1.3.2
Members of the RcnR/CsoR transcriptional repressors family detect anomalous cellular concentrations of several metals (e.g. Cu(I) or Ni(II)/Co(II) for CsoR and RcnR, respectively) in order to trigger a correct response. Allostery is exploited to link cognate metal-binding to the release of DNA operator-promoter region located upstream of the gene or operon regulated by these proteins.
According to the classical “concerted model”, also named after the researchers who first described it (Monod-Wyman-Changeux, MWC model), the regulated protein exists in two different states depending on the presence or absence of the ligand molecule acting as an effector (Monod et al. 1965). Ligand-binding to a certain subunit of the protein will trigger the conformational change that, once propagated to the other protein subunits, will result in an alteration of ligand-binding affinities for these subunits (Monod et al. 1965).
Coupling allosteric free energy ΔGc is a parameter which reports the magnitude of the allosteric
driving force in a quantitative way (Grossoehme & Giedroc 2009; Reinhart 2004). Depending on the system and protein regulator characteristics, it can be calculated by measuring metal- binding affinity of free and DNA-bound protein or otherwise by measuring DNA-binding affinity of apo- and holo-protein forms (Grossoehme & Giedroc 2009; Reinhart 2004).
23 | P a g e Grossoehme and coworkers (Grossoehme & Giedroc 2009) reported the coupled thermodynamic equilibria (assuming a closed system) which describes in a simplicistic way the relationship between protein regulator (P), n atoms of metal (n M) and a single DNA operator (D) (see Figure 3. 22A). The scheme represents the four end-states that a metalloregulator protein can adopt in the presence of its cognate metals and a limited amount of DNA (Grossoehme & Giedroc 2009). Each of these four equilibria is described by the respective affinity constant. The coupling constant (Kc) (Figure 3. 22B) is a dimensionless quantity which
represents the magnitude of allosteric regulation and can be determined either from the ratio
K4/K3 (KD Mn·P /KD P ) or the ratio K2/K1 (KM D·P /KM
P) and used to calculate ΔG
c using the standard
thermodynamic equation (Equation 1):
ΔGc = -RTlnKc Equation 1
where R is the ideal gas constant (8.314 J K−1 mol−1) and T is 298 K (temperature at which the experiments is performed). In the case of DNA-binding de-repressors such as RcnR and CsoR, where the conformational change induced by cognate metal binding leads to an assembly state with a diminished DNA-binding affinity, the ternary complex (P·M)D is less stable than P·M and free P (K3 < K1 ;K4 < K2) and Kc > 0. As a consequence the ligand exchange equilibrium
shown in Figure 3. 22B is shifted to the left (ΔGc > 0) (Grossoehme & Giedroc 2009). This
approach suggests that in de-repressors a positive increment of allosteric free energy is necessary in order to link metal-binding to the structural change in protein assembly that will result in the release of the DNA operator-promoter.
An exemplary case is given by ZiaR, the Zn(II)-sensor present in Synechosystis PCC6803. Although the organism possesses also the Ni(II)-sensor InrS, which has a KZn(II) comparable to
the tightest sites of ZiaR, the allosteric response elicited by Zn(II) in ZiaR is greater to that observed in InrS. Since both proteins are derepressors this conclusion can be derived from the simple comparison of their ΔGc
Zn(II)
. As a result, InrS is less effective than ZiaR in derepressing its target promoter region upon Zn(II)-binding, therefore to achieve a comparable degree of derepression it would need a higher concentration of Zn(II) (Foster et al. 2014b).
Another exemplary case occurs in Mycobacterium tuberculosis: the DNA-binding transcriptional repressor NmtR, senses surplus cobalt and nickel but has tighter affinity for zinc, while a related protein SmtB, detects zinc but does not detect nickel. Both of them contain a helix-turn-helix motif and the metal-sensing residues are located in the same regions of the respective folds but SmtB has only four binding residues whereas NmtR possesses six metal- binding residues. In this case the specificity for metal sensing is given by the coordination geometry preferred by the metals determining allostery. Nickel and cobalt in NmtR bind to six residues in an octahedral geometry, the geometry necessary to lead to the conformational
24 | P a g e change in the sensor that alters its binding to DNA. However, binding of zinc takes place through only four of the six residues in a tetrahedral geometry and although zinc binds to NmtR tightly it does not produce the conformational change required to drive allostery (Cavet et al. 2002; Pennella et al. 2003) and as a result repression in vivo is not reduced. Experiments conducted in vivo showed no detection by NmtR of zinc (Cavet et al. 2002; Pennella et al. 2003).
Channeling
1.3.3
Another strategy for controlling metal availability involves protein-protein contact, for example there exist metallochaperones which traffic metals acting as intermediaries providing the right metal ion to the protein. The metal acquired by a protein from the carrier protein depends more on the protein-protein interactions than the affinity of the protein for the metal. For instance in
Synechocystis sp. PCC 6803 the zinc exporter ZiaA binds copper(I) more tightly than zinc(II)
but the copper-chaperone Atx1 does not interact with ZiaA, therefore the specific Atx1 interactions prevent ZiaA from gaining access to copper(I).
Another example of how thermodynamics cannot be applied to explain metal-sensing is provided by Synechocystis CoaR. CoaR is a Co(II)-sensing activator belonging to the MerR family of transcriptional regulators (Section 1.2.2.1) and regulates the transcription of CoaT, a P1-type ATPase Co(II) exporter (Rutherford et al. 1999). Co(II)-affinity of CoaR tightest site was estimated to be approximately within the range of 8.64 nM < KCo(II) < 1 mM by UV-vis
spectroscopy and competition with the ratiometric fluorescent metal chelator Fura-2 (Fura-2;
KCo(II) = 8.64 x 10 -9
M) (Patterson et al. 2013). Since CoaR is an activator, and hence binds Co(II) when on its operator promoter, KCo(II) of the DNA-bound form of CoaR was also
investigated, however, it did not appear tighter than that of the free-form (Patterson et al. 2013). Since Synechocystis PCC 6803 possesses also metal sensors from other families (ZiaR from ArsR/SmtB family, InrS from CsoR/RcnR family, and ZuR from Fur family) which all possess a tighter Co(II)-binding affinity of their tightest sites (Patterson et al. 2013), the specificity of Co(II)-binding by CoaR cannot be explained by relative affinity. In the same work, Patterson and colleagues show by fluorescence anisotropy that Co(II) may act as an allosteric effector in
vitro also for ZiaR, InrS and Zur, in addition to CoaR (Patterson et al. 2013).
CoaR possesses a precorrin isomerase-like domain (Rutherford et al. 1999), therefore the possibility that Co(II) ions are channeled to the metal binding site of the protein was invoked to explain the specificity of the detection of cellular Co(II) by the sensor. In fact, precorrin isomerases catalyse the methyl isomerization in cobalamin biosynthesis and their substrates are cobalt-binding tetrapyrroles (Moore & Warren 2012). This observation, along with the weak Co(II)-binding affinity, suggests that CoaR may not solely detect free Co(II) ions, but either a tetrapyrrole which may aid Co(II) insertion into the protein by enhancing the KCo(II) of the site,
25 | P a g e or a pre-formed Co(II)-tetrapyrrole adduct (Patterson et al. 2013). These hypotheses were investigated by Patterson and coworkers by mutating the residues involved in the binding of tetrapyrrole substrates in precorrin isomerases and showing the consequent loss of activation of the Co(II)-dependent expression from the coaT promoter (Patterson et al. 2013).