Pheno- and Genotyping of Three Novel Bacteriophage Genera That Target a Wheat Phyllosphere Sphingomonas Genus

Bacteriophages are viral agents that infect and replicate within bacterial cells. Despite the increasing importance of phage ecology, environmental phages—particularly those targeting phyllosphere-associated bacteria—remain underexplored, and current genomic databases lack high-quality phage genome sequences linked to specific environmentally important bacteria, such as the ubiquitous sphingomonads. Here, we isolated three novel phages from a Danish wastewater treatment facility. Notably, these phages are among the first discovered to target and regulate a Sphingomonas genus within the wheat phyllosphere microbiome. Two of the phages displayed a non-prolate Siphovirus morphotype and demonstrated a narrow host range when tested against additional Sphingomonas strains. Intergenomic studies revealed limited nucleotide sequence similarity within the isolated phage genomes and to publicly available metagenome data of their closest relatives. Particularly intriguing was the limited homology observed between the DNA polymerase encoding genes of the isolated phages and their closest relatives. Based on these findings, we propose three newly identified genera of viruses: Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi, following the latest ICTV binomial nomenclature for virus species. These results contribute to our current understanding of phage genetic diversity in natural environments and hold promising implications for phage applications in phyllosphere microbiome manipulation strategies.


Introduction
To date, bacteriophages (commonly also known as "phages") are recognized as the most abundant and diverse biological entities in the biosphere [1]. These prokaryotic viruses exist in all known habitats that support bacterial growth, including terrestrial and aquatic environments [2,3]. Viral communities play a significant ecological role. With the unique ability of phages to infect and replicate within their host, typically with a narrow host range targeting only specific bacteria, they become important players in modulating the diversity of natural bacterial communities, thus having a notable impact on bacterial ecology and population dynamics [4,5].
The distinct host specificity of bacterial viruses makes the practical application of phages as biocontrol and biomanipulation agents highly attractive. Food industrial and agricultural practices are known to benefit from the ability of phages to alter the composition of bacterial communities [4,6,7], which makes the discovery and characterization of novel phages highly relevant.
Bacterial viruses are highly prevalent in natural environments, leading to phages constituting one of the greatest genomic reservoirs on the planet [8]. Yet, isolates of some phages remain surprisingly underexplored. Even with recent advances in high-throughput DNA-sequencing techniques, current genomic databases lack high-quality phage genome sequences that can be linked to specific environmentally important bacterial species.
Bacterial members of the genus Sphingomonas, belonging to the Sphingomonadaceae family, are commonly known as an ecologically important group of bacteria that are ubiquitously distributed in diverse aquatic and terrestrial environments, as well as soil [9,10]. Other sources of Sphingomonas include the rhizosphere as well as the surfaces of various plants on which sphingomonads have been detected in rather high cell densities [11][12][13].
Interestingly, recent studies have identified the vital role of some Sphingomonas species in promoting plant growth during various abiotic stress conditions, a role attributed to their potential to produce plant growth hormones [14][15][16][17].
Overall, species belonging to the Sphingomonas genus are known to possess multifaceted functions. Importantly, they possess an extensive metabolic versatility with a prominent ability to degrade and transform a wide variety of natural organic compounds as well as environmental contaminants, i.e., xenobiotically produced compounds, such as biphenyl, chlorinated phenols, polycyclic aromatic hydrocarbons (PAHs), and various types of herbicides and pesticides [18][19][20][21]. On that note, members of the Sphingomonas genus have gained recent attention as being strong candidates for a wide range of biotechnological applications, with current utilization in bioreactors for enzyme production and in situ bioremediation practices [22][23][24].
Despite bacterial members of the genus Sphingomonas being widely prevalent in multiple environmental habitats, the isolation and characterization of phages targeting species of this genus remain sparse. Currently, there is a paucity of sequenced genome data of phages targeting Sphingomonas species and only a few have been fully sequenced [25].
The isolation of Sphingomonas-infecting phages originating from aquatic environments has been described in two studies with the main focus on morphology and infection kinetics [26,27], whereas morphology, host range, and phylogenetic diversity were described for Sphingomonas phages isolated from floodwater of a Japanese paddy field [28]. Recently, an in-depth genomic and proteomic characterization of the phage Lacusarx, which infects a Sphingobium species of the Sphingomonadaceae family, was reported [29]. Finally, physiological and genomic characteristics of the lytic phage vB_StuS_MMDA13, shown to infect Sphingomonas turrisvirgatae, revealed a very low similarity to other known phage genomes [25].
In this study, we describe the isolation and molecular characterization of the three newly discovered lytic phages Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi, which were isolated from a Danish wastewater treatment facility. All phages possess the ability to infect and replicate on a bacterium from the Sphingomonas genus originating from the wheat phyllosphere, chosen based on the current status of wheat being an increasingly important crop worldwide [30]. The wheat phyllosphere microbiome, of which sphingomonads are potential keystone species [31,32], is believed to play a crucial role in extending the plant phenotype [33], hence implicating plant productivity and health. Yet, little is known about the microbial components that target and regulate the wheat phyllosphere microbiome, such as phages.
All three phage genomes were described bioinformatically, whereas morphological features were described for only two phages due to the instability and consequent loss of the phage Vexovirus birtae. In brief, intergenomic studies revealed very little nucleotide sequence similarity between the isolated phages, and also to publicly deposited phage genomes, suggesting that the isolated phages represent three novel genera of viruses. Both the Longusvirus carli and Molestusvirus kimi phages were found to display a non-prolate Siphovirus morphotype. Despite the physical loss of the phage Vexovirus birtae, we do believe that these results contribute to a growing body of knowledge on the taxonomic diversity and ecological role of phages in natural environments-in particular, of those that might assist in regulating keystone bacterial members of the wheat phyllosphere microbiome.

Materials and Methods
The following describes the isolation and genome sequencing of both the bacterial Sphingomonas sp. host and the phages targeting this. We proceeded with Transmission Electron Microscopy (TEM) examination of phage morphology, bacteriophage infection kinetics, and host range determinations of the isolated vira. Detailed information on the methodology follows in the respective subsections below.

Isolation, Growth Conditions, and Genome Sequencing of the Bacterial Sphingomonas sp. Host
The bacterial host, Sphingomonas sp. strain LR55, was isolated in June 2021 from the surface of flag leaves of the winter wheat cultivar Sheriff, grown in an experimental field in Hoeje Taastrup, near Copenhagen, Denmark (55 • 40 12.2 N 12 • 18 13.0 E). Wheat flag leaves were picked, pooled, and washed in 0.9% NaCl + 0.05% Tween80 prior to plating on Sphingomonas selective L9 agar minimal medium [34]. Incubation was performed aerobically at 20 • C for 48 h. Colonies were re-streaked for purification on the same medium and with the same conditions at least three times and then stored at −80 • C in 20% (v/v) glycerol. Further inoculation of the host strain for phage isolation, phage host range determination, morphological characterization, and infection kinetics was performed on R2A medium [35].
Prior to DNA sequencing, DNA was extracted from 2 mL of liquid culture of the host strain in R2A medium using the Genomic Mini AX Bacteria kit (A & A Biotechnology, Gdynia, Poland). DNA long-read sequencing was performed using a PromethION48 sequencer from Oxford Nanopore Technology (ONT) with PromethION R9.4.1 flow cell chemistry. For library building, the Rapid Barcoding Kit 96 V14 (SQK-RBK114.96) was used (Oxford Nanopore Technologies, Oxford, UK). Base calling was performed with Guppy (version 6.4.6). Draft genomes were assembled using Flye (version 2.9.2-b1786) [36], annotated with PROKKA (version 1.13.3) [37] and online with BlastKOALA [38]. Assembled and annotated genomes were submitted to the NCBI GenBank [39] under the Bioproject PRJNA976731.

Isolation, Purification, and Sequencing of Phages
Samples for phage isolation were collected from the Danish wastewater (WW) treatment plant, Lynetten, near Copenhagen. To enrich for phages infecting the bacterial host, a modified version of a previously described phage enrichment protocol was used [40]. In brief, ten mL of WW sample was centrifuged at 5000× g for 15 min to remove particulates and added to 50 mL of liquid R2A medium [35] supplemented with 10 mM of CaCl 2 and MgCl 2 together with one mL of overnight culture of the bacterial Sphingomonas sp. host strain LR55. Following a 48 h incubation step at room temperature (approximately 22 • C), the culture was centrifuged at 5000× g for 15 min to remove bacterial debris and the supernatant, now containing enriched phage particles, was filtered using a 0.45 µm syringe PVDF filter (Merck Millipore, Darmstadt, Germany). A total 1 mL of overnight culture of the bacterial Sphingomonas sp. host strain LR55 (~10 9 CFU/mL) was mixed with 100 µL of different dilutions of the filtered supernatant and plated using a slightly modified double agar overlay assay protocol [41] with R2A medium supplemented with 10 mM CaCl 2 and MgCl 2 and 0.4% agarose as top agar, and R2A medium as bottom agar.
To ensure single phage isolates, five rounds of plating and picking individual phage plaques were conducted using the double agar overlay assay as described above [41].
Phage DNA was extracted using a method modified from previous work [42]. Briefly, 10 U of DNase I (Thermo Scientific, Waltham, MA, USA) was added to 400 µL of phage lysates (~10 10  The phage solution was incubated for 1 h at 55 • C to open phage capsids. The temperature was hereafter raised to 70 • C for 10 min to inactivate proteinase K. The DNA was purified and concentrated using a DNA Clean & ConcentratorTM-5 kit (Zymo Research, Irvine, CA, USA) according to the manufacturer's protocol and eluted with 20 µL of PCR-grade water (Sigma-Aldrich, St. Louis, MO, USA).
DNA sequencing libraries were prepared using the NEBNext Ultra II FS DNA Library Prep kit for Illumina (Illumina, San Diego, CA, USA) according to manufacturer's instructions. Prepared libraries were sequenced in a 151 paired-end sequencing run using the Illumina iSeq platform (Illumina, San Diego, CA, USA).

Transmission Electron Microscopy (TEM) Examination of Phage Morphology
For morphology analyses of the phage particles, transmission electron microscopy (TEM) imaging was performed. Phages Longusvirus carli and Molestusvirus kimi, stored at a high titer in SM buffer [43], were purified for TEM examination by caesium-chloride (CsCl) gradient ultracentrifugation in an Optima XE-90 ultracentrifuge (Beckman Coulter, CA, USA) at 28,000 rpm for 90 min at 10 • C using three density layers (1.3, 1.45, and 1.7 g/cm 3 ) [44] and concentrated using Amicon Ultra-4 centrifugal cellulose units with a pore size of 50,000 NMWL (Merck Millipore, Burlington, MA, USA). The CsCl-purified phage samples were pipetted on 100-mesh copper grids coated with 17.5 nm carbon for absorption for 20 min. Subsequently, the carbon films were washed twice with deionized water and negatively stained for a few seconds with 2% uranyl acetate using a protocol modified from previous work [45]. Electron microscopy was performed on a Talos L120C transmission electron microscope (ThermoFisher Scientific, Eindhoven, The Netherlands) using a 4 k × 4 k Ceta camera (ThermoFisher Scientific, Eindhoven, The Netherlands) set to an acceleration voltage of 80 kV.
Phage head and tail sizes were analyzed using the TEM analysis software Velox v3.9.0 (ThermoFisher Scientific, Eindhoven, The Netherlands). The mean and standard deviation of at least 10 measured virions per phage sample were determined. The TEM pictures were manually improved using GIMP v2.10.32 for contrast increasing and brightness adjustment.

One-Step Growth Curve Experiments
One-step growth curve experiments were conducted using a previously described, but slightly modified, protocol [46] to determine the infection kinetics (i.e., burst size and latent period) of phages Longusvirus carli and Molestusvirus kimi. In brief, the bacterial host Sphingomonas sp. strain LR55 was grown in liquid R2A medium [35] in a shaking water bath at 25 • C, 120 rpm, to an OD 600 of 0.24 (determined to correspond to 2.5 × 10 8 CFU/mL) and infected at a multiplicity of infection (MOI) of 0.06 and 0.024 with phages Longusvirus carli and Molestusvirus kimi, respectively. The cultures were incubated with shaking at 120 rpm for 40 min at 25 • C to allow adsorption of phages to host cells. Following adsorption, 1 mL of each infected culture was centrifuged at 8000× g for 3 min to remove non-adsorbed phages. Pellets were resuspended in 1 mL of SM buffer [43], diluted by a factor E-04 in 30 mL of fresh liquid R2A medium [35], and incubated in triplicates in a water bath at 25 • C and at 120 rpm shaking. To determine the latency period and burst size for each phage, aliquots of 0.2 mL were removed over time, for a total of 10 h, and phage titer was determined in triplicate using the double agar overlay technique [41]. The latency periods were measured excluding the 40 min of phage adsorption. The average burst size for each phage (i.e., the average number of phages released per infected host cell) was computed as the ratio of the average phage titer after the rise period to the average phage titer during the latency period (directly representing the number of infecting phage particles due to removal of non-adsorbed phages prior to dilution) [47].
The adsorption efficacy of each phage was measured by counting unattached PFUs relative to total PFUs using the double agar overlay assay as described above. This was achieved by centrifugation of the samples at 12,000× g for 3 min at T = 0 min and T = 40 min and titering the PFUs of the supernatant (representing unattached phages) compared with the total PFUs without centrifugation (representing total phages). Measurements were performed in triplicates.

Determination of Phage Host Range
Host range determination was performed by spotting 20 µL of 100-fold dilutions (undiluted up to E-08 factor) of the phages Longusvirus carli and Molestusvirus kimi (~10 10 PFU/mL) on lawns of different Sphingomonas wheat flag leaf isolates from our Laboratory collection as well as on lawns of 3 previously characterized Sphingomonas strains (Table 1; Section 3.5) using double agar overlay assays [41] with the same media and growth conditions as described in Section 2.2. The Sphingomonas wheat flag leaf isolates were obtained in June 2021 or 2022 from Danish experimental fields at various locations by a procedure similar to that of the Sphingomonas sp. host strain LR55 (see Section 2.1). Initial growth medium used for isolation included L9 minimal medium [34], R2A minimal medium [35], VL55 medium [48], and a plant extract medium consisting of a basal medium ((NH 4 ) 2 HPO 4 (0.05 g/L), CaCl 2 (0.02 g/L), NaCl (0.1 g/L), MgSO 4 (0.2 g/L), K 2 HPO 4 (0.4 g/L), KH 2 PO 4 (0.4 g/L), and Glucose (2.0 g/L), pH 7.2.) to which 30 g of crushed wheat leaves in 300 mL of water, filtered and autoclaved, was added. The genomes of the Sphingomonas wheat flag leaf isolates were nanopore sequenced using protocols similar to that of the Sphingomonas sp. host strain LR55 (see Section 2.1). Details on sampled locations are stated in Table 1; Section 3.5.
Sensitivity to phage infection was assessed by observing either the formation of individual plaques or the presence of clear zones of phage lytic activity after 48 h of incubation at 20 • C of the spotted dilutions. When individual plaque formation was observed, the number of PFUs was titered and further subjected to relative efficiency of plating (EOP) analysis using triplicates. EOP was calculated as follows: average PFU on target bacteria/average PFU on host bacteria [49].

Genomic Analysis of the Isolated Phages
Clean Illumina reads of phage genomes were assembled using CLC Genomic Workbench V22 (QIAGEN, Aarhus, Denmark), as described previously [42]. Open reading frame (ORF) identification and annotation of assembled genomes was performed using RASTtk annotation server version 2.0 [50] with the modifications described elsewhere [42]. Assembled and annotated genomes of phages Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi were assigned to GenBank with accession numbers OR225223, OR225224, and OR225222, respectively.
Intergenomic similarities were calculated and visualized using the VIRIDIC web server [51] based on pairwise nucleotide comparisons between the five closest relatives to each phage identified using BLASTN [52]. BLASTP [53] was used to identify the five genes with the highest nucleotide sequence similarity to the DNA polymerase encoding genes as well as the large terminase genes of Molestusvirus kimi and Vexovirus birtae in the publicly available databases. To inspect the gene cluster conservation (or lack thereof), phages carrying these genes were compared using the clinker toolkit V0.0.27 [54].

Results and Discussion
3.1. Isolation, Sequencing, and Identification of the Bacterial Host, Sphingomonas sp. Strain LR55 The bacterial host strain (LR55), a member of the Sphingomonas genus, was isolated on selective L9 media plates as orange-pigmented colonies from the surface of wheat flag leaves collected at a Danish experimental wheat field. Whole genome sequencing (WGS) using Oxford Nanopore Technology (ONT) yielded a total of 457 MB of data with an N50 value of 3,960,318 bp and a 111× sequencing coverage of the 4.07 Mbp genome, consisting of two contigs with a GC-content of 65.3%. Following taxonomical classification of the assembled genome using the GTDB-Tk database [55], the host strain, LR55, was classified as Sphingomonas sp. without being assigned to any known species level, hence representing a putatively novel Sphingomonas species not currently available in the genome taxonomy database. Prodigal gene prediction resulted in a total number of 4097 genes including 4024 protein coding sequence (CDS) regions and 73 non-coding RNAs (60 tRNAs, 12 rRNAs, and one tmRNA). No complete prophage genomes were detected in the host genome using PHASTER [56].
As detected by BlastKOALA (providing an annotation of 41.2% of the detected CDS), Sphingomonas sp. strain LR55 was found to harbor several catabolic genes potentially involved in the biodegradation and metabolism of xenobiotic compounds, as commonly seen in sphingomonads [23]. Multiple genes involved in the degradation of fluorinated and chlorinated organic compounds as well as aromatic hydrocarbons were found in the annotation; also, one gene that was possibly associated with the degradation of the persistent organic pollutant dioxin, a suspected cancer-causing agent [57], was detected. Interestingly, one strain of the Sphingomonas genus, Sphingomonas wittichii RW1, was previously described as a potent dioxin-degrading bacterium [58,59].

Identification of Phages Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi and General Features of Phage Genomes
Following a standard enrichment procedure for isolating phages from a Danish wastewater source, as described above using the wheat phyllosphere Sphingomonas sp. strain LR55 as host, several clearly visible plaques were obtained, of which 16 were chosen for further purification. A sequencing-based pre-screening revealed that two of the propagated phages belonged to the Molestusvirus kimi genotype, one belonged to the Vexovirus birtae genotype, whereas the rest belonged to the Longusvirus carli genotype.
For an in-depth genomic analysis, sequencing of genomic DNA extracted from phages Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi resulted in fully assembled genomes of sizes 57,382 bp; 45,617 bp; and 43,699 bp, respectively. The genomes of the three phages were all terminally redundant; however, no clear bias in the start position of reads could be found and, thus, the correct start/end possession of the genomes or any direct terminal repeats could not be identified. For Vexovirus birtae and Molestusvirus kimi, a putative small terminase was identified and chosen as the start site of the genomes. The genomic GC-content of the three phages ranged from 55.9% to 63.0% (Longusvirus carli, 63.0%; Vexovirus birtae, 57.2%; and Molestusvirus kimi, 55.9%), which is lower than the GC-content of the bacterial host (65.3%), as is often seen in phages [29,60,61].
The phage Longusvirus carli shares very little nucleotide sequence similarity (less than 10%) to any known phage in publicly available databases and is the first representative of a completely new group of phages ( Figure 1, BLASTN [52]). Only 7 of 87 open reading frames could be successfully annotated. The annotated genes are primarily involved in DNA replication and phage DNA packaging or represent putative structural genes. Because of the low sequence similarity to known phages and the lack of annotated genes, further genetic analysis was not attempted.
The genomes of Vexovirus birtae and Molestusvirus kimi can be divided into three gene clusters. The first cluster (Figure 2, green line) contains mostly structural genes as well as genes that are typically expressed late in the phage life cycle. Here, genes encoding the major capsid, small and large terminases, as well as various putative structural genes can be found. Most of the genes in this cluster are highly conserved between relatives but some of the genes towards the end of the operon are less conserved. The less-conserved genes contain some putative tail genes and might be involved in host recognition, which could partly explain their lack of conservation.  [51] showing the intergenomic similarity of the isolated phages, Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi, compared with the five closest relatives of each phage. Relevant pairwise nucleotide comparisons were identified and made using BLASTN [52]. Phages isolated in this study are colored green.
The genomes of Vexovirus birtae and Molestusvirus kimi can be divided into three gene clusters. The first cluster (Figure 2, green line) contains mostly structural genes as well as genes that are typically expressed late in the phage life cycle. Here, genes encoding the major capsid, small and large terminases, as well as various putative structural genes can be found. Most of the genes in this cluster are highly conserved between relatives but  [51] showing the intergenomic similarity of the isolated phages, Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi, compared with the five closest relatives of each phage. Relevant pairwise nucleotide comparisons were identified and made using BLASTN [52]. Phages isolated in this study are colored green.
The second cluster (Figure 2, purple line) contains a number of genes involved in DNA metabolism and DNA replication including the DNA polymerase, primase, and helicase encoding genes. Unlike the other two gene clusters, genes belonging to this cluster are mostly reversely orientated (i.e., encoded on the antisense strand). Most of the genes in this cluster are conserved with their closest relatives ( Figure 2). However, a noticeable exception to this is the DNA polymerase encoding gene. The DNA polymerase encoding gene of Vexovirus birtae and Molestusvirus kimi does not share any homology to the DNA polymerase encoding genes of their closest relatives, the Sphingomonas phages Kharn (Accession: MN734439.1), Lucius (Accession: MN734438.1), and Eidolon (Accession: MN734437.1) (Figure 2, red arrow). Interestingly, notable homology is shared with the DNA polymerase encoding genes of their more distant relatives (Figure 2, BLASTP [53]). This might indicate a recent event of horizontal gene transfer of the DNA polymerase encoding gene. Interestingly, some of the close relatives contain a roughly 5 Kbp insertion of an entire queuosine biosynthesis operon into this gene cluster. This operon has previously been shown to be involved in DNA modification of phage DNA and protection against host restriction enzymes [62][63][64].

Phage Plaque and Virion Morphology Characteristics
Following the initial isolation and DNA sequencing of Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi, stability issues were unfortunately observed for Vexovirus birtae, which resulted in loss of infectivity of the phage lysate during storage, preventing further propagation and morphology analyses of this phage. However, prior to the physical loss of Vexovirus birtae, the phages Longusvirus carli and Vexovirus birtae were found to share a similar plaque morphology, both forming clear and circular plaques of around 0.8-2 mm in diameter on a bacterial lawn after incubation for two days at 20 °C ( Figure  3A, top insert). In comparison, the phage Molestusvirus kimi formed larger, less clear, and circular plaques of around 1-4 mm in diameter on a bacterial lawn using the same incubation conditions ( Figure 3B, top insert). Even though this is a large difference, it is not evident that the difference in gene size is a result of mis-annotation caused by stop codon read-through. No tRNA genes or other signs of stop codon read-through were identified.

Phage Plaque and Virion Morphology Characteristics
Following the initial isolation and DNA sequencing of Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi, stability issues were unfortunately observed for Vexovirus birtae, which resulted in loss of infectivity of the phage lysate during storage, preventing further propagation and morphology analyses of this phage. However, prior to the physical loss of Vexovirus birtae, the phages Longusvirus carli and Vexovirus birtae were found to share a similar plaque morphology, both forming clear and circular plaques of around 0.8-2 mm in diameter on a bacterial lawn after incubation for two days at 20 • C ( Figure 3A, top insert). In comparison, the phage Molestusvirus kimi formed larger, less clear, and circular plaques of around 1-4 mm in diameter on a bacterial lawn using the same incubation conditions ( Figure 3B, top insert).  Transmission electron microscopy (TEM) imaging of at least 10 measured virions per phage sample revealed virion morphological features for the phages Longusvirus carli and Molestusvirus kimi typical of a non-prolate (B1) Siphovirus morphotype belonging to the class Caudoviricetes, with icosahedral heads and flexible, non-contractile tails of various lengths [65]. The phage Molestusvirus kimi has a short tail with a more rounded tip and is less pointy (Figure 3B, bottom inserts) compared with the phage Longusvirus carli, which displays a longer tail with a regular tail tip ( Figure 3A, bottom inserts). For none of the phages were visible fibers detected ( Figure 3A,B, bottom inserts; Figure 3C). Transmission electron microscopy (TEM) imaging of at least 10 measured virions per phage sample revealed virion morphological features for the phages Longusvirus carli and Molestusvirus kimi typical of a non-prolate (B1) Siphovirus morphotype belonging to the class Caudoviricetes, with icosahedral heads and flexible, non-contractile tails of various lengths [65]. The phage Molestusvirus kimi has a short tail with a more rounded tip and is less pointy ( Figure 3B, bottom inserts) compared with the phage Longusvirus carli, which displays a longer tail with a regular tail tip ( Figure 3A, bottom inserts). For none of the phages were visible fibers detected ( Figure 3A,B, bottom inserts; Figure 3C).
To our knowledge, however, the phages Longusvirus carli and Molestusvirus kimi are among the first characterized phages with a non-prolate Siphovirus morphotype that infect a member of the Sphingomonas genus. Recently, the phage vB-StuS_MMDA13, also targeting members of the Sphingomonas genus, was found to display a Siphovirus morphotype, yet with an icosahedral elongated prolate head [25]. Finally, another recent phage called Lacusarx, which targets members of the Sphingobium genus, was also found to display the head-tail geometry of the Siphovirus morphotype but with a notably elongated prolate capsid and more unusual morphology [29].

Phage Infection Kinetics
One-step growth curve experiments were conducted to determine the infection kinetics (i.e., burst size and latent period) of the phages Longusvirus carli and Molestusvirus kimi.
Overall, both phages were shown to display latency periods of approximately 150 min, which is excluding the 40 min of adsorption prior to dilution and incubation of the phagehost infection mixtures. Interestingly, the latency period of the phage Longusvirus carli is followed by a long rise period of approximately the same duration as the latency period, whereas the rise period seems shorter for the phage Molestusvirus kimi (~30 min) (Figure 4). It can, however, be argued whether a slight increase in phage titer continues after 180 min for this phage, leading to a possibly extended rise period as well. These long latency and rise periods are not uncommon for slow-growing bacterial host strains (in this study, estimated generation time~180 min), and previous studies have reported similar data for other Siphoviridae infecting, slow-growing Alphaproteobacteria [25,71,72].
The average burst size of the phage Longusvirus carli was determined to be approximately 117 virions per infected cell, whereas the average burst size of the phage Molestusvirus kimi was about 17 virions per infected cell. For both phages, computation of burst size included the last three and the first three sampling points for calculations of average phage titer after the rise period and during latency, respectively. It can be debated whether more or less sampling points should be considered for burst size calculations. Such changes, however, did not lead to any significant alterations in the range of reported burst sizes.
Burst size estimates of prokaryotic cells in situ are highly variable and typically found in the range from a few to around 500 phages [73]. Generally, burst size is believed to be influenced by several factors, such as bacterial/viral size, metabolic activity of the host, as well as the characteristics of the phage and host [73]. Obviously, burst size differs between species; however, phages isolated from the same bacterial host and cultured in a similar nutrient-rich medium might also display different burst sizes [74].
For phages infecting bacterial members of either the Sphingomonas genus or the Sphingobium genus, burst sizes were previously estimated to an average of 30 and 46.5 virions per infected cell [25,29], respectively, which resembles similar orders of magnitude as seen in this study.
Interestingly, the adsorption efficacy of the phages Longusvirus carli and Molestusvirus kimi was determined to be 62.9% ± 17.5 (SD) and 49.8% ± 16.3 (SD), respectively, following 40 min of incubation, which represents a slow rate of adsorption. Therefore, the rise period of Molestusvirus kimi likely represents the slow attachment rather than variation in eclipse periods between individual virocells. The long rise period of Longusvirus carli can only be partly explained by the slow attachment. period, whereas the rise period seems shorter for the phage Molestusvirus kimi (~30 min) (Figure 4). It can, however, be argued whether a slight increase in phage titer continues after 180 min for this phage, leading to a possibly extended rise period as well. These long latency and rise periods are not uncommon for slow-growing bacterial host strains (in this study, estimated generation time ~180 min), and previous studies have reported similar data for other Siphoviridae infecting, slow-growing Alphaproteobacteria [25,71,72]. The average burst size of the phage Longusvirus carli was determined to be approximately 117 virions per infected cell, whereas the average burst size of the phage Molestusvirus kimi was about 17 virions per infected cell. For both phages, computation of burst size included the last three and the first three sampling points for calculations of average phage titer after the rise period and during latency, respectively. It can be debated whether more or less sampling points should be considered for burst size calculations. Such changes, however, did not lead to any significant alterations in the range of reported burst sizes. Typically, phage adsorption rates are functions of several parameters, such as virion diffusion rates, host size and concentration as determining virion collision rates, phage affinity for attachment receptor sites, and likelihood of reversible attachment [75]. Hence, adsorption kinetics are expectedly highly variable, depending on the specific case of phagehost interaction.

Determination of Phage Host Range
A total of sixteen strains, including thirteen recently sequenced Sphingomonas isolates obtained from the wheat phyllosphere from Danish wheat trial fields, and three Sphingomonas strains previously characterized as known degraders of chlorinated phenoxy herbicides (Sphingomonas sp. PM2 [76] and Sphingomonas herbicidovorans MH [77]) as well as dibenzofuran and dibenzo-p-dioxin (Sphingomonas wittichii RW1 [59]), were used to study the host range of the phages Longusvirus carli and Molestusvirus kimi (Table 1). Several wheat isolates obtained in a similar manner as the Sphingomonas sp. host (strain LR55) were chosen with the aim of detecting an increased phage susceptibility. The results, however, demonstrated that the lytic potential of the phage Longusvirus carli is exclusively restricted to the original Sphingomonas sp. host, whereas all other tested strains were insensitive to infection, suggesting a narrow host range of this phage.  [35]; VL55-VL55 medium [48]. Isolates were obtained from either the Sheriff, Rembrandt, Kvium, or Heerup winter wheat cultivars. c Assembled and annotated genomes were submitted to NCBI GenBank [39] under the Bioproject PRJNA976731. d Clearing zones on lawns of these Sphingomonas isolates were observed at concentrations of~10 8 PFUs/mL and above, but formation of individual plaques was not observed.
Surprisingly, the phage Molestusvirus kimi formed clear, individual plaques at concentrations of~10 PFUs/mL and above when tested on one of the Sphingomonas aerolota wheat phyllosphere isolates (strain no. 22S1L9-1; Table 1) sampled from a different location and year compared to that of the original host. The EOP for this isolate was 1.9 ± 0.2, indicating a slightly better infection potential of the phage Molestusvirus kimi on this Sphingomonas isolate compared to the original host.
In addition, the phage Molestusvirus kimi was able to generate clear zones of phage lytic activity at concentrations above 10 8 PFUs/mL on two of the tested wheat phyllosphere Sphingomonas isolates (Table 1), suggesting lysis from without or abortive infection [78].
Both phages were tested against an additional 15 Sphingomonas aerolata strains, isolated from different wheat trial fields; however, none showed susceptibility to either of the phages, which further confirms the complexity of phage-host interactions as well as the specificity and narrow host range of these phages.
These observations resemble previous studies where narrow host ranges were reported for phages infecting bacterial members of both the Sphingomonas genus [25] and the Sphingobium genus [29]. Typically, most phages targeting other types of environmental bacteria are found to be host specific without the ability to infect a wider range of bacteria across species and genus, i.e., to trespass generic boundaries [73,[79][80][81].

Conclusions
Bacterial members of the Sphingomonas genus hold immense environmental significance and have garnered considerable attention as candidates for a wide range of biotechnological applications. These bacteria are widely distributed across diverse natural environments, playing a vital role as keystone members of plant phyllosphere microbiomes. Exploring and understanding the intricate interactions between bacteriophages and Sphingomonas species not only enhances our fundamental comprehension of natural microbial systems but also unveils promising avenues for future microbiome manipulation strategies.
Currently, the repertoire of phages specifically targeting bacterial members of the Sphingomonas genus remains relatively limited. However, in this study, we present the isolation and characterization of three novel phages-Longusvirus carli, Vexovirus birtae, and Molestusvirus kimi-which, for the first time, are shown to target and regulate a Sphingomonas member found within the wheat phyllosphere microbiome.
Phenotypically, two of the phages were found to display a non-prolate (B1) Siphovirus morphotype, which is common among phages isolated from various environmental habitats but less widespread among phages infecting members of the Sphingomonas genus. In addition, the isolated phages had a narrow host range like several other phages, and, during infection, both phages displayed long latency periods, which is not uncommon for slow-growing bacterial host strains like the Sphingomonas host isolated in this study.
What is particularly intriguing is that the genotypic analyses of these phages revealed distinct features that distinguish them significantly from known phages cataloged in publicly available databases. Together with other recently reported findings of novel phages isolated from natural environments, these findings not only underscore the tremendous genetic diversity among ecological phages but also highlight the fact that our current understanding barely scratches the surface of this vast landscape.
Hence, the description of these phages represents a valuable addition to the growing database of environmental phage genomes and contributes to the exciting expansion of our taxonomic knowledge concerning bacterial viruses.