Skip page top navigationFDA Logo--links to FDA home page Logo of and Link to start page of Office of Regulatory Affairs, U.S. Food and Drug Administration U.S. Food and Drug Administration Office of Regulatory Affairs HHS Logo and link to Department of Health and Human Services website

FDA Home Page | Federal-State | Import Program | Compliance | Inspection | Science | ORA Search

Laboratory Information Bulletin

DFS/ORO/ORA No. 3982
Page 1 of 4



In recent years, several District offices have received consumer complaints about Guanabana Juice, a product made from a tropical fruit, Annona muricata, L., also known as Soursop. The complaints involved over 65 documented cases of illness (vomiting within 1 hr of ingesting the product). The following method was developed and has subsequently been used with consistent success by several District labs, foreign industry labs, and foreign government labs.


Initially, a wide range of analyses were performed on several lots which had been implicated in a number of complaints of illness. These analyses included:

1. Microbiological exam by direct smear performed by analysts in Micro Lab: no organisms found

2. pH analysis performed by analysts in Micro Lab: 3.39 - 3.70 ("normal")

3. Staph enterotoxin performed by analysts in Micro Lab: none found

4. Pesticide analysis for chlorinated hydrocarbons, organophosphates, and synthetic pyrethroids performed by analysts in Pesticide Lab: none found

5. Heavy Metals performed by Metals Analyst Richard Baetz: no significant levels of lead, cadmium, copper, or zinc (one code had 200 ppm tin)

6. Sulfites performed by Food Lab Analyst Antoine Major: none found

7. Microscopic exam for Mold performed by analysts in Food Lab: no significant amounts found

8. Can abnormalities examined by analysts in Micro Lab: none found (except for etched cans in the code with elevated tin levels)

An investigation into the nature of the fruit which involved literature searches and phone conversations with research scientists and toxicologists revealed that the seeds are toxic and contain a natural emetic. In addition, the seeds are used by natives in Central America and the Caribbean as a pesticide and fish poison.1 It was suspected that some of the seeds had become blended into the juice and were the source of the illnesses. A chlorzinc iodine stain (Hertzberg's Stain)2 which differentiated the various cells present in the seeds has been reported3 and was used to study the suspect sample. A sample of fresh guanabana fruits was collected for use as a control sample and for comparative purposes.

Discussion and Conclusion

Fresh pulp with ground seeds added and pulp without seeds was compared to the sampled product. Microscopic exam at 200X of the "suspect" codes and the controls revealed three types of cells seen in the suspect codes as well as the fresh pulp with added ground seeds. The cells were not seen in the pulp without seeds. In addition, these cells were abundant in samples of finely ground seeds and were tentatively identified as epidermal cells from the outside of the seed coat and perisperm and endosperm cells from the inside of the seed, Figures 1-4. The epidermal cells stained reddish-brown and had thick cells walls. The endosperm cells had a structure similar to the epidermal cells except these cells stained navy blue or black. Perisperm cells stained yellow to amber with thinner cell walls. None of these cells contained nuclear material.


Preparation of KI/I2 stain:

0.5 g Iodine + 10.5 g KI in 25 mL H2O

Preparation of Hertzberg's Stain:

A saturated solution of ZnCl2 was prepared at 70C (approximately 50 g in 10 mL). It generally took 3-4 hours for the ZnCl2 solution to become saturated. An equal volume of KI/I2 solution was mixed with the ZnCl2 solution and allowed to stand until a clear reddish-brown color.

Preparation of Samples:

Approximately 40 mL of product was centrifuged at RCF = 655 x g for 15 minutes. The supernatant was decanted and 0.5 g of sediment was thoroughly mixed with 1.0 mL of Hertzberg's Stain. Samples were examined immediately (1 - 2 drops) at 200X utilizing a Howard Mold Count slide. Two slides were examined from each can of sample and at least 3 cans from each sample were examined. Each sample consisted of one lot (also known as a "code").


Samples which were known to have caused illness were found to have at least 2 fields (of the 2 slides examined from each can) with seed cells present. The seed cells consisted of a group of at least 10 cells or more. Recently, a sample was received in which the product appeared to have been blended to such an extent that groups of seed cellular material consisted of only 2 to 4 cells, but a total of 20 or more of these cells were still found on microscopic examination of 2 slides. However, the presence of any seed cells, regardless of number found, was considered conclusive evidence that the seed toxins were present and that the product was unsafe for consumption since the seeds are known to be extremely toxic.


1. Morton, Julia F. Atlas of Medicinal Plants of Middle America, Bahamas to Yucatán, Charles Thomas Publishers, Springfield, Ill., 1981, pp. 223-225.

2. Gardner and Cooke, Chemical Synonyms and Trade Names, CRC Press, Cleveland, Ohio, 6th Edition, 1968, p.292.

3. Winton, K.B. and A.L., Structure and Composition of Foods, Vol. II, John Wiley and Sons Publishers, New York, 2nd printing, 1945, pp. 526-529.

Figure 1. Soursop. Seed in longitudinal section. S spermoderm; aep outer epiderm; sub subepiderm; sc1longitudinally arranged sclerenchyma cells; sc2 transversely arranged sclerenchyma cells; iep inner epiderm. N perisperm of empty, more or less compressed cells; ol secretion cell. Eendosperm containing al aleurone grains. X160. Inset, seed X1. (K. B. W.)


Figure 2. Epidermal cells (dark reddish Brown color)

Figure 3. Perisperm Cells (yellow to amber color)

Figure 4. Endosperm Cells (navy blue to black color)