Balb/c mice were supplied by the Laboratory Animal Unit of Sun Yat-sen University. All animal experiments were approved by the Committee on the Use of Live Animals in Teaching and Research of Zunyi University and conformed to institutional and governmental guidelines and regulations.

All S. pneumoniae strains used in this study were derivatives of the parental S. pneumoniae D39 strain. These S. pneumoniae strains were cultured in Todd-Hewitt broth (Oxoid, United Kingdom) containing 0.5% yeast extract (THY) or grown on Columbia agar (Difco, United States) containing 5% sheep blood (Ruite, China) at 37^∘C in a 5% CO₂ incubator. Escherichia coli DH5α was grown in LB medium (1% tryptone, 0.5% yeast extract, 1% NaCl) at 37^∘C with shaking at 200 rpm. When required, growth media for S. pneumoniae were supplemented with tetracycline (Tet, 3.5 μg/mL) and chloramphenicol (Cm, 4 μg/mL), and media for E. coli were supplemented with ampicillin (Amp, 100 μg/mL) and chloramphenicol (Cm, 20 μg/mL). The iron-restricted medium was prepared by adding 5% Chelex-100 (Bio-Rad, United States) to THY for 8 h under continuous stirring, then filtering sterilization to remove the Chelex-100 and supplementing with 100 μM CaCl₂ and 2 mM MgCl₂. When necessary, 20 μM FeCl₃, hemin or ferrichrome (Fch) was added to the iron-restricted medium.

Comparative analysis between the amino acid sequence of S. pneumoniae SPD_1609 protein without the signal peptide (35−355 AA) with multiple protein sequences of (iron) ABC transporter substrate-binding proteins from Streptococcus pseudopneumoniae, Streptococcus mitis, Streptococcus infantis, Streptococcus suis, Clostridium cadaveris, Abiotrophia defectiva, Granulicatella adiacens, Hungatella hathewayi, Roseburia intestinalis, Bacillus cereus, Paenibacillus sp. Y412MC10, Bacillus sp. FJAT-14515, Trueperella pyogenes, and Lactobacillus heilongjiangensis and the PitA protein of S. pneumoniae TIGR4 were performed. The sequence similarity searches of SPD_1609 were initiated by a Position-Specific Iterated BLAST (PSI-BLAST) search and analyzed by Clustal X2.1. To reflect exactly the similarity of these proteins, the alignment was performed only with the mature part of the lipoprotein, without the signal peptide, and the signal peptide was predicted using Signal P 4.1 soft.

A long flanking homology-polymerase chain reaction (LFH-PCR) process was used to generate the spd-1609- mutant strain (1609- mutant) (Wach, 1996). Briefly, the 500 bp region upstream of spd-1609 was amplified using primers spd-1609–P1 and spd-1609–P3, while the 500 bp region downstream of spd-1609 was amplified using primers spd-1609–P2 and spd-1609–P4. The tetracycline gene was amplified using primers tet-F and tet-R. The three PCR fragments generated were joined together by overlap extension PCR using primers spd-1609–P1 and spd-1609–P2 to form an approximately 2.5-kb final linear DNA construct containing the deletion fragment. Then, linear DNA construction was used for homologous recombination and transformed into S. pneumoniae D39. Transformants were selected with agar plates containing 3.5 μg/mL tetracycline, and mutants were confirmed by PCR and Western blotting. All mutations were stable after six sequential passages in 0.5% THY without antibiotic selection. The primers used for the construction of the mutants are listed in Supplementary Table S1.

To confirm the link between the 1609- mutant strain and the phenotype exhibited by the S. pneumoniae 1609- mutant, a complement strain (1609 complement) that contains a recombinant shuttle plasmid pIB169 (Biswas et al., 2008) with the coding sequence of the full-length spd-1609 gene in the isogenic 1609- mutant background was used. The full spd-1609 gene, including a C-terminal 6 × His-tag, was amplified with primers pIB169-1609-F and pIB169-1609-R by PCR from the D39 genomic DNA, digested with Sac II and Kpn I and then ligated into the digested vector to generate the complementation plasmid, pIB169-1609. The recombinant plasmid was screened by using 20 μg/mL Cm in E. coli and 4 μg/mL Cm in S. pneumoniae and confirmed by PCR, DNA sequencing and Western blotting.

Streptococcus pneumoniae D39 wild type, 1609- mutant and 1609 complement strains were inoculated into normal THY medium or iron-restricted Chelex-THY medium at equal inoculation at 37 ^∘C with 5% CO₂, OD₆₀₀ was monitored every 2 h for 12 h. All data was conducted using GraphPad Prism 5.0.

The spd_1609 gene encoding the SPD_1609 protein without a signal peptide in the S. pneumoniae D39 strain was amplified by PCR and ligated into the prokaryotic expression vector pGEX-4T-1 to generate the recombinant plasmid pGEX-4T-1609. The fusion protein GST-1609 was induced by IPTG expression in Escherichia coli BL21 (DE3) and purified by GST affinity chromatography. After the GST-tag was cleaved by thrombin, the SPD_1609 protein was obtained using GST affinity chromatography. Balb/c mice were immunized with the purified SPD_1609 protein without a GST tag to generate SPD_1609 polyclonal antibodies according to the previous study (Zhang et al., 2015). For Western blotting analysis, bacteria were collected when the OD₆₀₀ reached 0.5 and lysed in SDS lysis buffer with sonication. After separation on a 12% SDS-polyacrylamide gel (SDS-PAGE), protein samples were transferred onto polyvinylidene difluoride (PVDF) membranes (Bio-Rad, United States). The membranes were then blocked with 5% (w/v) skim milk and probed with the mouse primary antibody SPD_1609. Horseradish peroxidase (HRP)-conjugated anti-mouse IgG (Promega, United States) was used as the secondary antibody, and the results were visualized using Clarity Western ECL Substrate (Bio-Rad, United States). SDS-PAGE of total proteins was used as the loading control for the Western blotting experiments.

Streptococcus pneumoniae D39 wild type, 1609- mutant and 1609 complement strains were grown in normal THY medium or iron-restricted Chelex-THY medium, respectively. The logarithmic phase cultures were centrifuged and washed three times with prechilled phosphate-buffered saline (PBS) that had been treated with Chelex-100. The cell pellets were dried in a Scanvac Freeze Dryer (Labgene Scientific, Switzerland), and the dry weights were measured. The dried cells were digested in neat trace metal-grade nitric acid for 20 min at 95^∘C. The acid solution was diluted with ultrapure water and centrifuged at 13,200 g for 30 min. The supernatants were collected and submitted for ICP-MS analysis. The results were expressed as ng of Fe per mg dry weight of cells (Jacobsen et al., 2011; Waterman et al., 2012). Three independent biological experiments were repeated.

A549 human alveolar epithelial cell lines (ATCC: CCL-185) were maintained in DMEM medium (Life Tech Technologies, United States) supplemented with 10% FBS (Gibco, United States) and incubated at 37 ^∘C in 5% CO₂. Cells were transferred to 24-well plates and cultivated to confluent cell layers (∼2 × 10⁵ cells/well) for adherence and invasion assays. Bacterial adherence and invasion assays were performed essentially as previously described with minor modifications (Quin et al., 2007; Johnson et al., 2015). Briefly, S. pneumoniae wild type, 1609- mutant and 1609 complement strains were grown to an optical density (OD₆₀₀) of approximately 0.3 and then diluted to 2 × 10⁷ CFU/mL with phenol red-free 1640 medium containing 1% FBS before adding to the monolayer at a multiplicity of infection (MOI) = 100 bacteria/cell and incubated for 1 h (adherence assays) or 2 h (invasion assays).

For adherence assays, the unbound bacteria were removed by washing with 1 × PBS three times, and the number of adhering bacteria was counted by lysing the A549 cells with 0.025% Triton X-100 and plating the lysate. For invasion assays, the monolayers were initially treated as the process for the adherence assay, but after removing the medium and washing with 1 × PBS three times, 1 mL of phenol red-free 1640 medium with 1% FBS containing 100 μg/mL gentamicin was added to the monolayers to kill bacteria that attached to the surfaces of the A549 cells. The plates were incubated for 1 h at 37^∘C in 5% CO₂. After this step, the monolayers were washed three times with 1 × PBS, and the number of invading bacteria was counted by lysing the A549 cells with 0.025% Triton X-100 and plating the lysate.

All adherence and invasion experiments were performed in triplicate and repeated three times, and the adherence and invasion abilities were expressed by determining the number of adherent or invasive bacteria per number of host cells.

Six-week-old female Balb/c mice were used for animal infection experiments. For the bacteremia infection model, mice were challenged with 5 × 10⁶ CFU of bacteria by intravenous (i.v.) injection through the tail vein and monitored for mortality over the next 14 days. Blood samples were collected via tail venous puncture at 24 h postinfection and properly diluted and plated in replicates on Columbia blood agar plates.

When wild-type S. pneumoniae, 1609- mutant and 1609 complement strains were grown to logarithmic growth phase (OD₆₀₀ = 0.3), total RNA from each strain was extracted with TRIzol reagent (Invitrogen, United States) following the manufacturer’s protocol. The purity and concentration of the isolated RNA were determined using a NanoDrop 2000 UV-VIS Spectrophotometer (Thermo Scientific, United States). cDNA was generated from 1 μg RNA without DNA contamination using the TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix Kit (TransGen Biotech, China) according to the kit’s specifications. RT-qPCR was carried out using EvaGreen Dye (Bio-Rad, United States) in a Miniopticon Real-Time PCR System (Bio-Rad, United States). The cycle threshold (Ct) value was recorded, and the relative quantification of specific gene expression was calculated using the 2^–ΔΔCt method (Livak and Schmittgen, 2001), with gyrB as an internal control. The results are shown as the 1609- mutant or 1609 complement against the wild-type (WT) strain. The primer sequences are shown in Supplementary Table S1. All data were evaluated with three independent biological experiments.

To characterize histopathology following bacteremia infection, lung biopsy samples were collected from mice infected with S. pneumoniae WT, 1609- mutant and 1609 complement strains. Mice were sacrificed at 48 h postinfection, and samples were collected and fixed in 4% paraformaldehyde. Fixed lung samples were embedded in paraffin according to procedures used for routine histology. Stained samples were examined using an Olympus microscope (Olympus, Japan).

Data from the ICP-MS, RT-qPCR, adhesion and invasion assays were analyzed using two-tailed unpaired Student’s t-test and expressed as the mean ± SEM. Data on the survival of mice for the virulence experiment were analyzed using the log-rank (Mantel-Cox) test, and the numbers of CFUs in the different experimental groups were compared using the Mann-Whitney test. Statistical analysis was carried out using GraphPad Prism 5.0, and significant differences were considered when p < 0.05.

To discover whether the SPD_1609 protein is conserved in gram-positive bacteria, an alignment was performed using (iron) ABC transporter substrate-binding proteins from S. pseudopneumoniae, S. mitis, S. infantis, S. suis, C. cadaveris, A. defective, G. adiacens, H. hathewayi, R. intestinalis, B. cereus, Paenibacillus sp., Bacillus sp., T. pyogenes, and L. heilongjiangensis with PSI-BLAST search and Clustal X2.1 soft. The alignment indicated that the sequences are homologous with more than 56% similarities (Table 1). As shown in Figure 1, 14 amino acids in the SPD_1609 protein are invariant residues in all included homologous proteins, with two glutamic acid (E), a tyrosine (Y) and an aspartic acid (D) being the iron-coordination residues in all well-characterized iron binding proteins (Sun et al., 2009; Cheng et al., 2013). In addition, SPD_1609 has no significant similarity with PiaA or PiuA but shares 23% identity and 40% positive with PitA when aligned with the BLAST 2 sequence program (Table 1 and Figure 1).

To explore the function of the SPD_1609 protein in S. pneumoniae, we constructed a 1609- mutant strain and a 1609 complement strain. The 1609- mutant and complement strains were confirmed by PCR and Western blotting (Figure 2 and Supplementary Figure S1). First, we detected the growth curves of the WT, 1609- mutant and 1609 complement strains under iron-abundant (THY) and iron-depleted conditions (Chelex-THY). As shown in Figure 3A, under the iron abundant (THY) condition, the growth curves showed almost no difference among the WT, 1609- mutant and 1609 complement strains. However, under iron-depleted conditions (Chelex-THY), compared with WT, the 1609- mutant strain grew slowly and had a lower maximal OD value, while the 1609 complement strain showed similar growth curves to WT. Then, we tested the growth of the 1609- mutant strain in normal THY media, Chelex-100 treated iron-restricted media (Chelex-THY), and iron-restricted media with the addition of 20 μM FeCl₃, hemin or ferrichrome (Fch). The growth of the 1609- mutant strain in iron-restricted media was reduced compared with that in normal THY media, and the addition of hemin could restore growth to a similar normal level; however, the addition of FeCl₃ or Fch did not restore growth (Figure 3B). These results indicated that lipoprotein SPD_1609 may be involved in FeCl₃ or ferrichrome uptake but not hemin uptake.

Although our previous study reported that the piaA/piuA/1609 triple mutant resulted in impaired iron acquisition compared to the piaA/piuA double mutant (Yang et al., 2016), the effect of the 1609- single mutant on iron uptake is unknown. Therefore, the intracellular levels of iron among the WT, 1609- mutant and 1609 complement strains were detected using ICP-MS. In normal THY media, the iron contents were not significantly different among the WT, 1609- mutant and 1609 complement strains (Figure 3C). However, in the iron-depleted media, the iron content in the 1609- mutant strain was obviously lower than that in the WT and 1609 complement strains (Figure 3C).

In addition to pneumolysin and the phagocytosis-inhibiting polysaccharide capsule, the virulence of S. pneumoniae is enhanced by the capacity of bacteria to adhere to and invade host cells and then diffusion into host tissue (Andrews et al., 2003; Zakrzewicz et al., 2016). Many surface proteins of S. pneumoniae have been shown to be involved in adherence and invasion processes (Balachandran et al., 2002; Uchiyama et al., 2009; Keller et al., 2013). Considering that SPD_1609 is a surface lipoprotein, we investigated whether SPD_1609 could affect pneumococcal adherence and invasion to human alveolar epithelial cells (A549) in vitro, and piuA- mutant as a positive control. As shown in Figure 4, the 1609- mutant strain exhibited an almost 60% decrease in A549 adherence and a 70% decrease in A549 invasion compared with the WT parent strain, and the defect was rescued upon complementation of the mutant with spd-1609 on a plasmid vector: 1609 complement strain. This result indicates that SPD_1609 is important for S. pneumoniae colonization of human lung carcinoma cells.

To further explore the reason that the 1609- mutation reduced the adherence and invasion of the bacterium to host A549 cells, the gene expression of a range of surface proteins including adhesins choline binding protein A (pcpA and cbpA), neuraminidase A (nanA) and piuA involved in colonization (Sanchez-Beato et al., 1998; Kadioglu et al., 2008) was detected using RT-qPCR. The RT-qPCR results indicated that of the pcpA and cbpA and nanA genes were downregulated in the 1609- mutant strain compared with the WT parent strain (Figure 5A), consistent with the observation that the spd-1609 deletion reduced bacterial adherence and invasion to host A549 cells. However, piuA was induced in the 1609- mutant strain compared to WT; one possible reason for the high expression of piuA in the 1609- mutant strain could be to compensate for iron uptake. In contrast, these genes showed no significant change between the WT and 1609 complement strains (Figure 5B).

We also dissected the importance of the SPD_1609 protein in vivo, and a mouse model of bacteremia infection was employed. For bacteremia infection, a group of six mice was inoculated i.v. with 5 × 10⁶ CFU of the WT, the 1609- mutant or the 1609 complement strain through the tail vein and monitored for mortality over a 14-day period. Mice infected with the 1609- mutant had significantly lower bacterial CFU in the blood at 24 h following infection than the mice infected with the WT or the 1609 complement strains (Figure 6A). Moreover, the mice infected with the 1609- mutant displayed significantly higher survival rates compared with the mice infected with the WT strain (Figure 6B). When the 1609- mutant strain was complemented with the spd_1609 gene, no significant difference was observed in survival (Figure 6B) or bacterial blood titers (Figure 6A) compared to the results for infection with the WT strain.

To further observe the effect of the SPD_1609 protein on bacteremia infection, histology analysis was performed. The lung biopsy samples were collected from the mice uninfected with bacteria (control), and the mice infected with wild-type S. pneumoniae, 1609- mutant and 1609 complement strains after 48 h of infection. Lung histological examination revealed that mice infected with WT, 1609- mutant or 1609 complement strains displayed alveolar disruption, with more inflammatory cell infiltration than control (Figures 6C–F). However, less alveolar damage and inflammatory cell infiltration was observed in mice infected with the 1609- mutant when compared with mice infected with the WT or 1609 complement strains (Figures 6D–F).

These results highlight the importance of the SPD_1609 protein for virulence in a mouse model of bacteremia infection, indicating that the SPD_1609 lipoprotein is important for bacteremia infection during in vivo growth.