Analysis of the temporal cultures shows that compositional variations of the in vitro oral communities was not entirely host-dependent – though some host-specific qualitative differences might have persisted through the culturing process – and that invariance associated with organismal succession seemed to have been captured in the in vitro cultures, from Streptococcus OTUs to Veillonella OTUs to a slight rise in Prevotella OTUs. Though it is difficult to argue, without data on the absolute cell numbers, that Veillonella OTUs definitively dominate the cultures with longer incubation times, we can conclude that the temporal shifts in relative abundances without changes in the medium or nutrients indicated a progression of organisms that is not wholly artificially induced, i.e. restricted to the laboratory. With this knowledge, our intention is to pursue whether this model can be preserved and repeatably cultured.Plaque samples were collected from three volunteer hosts, under protocols 3-18-0189 and 3-19-0119 approved by the UCSB Human Subjects Committee. The collection process was as follows: Supragingival plaque scrapes of molar teeth from three healthy adult hosts were collected with sterile metal curettes, after hosts had abstained from food, drink, and dental hygiene practices for 12 – 16 hours. For consistency, plastic seedling pots we limited the sites of the scrapes to the mucosal supragingival surfaces of 3 molar teeth, though with no restrictions on only upper or only lower molars.
The plaque from each host was used to inoculate 6mL of SHI media. SHI media was prepared according to methods in a paper by Tian and coworkers, and centrifuged at 8,000 x g for 5 minutes. Inoculated media was divided equally into three wells in a 24-well surface-modified plate such that each well received 1.98mL of liquid. Prior to receiving the inoculated media,all wells to be used were preconditioned with a pellicle layer. The pellicle was formed by adding 150µL of filtered clarified saliva to each well, incubating the plate at 37°C for 1 hour, and then sterilizing with UV radiation at 254 nm for 1 hour. Saliva was supplied as frozen fractions pooled from healthy human volunteers, and clarified after defrosting on ice by centrifuging at 6,000 x g for 3 minutes, mixing with 1X PBS, pH 8.0, and passing through 0.2-micron syringe filters. Inoculated SHI media from different hosts was pipetted into separate plates. Three negative control wells were also prepared in a separate plate, where wells received a salivary pellicle layer and 2.0mL of SHI media without inoculum from host plaque. After receiving either sterile or inoculated medium, all plates were incubated at 37°C in a sealed vessel in an anaerobic atmosphere of 85% nitrogen, 10% hydrogen, and 5% carbon dioxide. Every 24 hours, approximately 1.3mL of spent media was pipetted from the top of each well without disturbing the sedimented cells and replaced with 1.5mL of fresh SHI medium, followed by the addition of 20µL of 0.5% sucrose. Plates were then returned to the anaerobic atmosphere for continued incubation until harvesting at the 72-hour mark. These 72-hour cultures were termed “initial cultures”. Two of the three inoculated wells from the initial cultures were used in the subsequent preservation experiment. The third was pelleted at 16,000 x g for 5 minutes, flash-frozen in liquid nitrogen, and stored at -80°C for further analysis later.
We investigated several methods of preserving the microbial communities derived from the initial cultures. Two wells from the control plate and two wells from each host plate were selected out of the initial culture plates. The entire volume of a single well – approximately 1.75mL – was divided into five volumes, each measuring approximately 0.35mL. Sterilized glycerol was added to one aliquot to a final concentration of 20%, and this sample was flash-frozen in liquid nitrogen for comparative analysis later. The remaining four aliquots were subjected to the following four preservation conditions, respectively: 4°C for 1 day, 4°C for 3 days, cryopreservation with 40% glycerol at -80°Cfor 1.5 weeks, and cryopreservation with 40% glycerol at -80°C for 5.5 weeks. After preservation, we reserved 180µL for subsequent propagation experiments. The remaining approximately 170µL of the 350µL sample was divided into 2 equal aliquots, flash-frozen in liquid nitrogen with 40% glycerol, and stored at -80°C for comparative analysis later. These aliquots were termed “preserved cultures”. To assess how preservation affected the revival and growth of microbial communities, we used the reserved 180µL from each preservation condition to inoculate 6.5mL of sterilized SHI medium. The inoculated media was divided equally among 3 wells in surface-modified plates, having been conditioned by a pellicle layer as detailed above, and hosts were kept separate from one another. Plates were incubated at 37°C for 48 hours under the aforementioned anaerobic atmospheric conditions, wells were treated with procedures identical to those in the initial cultures, with replenishment of spent/partially spent media and supplementation of sucrose.
Of the 3 wells of the resulting 48-hour cultures, henceforth termed “propagated” cultures, one well was pelleted at 16,000 x g for 5 minutes and flash-frozen in liquid nitrogen as a backup. The other two wells were mixed with glycerol to a final concentration of 20%, pelleted at 16,000 x g for 5 minutes, flash-frozen in liquid nitrogen, and stored at -80°C until further processing.Genomic DNA from initial cultures, preserved cultures, and propagated cultures, positive controls, and negative controls, was extracted with the DNeasy PowerSoil kit . Negative controls included controls incubated alongside initial cultures, preserved cultures, and propagated cultures as well as sterile SHI medium and extraction controls, which consisted of 200µL of fresh 1X PBS. Positive controls consisted of 200µL of ZymoBiomics microbial community standard . The entire collection of samples, including culture controls and the positive control, was randomly divided into batches of 11. Each batch was processed with a separate extractioncontrol, making a total of 12 samples for each extraction run. For each run, frozen pellets were thawed on ice and washed with 1X PBS three times. Washed pellets were resuspended in 200µL of 1X PBS, and DNA was extracted according to manufacturer’s instructions. Estimation of post-extraction bacterial biomass was performed using the SSo Advanced Universal SYBR Green Supermix qPCR assay using gene-specific primers designed to amplify the V4 region from the 16S rRNA gene . Quantitation by qPCR showed that eighty out of eighty-two non-negative-control samples contained greater than 10,000 copies per µL, and that all negative controls contained 17 to 2885 copies per µL. These data confirmed that samples contained sufficient bacterial biomass for sequencing. The very low concentrations of 16S rRNA in the negative controls was consistent with expected low bacterial biomass. All samples were included for sequencing.16S rRNA gene amplicon libraries were constructed following the dual-index sequencing protocol developed by Kozich and coworkers using the 515F-805R primer pair designed by the earth microbiome project . This primer pair yields amplicons approximately 250 basepairs long, enabling full coverage of the region by a combination of the forward read and the reverse read . Amplicons were generated on 96-well plates using 1µL of template DNA, 1µL of each index primer and 17µL of Accuprime Pfx Supermix . Each plate contained a negative control consisting of 1µL of molecular grade water, and a positive control well consisting of 1µL of ZymoBiomics community DNA standard . The PCR process for amplicon generation was performed under the following conditions: Initial denaturation at 95°C for 2 minutes, followed by 30 cycles of denaturation at 95°C for 20 seconds, annealing at 55°C for 15seconds, and extension at 72°C for 1 minute, followed by a final extension at 72°C for 10 minutes. Amplicon purification was performed using AMPure XP beads, containers size for raspberries followed by normalization and pooling to equimolar amounts of each sample. Each sample was then sequenced in duplicate in different plates with PE300 V3 chemistry on the Illumina MiSeq platform in the genetics core of the Biological Nanostructures Laboratory within the California NanoSystems Institute at UCSB.The software package mothur was used to process the paired-end Illumina 16S rRNA reads. Briefly, read pairs were assembled into contigs, and contigs with ambiguous reads were excluded to obtain 13,249,077 total contigs. Contigs were then aligned to version 132 of the SILVA SSU reference non-redundant database, and those that did not align were removed. A denoising algorithm and chimera identification with UCHIME were performed for additional quality control, resulting in 11,662,068 high quality sequences .
The contigs were clustered at the 3% dissimilarity level to generate operational taxonomic units . Sequences were classified against the SILVA SSU reference database mentioned above using a naive Bayesian classifier . The ZymoBiomics Community DNA served as our mock community for sequencing, to calculate sequencing error rates. When these mock communities were examined using the above pipeline, an overall error rate of 0.019277% was observed and 8 OTUs were detected as expected from the 8 listed bacterial OTUs in the manufacturer’s handbook, indicating that the read processing pipeline has a low error rate and does not drastically overestimate diversity.All statistical analyses were performed with R in RStudio , with the phyloseq package . Before analyzing the cultures, we examined negative controls for potential contamination by comparing the number of reads and OTUs between negative controls and cultures. We also examined positive controls by comparing read counts and relative abundances of the DNA standard and those of the microbial standard with each other and with their respective expected distributions. After verifying that the negative and positive controls yielded expected results, we examined sequencing depth by host and by sample type. To account for differences in sequencing depth, we also examined the number of OTUs by host and sample types, and then rarefied samples to a uniform depth and observed the number of OTUs post rarefaction. To investigate potential underestimation of OTUs due to low read depth because of rarefaction, we constructed rarefaction curves for all samples, excluding one outlier with more than 700,000 reads and two samples with fewer than 1,000 reads . To determine whether any correlation existed between sequencing depth and the number of OTUs, we plotted the number of reads against three diversity indices while omitting controls, mock communities, samples with fewer than 1,000 reads, and an outlier with more than 700,000 reads. To determine the most common phyla across these same samples, we generated prevalence plots of sequencing depth vs. taxa prevalence expressed as a percentage. After examining the consequences of rarefaction, we opted to use the rarefaction step and the uniform depth of 240 reads for subsequent analyses of samples. To examine the distribution of OTUs, we plotted relative abundances according to preservation conditions, for each host. Principal coordinate analysis was performed on the relative abundances using the Bray-Curtis dissimilarity metric. A dot plot of relative abun-dances that were averaged by preservation conditions and categorized by host was used to inspect the five most prominent OTUs. Principal component analysis was also performed on the relative abundances of initial, preserved, and propagated cultures, and the PCA scree plot was used to assess the contribution to total variation from components 3 and beyond . To assess the validity of performing PCA on relative abundance data, we used centered-log-ratio and isometric log-ratio transformations on the relative abundance data and subjected the transformed data to PCA.Examination of negative culture controls shows that these controls contained fewer than 1,500 reads , with 28 out of 33 samples having fewer than 800 reads. The low read counts of the negative controls, combined with the high read counts of the cultures , indicate that both external and internal contamination was low. In the controls that contained hundreds to more than 1,000 reads, the most prominent OTUs were also found in cultures. Like other negative controls, the negative controls for propagation and PCR yielded substantially lower read counts than the plaques and cultures, regardless of preservation condition . Two sets of mock communities served as positive controls for the extraction, amplification, and sequencing steps. The ZymoBIOMICS Microbial Community Standard served as a positive control for all three steps while the ZymoBIOMICS Microbial Community DNA Standard served as a positive control only for the amplification and sequencing steps. Expected relative distributions of both standards, generated based on manufacturer’s specifications, are shown in Appendix C.3, Figure 47. Since 16S rRNA sequencing precludes eukaryotic reads, two organisms from the DNA standard were excluded from the analysis.