Abstract
Bacterial chemotaxis can enhance the bioremediation of contaminants in aqueous and subsurface environments if the contaminant is a chemoattractant that the bacteria degrade. The process can be promoted by traveling bands of chemotactic bacteria that form due to metabolism-generated gradients in chemoattractant concentration. We developed a multiple-relaxation-time (MRT) lattice-Boltzmann method (LBM) to model chemotaxis, because LBMs are well suited to model reactive transport in the complex geometries that are typical for subsurface porous media. This MRT-LBM can attain a better numerical stability than its corresponding single-relaxation-time LBM. We performed simulations to investigate the effects of substrate diffusion, initial bacterial concentration, and hydrodynamic dispersion on the formation, shape, and propagation of bacterial bands. Band formation requires a sufficiently high initial number of bacteria and a small substrate diffusion coefficient. Uniform flow does not affect the bands while shear flow does. Bacterial bands can move both upstream and downstream when the flow velocity is small. However, the bands disappear once the velocity becomes too large due to hydrodynamic dispersion. Generally bands can only be observed if the dimensionless ratio between the chemotactic sensitivity coefficient and the effective diffusion coefficient of the bacteria exceeds a critical value, that is, when the biased movement due to chemotaxis overcomes the diffusion-like movement due to the random motility and hydrodynamic dispersion.
Similar content being viewed by others
References
Adler J (1966) Chemotaxis in bacteria. Science 153(3737):708–716
Alexandre G, Greer-Phillips S, Zhulin I (2004) Ecological role of energy taxis in microorganisms. FEMS Microbiol Rev 28(1):113–126
Bennett S, Asinari P, Dellar PJ (2012) A Lattice Boltzmann model for diffusion of binary gas mixtures that includes diffusion slip—Bennett—2011. Int J Numer Method Fluid 69(1):171–189
Bosma T, Schnoor JL, Schraa G (1988) Simulation model for biotransformation of xenobiotics and chemotaxis in soil columns. J Contam Hydrol 2(3):225–236
Brosilow B, Ford R, Sarman S, Cummings P (1996) Numerical solution of transport equations for bacterial chemotaxis. SIAM J Appl Math 56:1639–1663
Budrene EO, Berg HC (1995) Dynamics of formation of symmetrical patterns by chemotactic bacteria. Nature 376(6535):49–53
Chen S, Doolen G (1998) Lattice Boltzmann method for fluid flows. Ann Rev Fluid Mech 30:329–364
d’Humières D (1992) Generalized lattice-Boltzmann equations. Rarefied gas dynamics—theory and simulations. In: Proceedings of the 18th international symposium on rarefied gas dynamics, Univ of British Columbia, Vancouver, pp 26–30
d’Humières D, Ginzburg I, Krafczyk M, Lallemand P, Luo L (2002) Multiple-relaxation-time lattice Boltzmann models in three dimensions. Philos Trans R Soc Lond Ser A 360(1792):437–451
Dillon R, Fauci L (2000) A microscale model of bacterial and biofilm dynamics in porous media. Biotechnol Bioeng 68(5):536–547
Duffy K, Cummings P, Ford R (1995) Random walk calculations for bacterial migration in porous media. Biophys J 68:800–806
Fenchel T (2002) Microbial behavior in a heterogeneous world. Science 296(5570):1068–1071
Grimm A, Harwood C (1997) Chemotaxis of \({P}seudomonas\) spp. to the polyaromatic hydrocarbon naphthalene. Appl Environ Microbiol 63(10):4111–4115
Hawkins AC, Harwood CS (2002) Chemotaxis of Ralstonia eutropha JMP134 (pJP4) to the herbicide 2, 4-dichlorophenoxyacetate. Appl Environ Microbiol 68(2):968–972
He X, Zou Q, Luo L, Dembo M (1997) Analytic solutions of simple flows and analysis of nonslip boundary conditions for the lattice Boltzmann BGK model. J Stat Phys 87(1–2):115–136
Hillen T, Painter K (2009) A user’s guide to PDE models for chemotaxis. J Math Biol 58:183–217
Hilpert M (2005) Lattice-Boltzmann model for bacterial chemotaxis. J Math Biol 51(3):302–332
Horstmann D (2003) From 1970 until present: the Keller–Segel model in chemotaxis and its consequences I. Jahresbericht DMV 105(3):103–165
Horstmann D, Stevens A (2004) A constructive approach to traveling waves in chemotaxis. J Nonlinear Sci 14(1):1–25
Jabbarzadeh E, Abrams CF (2005) Simulations of chemotaxis and random motility in finite domains. Cambridge University Press, Cambridge
Jabbarzadeh E, Abrams CF (2007) Simulations of chemotaxis and random motility in 2D random porous domains. Bull Math Biol 69(2):747–764
Junk M, Yang Z (2005) One-point boundary condition for the lattice Boltzmann method. Phys Rev E 72(6):066–701
Kalinin YV, Jiang L, Tu Y, Wu M (2009) Logarithmic sensing in Escherichia coli bacterial chemotaxis. Biophys J 96(6):2439–2448
Keller E, Segel L (1971a) Model for chemotaxis. J Theor Biol 30:225–234
Keller EF, Segel LA (1971b) Traveling bands of chemotactic bacteria: a theoretical analysis. J Theor Biol 30(2):235–248
Long W, Hilpert M (2007) Analytical solutions for bacterial energy taxis (chemotaxis): traveling bacterial bands. Adv Water Resour 30(11):2262–2270
Long W, Hilpert M (2008) Lattice-Boltzmann modeling of contaminant degradation by chemotactic bacteria: exploring the formation and movement of bacterial bands. Water Resourc Res 44(9):W09,415
Marcos FuHC, Powers TR, Stocker R (2012) Bacterial rheotaxis. Proc Natl Acad Sci USA 109(13):4780–4785
Marsily GD (1986) Quantitative hydrogeology: groundwater hydrology for engineers. Academic Press, New York
Marx RB, Aitken MD (1999) Quantification of chemotaxis to naphthalene by Pseudomonas putida G7. Appl Environ Microbiol 65(7):2847–2852
Marx RB, Aitken MD (2000) Bacterial chemotaxis enhances naphthalene degradation in a heterogeneous aqueous system. Environ Sci Technol 34(16):3379–3383
McCracken M, Abraham J (2005) Multiple-relaxation-time lattice-Boltzmann model for multiphase flow. Phys Rev E 71(036):701
Mei R, Shyy W, Yu D, Luo LS (2000) Lattice Boltzmann method for 3-D flows with curved boundary. J Comput Phys 161(2):680–699
Myerscough MR, Maini PK, Painter KJ (1998) Pattern formation in a generalized chemotactic model. Bull Math Biol 60(1):1–26
Nelson K, Ginn T (2001) Theoretical investigation of bacterial chemotaxis in porous media. Langmuir 17:5635–5645
Olson MS, Ford RM, Smith JA, Fernandez EJ (2004) Quantification of bacterial chemotaxis in porous media using magnetic resonance imaging. Environ Sci Technol 38(14):3864–3870
Othmer HG, Hillen T (2000) The diffusion limit of transport equations derived from velocity-jump processes. SIAM J Appl Math 61(3):751–775
Pan C, Prins JF, Miller CT (2004) A high-performance lattice Boltzmann implementation to model flow in porous media. Comput Phys Commun 158(2):89–105
Pan C, Luo L, Miller C (2006) An evaluation of lattice Boltzmann schemes for porous medium flow simulation. Comput Fluid 35(8–9):898–909
Pandey G, Jain RK (2002) Bacterial chemotaxis toward environmental pollutants: role in bioremediation. Appl Environ Microbiol 68(12):5789–5795
Parales RE, Ditty JL, Harwood CS (2000) Toluene-degrading bacteria are chemotactic towards the environmental pollutants benzene, toluene, and trichloroethylene. Appl Environ Microbiol 66(9):4098–4104
Park S, Wolanin PM, Yuzbashyan EA, Lin H, Darnton NC, Stock JB, Silberzan P, Austin R (2003) Influence of topology on bacterial social interaction. Proc Natl Acad Sci USA 100(24):13910–13915
Pedit JA, Marx RB, Miller CT, Aitken MD (2002) Quantitative analysis of experiments on bacterial chemotaxis to naphthalene. Biotechnol Bioeng 78(6):626–634
Qian Y, Succi S, Orszag S (1995) Recent advances in lattice Boltzmann computing. Ann Rev Comput PhysIII:195–242
Rivero M, Tranquillo R, Buettner H, Lauffenburger D (1989) Transport models for chemotactic cell populations based on individual cell behaviour. Chem Eng Sci 44(12):2881–2897
Samanta SK, Bhushan B, Chauhan A (2000) Chemotaxis of a Ralstonia sp. SJ98 toward different nitroaromatic compounds and their degradation. Biochem Biophys Res Commun 269(1):117–123
Saragosti J, Calvez V, Bournaveas N, Buguin A, Silberzan P, Perthame B (2010) Mathematical description of bacterial traveling pulses. PLoS Comput Biol 6(8):1–12
Saragosti J, Calvez V, Bournaveas N, Perthame B, Buguin A, Silberzan P (2011) Directional persistence of chemotactic bacteria in a traveling concentration wave. Proc Natl Acad Sci USA 108(39):16235–16240
Singh R, Olson MS (2011) Transverse mixing enhancement due to bacterial random motility in porous microfluidic devices. Environ Sci Technol 45(20):8780–8787
Taylor JR, Stocker R (2012) Trade-offs of chemotactic foraging in turbulent water. Science 338(6107):675–679
Tindall MJ, Maini PK, Porter SL, Armitage JP (2008) Overview of mathematical approaches used to model bacterial chemotaxis II: bacterial populations. Bull Math Biol 70(6):1570–1607
Tyson R, Stern LG, LeVeque RJ (2000) Fractional step methods applied to a chemotaxis model. J Math Biol 41(5):455–475
Wang X, Long T (2012) Bacterial chemotaxis toward a NAPL source within a porescale microfluidic chamber. Biotechnol Bioeng 109(7):1622–1628
Wang ZA (2013) Mathematics of traveling waves in chemotaxis-review paper. Discret Contin Dyn Syst Ser B 18(3):601–604
Widman MT, Emerson D, Chiu CC, Worden RM (1997) Modeling microbial chemotaxis in a diffusion gradient chamber. Biotechnol Bioeng 55(1):191–205
Yan Z, Bouwer E, Hilpert M (2014) Coupled effects of chemotaxis and growth on traveling bacterial waves. J Contam Hydrol 164:138–152
Acknowledgments
This work was supported by NSF #0911425. We would like to thank our colleagues Dr. Yaqi You and Dr. Shao-Yiu Hsu for their helpful suggestions and comments.
Author information
Authors and Affiliations
Corresponding author
Appendix: Chapman–Enskog Expansion
Appendix: Chapman–Enskog Expansion
Here, we present the elaborate derivation that shows the LB equations (8) and (9) approximate solutions to the continuum-scale equations (4) and (5) via a Chapman–Enskog expansion (McCracken and Abraham 2005). To simplify the notation, we first consider a generic reactive and chemotactic species that could represent either bacteria or substrate. For the special case of a substrate, \(\hat{\mathbf{v}}_\mathrm{c}= 0\) is used. We also set the reaction terms to zero, \(R_{\mathrm{b,s}}=0\). The case of \(R_{\mathrm{b,s}} \ne 0\) can be dealt with as described by Hilpert (2005).
The generic MRT-LB equation can be written as
where \(\mathsf{\Omega }\) is the collision matrix in particle space. In order to recover the MRT-LB equations (8) and (9), one just needs to insert the identity matrix \(\mathsf{M}^{-1}\mathsf{M}\) before and behind \(\mathsf{\Omega }\), and then obtains
where \(\mathsf{S}=\mathsf{M}\mathsf{\Omega }\mathsf{M}^{-1}\) is the collision matrix in moment space. This matrix is constructed to be diagonal,
In order to relate Eq. (17) to a macroscopic equation, we apply a Taylor-series expansion to the particle distribution function \(f\),
and further expand the particle distribution function
and the time derivative
We ensure that the expansion conserves mass locally by requiring that
where \(C\) is the dimensionless concentration of the bacteria or substrate, and \(N\) is the number of lattice vectors. By substituting Eqs. (19–21) into the LB equation (17), we obtain
Let us order this equation after powers of \(\delta \) and assume that the coefficients of the \(\delta ^n\) vanish. To the first order of \(\delta \), we obtain
To the second order of \(\delta \), we obtain
Multiplying (24) and (25) by the transformation matrix \(\mathsf{M}\) converts them into moment space:
Here \(\text{ A }^{(\alpha )}\) is a constant \(N\times N\) matrix, which is defined for all Cartesian directions \(\alpha \) and the components of which are given by
where \(\mathbf{e}_{i,\alpha }\) is the \(\alpha \)th Cartesian component of the \(i\)th lattice vector.
For the D3Q19 model, the lattice vectors are
We used methods from d’Humières et al. (2002)’s paper to construct the transformation matrix \(\mathsf{M}\), although we selected a different order of the lattice vectors in order to simplify numerical implementation.
This matrix transfers the particle distribution function \(f\) to the hydrodynamic moment vector \(|m\rangle \),
We only consider the first, fourth, sixth, and eighth moments, which represent the concentration and the mass flux density in the continuum-scale equations (4) and (5). Since \(\mathbf{e}_i\) and \(\mathsf{M}\) are now known, we can calculate the three \(\mathsf{A}_{}^{(\alpha )}\) matrices according to Eq. (28),
Now we evaluate the first component of Eq. (26) and get
where \(C^{(0)} = C\). To ensure that this first-order mass balance equation describes the advective transport of the concentration \(C\), the equilibrium mass flux vector must be given by
where \(\hat{\mathbf{v}}\) is the known dimensionless fluid velocity, and \(\hat{\mathbf{v}}_\mathrm{c}\) is the dimensionless bacterial chemotactic velocity. For the substrate, \(\hat{\mathbf{v}}_\mathrm{c}= 0\). The concentration change due to diffusion is quantified by the first component of Eq. (27),
In order to determine \(j_x^{(1)}\), \(j_y^{(1)}\) and \(j_z^{(1)}\), we evaluate the fourth, sixth, and eighth components of Eq. (26),
By solving Eqs. (32), (33) and (34) for \(j_x^{(1)}\), \(j_y^{(1)}\) and \(j_z^{(1)}\), and substituting the results into Eq. (31), we obtain
In order to model isotropic diffusion, mixed spatial quantities must vanish, thus \(p_{zx}^{(0)} = p_{yz}^{(0)} = p_{zx}^{(0)} = 0\). Furthermore, for the sake of symmetry, we set \(p_{xx}^{(0)} = p_{ww}^{(0)} = 0\) and use the same collision rates in all three Cartesian directions, \(s_4 = s_6 = s_8 := 1/\tau \). \(s_1=0\) is required to ensure mass conservation. All other elements of \(\mathsf{S}\) are free to choose, and they only affect numerical stability rather than mass conservation. The only remaining unknown parameter is \(e^{(0)}\). In order to obtain a similar expression for the relaxation time as the SRT-LB model, d’Humières et al. (2002) chose \(e^{(0)} = - 11C\) when applying the MRT-LB model to simulate a velocity field. However, this value makes our simulations unstable, which could be due to the fact that the momentum of neither the bacteria nor the chemoattractant is generally conserved. As a result, we choose \(e^{(0)} = - 1.5C\), which has the best stability in our parameter tests. Therefore,
Like Hilpert (2005), we assume that the velocity fields, \(\hat{\mathbf{v}}\) and \(\hat{\mathbf{v}}_\mathrm{c}\), vary slowly in both time and space dimension, such that \(\partial _0 \hat{\mathbf{v}}/ \partial t \approx 0\), \(\partial _0 \hat{\mathbf{v}}_\mathrm{c}/ \partial t \approx 0\), \(\hat{{\varvec{\nabla }}} \cdot \hat{\mathbf{v}}\approx 0\), and \(\hat{{\varvec{\nabla }}} \cdot \hat{\mathbf{v}}_\mathrm{c}\approx 0\). We can then use Eq. (29) to rewrite the mixed derivative on the right-hand-side of Eq. (36), \(\hat{{\varvec{\nabla }}} \cdot \partial _0 \mathbf{j}^{(0)}/ \partial t = - ||\hat{\mathbf{v}}+\hat{\mathbf{v}}_\mathrm{c}||^2 C\). By multiplying Eq. (36) by \(\delta \) and adding the result to Eq. (29), we obtain the transport equation accurate to second order,
where
In LBM simulations, one needs to guarantee \(||\hat{\mathbf{v}}+\hat{\mathbf{v}}_\mathrm{c}|| \ll 1\), which can be achieved by choosing a suitable typical velocity to non-dimensionalize the transport equation. Otherwise, the simulations become numerical unstable. Therefore, the second term on the right-hand side of Eq. (37) is much smaller than the first term on the right-hand side and can be ignored. When \(\delta \rightarrow 0\), the LB model achieves solutions to the following dimensionless advection diffusion equation:
where \(\hat{D} = 1/\text{ Pe }\) as in Eqs. (4) and (5). We still need to make choices for \(\psi ^{(0)}\), \(q_{x,y,z}^{(0)}\) and \(m_{x,y,z}^{(0)}\). Like d’Humières et al. (2002), we choose \(\psi ^{(0)} = -C\), \(q_x^{(0)}=-7j_x^{(0)}{}/3\), \(q_y^{(0)}=-7j_y^{(0)}/3\), \(q_z^{(0)}=-7j_z^{(0)}/3\), and \(m_{x,y,z}^{(0)}=0\). For clarity, we display the entire equilibrium moment vector of the bacteria and the substrate:
Rights and permissions
About this article
Cite this article
Yan, Z., Hilpert, M. A Multiple-Relaxation-Time Lattice-Boltzmann Model for Bacterial Chemotaxis: Effects of Initial Concentration, Diffusion, and Hydrodynamic Dispersion on Traveling Bacterial Bands. Bull Math Biol 76, 2449–2475 (2014). https://doi.org/10.1007/s11538-014-0020-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11538-014-0020-1