The insertion sequence IS1245 is a novel mycobacterial repetitive element identified in Mycobacterium avium. It encodes a transposase which exhibits a 64% amino acid similarity with IS1081, an insertion element present in the M. tuberculosis complex. The host range of IS1245 appears limited to M. avium as this element was not identified in M. intracellulare or in any other of 18 mycobacteria species tested. When IS1245 was used for restriction fragment length polymorphism (RFLP) analysis, human isolates characteristically presented a high number of copies (median, 16; range, 3 to 27) and a diversity of RFLP patterns comparable to that found by pulsed-field gel electrophoresis. Isolates from nonhuman sources differed both in number of copies and in RFLP pattern diversity: while swine isolates shared the characteristics of human strains, those from several avian sources exhibited a very low copy number of IS1245 and appeared clonal on the basis of RFLP.

Mycobacterium avium was recovered from 21 birds and 10 pigs. Bird isolates carried IS901 and a few copies of IS1245 and appeared highly related by pulsed-field gel electrophoresis. Pig isolates showed features previously described in human isolates: a lack of IS901, a high copy number of IS1245, and marked polymorphism by pulsed-field gel electrophoresis. Mycobacterium avium represents an important pathogen in both humans (5, 7, 12, 19, 25) and animals (8, 16). For humans, water has been identified as a definitive source of infection (24). However, the observed genetic diversity of M. avium strains recovered from AIDS patients (1) and the occurrence of M. avium in animals and various environmental samples suggest the existence of additional reservoirs (17). The purpose of this study was to characterize M. avium strains recovered from animal sources within a defined geographical area-the German-speaking part of Switzerland-by using a number of genotype markers, namely, the presence of insertion elements IS900 (9), IS901 (15), and the novel element IS1245 (10), and by evaluating strain relatedness by pulsed-field gel electrophoresis (PFGE). Materials investigated (n ϭ 398) included mandibular and mesenteric lymph nodes from 120 slaughtered pigs collected from two different abattoirs; 103 samples from 25 chickens and 15 eggs from eight different flocks; organs (n ϭ 21) from seven additional birds, including five exotic birds, in which acid-fast bacilli had been detected (Institute for Veterinary Bacteriology, University of Berne); and environmental samples, such as those from soil (n ϭ 8), pig and chicken litter (n ϭ 9), pondwater (n ϭ 1), and chicken feed (n ϭ 1). Samples were homogenized, decontaminated with sodium dodecyl sulfate-sodium hydroxide, and neutralized, and the sediments were resuspended in phosphate-buffered saline (20). One BACTEC 12B vial supplemented with Panta Plus (Becton Dickinson Diagnostic Instrument Systems, Sparks, Md.), one pyruvate-containing Löwenstein-Jensen slant, and one slant of Herrold's medium were inoculated. Slants were incubated at 37ЊC for 12 weeks and inspected for growth weekly. BACTEC 12B vials were processed for 12 weeks with the BACTEC-460 TB apparatus (Johnston Laboratories Inc., Sparks, Md.) according to the manufacturer's recommendations. For identification, mycobacterial isolates underwent PCR-restriction enzyme analysis of the gene coding for the 65-kDa heat shock protein of mycobacteria (21) and conventional biochemical tests (13).

The application of molecular techniques to investigate strain relatedness may help define the local epidemiology of Mycobacterium avium infection, and, by identifying false isolates (i.e. neither pathogens nor colonizers) resulting from contamination, may serve as a tool for quality control in the laboratory. For this purpose, isolates from all patients (n = 129) with Mycobacterium avium infections identified over a two-year period were investigated by pulsed-field gel electrophoresis (PFGE). Of 38 PFGE patterns identified, 34 corresponded to unique strains or to isolates present in no more than two or three individuals. One prevalent strain was identified among HIV-infected patients and three patterns were related to culture contamination events. PFGE (i) established the diversity of Mycobacterium avium strains in a community; (ii) identified the existence of a unique strain that may account for one-fifth of Mycobacterium avium isolated from HIV-infected patients locally; (iii) documented the extent and resolution of a suspected pseudo-outbreak; and (iv) uncovered an additional-unsuspected contamination event.