Although microbial marine communities are the essential players of the marine ecosystem functioning, being the main drivers of the major marine biogeochemical cycles (Azam et al., 1983; Karl, 2002; Pomeroy et al., 2007), little is known about the ecological and oceanographic drivers that shape their distributions. Some known factors that affect the microbial community structure include geographical position from tropical to polar regions (Fuhrman et al., 2008; Brown et al., 2012; Ghiglione et al., 2012) and the physical-chemical parameters of the marine environment such as temperature, salinity, depth and chlorophyll-a (Chl-a) concentration (Crump et al., 2004; DeLong, 2006; Falcón et al., 2008; Campbell and Kirchman, 2013; Fortunato et al., 2013; Sunagawa et al., 2015; Herlemann et al., 2016). Furthermore, oceanographic features such as water masses, oceanic fronts, and advection have also been described as physical barriers that can limit the dispersal of microorganisms and, at the same time, influence the structure of the communities at the ocean scale (Galand et al., 2010; Agogué et al., 2011; Wilkins et al., 2013a).

The Southern Ocean (SO) plays an important role in global ocean circulation, biogeochemical cycles, and climate (Broyer and Koubbi, 2014). The surface of the SO is composed of several distinct zones defined by transitions in surface water temperatures and density (Orsi et al., 1995; Sokolov and Rintoul, 2002). The Antarctic Circumpolar Current (ACC) divides the SO into three major zones: the Subantarctic Zone (SAZ), the Polar Frontal Zone (PFZ), and the Antarctic Zone (AAZ) (Pollard et al., 2002). These three major zones are separated from each other from north to south by the Subtropical Front (STF), the Subantarctic Front (SAF), the Polar Front (PF), and the Southern Fronts (SF) (Whitworth and Nowlin, 1987; Orsi et al., 1995; Sokolov and Rintoul, 2002). The Southern Fronts (SF) belonging to the ACC are divided into (i) the southern ACC front (sACCf) and (ii) the Southern Boundary of the ACC front (sbACC) (Sokolov and Rintoul, 2002).

Oceanic fronts are regions where environmentally distinct water masses meet, creating sharp physicochemical gradients over fine spatial scales (Belkin et al., 2009). Therefore, each zone and fronts possess characteristic temperature, salinity, and productivity levels (Pollard et al., 2002; Sokolov and Rintoul, 2002) and tend to have distinctive biological communities (Sokolov and Rintoul, 2002). However, as a consequence of climate change, waters toward the pole side of the ACC waters have become warmer and more saline, while those to the north are cooler and fresher (Ning et al., 2008; Haumann et al., 2016).

The West Antarctic Peninsula (WAP) and the South Shetland Islands in the AAZ are characterized by varied and complex coastal ecosystems and oceanographic processes (Prezelin et al., 2004; Schofield et al., 2010). This coastal ecosystem is subjected to strong seasonal and interannual variability driven by global atmospheric and oceanic changes (Morley et al., 2020). The increasing atmospheric and ocean temperatures registered along the Antarctic Peninsula (AP) have contributed to the collapse of numerous ice shelves and the retreat of many glaciers (Kim and Ducklow, 2016; Morley et al., 2020). Some of the consequences of ice melting and the resulting freshwater input for the AP coastal ecosystems include salinity decrease and stratification of the water column, limited light penetration, increased iron availability, and opening of new areas for primary production (Schofield et al., 2010; Höfer et al., 2019; Hopwood et al., 2020; Morley et al., 2020). More specifically, along King George Island, many semi-enclosed embayments such as Maxwell Bay (MB) and its tributary Marian Cove (MC) are exposed to freshwater input from glaciers surrounding the bay and oceanic water exchange with the Bransfield Strait (BS) (Llanillo et al., 2019). The evidence indicates that this area is experiencing accelerated warming due to the retreat of glaciers such as the glacier associated with MC, which has retreated 1.7 km from 1956 to 2012 (Ahn et al., 2016). As a consequence of the retreat of the MC glacier, the structure and function of the megabenthic epifauna have been affected and benthic diatom blooms have been recorded as unusual in this area (Moon et al., 2015; Ahn et al., 2016).

Several studies of diversity and community structure on bacterioplankton at regional scale have been carried out in the SO (Piquet et al., 2011; Signori et al., 2014; Hernández et al., 2015; Luria et al., 2016; Dinasquet et al., 2017; Picazo et al., 2019; Alcamán-Arias et al., 2021). However, only a few studies have reported results from a large-scale investigation (Wilkins et al., 2013a,b; Luria et al., 2016), and none has targeted the active fraction of the microbial communities. The seasonal dynamics of some heterotrophic prokaryotic communities have, for instance, been described in association with phytoplanktonic blooms in nutrient-enriched areas of the SO (Obernosterer et al., 2011; Kim and Ducklow, 2016; Luria et al., 2016; Liu et al., 2019, 2020). In open oceanic areas, physical parameters such as temperature, depth, ice melting, advection, oceanic fronts, and water circulation have been described as the major drivers shaping prokaryotic community structure (Piquet et al., 2011; Wilkins et al., 2013a,b; Alcamán-Arias et al., 2021). In some coastal areas of the Antarctic region such as bays and fjords, the combination of environmental parameters, including salinity, nutrients, and dissolved/particulate organic/inorganic matter, are described as major structuring factors (Moreno-Pino et al., 2016; Picazo et al., 2019; Kim et al., 2020). While the composition of the bacterioplankton community in the SO shows some variability across the different areas, members of the classes Gammaproteobacteria, Alphaproteobacteria, and Bacteroidetes tend to dominate the number of sequences (Abell and Bowman, 2005; Ghiglione et al., 2012; Wilkins et al., 2013b; Baltar et al., 2016; Liu et al., 2019; Logares et al., 2020). Among Alphaproteobacteria, SAR11 often dominates in both coastal and open ocean areas (López-García et al., 2001; Murray and Grzymski, 2007; Giebel et al., 2009; Piquet et al., 2011; Ghiglione et al., 2012). The clade is more abundant in Subantarctic and PF areas where high nitrogen and low chlorophyll conditions prevail, giving it a competitive advantage compared to the Antarctic zone where phytoplankton blooms result in increased concentrations of high molecular weight dissolved organic matter that has been found to decrease their abundance (Giebel et al., 2009; Ghiglione and Murray, 2012). For Gammaproteobacteria, the order Alteromonadales includes heterotrophs with broad substrate preferences (Bowman et al., 1997). Within this order, Cowellia genera is a common member of polar systems with adaptation to cold environments through the production of cold active extracellular enzymes (Methe et al., 2005). Among Bacteroidetes, members of the class Flavobacteria are the major components of the planktonic communities in the SO, especially prevalent in particle attached communities and in association with phytoplankton blooms (Abell and Bowman, 2005). This class has a well-described ability to degrade organic matter, suggesting an important role in the remineralization of primary production products (Kirchman, 2002). The predicted climate changes in different areas that comprise the SO, such as water temperature increase, ocean acidification, changes in salinity, and water stoichiometry among others, are expected to influence the composition of the microbial communities in terms of abundance and distribution (Pearce, 2008), demanding a better understanding of the key patterns of microbial ecology and biogeography as well as the specific ecosystem of the SO.

The objective of this study was to test how geographical distance and environmental parameters shape prokaryotic community composition in the SO. We explored the distribution, abundance, and diversity of prokaryotic communities across different oceanographic areas in the surface layer of the SO. We then analyzed both the total and active fraction of the communities, based on the 16S rRNA gene and 16S rRNA-based, using samples collected over a spatial transect spanning over 6,500 km of the SO and the WAP continental shelf, terminating in a coastal fjord system on King George Island. Our study traverses large regions of the ocean but also includes small spatial gradients (fjords), which should be able to identify the impact of both physical (e.g., geographical distance, oceanic fronts) and environmental barriers in the dispersal and structuring of bacterioplankton communities in the SO.

Our study reports the biogeography of marine microbes in a vast and sparsely studied area of the SO. The patterns of community composition show that both the active fraction and the standing stock of bacterioplankton assemblages are structured by both key environmental parameters (temperature, salinity, POC, PON, and silicic acid) (27% of the variance) and by geographical distance (10% of the variance). The estimated amount of variance in our study was similar to the values reported in a review by Hanson et al. (2012). They point out that the average variance explained by environmental effect is 27% compared to 10% for a distance effect, highlighting that the contemporary selection may turn out to have a greater effect on microbial composition than historical processes on the biogeography of microorganisms. Among the factors that could have explained microbial community composition but that we could not measure in this study, the selection, drift, dispersal, and diversification (mutation) are the fundamental ecological processes (Hanson et al., 2012; Nemergut et al., 2016). Dispersal and drift have been identified as the important ecological processes in biogeographical studies conducted in marine environments (Lindström and Langenheder, 2012; Wilkins et al., 2013b).

Spatial distance has been proposed as a barrier for dispersal in marine microorganisms (Cho and Tiedje, 2000; Ramette and Tiedje, 2007). In most of the combinations tested in this study, dissimilarity increased as the geographical distance between communities increased (Figure 5). This strong distance–decay relationship can be associated with decreasing similarity in environmental parameters and hydrography with the increasing geographical distance. However, the amount of variance explained by distance vs. environmental factors will depend on the sampling scale (Martiny et al., 2006). A large spatial geographical distance was covered in our study (~6,500 km) for which distance would be expected to dominate over the environmental variation. However, the environmental characteristics of each sampled zone were stronger in driving the variability of the bacterioplankton communities, pointing out that the observed biogeographical patterns are a combination of both distance and environmental factors.

Strong distance–decay patterns were also encountered in communities that were not geographically distant. This was the case for SA vs. PF zones, AS vs. WAP zones, and BS vs. AS zones, all of them separated by <250 km of distance. Several studies indicate that oceanographic properties of water masses and fronts play a significant role in the structure of bacterioplankton communities (Galand et al., 2010; Wilkins et al., 2013a; Baltar et al., 2016; Djurhuus et al., 2017). The strong shift in community composition observed here within short distances could be due to the presence of different water masses. In particular, some samples obtained within the SA zone were collected in the southern branch of SAF and PF; meanwhile, the majority of the samples from the PF zone were obtained in the sACCf (Supplementary Figure 1). Likewise, the samples collected for the AS zone did not cross any front, unlike those collected in the WAP zone, which mostly established themselves in the sbACC (Supplementary Figure 1). In the case of BS, the samples obtained were located within the sACCf. This result demonstrates that SO fronts can act as significant biogeographical boundaries, generating significant changes in the community composition of spatially near bacterioplankton communities.

The bacterioplankton communities were heterogeneously distributed along the SO gradient being structured by the environmental parameters of each zone (Figure 8). According to the environmental parameters, the bacterioplankton communities of the seven zones were divided into two major groups: (i) oceanic areas of the SO transect composed of SA, PF, AS, and WAP zones and (ii) coastal areas of the SO composed of BS, MB, and MC.

The temperature was the main environmental factor that correlated with community composition among the zones possibly restricting the abundance of bacteria such as Alteromonadales and SAR11 (Supplementary Table 4) while promoting the appearance of Cyanobacteria (Figure 8). In agreement with these findings, it has been shown that the distribution of SAR11 (a globally distributed bacteria) can be significantly affected by temperature changes (Brown et al., 2012). The low temperatures recorded in BS, MB, and MC could act as a growth factor for SAR11 for which high abundances are found at temperatures above 0°C (Brown et al., 2012). Conversely, Alteromonadales are cold-adapted heterotrophs that are normally found in cold waters of the SO related to ice melt (Piquet et al., 2011). This may explain their high abundance as indicator taxa in the last part of the transect, a zone associated with ice formation and with the lowest temperatures registered (Figure 8). Furthermore, one of the species that contributed most to the differentiation of the SA community from the rest of the zones was Cyanobacteria, which dominated in this zone despite being a minor taxonomic component (0.57%) of the entire data set (Figure 8). The first sample of the SA, where the number of sequences of Cyanobacteria was the most abundant, was collected in waters belonging to the southern branch of the SAF (−52.1400°S) (Figure 8). High abundances of Cyanobacteria have previously been reported further north of the STF where subtropical waters exhibit higher temperature and salinity (Wilkins et al., 2013a; Liu et al., 2019). Additionally, water surface temperature has been reported as a determining factor in the distribution of Cyanobacteria in the SO (Marchant et al., 1987). The water temperature at which the highest abundance of Cyanobacteria was found corresponds to 12°C, which represents the highest water temperature recorded in the entire transect, suggesting that the presence of Cyanobacteria is related to the water temperature in accordance with the characteristics described for the SAF.

The bacterioplankton communities of the PF, AS, and AP zones were mainly structured by variations in the concentrations of silicic acid, POC, and PON (Figure 8) and second, Chl-a, DON, and ammonium concentrations. The most important nutrients for phytoplankton abundances are nitrogen and phosphorus and, in particular, silicate for diatoms (Sidabutar, xbib2016). However, ammonium has been described to be an important nutrient able to alleviate the photoinhibition or the light-limitation at which phytoplankton is exposed in Antarctic waters (Agustí et al., 2009). Elevated ammonium concentrations (~3.3 μm) were recorded for the AP zone compared to the rest of the transect. Ammonium concentrations are seasonally variable in the WAP with peak concentrations (>3.5 μm) recorded in early autumn (March–May) (Henley et al., 2017). Previous studies have established successional patterns between phytoplankton blooms and some specific bacterioplankton members (Abell and Bowman, 2005; Obernosterer et al., 2011; Kim et al., 2014; Choi, 2016; Luria et al., 2016; Liu et al., 2019). In the PF, AS, and AP zones, the generalist Oceanospirillales, Rhodobacterales, and Flavobacteriales dominated the indicator taxa. Flavobacteria, Rhodobacterales, and Oceanospirillales have been positively correlated with the abundance of large and small diatoms in the SO (Kim et al., 2014; Luria et al., 2016; Liu et al., 2019). Considering the environmental parameters and the indicator OTUs selected, it is possible to indicate that the composition of the bacterioplankton community in the AS and AP are tightly related and structured by phytoplankton blooms.

In BS, MB, and MC, the community structure of bacterioplankton was mainly driven by salinity, POC, and silicic acid (Figure 8, Supplementary Figure 4, and Supplementary Table 4). The BS is characterized by a mixture of different bodies of water coming from different sectors of the WAP. The surface layer during winter is characterized by temperatures lower than 1°C and salinity of 34.0 which is defined by Antarctic Surface Water (Hofmann et al., 1996). Meanwhile, MB and MC fjord present a typical estuarine circulation with superficial colder and less saline water (Llanillo et al., 2019). Furthermore, MC registered high concentrations of silicic acid and POC probably associated with meltwater streams with terrigenous particles that persisted until late March (Yoo et al., 2015). The bacterioplankton community composition observed for BS, MB, and MC zones were similar to the composition reported by other studies in the same area, relating the differences in the composition to the environmental gradients rather than spatial distance (Zheng et al., 2011; Moreno-Pino et al., 2016; Kim et al., 2020). In particular, Campylobacterales (Epsilonproteobacteria) and Puniceicoccales (Verrucomicrobiales) were selected as the indicator taxa of MC and MB, respectively. Campylobacterales was detected in MC from head to mouth and from 0 to 25 m unlike that reported by Kim et al. (2020), who report a higher abundance of this order in areas near the mouth of the fjord influenced by water from MB. Furthermore, Verrucomicrobiales and Campylobacterales have been described as the common members of particle-associated communities having the ability to degrade particulate organic carbon sources ((Crespo et al., 2013; Fontanez et al., 2015; Duret et al., 2019)). Differences in the hydrographical properties of the water column, such as melt water input or mixed wind can lead to variation in the microbial community composition of adjacent zones (Zheng et al., 2011; Moreno-Pino et al., 2016). Consequently, differences in the morphological characteristic, sediment load, and meltwater discharge could be responsible for the observed differences in community composition between the zones BS, MB, and MC.

Finally, our results indicate the differences between the total (16S rRNA gene) and active (16S rRNA-based) microbial community composition among the complete data set. The 16S rRNA-based fraction appeared to be more resolutive for describing and explaining the bacterioplankton assemblages of each community compared to the 16S rRNA gene. In addition, in our study, higher α-diversity values were observed for the active community compared to the total community (Figure 4). Furthermore, a higher degree of dissimilarity between the two fractions was observed across the transect (Supplementary Figure 3). The high compositional dissimilarity between the 16S rRNA gene and 16S rRNA-based community may be due to differential changes in the relative abundance of certain OTUs. Traditionally, ribosomal RNA genes are used to identify microorganisms present in a specific environment regardless of their metabolic state, whereas ribosomal RNA has been consistently used to characterize the growth state or activity of microorganisms in cultured and mixed microbial communities (Campbell et al., 2009; Gaidos et al., 2011; Campbell and Kirchman, 2013; Hunt et al., 2013). This is because a positive relationship between the response of bacteria to resource availability and the number of rRNA operons has been shown (Klappenbach et al., 2000), and ribosomal content has been found positively correlated with the growth rate for many bacterial groups (DeLong et al., 1989; Schäfer et al., 2001; Troussellier et al., 2002). The number of 16S rRNA gene operons has been described as the variable in bacteria reflecting different ecological strategies of growth and activity (Moeseneder et al., 2005; Rastogi et al., 2009). However, the number of ribosomes can be even more abundant than rRNA operons depending on the growth rate of the bacteria (Kerkhof and Ward, 1993). Therefore, it could be assumed that in oligotrophic marine environments, such as SO, where bacteria tend to show a slow growth rate and a low number of rRNA operons, some marine bacteria might not be detected at the rRNA gene level but could be still detected at the rRNA level (Fegatella et al., 1998; Moeseneder et al., 2005). This could explain the higher α-diversity observed in the active fraction compared to the total fraction. Furthermore, it has been described that in heterogeneous environments, such as estuaries, environmental variables such as salinity and Chl-a may affect the differential growth rate but not the abundance of individual taxa (Campbell and Kirchman, 2013). Accordingly, we can expect that the bacterioplankton communities in heterogenous gradients will show higher differences between the active and total community in comparison with bacterioplankton communities living in more stable environmental gradients.

In conclusion, our results show that there were differences in the bacterioplankton community composition across the SO transect from New Zealand to WAP. The differences in taxonomic composition among the seven zones analyzed were strongly correlated with environmental parameters and with geographical distances. The bacterioplankton community turnover was mainly mediated by temperature, silicic acid, PON, salinity, and POC. However, ocean fronts also shaped microbial community composition by increasing physicochemical and biological dissimilarities within restricted geographical areas and thus promoting distinct communities within short distances. Consequently, the biogeographical patterns observed in our study originate from a combination of environmental gradients and oceanographical discontinuities produced by oceanic fronts. In summary, the main contribution of this study is to unveil the complex spatial patterns of the SO marine bacterioplankton communities, at small and large geographical scales (from fjord to ocean), based on the RNA analysis. RNA appeared as a useful tool to explore bacterioplankton community composition across contrasted marine areas. It provided a precise description of complex microbial community dynamics in a changing environment.

The original contributions presented in the study are included in the article/Supplementary Material. The genomic data presented in the study are available at the European Nucleotide Archive database (ENA) under project identification number PRJEB45020 and sample accession numbers ERS6476988 and ERS6442804.

CM-M, PG, and HG wrote the original draft. JI and CM-M collected the microbial diversity samples. JI contributed to review the manuscript. CM-M prepared and processed the 16s rRNA sequencing data. CM-M and PG performed the statistical analysis. S-YH, BL, and I-YA collected and processed the physicochemical water parameters. MC collected the chlorophyll samples and reviewed the manuscript. MV processed the chlorophyl samples and contributed to review the manuscript. CF provided the resources for sampling collection and contributed to review the manuscript. HG and I-YA provided funding and resources for field campaign. All authors have read and agreed to the published version of the manuscript.

This research was conducted as a part of the project Carbon cycle change and ecosystem response under the Southern Ocean warming (PE22110) and supported by the Korea Polar Research Institute (KOPRI), Incheon. JI and CM-M were partially supported by the National Agency for Research and Development (ANID) through project FONDAPIDEAL (Grant No. 15150003). CM-M was also financed through Ph.D. ANID Scholarship Doctorado Nacional 2016 (No. 21160201). CF was financed by the National Agency for Research and Development (ANID) through Basal Funding COPAS (FB2100021). MV and MC were funded by the US National Science Foundation–Office of Polar Programs (Award No. 1822289). The data presented are part of the Ph.D. Thesis of CM-M.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.