Response of Microbial Communities to Naturally Occurring Radioactive Material–Contaminated Sediments: A Microcosm‐Based Study

There is a growing need to understand the potential ecological impacts of contaminants in offshore oil and gas infrastructure, especially if that infrastructure is to be left in situ as a decommissioning option. Naturally occurring radioactive material (NORM) is one type of contaminant found in solid deposits on internal surfaces of infrastructure that poses potential ecological harm if released into the marine environment. Microbes are important components of marine sediment ecosystems because they provide ecosystem services, yet the impacts of NORM contamination to these communities are not well understood. The present study aimed to investigate the response of benthic microbial communities to NORM‐contaminated scale, collected from an offshore oil and gas system, via controlled laboratory microcosm studies. Changes to microbial communities in natural sediment and sediments spiked with NORM at radium‐226 activity concentrations ranging from 9.5 to 59.8 Bq/kg (in partial equilibria with progeny) over 7 and 28 days were investigated using high‐throughput sequencing of environmental DNA extracted from experimental sediments. There were no significant differences in microbial community composition between control and scale‐spiked sediments over 7 and 28 days. However, we observed a greater presence of Firmicutes in the scale‐mixed treatment and Chloroflexi in the scale‐surface treatments after 28 days. This could suggest selection for species with contaminant tolerance or potential resilience to radiation and metal toxicity. Further research is needed to explore microbial tolerance mechanisms and their potential as indicators of effects of radionuclide‐contaminated sediments. The present study demonstrated that microcosm studies can provide valuable insights about the potential impacts of contamination from oil and gas infrastructure to sediment microbial communities. Environ Toxicol Chem 2024;43:1648–1661. © 2024 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.


INTRODUCTION
Over the last decade, there has been increased concern about the environmental risks of contaminants in offshore oil and gas infrastructure that may be left in the ocean (in situ) as a decommissioning option (Cresswell et al., 2021;MacIntosh et al., 2021MacIntosh et al., , 2022;;Melbourne-Thomas et al., 2021).The decommissioning of offshore facilities and their subsea infrastructure may result in the release of contaminants into the marine environment from various sources, either immediately or over an extended period postoperation (Watson et al., 2023).Assessing the ecological impacts of contaminants has been recognized as a global research priority with a need to further understand the impacts of offshore petroleumassociated contaminants on the marine environment (Watson et al., 2023).Naturally occurring radioactive material (NORM) is one group of contaminants that accumulate in oil and gas production infrastructure, often as solid deposits of inorganic crystalline salts (typically dominated by barium sulfate, barite [BaSO 4 ]), known as "scales" (Koppel et al., 2023).
Naturally occurring radioactive material can become incorporated into these scales and contain radionuclides from the uranium ( 238 U) and thorium ( 238 Th) decay series, including radium ( 226 Ra, 228 Ra), lead ( 210 Pb), and polonium ( 210 Po).Other metal contaminants, such as Pb and copper (Cu), are also found in NORM from corrosion products of the carbon steel pipelines.Naturally occurring radioactive material-contaminated scale can be present for years beyond cessation of operations (tens to thousands of years) because of the long physical halflife of 226 Ra (1602 years) and, if left in situ with associated infrastructure, may cause ecological harm when released to the marine environment such as after the infrastructure corrodes (Grung et al., 2009;Jensen et al., 2016;MacIntosh et al., 2021;Olsvik et al., 2012).
Contaminant impacts need to be considered in the decommissioning decision-making process.Recent research has demonstrated considerable changes in microbial communities in response to different environmental disturbances, including NORM contamination, which may have consequences for the functionality of the ecosystem (Feng et al., 2022).Numerous molecular studies have shown contamination, from the presence of elevated levels of anthropogenic radionuclides (such as cesium [ 137 Cs] and plutonium [ 239 Pu]), can impact the structure and function of bacterial communities by exerting pressure on the local microbial community (Hoyos-Hernandez et al., 2019;Theodorakopoulos et al., 2017).For example, environmental DNA (eDNA) metabarcoding has revealed the ability of some microorganisms to survive in soils contaminated with 137 Cs at Chernobyl and strontium-90 ( 90 Sr) at Fukushima, with the communities being preferentially selected for radionuclide-tolerant species (Hoyos-Hernandez et al., 2019;Theodorakopoulos et al., 2017).This phenomenon has also been observed at mining-impacted sites, including mine tailings and other waste sites (Chandler et al., 2021;Kavehei et al., 2022;Sutcliffe et al., 2017).However, no studies have utilized molecular techniques to investigate the potential effect of NORM-contaminated products from offshore petroleum infrastructure (i.e., combination of metals and radionuclides) on the local sediment microbial communities and if selective bacterial taxa may in fact play a vital role in the sequestration and biotransformation of radionuclides.
In anoxic marine conditions, there may be abiotic and biotic dissolution of NORM scales (Koppel et al., 2023).This may release insoluble barium and radium into anoxic porewaters that may become soluble to then reprecipitate.Furthermore, sulfide, generated from microbial sulfate reduction, will react with transition metals (Cu, Pb) to form metal sulfide precipitates that are insoluble.In terms of the microbial-radionuclide interaction mechanisms, radionuclides exhibit species-specific behavior in their interactions with microorganisms (Brookshaw et al., 2012).Therefore, an understanding of microbial-radionuclide speciesspecific behavior is vital to understand the functioning of microbes in NORM-contaminated marine sediments.While determining the short-and long-term ecological effects of elevated NORM contaminants on marine macroflora and fauna is a growing research area, few studies have considered or explored an ecotoxicological approach for understanding the importance of marine sediment microbes when assessing NORM-based contaminants (MacIntosh et al., 2021).
The aim of the present study was to examine the response of benthic microbial communities to the presence of NORM-contaminated scale in a controlled laboratory microcosm by examining differences in microbial composition and structure between natural sediment and scale-contaminated sediments over (a) 7-day exposure, and (b) 28-day exposure, in the presence of benthic macroinvertebrates.We hypothesized that over longterm (28-day) exposure, the presence of NORM-contaminated scale in marine sediments would lead to changes in the microbial communities.To test this, we used high-throughput sequencing of eDNA targeting the 16S ribosomal RNA (rRNA) gene to assess the microbial communities (Bacteria and Archaea) of natural marine sediments and NORM-spiked sediments with baritecontaminated scale.Finally, we discuss the suitability of eDNA metabarcoding as an ecological line of evidence in ecotoxicological studies for environmental risk assessments and monitoring offshore petroleum infrastructure postdecommissioning.

MATERIALS AND METHODS
The present study was conducted in conjunction with MacIntosh et al. (2023), where experimental sediment-based microcosms were used to understand the interactions and biogeochemical behavior of NORM-contaminated barite scale in marine sediments.A detailed chemical and radiochemical characterization of the scale, in addition to scale preparation and analysis, sediment collection, and experimental setup, is provided in MacIntosh et al. (2023).
In brief, the pulverized scale contained BaSO 4 (56.9%Ba and 13.8% S; barite), with high concentrations of the metals Cu, chromium (Cr), manganese (Mn), nickel (Ni), and Pb and activity concentrations of 226 Ra, 210 Pb, 210 Po, 228 Ra, and 228 Th as reported in MacIntosh et al. (2023).Two sediments were collected from Bonnet Bay (silty sediment with higher organic matter, 4.5%; Woronora River, Sydney, Australia) and Cronulla beach (sandy sediment with low organic matter; Oak Park, Sydney, Australia) and manually mixed in a 10:1 (720 g:80 g) sand:sediment ratio to create a sediment mixture.
For the 28-day experiment, two scale exposure treatments were tested.In the first treatment, scale was mixed throughout the sediment (scale mixed ); and in the second treatment, scale was added as a surficial layer of the sediment (scale surface ).For the 7day experiment, only a control and scale mixed treatment were tested.In both experiments the scale exposure treatments comprised 0.8 g of scale added to 800 g wet sediment (i.e., a 1:1000 dilution) per microcosm chamber.Each treatment was assigned to five replicate microcosms.Microcosms consisted of 2.35 L acid-washed and rinsed sealed plastic cylinders with the sediment and overlain by 1 L of constantly aerated filtered seawater (33 ± 2 ppt).Sediment preconditioning occurred 3 days prior to the commencement of the experiment to equilibrate without disturbance and allow the suspended particles to settle under aeration, before the start of the experiment; microcosms were incubated under ambient light conditions on a 1:23-h light: dark rotational cycle in a controlled laboratory temperature of 22 °C (MacIntosh et al., 2023).
Following the sediment preconditioning stage and prior to adding any animals, the scale mixed microcosms were prepared.Specifically, sediments were removed from microcosms and placed into clean containers.Each replicate then had approximately 0.8 g (dry wt) of pulverized (<30 μm) BaSO 4 scale added and rigorously mixed (manually using a spade) for a total of 2 h for thorough homogenization.The scale mixed sediments were then placed back into their respective microcosm with aerated overlying filtered seawater (1 L).Therefore, the scale mixed sediment treatments represented scale within well-oxygenated sediment at the commencement of the bioaccumulation experiment.The control and scale surface sediments were taken out of their respective microcosms and mixed in individual containers.After mixing, the sediments were placed back into their respective microcosm, and 1 L of overlying seawater was added to only the control microcosms (seawater was added to scale surface treatments after animals, see below).
All animals were individually weighed to create a constant biomass (total mass of organisms per taxon per replicate sediment microcosm) of each species across all the microcosms: Tellina deltoidalis 3.5 ± 0.4 g wet weight (n = 8 per chamber for 7-day test; n = 3-4 for 28-day test), Victoriopisa australiensis 0.8 ± 0.2 g wet weight (n = 8 per chamber for 7-day test; n = 3 for 28-day test).They were then added to the separate microcosms.Only Nephtys australiensis 0.9 ± 0.2 g wet weight (n = 5) was added into the microcosms during the 28-day test.The animals were given 30 min to burrow and settle into the sediment.After observed settlement (e.g., no T. deltoidalis were present on the surface) in the scale surficial microcosms, 0.8 g of the pulverized scale was weighed and added to 25 mL of seawater to create a slurry.A 5-mL aliquot of the slurry was gently pipetted into each of the scale surficial assigned microcosms (containing only 50 mL of overlying seawater when the animals were introduced).After the slurry was added, the remaining 950 mL of seawater was added and aerated.There were no observations of resuspended and floating scale.
The 7-and 28-day exposure tests were performed according to Simpson & Spadaro (2016), with modifications in relation to the presence of scale.All microcosms were maintained at approximately 22 °C for 28 days at a 1:23-h light: dark rotational cycle under ambient light conditions to simulate marine sediment conditions at >100 m water depth with aeration.Lights were only operational during monitoring or sample collection.Physicochemical parameters of overlying seawater were measured daily using a water quality probe (YSI ProQuatro Multiparameter meter) according to the manufacturer's instructions and included temperature (22 ± 1 °C), salinity (33 ± 2 ppt), dissolved oxygen saturation (>85%), and pH (8.0 ± 0.2).Nitrite (0-5 ppm), nitrate (0-5 ppm), and total ammonia (<3 mg) were measured using a rapid marine test kit (API Fish Care; LR8600).
At the completion of each experiment, sediment subsamples were collected for eDNA extraction and microbial analysis.

Description and composition of scale
The BaSO 4 scale (56.9% Ba and 13.8% S; barite) used in the present study was recovered from a decommissioned subsea pipe (well tubulars) and prepared per MacIntosh et al. (2023).The pulverized scale material was subjected to elemental and radiological analyses, with qualification of the crystalline structure via X-ray diffraction previously reported by MacIntosh et al. (2022MacIntosh et al. ( , 2023)).The scale contained high concentrations of the metals Cu (448 mg/kg), Cr (151 mg/kg), Mn (21.7 mg/kg), Ni (7.35 mg/kg), and Pb (198 mg/kg; i.e., corrosion products from the structural lining of the pipes).The total contained activity of the radionuclides 226 Ra, 210 Pb, 210 Po, 228 Ra, and 228 Th was 410 Bq/g (Table 1), with individual radionuclides exceeding the International Atomic Energy Agency (IAEA) low-level radioactive waste exemption criterion of 1 Bq/g (IAEA, 2004).

Sediments used in exposure experiments
Natural sediments were collected from two geographic areas in New South Wales, Australia (mentioned above), that had sedimentary characteristics comparable to continental shelf marine sediments where oil and gas installations are usually found (Heap & Sbaffi, 2008;Radke et al., 2017); sand and silt and were mixed to create a sediment mixture for the experiments.In the context of the present study, the sediment microcosms were preconditioned with reference sediment prior to the introduction of contaminants and organisms.This preconditioning step aimed to establish a baseline sediment environment that would allow for the settlement of suspended particles and the development of microbial communities under controlled conditions.
Experimental design for barite-contaminated sediment exposure 7-day barite sediment exposure.A 7-day-exposure, twophase experiment (7-day bioaccumulation and 7-day depuration phases) was conducted using sediment microcosms to investigate the bioaccumulation of metals and radionuclides from NORM-contaminated scale to two marine species (V.australiensis, amphipod; T. deltoidalis, bivalve; Supporting Information S1, Figure 2).This consisted of an initial 7-day bioaccumulation phase where, firstly, V. australiensis and T. deltoidalis were exposed to either control (no barite, n = 5) or barite-contaminated (scale mixed , n = 5) sediments for 7 days, with a subset being transferred to clean filtered seawater for 48 h to depurate for an additional 7 days.After the 7-day bioaccumulation exposure, four sediment cores were collected from each replicate of the control (n = 5) and baritecontaminated treatment (n = 5) groups.Four cores were taken from each replicate to ensure a representative sample of the whole microcosm was taken.Each core was briefly homogenized using a vortex mixer before being composited into a single sterile 50 mL for further homogenization (Supporting Information S1, Figure 1).Fifteen grams (wet mass) from each replicate was weighed and frozen in −80 °C within 1 h of collection.
28-day barite sediment exposure.A 28-day sediment exposure experiment was performed to test the potential ecological effects of NORM-contaminated scale to marine organisms (V.australiensis, amphipod; T. deltoidalis, clam, and N. australiensis, catworm) in marine sediments spiked with low-level concentrations of barium sulfate scale (MacIntosh et al., 2023).
At the end of the 28-day experiment reported in MacIntosh et al. ( 2023), four sediment cores (10 cm depth) were randomly subsampled in a sterile 15-mL tube from each replicate of the scale treatments (scale surface n = 5, scale mixed n = 5) and control group (no barite scale added, n = 5; Supporting Information S1, Figure 1).Per the 7-day barite sediment exposure above, each core was briefly homogenized using a vortex mixer before being composited into a single sterile 50 mL for further homogenization.Approximately 10 g (wet mass) from each pooled subsample was weighed and stored frozen at −80 °C prior to DNA extraction in-house.The remaining sediment was then dried overnight at 45 °C prior to being sent for trace metal and radiochemical analysis, per MacIntosh et al. (2023).

Analytical methods for metals and radionuclides in sediments
At the completion of both experiments, four sediment subsamples were collected from each microcosm chamber in both exposure experiments with acid-washed 20-mL vials and then transferred to 50-mL centrifuge tubes, homogenized, and mixed with a vortex prior to analysis for total recoverable metals and radionuclides (28-day n = 15, 7-day n = 10).Total recoverable metals (TRMs) were determined after microwaveassisted (MARS Express) reverse aqua regia digestion (3:1 HNO 3 :HCl), per Simpson & Spadaro (2011).
Sediment 226 Ra, 210 Po, and 228 Th were analyzed by alpha spectrometry, per MacIntosh et al. (2023).In brief, sediment samples were prepared according to the in-house Australian Nuclear Science and Technology Organisation's Environmental Radioactive Measurement Centre (ERMC) certified method (I-3335 and I-3022).After sediments were acid-digested, a 20-mL subsample of the digest solutions was weighed and divided into four separate aliquots for radionuclide sequential extractions and analyses: 8 mL for 226 Ra and 210 Po and 8 mL for 228 Th.A known quantity of in-house synthetic tracers as the internal standard yield was added to the respective aliquots for each of the radionuclides: 0.2 g of 209 Po (0.27 ± 0.001 Bq/g), 0.2 g of 133 Ba (842.6 ± 20.3 Bq/g), and 0.1 g of 229 Th (0.2666 ± 0.08 Bq/g), following the ERMC Po chemical isolation method and the ERMC Ra chemical isolation method (I-330), described in MacIntosh et al. (2023).For 226 Ra, eluent was separated using cation exchange and resin with a vacuum box system, followed by colloidal precipitate of BaSO 4 on a membrane filter using a lock-seal Gelman filter apparatus and analyzed via gamma spectrometry.All samples were left for 21 days to allow for radon-222 ( 222 Rn) ingrowth and then radioanalyzed for at least 21 h using HPGe detectors (Ortec).The activity of 226 Ra was determined by the average activity of 222 Rn daughters: 214 Pb (using energy line 351.60 keV) and 214 Bi (using energy line 609.3 keV).Polonium-210 was autodeposited onto silver disks, and samples were counted via alpha spectrometry for up to 200,000 using silicon surface barrier detectors and Maestro ® -32 MCA Emulator software to obtain sample spectra.Polonium-210 activities were calculated from the peak counts of 210 Po and 209 Po.The resulting spectra were then analyzed using Ortec ® AlphaVision.For quality assurance/ quality control (QA/QC) measures, the chemical recoveries of the samples were measured by detecting the 133 Ba tracer via gamma spectrometry to measure the recovery rate of 133 Ba (302 and 356 keV peaks).A certified 133 Ba standard disc was counted and compared with each sample to determine the chemical recovery.Recoveries for both 226 Ra and 210 Po ranged between 80% and 110% and were deemed reliable.For QA/ QC, a blank, known control solution and standard reference material from the IAEA (gamma spectrometry standards) were counted on each detector and the results assessed against the certified activities from IAEA marine sediment.

DNA extraction
The frozen sediment samples mentioned above (see section 7-day barite sediment exposure; n = 15) were removed from the -80 °C freezer and gently thawed at room temperature (22 °C), and approximately 0.25 g of sediment was used for DNA extraction using DNeasy PowerSoil Pro Kits (Qiagen) according to the manufacturer's instructions with the following modifications: Fast Prep-24 (MP Biomedicals) was used to lyse the samples for 30 s, with the speed set to 4.5; DNA was eluted in 2 × 50 µL of elution buffer.Success of extraction and DNA yield were measured on a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific) at 260 nm.Extracted DNA was then sent along with the frozen sediment samples (n = 20) on dry ice to eDNA Frontiers at Curtin University for amplification and sequencing.

PCR amplification and sequencing analysis
Samples were provided to eDNA Frontiers as (1) frozen sediment samples from the 7-day exposure experiment (see section 7-day barite sediment exposure), and (2) extracted DNA from the sediments from the 28-day exposure experiment (see section 28-day barite sediment exposure).
Polymerase chain reactions (PCRs) were performed in duplicate on each extraction to determine the required dilution for optimal amplification by adding DNA template at neat concentration and after dilution to both 1/10 and 1/100 concentrations.The PCRs were run using unique, single-use combinations of 8-bp multiplex identifier-tagged (MID-tag) primers, as described in Koziol et al. (2019) and van der Heyde et al. (2020).
The PCRs were performed at a final volume of 25 μL, where each reaction was comprised of 1 Å~PCR Gold Buffer (Applied Biosystems), 0.25 mM deoxynucleotide triphosphate mix (Astral Scientific), 2 mM MgCl 2 (Applied Biosystems), 1 U AmpliTaq Gold DNA polymerase (Applied Biosystems), 0.4 mg/mL bovine serum albumin (Fisher Biotec), 0.4 μM forward and reverse primers, 0.6 μL of a 1:10,000 solution of SYBR Green dye (Life Technologies), and 2 μL template DNA.Master mixes were prepared using a QIAgility instrument (Qiagen) in an ultraclean laboratory facility, with negative and positive PCR controls (n = 2).All PCRs were run with negative and positive DNA controls to ensure the validity of results: 2 to 5 μL of PCR-grade water (Molecular Biology Reagent; Sigma-Aldrich) was substituted for DNA in negative controls, the positive for Bacteria was a synthetic oligo of Leuconostoc carnosum.
The PCRs were performed on StepOne Plus instruments (Applied Biosystems) with the following cycling conditions: 95 °C for 5 min, followed by 50 cycles of 95 °C for 30 s, 54 °C (Bacteria) or 55 °C (Archaea) for 30 s, 72 °C for 45 s, with a final extension stage at 72 °C for 10 min.Only samples from the dilution that showed optimal amplification were taken through for further processing.Using unique, single-use combinations of 8-bp MID-tag primers, fusion-tag PCRs were run in duplicate on the dilutions as determined above for the Bacteria assays and a fusion-tag quant for the Archaea assay.The PCRs and cycling conditions were unchanged.
A sequencing library was created by combining samples into minipools based on the PCR amplification results from each sample.The minipools were then combined in approximately equimolar concentrations to form each library.Each library was then size-selected (250-600 bp cutoff) using a Pippin Prep instrument (Sage Sciences) with 2% dye-free cassettes, cleaned using a QIAquick PCR purification kit, quantified on a Qubit (Thermo Fisher), and diluted to 2 nm.Libraries were sequenced on an Illumina MiSeq (PE-250; Illumina) using 500-cycle kits with custom sequencing primers.Laboratory extraction and PCR controls were included to test for contamination.

Bioinformatics and taxonomic assignments
Sequenced data were processed using the Greenfield Hybrid Amplicon Pipeline as an in-house clustering and classification pipeline built around tools from USEARCH (Edgar, 2010) and the Ribosomal Database Project (RDP; Cole et al., 2014) combined with locally written tools for demultiplexing and generating operational taxonomic units (OTU) tables.The poor-quality tails of each read were trimmed prior to pair-merging using an in-house windowed qual-scorebased technique with a fastQC qual score cutoff of 25.The merged reads were then filtered, discarding any with lengths outside the target range of 333 to 351 for 16S Archaea and 213 to 227 for 16S Bacteria.These reads were then clustered at 97% similarity using the cluster_otus command in USEARCH (Ver.11.0.667;Edgar, 2010).Representative OTU sequences from each OTU were assigned a taxonomic lineage using the RDP Naive Bayesian Classifier with a minimum confidence threshold of 50% (Cole et al., 2014) and matched to its closest sequence in a RefSeq-based set of reference sequences.
After processing and prior to statistical analyses, the data were separated by experiment and filtered to remove potentially erroneous sequences based on the proportion of contamination in the positive controls.The filtering process was performed according to the methods outlined by Gillmore et al. (2021).In brief, for all of the separate data sets, the proportion of contamination in the positive controls was determined.The positive controls amplified in the PCR were also used for screening of successful amplification and sequencing and to check for cross-contamination in the 16S libraries.The proportion of contamination was low in all data sets (0.001% for prokaryotes), and these values were set as the cutoff for filtering the data set.The proportion of read counts for each OTU in each sample was determined (the read count for each OTU divided by the total read count for that sample).If the proportion of read counts for each OTU per sample was less than the proportion of contamination (0.001%-0.01%), then those reads were removed from the data set.After QC checks were complete, controls were removed from the data set.Any OTUs that appeared in fewer than two samples or had a match of <80% were removed.The positive controls and any OTUs that did not meet the QC checks were removed from the data set, and these final data sets were used in the statistical analysis.The OTUs sequenced and assigned to the macroinvertebrates used in the experiments were removed from the final analysis.Prior to statistical analyses, Bacteria and Archaea were aggregated together and are referred to as the "prokaryotic community."

Statistical analyses
Species richness (Shannon's and Simpson's diversity indices) and Pielou's evenness were calculated for each treatment group for the two experiments using the phyloseq package in RStudio (Ver.4.1.2;2022).The diversity metrics were obtained using the "estimate_richness" function.Data were assessed for normality using Q-Q plots, and homogeneity of variance by Shapiro-Wilk test and Levene's test was tested prior to hypothesis testing.For each alpha diversity measurement, a one-way analysis of variance (ANOVA) was conducted on the normalized data to compare the mean alpha diversity matrixes between treatment groups for both sediment exposure experiments.Tukey's post hoc pairwise comparisons were used to test for differences between the mean diversity measurements between treatment groups at a significance level of p < 0.05.
The relative abundances (percentage) of bacterial communities were studied at the phylum and class levels among treatment groups.At each level, taxa were considered dominant if the average relative abundance was >1%.To determine whether phyla relative abundances were significantly different at the treatment level, a one-way ANOVA was performed on the nontransformed data, with additional post hoc Tukey-Kramer's tests if applicable.Prior to multivariate analysis of the community composition and structure data, all species abundance data from the final filtered data sets were checked for normality of the residuals and homoscedasticity of data (Levene's test); and when one or both of these conditions were not met, the data were Hellinger's-transformed (Chariton et al., 2015).Ordination of compositional data was performed by nonmetric multidimensional scaling (NMDS) using the Bray-Curtis similarity coefficient to visualize the community assemblages at the OTU level.We tested for treatment effects on microbial community dissimilarity (at the OTU level) with a one-way analysis of similarity (ANOSIM) to determine the differences among treatment groups, with a pairwise a posteriori test based on 9999 random permutations used to identify similarities and differences among the treatments, if applicable.The ANOSIM yields an R value, whereby values close to R = 1 indicate a high separation between treatment groups, while values close to R = 0 indicate a low group separation.The ANOSIM tests were performed on the same transformed data used for the NMDS plots, using the function anosim, from the vegan package.In addition, permutation tests for homogeneity in multivariate dispersion (PERMDISP) were performed using the betadisper function in the vegan package to evaluate between-group differences and within-group dispersions, respectively.All analyses were conducted in RStudio (Ver.4.3.2).

Sediment chemistry from scale-exposure experiments
The sediments from the 28-day scale-exposure experiment had TRM concentrations (arsenic, Cd, Cr, Cu, Mn, Pb, zinc [Zn]) below their respective default guideline value (DGV) or guideline value-High values (MacIntosh et al., 2023).The 7-day scale exposure also had metal concentrations below their DGV.However, a few metals were elevated (Cu, Ni, Pb, Zn) with the sum quotient of these metals exceeding 1 (i.e., sum quotient is the summed ratio of all metal mixture components that may provoke an effect on a test organism; each of the metal concentrations in the sediment divided by the concentration of their respective sediment quality DGV), which is a common metric to assess mixture toxicity and therefore could indicate potential for metal mixture toxicity (Meyer et al., 2015).The only significant difference in metal concentrations between the control and scale-spiked treatment groups was for Ba (7-day, F (1,12) = 67.7,p = 2 × 10 −17 ; 28-day, F (8,16) = 9.42, p = 0.003; Table 2), with no significant differences in the TRM concentrations of corrosion-specific metals (Cr, Cu, Mn, Pb, Zn) in the scale-spiked sediments relative to controls (Table 2).
The 7-day exposure had higher radionuclide activity concentrations than the 28-day test (Table 3), potentially related to discrepancies in measurement techniques (alpha vs. gamma spectrometry) and the duration of the tests.Both 210 Po and 226 Ra were significantly elevated in the scalecontaminated sediments compared with the control sediments; therefore, the presence of pipe scale was sufficient to increase the activity in the scale-contaminated relative to control sediments.b Data retrieved from sediments (n = 10) analyzed from the 7-day barite-exposed sediments.
Significant differences between treatment groups are indicated by the following: *p < 0.05, **p < 0.01.
Beta diversity: Community composition.The comparison of the prokaryotic communities by NMDS analysis did not show differences between the sediments from the 7-day exposure (Figure 2; R = 0.03, p = 0.17).There was no distinct grouping of the microbial compositions, with overlap between treatment groups indicating a low degree of variation between the prokaryotic assemblages.The PERMDISP results did not show any significant differences among and between the treatments in dispersion magnitude (F = 0.05, p = 0.82).This suggests that short-term exposure to contaminated sediments with barite scale did not result in a shift in microbial community composition and diversity.
There was a difference in the dominant phyla community composition between the treatment groups attributed to the significantly higher presence of Chloroflexi (15%) appearing in the scale-surface treatment and Firmicutes only appearing in the scale-mixed treatment (26%), compared with the control (ANOVA, F (41,5871) = 2.35, p < 0.01).This suggests selection of more species for contaminant tolerance, with the notable presence and absence of certain taxa differing between treatments.
On visual inspection, comparisons of the relative abundance of prokaryotes among treatment groups at the abovedescribed taxonomic level (i.e., class) revealed that the scalemixed treatment had the most dissimilar microbial community with the largest number of different classes of prokaryotes (i.e., Bacilli, Anaerolineae, and Clostridia; Figure 3).

Bacterial alpha diversity.
For the 28-day exposure there were no differences in the mean total OTU richness between treatments; however, the scale treatments had a relatively lower mean Simpson diversity index, signifying lower microbial OTU diversity (Supporting Information S1, Figure 3).However, there were no significant differences observed in total OTU richness (ANOVA, F (2,12) = 5.12, p = 0.52) or for diversity (Simpson index, ANOVA, F (2,12) = 1.27, p = 0.51).b Data retrieved from sediments (n = 10) analyzed from the 7-day barite-exposed sediments.
Significant differences between treatment groups are indicated by the following: *p < 0.05, **p < 0.01.
Beta diversity: Community composition.The prokaryotic assemblages from the 28-day exposure sediments did not show distinct dissimilarities in the microbial community composition between the treatments (Figure 4).The ANOSIM did not reveal differences in the microbial communities between treatment groups (R = 0.06, p = 0.15), with the NMDS plots showing overlap among the scale-spiked treatment groups.There appeared to be greater dissimilarity (i.e., controls were more dispersed than the scale-spiked treatments) between the controls compared with the scale-spiked treatments.This indicates the sediment microbial communities were different between each individual control microcosm chamber and suggests that the presence of NORM scale contamination had less variability between microbial communities through selection of more tolerant species.While there appeared to be clear variation between the sediment microbe community composition in the control mesocosms (Figure 4), PERMDISP results did not show any significant differences among and between the scale-spiked treatments in dispersion magnitude (F = 0.75, p = 0.5).Based on the small clustering of the scale-spiked treatment groups, the greater dispersion of the control group, and the insignificant statistical differences, we can infer that there may be plausible signs of community shifts in response to the treatments.

DISCUSSION
Microbial communities can be used in microcosm-based experiments to elucidate the potential ecological effects of contaminants in the marine environment.While previous studies on the effect of petroleum-associated contaminants have often focused on polycyclic aromatic hydrocarbons (Lin et al., 2023;Näkki et al., 2021;Premnath et al., 2021), total petroleum hydrocarbons (Li et al., 2019;Saul et al., 2005), or nonaromatic compounds, impacts of offshore petroleumassociated NORM-based contaminants on microbial communities in sediment are largely unknown.We addressed this in the present study by investigating the effects of NORM barite scale on marine sediment microbial communities.

Response of microbial communities to the presence of barite scale
Our study detected no clear disparities in the microbial community compositions between the control and scalespiked contaminated sediments for the 7-day experiment and 28-day experiment.The 7-day exposure had higher radionuclide and metal activity concentrations than the 28-day exposure.The radionuclide differences may be a methodological difference, with gamma spectrometry being more reliable for low-level environmental measurements in sediments.The ratio of 226 Ra to 210 Po was 2.8 in the scale, 0.2 in the 7-day test, and 2.3 in the 28-day test.Based on this, we expect that 226 Ra activity concentrations could be higher than reported.Metal concentrations were low in the 28-day test but relatively higher in the 7-day exposure.While all metals were below their DGVs, the summed quotient of Cu, Pb, Ni, and Zn was 1.7, which could indicate a risk of toxicity from the metal mixture exposure.
While there were no statistical differences in the bacterial community composition, there was a distinct change in the relative abundance of dominant phyla in the scale-spiked treatments in the 28-day exposure, in relation to the control.Bacillota and Firmicutes became more dominant in the scalemixed treatment and Chloroflexi in the scale-surface treatment, with a notable absence in the controls.The changes in Bacillota, Firmicutes, and Chlorflexi abundance are consistent with previous findings in which those taxa have been isolated from soils and sediments enriched with 238 U, 232 Th, and other radionuclides (Hug et al., 2013;Theodorakopoulos et al., 2017).As a result, evidence suggests that these bacteria are tolerant of radiation, metal contamination, and radiotoxicity (Lopez-Fernandez et al., 2021;Shukla et al., 2017).
Firmicutes have been well documented to survive and compete in contaminated environments, including uraniumcontaminated sites (Shukla et al., 2017;Sutcliffe et al., 2017;Theodorakopoulos et al., 2017).Within this phylum, endospore-forming genera have greater survival in environmental conditions with gamma radiation and might tolerate elevated concentrations of radionuclides (Enyedi et al., 2019;Farkas et al., 2002).This could indicate that radiation-resistant microbiota are present in the scale-spiked sediments in our study, suggesting an ecological and functional importance within marine contaminated ecosystems.Chloroflexi have been reported to become dominant following acute gamma irradiation of soils and have been detected in natural uranium cores and at contaminated sites (Hug et al., 2013;Theodorakopoulos et al., 2017).This phylum has the potential to tolerate uranium and/or its associated radioactivity.In our study, the higher abundance of OTUs affiliated with Chloroflexi in the scalesurface sediments may be a response to exposure to levels of elevated radionuclide activity concentrations within the sediments.It is worth noting that these phyla are ubiquitous in diverse environments, including being involved in carbon cycling on the subsurface of sediments (Hug et al., 2013;Lopez-Fernandez et al., 2021).They are important degraders of complex organic matter and provide organic acids to other micro-and macroorganisms in the sediment (Hug et al., 2013).The abundance of OTUs affiliated with Chloroflexi in the scalesurface treatment suggests they could have distinct ecological niches and physiological characteristics, such as tolerance to radionuclide exposure.
Microbes living in contaminated environments could have the potential to tolerate the presence of elevated ionizing radiation, in addition to having the desired functions to resist the chemical and radiotoxicity of radionuclides (Lopez-Fernandez et al., 2021;Shukla et al., 2017).Although beyond the scope of the present study, our findings could initiate research into potential indicator species indicative of radionuclide-contaminated sediments in the marine environment and tolerance genes related to the presence of radiation and radionuclides.

Limitations
The low replication of the experimental design reduced the likelihood of being able to detect small changes in community structure.In addition, the length of the experiment was constrained further, limiting the ability to observe any changes that might occur following >28-day exposure.Future experiments could increase the number of replicates, increase the length of exposure, and/or increase the exposure concentration to confirm whether NORM contamination can impact microbial communities.Based on the controlled laboratory conditions in our study, we expected to see impacts of the presence of NORM scale on the sediment microbial community over a long-term exposure (28 days).The activity concentrations of the NORM spiked sediments for the 28-day test were broadly equivalent to the most conservative threshold value (0.009 Bq/g for 226 Ra) derived by Koppel et al. (2023), which was intended to be a conservative value aligned to the 10 µGy/h dose rate reference level commonly used as a first-step screening value in environmental risk assessments.
Although our results provide experimental responses to the presence of NORM scale, directly relating these results to field scenarios is difficult.Controlled laboratory conditions attempt to simulate field environmental conditions of sediment, ocean temperature, water quality, light exposure, and the natural biotic community composition.Therefore, use of sediment microcosms can be a unique method for bridging laboratory experiments and field observations because it allows for testing of dose-response scenarios under controlled conditions that replicate key field conditions such as the physicochemical variables and baseline sediment community composition.It also allows the investigation of a single stressor (i.e., NORM scale) to better understand potential toxic effects which may be difficult to elucidate from a field study.

Future studies
The present study investigated the impact of NORM exposure to sediment microbial communities at one fixed concentration and two time points.Future studies could replicate these experiments to investigate microbial community changes at different time points and across a range of NORM scale concentrations, such as 10, 100, and 1000 times dilution by mass with sediment.Testing a concentration series would allow for dose-response relationships to be investigated (George & Wan, 2019;Mahamoud Ahmed et al., 2020;Mayor et al., 2013).This would also provide valuable insights for environmental risk assessment and management of decommissioned offshore structures.
The duration of the exposure is important to consider; it is contingent on the health and adaptability of the microbial community within the laboratory conditions.Reproducing the present study under varying exposure durations can allow the identification of temporal trends and determine the resistance or vulnerability of microbial communities over the short or long term (Shade et al., 2012).Short-term exposure can lead to immediate shifts in microbial communities, while long-term exposure may result in altered compositions and reduced diversity.Coincidentally, radiologically based contaminants, such as NORM scale, have the potential to have positive and negative effects on sediment microbe communities (Reisz et al., 2014;Shukla et al., 2017).This is largely dependent on the concentration, exposure duration, and the specific characteristics of the microbial communities involved (Reisz et al., 2014;Shukla et al., 2017).However, there could be differences in the microbial community compositions if the experiments were run for longer durations, and with a greater number of replicates we can increase our ability to detect differences in those communities.While our data demonstrated no short-term (7-day) effects of NORM scale contamination on the microbial structure of marine sediments, the results suggest the potential for shifts in the microbial community structure over longer durations of exposure (>28 days) to NORM scale contamination.eDNA as a line of evidence to assess the risk of NORM scale The measurement of NORM scale and assessing the potential risk to the marine environment require an evidencebased integrated approach to evaluate potential ecological impacts.Several tools and methodologies exist that can be employed to assess NORM risk, including advanced nuclear and imaging techniques to characterize NORM-contaminated scale (MacIntosh et al., 2024), the use of radiological biota dose modeling software (MicroShield Pro, ERICA Tool) using scale-specific parameters for different environmental exposure scenarios (MacIntosh et al., 2022), and use of microcosms to perform a series of marine infauna-exposure tests to understand the bioaccumulation of scale-associated constituents (MacIntosh et al., 2023).We acknowledge the presence of sediment-dwelling organisms during the experiments and the potential influence of benthic species on microbial communities.Studies on clams and amphipods have shown their ability to stimulate the activity of the microbial community in surface sediments (i.e., inhibition or activation of different bacterial phyla groups), including the biodeposition of feces or bioengineering activities (e.g., burrow walls, irrigation; Karlson et al., 2007;Lukwambe et al., 2020;Mermillod-Blondin et al., 2004;Zilius et al., 2022).While this is outside the scope of the present study, future studies could investigate the microbe-benthic macrofauna interactions and the collective contributions in NORM-contaminated sediments that provide an additional line of evidence that allows for microcosms to be used to assess impacts on multiple species at the microand macrofauna levels for a potentially impacted ecosystem.
Microcosms can be used as a form of "direct toxicity assessment" that offers several benefits in the context of risk assessment, when it comes to evaluating the toxicity of NORMbased materials (Benton et al., 2007).Microcosms can facilitate the simulation of realistic environmental conditions, replicating the key dynamics of marine ecosystems, and allow for the assessment of the multiple effects of contaminant mixtures, presenting a more realistic representation of environmental exposure scenarios (Alexander et al., 2016;Benton et al., 2007;Caquet et al., 2000).In contrast to certain other modes of toxicity assessment, microcosms can complement dose modeling or assumptions regarding exposure scenarios.Organisms are exposed to the actual material in a controlled environment, eliminating the requirement for extrapolations or assumptions regarding concentrations and exposure pathways, and can be used to verify or support models and extrapolation of other data.
Bacteria play pivotal roles in the marine environment, and the use of eDNA technology provides a valuable tool for quantifying their presence and absence (Nwachukwu & Babalola, 2022).This method offers additional insights, especially regarding NORM scale impacts on microbial organisms that may not be easily cultured in the lab and may extend beyond the scope of traditional testing (Nwachukwu & Babalola, 2022).Environmental DNA technology facilitates bacterial communities becoming valuable indicators and key environmental functions/cycles (e.g., nitrogen, carbon, phosphorus cycling) in the presence of NORM scale, offering a proactive approach to environmental monitoring.This early screening system is crucial for managing and mitigating potential impacts on marine ecosystems.

CONCLUSION
Microbial communities are crucial for understanding the interactions between contaminants and the marine environment because of their roles in nutrient cycling, biodegradation, and ecological balance.Our study investigated the effects of NORM barite scale on these communities and found no significant differences in community composition between control and scale-spiked sediments for both short-term (7-day) and medium-term (28-day) experiments.However, there were notable changes in the dominant bacterial phyla in the scalespiked treatments following the 28-day exposure, suggesting the potential for selection for contaminant tolerance from exposure to radiation and metal toxicity.Further research is needed to understand the tolerance mechanisms of microbial communities to NORM-associated contaminants and their potential as indicator species for radionuclide-contaminated sediments.Controlled laboratory studies like ours provide valuable insights but may not fully replicate field conditions.Future studies should explore different exposure concentrations and durations to elucidate dose-response relationships and assess long-term impacts on microbial communities.Microcosms offer a unique method for bridging laboratory experiments with field observations, allowing for more realistic simulations of environmental conditions and assessment of contaminant effects on multiple species.Finally, microcosms, along with the analysis of microbes and eDNA, serve as valuable indicators for assessing exposure to NORM within marine ecosystems, providing insights into the intricate dynamics of radioactive contamination and its potential impacts on biological communities.
Supporting Information-The Supporting Information is available on the Wiley Online Library at https://doi.org/10.1002/etc.5887.
" represents scale added to the surficial layer to a final concentration of 0.1% relative to the sediment mass.ANZG = Australian and New Zealand Sediment Quality Guidelines (ANZG, 2018); DGV = default guideline value; GV = guideline value.a Data retrieved from sediments (n = 15) analyzed by MacIntosh et al. (2023).

FIGURE 1 :
FIGURE 1: Relative abundance and composition of the (A) bacterial phyla and (B) bacteria assigned to class level determined by 16S ribosomal RNA gene amplicon sequencing across sediment samples from the 7-day exposure experiment with control (no barite scale; n = 5) and scale-spiked exposure (n = 5).Only bacteria with an average abundance of >1% in the entire set are shown.For the scale-spiked treatment, barium sulfate scale was added at 0.1% of the total wet sediment (0.8 g of scale: 800 g marine sediment) to create a 1:1000 dilution of pipeline scale.

FIGURE 2 :
FIGURE 2: Nonmetric multidimensional scaling visualizations of operational taxonomic units (OTU) of community composition analysis based on 97% similarity OTU abundance data for the 7-day exposure experiment with control (no barite scale; n = 5) and mixed scale exposure (n = 5); beta diversity was measured by OTU community composition determined by Bray-Curtis distance matrix.If present, ellipses represent 95% confidence intervals comparing treatment groups.Data set was standardized using Hellinger's transformation.NMDS = nonmetric multidimensional scaling.

FIGURE 3 :
FIGURE 3: Relative abundance and composition of the (A) bacterial phyla and (B) bacteria assigned to class level determined by 16S ribosomal RNA gene amplicon sequencing across sediment samples from the 28-day exposure experiment with control (no barite scale; n = 5), scale-surface exposure (0.1% of the total sediment mass scale added to only the surficial layer), and scale-mixed exposure (0.1% of the total sediment mass scale added to the sediment and mixed).Only bacteria with an average abundance of >1% in the entire set are shown.For the scale-spiked treatments, barium sulfate scale was added at 0.1% of the total wet sediment (0.8 g of scale: 800 g marine sediment) to create a 1:1000 dilution of pipeline scale.

FIGURE 4 :
FIGURE 4: Nonmetric multidimensional scaling visualizations of operational taxonomic units (OTU) community composition analysis based on 97% similarity OTU abundance data for the 28-day exposure experiment with control (no barite scale; n = 5), surficial scale exposure (0.1% of the total sediment mass scale added to only the surficial layer), and mixed scale exposure (0.1% of the total sediment mass scale added to the sediment and mixed).Beta diversity was measured by community composition determined by Bray-Curtis distance matrix.Data set was standardized using Hellinger's transformation.NMDS = nonmetric multidimensional scaling.

TABLE 1 :
Mean activity concentrations of radionuclides with >10-day half-lives determined in pulverized barite scale samples from the well tubular (±SE; n = 3)

TABLE 2 :
Mean metal concentrations of sediments from the microcosms after the 28-day sediment exposure experiment (n = 15) and the 7-day scale-spiked sediment exposure experiment (n = 10) presented as total recoverable metals Data are presented as mean ± SE to 2 significant figures.Control is indicative of control sediments with no scale, "Scale mixed " represents scale mixed with the sediment to a final concentration of 0.1% relative to the sediment mass, and "Scale surface

TABLE 3 :
Mean activity concentrations of radionuclides with >10-day half-lives in sediments after the 28-day sediment exposure experiment (n = 15) and the 7-day scale-spiked sediment exposure experiment (n = 10) Data are presented as mean ± SE to 2 significant figures.Control is indicative of control sediments with no scale, "Scale mixed " represents scale mixed with the sediment to a final concentration of 0.1% relative to the sediment mass, and "Scale surface " represents scale added to the surficial layer to a final concentration of 0.1% relative to the sediment mass.