**1. Introduction**

More than 80 years after the first reported case of sulphonamide-resistant bacterial strains [1], antibiotic resistance is now a global threat to human health, projected to cause 10 million deaths annually by 2050 [2,3]. Historically, microorganisms have been a source of over 22,000 biologically active metabolites, including antibiotics, isolated from terrestrial and marine strains [4]. In particular, the order Actinomycetales (actinomycetes) has been shown to produce structurally diverse specialised metabolites which exhibit a wide range of biological activities. Indeed, more than 7000 metabolites have been isolated from the genus *Streptomyces* and approximately 3000 metabolites from the "rare" (due to their lower isolation frequency) actinomycete genera [5]. Actinomycetes have an average genome size of over 5 Mb. However, this number varies greatly between genomes which can reach up to 12 Mb in some *Streptomyces* species [6]. Actinomycetes dedicate 0.8–3.0 Mb of their whole genome to specialised metabolite production, which has been shown to result in 20–50 Biosynthetic Gene Clusters (BGCs) per strain [7]. A recent study of 21 rare marine actinomycetes isolated from temperate and sub-tropic marine environments unveiled

**Citation:** Soldatou, S.; Eldjárn, G.H.; Ramsay, A.; van der Hooft, J.J.J.; Hughes, A.H.; Rogers, S.; Duncan, K.R. Comparative Metabologenomics Analysis of Polar Actinomycetes. *Mar. Drugs* **2021**, *19*, 103. https://doi. org/10.3390/md19020103

Academic Editor: Max Crüsemann

Received: 21 January 2021 Accepted: 8 February 2021 Published: 10 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

diverse and numerous BGCs, with *Actinomadura* spp. and *Nocardia* spp. containing 44 and 38 BGCs per strain, respectively. Interestingly, only three percent of the BGC families overlapped between the *Streptomyces* strains and the rare actinomycetes, suggesting an exciting resource for biosynthetic novelty [8]. Moreover, analysis of 119 genomes of the rare actinomycete genus *Salinispora* derived from various sub-tropic and tropic locations revealed 176 BGCs, of which only 24 were linked to their respective products [9], indicating a potential resource for novel metabolites.

Although actinomycetes from the (sub-)tropics have been extensively studied and shown to be a promising source of biological and chemical novelty, Actinobacteria isolated from Polar regions also show potential for affording new biologically active specialised metabolites. Actinobacteria accounted for 5% of the total microbial community present in Antarctic sediment samples as revealed by high-throughput 16S rRNA gene sequencing, with the *Salinibacterium* genus being amongst the most abundant [10]. In terms of isolating rare actinomycetes from Polar habitats, a study of ancient Antarctic and sub-Arctic sediment samples yielded 50 bacterial strains, of which 39 belonged to rare actinomycetes genera (*Microbacterium*, *Dietzia*, *Rhodococcus*, and *Pseudonocardia*) [11]. Regarding the chemical potential, molecular networking indicated rare actinomycetes from sub-Arctic and Antarctic sediments to be a rich source of metabolites [11,12]. Since 2001, a total of twenty-nine new metabolites have been isolated from Antarctic and sub-Arctic bacteria, with 13 being produced by marine actinomycetes [13]. For example, the rare actinomycete *Nocardia dassonvillei* BM-17 isolated from an Arctic sediment sample yielded a new phenazine derivative with significant antifungal and cytotoxic activity [14]. Two new α-pyrones were isolated from an Antarctic *Nocardiopsis* strain [15], whereas the seaweed-derived *Nocardiopsis* sp. 03N67 collected in the Arctic Ocean afforded a new diketopiperazine, cyclo-(L-Pro-L-Met), which showed promising anti-angiogenesis activity [16].

It is widely known that variations in cultivation parameters can induce the expression of so-called 'silent' BGCs, for which the biosynthetic enzymes have been identified but no natural product has been isolated from laboratory cultures [17]. Hence, changes in abiotic factors such as nutrient availability (carbon, nitrogen, trace-elements), temperature, salinity, and pH have been shown to influence the production of such 'cryptic' specialised metabolites. Moreover, the chemical profile of a microorganism can depend on the culture vessels, shaking conditions, and aeration, as well as on co-culturing techniques [18]. For example, when glucose was substituted for glycerol in the ISP2 growth medium, the liquid culture of *Streptomyces* sp. C34 isolated from the Atacama Desert, yielded ansamycintype polyketides [19]. Furthermore, the marine-derived *Streptomyces* sp. CHQ-64 was found to produce new biologically active polyene-polyols and hybrid isoprenoid alkaloids when cultured under shaking, whereas static fermentation yielded only one new metabolite [20,21]. Therefore, the "One Strain Many Compounds" (OSMAC) approach [22] has been a successful addition in the microbial drug discovery pipeline

Mass spectrometry approaches are often used to compare multi-strain metabolomics. Molecular networking using the Global Natural Products Social Molecular Networking (GNPS) infrastructure [23] has been deemed a valuable tool in the discovery of new metabolites [24] as it provides rapid dereplication [25] and identification of unknown parent ions. Clarinoside, a new pentalogin from the plant *Mitracarpus scaber* Zucc [26], retimycin A, a non-ribosomal peptide from *Salinispora arenicola* [27], and deoxyphorbol ester derivatives from *Euphorbia dendroides* [28] are a few of the specialised metabolites discovered using MS-guided isolation based on the GNPS platform. Recently, further analysis tools have been implemented on the GNPS infrastructure, such as MS2LDA which provides fragmentation patterns of commonly co-occurring mass fragment peaks and/or neutral losses that often represent molecular substructures (Mass2Motifs) [29], and Network Annotation Propagation (NAP) [30] which improves in silico fragmentation of the input data. The MolNetEnhancer workflow was introduced to combine the outputs of the above-mentioned tools and add (putative) chemical class annotations to molecular families in the molecular network [31].

The sequencing of the first *Streptomyces* genome in 2002 [32] paved the way for the discovery of further microbial natural products based on genomic data. The continuous development of genome sequencing technology has led to a wealth of genomic data which has motivated the development of sophisticated mining tools that can augment the search and discovery of novel specialised metabolites [33,34]. Several of these [35] are publicly accessible, enabling thorough and targeted genome mining of complex bacterial genomes, including antiSMASH for the identification of secondary metabolites BGCs [36], ARTS for high-throughput screening of bacterial genomes in reference to antibiotic production [37], and BiG-SCAPE for clustering BGCs into Gene Cluster Families (GCFs) [38].

The linking of genome and metabolome mining outcomes to accelerate natural products discovery has shown great promise over the last decade, with several tools contributing to bridging the gap between BGCs and mass spectra [39]. Although the term "metabologenomics" was officially introduced in 2016 to describe correlations between BGCs and metabolites [40], research on this approach started earlier. A study of actinomycete strains in which their genomic data were linked with MS profiles, led to the identification of GCFs for the previously reported natural products desertomycins and oasamycins, for which the corresponding BGCs were unknown [41]. The authors were also able to isolate and characterise a new chlorinated metabolite, tambromycin, and correlate it with its BGC in 11 actinomycete strains using metabologenomics [40]. Another successful study used a combination of molecular networking and pattern-based genome mining approach from which arenicolide A was linked to an uncharacterised BGC (PKS28) and the new metabolite, retimycin A was identified, characterised and linked to the known NRPS40 pathway [27].

Herein, we introduce a novel unsupervised -*omics* integration method to link tandem mass spectrometry data to BGCs to accelerate the analysis of large microbial natural products datasets. NPlinker, a newly introduced software framework [42] was applied for the first time to bridge the large metabolomics and genomics datasets of marine Polar Actinobacteria. With the aid of the novel approach Rosetta, links between spectra and BGCs for chloramphenicol and ectoine were established. Molecular networking of the 100 metabolite extracts derived from applying the OSMAC approach, showed growth media specificity and potential chemical novelty was suggested. Moreover, the metabolite extracts were screened for antibacterial activity and promising selective bioactivity against drug-persistent pathogens such as *Klebsiella pneumoniae* and *Acinetobacter baumannii* was observed.

#### **2. Results**
