The thermal impulse response of Escherichia coli

Abstract

Swimming Escherichia coli responds to changes in temperature by modifying its motor behavior. Previous studies using populations of cells have shown that E. coli accumulate in spatial thermal gradients, but these experiments did not cleanly separate thermal responses from chemotactic responses. Here we have isolated the thermal response by studying the behavior of single, tethered cells. The motor output of cells grown at 33 degrees C was measured at constant temperature, from 10 degrees to 40 degrees C, and in response to small, impulsive increases in temperature, from 23 degrees to 43 degrees C. The thermal impulse response at temperatures < 31 degrees C is similar to the chemotactic impulse response: Both follow a similar time course, share the same directionality, and show biphasic characteristics. At temperatures > 31 degrees C, some cells show an inverted response, switching from warm- to cold-seeking behavior. The fraction of inverted responses increases nonlinearly with temperature, switching steeply at the preferred temperature of 37 degrees C.

Fig. 1.

CCW bias versus temperature. The motor output of single, tethered, wild-type cells, grown at 33°C, was measured for at least 10 minutes. Each data point is the mean of the average CCW bias per cell. The error bars are the weighted standard deviation. The number of cells assayed at each temperature, from left to right, are: 82, 69, 72, 53, 60, 68, 45, and 59.

Fig. 2.

Impulse responses at 31°C. (a) The points are the CCW biases averaged per wild-type cell in response to an impulse of heat. The thick curve is the CCW bias averaged over all cells. The ratio of the area of the positive to the area of the negative lobe is 0.4. The data were computed from 802 cycles from 21 cells. Each cycle lasted 10 seconds, and the laser pulse started at 0 s and lasted for 50 ms. (b) The double impulse response. The open diamonds show the response when the laser was pulsed twice, once at t=0 sec and again at t=0.66 sec. The response from a single pulse is shown by dots. The expected double impulse response assuming linearity is shown as a solid line. The single impulse response data were computed from 24 cells with a total of 1,075 cycles. The double impulse response data were computed from 34 cells with a total of 1,378 cycles. The laser power was half of that used in a. (Inset) Step response was computed from 16 cells with a total of 359 cycles. Bar indicates when laser was turned on. Laser power was approximately 1/60th the power used in a.

Fig. 3.

Inverted impulse responses. The open diamonds indicate the impulse response from wild-type cells adapted to 0.1 mM α-methyl-dl-aspartate and 0.1 mM l-serine at 31°C (9 cells, 328 cycles). The dots indicate the impulse response from wild-type cells adapted to a temperature of 40°C (23 cells, 683 cycles).

Fig. 4.

Impulse response of a single cell at 30°C and 40°C. The impulse response of a single wild-type cell at 30°C is shown by the dots, and at 40°C by the open diamonds. Note that by looking at the direction of the first lobe, the response at 30°C is positive while the direction of the response at 40°C is negative. A total of 94 cycles were measured. Cycles 1–29 and 60–94 were taken at 40°C, while cycles 30–59 were taken at 30°C. The system was given 15 minutes to equilibrate between temperature changes. Data were smoothed with a 120-ms averaging window.

Fig. 5.

Directionality of thermal response versus temperature. Dots indicate the fraction of cells with a positive-directed (warm-seeking) thermal impulse response, a positively directed first lobe. Diamonds indicate the fraction of cells with a negative-directed response, which is simply 1 − (positive fraction) but is plotted to help guide the eye. If the directionality of the impulse response was ambiguous, those data were not considered. The number of cells assayed at each temperature, from left to right: 20, 19, 21, 37, 25, 34, 32, 20, 23, 8.