Choice of bead-beater instrument can result in significant differences in the outcome of host-associated microbiome studies

Oyster tissue DNA extraction

Oysters were collected from the aquaculture research lease of the Wakulla Environmental Institute located in Oyster Bay [30.0342702, -84.3558844], near Panacea, Florida. Oysters were grown in plastic cages, suspended from lines attached to pilings at depths of 1-2 meters (Australian Adjustable Long-Line System). At each sampling period (December 2020, January 2021 and March 2021), n=15 adult oysters of approximately the same size were collected over ice and transported to the laboratory at FAMU, where they were rinsed, shucked, tissues harvested and stored at 4°C for the DNA extraction to be carried out the next day. Genomic DNA from minced oyster tissues were extracted using DNeasy PowerLyzer PowerSoil DNA extraction kit was according to the manufacturer’s protocol but the samples were processed in two separate batches: one that was processed for lysis using the MSP FastPrep method (samples labeled from 1-6) and the second set was processed using the Bead Ruptor Elite method by Omni International (samples labeled as A-F), respectively. In this manner, six DNA extracts were obtained for each group (Table 1), and the extracted DNA was analyzed for quality and quantity using a micro-volume spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA).

Gene sequencing and bioinformatics analysis

We used the 515F/926R primer pairs, containing the CS1 and CS2 common sequence linkers (Walters et al.³¹), to amplify the V4-V5 variable regions of the 16S rRNA gene. The PCR amplicon generation was performed in two independent steps: ST1 and ST2, as described previously.³² The first step (ST1) basically involves the generation of the amplicon products and the attachment of the common sequence linkers (CS1 and CS2) at the 5’ end of all the DNA molecules. The reaction mixture in this step comprised 9 μL of the PCR reagent and 1 μL of the Genomic DNA template. The reagent comprised 5 μL repliQa HiFi ToughMix (PCR mastermix, Quantabio), 3μL DNA-free water, and 0.5 μL each of the gene specific forward and reverse primers containing the 5’ linkers. The ST1 PCR condition comprises initial denaturation at 98°C for 2 min, followed by 28 cycles of denaturation at 98°C for 10min, annealing at 50°C for 1 min, and elongation at 68°C for 1 min.

The second step PCR (ST2) was basically to incorporate sequence adapters and sample-specific barcodes into the ST1 amplicons. The ST2 reaction mixture comprised 3 μL of repliQa HiFi ToughMix (PCR mastermix, Quantabio), 4 μL of DNA-free water, 1 μL each of the forward and reverse sample-unique Access Array Primer pairs (Fluidigm, South San Francisco, CA; Item# 100-4876), and 1μL of ST1 PCR product, as the template. The forward Access Array primer contained the 3’ common sequence linker (CS1), and the adapter sequence, while the reverse Access Array primer contained the sample-specific barcode in addition to the 3’ CS2 and the adapter sequence. The ST2 PCR condition was like the ST1, only that the ST2 had 8 cycles; besides, the annealing process of the ST2 was performed at 60°C.

The amplicons were pooled, and the library preparations and sequencing using the Illumina MiSeq based on V3 chemistry (600 cycles) were carried out at the Genomic and Microbiome Core Facility at Rush University. Meanwhile, to address the issue of low sequence complexity with 16SrRNA amplicons, the pooled library was spiked illumina Phix viral DNA library (a prepared high complexity library) at a 15% volume ratio.

The MicrobiomeAnalyst pipeline was applied for the relative abundance analysis and to create the metagenomics plots using data processed from QIIME2.

Metagenomic sequence accession numbers

The metagenomic sequences obtained from the FastPrep and the Bead Ruptor Elite samples have been deposited in NCBI’s Sequence Read Archive under BioProject PRJNA986007 with the accession numbers SRR25013903 and SRR25013902, respectively, and BioSamples SAMN35839752 and 35839753.

DNA concentration analysis

The extracted DNA was quantified using a micro-volume spectrophotometer (NanoDrop Technologies) and the results are show in Table 1. The statistical significance of difference in DNA concentrations between DNA isolated using the FastPrep lysis (MP Inc.) versus the Bead Ruptor Elite (Omni international Inc.) methods was assessed with the Mann-Whitney U test.³³ The distribution of the DNA quantities was not normal, and the value of the variable was continuous, making a non-parametric test suitable for the analysis. The Mann-Whitney U test is used to determine whether there is a statistically significant difference between the medians of two independent groups. The Mann-Whitney U statistic was selected as the smallest of the two U values for the two groups calculated following the U statistic formula,³³ and the U value was compared with the standard Mann-Whitney U critical value. At 0.05 significant level, the U statistic was 11 while the U critical was 5, suggesting there was no evidence to support the null hypothesis, which states that the DNA concentrations from the two groups were equal (Table 2). In other words, the quantity of DNA was affected by the extraction protocol. This finding mirrors previous reports, which suggests that the nature of DNA lysis equipment can be a significant factor with impacting the outcome of metagenomics studies. For example, Zhang et al.³⁴ reported that the bead-beating intensity in DNA lysis affected the DNA yield in gut samples. In line with this report, our results also suggest that the FastPrep lysis method provided the desired beating intensity than the Bead Ruptor Elite, hence the average DNA yield from the FastPrep was higher than the Bead Ruptor Elite.

Assessment of the oyster microbiota

Metagenomic-based microbial taxonomic analysis of the eastern oyster

The actual abundance of microbiomes in the oyster tissue at the phylum level is shown in Figure 1. Vertical bars represent the abundance of organisms in each sample and the colors represent the identified phyla. The samples were mostly populated by the Protobacteria phyla, regardless of lysis method. The samples treated by the FastPrep lysis method also harbored Firmicutes and Bacteroidota, which were not identified in the Bead Ruptor Elite lysis method. Traces of Actinobacteria were found in both groups. This is indicative that the DNA extraction was impacted by the lysis method such that MP FastPrep is a better choice for the oyster tissue samples used in this study. The absence of the Gram-positive bacteria, Firmicutes, from the Bead Ruptor Elite group indicates that this method could not provide the adequate bead beating intensity required to break the complex cell wall of Gram-positive bacteria.³⁴ On the other hand, abundance of Bacteroidetes and Firmicutes in the FastPrep group strongly indicates that this method could lyse both the Gram-negative and the Gram-positive bacteria in oyster samples.³⁴

Figure 1. Stacked bar plot of bacterial abundances identified at the phylum level.

Samples were isolated by the same DNA extraction protocol but differed in their lysis method: A, when samples were lysed by the FastPrep method and B, when samples were lysed by the Bead Ruptor Elite method.

The genus-level taxonomic groups identified are shown in Figure 2, which showed that the Delftia genus dominated up to 80-85% of all the genera identified across all samples when the lysis was performed using the FastPrep method. Interestingly, when the lysis was carried out using the Bead Ruptor Elite lysis method, the predominant genera at 80-90% was Burkholderia_Caballeronia_Paraburkholderia. Mycoplasma was another genus identified from the samples treated with the FastPrep method but were mostly absent in the Bead Ruptor Elite treated samples. This shows that two different lysis protocols can result in different taxonomic data and thus researchers need to be cognizant of DNA extraction process being followed for such metagenomics studies, especially those that involve host tissues containing a high amount of mucous and other inhibitory materials. Both Delftia and Burkholderia belong to the Betaproteobacteria class and Pseudomonadota phylum. However, the bacterial genera identified from the FastPrep method were more diverse, thus suggesting that this lysis method provided a better recovery of the oyster microbiota than the Bead Ruptor Elite method. In addition, the presence of Burkholderia, a Gram-negative bacteria with a fairly easy-to-lyse cell wall, in the Bead Ruptor Elite group further suggests that this method could not provide enough intensity to rupture the cell walls of more robust bacteria in the oyster samples, lending further support that DNA lysis can significantly impact the outcome of metagenomics assessment, especially in complex samples from host tissues and the gut ecosystem.³⁴

Figure 2. Stacked bar plot of bacterial abundances identified at the genus level.

Samples were isolated by the same DNA extraction protocol but differed in their lysis method: A, when samples were lysed by the FastPrep method and B, when samples were lysed by the Bead Ruptor Elite method.

Delftia is a genus of Gram-negative bacteria named after the city of Delft in The Netherlands, where the genus was first isolated. Delftia is represented by five recognized species, including D. acidovorans, D. tsuruhatensis, D. lacustris, D. litopenaei, and D. deserti. Delftia species have wide geographical distribution and they have been found in fresh and marine water, plants, soil, and sludge.³⁵ Banach et al.³⁶ found Delftia to be associated with aquatic ferns and synthesized plant growth-promoting substances. Delftias are non-sporulating and are known for their unique metabolic ability to break down or transform different forms of environmental pollutants. For example, De Gusseme et al.³⁷ reported that Delftia tsuruhatensis and Pseudomonas aeruginosa were solely responsible for the total removal of acetominophen, an analgesic and antipyretic drug, from a wastewater treatment plant. According to Zhang et al.,²⁹ Delftia tsuruhatensis can degrade several chloroanilines through the process of mineralization, which is an important step in nitrogen cycle. Wu et al.³⁸ reported that a strain identified as Delftia lacustris was resistant to heavy metals, such as Cr, Pb, Hg, and Cd. They also reported that the strain could degrade aromatic compounds like naphenol, naphthalene, 2-methylnaphthalene, and toluene. Aside from the ability to remove pollutants, Delftia species can play significant roles in nitrogen cycling processes: Delfta tsuruhatensis performed both heterotrophic nitrification and aerobic denitrification functions in a sewage treatment ecosystem.³⁹ Chen et al.⁴⁰ isolated a heterotrophic and aerobic nitrifying-denitrifying strain of Delftia tsuruhetensis from a co-culture of rice husk and swine manure. The strain had ammonium removal rate of about 98% from a piggery wastewater. We are tempted to predict that Delftia species likely play several roles related to ecosystem services withing the Eastern oysters.

The genera Burkholderia is a Gram-negative bacterium of the Betaproteobacteria class with about 80 recognized species. Burkholderia has diverse bioremediation properties and can synthesize a wide range of antimicrobial compounds.⁴¹ Interestingly, Burkholderia predominated amongst the marine sponge-associated bacteria.⁴² It has also been shown that Burkholderia species are active in the denitrification processes. Vitale et al.⁴³ stated that sites occupied by Burkholderia are characterized by limited oxygen where the organisms generate energy by denitrification; furthermore, several genes required for growth under denitrification condition abound in Burkholderia thailandensis.

Species diversity within the oyster tissue sample

Two types of diversity analysis were applied to estimate the species distribution within the community based on operational taxonomy units (OTUs). The observed diversity or richness is a measure of the individual number of species found in each sample. The richness does not account for the relative proportion or abundance of the individual species in a sample. In Figure 3(A), the OTUs from the FastPrep method were about 40% more diverse than the OTUs from the Bead Ruptor Elite method. The observed difference in OTUs suggests that the DNA extraction kits significantly impacted the richness of the oyster tissue microbiome community.

Similarly, the Shannon diversity or equitability index considers the proportion or evenness of species in a community. The Shannon index ranges from 0 to 1⁴⁴; a perfectly distributed community takes the index of 1. The Shannon equitability of the OTUs from the two DNA lysis methods kit is shown in Figure 3(B). By this measure, species from the FastPrep method were about 70% more evenly distributed than the Bead Ruptor Elite method. In other words, the FastPrep method resulted in richer and evenly diversified bacterial community relative to the Bead Ruptor Elite method.

Figure 3. The diversity plot of species-level OTUs obtained from the oyster samples using the two DNA lysis methods.

Shown are A, the observed diversity index and B, the Shannon equitability index.

Species diversity on principal coordinate axes

The plot of the principal coordinate analysis of the species from the oyster tissue samples grouped by the two DNA lysis extraction protocols is shown in Figure 4. The first principal coordinate axis accounted for 54.8% of the total variation, while the second principal coordinate axis accounted for 16% of the total variation. Projecting the sample points on the first principal coordinate axis (the horizontal axis), the bacterial communities from the FastPrep method were completely separated from the Bead Ruptor Elite method, indicating that the DNA extraction method significantly affected the identification of bacterial species from the oyster tissue samples.

Figure 4. Principal coordinate analysis (PCA) plot the oyster tissue microbiota at species level grouped as a function of the DNA lysis methods.

[SPERMANOVA] F-value: 2880.2; R-squared: 0.99654; p-value < 0.003.

The accompanying permutational multivariant analysis of variance (PERMANOVA) was carried out to investigate the statistical significance of the difference in microbial compositions from the two groups of oyster samples as a function of the DNA lysis protocol, such that the predictor of the species diversity model was the DNA extraction method. The result indicate that the R² value was 0.9965, p value of < 0.003, meaning that the effect of DNA extraction method was significant and that it accounted for about 99.7% of the total species diversity in the two groups.

A heatmap showing the variability of identified microbiome at the family level across different oyster tissue samples is shown in Figure 5. Specifically, this heatmap analysis indicated Burkholderia species to be overrepresented in the Bead Ruptor lysis method, which goes well with the findings that these were the predominant genera at 80-90% abundance using this lysis method. Similarly, Pseudoalteromonadaceae were most abundant in the Bead Ruptor lysis method. Overall, Figure 5 shows those bacteria in red arrows that were overrepresented in the Bead Ruptor Elite method and blue arrows represent bacterial communities that were found to be overrepresented in the FastPrep method, respectively.

Figure 5. The Heatmap of oyster tissue microbiota at family level.

Each column represents a sample, and each row represents a bacterial family. Red arrows represent the overrepresented bacteria in the Bead Ruptor Elite method and blue arrows represent bacterial communities that were found to be overrepresented in the FastPrep method.

Differential abundance of the oyster microbiomes

The result of the MetagenomeSeq differential analysis is shown in Figure 6. Ten different genera from the entire samples were differentially abundant between the two lysis methods tested. These included Alivibrio, Vibrio, Fluvicoda, Rosemainus, Delftia, Serratia, Cutibacterium, Pseudoalteromonas, and Mycoplasma genera were differentially abundant in the FastPrep method. Conversely, Burkholderia-Cabelleronia-Paraburkholderia bacterial group was differentially abundant in the Bead Ruptor Elite method. In other words, the DNA extraction lysis protocol significantly influenced the type and abundance of microbial genera in the oyster tissue samples.

Figure 6. MetagenomeSeq analysis identifying microbiomes at genus level in the oyster tissue samples that were differentially abundant between DNA Extraction Protocol identified using the FastPrep lysis versus the Bead Ruptor Elite method.