No commercial vaccine exists against Group A streptococci (GAS; Streptococcus pyogenes) and only little is known about anti-GAS protective immunity. In our effort to discover new protective vaccine candidates, we selected 21 antigens based on an in silico evaluation. These were all well-conserved among different GAS strains, upregulated in host-pathogen interaction studies, and predicted to be extracellular or associated with the surface of the bacteria. The antigens were tested for both antibody recognition and T cell responses in human adults and children. The antigenicity of a selected group of antigens was further validated using a high-density peptide array technology that also identified the linear epitopes. Based on immunological recognition, four targets were selected and tested for protective capabilities in an experimental GAS infection model in mice. Shown for the first time, three of these targets (spy0469, spy1228 and spy1801) conferred significant protection whereas one (spy1643) did not.


Group A streptococci (GAS; Streptococcus pyogenes) are major human pathogens causing a wide variety of diseases ranging from uncomplicated infections like pharyngitis and impetigo to life-threatening invasive diseases1. A conservative estimate propose that GAS infections and their sequelae account for more than 500,000 annual deaths2. There are no licensed vaccines against GAS.

The most studied antigen is the surface M protein where both the variable N-terminal and its conserved C-terminal region have been proposed as vaccine candidates3,4,5,6. However, the presence of more than 220 different emm types7and the extensive amount of evidence that immune responses against the M protein are associated with development of post-strep sequelae8,9 have made the investigation of conserved non-M protein antigens attractive. Some protective non-M protein antigens, such as the streptococcal C5a peptidase (ScpA), the IL-8 serine protease (SpyCEP) and fibronectin-binding proteins have been identified, although they have yet to enter clinical trials (reviewed in10,11).

The goal for this study was to identify new protective non-M protein vaccine candidates. The first selection criterion in our strategy was that the antigens should be upregulated following interaction with the host. The next requirement was an immunological recognition in human adults and children. Based on recent data, we decided that our vaccine candidates should be recognized by both antibodies and T cells. It is already well-established that antibodies have protective capacity12,13,14,15,16,17 but cellular responses have also been suggested to possess antibody-independent protective capacity in a murine GAS infection model18,19,20. In addition, in a recent study we showed that the majority of both children and adults not only possess antibody responses, but also strong Th1 responses against GAS antigens21. Besides immunological recognition, we also added the criteria that antigens should be conserved among GAS stains and be extracellular or associated with the surface of the bacteria.

Although our strategy in part overlaps with approaches used in previous GAS antigen discovery studies22,23,24,25,26,27, we succeeded in identifying three undescribed GAS antigens that were all recognized by T- and B cells. We show that the three antigens were able to protect against infection with GAS in a murine infection model.


Selecting GAS antigens

The aim of this study was to identify and characterize protective GAS antigens that constitute both T- and B cell targets. Our approach was to select antigens that displayed increased gene expression during interaction with the host, as these are likely to represent key factors in the establishment of an infection. Several transcriptome studies of GAS bacteria recovered after interacting with the host have been published. These include experimental pharyngitis in cynomolgus macaques28, growth in human blood29, soft tissue infection in mice30 and phagocytosis by human polymorphonuclear leukocytes31. Among the upregulated genes in these studies we selected a subset that were all i)conserved with over 90% identity in most of the 26 (gap-free) GAS genome sequences that were available at the time of the study (representing 16 emm types, see Supplementary table 1) and ii)predicted to have an extracellular location with the pSORT v3.0 online software or to be integral to the membrane with extracellular domains using the TMpred server32. In total, we selected 21 GAS antigens (table 1) and all of these were expressed as recombinant proteins. An overview of the individual design for each recombinant antigen can be found in Supplementary table 2.

Table 1: Overview of selected Streptococcus pyogenes antigens.

Antigens recognized by human IgG

All antigens were next tested for recognition by antibodies and T cells in humans, starting with recognition by IgG in human plasma. We first performed an initial screen for IgG reactivity in three pools of human plasma from 32 healthy adults (10–12 adults in each pool). This provided an indication of the “immune reactive” antigens, from which antigens would be selected for a more detailed analysis in individual donors. Spy0269, spy0469, spy1228, spy1643, spy1801 and spy2010 showed the highest IgG responses (Fig. 1, and Supplementary figure 1 that shows the plasma titration curves used to calculate the EC50 values). Spy2010 (C5a peptidase; ScpA) and spy0269 are well-known protective vaccine antigens23,33, but the other antigens were undescribed as protective antigens in the literature, and were therefore selected for further analysis in individual plasma samples. Spy0269 was included for comparison.

Figure 1: Initial screening of antigen specific IgG in human plasma pools.