Layout 1 INTRODUCTION Freshwater ecosystems are tightly linked to the sur- rounding terrestrial landscape, which continuously ex- ports soil material and microorganisms to the water. Understanding the functioning and structuring of lake and river bacterioplankton communities therefore requires taking into consideration these potential linkages between habitats. However, most current work on microbial bio- geography is restricted to single types of ecosystems (i.e., only lakes, only soils, only rivers), thus disregarding the potential dispersal of taxa between communities from similar or different habitats, and rendering a rather frag- mented view of microbial biogeography. The few recent studies considering such land-water linkages have shown not only that this connectivity with the terrestrial environment may influence aquatic micro- bial communities by providing nutrients and dissolved or- ganic matter of different quality (Besemer et al., 2013; Berggren and del Giorgio, 2015; Ruiz-González et al., 2015a; Wilhelm et al., 2015), but also that the immigra- tion of microbes from the surrounding catchment can largely explain the local composition of the receiving community. For example, terrestrial runoff causes the ad- vection of high numbers of bacterial taxa to aquatic bod- ies, and consequently the systems most tightly connected to the landscape will be more strongly influenced by this transport of terrestrial bacteria. Although headwater streams have been suggested to act as integrators and vec- tors of microbial diversity from soil sources (Crump et al., 2007; Nelson et al., 2009; Besemer et al., 2012, 2013), only a few studies have directly explored this (Savio et al 2015; Ruiz-González et al., 2015a; Niño-García et al 2016a), showing that most of these dispersed bacteria, which dominate communities in headwaters, gradually disappear along the river continuum. This results in a di- Advances in Oceanography and Limnology, 2017; 8(2): 222-234 ARTICLE DOI: 10.4081/aiol.2017.7232 This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). Contrasting dynamics and environmental controls of dispersed bacteria along a hydrologic gradient Clara Ruiz-González,1* Juan Pablo Niño-García,2 Martin Berggren,3 Paul A. del Giorgio4 1Institut de Ciències del Mar (ICM-CSIC), Passeig Marítim de la Barceloneta 37-49, 08003, Barcelona, Spain; 2Escuela de Microbiología, Universidad de Antioquia, Ciudad Universitaria Calle 67 No 53-108, Medellín, Colombia; 3Department of Physical Geography and Ecosystem Science, Lund University, Sölvegatan 12, SE-223 62, Lund, Sweden; 4Groupe de Recherche Interuniversitaire en Limnologie et en Environnement Aquatique (GRIL), Département des Sciences Biologiques, Université du Québec à Montréal, Case Postale 8888, succursale Centre-Ville, Montréal, QC, H3C 3P8, Canada *Corresponding author: clara.ruiz.glez@gmail.com ABSTRACT Freshwater bacterioplankton communities are influenced by the transport of bacteria from the surrounding terrestrial environments. It has been shown that, although most of these dispersed bacteria gradually disappear along the hydrologic continuum, some can thrive in aquatic systems and become dominant, leading to a gradual succession of communities. Here we aimed at exploring the environmental factors driving the structure of such contrasting bacterial populations as well as their functional properties. Using Illumina sequencing of the 16S rRNA gene, we characterized the taxonomic composition of bacterioplankton communities from 10 streams and rivers in Québec spanning the whole hydrologic continuum (river Strahler order 0 to 7), which were sampled in two occasions. With the aim to understand the fate and controls of the transported bacteria, among the taxa present at the origin of the hydrologic gradient (i.e., in the smallest headwater streams) we identified two types of dynamics: i) ‘Tourist’ taxa, which were those that decreased in abundance from the headwaters towards the largest rivers, and ii) ‘Seed’ taxa, those that increased their abundances along the hydrologic continuum. Communities changed gradually from the fast-flowing headwater streams dominated by ‘Tourist’ taxa (ca. 95% of the sequences) towards the largest rivers (Strahler order 4-7) where ‘Seed’ taxa comprised up to 80% of community sequences. Variation in taxonomic composition of the communities dominated by ‘Tourist’ taxa in streams seemed related to different degree of terrestrial inputs, whereas compositional changes in ‘Seed’ communities in the large rivers were linked to differences in autochthonous processes. Finally, the two types of communities differed significantly in their metabolic potential assessed through Biolog Ecoplates. All this suggests that hydro- logic transport modulates the gradual replacement of two contrasting population types subjected to different environmental controls and with different metabolic potentials. Moreover, we show that the separate exploration of the two pools of taxa allows unveiling en- vironmental drivers and processes operating on them that remain hidden if explored at the whole community level. Key words: Dispersal; river continuum; environmental sorting; aquatic bacterial communities; Illumina sequencing. Received: December 2017. Accepted: December 2017. No n- co mm er cia l u se on ly 223C. Ruiz-González et al. rectional structuring of aquatic communities and a gradual decrease in taxonomic richness that has been observed both along individual rivers (Savio et al 2015) or across multiple unconnected aquatic sites (Ruiz-González et al., 2015a; Niño-García et al 2016a). More importantly, how- ever, some of these transported soil bacteria seem to be able to grow during transit in the water and to dominate freshwater bacterioplankton communities when the water residence time is long enough, i.e., in large rivers and lakes (Crump et al 2012, Ruiz-González et al., 2015a). The residence time and the distance travelled by a given parcel of water will thus determine the relative pro- portions of these two contrasting pools of communities that move along the aquatic continuum. The most basic distinction is between the ones that thrive and adapt to local conditions (i.e., ‘seed’ taxa, sensu Ruiz-González et al., 2015a) and those that are mal-adapted to the aquatic environment and which therefore decrease in abundance along the hydrologic gradient due to death or dilution (i.e., ‘tourist’ taxa, sensu Newton et al., 2011). In support of this idea, several studies have highlighted the role of hy- drology, and water residence time in particular, as a major driver of the assembly of aquatic communities in fresh- water systems. For example, it was shown that the differ- ence between bacterioplankton communities in lakes and the immigrant assemblages arriving through their inlets varies as a function of lake water residence time (Lind- ström and Bergström 2004; Lindström et al., 2006). In this regard, Crump et al. (2007) suggested that bacteria transported by advection from upstream sources should constitute a small fraction of the total cells in systems with water residence times longer than bacterial doubling times. Other studies have explicitly explored the dynamics of different groups of bacterial taxa, showing that, whereas communities at the origin of the hydrologic con- tinuum (i.e., in the smallest headwater streams) are mostly dominated by taxa washed from soils that likely cannot thrive in aquatic systems, downstream systems with longer water residence times mostly comprise typical freshwater taxa adapted to local conditions (Savio et al., 2015; Ruiz-González et al., 2015a; Niño-García et al., 2016a). Related to this, Niño-García et al., (2016a) found that, although pH seemed to shape bacterial communities along the whole aquatic continuum in the boreal land- scape (ca. 400 streams, rivers and lakes), its role as a driver of taxonomic differences was much more important in large rivers and lakes than in headwater streams where water residence times are shorter. The observed response to pH in fast-flowing streams was interpreted as a signa- ture of the landscape past environmental conditions (Niño-García et al., 2016a) rather than to in-stream selec- tion. Since this was done considering the overall commu- nities, which harbour both types of populations, it is possible that more clear links between environmental drivers and taxonomy will appear considering the two populations separately. For example, catchment properties should be more important in shaping the observed taxo- nomic composition of ‘Tourist’ taxa in headwater streams communities (Lear et al., 2013), whereas communities from large rivers or lakes should be mostly structured by local aquatic conditions (Logares et al., 2013; Souffreau et al., 2015). The corollary of the above is that the succession and replacement of populations with increasing water resi- dence time should be accompanied by shifts in the func- tional capacities of the communities, since the selection and growth of ‘Seed’ taxa along the aquatic continuum should be mostly driven by environmental factors that re- quire a specific functional response. This link between the functional and taxonomic successions that occur along the aquatic continuum may be blurred by the presence of a large fraction of ‘Tourist’ taxa that do not necessarily con- tribute to community functioning. In this regard, Niño- García et al. (2016b) showed that the presence of most rare taxa across lake bacterioplankton assemblages was due to hydrologic transport from the associated rivers and, although they may not be functional, they comprised most of the species richness detected across lakes. This may ex- plain why, in a previous study where we explored the large-scale spatial patterns in the metabolic profiles of bo- real lake and river bacterioplankton communities (Ruiz- González et al., 2015b), we found little or no coupling between changes in structure and function. This of course has major implications on our capacity to predict how variations in the observed taxonomic composition may impact community metabolism or their responses to en- vironmental perturbations, and calls for the need to parti- tion assemblages into components that are subjected to different controls, in order to explore the links between the environment, taxonomic identity, and community functioning in these freshwater microbial assemblages. Here we aimed at exploring the fate, the environmen- tal drivers and the metabolic profiles of the dispersed bac- terial populations along a hydrologic gradient. To do so, using Illumina sequencing of the 16S rRNA gene, we as- sessed the taxonomic composition of bacterioplankton communities inhabiting 10 different rivers from two tem- perate regions in Québec, ranging from small headwater streams (Strahler order 0 or 1) to large rivers (Strahler order 7). The selected rivers flow through very different landscapes, including forested areas, wetlands and scrub- lands (Berggren and del Giorgio, 2015), which had been previously shown to influence the patterns of microbial metabolism in these same rivers (Berggren and del Gior- gio, 2015). We searched for two types of dynamics among headwater bacterial taxa: i) those that decreased in abun- dance along the hydrologic gradient (i.e., ‘Tourist’ taxa, sensu Newton et al., 2011), which thus represent taxa that No n- co mm er cia l u se on ly 224 Dispersed bacteria in a river continuum may not be able to thrive in downstream waters, and ii) those that increased in abundance likely due to in-river selection and growth (i.e., ‘Seed’ taxa, sensu Ruiz- González et al., 2015a). In order to understand the eco- logical underpinnings of the two populations, we explored the environmental variables explaining compositional changes of these communities at both extremes of the hy- drologic gradient (i.e., where they were dominant), as well as their metabolic potential using Biolog Ecoplates. We hypothesized that there should also be a watershed-spe- cific signature in the taxonomic composition of the com- munities dominated by ‘Tourist’ taxa, which reflect the broad catchment differences, whereas assemblages dom- inated by ‘Seed’ taxa should reflect in-river physicochem- ical conditions. METHODS Study sites, sampling and basic parameters The sampling design has been previously described in Berggren and del Giorgio (2015). Briefly, we sampled 10 streams and rivers for the characterization of bacterio- plankton communities with Illumina sequencing (Illumina Inc., San Diego, CA, USA) of the 16S rRNA gene. The streams were located in two regions of Quèbec, Canada, and spanned the whole hydrologic gradient ranging from small headwater streams of Strahler order 0 or 1 to large rivers (order 6 or 7); they were also selected to maximize the differences in catchment type (peat bogs versus a range of forest soils) and total upstream distance of each site (Tab. 1). Additional details on the environmental char- acteristics of the sites can be found in Tab. 1 of Berggren and del Giorgio (2015). Each site was sampled twice be- tween 31 May and 24 August 2010, during late spring and summer low-flow conditions, in order to assess the con- sistency of the patterns observed. Two of the rivers could be sampled only once (Tab. 1). Water was collected at depths of 20-30 cm and all samples were stored in cooling boxes in the dark until analysis. Water temperature, dissolved oxygen (DO), pH and conductivity were measured in situ with an YSI probe. At each site, a water sub-sample was filtered in situ through 0.45 µm (PES cartridge, Sarstedt) and stored in acid-washed glass vials for dissolved organic carbon (DOC), nutrients and optical analyses, and another sample was kept in acid-rinsed bottles for DNA processing in the laboratory and to measure the substrate utilization profiles using Biolog EcoplatesTM (see below). DOC concentration was measured on an OI1010 TOC analyser. The optical properties of DOC were measured as indices of its composition and source. Coloured dissolved organic matter (cDOM) was quantified as the absorbance at 440 nm using an Ultrospec 3100 spectrophotometer. DOC composition was described on the basis of fluores- cence absorption/emission spectra (EEMS) measured in a Shimadzu RF5301 PC spectrofluorophotometer across ex- citation/emission wave- lengths of 275-450 nm and 280- 600 nm, respectively. Six main fluorescence components were recovered from the EEMS using parallel factor analy- sis (PARAFAC): Components 1 to 3 (C1-C3), previously related to refractory, humic material of presumably terres- trial origin, and C4-C6, associated with more biolabile freshly produced DOM (for details see Lapierre and del Giorgio, 2014; Stubbins et al., 2014). C5 and C6 have been linked to autochthonous processes, having C6 a fluores- cence signature typical of protein-like material. The per- centage contribution of each component was calculated relative to the total fluorescence of the six PARAFAC com- ponents. Geographical analyses River length, total upstream distance and Strahler order, catchment areas and elevation of the sampled sites were obtained using the ARCMAP 10 and ARCGIS V10 Tab. 1. Characteristics of the different sampling sites ordered based on their distance to their respective headwaters. Most rivers were sampled in two occasions, but two sites were sampled only once. Name Region Season Strahler river order Upstream distance (m) Catchment type Stream # 09 Abitibi Spring/summer 0 454 Broad-leaf forest Stream # 08 Abitibi Spring/summer 1 1193 Broad-leaf forest Stream # 07 Abitibi Spring/summer 1 1828 Broad-leaf forest Petit Nord Suivit Baie James Spring/summer 2 2205 Bog Petit Aiguebelle Suivit Abitibi Spring/summer 2 2951 Mixed forest Ruisseau Brunet Abitibi Spring/summer 4 7149 Mixed forest Inter-Nord Suivit Baie James Spring 4 19575 Bog Magousi Abitibi Summer 6 66488 Clay soils Kinojevis Abitibi Spring/summer 7 172271 Clay soils Harricana Baie James Spring/summer 7 232654 Bog No n- co mm er cia l u se on ly 225C. Ruiz-González et al. software (ESRI Inc., Redland, CA, USA) applied on Dig- ital Elevation Models (DEM) derived from (1:50 000) maps. For each catchment, we averaged slope based on DEM, and determined various landcover properties from Geobase (2009). Bacterial community taxonomic composition Bacterial biomass was collected filtering 300–500 mL of water through 0.22 mm (47 mm, GSWP, Millipore). Bacterial DNA was extracted using the MoBio (Carlsbad, CA, USA) DNA extraction kit, following the manufac- turer’s instructions. Tagged amplicons of the 16SrRNA gene (V3-V4 region) were obtained with the primers 515F and 806R using a two-step PCR and the Access Array bar- code library (Fluidigm), and sequenced on an Illumina MiSeq2000 using a paired-end approach (Caporaso et al., 2012). Paired-end reads were assembled with FLASH (Magoc and Salzberg, 2011) and sequences between 250 and 290 bp were used for downstream analyses in QIIME to remove primers, low-quality, archaeal and chloroplast reads (Caporaso et al., 2010). Quality sequences were binned into operational taxonomic units (OTUs, ≥97% similarity) using UCLUST v1.22q (Edgar, 2010) and RDP classifier (Wang et al., 2007). Representative sequences were then aligned against the SILVAv108 reference align- ment (Pruesse et al., 2007). To ensure that rare bacteria were not the result of sequencing errors, we discarded all OTUs showing <4 sequences. To enable comparisons be- tween samples, the OTU table was randomly subsampled to ensure an equal number of sequences per sample, based on the sample with the least number of reads (72,798 se- quences). Identification of ‘Seed’ and ‘Tourist’ OTUs In order to explore the dynamics of the taxa dispersed along the hydrologic continuum, we first selected all the OTUs present in the smallest streams of Strahler order 0 (those that did not appear in topographic maps because they are too small or temporary) or order 1. Among those, we further categorized OTUs as i) ‘Tourists’ (sensu Newton et al., 2011), if their abundances were higher in streams than in the largest rivers of order 6 or 7 (which may represent terrestrial taxa washed from soils that cannot thrive in aquatic conditions), or ii) ‘Seeds’, if their abundances were higher in the largest rivers than in headwaters of order 0 and 1 (and thus might represent taxa seeding aquatic com- munities; Ruiz-González et al., 2015a). Bacterial substrate utilization patterns The capacity of the riverine bacterial communities to respire 31 carbon substrates was assessed using the Biolog EcoplateTM (CA, USA) as detailed in Berggren and del Giorgio (2015). Briefly, one plate per site was incubated in the dark and at 20°C, and the absorbance was measured periodically (every 6 or 12 hours depending on colour de- velopment rates) until an asymptote was reached, between 2 and 9 days. At each time point, the overall colour de- velopment of the plates was expressed as the average well colour development (AWCD), and the mean colour de- velopment of each compound was calculated as the blank- corrected mean absorbance of each substrate measured at the time when the AWCD was closest to 0.5 (Garland et al., 2001), usually between days 1 and 4. All Biolog in- cubations were initiated 2-6 h after field sampling. Statistical analyses Differences in bacterial taxonomic composition be- tween ecosystem types were tested with ANOSIM (Clarke, 1993). The Shannon index was calculated as an estimate of bacterial taxonomic diversity. Bray-Curtis dis- tance was used as an estimator of taxonomic dissimilarity between samples, and was visualised by non-metric mul- tidimensional scaling (NMDS) analysis. We performed least squares linear regression to establish relationships between individual variables. All analyses were per- formed in JMP 9.0.1 (SAS Institute, NC, USA) or R 3.0.0 software (R Core Team, 2013; Vegan package). RESULTS After rarefaction, sequencing of the 16S rRNA gene from the 18 samples resulted in 1,310,364 quality se- quences, which clustered into 64,489 OTUs. Altogether, 59 bacterial phyla were detected, yet sequences were mainly dominated by the phylum Proteobacteria (45% of the total number of sequences), followed by Actinobacte- ria (14%), Bacteroidetes (12%), and OD1 (7%) and Ver- rucomicrobia (4%). The number of OTUs per site ranged from 4376 to 17604, and we observed pronounced de- creases in OTU richness and taxonomic diversity (i.e., Shannon index) along a gradient of increasing distance from the headwaters (Fig. 1 a,b). Interestingly, this pattern was also reflected in a clear structuring of communities depending on the distance travelled (Fig. 1f) regardless of the season, region, or catchment type (Figs 1 c-e). In order to explore the spatial dynamics of the taxa dispersed along the hydrologic continuum, we identified, among the OTUs present at the origin of the hydrologic gradient (i.e., in the smallest headwaters), two groups of taxa: ‘Tourists’, if their abundances decreased from head- waters to large rivers, or ‘Seeds’ if their abundances in- creased towards downstream waters (see Methods). We assume that whereas the former represent taxa washed from soils that cannot thrive in the aquatic environment, the latter comprise taxa (presumably also washed from soils, see Ruiz-González et al., 2015a) able to thrive in the aquatic No n- co mm er cia l u se on ly 226 Dispersed bacteria in a river continuum environment. Doing so, 38,725 OTUs were categorized as ‘Tourists’, and only 6026 as ‘Seeds’ (Fig. 2). Together, these two groups represented 92.4% of the total se- quences in the dataset. The remaining taxa, 19,738 OTUs, could not be clearly categorized in any of the two groups due to extreme rareness and lack of any dis- cernible spatial pattern, but these uncategorized taxa rep- resented a very small fraction of the total sequences (7.4%), meaning that these simple criteria recovered most of the diversity across the sampling sites. We observed clear contrasting patterns between the two pools of taxa (Fig. 2): ‘Tourist’ OTUs largely domi- nated headwater communities in all cases but showed pro- nounced decreases towards the largest rivers (Fig. 2a), whereas ‘Seed’ OTUs shifted from making up less than 9% of headwater sequences to numerically dominate the largest rivers (ca. 80% of sequences in rivers of order 6 and 7, Fig. 2b). The decrease in ‘Tourist’ OTU abundance was clearly due to a comparable reduction in the number of OTUs (Fig. 2c), but the dominance of ‘Seeds’ in down- stream ecosystems seemed caused by large increases in abundance of a small number of OTUs dispersed from the headwaters (Fig. 2d). Interestingly, in both cases we ob- served that such shifts in abundance stabilized after a cu- mulative distance of around 7 km from the headwaters, which corresponded to rivers of Strahler order 4 (Fig. 2). Beyond that point, both the ‘Tourist’ and ‘Seed’ OTUs did not show further changes in relative abundance regardless of the very large differences in the total upstream distance. In terms of the taxonomic composition, the two taxa pools differed largely from each other (Fig. 3), and whereas downstream rivers were numerically dominated by groups like Betaproteobacteria, Actinobacteria and Cyanobacteria, in headwater streams we found higher abundances of Alpha-, Delta- and Gammaproteobacteria, OD1 or OP3, among others (Fig. 3). Interestingly, this dis- tribution of taxonomic groups was remarkably similar be- tween the two sampling occasions. When the two taxa pools were visualized separately in an NMDS, we found that in both cases the communities were also structured following the patterns as a function of distance from the headwaters described above for the whole community (Fig. 4), and again this pattern was in- dependent of the season, region or catchment type (details not shown). Interestingly, moreover, we observed that, in those sites where either ‘Tourist’ or ‘Seed’ OTUs were most abundant, we could detect two different groups of communities based on their taxonomic composition (t1 and t2 in ‘Tourist’-dominated communities and s1 and s2 in ‘Seed’-dominated communities, see arrows in Fig. 4). Fig. 1. Change in OTU number (a) and taxonomic diversity (Shannon index, b) as a function of the measured distance between each sampling point and its headwaters. The lines are a best-fit smooth curve through the centre of the data calculated using weighted least squares. Each pair of dots represents two sampling points (spring, black; summer, grey) of the same site. c-f) Non-metric multidimensional scaling (NMDS) plots based on Bray-–Curtis distances of taxonomic composition of all communities, color-coded by season (c), catchment type (d), region (e) and distance from the headwaters (f). The size of the dot is proportional the number of OTUs per community. No n- co mm er cia l u se on ly 227C. Ruiz-González et al. Fig. 3. Taxonomic composition of ‘Seed’ and ‘Tourist’ OTUs along a gradient of increasing distance from the headwaters. Data are shown separately for the two sampling times, and are presented as the sum of total sequences associated to each group. The classification was performed at the phylum level in most cases, but the Phyla Proteobacteria (Prot) and Bacteroidetes (Bact) were split into classes. Fig. 2. Contribution of headwater OTUs (i.e., those OTUs detected in streams of order 0 and 1, see Methods) to total bacterial sequences (a,c) or OTUs (c,d) in aquatic communities as a function of the measured distance between each sampling point and its headwaters. Headwater OTUs were divided depending on their dynamics along the hydrologic continuum into ‘Tourist’ OTUs (OTUs whose abun- dances decreased towards rivers, a,c) and ‘Seed’ OTUs (OTUs whose abundances increased towards rivers, b,d). The lines are a best- fit smooth curve through the center of the data calculated using weighted least squares. Each pair of dots represents two sampling points (spring, summer) of the same site. No n- co mm er cia l u se on ly 228 Dispersed bacteria in a river continuum It is important to note that these two groups of sites did not differ largely in the relative abundance of either ‘Seeds’ or ‘Tourist’ OTUs regardless of the differences in distance travelled (Figs. 2 and 4). Therefore, the observed changes in composition between t1 and t2, and between s1 and s2, must be due to changes on the group member- ship as a response to environmental factors and not only to direct hydrologic transport. In order to assess so, using envfit analysis we explored the environmental variables that best explained the taxo- nomic differences between those two groups of sites (t1 vs t2 and s1 vs s2) for each of the taxa pools (i.e., the smallest headwater streams in the case of communities dominated by ‘Tourist’ OTUs, and the largest rivers in the case of communities dominated by ‘Seed’ OTUs, Fig. 5). We observed that the compositional differences between the two groups of communities dominated by ‘Tourist’ OTUs (t1 and t2) were associated with differences in the ratio of cDOM to DOC, which is a proxy of terrestrial in- fluence, as well as in the proportions of the DOM fluo- rescent components C1 (associated to processed humic-like material) and C3 (freshly produced humic-like material, Fig. 5a). Conversely, variation between the two groups of ‘Seed’-dominated communities in the largest rivers (s1 and s2) seemed to be due to differences in au- tochthonous processes, since chlorophyll a concentration and % of C5 (associated to algal material) component ap- peared as the variables most clearly explaining the ob- served pattern, together with another humic-like component C2 (Fig. 5b). In general, the taxonomic differ- ences between ‘Seed’ dominated communities were larger than those between communities dominated by ‘Tourist’ OTUs (see axes in Fig. 5). Finally, we investigated differences between the meta- bolic profiles (assessed by the Biolog Ecoplates) of all communities (Fig. 6). Interestingly, we observed clear metabolic differences between communities dominated by ‘Seed’ and ‘Tourist’ OTUs, with the exception of one site that corresponded to the only river of order 6, which grouped together with the headwater streams (Fig. 6a). These differences seemed mostly caused by variations in the use of four carbohydrates (D-Cellobiose, D-Lactose, Glucose, Erythritol), one carboxylic acid (2-Hydroxy Benzoic Acid) and one amino acid (L-Serine, Fig. 6a). In general, communities dominated by ‘Seed’ OTUs (i.e., large rivers) displayed higher capacity to respire L-Serine and Erythritol and lower capacity to use D-Cellobiose, D- Lactose and 2-Hydroxy Benzoic Acid than communities dominated by ‘Tourist’ OTUs. However, ‘Tourist’-domi- nated communities showed larger metabolic differences between groups t1 and t2 than ‘Seed’ communities, which showed less clear differentiation between s1 and s2 (Fig. 6a). Interestingly, we observed a gradient of increasing metabolic dissimilarity along a gradient of river order (Fig. 6b), suggesting that communities from the small headwater streams were more functionally similar to each other than communities located towards the lower end of the hydrologic continuum. Fig. 4. Non-metric multidimensional scaling (NMDS) plots based on Bray-Curtis distances of taxonomic composition of both the ‘Seeds’ and ‘Tourist’ OTUs (see legend of Fig. 1). The color gradient indicates the distance between each sampling point and its head- waters, and the size of the dot is proportional to the relative contribution of either ‘Seed’ or ‘Tourist’ OTUs to each community’ sequences. For each taxa pool, the two arrows indicate the two groups of communities where each type of OTU dominated (t1 and t2 in ‘Tourist’- dominated communities and s1 and s2 in ‘Seed’-dominated communities), which were compared in order to understand the environ- mental drivers explaining the observed changes in composition but not in abundance (see Figs. 5 and 6). No n- co mm er cia l u se on ly 229C. Ruiz-González et al. DISCUSSION Our results show a sequential structuring of riverine bacterioplankton communities, which changed from highly diverse communities inhabiting the smallest head- water streams, towards assemblages with much lower tax- onomic richness and diversity inhabiting the largest rivers (Fig. 1 ,b). This pattern was also reflected by an organi- zation of communities along a gradient of cumulative dis- tance from the headwaters (i.e., travelled distance, Fig. 1f) that coincided with concomitant increases in river order (details not shown) and likely also water residence time, although the later was not directly measured. The fact that the observed structure was independent from the geographic location of the rivers, the sampling period, and the catchment type (Fig. 1), supports previous results showing a major role of hydrology and landscape connec- tivity in shaping the structure of freshwater bacterioplank- ton communities (Crump et al., 2007; Nelson et al., 2009; Niño-García et al., 2016a). The high numbers of OTUs that we observed in the smallest headwater streams agrees with the notion that these systems act as integrators and dispersers of micro- bial diversity originating from the surrounding terrestrial landscape (Savio et al., 2015; Hauptmann et al., 2016; Ruiz-González et al., 2015a; Niño-García et al 2016a). Moreover, whereas the observed decrease in OTU rich- ness suggests that many of these dispersed bacteria cannot thrive in the water, the reduction in taxonomic diversity further implies that there is a progressive dominance of taxa adapted to aquatic conditions, as previously reported (Savio et al., 2015; Niño-García et al 2016a). In order to check whether we could differentiate such contrasting dy- namics within the dispersed taxa, we identified those OTUs present at the origin of the hydrologic continuum (i.e., streams of order 0 or 1) that i) decreased in abun- dance (i.e., ‘Tourist’ OTUs) or ii) increased in abundance (‘Seed’ OTUs) towards the largest rivers. Doing so, 58% of all OTUs were categorized as ‘Tourist’, and only 9% as ‘Seeds’ (Fig. 2), but together represented 92.4% of the total sequences in the dataset, suggesting that these simple criteria captured the most important community dynamics across the studied riverine ecosystems. We observed pronounced decreases in OTU number and relative contribution of ‘Tourist’ taxa from the head- waters to the largest rivers. These two parameters (‘Tourist’ OTU richness and abundance, Fig. 2 a,c) showed a very similar pattern and indeed largely covaried (R2=0.93, n=18, P<0.0001), suggesting that the observed abundance de- crease of ‘Tourist’ OTUs was likely the result of dilution or removal of taxa that could not grow in the aquatic envi- ronment. The presence of typical soil or groundwater groups such as Acidobacteria, Deltaproteobacteria, Thermi, OD1, OP3, Verrucomicrobia within the ‘Tourist’ group sup- ports the fact that their presence is largely due to advection from the catchment (Janssen et al., 2002; Lauber et al., 2009; Barberán et al., 2014; Hiller et al., 2015). Conversely, ‘Seed’ OTUs displayed pronounced increases in abundance along the hydrologic gradient that were not accompanied by comparable increases in the number of OTUs (Fig. 2 Fig. 5. PCA analysis of the communities where ‘Seed’ (a) and ‘Tourist’ (b) OTUs dominated, color-coded by the two groups of sites (t1 and t2 in ‘Tourist’-dominated communities and s1 and s2 in ‘Seed’-dominated communities) observed in Fig. 4. The vectors show the environmental variables that best fitted the ordination space (using the R envfit funtion). The size of the vector is pro- portional to the strength of the correlation of each variable. cDOM/DOC, ratio of coloured dissolved organic matter to dis- solved organic carbon concentration (a proxy of terrestrial inputs); Chla, Chlorophyll a concentration; %C1, %C2, %C3,%C5, per- centage contribution of the fluorescent parallel factor analysis components C1,2,3 and C5 to the total fluorescence. No n- co mm er cia l u se on ly 230 Dispersed bacteria in a river continuum b,d). This indicates that ‘Seed’ OTUs comprise a smaller pool of taxa, also present at the origin of the hydrologic continuum, which grow during transit through riverine net- works and dominate downstream assemblages. Based on previous studies, it is likely that most of these ‘Seeds’ are actually washed from the surrounding soils, where they can persist at low abundances (Crump et al., 2012; Ruiz- González et al., 2015a), maybe dormant (Lennon and Jones 2011). Moreover, this distinction between ‘Seeds’ and ‘Tourist’ OTUs is in accordance with the partition of boreal lake bacterioplankton communities into ‘core’ (adapted to local conditions) and ‘accidental’ (linked to hydrologic transport) components based on their large-scale spatial dy- namics (Niño-García et al., 2016b), and supports the notion that the main community assembly mechanisms change gradually along hydrologic gradients, shifting from com- munities strongly subjected to mass effects in the fast flow- ing headwater streams that are highly influenced by terrestrial inputs, to assemblages where local sorting of species will prevail (Lindström et al., 2006; Crump et al., 2007; Nelson et al., 2009; Ruiz-González et al.,, 2015; Niño-García et al., 2016a). Interestingly, we observed that such increases and de- creases in the abundances of ‘Seed’ and ‘Tourist’ OTUs, respectively, tended to stabilize beyond a travelled dis- tance of ca. 7 km (Fig. 2 a,b), corresponding to the rivers of order 4. This coincides with our own previous obser- vation, in a different region of boreal Québec, that the pro- portion of sequences associated to the soil-derived taxa able to grow in the water increased progressively from the smallest streams until rivers of order 4, beyond which their contribution to total sequences tended to plateau at ca. 70% (Ruiz-González et al., 2015a). Since this stabi- lization corresponded approximately to a water residence time of 10 days, it was suggested that beyond 10 days hy- drology no longer constrains the growth of taxa, and thus any further variations in community composition will be due to local selection driven by changes in aquatic physic- ochemical or biological conditions (Ruiz-González et al., 2015a; Niño-García et al., 2016a). Accordingly, here we observed that although beyond 7 km the relative contribution of ‘Seed’ OTUs remained re- markably constant (around 80% of the community se- quences), there were pronounced changes in taxonomic composition between those communities (likely due to changes in river physicochemical conditions), which showed two clear groups of sites corresponding to rivers of order 4 (s1) and rivers of order 6-7 (s2, Fig. 4b). Similarly, Fig. 6. a) PCA based on the metabolic profiles (Biolog Ecoplates) of all communities. Different shapes indicate whether ‘Tourist’ or ‘Seed’ OTUs dominated the communities (see arrows in Fig. 4). The vectors represent the carbon substrates that significantly explained the observed ordination; CA, carboxylic acid; CH, crbohydrate; AA, amino acid. b) Increases in metabolic dissimilarity (assessed as Euclidean distance between samples characterized through Biolog profiles) as a function of Strahler river order. No n- co mm er cia l u se on ly 231C. Ruiz-González et al. at the other end of the hydrologic gradient, communities dominated by ‘Tourist’ OTUs (i.e., those in the smallest headwater streams, order 0-2) also clustered into two dif- ferent groups of sites (t1 and t2, Fig. 4a). We expected that such compositional differences in community composition would be driven by different environmental factors at each extreme of the hydrologic gradient, since ‘Tourist’ taxa should mainly reflect the terrestrial environments where they originate, whereas ‘Seed’ taxa should be subjected to species selection driven by local aquatic environmental conditions. To assess so, we split the dataset into commu- nities dominated by ‘Tourist’ or ‘Seed’ OTUs, respectively, and explored their variability in relation to all the measured variables (Fig. 5). Variables associated to DOM quality emerged as the most strongly related to the ordination in both cases, but they differed between ‘Seed’- and ‘Tourist’- dominated communities (Fig. 5 a,b). For example, taxo- nomic differences between the two groups of communities dominated by ‘Tourist’ taxa in streams (t1 and t2) seemed to be related to differences in the ratio of cDOM to DOC (a proxy of terrestrial influence; Lapierre and del Giorgio, 2014) or in the proportion of the DOM fluorescent compo- nents C1 (associated to processed humic-like material) and C3 (freshly produced humic-like material, Fig. 5a). Al- though we could not see a clear clustering of communities based on the different catchment types shown in Tab. 1, not even for the headwater assemblages (Fig. 1), this observa- tion supports the hypothesis that headwater streams reflect to some extent a signature of the surrounding catchment, which in these same rivers had been shown to provide DOC of different quality (Berggren and del Giorgio, 2015). In this regard, Niño-García et al., (2016) observed a strong structuring effect of pH on boreal bacterioplankton com- munities, but attributed this to a regional landscape signa- ture on the immigrant pool rather than to species sorting due to pH itself. Similarly, Lear et al., (2013) found that the biogeography of stream biofilm bacterial communities was related to variability in catchment land use, supporting a strong influence of the surrounding terrestrial environment in systems with high connectivity with land. On the other hand, taxonomic variation across ‘Seed’ communities in the largest rivers appeared to be driven by differences in local aquatic processes, since chlorophyll a concentration and %C5 (associated to freshly-produced algal material) component appeared as significant envi- ronmental variables (Fig. 5b) and varied largely between the two groups of sites s1 and s2. Chlorophyll concentra- tions were higher in larger rivers independently of the sur- rounding catchment type (details not shown), which agrees with the expected pattern of downstream increase in planktonic chlorophyll, associated with longer resi- dence times and increases in nutrients (Sabater et al., 2008), and this likely triggered the growth of bacterial taxa adapted to more productive conditions in s2 than in s1 sites. In any case, the observation that all these ‘Seed’ taxa, regardless of their taxonomic identity, were detected in the headwater streams supports that landscapes in these regions contain a reservoir of diversity able to colonize aquatic ecosystems if the water residence time is long enough (Ruiz-González et al., 2015a). Along this same line, it is remarkable to note that the studied streams and rivers were not physically connected; all of them were located in different watersheds in two dif- ferent geographic regions, separated by more than 350 km. Therefore, the fact that those OTUs present across the sam- pled headwaters included taxa that would become dominant in all of the larger rivers suggests that there must be a large regional pool of taxa from which aquatic communities re- cruit, which is not subjected to dispersal limitation. Inter- estingly, this idea of the landscape reservoir of aquatic diversity has been proposed at scales much larger than the present study (Niño-García et al., 2016a; Ruiz-González et al., 2015a, 2017) and is in accordance with the seed-bank concept, a standing reservoir of dormant microbes (i.e., mi- crobes in a reversible state of very low metabolic activity) that persist at low abundances but that can grow and be- come abundant upon changes in environmental conditions (Lennon and Jones, 2011). This further highlights that mi- crobial seed banks may transcend ecosystem types, such that viable microbes can persist under unfavourable condi- tions in some ecosystems until they are transported to a suit- able environment (Sjöstedt et al., 2012; Lee et al., 2013; Comte et al., 2014; Ruiz-González et al., 2015a, 2017). Soils in particular are known to contain a large fraction of dormant cells (Lennon and Jones, 2011), a fraction of which seems able to resuscitate and grow upon advection to the water. In any case, this fact challenges our current ecosystem-specific approaches to study bacterial commu- nities, and highlights that it will not be possible to under- stand how communities are assembled or structured without considering at least the most important potential sources of microbial dispersal. Finally, the fact that such a large fraction of the diver- sity may represent taxa that are not adapted to local con- ditions has important implications for our understanding of community functioning. This agrees with recent evi- dence suggesting that a potentially large fraction of the microbial diversity detected in biogeographic studies can be either dead, in a dormant state, or present as extracel- lular DNA (Lennon and Jones, 2011; Blagodatskaya and Kuzyakov, 2013; Carini et al., 2016). In this regard, ex- ploring the dynamics of different taxa across ca. 200 lakes Niño-García et al. (2016a) revealed that the presence of most rare bacterial taxa was due to hydrologic transport from the associated networks, and thus these taxa likely represent a species pool that is not contributing to the functioning of communities but that is detected in all se- quencing studies. For these reasons, we expected to find No n- co mm er cia l u se on ly 232 Dispersed bacteria in a river continuum a clearer link between taxonomy and the metabolic pro- files of the studied communities when considering sepa- rately the pool of ‘Seed’ and ‘Tourist’ taxa than when considering the whole communities, and in addition, we expected to find a stronger correlation between taxonomy and function for ‘Seeds’ than for ‘Tourist’ OTUs. How- ever, this was not the case: changes in taxonomic compo- sition were weakly but significantly correlated to changes in metabolic profiles in all three cases, i.e., when the whole community was considered (Mantel R=0.189, P<0.05), or when considering ‘Seeds’ (Mantel R=0.201, P<0.05) or ‘Tourist’ OTUs (Mantel R=0.215, P<0.05) sep- arately, and the strength of the correlation was very simi- lar in all cases. This may be due to the fact that the Biolog Ecoplates used for assessing the metabolic profiles do not measure in situ activity levels but rather a functional po- tential, because it is a method that relies on the growth of microorganisms on the selected substrates during an in- cubation of several days (Garland et al., 2001). Thus, the microorganisms that will grow on the plates may be rare in the original communities (Smalla et al., 1998), partic- ularly in headwaters where taxa do not have the time to attain high abundances due to the fast hydraulics. How- ever, we did observe clear metabolic differences between communities dominated by ‘Seeds’ and ‘Tourist’ OTUs (Fig. 6a), suggesting that the functional structure or po- tential of these communities actually differs. Interestingly, moreover, whereas communities from the headwater streams were functionally more similar to each other, as- semblages in large rivers seemed much more heteroge- neous metabolically (Fig. 6b). This agrees with a previous study where we analyzed the metabolic profiles of ca. 300 freshwater bacterioplankton assemblages and found that communities from systems more tightly connected to the land were more similar metabolically than communities from larger lakes or rivers, and this increase in metabolic dissimilarity followed a gradient of increasing concentra- tions of autochthonously produced organic carbon (Ruiz- González et al., 2015b). This pattern may suggest that as systems depart from land, the structural diversification of the available DOM pool due to local processes (Kothawala et al., 2014) causes a functional diversifica- tion, whereas terrestrial DOM is more homogeneous or simply the pool of headwater taxa that can grow on the substrates is more restricted that in true aquatic commu- nities. It is important to note that this diversification of function along the aquatic flow path does not come in par- allel to a taxonomic diversification; in fact, there was an overall decline in OTU richness along this same contin- uum. In a previous study on these same rivers (Berggren and del Giorgio, 2015) the type of catchment (mainly for- est headwater versus peat influenced) was shown to be the strongest determinant of multiple facets of bacterial metabolism (e.g., biomass production, respiration, growth efficiency) presumably driven by watershed-scale differ- ences in DOC quality and quantity. Indeed, we found that catchment type was the variable explaining most of the metabolic differences between communities (Tab. 2). However, we also found weaker but significant differ- ences based on the four types of communities (Tab. 2), suggesting that the taxonomic composition also plays a role in determining the metabolic potential of communi- ties in these heterogeneous rivers. CONCLUSIONS Overall, our results show that bacterioplankton com- munities from lotic systems may comprise varying pro- portions of taxa adapted and not adapted to live in aquatic conditions, and this balance is largely driven by the time that bacteria have spent in a given parcel of water. Since the assembly mechanisms of both components differ, the environmental variables explaining shifts in taxonomic composition also change between the two pools of taxa, and whereas variables related with terrestrial inputs from the catchment explained differences in communities dom- inated by ‘Tourist’ OTUs, varying concentrations of chlorophyll and the associated DOC appeared to be re- lated to compositional changes in ‘Seed’-dominated com- munities. Finally, we found that the metabolic potential associated to these contrasting taxa pools also differed. All this highlights the need to consider the landscape con- nectivity between assemblages and the resulting dynamics of specific populations in order to fully understand the as- sembly mechanisms and environmental factors shaping these flowing communities. ACKNOWLEDGEMENTS We thank the whole CarBBAS team for their contri- bution to the field and laboratory components of this re- Tab. 2. ANOSIM R values indicating significant differences in Biolog metabolic profiles between different categories. All R values for Euclidean distance matrices were calculated using 9999 permutations. Significance codes for R values Community group refers to the 4 groups of communities where either ‘Seed’ or ‘Tourist’ OTUs dominated (see arrows in Fig. 4). ANOSIM R values By season ns By region ns By catchment type 038** By river order 0.19* By community group 0.16* ns, not significant; **P<0.0001; *P<0.05. No n- co mm er cia l u se on ly 233C. Ruiz-González et al. search. In particular, we thank J.F. Lapierre for providing the PARAFAC data. This study is part of the program of the Carbon Biogeochemistry in Boreal Aquatic Systems (CarBBAS) Industrial Research Chair, co-funded by the Natural Science and Engineering Research Council of Canada (NSERC) and Hydro-Quebec. We acknowledge funding from FRQNT, NSERC, and the Canada Research Chairs program. CRG benefited of a Juan de la Cierva post-doctoral fellowship from the Spanish Ministry of In- dustry, Economy and Competitivity. REFERENCES Barberán A, Ramirez KS, Left JW, Bradford MA, Wall DH, Fierer N, 2014. Why are some microbes more ubiquitous than others? Predicting the habitat breadth of soil bacteria. Ecol. Lett. 17:794-802. Berggren M, del Giorgio PA, 2015. Distinct patterns of micro- bial metabolism associated to riverine dissolved organic car- bon of different source and quality. J. Geophys. Res. Biogeo.120:989-999. Besemer K, Singer G, Quince C, Bertuzzo E, Sloan W, Battin TJ 2013. Headwaters are critical reservoirs of microbial di- versity for fluvial networks. P. Roy. Soc. B. 280:20131760. Besemer K, Peter H, Logue JB, Langenheder S, Lindstrom ES, Tranvik LJ, Battin TJ, 2012. Unraveling assembly of stream biofilm communities. ISME J. 6:1459-1468. Blagodatskaya E, Kuzyakov Y, 2013. Active microorganisms in soil: Critical review of estimation criteria and approaches. Soil Biol. Biochem. 67:192-211. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Huntley J, Fierer N, et al. 2012. Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq plat- forms. ISME J. 6:1621-1624. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. 2010. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods 7:335-336. Carini P, Marsden PJ, Leff JW, Morgan EE, Strickland MS, Fierer N, 2016. Relic DNA is abundant in soil and obscures esti- mates of soil microbial diversity. Nat. Microbiol. 2:16242. Clarke KR, 1993. Non-parametric multivariate analysis of changes in community structure. Australian J. Ecol. 18:117-143. Comte J, Lindström ES, Eiler A, Langenheder S, 2014. Can ma- rine bacteria be recruited from freshwater sources and the air? ISME J. 8:2423-2430. Crump BC, Amaral-Zettler LA, Kling GW, 2012. Microbial di- versity in arctic freshwaters is structured by inoculation of microbes from soils. ISME J. 6:1629-1639. Crump BC, Adams HE, Hobbie JE, Kling GW, 2007. Biogeog- raphy of bacterioplankton in lakes and streams of an Arctic tundra catchment. Ecology 88:1365-1378. Edgar RC, 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26:2460-2461. Garland JL, Mills AL, Young JS, 2001. Relative effectiveness of kinetic analysis vs single point readings for classifying environmental samples based on community-level physio- logical profiles CLPP. Soil Biol. Biochem. 33:1059-1066. Hiller KA, Foreman KH, Weisman D, Bowen JL, 2015. Perme- able reactive barriers designed to mitigate eutrophication alter bacterial community composition and aquifer redox conditions. Appl. Environ. Microbiol. 81:7114-7124. Janssen PH, Yates PS, Grinton BE, Taylor PM, Sait M, 2002. Improved culturability of soil bacteria and isolation in pure culture of novel members of the divisions Acidobacteria, Actinobacteria, Proteobacteria, and Verrucomicrobia. Appl. Environ. Microbiol. 68:2391-2396. Kothawala DN, Stedmon CA, Müller RA, Weyhenmeyer GA, Köhler SJ, Tranvik L, 2014. Controls of dissolved organic matter quality:evidence from a large-scale boreal lake sur- vey. Glob. Change Biol. 20:1101-1114. Lapierre JF, del Giorgio PA, 2014. Partial coupling and differ- ential regulation of biologically and photochemically labile dissolved organic carbon across boreal aquatic networks. Biogeosciences 11:5969-5985. Lauber CL, Hamady M, Knight R, Fierer N, 2009. Pyrosequenc- ing-based assessment of soil pH as a predictor of soil bacte- rial community structure at the continental scale. Appl. Environ. Microbiol. 75:5111-5120. Lear G, Washington V, Neale M, Case B, Buckley H, Lewis G, 2013. The biogeography of stream bacteria. Global Ecol. Biogeogr. 22:544-554. Lee JE, Buckley HL, Etienne RS, Lear G, 2013. Both species sorting and neutral processes drive assembly of bacterial communities in aquatic microcosms. FEMS Microbiol. Ecol. 86:288-302. Lennon JT, Jones SE, 2011. Microbial seed banks:the ecological and evolutionary implications of dormancy. Nat. Rev. Mi- crobiol. 119:119-130. Lindström ES, Bergström AK, 2004. Influence of inlet bacteria on bacterioplankton assemblage composition in lakes of different hydraulic retention time. Limnol. Oceanogr. 49:125-136. Lindström ES, Forslund M, Algesten G, Bergström A, 2006. Ex- ternal control of bacterial community structure in lakes. Limnol. Oceanogr. 51:339-342. Logares R, Lindström ES, Langenheder S, Logue JB, Paterson H, Laybourn-Parry J, et al. 2013. Biogeography of bacterial communities exposed to progressive long-term environmen- tal change. ISME J. 7:937-948. Magoc T, Salzberg S, 2011. FLASH: Fast length adjustment of short reads to improve genome assemblies. Bioinformatics 27:2957-2963. Nelson CE, Sadro S, Melack JM, 2009. Contrasting the influ- ences of stream inputs and landscape position on bacterio- plankton community structure and dissolved organic matter composition in high-elevation lake chains. Limnol. Oceanogr. 54:1292-1305. Newton RJ, Jones SE, Eiler A, McMahon KD, Bertilsson S, 2011. A guide to the natural history of freshwater lake bac- teria. Microbiol. Mol. Biol. Res. 75:14-49. Niño-García JP, Ruiz-González C, del Giorgio PA, 2016a. In- teractions between hydrology and water chemistry shape bacterioplankton biogeography across boreal freshwater net- works. ISME J. 10:1755-1766. Niño-García JP, Ruiz-González C, del Giorgio PA, 2016b. Land- scape-scale spatial abundance distributions discriminate core from random components of boreal lake bacterioplankton. Ecol. Lett. 19:1506-1515. No n- co mm er cia l u se on ly 234 Dispersed bacteria in a river continuum Pruesse E, Quast C, Knittel K, Fuchs BM, Ludwig W, Peplies J, et al. 2007. SILVA:a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res. 35:7188- 7196. Ruiz-González C, Niño-García JP, del Giorgio PA, 2015a. Ter- restrial origin of bacterial communities in complex boreal freshwater networks. Ecol. Lett. 18:1198-1206. Ruiz-González C, Niño-García JP, Lapierre JF, del Giorgio PA, 2015b. The quality of organic matter shapes the functional biogeography of bacterioplankton across boreal freshwater ecosystems. Global Ecol. Biogeogr. 11:11-12. Ruiz-González C, Niño-García JP, Kembel SW, del Giorgio PA, 2017. Identifying the core seed bank of a complex boreal bacterial metacommunity. ISME J. 11:2012-2021. Savio D, Sinclair L, Ijaz UZ, et al., 2015. Bacterial diversity along a 2600 km river continuum. Environ. Microbiol. 17:4994-5007. Sjöstedt J, Koch-Schmidt P, Pontarp M, Canbäck B, Tunlid A, Lundberg P, et al. 2012. Recruitment of members from the rare biosphere of marine bacterioplankton communities after an environmental disturbance. Appl. Environ. Microbiol. 78:1361-1369. Smalla K, Watchtendorf U, Heuer H, Liu W, Forney L, 1998. Analysis of BIOLOG GN substrate utilization patterns by mi- crobial communities. Appl. Environ. Microbiol. 64:1220-1225. Souffreau C, Van der Gucht K, van Gremberghe I, Kosten S, Lacerot G, Lobão LM, et al., 2015. Environmental rather than spatial factors structure bacterioplankton communities in shallow lakes along a >6000 km latitudinal gradient in South America. Environ. Microbiol. 17:2336-2351. Stubbins A, Lapierre JF, Berggren M, Prairie YT, Dittmar T, del Giorgio PA, 2014. What’s in an EEM? Molecular signatures associated with dissolved organic fluorescence in boreal Canada. Envir. Sci. Tech. 48:10598-10606. Wang Q, Garrity GM, Tiedje JM, Cole JR, 2007. Naïve bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl. Environ. Microbiol. 73:5261-5267. Wilhelm L, Besemer K, Fragner L, Peter H, Weckwerth W, Bat- tin TJ, 2015. Altitudinal patterns of diversity and functional traits of metabolically active microorganisms in stream biofilms. ISME J. 9:2454-2464. No n- co mm er cia l u se on ly