Bacillus subtilis Phages Related to SIOphi from Desert Soils of the Southwest United States

Abstract

Background:

Phages impact Bacillus subtilis microbial dynamics, but our understanding of the diversity of phages that can lyse this model organism is limited.

Materials and Methods:

Phages were isolated from soil collected from two Southwest United States desert sites on wild B. subtilis strains. Phage genomes were assembled and bioinformatically characterized, and phage host range was assessed.

Results:

Four myoviruses with high nucleotide and amino acid similarity to each other (>97%) and to three phages in GenBank (>82%), including SIOphi, were isolated. These phages have double-stranded DNA genomes (153,882–156,577 bp), low GC content, and 256–270 putative protein coding genes (28–30% with predicted functions). Comparative genomics revealed differences in putative lysis and replication genes.

Conclusions:

Comparative approaches provide insight into phage evolution, identifying unique genes shared within phage clusters. Better characterization of Bacillus phages will aid in linking genetic differences among phages (e.g., lysin genes) to phenotype (e.g., host range).

Introduction

Bacteriophages (phages) are abundant in all ecological communities and are important drivers of ecological and evolutionary processes.^1–8 Phages are diverse, with a wide range of genome sizes and a high frequency of novel genes due to the diversity of their bacterial hosts, as well as the speed of bacteria and phage evolution.^2,9,10 Despite the importance of Bacillus species to microbial communities,^11–13 their impact on plant community structure,¹⁴ and the role of Bacillus subtilis as a model organism,^9,15,16 only 408 sequenced phages (as deposited in NCBI in July 2023) have been isolated on various species of this genus. This is a small number in contrast to the more than 2000 sequenced phages isolated on Mycobacterium smegmatis, largely due to the efforts of the SEA-PHAGES program.¹⁷ Our research is focused on describing phages specifically isolated on wild strains of B. subtilis. Here, we describe four such phages that form a distinct subcluster, sharing high genetic similarity to only three previously described phages.⁹

Materials and Methods

Approximately 1 g of soil from samples collected either at Tumamoc Hill in Tucson, AZ, or Death Valley National Park (Table 1) was added to 20 mL LB broth and incubated at 37°C, shaking at 250 rpm for 4 h, then filtered (0.22 μm). Filtrates were plated either on B. subtilis wild strain T89-30 or TG115, strains isolated by Istock et al.^18,19 and kept as spore stocks in our lab. Two individual plaques were isolated from each soil sample and were single-plaque purified three times on lawns made from spores of the isolation host. High titer lysates were prepared by flooding with LB broth multiple webbed plates. Lysates were filtered (0.22 μm) and DNA was extracted using the Promega Wizard DNA Clean-Up System. A high-titer lysate of one of our isolated phages was used for visualization on a Philips CM100 transmission electron microscope at an accelerating voltage of 100 kV by staining with 2% phosphotungstic acid.

Table 1.

Basis Genomic Characteristics of Four Novel Bacillus subtilis Phages and Their Close Relatives

Phage	Isolation strain	Genome size (bp)	DTR (bp)	%GC	No. of ORFs	Collecting site and date	GPS coordinates	Accession number
000TH008	T89-30	155,246	3014	38.6	266	Tumamoc Hill, AZ, September 2011	32.2180N 111.0036W	MW419084
000TH009	TG115	155,191	3012	38.6	266	Tumamoc Hill, AZ, September 2011	32.2180N 111.0036W	MW419085
015DV002	TG115	153,882	2802	38.7	256	Death Valley NP, CA, May 2014	36.1485N 116.7702W	MW419086
015DV004	TG115	156,577	3080	38.6	270	Death Valley NP, CA, May 2014	36.1485N 116.7702W	MW419087
SIOphi	Bacillus subtilis 53	146,698	N/A	39.0	206	N/A	N/A	KC699836
vB_BspH_Mawwa	Bacillus sp.54	149,014	N/A	39.0	230	Springville, UT	N/A	MW749002
vB_BspH_TimeGriffin	Bacillus sp.54	148,525	N/A	39.1	235	Provo, UT	N/A	MW749007

Genome size includes the DTR, whose length is also given. ORF for either known or putative gene; NP; N/A based on information associated with accession for these phages. Samples from Death Valley were collected under the permit DEVA-2014-SCI-0026. In text, we refer to the two phages isolated from Death Valley as DV phages and the two phages isolated from Tumamoc Hill as TH phages.

DTR, direct terminal repeat; N/A, not available; NP, National Park; ORF, open reading frame.

DNA samples were sent to North Carolina State University's Genomic Science Laboratory for sequencing on Illumina MiSeq platform (v3 150 SE flow cell) from libraries prepared with Illumina Truseq Nano DNA library prep kit. We aligned and assembled reads using GS de novo assembler v2.9²⁰ and verified quality with Consed v29.²¹ Each genome assembled as a single contig with >1000 × coverage and consensus sequences were of high quality with no additional sequencing necessary. Genome ends were determined with PhageTerm.²²

Finished sequences were imported into DNAMaster v5.22.22²³ for annotation. Putative genes were called based on both Glimmer v3.0 and GeneMark v2.5 algorithms.^24,25 Putative protein functions were predicted using BLAST v2.12²⁶ (if E-value below 10⁻⁵) and HHpred (if probability <85%, coverage >50%, and E-value <10⁻⁵).²⁷ The absence of tRNA genes was confirmed with Aragorn.²⁸ Phage lifecycle was confirmed by analyzing annotations with PhageLeads²⁹ and BACPHLIP.³⁰ Default settings were used in all programs. All genome annotations have been submitted to NCBI (accession numbers in Table 1).

A Phamerator database³¹ was constructed with 312 Bacillus phages with complete sequences available on NCBI as of May 1, 2022 (Bacillus v5; https://phamerator.org).^32,33 This database groups phages into clusters and subclusters and was used to generate genome maps illustrating relationships between phages based on both amino acid (aa) and nucleotide similarity. As part of the Phamerator clustering protocol, gene products are first grouped into phage protein families, or “phams,” using the PhaMMseqs pipeline; then phages that share 35% or more of their phams are classified as belonging to the same cluster.^32,33

Additional genomic comparisons were done by (1) calculating Average Nucleotide Identity (ANI) and Average Amino acid Identity (AAI) using algorithms from the enve-omics lab,³⁴ (2) generating genome dotplots with Gepard,³⁵ (3) conducting BLASTn (nr/nt database) or BLASTp (nr database) searches (performed July 2023), either against the complete database or the database for tailed phages (taxid: 10699, 10662 and 10744), using default parameters, or (4) aligning aa sequences of homologous genes using ClustalW within MEGA11.³⁶ Alignment figures were made with ESPript3.0.³⁷

We tested our phages' ability to lyse seven B. subtilis strains: two lab strains obtained from the Bacillus Genetic Stock Center—the model strains 168 (BGSCID 1A1) and W23SR (BGSCID 2A3)—and five wild strains in our lab collection originally from Istock et al.^18,19 A microtiter plate-based assay³⁸ was set up with initial titered concentrations of 90.91 × 10⁶ PFU/mL and 90.91 × 10⁷ CFU/mL for a multiplicity of infection of 1:10. Two hundred twenty microliters of LB2 broth with bacteria and phage was added to wells of a 96-well plate with each phage-bacteria combination replicated five times. Wells with only bacteria served as controls. Plates were incubated for 20 h at 37°C auto mixing for 10 s every 5 min with absorbance reading at 600 nm every 30 min. For a conservative estimate of host range, we scored a phage as impacting bacterial growth if at least four out of five growth curves for bacteria with phage were below all the growth curves for bacteria alone.

Results

Basic genomic characteristics

Four B. subtilis phages were isolated from soil samples from Tumamoc Hill in Tucson, AZ, or Death Valley National Park (Table 1). These phages were isolated on wild bacterial strains of B. subtilis. Their double-stranded DNA genomes ranged from 153,882 to 156,577 bp, including direct terminal repeat (DTR) ends. They have 256–270 putative protein coding genes, no tRNA genes, and a GC content (38.6–38.7%) lower than typical of their B. subtilis host, which is usually around 43%.³⁹ These phages produced clear plaques on their isolation host, suggesting they are lytic, a conclusion supported by annotations of their genomes and by genomic analyses using the programs PhageLeads²⁹ and BACPHLIP.³⁰

As is typical of myoviruses, these phages have a long contractile tail (Fig. 1) and a genome greater than 125 Kb.^2,40 Gene order in myoviruses is less conserved than in siphoviruses or podoviruses; however, two groupings of structural genes can still be identified (Fig. 2; Supplementary Fig. S1).^40,41 The first includes the terminase, portal, and head proteins, while the second encodes the baseplate and tail proteins.⁴⁰ Overall, a function could be assigned to less than 30% of the putative genes (Fig. 2; Supplementary Table S1). These phages show high nucleotide and amino acid similarity to each other—greater than 97% for all pairwise comparisons (Table 2; Supplementary Fig. S2).

FIG. 1.

Representative brightfield TEM image of 015DV002 stained with 2% phosphotungstic acid. Image was taken at 64,000 × magnification using a Philips CM100 TEM. TEM, transmission electron microscope.

FIG. 2.

Genome map of 015DV004, chosen because it is the longest genome with the most genes among our four W7 phages. The ruler shows genome length (in kilobases) with the forward genes shown above the ruler. Function or putative functions are listed above genes for genes on the forward strand and below genes for those genes on the reverse strand (see also Supplementary Table S1). Map was created using Phamerator.³¹

Table 2.

Average Nucleotide Identity (Above the Diagonal) and Average Amino Acid Identity (Below the Diagonal) for Our Seven Novel Phages and SSP1, with Each Phage Compared to the Other Phages in Its Cluster

	000TH008	000TH009	015DV002	015DV004	SIOphi	vB_BspH_Mawwa	vB_BspH_TimeGriffin
000TH008	—	100	98	98	87	87	87
000TH009	100	—	98	98	87	87	87
015DV002	97	97	—	98	87	87	87
015DV004	97	97	98	—	87	87	87
SIOphi	82	82	82	82	—	97	98
vB_BspH_Mawwa	82	83	83	82	95	—	98
vB_BspH_TimeGriffin	82	83	83	83	96	97	—

Data presented as % and calculated using algorithms from Rodriguez-R and Konstantinidis.³⁴

These phages belong to subcluster W7

Based on the best BLASTn searches (see below), as well as BLASTp searches that were conducted during our annotations, we identified only Bacillus phages as the best matches with respect to genome or protein similarity; therefore, our subsequent analyses focused only on Bacillus phages. A BLASTn search using 015DV002 identified three Bacillus phages with significant nucleotide similarity to our phages: SIOphi (query coverage/identity 71%/91.3%), vB_BspH_Mawwa (74%/88.1%), and vB_BspH_TimeGriffin (71%/88.4%). The next closest match is to Bacillus phage SBSphiJ5 with only 11% query. Based on Phamerator clustering,^31,32 these seven phages belong to the W7 subcluster (referred to as J2 in an older clustering analysis with a database of 83 phages⁹). These W7 phages show at least 87% nucleotide or 82% aa similarity for all pairwise comparisons (Table 2; Supplementary Fig. S2).

Based on pham assignment, there are 21 genes that are unique to these seven phages (i.e., not found in any other Bacillus phages included in our Phamerator database). All these genes are small (range 129–505 bp) and could not be assigned a putative function. In addition, there are 28 genes only found in our W7 phages, 6 genes only found in the 2 TH phages, and 9 genes only found in the 2 DV phages. The majority of these genes also have no known function.

Comparative genomics of our W7 phages

The genes with functions that are conserved across the four W7 phages are genes involved with DNA metabolism, cell lysis, and virion structure, which is expected as these are well-conserved genes in Bacillus phages.⁹ However, there are some differences worth noting, particularly for genes involved in phage attachment to their host and host lysis.

As is typical of phage lysins, the N-terminus is the enzymatically active domain, while the C-terminus contains the cell wall binding domain.⁴² For the lysin of our four W7 phages, the C-terminus shows high amino acid conservation and contains a domain of unknown function (DUF3597; pfam12200; Supplementary Fig. S3). In contrast, the N-terminus differs between 000TH008 (N-acetylmuramoyl-l-alanine amidase domain) and the other three phages (000TH009, 015DV002, and 015DV004; l-alanyl-d-glutamate peptidase domain), resulting in the lysins being classified in different phams. This suggests that these two lysins may differ in how they cleave the peptides that link the polysaccharide layers of the peptidoglycan cell wall.⁴³ The lysin in 000TH008 belongs to a pham found in 21 other phages in our database, all members of the W cluster, including SIOphi, vB_BspH_Mawwa, and vB_BspH_TimeGriffin, while the other lysin is only found in our other three phages. The l-alanyl-d-glutamate peptidase domain belongs to the M15 metallopeptidase family and, outside of our database, the best DELTA-BLASTp (Domain Enhanced Lookup Time Accelerated BLAST) match is to a metallopeptidase in Edaphobacillus lindanitolerans (query coverage 95%; identity 41.6%, E-value = 1 × e⁻⁹²).

Another genomic region of expected conservation is the section that encodes the baseplate and tail proteins.⁴⁰ These genes must coevolve with each other to maintain necessary interactions.⁴⁴ This high conservation is observed across the tail proteins of our W7 phages, except for one tail protein in 015DV004 (gp 178; Supplementary Fig. S4). Amino acid and nucleotide alignments of the tail protein homologs (Supplementary Fig. S4) reveal three regions: an N-terminus with high amino acid conservation not reflected in the nucleotide sequence, a middle region with both high amino acid and nucleotide conservation, and a C-terminus with less aa and nucleotide conservation. The N-terminus and middle region of higher conservation corresponds to a siphovirus protein of unknown function (DUF859, pfam 05895).

All four W7 phages encode the alpha and beta subunits of a ribonucleoside-diphosphate reductase (RNR), which has been observed to be the most highly conserved Bacillus gene product.⁹ However, in 015DV002, the alpha subunit is split into two genes (genes 131 and 133) by an HNH endonuclease (gene 132) (Supplementary Fig. S1; Supplementary Table S1). This phage also possesses an additional homing endonuclease (gene 136) that is not found in our other three W7 phages. Homing endonucleases are typically found as stand-alone genes in phage genomes but can also be inserted as self-splicing introns or inteins into coding genes.^45,46 In particular, in Aeromonas phage Aeh1, the alpha subunit (nrdA) is also interrupted by a HNH family homing endonuclease which did not inactivate RNR function; the spliced alpha subunits could still assemble with each other and with the beta subunit.⁴⁶

A number of genes with specific functions are only called in 000TH008 and 000TH009. These include MqsA-like antitoxin, rubredoxin reductase, HNH restriction endonuclease, and uracil DNA glycosylase. In contrast, a gene that encodes a glycosyl transferase is only called in the two DV phages. These differences are interesting and remaining to be explained. In particular, the MqsA-like antitoxin is found in bacteria and prophages and is part of a toxin/antitoxin (TA) system that provides stability to plasmids and plays a role in biofilm formation and phage inhibition.⁴⁷ Thus, this gene may be useful in competition with other phages and was likely incorporated in the 000TH phages by horizontal gene transfer.

Repeated sequences in the W7 phages

Based on Phamerator maps, there are many repeated sequences between these genomes (Supplementary Fig. S1). These repeated sequences appear primarily in the DTR regions and the early region of the genome (first 50 genes). Most of these repeats appear to be intergenic, though a few cases of intragenomic repeats appear in genes that encode the minor capsid protein (gene 165 in 015DV004), a hypothetical protein (gene 196 015DV004), and the lysin mentioned above. Intergenic repeat sequences have been observed in Gordonia phages of cluster DJ (Personal communication; Pollenz, 2023), but little is known about their function.

Host range

We tested these phages against seven B. subtilis strains in our collection, including lab strain 168 and W23 (Supplementary Table S2). The DV phages had limited ability to lyse 168, while the TH phages could not lyse this strain; none of these phages could lyse strain W23. In contrast, all four phages could lyse all of the wild strains that we tested them against to varying degrees. For example, the impact of 015DV004 on the growth of T89-05 was smaller compared to the other three phages which showed more pronounced effects on bacteria growth (Supplementary Table S2). Notably, this is the phage with the less-conserved tail protein (gp 178) discussed above.

Discussion

Myoviruses typically have large genomes (>125 Kb) and have been studied for their ability to use multiple infection and lysis strategies for replication.⁴⁸ The phages described here represent a novel subcluster of Bacillus phages with high nucleotide and amino acid similarity among themselves, and with the genomes of three other Bacillus phages. The number of unique genes without an identifiable putative function in these phage is consistent with previous studies, as dsDNA phage genomes are known to be a reservoir of putative coding genes of unknown function.⁴⁹ These unique genes of unknown function help to delineate clusters and subclusters of phages, potentially elucidating evolutionary boundaries.⁹ It is also possible that discovery and description of additional phages may blur these boundaries.⁵⁰ In addition, comparisons between multiple related phages help us better understand the host networks through which horizontal gene transfer occurs. Our study points to two such possibilities: the acquisition of the lysin l-alanyl-d-glutamate peptidase domain and the MqsA-like antitoxin in these phages.

A major impetus of phage genomic description is to link genetic differences among phages (e.g., lysin genes) to phenotype (e.g., host range). This is not an easy task. First, all methods for assessing host range have limitations.⁵¹ For example, plaques observed with spot testing can be caused by lysis from without. Second, we need a better understanding of the diversity of bacterial hosts that phages interact with in soil communities, so that phages can be tested against relevant bacterial hosts rather than lab strains they have not evolved with. Our four phages could either not lyse lab strains 168 and W23 or had limited ability to do so. Lab-propagated bacterial strains are known to undergo genome streamlining, and it is therefore possible that the bacterial genes required for productive phage infection were lost, as such genes may not be adaptive in a homogenous lab environment.⁵² Finally, phage research needs to further develop quantitative measures, in addition to qualitative (yes/no) measures, of phenotypic traits such as infectivity or host range. These will be necessary to understand when and how differences among gene complement and gene sequence affect phage fitness.

Authors' Contributions

Conceptualization, V.A.D. and G.P.K.; Methodology, V.A.D., G.P.K.; Validation, V.A.D., G.P.K., and L.H.M.; Formal analysis, V.A.D., G.P.K., and L.H.M.; Investigation, V.A.D., G.P.K., L.H.M., A.C.V., B.E.C., K.B.L., M.S.S., A.A.G., and J.M.D.; Resources, V.A.D. and G.P.K.; Data curation, V.A.D., G.P.K., L.H.M; A.C.V., B.E.C., K.B.L., M.S.S., A.A.G., and J.M.D.; Writing—original draft preparation, V.A.D., G.P.K., and L.H.M.; Writing—review and editing, V.A.D., G.P.K., L.H.M., A.C.V., M.S.S., B.E.C., J.M.D., and A.A.G.; Visualization, V.A.D., G.P.K., and L.H.M.; supervision, V.A.D. and G.P.K.; Project administration, V.A.D. and G.P.K; Funding acquisition, V.A.D. and G.P.K.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This research was supported by a grant to Gettysburg College from the Howard Hughes Medical Institute through the Precollege and Undergraduate Science Education Program (Grant No. 52007540) and by Research and Professional Development grants from Gettysburg College to V.A.D. and G.P.K.