Bacterial Thermotaxis by Speed Modulation

Abstract

Naturally occurring gradients often extend over relatively long distances such that their steepness is too small for bacteria to detect. We studied the bacterial behavior in such thermal gradients. We find that bacteria migrate along shallow thermal gradients due to a change in their swimming speed resulting from the effect of temperature on the intracellular pH, which also depends on the chemical environment. When nutrients are scarce in the environment the bacteria's intracellular pH decreases with temperature. As a result, the swimming speed of the bacteria decreases with temperature, which causes them to slowly drift toward the warm end of the thermal gradient. However, when serine is added to the medium at concentrations >300 μM, the intracellular pH increases causing the swimming speed to increase continuously with temperature, and the bacteria to drift toward the cold end of the temperature gradient. This directional migration is not a result of bacterial thermotaxis in the classical sense, because the steepness of the gradients applied is below the sensing threshold of bacteria. Nevertheless, our results show that the directional switch requires the presence of the bacterial sensing receptors. This seems to be due to the involvement of the receptors in regulating the intracellular pH.

Introduction

Microorganisms such as bacteria sense changes in their environment, including chemical and thermal, and respond by changing their swimming pattern to facilitate their migration toward their favored region (1–8). Escherichia coli bacteria have five receptor species, which detect various chemicals (1–3) and respond differently to temperature (9–21). When swimming in a chemical (or thermal) gradient, bacteria continuously detect changes in the concentration of chemicals (or temperature). If they sense an improvement in the environmental conditions along their swimming trajectory, they respond by extending their swim in that direction, a process known as chemotaxis (or thermotaxis). This response occurs almost instantaneously over timescales as short as seconds. However, bacterial sensing and responding through this signal transduction pathway is limited, i.e., bacteria are not able to sense or respond to changes, either chemical (2) or thermal, below a certain threshold. In the case of temperature for example, when bacteria are exposed to a temperature gradient with steepness <0.02°C/μm that extends over short distances (∼100 μm), no response or directed migration is observed (Fig. S1 in the Supporting Material). Such shallow gradients are frequently found in nature, albeit they often extend over much longer distances. When encountering such conditions, the response mechanism described previously will not drive bacteria toward their favored environment.

Nonetheless, temperature and chemicals influence the cellular behavior in other regards. Temperature changes the rate of enzymatic activity, biochemical reactions, proteins' conformation and binding affinity, as well as the viscosity and pH of the environment. Certain chemicals can also affect the rate of enzymatic activity (if they are sources of energy), and can change the environment's pH. In addition, the bacterial cell is propelled by the rotation of long flagella that have a helical shape and are attached to rotary motors embedded in the cell wall. The rotation of these motors is driven by the flux of protons across the membrane. To maintain this flux, the cell needs to sustain a pH difference between the interior and exterior of the cell, which in turn requires energy. Therefore, it is expected that the speed of bacteria would be affected by all factors that influence the pH or cellular metabolism, including temperature and chemicals.

In a previous work (22), we have shown that in a shallow temperature gradient, E. coli bacteria still exhibit a directed migration that occurs over timescales as long as tens of minutes. We also showed that the direction of bacterial migration in such shallow temperature gradients depends on the chemical environment. The direction of the bacterial migration and their favored environment, however, could not be attributed to their chemo- and thermotaxis system in the classical sense as explained previously, because bacteria lacking all of their methyl accepting receptors also exhibited directed migration under such conditions. In that study, we speculated that the reason for the observed migration pattern of bacteria in shallow temperature gradients is the effect of temperature on the bacterial swimming speed, which changes their diffusivity. However, we did not have an explanation to why the effect of temperature on the speed is different in different chemical environments and how the speed can be responsible for the observed density profile along the gradient.

In this work, we simplify the problem by focusing on the effect of one chemical, serine, on the bacteria's swimming speed. Our choice of serine is because serine is known to be one of a few amino acids that allow E. coli to maintain its motility under anaerobic conditions (23,24) and has been shown to change the swimming speed of E. coli (25). In addition, serine is a strong attractant that is sensed by the most abundant chemoreceptor Tsr (9,26), and it is known to be a rich source of carbon. Our results reveal an interesting and previously undetected change in the bacterial intracellular pH that is seemingly regulated, at least in part, by the chemical- and heat-sensing receptors. Through accurate quantitative measurements, we show that this change in pH causes speed modulations that mediate bacterial thermotaxis in shallow temperature gradients. We also show that even the directional switch from heat-seeking to cold-seeking under such shallow gradients is due to the effect of the chemical environment on the bacterial swimming speed through the intracellular pH. Our measurements of the effects of temperature and serine on the swimming speed reveal a previously undetected feature. We find that the speed is a sigmoidal function of serine concentration and the difference between the maximum and minimum of the sigmoid increases with temperature. We are able to describe these results using a simple phenomenological model that separates between thermal and chemical effects.

These findings are the first experimental results, to our knowledge that suggest an alternate thermotaxis method to the classical signal transduction pathway. They also reveal the importance of the physical environment effects on cellular processes in controlling the behavior of microorganisms. Finally, the detected change in the intracellular pH could have significant consequences to the field of cellular biology due to the importance of pH in regulating many cellular and molecular processes.

Materials and Methods

Bacterial culture preparation

In all experiments described here (unless otherwise stated), bacteria carrying a plasmid expressing yellow fluorescent protein (YFP) constitutively were grown in M9 minimal medium supplemented with 1 g/L casamino acids and 4 g/L glucose (M9CG) at 30°C while shaking at 240 rpm. When the culture reached optical density (OD_600nm) of 0.1 (mid-exponential phase), the cells were washed once with motility buffer (MB: 10 mM potassium phosphate, 10 mM sodium lactate, 0.1 mM EDTA, and 1 μM L-methionine, pH = 7.0), resuspended in MB supplemented with the required concentration of serine and incubated at room temperature for ∼30 min before each experiment.

Experimental setups

Bacteria were loaded into long, thin channels (800 μm wide, 20–25 μm deep) fabricated in polydimethylsiloxane and attached to a microscope glass slide. After sealing the channels at both ends with epoxy glue, the slide was mounted onto the temperature control apparatus. To apply a constant homogeneous temperature, we used a Peltier device (25 mm × 25 mm) attached to the microscope slide via a thin copper plate (∼2 mm thick × 25 mm × 75 mm). All contacts were made using thermal grease (type Z9 heat sink compound, GC electronics). The other side of the Peltier device was attached to an aluminum block acting as a heat sink, whose temperature was fixed by water circulation.

For applying a temperature gradient, two Peltier devices separated with a 10 mm Plexiglas piece, were used to set the lower and higher temperature ends of the gradient. As in the case of the homogeneous temperature, the Peltier devices and the Plexiglas piece were attached to the microscope slide on one side, whereas the other was attached to a heat sink (Fig. 1). The linear temperature gradient was calibrated using the temperature and pH sensitive dye BCECF (Molecular Probes,) (27).

Figure 1.

Experimental setup. Two Peltier devices, 10 mm apart, were attached to a microscope slide via thin copper plates on one side and to a heat sink on the other side. The heat sink was simply composed of an aluminum plate through which water was circulated by a water bath to maintain its temperature. The polydimethylsiloxane channels were attached to the other side of the microscope slide and all the exposed spaces of the slide were then covered with Plexiglas for better thermal isolation. All contacts were made using thermal grease.

Data acquisition and analysis

To measure the bacterial speed, videos of swimming bacteria were acquired at the various conditions by fluorescence microscopy using an inverted microscope (Zeiss Axiovert 40 CFL) with a 20× objective. 20-s videos were recorded at a rate of 5 fps using a charge-coupled device camera (Progress MF, Jenoptik). Swimming trajectories were analyzed using ImageJ Particle Tracker plugin (23,28) and custom MATLAB scripts (The MathWorks, Natick, MA). The speed was then calculated from these trajectories for each condition by simply calculating the average distance traveled by the bacteria between consecutive frames. Each data point was calculated from several thousand trajectories acquired in at least three different experiments carried out under the same conditions.

To measure distribution of bacterial concentration in the temperature gradient, images of the fluorescent bacteria were acquired at different locations along the gradient using a 10× objective. The number of bacteria at each location was counted using ImageJ Particle Analyze module and then used to calculate the bacteria's concentration at that location.

pH measurement

The change in the bacterial intracellular pH was determined by measuring the change in the fluorescence intensity of YFP, which is pH sensitive as demonstrated previously (29). The change in the medium's pH was measured using the BCECF dye, whose fluorescence increases with pH. BCECF dye was added to the medium at a concentration of 200 nM. The fluorescence intensity of the BCECF dye and YFP at the different experimental conditions was measured using an Eppendorf Mastercycler ep realplex, which was used to change the samples' temperature and measure the BCECF dye and YFP fluorescence intensity simultaneously. For additional details of the measurements and calibration processes, see the Supporting Material.

Membrane potential measurement

Proteorhodopsin Optical Proton Sensor (PROPS) was expressed in bacteria using the pBAD plasmid obtained from Dr. Cohen's Laboratory (30) and following their protocol; Wild-type Escherichia coli RP437 expressing PROPS were grown in 50 mL of LB medium at 30°C while shaking at 240 rpm to early log phase (OD_600nm = 0.3–0.4). Arabinose and all-trans retinal were then added to the culture to a final concentration of 1.3 mM and 5μM, respectively. The culture was further grown in the dark for 3–4 h. Cells were then washed with 10 mM potassium phosphate buffer (pH 7) by centrifugation at 4°C, 10,000 rpm, for 5 min, resuspended in 3 ml of the testing buffer, and incubated on ice for ∼30 min.

Before the measurement, nigericin was added to the culture to a final concentration of 1 μM to eliminate the effect of the pH (31), and cells were incubated for at least 7 min at each temperature to allow them to reach steady state. At the end of each run, carbonyl cyanide 3-chlorophenylhydrazone was added to the mix at a final concentration of 50 μM to set the membrane potential to zero (32) as a reference point. Spectroscopic measurements of PROPS were carried out using the Tecan Infinite M200 microplate reader. The cells were excited at 570 nm and the fluorescence emission was scanned between 630 and 780 nm. Maximum emission was observed at 718 nm (±6 nm) and was used to estimate the membrane potential following the calibration curve described in the Supporting Material.

Results and Discussion

The effect of serine on the bacterial behavior in a shallow temperature gradient

The concentration of different attractants in the cell's surrounding can alter its response to temperature changes by increasing the receptors' methylation state (15–19). For example, at serine concentrations as high as 600 μM, bacteria do not respond to temperature changes (Fig. S1), whereas without nutrients in the medium bacteria always go to the high temperature (Fig. S1). On the other hand, when serine concentration is increased to 2 mM, and aspartate is added to a concentration of 3 mM, the bacteria escape from the heated region toward lower temperature (Fig. S1).

In addition to chemicals, the behavior of bacteria in a temperature gradient is affected by the steepness of the gradient. When the steepness of the gradient that extends over 100 μm is <0.02°C/μm, no migration is observed toward the warm end of the gradient indicating that cells do not respond to temperature changes on that scale (Fig. S1). However, in a previous study we have observed directional migration of bacteria in temperature gradients whose steepness is one order of magnitude smaller, but which extend over longer distances (∼10 mm) (22). The main difference between these two scenarios is the distance over which the gradient extends. In the first scenario, the difference in temperature between the two extremes of the gradient is too small to cause any significant change in the cellular processes. On the other hand, when the gradient extends over long distances of the order of magnitude of ∼10 mm, the difference in temperature that bacteria experience along the gradient is as high as 20°C. This can in turn cause a significant change in many cellular processes as explained earlier.

We have studied the behavior of the wild-type E. coli bacteria RP437 in a simple chemical environment (MB) without any nutrients added, under a shallow temperature gradient between ∼22 and ∼43°C extending over a distance of 10 mm as described in the Materials and Methods. Under such conditions the steepness of the gradient is 0.002°C/μm. Consistent with previous observations, the bacteria accumulate at the warm end (Fig. 2 A). However, when serine is added to the medium, the direction of bacterial migration in the temperature gradient is inverted. At serine concentrations higher than 300 μM, the bacteria accumulated at the cold end of the gradient (Fig. 2, B and C). This response shift is not due to the change in the methylation state of the receptors, because at serine concentrations of ∼300–600 μM, bacteria lose their ability to sense temperature changes (Fig. S1). In addition, when the nonmetabolizable form of serine (α-methyl-DL-serine) is added to the medium, the response is not altered (results not shown).

Figure 2.

The effect of serine concentration on the direction of bacterial migration in a shallow temperature gradient. Examples of concentration profiles of the bacteria, without serine added to the medium (A) and with 600 μM serine (B) measured at different times after applying the temperature gradient. The concentrations were measured as explained in the Materials and Methods. All measurements were normalized by the initial concentration at each location to allow better comparison of different experiments. (C) The shift of the bacterial population's center of mass as a function of serine concentration calculated after ∼45 min from applying the temperature gradient. A positive shift indicates migration to the right (higher temperature), whereas a negative shift indicates migration to the left (lower temperature). Each point on the graph is the average of at least three different experiments, and the error bars represent the standard deviation between experiments. Note that the error bar of the 300 μM measurement is very large. That is because some of the experiments exhibited a shift toward higher temperature, whereas in other experiments the bacteria migrated toward a lower temperature.

To understand this phenomenon, we consider the bacteria as simple random walkers whose run time and speed depend on temperature. Theoretical analysis of such systems in one dimension revealed that the steady-state concentration profile of the random walkers is inverse proportional to their speed profile (33,34), i.e.,

$$\rho \left(x\right)={\rho }_0\frac{v_0}{v\left(x\right)}.$$ (1)

This suggests that it is possible that the bacterial distribution depicted in Fig. 2 is due to the effect of temperature on the swimming speed of the bacteria. To test this hypothesis, we measured the bacterial swimming speed as a function of serine concentration in the medium at different background temperatures as described in the Materials and Methods. Note that our method for measuring the bacterial speed is different from that used by other groups (20,35). We do not separate the bacterial motion into two components of run and tumble because even when a bacterium is changing direction it can still be moving, and the speed of the bacteria during that period should also be considered for the purpose of comparing with Eq. 1. Our results, presented in Fig. 3 A, show that the bacterial swimming speed exhibits a sharp increase as a function of the serine concentration for all temperatures. The increase in the speed occurs always at ∼250 μM serine and can be described to a very good approximation by a sigmoidal function:

$$v\left(T,S\right)=v_{\mathit{min}}\left(T\right)+\frac{v_{\mathit{max}}\left(T\right)-v_{\mathit{min}}\left(T\right)}{1+\text{exp}\left[\frac{-\left(S-S_H\right)}{S_0}\right]},$$ (2)

where v_min(T) and v_max(T) are temperature-dependent functions that describe the speed of the bacteria at low and high serine concentration, respectively. Fitting this equation to our experimental data results in the values for S_H and S₀ given in the caption of Fig. 3. Even though S_H and S₀ are constants that do not depend on temperature, the speeds of the different modes, v_min(T) for low serine concentration and v_max(T) for high serine concentration, change differently with temperature (Fig. 3 B). At high serine concentration, the speed increases continuously with temperature, whereas at low concentration the speed increases initially with the temperature but decreases when the temperature becomes higher than ∼30°C. The increase of the speed as a function of serine concentration has been reported before (25). This increase is not due to widening of the speed distribution, which would indicate an increase in the speed of a fraction of the population or an increase in the speed of the whole population for part of the time, but rather a shift in the distribution (Fig. S2), which suggests that the rotational speed of all flagellar motors increases with the serine concentration.

Figure 3.

The effects of serine concentration and temperature on the bacterial swimming speed. (A) The swimming speed of bacteria as a function of serine concentration for different temperatures as indicated in the legend. Note that the increase in the speed occurs around the same serine concentration for all temperatures. The lines depict the function: $v\left(T,S\right)=v_{min}\left(T\right)+[v_{max}\left(T\right)-v_{min}\left(T\right)]/[1+\exp \left[-\left(S-S_H\right)/S_0\right]],$ with S_H ≈ 250 mM, S₀ ≈ 30 mM, and v_min(T) and v_max(T) are temperature-dependent functions that represent the speed of the bacteria at low and high serine concentration, respectively. Here, v_min(T) was calculated by averaging the speed measured at 0 and 100 μM serine, whereas v_max(T) was calculated by averaging the speed measured at 400, 500, and 600 μM serine. To get a better sense of how v_min(T) and v_max(T) depend on temperature, the speed of bacteria as a function of temperature (in MB without serine and with 600 μM serine) is presented in (B). Each point in these graphs was calculated from a few thousand trajectories acquired in at least two different experiments.

Finally, a comparison between the speed profile as a function of temperature at the different serine concentrations and the bacterial concentration profiles presented earlier in Fig. 2 reveals a consistency between our experimental data and the relationship presented in Eq. 1 (Fig. S3). The similarity between the inverse speed profile and the concentration profile, suggests that the effect of temperature on the bacterial swimming speed could indeed be the driving mechanism of bacterial thermotaxis in shallow temperature gradients, where the steepness of the gradient is lower than the sensitivity threshold of the bacteria. However, the question remains: what is the reason for the different temperature dependence of the swimming speed at low and high serine concentrations?

The effect of serine and temperature on the bacterial membrane potential and intracellular pH

To answer the question presented at the end of the previous subsection, we invoke the fact that the swimming speed of the bacteria is dependent on the rotational speed of the flagellar motor. We also consider that the flagellar motor is driven by the proton motive force (PMF), which is a combination of the membrane potential and the intracellular-extracellular pH difference (36):

$$\Omega =\alpha \left(f\right)\left[\Delta \psi -2.3\left(\frac{RT}{F}\right)\Delta pH\right],$$ (3)

where Ω is the angular velocity of the motor, α(f) is a constant that depends on the frictional drag coefficient, f, Δψ is the membrane potential, R and F are the gas and Faraday constants, respectively, T is the temperature in °K, and $\Delta \psi -2.3\left(RT/F\right)\Delta pH$ is the PMF. Therefore, a change in the swimming speed of the bacteria is a result of a change in the membrane potential and/or the intracellular-extracellular pH difference.

To determine what changes in the membrane potential and/or pH, are responsible for the observed change in the swimming speed, we measured both as a function of temperature and serine concentration. Fig. 4 shows that the membrane potential (measured using PROPS as described in the Materials and Methods) does not change when the serine concentration is changed, but it increases with the temperature up to ∼35°C and decreases afterward (see also (32), Fig. 10)). On the other hand, our measurements of the intracellular and extracellular pH, using YFP and BCECF, respectively, as described in the Materials and Methods, show that although the extracellular pH is affected by temperature only, the intracellular pH changes with serine as well. Fig. 5 shows that the intracellular pH exhibits a sigmoidal-like behavior as a function of serine concentration similar to the behavior detected in the swimming speed (Fig. 3 A). The increase in the intracellular pH occurs around similar values of serine concentration ∼250 μM, and the difference between the intracellular pH at high and low serine concentrations is also temperature dependent as exhibited by the swimming speed. Moreover, the intracellular-extracellular pH difference for low and high serine concentrations are almost the same at low temperature up to ∼30°C, but the difference between the two increases for temperatures higher than that (Fig. 6 A). Finally, the PMF, calculated using the measurements of the membrane potential and pH, show that at low serine concentration it increases with temperature up to 30°C and decreases afterward, whereas for high serine concentration it continuously increases with temperature in good agreement with the measurements of the bacterial swimming speed (Fig. 6 B and Fig. 3 B, respectively). The bacterial swimming speed calculated using these values of the PMF is in good agreement with the measured values presented in Fig. 3 B (Fig. S4).

Figure 4.

The effects of serine and temperature on the membrane potential. (A) The membrane potential of bacteria as a function of serine concentration for different temperatures as indicated in the figure legend. Note that the membrane potential is not affected by an increase in the serine concentration at different temperatures. (B) The membrane potential of bacteria as a function of temperature at different serine concentrations as indicated in the figure legend. The membrane potential increases up to 35°C and decreases above that temperature for all serine concentrations. All measurements were carried out as described in the Materials and Methods.

Figure 5.

The effect of serine on the intracellular pH. (A) The intracellular pH at different temperatures as a function of serine concentration in the medium, measured as described in the Materials and Methods. (B) The difference between the intracellular and extracellular pH as a function of serine at different temperatures calculated using the measurements in (A) and Fig. 6A inset. The lines in the graphs are to guide the reader.

Figure 6.

The effect of temperature on the pH difference and the PMF. (A) The intracellular pH as a function of temperature with and without 600 μM serine. The pH in both cases is almost the same up to 30°C, after which it starts increasing with temperature when serine is present in the medium and decreasing when no nutrients are added. The extracellular pH on the other hand decreases with temperature exactly the same with and without serine as depicted in the inset. (B) The PMF calculated as defined in Eq. 3 using the measurements of the membrane potential and the pH described previously in Fig. 4B and Fig. 6A, respectively.

The change in the intracellular pH detected here is very surprising, and it seems contradictory to previous observations of cellular pH homeostasis (37). However, it is important to note that the MB used in all measurements does not contain a carbon source and does not support cellular growth or protein production, and therefore, it is possible that the cell in such an environment is not capable also of regulating and maintaining a stable cytoplasmic environment. As a result, in MB the intracellular pH behaves in a similar manner to the extracellular pH and therefore decreases with temperature (Fig. 6 A, inset). On the other hand, serine as mentioned earlier is known to be a bacterial source of carbon that can be used to maintain bacterial motility under anaerobic conditions (23,24). Therefore, when it is added to the medium perhaps the bacteria are able to maintain a stable intracellular environment and increase their pH close to its normal values. This increase however, occurs only for temperatures higher than 24°C possibly because of the reduced rate of metabolism at a lower temperature. This hypothesis is further supported by the fact that this effect is not detected when a nonmetabolizable form of serine is used. To understand the exact mechanism that controls this observed change in the intracellular pH, further investigation is required, but these results provide an important insight that can also be useful for understanding the mechanism bacteria use to regulate their intracellular pH.

Summary and Conclusions

Temperature is one of the most important factors that influence living organisms. All biochemical reactions are affected by temperature, and therefore, it is important for all living organisms to maintain a stable body temperature optimal for their survival. Multicellular organisms are generally able to regulate their body temperature and thus are able to thrive in a wide range of temperatures. On the other hand, single-cell microorganisms whose body temperature is set by the environment strive to migrate toward regions where the temperature is optimal for their growth and metabolism. For that purpose, single-cell microorganisms, such as E. coli bacteria have developed sensory networks that allow them to sense and respond to thermal cues and direct their migration toward their favored environment. However, active response is not the only mechanism whereby bacteria can migrate in a temperature gradient.

We have shown here that in a shallow temperature gradient, where bacteria are not able to respond to changes in temperature along their run path through the known chemotaxis and thermotaxis signal-transduction pathway, they still exhibit a directed migration along the temperature gradient. This migration we find is due to the effect of temperature on the bacterial swimming speed. Because bacteria spend more time where their speed is lower they tend to accumulate in that region. Moreover, we find that the direction of bacterial migration is affected by the concentration of serine in the environment. Our results show that the bacterial swimming speed increases with the serine concentration following a sigmoidal function with constant kinetic parameters, S_H and S₀ (Fig. 3), and a temperature-dependent minimum and maximum, v_min(T) and v_max(T). The temperature-dependent nature of v_min(T) and v_max(T), at high and low serine concentration, respectively, is the reason for the change in the direction of bacterial migration. We also find that the difference observed between v_min(T) and v_max(T) is due to the increase in the bacteria's intracellular pH caused by serine at high temperature.

A question that these results immediately raise is whether receptorless bacteria respond similarly to such temperature gradients under the same chemical conditions. In a previous study we showed that indeed the mutant strain bacteria UU2612 (38) lacking all chemoreceptors migrate toward high temperature in MB. However, when serine is added to the medium we do not observe a similar migration toward low temperature. This could indicate that the observed migration toward low temperature is not a result of the speed dependence on temperature but rather due to direct sensing of the temperature gradient via the chemoreceptors. Yet, measuring the change in the intracellular pH of this mutant strain as a function of temperature and serine, reveal that the intracellular pH does not exhibit a similar behavior. Our results show that in the absence of the chemoreceptors, the intracellular pH decreases with temperature similarly with and without serine in the medium (results not shown). Tsr, alongside Tar and Aer, have been shown to respond to changes in the intracellular pH (39–41). However, this is the first observation, to our knowledge, that suggests that the chemoreceptors might be involved in regulating the intracellular pH, as well. The mechanism underlying the intracellular pH regulation and the role of the receptors in this process are still unclear and subject to an ongoing investigation. Nevertheless, this finding is very important as it suggests that the chemoreceptors could be involved in alternate signal transduction pathways different than the one already known.

In conclusion, our results show that there is more than one way for bacteria to sense temperature gradients. When the temperature gradient is not steep enough to be detected by the classical bacterial signal transduction pathway, the change in the speed (as a result of changes in the intracellular pH, which also seems to be regulated by the chemoreceptors) allows the bacteria to migrate in such environments. This improves the ability of bacteria to respond to thermal gradients.