Avian MRP126 Restricts Microbial Growth through Ca(II)-dependent Zn(II) Sequestration

Abstract

The calgranulins form a class of S100 proteins in higher vertebrates that innate-immune cells release in abundance at infection sites. These proteins function by binding transition metal ions to prevent microbial pathogens from obtaining those essential nutrients. Mammals express three distinct members of this family, S100A8 (calgranulin A), S100A9 (calgranulin B, which heterooligomerizes with S100A8 to form calprotectin), and S100A12 (calgranulin C), that exhibit Ca(II)-dependent transition-metal-binding properties. Human calprotectin effectively sequesters Mn(II), Fe(II), Ni(II), and Zn(II), whereas human S100A12 selectivity sequesters Zn(II) over these other metal ions. Birds and reptiles express a single calgranulin homolog named MRP126, which we reasoned could have properties more similar to either calprotectin or S100A12. Here we present the purification and biophysical characterization of chicken MRP126 and, to the best of our knowledge, provide the first assessment of the metal-binding and antimicrobial properties of an avian MRP126. We show that MRP126 is a homodimer that selectively sequesters Zn(II) and restricts the growth of certain microbes. MRP126 binds Zn(II) at two canonical His₃Asp sites. The presence of excess Ca(II) increases the affinity of the His₃Asp sites from the low-nanomolar to the low-picomolar range, thereby enhancing antimicrobial activity. Chicken MRP126 also binds additional Zn(II) equivalents with low-nanomolar affinity at two non-conserved dicysteine sites and with high-nanomolar affinity using a histidine-rich C-terminal tail that is a hallmark of this clade of calgranulins. Our results with chicken MRP126 suggest that Ca(II)-dependent Zn(II) sequestration was a role of the last common ancestor of calgranulin proteins, with mammalian calprotectin subsequently evolving a broader metal-binding repertoire.

Graphical Abstract

Introduction

The S100 protein family comprises numerous small (10–15 kDa) α-helical proteins, typically found as dimers, that use EF-hand motifs to bind Ca(II).¹ Specific to chordates, S100 proteins fulfill a variety of biological functions, including intracellular regulatory roles and extracellular cell-to-cell signaling that leverage the ability of these proteins to sense Ca(II).^{2, 3} The calgranulins, which form a subgroup of the S100 family, include the mammalian homologs S100A8 (calgranulin A, also known as myeloid-related protein (MRP) 8), S100A9 (calgranulin B, MRP14), and S100A12 (calgranulin C).⁴ The S100A8 and S100A9 polypeptides can heterodimerize and -tetramerize to form calprotectin (CP).⁵ Many bird and reptile species possess a single calgranulin homolog named MRP126.⁶ These proteins likely all diverged from a single protein present in the last common ancestor of higher vertebrates, although the precise order of gene duplication and speciation leading to the current distribution of homologs remains uncertain.⁶

Calgranulins are abundantly expressed in and secreted by white blood cells, especially neutrophils, where they contribute to the innate-immune response to infection.^{7, 8} One conserved function involves the activation of pro-inflammatory signaling pathways and thus the recruitment of additional innate-immune cells to infection sites.^9–12 For example, S100A12 and CP (as well as S100A8 and S100A9 homodimers) have been reported to activate Toll-like receptor 4 (TLR4)^13–15 and the receptor for advanced glycation end products (RAGE)^15–18 to ultimately stimulate cytokine production.

As a second important function, calgranulins inhibit the growth of microbial pathogens by restricting the availability of essential transition metal nutrients as part of the innate-immune system’s strategy of “nutritional immunity.”^{5, 19–23} Studies with human (hCP) and murine (mCP) homologs of CP demonstrated its ability to bind numerous divalent first-row transition metal ions (including Mn, Fe, Ni, and Zn) with high affinity and thereby starve pathogens of those essential nutrients.^24–31 This versatility depends on a biologically unprecedented hexahistidine (His₆) metal-binding site formed at the heterodimer interface.^{26–28, 30, 32, 33} CP also contains a second high-affinity metal-binding site comprised of three histidines and an aspartate (His₃Asp) that sequesters Zn(II) but not Mn(II), Fe(II) and Ni(II).^{25, 28, 29, 30} The S100A12 homodimer contains two His₃Asp metal-binding sites^{34, 35} that provide Zn(II)-selective metal sequestration.³⁶ Because of its ability to sequester multiple nutrient metal ions, CP restricts the growth of a broader range of microbes than does S100A12.^{24, 25, 28, 36}

Studies of CP and S100A12 show that Ca(II) ions play an important role in the behavior of these proteins. In particular, coordination of Ca(II) ions to the EF-hand domains enhances the metal-binding affinities and antimicrobial activity of both CP and S100A12. These observations form the basis of a model where physiological extracellular concentrations (≈2 mM) of Ca(II) increase the affinity of CP and S100A12 for transition metal ions^{25, 29, 33, 37} and thereby improve their antimicrobial potency.^{25, 28, 36} For CP, Ca(II) binding at the EF-hands or transition-metal binding at the His₆ site also causes two heterodimers to dimerize, thus forming a heterotetramer.^{25, 38, 39} The Ca(II) concentration required for tetramerization varies between mCP and hCP, with a significantly higher amount needed for mCP.⁴⁰ For S100A12, higher-order oligomerization states have also been observed in the presence of Zn(II) (tetramers) and Zn(II) and Ca(II) (higher-order oligomers, which may include tetramers or hexamers).^{41, 42}

To improve our understanding of calgranulin evolution, we aimed to evaluate the biochemical, biophysical properties, transition-metal binding ability, and antimicrobial activity of a bird or reptile MRP126 and thereby help establish the biological role of that branch of the calgranulin clade. As birds and reptiles typically have only a single calgranulin homolog, we reasoned that MRP126 likely functions analogously to either mammalian CP or S100A12.

We selected chicken MRP126 for this work because previous in vivo studies suggested it had a role in the chicken innate-immune response to bacterial infection. In particular, MRP126 was found to be highly abundant in splenic heterophils (the functional equivalents of neutrophils in birds) and macrophages.^43–45 In addition, MRP126 RNA and MRP126 protein expression increased 40- and 300-fold, respectively, in the chicken cecum following oral infection with Salmonella enterica serovar Enteritidis.^{44, 46} Inoculation of chicks with adult chicken microbiota also stimulated MRP126 expression in the cecum.⁴⁷ Although we are unaware of any direct evidence to date demonstrating an antimicrobial function for MRP126, these findings are consistent with MRP126 performing a similar innate-immune function to that of mammalian CP and/or S100A12. Moreover, chicken MRP126 was previously purified from Escherichia coli using a polyhistidine affinity tag⁶ and shown to stimulate chicken TLR4 to activate NF-κB, suggesting a conservation of a pro-inflammatory signaling role within the calgranulin family.⁶

From sequence comparisons of chicken MRP126 to mammalian S100A8, S100A9, and S100A12 proteins (Figure 1A), each chicken MRP126 monomer contains two Ca(II)-binding EF-hand motifs. Each MRP126 homodimer likely possesses two transition-metal-binding sites at the dimer interface, as the three histidine and aspartate residues that form the His₃Asp site in CP and S100A12 are present in MRP126. In addition, chicken MRP126 has a histidine-rich C-terminal tail, a common feature of avian/reptile calgranulin homologs (Figure S1). We hypothesized this tail could allow MRP126 to form a six-coordinate metal-binding site analogous to the His₆ site of CP, which includes four histidine residues from the S100A8/S100A9 interface (His₄ site) and two histidine residues from the C-terminal tail of S100A9. To further explore potential MRP126 metal-binding sites, we generated a homology model (Figure 1B) of the protein with SWISS-MODEL⁴⁸ using an NMR solution structure of the Ca(II)-bound human S100A9 homodimer as the template (PDB: 5I8N).⁴⁹ From this structural model, it appears likely that in MRP126, His26, Asp36, His96, and His100 form two His₃Asp metal-binding sites at the dimer interface, whereas the histidine-rich C-terminal tail (containing His105, His109, His111, His113, His115, His117, and His119) could either contribute additional ligands to these sites or form separate metal-binding sites, perhaps also including His98 and Cys102. In this work, we aimed to test these hypotheses and assess how the functional properties of MRP126 compare to previously-studied mammalian calgranulins.

Figure 1: MRP126 is an avian homolog of mammalian S100A8, S100A9, and S100A12.

(A) Sequence alignment of human (h) and mouse (m) S100A8, S100A9, human S100A12, and chicken (c) MRP126. The locations of the helices seen in calprotectin (PDB: 5W1F³⁰) and S100A12 crystal structures (PDB: 2WC8³⁵) are indicated above the alignment. Known/expected transition-metal-binding residues are highlighted in orange, Ca(II)-binding residues are purple, all cysteines as well as H98 in MRP126 are cyan, and the seven C-terminal tail histidine residues in MRP126 are green. (B) Homology model of homodimeric MRP126 generated with SWISS-MODEL⁴⁸ using an NMR solution structure of the Ca(II)-bound human S100A9 homodimer as the template (PDB: 5I8N⁴⁹). The predicted His₃Asp sites are highlighted in orange in the left panel and the N- and C-termini of each monomer are indicated. In the right panel, a zoomed-in view of the top half of the homology model includes side chains for the residues highlighted in the sequence alignment (same color code) as well as a cyan sphere for the position of N-terminal residue (G9), which we highlight because C5 precedes this position and thus is not included in the model. See also Figure S1 for a sequence logo generated from 118 avian/reptile MRP126 homologs.

Here we present the purification and characterization of chicken MRP126. We show that MRP126 is an α-helical homodimer that sequesters Zn(II) and exhibits similar antimicrobial properties to human S100A12. MRP126 fulfils this role by binding two equivalents Zn(II) at the conserved His₃Asp sites with Ca(II)-dependent affinities. MRP126 also binds additional equivalents of Zn(II) with lower affinity at non-conserved sites involving two cysteines and its histidine-rich C-terminal tail. These results suggest that Zn(II) sequestration is the ancestral role in this clade of the S100 family, with mammalian CP likely developing the broader metal-withholding repertoire that underlies its more potent antimicrobial ability after the evolutionary split between the ancestors of modern mammals and those of birds and reptiles.

Materials and Methods

Complete experimental details are provided as Supporting Information.

Results

Purification and characterization of chicken MRP126 and variants.

Guided by our prior work with mammalian S100 proteins,^{25, 32, 36, 40, 50} we overexpressed MRP126 in E. coli from a codon-optimized gene inserted into pET41a. SDS-PAGE analysis showed that the expressed protein was soluble, and that overnight expression at 25 °C after induction with 0.5 mM IPTG maximized protein production. Starting from a cell pellet obtained from 2 L of culture, we combined the soluble fractions from two rounds of sonication and centrifugation and added 50% ammonium sulfate to precipitate contaminating proteins. The protein was subsequently dialyzed and purified using cation exchange chromatography and size-exclusion chromatography (SEC). The pure protein was then dialyzed against Chelex resin to remove contaminating metal ions. Typical yields of MRP126 were ≈60–90 mg / 2 L of culture.

For interrogating predicted metal-binding residues, we also designed and purified several variants of MRP126. We used site-directed mutagenesis to make the quadruple mutant ΔHis₃Asp (H26N/D36S/H96N/H100N), in which the four conserved metal-binding residues in the putative His₃Asp site are replaced with shorter hydrophilic residues. This variant was less stable than the native protein and we therefore reduced the ammonium sulfate concentration to 20% during the precipitation step and added 75 mM NaCl during the subsequent dialysis steps to keep it soluble during purification. We also generated a truncated variant of MRP126 named ΔTail in which the histidine-rich tail was removed by replacing the codon encoding H105, the first of the seven histidine residues in the C-terminal tail, with a stop codon. We also generated three more single-point mutants: C5S and C102S, each missing one of the two native cysteines that could contribute to a non-canonical (among the calgranulins) metal-binding site, and D36H, in which the putative metal-binding site aspartate is replaced with a histidine to create a potential His₄ site similar to the one found in CP (or His₆ site if tail histidine residues also contribute). These variants were expressed and purified analogously to the native protein except that the ΔTail variant required a different salt gradient for cation exchange chromatography. Typical yields for these variants were 60–100% of that for native MRP126.

Native MRP126 and all variants were produced in high purity as determined by SDS-PAGE (Figure S2). Mass spectrometry of the purified proteins indicated that MRP126 and all variants had the expected molecular weights for the species lacking the N-terminal methionine, a common post-translational modification in E. coli (Table S2). ICP-MS analysis confirmed that the purified proteins were obtained in the apo form with only trace amounts of contaminating transition metal ions (Table S3).

Chicken MRP126 is an α-helical homodimer.

Circular dichroism spectroscopy indicated a predominantly α-helical secondary structure for MRP126 and variants (Figure S3) with negligible change upon Ca(II) addition, as expected based on the structures of other S100 proteins. In a thermal denaturation assay in which the molar ellipticity was monitored at 226 nm as the temperature was increased from 35 °C to 95 °C in 1-°C intervals (Figure S4), we detected an apparent melting temperature (T_m) of ≈81 °C for MRP126 (8 μM) in both the absence and presence of 2 mM Ca(II). Thus, in contrast to previous studies of CP and mCP that showed excess Ca(II) ions increased the melting temperature by 10–20 °C,^{25, 40, 51} Ca(II) did not appear to affect the thermal stability of MRP126 in this assay.

We next employed analytical SEC to evaluate the quaternary structure of MRP126 and variants. We found that apo MRP126 and variants eluted at 12.1 mL (75 mM HEPES, 100 mM NaCl, pH 7.0 + 0.2 mM TCEP), which corresponds to a molecular weight of 26 kDa and is consistent with a homodimer under these conditions (Figure S5 and Table S4). Adding 2 mM Ca(II) to the sample and running buffer had negligible effect on the elution volume, indicating that Ca(II)-bound MRP126 and variants are also homodimers (Figures S3, S4, and S5). Thus, MRP126 does not tetramerize in the presence of Ca(II) as observed for hCP and mCP.^{25, 38, 39, 40, 51 ,52} In this regard, MRP126 behaves more similarly to S100A12, which exists as homodimers in the absence and presence of excess Ca(II)).³⁶

MRP126 selectively depletes Zn from bacterial growth medium using its His₃Asp sites.

We next investigated the metal-sequestering ability of MRP126 using an unbiased metal-depletion assay previously reported by our lab.²⁸ We treated Tris:TSB medium with 8 μM of MRP126 or a variant at 30 °C for 20 h with shaking at 150 rpm and then spin-filtered each sample in a 10 kDa-molecular weight cut-off concentrator to remove MRP126 or variant and any bound metal ions. ICP-MS analysis of the medium revealed that treatment with MRP126 had negligible effect on the concentrations of Mn, Fe, Co, Ni, or Cu in Tris:TSB (Figure 2A–E), whereas the amount of Zn in the medium was reduced ≈20–30 fold (Figure 2F). Moreover, similar metal depletion occurred regardless of whether 2 mM Ca(II) and/or 3 mM BME was included in the medium. These conditions were evaluated because both Ca(II) and BME affect the metal-depletion profile of hCP.²⁸ These results indicate that MRP126 effectively depletes Zn, but not other first-row transition metal ions, from Tris:TSB medium. The metal-depletion profile is similar to that of S100A12, which also provided Zn-selective depletion, and contrasts with the multi-metal depletion observed for CP (Figure 2). Thus, the metal-depletion profiles are consistent with those of characterized His₃Asp sites and differ from that of the His₆ site in CP.

Figure 2. MRP126 selectively depletes Zn from bacterial growth medium using its His₃Asp sites.

Depletion of (A) Mn, (B) Fe, (C) Co, (D) Ni, (E) Cu, and (F) Zn. Tris:TSB medium was incubated with 8 μM of MRP126, ΔTail, ΔHis₃Asp, D36H, or 10 μM hCP for 20 h at 30 °C. Following spin filtration to remove proteins > 10 kDa, the metal content of the filtrate was determined by ICP-MS (mean ± SEM, n ≥ 4). The medium contained no supplement (white bars), a 3-mM BME supplement (light gray bars), a 2-mM Ca(II) supplement (medium gray bars), or both 2-mM Ca(II) and 3-mM BME supplements (dark gray bars). Raw ICP-MS data are provided in Table S5.

We also employed the metal-depletion assay to evaluate MRP126 variants and obtain insight into the site(s) contributing to Zn depletion. Treatment with the ΔHis₃Asp variant reduced the Zn content of the medium by only ≈2 fold (Figure 2F), indicating that the conserved His₃Asp site is required for maximal Zn depletion by MRP126. In contrast, the ΔTail variant depleted Zn similarly to the native protein (Figure 2F), indicating that the histidine-rich tail is not essential for this process. Lastly, the D36H variant greatly reduced the Zn content of the medium, leaving slightly more Zn in the medium than seen with MRP126. The data also show that D36H, which likely has two His₄ sites at the homodimer interface, was more effective at sequestering Ni and perhaps Cu than native MRP126 (Figure 2D–E). Nevertheless, this mutation did not enhance Mn or Fe depletion (Figure 2A–B). Taken together, these data suggest that neither MRP126 nor the D36H variant create a 6-coordinate site using the His₃Asp/His₄ motif and two histidine residues from the tail region. Alternatively, if a 6-coordinate site forms, it exhibits a different metal-depletion profile from that of the His₆ site of CP.

In summary, these experiments demonstrated that MRP126 selectively sequesters Zn from Tris:TSB medium and that the conserved His₃Asp site, but not the histidine-rich tail, is required for full Zn depletion. These data suggest that the His₃Asp sites are high-affinity Zn(II) sites in MRP126. Moreover, the ΔHis₃Asp variant depleted some Zn from the medium, which indicates that additional sites in MRP126 provide relatively high-affinity Zn binding.

MRP126 restricts the growth of select microbes in a Ca(II)-dependent manner.

Because the mammalian calgranulins exhibit antimicrobial activity that is attributed to metal sequestration, we next investigated the growth inhibitory activity of MRP126 and variants against a panel of microbes that were previously shown to be susceptible to S100A12 and/or hCP.^{25, 27, 28, 36} These organisms include the Gram-negative bacteria E. coli and Acinetobacter baumannii, the Gram-positive bacteria Staphylococcus aureus, Listeria monocytogenes, and Lactobacillus plantarum, and the opportunistic fungal pathogen Candida albicans. In addition, we tested two serovars of the Gram-negative pathogen Salmonella enterica, one of which (S. Enteritidis) was previously shown to stimulate MRP126 expression during infection in chickens.^{44, 46}

We performed standard antimicrobial activity assays with mid-log phase cultures in medium prepared with and without a 2-mM Ca(II) supplement using MRP126 as well as hCP and S1000A12 (Figure 3). MRP126 inhibited the growth of several microbes, most notably L. monocytogenes (Figure 3G), L. plantarum (Figure 3H), and C. albicans (Figure 3I). In contrast, MRP126 had a modest inhibitory effect on A. baumannii growth (Figure 3E) and it failed to inhibit the growth of E. coli (Figure 3A), S. Typhimurium (Figure 3C), S. Enteritidis (Figure 3D), and S. aureus (Figure 3F). Nevertheless, MRP126 prevented the growth of E. coli ΔznuA, which lacks a functional high-affinity ZnuABC Zn(II)-uptake system (Figure 3B). This result indicates that the antimicrobial activity of MRP126 comes from Zn(II) deprivation. Overall, the antimicrobial activity profile of MRP126 is similar to that of S100A12, which selectively sequesters Zn from Tris:TSB and other media.^{36, 53} In contrast, hCP restricted the growth of all evaluated organisms, at least to some degree, which we attribute to its broader metal-sequestration ability (Figure 2). We reason that the organisms exhibiting susceptibility to S100A12 and MRP126 are more sensitive to Zn(II) limitation than the organisms that only show susceptibility to hCP, at least under these experimental conditions.

Figure 3. MRP126 inhibits the growth of select microbes in a Ca(II)-dependent manner.

Growth inhibitory activity of hCP (black), S100A12 (blue), and MRP126 (red) against (A) E. coli K12, (B) E. coli ΔznuA (C), S. Typhimurium IR715,⁵⁴ (D) S. Enteritidis P125109,⁵⁵ (E) A. baumannii 17961, (F) S. aureus USA300 JE2, (G) L. monocytogenes 19115, (H) C. albicans SC5314, and (I) L. plantarum WCSF1. The ΔTail (orange), and ΔHis₃Asp (purple) MRP126 variants were also tested for activity against L. plantarum. Microbes were grown in the appropriate AMA medium ± 2 mM Ca(II) at 37 °C for 8 h for all bacteria or 20 h or C. albicans before the OD₆₀₀ was measured (mean ± SEM, n ≥ 4). Asterisks (*) indicate the concentrations of MRP126 where statistically significant differences in culture OD₆₀₀ were observed compared to the no protein condition (p < 0.05 in a two-tailed unpaired t-test). See Table S6 for additional microbial strain and growth information.

These data also revealed a Ca(II) effect for some organisms. In particular, supplementation of the medium with Ca(II) was critical for MRP126 antimicrobial activity against L. monocytogenes and L. plantarum (Figures 3G and I). Thus, in terms of Ca(II) dependence, MRP126 is similar to S100A12, hCP, and mCP.^{25, 27, 28, 36, 40}

Lastly, we compared the ability of the ΔTail and ΔHis₃Asp variants to suppress the growth of L. plantarum (Figure 3I). The ΔTail variant inhibited growth of this species, but not as effectively as native MRP126 at an intermediate concentration (16 μM). In contrast, the ΔHis₃Asp variant showed no growth inhibitory activity at any tested concentration. Thus, the canonical His₃Asp site of MRP126 is required for full antimicrobial activity against L. plantarum – and likely other organisms – under these conditions. Taken together with the metal-depletion studies and similarities to S100A12, these data further support the notion that the antimicrobial ability of MRP126 results from its ability to restrict Zn availability.

MRP126 homodimers bind multiple equivalents Zn(II) with Ca(II)-dependent affinity.

Guided by the metal-depletion and antimicrobial activity assay results, we examined the effect of Zn(II) binding on the biophysical properties of MRP126. Providing an excess of Zn(II) (12.5 equivalents) did not change the CD spectrum of MRP126 in the absence or presence of 2 mM Ca(II) (Figure 4A), indicating that major changes to the predominantly α-helical secondary structure did not occur. Nevertheless, Zn(II) stabilized the folded protein in the thermal denaturation assay (Figure 4B), as heating to 95 °C failed to fully denature the Zn(II)-bound protein. In addition, pre-incubation of MRP126 with Zn(II) (4.4 equivalents) had negligible impact on the analytical SEC elution volume (Figure 4C), indicating that the protein likely remains a homodimer after binding Zn(II).

Figure 4. Zn(II) has negligible effect on the secondary structure and oligomerization state of MRP126 but does enhance its thermal stability.

(A) CD spectra of MRP126 (8 μM) in 1 mM Tris-HCl, pH 7.5, with 0.2 mM TCEP at 25 °C and no metal added (black), 2 mM Ca(II) added (red), 100 μM Zn(II) added (blue), and 2 mM Ca(II) + 100 μM Zn(II) added (purple). Spectra are averages of three measurements and are representative of at least two independent experiments. (B) CD thermal denaturation of MRP126. Protein (as in panel A) was heated from 35 °C to 95 °C in 1 °C increments, and the CD magnitude was recorded at 226 nm. The relative change between the starting (T ≤ 40 °C) value and ending (T ≥ 90 °C) value normalized to the change observed for the no metal condition is plotted. Data are averages of three independent experiments. (C) Analytical SEC chromatograms of MRP126 (80 μM) in 75 mM HEPES, 100 mM NaCl, pH 7.0, with 0.2 mM TCEP at 4 °C and no metal added (black), 2 mM Ca(II) added to sample and running buffer (red), 350 μM Zn(II) added to sample (blue), and 2 mM Ca(II) added to sample and running buffer with 350 μM Zn(II) added to sample (purple). Elution volumes are listed in Table S4. Data are representative of at least two independent experiments.

Next, to obtain insight about the stoichiometry of Zn(II) binding and the apparent Zn(II) affinities, we performed a series of Zn(II) competition experiments using three small-molecule Zn(II) sensors with varying Zn(II) affinities. We found that MRP126 outcompeted Zincon (K_d,Zn(II) ≈ 10 μM^{56, 57}) for approximately six equivalents Zn(II) in both the absence and presence of 2 mM Ca(II) (Figure 5A–B). MRP126 outcompeted MF2 (K_d,Zn(II) ≈ 30 nM^{58, 59}) for approximately four equivalents Zn(II) in the absence of Ca(II) (Figure 5C). At higher Zn(II) equivalents, the titration curve indicated competition between MRP126 and MF2 for the metal ion (Figure 5C). MRP126 could not compete with the Ca(II)-insensitive Zn(II) sensor ZP4 (K_d,Zn(II) = 650 pM⁶⁰) in the absence of Ca(II) (Figure 5D); however, when excess Ca(II) ions were present in the buffer, MRP126 outcompeted ZP4 for approximately two equivalents Zn(II) before competition was observed for approximately one more equivalent Zn(II) (Figure 5E).

Figure 5. Zn(II)-competition titrations for MRP126 and variants inform the Zn(II)-binding sites and apparent affinities.

All experiments were performed in 75 mM HEPES, 100 mM NaCl, pH 7.0, with 0.2 mM TCEP, with or without 2 mM Ca(II) as indicated. (A-B) Zincon (K_d,Zn(II) ≈ 10 μM^{56, 57}) competition experiments with 20 μM Zincon and 8 μM protein in the absence (A) or presence (B) of 2 mM Ca(II). Zn(II) was added 24x in 2.5 μM increments (0.31equivalents), 2x in 5 μM increments (0.63 equivalents), and 2x in 10 μM increments (1.25 equivalents). The absorbance at 621 nm was monitored to detect formation of the Zn(II)-Zincon complex. (C) MF2 (K_d,Zn(II) ≈ 30 nM^{58, 59}) competition experiments with 10 μM MF2 and 8 μM protein in the absence of Ca(II). Zn(II) was added 28x in 2.5 μM increments (0.31 equivalents). The absorbance at 325 nm was monitored to detect formation of the Zn(II)-MF2 complex. (D-E) ZP4 (K_d,Zn(II) = 650 pM⁶⁰) competition experiments with 1.5 μM ZP4 and 4 μM protein in the absence (D) or presence (E) of 2 mM Ca(II). Zn(II) was added 30x in 0.5 μM increments (0.13 equivalents) and 5x in 1 μM increments (0.25 equivalents). The increase in ZP4 fluorescence emission (λ_ex = 495 nm) indicated formation of the Zn(II)-ZP4 complex and was integrated from 505–650 nm. All data are averages ± SEM, n = 3. NoP = no protein control titrations. Solid lines represent nonlinear regression fits to the data using Dynafit.⁶¹ See Table 1 for a summary of these results, including calculated K_d,Zn(II) values for MRP126 and variants. See also Figures S6–S10 and S11–S15 for simulated fits with higher and lower K_d values to these data.

In combination, these results suggest that MRP126 has three distinct pairs of Zn(II)-binding sites, and that Ca(II) binding has a marked effect on Zn(II) binding to the two highest-affinity pairs of sites. To probe the Zn(II) affinities for these different sites, we fit the competition titration data to binding equilibria models (see Supporting Information, Figures 5C, 5E, S6–S15 and Table 1).

Table 1.

MRP126 variant Zn(II)-binding stoichiometries and affinities from competition experiments.^a

Zincon	MRP126	ΔTail	ΔHiS3Asp	C5S	C102S
− Ca(II), equiv Zn(II)	6.9	4.1	4.7	6.6	5.6
+ Ca(II), equiv Zn(II)	6.3	4.1	4.4	5.6	5.3
MF2, − Ca(II)	MRP126	ΔTail	ΔHiS3Asp	C5S	C102S
Kd,zn(II) Sites 1 + 2	1 ± 0.3 nM	1.6 ± 0.5 nM	7 ± 2 nM	5 ± 1 nM	4.3 ± 0.7 nM
equiv Zn(II)	4	4	2	4	4
Kd,zn(II) Site 3	80 ± 10 nM	--	190 ± 40 nM	130 ± 40 nM	150 ± 40 nM
equiv Zn(II)	2	0	2	2	2
ZP4, + Ca(II)	MRP126	ΔTail	ΔHiS3Asp	C5S	C102S
Kd,Zn(II) Site 1	23 ± 4 pM	67 ± 8 pM	--	42 ± 4 pM	49 ± 5 pM
equiv Zn(II)	2	2	0	2	2
Kd,Zn(II) Site 2	3.6 ± 0.4 nM	13 ± 2 nM	1.9 ± 0.2 nM	--	--
equiv Zn(II)	2	2	2	0	0

^aAssays were performed as described in the legend for Figure 5. For Zincon, the equiv Zn(II) reported corresponds to the equivalents Zn(II) added to each MRP126 variant before A₆₂₁ exceeded 0.032 (which corresponds to halfway between the A₆₂₁ of the No Protein sample before and after the first Zn(II) addition). For MF2 and ZP4, the K_d,Zn(II) values are reported from fitting the titration data to a model in which the indicated binding sites have the capacity for the equivalents Zn(II) listed below. Site 1 corresponds to the His₃Asp site, Site 2 corresponds to the Cys₂His₂ site, and Site 3 corresponds to the His_x site (see below).

From the MF2 competition data, we calculated that, in the absence of Ca(II), MRP126 binds four equivalents Zn(II) with low-nanomolar affinity (K_d,Zn(II) ≈ 1 nM) and an additional two equivalents with approximately 100-fold lower affinity (K_d,Zn(II) ≈ 100 nM). From the ZP4 competition data, we calculated that MRP126 binds two equivalents Zn(II) with picomolar affinity (K_d,Zn(II) ≈ 20 pM) and an additional two equivalents Zn(II) with low-nanomolar affinity (K_d,Zn(II) ≈ 4 nM). We note that there is some discrepancy between the calculated high-affinity values from the MF2 competition titrations (K_d,Zn(II) ≈ 1 nM) and our observation that MRP126 could not effectively compete with ZP4 in the absence of excess Ca(II). As a result, these values should be taken as estimates rather than precise values for the apparent Zn(II) dissociation constants.

MRP126 has three distinct pairs of Zn(II)-binding sites.

Having established that MRP126 binds three pairs of Zn(II) ions, each with a distinct tier of affinities, we next aimed to identify which amino acid residues comprise each pair of binding sites. We therefore compared Zn(II) competition titrations of the ΔHis₃Asp, ΔTail, C5S, and C102S MRP126 variants to what we observed for the native protein (Figure 5 and Table 1). With Zincon both with and without Ca(II) (Figure 5A–B), ΔHis₃Asp shifted the titration curve by ≈2 equivalents from the WT, which indicates that each His₃Asp site coordinates one Zn(II) ion. ΔTail also shifted the titration curve such that a break occurred after ≈4 equivalents Zn(II). This result suggests that the C-terminal tail histidine residues contribute to at least one pair of Zn(II)-binding sites. In addition, C5S behaved similarly to the native protein, and C102S shifted the curve by ≈1 equivalent compared to native MRP126.

We obtained similar results from Zn(II) competition titrations with MF2 in the absence of Ca(II) (Figure 5C). ΔHis₃Asp outcompeted MF2 for ≈2 equivalents Zn(II) and thus shifted the titration curve by ≈2 equivalents from the native protein. ΔTail outcompeted MF2 for 4 equivalents Zn(II); however, compared to the native protein and the other variants, the rise in A₃₂₅ after the break point was much steeper for ΔTail. This result indicates that the competition after four equivalents Zn(II) seen for native MRP126 arises from Zn(II) binding to sites that require the C-terminal histidine-rich tail. In addition, both the C5S and C102S variants had somewhat reduced competitiveness over the first four equivalents Zn(II) compared to the native protein, suggesting that both cysteine residues may contribute to the highest-affinity sites in the absence of Ca(II).

With ZP4 in the presence of Ca(II), native MRP125, ΔTail, C5S, and C102S all clearly outcompeted the dye for two equivalents Zn(II), whereas ΔHis₃Asp did not (Figure 5E). Thus, the His₃Asp site indeed constitutes the highest-affinity Zn(II)-binding site in MRP126 in the presence of Ca(II). However, ΔHis₃Asp provided some competition over the course of the first two equivalents Zn(II), which demonstrates that other binding sites provide low-nanomolar Zn(II) affinity. Indeed, comparing the competition titrations after addition of 2 equivalents Zn(II) informs the composition of this middle-tier affinity site. Both C5S and C102S variants featured sharp increases in ZP4 fluorescence after the break point, suggesting that the residual competition with the high-affinity dye seen with the native protein after this point depends on the presence of both native cysteine residues. The ΔTail competition titration curve featured an intermediate-steepness increase after two equivalents Zn(II), falling between the cysteine-to-serine variants and MRP126. This comparison suggests that one or more C-terminal tail histidine residues contribute to Zn(II) binding at the second-tier sites, but that it/they are not as critical as either C5 or C102.

Overall, from these titration curves and fitting the data to various models (see Supporting Information), we can conclude: (i) The conserved His₃Asp motifs constitute the two picomolar-affinity Zn(II)-binding sites (defined as Site 1 in our binding models) in the presence of Ca(II). (ii) Two non-conserved cysteines anchor a second pair of low-nanomolar Zn(II)-binding sites (Site 2) and the affinity of these sites also increases when excess Ca(II) ions are present. Also, these sites likely include at least one histidine residue from the C-terminal tail. (iii) The histidine-rich C-terminal tail provides a third-tier of Zn(II)-binding sites (Site 3) that coordinate the ion with high-nanomolar affinity and for which a Ca(II) effect was not observed.

Co(II)-binding titrations further inform the identities of the Zn(II)-binding sites.

To further probe the coordination environments at the multiple Zn(II)-binding sites identified from the Zn(II) competition titrations, we titrated MRP126 and variants with Co(II) and monitored binding by optical absorption spectroscopy (Figures 6 and S16). For MRP126, the spectra (Figure 6A and F) were reminiscent of those previously reported for S100A12 and hCP as well as S100A7,^{25, 36, 50} although with additional spectral features not seen for the human proteins as described below. For instance, in the presence of Ca(II) and ≈3 equivalents Co(II) (Figure 6F, bold trace), the solution of MRP126 was magenta and the optical absorption spectrum of this sample in the region corresponding to d–d transitions exhibits a broad signal centered at 563 nm with a shoulder at 525 nm, which is consistent with a 4- or 5-coordinate Co(II) species and closely resembles the Co(II) spectra assigned to the His₃Asp sites of hCP and S100A12.^{25, 36} Thus, this result indicated that MRP126 indeed coordinates transition metal ions using its His₃Asp sites. It also rules out a six coordinate site comprised of the His₃Asp site and two additional ligands.

Figure 6. Co(II)-binding titrations indicate different metal-binding motifs for MRP126 variants.

Optical absorption spectra for Co(II)-binding titrations (0–10 equivalents) of MRP126 and variants (240 μM) in 75 mM HEPES, 100 mM NaCl, pH 7.0, with 0.2 mM TCEP with (A-E) no Ca(II) added or (F-J) 6 mM Ca(II) added. (A,F) MRP126, (B,G) ΔTail, (C,H) ΔHis₃Asp, (D,I) C5S, (E,J) C102S. Each line represents an addition of 150 μM Co(II) (0.625 equivalents). In each panel, the spectrum corresponding to 3.1 equivalents Co(II) is shown as a bolder line. The ε values were calculated assuming that all metal-binding sites occur in pairs in the MRP126 homodimer; these values therefore encompass at least two distinct Co(II)-protein interactions for most variants. All panels are representative of at least two independent experiments. See also Figure S16 for versions of these plots that include the 300–400 nm range.

In the presence of Ca(II), the spectra of the ΔTail, C5S, and C102S variants (Figure 6G, I, and J) all looked nearly identical to that of MRP126 after ≈3 equivalents Co(II) were added, indicating that these variants also bind transition metal ions with the His₃Asp motif. In contrast, the spectra for ΔHis₃Asp (Figure 6H) exhibited a completely different shape than observed for MRP126 and other variants. In the presence of Ca(II), ΔHis₃Asp had a saturable overall absorption peak centered at around 625 nm and declining to the baseline by 700 nm. This peak shape is similar to what has been reported for Co(II)-substituted Cys₂His₂ Zn(II)-binding sites in zinc-finger proteins,^62–64 whereas similar shoulders/side-peaks (although of different relative magnitudes than ours) were also seen around 575 nm and 675 nm in another zinc-finger Cys₂His₂ site.⁶⁵ We spectroscopically detected this second-tier Co(II)-binding site for MRP126 after >3 equivalents Co(II) were added in the presence of Ca(II) (Figure 6F). Indeed, the spectra of MRP126 with 10 equivalents Co(II) in the presence of Ca(II) appears to be a near summation of the absorption spectra of MRP126 with three equivalents Co(II) and that of the ΔHis₃Asp variant with 10 equivalents Co(II). Moreover, the spectral features between 600 and 700 nm that we attribute to likely Cys₂His₂ sites in MRP126 and the ΔHis₃Asp variant were less pronounced for Co(II)-bound ΔTail (Figure 6G) and especially for the Co(II)-bound C5S and C102S variants (Figure 6I and J). These results are consistent with a Cys₂His₂ site; for each variant, we expect that one ligand is lost, changing the primary coordination sphere.

Additional spectral evidence of the Cys₂His₂ site was seen in the 300–400 nm range, where we detected a sulfur-to-cobalt charge-transfer transition in the spectra of all five variants (Figure S16). As expected, the intensity of this transition was reduced by approximately 50% for the C5S and C102S variants that each lack one of the native cysteine residues. From spectral comparisons, we could not identify any features that definitively correspond to Co(II) binding at the proposed third-tier affinity Zn(II) sites composed primarily of the C-terminal tail histidine residues, which may reflect either the poor affinity of Co(II) for that site and/or that the magnitude of the extinction coefficient for this Co(II) site is relatively low. The latter effect was previously seen with the ΔHis₃Asp variant of hCP that binds Co(II) exclusively at its His₆ site with an extinction coefficient of ≈40 M⁻¹cm⁻¹.²⁵

Ca(II) prioritizes metal-binding and delays metal-exchange at the His₃Asp sites.

For all MRP126 variants, the Co(II) optical absorption spectral peaks did not show saturation behavior as quickly (if at all) in the absence of Ca(II) (Figure 6A–E), which likely reflects a lower affinity for Co(II) under these conditions that is consistent with our findings with Zn(II) (Figure 5D–E). To more easily compare the changes in the Co(II) spectra for each variant in the absence and presence of Ca(II), we plotted ε₅₆₃ versus the Co(II) / MRP126 molar ratio to monitor the transition corresponding to Co(II) binding at the His₃Asp sites and ε₃₅₀ versus the Co(II) / MRP126 molar ratio to monitor the transition that we attributed to sulfur-to-cobalt charge transfer at the Cys₂His₂ sites (Figure 7). In the presence of Ca(II) (Figure 7B and D), we detected an ordered binding process for MRP126 as well as the ΔTail, C5S, and C102S variants. For the first ≈3 equivalents Co(II) added, ε₅₆₃ increased linearly and negligible change in ε₃₅₀ occurred, whereas at >3 equivalents Co(II) added, ε₅₆₃ plateaued and ε₃₅₀ increased dramatically. Thus, in the presence of excess Ca(II) ions, the His₃Asp sites preferentially filled with Co(II) and only after they were saturated was binding to the Cys₂His₂ sites detected. As expected, the ΔHis₃Asp variant immediately began to increase at ε₃₅₀, reflecting the Co(II) binding at the Cys₂His₂ sites. In contrast, in the absence of Ca(II) (Figure 7A and C), we did not observe any clearly ordered binding for the native protein, as both ε₅₆₃ and ε₃₅₀ simultaneously increased at the start of the Co(II) titration. Thus, Ca(II) appears to be essential for directing transition metal ions preferentially to the His₃Asp sites, as the His₃Asp and Cys₂His₂ sites appear to fill indiscriminately with Co(II) in the absence of Ca(II). These titrations also show that the binding of Ca(II) increases the transition-metal affinity of the His₃Asp sites, both in absolute terms and relative to the affinity of the Cys₂His₂ site. This behavior is reminiscent of prior studies of hCP, which showed that the presence of excess Ca(II) ions directs transition metals (e.g. Mn(II), Fe(II)) to the His₆ site.^{33, 37}

Figure 7. Optical absorption data indicate His₃Asp site fills first in presence of Ca(II).

Optical absorption values for the Co(II) d–d transition at 563 nm corresponding to binding at the His₃Asp site (A-B) and a Co(II)–S charge transfer transition at 350 nm corresponding to binding to the Cys₂His₂ site (C-D), taken from the titrations seen in Figures 6 and S16 (mean ± SEM, n ≥ 2).

Lastly, we investigated metal substitution by adding 10 equivalents Zn(II) to the Co(II)-bound proteins (Figure 8). In the presence of Ca(II), Zn(II) addition to the native protein resulted in disappearance of the spectral features for the Co(II)-bound protein between 600–700 nm that we attribute to the Cys₂His₂ site within 2 min, whereas the peak centered on 563 nm corresponding to the His₃Asp site decayed over the course of approximately 1 h (Figure 8A). Similar behavior was observed for the ΔTail, C5S, and C102S variants (Figure 8C, D, and E), with a noticeable acceleration in metal substitution for ΔTail. In contrast, for ΔHis₃Asp (Figure 8B), all signals for d–d transitions were lost within 2 min following Zn(II) addition. For the ΔTail variant in the absence of Ca(II), Zn(II) addition also eliminated all d–d transitions in a 2-min timeframe (Figure 8F) For the other Co(II)-bound variants in the absence of Ca(II), Zn(II) addition caused the protein to precipitate, precluding any time course measurements of metal substitution. Taken together, these results indicate that the Cys₂His₂ site has relatively fast metal substitution in either the absence or presence of Ca(II), whereas the His₃Asp site displays relatively slow metal exchange in the presence of Ca(II), but not in its absence. Overall, these observations are in accord with the results from the metal depletion and Zn(II) competition experiments and further indicate that Ca(II) binding helps prevent the loss of captured transition metal ions at the His₃Asp sites of MRP126.

Figure 8. His₃Asp sites kinetically trap transition metals in the presence of Ca(II).

Optical absorption spectral time course of Zn(II) (10 equivalents) replacement of Co(II) (10 equivalents) in 75 mM HEPES, 100 mM NaCl, pH 7.0, with 0.2 mM TCEP with 6 mM Ca(II) added (A-E) or no Ca(II) added (F). The lines represent spectra before Zn(II) addition (bold line) and after 2 min, 6 min (bold red line), 20 min, 60 min, and 120 min. (A) MRP126, (B) Δ His₃Asp , (C) ΔTail , (D) C5S, (E) C102S, (F) ΔTail without Ca(II). Only 2 and 6 min time points were collected for ΔTail without Ca(II). Adding Zn(II) to the other Co(II)-bound MRP126 variants in the absence of Ca(II) caused immediate protein precipitation, which precluded similar time course measurements. ε values were calculated assuming two binding sites per dimer. All panels are representative of at least two independent experiments.

Discussion

In this work, we evaluated the biophysical, metal-binding, and antimicrobial properties of an avian calgranulin homolog of mammalian CP and S100A12 to provide insights into the ancestral function of this protein family and better understand the subsequent divergence in function among extant homologs (as summarized in Figure S17). We selected chicken MRP126 because of previous studies implicating it in the innate immune response to bacterial infection^43–47 and identifying a conserved role in inflammatory signaling.⁶ Our studies reveal several key similarities and differences between MRP126 and its mammalian homologs. Unlike CP, which forms tetramers,^{25, 38, 39, 40, 52, 66, 67} and S100A12, which forms higher-order oligomers (possibly tetramers or hexamers),^{35, 41, 42} MRP126 exists exclusively as a dimer under the conditions employed in this study (Figure 4). This result suggests that the ancestral homolog was likely dimeric, with the higher-order oligomerization properties of mammalian homologs evolving more recently. The Zn-selective metal-depletion profile displayed by MRP126 (Figure 2) mirrors the behavior of S100A12,³⁶ rather than the broader metal-depletion profile (including Mn, Fe, and Ni) seen with CP^{28, 40} Parsimony therefore suggests that Zn(II) sequestration was the ancestral function of the calgranulin clade, and mammalian CP evolved its more promiscuous metal-sequestration ability after the homologs had diverged (and perhaps required the subsequent divergence and complementary evolution of S100A8 and S100A9). D36H, an MRP126 variant engineered to have a His₄ site similar to the His₄ site of CP, showed improved Ni depletion compared to the native protein (Figure 2D), providing some insight into how evolution may have begun to expand the selectivity profile of the ancestral protein. However, unlike CP, the D36H variant depleted neither Mn nor Fe. Previous work with variants of CP lacking either or both histidine residues in the S100A9 C-terminal tail that complete the His₆ site coordination sphere showed importance of those residues for Mn(II) and Fe(II) capture.^{26, 28, 32, 68} These results therefore suggest that the D36H variant of MRP126 does not form a CP-like His₆ site. Thus, further protein evolution that enabled the histidine-rich C-terminal tail to participate in a transition-metal-binding site at the dimer interface was likely necessary to impart this advantageous property to CP.

The apparent Zn(II) selectivity of MRP126 likely results in its diminished antimicrobial activity profile compared to hCP (Figure 3). Along these lines, in the antimicrobial activity assays, MRP126 performed similarly to S100A12, which sequesters Zn(II) but not other metal nutrients like Mn(II) and Fe(II).³⁶ Bird and reptile species each express this lone calgranulin homolog, and it remains an open question whether these organisms have evolved any compensatory mechanism – perhaps involving unrelated proteins or more-distantly-related S100s – to provide the missing Mn(II)/Fe(II) sequestration ability imparted by CP. If those species indeed lack such compensatory proteins, our antimicrobial activity experiments with MRP126 (Figure 3) would suggest that the bird/reptile innate-immune system has a less-potent arsenal than do mammals for implementing a nutritional-immunity defense against potential invaders.

As was previously seen for CP and S100A12,^{25, 36, 40} the presence of excess Ca(II) ions at a concentration that mimics the extracellular environment enhanced the antimicrobial activity of MRP126, most notably against L. monocytogenes and L. plantarum (Figure 3G and I). This finding agreed with the Ca(II)-dependent increase in Zn(II)-binding affinity of MRP126 as measured by competition with ZP4 (Figure 5D and E). It is also reminiscent of prior observations of Ca(II) enhancing the Zn(II) affinities of the His₃Asp and His₆ sites of hCP.²⁵ Therefore, the ability of Ca(II) to enhance transition-metal binding affinity appears to be an ancestral property of the calgranulin family, although other effects of Ca(II) binding (such as changes to quaternary structure) have clearly diverged. S100A7 (psoriasin), a more distantly-related S100 protein, also has two His₃Asp sites, sequesters Zn(II) and performs an apparent innate-immune role.^{50, 69, 70} Whereas Ca(II) binding affected the midpoint potential of two intramolecular disulfide bonds in the S100A7 homodimer, Ca(II) did not as clearly enhance the Zn(II) affinity or antimicrobial potency of this protein.⁵⁰ This comparison suggests that the ancestral calgranulin evolved the ability to use Ca(II) to modulate its transition-metal-binding sites, thereby leveraging higher extracellular Ca(II) concentrations to convert to a higher-affinity state upon release from the cytoplasm at infection sites.

In contrast to hCP and S100A12, which each have two high-affinity Zn(II) sites per dimer,^{25, 27, 36} MRP126 binds additional equivalents Zn(II) (Figure 5). Based on the Zn(II) competition titrations (Figure 5) and Co(II)-binding studies (Figure 6), we present a model for Zn(II) binding to this protein (Figure 9). In the absence of Ca(II), Zn(II) binds with similar low-nanomolar affinity to the two conserved His₃Asp sites at the homodimer interface as well as the two Cys₂His₂ sites comprised of Cys5 from one monomer, Cys102 from the second monomer, and (likely) His98 (based on its location on the same face of the helix one turn below C102) and a C-terminal tail His from the second monomer. Additional Zn(II) ions can then bind with high-nanomolar affinity to two sites primarily or entirely comprised of C-terminal tail histidine residues (His_x sites) from each monomer. In the presence of excess Ca(II), the Zn(II) affinities of the His₃Asp and Cys₂His₂ sites are enhanced. The first two Zn(II) equivalents bind to the His₃Asp sites with low-picomolar affinity, the next two Zn(II) equivalents bind to the Cys₂His₂ sites with improved but still low-nanomolar affinity, and two additional Zn(II) equivalents then bind with high-nanomolar affinity to the His_x sites. This ability of Ca(II) to prioritize one type of transition-metal-binding site over another echoes previous findings with hCP where providing Ca(II) imparts a preference for Fe(II) to bind to the His₆ site over the His₃Asp site³⁷ and Mn(II) to bind to the His₆ site over either the His₃Asp site or the EF-hand sites.³³

Figure 9. Model of MRP126 Ca(II)-dependent Zn(II) binding.

Each MRP126 homodimer contains three distinct pairs of Zn(II)-binding sites. Site 1 is a canonical calgranulin His₃Asp site comprised of His26 and Asp36 from one monomer and His96 and His100 from the other monomer. Site 2 is a dicysteine site that likely consists of Cys5 from one monomer and Cys102, His98, and a C-terminal tail histidine from the other monomer, which we denote Cys₂His₂. Site 3 likely consists of multiple histidines from the C-terminal tail of one monomer, which we call His_x. In the absence of Ca(II), Zn(II) binds to MRP126 with similar low-nanomolar affinity at either Site 1 (His₃Asp) or Site 2 (Cys₂His₂), then with high-nM affinity at Site 3 (His_x). Ca(II) dramatically increases the affinity of Site 1 and increases the affinity of Site 2 to a lesser degree. Thus in the presence of Ca(II), Zn(II) preferentially binds with low-pM affinity to Site 1 (His₃Asp), then with lower-nanomolar affinity to site Site 2 (Cys₂His₂), and finally with high-nanomolar at Site 3 (His_x). The primary coordination sphere for the His_x site is unknown and it is drawn as a His₄ site. The dashed line indicates the approximate location of the point of truncation for the ΔTail variant.

The importance of these additional Zn(II)-binding sites to the Zn(II)-sequestering, antimicrobial role of MRP126 is unclear. Neither of the cysteine residues that participate in Zn(II) binding is conserved within bird and reptile MRP126 homologs (Figure S1); however, these cysteine residues always appear as a pair in the homologs that do have them. It is also possible that, in an oxidizing environment, these cysteine residues could form a disulfide bond and thereby eliminate the Cys₂His₂ Zn(II)-binding site and perhaps alter the Zn(II) affinity of the His₃Asp site. In contrast, a histidine- (and glutamine-) rich C-terminal tail does appear to be a hallmark feature of this clade of the calgranulin family (Figure S1), which may suggest that the moderate-affinity Zn(II) binding it affords confers a functional advantage to the protein. The ΔHis₃Asp variant depleted some Zn from bacterial growth medium (Figure 2F) but did not affect the growth of L. plantarum (Figure 3I), whereas the ΔTail variant depleted Zn similarly to the native protein (Figure 2F) and had a slightly attenuated antimicrobial effect on L. plantarum (Figure 3I). From these results, it appears that the intact His₃Asp sites are necessary and sufficient for MRP126 to function as a Zn(II)-sequestering protein, but future studies with additional variants could help determine whether the presence of the additional Zn(II)-binding sites offers functional advantages.

Conclusions

In conclusion, we demonstrated that an avian calgranulin homolog, chicken MRP126, sequesters Zn(II) in a Ca(II)-dependent manner to restrict the growth of certain microbes. MRP126 shares these functional properties with mammalian S100A12 and the His₃Asp site of CP, which indicates that an ancestral calgranulin had already evolved similar traits in the last common ancestor of higher vertebrates.

Abbreviations

AMA

antimicrobial activity

ATCC

American Type Culture Collection

BHI

brain heart infusion medium

BME

β-mercaptoethanol

CD

circular dichroism

CP

calprotectin

hCP

human calprotectin (including hCP-Ser, the Cys → Ser variant)

mCP

murine calprotectin (including mCP-Ser, the Cys → Ser variant)

DTT

dithiothreitol

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

ICP-MS

inductive-coupled plasma mass spectrometry

IPTG

isopropyl β-d-1-thiogalactopyranoside

LB

Luria–Bertani medium

LC-MS

liquid chromatography mass spectrometry

MF2

Mag-fura-2

MRS

De Man, Rogosa, and Sharpe medium

PDB

Protein Data Bank

PMSF

phenylmethylsulfonylfluoroide

SEC

size exclusion chromatography

SEM

standard error of the mean

TCEP

tris(2-carboxyethyl)phosphine

Tris

tris(hydroxymethyl)aminomethane

TSB

tryptic soy broth medium

YPD

yeast extract peptone dextrose medium

ZP4

Zinpyr-4

ΔHis₃Asp

H26N/D36S/H96N/H100N quadruple mutant variant of MRP126

ΔTail

H105Stop variant of MRP126