Food Safety and Biotechnology

Michael W Pariza. Cambridge World History of Food. Editor: Kenneth F Kiple & Kriemhild Conee Ornelas, Volume 2, Cambridge University Press, 2000.

The ready availability of safe, wholesome food is often taken for granted by citizens of modern societies. However, maintaining the safety of a large, diverse food supply is a challenging undertaking that requires coordinated effort at many levels. In this chapter, the principles of food safety are discussed first with regard to traditional foods and then again as they concern novel foods developed through genetic modification.

Definitions and Priorities

The term “food safety,” as used today, encompasses many diverse areas, including protection against food poisoning and assurance that food does not contain additives or contaminants that would render it unsafe to eat. The term evolved mainly in the context of preventing intoxication by microbial poisons that act quickly (within hours to a day or two of exposure) and often induce such serious symptoms as convulsive vomiting and severe diarrhea, or respiratory failure and death (Cliver 1990).An example of the former is staphylococcal food poisoning, caused by the proteinaceous enterotoxins of the pathogenic bacterium Staphylococcus aureus (Cliver 1990). Botulism is an example of the latter, caused by the neurotoxins synthesized by Clostridium botulinum (Cliver 1990).

Both staphylococcal food poisoning and botulism result from ingesting toxins that are preformed in the implicated foods. For illness to ensue, it is necessary only to ingest the toxin, not the microbe itself. Food poisoning may also follow the ingestion of certain pathogenic bacteria, such as Salmonella, which produce gastrointestinal (GI) infections, and Escherichia coli 0157:H7, which produces an infection and a potent toxin within the GI tract (Cliver 1990). The infection results in bloody diarrhea, while the toxin enters the bloodstream and induces kidney damage.

Given the foregoing, it should come as no surprise that the Food and Drug Administration (FDA), the lead federal agency charged with ensuring the safety of the food supply in the United States, considers foodborne microbial pathogens and the toxins they produce to be the most important of the various food safety risks (Table VII.8.1). Unfortunately, evidence indicates that the public often tends to focus too much on pesticide residues and food additives, wrongly believing that the risk from these materials is as great as that from pathogenic microorganisms (Pariza 1992c).

Table VII.8.1. Ranking food safety risks Source: Young (1989).

1. Food-borne disease.
2. Environmental contaminants.
3. Naturally occurring toxins.
4. Food additives.
5. Pesticide residues.

Table VII.8.2. Summary of reported food-borne disease outbreaks in the United States, 1983-7a Source: Bean et al. (1990).

Percentage Percentage
Etiologic agent of cases of outbreaksb
Pathogenic bacteriac 92 66
Virusesc 5 5
Parasitesc <1 4
Chemical agentsd 2 26

a During this time period, 91,678 cases and 2,397 outbreaks of food-borne illness were reported to the Centers for Disease Control (CDC). The etiologic agent was identified in 38 percent of the outbreaks. These data are summarize in this table.

b An outbreak is the occurrence of two or more cases of illness transmitted by a single food.

c Pathogenic bacteria, viruses, and parasites are all classified as microorganisms.

d Some instances of chemical poisoning were actually due to chemical toxins produced by microorganisms. No instances of illness traced to synthetic chemicals were identified in the report.

Microbial Foodborne Illness

Table VII.8.2 shows a summary of foodborne illness in the United States, compiled by the Centers for Disease Control (CDC), from 1983 to 1987 (Bean et al. 1990). But according to both the CDC and the FDA, the reported cases and outbreaks account for only a tiny fraction of the total foodborne illnesses that occurred during that period. It is estimated that several million cases of microbial food poisoning actually occur in the United States each year (Archer and Kvenberg 1985).

Foodborne illness is caused by a relatively small number of bacterial species. All raw foods contain microorganisms, some of which may be pathogens. Whether or not this becomes a problem is determined largely by how the food is handled prior to consumption. Pathogen growth depends on a number of factors. The most important of these are (1) the levels of salt and other electrolytes present in the food; (2) the types and amount of acid in the food; (3) the amount of available water; (4) the presence of microbial inhibitors (both naturally occurring and synthetic); (5) storage conditions; and (6) cooking conditions. Safe food processing depends on understanding these factors and effectively utilizing the basic principles of microbiology as they apply to foods.

In addition to precipitating illness and soaring medical costs, foodborne illness can also bring economic devastation for producers of an implicated product (Todd 1985).A dramatic example occurred in 1982, when two Belgians developed botulism after sharing a can of Alaskan salmon. One of them subsequently died of the illness.

The problem was traced to a can press that was used at the processing plant to form the can in question. The press was defective and had produced a tiny rupture in the wall of the implicated can. The rupture, which was virtually invisible, nonetheless permitted a small amount of nonsterile water to be drawn into the can during the cooling step that followed heat processing. The water was contaminated with spores of Clostridium botulinum which, once inside the can, germinated and produced the toxin that induced the illness. Fears that other cans might have been similarly damaged prompted voluntary recalls of much of the canned Alaskan salmon produced that year. In all, 60 million cans were recalled and discarded (Hayes 1983).

Foods provide a medium for the growth of specific pathogens, but they can also be passive carriers for virtually any pathogen capable of producing infection via the oral route. In this case the pathogens do not actually have to grow in the food. They need only to survive. Good examples are the foodborne viruses, such as the virus that causes hepatitis A.These cannot grow in the foods that carry them, yet illness follows ingestion of the contaminated food.

It should be noted that food spoilage and food safety are not related in a direct manner. A food may be spoiled but nonetheless safe. Alternatively, however, a food can contain bacterial toxins at deadly levels yet appear perfectly “normal.” And finally, it is important to recognize that not all microbes in food are bad. Rather, harmless microorganisms are an integral part of many traditional foods. One could not produce bread, cheese, yoghurt, soy sauce, sauerkraut, beer, or wine without specific yeasts, bacteria, and molds. Such microorganisms are needed to induce desirable changes in the food through a process called “fermentation.” Besides providing characteristic flavors, fermentation is also an important, effective method of food preservation. The harmless fermentation microorganisms produce microbial inhibitors and otherwise change the ecology of the food in ways that discourage subsequent pathogen growth.

Risks from Synthetic Chemicals

The data summarized in Table VII.8.2 indicate that the risk of developing foodborne microbial illness is far higher than the risk of becoming ill because of exposure to toxic chemicals in food. The instances of chemical poisoning reported in Table VII.8.2 represent almost exclusively naturally occurring toxins or contaminants. Only one approved food additive (monosodium glutamate) was reported as possibly causing illness in two outbreaks that occurred in 1984, involving a total of seven persons. In no instance was a synthetic chemical identified in the CDC report as a cause of illness (Bean et al. 1990).The data summarized in Table VII.8.2 refer to clinical illness, almost all of which is acute (that is, there is only a short time interval between exposure and onset of symptoms). Of course, questions about risks from “chemicals” in food often center on chronic effects, such as cancer, which in humans takes many years to develop.

The causes of chronic illness are much more difficult to determine because of the long time intervals between initial exposure and the onset of symptoms. Nonetheless, there is an extensive toxicology data base against which hypotheses can be tested. For example, based on these data, one may conclude that it is highly unlikely that synthetic pesticide residues or food additives pose a cancer risk to consumers (Ames, Profet, and Gold 1990; Pariza 1992a, 1992b).

Risks of Naturally Occurring Pesticides

Plants lack the defenses that animals possess (mobility, claws, and teeth), and so they must rely on passive defenses that are effective in situ. Accordingly, evolution has provided plants with the ability to synthesize naturally occurring toxins. In a very real sense, plants are practitioners of chemical warfare against potential predators.

The natural toxins that plants produce are usually nonspecific in nature; that is, they act against a variety of potential pests, including mammals as well as insects. The pathways regulating the synthesis and expression of these toxins tend to be under general control, meaning that any number of stresses, from insect infestation to drought, can “turn on” the entire spectrum of chemical and biological defenses that a plant possesses.

In this context, crop breeding can be seen as a means of controlling unwanted wild traits, such as toxin production, in plants destined for the table. Experience has shown that this is a relatively easy matter, even without any knowledge of the underlying science, doubtless because toxin production is not an inherent part of the growth and reproductive processes. Humans select and artificially propagate variants that would not otherwise survive but that nonetheless exhibit desirable traits. Because of the genetic linkage within the generalized protective mechanism pathways, increased expression of traits that humans deem desirable will usually be accompanied by decreased toxin expression (International Food Biotechnology Council 1990).

Unfortunately, selected traits that make a plant desirable as human food also make it desirable to insects. This, in fact, is why synthetic pesticides are used: They replace the naturally occurring pesticides and related survival traits that have been bred out of food plants.

Yet, although the natural pesticide levels of food plants have been lowered considerably through breeding, they have not been totally eliminated. Rather, the data in Table VII.8.3 show that many commonly consumed food plants and plant products contain naturally occurring pesticidal carcinogens. Moreover, unlike synthetic chemicals, which must be rigorously tested before regulatory agencies will approve them for food-related use, the toxins made naturally by plants are not ordinarily subjected to any sort of testing (Ames et al. 1990). For some individuals, however, it is unsettling to learn that many of those substances that have been tested have proved to be carcinogenic in rat or mouse feeding studies.

Yet it is important to put these findings into perspective by a consideration of (1) the relative exposure of the public to natural versus synthetic carcinogenic pesticide residues, and (2) the relative risk of cancer from these exposures.

With regard to exposure, FDA data indicate that the average daily intake of synthetic pesticides per person in the United States is 0.09 milligrams (mg) (Ames et al. 1990). About half of this intake (0.04 mg) comes from four chemicals that were not carcinogenic in high-dose rodent feeding studies. Thus, assuming that all of the remaining exposure is to carcinogenic pesticides (an unlikely assumption), the maximum exposure to synthetic carcinogenic pesticide residues is 0.05 mg per person per day.

Table VII.8.3. Some natural pesticidal carcinogens in food Source: Ames, Profet, and Gold (1990).

Rodent carcinogen Plant food Concentration,
5-/8-methoxypsoralen Parsley 14
Parsnip, cooked 32
Celery 0.8
Celery, new cultivar 6.2
Celery, stressed 25
þ-hydrazinobenzoate Mushrooms 11
Glutamyl r-hydrazinobenzoate Mushrooms 42
Sinigrin (allyl isothiocyanate) Cabbage 35–590
Collard greens 250–788
Cauliflower 12–66
Brussels sprouts 110–1560
Mustard (brown) 16000–72000
Horseradish 4500
Estragole Basil 3800
Fennel 3000
Safrole Nutmeg 3000
Mace 10,000
Pepper, black 100
Ethyl acrylate Pineapple 0.07
Sesamol Sesame seeds (heated oil) 75
α-methylbenzyl alcohol Cocoa 1.3
Benzyl acetate Basil 82
Jasmine tea 230
Honey 15
Catechol Coffee (roasted beans) 100
Caffeic acid Apple, carrot, celery, cherry, eggplant,
endive, grapes, lettuce, pear, plum, potato 50–200
Absinthe, anise, basil, caraway, dill, marjoram,
rosemary, sage, savory, tarragon, thyme >1000
Coffee (roasted beans) 1800
Chlorogenic acid (caffeic acid) Apricot, cherry, peach, plum 50–500
Coffee (roasted beans) 21600
Neochlorogenic acid (caffeic acid) Apple, apricot, broccoli, brussels sprouts,
cabbage, cherry, kale, peach, pear, plum 50–500
Coffee (roasted beans) 11600

By comparison, it is estimated that Americans consume 1.5 grams of natural pesticides per person per day, or about 15,000 times more than the amount of synthetic pesticides. Moreover, many of these natural pesticides are also carcinogenic in high-dose rodent feeding studies (Table VII.8.3). Clearly, the public is exposed to far higher levels of naturally occurring pesticides, and, in fact, when compared to naturally occurring pesticides, synthetic pesticide exposure is trivial (Ames et al. 1990).

Moreover, toxicology is not just about exposure levels to toxic substances, because animals, including humans, have numerous biochemical and physiological resources to defend against such exposure. These defenses include enzymes that detoxify toxins and repair toxin-induced damage to DNA and other important biological structures, as well as the continual shedding of cells at epithelial surfaces that are most exposed to toxic insult (that is, cells at the surface of skin, eyes, and gastrointestinal tract). In short, humans have layer on layer of protection against the toxic dangers of “natural” environments.

Because of the immense number and variety of potential toxins, the defense systems must by necessity be general in nature. In this regard, there is nothing unique about synthetic pesticides. The protective mechanisms that act against naturally occurring toxins also act against synthetic toxins. These protective mechanisms are, of course, even more effective when the dose is exceedingly low as is the case with exposure to synthetic pesticide residues on produce. Moreover, it is important to note that the biological effects produced in animals by naturally occurring pesticides and other potential toxins are not always adverse. Rather, paradoxically, many of the substances shown in Table VII.8.3 that cause cancer when fed to rodents at excessively high levels seem to protect against cancer when fed at the lower, more realistic levels encountered in a normal diet (Pariza 1993).

One implication of this paradox is that unrealistically high dose cancer tests have been responsible for indicting the naturally occurring substances shown in Table VII.8.3. In fact, this consideration has led many experts to reexamine the entire framework through which we detect and regulate “carcinogens” (Ames et al. 1990; Pariza 1992c; National Research Council 1993).

The Safety of Novel Foods

There are many kinds of novel foods. In the simplest sense, a food may be novel in one locale and a dietary staple in another. Examples include the many foods that crossed the Atlantic Ocean between Europe and the Americas in the years following Columbus’s voyages (Goldblith 1992). Corn, potatoes, tomatoes, and cacao (chocolate), which today are common throughout the world, were novelties for everyone on earth, save the Native Americans, just 500 years ago.

At the other end of the spectrum are novel foods developed by using recombinant DNA technology (for example, Calgene’s flavr savr tomato).The recombinant DNA process is defined as “cutting and recombining DNA molecules to remove segments from or otherwise modify an organism’s genetic material, or to combine segments of DNA from different types of organisms” (International Food Biotechnology Council 1990: ).

Discovered and refined only since the 1970s, recombinant DNA methods represent a revolutionary means of controlling the introduction of new traits into plants, animals, and microorganisms. It should be stressed, however, that the “revolution” is not the process itself. There is ample evidence that gene transfer between different species has occurred throughout the history of life on this planet (International Food Biotechnology Council 1990). Rather, the revolution is that humans are able to control the process. For the first time in history, it is possible to direct the transfer of DNA from one species to another. It is also possible to create entirely new genes that, seemingly, have not existed before and transfer them into living hosts where the new genes will then replicate along with the host’s genetic material.

The incredible power of this technology has led some to conclude that to prevent misuse, the process itself must be carefully regulated. Others argue that the process is simply a scientific tool and that the focus of regulation should be on the products of biotechnology, not on the process.

With specific regard to food, most experts agree that the focus of safety evaluation should be on the novel foods themselves, and not on the process whereby novel foods are developed (International Food Biotechnology Council 1990; Kessler et al. 1992).The rationale for this reasoning is that recombinant DNA technology is no less safe than plant breeding or any other traditional means of genetic manipulation. In all cases, one is (or should be) concerned with the expression of genetic information.

For example, in traditional plant breeding, one should determine if unexpected traits that bear on safety are expressed by the new cultivar. Such unexpected traits may arise directly from the genetic material itself that has been transferred in the cross may result because the transferred genetic material affected a regulatory element, thereby inducing the expression of an undesirable new trait (for example, the expression of a toxin).

In any attempt to establish the safety of a novel food, the same procedures should be followed to assure that the gene being transferred produces a safe product and that the gene does not induce changes in the recipient that bear on health and safety issues.

Addressing the first concern (whether the gene that is being transferred produces a safe product) is a relatively straightforward matter. This is because the gene can be completely defined. If the gene codes for a protein, it, too, can be completely defined and studied directly for toxigenic potential.

The second concern (whether the transferred gene induces changes in the recipient that bear on health and safety issues) is somewhat more problematic, as it is with traditional methods of genetic modification.

For example, the data of Table VII.8.3 indicate that many food plants naturally produce potentially carcinogenic compounds. As already discussed, the expression of such natural pesticides does not occur at a constant rate irrespective of other considerations. Rather, these substances are under genetic control, which in turn is influenced by environmental factors, such as insect infestation and drought. A pertinent example shown in Table VII.8.3 concerns celery.

Celery produces a class of natural pesticides called psoralins, which are light-activated mutagenic carcinogens. Ordinarily, the concentration of psoralins in celery is about 0.8 parts per million, which is too low to produce evidence of adverse effects in persons who consume the vegetable. However, through the use of traditional breeding methodology, an insect-resistant celery cultivar was developed that contained about 10 times the concentration of psoralin typically found (Table VII.8.3). This level was sufficiently high to produce skin irritations in consumers (Ames et al. 1990). Ordinary celery, subjected to stress, contains still higher levels of psoralin.

The increased psoralin content of the new celery cultivar was the result of genetic manipulation via traditional crop breeding techniques. It represents a reversal of the more commonly observed trend of decreased levels of naturally occurring toxicants in plants as they are “domesticated” through crop breeding (International Food Biotechnology Council 1990). In this particular case, however, the breeder’s goal was to select for enhanced insect resistance so as to minimize the need for synthetic pesticides. Hence, the outcome of increased levels of psoralin (celery’s natural insecticide) is not surprising. It should be noted that when the excessive psoralin problem was discovered, the cultivar was quickly withdrawn from the market (Ames et al. 1990).

This example offers insight into the evaluation of the safety of novel foods. The issue is whether introducing the new gene into the recipient triggers a response by the recipient that bears on health and safety. One of the main considerations is the possibility of increased synthesis of a toxicant synthesized naturally by the recipient plant (such as psoralin in a hypothetical novel celery).

There is considerable information on naturally occurring toxicants that are produced by food plants (International Food Biotechnology Council 1990). Such information can provide the framework for comparing the natural toxin levels in a novel food with those in its traditional (unmodified) counterpart—for instance, the psoralins in the hypothetical novel celery versus those in ordinary celery. A given plant species produces only a limited number of the known naturally occurring toxicants (usually restricted to a specific chemical class). Hence, it is not necessary to analyze each plant for every known, naturally occurring toxin. Rather, effort can be concentrated on the potential toxicants with which the plants in question are known to be associated. Such assays can take the form of specific chemical analyses, possibly coupled with judicious, limited animal testing (International Food Biotechnology Council 1990; Kessler et al. 1992).

The concept of comparative toxicology—assessing safety of a novel food by comparing it with its traditional counterpart—is central to the food safety assessment schemes developed by the International Food Biotechnology Council, a food and biotechnology industry–sponsored group (International Food Biotechnology Council 1990) and the FDA (Kessler et al 1992). In addition to assays for specific toxins, one should also consider changes in composition that might affect pathogen growth (Pariza 1992b).


With the advent of biotechnology, the production and availability of food is entering a new era of great promise for improved human welfare. Procedures and methodologies are available to ensure the safety of the novel foods that have been and will be developed through biotechnology (International Food Biotechnology Council 1990; Kessler et al. 1992). These procedures and methodologies are not fundamentally different from those used now to assure the safety of traditional foods developed though traditional means of breeding and improvement.