A genome-wide overexpression screen reveals Mycobacterium smegmatis growth inhibitors encoded by mycobacteriophage Hammy

Abstract

During infection, bacteriophages produce diverse gene products to overcome bacterial antiphage defenses, to outcompete other phages, and to take over cellular processes. Even in the best-studied model phages, the roles of most phage-encoded gene products are unknown, and the phage population represents a largely untapped reservoir of novel gene functions. Considering the sheer size of this population, experimental screening methods are needed to sort through the enormous collection of available sequences and identify gene products that can modulate bacterial behavior for downstream functional characterization. Here, we describe the construction of a plasmid-based overexpression library of 94 genes encoded by Hammy, a Cluster K mycobacteriophage closely related to those infecting clinically important mycobacteria. The arrayed library was systematically screened in a plate-based cytotoxicity assay, identifying a diverse set of 24 gene products (representing ∼25% of the Hammy genome) capable of inhibiting growth of the host bacterium Mycobacterium smegmatis. Half of these are related to growth inhibitors previously identified in related phage Waterfoul, supporting their functional conservation; the other genes represent novel additions to the list of known antimycobacterial growth inhibitors. This work, conducted as part of the HHMI-supported Science Education Alliance Gene-function Exploration by a Network of Emerging Scientists (SEA-GENES) project, highlights the value of parallel, comprehensive overexpression screens in exploring genome-wide patterns of phage gene function and novel interactions between phages and their hosts.

Introduction

Bacteriophages and bacteria are locked in an ancient, high-stakes arms race, giving rise to myriad genetic strategies for infecting or thwarting infection, respectively (Hampton et al. 2020). Phage genomes encode a core set of gene products that form the virion structure or interact with host factors to carry out key steps of infection, including genome replication, gene expression, and cell envelope disruption. Typically, phages also encode a suite of highly variable accessory genes thought to confer an advantage under specific conditions (Dedrick et al. 2013, 2017; Hatfull 2020). Comparative analyses of many sequenced phages have revealed a continuum of genetic diversity, with mosaic genomes encoding thousands of different phamilies, or groupings of related gene products sharing >25% amino acid identity; most phamilies have no recognizable sequence features and cannot be assigned function (Cresawn et al. 2011; Pope et al. 2015, 2017; Hatfull 2020; Gauthier et al. 2022). In turn, bacteria have evolved a multitude of systems that sense and respond to phage infection via either abortive or nonabortive mechanisms (Hampton et al. 2020; LeRoux and Laub 2022; Millman et al. 2022). Recent work has revealed that several of these abortive systems are specifically triggered by core phage products including capsid structures (Tal et al. 2021; Zhang et al. 2022; Huiting et al. 2023) and DNA replication machinery (Stokar-Avihail et al. 2023). Phages are also in vicious competition with each other, and increasingly, many phage products are thought to interact with host factors to block infection by competitors (Dedrick et al. 2017; Ko and Hatfull 2018). It is evident that the interactions between phages and bacteria are complex, and further work is needed to elucidate the full expanse of genetic functions that both sides have evolved.

One powerful strategy for investigating phage gene function is overexpression in the bacterial host, connecting genes to phenotypes like inhibition of bacterial growth and serving as a first experimental filter to identify phage products that interact with key host factors (Liu et al. 2004; Molshanski-Mor et al. 2014; Ko and Hatfull 2018). Recent overexpression screens have shown that genes capable of disrupting mycobacterial growth are abundant within the mycobacteriophage population (Ko and Hatfull 2020; Heller et al. 2022). Systematic overexpression of each gene encoded by Cluster K mycobacteriophage Waterfoul revealed that roughly one-third of gene products hindered Mycobacterium smegmatis growth (Heller et al. 2022). Given the remarkable size and diversity of the phage population, additional phenotypic mining of phage genomes is sure to uncover many more genes capable of modulating bacterial behavior.

Here, we report the results of a genome-wide overexpression screen for Hammy, a temperate siphovirus that infects M. smegmatis. Based on gene content similarity, Hammy is classified as part of Cluster K (Subcluster K6) and is related to Waterfoul and several clinically relevant mycobacteriophages such as ZoeJ (Subcluster K2) (Russell and Hatfull 2016; Anders et al. 2017; Dedrick, Guerrero-Bustamante, Garlena, Russell, et al. 2019; Guerrero-Bustamante et al. 2021). The Hammy genome contains 95 predicted protein-coding genes, only ∼35% of which have annotated functions (Fig. 1) (Anders et al. 2017). Out of 94 Hammy genes evaluated in our systematic screen, a diverse set of 24 genes, representing ∼25% of the genome, were found to interfere with host growth; 12 of these diverse products caused near complete growth abolition. This study, done as part of the Science Education Alliance Gene-function Exploration by a Network of Emerging Scientists, or SEA-GENES project (Heller and Sivanathan 2022), contributes novel hits to the growing list of known antimycobacterial proteins while also providing insight into the conservation of host inhibition across related phage genomes.

Fig. 1.

The genome of phage Hammy. The Hammy genome is shown as a line with kbp markers and genes represented by boxes—those above the line are transcribed rightwards and those below are transcribed leftwards. Numbers inside the box correspond to gene numbers and predicted functions are indicated above or below each gene. Box shading corresponds to cytotoxicity scoring, with white boxes designating genes found to have no effect on M. smegmatis growth (cytotoxicity score 0), hatched box indicating omitted gene 19 not tested in this study, and blue representing observed toxicity in our assay. The saturation of blue boxes corresponds to the severity of growth inhibition using the following scores: light blue (score 1; reduction in colony size; genes 12, 29, 36, 53, 60, 61, 63, 67, and 84), medium blue (score 2; 1–3 log reduction in viability; genes 32, 90, and 92), and dark blue (score 3; >3-log reduction in viability; genes 9, 20, 34, 50, 51, 54, 56, 58, 68, 69, 77, and 78).

Materials and methods

Growth of mycobacteria and mycobacteriophage

M. smegmatis mc²155 was grown in Middlebrook 7H9 (Difco) broth supplemented with 10% AD (2% w/v dextrose, 145 mM NaCl, 5% w/v albumin fraction V), 0.05% Tween80, and 10 µg/ml cycloheximide (CHX) or on Middlebrook 7H10 (Difco) or 7H11 (Remel) agar supplemented with 10% AD, 10 µg/ml CHX, and Kanamycin 10 µg/ml (GoldBio) as needed for selection of pExTra plasmids. To transform M. smegmatis mc²155, electrocompetent cells were electroporated with ∼50–100 ng of pExTra plasmid DNA, recovered in 7H9 broth for 2 h at 37°C with shaking, and transformants selected on 7H10 or 7H11 agar supplemented with 10 µg/ml Kanamycin (GoldBio). After 4 days of incubation at 37°C, colonies of transformants were used directly in plate-based cytotoxicity assays or to inoculate cultures for liquid growth assays. Hammy was propagated on M. smegmatis mc²155 grown at 25°C in the presence of 1 mM CaCl₂ and no Tween in Middlebrook media and top agar.

Construction of the pExTra-Hammy library

Each Hammy gene was cloned into the pExTra shuttle vector (Heller et al. 2022) downstream of an anhydrotetracycline-inducible promoter, pTet (Ehrt et al. 2005), to control gene expression, and upstream of linked mcherry transcriptional reporter (Fig. 2a). Genes were PCR-amplified (New England Biolabs Q5 HotStart 2× Master Mix) from a high-titer Hammy lysate using a forward primer complementary to the first 15–25 bp of each gene sequence (Integrated DNA Technologies), introducing a uniform ATG start codon, and a reverse primer complementary to the last 15–25 bp of the gene sequence, including a uniform TGA stop codon (Supplementary Table 1). All forward primers contained a uniform, RBS-containing 5′ 21 bp sequence and all reverse primers contained a separate 5′ 25 bp sequence; these added sequences have identity to the pExTra plasmid flanking the site of insertion. Linearized pExTra plasmid was prepared via PCR (NEB Q5 HotStart 2× Master Mix) of pExTra01 (Heller et al. 2022) using divergent primers pExTra_F and pExTra_R and assembled with each gene insert by isothermal assembly (NEB HiFi 2× Master Mix). D29 10 and codon-optimized Kostya 33 inserts were synthesized with 5′ and 3′ flanking sequences (Integrated DNA Technologies) and assembled directly with linearized pExTra plasmid. Recombinant plasmids were recovered by transformation of Escherichia coli NEB5α F’I^Q (New England Biolabs) and selection on LB agar supplemented with 50 µg/ml Kanamycin. The pExTra-Waterfoul29 plasmid was described previously (Heller et al. 2022).

Fig. 2.

Expression of phage genes from the pExTra plasmid. a) Recombinant pExTra plasmids constructed in this study encode Hammy gene sequences downstream of the pTet promoter and upstream of mcherry. The two genes in this pExTra operon are transcriptionally linked, each bearing their own translational signals. b) Results of representative cytotoxicity assays are shown to demonstrate the range of observed growth defects. In each assay, colonies of M. smegmatis mc²155 transformed with the specified pExTra plasmid were resuspended, serially diluted, and spotted on 7H11 Kan media containing 0, 10, or 100 ng/ml aTc. Triplicate colonies (a, b, c) were tested for each gene alongside a positive control strain (+) transformed with pExTra02 (expressing wildtype Fruitloop 52) and a negative control strain (−) transformed with pExTra03 (expressing Fruitloop 52 I70S).

The inserted genes for all recovered pExTra plasmids were sequence-verified by Sanger sequencing (Azenta) using sequencing primers pExTra_universalR and pExTra_seqF; longer genes were also sequenced with internal sequencing primers listed in Supplementary Table 1. All plasmid inserts were found to match the published genome sequence.

Cytotoxicity screening and phenotype scoring

To assess cytotoxicity, pExTra-transformed colonies were resuspended and serially diluted in 7H9 broth then spotted on 7H10 or 7H11 plates supplemented with 10 µg/ml Kanamycin and 0, 10, or 100 ng/ml anhydrotetracycline (aTc; Alfa Aesar). Each strain was tested in triplicate alongside the pExTra02 positive control plasmid, encoding cytotoxic gene Fruitloop 52, and the pExTra03 negative control plasmid, encoding a nontoxic mutant allele of Fruitloop 52 (I70S) (Ko and Hatfull 2018; Heller et al. 2022). Phage Hammy was isolated at 25°C (Anders et al. 2017) and infection is restricted to temperatures <37°C (data not shown); however, in our screen, gene-mediated impacts on growth were monitored at 37°C, the temperature previously used to evaluate Waterfoul gene cytotoxicity (Heller et al. 2022) and determined to be optimal for growth of M. smegmatis strains transformed with pExTra-Hammy plasmids. At 37°C, spot growth was typically visible after 2 days, with effects on colony size and appearance more apparent after 3–5 days of incubation. Cytotoxic phenotypes were scored by comparing the spot dilution out to which cells grew in the presence vs the absence of aTc inducer and classified as either having no effect (score 0), being moderately cytotoxic with a 1–3 log reduction in cell viability (score 2), or being highly cytotoxic, causing complete or near complete (>3-log) inhibition of growth (score 3). Strains were also evaluated for aTc-dependent size reduction in individual colonies (score 1) as compared to the Fruitloop 52-I70S negative control strain on the same aTc plate and the same strain on plates without inducer. Pink colony color from mcherry gene expression provided a visual indicator of gene expression through the pTet operon.

All genes were tested in at least two independent cytotoxicity experiments; one round of screening was done on 7H10 agar and another on 7H11 agar, with good agreement observed between the different growth media. Unless otherwise indicated, the results and images reported in this study are all from experiments done on 7H11 agar. Reported cytotoxic genes were found to cause growth inhibition in more than two independent experiments with strong agreement between triplicate samples within each experiment. The magnitude of cytotoxicity for some genes was observed to vary slightly between experiments, likely due to minor variations in media or growth conditions (e.g. compare Hammy 12 results in Fig. 2b and Supplementary Figure 1); variability was more pronounced in those strains observed to have milder toxic effects, which are potentially at the detection limit of our assay. If differences in magnitude were observed between experiments, genes were scored based on the more conservative result. The inclusion of the Fruitloop 52 control strains on each screening plate aided in the evaluation of relative gene-mediated effects, and results were ultimately scored based on observations on 100 ng/ml aTc.

For liquid growth assays, transformants were grown in Middlebrook 7H9 broth with 10 µg/ml Kanamycin until saturated. Cultures were back diluted in fresh medium and induced with 100 ng/ml aTc at an O.D. 600 of ∼0.2–0.4; growth was monitored at 37°C with shaking.

Hammy genomic analysis

The Hammy genome map was created using the web-based tool Phamerator (phamerator.org). Reported gene functions are based on those available in the Hammy GenBank record (Accession KY087993). In several cases, these GenBank annotations were supplemented by comparison of functional assignments for other phamily members on PhagesDB (Russell and Hatfull 2016), especially those from more recently annotated Cluster K mycobacteriophage genomes; these designations were confirmed using HHPRED (PDB_mmCIF70, SCOPe70_2.08, Pfam-A_v35, NCBI_Conserved_Domains(CD)_v3.19) (Gabler et al. 2020), NPS Helix-Turn-Helix predictor (https://npsa-prabi.ibcp.fr/), and DeepTMHMM (Hallgren et al. 2022). Gene content comparison between Hammy and Waterfoul genomes was performed using the gene content comparison tool on phagesDB (https://phagesdb.org/genecontent/), with phamily designations downloaded from the database on February 27, 2023. Reported protein similarities were determined by multiple sequence alignment with Clustal Omega (Madeira et al. 2022).

Results and discussion

A system for studying Hammy gene overexpression in M. smegmatis

To systematically investigate the effects of Hammy gene overexpression on mycobacterial growth, an arrayed library of Hammy protein-coding genes was generated. Ninety-four genes were included in this analysis, excluding only gene 19, the long isoform of the tail assembly chaperone produced through programmed frameshifting; this analysis also excluded the single tRNA encoded in the Hammy genome. Each gene was cloned into the multicopy expression vector, pExTra, under the control of aTc-inducible promoter pTet and transcriptionally linked to fluorescent reporter gene mcherry (Fig. 2a; Heller et al. 2022). Sequence-verified plasmids were used to transform M. smegmatis mc²155, and transformants for all 94 strains were analyzed in a semiquantitative spot dilution assay to measure the impacts of overexpressing each gene on bacterial growth.

As illustrated in Fig. 2b, for each Hammy gene, dilution series for three transformed colonies were prepared and spotted on increasing concentrations of aTc inducer alongside control cells expressing a wildtype (pExTra02) or mutant (pExTra03) allele of phage Fruitloop gene 52. Previous work has established that overproduction of wildtype Fruitloop gp52 severely impairs growth of M. smegmatis, whereas comparable levels of the gp52-I70S variant cause no reduction in growth (Ko and Hatfull 2018). An aTc-dependent increase in pink coloration of spots was observed for the negative control pExTra03 strain and several experimental strains, confirming previous reports of tunable expression from pTet (Ehrt et al. 2005; Parikh et al. 2013; Heller et al. 2022). For Hammy genes 46 and 71, aTc-independent mcherry expression was observed (Supplementary Figure 1), suggesting that these sequences may harbor promoter sequences that can be recognized by the host transcriptional machinery.

Our screening strategy allowed for observation of a range of impacts on host growth, from mild reduction in colony size to multilog reduction in colony number, and growth effects were scored using an index from 0 (no effect on bacterial growth) to 3 (near complete abolition of host growth) (see Fig. 2b for representative examples). In our system, cytotoxicity was observed to increase for some genes in a dose-dependent manner on 10 and 100 ng/ml aTc (e.g. Hammy 29; Supplementary Figure 1), whereas for others, full cytotoxic effects were seen at the lower aTc concentration (e.g. Hammy 67; Supplementary Figure 1). Final cytotoxicity scores were assigned based on observations from the higher concentration of inducer.

Overexpression of most Hammy genes does not inhibit host growth

Seventy four percent of the 94 Hammy genes screened caused no appreciable reduction in bacterial growth in our assay (cytotoxicity score 0), with comparable viability and colony size in the presence or absence of aTc inducer (Fig. 2b; Supplementary Figure 1). This figure is in line with a previous systematic screen of genes encoded by related mycobacteriophage Waterfoul, 66% of which had no measurable impact on M. smegmatis growth (Fig. 3a; Heller et al. 2022), as well as a previous study by Ko and Hatfull (2020) that found that roughly 75% of a collection of genes with no known function (NKF) from assorted mycobacteriophages were not toxic when overexpressed in M. smegmatis. Out of the 70 Hammy genes found to be nontoxic in our assay, 60 displayed varying degrees of pink spot coloration on the highest concentration of inducer (Supplementary Figure 1). This visual confirmation of transcription and translation occurring through the pTet operon suggests that phage gene expression is occurring in most strains; however, we cannot rule out that some of these gene products may not accumulate within cells in our assay.

Fig. 3.

Comparison of Hammy and Waterfoul patterns of cytotoxicity. a) The proportion of Waterfoul (top) and Hammy (bottom) genes tested in this study or Heller et al. (2022) that were assigned each score (0–3) is represented as a stacked bar chart. Heller et al. scored two cytotoxic genes (Waterfoul 8 and 86) as 3* as recovery of pExTra trasformants was inhibited by presence of these gene inserts even in the absence of aTc. b) The proportion of Waterfoul (top) and Hammy (bottom) cytotoxic genes (score 1–3) that have NKF or that fall into various functional classes is represented as a stacked bar chart with different colors indicating functional class as described in the key. c) Shown is a chart listing the 45 out of 46 gene phamilies shared by Hammy and Waterfoul that were tested in both studies. Phamily number designations and functions are listed (NKF, no known function; TMD, transmembrane domain) next to representative homologous genes from Hammy and Waterfoul, with boxes shaded by cytotoxicity score. A binary indicator of whether homologous genes were both classified as toxic or nontoxic is illustrated by black or red shading, and % protein identity is indicated by the purple gradient boxes. d) For both Hammy (top) and Waterfoul (bottom), cytotoxicity scores (0–3) for each gene are plotted in genomic order along the horizontal axis, showing clustering of cytotoxic genes in the early and late lytic regions. e) Aligned maps of Hammy and Waterfoul provide a zoomed-in view of two genomic regions harboring clusters of cytotoxic genes, with boxes numbered by gene and shaded corresponding to the cytotoxicity scores (Heller et al. 2022) as in panels a and c. Gene phamilies found in both genomes are labeled by phamily numbers (taken from phagesDB.org as of February 27, 2023) with asterisks indicating those gene phamilies that were found to be cytotoxic (scores 1–3) or nontoxic consistently across both genomes.

A diverse set of Hammy gene products inhibit M. smegmatis growth

Out of 94 Hammy protein-coding sequences, 24 genes were found to reproducibly reduce growth of M. smegmatis in an aTc-dependent manner (Table 1, Fig. 1, and Supplementary Figure 1). Within this set, 9 genes caused effects characterized as mild, with notable reduction in colony size but no measurable decrease in plating (cytotoxicity score 1); the remaining 15 genes caused severe growth defects, with 3 causing a 1–2 log reduction in colony number (cytotoxicity score 2), and 12 resulting in >3-log reduction in colony number or almost complete inhibition of growth (cytotoxicity score 3) (Fig. 3a). These 24 cytotoxic gene products range in size from 84 amino acids to the largest gene product encoded in the genome, the 4,140 amino acid tape measure protein gp20 (Table 1).

Table 1.

Hammy genes observed to inhibit Mycobacterium smegmatis growth upon overexpression.

Gene	Phamilya	Length (aa)	Predicted function and features	Cytotoxicityb
12	68545	930	Major capsid protein	1
29	68638	1704	Lysin A	1
36	68706	420	NKF; HTH-DNA-binding	1
53	68569	84	WhiB family transcription factor	1
60	68645	289	Cas4 family exonuclease	1
61	3651	101	NKF	1
63	67008	213	NKF	1
67	60651	336	HNH endonuclease	1
84	69022	204	NKF	1
32	67117	339	NKFc; TMDd	2
90	14302	627	NKF	2
92	59346	699	NKF	2
9	68668	2574	NKF	3
20	67034	4140	Tape measure protein; TMD	3
34	65534	1134	MRE11 double-strand break endo/exonuclease	3
50	44663	165	NKF; TMD	3
51	850	258	NKF	3
54	68730	735	NKF; DUF3310	3
56	68998	468	NKF	3
58	67001	555	DnaQ-like (DNA polymerase III subunit)	3
68	607	2622	DNA primase/helicase	3
69	749	678	RusA-like resolvase	3
77	3303	390	NKF	3
78	596	834	NKF; TMD; HTH-DNA-binding	3

^a Phamily numbers were recorded from phagesdb.org as of February 27, 2023.

^b Growth of strains on media supplemented with 10 ng/ml or 100 ng/ml aTc was compared to the same strains plated on media without aTc. For those scored as 0, aTc-dependent reduction in host growth was not observed; those scored as 1 exhibit an aTc-dependent reduction in colony size, those denoted as 2 demonstrate a ∼1–3-log difference in growth in the presence of aTc, and those denoted as 3 demonstrate a severe >3-log reduction in growth in the presence of aTc.

^c NKF, no known function.

^d TMD, transmembrane domain.

Only a subset of these gene products have putative functional assignments or recognizable sequence features (Table 1; Figs. 1 and 3b). Two are structural proteins, with overproduction of the major capsid subunit gp12 causing mild toxicity (score 1) and overproduction of the tape measure protein gp20 causing a severe growth defect (score 3). Interestingly, overproduction of the Lysin A protein from Hammy (gp29) was found to cause a mild growth defect (Supplementary Figure 1). Efficient phage-mediated lysis of the mycobacterial cell is thought to require coordination between the phage holin, peptidoglycan-hydrolyzing Lysin A, and Lysin B esterase protein to disrupt the multilayered cell envelope (Catalão and Pimentel 2018). Expression of lysA genes by themselves is typically not toxic (Payne et al. 2009; Payne and Hatfull 2012; Catalão and Pimentel 2018; Heller et al. 2022); however, a small number of mycobacteriophage lysA gene phamilies have been reported to cause varying degrees of host cell death in a holin-independent manner when independently overexpressed from plasmids (Payne and Hatfull 2012).

Six gene products with predicted functions related to DNA replication or metabolism were observed to hinder host growth, including the predicted MRE11 double-strand break exo/endonuclease (gp34), Cas4 family exonuclease (gp60), RusA-like resolvase (gp69), one of two Hammy HNH endonucleases (gp67), and candidate replisome components DNA primase/helicase (gp68) and DnaQ (gp58). Several of these were among the most toxic gene products scored in our assay, with overproduction of the MRE11 double-strand break exo/endonuclease (gp34), DnaQ (gp58), DNA primase/helicase (gp68), and RusA (gp69) causing near complete abolition of host growth (Table 1; Supplementary Figure 1). Another highly cytotoxic gene product, NKF protein Hammy gp54, contains a conserved domain of unknown function (DUF3310), which was previously implicated in nucleotide kinase activity for the T7 coliphage protein gp1.7 (Tran et al. 2008, 2012, 2014).

Three genes with putative DNA-binding functions were observed to inhibit host growth (Table 1). This includes the WhiB transcription factor gp53 and predicted helix-turn-helix DNA-binding protein, gp36, both of which caused a reduction in colony size upon overproduction (score 1). The potent growth inhibitor gp78 (score 3) is predicted to harbor a C-terminal helix-turn-helix DNA-binding domain as well as four putative transmembrane helices toward the N-terminus of the protein. Two additional gene products (gp32 and gp50) are predicted by DeepTMHMM to contain transmembrane domains (TMD). Hammy gp32 contains a single predicted TMD at its N-terminus and is encoded immediately downstream of the candidate holin protein, gp31, which has four predicted TMDs. Phage lysis cassettes often encode multiple holin-like TMD proteins (Hatfull 2018; Pollenz et al. 2022), and gp32 may thus play a role in disruption of the cell envelope. The small, 54-amino acid protein gp50 is predicted to have a 20-amino acid TMD at its C-terminus such that a short cytoplasmic domain at the N-terminus would be anchored in the cell membrane. Predicted membrane proteins, notorious for impairing cell growth when produced at high levels in bacteria (Miroux and Walker 1996; Wagner et al. 2006), were not uniformly toxic in our cytotoxicity screen. In total, Hammy encodes seven proteins with predicted TMDs. Overproduction of three of these (gp31, gp35, gp44) had no detectable effect on host growth in our assay, though we note that obvious pink coloration of spots was only observed for one of these (gp44; Supplementary Figure 1).

Overall, our systematic screen revealed that one-quarter of Hammy genes are deleterious to M. smegmatis growth when overexpressed in our system, representing a set of mycobacterial growth inhibitors that are diverse in terms of their size, sequence features, and magnitude of effect. As with any overexpression screen, it is possible that some of the observed growth defects, especially those milder defects seen only at the higher level of induction, may be a consequence of artificially high protein levels. It is encouraging though that half of the hits described here caused near complete abolition of host growth, with severe growth defects observed even on the lower concentration of aTc (Supplementary Figure 1). These products offer promising candidates for further elucidation of the interactions between phage factors and key cellular complexes.

Conservation of mycobacterial growth inhibition by related phage genes

To facilitate gene content comparisons across highly mosaic phage genomes, gene products encoded within sequenced actinobacteriophage genomes are grouped into gene phamilies based on amino acid similarity (Cresawn et al. 2011; Gauthier et al. 2022), and each gene in the Hammy genome belongs to a designated phamily with known homologues in other phages (Russell and Hatfull 2016). A systematic screen using the same pExTra overexpression system, plate-based cytotoxicity assay, and scoring index was recently reported for related Cluster K mycobacteriophage Waterfoul (Heller et al. 2022), enabling a comparison of phenotypes conferred by shared gene phamilies encoded by the two related phages. Waterfoul and Hammy encode 46 phamilies in common (Cresawn et al. 2011; Russell and Hatfull 2016), and 34 of these were consistently classified as toxic or not toxic in this study and the Waterfoul report (Fig. 3c). Nine shared gene phamilies were observed to inhibit mycobacterial growth in both studies, including three highly cytotoxic phamilies encoded in both genomes represented by Hammy NKF genes 9, 54, and 78 (Fig. 3c). Consistent cytotoxic effects by homologous gene products further support these phamilies as strong candidates for factors targeting essential mycobacterial processes, and in general, functional comparisons offer opportunities to probe the conserved sequence determinants of host interference.

Most of the phenotypic discrepancies between shared Hammy and Waterfoul phamilies are for genes causing only reduction in colony size (score 1) in one of the two reports (Fig. 3b), suggesting that these milder growth defects may be at the detection limit of our assay. However, a more notable discrepancy in phenotype was observed for three shared gene phamilies: Hammy genes 22, 52, and 88 were found to have no impact on host growth when overexpressed, whereas the homologous genes from Waterfoul (genes 21, 46, and 88) caused severe reduction in colony number (score 2 or 3) upon overexpression (Fig. 3c). It remains to be seen whether these discrepancies are due to variation in biological function or are a result of expression differences in our system. For example, major tail protein Hammy gp21 is 78% identical to cytotoxic protein Waterfoul gp22, but spots overproducing Hammy gp21 displayed no pink coloration on aTc, suggesting that the lack of toxic phenotype may be due to insufficient protein levels (Supplementary Figure 1).

Interestingly, consistent toxicity was observed for two conserved structural components—the related major capsid proteins Hammy gp11 and Waterfoul gp12 (80% identical) and TMD-containing tape measure proteins Hammy gp20 and Waterfoul gp19 (58% identical). Virion structures are typically thought to function outside or at the surface of the cell making their cytoplasmic cytotoxicity a somewhat surprising result; however, these structural components are highly expressed in the cell during lytic infection (Dedrick et al. 2013, Dedrick, Guerrero Bustamante, Garlena, Pinches, et al. 2019), and recent work has demonstrated their potential to trigger diverse cellular responses (Pedulla et al. 2003; Zhang et al. 2022; Huiting et al. 2023; Stokar-Avihail et al. 2023). Thus, further examination of the cytotoxicity conferred by these structural genes as well as impacts caused by other mycobacteriophage structures is warranted.

The mild impact on growth caused by WhiB protein Hammy gp53 is consistent with a previous finding that related phamily member, TM4 gp49 (45% identical), disrupts cell division and inhibits growth when overproduced in M. smegmatis (Rybniker et al. 2010). Rybniker et al. (2010) proposed that the TM4 WhiB impaired growth by downregulating expression of the essential host transcription factor whiB2 through competitive binding of the whiB2 promoter, and the Hammy WhiB may hinder host growth through a similar mechanism.

Phenotypic comparisons can also be made between genes that belong to distinct gene phamilies but are predicted to carry out common functions. Different phamilies with putative roles in Hammy or Waterfoul genome replication were identified as cytotoxic, including the Waterfoul DNA primase/polymerase gp65, the Waterfoul DNA helicase protein gp66, the Hammy DNA primase/helicase gp68, as well as predicted sliding clamp subunit Waterfoul gp48, and DnaQ protein Hammy gp58 (Fig. 3b–e). We note that the Waterfoul DnaQ protein gp54, which is categorized in a separate phamily from Hammy DnaQ gp58, was not observed to be toxic, and thus, cytotoxicity may be specific to the Hammy gp58 phamily rather than a general consequence of overproduction of phage DnaQ proteins. Here too, additional investigation is needed to understand whether cytotoxicity by phage replication proteins is due to interference with the host DNA or replication machinery or because of some other mechanism. It is worth noting that in recent explorations of phage–host dynamics, phage-encoded replication proteins have also been observed to elicit diverse cellular defense responses (Stokar-Avihail et al. 2023).

Hammy and Waterfoul vary in their lysis cassettes, with divergent lysA genes adjacent to lysB and holin-encoding genes from common gene phamilies (Fig. 3e). Overproduction of related Lysin B (75% identical) and holin (78% identical) proteins from Wateroul and Hammy had little to no effect on host growth (Fig. 3b), whereas the distinct lysA genes had differing impacts on host growth (Fig. 4a). Hammy lysA overexpression was mildly toxic, causing a reduction in colony size, whereas Waterfoul lysA overexpression had no visible impact on host growth. Most Lysin A proteins tested to date, including the Hammy Lysin A homolog in phage Bxz2 (48% amino acid identity) do not cause cell lysis independent of holin activity; however, a small number of mycobacteriophage lysA genes have previously been shown to cause holin-independent lysis (Payne and Hatfull 2012). Using an M. smegmatis liquid growth and ATP release assay, Payne and Hatfull reported that overproduction of the D29 Lysin A protein gp10 caused pronounced ATP release and cell lysis, whereas overproduction of other Lysin A proteins from other gene phamilies, including Waterfoul gp29 homologue, Kostya gp33, had only mild effects.

Fig. 4.

Overproduction of multidomain mycobacteriophage lysin a proteins. a) lysA genes from various mycobacteriophages were expressed from pExTra and tested in the plate-based cytotoxicity assay alongside pExTra02 (+) and pExTra03 (−) controls. b) The same strains were evaluated in an endpoint liquid growth assay. Two colonies of each transformed M. smegmatis strain were grown until saturation and subcultured in duplicate 24-well plates. Upon reaching an OD600 of ∼0.2, one set of cultures was induced by addition of 100 ng/ml aTc and the other left as an uninduced control. Growth was monitored over 24 h of induction at 37 °C with shaking. Plate images from a representative experiment showing lysis and culture appearance after 24 h with (right image) or without induction (left image) are shown. c) The tested Lysin A proteins harbor various enzymatic domains as predicted by HHPRED and as previously described (Payne and Hatfull 2012), including two putative peptidase domains, P1 (1CV8_A; probability 99.2%) and P2 (3NPF_B; probability >88%), a G19-like glycosyl hydrolase domain (5H7T_A; probability >92%), and a G25-like glycosyl hydrolase domain (cd06419; probability >99%). Relationships between these Lysin A proteins are represented by the phylogenetic tree (left; Clustal Omega and Splitstree (Huson and Bryant 2006)) and % amino acid identity values on the right.

To directly compare the impacts of Hammy lysA overexpression on M. smegmatis growth to those of previously studied mycobacteriophage lysA genes, we tested several lysA genes in both our plate-based cytotoxicity assay and a liquid endpoint growth assay. Consistent with the holin-independent lysis results reported by Payne and Hatfull (2012), D29 lysA overexpression completely inhibited growth in both assays (Fig. 4a and b), with signs of cell lysis, including cleared culture and visible cell debris, observed in liquid after 24 h induction. The effect mediated by D29 lysA expression was much more pronounced than that observed for the Hammy or Kostya lysA genes, both of which only caused an aTc-dependent reduction in colony size on plates (Fig. 4a) but no notable signs of cell death in liquid (Fig. 4b). Pink coloration of cultures expressing the Hammy lysA in the presence of inducer suggests that expression through the pTet operon is occurring in cells. Interestingly, even in the absence of aTc, cells transformed with the pExTra-Kostya33 plasmid turned pink and formed smaller colonies, suggesting that this lysA sequence may harbor an internal gene start (Catalão et al. 2011). As seen previously, despite being 66% identical to the Kostya Lysin A protein, overproduction of the Waterfoul Lysin A had no effect on host growth on plates (Fig. 4a) or in liquid (Fig. 4b), though only faint pink color was observed in the presence of aTc, meaning that the Waterfoul lysA may be poorly expressed in our system. These data imply that although mildly toxic when produced independently on plates, the Hammy Lysin A is not mediating pronounced holin-independent cell lysis in our system like is seen with the D29 Lysin A. Mycobacteriophage Lysin A proteins are multidomain proteins, with putative peptidase and glycoside hydrolase functions (Fig. 4c; Payne and Hatfull 2012). The additional observation of holin-independent cytotoxicity presented here may prove useful in genetically dissecting the determinants of these lytic and nonlytic phenotypes.

Finally, 12 of the 24 growth inhibitors identified in our Hammy screen (Hammy 32, 50, 51, 56, 58, 61, 68, 69, 77, 84, 90, 92) represent gene phamilies for which, to the best of our knowledge, cytotoxic effects have not been previously reported, thus also adding many novel hits to the set of known phage-encoded growth inhibitors.

Insights into genomic patterns of mycobacterial growth inhibition

Beyond phenotypic comparisons of shared gene phamilies, our Hammy results also lend support for several genome-wide patterns first observed for Waterfoul (Heller et al. 2022). Cytotoxic genes are abundant within phage genomes: one-quarter of Hammy genes and one-third of Waterfoul genes were found to inhibit growth of the host when overexpressed, with a similar fraction of genes in each genome capable of causing severe growth defects in the host (Fig. 3a). Both NKF genes and genes with predicted functions inhibit host growth (Fig. 3b), and while it is unlikely that all are employed during infection to kill the host cell, this high frequency of host interference underscores the complexities of host–phage dynamics. Indeed, recent work on the diverse interactions that occur between phages infecting a common host, suggests that some of these cytotoxic phage products may bind cellular factors with the purpose of blocking secondary infection by other phages (Dedrick et al. 2017; Ko and Hatfull 2018; Gentile et al. 2019; Montgomery et al. 2019).

We also find that cytotoxic genes are distributed throughout both genomes and are often found near other cytotoxic genes (Fig. 3d). Transcriptomic data from related Cluster K mycobacteriophage ZoeJ suggests that many of these cytotoxic genes are likely co-expressed over the course of the Hammy or Waterfoul life cycle (Dedrick, Guerrero Bustamante, Garlena, Pinches, et al. 2019), and it is plausible that cytotoxic products encoded by adjacent genes target related host processes. The predicted early lytic regions of both genomes are enriched with cytotoxic genes, including conserved DUF3310 genes as well as various genes predicted to function in transcription, DNA replication, and DNA metabolism (Fig. 3e). As some of the first gene products made during infection, these are good candidates for factors involved in the coordinated takeover of critical cellular processes. A second cytotoxic gene cluster is found between the holin and integrase genes in both Hammy and Waterfoul (Fig. 3e), with cytotoxic genes encoding poorly conserved TMD proteins Hammy and Waterfoul gp32 (19% amino acid identity) flanked by conserved holin genes and cytotoxic nucleases, Hammy and Waterfoul gp34. Phage lysis cassettes are highly mosaic (Hatfull 2018; Pollenz et al. 2022), and phenotypic profiling such as that described here may identify convergent functions within these cassettes and help to better characterize the mycobacteriophage lysis pathway.

In conclusion, only a small fraction of the thousands of known mycobacteriophage gene phamilies have been experimentally explored to date (Rybniker et al. 2008, 2010, 2011; Mehla et al. 2017; Ko and Hatfull 2018, 2020; Heller et al. 2022). Here, we show that systematic screening of phage genomes in a simple, plate-based overexpression assay can identify a diverse set of bacterial effectors, offering many promising candidates for elucidation of novel interactions between phages and their mycobacterial hosts. Extension of this screening strategy to additional diverse mycobacteriophage genomes will undoubtedly uncover many more novel effectors and provide additional insight into the patterns of phenotypic conservation examined here.

Data availability

All plasmids and plasmid sequences reported in this study are available upon request. The authors affirm that all data necessary for confirming the conclusions of this article are represented fully within the article and its tables and figures.

Supplemental material available at G3 online.

Funding

This work was conducted as part of the Science Education Alliance GENES (Gene function Exploration by a Network of Emerging Scientists) project supported by the Howard Hughes Medical Institute.