Role Played by the Fibronectin‐Binding Protein SfbI (Protein F1) ofStreptococcus pyogenesin Bacterial Internalization by Epithelial Cells

We read with great interest the article by Jadoun et al. [1], which confirms our previous findings [2] that the fibronectin-binding protein I (SfbI) of Streptococcus pyogenes mediates not only bacterial attachment but also internalization into nonphagocytic cells. For the benefit of the readers, we must clarify that the SfbI protein studied by our group [3] is, as determined by sequence analysis, an allelic variant of the protein that Hanski et al. [4] identified in parallel studies and named “protein F,” which was recently renamed “protein F1” [5]. Jadoun et al. [1] provided important insights regarding the capacity of this protein as invasin, using a genetic approach. The characterization of F1- and M6-deficient mutants generated in an S. pyogenes M6 background allowed Jadoun et al. [1] to conclude that M6 is the main adhesion factor and that both proteins are required for efficient internalization. Of interest, in the initial part of their paper, the authors stated that “nothing was known regarding the gene(s) or the underlying mechanism(s)” involved in S. pyogenes invasion. We would like to point out that our studies published in April 1997 [2] demonstrated that not only S. pyogenes internalization, but also attachment, was blocked either by treating HEp-2 cells with purified SfbI protein or by using antibodies against the full-length SfbI protein or its fibronectin-binding domain. In addition, inert latex beads coated with purified recombinant SfbI protein invade eukaryotic cells, demonstrating that SfbI alone is enough to trigger the internalization process without requiring any additional bacterial protein. Therefore, the major role played by M6 protein in bacterial attachment and the requirement of both M6 and F1 for the invasive phenotype reported by Jadoun et al. [1] may be strain-specific or related to isolates of the M6 serotype. Although SfbI/F1 protein has the capacity of mediating invasion, we consider incorrect the general statement that “protein F1 is required for efficient entry of S. pyogenes into epithelial cells.” In fact, although they neither bind fibronectin nor possess the prtF1 gene, the majority of the isolates reported by Jadoun et al. [1] are able to invade eukaryotic cells efficiently. We have recently provided epidemiologic evidence regarding the relationship between the in vitro internalization of S. pyogenes isolates and their clinical source [6]. Testing the invasiveness of 96 clinical isolates (56 from throat or skin and 40 from blood), we have identified 6 isolates belonging to different M serotypes that, although sfbI-negative and unable to bind fibronectin, showed high in vitro invasiveness. Furthermore, Cue et al. [7] reported that the in vitro invasiveness of an M1 serotype isolate can be activated by mammalian proteins and RGD peptides. Therefore, we would like to stress that the internalization of S. pyogenes appears to be a very complex, strain-dependent process. Although SfbI/F1 protein can mediate internalization via fibronectin, there are certainly other bacterial structures able to trigger uptake by nonphagocytic cells. S. pyogenes can infect different host tissues, causing a broad range of diseases. Since extracellular matrix components differ in specific tissues, bacteria might produce different adhesins, invasins, or both according to the specific target. It has also been suggested that streptococci use the invasion of epithelial cells to gain access to deeper tissues or to the bloodstream [8]. In agreement with Jadoun et al. [1], we observed higher in vitro invasiveness in isolates recovered from patients with noninvasive diseases, such as pharyngitis [6]. To provide further hints concerning the role of SfbI in bacterial invasion, we studied the phenotypic expression of SfbI in 61 isolates from different sources. The sfbI gene is present in 90% and 70% of the blood and throat or skin isolates, respectively [9]. Of interest, blood isolates exhibit poor attachment and internalization, and only 27% of the 36 sfbI-positive isolates tested expressed SfbI [6]. In contrast, 95% of the 25 tested throat or skin isolates containing the sfbI gene expressed SfbI, and 60% were able to invade efficiently HEp-2 cells in vitro. Not only should genotypic and phenotypic studies always be carried out in parallel, but speculation about the significance of a particular phenotype and its potential relevance in natural infections should be based on analysis of a large number of S. pyogenes isolates from different sources.