The sequence of small-subunit rRNA varies in an orderly manner across phylogenetic lines and contains segments that are conserved at the species, genus, or kingdom level. By directing oligonucleotide primers at sequences conserved throughout the eubacterial kingdom, we amplified bacterial 16S ribosomal DNA sequences with the polymerase chain reaction. Priming sites were located at the extreme 5' end, the extreme 3' end, and the center of 16S ribosomal DNA. The isolates tested with these primers included members of the genera Staphylococcus, Coxiella, Rickettsia, Clostridium, Neisseria, Mycobacterium, Bilophila, Eubacterium, Fusobacterium, and Lactobacillus and the family Enterobacteriaceae. Initially, the yields from the reactions were erratic because the primers were self-complementary at the 3' ends. Revised primers that were not self-complementary gave more reproducible results. With the latter primers, 0.4 pg of Escherichia coli DNA consistently gave a visible band after amplification. This method should be useful for increasing the amounts of bacterial 16S ribosomal DNA sequences for the purposes of sequencing and probing. It should have a broad range of applications, including the detection and identification of known pathogens that are difficult to culture. This approach may make it possible to identify new, nonculturable bacterial pathogens.

Human colonic biota is a complex microbial ecosystem that serves as a host defense. Unlike most microbial ecosystems, its composition has been studied extensively by relatively efficient culture methods. We have compared an established culture-based method with direct amplification and partial sequencing of cloned 16S rRNA genes from a human fecal specimen. Nine cycles of PCR were also compared with 35 cycles. Colonies and cloned amplicons were classified by comparing their ribosomal DNA (rDNA; DNA coding for rRNA) sequences with rDNA sequences of known phylogeny. Quantitative culture recovered 58% of the microscopic count. The 48 colonies identified gave 21 rDNA sequences; it was estimated that 72% of the rDNA sequences from the total population of culturable cells would match these 21 sampled sequences (72% coverage). Fifty 9-cycle clones gave 27 sequences and 59% coverage of cloned rDNAs. Thirty-nine rDNAs cloned after 35 cycles of PCR gave 13 sequences for 74% coverage. Thus, the representation of the ecosystem after 35 cycles of PCR was distorted and lacked diversity. However, when the number of temperature cycles was minimized, biodiversity was preserved, and there was good agreement between culturing bacteria and sampling rDNA directly.

No abstract

We describe an immunocompromised renal transplantation patient with opportunistic lung infection due to Bartonella henselae (formerly Rochalimaea henselae) and provide evidence suggesting transmission from a pet cat. Computed tomographic scans of the chest and lung biopsies provided material for diagnosis. The etiology was established by polymerase chain reaction and sequencing of a 16S ribosomal DNA segment from infected lung tissue. Histopathologic and serological evidence supported the molecular data. B. henselae was isolated from the blood of eight of the patient's many cats. The patient responded to prolonged therapy with doxycycline, and relapse did not occur during a 1-year follow-up. B. henselae joins a long list of pathogens that can cause lung infections in association with cell-mediated immunodeficiency states. Molecular methods are useful in diagnosis of this infection in light of the bacterium's fastidious growth characteristics. If an immunocompromised patient has lung nodules and a history of exposure to cats, B. henselae should be sought in biopsy specimens.