Mining the microbiome of Lake Afdera to gain insights into microbial diversity and biosynthetic potential

Abstract Microorganisms inhabiting hypersaline environments have received significant attention due to their ability to thrive under poly-extreme conditions, including high salinity, elevated temperatures and heavy metal stress. They are believed to possess biosynthetic gene clusters (BGCs) that encode secondary metabolites as survival strategy and offer potential biotechnological applications. In this study, we mined BGCs in shotgun metagenomic sequences generated from Lake Afdera, a hypersaline lake in the Afar Depression, Ethiopia. The microbiome of Lake Afdera is predominantly bacterial, with Acinetobacter (18.6%) and Pseudomonas (11.8%) being ubiquitously detected. A total of 94 distinct BGCs were identified in the metagenomic data. These BGCs are found to encode secondary metabolites with two main categories of functions: (i) potential pharmaceutical applications (nonribosomal peptide synthase NRPs, polyketide synthase, others) and (ii) miscellaneous roles conferring adaptation to extreme environment (bacteriocins, ectoine, others). Notably, NRPs (20.6%) and bacteriocins (10.6%) were the most abundant. Furthermore, our metagenomic analysis predicted gene clusters that enable microbes to defend against a wide range of toxic metals, oxidative stress and osmotic stress. These findings suggest that Lake Afdera is a rich biological reservoir, with the predicted BGCs playing critical role in the survival and adaptation of extremophiles.


Introduction
The microbiome of hypersaline environments is consists of specialized micr oor ganisms known as halophiles.(Waditee-Sirisattha et al. 2016 ).Halophiles occupy a diverse ecological niches, including microbial mats, saline soils, hypersaline soda lakes, brine pools, and salt Lakes (Weimer andRompato 2009 , Chen et al. 2020 ).To survive in these en vironments , halophiles ha v e de v eloped adaptations to multiple stressors, such as ionic stress, osmotic stress, desiccation stress and carbon-poor conditions (Corral et al. 2020 ).These adaptations include various genetic mechanisms , for example , efflux pumps , whic h help micr obes to acclimatize to their envir onment (Or en 2015 ).Additionall y, man y surviv al mec hanisms involv e the pr oduction of secondary metabolites, whic h potentiall y serv e as sources of biotec hnologicall y valuable molecules with a wide range of applications.For instance, halophiles produce stable enzymes beneficial in different industrial pr ocesses, r anging fr om biopol ymer pr oduction to bior emediation (Makar ov a et al. 2019 ), cr eation of macr omolecule stabilizers and biofertilizers (Dassarma et al. 2010, Amoozegar et al. 2017 ) to a large number of applications in fermented food products.(Yin et al. 2015, Kiadehi et al. 2018 ).Despite their commercial potential, halophiles r emain r elativ el y under explor ed in terms of their capacity to produce antimicro-bial and anticancer drugs (Charlesworth andBurns 2015 , Corral et al. 2020 ).
Biosynthetic gene clusters (BGCs) are a locally clustered group of two or more genes that confers a competitive adv anta ges to micr oor ganisms by encoding secondary metabolites (Medema et al. 2015 ).These clusters encompass a variety of chemical and biological groups, including non-ribosomal peptide synthetases (NRPS), polyketide synthases (PKS), terpenes, and bacteriocins (Wang et al. 2019 ).NRPS and PKS are particularly important targets for natur al pr oduct discov ery as they ar e known to synthesize a div erse array of antimicrobials and pharmaceutical products (Tillett et al. 2000, Wang et al. 2014, Chen et al. 2020 ).The condensation (C) and ketosynthase (KS) domains within these clusters serve as conserved genomic markers, essential to distinguish between differ ent NRPS/PKS natur al pr oduct pathways (Ziemert et al. 2012 ).Pr e viousl y, the detection of these pathways has led to the identification of compounds with unique c hemistry, suc h as lactocillin (Donia et al. 2014 ) and salinilactam (Udwary et al. 2007 ) antibiotics.Curr entl y, r esearc h on undiscov er ed BGCs involv es an increasing focus on genomic and metagenomic analysis (Makarova et al. 2019 ).These a ppr oac hes enhance the r econstruction of nearcomplete genomes de novo, enabling the potential r ecov ery of novel BGCs in both culturable and uncultivable microorganisms.One such interesting niche to discover bioactive compounds using this method is the brine pool of Afdera.The microbiology of this lake r emains unexplor ed although halophilic micr oor ganisms have been detected by molecular studies carried out in commercial salt from the salterns bordering the lake (Gibtan et al. 2017 ).
The brines in the Afar depression, including those in Lake Afder a, ar e enric hed with differ ent metals, suc h as lithium, whic h is of interest for industrial applications and r epr esents an additional selectiv e pr essur e (Bekele and Sc hmer old 2020 ).Lake Afder a, the lar gest of brine lakes in the Afar depression, is characterized for its maximum depth (80 m), volume (2.4 km 3 ) and ele v ation (-112 m) (Fig. 1 ).Pr e vious studies have isolated and c har acterized thermostable amylase-pr oducing bacteria (Yassin et al. 2021 ) and fungi (Welday et al. 2014 ) from soil samples of Afdera.Ho w ever, no studies have yet explored the potential significance of BGCs in this lake .T he present study aims to determine the diversity of BGCs from the microbial populations recov er ed fr om Lake Afder a to pr ovide insights into their genetic potential and identify the adaptation functions of these BGCs in this ecosystem.To ac hie v e this, we conducted in-depth analyses of metagenome assembled genomes (MAGs) using antibiotics and Secondary Metabolites Analysis SHell (antiSMASH), BActeriocin GEnome mining tooL (BAGEL4) and Natur al Pr oduct Domain Seeker (NaPDoS) pipelines.We present the microbiome composition of Lake Afder a. Additionall y, we hav e identified biotec hnologically significant BGCs in the metagenomic data that are pro-posed to play important roles in adaptation to poly-extreme environments.

Sampling and measurement of physicochemical parameters
During two separate field trips, triplicate water (4000 mL) samples were collected directly from the edge of Lake Afdera.The first field trip took place in April 2021 at 0703016 E/1462263 N and samples were collected across transects of the lake.In the second field trip in October 2021, additional samples were collected at 0703009 E/1462256 N. Physicochemical parameters (pH, total dissolved solid (TDS), temperature and salinity) were measured in situ using a portable r efr actometer (HI-9829-02 advanced portable m ulti-par ameter pH/ISE/EC/DO/Turbidity water pr oof meter, Eden Way, United Kingdom).All sampling points and mapping data wer e geor efer enced using a Garmin ® handheld GPSMAP64.The samples were collected under stringent aseptic conditions and sterilized glass bottles, which were then capped and sealed with par afilm ta pe.Bottles wer e tr ansported in an ice box and stor ed at −20 • C until further analysis.Samples for culture assays were k e pt at 4 • C.

Determination of elemental composition
The elemental composition of the lake was determined via use of an Inductiv el y Coupled Plasma Optical Emission Spectrome-ter (ICP-OES), (Agilent 5100 SVDV ICP-OES from the USA, conforming to ES ISO 11885:2007 standards).The detection limit was 0.01 μg L −1 .The standar d w orking parameters were selected and a published outlined pr ocedur es wer e follo w ed (Wiel 2003 ).Prior to analysis, the samples (50 mL) were subjected to digestion at 80 • C with 10 mL of nitric acid, cooled, filtered and diluted to 100 mL with distill water.

Enrichment media and isolation of halophilic bacteria
Culture media was formulated based on a classical halophile miner al gr owth medium, incor por ating the physico-c hemical c har acteristics of Lake Afdera's en vironment.T his media design follo w ed a modified published method (Kiki 2016, Belilla et al. 2019 ).A nutrient broth (NB 1090 media), using a ratio of 10:90 samples to broth, w as emplo y ed.The nutrient broth w as pr epar ed using artificial sea water, comprising of NaCl 100 g, MgCl 2 8 g, MgSO 4 20 g, CaCl 2 0.5 g, KCl 2.5 g, FeSO 4 1 g, dissolved in sterilized distilled water to ac hie v e a final volume of 1000 ml.The mixture was then autoclav ed, cooled and pour ed into sterilized falcon tubes containing the samples, the falcon tube containing the mix were subjected to incubation at 37 • C for 7 days in an orbital shaker.For bacterial isolation, 100 μL of the broth was streaked onto an agar medium composed of starch 15 g, glucose 2.5 g, yeast extract 2.5 g, agar 20 g and sterilized artificial seawater 1000 ml, pH adjusted at 5.8.The plates were incubated at 37ºC for a period of 7-14 days.

Genomic DNA extraction and 16S rRNA gene sequencing
Genomic DN A w as extracted using GeneMark bacterial DN A purification kit.Appropriate liquid media cultures for bacterial DNA extr actions wer e pr epar ed and incubated at 37 • C, 150 r pm for 7 da ys .T he extraction w as completed follo wing the manufacturer's instructions provided in the kit.The 16S rRNA gene was sequenced using ABI 3730XL platform, service provided by Inqaba Biotec, Pr etoria, South Africa.Sequence c hr omatogr am anal ysis was performed using Finc hTV anal ysis softwar e. Taxonomic identification was conducted using the NCBI BLASTN server.A phylogenetic tree was constructed using maximum likelihood method using Kim ur a-2 par ameter model.

En vironmental DN A extr action and metagenomic sequencing
The water samples (1000 mL) were sequentially filtered through 0.45 and 0.22 μm GE ® polycarbonate filter membrane .T he membr anes fr om 0.22 μm wer e cut into small pieces in sterilized condition and DNA extraction was performed using an optimized CTAB method (Zhou et al. 1996 ).The DN A extractions w ere performed in triplicates and r eplicates wer e later pooled prior to metagenome sequencing.DNA quality and quantity was examined with a Thermo Scientific NanoDrop 3300 Fluorospectrometer (Thermofisher Scientific, USA).The extracted DN A w as randoml y shear ed into short fr a gments and ligated with Illumina adapters to construct a library.The libraries were pooled, barcoded and subsequently shotgun sequenced on one lane of a flow cell using a 150 bp paired-end run on a NovaSeq PE150 instrument (Illumina) at Novo-gene (Hong Kong).The sequences were de-m ultiplexed using Cassav a v.2.0 and FastQC was used for quality control checks on the sequence composition of paired-end raw reads.Trimmomatic v0.36 (Q-value ≤ 38; N > 10 bp; reads overlap with adapter > 15 bp) w as emplo y ed to r emov e low quality bases and any adapter contamination.

Assembl y, metagenome assembl y of genomes (binning) and annotation
The assembly of the reads was initially conducted using MEGAHIT (Li et al. 2015 ).Scaffolds containing "N" were removed and scaftigs wer e subsequentl y formed (Mende et al. 2012, Nielsen et al. 2014 ).The cleaned reads were then mapped to assembled scaftigs using Bowtie2 and the unutilized paired end reads were collected (Langmead and Salzberg 2012 ).Mixed assembly was carried out on unutilized reads and after which reads shorter than 500 bp wer e trimmed fr om the scaftigs and the mixed assembled units (Li et al. 2014 ).The assembled contigs were further binned into metagenome assembled genomes (MAGs) using the method describe by Yang and co-authors (Yang et al. 2021 ).MAGs (binning) was conducted using MetaBAT2 (Kang et al. 2019 ), CON-COCT (Alneberg et al. 2014 ), and MaxBin (Wu et al. 2016 ).The retrie v ed MAGs wer e pooled with DAS Tool (v1.1.1)(Sieber et al. 2018 ) and their completeness and contamination were assessed using CheckM ( ≥ 80% completeness and ≤ 10% contamination) (Parks et al. 2015 ).The metagenome was annotated using NCBI GenBank annotation pipeline (Altschup et al. 1990 ) and RAST (Rapid Annotation using Subsystem Technology) tools, employing the classic RAST annotation scheme (Aziz et al. 2008 ).Furthermore, functional classification of the predicted genes was conducted using the Cluster of Orthologous Groups (COG) framework (Tatusov et al. 2003 ).

Taxonomic Assignment of Contigs
The taxonomic diversity of the assembled contigs was performed by comparing metagenomic reads based on sequence or phylogenetic similarity to the database sequence of taxonomically informative gene families (microNR database) (Buchfink et al. 2014 ).The taxonomic annotation of each metagenomic homolog was then carried out using MEtaGenome Anal yzer comm unity edition (MEGAN) (Huson et al. 2011 ).MEGAN allocated the reads onto the NCBI taxonomy using settings of the Lo w est Common Ancestor (LCA) algorithm.Tree file extracted from MEGAN was uploaded to an online Inter activ e Tr ee Of Life (iTOL) v ersion 5.0 (Letunic and Bork 2021 ) and circular phylogenetic tree was constructed.

Detection of BGCs
The MAGs were subjected to antiSMASH 6.0.1 (Blin et al. 2021 ) to mine BGCs and contigs of size equal to or larger than 1000 bp were utilized.The RAST-web tool server and NCBI Gen-Bank annotation pipeline were utilized to identify various proteins and genes responsible for adaptation to extreme en vironment.T he identified BGCs were compiled, and a stacked bar chart was generated using R studio, with visualization created in ggplot2 and R Color Br e wer pac ka ges.

Detection of bacteriocins and domains of NRPS and KS
B AGEL4 w as used to assess the bacteriocin and ribosomally synthesized and posttranslational modified peptides (RiPPs) within the investigated MAGs (Heel et al. 2018 ).Potential clusters and annotated classifications were identified as areas of interest (AOI).In addition, NaPdoS was used to search potential known natural product biosynthetic domains, specifically C and KS, by comparing them to a domain database of pr e viousl y c har acterized natur al products (Ziemert et al. 2012 ).Subsequently, a circular phylogenetic tree was constructed for all C and KS domains using NaP-DoS.The r esulting tr ees wer e visuall y pr esented and annotated Table 1.Physico-c hemical measur ements and elemental content (mg/L) of Lake Afdera, water samples taken in April, 2021.

Measured parameters Measured value
Temper atur e (

Physico-chemical measurements, elemental composition and enrichment media
Physico-c hemical measur ements (Table 1 ) indicate that Lake Afder a is pr edominantl y saline, with a slightl y acidic pH.The GPS meter show that the lake is one of the lo w est land areas in Africa.The elemental composition r e v ealed that the lake is contaminated with heavy metals (Table 1 ), including iron (Fe), lead (Pb), zinc (Zn), and nickel (Ni).Additionall y, a ppr eciable amount (15.7 mg/L) of lithium (Li) was detected in the lake.In this study, bacterial species similar to Bacillus were isolated using an optimized halophilic enrichment media called NB 1090.The sequencing results from 16S rRNA gene identified EAS001 as closel y r elated to Lysinibaccilus fusilformis and EAF001 as Bacillus cereus .The phylogenetic tree also displayed closely related species (Fig. 2 ).

Metagenome data analysis and microbial community compositions
A total of 45081590 reads were generated with GC content of 56% (Table 2 ).These r eads wer e subsequentl y assembled into 93220 contig sequences, with a total contig length of 108548798 bp (Ta-ble 2 ).From the metagenomic assembly (binning), 17 MAGs were generated and following inspection using CheckM, ten candidate MAGs r epr esenting differ ent assigned phyla wer e selected.These wer e c hosen for detection and investigation of BGCs and environmental adaptation mechanisms ( Table S7 ).
Additionall y, the arc haeal population was r epr esented in the metagenome sequences, with archaeal phyla such as Candidatus Nanohaloarc haeota, Cr enarc haeota, and Euryarc haeota being detected.In addition to these phyla, we identified a small percentage (less than 1%) of the community at the class le v el.These include Nanohaloarc haea, Thermopr otei, Halobacteria, Methanobacteria, Methanomicrobia, and Thermococci.

Secondary metabolism BGCs
A total of 68 BGCs were detected from 10 MAGs using antiSMASH (Table 3 and Table S1 ).Among these, 20.6% of the BGCs encoded for non-ribosomal peptide-synthetase (NRPS), showing 100% gene similarity to bacitracin and lichenysin from the MIBIG most similarity cluster database index ( Table S1 ).

Analysis of KS and C Domains
Compounds such as fengycin, bacitracin and lichenysin contributed over 80% of the total C domains detected.In addition, se v er al other biologicall y activ e compounds, lik e syringom ycin, complestatin, mycosubtilin, microcystin, nostopeptolide, bacillibactin, p y over dine w ere detected in the MAGs assigned for Bacilli, Gamma pr oteobacteria and Cyanobacteria (Fig. 7 , Table S4 ).F igure 4. Cir cular phylogeny tree for metagenomically detected microbes from water samples collected from Lake Afdera Table 3.The BGC types detected in the water samples of Lake Afdera and their respective biotechnological and industrial applications.The KS domain from the analyzed MAGs aligned closely with aryl pol yene, c haetoglobosin, iter ativ e cis-AT and modular cis-AT as well as Fatty acid synthesis (FAS) (Fig. 8 , Supplementary Table S3 and S5 ).

Genomic insights into microbial survi va bility mechanisms
Through annotation using NCBI GenBank pipeline, RAST tool and COG database, we identified various putative microbial genes, proteins and compounds associated with metal resistance and adaptation to the metal-rich environments of Lake Afdera.Genes encoding copper oxidases ( copA ) and copper resistance proteins ( copB, copC, copD, copG ) were detected ( Table S6 ).Additionally, the presence of genes such as ferric uptake regulator ( fur ) and Febacillibactin uptake systems ( F euABC ) suggests r egulatory mec hanisms for managing excess concentration of iron ( Tables S6 , S8 ).Our sequence analysis also revealed multiple gene clusters encoding pr oteins involv ed in mercury r esistance, including mercuric ion reductase ( merA ) and genes regulating transcriptional  S6 ).Furthermore , multi-hea vy metal efflux proteins such as magnesium-cobalt efflux protein CorC and the cobalt-zinc-cadmium resistance protein CzcD were detected ( Tables S6 , S8 ).

Discussion
The advent of genome mining tools for studying microbial genomes has led to the widespread identification of BGCs across the bacterial domain (Cimermancic et al. 2014 ).BGCs are distributed in various gene cluster families and play crucial role in pr oviding micr obial comm unities with the ability to adapt to ex-tr eme envir onments (Wang et al. 2019 ).Her ein, we pr esent the first metagenome mining study of BGCs from understudied micr oor ganisms in Lake Afdera, shedding lights into their chemical potential as well as their adaptation mechanisms within the pol y-extr eme habitat.Mor eov er, we hav e de v eloped an inexpensive classical halophile mineral growth medium, NB 1090, by modifying a published method (Kiki 2016, Belilla et al. 2019 ).Such isolation media may be instrumental for cultivation of microbes from hypersaline en vironments .Our optimized method enriched micr obial species r elated to B. cereus and L. fusiformis as evidenced from the sequenced 16S rRNA gene data.This finding aligns with pr e vious r esearc h on Afder a soil, whic h r eported thermostable amylases from culturally isolated Bacillus species (Yassin et al. 2021 ).
The microbiome of Lake Afdera mainly consists of the bacterial phyla such as Pseudomonadota, Actinomycetota, Bacilliota, Bacteroidota and Cyanobacteria.This finding is consistent with similar studies conducted in hyper-arid regions of Atacama desert, where these phyla were frequently observed (Azua-Bustos et al. 2015, Orellana et al. 2018 ).Additionally, a 16S rRNA gene survey of commercial salts extracted from Lake Afdera revealed an abundance of Pseudomonadota, Bacteroidota, Actinomycetota and Bacilliota (Gibtan et al. 2017 ) .A broad bacterial (and to a lesser extent, arc haeal) div ersity was identified in the saline ric h m ulti-extr eme envir onment of the Afar Depr ession.This div ersity might r esult fr om m ultiple independent ada ptation mec hanisms within the archaeal community, which appear to contrast with the extensive molecular adaptations observed in the bacterial do-Figure 8. Phylogenetic tree constructed to analyze ketosynthase (KS) domains against the NaPDoS domain database using maximum likelihood method.The outer ring of the tree represents natural products, which are shaded according to their bioactivity and domains also color coded.Most KS domains are aligned closely to fatty acid synthesis, antitumor and antibacterial compounds.Novel Biosynthetic Gene Clusters offer k e y role to discover unique natural products in extreme Halophilic environments main (Hallsworth et al. 2007, Ste v enson et al. 2015, Lee et al. 2018 ).An example of such an adaptation is the intracellular accumulation of K + (the 'salt-in' str ategy), whic h m ust function in conjunction with intracellular proteins adapted to these harsh conditions.Prior to the present study, it has not been shown whether the identified phylum from Lake Afdera possess biosynthetic capacity for natural compound discovery.
The MAGs were thoroughly analyzed using antiSMASH 6.0.1 to predict BGCs and a total of 68 BGCs were detected with NRPS, bacteriocins and NRP-metallophore being the most abundant among them.Among the identified taxonomies, the highest number of BGCs was detected in the MAGs associated with Actinomycetota and Bacilliota, supporting their reported biosynthetic capacity (Fig. 5 ).An antiSMASH survey of these phyla harbored the most diverse BGCs, to mention few NRPS , PKS , saccharide , β-lactone , RiPP like, CDPS ( Table S1 ).Saccharides enable bacteria to form biofilms thereby providing adaptation mechanisms from toxins and dehydration (Wolferen et al. 2018 ).CDPS and β-lactones exhibited a broad range of biological activities, including antimicrobial and anticancer activities, making them promising candidates for de v eloping ne w pharmaceuticals to combat the gr owing antibiotic resistance crisis (Li andRebuffat 2020 , Singh et al. 2021 ).Other identified BGCs and their associated functions further underscores the microbiome's potential in Lake Afdera for biotechnological applications and adaptation mechanisms, such as RiPP like (Chan and Burrows 2021 ), NRPS, PKS (Burns et al. 2005 ), terpene (Medema et al. 2015 ), NAGGN (Sagot et al. 2010 ) and ectoine (Jorge et al. 2016 ).Ectoine, in particular, enables halophiles to endure hypersaline conditions, suggesting that "salt-out" mechanism likely contributes to the survival of microbes in Lake Afdera's high salinity (Ma et al. 2010 ).
Bacteriocin BGCs were also detected using BAGEL4 ( Table S2 ).This survey has yielded se v er al catalogues of bacteriocin clusters (Fig. 6 ) with antibiotic potential (Cotter et al. 2013 ).An earlier survey of hypersaline environments revealed multiple BGCs including bacteriocins, and these clusters were hypothesized to play crucial functions in the survival of microbial community (Crawford et al. 2016, Ziko et al. 2019 ).It is possible that microbes inhabiting Lake Afdera utilize similar survival mechanisms.Further analysis of the metagenome for C and KS domains, using NaPDoS pipelines, identified antimicrobial peptides such as lichenysin, bacitr acin, bacillibactin and fengycin.Lic hen ysin, r enowned for m ulti-functionality, serv e as an efficient ion chelator and surfactant that maintains stability e v en under extreme conditions (Yeak et al. 2022 ).Additionall y, C domains originating fr om Gamma pr oteobacteria aligned with syringomycin, recognized for its antimicrobial and biosurfactant functions (Raaijmakers et al. 2010 ).Sim ultaneousl y, the majority of KS domain sequences were found to closely align with compounds related to fatty acid synthesis (F AS) (Fig. 8 ).F AS compounds support the survival of organisms in extr eme envir onments by pr oviding photopr otectiv e activity a gainst UV r adiation and shielding fr om ROS (Chen et al. 2020 ).Consequentl y, widespr ead distribution of BGCs plays vital roles in micr obial ada ptations to high UV r adiation, extr eme temperatures, salinity and desiccation stress (Wong et al. 2018, Wang et al. 2019 ).
In this study, metagenome analysis predicted multiple genes conveying metal tolerance in microorganisms.Metals serve a k e y role within these organisms, acting as catalysts, co-factors for enzymes, stabilizers for proteins and participate in several redox reactions either by donating or accepting electrons (Dopson and Holmes 2014 ).Ho w e v er, Lake Afder a is typicall y enric hed with high metal concentr ations, whic h can be toxic to micr oor ganisms.Se v er al open reading frames (ORFs) have been identified to code for putativ e pr oteins associated with copper r esistance.Notabl y, the multi-copper oxidase, copA, has been detected and found to play an important role in copper detoxification by reducing oxygen to water (Quintanar et al. 2007 ).Further metagenome analyses unveiled the presence of copper-resistance proteins, CopB, CopC, CopD and CopG, whic h ar e r ecognized for their r ole in the uptake and transport of copper to cytoplasm for subsequent expulsion (Benison 2019 ).
Other genes relating to heavy metal tolerance detected in this environment include mercuric reductase ( merA ) and its regulatory proteins .T hese proteins catalyzes the reduction of Hg(II) to volatile Hg(0), effectiv el y detoxifying the immediate micr obial environment (Barkay et al. 2010 ).The Lake Afdera metagenome also contained znuABC genes, which have been identified as involved in the uptake and intracellular regulation of Zn 2 + ions across the cell membrane (Mikhaylina et al. 2018 ), and genes encoding for multimetal resistance protein such as cobalt-zinc-cadmium CzcD and ma gnesium-cobalt efflux pr otein CorC.These r esistance systems are emplo y ed b y cells to expel multi-heavy metals out of the cell membrane (Dopson and Holmes 2014 ).
Meta genome anal ysis has identified various copies of genes r esponsible for gener ation of se v er al anti-o xidati ve enzymes, whic h ar e involv ed in pr otecting a gainst o xidati v e str ess .T his includes super oxide dism utase, glutathione per oxidase (GPx), glutathione (GSH), catalases , reductases , glutaredo xin and pero xiredoxin ( Supplementary Table S6 ).These enzymes enable bacteria's DNA, proteins, lipid membranes as well as metabolic system and other cell compartments to function efficiently (Kumar et al. 2019, Abdel-Mageed et al. 2020 ).Among other genes detected for oxidativ e str ess mana gement ar e Rbr and the alkyl hydr oper oxide r eductase subunit C, whic h conv ert endogenousl y gener ated hydr ogen per oxide into water (Imlay 2013 ).Other compounds pr edicted to be effective for the management of o xidati ve stress include trehalose, sugar molecule known for its role as stress protectant against the damage caused by en vironmental stresses , such as osmotic stress (Chen et al. 2017 ).
In hypersaline en vironments , osmotic stress significantly disturbs the internal osmotic balance of microbial life.Genome mining has identified various genes and solutes that are crucial for neutralizing osmotic stress.Among these k e y genes betT, opuAC and proW ar e r esponsible for the uptake of c holine, glycine/betaine and glycine betaine/L-proline respectively.These genes help to counteract osmotic pressure across the membrane (Kempf and Bremer 1998 ).In addition, other predicted compounds include peptide such as NAGGN and ectoines (Sagot et al. 2010 ).The presence of these genes and compounds underscores their potential importance as osmoregulators to counteract osmotic stress at Lake Afdera.

Conclusion
Our findings present the first metagenomic study that identified BGCs involved in synthesizing various classes of compounds in the microbiome of Ethiopia's Lake Afdera.These include NRPS, NRP-metallophore , bacteriocins , RiPP like and ectoine .T hese iden-tified BGCs possess a wide range of potential uses in industry and could also hold significant potential in environmental and medicinal fields.Additionally, the study isolated and identified strains of Bacillus .The sequencing analysis also highlighted BGCs associated with microbial adaptation to harsh environmental conditions .T hese genes equip micr oor ganisms to withstand challenges such as salinity, extreme temperatures, high metal concentrations, intense radiation and desiccation.Consequently, the analysis of these BGCs has r e v ealed natur al pr oducts that could be significant for the envir onmental ada ptation of Lake Afdera micr obiome, Afar Depr ession.Further r esearc h is needed to gain a detailed understanding of the molecular mechanisms underlying pol y-extr emophiles in Lake Afdera and to elucidate the anti-stress mechanisms these organisms employ.

Figure 1 .
Figure 1.(A) Satellite image of Lake Afdera, Afar Depression, northern Ethiopia and (B) Panoramic view showing the sampling location site

Figure 2 .
Figure2.Phylogenetic tree constructed using maximum likelihood method using Kim ur a-2 par ameter model.The robustness of tree was e v aluated by Bootstr a p method (1000 r eplication).Onl y bootstr a p v alues gr eater than 50% ar e shown, and the scale bar indicates 0.02 substitutions per site.

Figure 3 .
Figure 3. Stacked bar plot showing the r elativ e abundance of microbial communities in brine pool habitats of Lake Afdera, according to family-level taxonomic distribution.

Figure 5 .Figure 6 .
Figure 5. Stacked bar chart of various BGC types identified by antiSMASH in the MAGs of different assigned taxa at class level and only taxa with predicted BGCs was plotted.Gene clusters are arranged left to right based on average proportion contribution on each MAGs

Figure 7 .
Figure 7. Phylogenetic tree constructed using maximum likelihood method to analyze condensation (C) domains against the NaPDoS domain database .T he outer ring of the tree represents natural products, which are shaded according to their bioactivity.Various C domains aligned most closely to pathways encoding for antibacterial and antifungal compounds.

Table 2 .
Metagenome sequence analysis and assembly data.