The Cooking Makes a Difference


OOKING practices can cause large variations in the total mutagenic activity and in the amount of specific mutagens present in muscle-containing foods. For example, the amount of mutagens in a cooked hamburger from a restaurant varies considerably from one vendor to another and is often several-fold lower than that in a hamburger prepared in our laboratory (and presumably at home). The variation has much to do with the details of food preparation, such as cooking temperature and cooking time. It is becoming increasingly clear that there can be many different routes and rates of formation for the different mutagens we are investigating. Thus, a major concern is to identify the precursors and specific reaction conditions that lead to the formation of mutagens during cooking. With this information, it may be possible to devise strategies to reduce or prevent the formation of mutagens.


Figure 7. Modeling the formation of the potent mutagen PhIP. We combined two precursors, phenylalanine and creatine, in amounts naturally found in raw beef. After simple dry heating, PhIP was produced in yields comparable to those we obtain in beef after the cooking process. We have also labeled the two precursors with heavy isotopes to track the incorporation of specific atoms within each precursor molecule into the PhIP molecule. Such work shows unequivocally the source of atoms that make up the mutagenic product.


Precursors
The reactions that produce mutagens in cooked food are not merely the random coalescence of small fragments. We now know that the heterocyclic amines can be formed from single amino acids (the building blocks of proteins) or proteins when these precursors are heated alone. However, the temperatures required to produce mutagens from amino acids or proteins by themselves are higher than those normally used in cooking.
Muscle meats contain creatine and creatinine. At more typical cooking temperatures (greater than 150°C), one or both of these two precursors react with the free amino acids and, in some cases, sugars to form a series of heterocyclic amines more easily.

Modeling the Formation
We have modeled the formation of the important mutagen, PhIP (pronounced "fip"), starting with the amino acid phenylalanine mixed with either creatine or creatinine, both of which are found naturally in animal muscle. When phenylalanine and creatine are mixed in the proportion normally found in raw beef and dry heated at 200°C, PhIP is produced in amounts comparable to those found after cooking beef. Figure 7 shows the structures of phenylalanine and creatine and of the PhIP molecule that is produced.
We have modeled the formation of several other food mutagens in additional laboratory experiments. For example, the mutagen IQ can be formed with creatine, creatinine, and any of four different amino acids, again suggesting many different possible routes of formation.
Model reactions can help us identify new mutagens as well. In one case, dry heating three precursors known to be present in meat led us to identify a mutagen with two amino and two methyl groups and a molecular weight of 244. However, the presence of this new mutagen in food has not been verified.

Variations in Cooking
During the actual cooking of meat patties, water and precursors move to the hot, drying contact surfaces of the meat where reactions occur. Such migration, with water serving as the transport vehicle, may account for the concentration of precursors near the meat surface, which we have observed in several investigations. However, different cooking practices can lead to very different results. For example, some mutagens are produced at all frying temperatures, whereas others may require higher temperatures. Furthermore, when hamburger patties are grilled at high temperature over an open flame, we can account for less than 30% of the mutagens in the meat. When cooking over an open flame, polycyclic aromatic hydrocarbons (different from AIA food mutagens) arise from fat that drips from the meat--this is an entirely different mechanism than those that produce heterocyclic amines from heated muscle tissue itself. Thus, the formation of mutagens is complex and highly dependent on the details of cooking.

Preparation Principles
Given this complicated picture, what statements about food preparation can we make with any certainty? Here is a summary of some of the important things we have learned about the cooking process:

  • Food mutagens can be produced both with and without water present. Early reports suggested that water is essential to produce food mutagens. In later studies, dry heating actually gives a greater percentage of certain types of mutagens compared with aqueous heating. We know, for example, that the mutagen PhIP is formed relatively efficiently in dry heating reactions. We have also found that water tends to inhibit the formation of IQ-type mutagens.
  • Microwave pretreatment of meat can reduce the formation of heterocyclic amine mutagens. When meat is microwaved for a few minutes, a clear liquid is released, which contains many of the precursors of mutagens. When the resulting liquid is drained off before frying, our studies show that mutagenic activity, as measured by the Ames test, and the amount of heterocyclic amine are 90 to 95% lower than they are in meat samples that are not pretreated by microwave cooking. (Reference 6) The box discusses this and other methods that have been tested to reduce the formation of mutagens.
  • Different cooking methods produce quite different results. In general, frying, broiling, and flame grilling muscle meats produce more heterocyclic amines and mutagenic activity than other methods. Stewing, steaming, and poaching produce little or no mutagenic activity. Roasting and baking have variable responses.
  • Heating temperature is extremely important as is the time of cooking at a given temperature. Our extensive findings on this important topic are best discussed according to the type of food product.


  • Cooked Muscle Meats
    Fried beef patties appear to be the most commonly eaten cooked meat with the highest mutagenic activity. Because of the high intake of fried beef (based on surveys from the U.S. Department of Agriculture and the Department of Health, Education, and Welfare), this food may be the most important single source of heterocyclic amines in the typical American diet. However, several other popular cooked meats, including fish, chicken, and pork, have been shown to produce a potent response in the Ames test.
    Of the several different heterocyclic amine mutagens now identified, we wanted to know which ones are most important (that is, most abundant by mass) in cooked muscle meats. To help answer this question, we compared the results of many studies from LLNL and elsewhere. Specifically, we compared the mass percentages of different mutagens in cooked muscle meats, including fried beef, broiled fish, and commercially prepared beef extract. We found that the results were generally consistent among different laboratories even when different analytical methods were used.
    First, we did not detect significant levels of three mutagens, Trp-P-1, Trp-P-2, and Glu-P-1, in any of our meat samples. Second, we found that four compounds, IQ, 8-MeIQx, 4,8-DiMeIQx, and PhIP, contribute about 80% of the mutagenic activity in the cooked muscle foods that were studied. Third, we found that PhIP alone can account for a startling 83 to 93% of the mass of these four mutagenic compounds. Clearly, the analysis of PhIP is important because it appears to be, by far, the most abundant heterocyclic amine by mass in commonly eaten cooked meats. Because PhIP is as carcinogenic as the other mutagens, its analysis becomes even more essential. We examined the production of PhIP and other mutagens in beef at different cooking temperatures and times. The box gives the details on how we prepare our fried beef. Figure 8 shows the mutagenic activity, as measured by the Ames test, of all the mutagens combined in a gram of beef patty fried at 150°, 190°, and 230°C. We found no detectable heterocyclic amines after frying at 150°C for 2 to 4 minutes. In general, increasing either the temperature or time of cooking (specifically, frying on a solid metal restaurant-type grill) causes a dramatic increase in both the mutagenic activity and the total amount of mutagens produced, especially PhIP and 8-MeIQx. For the most part, as shown in Table 3, the amount of individual mutagens in fried beef increases proportionately with the cooking temperature. A clear exception to this trend is the compound PhIP, which is produced at much greater concentrations at higher temperatures relative to the other mutagens we have studied. When the cooking temperature and time are increased, the PhIP content of fried beef patties increases nearly exponentially.

    Figure 8. A graph of the mutagenic activity in beef patties fried at three different temperatures. The essential point is that mutagenic activity increases with both frying temperature and time.














    Mutagens from Grain?
    We also recently used the Ames test to assess the mutagenic activity in many heated foods derived from grain products. Our studies include cooked breads (white, pumpernickel, crescent rolls, and pizza crust), breadsticks, heated flour from many different grain sources, breakfast cereals, graham crackers, and meat-substitute patties after frying. These foods were either tested as purchased without additional cooking (for example, graham crackers and a grain beverage powder) or were cooked according to package instructions. In some studies, we deliberately overcooked the grain products for twice the cooking time at the specified temperature setting to see if the mutagen content would increase with continued cooking, as it does in muscle meats.

    Figure 9. The mutagenic activity of wheat gluten increases dramatically when heated at 210°C for up to 2 hours. This potent response tells us that one or more highly mutagenic chemicals, still unidentified, are formed with continued cooking at high temperature.












    Our studies generally demonstrate increased mutagenic activity in grain foods with cooking time, but the exact composition of the food is important. For example, when wheat gluten (the protein in wheat seed) is heated alone at 210°C in a beaker, it shows a potent, time-dependent mutagenic response (Figure 9). Because breadsticks are high in wheat gluten, they also show some activity when heated normally and much higher activity when overcooked. In fact, the mutagenic activity of breadsticks cooked for double the regular heating time is 20% that of a hamburger fried 6 minutes per side at 210°C. In all cases, overcooking grain foods led to much higher mutagenic activity. Cooked garbanzo bean flour and the grain beverage powder, which we tested as purchased, had relatively high mutagenic activity. Cooked rice and rye flour (containing no gluten), on the other hand, showed no detectable activity, and rice cereal showed very little. Fried tofu (soy bean curd) was not mutagenic, and the measured level of activity in meat-substitute patties (which are made from vegetable proteins) after frying was about 10% or less than that of a beef patty cooked under the same conditions.
    Table 4 summarizes the mutagenic activity, as measured by the Ames/Salmonella test, for a variety of cooked-grain food products. The results are expressed as mutagenic activity from the Ames test, so they cannot be directly compared with those in Table 3. (Recall that the numbers in Table 3 represent a different measure, namely the content by weight of individual mutagens expressed as nanograms of mutagen per gram of beef.) Because we do not yet know the identity of the mutagens present in cooked grain products, we cannot provide their content by weight. However, to allow for some comparison between cooked grains and meat, we have included the values of mutagenic activity for hamburger cooked for three different times at the end of Table 4.
    Overall, the level of mutagenic activity measured in heated nonmeat foods is lower than that in cooked meats. It is important to recognize that the cooked grains we studied lack the creatine and creatinine levels that explain the formation of mutagens in muscle meats during cooking. We are currently investigating the question of why foods high in gluten are quite mutagenic in the absence of creatine and creatinine. We suspect that the amino acid, arginine, can substitute for the creatine and creatinine precursors found in meat, but it may be a less efficient mutagen precursor in cooked grain products.
    Before we can evaluate the risk associated with cooked grains, we need to determine the mass of mutagens in each food and to identify the specific mutagenic compounds that are present. Except for very low levels of PhIP in wheat gluten (accounting for only 4% of its mutagenic activity), our analysis did not reveal any of the other mutagens found in cooked meat or listed in Table 2. Because the mutagens in cooked grain appear to be as potent as the heterocyclic amines--and such potency is unusual, we suspect that the mutagenic compounds may be new heterocyclic amines similar to those we have found in cooked meats. However, more work needs to be done before we understand the entire picture.



    What About Fumes?
    Some studies have suggested the possibility of an increased risk of respiratory tract cancer among cooks and bakers. When foods rich in protein are heated, the fumes that are generated sometimes contain many different known carcinogens, including polycyclic aromatic hydrocarbons and heterocyclic amines. Working with colleagues at the University of California at Davis, we recently studied the mutagenicity of fumes generated when beef is fried at high temperatures.
    We collected smoke from cooking by using a special sampling system consisting of a condenser, Teflon filters, and absorbent tubes containing polyurethane foam and a resin. We found that the main volatile compounds generated during frying were alcohols, alkanes, aldehydes, ketones, phenols, and acids. Their presence--we measured 34 different components--may account for much of the toxicity of fume samples in bacterial tests. Two mutagens, PhIP and A(alpha)C, were the most abundant of the heterocyclic amines we measured in smoke, with AåC accounting for 57% of the total weight of mutagens in the recovered samples. However, even though A(alpha)C seems to be the most volatile of our quantified heterocyclic amines in smoke, its actual contribution to the mutagenicity of fumes is negligible because its mutagenic potency is lower than that of some other heterocyclic amines in smoke. We also detected significant levels of MeIQx and DiMeIQx.
    In a modified Ames test, one that is much more sensitive than our standard assay and uses two different strains of bacteria, the fried meat extracts had 30,700 revertants per gram (see box, p. 16 for a definition of "revertants"), whereas the fumes produced by frying had 10,400 revertants per gram of fried meat. Thus, the fumes generated during the cooking of meat represent about one-third of the mutagenic activity measured in the fried meat itself. It is important to recognize that the amount of mutagens inhaled is very low compared to consuming solid, cooked meat. Nevertheless, the presence of toxic compounds in meat fumes, even at relatively low levels, could pose some risk to food preparers who are exposed to them for long periods over many years.

    Cook to Manage Mutagens
    Our research on food mutagens is not specifically designed to generate practical advice for diet- and health-conscious individuals. Many questions remain unanswered in this highly complex field of investigation. Although food mutagens are extremely potent, our preliminary estimates of risk are not alarming primarily because of their low concentrations in food. Nevertheless, the amount of mutagens ingested can be reduced by choice of diet and by modifying cooking practices.



    Cooking Tips Summary

  • Fried beef has very high mutagenic activity. Its popularity suggests that this food may be the most important source of heterocyclic amines in the typical Western diet.
  • Most, but not all, of the mutagenic activity in fried beef can be accounted for by known heterocyclic amines. The single mutagen PhIP accounts for most of the combined mass of mutagens in fried beef cooked well-done.
  • The fumes generated during the cooking of beef have about one-third the mutagenic activity measured in the fried meat itself. Occupational exposure over long periods could pose some risk, but probably much less than that from consuming the meat.
  • The fat content and thickness of meat have little effect on mutagen content, whereas the method and extent of cooking have major effects. Frying, broiling, and barbecuing muscle meats produce more heterocyclic amines and mutagenic activity, whereas stewing, steaming, and poaching produce little or no mutagenic activity. Roasting and baking show variable responses.
  • Both cooking temperature and time can be manipulated to minimize the formation of mutagens. Increasing the frying temperature of ground beef from 200 to 250°C increases mutagenic activity about six- to sevenfold. Reducing cooking temperature and time can significantly lower the amounts of mutagens generated and subsequently consumed in the diet.
  • Microwave pretreatment of meat, followed by pouring off the clear liquid before further cooking, can substantially reduce the formation of heterocyclic amine mutagens, even if the meat is cooked well-done.
  • Most nonmeat foods, including cooked grain products, contain lower levels of mutagens than cooked meats.
    At least in rodents, we know that food mutagens trigger cancer in several different target tissues, such as the liver, colon, breast, and pancreas. In a follow-up installment in Science and Technology Review, we will address the health risks to humans that may arise from exposure to heterocyclic amines. For this intriguing part of the story, we will show how these highly toxic compounds can react with the most critical macromolecule of all, DNA. With a connection established between food mutagens, DNA damage, and the potential for cancer, we will then try to make sense of what all the numbers on mutagenic activity and mutagen content in food mean for the average person.

  • Return to Introduction Page//Return to July 1995//Return to S&TR 1995