Kluyveromyces lactis, Penicillium roque-forti, P. camenberti, and A. niger, or A.awamori, the same hazard and risk assessmentcan be applied with similar results.

CONCLUSION

Fermented foods are considered by consumersas natural and healthy.7'8'14'22 The introduction ofgenetically modified plants, microorganisms, orenzymes into these products needs careful discus-sion with authorities and consumer organizations.Many fermented foods contain living microorgan-isms, and if these organisms are genetically modi-fied, one should consider them as a potentialsource of DNA for horizontal gene transfer be-cause a number of studies provide circumstantialevidence that, in rare cases, genes have been later-ally transmitted among Eubacteria, Archaea, andEukarya. However, extensive studies byPuhler's6 group showed that this transfer occursat extremely low frequency.

REFERENCES

1. Barra, D. & Simmaco, M. (1995). Amphibian skin: apromising resource for antimicrobial peptides. Tib Tech13,205-209.

2. Beuchat, L. R. (1983). Indigenous fermented foods. InBiotechnology. Vol. 5, pp. 477-528. Edited by H. -J. &G. Reed. Weinheim: Verlag Chemie.

3. Casteels, P. & Ampe, C. (1989). Apidaecins: antibacte-rial peptides from honeybees. Embo 8, 2387-2391.

4. Chervitz, S. A., Aravind, L., Sherlock, G., Ball, C. A.,Koonin, E. V., Dwight, S. S., Harris, M. A., Dolinski,K., Mohr, S., Smith, T., Weng, S., Cherry, J. M. &Botstein, D. (1998). Comparison of the complete proteinsets of worm and yeast: orthology and divergence. Sci-ence 282, 2022-2033.

5. Cornelissen, B. J. C., Hooft van Huijsduijnen, R. A. M.& BoI, J. F. (1986). A tobacco mosaic virus-induced to-bacco protein is homologous to the sweet-tasting proteinthaumatin. Nature 321, 531-532.

6. Droge, M., Punier, A. & Selbitschka, W. (1998). Hori-zontal gene transfer as biosafety issue: a natural phe-nomenon of public concern. J Biotech 64, 75-90.

7. Gasser, F. (1994). Safety of lactic acid bacteria and theiroccurrence in human clinical infections. Bull Inst Pas-teur 92, 45-667.

8. Gilliland, S. E. (1990). Health and nutrition benefits forlactic acid bacteria. FEMS Microbiol Rev 87, 175-188.

9. Hamstra, A. M. (1993). Consumer acceptance of foodbiotechnology. SWOKA Research Report 137, TheHague, The Netherlands.

In these discussions, the proven benefits to theconsumer should be explained in clear terms.The risk assessment should show that the safetyhazard related to the newly introduced gene(s) iszero, and consequently, the risk for consumer iszero. Whenever possible, without losing theconsumer benefit, the environmental hazardshould also be zero. If that is not possible, therisk assessment should show that the risk is ex-tremely low, preferably less than 10~8 per prod-uct unit produced. Consumer studies show thatprovided there is a clear benefit and risk is ab-sent, most consumers will accept these products.To gain the confidence of the consumer for fer-mented foods in which recombinant DNA tech-nology plays a role, clear and transparent label-ing and public relations will be required.

10. Hoffmann, T., GoIz, C. & Schieder , O. (1994). ForeignDNA sequences are received by a wild-type strain ofAspergillus niger after co-culture with transgenic higherplants. Curr Genet 27, 70-76.

11. Jorgensen, R. B. & Andersen, B. (1994). Spontaneoushybridisation between oil seed rape (Brassica napus)and weedy B. campestris: a risk of growing geneticallymodified oil seed rape. Am J Botany 81, 1620-1626.

12. Klijn, N., Weerkamp, A. H. & de Vos, W. M. (1995).Biosafety assessment of the application of geneticallymodified Lactococcus lactis spp. in the production offermented milk products. System Appl Microbiol 18,486^192.

13. Koonin, E. V., Tatusov, R. L. & Galperin, M. Y. (1998).Beyond complete genomes: from sequence to structureand function. Current Opinion in Structural Biology 8,355-363.

14. Lee, Y. -K. & Salminen, S. (1995). The coming of age ofprobiotics. Trends Foods Sd Technol 6, 241-245.

15. Lorenz, M. G. & Wackernagel, W. (1994). Bacterialgene transfer by natural gene transformation in the envi-ronment. Microbiol Rev 58, 563-602.

16. Mann, C. C. (1999). Crop scientists seek a new revolu-tion. Science 283, 310-314.

17. Mann, C. C. (1999). Genetic engineers aim to soup upcrop photosynthesis. Science 283, 314-316.

18. Miller, R. V. (1998, January). Bacterial gene swappingin nature. Sd Am 46-5 1 .

19. Pedrosa, M. C., Golner, B. B., Goldin, B. R., Barakat, S.,Dallal, G. & Russel, R. (1995). Survival of yoghurt-con-taining organisms and Lactobacillus gasseri (ADH) andtheir effect on bacterial enzyme activity in the gas-trointestinal tract of healthy and hypochlorhydric eld-erly subjects. Am J Clin Nutr 61, 353-358.

20. Ronald, P. C. (1997, November). Making rice diseaseresistant. SdAm 68-73.

21. Rowland, L, Wiseman, H., Sanders, T., Adlercreuz, H.& Bowey, E. (1999). Metabolism of oestrogens andphytooestrogens: role of the gut flora. Biochem SocTrans 27, 304-308.

22. Salminen, S., Bouley, C., Bouron-Ruault, M. -C.,Cummings, J. H., Franck, A., Gibson, G. R., Isolauri, E.,Moreau, M. -C., Roberfroid, M. & Rowland, I. (1998).Functional food science and gastrointestinal physiologyand function. Br J Nutr 80, S147-S171.

23. Salminen, S., Isolauri, E. & Salminen, E. (1996). Clini-cal uses of probiotics for stabilizing the gut mucosal bar-rier: successful strains and future challenges. Ant vanLeeuwenhoek 70, 347-358.

24. Schliiter, K., Futtere, J. & Potrykus, I. (1995). Horizon-tal gene transfer from a transgenic potato line to a bacte-rial pathogen (Erwinia chrysanthemi) occurs — if atall — at extremely low frequency. Biotechnol 13, 1094-1098.

25. Schubbert, R., Renz, D., Schmitz, B. & Doerfler, W.(1997). Foreign (M13) DNA ingested by mice reachesperipheral leukocytes, spleen, and liver via the intestinalwall mucosa and can be covalently linked to mouseDNA. Proc NatAcadSci USA 94, 961-966.

26. Stiles, M. E. (1996). Biopreservation by lactic acid bac-teria. Ant van Leeuwenhoek 70, 33 1-345.

27. Teuber, M. (1990). Production and use of chymosinfrom genetically altered microorganisms. LebensmittelIndMilchwirtschaftlS, 1118-1123.

28. Verrips, C. T. (1989). Growth of microorganisms incompartmentalized products. In Mechanisms of Actionof Food Preservation Procedures, pp. 363—393. Editedby G. W. Gould. London: Elsevier Applied Science.

29. Verrips, C. T. (1995). Structured risk assessment ofrDNA products and consumer acceptance of these prod-ucts. In Biotechnology. Vol. 12, pp. 157-196. Edited byH. -J. Rehm & G. Reed. Weinheim, Denmark: VCHVerlaggesellschaft mbH.

30. Verrips, C. T. & Van den Berg, D. J. C. (1996). Barriersto application of genetically modified lactic acid bacte-ria. Ant van Leeuwenhoek 70, 299-3 16.

31. Wells, J. M., Robinson, K., Chamberlain, L. M.,Schofield, K. M. & LePage, R. W. F. (1996). Lactic acidbacteria as vaccine delivery vehicles. Ant vanLeeuwenhoek 70, 317-330.