Endogenously produced H2O2 is intimately involved in iron metabolism in Streptococcus pneumoniae

ABSTRACT In the presence of molecular oxygen, the human pathogen Streptococcus pneumoniae produces and secretes large amounts of hydrogen peroxide (H2O2), which can readily interact with free and heme-bound iron. Here, we investigated the role of the endogenously produced H2O2 in iron acquisition. The data revealed that S. pneumoniae uses H2O2 to liberate iron from met-hemoglobin (Hb-Fe3+) extracellularly, allowing the bacterium to import and grow on free iron even when cultivated on met-hemoglobin as the only iron source. The loss of H2O2 production leads to a dramatic pneumococcal intake of heme and is associated with a robust upregulation of most iron uptake machinery (indicating an iron starvation signal). These and other data reveal a close and previously unexplored relation between H2O2 production and iron metabolism in S. pneumoniae. The data also show that, in addition to extracellular degradation, pneumococci are armed with H2O2-independent mechanisms for intracellular heme catabolism. IMPORTANCE Heme degradation provides pathogens with growth essential iron, leveraging on the host heme reservoir. Bacteria typically import and degrade heme enzymatically, and here, we demonstrated a significant deviation from this dogma. We found that Streptococcus pneumoniae liberates iron from met-hemoglobin extracellularly, in a hydrogen peroxide (H2O2)- and cell-dependent manner; this activity serves as a major iron acquisition mechanism for S. pneumoniae. Inhabiting oxygen-rich environments is a major part of pneumococcal biology, and hence, H2O2-mediated heme degradation likely supplies iron during infection. Moreover, H2O2 reaction with ferrous hemoglobin but not with met-hemoglobin is known to result in heme breakdown. Therefore, the ability of pneumococci to degrade heme from met-hemoglobin is a new paradigm. Lastly, this study will inform other research as it demonstrates that extracellular degradation must be considered in the interpretations of experiments in which H2O2-producing bacteria are given heme or hemoproteins as an iron source.

The total iron pool in human adults is around 3-4 g of which about 75% is in the form of heme mostly bound to hemoglobin (6).To avoid toxicity and overcome solubility problems, the host sequesters virtually all the iron in the extracellular and intracellular milieus using specialized proteins that transport, bind, or store iron or heme.During infection, iron availability is further decreased due to a coordinated host response named nutritional immunity (7,8).Hence, invading bacteria that require iron, such as S. pneumoniae, rely on dedicated mechanisms to capture the metal from host proteins.We recently showed that heme, hemoglobin, and several other host heme sources restore pneumococcal growth in an iron-depleted medium.Notably, hemoglobin promotes growth in vitro to a greater capacity than other iron or heme sources and supported growth in concentrations that were toxic with equivalent amounts of free heme (9).In addition to facilitating vigorous planktonic proliferations, hemoglobin induces early and robust pneumococcal biofilms in vitro and drives global transcriptome changes enabling the pathogen adaptation to the host mucosal environment (10).These and other observations established hemoglobin as a principal nutrient and host signal for pneumococci (9).
Gram-positive pathogens often employ protein-relay machinery consisting of surface proteins that capture the heme from the host and deliver it across the thick cell wall and eventually to the substrate-binding component of a membrane-bound ABC transporter (11).The molecular mechanisms for heme or iron acquisition are not fully explained in S. pneumoniae.Surface receptors such as those that shuttle heme across the cell wall in other Gram-positive bacteria have not been described in this human pathogen.Several ABC transporters involved in iron uptake were identified, but the in vivo ligand of some transporters remained undetermined.The two pneumococcal ABC systems proposed to import ferric iron are the pit1ADBC (spd_0223-0227) and pit2ABC (spd_1607-1609) transporters (12,13).In the D39 strain, the substrate-binding component, pitA1 is truncated, but the permeases and ATP-binding protein are intact.S. pneumoniae also encodes the pia transporter of which PiaA, the ligand-binding protein, was crystalized with ferrochrome (14).Previous studies also demonstrated that the Pia proteins support pneumococcal growth on heme iron (15).Similarly, the PiuBCDA transporter promotes streptococcal binding to heme and hemoglobin (16), but the substrate-binding component, PiuA, binds in vitro norepinephrine and enterobactin in higher affinity than heme (17).A recent study implicated an additional pneumococcal ABC transporter, spd_0088-0090, in heme uptake (18).Due to the redundancy of iron uptake mechanisms, genetic knockouts in multiple transport systems are required before attenuated pneumococcal growth is observed (13).
Several studies reported the presence of free hemoglobin in the mucosa of the upper respiratory tract due to bleeding and erythrocyte lysis (19).In intact erythrocytes, hemoglobin iron is found in its reduced form (ferrous, Fe +2 ) bound to oxygen (oxyhe moglobin) or carbon dioxide.Upon erythrocytes lysis, the iron in oxyhemoglobin is autoxidized to ferric (Hb-Fe +3 ) forming met-hemoglobin, which cannot bind oxygen (20).Oxyhemoglobin interactions with hydrogen peroxide (H 2 O 2 ) lead to heme degradation; these reactions start with the formation of ferryl hemoglobin (Hb-Fe +4 =O), which in turn mediates one-electron oxidation of H 2 O 2 creating a superoxide radical in the heme-binding pocket.Subsequent reactions with the porphyrin ring degrade the heme and release free iron and two fluorescent products (20,21).On the other hand, methemoglobin interactions with H 2 O 2 do not lead to heme degradation but catalytically consume the H 2 O 2 (20).In these catalase-like reactions, one H 2 O 2 molecule serves as a two-electron acceptor leading to the formation of oxo-ferryl complex (Fe +4 =O) and transient apoprotein radicals (e.g., *Hb-Fe +4 =O) while a second H 2 O 2 molecule serves as a two-electron donor that reduces the ferryl hemoglobin back to met-hemoglobin while producing molecular oxygen (O 2 ) (20,22,23).
In the human host, pneumococcus is exposed to oxygen levels that vary from 20% to 5% in the upper to lower respiratory tracts to virtually no free oxygen in the blood (24,25).In the presence of oxygen, S. pneumoniae produces H 2 O 2 as a byproduct of the enzymatic reactions of the pyruvate oxidase (spxB) and lactate oxidase (lctO) (26,27).The SpxB and LctO enzymes are responsible for ~80% and ~20% of the pneumococcal H 2 O 2 production, respectively (28).SpxB activity augments the pneumococcal release of Ply, a pore-forming toxin, which promotes erythrocyte lysis (29), likely increasing the availability of free hemoglobin in the immediate bacterial environment.Our group demonstrated that spxB is highly expressed during growth with hemoglobin (9).We showed that endogenously produced H 2 O 2 catalyzes the oxidation of oxyhemoglobin to met-hemoglobin and provided spectroscopic data supporting the H 2 O 2 -mediated degradation of the heme in oxyhemoglobin from rupturing erythrocytes (30).Still, the impact of H 2 O 2 on met-hemoglobin, the dominant form of hemoglobin in the extracellu lar compartment, and on pneumococcal iron acquisition overall was not described.In this study, we aim at these knowledge gaps and show that S. pneumoniae uses H 2 O 2 to acquire iron from met-hemoglobin and that H 2 O 2 production is closely linked to iron and heme metabolism in this key human pathogen.

The interactions between H 2 O 2 and met-hemoglobin promote their removal from the medium during pneumococcal growth
Since free oxyhemoglobin is rapidly oxidized by H 2 O 2 (31,32), we rationalized that S. pneumoniae encounters mostly met-hemoglobin in the respiratory mucosa.To begin investigating the impact of H 2 O 2 production on pneumococcal interactions with met-hemoglobin and iron uptake, we constructed a double ∆spxB∆lctO mutant and a complemented strain (expressing spxB and lctO genes from a heterologous location in the chromosome) in the background of the D39 strain.We monitored the H 2 O 2 levels in the culture media during growth and confirmed that the ∆spxB∆lctO mutant did not produce H 2 O 2 at detectable levels, while we found 5-7 mM of H 2 O 2 in the culture supernatant of wild-type (WT) and the complemented strains (Fig. 1A).
We next examined the H 2 O 2 amount in pneumococcal cultures growing in the presence of met-hemoglobin (Fig. 1B).We solubilized lyophilized hemoglobin (Sigma) in phosphate buffer saline (PBS) and determined the absorbance spectrum following incubation with oxidative (potassium ferricyanide) or reducing (DTT) agents.The Soret peak of the hemoglobin incubated with or without potassium ferricyanide overlapped.In the presence of DTT, however, the absorbance maximum shifted right from 408 to 418 nm, indicating iron reduction (Fig. 1B, insert).These spectral characteristics indicate that the hemoglobin in our stock solution is in the oxidized (met-hemoglobin) state.Analysis of the H 2 O 2 levels in the culture medium revealed that when the WT and complemented strains were grown in Todd-Hewitt broth (THYB) supplemented with 20-µM met-hemoglobin, H 2 O 2 was not detectable at the 3-and 6-hour time points (Fig. 1B).Hence, the H 2 O 2 produced by pneumococci is removed, likely due to reactions with the medium met-hemoglobin.Low amounts of H 2 O 2 (10%-25% compared to THYB) were detected after 18 hours suggesting met-hemoglobin depletion during incubation allows for some accumulation of H 2 O 2 at the later time point.
While the hemoglobin remains soluble in un-inoculated THYB, visual inspection suggested that some of the met-hemoglobin is falling out of solution, with more precipitation occurring in the WT cultures compared to the ∆spxB∆lctO mutant.Hemo globin precipitation was confirmed by centrifugation and washes of the cell pellet.We used spectroscopic measurements to follow the levels of the hemoglobin that remained in solution during growth.The hemoglobin displayed a somewhat broad Soret peak, suggesting that some of the heme iron in the hemoglobin was reduced by the medium.Nevertheless, we did not see distinct changes in the Soret maximum during incubation in regular THYB even after 18 hours of incubation.On the other hand, in a medium that was inoculated with the WT S. pneumoniae, the hemoglobin spectrum maximum was decreased and left shifted in the first 3 hours forming the narrow peak at 408 nm, which is typical of oxidized hemoglobin (met-hemoglobin).The Soret peak decreased dramati cally within the next 3 hours of incubation and eventually was below the detection level by the 18-hour time point (Fig. 1C).In the ∆spxB∆lctO culture, the changes in the Soret happened at a much lower rate; no changes in absorption were observed in the first 3 hours of incubation.While the Soret peak decreased during longer incubation, it was significantly higher compared to the Soret maximum with the WT strain at the 6-hour time point.Altogether, the data suggest that H 2 O 2 produced by S. pneumoniae interacts with the met-hemoglobin in the medium via reactions that remove both H 2 O 2 and the protein.Still, mechanisms that are independent of H 2 O 2 also contribute to the reduction in the levels of soluble met-hemoglobin in the pneumococcal growth medium.

H 2 O 2 reaction mediates iron release from met-hemoglobin in a cell-depend ent manner
We next tested if the interactions between the endogenously produced H 2 O 2 and met-hemoglobin release iron into the culture supernatant.THYB supplemented with 20-µM met-hemoglobin was allowed to incubate at 37°C.The hemoglobin was then removed by filtration, and the concentration of free iron in the medium was determined at different time points using the chromophore ferrozine (33).The addition of met-hemo globin to THYB led to an immediate release of small amounts of iron (~20 nM), but no significant change in iron levels was observed from that point during incubation (insert in Fig. 2A).Hence, it seems possible that a minute fraction of the hemoglobin solution contains free iron.Still, the hemoglobin was generally stable in THYB and did not spontaneously release iron during incubation in THYB at 37°C.Notably, when THYB supplemented with 20-µM met-hemoglobin was inoculated with the WT D39 cells, the free iron concentration continued to rise above the baseline observed with an un-inoculated medium.The addition of catalase prevented this rise in iron level in the early time points, and minuscule amounts of additional iron were detected after 18 hours (Fig. 2A).Moreover, we did not observe iron release from met-hemoglobin into the medium in the ∆spxB∆lctO culture, while the complemented strain released iron in higher amounts compared with the WT strain (Fig. 2A).Hence, H 2 O 2 produced by the growing pneumococci acted to release iron from met-hemoglobin into the extracellular environment.
To examine if H 2 O 2 was sufficient to cause iron release in the pneumococcal culture, we added met-hemoglobin into the supernatant samples collected from overnight cultures of the WT (which contains ~7-mM H 2 O 2 , Fig. 1A).Spent medium collected from ∆spxB∆lctO culture served as a negative control.Interestingly, we did not see an increase in free iron during incubation of met-hemoglobin in either spent media (Fig. 2B).Hence, endogenously produced H 2 O 2 aids iron release, but it is not exclusively sufficient to release iron from the externally added met-hemoglobin as this process also requires the bacterial cells.A 20-µM met-hemoglobin was added to supernatant samples collected from overnight cultures of D39 WT or ∆spxB∆lctO strains, and the samples were allowed to incubate at 37°C.The iron concentration in each sample was determined after the hemoglobin was removed by filtration (as in A).Insert shows the concentration of free iron in the reactions.Subtracting the iron concentration found in un-inoculated THYB with 20-µM met-hemoglobin (insert in A) from those found in spent media incubated with hemoglobin revealed no net iron release.Data were derived from experiments done in duplicates and repeated twice and analyzed by student t-test, where * indicates P ≤ 0.05 and ** indicates P ≤ 0.01.

S. pneumoniae grown on met-hemoglobin as an iron source imports mostly free iron
To better understand how H 2 O 2 production impacts S. pneumoniae use of met-hemoglo bin as an iron source, we aimed to determine the heme and iron content in cells growing on met-hemoglobin iron.For this, we explored two separate growth assays in which we depleted the free iron from the medium by either using THYB pretreated with Chelex-100 resin (THYB CHX , Fig. 3A) or in THYB containing the iron chelator, nitrilotriacetic acid (NTA) (THYB NTA , Fig. 3B).The main difference between these two growth assays is that with THYB CHX , the iron is removed ahead of inoculation from the medium.Hence, when met-hemoglobin is added (Fig. 3C and D), the bacteria can use both the heme and free iron (if it is freed into the supernatant during growth).In THYB NTA , however, the only source of iron that is available for pneumococcus is the met-hemoglobin (Fig. 3E and F), since NTA is in excess and will chelate any iron released into the medium.
Chelating the iron in THYB either by Chelex or by NTA impaired pneumococcal growth (Fig. 3, empty symbols).The addition of 5-and 20-µM met-hemoglobin to the media restored the growth of both the WT and the ∆spxB∆lctO strains in a dose-depend ent manner (Fig. 3C-F).Hence, growth arrest in THYB CHX or THYB NTA resulted from iron depletion.Moreover, ∆spxB∆lctO growth in an iron-depleted medium supplemented with met-hemoglobin demonstrates that H 2 O 2 production is not essential for pneumococcus to use met-hemoglobin as an iron source and the pathogen has additional means to liberate iron from heme.
We next measured the heme content in cells collected at different time points during cultivation in THYB CHX or THYB NTA supplemented with 20-µM met-hemoglobin (Fig. 4A  and B).Respectively, the tested strains accumulated heme with the heme levels peaking at the logarithmic phase of growth (6 hours) and then reduced by the 18-hour time point.The WT strain contained more heme when grown in THYB NTA compared with THYB CHX , exhibiting four times more heme at the late logarithmic phase of growth (Fig. 4B and A, respectively).We also tested heme accumulation by a knockout of the piuBCDA system, previously implicated in heme import.The ∆piuBCDA strain imported signifi cantly less heme than the WT, when grown in THYB NTA (Fig. 4B), confirming, for the first time, that the Piu proteins contribute to heme accumulation in vivo.Interestingly, cellular heme levels in the WT and ∆piuBCDA were comparable during growth in THYB CHX (Fig. 4A).
The most noticeable difference in heme content was exhibited by the ∆spxB∆lctO mutant, which accumulated much more heme than the other strains (fivefold more than the WT in THYB CHX and twofold more in THYB NTA , Fig. 4A and B).Moreover, the cellular levels of heme in this strain remained high even after 18 hours of incubation.Like the WT and ∆piuBCDA, the ∆spx∆lctO strain also imported more heme when grown in THYB NTA compared with THYB CHX, Despite the big difference in heme content between the WT and the ∆spxB∆lctO strains when grown on met-hemoglobin iron in either THYB CHX or THYB NTA , inductively coupled plasma mass spectrometry (ICP-MS) analysis (done at the 6-hour time point) did not reveal meaningful difference in cellular iron content between these two strains (Fig. 4C).We observed an upward trend in total iron in cell samples collected from THYB NTA compared to those collected from THYB CHX , but the difference was not statistically significant.Therefore, the WT strain imported mostly free iron when grown on methemoglobin iron.In the absence of extracellular heme degradation, the ∆spxB∆lctO imported heme to fulfill its iron needs.Importantly, the two strains balance iron and heme uptake and keep total cellular iron levels at a steady state regardless of the iron source available to them.
We also evaluated the total heme concentration (i.e., bound to met-hemoglobin and free) in the medium during growth of the WT, ∆piuBCDA, or ∆spxB∆lctO strains in THYB NTA supplemented with met-hemoglobin (Fig. 5A).In these experiments, we removed the cells and then extracted the heme from the medium using acidified chloroform as described (34).In the cultures of the WT and ∆piuBCDA strains, about 30% of the heme found in the medium after 3 hours of growth was spent by the 6-hour time point and was almost completely diminished after 18 hours of incubations.The heme was also removed during the growth of the ∆spxB∆lctO cells but at a lower rate.The culture supernatant of this strain contained ~33% more heme than those of the WT and ∆piuBCDA strains even after 18 hours (Fig. 5A).To determine the fraction of free heme (unlike heme that is bound to met-hemoglobin), we repeated the assay after removing the met-hemoglobin by filtration (Fig. 5B).We found only minute amounts (~4% of the total heme) of free heme in the medium of all cultures, indicating that most of the heme remaining in the supernatant at any time point is still bound to met-hemoglobin.

Inactivation of H 2 O 2 production drives overexpression of iron and ironcomplex importers
The growth of the ∆spxB∆lctO mutant was significantly impaired in THYB CHX and in THYB NTA compared with regular THYB (Fig. 3A and B, gray symbols,).Still, the ∆spxB∆lctO mutant exhibited limited growth in these iron-depleted media (gray empty symbols), where the growth of the WT strain was completely diminished (blue empty symbols), indicating that it is more resistant to iron depletion.To compare iron metabolism in these two strains, we compared the expression of iron-related genes during growth in regular THYB.RNA was prepared from cell samples collected at the mid-logarithmic phase, and gene expression was analyzed by reverse transcription PCR (RT-PCR).We observed strong activation of several iron and heme importers in the ∆spxB∆lctO strain (Fig. 6).This includes the activation of pitA2 (14-fold), pitD1 (8.7-fold), spd_0090 (5.2-fold), and piaA (1.6-fold).The expression of spd_0310 (35), encoding a putative heme shuttling protein, was also induced 1.6-fold.Curiously, the piuB gene was strongly downregulated (0.067fold) in the ∆spxB∆lctO mutant.Hence, inactivation of the spxB and lctO has a significant impact on iron homeostasis in S. pneumoniae.

Pneumococci growing on met-hemoglobin in THYB NTA experience iron depletion compared with those growing in THYB CHX
We noticed that all three strains tested for growth on met-hemoglobin iron grew faster and reached higher biomass in THYB CHX repleted with met-hemoglobin compared to repleted THYB NTA (Fig. 3C and D compared with Fig. 3E and F).To gain more insights into pneumococcal physiology during cultivation using met-hemoglobin as a single source of iron, we compared the expression of the pneumococcal genes implicated in iron or heme uptake between cells grown in THYB CHX or THB NTA supplemented with limiting amounts (5 µM) met-hemoglobin.RNA was prepared from samples collected at the midlogarithmic phase of growth, and the expression of selected iron homeostasis genes was analyzed by RT-PCR (Fig. 7).Although both media were supplemented with equal amounts of met-hemoglobin, we observed differential expression of several transporters.The most pronounced change was the 10-fold induction of pitA2 (spd_1609) during growth in THYB NTA compared with the Chelex-treated THYB.Smaller but more significant induction of the piuB (1.9-fold) and pitD1 (spd-0225, 1.7-fold) was also observed.The elevated expression of pitD1 during growth in THYB NTA compared with THYB CHX is consistent with the observation that the ∆piuBCDA strain imported less heme than the WT during growth in this medium, while the piu loss did not impact heme content during growth in THYB Chx .Interestingly, cells grown on hemoglobin iron in the presence of NTA downregulate the expression of spd_0090 (0.54-fold), which is part of a recently described heme importer, and of piaA (0.43-fold), a ferrochrome-binding protein (18).The repression of spd_0090 and piaA genes suggests different regulation modalities for these transporters.
FIG 5 The heme in culture medium containing met-hemoglobin is spent faster by H 2 O 2 -producing S. pneumoniae.Cells of the isogenic strains D39 WT (blue), ∆piuBCDA (orange), and ∆spxB∆lctO (gray) were allowed to grow in THYB NTA supplemented with 20-µM met-hemoglobin (as described in Fig. 3).Culture samples (2 mL) were collected along the growth, the cells were removed by centrifugation, and total heme content was determined by the acidified chloroform method (A), or the hemoglobin was removed by filtration prior to determination of the free heme concentration (B).The data were derived from experiments done in duplicates and repeated twice.The Student t-test was used to determine statistical significance, where * indicates P ≤ 0.05 and ** indicates P ≤ 0.01.

FIG 6
The ∆spxB∆lctO mutant overexpress iron uptake genes compared with the WT strain.Shown are the fold changes in the expression of iron uptake and shuttling genes in the ∆spxB∆lctO mutant compared to the isogenic WT strain as determined by qRT-PCR.RNA was prepared from cells grown in THYB up to the mid-log phase of growth.Data derived from two bioreplicates and processed in duplicates are shown in a bar graph and table format.

DISCUSSION
The human pathogen, S. pneumoniae, requires iron and can readily obtain the metal from heme and host hemoproteins, with hemoglobin the most growthbeneficial source compared with other iron sources (9).Still, the molecular mechanisms by which pneumococci obtain and import heme are only partially described, and the literature contains inconsistent reports regarding the function of transporters implicated in iron acquisition (15)(16)(17)36).Additionally, it is unknown how this important pathogen degrades heme to liberate iron.One confounding factor that was largely overlooked in previous studies of iron metabolism in S. pneumoniae is that in the presence of oxygen, the bacterium produces and releases to the extracellular environment copious amounts of H 2 O 2 , which can rapidly interact with free and hemoglobin-bound iron.While H 2 O 2 degrades both ferrous heme and ferric heme that are free in solution, the fate of the heme that is bound to hemoglobin depends on the iron redox state (20-23, 31, 32).In this study, we began describing the role of H 2 O 2 production in iron acquisition by S. pneumoniae, uncovering intimate relationships between iron metabolism and H 2 O 2 production as illustrated in Fig. 8.
Examining the impact of met-hemoglobin on pneumococcal cultures showed that the millimolar concentration of H 2 O 2 produced by the pneumococcal SpxB and LctO enzymes is removed from the medium in the presence of micromolar amount of methemoglobin (Fig. 1B).This catalytic removal of H 2 O 2 is consistent with the catalase-like activity of met-hemoglobin, whereby H 2 O 2 acts as both an electron donor and acceptor in an oxidation-reduction cycle of Hb-Fe +3 and Hb-Fe +4 (32).H 2 O 2 removal likely reduces the consequential oxidative stress pneumococcus experiences during growth likely contributing to the dramatic growth benefits we observed when S. pneumoniae is cultivated in the presence of met-hemoglobin (9).Interestingly, spxB expression aug ments the release of Ply by S. pneumoniae (29), which in turn promotes erythrocyte lysis and the release of hemoglobin to the medium (29).Hence, it is plausible that in vivo, H 2 O 2 -producing pneumococci (such as those colonizing the respiratory tract) also benefit from the removal of H 2 O 2 in their vicinity by met-hemoglobin.Met-hemoglobin is stable in a cell-free medium, but it is falling out from the medium during incubation with growing streptococci.We notice a faster decrease in soluble methemoglobin level in the WT cultures compared with the ∆spxB∆lctO mutant (Fig. 1C).Oxidation of met-hemoglobin by H 2 O 2 produces reactive amino acid radicals, destabil izes the globin chain, and promotes met-hemoglobin conversion to hemichrome (methemoglobin complexes in which a histidine residue, distal or external, binds to the iron sixth position).Since hemichrome readily precipitates out of solution forming Heinz bodies in vivo (37), it is likely that hemichrome formation in the presence of H 2 O 2 enhances the loss of met-hemoglobin from the solution.We observed met-hemoglobin precipitation also in ∆spxB∆lctO cultures although less than with the WT.Hence, addi tional streptococcal-dependent but H 2 O 2 -independent mechanisms, possibly the reduction in the medium pH during bacterial growth, also lead to met-hemoglobin denaturation (38).The hemoglobin that falls out of the solution likely supports the robust biofilm growth we observe when S. pneumoniae is grown in the presence of methemoglobin (10).
It is well documented that H 2 O 2 degrades free heme and heme that is bound to oxyhemoglobin (20,22,23,31), while its reactions with met-hemoglobin heme do not lead to iron release (23).Therefore, we were surprised to find that incubation of met-hemoglobin with H 2 O 2 -producing streptococci led to accumulation of free iron in the medium, albeit in low amounts (Fig. 2A).Extracellular iron accumulation was not observed with the ∆spxB∆lctO mutant, and externally added catalase blocked this buildup in the WT cultures (Fig. 2A).These observations suggest that iron is released extracellularly from met-hemoglobin heme by a H 2 O 2 -dependent process.Incubation of met-hemoglobin in a spent medium that contains H 2 O 2 did not lead to iron release from met-hemoglobin (Fig. 2B), indicating that extracellular degradation of the heme in met-hemoglobin also requires the presence of pneumococcal cells.
Pneumococcus has two surface-exposed thioredoxin proteins, Etrx1 and Etrx2, that help them cope with oxidated stress by maintaining a reductive outward environment (32).It seems possible that these enzymes reduce the ferric iron in met-hemoglobin molecules, making it suspectable to degradation by H 2 O 2 .Alternatively, binding of met-hemoglobin to pneumococcal receptors may prompt local heme release, which is then degraded by H 2 O 2 .The contribution of H 2 O 2 -mediated degradation of free heme that was pumped out by pneumococci cannot be overruled as well.Still, the low amounts of iron found in the medium (Fig. 2A) compared with the relatively high concentration of met-hemoglobin present (Fig. 5) support a model by which there is a localized iron release from met-hemoglobin, which is in turn taken up by the cells.
To examine how extracellular heme degradation impacts pneumococcal iron metabolism, we compared growth and cellular heme and iron content among pneu mococci grown on met-hemoglobin iron.All tested strains were able to grow (Fig. 3) and accumulated heme (Fig. 4A and B) in either THYB CHX or THYB NTA supplemented with met-hemoglobin.The piuBCDA mutant accrued significantly lower heme amounts compared with the parent WT strain when grown in THYB NTA (orange and blue symbols, Fig. 4B).This phenotype indicates that active uptake promotes cellular heme accumula tion and establishes the Piu proteins as an important mechanism when S. pneumoniae is dependent on external heme supply.Despite these differences in heme content, the growth of the ∆piuBCDA mutant was like that of the WT parent (Fig. 3), consistent with the presence of additional transporters whose activity compensates for the loss of the Piu proteins.
Cellular heme levels peaked during the mid-logarithmic phase of growth in most cultures and then decreased.We suggest that intracellular heme catabolism and possibly heme export facilitated this decrease in heme levels during long incubation.Since the free iron in the medium is chelated, pneumococci cultivated in THYB NTA with met-hemo globin must degrade the heme in the intracellular compartment to obtain the metal.Hence, the growth of the ∆spxB∆lctO on met-hemoglobin in THYB NTA indicates the presence of pneumococcal mechanisms for intracellular heme degradation that are independent of H 2 O 2 .Interestingly, the only sample in which the heme level did not decline after prolonged incubation was that of ∆spxB∆lctO cells grown on met-hemoglo bin in THYB CHX .This observation suggests that H 2 O 2 may also contribute to intracellular heme degradation.It is also possible that H 2 O 2 -independent heme degradation comes into play only in the absence of H 2 O 2 production and without a free iron source (such as in THYB NTA ).The pneumococcal mediators of catalytic heme degradation are unknown at this time, as heme-degrading enzymes were not described, and the pathogen does not encode homologs of recognized heme-degrading proteins (39).Nevertheless, catalytic heme degradation is likely critical during infection in sites with low oxygen tension (where pneumococci do not produce H 2 O 2 ).
We found less heme in the WT compared with the ∆spxB∆lctO mutant when the bacteria were grown on met-hemoglobin iron (blue and gray symbols, Fig. 4A and  B).These differences in heme content were bigger when the cells were cultivated in THYB CHX than in THYB NTA (5-fold versus 2.5-fold, respectively) in which pneumococcal growth was dependent on heme import.Despite the difference in cellular heme, both strains contained equal amounts of iron (Fig. 4C).These findings suggest that the WT strain incorporates mostly free iron under our experimental growth conditions.Intracellular heme degradation or export may be enhanced in the H 2 O 2 -producing cells.Still, the reduction in heme content (about a quarter) within the WT strain when grown in THYB CHX compared with THYB NTA (blue symbols, Fig. 4A and B) supports the notion that the WT strain takes up free iron although it is cultivated in THYB CHX with met-hemoglobin as the only source of iron supporting extracellular heme degradation.
We found that inactivation of the genes for the H 2 O 2 -producing enzymes, spxB and lctO, resulted in wide activation of transporters for iron (pitD1 and pitA2, 8.7-9-fold), heme (spd_0090, 5.2-fold), and an unknown iron complex (piaA, 1.6-fold; Fig. 6).We hypothesize that the likely increase in iron import contributes to the resistance to iron stress exhibited by the ∆spxB∆lctO mutant (Fig. 4A and B).The strong repression of the piuB expression (0.067) is intriguing and may reflect the complexity of piu regulation.The transcription of the piu genes is regulated by multiple proteins including CodY and the orphan response regulator RitR (40,41) and environmental conditions such as oxygen tension, oxidation stress, and nutrition status.Nevertheless, the significant activation of several iron-related transporters in the ∆spxB∆lctO strain strongly associates H 2 O 2 production and iron metabolism in S. pneumoniae.
All pneumococcal strains grew to higher cell density on met-hemoglobin iron in THYB CHX compared with THYB NTA (Fig. 3C-F).In addition, WT pneumococci grown on 5-µM met-hemoglobin as an iron source exhibited 10-fold activation of the metal iron transporter, pitA2, and a small increase in the expression of pitD1 and piuB genes (1.7and 1.9-fold respectively) when grown in THYB NTA compared with THYB CHX (Fig. 7).This observation indicates that the transcription of the pitD1, pitA2, and piuB genes is responsive to iron (whose availability in the medium is limited in THYB NTA ) and not heme.The molecular mechanism for such iron regulation remains elusive; however, since S. pneumoniae does not code for obvious homologs of the iron-dependent repressors from the Fur families, a potential DtxR family member named SmrB has been identified.SmrB resembles the regulator of Mn uptake in Treponema pallidum, TroR, but its role in iron metabolism is unknown (42).Both media were supplemented with limiting but equal amounts of met-hemoglobin.Hence, it seems that pneumococci that are not allowed to benefit from external heme degradation by H 2 O 2 perceive more severe iron stress compared with those that grow without an external chelator.Together, these observations provide additional support to the idea that extracellular heme degradation by H 2 O 2 provides pneumococci with nutritional iron from met-hemoglobin.
S. pneumoniae is a versatile pathogen that can spread from the upper respiratory tract and survive in multiple niches in the host (2).In this study, we establish that H 2 O 2 is a key mediator in pneumococcal use of the host heme and hemoproteins as iron sources.We expect that in vivo, the H 2 O 2 -dependent pathway is important particularly during nasopharyngeal and lung colonization, while during invasive progressions, such as with bacteremia, pneumococci likely rely on additional mechanisms for iron acquisition that are independent of H 2 O 2 .Sickle cell anemia (SCA) is another in vivo condition where the interactions between pneumococcal-produced H 2 O 2 and hemoglobin described in this study may contribute to the infection outcome.Patients with SCA suffer from invasive pneumococcal disease at a 100 times higher rate than non-SCA patients despite not having distinguishable differences in their immune system (43).Interestingly, SCA hemoglobin (HbS) is more sensitive to H 2 O 2, with the ferrous form (HbS-Fe 2+ ) auto-oxi dizing to the ferric (HbS-Fe 3+ or met-HbS) at a nearly double the rate of normal hemoglo bin (43).It is, therefore, likely that endogenously produced H 2 O 2 enhances iron release, increases hemichrome generation, and promotes pathology in these patients.

Bacterial growth and media
Frozen stocks of S. pneumoniae were preserved in skim milk-tryptone-glucose-glycerin (STGG) as described (44) and stored at −80°C.S. pneumoniae cells from STGG stock were plated on tryptic soy blood agar plates (BAPs) (Becton Dickinson) and incubated at 37°C.Cells collected from BAPs following overnight incubation were used to inoculate fresh medium in a starting OD 600 of 0.05.Pneumococci were grown in THYB containing 0.5% (wt/vol) yeast extract (Becton Dickinson), iron-depleted THYB (THYB CHX or THYB NTA ), or iron-depleted THYB supplemented with hemoglobin.THYB CHX was prepared by adding 5% (wt/vol) Chelex-100 resin (Sigma-Aldrich) to THYB and incubating with continued mixing overnight.The resin was removed using a 0.45-micron filter, and the medium was supplemented with 2-mM MgCl 2 and 100-µM CaCl 2 .THYB NTA was prepared by adding 3-mM NTA (Sigma-Aldrich), 0.55-mM ZnSO 4 , and 0.55-mM MnCl 2 to fresh THYB followed by filter sterilization (0.45-micron filters).The metal composition of the chelated media was determined experimentally by supplementing back with different transition metals and testing growth; we chose eventually to include only the metals whose absence negatively impacted growth.We cannot exclude the possibility that a reduction in non-growth essential metals might have impacted the finding.Hemo globin stock solutions were prepared fresh by resuspending the lyophilized power of human hemoglobin (Sigma-Aldrich) in 1× PBS (1 mM).Hemoglobin was then added at the indicated concentration, and the medium was then filter sterilized (0.45-micron filters).Pneumococci were grown in 96-well flat bottom tissue culture plates at 37°C in a Multiscan SkyHigh Spectrophotometer (Thermo Scientific).Optical density was measured hourly after brief shaking.

Hemoglobin spectroscopic measurements
Absorbance spectra (250-700 nm) of hemoglobin samples in PBS or THYB were determined using a DU730 LifeScience UV/VIS or Beckman-Coulter DU 730 UV/Vis spectrophotometer in a quartz cell with an optical path length of 10 mm.In some experiments, 10-µM hemoglobin was incubated with 10-mM DTT (Sigma-Aldrich) or 30-µM potassium ferricyanide (Thermo Fisher) for 1 hour in PBS prior to spectroscopic measurements.

Bacterial strains and mutant construction
The bacterial strains and plasmids used in the study are listed in Table 1 and the primers in Table 2.The ∆spxB∆lctO knockout and its isogenic ΩspxB-lctO complemented strain were engineered S. pneumoniae D39 strain as described for the construction of this mutation and complementation in TIGR4 strain (45).The ∆piuBCDA mutant was created by replacing the piuBCDA gene cluster with the ermC gene from plasmid pJRS233 (46).A chimeric fragment consisting of the ermC gene, expressed from its native promoter, preceding with a 1,013-bp fragment of D39 piuB upstream region (including the first 213 bp of piuB open reading frame [ORF]) and proceeding with an 818-bp fragment with piuA downstream area (including the last 534 bp of piuA ORF), was cloned into pAF103 by Gibson cloning using the GeneArt Seamless Cloning Kit (Invitrogen) and transformed into One Shot TOP10 Escherichia coli strain.A linear fragment (1,252 bp) containing ermC flanked with S. pneumoniae chromosomal fragments was amplified from pAF103 and introduced into competent D39 WT cells.Pneumococcal clones harboring the ∆piuBCDA mutation were selected on erythromycin (0.5 µg/mL), and the chromosomal deletion of piuABCD genes was confirmed by PCR analysis.

Measurement of free iron concentration in pneumococcal growth media
We determined the concentration of free iron in the medium samples using an optimized ferrozine (3-(2-pyridyl)-5,6-diphenyl-1,2,4-triazine-p,p′-disulfonic acid monosodium salt hydrate, Milli Sigma) based assay as described (33).Briefly, hemoglobin was incubated in 2-mL THYB or in spent THYB (collected from overnight pneumococcal cultures) in 12-well flat bottom microtiter plates, at 37°C.Hemoglobin-free samples collected at different time points were treated with 50-mM ascorbic acid and incubated with the 50-mg ferrozine/mL and 500-mM potassium acetate (pH 5.5) in 96-well flat bottom plates for 2.5 hours prior to reading.Absorbance at 562 nm was determined after 135 minutes of incubation at 37°C, and iron concentration in the medium was calculated using a standard curve.The same was done with the cell-and hemoglobin-free supernatant samples collected from pneumococcal cultures grown in THYB with hemoglobin with or without 200-U/µL catalase (Sigma-Aldrich).Net iron released into the culture during growth was calculated by subtracting the iron values determined in un-inoculated THYB containing hemoglobin.

Determination of heme content
Fresh THYB CHX or THYB NTA was supplemented with 20-µM hemoglobin, inoculated with S. pneumoniae OD 600 0.05 in 12-well flat bottom microtiter plates and incubated at 37°C.Six-milliliter culture samples (standardized to OD 600 1.0) were collected at different time points.The cells were harvested by centrifugation, washed three times with PBS, resuspended in 2-mL DMSO, and sonicated (20% amplitude for 30 seconds).Heme was extracted by acidified chloroform, and its concentration was determined using a standard curve made with hemin solutions in DMSO as described (34).Briefly, 2 mL of 50-mM glycine buffer, pH 2.0, 0.1 mL of 4 N HCl (pH 2.0), 0.2 mL of 5-M NaCl (pH 2.0), and 2 mL of chloroform:isopropanol were added to the cell lysates.The reactions were mixed vigorously and were allowed to incubate at room temperature for 1 minute.The absorbance of the organic phase at 388, 450, and 330 nm was recorded and fed into the correction equation Ac = 2 × A388 − (A450 + A330).Hemoglobin (if it was present in the culture medium) was removed by filtration before heme extractions.

Total iron by ICP-MS
Fresh THYB CHX or THYB NTA supplemented with 20-µM hemoglobin was inoculated with S. pneumoniae grown on BAPs (starting OD 600 0.05) and allowed the culture to grow in 12-well microtiter plates at 37°C for 6 hours.Six-milliliter culture samples (standardized to OD 600 1) were harvested, washed three times with phosphatebuffered saline, and sent for ICP analysis (Center for Applied Isotope Studies, University of Georgia, Athens, GA) as described (9).

qRT-PCR analysis
Quantitative RT-PCR (qRT-PCR) analysis was carried out using the Power SYBR Green RNA-to-Ct 1-Step Kit (Applied Biosystems) and StepOne DNA PCR machine (Applied Biosystems) according to the manufacturer's specifications.A 25-ng RNA was used per qRT-PCR reaction, and each reaction was done in duplicates.Primers used for qRT-PCR are listed in Table 2.The relative expression was normalized to the endogenous control gyrB gene, and fold changes were calculated using the comparative 2 −ΔΔCT method.

FIG 1
FIG 1 H 2 O 2 and met-hemoglobin consumption during pneumococcal growth on met-hemoglobin iron.Shown are H 2 O 2 concentrations in the culture supernatant of D39 WT, ∆spxB∆lctO mutant, and complemented (ΩspxB-lctO) strains grown in THYB, (A) or (B) in THYB supplemented with 20-µM met-hemoglobin (Hb).The insert contains the absorbance spectra of 10-µM hemoglobin stock solutions in PBS (Hb), PBS with 30-µM potassium ferricyanide (Hb FeCN) or 10-mM dithiothreitol (Hb DTT).(C) The absorbance spectra of un-inoculated THYB containing 10-µM Hb, THYB 10-µM Hb inoculated with the D39 WT or the ∆spxB∆lctO mutant incubated at 37°C for up to 18 hours (H), spent THYB medium used as blank.Cells were removed by centrifugation prior to absorbance reading.The data were derived from experiments done in duplicates and repeated twice and analyzed by analysis of variance, where * indicates P ≤ 0.05 and ** indicates P ≤ 0.01.

FIG 2
FIG 2 H 2 O 2 produced by S. pneumoniae acts to liberate iron from met-hemoglobin in a cell-dependent manner.(A) Un-inoculated THYB supplemented with 20-µM met-hemoglobin or medium inoculated with isogenic D39 WT, ∆spxB∆lctO, and the complemented strain (ΩspxB-lctO) were incubated at 37°C.WT pneumococci were used to inoculate also THYB with 20-µM met-hemoglobin containing 200-U/µL catalase (WT catalase).Free iron levels in media samples collected at different time points were determined after removing the cells by centrifugation and the hemoglobin by filtration.Insert shows iron concentrations in un-inoculated THYB incubated with 20-µM met-hemoglobin.Net iron released into the culture media was calculated by subtracting the iron concentration found in un-inoculated medium at the same time point.(B)

FIG 3 S
FIG 3 S. pneumoniae growth using met-hemoglobin as the sources of iron.Cells of the isogenic strains D39 WT (blue), ∆piuBCDA (orange), and ∆spxB∆lctO (gray) were used to inoculate fresh media in microtiter plates at OD 600 0.05.The cultures were incubated at 37°C, and the optical density was recorded.S. pneumoniae was cultivated in regular THYB or iron-depleted THYB supplemented with met-hemoglobin (Hb) (full symbols) or in iron-depleted THYB (empty symbols).Iron was depleted from the medium by treating it with Chelex-100 (THYB CHX ) or by the addition of 3 mM of the iron chelator, NTA (THYB NTA ).Shown is pneumococcal growth in THYB and THYB CHX (A), THYB and THYB NTA (B), and THYB CHX and THYB CHX with (C) 5-or (D) 20-µM met-Hb and THYB NTA and THYB NTA with (E) 5-or (F) 20-µM met-Hb.Experiments were done in triplicates and repeated at least two more times.Each curve shown is derived from an average of three bioreplicates from a representative experiment.

FIG 4
FIG4 Heme and iron cellular content in pneumococci grown using met-hemoglobin as the source of iron.Cells of the isogenic strains D39 WT (blue), ∆piuBCDA (orange), and ∆spxB∆lctO (gray) were allowed to grow in THYB CHX or THYB NTA supplemented with 20-µM met-hemoglobin (as described in Fig.3).Cells harvested at different time points were washed, and the heme contants were determined by the acidified chloroform method.Shown are the cellular heme contents in pneumococci grown in met-hemoglobin supplemented THYB CHX (A) or THYB NTA (B).Intracellular iron content measured by ICP-MS in cell samples collected at the 6-hour time point is shown in C. The data (normalized to optical density) are derived from experiments done in duplicates and repeated twice.The Student t-test was used to determine statistical significance, where * indicates P ≤ 0.05 and ** indicates P ≤ 0.01; NS indicates not significant.

FIG 7
FIG 7The expression of iron uptake genes is induced in cells grown using met-hemoglobin as the source of iron in THYB NTA compared with cells cultivated in THYB CHX .Cells grown in THYB CHX or THYB NTA supplemented with 5-µM hemoglobin were harvested at the mid-log phase (OD 600 ~0.4), RNA was prepared, and gene expression was determined by qRT-PCR.Data derived from two bioreplicates processed each in duplicates are shown in a bar graph and a table format.Values smaller than one indicate downregulation.

FIG 8 A
FIG8 A proposed model for pneumococcal growth on met-hemoglobin as an iron source.The left side represents aerobic environments where the SpxB and LctO are producing H 2 O 2 that escapes to the extracellular milieu.We propose that met-hemoglobin (Hb 3+ ) is captured on the surface and is either reduced (to Hb +2 ) by surface reductases or heme is released by receptor-mediated processes.Subsequently, reduced hemoglobin and liberated heme are vulnerable to H 2 O 2 attack resulting in the release of iron that is imported by iron-dedicated transporters.Extracellular met-hemoglobin cells form hemogloblin precipitants (hemichrome) following reactions with H 2 O 2 .The right side represents anoxic environments where pneumococcus does not produce H 2 O 2 .Here, pneumococcus binds met-hemoglobin and imports heme as a primary iron source via heme transporters.Image was generated by Biorender.com.

TABLE 1
Strains and plasmids

of H 2 O 2 concentration in pneumococcal growth media
Pneumococcal cells collected from BAP following overnight growth at 37°C were used to inoculate fresh THYB or THYB supplemented with 20-µM met-hemoglobin at OD 600 0.05 (6 mL in 15-mL falcon tubes).The supernatant was prepared from culture samples by centrifugation and filtration (0.45 micron), and the hemoglobin was removed by filtration with Ultra-15 centricon filters (molecular weight cutoff of 30,000, Amicon).The sample H 2 O 2 content was measured using the Quantitative Peroxide Assay Kit (Thermo Fisher) per the manufacturer's instructions.A serial dilution of 30% H 2 O 2 in THYB was used to generate a standard curve, from which we derived the H 2 O 2 concentration in media samples.

TABLE 2
Primers used in this study