E. coli were grown at 37°C aerobically (225 rpm) in Luria-Bertani (LB) broth or agar, supplemented with appropriate antibiotics. GAS was grown statically in Todd Hewitt yeast broth (THYB, 5 ml of media in 15-ml screw-top tubes, Thermo Scientific #33965) or agar (THYA) at 37°C. L. lactis was grown statically in GM17 at 30°C (10 ml of media in 15-ml screw-top tubes). Plasmid extractions were performed using the Promega Wizard Miniprep kit (PR-A7510) or the Qiagen Midiprep kit (12123). Genomic DNA was harvested with Invitrogen PureLink (K1820-01). Unless otherwise specified, chemicals were purchased from Sigma.

A list of strains and plasmids can be found in Table 1 , and primer sequences are shown in Table 2 . Plasmid engineering was confirmed with restriction digest and PCR analyses.

A hupZ::cm^R (chloramphenicol acetyltransferase) allele with flanking regions of chromosome homology was assembled in pUC19 using NEBuilder HiFi Assembly Kit (#E5520) generating plasmid pKV111. The chromosomal hupZ upstream region was cloned from GAS strain NZ131 using primers ZE844 and ZE845. The downstream arm using ZE842 and ZE843 and the cm^R gene was amplified from pDC123 using primers ZE840 and ZE841. The hupZ::cm^R allele was amplified from pKV111 with ZE876 and ZE787 primers, adding flanking EcoRI sites to move the fragment into the temperature-sensitive vector pJRS700, generating pKV117.

The shuttle vector, pKV138, expressing hupZ under GAS recA promoter, was generated for complementation. The hupZ gene was amplified from NZ131 gDNA and ligated into pLC007. The empty vector control (pKV141) was generated by cutting pLC007 with HindIII and self-ligating.

The vector pKV105, expressing hupZ under nisin regulation, was used for heterologous expression in Lactococcus lactis. The gene was cloned from NZ131 strain using primers ZE685 and ZE686. The insert and pNZ8008 were digested, ligated, and electroporated into MC4100 E. coli. Competent MG1363 L. lactis that already contained pXL14, which codes the nisin response regulator, were electroporated with either pKV105 or pNZ8008 (as a negative control).

Plasmid pKV102 expressing HupZ with a C-terminal Strep-tag was purchased from Vector Builder. The HupZ sequence was placed in a pET bacterial protein expression vector that contains a T7 promoter and pBR322 origin of replication.

Maltose binding protein (MBP) fusion (pKV130 or pKV135) was generated with ligation-independent cloning into pET/His₆/MBP/TEV (Addgene plasmid #29656 or pET/MBP/TEV (Addgene plasmid #48311) expression vector as previously described (Porter and Christianson, 2019). Briefly, hupZ was cloned from NZ131 with primers ZE978 and ZE979 that added the following upstream, 5’TACTTCCAATCCAATGCA3’, and downstream, 5’TTATCCACTTCAATGTTATTA3’, sequences to the insert. The purified PCR products were incubated with T4 DNA polymerase (Invitrogen 1800-5017), bovine serum albumin (BSA), dithiothreitol (DTT), dCTP, and T4 polymerase buffer. The vector pET/His₆/MBP/TEV or pET/MBP/TEV was linearized with SspI restriction enzyme and then incubated with T4 DNA polymerase, BSA, DTT, dGTP, and T4 polymerase buffer. The reactions were cleaned using ethanol precipitation and resuspended in 12 µl of diH₂O. One microliter of the vector was incubated with 4 µl of insert for 30 min at room temperature. One microliter of 25 mM EDTA was added to the solution and allowed to sit for an additional 15 min and then transformed into E. coli.

The GAS ΔhupZ::cm ^R mutant was generated using insertion inactivation to knock out hupZ in the GAS M49 strain NZ131. Competent NZ131 was transformed with the temperature-sensitive pKV117 (harboring the hupZ::cm^R allele and flanking chromosomal region) and plated on THYA with kanamycin (the vector marker) at 30°C. Colonies were then passed three times in THYB with kanamycin at 37°C and were plated with either chloramphenicol or no antibiotic at 30°C. The resulting GAS clones harboring an integrated pKV117 were confirmed by PCR. Daily, individual colonies were propagated at 37°C and screened through replica plating to ensure the loss of the vector marker and the maintenance (knockout mutant) or loss of the cm^R gene (wild-type rescue). The replacing of hupZ with hupZ::cm^R allele in M49Lyles or the regeneration of the wild-type hupZ allele in M49Rescue was confirmed with PCR.

The pKV135 (expressing MBP-HupZ) plasmid was transformed into Invitrogen Chemically Competent BL21(DE3) cells before each expression. All others were grown overnight from a glycerol stock and then diluted in fresh LB media. Cultures were grown at 37°C with 225 rpm until they reached an OD₆₀₀ of 1, induced with 1 mM isopropyl β-D-1-tiogalactopyranosie (IPTG), and incubated overnight at 20°C with 180 rpm. The following day, cells were harvested by centrifugation. HupZ-His₆- or MBP-HupZ-expressing cells were resuspended in 20 mM Tris (pH 8.0), 100 mM NaCl, and 0.1% Triton X-100. HupZ-Strep cells were resuspended in 20 mM Tris/HCl and 500 mM NaCl. One cOmpleteEDTA-free Protease inhibitor tablet (Roche #1183670001) per 500 ml of grown culture was added before sonification. The cellular debris was pelleted by centrifugation at 20,000 × g for 30 min at 4°C, and the lysate was filtered using a 0.22 µM filter unit. Protein was purified on an AKTA FPLC using Cytiva HisTrap HP, MBPTrap HP, or StrepTrap HP Sepharose columns.

Protein purification and size were confirmed using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, Figure 1A ). Protein concentration was determined by the ThermoScientific Lowry Protein Assay Kit (23240). The buffer used for reconstitutions and degradations consisted of 20 mM sodium phosphate and 500 mM NaCl (pH 7.4). Arginine (30 mM) and 0.1% glycerol were added to the buffer for HupZ-Strep protein to promote solubility.

Purified TEV protease (expressed from pRK793) was generously provided by Dr. Nicholas Noinaj of Purdue University (Tropea et al., 2009). Purified His-MBP-HupZ was incubated with TEV protease at 100 μg/1 μg, respectively, in 200-μl aliquots overnight statically. Cut protein circulated for 1 h at 4°C with resin from three NEBExpress Ni Spin Column rotating head to head in a clean gravity column to remove free His-MBP. Flow-through was collected and subsequently processed via an AKTA-FPLC system using Cytiva MBPTrap HP columns to further remove His-MBP. Collected flow-through of tag-free HupZ was concentrated using Amicon^® Ultra-15 Centrifugal Filter Unitfilters. Purity was assessed by SDS-PAGE ( Figure 1B ) and quantified by Lowry.

Female outbred CD1 (Charles River) mice aged 6 to 8 weeks were used for all experiments. Experiments were performed as previously described (Patras and Doran, 2016; Cook et al., 2018). A day prior to inoculation (day −1), mice were given an intraperitoneal injection of 0.5 mg of β-estradiol valerate (Alfa Aesar) suspended in 100 μl of filter-sterilized sesame oil (Acros Organics MS) to synchronize estrus. On day 0, WT and ΔhupZ::cm^R mutant strains were grown to an OD₆₀₀ = 0.4 and mixed 1:1. Mice were vaginally inoculated with the mixed culture in 10 μl of PBS containing 10⁷ CFU. On days 1, 2, 3, and 5, the vaginal lumen was washed with 50 μl of sterile PBS, using a pipette to gently circulate the fluid approximately 6–8 times. The lavage fluid was then collected and placed on ice for no more than 30 min. Vaginal lavage was serially diluted in PBS and plated on CHROMagar StrepB (WT) or CHROMagar StrepB with chloramphenicol (ΔhupZ) plates to obtain CFU counts (Cook et al., 2018). Murine colonization studies were reviewed and approved by Binghamton University Laboratory Animal Resources (LAR) and by the Binghamton Institutional Animal Care and Use Committee (IACUC) under protocols 803-18 and 857-21.

We previously showed that a recombinant HupZ protein containing a C-terminal fusion to His₆ tag (HupZ-His₆) binds and degrades heme in vitro (Sachla et al., 2016). HupZ-His₆ crystalized as a homodimer with a split β-barrel conformation, a fold also seen in FMN-binding heme-degrading proteins described in several bacterial species (Sachla et al., 2016; Lyles and Eichenbaum, 2018). Additional investigations revealed that this recombinant HupZ protein interacts with heme in vitro by its His₆ tag, leading to a higher-order oligomeric structure, heme stacking, and degradation (Traore et al., 2021). These findings cast doubts about the function and the role of HupZ in heme metabolism. To reexamine heme-binding by HupZ, we constructed a new recombinant C-terminal fusion replacing the His₆ with a Step-tag (HupZ-Strep) to facilitate purification. We expressed and purified the recombinant HupZ-Strep ( Figure 1A ), but the purified protein was not stable and precipitated out of the solution. We added 1% glycerol and 30 mM arginine to the buffer to increase stability for later heme titration experiments (Vagenende et al., 2009; Kudou et al., 2011). Titration of HupZ-Strep with externally added heme revealed the formation of a growing UV-VIS absorption peak at 404 nm that is indicative of heme-binding ( Figure 1B ). However, the heme bound form of HupZ-Strep exhibited a shift in absorption maxima compared to the holo HupZ-His₆, which has a 414-nm Soret peak ( Figure 1C ). Additionally, unlike HupZ-His₆, the absorption spectrum of HupZ-Strep did not include the α and β bands between 500 and 600 nm, implying HupZ-Strep binds heme without an axial heme ligand that coordinates the iron (Giovannetti, 2012).

The crystal structure of HupZ-His₆ indicates that the protein’s C-terminus is close to where the HupZ dimers form a quaternary β-barrel (Sachla et al., 2016). To avoid the possibility that a C-terminal addition may interfere with the function and stability of HupZ, we constructed N-terminal fusions to two different MBP-containing vectors (His-MBP-HupZ and MBP-HupZ) and purified the proteins ( Figure 1A ). We generated a tag-less HupZ by cleaving His-MBP-HupZ with TEV protease ( Figure 1B ), leaving a serine residue after cleavage. We assayed heme binding by incubating 10 μM of tag-less HupZ with increasing heme concentration ( Figure 2A ). To confirm heme binding, the protein was allowed to set with 2× concentration of heme overnight and then passed through a PD-10 desalting column to remove any upbound heme ( Figure 2B ). Overall, tag-less HupZ exhibited a heme-binding spectrum similar to HupZ-Strep and different from HupZ-His₆ ( Figures 1D, E ). Tag-less HupZ’s Soret was broad and peaked at 385 nm. The 500- through 600-nm range lacked the α and β bands that indicated coordination of the central heme-iron.

Tag-less HupZ was not highly stable and prone to precipitate. Due to the solubility problems, further testing used a fusion to MBP (MBP-HupZ), which often aids in protein solubility. Unlike HupZ-Strep and tag-less HupZ, MBP-HupZ is soluble. Heme titration and reconstitution experiments demonstrated that the holo MBP-HupZ’s UV-VIS spectrum ( Figures 3A, C ) is similar to heme binding by tag-less HupZ. We also assess the heme-binding of a purified MBP protein as a negative control. MBP bound only a negligible amount of heme, exhibiting a vastly different absorption spectrum during heme titrations and reconstitutions ( Figures 3B, D ). MBP-HupZ also migrated as a monomer on native PAGE, indicating that it does not assemble into a high oligomeric state in vitro like HupZ-His₆ ( Figure 1C ) (Traore et al., 2021). Therefore, all three recombinant HupZ proteins bind heme, albeit without an axial heme ligand. These findings suggest that heme-binding is native to the HupZ protein and independent of the tag, the location of the fusion, or the multimeric state.

Histidine is a common residue in short peptides that bind heme and often functions as the heme axial ligand in hemoproteins (Wissbrock et al., 2019). The imidazole moiety of histidine interacts with iron and other transition metals during binding. Since holo tag-less and MBP-HupZ did not exhibit iron coordination, we tested if exogenous imidazole could serve this function in vitro ( Figures 2B and 3C ). In both recombinant proteins, the addition of 1.6 mM imidazole caused a shift in the Soret peak, the generation of α and β bands, and the protein solution turned red ( Figures 2B inset and 3C inset). As a control, we also measured the UV-VIS spectrum of MBP after incubation with heme, which exhibited minor spectral changes and still lacked α and β bands ( Figure 3D ). The UV-VIS spectra of free heme did not change with the addition of imidazole ( Figure 2C ). These observations suggest that exogenous imidazole can coordinate the iron in holo tag-less and MBP-HupZ but not in MBP or free heme. Together, this indicates that HupZ binds heme without an axial heme ligand but can readily interact with an exogenous imidazole group to coordinate the iron.

HupZ-His₆ degrades heme in vitro, releasing CO, free iron, and a chromophore (Sachla et al., 2016). Since externally added imidazole can coordinate the heme iron in holo-MBP-HupZ, we tested if MBP-HupZ can also break down the heme under these conditions. Holo-MBP-HupZ was incubated with 1.6 mM imidazole, ferredoxin (as a reducing agent), NADPH, an NADPH regeneration system (glucose-6-phosphate and glucose-6-phosphate dehydrogenase), and catalase (to control for non-enzymatic degradation of heme by hydrogen peroxide) and allowed to run for 6 h ( Figure 4A ). As indicated by the arrows, the Soret and the α and β bands decreased steadily during incubation, indicating heme degradation ( Figure 4B ). Still, the MBP-HupZ reaction did not result in an absorption peak at 600–700 mm, indicating the formation of biliverdin or similar molecules. In vitro heme degradation by some heme oxygenases (e.g., HemO and HemO/PigA) can result in ferric-biliverdin, which has no absorption properties (Zhu et al., 2000; Ratliff et al., 2001). To liberate the iron, we treated the MBP-HupZ reaction with acid. As with HemO, the reaction acidification caused the reduction of the α and β bands and formed a chromophore, although at 660 nm and not 680 nm as with HemO. Hence, holo MBP-HupZ can degrade heme only when an externally provided imidazole group is present to coordinate the iron. No significant spectral changes were observed in the absence of imidazole ( Figure 4B ), indicating that holo-MBP-HupZ did not degrade the heme.

All of the heme-degrading proteins described to date catalyze the breakdown of heme b. The one exception is Pden_1323, from Paracoccus denitrificans, who also belongs to the HugZ family. Pden_1323 lacks the C-terminal loop that contains the axial heme ligand (His245 in HugZ) but degrades heme c bound to a cytochrome fragment (MP11) (Li et al., 2021). In cytochrome c, the heme iron is coordinated by a histidine in proteaceous region attached to heme c and that histidine is retained in MP11 (Kranz et al., 2009). Since MBP-HupZ can degrade heme if the iron is coordinated by exogenous imidazole, we also tested if it could degrade heme c provided by MP11. MBP-HupZ bound MP11 with a 406-nm Soret and had α and β bands ( Figure 4D ). MBP-HupZ bound to MP11 was then tested for degradation with ferredoxin, NADPH, and NADPH regeneration system (in the presence of catalase). The reaction resulted in a progressive, though limited, decrease in the Soret and the α and β bands ( Figure 4C ). Like with heme b, the HupZ reaction did not produce an absorption peak between 600 and 700 mm. However, subsequent acidification of the solution led to forming a 615-nm chromophore ( Figure 4D ). Together, the data show that using ferredoxin, HupZ only moderately degrades heme b and heme c in vitro. It is possible that the fusion with MBP may hinder the degradation or in vitro conditions are not optimal.

To test if the observed heme c in vitro degradation by HupZ is biologically relevant, we constructed a ΔhupZ mutant in GAS by insertion inactivation and assessed the ability of both the wild-type rescue and ΔhupZ strains to use a fragment of heme c (MP11) as an iron source ( Figure 5 ). Inactivation of hupZ had a small positive impact on growth in the regular laboratory THYB ( Figure 5A ). Adding the iron chelator dipyridyl to THYB (THYB-DP) restricted growth in both strains to less than 20% and supplementing the medium with a range of 5–20 μM of MP11 could not significantly restore growth in either strain of bacteria, as indicated by a 1-way ANOVA across treatment types within each strain. There was also no significant difference between wild-type and ΔhupZ strains when comparing identical treatment conditions using a Student’s t-test ( Figure 5 ). A 2-way ANOVA covering aggregated data points for wild-type percentage growth to aggregated ΔhupZ percentage growth indicated that there was a significant mean difference in the wild type compared to the mutant in the iron-depleted media (17.5% to 14.2%, respectively). Indicating that while GAS does not utilize MP11 as an iron source, HupZ may provide relief from heme c toxicity.

To further evaluate the role of HupZ in heme use in vivo, we examined the mutant and the wild-type rescue strain for heme use and sensitivity. Unlike with MP11, supplementation of THYB-DP with hemoglobin at a range of 2.5 to 20 μM restored GAS growth in THYB-DP, indicating that both strains can use hemoglobin as an iron source ( Figures 6A, B ). The hemoglobin dose–response was delayed in the ΔhupZ strain compared to the wild-type rescue strain. For complementation, we expressed hupZ in trans from GAS recA promoter. Comparing the complemented and control (empty vector) strains revealed a small but statistically significant difference between the strains when grown on hemoglobin iron ( Figures 6C, D ). First, data show that losing hupZ reduces GAS’ ability to use heme iron.

The moderate phenotype of the hupZ mutant suggests redundancy in mechanisms to gain iron by GAS. Hence, we also tested the influence of hupZ expression on hemoglobin iron use by a heterologous host. HupZ was expressed from the P_nis promoter in L. lactis ( Figure 6E ). Lactococcal growth in GM17 was inhibited with 10 mM DP. Adding hemoglobin to the iron-restricted medium restored growth, indicating that L. lactis can use hemoglobin as an iron source. L. lactis expressing hupZ responds to a lower hemoglobin concentration and reaches a higher maximal density than the L. lactis harboring an empty vector. Second, HupZ promotes the use of heme iron in both GAS and L. lactis.

We tested the sensitivity of the GAS mutant and the wild-type rescue strains to free heme ( Figure 6F ). The addition of 5 mM heme to THYB restricted the growth of the hupZ mutant while it had no impact on the rescue strain. The addition of 10 μM heme inhibited the growth of both strains, but the wild-type rescue strain grew better than the hupZ mutant. Third, hupZ helps GAS manage heme toxicity at a low heme concentration.

Colonization of the host mucosa constitutes an important first step in GAS pathogenesis in vivo (Walker et al., 2014; Zhu et al., 2020). We used a murine colonization model to determine whether hupZ plays a role in the ability of GAS to colonize the vaginal mucosa. When wild-type rescue and ΔhupZ mutant cells were mixed at a 1:1 ratio and inoculated intravaginally into mice, they were initially able to colonize in approximately equal ratios ( Figure 7 ). Following the initial vaginal lavage on day 1, the mutants were no longer able to compete with the WT cells, and the ratio of recovered WT:ΔhupZ mutant cells increased dramatically. The data indicate that hupZ is important in allowing GAS to maintain host surface colonization.