Large shifts among eukaryotes, bacteria, and archaea define the vertical organization of a lake sediment

Sediments are depositional areas where particles sink from water columns, but they are also microbial hotspots that play an important role in biogeochemical cycles. Unfortunately, the importance of both processes in structuring microbial community composition has not been assessed. We surveyed all organismic signals of the last ca. 170 years of sediment by metabarcoding, identifying global trends for eukaryotes, bacteria, archaea, and monitored 40 sediment parameters. We linked the microbial community structure to ongoing and historical environmental parameters and defined three distinct sediment horizons. This not only expands our knowledge of freshwater sediments, but also has profound implications for understanding the microbial community structure and function of sediment communities in relation to future, present, and past environmental changes.


Introduction
Box 1. Definition of "present" and "past" sediment parameters. We define present parameters as the principal components of all context data derived from a) pore water analysis, which indicates that chemical gradients are caused by the consumption and production of ongoing biological processes (e.g., sulfate and methane), as well as from b) directly measured parameters of microbial activities (e.g., bacterial protein production). The present parameters are therefore an expression of recent microbial processes.
Past parameters are the principal components of conservative parameters, which once introduced into the sediments will not change significantly and are therefore an expression of the lake's history (e.g., heavy metals). Here, we also categorize the total amount of elemental carbon, nitrogen, hydrogen, and sulfur as mainly conservative parameters. The past parameters are therefore an expression of historical changes. 75

76
The sediment cores were black in color, with no visible lamination, and they had a high 77 water content (93 − 97%). Macrozoobenthos was visually absent, and metabarcoding 78 (see below) detected only the presence of nematodes ( Figure S1). A consistent pattern 79 across all four sites was the exponential increase of dissolved refractory carbon with  [34]. 82 Mean prokaryotic cell numbers were 1.8 ± 0.5 · 10 9 per ml of wet sediment and 83 were higher in the uppermost sediment horizons. Bacterial biomass production (BPP) 84 (range: 0 − 282 µg · ml −1 d −1 ) decreased rapidly with depth, approaching zero below 85 10 cm. Total DNA concentration (range< 0.3 − 17.6 µg · ml −1 sediment) decreased 86 exponentially with depth and was negatively correlated with FI (r = −0.886). 87 DNA half-life was inferred to be t 1/2 = 22 a (corresponding to 5.4 cm; f (DN A) = 88 13.9 · e −0.128x , r 2 = 0.81). The RNA content of the sediment was lower than DNA at   Figure S3), but it reached significance only at the deepest sampled layer (26 − 30 cm 114 depth). Taxon turnover/replacement was consistently high and significant for multiple 115 sediment depths within the first 14 cm (Table 1).

116
The OTUs that were the most influential in shifting the community structure 117 contribution to ß-diversity ( Figure S4) [35]. The identities of these 96 "structuring"

119
OTUs indicate an interrelationship of all three domains, with the most influential 120 phyla being Euryarchaeota and Thaumarchaeota in the archaea and Chloroflexi, 121 Proteobacteria, and Phycisphaerae in the bacteria. Redox-dependent groups (13% 122 of the taxa could be clearly assigned to redox processes by their classification, e.g.,

125
Out of these 96 "structuring" OTUs, we identified the OTUs that are significantly 126 elevated in each of the three zones ( Figure 2). While we see eukaryotic and bacterial   Figure 2. Overview on Lake Stechlin's sediment structure. The cluster analysis separates three depth horizons: The redox-stratified zone (0 − 5 cm), which includes a thin layer of oxygen. A few fauna species exist in this zone, i.e., Nematoda, Gastrotricha, and microeukaryotes (e.g., Ciliophora), in addition to large numbers of highly active bacteria. Below 5 cm, where 50% of the DNA is already decomposed, the system enters the transition zone. This zone is situated below the sulfate-methane transition. Below 14 cm, we find the depauperate horizon, which extents in the deeper sediment, in which archaea dominate the community. In an extrapolation of the richness component of the community structure, the loss of richness would completely dominate (100%) the microbial community at 1 m depth (approx. 500 a). Following the decay curve of the DNA, 99.99999% of the DNA would be transformed at that depth. On the right side, the 10 most structuring OTUs (from Figure S4) are listed, which were significantly elevated in the corresponding horizon (only results with p < 0.01 in the TukeyHSD PostHoc Test, were included). The brackets ab and bc mark those OTUs that were elevated in the upper two or lower two zones, respectively. Only two OTUs were elevated in the transition zone. The grey box marks the single taxon that was significantly different in all three horizons. Taxon names are color coded according to their classification or phototrophy if applicable: phototrophic organism (green), eukaryotes (black), bacteria (red), archaea (blue). OM, the sediment habitat is considered to be an autonomous system in terms of 172 species diversity and community structure [17,18], with the exception of surface 173 supply of planktonic OM, including decaying eukaryotes ( Figure S4). Indications 174 that a high species turnover, as suggested by our results, may be a common feature 175 of vertical sediment profiles have been reported for bacterial taxa in coastal marine 176 sediments [38], for marine archaea and bacteria [39], and for freshwater archaea [20].

177
Previous freshwater sediment studies that found a more moderate species turnover 178 were restricted by the use of low-resolution methods [15,20,40]  Our results pointed to three depth clusters. Based on the high replacment (Table   184 1) and measurable microbial production in the first two depth clusters (cluster a, b) 185 (Figure 1), we decided to classify them into one overarching horizon, which we called 189 Box 2. Sediment zonation according to taxonomic clustering, ß-partitioning, and context data.
• I. The replacement horizon (1 − 14 cm|ca. 0 − 60 a) is subdivided into two zones, delineated by the sulfate-methane transition, which falls approximately within or correlates with a local prokaryotic cell maximum.
-Cluster a) redox-stratified zone (0 − 5 cm|ca. 0 − 20 a). This zone encompasses the redox-stratified zone and the typical sequence of electron acceptors from oxygen to sulfate. N H + 4 and SRP increases. It is characterized by high microbial activity, cell numbers, taxa turnover, spatial variability, low DNA:RNA ratios, and low FI values (1.7 − 1.8). Potentially 50% of the sedimented organic matter is metabolized in this horizon, aligning with the rapid decrease of eukaryotes. Bacteria dominate this horizon.
-Cluster b) transition zone (5−14 cm|ca. 20−60 a). This zone is characterized by strong gradients. Microbial cell numbers drop off and activity decreases, diversity decreases, the taxa turnover stays high, the DNA:RNA ratio doubles, and FI values increase (1.8 − 1.9). Methane concentrations rise, and CO 2 has a local minimum. Potentially, another 35% of the sediment organic matter is turned over in this horizon. Eukaryotes approach zero, bacteria decline, and archaea rise rapidly in their community contribution.
Here, we also see a maximum of taxa with no close relatives to known database entries.
• II. The depauperate horizon (14−x cm|ca. > 60 a) is very distinct from horizon I, and it is unclear how far this horizon reaches. It appears to be archaea-dominated and is characterized by a loss in richness and a shift toward the dominance of single taxa. The replacement horizon Bacterial activity was highest in the redox-stratified 191 zone (cluster a), which is where most of the settled OM is available. As sulfate is 192 13/40 depleted at 5 cm depth, most redox processes will take place above this. The majority 193 of freshwater sediment studies examine this zone in great detail (e.g., [4,5,10]), 194 including the identification of redox processes at the millimeter scale [41,42]. Highly 195 active decomposition processes lead to high prokaryotic cell numbers close to the 196 sediment surface (e.g., [5,24,43], as well as a constantly lower abundance in deeper 197 horizons [15]. There was also an initial loss of many taxa in the highly active oxycline serve as an electron donor and acceptor [46,47] and because OM quality is known 214 to modulate microbial redox processes [48], with apparent consequences for carbon 215 turnover rates [33,47]. be high in sediments [50,51]. In particular, viral lysis of cells is supposed to be very 241 important in sediments [52,53]. We found indications for cellular recycling caused by 242 the predatory Bacteriovoracaceae (cf. [54]), which is one of the structuring bacterial 243 lineages identified in Figure S4. Another potential mechanism -one that may be most as they occur in sediments [58]. Deep sediment layers may offer specific energy-288 poor microniches favoring a high variety of syntrophic microorganisms [59,60]. The

289
Dehalococcoidales (Chlorflexi [61,62]) and the Miscellaneous Crenarchaeotic Group 290 (MCG, [63,64]) are promising candidates for such hybrid forms of energy harvesting 291 and are both among the most influential lineages in our data ( Figure S4). In our 292 data, the MCG co-occur with Dehalococcoidales, similar to their presence in Lake Baikal's (> 1500 m water depth) methane hydrate-bearing sediment [65]. Another lineage can exhibit a cell-to-cell coupling, allowing for thermodynamic processes that would otherwise not be possible [66].  [67] in Lake Biwa (Japan) 307 point to an increase in the proportion of archaea with sediment depth. However, the 308 results obtained by quantitative PCR for Lake Taihu (China; [14]) and Lake Pavin 309 (France; [20]) fail to show such a relationship. Archaea have, on average, compact 310 genomes [69] and a lower ribosomal copy number than bacteria [70], which may 311 lead to underestimates of the actual archaeal abundance. Similar to our results, [20] 312 found three sequential depth clusters in the archaeal community structure within are suspected to metabolize detrital proteins ( [19], discussed in more detail below).

319
Interestingly, we also identified a MCG-B as structuring OTU for the transition zone 320 (Figure 2), a group which was recently described as eukaryotic progenitor from a 321 18/40 hydrothermal vent field (Lokiarchaeota, [71]). Also, several MCG OTUs belonged to 322 the top structuring taxa. MCG was recently named as Bathyarchaeota by [72] for 323 its deep-branching phylogeny and its occurrence in deep subsurface environments 324 -environmental conditions that our cores (30 m water depth and 30 cm length) did not 325 meet. Our results suggest that the specific niche adaptation of these microbes is not 326 necessarily related or restricted to the deep biosphere but rather to a cellular state of 327 "low activity" [73]. In this context, it is interesting that single MCG OTU sometimes important role, and we were able to identify important sediment taxa for each horizon. 349 We put a spotlight on the largely unexplored freshwater sediments and confirmed earlier findings that were previously described only for marine sediments, such as the  To determine the DNA:RNA ratio, we extracted total nucleic acids using a phenol-429 chloroform protocol from 200 − 400 µl sediment, as described by [82].