Streptomyces phage sequencing and genome comparisons
Previous investigations in our lab characterizing chemical anti-phage defence7 used a collection of Streptomyces coelicolor phages (ϕScoe phages) isolated from geographically diverse soil samples. Negatively stained electron microscopy images showed that these phages belong to the order Caudovirales, with double-strand DNA genomes contained inside icosahedral heads and long, non-contractile tails. To gain insight into this phage collection, we determined their genome sequences using Illumina NextSeq 2000 short-read technology. Each genome was assembled into a single contig, with > 86% of the reads mapped to each sequence. The genomes were annotated using automated gene predictions followed by careful manual inspection using our in-house phage database for functional annotations.
DNA-DNA comparisons performed using Pyani18 revealed that the twelve ϕScoe phage genomes shared pairwise nucleotide identities ranging from 73–91% across the complete genome sequences (Fig. 1, Supplementary Fig. S1a). The ϕScoe phage genome lengths varied in size from 48–51 kb, and each was found to contain between 75 and 84 predicted open reading frames (ORFs) (Table 1). Of these ORFs, 51 were found to be conserved among all twelve phages (Fig. 2, Supplementary Fig. S1b, Supplementary Table S1). The genome organization of these phages is similar to previously characterized Streptomyces phages, encoding the functional modules necessary for Siphophage particle formation (DNA packaging, capsid and tail assembly), cell lysis, DNA replication, as well as an immunity repressor and serine integrase that suggest a temperate lifestyle for these phages (Supplementary Table S2). Genome comparisons with previously characterized phages revealed striking similarity to Streptomyces phages R419 and ϕJoe20, with average pairwise nucleotide identities of 71–78% and 73–89% respectively (Fig. 1, Supplementary Fig. S1). The R4-like group of phages appears to be the most abundant group of Streptomyces phages21, falling within the BD family at the Actinobacteriophage Database22.
Table 1
Phage | Genome Size (kb) | GC Content (percent) | Predicted ORFs | Lifestyle | Genbank Accession |
ϕScoe1 | 51 | 66.83 | 80 | Temperate | OQ672741 |
ϕScoe2 | 49.4 | 65.77 | 79 | Temperate | OQ672744 |
ϕScoe3 | 48.3 | 65.6 | 81 | Temperate | OQ672747 |
ϕScoe10 | 51.1 | 65.23 | 80 | Temperate | OQ672742 |
ϕScoe15 | 49.4 | 66.07 | 80 | Temperate | OQ672743 |
ϕScoe23 | 49.8 | 65.84 | 75 | Temperate | OQ672745 |
ϕScoe25 | 49.4 | 66.13 | 84 | Temperate | OQ672746 |
ϕScoe44 | 48.8 | 65.08 | 80 | Temperate | OQ672748 |
ϕScoe45 | 48.8 | 65.57 | 80 | Temperate | OQ672749 |
ϕScoe54 | 48.7 | 65.83 | 78 | Temperate | OQ672750 |
ϕScoe55 | 48.5 | 65.17 | 77 | Temperate | OQ672751 |
ϕScoe56 | 48.4 | 65.47 | 78 | Temperate | OQ672752 |
Conserved Streptomyces phage proteins
Both R4 and ϕJoe are temperate phages, which means that they can pursue either a lytic or lysogenic lifestyle following phage infection. Temperate phages require an immunity repressor to control the balance between lysis and lysogeny. They do so by downregulating promoters required for the lytic pathway. Each of the ϕScoe phages was found to encode a repressor protein between the head-tail-lysis genes and the putative early region genes. These repressor proteins share sequence identities of 64 to 90% with the repressor proteins found in Streptomyces phages R4 and ϕJoe (Supplementary Fig. S2). Phage R4 has been shown to have an atypical regulatory mechanism that silences transcription from the prophage, known as the repressor-stoperator regulatory system21. This type of repressor regulates transcription initiation at an early lytic promoter, but also acts at many sites across the phage genome, blocking the movement of RNA polymerase and thereby helping to maintain lysogeny. Bioinformatic analyses revealed eight or more R4 stoperator binding sites in intergenic regions of the ϕScoe phages (Supplementary Table S3), suggesting that this mechanism of repression is conserved across this family of phages.
Most temperate phages encode an integrase protein whose activity mediates integration and excision of the prophage during the lysis-lysogeny decision. The integrases of many Streptomyces phages have been shown to belong to the large serine recombinase family23. In keeping with this, the ϕScoe phages characterized in this work each encode a serine integrase. These proteins are highly variable in sequence across the twelve phages, with pairwise sequence identities ranging from 13% – 91% (Supplementary Fig. S3a). Two groups of phages have more closely related integrases: phages ϕScoe2, ϕScoe15, and ϕScoe55 share 75% – 91% identity with each other and the previously characterized ϕJoe integrase, while ϕSoce25, ϕScoe56, ϕScoe23, ϕSoce45, and ϕScoe10 share 64% – 83% identity among their integrases. During phage integration, the serine integrase protein acts at specific sites, the attP site in the phage and the attB site in the bacterial chromosome. Phages that encode similar integrases and attP sites have been shown to use the same bacterial attB sites for integration24. To identify potential attP sites in the ϕScoe phages we queried their sequences with known attP sequences from previously characterized actinophages R425, ϕJoe20, ϕC3126, ϕBT127, TG128, and SV114 (Supplementary Fig. S3b). Sequences related to the ϕJoe attP site were identified upstream of the integrase genes in phages ϕScoe2, ϕScoe15, and ϕScoe55 (Supplementary Fig. S3c). These four phages encode integrases that share high sequence identity and are more distantly related to the other serine integrases found in the ϕScoe group of phages. AttP sites were not identified for the remaining phages.
The genes that are required for the assembly of the head and long, non-contractile tail of the siphophage virion are typically found in a cluster with a conserved gene order. As previously described29, we combined analysis by HHpred30 and gene position to annotate genes encoding morphogenetic functions, such as the large and small terminases, the capsid and portal proteins, head-tail joining proteins, and tail tube protein, that are conserved among most long-tailed phages (Supplementary Table S2). These functions were named according to Casjens et al29. Proteins comprising the distal end of the tail, referred to as the baseplate or tail tip complex, vary among different types of long-tailed phages. In this region, the ϕScoe phages resemble other siphophages that infect Gram-positive species, such as Lactococcus phage TP901-1. Accordingly, the first conserved gene in the baseplate region encodes the Distal Tip (Dit) protein and the second gene encodes a protein that structurally resembles the baseplate hub protein (BH2) of myophages31 and is often referred to as Tal in the Gram-positive siphophage literature32. Proteins encoded by genes CG21, CG22, and AG22 are predicted by HHpred to resemble receptor binding proteins found in other siphophages; thus, these proteins have been assigned that function (Supplementary Fig. S1). These receptor binding proteins share average pairwise sequence identities of 34–70%. The greatest variability observed in the morphogenetic region was noted in the proteins that comprise the tail tip (Fig. 3). The variability in the tail tip proteins may contribute to the bacterial host range differences noted for these phages.
Non-conserved phage proteins
In addition to the highly conserved genes required for replication and assembly of the phage particle, the ϕScoe phages encode many genes that are not conserved among all members of this group, which we refer to as accessory genes (AGs). Between the twelve phages there are a total of 96 different accessory genes that are present in at least one phage. Eleven accessory genes are present in every phage except ϕScoe1. Nine of these genes are of unknown function, and the remaining two encode for Lsr2 and Ocr (Overcome classical restriction) proteins (Fig. 2, Supplementary Table S1). Lsr2 is a nucleoid-associated protein that binds AT-rich regions of DNA; it has been shown to play a role in silencing prophages in Streptomyces venezuelae33 and Corynebacterium glutamicum34. As the ϕScoe phages encode repressor proteins that maintain lysogeny, the Lsr2 proteins they encode may play a role in microbial warfare by inhibiting the host encoded Lsr2 proteins. Lsr2 has also been shown to be influence the dynamics of the replication machinery and the duration of DNA synthesis in Mycobacteria35, suggesting a possible role for the phage Lsr2 proteins in disrupting host processes to subvert them in a way that benefits the phage. Ocr is a protein found in phage T7 that counters restriction-modification and BREX anti-phage defences36,37 and likely plays a similar role in these phages, perhaps to counter the BREX phage growth limitation (Pgl) system38 found in S. coelicolor.
Fifty-nine genes are found in only one phage (Fig. 2, Supplementary Table S1). Of these genes, only two have predicted functions; ϕScoe23 encodes a DNA polymerase III (AG58), and ϕScoe45 contains an endonuclease (AG62) in addition to the HNH-endonuclease that all twelve phages share. Phage ϕScoe10 contains a cluster of five accessory genes in the morphogenetic region, between the portal protein and capsid maturation protease (Fig. 2). Four of these genes are annotated as hypothetical proteins, while AG17 shares 62% sequence identity with a tail fibre protein from an Agrobacterium phage, suggesting a potential role in bacterial host range. Most accessory genes in this group of phages are located downstream of the morphogenetic region. These genes share synteny, but there are several local rearrangements in this region (Fig. 2). For example, AG67 in ϕScoe23 and AG76 in ϕScoe2 are present in different relative locations than in the remaining phages that encode these proteins. This pattern is observed for conserved genes as well. For example, conserved gene CG47 appears at different locations in phages ϕScoe1, ϕScoe2, and ϕScoe23 as compared to the rest of the phages.
Phage infection induces secondary metabolite production and early sporulation
While the ϕScoe group of phages share many genes across the full length of the genome, the variable accessory gene complements suggested that they may have different host ranges. To address this question, the twelve phages were propagated to high titer in S. coelicolor and their plating efficiencies were assessed on a collection of 21 Streptomyces species. Each phage showed a unique infection profile (Fig. 4a), with all phages able to efficiently replicate in more than half of the species tested. In some cases, we observed clearings that did not resolve into plaques (grey shading). This suggests that the phages infect but are unable to proceed through the replication cycle, perhaps because the strain lacks a host factor that the phage needs to replicate, or because it encodes a defence system that targets the phage intracellularly. As we previously determined7, some species are highly resistant to phage infection, while others are very sensitive. No patterns between phage gene conservation and host range could be discerned. However, we noted that some Streptomyces species appear to upregulate secondary metabolite production in the presence of phage challenge as shown by the production of coloured compounds (Fig. 5, 6).
Secondary metabolite production39 and protein-based anti-phage defences40–42 have been shown to be upregulated at high cell density. Streptomyces have a complex developmental lifecycle, in which they grow from spores into a branching network of young mycelia. The young mycelia grow into mature mycelia over time, moving into exponential growth approximately 12–14 hours post-inoculation43,44. The cells move into a transitionary phase after approximately 18 hours of growth and undergo a developmental switch, after which they begin producing specialized metabolites and aerial hyphae which mature into spores45,46. To determine how this developmental process affected the ϕScoe phages, we determined how efficiently they could initiate infection and replicate in S. coelicolor over time. The phage lysates were diluted to approximately 3x106 plaque forming units (PFU) per mL and plated on cells that had been grown for various amounts of time following spore germination. Most of the phages maintained normal plating efficiency until ~ 16h, with some (e.g. ϕScoe2 and ϕScoe45) still able to plate on cells that had been grown for 18 hours before the phages were applied (Fig. 4b). By contrast, ϕScoe56 formed plaques poorly at 15 hours, suggesting that this phage was more sensitive to some cellular factor that is differentially expressed during the Streptomyces life cycle. None of the phages were able to efficiently replicate in mature mycelia, which is consistent with previous work showing that Streptomyces phages replicate best in exponentially growing cells17,45,46.
We also investigated the effects of ϕScoe phage infection on bacterial growth. In this assay ~ 100 phages were applied to a DNB-agar plate adjacent to a spot containing ~ 1000 S. coelicolor spores, so that the two samples were partially overlapping (Fig. 5a). These cultures were incubated at 30°C and bacterial growth was monitored using a high-resolution scanner, with images captured every 12 hours for a 120-hour incubation period. Bacterial growth was initially inhibited by phage infection where the bacterial spores and phages overlapped (Fig. 5b). At 24–36 hours bacterial growth began to appear where the initial phage infection had taken place, radiating out from the larger uninfected colony into the zones of clearing. The cells that grew in these areas, which had initially been completely lysed by the activity of the phages, appeared to be resistant to phage infection – this could be either through mutations in the bacteria that provide resistance, through the formation of lysogens, or perhaps through a developmentally mediated resistance mechanism. To determine if bacterial mutations or lysogen formation was the reason for the observed resistance, we isolated colonies from these resistant cells and prepared spore stocks to allow us to test for resistance to further phage infection. Following germination of the spores, no phage resistance was observed and PCR-based screening of the previously phage resistant cells showed that they did not contain prophages. These data suggest that resistance that arises following phage challenge is through some developmentally regulated factor.
The infection and replication of the ϕScoe phages in S. coelicolor colonies triggered the production of a red-coloured secondary metabolite in response to the phage challenge (Fig. 5b). S. coelicolor is known to make two red metabolites, actinorhodin47 (red or blue depending on pH) and undecylprodigiosin48. To determine which of these metabolites was being induced by phage infection, we monitored growth of S. coelicolor strains that harboured mutations in actI49 and redD50, which do not produce actinorhodin or undecylprodigiosin, respectively. Infection of wild type S. coelicolor by all ϕScoe phages induced high levels of red metabolite at 36 hours (Fig. 5b, c). This phenotype was greatly abrogated in the actI mutant, suggesting that actinorhodin is the major contributor to this phenotype (Fig. 5c). This result is consistent with recent work that showed several other Streptomyces phages induced actinorhodin production upon phage infection51. Infection of S. coelicolor by several of the ϕScoe phages also appeared to trigger early sporulation in infected colonies as shown by the appearance of a fuzzy white colony morphology at later timepoints (Fig. 5b).
To determine if these metabolite and sporulation effects are a general response to phage infection, we challenged 15 other Streptomyces species with phages that were able to plate to high titer and looked for visible cellular responses. Phage infection induced production of a dark brown metabolite in WAC170 (Fig. 6a). Notably, induction of this metabolite appears to be phage-specific; although ϕScoe1 visibly infected and killed a portion of the bacterial colony it did not induce production of this secondary metabolite. Strains WAC178 and WAC288 both showed signs of early sporulation in response to phage infection (Fig. 6b,c). As Streptomyces phages replicate only in exponentially growing cells17,45,46, this hastening of sporulation may be an adaptive response to allow cells within the colony to become phage resistant and limit the phage epidemic.