A calculator for metalation inside a cell

Biotechnology is heavily dependent upon metal-containing enzymes since almost a half of all enzymes are estimated to require one or more bound metal. Sometimes the metals are embedded within cofactors, such as cobalt in vitamin B12, iron in heme or nickel in cofactor F430. Loading metalloenzymes, or loading the proteins that supply metals to cofactors, with the correct metal is a remarkable achievement of all living cells. Our long-term goal has been to understand intracellular metalation and to make this process predictable and exploitable. An easy-to-use, computer-based, calculator of metalation in vivo will be produced and tested.

The availabilities of metals inside cells is a dominant factor in determining which metals bind to which proteins (Nature 2008 455: 1138-1142). The intracellular milieu buffers some metals to lower availabilities than other metals and DNA-binding metal sensors are tuned to these availabilities (Nature Chemical Biology 2017 13: 409-417). Over the course of seven years, supported by the BBSRC (plus Procter and Gamble), a series of thermodynamic values were collected for a set of DNA-binding metal sensors, making it possible to calculate the intracellular availabilities of metals in a bacterium (Nature Chemical Biology 2019 15: 241-249). By reference to these values for intracellular metal-availability, it has become possible to make coarse-grain predictions about the metalation of metalloproteins, specifically identifying the right metal for a protein and not merely the tightest-binding metal. Now we will refine (and test) these predictions to understand how metalation and mis-metalation changes as a function of culture conditions. An ability to make such predictions is relevant to the optimisation of metalation in synthetic biology, is of value to industrial biotechnology, but also relevant to basic understanding of cells in systems biology and more generally.

Calculation of intracellular metal availabilities will exploit three conceptual advances: First that the cell biology of metals is rapid and associative via a labile buffer; second that the thermodynamic cycles of metal-sensors are mathematically tractable if metals are buffered (hence excluded from mass balance calculations); third that reference to these values for available metal enables calculation of the in vivo metalation state of a protein (or other molecule) that can account for inter-metal competition.

The PI and CoI Warren directed a phase I, and now phase II, BBSRC Network in Industrial Biotechnology and Bioenergy with the purpose of accelerating the exploitation of advances in metals in biology research to support bio-manufacturing and other facets of biotechnology. A purpose of the work in this proposal is to make the optimisation of metalation accessible to a platform of industries, as well as to academics, who need not be experts in the sub-discipline. Applications could include the optimisation of metalation in bioprocesses or the optimisation of mis-metalation by antimicrobial ionophores, as examples.

We will use three different systems to test three distinct types of predictions of the calculator. We have worked with all of the systems previously and obtained prior data supporting their choice for this work. One of the systems is E. coli that has been engineered by Warren to heterologously produce non-native tetrapyrroles: These tetrapyrroles include cofactor F430 and vitamin B12. The latter has immediate practical relevance. We have evidence (unpublished) that in the synthetic E. coli system the cobalt delivery protein for B12 becomes mis-metalated with zinc but that this can be overcome when cells are supplemented with surplus cobalt. A secondary purpose of this work is to manipulate this strain to enhance the supply of vitamin B12 for the increasing number of individuals turning to low meat diets (plants neither make nor require vitamin B12), in turn exemplifying the merit of the metalation calculator.

An easy-to-use, computer-based, calculator of in vivo metalation will be produced. Calculations will be based on differences between the free energy (deltadeltaG) for metal complexation with a protein, or other molecule, of interest and the free energy for (associative) available, intracellular metal: The latter will be read-out from the responses of (thermodynamically calibrated, Nature Chemical Biology 2019 15: 241-249) DNA-binding, metal-sensors; the former will be calculated from the metal affinities of a protein (or other molecule) of interest using the standard relationship, deltaG = -RTlnKA.

The calculator will predict fractional occupancies of proteins (and other molecules) with magnesium, manganese, iron, cobalt, nickel, zinc and copper. It will take into account inter-metal competition for a binding site based upon a comparison of free energy differences, deltadeltaG for each metal.

Expression of genes regulated by the metal sensors will be monitored by qPCR in media supplemented, or depleted, in each metal to calculate intracellular metal availabilities under specific growth conditions (conditional cells).

Predictions of the calculator will be tested in three different ways optimised for tight (part 2.3), intermediate (part 2.4) and for weak (part 2.5) binding metals. In each case metallo-proteins will be used that we have studied previously, Atx1, CobW, CfbA and MncA (OxdC) and for which there is already evidence to suggest mis-metalation when expressed in E. coli. Moreover, for each of the selected proteins there is also already evidence to suggest that metalation in vivo can be conditional as a function of the metal-status of the growth medium. Occupancies will be read-out directly by ICP-MS (Atx1), indirectly from metal-dependent biosynthetic products (CobW and CfbA) and directly after kinetic trapping (MncA or OxdC).

We will provide a thermodynamic framework for studying factors affecting, and for optimising, metalation.