Agency Response Letter GRAS Notice No. GRN 000091
CFSAN/Office of Food Additive Safety
January 7, 2002
Stuart M. Pape
Patton Boggs LLP
2550 M Street, N.W.
Washington, DC 20037
Re: GRAS Notice No. GRN 000091
Dear Mr. Pape:
The Food and Drug Administration (FDA) is responding to the notice, dated November 30, 2001, that you submitted on behalf of Marlow Foods Ltd. (Marlow) in accordance with the agency's proposed regulation, proposed 21 CFR 170.36 (62 FR 18938; April 17, 1997; Substances Generally Recognized as Safe (GRAS); the GRAS proposal). FDA received the notice on December 3, 2001, and designated it as GRAS Notice No. GRN 000091.
The subject of the notice is mycoprotein. The notice informs FDA of the view of Marlow that mycoprotein is GRAS, through scientific procedures, for use in food in general, except meat products, poultry products, or infant formula. For simplicity, in the remainder of this letter FDA refers to the intended use of mycoprotein as "use in food."
Marlow's notice includes the report of a panel of individuals (Marlow's panel) who evaluated the data and information that are the basis for Marlow's GRAS determination. Marlow's notice also includes a publication derived from that report. Marlow considers the members of its panel to be qualified by scientific training and experience to evaluate the safety of substances added to food. In its report, Marlow's panel concludes that mycoprotein is a safe and suitable ingredient for use in food as a source of protein in the diet.
As discussed more fully below, the agency has no questions at this time regarding Marlow's conclusion that mycoprotein is GRAS under its intended conditions of use, provided that neither trichothecene mycotoxins (represented by nivalenol, deoxynivalenol, 3-acetyldeoxynivalenol, diacetoxyscirpenol, fusarenone X, and neosolaniol) nor fusarin mycotoxins are detectable in the ingredient. As also discussed more fully below, FDA is continuing to proceed to publish an order in response to a petition (FAP 6A3930) that Marlow previously submitted to FDA, requesting that FDA approve the use of mycoprotein in food as a food additive.
Data and information that Marlow presents to support its GRAS determination
Mycoprotein is the processed cellular mass that is obtained from the filamentous fungus Fusarium venenatum strain PTA-2684. F. venenatum strain PTA-2684 derives from a soil sample obtained from Buckinghamshire, United Kingdom. Because the predominant component of the cellular mass derived from F. venenatum strain PTA-2684 is protein, and because "myco" is a prefix that means fungus, Marlow calls the processed cellular mass "mycoprotein." In its notice, Marlow describes the development of mycoprotein during the 1970's and 1980's and cites a published report about that development. Mycoprotein has been commercially available in the United Kingdom since 1985 and in other countries in Europe since 1991.
The protein-rich cellular mass derived from microorganisms is commonly known as "single cell protein." Fungal sources of single cell proteins have a relatively high content of ribonucleic acid (RNA) compared to traditional sources of protein, such as meat, fish, or soybeans (Refs. 1 and 2). Ingested RNA is ultimately broken down to purines and pyrimidines. Purines are subsequently converted to uric acid, which may add to the serum uric acid pool produced from metabolism of endogenous purines. A United Nations Protein Advisory Group recommended in 1972 that single cell proteins intended for human consumption should not provide more than 2 grams of RNA per day because elevated levels of uric acid could increase the risk of developing gout and kidney stones in susceptible individuals. Consistent with the 1972 recommendations of the United Nations Protein Advisory Group, Marlow processes the cellular mass that is obtained from F. venenatum strain PTA-2684 to reduce the content of RNA.
Marlow describes published information about the composition and properties of mycoprotein. As a member of the class Fungi-Imperfecti, F. venenatum strain PTA-2684 has a structure and composition typical of micro-fungi, except for a higher protein content. The filamentous nature of the cells gives the cellular mass a meat-like texture that makes it suitable for a variety of applications in food, including use as a muscle replacer in meat-alternative products. Mycoprotein also can be used as a fat replacer in certain dairy products and as a cereal replacer in products such as breakfast cereals or puffed snacks. Commercially available products in Europe include meat-free burgers and fillets and prepared meals (e.g., stir-fries, curries, and pasta dishes in which mycoprotein is the central component). Mycoprotein is also sold in Europe as a food ingredient for home use. Marlow uses information from several sources to estimate that dietary intake from consumption of mycoprotein would be within the range of 6 to 11 grams per person per day (g/p/d) on a dry weight basis (equivalent to 26 to 46 g/p/d on a wet weight basis) for the general population and 14 to 28 g/p/d on a dry weight basis (equivalent to 56 to 112 g/p/d on a wet weight basis) for consumers who avoid meat.(1),(2)
Marlow describes generally available information, including a published report, about the nutritional characteristics of mycoprotein. The major components of mycoprotein are protein, fiber, and lipids.
The protein content of mycoprotein typically ranges from 42 to 50 percent on a dry weight basis. Marlow considers that mycoprotein is a high quality protein because mycoprotein contains all of the essential amino acids and because unpublished studies in rats demonstrate that the protein efficiency ratio (PER) and net protein utilization (NPU) are both greater than 85 percent of the corresponding values for casein. Marlow considers that the results of a published study in human volunteers confirms that the results of these rat studies could be extrapolated to man.
Mycoprotein contains constituents that derive from the cell wall and typically contribute 22 to 28 percent to mycoprotein on a dry weight basis. The cell wall is composed of chitin (poly N-acetyl glucosamine) and beta-glucans (beta-1:3 and beta -1:6 glucosidic linkages). Marlow describes unpublished studies as support for its conclusion that the cell wall components of mycoprotein possess characteristics of dietary fiber and that this dietary fiber is largely insoluble.
The fat content of mycoprotein typically ranges from 12 to 14 percent on a dry weight basis. Marlow's notice includes a publication that describes the fat content of mycoprotein. Marlow considers that this fat content is more like vegetable fat than animal fat because it has a low proportion of saturated fatty acids and a high proportion of mono- and polyunsaturated fatty acids.
Marlow describes published information about the ability of Fusarium species to produce trichothecene mycotoxins and fusarin mycotoxins. Marlow also cites published information to support its view that it is generally known that such mycotoxins are produced only when growth is limited by imbalanced nutritional or physical conditions such as a high ratio of carbon to nitrogen, low oxygen tension, and incomplete nutrient requirements. Marlow emphasizes that the conditions that it uses to grow F. venenatum strain PTA-2684 are not conducive to the production of mycotoxins.
Marlow describes the manufacturing process for mycoprotein and notes that this process is described in its panel report and in a recent publication. Mycoprotein is produced by the continuous fermentation of F. venenatum strain PTA-2684. The fungus is grown aerobically under steady-state conditions maintained by a continuous feed of nutrients and simultaneous removal of the culture. To prevent mycotoxin synthesis, the production strain is grown at a high rate without any nutritional limitations. To ensure such conditions, the culture is supplied with a nutritionally balanced chemically defined fermentation medium containing easily metabolizable nutrients, including glucose as a sole carbon source. The medium is provided at a rate that allows the cells to grow at a specific growth rate of at least 0.17 per hour. The biomass removed from the fermenter is rapidly heated by injection of steam. The rapid heating process kills the cells. The fermentation broth is subsequently separated from the cellular mass by centrifugation. The combination of rapid heating and centrifugation reduces the content of RNA in the cellular mass from about 10 percent in viable cells to about 0.5 to 2 percent in mycoprotein on a dry weight basis. The processed cellular mass is subjected to numerous tests to demonstrate compliance with specifications.
Marlow describes the process controls that it uses to assure that the manufacturing conditions for mycoprotein are not conducive to the production of mycotoxins. Marlow has elaborated and validated a method for monitoring mycoprotein to assure that trichothecene or fusarin mycotoxins are not present at detectable levels. The method for detection of trichothecene mycotoxins elaborated by Marlow is based on a published method for detection of trichothecenes in cereals and spices. Marlow's method is applicable to the following trichothecene mycotoxins: nivalenol, deoxynivalenol, 3-acetyldeoxynivalenol, diacetoxyscirpenol, fusarenone X, and neosolaniol. Detection limits for individual trichothecenes are no greater than 2 micrograms per kilogram (on a wet weight basis). Marlow also has elaborated a method for detecting fusarin mycotoxins using high performance liquid chromatography with mass spectrometric detection. Detection limits for individual fusarins are no greater than 5 micrograms/kg (on a wet weight basis).
The specifications established by Marlow for mycoprotein stipulate that the level of RNA must not exceed 2 percent on a dry weight basis. Given that the highest estimate of dietary intake of mycoprotein is 17 to 33 g/p/d on a dry weight basis, the intake of RNA from consumption of mycoprotein would be no more than 0.35 to 0.7 g/p/d, which is well below the level recommended by the United Nations Protein Advisory Group. The specifications also stipulate that trichothecene and fusarin mycotoxins are not detectable in mycoprotein (on a wet weight basis) and that the level of the heavy metal (lead, arsenic, mercury and cadmium) is no greater than 0.1 mg/kg (on a dry weight basis). Marlow also established specifications for protein and ash. The level of protein in mycoprotein must be not less that 41 percent (on a dry weight basis) and the level of ash not more than 5 percent (on a dry weight basis).
Marlow describes animal studies that were evaluated by its panel and reported in the publication derived from the panel's report. These studies include four subchronic studies conducted in rats (with durations ranging from 13 weeks to 22 weeks), a 13-week study conducted in baboons, a one-year study conducted in dogs, a two-year carcinogenicity study conducted in rats with an in-utero phase, and a two-generation study conducted in rats with a teratology phase. Marlow cites the conclusions of its panel that these studies demonstrate that mycoprotein does not cause chronic toxicity, is not a reproductive toxicant, is not a teratogen, and is not carcinogenic.
Marlow describes unpublished studies designed to investigate potential effects of mycoprotein on mineral absorption. Marlow concludes that studies conducted in mature rats that were beyond the rapid growth phase, or in young, fast-growing rats, showed no significant adverse effects on mineral balance with regard to calcium, phosphorus, magnesium, iron, copper and zinc. Marlow also concludes that an examination of mineral absorption during a digestibility study involving ileostomy patients showed no significant effects on the apparent absorption of calcium, magnesium, phosphorus, iron and zinc in comparison with a polysaccharide-free diet and wholemeal bread.
Marlow describes human studies, which were designed to assess human tolerance to mycoprotein, that were evaluated by its panel and reported in the publication derived from the panel's report. Marlow concludes that these studies demonstrate that mycoprotein is well-tolerated by humans. Marlow also provides information about reports from European consumers about adverse reactions to products containing mycoprotein. For example, Marlow received 92 communications regarding adverse events in 1999, and 89 such communications in 2000. Marlow calculates that these figures equate to an incidence rate per the estimated number of consumers of 1 in 130,000 and 1 in 146,000 and compares these estimated rates to published estimates of the rates of adverse reactions in the United Kingdom to soy (1 in 350) and to fish/shellfish (1 in 35). Marlow also notes that the reported adverse reactions are primarily gastrointestinal. Marlow cites a published report about an investigation of possible adverse reactions to mycoprotein to support its conclusion that only a very small percentage of these reactions (approximately 5 percent) are true food allergies (i.e., are mediated by Immunoglobulin E) and that the rest are likely food intolerances. Marlow cites the conclusions of its panel that the level of allergic reactions is well below that of products that are mainstays of the human diet.
Marlow cites a publication that described a human study designed to test the nutritional value of two microfungal foods (Ref. 1). In that study, human subjects consumed diets that contained mycoprotein, a single cell protein (called "PVP") from another microfungal source (i.e., Paecilomyces variotii), or cow's milk protein as the sole source of protein. Although the mycoprotein used in the study had been treated to reduce the content of RNA, the RNA content of the microfungal protein PVP had not been reduced. The study authors noted that the serum uric acid of subjects who consumed PVP, but not subjects who consumed mycoprotein or cow's milk protein, increased from 5.5. mg/dl at the beginning of the study to 9.2 mg/dl at the end of the study. The study authors concluded that these results supported the recommendations to reduce the nucleic acid content of microfungal proteins if they are to be used as a major source of protein in the diet.
FDA's response to Marlow's GRAS notice
Based on the information provided by Marlow, as well as the information in FAP 6A3930 and other information available to FDA, the agency has no questions at this time regarding Marlow's conclusion that mycoprotein is GRAS under the intended conditions of use, provided that neither trichothecene mycotoxins (represented by nivalenol, deoxynivalenol, 3-acetyldeoxynivalenol, diacetoxyscirpenol, fusarenone X, and neosolaniol) nor fusarin mycotoxins are detectable in the ingredient. The agency has not, however, made its own determination regarding the GRAS status of the subject use of mycoprotein. As always, it is the continuing responsibility of Marlow to ensure that food ingredients that the firm markets are safe, and are otherwise in compliance with all applicable legal and regulatory requirements.
In accordance with proposed 21 CFR 170.36(f), a copy of the text of this letter, as well as a copy of the information in the notice that conforms to the information in proposed 21 CFR 170.36(c)(1), is available for public review and copying on the homepage of the Office of Food Additive Safety (on the Internet at http://www.cfsan.fda.gov/~lrd/foodadd.html).
Marlow's 1986 petition to FDA to approve the use of mycoprotein as a food additive
Under sections 201(s) and 409 of the Federal Food, Drug, and Cosmetic Act (the FFDCA), mycoprotein that would be added to food is a food additive that is subject to premarket review and approval by FDA unless mycoprotein is generally recognized, by experts qualified by scientific training and experience to evaluate the safety of substances added to food, to be safe under the conditions of its intended use. In 1986, Marlow requested that FDA approve the intended use of mycoprotein as a food additive (FAP 6A3930). In 1999, Marlow submitted, as an amendment to FAP 6A3930, a copy of the panel report that accompanies Marlow's GRAS notice.
In 1986, when Marlow submitted FAP 6A3930, most of the data and information that Marlow relies on to conclude that the intended use of mycoprotein is safe were not generally available to the public. Given the circumstances in 1986, there was no basis at that time to consider whether the intended use of mycoprotein could be GRAS. In contrast, at the present time most of the data and information that Marlow relies on to conclude that the intended use of mycoprotein is safe are generally available to the public. In addition, these data and information have been reviewed by a panel of individuals that Marlow considers qualified by scientific training and experience to evaluate the safety of substances added to food. Given the circumstances at present, FDA has no questions about Marlow's view that mycoprotein is GRAS under the conditions of its intended use.
Importantly, the specific circumstances, described above, that surround Marlow's submission of GRN 000091 were the principal factor in the agency's decision, as a matter of its discretion, to file and evaluate Marlow's GRAS notice while a food additive petition for the same use of the same ingredient remains under review at the agency. Marlow discussed its intent to submit a GRAS notice with the agency before it did so (Refs. 3 and 4). The agency's review, in the context of the pending food additive petition, of the specific data and information that Marlow is relying on to demonstrate safety was substantially complete. FDA was aware that Marlow had convened a panel to independently evaluate those data and information. During meetings on October 3 and 18, 2001, Marlow discussed the fact that most of the data and information that are the basis for its conclusion that its intended use of mycoprotein is safe have now become generally available. Given these circumstances, FDA consented to evaluate, under the rubric of the proposed voluntary GRAS notification program, Marlow's view that the intended use of mycoprotein had become GRAS.
Under different circumstances, FDA likely would have declined to devote resources simultaneously to both a GRAS notice and a food additive petition for the same intended use of the same substance. For example, FDA could have requested that Marlow withdraw FAP 6A3930 before FDA filed Marlow's GRAS notice. However, at the meetings on October 3 and 18, 2001, FDA and Marlow agreed that it would be beneficial for FDA to complete its evaluation of FAP 6A3930. For example, FDA believes that it would be valuable for the procedures that Marlow has developed to assure the safe manufacture of mycoprotein to be available in a concise and convenient form through incorporation by reference in the agency's regulations. Consistent with the GRAS proposal, the mechanism to obtain a regulation is the food additive petition process. Accordingly, FDA is continuing to proceed to publish an order in response to FAP 6A3930.
Alan M. Rulis, Ph.D.
Office of Food Additive Safety
Center for Food Safety and Applied Nutrition
1. Udall J.N., et. al. 1984. The tolerance and nutritional value of two microfungal foods in human subjects. American Journal of Clinical Nutrition 40: 285-292.
2. Elizabeth Moore-Landecker, Fundamentals of the Fungi. Prentice Hall, Upper Saddle River, NJ. Fourth Edition. 1996. Pp. 543-545.
3. Memorandum of Meeting held on October 3, 2001, between Stuart Pape of Patton Boggs and Alan Rulis, Laura Tarantino, and Linda Kahl of FDA.
4. Memorandum of Meeting held on October 18, 2001, between Stuart Pape of Patton Boggs, John Pinkney and Nick Hughes of Marlow Foods, and Laura Tarantino and Linda Kahl of FDA.
(1)Prior to receiving GRN 000091, FDA evaluated dietary exposure to mycoprotein in the context of its review of FAP 6A3930. FDA estimates that dietary exposure to mycoprotein would be within the range of 17 to 33 g/p/d on a dry weight basis (equivalent to 70 to 139 g/p/d on a wet weight basis). This exposure estimate is based on meat consumption provided in the USDA 2-day 1994-1996 Continuing Survey of Food Intake by Individuals and assuming that mycoprotein would entirely replace meat in the diet. The exposure estimate of 17 to 33 g/p/d on a dry weight basis was considered in Marlow's panel report.
(2)Mycoprotein that is prepared for use as a food ingredient typically has a solids content of about 25 percent. In its notice, Marlow refers to this form of mycoprotein as "wet." Throughout its notice, Marlow provides information about mycoprotein based both on this "wet weight" and on "dry weight" (i.e., 100 percent solids).