Introduction: Staphylococcus aureus is a major human pathogen. Infections by this bacterium range from minor skin and soft tissue infections, to more invasive and life threating infections like sepsis, osteomyelitis, and pneumonia. In order to successfully infect the mammalian host, S. aureus has to overcome iron scarcity within the host. As such, S. aureus is thought to produce toxins that lyse erythrocytes, releasing hemoglobin, a critical iron source for S. aureus in mammals. The bi-component β-barrel pore-forming leukocidins kill human neutrophils and were mostly considered as virulence factors fighting the host innate immune system. We show here thatHlgAB and LukED are two potent hemolytic leukocidins against human erythrocytes. Furthermore, we describe the identification of the Duffy antigen receptor for chemokines (DARC) as the red blood cell receptor for both LukED and HlgAB leukocidins.
Results: We took advantage of erythrocyte samples from different Duffy (Fy) genotyped individuals from the French National Blood Transfusion Institute. We evaluated the receptor expression on the surface of erythrocytes (Figure 1). DARC-negative erythrocytes (phenotype Fya-/b-) were fully resistant to HlgAB and LukED. Compared to individuals expressing normal levels of DARC (phenotypes Fya+/b+, Fya+/b+weak, Fya+/a+, and Fyb+/b+), individuals expressing intermediate (phenotype Fyb+weak/b+weak) or very low levels of DARC (phenotype Fyb-/b+weak) showed intermediate susceptibility to both HlgAB and LukED (Figure 1). These data demonstrate that the hemolytic activity of HlgAB and LukED is dependent on the expression of DARC at the red blood cell surface. To better understand the interaction of HlgAB and LukED with DARC, we screened HEK293T cells transfected with plasmids encoding DARC mutants (Tournamille et al, 2003). We demonstrate that while both HlgAB and LukED bind to DARC with nanomolar affinities, they do so by recognizing different domains of the receptor. To directly evaluate whether LukED and HlgAB can promote bacterial replication as a result of erythrocyte lysis, S. aureus was grown in iron-starved medium supplemented with cell-free extracts of erythrocytes treated with LukED or HlgAB. We observed that HlgAB and LukED were each capable of promoting S. aureus growth in a DARC and hemoglobin scavenging system (IsdBH)-dependent manner. These in vitro studies were supported by a murine bacteremia model.
Discussion: By combining human studies of DARC polymorphisms with gain and loss of function experiments and biochemical analyses, we demonstrate that DARC is necessary and sufficient to render host cells susceptible to LukED and HlgAB. By targeting DARC, HlgAB and LukED support S. aureus growth in a hemoglobin-acquisition dependent-manner. Thus, these findings provide the missing link of how S. aureus targets and lyses erythrocytes to release one of the scarcest nutrient within the mammalian host. Human epidemiological studies comparing the severity of S. aureus infection in patients with DARC positive or DARC negative erythrocytes are now required to evaluate the contribution of DARC-mediated hemolysis in human staphylococcal diseases. Given the resistance of DARC negative erythrocytes to the parasites Plasmodium vivax and P. knowlesi, and now to the hemolytic activity of the bacterium S. aureus, our findings suggest the possibility of a positive selection event in response to these important human pathogens.
(A) Susceptibility of human erythrocytes to S. aureus β-barrel pore forming toxins. The dashed line indicates 50% hemolysis. n = 6.
(B) Levels of DARC and CD55 on erythrocytes of donors with different Fy phenotypes. The dashed line indicates the detection threshold. n = 2-7 ± SEM.
(C,E) Susceptibility of human erythrocytes with different Fy phenotypes to HlgAB (C) and LukED (E). The dashed line indicates 50% hemolysis. n = 2-7 ± SEM.
(D, F) Correlation of half-maximal effective concentrations (EC50) of HlgAB (D) and LukED (F) with the total number of receptors expressed on the erythrocyte surface.
For Fyb+weak/Fyb- donors, LukED EC50 could not be calculated. n = 2-7 ± SEM.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.