The protein sequence of SPD_0090 from S. pneumoniae D39 was used as a seed to query the UniProt database using the Protein BLAST tool. Then, multiple sequence alignment and cluster analysis were performed with high-scoring proteins using the software package Clustal-X 2.1 (Zheng et al., 2021).

S. pneumoniae D39 wild type (WT) was cultured in THY medium consisting of Todd-Hewitt broth (Oxoid, United Kingdom) supplemented with 0.5% yeast extract (Oxoid, United Kingdom) or grown on Columbia agar (Difco, United States) containing 5% sheep blood (Ruite, China) in an incubator containing 5% CO₂ at 37°C. Escherichia coli BL21 was cultured in Luria-Bertani (LB) medium in a shaking incubator at 37°C. The transformed strains were selected for growth in the presence of 4 μg/ml of chloramphenicol (Cm; Sigma, United States), or 100 μg/ml of ampicillin (Amp; Sigma, United States) were added to the medium.

The iron-restricted medium was prepared by adding 5% Chelex-100 resin (Bio-Rad, United States) to THY with continuous stirring for 8 h, then filtered to remove the resin, sterilized, and supplemented with 100 μM of CaCl₂ and 2 mM of MgCl₂. If necessary, an iron source, such as hemin or FeCl₃, was added to the medium. All the bacterial strains and plasmids used in this study are listed in Table 1.

D39Δspd0090 was constructed using the flanking homology polymerase chain reaction (LFH-PCR) (Wach, 1996) method, replacing the target gene spd-0090 with the antibiotic resistance cassette gene (Tet, tetracycline, Sigma, United States). The LFH-PCR containing the upstream (705 bp) and downstream (652 bp) sequences of spd-0090 and the Tet gene (1,501 bp) products were introduced into S. pneumoniae D39 and incubated at 37°C for 2 h. The primers used for this work are listed in Table 2. The transformants were transferred to Columbia sheep blood agar plates containing Tet (3.5 μg/ml), and positive strains were selected in which the spd_0090 gene had been completely replaced by the tet gene. D39Δspd0090 strain was stabilized after 7–8 consecutive passages in THY medium containing 3.5 μg/ml Tet. To construct the spd-0090-complement mutant strain, the spd-0090 gene was inserted into the pIB169 plasmid using the ClonExpress II One Step Cloning Kit (Vazyme, China) method (Miao et al., 2018). The recombinant plasmid pIB169-spd-0090 containing a Cm resistance cassette was transformed into strain D39Δspd0090. The plasmid pIB169 was also transformed into strain D39Δspd0090 to construct as vector control. Transformants were screened with a Columbia sheep blood agar plate containing 4 μg/ml Cm. All mutant strains were further confirmed by Western blot.

PsaA, PiuA, and PiaA antibodies were derived from the previous work (Yang et al., 2016). The purified SPD_0090 protein was used as an antigen for the immunization experiments to generate multiclonal antibodies in mice according to the previous report (Brown et al., 2001b). Equal amounts of proteins were analyzed by 10–12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, United States), which were incubated with the corresponding antibodies at 4°C. This was followed by incubation with horseradish peroxidase (HRP)-conjugated goat anti-mouse as a secondary antibody. Results were visualized with Clarity Western ECL substrate (Bio-Rad, United States) and captured using Image Master 2D Platinum 6.0 (GE Healthcare, United States).

S. pneumoniae D39, D39Δspd0090, and D39Δspd0090⁺ strains at OD₆₀₀ ∼ 0.6 in THY medium were collected by centrifugation and treated by 20 mg/ml lysozyme at 37°C for 30 min. The extraction of RNA and the synthesis of cDNA follow the previous methods (Dong et al., 2021).

Real-time quantitative PCR (RT-qPCR) was performed using the EvaGreen Dye (Bio-Rad, United States) with the TransStart Tip Green qPCR supermix kit (TransGen Biotech, China). The cycle threshold (CT) value was recorded, and the relative quantification of gene expression was calculated using the 2^–ΔΔCT method, with 16S rRNA as an internal control. The results were compared to gene expression in the D39Δspd0090 or D39Δspd0090⁺ with gene expression in the D39 strain. All data were obtained from three independent biological experiments. The primers used for RT-qPCR are shown in Table 2.

The spd_0090 gene, encoding the SPD_0090 protein without a signal peptide, was cloned from S. pneumoniae D39 by PCR, and the pGEX-4T-1 vector was digested with SalI and NotI (TaKaRa, Japan) and ligated with T4 DNA ligase (TaKaRa, Japan). The fused vector pGEX-4T-1-spd-0090 was confirmed by DNA sequencing (Tsingke, China) and then transformed to Escherichia coli BL21 followed by culturing in an LB medium containing 100 μg/ml Amp at 37°C in a shaking incubator. The expression and purification of SPD_0090 protein were performed according to the previous method (Zheng et al., 2021).

S. pneumoniae D39 was cultured in THY to OD₆₀₀ ∼ 0.6, collected by centrifugation, and washed three times with phosphate-buffered saline (PBS). The collected cells were resuspended in 800 μl PBS with 1 mM phenylmethylsulfonyl fluoride (PMSF) and 0.5 mg/ml lysozyme, digested at 37°C for 30 min, then centrifuged at 3,000 × g for 5 min, and the pellet was discarded. A final concentration of 2% Triton X-114 was added to the supernatant and incubated at 4°C for 1 h (Bordier, 1981; Khandavilli et al., 2008). The sample was then transferred to a water bath at 37°C and incubated for 10 min to induce phase separation. The membrane proteins were separated by centrifugation at 10,000 × g for 10 min at room temperature, and the upper aqueous phase and the lower membrane phase were used separately for SDS-PAGE.

The purified SPD_0090 (50 μM) was incubated with an excess of hemin (200 μM) for 2 h at room temperature and passed through a desalination column and a 10-kDa ultrafiltration column (Millipore, United States) (Cao et al., 2018). Then, the sample was washed six times with 20 mM Tris-HCl (pH 7.4, containing 100 mM NaCl) in a 10-kDa ultrafiltration column by centrifugation to remove the hemin unbound to SPD_0090, and the remaining sample was the hemin-saturated SPD_0090 protein, named SPD_0090 + hemin. The hemin solutions were prepared in 20 mM Tris-HCl containing 100 mM NaCl (pH 7.4) at a concentration of 25 μM, and 500 μl of SPD_0090 and SPD_0090 + hemin solutions were scanned from 200 to 600 nm by UV spectrophotometer (Evolution 300, Thermo Fisher Scientific, United States), and each experiment was repeated three times.

The binding of SPD_0090 to hemin was detected by a fluorescence spectrometer (F7000, Hitachi, Japan). A concentration of 2 μM SPD_0090 was prepared with 20 mM Tris-HCl (pH 7.4, containing 100 mM NaCl) buffer, and a high concentration of 1.2 mM hemin solution was prepared with the same buffer. The parameter settings are the same as earlier (Miao et al., 2018). The fluorescence spectra of SPD_0090 from 300 to 450 nm were recorded after that 1.2 mM hemin was gradually added to 1.5 ml of 2 μM SPD_0090 until the fluorescence intensity stabilized. Each scan was repeated three times and the titration curves were analyzed in Origin version 9.0, and the binding affinity (Ka) was calculated using Hill plots.

A buffer of 20 mM Tris-HCl (pH 7.4, containing 100 mM NaCl) was used to prepare 400 μl each of 100 μM hemin and SPD_0090 + hemin solution (SPD_0090: hemin = 1:1). The electron paramagnetic resonance spectra of hemin and SPD_0090 + hemin solution were measured by a Bruker spectrometer (Elexsys E580, Bruker, Germany) with the following parameters: frequency, 9.4 GHz; power, 4 mW; modulation amplitude, 3 G; modulation frequency, 100 kHz; and temperature, 10 K. The obtained data were processed and plotted using the Origin version 9.0 software.

The A549 human alveolar epithelial cell line (ATCC: CCL-185) was cultured in Dulbecco’s modified eagle medium (DMEM) (Life Tech Technologies, United States), supplemented with 10% fetal bovine serum (FBS, Gibco, United States), and incubated at 37°C in 5% CO₂. Cells were transferred to 12-well plates and cultured to uniformly full cell layers (2 × 10⁵ cells/well) for adhesion and invasion assays. Bacterial adhesion and invasion assays were performed essentially, as previously described, with minor modifications (Quin et al., 2007; Yang et al., 2019). Briefly, D39 and D39Δspd0090 strains were cultured to OD₆₀₀ ∼ 0.6 and then diluted to 1 × 10⁷ CFU/ml. The collected bacteria were resuspended in a DMEM containing 1% FBS and added to cell plates and incubated in a CO₂ incubator for 2 h.

For the adhesion assay, A549 cells were washed three times with sterile PBS to wash away unadhered bacteria. Cells were lysed by adding 1% Triton X-100 for 30 min on ice, diluted in a gradient, and spread on Columbia agar plates containing 5% sheep blood.

For the invasion assay, DMEM contained 100 μg/ml of gentamicin (Gen, Sigma, United States) and 1% FBS was added and incubated for 30 min at 37°C, followed by three washes with sterile PBS. Cells were lysed by adding 1% Triton X-100 for 30 min on ice, diluted from 10¹ to 10⁵ times in a gradient, and spread on Columbia agar plates containing 5% sheep blood. Three independent biological experiments were repeated.

All animals used in the experiments were purchased from the Experimental Animal Management Center of Southern Medical University (Guangdong, China), Female BALB/c mice between 6 and 8 weeks of age were used in our experiments. For animal infection experiments, the D39 and D39Δspd0090 strains were grown in THY to logarithmic growth phase, collected, and washed three times with PBS. The bacterial solution (2.5 × 10⁸ CFU/20 μl) was dropped into the nose of the mice. The infected mice were observed for survival transitions; nasal lavage and lung tissue were collected at 12 or 24 h. The number of viable bacteria was determined by serial dilutions applied to a Columbia medium containing 5% sheep blood. Survival times were calculated using the GraphPad Prism software and analyzed by the log-rank (Mantel–Cox) test (Robb et al., 2017). The number of bacteria in the nasal lavage solution and lungs was analyzed by the Mann–Whitney test. A value of p < 0.05 was considered statistically significant.

The growth curves of the bacteria in restriction medium C + Y were determined by UV-visible spectroscopy (Evolution 300, Thermo Fisher Scientific, United States) (Miao et al., 2018). The medium was configured at room temperature and filtered through a sterile 0.22-μm filter membrane to remove bacteria. S. pneumoniae D39, D39Δspd0090, D39Δspd0090¹⁶⁹⁺, and D39Δspd0090⁺ strains were inoculated into the C + Y medium (no sugar), with galactose (10 mM), glucose (8 mM), and lactose (8 mM) at equivalent inoculation doses at 37°C in 5% CO₂ incubator. S. pneumoniae D39, D39Δspd0090, D39Δspd0090⁺, and D39Δspd0090¹⁶⁹⁺ strains were inoculated into normal THY medium, iron-restricted medium containing iron chelator 2,2′-bipyridine (1 mM), or iron-restricted culture supplemented with 25 μM of hemin. The iron-restricted THY medium was prepared by adding 1 mM 2,2-bipyridine (DP) (Sigma-Aldrich), which was filtered with a 0.22-μM filter to THY, and incubated at 37°C for 2 h. The optical density at 600 nm (OD₆₀₀) was continuously measured every 2 h by UV-visible spectroscopy) for 12 h. All data were analyzed by using GraphPad Prism version 8.0.2 (GraphPad Software, United States).

Total proteins were extracted from D39 and D39Δspd0090 in the exponential growth phase when they were cultured in the THY medium. A volume of 500 μg of total protein was taken from each sample and digested with trypsin (Promega, United States) at 37°C for 18 h and lyophilized. Then, an iTRAQ Reagent multiplex kit (AB Sciex, United States) was used to label peptide samples according to the manufacturer’s protocol. The two groups of samples were labeled separately at room temperature for 1 h and then the labeled samples were mixed in equal proportions and lyophilized (Group 1: “115” labels D39, “113” labels D39Δspd0090; Group 2: “119” labels D39, “113” labels D39Δspd0090).

The labeled samples were grouped using high-performance liquid chromatography (HPLC; WuFeng, China), referring to the previous description for the specific experimental steps (Zheng et al., 2021). Then, the samples were desalted according to the instructions for the MonoTip C18 (SHIMADZU, Japan) desalting column. Desalted peptides were resuspended with buffer A (5% acetonitrile, 0.1% formic acid) and detected using an Orbitrap Fusion Lumos mass spectrometer (Thermo, United States). Then, the raw data obtained were identified and quantified by using the Proteome Discoverer software version 2.1 (Thermo Fisher Scientific, United States) and searched using the S. pneumoniae D39 database. The specific parameters were as follows: Sample Type, iTRAQ 8plex (peptide labeled); Digestion, Trypsin; instrument, Orbitrap Fusion Lumos; Variable modifications, Oxidation (M), Acetyl (Protein N-term); search effort, thorough; and detected protein threshold [unused ProtScore (Conf)] > 1.30 (95.0%). In this study, the cutoff threshold for increased and decreased abundance of the differentially expressed proteins was defined using the population statistics applied to the biological replicates reported by Gan et al. (2007). The finalized cutoff values for increased (1.4) and decreased (0.71) protein abundance were determined as the threshold. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were performed in Glue GO plug-in of Cytoscape (version 3.9.0) with p < 0.05 to analyze the biological processes of differentially expressed proteins (Shannon et al., 2003).

According to the protein sequence of SPD_0090 as indicated by NCBI¹, the analysis of SPD_0090 localization by using PSORTb version 3.0.3² (Yu et al., 2010) and the analysis of SPD_0090 signal peptide prediction by using SignalP - 5.0³ (Armenteros et al., 2019) were performed.

The expression and purification of GST_SPD_0090 and Glutathione-S-transferase (GST) protein were carried out by the above-mentioned method (Zheng et al., 2021). GST_SPD_0090 solubilized to glutathione-Sepharose (GE Healthcare, United States) beads was incubated with D39 cell lysate in PBS buffer overnight at 4°C with the GST as the negative control. Beads were washed with IP Lysate buffer (Beyotime, China) to remove the unbound proteins. Proteins bound to the beads were collected in PBS buffer by centrifugation and heated at 100°C. The isolated proteins were separated by SDS-PAGE, stained by AgNO₃, then digested by trypsin, and identified by mass spectrometry (MS) (Zhou et al., 2021) (Orbitrap Fusion Lumos; Thermo Fisher Scientific).

Data were statistically analyzed by using GraphPad Prism version 8.0.2. The differences between the two groups were analyzed by two-tailed, unpaired Student’s t-tests, and data are expressed as mean ± standard (SD). Results were considered significant at p ≤ 0.05 (denoted by *), p ≤ 0.01 (denoted by ^**), p ≤ 0.001 (denoted by ^***), and p ≤ 0.0001 (denoted by ^****). The data on the survival of mice for the virulence experiment were analyzed using the log-rank (Mantel–Cox) test, and a number of CFUs in the different experimental groups were compared using the Mann–Whitney test.

The MS proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (Deutsch et al., 2020) partner repository with the dataset identifier PXD030183⁴ (username: reviewer_pxd030183@ebi.ac.uk; password, 5LSbJZVk).

To assess the conservation of SPD_0090 in bacteria, amino acid sequence alignment was carried out in this study (Figure 1). The result showed that the SPD_0090 homologous proteins are highly conserved in Gram-positive bacteria. Several highly homologous proteins of SPD_0090, e.g., SGO_1763, BCR26_01130, and CBF35_08275, are annotated as sugar ABC transporter substrate-binding proteins in the Uniprot database, but no experimental data studies reveal their function so far. As mentioned in the introduction section, our previous study has suggested that SPD_0090 may be associated with iron transportation in S. pneumoniae (Yang et al., 2016). In this study, we used a biochemical method combined with quantitative proteomics to investigate the potential biological function of SPD_0090 in S. pneumoniae.

To find out the detailed biological functions of SPD_0090 in S. pneumoniae, the D39Δspd0090 strain was first constructed by homologous substitution, and the complemented mutant D39Δspd0090⁺ was obtained by supplementing the plasmid pIB169-spd_0090 into the D39Δspd0090 strain; pIB169 was introduced to the D39Δspd0090 strain to obtain D39Δspd0090¹⁶⁹⁺. Western blot showed that the SPD_0090 protein was highly expressed in D39Δspd0090⁺ (Figure 2), while it was undetectable in D39Δspd0090 and D39Δspd0090¹⁶⁹⁺. These results indicated the successful construction of the gene knockout and complement strains.

We first checked whether SPD_0090 affects the growth of S. pneumoniae. The growth curves showed that D39, D39Δspd0090, D39Δspd0090¹⁶⁹⁺, and D39Δspd0090⁺ strains barely grew in the C + Y medium without any sugar source (Figure 3A). However, previous studies have speculated that SPD_0090 is a galactose ABC transporter protein (Durmort and Brown, 2015), in the presence of glucose, galactose, and lactose; the growth rate of D39Δspd0090 and D39Δspd0090¹⁶⁹⁺ was lower than that D39 (Figures 3B–D), but the growth rate of D39Δspd0090⁺ growth was close to that of D39. We counted the data of D39 and D39Δspd0090 in the logarithmic phase of growth and found that the growth of D39Δspd0090 is significantly slower than that of D39 cultured in three sugar sources even at a high concentration of sugars (Supplementary Figure 1). The above results suggest that the knockout of the spd_0090 gene resulted in a restricted growth of S. pneumoniae.

To study the relationship between SPD_0090 and iron metabolism in S. pneumoniae, we detected the SPD_0090 protein level in D39 and the triple mutant (ΔpiuA/ΔpiaA/ΔpitA) in which three genes encoding major iron transporters were simultaneously knocked out. There were significantly higher levels of SPD_0090 protein in the triple mutant, relative to D39 (Figure 4A), which is consistent with the previous iTRAQ-based quantitative proteomics results of our lab (Yang et al., 2016). Furthermore, PSORTb prediction showed that SPD_0090 is not an intracellular protein (Figure 4B). Then, we used signal prediction and found that SPD_0090 is a lipoprotein anchored in the cell membrane, which contains a cleavage site of typical lipoprotein signal peptide (between positions 23 and 24 amino acids of the SPD_0090 protein sequence) (Figure 4C). We isolated membrane proteins from S. pneumoniae D39 by using Triton X-114 and carried out a Western blot analysis (Bordier, 1981). A well-known membrane protein PiuA was used as a positive control, and the soluble protein Ply was present in the aqueous phase as a negative control. As shown in Figure 4D, SPD_0090 is present in the whole cell lysate and membrane protein fractions, but not in the upper aqueous phase, demonstrating that SPD_0090 is a membrane protein.

To further study which form of iron is associated with SPD_0090, two major proteins of PiuA (hemin transporter) and PiaA (ferrichrome transporter) were detected in D39 and D39Δspd0090 cultured in the THY medium by Western blot. There was a higher level of PiuA in D39Δspd0090, relative to D39, while there was no significant difference in PiaA protein levels between the two strains (Figure 4E). As PiuA is responsible for hemin uptake on the cell surface, we speculated that SPD_0090 may be involved in hemin transportation. We detected the growth curves of D39, D39Δspd0090, D39Δspd0090¹⁶⁹⁺, and D39Δspd0090⁺ under iron-abundant (THY) and iron-restricted conditions (1 mM 2,2′-bipyridine, DP). As shown in Figures 4F,G in the THY medium, there was almost no difference in the growth of D39 and D39Δspd0090, D39Δspd0090¹⁶⁹⁺, and D39Δspd0090⁺. In the condition of the iron restriction, the growth of D39Δspd0090 was significantly lower than that of D39 (Figure 4F). The growth of D39Δspd0090¹⁶⁹⁺ was significantly lower than that of D39Δspd0090⁺ (Figure 4G). When hemin was supplemented in the iron-restricted medium of D39 and D39Δspd0090, the growth of D39 returned close to that in the THY medium when the growth of D39Δspd0090 was only partially recovered (Figure 4F). This result indicated that the deletion of spd_0090 may affect the ability of S. pneumoniae to use hemin and thus result in diminished growth. Then, we examined whether the SPD_0090 protein is regulated by hemin in the medium. As shown in Figure 4H, SPD_0090 protein had higher levels in iron-restricted culture conditions relative to iron-replete conditions, suggesting that SPD_0090 is evidently regulated by iron concentration. Then, we supplemented hemin and FeCl₃ as an iron source to the iron-restricted medium. The Western blot analysis showed that the level of SPD_0090 protein was decreased to a quarter of the original level (in the iron-restricted medium) after the supplement of 25 μM hemin, while it remained almost unchanged upon 25 μM FeCl₃ addition (Figure 4H). These changes in SPD_0090 abundance in response to hemin are similar to those of PiuA (Figure 4H). Therefore, based on the above results, we hypothesized that SPD_0090 is a potential hemin transporter located on the cell surface.

Next, we expressed, purified SPD_0090 protein, and in vitro characterized its hemin binding properties (Figure 5A). In the UV/vis spectra of Hemin, a broad absorption peak at around 385 nm was observed which is the typical absorption peak of hemin. In the UV spectra of SPD_0090-hemin, the characteristic absorption peak of the band of hemin was shifted from 385 to 410 nm which is a typical spectrum of hemin-binding protein (Figure 5B). Subsequent fluorescence spectroscopy detection showed that hemin binding to SPD_0090 resulted in fluorescence quenching at 340 nm. The fluorescence titration curve was fitted to a Hill plot, and the binding constant (Ka) of SPD_0090 to hemin was calculated to be 4.6 × 10⁴M^–1 (Figure 5C). Then, we further used EPR to investigate the nature of hemin binding to SPD_0090. As shown in Figure 5D, the single peak signal in the high-spin (HS) region [HS, 900–1,900 G] at g = 6.0 and the low-spin (LS) region (LS, 2,300–4,300 G) at g = 2.0 is a characteristic peak of hemin (Carette et al., 2006; Lu et al., 2010). After binding with SPD_0090, the EPR feature of hemin was changed at the peaks (Figure 5D). The above results suggested that SPD_0090 can bind hemin in vitro.

Some iron transporters of pathogenic bacteria also directly or indirectly affect adhesion and invasion of the host cells (Durmort and Brown, 2015; Nguyen and Götz, 2016). Therefore, we investigated whether SPD_0090 is also important to S. pneumoniae infection in human lung cancer cells (A549). The adhesion and invasion ability of D39 and D39Δspd0090 to A549 cells were measured. As shown in Figures 6A,B, compared to D39, D39Δspd0090 showed a fourfold and a fivefold increase in the number of adherent and invasive bacteria, respectively (Figures 6A,B). This result indicated that SPD_0090 negatively contributes to S. pneumoniae infection of human lung carcinoma cells.

To fully dissect the effect of SPD_0090 on the ability of S. pneumoniae to infect the host, we used intranasally infected mice (pneumonia model) to assess bacterial virulence. After infection with D39, mice began dying after 36 h, while their survival rate at 72 h remained at 16%. The infection of the knockout strain was completely different from that of the WT. After infection with D39Δspd0090, mice began to die by 12 h and all died at 48 h. This result indicated that the mortality rate of infection with D39Δspd0090 is significantly higher than that with D39 (Figure 6C). To further test the ability of D39 and D39Δspd0090 strains to invade host tissue cells, we collected nasal lavage and lungs from mice at 12 and 24 h post-infection and plated them on the agar, followed by CFU counting. At both time points, bacterial counts of D39Δspd0090 in nasal and lung were higher than that of D39 (Figures 6D,E). This invasion resulting in the animal model is consistent with the conclusion derived from the infection of A549 cells. The combined results emphasized that SPD_0090 protein negatively regulates the virulence of S. pneumoniae which is different from the most known iron transporters.

To investigate the mechanism of the SPD_0090 effect on the virulence of S. pneumoniae, we performed iTRAQ-based proteomics to find out the biological processes changed by the spd-0090 gene knockout. In total, 1,037 proteins were identified in D39 and D39Δspd0090. The threshold for defining differential expressed proteins in this experiment was defined using the biological replicate method reported by Gan et al. (2007). The analysis of protein coefficient of variation was performed for the proteins identified in this experiment, and 88% of the proteins had coefficients of variation ≤ 40% (Figure 7A), so the biological variation threshold for this experiment was set at 40%. The threshold for fold change of differentially expressed proteins was defined as ≥1.4 for increased abundance or ≤0.71 for decreased abundance with p < 0.05 (Figure 7B). In the two biological replicates, there were 52 differentially expressed proteins in D39Δspd0090 compared to D39 (Table 3). The GO analysis indicated that 52 differentially expressed proteins were enriched in several metabolic pathways, including galactose metabolism, amino sugar and nucleotide sugar metabolism, iron transport, and other glycan degradations (Figure 7C).

During the colonization and infection process, S. pneumoniae must obtain a sugar source from the host environment to support their survival (Palmer and Skaar, 2016; Lopez and Skaar, 2018). The quantitative proteomics data showed that many proteins that participated in carbon metabolism and sugar transportation had a lower abundance in D39Δspd0090 compared to that in D39. For example, galactokinase (GalK), aldose-1-epimerase (GalM), galactose-1-phosphate (GalT), galactose-6-phosphate isomerase subunit (LacB), tagatose 1,6-diphosphate aldolase (LacD), and galactose transport [three components of phosphotransferase (PTS) system] (Table 3). This result indicated that the deletion of the spd_0090 gene reduced the expression of catabolic pathways responsible for energy production which is why the knockout stain grew slow.

Notably, 1,4-beta-N-acetylmuramidase (LytC), oligopeptide-binding protein (AliA), and amino acid-binding protein (LivJ) were present at relatively higher levels in D39Δspd0090 and known to be involved in nasopharyngeal epithelial colonization as well as S. pneumoniae invasion (Kerr et al., 2004; Basavanna et al., 2009; Corsini et al., 2021), suggesting that their elevated protein levels may be associated with the increased virulence of the knockout strain. To validate the quantitative result of proteomics and further explore the effect of SPD_0090 on the virulence of S. pneumoniae, the gene expression levels of several virulence factors were determined by RT-qPCR in D39, D39Δspd0090, and D39Δspd0090⁺. As shown in Figure 7D, the mRNA expression levels of virulence factors livJ, lytC, and aliA of D39Δspd0090 were significantly higher than that of D39 (Figure 7D), but their expression in the complemented strain was close to that in the WT (Figure 7E). This result is consistent with the abundance profile of these virulence factors. The increased expression of virulence factors explains why the spd_0090 gene knockout strain has stronger adhesion and invasion ability to A549 cells than the WT strain.

We also found that two known iron transporters PiuA and SPD_1609 were more abundant in D39Δspd0090. Previous studies have reported that PiuA is an important hemin transporter, and lipoprotein SPD_1609 may be involved in FeCl₃ or ferrichrome uptake. The alteration of these two genes at the mRNA level was validated by RT-qPCR (Figures 7D,E). The mRNA of piuA and spd_1609 in D39Δspd0090 was about three and twofold higher than that in the D39, respectively, but there was no significant change in the complement strain (Figure 7E). The upregulation of the iron transports after knockout of the spd_0090 gene could be to compensate for iron uptake. Importantly, PiuA and SPD_1609 also contribute to the virulence of S. pneumoniae^11,14 which may be another reason for the increased virulence of D39Δspd0090.

To further investigate how SPD_0090 affects S. pneumoniae virulence, we performed GST pull-down to identify proteins interacting with SPD_0090. A total of 138 interacting proteins were identified by MS after deducting the proteins in the control group. The detailed information on these proteins is listed in Supplementary Table 1. Some cytoplasmic proteins were also identified possibly due to two reasons, i.e., (1) some cytoplasmic proteins bind with proteins that interact with SPD_0090, and (2) the cytoplasmic proteins with high abundance are difficult to be fully removed from the preparation. Among them, three proteins were associated with virulence, including LytC, endopeptidase (PepO), and pneumolysin (Ply). Coincidentally, the elevated protein level of LytC was also detected in D39Δspd0090 in quantitative proteomics. We speculated that SPD_0090 interacting with LytC may inhibit LytC function and negatively contribute to bacterial virulence in D39, which requires further validation. In addition, glyceraldehyde-3-phosphate dehydrogenase (Gap), enolase (Eno), and beta-N-acetylhexosaminidase (StrH) are associated with bacterial metabolism, which is also consistent with the results identified in the proteomics implying that SPD_0090 affects carbon metabolism in S. pneumoniae.