The paper in Nature Communications is here: http://go.nature.com/2BBHYw0

Fine control of metal concentrations is necessary for cells to discern zinc from cobalt    

The Robinson lab at Durham University has been finding DNA-binding metal-sensing transcriptional regulators and exploring how they work for twenty five years (Molecular Microbiology 1993, 7 177-187). Now we understand how they discern one metal from another despite being unable to do so at metal-binding. Metal sensors support the control of cellular metal levels which is necessary for correct protein metalation (Nature 2009, 460 823-830). Helping metalloproteins to bind the right metals is a vital, underappreciated, role of all living cells.  

Implications and applications of understanding and manipulating protein metalation are supported by an Industry-Academia network led from the Universities of Durham and Kent in the UK (Metals in Biology BBSRC NIBB).

About a half of the reactions of life are catalysed by metals but bioinorganic chemistry poses a challenge. Because proteins are somewhat flexible they are imperfect at selecting metals such that all proteins bind metals with the same order of affinity, regardless of which metal(s) they need. This binding order, the Irving-Williams series, was reported some sixty years ago (Nature 1948, 162 746-747): For example copper and zinc will bind thousands of times more tightly than magnesium, manganese or ferrous iron.

Mg²⁺ < Mn²⁺ < Fe²⁺ < Co²⁺ < Ni²⁺ < Cu²⁺ (Cu⁺) > Zn²⁺

The Irving Williams series from weak binding Mg²⁺ to tight binding Cu²⁺

Two previous observations from our research group are that metal availability where proteins fold determines which metals bind, not absolute metal preference (Nature 2008, 455 1138-1142), and secondly that metal sensors are tuned to these vital, buffered metal concentrations (Nature Chemical Biology 2017, 13 409-414). Cells maintain less competitive metals at higher concentrations than more competitive metals to ‘level the playing field’. But this also means that many proteins, including the sensors, are liable to mismetalation rendering cells susceptible to disruption of metal-availability. We wonder if this might be a microbial ‘Achilles heel’ explaining why metals, chelants and ionophores have so often been chosen, albeit empirically, as antimicrobials in agriculture, consumer goods and healthcare, plus indeed by evolution in so-called nutritional immunity (Advances in Microbial Physiology 2017, 70 entire volume).

The factors that determine which metals a sensor detects have been (alliteratively) summarised as affinity, allostery, access and abundance. Quantitatively combining these parameters has been a challenge. Our paper sets out computational methods and scripts to do this. We took advice from knowledgeable colleagues such as David Giedroc (Indiana) and Tony Wedd (Melbourne) in order to carefully measure multiple thermodynamic parameters of two zinc sensors, a cobalt sensor, a formaldehyde sensor and a mutant of the formaldehyde sensor from Salmonella Typhimurium. As a word of caution, many reported protein-metal affinities are erroneous and we encourage you to take a look at Natural Products Reports 2010, 27 768-789 for some of the reasons why. We worked with skilful collaborators at Procter and Gamble in Cincinnati to determine the number of molecules of each sensor in Salmonella cells. Data in hand, it was then possible to solve a series of coupled equilibria and calculate how much of each sensor should bind to DNA at different intracellular concentrations of cobalt or zinc. At first glance the results appeared wrong with the cobalt sensor about a hundred times more sensitive to zinc than to cobalt. However, we now understand that metal-sensors are ‘attuned’ to ‘buffered’ concentrations of their cognate metals (Nature Chemical Biology 2017, 13 409-414), and zinc is buffered below the zinc-sensitivity of the cobalt sensor. Equivalent calculations for the zinc sensors reveal the lower buffered concentration of zinc.

At last, metal specificity becomes comprehensible with the correct sensor for each metal being the one attuned to the buffered concentration of that metal. This tuning can be predicted from equilibrium thermodynamics, suggesting a swift associative cell biology of metals and paving the way to discover the buffered concentrations of all metals inside a bacterial cell.

The calculations show that the cobalt sensor should respond to zinc, and the zinc sensors should respond to cobalt, if the intracellular concentrations increase above the buffered concentrations by just an order of magnitude or so. Suggestions from a referee encouraged us to do quantitative PCR which confirmed that short metal-shocks cause aberrant responses to the wrong metals exactly as predicted: Care must also be taken when assigning metal specificity to metal sensors! As reflected in the title of the paper, for cells to discern between metals there must be fine control of the buffered metal concentrations inside cells and this is susceptible to subversion.      

The work is a collaboration between Bioscientists and Chemists at Durham University along with Industry researchers at Procter and Gamble, funded by the Biotechnology and Biosciences Research Council (BBSRC Industrial Partner award BB/J017787/1).     

Deenah Osman and Nigel Robinson, Departments of Biosciences and Chemistry, Durham University, UK