Meta-analysis of quantification methods shows that archaea and bacteria have similar abundances in the subseafloor

Abstract

There is no universally accepted method to quantify bacteria and archaea in seawater and marine sediments, and different methods have produced conflicting results with the same samples. To identify best practices, we compiled data from 65 studies, plus our own measurements, in which bacteria and archaea were quantified with fluorescent in situ hybridization (FISH), catalyzed reporter deposition FISH (CARD-FISH), polyribonucleotide FISH, or quantitative PCR (qPCR). To estimate efficiency, we defined "yield" to be the sum of bacteria and archaea counted by these techniques divided by the total number of cells. In seawater, the yield was high (median, 71%) and was similar for FISH, CARD-FISH, and polyribonucleotide FISH. In sediments, only measurements by CARD-FISH in which archaeal cells were permeabilized with proteinase K showed high yields (median, 84%). Therefore, the majority of cells in both environments appear to be alive, since they contain intact ribosomes. In sediments, the sum of bacterial and archaeal 16S rRNA gene qPCR counts was not closely related to cell counts, even after accounting for variations in copy numbers per genome. However, qPCR measurements were precise relative to other qPCR measurements made on the same samples. qPCR is therefore a reliable relative quantification method. Inconsistent results for the relative abundance of bacteria versus archaea in deep subsurface sediments were resolved by the removal of CARD-FISH measurements in which lysozyme was used to permeabilize archaeal cells and qPCR measurements which used ARCH516 as an archaeal primer or TaqMan probe. Data from best-practice methods showed that archaea and bacteria decreased as the depth in seawater and marine sediments increased, although archaea decreased more slowly.

Fig 1

Yields for seawater (A) and sediments (B) obtained by FISH, CARD-FISH with proteinase K permeabilization for archaea, CARD-FISH with other permeabilization methods for archaea, and polyribonucleotide FISH. The solid line is the 1:1 line, and the numbers of data points in each method group are listed in the corresponding color in each panel. In all CARD-FISH analyses, lysozyme was used to permeabilize bacteria.

Fig 2

Sediment FISH and CARD-FISH yield box plots colored by archaeal permeabilization method and ordered by the decreasing yield of each individual study. Black line, median seawater yield; blue shading, interquartile range (i.e., the range bounded by the 25th and 75th percentiles of the yields). Samples from >250-m water depths are labeled with “deep-sea” above the data. Numbers in the color ribbon at the top indicate the number of data points in each box plot, and numbers in brackets after each core description correspond to the source citations in Table S1 in the supplemental material. Data from studies with only one data point and studies in intertidal environments (see Fig. S3 in the supplemental material) were excluded from the plot. HMMV, Haakon Mosby Mud Volcano.

Fig 3

Evaluation of effects of archaeal permeabilization method on CARD-FISH yield (A and B) and fraction of archaea (C and D) for all data (A and C) and sediments shallower than 1 m (B and D). Letters under the box plots indicate statistically indistinguishable groups, based on ANOVA on rank-ordered data with the Tukey honestly significant difference post hoc analysis (P < 0.001), and the numbers of data points and individual studies are listed above each box plot (box plot values are listed in Table S5 in the supplemental material). Colors and shadings match those in Fig. 2.

Fig 4

Evaluation of qPCR counts relative to cell counts. (A) Sum of separately measured bacterial and archaeal 16S rRNA gene copy numbers versus cell counts colored by core. Gray-shaded region, 95% prediction interval for the data in aggregate (i.e., 95% of future measurements are predicted to be within this interval); green-shaded areas, known variation in the range of 16S rRNA gene copy numbers per genome (dark green = 3.04 per genome; light green = 24 per genome). Each sediment core is coded by color and has its own fit line; data source citations are in Fig. S4 and Table S1 in the supplemental material. (B) Standard deviation (Std. dev.) of qPCR counts (open blue circles) or yields by CARD-FISH with proteinase K archaeal permeabilization (closed red circles) versus standard deviation for cell counts for each core. (C) Sum of separately measured 16S rRNA gene copy numbers of archaea and bacteria versus a single measurement obtained with universal primers, with the 95% confidence interval shown in gray. In all panels, solid black lines are the 1:1 line and dashed black lines are linear regressions of the data.

Fig 5

Results of best-practice quantification methods. (A to C) CARD-FISH and FISH quantifications of bacteria (A), archaea (B), and the fraction of archaea relative to the sum of bacteria and archaea (C) in seawater; (C to F) cell densities measured with CARD-FISH with proteinase K for archaeal permeabilization (red) or estimated from relative qPCR measurements and total cell counts (blue) for bacteria (D), archaea (E), and the fraction of archaea relative to the sum of bacteria and archaea (F) in marine sediments. Dotted lines, linear regressions of data of the same color; gray shading, 95% confidence interval. Fit parameters, breakpoints, and statistical parameters are provided in Table S7 in the supplemental material.