Rhodobacter sphaeroides showed chemotaxis towards L-alanine but not towards the analog 2-aminoisobutyrate. 2-Aminoisobutyrate and alanine were shown to share a common transport system, but 2-aminoisobutyrate was not metabolized. Chemotaxis towards alanine was inhibited by structurally unrelated metabolites, suggesting cross-inhibition by common metabolic intermediates.Chemotactic sensing in the purple photosynthetic bacterium Rhodobacter sphaeroides is very different from that in enteric bacteria. Membrane-spanning chemoreceptors (methyl-accepting chemotaxis proteins) (1, 8) have not been found (3,17), and all identified chemoattractants are also metabolites (4, 9). We have shown that attractants must be transported into the cell for a response to occur, but we had not conclusively identified whether the energetic process of transport caused the chemotactic signal or whether subsequent metabolism of the effector was required (9, 12, 13).In Escherichia coli, the nonmetabolizable alanine analog 2-aminoisobutyrate is a strong chemoattractant, competing with alanine for methyl-accepting chemotaxis protein Tsr (11). We therefore examined the chemotactic response, transport, and metabolism of alanine and 2-aminoisobutyrate in R. sphaeroides.Chemotactic behavior. R. sphaeroides WS8 was grown photoheterotrophically as previously described (4) but with succinate, Casamino Acids, and ammonium sulfate replaced by 5 mM L-alanine. Cells, washed and resuspended in N2-sparged Na-HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) buffer (10 mM, pH 7.2), were starved for at least 1 h before 0.6 ml was added to the bottom chamber of a chemotaxis well and covered with an 8-pumpore-size polycarbonate membrane and an attractant solution was added to the upper chamber (5). Figure 1 shows that after 90 min there was a threefold increase in bacteria swimming through the membrane in response to 10 mM alanine, whereas there was no chemotactic response to the alanine analog 2-aminoisobutyrate. The response saturated above 10 mM, and the reduced response at 50 mM indicated that the response was chemotaxis rather than chemokinesis (14,15).This result was confirmed in plug plate assays (Table 1) in which washed cells resuspended in HEPES as described above were mixed to approximately 2 x 109/ml with an equal volume of 0.5% agar containing 50 ,ug of chloramphenicol per ml to prevent growth. To test for competition, backgrounds of 10 or 20 mM metabolites were added to the agar. Plugs of attractant in 2.5% agar were placed in the soft agar plates, and chemotaxis was measured as cell accumulation around the plugs after approximately 2 and 5 h.R. sphaeroides showed a strong response to alanine and * Corresponding author.