Background

The genus Geobacillus has served as an epicentre for biotechnological exploitation of thermophilic taxa [1, 2]. First described following the 16 s rRNA gene-based reclassification of previously recognised thermophilic clustering (group 5) Bacillus spp. [3], the genus currently comprises 12 validly described species [4]. Members are Gram-positive, aerobic or facultatively anaerobic, spore forming rods that are characterised by their thermophilicity, being capable of growth at temperatures ranging between 37–80˚C [5]. Key taxa of biotechnological value include Geobacillus stearothermophilus, G. thermoleovorans and G. thermodenitrificans [2, 6]. These and other taxa in the genus have been the topic of research and commercial development in a wide range of whole-cell applications, including bioremediation, crude oil recovery and refinement, textile processing, synthesis of nanoparticles, production of antibiotics and production of value added chemicals such as biodiesel, lactate and ethanol [2, 5, 6]. Geobacilli further serve as a source of various thermostable enzymes which present comparably more cost-effective, rapid, non-toxic and environmentally friendly alternatives to whole-cell or abiotic processes that support diverse industries [5, 6]. The application of thermophile derived enzymes has become more prevalent due to their greater thermostability, pH tolerance, catalytic efficiency and reduced cost and contamination rates associated with thermophilic operation [7]. Geobacillus-derived enzymes including α-amylases, α-glucosidases, cellulases, lipases, pectinases, xylanases have received extensive interest for their applicability towards agricultural, biofuel, food, paper, petrochemical, pharmaceutical and textile industries [2, 5, 6].

The application of whole genome phylogenetic approaches highlighted the clustering of Geobacillus taxa in two distinct clades, which were further distinguished based on GC content, resulting in the establishment of the genus Parageobacillus [8]. This genus currently comprises six validly described species which are readily isolated from diverse and globally distributed high temperature environments including hot springs, oil wells, hot composts and geothermal sites and sediments [5]. Another sister genus of both Geobacillus and Parageobacillus, Saccharococcus was established in 1984 and originally comprised a single species, S. thermophilus, isolated from beet sugar extracts [9]. A second thermophilic and xylanolytic species isolated from soil in Australia, S. caldoxylosilyticus, was subsequently described [10] but its taxonomic status was short lived, shifting first to the genus Geobacillus and subsequently the genus Parageobacillus [8].

In congruence with their wide and varied distribution, the genera Parageobacillus and Saccharococcus encompass a broad range of microorganisms with versatile metabolic potential, encoding a range of robust thermostable and thermoactive enzymes, many of which may be of biotechnological value [2, 5]. While some research has focused on the biotechnological potential of P. thermoglucosidasius, the inherent capacity of the genera Parageobacillus and Saccharococcus as a whole, in comparison to the sister genus Geobacillus, remains relatively underexplored. Here we have made use of whole genome sequence data and phylogenomic approaches to establish the relationship of taxa the genera Parageobacillus and Saccharococcus and demonstrate the clustering of S. caldoxylosilyticus with S. thermophilus in the latter genus. Further, using comparative genomic and pan-genome analyses, we provide an in depth characterisation of the biotechnological potential of these key thermophilic taxa.

Results & discussion

Phylogenomic analysis delineates Parageobacillus and Saccharococcus as two distinct genera

The genus Parageobacillus was resolved from the genus Geobacillus using phylogenomic analysis, and comprises six distinct species, including Parageobacillus caldoxylosilyticus [8]. However, the taxonomic status of the latter species remains contentious, having first been assigned to the genus Saccharococcus [10], subsequently the genus Geobacillus and finally the genus Parageobacillus [8]. In this study a core genome maximum likelihood phylogeny was constructed on the basis of 1,784 single-copy orthologous proteins conserved among 34 Parageobacillus strains, the Saccharococcus thermophilus DSM 4749 T genus type and the outgroup strain Geobacillus thermodenitrificans DSM 465 T. This phylogeny showed the clear delineation of the taxa in two distinct clades (Fig. 1), with the nine P. caldoxylosilyticus strains and Parageobacillus genomosp. 1 NUB3621 clustering with S. thermophilus DSM 4749 T, indicating they belong to the genus Saccharococcus. This is further supported by the Average Nucleic acid Identity (ANI) and digital DNA-DNA Hybridisation (dDDH) phylogenomic metrics, where intraclade ANI values of 92.22 and 96.16% and dDDH values of 59.19 and 73.72% are observed for the Parageobacillus and Saccharococcus clades, respectively, while interclade values are 83.57% (ANI) and 27.57% (dDDH) (Additional file 2: Table S1). Two Parageobacillus strains with species designation, namely KH3-4 and W-2, demonstrate dDDH (average 44.04%) and ANI (90.01%) values below the 70% and 96% threshold that constitute the species boundaries [8] and as such, they form a novel genomospecies, Parageobacillus genomosp. A.

Fig. 1
figure 1

Core genome ML phylogeny of the genera Parageobacillus and Saccharococcus. The phylogeny was constructed on the basis of 1,784 SCOs, with the concatenated alignment comprising 499,928 amino acid positions, of which 81,368 were deemed parsimony informative and 58,513 represented single sites. The ML phylogeny was constructed using the optimal evolutionary model JTT + F + I + G4 with ultrafast bootstrap support (n = 1,000 replicates). G. thermodenitrificans DSM 465T was used as outgroup

Full size image

The genomes of members of both Parageobacillus and Saccharococcus are similar in size (average: 3.763 and 3.742 Mb, respectively), while the genomic G + C contents of members of the genus Saccharococcus are on average 0.94% greater than their Parageobacillus counterparts (Table 1). The genomes of taxa in both genera code for a similar number of proteins (3,704 and 3,719, respectively), with the most proteins encoded on the genome of S. caldoxylosilyticus B4119 (3,986), followed by three P. thermoglucosidasius strains. In general, less proteins are encoded on the genomes of P. toebii strains (average 3,461 proteins). The least proteins (3,085) are encoded on the genome of S. thermophilus DSM 4749 T, with a genome that is also ~ 650 kb smaller than the other comparator taxa on average. Analysis of the COG functions associated with the proteomes of each strain showed that slightly more proteins (~ 2% or 84 proteins on average) involved in metabolism are encoded on the genomes of Saccharococcus taxa, which can primarily be attributed to the COG categories amino acid (E), nucleotide (F) and lipid (I) transport and metabolism (Additional file 1: Figure S1). By contrast, the outgroup taxon G. thermodenitrificans DSM 465 T codes for substantially fewer (2.5%) proteins involved in information processing and storage (primarily in COG category L—replication, recombination and repair) and a greater (3.2%) number of proteins of unknown function than the two comparator genera (Additional file 1: Figure S1).

Table 1 Metadata of the taxa and genomes used for comparative genomic and phylogenomic analysis
Full size table

Parageobacillus and Saccharococcus have open pan-genomes with P. toebii as a key driver of novel gene accrual

The core (conserved among all taxa in a set), accessory (conserved among some taxa or unique to specific taxon in set) and pan-genome (combination of core and accessory fractions) for the genera Parageobacillus and Saccharococcus were determined. The overall pan-genome of both genera combined (taxa) comprises 9,082 orthogroups, of which 1,950 (21.5%) are core to all taxa (Fig. 2A). A total 37.1% and 15.4% of the orthogroups are unique to the genera Parageobacillus and Saccharococcus, respectively. Analysis of the functions of the core and Parageobacillus- and Saccharococcus-unique fractions showed that carbohydrate transport and metabolism (COG category G), in particular, is overrepresented in the genus-specific proteome datasets, suggesting distinct metabolic capacities for the two genera. Furthermore, the synthesis of secondary metabolite biosynthesis (COG category Q) and defense mechanisms (COG category V) are largely genus-specific traits (Fig. 2B). Only eleven and twenty-one orthogroups are core to all Parageobacillus and Saccharococcus taxa in each set, respectively. The Parageobacillus-unique core proteins are dominated by transcription regulators (four proteins), while the Saccharococcus-unique core proteins include three proteins involved in amino acid transport and three proteins involved in copper resistance (CotA, CopC and YcnI) (Fig. 2B).

Fig. 2
figure 2

Protein conservation and function in the genera Parageobacillus and Saccharococcus. A Proportion of conserved and genus-specific orthogroups for the two genera. Numbers in red indicate those proteins conserved in all taxa within each dataset, while those in brackets reflect those that are present in some, but not necessarily all taxa within each genus or the combination of the genera. B Conserved Orthologous Group functional categories for the proteins conserved among all taxa in both genera, those specific to each genus and those that are conserved among all taxa in each genus. The graphs represent the relative proportions as calculated for those proteins categorized for a COG function (with the exception of category S: function unknown). The number of proteins assigned to COG functional categories are shown in brackets above each bar

Full size image

Pan- and core genome graphs were constructed for the genera Parageobacillus and Saccharococcus and extrapolated to encompass 100 genomes/genus (Fig. 3A). Both genera display an open pan-genome, with that of Parageobacillus being slightly larger than the genus Saccharococcus. Similar numbers of new genes (24.4 and 24.8) are predicted to be added to the pan-genome when the 100th genome of Parageobacillus and Saccharococcus is sequenced. When considering genome conservation, the core genome of Saccharococcus is predicted to be slightly larger (2,332) than that of Parageobacillus (2,171) across 100 genomes.

Fig. 3
figure 3

Pan- and core-genome graphs. A Pan- and core genomes for the genera Parageobacillus (blue) and Saccharococcus (red). Larger shapes indicate the actual values, while smaller shapes depict the extrapolated values. B The pan- and core genomes of P. thermoglucosidasius (blue), P. toebii (red) and S. caldoxylosilyticus (green)

Full size image

To evaluate the pan-genome dynamics of individual species within each genus, the pan- and core-genomes of three species for which ≥ 7 genomes are available (P. thermoglucosidasius—14 genomes, P. toebii—7 genomes and S. caldoxylosilyticus—8 genomes), were extrapolated (Fig. 3B). All three species display open pan-genomes. Similar pan- and core-genome trends were observed for P. thermoglucosidasius and S. caldoxylosilyticus, with the core genome approaching a predicted average of 3,012 orthogroups when 100 genomes are sequenced, while the 100th genome would add 15 novel proteins to the pan-genome of both species. By contrast, a much larger pan-genome (~ 2,800 more orthogroups when considering 100 genomes) was observed for P. toebii than the other species, with 24.4 novel proteins added by the 100th taxon genome included in the analysis. This species further has a substantially smaller core genome, with almost 600 core orthogroups less than the other two species. This suggests that P. toebii has a more unstable pan-genome than P. thermoglucosidasius and S. caldoxylosilyticus and that this species may be capable of greater ecological, metabolic and functional diversification than the two latter species [11]. This is further supported when considering the genomes incorporated in this study, where P. toebii-specific orthogroups (seven genomes) contribute 14.6%, while P. thermoglucosidasius-specific proteins (with double the number of genomes analysed) contribute 18.5% (Additional file 1: Figure S2A). The largest proportion of proteins involved in the supra-functional category information storage and processing (38.4%) is observed for the P. toebii-specific protein complement, while the P. thermoglucosidasius-specific proteins are primarily involved in metabolism (Additional file 1: Figure S2B). These species-specific datasets are predominated by proteins involved in DNA replication, recombination and repair (25.9%) and carbohydrate (10.4%) as well as amino acid (10.1%) transport and metabolism, respectively (Additional file 1: Figure S2C).

Plasmids, bacteriophages and transposable elements are key drivers of Parageobacillus and Saccharococcus diversification

Plasmid replicons, prophages and transposable elements were predicted for the comparator Parageobacillus and Saccharococcus taxa. Plasmid replicons occur in 75% and 54.5% of the taxa in each genus, respectively (Table 1). Half of the plasmid-bearing Parageobacillus taxa incorporate two plasmids, while S. thermophilus DSM 4749 T harbours two plasmids and S. caldoxylosilyticus B4119 is predicted to carry three distinct plasmids. The plasmids vary substantially in size, with the smallest (1,080 nucleotides) and largest (~ 105 kilobases) both occurring in P. thermoglucosidasius G08C001. These plasmids contribute up to 3.81% and 4.31% of the total genome and protein complement (highest for both observed in P. thermoglucosidasius 23.6) (Table 1). Prophage elements are more prevalent in both genera, with between one and eight (Parageobacillus genomosp. A. W-2) elements per genome (Table 1). In most cases these prophage elements are predicted to be incomplete, but three complete phage elements are predicted on the genome of Parageobacillus genomosp. A W-2 and phage-proteins contribute 6.29% of the total proteins encoded on the genome of the latter strain.

Between 29 (Saccharococcus genomosp. A NUB3621) and 263 (P. toebii WCH70) transposases (belonging to 74 distinct orthogroups) were predicted per genome. Notably, P. toebii incorporate an average of 124 transposases per genome, while P. thermoglucosidasius and S. caldoxylosilyticus genomes incorporate an average 50 and 63 transposases, respectively, indicating a key role for transposition in the diversification of P. toebii.

When considering plasmids, prophages and transposases in combination, these elements contribute 6.1% and 5.1% of the total genomic protein contents for Parageobacillus and Saccharococcus, respectively, while for the comparator G. thermodenitrificans DSM 465 T, these elements encompass only 3.8% of the total proteome. Stand-out taxa include P. thermantarcticus DSM 9572 T and P. toebii WCH70, where these elements in combination, contribute 9.7% (primarily prophage elements) and 10.1% (primarily transposases) of the total protein content, highlighting the combined role of these elements in shaping the highly versatile genera Parageobacillus and Saccharococcus. Given the genomic versatility and extensive core genome the Parageobacillus and Saccharococcus genome dataset was evaluated for proteins of potential biotechnological value.

Mining the Parageobacillus and Saccharococcus pan-genome for biotechnology

Parageobacillus as a source of novel antimicrobials

The emergence and rapid spread of antibiotic resistance among clinically relevant pathogens has driven the continued search for novel natural products to combat these pathogen [12]. To this extent, the geobacilli have been receiving increasing attention, with several studies identifying bacteriocins and bacteriocin-like inhibitory substances effective against a range of different pathogenic microorganisms [13,14,15]. antiSMASH [16] predicted on average 5.3 and 6.4 secondary metabolite biosynthetic loci in members of the genera Parageobacillus and Saccharococcus, respectively. Included among these are loci for the synthesis of metallophores (three types), betalactones (three types), betalactones (two types), a ladderane, a spore-killing factor and eight distinct bacteriocin biosynthetic loci. The latter loci were further confirmed and characterised using the BAGEL 4 [17] and RiPPMiner-Genome [18] servers.

A collection of six Class I and two Class II bacteriocins are distributed across the genome dataset. The Class I bacteriocins comprise four lantibiotic loci, a linear azole-containing peptide and a thiopeptide biosynthetic locus. The best-known Geobacillus antimicrobials are the lantibiotics geobacillin I and II of G. thermodenitrificans, effective against vancomycin-resistant Enterococcus faecium/methicillin-resistant Staphylococcus aureus and Bacillus cereus/B. subtilis, respectively [13]. The geobacillin I locus comprises ten genes, including geoAI which codes for the bacteriocin peptide, while the geobacillin II locus comprises three genes, with geoAII encoding the bacteriocin peptide [13, 15]. A complete geobacillin I locus was identified in a single taxon in our dataset, namely P. thermantarcticus DSM 9572 (84.6% average amino acid identity across 10 proteins; 92.9% average amino acid identity (AAI) for GeoAI bacteriocin peptide to G. thermodenitrificans NG80-2) (Fig. 4). Of note, 19/23 of the other Parageobacillus strains encode orthologues of geobacilin I self-immunity (geoEFGI) and two-component regulatory systems (geoKR) [13], suggesting they have immunity to the geobacillin I lantibiotic but are unable to produce it themselves. Only a single taxon in the dataset, P. toebii B4110, incorporates a geobacillin II locus (Fig. 4), which was previously shown to be more restricted in distribution than geobacillin I (only in two G. thermodenitrificans strains). The locus encodes all three proteins produced by the G. thermodenitrificans NG80-2 geobacillin II locus (99.9% average AAI). Downstream of the P. toebii locus are three genes coding for orthologues of erythromycin-like esterases (cd14728 – ere-like), which provide resistance to macrolides [19] and may potentially serve as a self-immunity mechanism for geobacillin II.

Fig. 4
figure 4

Parageobacillus and Saccharococcus antimicrobial biosynthetic loci. Schematic diagrams of the bacteriocin I and II loci present on the genomes of select Parageobacillus and Saccharococcus taxa. Genes coding for the active bacteriocin peptide are coloured in red, while those coding for accessory biosynthetic proteins, regulation and transport are shown in light yellow, green and blue, respectively. Genes coding for proteins with a potential role in self-immunity are indicated by orange arrows. A 1 kb scale bar is shown

Full size image

Two further lantibiotic biosynthetic loci types, lantibiotic III and lV were predicted on the genomes of 4/7 P. toebii strains and S. caldoxylosilyticus B4119, respectively (Fig. 4). The lantibiotic III cluster was previously identified in silico as lantibiotic cluster 4/5 [15], while the predicted bacteriocin peptide is a predicted esterase/lipase (cd00312). The lantibiotic IV cluster, novel to this study, includes genes coding for a lantibiotic dehydratase (lanB), cyclase (lanC) and ABC transporter (lanT), showing limited homology to the subtilin biosynthetic proteins SpaBCT of Bacillus subtilis (P33115-6; P39774.2; 33.5% average AAI). Two predicted FDLD family class I lanthipeptides (sharing 69.3% AAI) are encoded upstream of the other biosynthetic genes (Fig. 4).

Linear azole containing peptides (LAPs) contain heterocyclic rings of thiazole and (methyl)oxazole [20]. With the exception of P. thermantarcticus DSM 9572 and S. thermophilus DSM 4749, all examined taxa incorporate a four gene LAP biosynthetic locus coding for a cyclohydratase (sagC), a maturase (sagD) and a dehydrogenase (sagB) as well as a 74–113 amino acid bacteriocin peptide (71.5% AAI among the compared taxa) belonging to the heterocycloanthracin/sonorensin family (TIGR03601). Heterocycloanthracin was identified in Bacillus cereus and Bacillus anthracis [20] and sonorensin from a marine Bacillus sonorensis isolate [21]. Sonorensin has been shown to effective against Listeria monocytogenes and Staphylococcus aureus, with anti-biofilm activity for the latter pathogen and could be used as a food biopreservative [21]. P. thermantarcticus DSM 9572 and Parageobacillus genomosp. A W-2, incorporate a locus coding for a predicted sactibiotic (Fig. 4). Sactipeptides incorporate post-translational modifications with intramolecular bridges of cysteine sulphur to α-carbon linkages [22]. The identical 49 aa peptide in the Parageobacillus strains share 67.4% AAI with the huazacin peptide in Bacillus thuringiensis serovar huazongensis BGSC 4BD1 (EEM79974.1), which shows activity against the food-borne pathogen L. monocytogenes [23].

Two distinct class II bacteriocin loci were also identified among the studied taxa. A 123 aa peptide (98.0% AAI) present in 15/24 Parageobacillus taxa (all P. thermoglucosidasius and P. toebii WCH70), but absent from all Saccharococcus strains, is predicted to belong to the lactococcin 972 family (pfam09683), produced by Lactococcus lactis and active against closely related organisms [24]. The second locus encodes a 48 aa peptide and is found on the plasmids of four P. thermoglucosidasius strains (100% AAI). It is predicted to belong to the aureocin A53 family (NF033881), which is produced by S. aureus and is active against L. monocytogenes [25].

Another potential group of antimicrobials are lactonases, which degrade or quench N-acyl-homoserine lactones (AHLs) that serve as chemical signalling molecules in Gram-negative pathogens and thereby inhibit AHL-regulated functions such as the production of virulence factors and biofilms [26]. One such lactonase, GcL (WP_017434252.1) was identified in S. caldoxylosilyticus DSM 14590 T [26]. Orthologues sharing 96.4% AAI are found in all Parageobacillus and five S. caldoxylosilyticus strains. A second predicted N-acyl-homoserine lactonase is found in all 35 comparator taxa and these share 68.2% AAI with the quorum quenching lactonase YntP of B. subtilis 168 (O34760.2). The latter lactonase inhibits streptomycin production in Streptomyces griseus [27]. Furthermore, orthologues (78.3% AAI) of a broad-substrate N-acyl-homoserine lactonase from G. kaustophilus HTA426 (GKL – 3OJG) [28] are encoded on 11/11 Saccharococcus genomes, as well as those of P. thermantarcticus DSM 9572 T and P. toebii DSM 14590 T. As such, given the increasing prevalence of antimicrobial resistance, thermostable N-acyl-homoserine lactonases produced by Parageobacillus and Saccharococcus should receive additional attention.

Parageobacillus and Saccharococcus as a source of bioindustrially relevant enzymes

With a projected market share of $ 16.9 billion by 2027 [29], enzymes and in particular their thermostable counterparts, form a cornerstone of a broad range of industries, including the production of food, detergents, textiles and bioenergy [7]. Using a range of in silico tools, the Parageobacillus and Saccharococcus genomes were screened for thermostable enzymes of potential biotechnological value.

Carbohydrate-active enzymes

Bacteria produce a range of carbohydrate-active enzymes (CAZymes) to degrade complex carbohydrate polymers into monomeric sugars, which from a biotechnological perspective can be further fermented into biofuels and a broad range of value-added chemicals [30]. A total of 2,130 CAZymes were predicted across the 35 compared genomes (average 61 CAZymes/genome) (Additional file 2: Table S2). These were predominated by glycoside hydrolases (GH: 44.6%) that hydrolyse or rearrange glycosidic bonds in carbohydrate chains, glycosyltransferases (GT: 43.0%) that form bonds in carbohydrate chains, and carbohydrate esterases (CE: 11.5%) that deacetylate ester-substituted carbohydrates [30, 31]. Biotechnological focus is on GH and CE classes, as well as less represented polysaccharide lyases (PL: only presented on 4/35 genomes) that catalyse the non-hydrolytic cleavage of glycosidic bonds in carbohydrate chains (Additional file 2: Table S2) [30, 31]. A total of 930 GHs were identified on the 35 compared genomes, with 57 of these (6%) predicted to be extracellularly secreted. Substantially greater numbers of GHs are encoded on the genomes of members of the genus Saccharococcus (average GHs: 34.5/genome) than those of Parageobacillus (average GHs: 23.8/genome). This could largely be attributed to several strains of S. caldoxylosilyticus, in particular KH1-5 and KH1-6 which both code for 44 GHs (Additional file 2: Table S2).

GHs are further classified into 186 GH families [31], each with their own hydrolytic mechanism and/or substrate. The Parageobacillus and Saccharococcus GHs cover 33 distinct GH families, eight of which are predicted to be secreted extracellularly. Of these families, two are uniquely represented in the genus Parageobacillus, while five families are restricted to Saccharococcus taxa. Between seven (P. toebii NEB718 and S. thermophilus DSM 4749) and twenty-six (S. caldoxylosilyticus DSM 12041 and KH3-5) of the 33 GH families are encoded on each individual strain genome, with only three GH families, namely GH13, GH18 and GH23, core to all 35 compared taxa (Additional file 2: Table S2). The latter two families are involved in peptidoglycan hydrolysis and play a role in spore germination [32] and cell wall remodelling and recycling [33], respectively. The GH13 α-amylase family, which degrades starch and its derivatives (e.g. amylopectin and pullulan) [34], is the most broadly represented of all GH families among the Parageobacillus and Saccharococcus taxa, with 260 members across the 35 genomes. Being the major storage carbohydrate of terrestrial plants, starch degrading enzymes are of value in the food, fermentation and pharmaceutical industries, in particular the thermostable variants as produced by Geobacillus and Parageobacillus species [5, 6].

The majority of GHs encoded on the Parageobacillus and Saccharococcus genomes are involved in the degradation of lignocellulosic biomass. Lignocellulose, comprised of cellulose, hemicellulose, lignin and minor fractions of lipids, proteins, pectin and soluble sugars, forms the predominant component of plant biomass and is one of the most abundant renewable substrates on Earth [30]. In geobacilli plant biomass degradation activity can be linked to the large, highly variable Hemicellulose Utilization System (HUS) locus, which incorporates hydrolytic enzymes, sugar transport systems and carboxylesterases to completely degrade and utilise the xylose backbone, arabinose, galactose and glucuronic acid side chains and methyl or acetyl group decorations [35]. Highly variable HUS loci were found in 14/35, which could further be subdivided into five types (I-V) (Fig. 5). Type I and II are restricted to P. thermoglucosidasius and P. thermantarcticus DSM 9572 (Type I) and P. thermoglucosidasius only (Type II) and target xylans decorated with glucuronic acid and arabinofuranose side chains, respectively. Unique to the Type I HUS loci is a gene coding for a GH5 endoglucanase, indicating that these taxa may also target the cellulose component of biomass. Type III HUS loci were found on the genomes of the two Parageobacillus genomosp. A isolates and three S. caldoxylosilyticus strains, and are predicted to target arabinose and glucuronic acid-containing xylans. The Type IV HUS locus, unique to S. caldoxylosilyticus VR-IP, likely also targets this hemicellulose substrate, but further incorporates genes coding for enzymes for the hydrolysis and metabolism of galactose (GH36), mannose (GH38_1, GH38_2 and GH38_3), N-acetylglucosamine (GH84) and fructofuranose (GH100) [31], suggesting this strain can degrade more complex plant biomass substrates. Finally, the Type V HUS locus of S. caldoxylosilyticus KH1-5 and KH1-6 encodes the cellular machinery for the degradation of rhamnogalacturonan I, with pathways for the degradation of the backbone as well as arabinan and glucuronic acid side chains. This pectic polymer forms a major part of the primary cell wall and middle lamella of most higher plants [36].

Fig. 5
figure 5

Parageobacillus and Saccharococcus Hemicellulose Utilisation (HUS) loci Schematic diagrams of the Type I-V hemicellulose utilisation loci on the genomes of select Parageobacillus and Saccharococcus taxa. Arrows corresponding to genes are coloured according to their predicted carbohydrate target. Grey arrows represent those genes coding for proteins of unrelated or unknown functions, while black arrows represent predicted transposase-associated elements. Lighter-shared arrows represent genes involved in carbohydrate transport, while cross-hatched arrows are those genes with predicted regulatory roles in carbohydrate metabolism. A 2 kb scale bar is shown

Full size image

The propensity of Parageobacillus and Saccharococcus taxa to degrade distinct and variously decorated plant biomass constituents offers excellent opportunities for biocomposting of plant biomass, potentially as mixed cultures [37, 38], or the production of value-added products such as oligosaccharides that could be used as prebiotics or food additives [39]. One component of plant biomass that affects the efficacy of enzymatic degradation is lignin. A lignin degrading laccase has been identified in Geobacillus sp. WSUCF1 (WP_011230630.1) [40]. Orthologues of this laccase are encoded on the genomes all 35 studied taxa (61.4% AAI), suggesting that they further incorporate the machinery to degrade this plant biomass constituent.

Lipases, carboxylesterases and proteases

Thermostable lipases and carboxylesterases are of growing interest in the food, pharmaceutical and fine-chemical industries, where their products of hydrolysis can be used for the synthesis of various chemicals [2, 6]. Where lipases degrade water-insoluble long chain triglycerides, carboxylesterases hydrolyse ester bonds in shorter chain acyl derivatives [6]. Comparison of the Parageobacillus and Saccharococcus proteomes against the Lipase Engineering Database (LED) [41] identified orthologues for twenty-four distinct homologous family groups (Additional file 2: Table S2). Of these, thirteen constituted alpha/beta hydrolases (abhydrolases – cl 21,494) for which no clear substrate/activity could be identified, while five distinct acetyl esterases are predicted to contribute to the removal of acetyl groups from lignocellulosic components (xylan and rhamnogalacturonan). Three distinct carboxylesterases are encoded on the genomes. p-Nitrobenzyl esterases need to be removed from oral beta-lactam antibiotics for their final synthesis, and the p-nitrobenzyl esterase (PbnA) of B. subtilis is effective in this activity [42]. Orthologues of this enzyme (P37967.2; 44.2% AAI) are present in 11/11 Saccharococcus strains and P. thermantarcticus DSM 9572. Orthologues of two characterised carboxylesterases from Geobacillus stearothermophilus (Est30; Pdb = 1TQH; 90% AAI) and G. thermodenitrificans CMB-A2 (EstGtA2; AEN92268.1; 72% AAI) are present in all 35 analysed taxa. Both of these thermostable enzymes show activity against p-nitrophenyl esters of different chain length [43, 44]. All 35 Parageobacillus and Saccharococcus taxa also encode orthologues of a lysophospholipase (YpA; COG 2267) as well as two distinct copies of GDSL-like lipases (pfam 13,472). However, the target triglycerides would need to be determined.

Microbial proteases and peptidases, in particular their thermostable counterparts, have a broad range of applications including the treatment of leather, as additives in detergents and in the food industry [2, 6]. Comparison of the proteome datasets against the MEROPS database [45] identified 4,765 distinct protein orthologues encoded on the 35 genomes. On average, slightly more (138) are encoded on the Saccharococcus genomes than on the Parageobacillus genomes (135), while 130 are encoded on the genome of G. thermodenitrificans DSM 465 T. The highest number of proteases/peptidases are encoded on the genome of S. caldoxylosilyticus B4119 (152) (Additional file 2: Table S2). The proteases/peptidases belong to 40 and 66 distinct MEROPS clans and families, respectively, with the highest numbers of families represented in Saccharococcus genomosp. A NUB3621 (60). The proteases/peptidases can be subdivided into 212 orthogroups, 91 of which (43%) are core to all compared taxa, while 33 (16%) occur only in a single taxon. A total of 40 and 21 protease/peptidase orthogroups are unique to either the genus Parageobacillus or Saccharococcus, respectively. Only a small proportion (23/212) of the protease/peptidase orthogroups are secreted extracellularly, with six each of these unique to Parageobacillus and Saccharococcus taxa, respectively.

Parageobacillus genomosp. nov. A KH3-4 and W-2 as well as Saccharococcus genomosp. nov. A NUB3621 (two copies) produce a predicted neutral thermolysin protease sharing 68.2% AAI (range 51.6–82.7%) with thermolysin from G. stearothermophilus (P43133.1). The latter protease (NprS) is commercially used to produce precursors for the artificial sweetener aspartame [46]. Serine proteases, particularly those of the subtilisin superfamily (S8), have a broad range of applications in the food, cosmetics and detergent industries, and in the treatment of sewage [47]. A total of 171 S8 family proteases are encoded across the Parageobacillus/Saccharococcus genomes, belonging to 16 distinct orthogroups (12/16 extracellularly secreted). Orthogroups of five and one subtilisin protease are unique to single strains of Parageobacillus and Saccharococcus, respectively, while a further three orthogroups are represented in Parageobacillus species only. While the S8 proteases in these taxa share between 27.1 and 43.6% AAI with subtilisin J of G. stearothermophilus NCIMB 10278 (P29142.1; 27), the S8 protease orthogroups in this study share < 50% AAI among them, indicating a broad underexplored set of proteases of potential biotechnological value among the genera Parageobacillus and Saccharococcus.

Enzymes for the molecular laboratory

Thermostable DNA-active enzymes encompass an expanding toolkit for numerous conventional molecular biotechnology applications, including PCR, genetic engineering, DNA sequencing, diagnostics and synthetic biology [5]. Several thermostable DNA polymerases have been derived and commercially developed from Geobacillus spp., most notably the Bst DNA polymerase, a family A DNA polymerase I with 5'-3' exonuclease activity isolated from G. stearothermophillus GIM1.543 [48]. All strains analysed possessed DNA polymerases of the families A (DNA PolI), C (DNA PolIII—α, τγ, δ, δ′ and β subunits), Y (DNA PolIV) and X (DNA PolX), represented by one orthogroup each (Additional file 2: Table S3). In addition, a putative DNA polymerase family B (PolB) orthologue, is encoded on the genomes of P. thermantarcticus DSM 9572 T and P. thermoglucosidasius DSM 21625. In addition to the DinB DNA polymerase IV orthologues (74.2% AAI; range 52.0–100%) encoded by all strains analysed, two P. toebii and eight S. caldoxylosilyticus strains encode putative UmuC DNA polymerase family Y (DNA PolV) orthologues (81.8% AAI) involved in UV-dependent and chemically-induced mutagenesis [49]. These polymerases may have application in inducing random mutagenesis for the purpose of directed evolution [50].

Thermostable restriction enzymes and their associated modification (RM) systems are used in various generic engineering strategies, sequencing and diagnostics [51]. Comparison to the REBASE database [52] identified 61 orthogroups incorporating restriction-modification (RM) components (Additional file 2: Table S4). These included twenty-seven Type I, eighteen Type II, nine Type III and six Type IV putative RM components. Most (59/61) of the identified RM components are encoded on the genomes of Parageobacillus spp., 43 of which are unique to the genus. Of these, twelve and twenty-three are specific to P. thermoglucosidasius and P. toebii, respectively. Saccharococcus genomes only encode 19/61 of the RM components, three of which are unique to S. caldoxylosilyticus strains. On average, ~ 7 and 3 RM components are encoded on the genomes of Parageobacillus and Saccharococcus, respectively, suggesting they, and in particular the former genus, represent a rich source for novel thermostable RM enzymes.

In addition to the native role CRISPR-Cas systems play in preventing foreign plasmid and nucleic acid transfer in prokaryotes [53], modified CRISPR-Cas systems have also been employed in various biotechnological and biomedical applications through targeted genome editing and gene regulation [54]. Recently, several Geobacillus Cas proteins have also received attention due to their thermostability and greater specificities when compared to the more frequently utilised mesophilic Cas9 systems [55]. Using CRISPRCasFinder [56], 34 distinct orthogroups were identified as Cas proteins of type I and type III CRISPR-Cas systems (Additional file 2: Table S5). Substantially more Cas proteins were identified in P. thermoglucosidasius (average Cas proteins: 16.96/genome) compared to Saccharococcus spp. (average Cas proteins: 6.1/ genome).

Whole-cell biotechnological applications for Parageobacillus and Saccharococcus

Applications of Parageobacillus and Saccharococcus in bioremediation

Aside from the biotechnological potential of their enzymes, there has also been extensive interest in whole cell biocatalysis with thermophilic Geobacilli (Fig. 6) [2]. Numerous Geobacillus (and Parageobacillus) strains have been investigated for their applicability towards various bioremediation applications, including degradation of xenobiotics, phenols and in particular long chain- and aromatic-hydrocarbons and petroleum hydrocarbons [6].

Fig. 6
figure 6

Schematic diagram showing potential whole-cell applications of the genera Parageobacillus and Saccharococcus. Coloured balls represent species containing at least one gene orthologue or pathway component for each respective system

Full size image

Analysis of the comparator protein dataset identified 84 distinct orthogroups associated with degradation of various xenobiotic compounds (Additional file 2: Table S6). The genomes of P. thermoglucosidasius taxa typically encode substantially more orthologues (61.4/genome) than either Saccharococcus spp. (43.6/genome) or P. toebii (28.9/genome; Fig. 6). The highest number of proteins involved in xenobiotic degradation occur in P. thermoglucosidasius 23.6 (69).

Phenol meta-cleavage pathway degradation loci (a twelve gene chromosomal and ten gene plasmid locus) have previously been identified in the genus Parageobacillus [57]. The full chromosomal phenol degradation operon is conserved among all P. thermoglucosidasius strains, 2/7 P. toebii, both Parageobacillus genomosp. A strains and 5/8 S. caldoxylosilyticus strains. (Additional file 2: Table S6). Furthermore, 8/14 P. thermoglucosidasius strains carry the complete plasmid-bound locus.

Crude and refined petroleum fractions may contain or release (upon combustion) high levels of organosulphur compounds, which are resistant to degradation and hazardous to the environment [58]. Consequently, biological means of reducing levels of organosulfur compounds either preventatively in processed petroleum products, or in the remediation of polluted systems, is desirable. Various thermophilic taxa have been observed capable of catabolising sulphur-rich petroleum compounds, including members of the genus Parageobacillus [58]. Thirty distinct orthogroups were associated with sulphur metabolism. Three desulphurization-associated gene clusters (1, 2 and 3; Additional file 2: Table S6), incorporating distinct monooxygenases, have previously been described in Parageobacillus thermoglucosidasius [58]. The genomes of all P. thermoglucosidasius and two S. caldoxylosilyticus strains incorporate all three complete desulphurisation clusters, while those of 7/9 S. caldoxylosilyticus strains and the two Parageobacillus genomosp. A isolates incorporate complete desulphurisation clusters 2 and 3 (Fig. 6). S. caldoxylosilyticus VR-IP and P. toebii WCH70 harboured only a complete desulphurisation cluster 3, while none were observed in the other P. toebii strains (Fig. 6).

Long-chain alkanes form a major component of crude oils. Several studies have identified the presence and activity of genes associated with variable length long-chain alkane catabolism in Geobacillus and Parageobacillus taxa [59]. Orthologues of LadAα (ART30136: 66.62–67.69% AAI range) and LadAβ (ART30139: 70.85–71.86% AAI range) and LadB (ART30142: 61.81–66.54% AAI range) that contribute to C10-C30 n-alkane utilisation in P. toebii B1024 [59] are encoded on the genomes of all P. thermoglucosidasius strains analysed, Saccharococcus genomosp. A KH3-5, S. caldoxylosilyticus ER4B. Both LadA orthologues, but no LadB orthologues are present in Parageobacillus genomosp. A KH3-4 and W-2 (Fig. 6; Additional file 2: Table S6). In accompaniment, at least one putative aldehyde dehydrogenase (ABO68462; 78.89–94.52% AAI range) and three alcohol dehydrogenase orthologues (ABO66657, ABO67118 and ABO68223; 76.77–79.42%, 79.76–86.10% and 86.7–91.3% AAI ranges, respectively), assumed to participate in LadA-initiated metabolism of long-chain alkanes in G. thermodenitrificans NG80-2 [59], were detected across all strains except S. thermophilus DSM 4749, which did not encode orthologues of ABO68462 or ABO67118.

Nitroalkanes are another group of highly recalcitrant compounds, utilized as fuels, solvents, herbicides and pesticides which are also toxic and carcinogenic [60]. Recently, three nitroalkane oxidizing enzymes (WP_064553126, WP_064551563 WP_064551165) were shown to variably degrade nitropropane and nitroethane in Parageobacillus genomosp. A W-2 [60]. Orthologues of each enzyme (Gt2929, Gt1378 and Gt1208; 55.20–100%, 76.96–100% and 88.73–100% AAI ranges, respectively) are encoded on the genomes of 34, 33 and all 35 of the analysed strains, respectively (Fig. 6; Additional file 2: Table S6).

Parageobacillus as a producer of green energy

In part due to its capacity for biomass degradation, as well as its fermentation pathways, P. thermoglucosidasius has received extensive interest for the production of biofuels (Fig. 6). In particular, ethanol production has been widely researched, but as a mixed acid fermenter with limited ethanol tolerance, metabolic engineering of this species is required [61].

Another P. thermoglucosidasius fermentation product of biotechnological interest is isobutanol, which can serve as biofuel, fuel additive or as a primer for the production of chemicals [6, 62]. The final step in isobutanol formation from isobutyraldehyde involves an isobutayraldehyde dehydrogenase, with two putative enzymes (AdhA and Geoth_3823) identified in P. thermoglucosidasius C56YS93 [62]. Orthologues of both enzymes, sharing 95.7% and 97.3%, are encoded on the genomes of all 35 and 33/35 comparator strains, respectively (Fig. 6), suggesting members of both genera could serve as targets for metabolic engineering for isobutanol production. P. thermoglucosidasius also produces 2,3-butanediol (2,3-BDO), which can be used as liquid fuel, fuel additive or chemically modified to produce high octane isomers for use in aviation fuels [6, 63]. Orthologues of one key enzyme involved in 2,3-BDO synthesis identified in P. thermoglucosidasius NCIMB 11955 [63], namely acetolactate synthase (ALS), were observed in all compared taxa (93.1% AAI; Fig. 6). The final enzyme in 2,3-BDO synthesis, butanediol dehydrogenase (BDH), was restricted to the genus Parageobacillus (97.9%), with a single copy encoded on the genome of 23/24 taxa, with the exception of P. toebii WCH70, where two copies exist (96.6% amino acid identity between copies).

Recent interest has focused on the production of hydrogen gas, an environmentally friendly and sustainable alternative energy carrier, from carbon monoxide-containing waste gases by P. thermoglucosidasius [64]. This biological water–gas shift reaction (WGS) involves an enzyme complex comprising a carbon-monoxide dehydrogenase (CODH) and hydrogen-evolving hydrogenase [64]. Previous analyses showed the CODH-hydrogenase locus to be restricted to P. thermoglucosidasius. Analysis of our annotated dataset showed homologous loci in all fourteen P. thermoglucosidasius taxa, with the CODH proteins CooCSF and hydrogenase proteins PhcABCDEFGHIJKL sharing 99.5% and 99.1% AAI, respectively among these taxa, while no orthologues were found in any other Parageobacillus or Saccharococcus taxa (Fig. 6). However, a recent study identified the CODH-hydrogenase locus on the genome of Parageobacillus sp. G301 [65]. ANI and dDDH values (97.24% and 76.3%, respectively with P. toebii DSM 14590 T) indicate that this strain belongs to the species P. toebii, and the CODH and hydrogenase proteins share an average AAI of 91.5% and 91.2% with those of the fourteen P. thermoglucosidasius taxa. As such, broader evaluation of the genera Parageobacillus and Saccharococcus for hydrogen-evolving systems of potential biotechnological value is warranted.

Conclusions

Phylogenomic analysis delineates Parageobacillus and Saccharococcus as two distinct genera, both of which present open pan-genomes. P. toebii in particular presents the greatest potential for novel gene accrual within Parageobacillus. Plasmids, bacteriophages and transposable elements are key drivers of genomic and functional, diversification of these genera. Both Parageobacillus and Saccharococcus harbour a wealth of biotechnological potential including potential novel antimicrobials and a range of thermostable enzymes. Functional and in vivo analyses of the novel antimicrobial peptides should serve to validate the potential of the studied taxa to contribute towards combatting antibiotic-resistant target bacteria. Similarly, the broad range of carbohydrate-, protein- and lipid-active enzymes, identified here and in previous studies, should be evaluated to expand the current repertoire of thermostable enzymes for a wide array of biotechnological applications. Our analyses have also further highlighted the potential for members of both Parageobacillus and Saccharococcus in a broad spectrum of whole-cell applications, including bioremediation of various xenobiotic compounds and environmental pollutants, the degradation of lignocellulosic biomass to generate various value-added products, as well as the use of these taxa to contribute towards the green energy market. Given the extensive genomic variability and the potential biotechnological pathways and enzyme complement, additional discovery and characterization, both genomic and functional, of novel Parageobacillus and Saccharococcus isolates will continue to expand the biotechnological toolkit of these intriguing genera.

Methodology

Genome assembly and annotation

The publicly available genome sequences of thirty-four Parageobacillus taxa, Saccharococcus thermophilus DSM 4749 T and G. thermodenitrificans DSM 465 T (used for comparative and outgroup purposes) were obtained from the NCBI genome assembly database [66]. Average Nucleotide Identity (ANI) values of all draft genomes were calculated with the OAT tool v. 0.9.1 [67]. The genome assemblies were subsequently improved using the MeDuSa genome scaffolder v. 1.6 [68], where the genome of the taxon sharing the highest ANI value (complete genome) was used as reference genome. All genomes were structurally annotated using Prodigal v.2.6.3 [69] and the proteome datasets were functionally annotated (and assigned to COG categories) using eggnog-mapper v. 2.1.12 [70] against the eggNOG v. 5.0 database [71]. The subcellular localisations of all proteins encoded on each genome were determined using PSORTb v. 3.0.3 [72]. Plasmids and transposable elements were identified on the basis of the eggNOG annotations, while phage elements were identified using the PHASTER server [73].

Biotechnologically relevant enzymes were identified and characterised using several pipelines. Secondary metabolite biosynthetic loci were identified using antiSMASH v. 7.0.1 [16] and further confirmed and characterised using the BAGEL 4 [17] and RiPPMiner-Genome [18] servers. CAZYmes were predicted from the protein datasets for each genome using the HMMer, Hotpep and DIAMOND tools of DbCAN3 [74] against the CAZYme database [31], where only those predictions made by ≥ 2 tools were considered as positive hits. Proteases/peptidases and lipases were identified and characterised by aligning the proteome datasets for each compared Parageobacillus and Saccharococcus strain against the MEROPS v. 11.0 database [45] and the Lipase Engineering Database (LED) v. 4.1.0 [41], respectively, using DIAMOND v. 2.1.8 [75]. CRISPR-Cas associated proteins were predicted through the CRISPRCasFinder tool v. 1.1.2—I2BC [56]. Other proteins of potential biotechnological relevance were identified by localized Blast analyses and alignment using Bioedit v. 7.7.1 [76]. Restriction-modification systems were tentatively identified on the basis of the eggnog-mapper annotations and confirmed through Blastp analysis against the REBASE database [52].

Phylogenomic analyses

The proteome datasets for each comparator strain (and G. thermodenitrificans DSM 465 T as outgroup) were compared and clustered into their orthologous groups using Orthofinder v. 2.5.5 [77]. Single copy orthologous (SCO) proteins conserved among all taxa (1,784 SCOs) were individually aligned using the M-Coffee implementation of T-Coffee v. 13.46.0.919e8c6b [78], concatenated and poorly aligned blocks were removed using GBlocks v. 0.91b [79]. The trimmed concatenated alignment was used to construct a maximum likelihood (ML) phylogeny using IQ-Tree v. 2.2.0 [80], with the optimal evolutionary model predicted using ModelFinder [81]. Branch support was provided using ultrafast bootstrap (UFBoot2) analysis (n = 1,000 replicates) [82]. Support for the core genome phylogeny and species delineation was provided by calculating the Average Nucleotide Identity (ANI) values with the OAT tool v. 0.9.1 [67] and digital DNA-DNA hybridization values (dDDH) were determined using the Genome-to-Genome Distance Calculator (GGDC 3.0) [83], where taxa sharing OrthoANI values > 96% and dDDH values > 70% were considered to belong to the same species [8, 67, 83].

Pan-genome analyses

The Orthofinder output was used to identify the core (conserved among all taxa), accessory (shared by several but not all compared strains) and unique (to a single taxon) proteome fractions of the compared Parageobacillus and Saccharococcus taxa. The presence (1) or absence (0) of each orthogroup was scored and the pan-genome of different datasets (Parageobacillus/Saccharococcus; P. thermoglucosidasius/P. toebii/S. caldoxylosilyticus) were used to determine the pan-genome using the bacterial pan-genome analysis (BPGA) pipeline [84] and extrapolated (to 100 genomes/per set of taxa) using PanGP [85]. The functions of the core, accessory and unique pan-genome fractions were determined by comparison of the pan-genome element-specific proteome datasets against the eggNOG v. 5.0 database [71] using eggnog-mapper v. 2.1.12 [70].