Development of a Novel Typing Scheme Based on the Genetic Diversity of Heme/Hemin Uptake System Hmu in Klebsiella pneumoniae Species Complex

Siderophore is a group of low molecular weight compounds with high affinity for ferric iron, which could facilitate bacterial iron consumption. Similarly, hemin/heme scavenged by the hemin uptake system HmuRSTUV usually act as another critical iron source for K. pneumoniae. ABSTRACT Iron is essential for the survival and reproduction of Klebsiella pneumoniae. Although K. pneumoniae employs multiple types of siderophores to scavenge iron during infections, the majority of host iron is retained within erythrocytes and carried by hemoglobin that is inaccessible to siderophores. HmuRSTUV is a bacterial hemin/hemoprotein uptake system. However, the genetic background and function of HmuRSTUV in K. pneumoniae remain unknown. We collected 2,242 K. pneumoniae genomes, of which 2,218 (98.9%) had complete hmuRSTUV loci. Based on the 2,218 complete hmuRSTUV sequences, we established a novel typing scheme of K. pneumoniae named hmST, and 446 nonrepetitive hmSTs were identified. In hypervirulent lineages, hmST was diversely distributed and hmST1 mainly existed in ST23 strains. In contrast, hmST was less diversely distributed among multidrug-resistant strains. hmST demonstrated greater genetic diversity in hypervirulent lineages and community-acquired and bloodstream-sourced strains. In vitro and in vivo experiments revealed that an intact hmuRSTUV was essential for hemin uptake, playing an important role in bloodstream infections. This study established a novel typing scheme of hmST based on hmuRSTUV providing new insights into identifying and monitoring the emergence of novel virulence evolution in K. pneumoniae. IMPORTANCE Siderophore is a group of low molecular weight compounds with high affinity for ferric iron, which could facilitate bacterial iron consumption. Similarly, hemin/heme scavenged by the hemin uptake system HmuRSTUV usually act as another critical iron source for K. pneumoniae. This study proved that Hmu system significantly promoted the growth of K. pneumoniae in the presence of hemin and played an important role in bloodstream infections. A novel typing scheme named hmST was established, and the genetic diversity of hmuRSTUV loci was analyzed based on a large number of genomes. This study provides new insights into identifying and monitoring the emergence of novel virulence evolution in K. pneumoniae.

Generally, ferric iron has a solubility of only 10 217 M at pH 7, while bacteria require iron at around 10 27 to 10 25 M to thrive (10). Iron is much less accessible in host (human), because it is frequently bound with proteins, such as ferritin, transferrin, and myoglobin (11). To satisfy the metabolic requirement, K. pneumoniae have evolved multiple strategies to steal iron from the environment, especially from the host during infections (7). Siderophore is one of the most important iron uptake systems for K. pneumoniae used to sequester iron. Siderophores are low-molecular-weight compounds with a high affinity for Fe 31 (12). K. pneumoniae can produce four types of siderophore, including enterobactin (Ent), salmochelin (Iro), yersiniabactin (Ybt), and aerobactin (Iuc) (13). These siderophores, especially aerobactin, significantly enhance the virulence of K. pneumoniae (14,15). Consequently, the ability to acquire iron has served as an essential marker for virulence (14,16).
Nevertheless, the majority of the host iron is located within erythrocyte hemoglobin (17), which comprises four subunits, each containing an iron atom bound to a heme group. This is a great "iron pool" that siderophores could not access. The mature erythrocyte remains in circulation for about 120 days until being engulfed and digested by phagocytes mainly in the liver, bone marrow, and spleen (17). Heme is then released from the phagolysosomes and degraded by the heme oxygenase-1, freeing iron. To prevent the toxicity caused by cell-free hemoglobin (CFH), haptoglobin removes CFH from plasma by forming high-molecular-weight haptoglobin-hemoglobin complexes (17,18). Once the hemoglobin binding capacity of haptoglobin is exceeded, heme can be released into the blood directly from free hemoglobin (19). Hemin is the oxidized version of heme. K. pneumoniae can scavenge hemin/heme as an iron source via the HmuRSTUV hemin uptake system (7). Like siderophore receptors, HmuR is a TonB-dependent outer membrane receptor that is required for the utilization of heme (Fig. 1). HmuTUV constitute an inner membrane ABC transporter involving in the transportation of hemin into the cytoplasm, while HmuS is essential for the degradation of hemin and iron releasing (20). The biological function of Hmu system has been well characterized in Yersinia spp. (20,21) but was far less explored in K. pneumoniae.
Molecular typing is critical for distinguishing bacterial strains and thus provides a good strategy for surveillance, outbreak investigation, and phylogenetic analyses. Traditional typing methods, including plasmid analysis, genome enzyme restriction, and pulsed-field gel electrophoresis, have been used for a long time for typing and clonal assignment (22). These methods require agarose gel electrophoresis to separate enzyme-restricted plasmid or genome DNA fragments. Multilocus sequence typing (MLST) based on the alleles of seven housekeeping genes (gapA, infB, mdh, pgi, phoE, rpoB, and tonB) is now widely performed in K. pneumoniae and serves as the standardized scheme used for parallel comparisons of strains from different sources. With the development of high-throughput sequencing techniques, large amounts of data generated by genome sequencing can be obtained in a short time at low cost. Core genome MLST (cgMLST) has been widely used to study the molecular epidemiological characteristics of K. pneumoniae owing to its higher resolution (23). However, these typing strategies are based on global bacterial genomes rather than specific aspects. To gain a deeper understanding of the genetic diversity of siderophores, novel typing schemes (including SmST, AbST, and YbST) based on siderophore-encoding loci have been developed and updated (https://bigsdb.pasteur.fr/klebsiella/) (24,25). These novel typing methods provide an important resource to identify and monitor the emergence of undiscovered virulence evolution in K. pneumoniae. Considering that hemin is one of the critical sources of iron, we hypothesized that Hmu system might be involved in the virulence of K. pneumoniae. In this study, we aimed to (i) research the genetic diversity of hmu loci in K. pneumoniae and accordingly establish a novel typing scheme capable of tracking and tracing the virulence evolution and to (ii) demonstrate the role of Hmu system in bloodstream infections.

RESULTS
Establishment of hmST based on hmuRSTUV locus. The hmu locus is located on the chromosome and is required for the hemin uptake in K. pneumoniae (Fig. 1). The genomes examined in this study included 1,970 genomes from Pathogenwatch (https:// pathogen.watch/) and 272 genomes from our previous study, which covered 20 tertiary hospitals in China (23). All public collections of genomes available at Pathogenwatch were included in the study. The 2,242 strains of the K. pneumoniae species complex were from 34 countries. Most of these were K. pneumoniae (2,212/2,242, 98.7%) strains, while K. variicola and K. quasipneumoniae accounted for only 0.7% (16/2,242) and 0.6% (14/2,242) of the genomes, respectively. Among the 2,242 K. pneumoniae species complex genomes, 2,218 (98.9%) had intact hmuRSTUV sequences, and 5 lacked hmuRSTUV. Among the remaining 19 strains, hmuR was absent in five strains, hmuS was absent in three, hmuT was absent in five, hmuU in five, and hmuV was absent in one. Based on the 2,218 complete hmuRSTUV allelic profiles, we established a novel typing scheme named hmST. Each  Table S1 in the supplemental material). Complete allelic profiles of hmuRSTUV sequences were shown in Table S1. hmuRSTUV locus diversity in K. pneumoniae. To explore the association between hmST and common lineages, we compared the distribution of hmSTs among the most common hypervirulent (including ST23, ST86, ST65, and ST25) and MDR (including ST11, ST258, ST512, ST15, and ST147) lineages. As shown in Fig. 2, hmST8 only existed in ST11, ST258, and ST512 MDR strains, all of which belong to CG258. hmST122 and hmST46 accounted for the majority of the hmSTs in ST15 and ST147 lineages, respectively. In contrast, hmST was diversely distributed among hypervirulent lineages. For example, though hmST1 was only observed in hypervirulent ST23 strains, multiple other hmSTs, including hmST2, hmST3, hmST7, hmST168, hmST209, hmST283, hmST351, hmST432, hmST436, and hmST439, were also found in ST23 strains. Interestingly, hypervirulent and MDR lineages had no shared hmST. Genomes from our previous work were selected to build phylogenetic trees (23). As shown in Fig. 3, the phylogenetic tree of the MDR lineages ST11 and ST4496 based on MLST reflected nearly the complete phylogeny of the hmuRSTUV allelic profile. For hypervirulent ST23, the nonconcordance of the phylogeny of hmST and MLST phylogeny were observed (ST23 data are presented separately in Fig. S1), indicating that hmST might possess higher discriminatory efficacy in classifying hypervirulent lineages.
We employed accumulation curves to assess the diversity of hmSTs among different lineages (hypervirulent and MDR lineages), sources (bloodstream, liver abscess, respiratory tract, and urinary tract), infection statuses (colonization and infection), and infection types (community onset and hospital acquired). As sampling frequency increased, more hmSTs were observed, resulting in the ascending accumulation curves. The slopes of the accumulation curves reflected the diversity of hmSTs. As shown in Fig. 4A, the accumulation curve for hypervirulent lineages was far steeper than that for MDR lineages, indicating that Similarly, hmSTs showed more diversity in bloodstream-sourced, infection-isolated, and community-onset strains compared to those from other infection sites (i.e., liver abscess, urinary tract, and respiratory tract), colonization and hospital-acquired strains ( Fig. 4B to D). Role of Hmu in hemin uptake. Although the function of HmuRSTUV in Yersinia spp. has been well elucidated, research on the Hmu system in K. pneumoniae is extremely limited. It was found that the Hmu system was expressed at greater levels in entS (encoding an enterobactin exporter) mutant CRKP strains, compared to strains with intact entS (8). However, the role of the Hmu system in K. pneumoniae was not determined. To explore the role of Hmu in hemin uptake, we constructed hmuR knockouts and hmuR complementations in strains D3 and ZKP51 to compare their growth under conditions where hemin served as the unique iron source. Clinical isolates D3 and ZKP51 represented hypervirulent and MDR lineages, respectively. Knockout and complementation were performed by using a CRISPR-Cas9-mediated genome-editing method. In M9 standard medium without any supplement, all of the strains showed no growth difference (Fig. 5). However, when M9 medium was supplemented with bovine hemin, D3DhmuR and ZKP51DhmuR strains grew more slowly during exponential phase and then reached a lower plateau compared to their corresponding wild types. The complemented mutants could restore the growth FIG 3 Phylogenies inferred from MLST (left) and hmST (right). Matching strains were connected using gray auxiliary lines. Strains were colored according to main sequence type based on seven-gene MLST. The figure was generated by using the ggtree package with R software.
Novel Typing Scheme Based on Hmu Microbiology Spectrum superiority (Fig. 5). However, the growth discrepancy between hmuR mutants and wild-type strains (as well as complemented mutants) became larger as the hemin supplement increased from 1.25 to 2.5 mg/mL. These results suggested that the Hmu system was essential for K. pneumoniae to thrive in the presence of hemin.
Role of Hmu in bloodstream infection. Since Hmu was essential in the hemin uptake process, we speculated that intact HmuRSTUV played an important role in bloodstream infections. A mouse lethality assay was performed using a bloodstream infection model for D3, D3DhmuR, and D3DhmuR::hmuR strains. As shown in Fig. 6, mice challenged with the D3DhmuR strain had lower risks of death than those challenged with the D3 and D3DhmuR:: hmuR strains (P = 0.008 [log-rank test]). Although D3 and D3DhmuR::hmuR strains exhibited comparable virulence (hazard ratio, 0.529; 95% confidence interval [CI] = 0.209 to 1.340; P = 0.179, with strain D3 as a reference), the D3DhmuR strain showed significantly reduced virulence compared to strains with intact hmuR (hazard ratio, 0.128; 95% CI = 0.032 to 0.502; P = 0.003, with strain D3 as a reference). These results indicated that an intact Hmu system played an important role in bloodstream infections.

DISCUSSION
To survive in iron-limited conditions, K. pneumoniae strains have evolved multiple iron uptake systems, which can be mainly summarized into three categories (26). The most important category is siderophores, such as salmochelin (IroBCDN), yersiniabactin (Irp1, Irp2, YbtAEPQSTUX, and FyuA), and aerobactin (IucABCDIutA), which are associated with enhanced virulence (25). The second category is iron ion uptake systems, including  (7). The third category is the HmuRSTUV heme/hemin uptake system. In the present study, we established a novel typing scheme for K. pneumoniae based on hmuRSTUV allelic profiles. Siderophore salmochelin is encoded by the iro locus (iroBCDN), aerobactin is encoded by the iuc locus (iucABCDiutA), and yersinibactin is encoded by the ybt locus (irp1, irp2, ybtAEPQSTUX, and fyuA). Lam et al. systematically analyzed the genetic diversity of these three siderophores and accordingly established novel typing schemes, namely, SmST, AbST, and YbST (24,25). iro and iuc were detected at low prevalence (,10%), and the ybt locus was detected in 40% of the K. pneumoniae genomes (24,25). These three siderophores were frequently associated with an elevated virulence of K. pneumoniae. In contrast, hmu was detect in most of K. pneumoniae genomes (98.9%) in this study, suggesting that the hmu locus was conserved in the genome of K. pneumoniae. Among 2,503 K. pneumoniae genomes, 62 AbSTs and 35 SmSTs were identified (25). Partly due to the relative higher prevalence and longer sequence of the ybt locus, 329 YbSTs were identified from 2,498 K. pneumoniae genomes (24). In our study, the hmu locus was much more prevalent, and 446 hmSTs were identified from 2,242 genomes, indicating that the hmST typing scheme might provide a novel strategy for monitoring and researching the molecular epidemiological characteristics of K. pneumoniae, regardless of lineage, virulence, or antimicrobial resistance. Besides via degradation of erythrocyte in circulation, host heme could be acquired from diet absorption. Heme is one of the predominant forms of iron in the human diet (17), especially in meat, poultry, and fish (27). Consequently, heme transporters have been identified in human intestinal epithelial cells (28). Therefore, it is important to understand how Hmu influences the fitness and survival advantages of K. pneumoniae strains in the gut and further causes colonization or infections. Interestingly, hmSTs are more diverse in hypervirulent lineages. In our previous study, we found that the average difference of core genomic profile in hvKP strains was significantly larger than that in MDR strains (23), indicating greater genome diversification among hvKP strains. This is coincident with our finding that hmST was more diversely distributed among hvKP strains than among MDR strains. The genetic backgrounds-including chromosomal recombination, surface polysaccharide locus diversity, pangenome, plasmid, and phage dynamics-were different between these two lineages (5). In this study, hmST1 was only  Similarly, other siderophore loci typing methods (such as AbST based on iuc loci and SmST based on iro loci) also revealed that hypervirulent and MDR lineages were associated with different allelic profiles. For example, hvKP lineages frequently harbored iuc1/iro1 and iuc2/iro2, whereas iuc5/iro5 were usually associated with MDR lineage CG258 (25). Also, in our previous study on the genetic diversity of siderophore loci, MDR lineage ST11 strains were associated with the yersiniabactin locus ybt9, while hypervirulent ST23 was associated with ybt1 (29). Compared to ybt, iro, and iuc, the hmu locus was significantly more common in K. pneumoniae and showed greater allelic diversity, which made hmST suitable for almost all strains of the K. pneumoniae species complex. To summarize, these evidences suggested that the hmST scheme was useful in typing and distinguishing hvKP from MDR K. pneumoniae.
For the use of hmST, all hmuRSTUV allelic sequences could be downloaded from the Ridom website (https://www.cgmlst.org/ncs/schema/). The software Ridom SeqSphere1 could be used to acquire the hmuRSTUV alleles of a new genome. To identify the hmST for a new genome, one can directly acquire the allelic profiles of hmuR, hmuS, hmuT, hmuU, and hmuV by Ridom SeqSphere1 or Blast. Each unique combination of alleles was assigned an hmST. By searching Table S1, one can get the hmST of the new genome. In the future, the development of a more user-friendly tool for hmST (such as a web-interactive tool) is needed.
As shown in the growth curves, the growth effect of the hmuR mutation is more significant in strain D3 than in strain ZKP51; this might also have resulted from a genetic Novel Typing Scheme Based on Hmu Microbiology Spectrum difference between the hvKP and MDR lineages. D3 is a hypervirulent strain producing four types of siderophores, i.e., enterobactin, salmochelin, yersiniabactin, and aerobactin, whereas ZKP51 is a CRKP strain that produced only enterobactin. We can therefore hypothesize that D3 required a much greater demand for iron. Thus, with hemin as the only source of iron, the inactivation of hmuR might have much more impact on the growth in D3 than the growth of ZKP51. The growth discrepancy between hmuR mutants and wild-type strains became larger as the hemin supplement increased from 1.25 to 2.5 mg/mL, suggesting the importance of an intact Hmu system in the presence of hemin. Therefore, in the bloodstream, where free heme could be released from hemoglobin, an intact Hmu system might facilitate bacterial growth and cause worse clinical outcomes in bloodstream infections.
In conclusion, HmuRSTUV is a critical iron uptake system in K. pneumoniae strains. The encoding locus hmuRSTUV was conserved in K. pneumoniae genomes, but genetic diversity was observed in hmuRSTUV allelic profiles. Based on the allelic profiles of hmuRSTUV, a novel typing scheme, hmST, was established. hmSTs were more diversely distributed in hypervirulent lineages than in MDR lineages. Interestingly, hypervirulent and MDR lineages had no shared hmSTs, suggesting that the typing scheme might be useful in distinguishing hvKP from MDR K. pneumoniae. The Hmu system played an important role in bloodstream infections, and hmST might provide an alternative strategy for monitoring and researching the molecular epidemiological characteristics of K. pneumoniae.

MATERIALS AND METHODS
Bacterial genomes and strains. The study genomes constituted of 1,970 genomes from Pathogenwatch (https://pathogen.watch/) and 272 genomes from our previous study, which covered 20 tertiary hospitals in China (23). All the 1970 genomes from Pathogenwatch (up to 20 April 2020) were downloaded as study genomes without selection. All of the information concerning these genomes (including isolating times, locations, infection types, and accession numbers) are listed in Table S2.
Isolates D3 and ZKP51 were selected to determine the function of HmuR. D3 and ZKP51 were representatives of the hypervirulent and MDR lineages, respectively. D3 was an ST23 and K1 strain, isolated from clinical liver abscess. D3 produced four types of siderophores, i.e., enterobactin, salmochelin, yersiniabactin, and aerobactin. ZKP51 was a K. pneumoniae strain isolated from a patient with bloodstream infection. ZKP51 belonged to ST11 and produced only enterobactin.
Knocking out and in situ complementation. To determine the function of Hmu, mutagenesis and complementation of hmuR were performed via a CRISPR-Cas9-mediated genome-editing method (30). Generally, the single-guide RNA (sgRNA) directs the Cas9 protein to a target sequence in the presence of a 59-NGG-39 protospacer adjacent motif (PAM). Cas9 nuclease then cleaves the target sequence to cause a double-strand break. To repair the break, homologous recombination occurs in the presence of exogenously supplied donor DNA repair templates.
Specifically, a 20-bp spacer sequence before a PAM site (59-NGG-39) in the target sequence was ligated into plasmid pSGKP between two reversed BsaI sites, which was located between the J23119 promoter and the sgRNA scaffold (30). For hmuR knockout, the donor DNA was generated by connecting the forward and reverse homology arms by overlap PCR. Cotransformation of the spacer introduced the pSGKP plasmid and the donor DNA into the L-arabinose-induced recipient cells harboring the pCasKP plasmid, which expressed the Cas9 protein and lambda Red recombination system. The LB agar plate containing 50 mg/mL apramycin and 100 mg/ mL hygromycin was used to screen the correct transformants at 30°C overnight. The successful knocking out was verified by PCR and Sanger sequencing. The process of in situ complementation was the same as for the knockouts. Of note, a silent mutation should be introduced in the PAM site on donor DNA to prevent the cutting of Cas9 when performing complementation. The primers used in this study are listed in Table 1, and the fragment for hmuR knockout and complementation is indicated in Fig. 1B. A detailed protocol on performing the knockout and in situ complementation is included in the supplemental material.
Plasmid pCasKP contained the temperature-sensitive replicon repA101(Ts) and pSGKP harbored the sucrose-sensitive gene sacB (30). To cure these two plasmids after successful editing, cells were streaked onto a Luria-Bertani (LB) agar plate containing 5% sucrose and incubated at 37°C overnight. Several Novel Typing Scheme Based on Hmu Microbiology Spectrum colonies were plated onto LB plates with or without apramycin or hygromycin supplementation, respectively. The colonies with the successful curing of both plasmids could only grow on the plate without antibiotic. Growth curve. The growth curve analysis was performed as previously described (31). Briefly, three independent cultures for each strain were grown overnight in Mueller-Hinton (MH) broth and diluted to 1:1,000 in M9 standard medium with bovine hemin supplement. To ensure that that only hemin could be used as an iron source, 250 mM 2,29-bipyridyl (BIP) was added to M9 medium. Three replicates of each culture (200 mL) were added to a flat-bottom 100-well plate. The plate was incubated at 37°C with agitation, and the optical density at 600 nm (OD 600 ) values of each culture were recorded every 5 min by using a Bioscreen C automated microbiology growth curve analysis system (Oy Growth Curves Ab, Ltd., Turku, Finland). OD 600 values of media without bacteria were used to normalize the initial OD for each culture.
Mouse lethality assay. D3, D3DhmuR, and D3DhmuR::hmuR strains were used to estimate the role of HmuR in bloodstream infection. We added 20 mL of overnight-cultured strain into 2 mL of MH broth, followed by incubation for 3 h until the samples reached the mid-log phase. All cultures were centrifuged at 6,000 rpm for 5 min, and the supernatants were discarded. We washed the cells with phosphate-buffered saline once and then resuspended them. Cell suspensions of 3 Â 10 6 CFU (0.2 mL) were injected into the caudal veins of 6-to 8-week-old BALB/c mice. Ten mice were used as a sample population for each strain. The mortality rates were recorded for 5 days postinjection. All procedures performed in studies involving animals were approved by the Institutional Animal Care and Use Committee of Sir Run Run Shaw Hospital.
Data availability. All the sequence accession numbers of the study genomes were available in the supplemental materials.