Comparison of motility of H. pylori in broth and mucin reveals the interplay of effect of acid on bacterium and the rheology of the medium it swims in

To colonize on the gastric epithelium Helicobacter pylori bacteria have to swim across a gradient of pH from 2-7 in the mucus layer. Previous studies of H. pylori motility have shown that at pH below 4 do not swim in porcine gastric mucin (PGM) gels. To separately assess the influence of gelation of PGM and that of pH on motors and pH sensitive receptors of H. pylori, we used phase contrast microscopy to compare the translational and rotational motion of H. pylori in PGM versus Brucella broth (BB10) at different pHs. We observed that decreasing pH leads to decreased fraction of motile swimmers with a decrease in the contribution of fast swimmers to the distributions of swimming speeds and length of trajectories. At all pH’s the bacteria swam faster with longer net displacement over the trajectory in BB10 as compared to PGM. While bacteria are stuck in PGM gels at low pH, they swim at low pH in broth, albeit with reduced speed. The body rotation rate and estimated cell body torque are weakly dependent on pH in BB10, whereas in PGM the torque increases with increasing viscosity and bacteria stuck in the low pH gel rotate faster than the motile bacteria. Our results show that H. pylori has optimal swimming under slightly acidic conditions, and exhibits mechanosensing when stuck in low pH mucin gels.

140 bacteria imaged at high magnification of 100X using Resistive Force Theory (RFT) following 141 the methods of Magariyama et al [24] and found that in PGM bacteria have to generate a higher 142 torque to overcome the increased viscosity, especially at low pH. 154 k is the Boltzmann constant, T is the temperature in Kelvin, and r is the radius of the particle.
155 Fig 1A shows the <MSD> averaged over all particles in BB10 and PGM at various pH on a log-156 log plot, with water as reference. We found that the <MSD> and the viscosity of BB10 are 157 similar to those of water, and not pH-dependent (Fig. 1A, C). In contrast in PGM, it is evident 158 from the log-log plot of <MSD> that equation 1 is no longer valid and instead MSD is 159 proportional to t α . By fitting the <MSD> vs time we obtained the exponent α < 1, indicating that 160 the mobility of the micro-particles embedded in PGM is sub-diffusive, i.e. hindered at low pH as 161 a result of the sol-gel transition that PGM undergoes as pH decreases below 4 [7,26]. As pH 162 decreased from 6.7 to 3.7 in PGM, α deceased from 0.8 to 0.6, implying increasing sub-163 diffusivity of particles in PGM as it gels (Fig. 1B). In comparison to PGM, the microparticles in 164 BB10 showed normal diffusion with α = 1 (Fig. 1B). Fig. 1C shows the viscosities of BB10 and 165 15mg/ml PGM calculated using Eqn. (2). In the case of PGM the effective viscosity was 166 estimated by using only the data in the long-time regime where α ~ 1. PGM is about 50 times 167 more viscous than BB10 at pH 6, and the viscosity of PGM increased rapidly, by a factor of 2, as 168 pH decreases from pH 6.1 to 3.7, whereas the viscosity of BB10 remains constant as pH 169 decreases (Fig. 1C). Furthermore, in PGM the exponent α is time dependent, varying between 170 0.5 to 1, reflecting the frequency dependence of the moduli of the viscoelastic PGM [7,22,23].
171 The complex viscoelastic moduli were calculated and are reported in Constantino [27]. The

MSD(t) = 4D o t
(1) = 6 (2) 211 from 6.7 to 3.7 indicative of increasing heterogeneity as pH decreases. The HR data in PGM 212 shows a peak at pH 5, we suspect that this reflects a biphasic behavior of HR, with two 213 characteristic trends corresponding to the gel phase at pH 4 and lower and the solution phase at 214 pH 5 and above. From AFM measurements [6] we have shown that PGM fibers which are 215 uniformly distributed at pH 6 begin to aggregate at pH 5, perhaps this leads to an increase in 217 phase, the particles are excluded from the regions with large concentration of fiber bundles and 218 exhibit hindered motion confined in the pores of the gel. In this case the inhomogeneity will 219 depend on the extent of crosslinking and degree of gel swelling/shrinking. 238 More trajectories in PGM exhibit a 2-dimensional random walk characteristic than in BB10.
239 Trajectories with turns and reversals can also be seen. The trajectories in PGM appear more 240 helical than in BB10. We further discuss this helical feature in the last section using single 241 bacteria imaging at 100 X magnification and fast frame rate to measure rotation of bacteria as 242 they translate, similar to that reported by Constantino et al [31]. Secondly, regarding the effect 243 of pH we note that in both BB10 and PGM there are considerably fewer trajectories at the low 244 pH's as compared to the images at pH > 5 in BB10 and pH > 4 in PGM. We found that in BB10 245 the bacteria swam over the entire range of pH 3 -6.3, with a decline in the percentage of motile 246 trajectories with decreasing pH, although some bacteria became immotile and coccoidal at pH 3.
247 In contrast to this, in PGM the bacteria swam only at pH 4 and higher; there were very few 248 swimmers (~5 in total) in the pH 3.5 sample, and this data was not analyzed, while at pH 3 there 249 were no swimmers; bacteria were stuck and observed to rotate in place at pH 3.5 and pH 3. The 250 percent of motile bacteria was counted by looking at randomly selected frames in the movies. By 251 this method we estimate that in BB10 there are about 40% motile bacteria at pH 3 and 4 and 252 around 60% at pH 5 -6, whereas in PGM there were only 12% motile bacteria at pH 4, and 253 around 20-25% motile bacteria at pH 4.5, 5 and 6. The reference bacteria from the sample in 254 BB10 were also examined at the same time as the PGM measurements were being conducted and 255 these remained motile confirming that bacteria were viable and that the immotility was due to 256 gelation of PGM not due to loss of motility in the sample.