A regulatory cascade controls Staphylococcus aureus pathogenicity island activation

Phage-inducible chromosomal islands are induced by phages and then exploit these for their own transfer. Here we identify a new group of inducible islands that is activated by other islands rather than phages.

Like Comment
Read the paper

Mobile genetic elements (MGEs) play a crucial role in microbial ecology and evolution. These include, among others, bacteriophages (phages), phage-inducible chromosomal islands (PICIs), integrative-conjugative elements (ICE), plasmids and transposons. They can encode for virulence determinants such as toxins and immune evasion factors, can confer new metabolic capabilities onto their host or enable them to resist antimicrobials. Bacteriophages (phages) are the most widely recognised MGEs and are the most abundant biological entities on the planet with an estimated number of more than 1031 phage particles in the biosphere1. They are viruses that infect bacterial cells and then either lyse their infected host (lytic cycle/lytic phages) or integrate into the host’s genome (lysogenic cycle/ temperate phages) where they reside in a dormant state. Phages participate in the dissemination of genetic material other than their own via several modes of horizontal gene transfer (HGT)2.

Phages and PICIs: a paradigm of molecular parasitism

Intriguingly, phages and their DNA packaging machinery are exploited by another class of MGEs, the PICIs3,4. In Staphylococcus aureus, PICIs are called Staphylococcus aureus pathogenicity islands or SaPIs and are integrated into the host chromosome at specific positions. Replication of these SaPIs can be induced by a (replicating) helper phage, either after phage infection or phage induction through the activation of the bacterial SOS response. The replicating phage produces a moonlighting protein that binds to the SaPI repressor and thereby removes it from its DNA target leading to the expression of SaPI replication and excision genes and the subsequent replication of the SaPI5. By employing this mode of activation, the SaPI limits the metabolic burden on the host and ensures replication is only induced in the presence of a helper phage that can package and disseminate the SaPI. SaPI induction frequently comes at a cost to the helper phage and its replication is reduced through a variety of SaPI-encoded interference mechanisms6.

SaPI3, an elusive pathogenicity island

Like temperate phages, replication of SaPIs needs to be tightly controlled to restrict the metabolic burden on the host and, in case of temperate phages, to prevent premature and unintended activation and host killing. In most SaPIs, replication is controlled by a regulatory switch, which is made up of a transcriptional repressor, Stl, that blocks the expression of the SaPI’s replication genes and a divergently transcribed transcriptional activator, Str (see Figure 1). However, in a SaPI family that includes SaPI3, and carries the toxin for food poisoning (SEB: enterotoxin B, a category 2 bioterrorism agent), the regulation module is different and has been precisely replaced with a module akin to that found in ICEBs1 of Bacillus subtilis7. This module consists of three rather than two genes encoding a repressor (ImmR) and a protease (ImmA) to one side of a divergent promotor and a transcriptional activator (Str’) to the other side. Although potentially fulfilling similar functional roles to those found in other SaPIs, no helper phages have been identified to this day leading to the accepted consensus that SaPI3 was a defective pathogenicity island.

Looking beyond the helper phage paradigm

And thus, we began our quest to understand more about this elusive and recalcitrant SaPI. We started out following two separate lines of investigation. First, we screened a large collection of phages to identify if any of them could induce SaPI3. Second, we were interested in the interactions that could occur between coresident SaPIs and how these could affect the dynamics of phage mediated transmission of these elements. Our first approach proved rather fruitless, the second however, showed that among all the SaPI combinations tested for interaction, pairs with SaPI3 and another SaPI showed altered transduction titres. Surprisingly, we saw that the presence of a second SaPI substantially increased transduction titres of SaPI3 far beyond levels observed via generalised (background) transduction. We were very excited by this discovery as it suggested that there were multiple layers of interactions occurring within the realm of phage-SaPI activation.

In search of the mechanism

We set out to identify the molecular mechanism underlying SaPI3 activation. Using a library of mutants8 in one of the helper SaPIs (SaPIbov1) that facilitated SaPI3 transfer, we quickly identified a single gene (orf16) encoding a protein of unknown function to be solely responsible for this. To our surprise, this gene was highly conserved in all staphylococcal SaPIs but could not be found in other bacterial species. We were able to show that many if not most members of this family were sufficient to promote SaPI3 replication and renamed the protein SaPI Inducer of SaPIs, Sis. When SaPIs are induced by a helper prophage, induction occurs through the interaction of the prophage-encoded inducer with the repressor of the induced SaPI resulting in a complex that is unable to repress SaPI replication. We reasoned that this was likely also the case in the interaction of the helper SaPI and SaPI3. Showing evidence of this interaction proved to be rather difficult as the helper SaPI-encoded inducer protein was highly toxic in Escherichia coli expression systems and none of the variants tested could be produced at sufficient levels to allow protein pulldown or bacterial two-hybrid interaction experiments. Having tried many different versions of the inducer protein in both assays, we decided to try super-resolution microscopy. Our hope was that we would see differences in intracellular localisation of fluorescently tagged SaPI3 repressor and fluorescently tagged Sis once they were expressed in the same cell. To our relief, this was exactly what happened, and we were able to show that only when both SaPI3 repressor and helper SaPI Sis were expressed in the same cell did both fluorescent tags localise in the same place confirming that an interaction between both proteins was responsible for SaPI3 induction.

Kickstarting the SaPI3 autoinduction circuit

However, SaPI3 is not only controlled by its ImmR repressor protein but also involves a protease, ImmA, encoded just downstream of the repressor (Figure 1). We showed that this protease acts directly on the SaPI3 repressor using a bacterial two-hybrid assay. We found that, while not essential for Sis mediated induction of SaPI3, it nevertheless played an important role in full activation of SaPI3 replication. Loss of the protease reduces the level of helper SaPI induced SaPI3 replication substantially. Additionally, we showed that the expression of the protease increases once SaPI3 is induced by Sis adding a first layer of positive feedback to the system.

As mentioned above, all SaPIs encode Sis and so does SaPI3. We were intrigue by this and wondered if the SaPI3 encoded Sis protein played an important role in the SaPI3 activation cycle. After all, it was the most active Sis in all the assays performed. We tested this by making a SaPI3 sis deletion mutant, expressed two Sis proteins from a plasmid and then tested whether this mutant could still be activated to the same extent as the wt. The SaPI3 sis mutant showed notably reduced SaPI3 transfer demonstrating that the SaPI3 Sis protein acted as a positive feedback mechanism that, once started, will keep the system going (Figure 1).

SaPI3 and beyond

Here we have identified a new regulatory cascade that involves three MGEs (Figure 1). A helper phage initially induces a helper SaPI. This SaPI then produces a protein, Sis, which can interact with the repressor protein of a satellite SaPI, SaPI3, kickstarting its replication. The regulatory system of SaPI3 differs from classic SaPI regulation not only because of its initial induction by a helper SaPI-encoded antirepressor, but also because it possesses a double auto-induction circuit. Autoinduction starts with the initial trans-derepression of the leftward immR promoter leading to the expression of immA. The ImmA protease then binds to the ImmR repressor and degrades it, which further upregulates the immR promoter. This process causes the activation of the rightward-facing str’ promoter, which also controls the expression of the SaPI3-encoded Sis, the most active Sis tested in our study. Since absence of SaPI3 sis reduces the transfer of SaPI3 irrespective of the Sis homologue kickstarting induction, SaPI3 sis is required for full induction of the SaPI3 replication cycle. In other words, once the helper-encoded Sis gets the system started, the double autoinduction circuit will do the rest (Figure 1).

An intriguing question that we are currently investigating is whether SaPI-encoded Sis can act on other MGEs that are regulated by the immA–immR system. If this is indeed the case, it is feasible that cross-regulating MGEs might have important biological effects on HGT, acquisition of virulence and antibiotic resistance. 


Model of SaPI3 activation
Figure 1 A regulatory cascade controls SaPI3 activation. (I) A replicating helper phage produces a protein that acts as antirepressor to the helper SaPI. (II) The phage-encoded antirepressor binds to the Stl repressor of the helper SaPI and removes it from the promoter region. The helper SaPI begins expressing genes to the right of its str promoter leading to the expression of the SaPI inducer of SaPIs (Sis). (III) Sis binds to the ImmR repressor of the satellite SaPI (SaPI3) and removes it from its target promoter. This leads to the expression of the ImmA protease, which degrades ImmR followed by expression of the SaPI3-encoded Sis homolog setting up a double-autoinduction loop that maintains satellite SaPI replication once it has been kickstarted by exogenous helper SaPI-encoded Sis.



1          Comeau, A. M. et al. Exploring the prokaryotic virosphere. Research in microbiology 159, 306-313, doi:10.1016/j.resmic.2008.05.001 (2008).

2          Chiang, Y. N., Penades, J. R. & Chen, J. Genetic transduction by phages and chromosomal islands: The new and noncanonical. PLoS Pathog 15, e1007878, doi:10.1371/journal.ppat.1007878 (2019).

3          Fillol-Salom, A. et al. Phage-inducible chromosomal islands are ubiquitous within the bacterial universe. ISME J 12, 2114-2128, doi:10.1038/s41396-018-0156-3 (2018).

4          Martinez-Rubio, R. et al. Phage-inducible islands in the Gram-positive cocci. ISME J 11, 1029-1042, doi:10.1038/ismej.2016.163 (2017).

5          Tormo-Mas, M. A. et al. Moonlighting bacteriophage proteins derepress staphylococcal pathogenicity islands. Nature 465, 779-782, doi:10.1038/nature09065 (2010).

6          Fillol-Salom, A., Miguel-Romero, L., Marina, A., Chen, J. & Penades, J. R. Beyond the CRISPR-Cas safeguard: PICI-encoded innate immune systems protect bacteria from bacteriophage predation. Curr Opin Microbiol 56, 52-58, doi:10.1016/j.mib.2020.06.002 (2020).

7          Auchtung, J. M., Aleksanyan, N., Bulku, A. & Berkmen, M. B. Biology of ICEBs1, an integrative and conjugative element in Bacillus subtilis. Plasmid 86, 14-25, doi:10.1016/j.plasmid.2016.07.001 (2016).

8          Ubeda, C. et al. SaPI mutations affecting replication and transfer and enabling autonomous replication in the absence of helper phage. Mol Microbiol 67, 493-503, doi:10.1111/j.1365-2958.2007.06027.x (2008).

Andreas Haag

Research Associate, University of Glasgow