RESUMO
The aims of this investigation were to enumerate coliforms in fresh mangoes, puree, cheeks, and cheeks-in-puree in order to determine the source of these organisms in the processed products, to determine methods for their control, and to identify coliforms isolated from cheeks-in-puree to determine whether they have any public health significance. Product from four processors was tested on two occasions. The retail packs of cheeks-in-puree having the highest coliform counts were those in which raw puree was added to the cheeks. Coliform counts in these samples ranged between 1.4 x 10(3) and 5.4 x 10(4) cfu/g. Pasteurisation reduced the coliform count of raw puree to < 5 cfu/g. Forty-seven percent of the 73 colonies, isolated as coliforms on the basis of their colony morphology on violet red bile agar, were identified as Klebsiella pneumoniae using the ATB 32E Identification System. Klebsiella strains were tested for growth at 10 degrees C, faecal coliform response, and fermentation of D-melizitose, to differentiate the three phenotypically similar strains, K. pneumoniae, K. terrigena and K planticola. Results indicated that 41% of K. pneumoniae isolates gave reactions typical of K. pneumoniae. A further 44% of strains gave an atypical reaction pattern for these tests and were designed 'psychrotrophic' K. pneumoniae. Klebsiella pneumoniae counts of between 2.1 x 10(3) and 4.9 x 10(4) cfu/g were predicted to occur in the retail packs of mango cheeks-in-puree produced by the processors who constituted this product with raw puree. In view of the opportunistic pathogenic nature of K. pneumoniae, its presence in these products is considered undesirable and steps, such as pasteurisation of puree, should be taken in order to inactivate it.