Search

Additional file 5: Figure S5. Microscopic images of differently fixed Gramella forsetii cells after cell sorting. Shown are overlay images of the HCR-FISH signal (green) and DAPI signal (blue). All images were taken with an epifluorescence microscope with a HC409LP (DAPI) and an ET500/LP (HCR-FISH) filter. The numbers in brackets indicate the exposure time for each image with the first number corresponding to the DAPI and the second to the HCR-FISH signals.

Additional file 8: Figure S8. Genomic quality estimation of Gramella forsetii and Maribacter forsetii sequencing products after MDA of 100 and 500 sorted cells using Quast. Shown are from left to right: the number of contigs longer than 1000 bp, the total assembly lengths, N50, number of misassemblies and number of mismatches per 100 kpb. Unhybridized samples were unfixed and were not subjected to HCR-FISH, but only sorted based on their DAPI signal. Maribacter forsetii unhybridized controls were taken from a cell aliquot and have not been filtered and washed off a filter like the other controls, including the Gramella forsetii unhybridized control. Significance thresholds (p-values) of pairwise t-tests are * < 0.05, ** < 0.01, *** < 0.001.

Additional file 6: Figure S6. Exemplary images of gel electrophoresis with MDA products from sorted isolates, fixed with formaldehyde, Lugol’s solution, ethanol or unfixed. The uppermost band of the used marker LambdaDNA Hind III corresponds to 23 kb. MDA products from formaldehyde fixed cells were either not detectable or in low amount.

Additional file 3: Figure S3. Four pure cultures (Gramella forsetii, Maribacter forsetii, Escherichia coli, Micrococcus sp.) were fixed with 10 different fixation methods (formaldehyde 4%, 1%, 0.25%, 0.1%, ethanol, Lugol’s solution with and without thiosulfate, glyoxal with and without ethanol and unfixed). HCR-FISH was done on filtered cells and signal intensity was measured after washing the cells off the filter and analyzing them in the flow cytometer. Plotted are the green fluorescence (530/40 nm) from HCR-FISH and blue fluorescence (450/60 nm) from DAPI staining. The fluorescence intensity is given in relative units on a logarithmic scale. The background fluorescence (dotted line) was defined for 10 RU. N.A. = not analyzed due to disrupted cells.

Additional file 1: Figure S1. The development of our pipeline was done in three steps before the application on an environmental sample. (1) In the first step, four bacterial isolates (GC content is given in mol%) were treated with four different fixatives (plus unfixed control) and their signal intensity was measured by flow cytometry (1a). Glyoxal was not further analyzed due to low signal intensities (red cross). Three isolates and three of the brighest fixations (plus control) were sorted (100 and 500 cells) and forwarded to MDA (1b). The MDA products of two isolates were sequenced and their genome quality assessed (1c). The best results in total (signal intensity and genome quality) were achieved with ethanol fixation and 500 cells (green star). (2) In the second step, the HCR-FISH protocol from Yamaguchi et al. [21] was adapted with different denaturation temperatures, hybridization buffers and amplification times. The signal intensities were assessed after HCR-FISH via microscopy. (3) In a third step, the optimized HCR-FISH protocol was tested for validation on isolates with ethanol fixation (plus unfixed control). There were no significant differences in assembly metrics between ethanol fixation and unfixed control samples after sequencing of MDA products from 500 sorted cells. (4) The optimized HCR-FISH protocol was tested on ethanol fixed and unfixed seawater samples. Bacteroidetes and the flavobacterial clade Vis6 were targeted by specific HCR-FISH probes, 500 cells sorted and sequenced. For comparison a whole community shotgun metagenome was prepared.

Additional file 24: Table S10. Bacterial strains used for benchmarking the influence of cell fixation and for HCR-FISH optimization.

Additional file 12: Figure S12. Sequencing read coverages across the reference genome of Gramella forsetii, Shewanella oneidensis and Pseudomonas putida. 500 cells from ethanol fixed (green) and unfixed (brown) samples were used as input for MDA.

Additional file 4: Figure S4. Blue fluorescence intensity (DAPI signal, 450/60 nm band-pass filter) of four isolates in dependency of cell fixation, measured by flow cytometry. The median of the signal population from flow cytometric analysis is shown. NA = not analyzed due to disrupted cells.

Additional file 25: Table S11. Probes used in FISH experiments. In HCR-FISH, the probe consists of a taxon specific sequence attached to the initiator sequence. The G’s in bold of the hairpin oligonucleotides are labelled with Atto488 (Biomers).

Additional file 9: Figure S9. Read coverages across the reference genome of Gramella forsetii and Maribacter forsetii cells. 100 and 500 cells were sorted from isolates, used as input for MDA and the products were sequenced. Unfixed, unhybridized cells of Maribacter forsetii were taken directly from the culture without being filtrated and washed off the filter, which explains the difference in coverage between the treatments that we did not see for Gramella forsetii.

Additional file 2: Figure S2. The background fluorescence in the green channel of flow cytometric measurements was set to 10 RU (blue line), based on comparisons betwen EUB-338 and Non-338 probes hybridized to 1% formaldehyde fixed Gramella forsetii (A, B) and Micrococcus sp. (C, D) samples. The Non-338 control of Gramella forsetii (A) was not stained with DAPI, the Non-338 control of Micrococcus sp. (C) was stained with DAPI. Green fluorescence was detected with a 530/40 nm filter, blue fluorescence with a 460/50 nm filter. 5000 events were recorded for Gramella forsetii (A, B) and 2000 for Micrococcus sp. (C, D).

Additional file 19: Figure S14. 16S rRNA consensus tree. The Vis6 cluster is targeted by the probes Vis6-814 and Vis6-871 and marked by the colored box. The closest cultured relative was Phaeocystitibacter luteus with appr. 90% sequence similarity. Nine 16S rRNA sequences, that have been binned to Vis6 MAGs from the Vis6 sorts, were placed in this tree. Six sequences were affiliated to Vis6 cluster C (marked in pink) and three sequences were affiliated to Flavobacterium ponti (in the outgroup). Flavobacterium ponti has appr. 86% 16S rRNA sequence identity to the Vis6 cluster.

Additional file 26: Figure S15. Exemplary scatter plots from flow cytometry showing the gating principle of seawater samples hybridized with HCR-FISH probe. (A) DAPI positive cells were selected in the forward scatter vs. blue fluorescence (DAPI signal) plot and (B) HCR-FISH positive cells were selected in the blue vs. green fluorescence (HCR-FISH signal) plot. Of (B) only those events that were also appearing in the gate in (A) were sorted.

Additional file 7: Figure S7. Crossing point times (CP, time of inflection point of real time amplification curve) of MDA reactions of three isolates with 100 or 500 cells input. Input samples were taken from Lugol’s fixed, ethanol fixed and unfixed cells. Additionally, unfixed cells that have not been subjected to FISH were used.

Additional file 10: Figure S10. Signal intensity of formaldehyde fixed Gramella forsetii cells, after HCR-FISH with different treatments. CARD-FISH and the former protocol (yellow bar: in hybridization buffer A, no denaturation, 45 min amplification) were used for comparison. 30 min denaturation at 65°C and 85°C were tested in combination with hybridization buffer A and hybridization buffer B for three amplification times: 15 min, 45 min and 120 min. The signal intensities were measured via microscopy and are given in RU. The choice of hybridization buffer did not make a significant difference to signal intensity, but HCR-FISH signals were slightly higher when buffer B was used (see materials and methods for details in composition). Increasing chain reaction amplification time from 15 or 45 min to 120 min enhanced fluorescence from 0.3 (15 min) and 0.4 RU (45 min) to 1.1 RU (average values).

Additional file 11: Figure S11. Signal intensity of four ethanol fixed isolates after HCR-FISH with 30 min denaturation (45-85°C) or without denaturation (-) and 2 h hybridization in comparison to CARD-FISH (CARD). Signal intensities were measured via microscopy and are given in RU.