the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 26 Mar 2024
Correlations among carbohydrate inventories, enzyme activities, and microbial communities in the western North Atlantic Ocean
Abstract. Heterotrophic bacteria process nearly half of the organic matter produced by phytoplankton in the surface ocean. Much of this organic matter consists of high molecular weight (HMW) biopolymers such as polysaccharides and proteins, which must initially be hydrolyzed to smaller sizes by structurally specific extracellular enzymes. To assess the relationships between substrate structure and microbial community composition and function, we concurrently determined carbohydrate abundance and structural complexity, bacterial community composition, and peptidase and polysaccharide hydrolase activities throughout the water column at four distinct stations in the western North Atlantic Ocean. Although the monosaccharide constituents of particulate organic matter (POM) were similar among stations, the structural complexity of POM-derived polysaccharides varied by depth and station, as demonstrated by polysaccharide-specific antibody probing. Bacterial community composition and polysaccharide hydrolase activities also varied by depth and station, suggesting that the structure and function of bacterial communities—and the structural complexity of their target substrates—are interlinked. Thus, the extent to which bacteria can transform organic matter in the ocean is dependent on both the structural complexity of the organic matter and their enzymatic capabilities in different depths and regions of the ocean.
- Preprint
(2380 KB) - Metadata XML
-
Supplement
(2000 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-615', Anonymous Referee #1, 29 Apr 2024
This manuscript, which is led by Chad Lloyd, of the Arnosti lab, presents a compelling dataset of mono- and polysaccharide concentrations and composition, bacterial abundances, production and community composition, and peptidase and polysaccharide hydrolase enzyme activity as profiles collected from 1 coastal shelf and 3 offshore stations in the western North Atlantic in the spring of 2019. The value of this contribution combines a new analytical approach, which provides a more comprehensive picture of the polysaccharide composition, with the hydrolases responsible for the breakdown of polysaccharides. Because environmental variability was demonstrated among the samples collected, authors hypothesize that the observations likely represent influences by varying environmental conditions on polysaccharides and relevant enzymatic activity.
General comments:
Double-check all references are in the reference list (the 2 Lloyd papers referenced in the text are not in the reference list)
One difficulty in evaluating the data and overall story presented here was due to the arrangement of the data in the figures presented. For instance, Figure 2a uses a stacked bar of relative abundances. In contrast, Figure 3 uses a light blue gradient of relative abundances as a heatmap. But these two figures could very easily use the same formatting to make a quicker comparison between mono- and polysaccharides. Additionally, some subplots don’t show data, but it’s hard to know if that is because the values were below the method’s detection limit or because no samples were collected.
I have strong reservations about the 16S data presented here for several reasons, as discussed in further detail below. (1) The volume listed for the extraction is 25mL, which is too small for use with standard kits. (2) Even if this is a typo-, the data presented don’t make sense with the lack of any archaea present in any samples. (3) Even if there is a logical reason to not show the archaea abundances, there isn’t any real connection between the POM composition/hydrolysis data presented here. It’s only mentioned that there are changes over depth, which is already commonly known. (4) The size-fraction collected (0.22 μm filters), are more likely representing a different community than those associated with the particles isolated here (see: https://doi.org/10.1038/ismej.2016.166 and https://doi.org/10.3389/fmicb.2020.01590). Bacteria are likely attached onto suspended particles, and these communities are likely significantly different from the free-living community. However, even if the suspended POM-associated community is captured in the water collected by a 0.2um filter, the free-living is so much more dominant than this particle-associated community that you won’t see the particle-attached taxa. Therefore, it’s my recommendation that this data be excluded from the manuscript because of the methods concern, poor integration of the data into the discussion, and lack of useful connection with the main message for this manuscript. The community data would only be useful to keep (if firstly all of the above concerns 1-4 were resolved) if you mention specific types of taxa that were associated with unique polysaccharide composition and/or enzymatic activity patterns. For your data to be relevant, you will also need to somehow show that certain taxa unique to certain composition/enzymes also have unique metabolic repertoire compared with other taxa (e.g., you can use BLAST and/or NCBI’s IMG databases to compare with your sequences).
Line by Line comments:
Line 73: Please mention that you collected water from “7 discrete depth horizons” to be more explicit. Define what depth(s) “surface” refers to. Usually, this is ~5m.
Line 86: It looks like you used leucine throughout but you mention here that you used bacterial production in carbon units. If you do use BP in carbon units, “A factor of 0.86” is not clear for how the leucine uptake rates were converted to carbon, in which case, please provide more explicit details about the conversion of pmol Leu hr-1 to μmol C L-1 per time.
Line 97: “pre-combusted” for how long at what temperature?
Line 104: Please include the detection limits for POC and PON
Line 105: You have presented quantitative measurements of monosaccharides here, but in Figure 2, you don’t show error bars. Either include the average measurement error in the methods or show the error in Figure 2. Also include the detection limit for this method.
Line 115: In the caption for Figure 3, you mentioned that “Note: the carbohydrate microarrays are only semiquantitative; while comparisons can be made between the individual stations and depths of a given epitope, they cannot be used to compare between different epitopes.” This detail seems more appropriate for the methods section.
Lines 130-135: 25mL is likely insufficient to capture representative DNA from seawater using amplicon sequencing. Usually, at least 1L is required when using kits to extract DNA from seawater (e.g., see Pascoal et al., 2023; https://doi.org/10.1038/s43705-023-00278-w). While a microvolume DNA extraction method has been presented for seawater (https://doi.org/10.1038/s43705-021-00079-z), this method uses a specific lysis protocol that differs from the kits, which does not appear to be followed in this manuscript. Also, samples are stored at -40 to -80 deg C in a stabilizing medium like sucrose lysis buffer. Please provide evidence that your methods (smaller volumes with kit and storage at -20 deg without buffer) truly capture a representative community compared with more common protocols.
Line 156: Could you specify here whether this was done at sea? I was curious if you brought autoclaved seawater to sea, in which case I assume you’re using coastal autoclaved seawater that you bring out to sea. If you’re doing it in the shore-based lab, could you provide details about how the seawater samples are preserved to bring them back to the lab? From the Hoarfrost and Arnosti (2017) reference, I assume that you’re doing this at sea, but I don’t understand why you need to use autoclaved seawater. Why can’t this method apply something like TCA for your “killed” control?
Line 172: Were these measurements also started at sea and then brought back to the shore-based lab for long-term incubations? If so, please state that samples were transferred from the ship to the shore-based lab within “x” amount of time for sample transport, during which time samples were at “x” temperatures.
Line 221: The first dataset presented in the text, after the physical context, is the POC concentrations. Please consider either showing these profiles in the main text or, instead, focusing first on the bacterial abundance/production, which you have shown in Fig. 1, along with the physical data. If you’d like to leave the bacterial data in this figure, you could consider making the TS plot smaller and including the POC concentration data. While I appreciate the authors' effort to separate the stations based on distinct physical characteristics; however, it appears that the TS plot adds little to your story; please consider moving this to the supplemental or appendix. Also, while the satellite data look informative, these plots are not currently mentioned in the results section, leaving little incentive for the reader to look carefully at the images. Please include text for this satellite data or remove the images from Fig. 1.
Line 224: Careful repeating the same data in figures and tables as for Bact. Cell Abundances in Fig. 1 and Table 2. Better to choose which form you like better and remove from the other. Also, at this line you write PON with reference to Table 2, but PON does not appear in Table 2.
Line 238: Since Chl-a concentrations provide context for presenting the results, you should include these plots in the main text. Consider including these in Figure 1.
Line 240: “Somewhat different monosaccharide constituents” is not clear. Revise to something more specific.
Line 248-260; Figure 3: The shading as relative abundance is not distinctive enough to tell when one station is truly different from another. Most boxes look about the same shade of blue. Consider converting this into a stacked bar plot where the y-axis is depth, and the x-axis is relative abundance. You have 8 different polysaccharide epitopes, where you could make each epitope a different color. Fig. 3 also presents 3 different extractions (H2O, EDTA, and NaOH), but this is not mentioned in the results, nor are they mentioned in the figure caption. Please highlight the relevance of these 3 different extractions in the results section here.
Line 263: Please consider moving up the description of cell abundances and bacterial productivity if you keep these data in Figure 1.
Line 271: You have jumped to cell-specific production and left out cell production. If you’d rather describe the cell-specific data, just put the bacterial production data in the supplement and the cell-specific data in the main text. Please add a sentence explaining why cell-specific data should be used.
Line 277-278: Concentrations and production rates for bottom waters are not discernable with this axis. If there are no lines on the figure, does that mean the rates/abundances were truly “zero”? Usually, BP and BA are presented as line graphs like you’ve done for Figure 2b. Note in the caption if you don’t have data for certain depths.
Line 280: You’ve presented 16S community for samples collected on a 0.22 μm filter. Very often these represent the “free-living” community, as opposed to the particle-associated, which is collected onto 1.2, 3.0, 5.0 μm filters (or larger). And very often, the bacterial communities below the winter mixed layer significantly differ from the surface communities (ref). It’s odd to see your 3,000m sample grouped with the surface sample in the NMDS plot. Also, could it truly be that you detected no archaea in any of your samples?
Figure 4: You don’t describe what the shaded areas refer to in the figure caption for the NMDS plots.
Lines 299 and 310: The overlapping shaded areas in your NMDS plot for all stations do not support these first sentences. Be more quantitative in how you describe community differences. There are packages in R that can help you to do this, which will generate p-values that you can use to support whether the stations were significantly different from one another.
Line 340: State in the Figure 5 caption whether samples without bars shown are “zero” or no data available
Line 419-420: Please link the specific figure to which you’re referring to each part of the sentence (e.g., polysaccharides comment refers to Fig 3, and monosaccharide comment refers to Fig 2).
Line 420-421: Please make this sentence clearer. I think what you mean is: that there is strong evidence for differences based on depth (please add a reference to support this), but previous monosaccharide evidence suggested little variation by location (Aluwihare reference). Note that the Aluwihare reference only refers to surface waters so you’ll need other reference(s) to support the depth trends.
Line 430: Would it be possible to assign each symbol in Figure 9 to a monosaccharide? If you have some polysaccharides that are of interest to this study, could you make some representation of those in this Figure? If possible, it could help to somehow summarize the main findings about polysaccharide persistence/removal observed here. For instance, do the polysaccharides leading to monosaccharide composition via specific enzymes in the surface differ in a simplistic way from those in the deepest depths?
Line 458: Could you list which end-acting enzymes you’re referring to here? It’s helpful for the reader not to need to flip back as you’ve done for the exo-acting enzymes.
Lines 468-480: You don’t actually need your 16S data for this discussion. This only mentions that there should be a connection between the POM composition/enzymes and the taxa.
Lines 483-513: This section is particularly compelling. Any other examples that you could dive into besides fucoidan and laminarin? I’d recommend, especially if the DNA descriptions are problematic, you expand the comparisons to other kinds of substrates/enzymatic between depths/locations. Additionally, you haven’t brought the cell-specific leu uptake nor cell abundance data back into the discussion, which do appear to provide a useful context to interpreting your sugar/enzymatic activity data.
Line 514: The use of “thus could” seems odd here, especially if someone jumps to read your conclusions without reading all of the prior text. Recommend removing this.
Citation: https://doi.org/10.5194/egusphere-2024-615-RC1 -
AC1: 'Reply on RC1', Chad Lloyd, 29 Jun 2024
We would like to thank the reviewer for the thoughtful comments. We have copied their comments and posted our replies in bold below each comment.
REVIEWER 1
This manuscript, which is led by Chad Lloyd, of the Arnosti lab, presents a compelling dataset of mono- and polysaccharide concentrations and composition, bacterial abundances, production and community composition, and peptidase and polysaccharide hydrolase enzyme activity as profiles collected from 1 coastal shelf and 3 offshore stations in the western North Atlantic in the spring of 2019. The value of this contribution combines a new analytical approach, which provides a more comprehensive picture of the polysaccharide composition, with the hydrolases responsible for the breakdown of polysaccharides. Because environmental variability was demonstrated among the samples collected, authors hypothesize that the observations likely represent influences by varying environmental conditions on polysaccharides and relevant enzymatic activity.
General comments:
Double-check all references are in the reference list (the 2 Lloyd papers referenced in the text are not in the reference list)
We thank the reviewer for catching this point – we will check all the references. These references have been added to the document.
One difficulty in evaluating the data and overall story presented here was due to the arrangement of the data in the figures presented. For instance, Figure 2a uses a stacked bar of relative abundances. In contrast, Figure 3 uses a light blue gradient of relative abundances as a heatmap. But these two figures could very easily use the same formatting to make a quicker comparison between mono- and polysaccharides. Additionally, some subplots don’t show data, but it’s hard to know if that is because the values were below the method’s detection limit or because no samples were collected.
We will add this information to the figures. For the polysaccharides, this data is semi-quantitative, and while we could map out the data in terms of relative abundance, it is not directly comparable to the monosaccharide data, which are quantitative. We will change the color scheme so that the figures are more easily discernible, and will distinguish between ‘no data’ and concentrations that are below the detection limit.
I have strong reservations about the 16S data presented here for several reasons, as discussed in further detail below. (1) The volume listed for the extraction is 25mL, which is too small for use with standard kits.
The DNeasy PowerWater Kit we used is designed for the extraction of low amounts of DNA from aquatic, especially marine, environmental samples. With 25 mL sample volume, we extract DNA from a maximum of 4x108 cells in surface waters of station 19 and 2.25 x105 cells in bottom water of station 20 – a sufficient number of bacteria to cover the full community. (See further replied to the line-by-line comments.) Additionally, as part of the bioinformatics processing, we created rarefaction curves to ensure that we had an adequate sampling of the microbial community; if we had too few reads, we would have removed them from the analysis, but there were not. Please note also that this part of the processing was done in the Molecular Ecology Department at the Max Planck Institute for Marine Microbiology, where we have decades of experience in this methodology. As an example, Reintjes et al., (2017) used only 10 mL in their DNA extractions and obtained sufficient DNA to adequately analyze the bacterial community (doi: 10.1038/ismej.2017.26).
(2) Even if this is a typo-, the data presented don’t make sense with the lack of any archaea present in any samples.
The primer pair (Bakt 314F and Bakt 805R) that was used in this study targets the amplification of bacteria 16S rRNA. Archaeal sequences were consequently not included for the analysis. We will comment specifically on this point in the Methods.
(3) Even if there is a logical reason to not show the archaea abundances, there isn’t any real connection between the POM composition/hydrolysis data presented here. It’s only mentioned that there are changes over depth, which is already commonly known.
As mentioned above, archaea were not sequenced in this study.
As discussed in the manuscript, one of the issues with analyzing marine carbohydrates is that monosaccharide composition alone does not provide key information about the structure of marine polysaccharides: in effect, the monosaccharide composition catalogs the different types of bricks, but provides no information about the building from which they are derived. The carbohydrate epitope data, however, does provide information about the structure of polysaccharides – some of the bricks are arranged in specific patterns. Moreover, especially in the absence of other means of determining polysaccharide structure, we can find clues about the structure of polysaccharides through information about the types of polysaccharide-hydrolyzing enzymes that are active in a given location: these enzymes hydrolyze specific polysaccharide linkages, and are produced by bacteria in order to hydrolyze those linkages. If we find that specific enzymes are active in a given location, it is likely that the corresponding structures are or were present in that location (also allowing for the possibility that the enzymes ‘leftovers’ and have been preserved by association with particles, etc etc etc). Enzymatic activities therefore provide a clue to the structure of organic matter that otherwise cannot currently be determined.
(4) The size-fraction collected (0.22 μm filters), are more likely representing a different community than those associated with the particles isolated here (see: https://doi.org/10.1038/ismej.2016.166 and https://doi.org/10.3389/fmicb.2020.01590). Bacteria are likely attached onto suspended particles, and these communities are likely significantly different from the free-living community. However, even if the suspended POM-associated community is captured in the water collected by a 0.2um filter, the free-living is so much more dominant than this particle-associated community that you won’t see the particle-attached taxa.
Since we used only a 0.2 um filter, our analysis represents the bulk microbial community (both free-living and particle-attached). We agree that the free living community is likely more dominant in this fraction, but given the effort involved to obtain the data presented in this manuscript, it was not feasible to sub-fractionate the water samples to analyze microbial communities in separate fractions. Note in any case that we do not attempt to make statistical links between POM composition and community composition.
Therefore, it’s my recommendation that this data be excluded from the manuscript because of the methods concern, poor integration of the data into the discussion, and lack of useful connection with the main message for this manuscript. The community data would only be useful to keep (if firstly all of the above concerns 1-4 were resolved) if you mention specific types of taxa that were associated with unique polysaccharide composition and/or enzymatic activity patterns. For your data to be relevant, you will also need to somehow show that certain taxa unique to certain composition/enzymes also have unique metabolic repertoire compared with other taxa (e.g., you can use BLAST and/or NCBI’s IMG databases to compare with your sequences).
Reviewer #2 thought that these data are important, and we agree. We therefore respectfully disagree with this reviewer on this point.
We note also that some of the suggestions listed above are beyond the current state of the field. With the exception of ‘selfish’ bacteria (Cuskin et al. 2015; Human gut Bacteroidetes can utilize yeast mannan through a selfish mechanism. Nature 517, 165-173; Reintjes et al. 2017 An alternative polysaccharide uptake mechanism of marine bacteria. The ISME Journal 11, 1640-1650), for example, it is currently not possible to unambiguously link specific bacterial taxa with hydrolysis of specific polysaccharides in seawater samples. (The selfish link is unambiguous because the bacteria are labeled by the fluorescent polysaccharides that they take up.) We agree that genomic/metagenomic/proteomic (etc) investigations can provide very useful information and strong indications of metabolic potential, but genetic potential may not be realized in situ, and many genes, transcripts, and proteins in any case cannot yet be specifically identified. Making progress on these points requires further advances in the field.
Line by Line comments:
Line 73: Please mention that you collected water from “7 discrete depth horizons” to be more explicit. Define what depth(s) “surface” refers to. Usually, this is ~5m.
We will add this information to the Methods.
Line 86: It looks like you used leucine throughout but you mention here that you used bacterial production in carbon units. If you do use BP in carbon units, “A factor of 0.86” is not clear for how the leucine uptake rates were converted to carbon, in which case, please provide more explicit details about the conversion of pmol Leu hr-1 to μmol C L-1 per time.
The conversion factor 0.86 is applied based on Simon and Azam (1989) and Kirchman, (2001), as cited in the Methods. We will re-write this sentence to improve clarity, especially since we presented the data as pmol Leu/hr instead of the converted values.
Line 97: “pre-combusted” for how long at what temperature?
400 C for 6 hrs. This will be added to the Methods.
Line 104: Please include the detection limits for POC and PON
We will add these to the Methods section.
Line 105: You have presented quantitative measurements of monosaccharides here, but in Figure 2, you don’t show error bars. Either include the average measurement error in the methods or show the error in Figure 2. Also include the detection limit for this method.
We will include the detection limits in the methods. Error for the analysis (i.e., monosaccharides) differed between monosaccharides, but this information will be added to the Methods section.
Line 115: In the caption for Figure 3, you mentioned that “Note: the carbohydrate microarrays are only semiquantitative; while comparisons can be made between the individual stations and depths of a given epitope, they cannot be used to compare between different epitopes.” This detail seems more appropriate for the methods section.
We will move this information into the Methods section, as suggested.
Lines 130-135: 25mL is likely insufficient to capture representative DNA from seawater using amplicon sequencing. Usually, at least 1L is required when using kits to extract DNA from seawater (e.g., see Pascoal et al., 2023; https://doi.org/10.1038/s43705-023-00278-w). While a microvolume DNA extraction method has been presented for seawater (https://doi.org/10.1038/s43705-021-00079-z), this method uses a specific lysis protocol that differs from the kits, which does not appear to be followed in this manuscript. Also, samples are stored at -40 to -80 deg C in a stabilizing medium like sucrose lysis buffer. Please provide evidence that your methods (smaller volumes with kit and storage at -20 deg without buffer) truly capture a representative community compared with more common protocols.
We respectfully disagree with the reviewer. Much smaller sample volumes are routinely used in oceanography. For example Reintjes et al., (2017) used 10 mL for extractions. We also note that Pascoal et al. 2023 tested filtered volumes of 1 – 1000 L for DNA extraction from seawater. They state that the volume of seawater filtered does not have a significant effect on prokaryotic and protist diversity, independently of the sequencing strategy. Since sample volumes below 1 L were not tested in their study, we cannot draw any conclusions from it about the effect on prokaryotic diversity of a sample volume of 25 mL. Moreover, the microvolume DNA extraction by Bramucci et al. 2021 is designed for volumes from 1 – 100 µl, a sample volume smaller by a factor of 2.5x103 – 2.5x105 than our sample volumes.
Line 156: Could you specify here whether this was done at sea? I was curious if you brought autoclaved seawater to sea, in which case I assume you’re using coastal autoclaved seawater that you bring out to sea. If you’re doing it in the shore-based lab, could you provide details about how the seawater samples are preserved to bring them back to the lab? From the Hoarfrost and Arnosti (2017) reference, I assume that you’re doing this at sea, but I don’t understand why you need to use autoclaved seawater. Why can’t this method apply something like TCA for your “killed” control?
We brought autoclaves to sea, and autoclaved the water from each station and depth at which we made the measurements. The autoclave approach is practical in that the microbial community is killed, and any enzymes are denatured by the heat and pressure. We find that adding chemicals to kill organisms does not necessarily inactivate enzymes that are already present in seawater. Moreover, bringing excess chemicals to sea (and generating chemical waste) should be avoided when possible.
Line 172: Were these measurements also started at sea and then brought back to the shore-based lab for long-term incubations? If so, please state that samples were transferred from the ship to the shore-based lab within “x” amount of time for sample transport, during which time samples were at “x” temperatures.
We will add this information to the Methods – the chilled incubations were transported in coolers and then stored at 4C; transport between the ship and shore lab took approximately 5 hrs.
Line 221: The first dataset presented in the text, after the physical context, is the POC concentrations. Please consider either showing these profiles in the main text or, instead, focusing first on the bacterial abundance/production, which you have shown in Fig. 1, along with the physical data. If you’d like to leave the bacterial data in this figure, you could consider making the TS plot smaller and including the POC concentration data. While I appreciate the authors' effort to separate the stations based on distinct physical characteristics; however, it appears that the TS plot adds little to your story; please consider moving this to the supplemental or appendix. Also, while the satellite data look informative, these plots are not currently mentioned in the results section, leaving little incentive for the reader to look carefully at the images. Please include text for this satellite data or remove the images from Fig. 1.
We will reduce the size of the TS plots, but not remove them as we believe that they add important context to the story. Even though many of them are somewhat overlapping, we have discussed the nuances in detail with physical oceanographers and believe that they are important (especially the distinctions in surface waters between the stations). We will add text for the satellite images into the results section and add the POC plots here as well.
Line 224: Careful repeating the same data in figures and tables as for Bact. Cell Abundances in Fig. 1 and Table 2. Better to choose which form you like better and remove from the other. Also, at this line you write PON with reference to Table 2, but PON does not appear in Table 2.
We have removed PON from the text. We will remove the cell counts from the table and will place them in a Supplemental Information table, as we think the juxtaposition of cell counts and bacterial productivity in Figure 1 is a striking result (which is discussed in the discussion).
Line 238: Since Chl-a concentrations provide context for presenting the results, you should include these plots in the main text. Consider including these in Figure 1.
This is a great suggestion. We will add a bar plot of Chl-a to Fig. 1 where we also show cell counts and bacterial productivity.
Line 240: “Somewhat different monosaccharide constituents” is not clear. Revise to something more specific.
We will change this phrase to add clarity to the statement.
Line 248-260; Figure 3: The shading as relative abundance is not distinctive enough to tell when one station is truly different from another. Most boxes look about the same shade of blue. Consider converting this into a stacked bar plot where the y-axis is depth, and the x-axis is relative abundance. You have 8 different polysaccharide epitopes, where you could make each epitope a different color. Fig. 3 also presents 3 different extractions (H2O, EDTA, and NaOH), but this is not mentioned in the results, nor are they mentioned in the figure caption. Please highlight the relevance of these 3 different extractions in the results section here.
Thank you for the suggestions. We will change the epitopes to have different colors so that the distinctions between stations are more evident. We will also include more details in the methods and the results to outline the importance of the sequential extraction as well as what was detected for each step.
Line 263: Please consider moving up the description of cell abundances and bacterial productivity if you keep these data in Figure 1.
We will make this Section 3.2, directly after presenting the physical oceanographic characteristics.
Line 271: You have jumped to cell-specific production and left out cell production. If you’d rather describe the cell-specific data, just put the bacterial production data in the supplement and the cell-specific data in the main text. Please add a sentence explaining why cell-specific data should be used.
We think that both the bulk data as well as cell-specific data are important. We will work to incorporate both of them better in the text, as cell abundance and bacterial productivity show differences between the stations.
Line 277-278: Concentrations and production rates for bottom waters are not discernable with this axis. If there are no lines on the figure, does that mean the rates/abundances were truly “zero”? Usually, BP and BA are presented as line graphs like you’ve done for Figure 2b. Note in the caption if you don’t have data for certain depths.
We will clearly distinguish between values below detection limit, and those where we had no sample and therefore no data (i.e., Stn. 17 300 m, 1500 m, 3000 m, bottom. We want the BP, cell counts, and now added in chl-a to be bar charts, so that they are easily comparable on the same plot style; additionally, having the bar charts helps provide distinct color/patterns that are more easily discernable than lines.
Line 280: You’ve presented 16S community for samples collected on a 0.22 μm filter. Very often these represent the “free-living” community, as opposed to the particle-associated, which is collected onto 1.2, 3.0, 5.0 μm filters (or larger). And very often, the bacterial communities below the winter mixed layer significantly differ from the surface communities (ref). It’s odd to see your 3,000m sample grouped with the surface sample in the NMDS plot. Also, could it truly be that you detected no archaea in any of your samples?
Using a 0.22 μm filter without a pre-filter will capture the entire community, particle-associated and free living. As commented above, the primers we used do not detect Archaea. Upon investigation, the 3000 m sample at Stn. 19 contains quite a similar community to the community at the DCM for that station (a contribution of Bacteroidetes as well as a similar selection of alpha- and gammaproteobacteria).
Figure 4: You don’t describe what the shaded areas refer to in the figure caption for the NMDS plots.
We will add this information to the caption.
Lines 299 and 310: The overlapping shaded areas in your NMDS plot for all stations do not support these first sentences. Be more quantitative in how you describe community differences. There are packages in R that can help you to do this, which will generate p-values that you can use to support whether the stations were significantly different from one another.
We will add p-values to the community composition to quantitatively show distinctions with depth as well as station. Thank you for the suggestion, as this was an oversight that we meant to include.
Line 340: State in the Figure 5 caption whether samples without bars shown are “zero” or no data available
We will add to the boxes to indicate if the values are present, zero, or low.
Line 419-420: Please link the specific figure to which you’re referring to each part of the sentence (e.g., polysaccharides comment refers to Fig 3, and monosaccharide comment refers to Fig 2).
This will be changed.
Line 420-421: Please make this sentence clearer. I think what you mean is: that there is strong evidence for differences based on depth (please add a reference to support this), but previous monosaccharide evidence suggested little variation by location (Aluwihare reference). Note that the Aluwihare reference only refers to surface waters so you’ll need other reference(s) to support the depth trends.
We will change this sentence to reflect the correct reference for Aluwihare, and find additional supporting references (if any) for the depth trends.
Line 430: Would it be possible to assign each symbol in Figure 9 to a monosaccharide? If you have some polysaccharides that are of interest to this study, could you make some representation of those in this Figure? If possible, it could help to somehow summarize the main findings about polysaccharide persistence/removal observed here. For instance, do the polysaccharides leading to monosaccharide composition via specific enzymes in the surface differ in a simplistic way from those in the deepest depths?
We will revise Fig 9, but it is important to retain the main point – hydrolysis of different polysaccharides can lead to the same collection of monosaccharides. The complex interrelationships of variations in communities, enzyme activities, and polysaccharide epitopes (polysaccharide structures) do not seem to lend themselves to simple conceptual figures.
Line 458: Could you list which end-acting enzymes you’re referring to here? It’s helpful for the reader not to need to flip back as you’ve done for the exo-acting enzymes.
We will add these to the text.
Lines 468-480: You don’t actually need your 16S data for this discussion. This only mentions that there should be a connection between the POM composition/enzymes and the taxa.
For lines 474 forward, we should have referenced figs. 4 and 6 here (community composition, polysaccharide hydrolase activities), both of which show distinctions with depth. We find depth-related distinctions, the references we cite found temporal distinctions in communities and genes and proteins (e.g. Teeling et al) as well as depth-difference (e.g. DeLong et al.) This part of the Discussion will be rewritten.
Lines 483-513: This section is particularly compelling. Any other examples that you could dive into besides fucoidan and laminarin? I’d recommend, especially if the DNA descriptions are problematic, you expand the comparisons to other kinds of substrates/enzymatic between depths/locations. Additionally, you haven’t brought the cell-specific leu uptake nor cell abundance data back into the discussion, which do appear to provide a useful context to interpreting your sugar/enzymatic activity data.
Fucoidan and laminarin appear to be ‘bookends’ in terms of not easily hydrolyzed/easily hydrolyzed polysaccharides – there are few comparable data in the literature on other polysaccharides. This manuscript is (to the best of our knowledge) the first to combine measurements of specific enzyme activities with more detailed compositional investigations of carbohydrates and information about microbial community composition – the project required significant collaborations and a lot of effort. We agree that further discussion of leucine uptake and cell abundance would enhance the discussion, so we will revise this part of the manuscript.
Line 514: The use of “thus could” seems odd here, especially if someone jumps to read your conclusions without reading all of the prior text. Recommend removing this.
This will be removed.
Citation: https://doi.org/10.5194/egusphere-2024-615-AC1
-
AC1: 'Reply on RC1', Chad Lloyd, 29 Jun 2024
-
RC2: 'Comment on egusphere-2024-615', Anonymous Referee #2, 29 Apr 2024
The authors present a very interesting study about the spatial and vertical variability of carbohydrate composition and utilization in the ocean using new analytical approaches, which is of extremely high value. The study also includes information about bacterial community composition, which the authors aim to relate with inventories and rates of utilization of such compounds. The dataset is of high quality, the manuscript is overall very well written, and the information is novel and relevant, yet there are some major issues that should be addressed. Particularly interesting is the observation of the disconnection between the composition of monomers and that of complex carbohydrates. A major flaw of the study is the insufficient effort to link chemical and biological information. The correlation between bacterial community composition and carbohydrates composition and enzymatic rates is surprisingly missing. The authors should include multivariate analysis statistically linking bot datasets (chemical a taxonomic) and include a detailed discussion on the links between bacterial composition and the carbohydrate composition and utilization rates. On the other side, the authors should also consider using an appropriate transformation of compositional data to avoid the associated biases (e.g. using the clr transformation). Another issue is related with the primers used for microbial taxonomic composition, the authors should mention that they are missing an important part of the microbial community, the archaea, which may numerically dominate in meso- and bathypelagic waters and may also play a relevant role in carbohydrate utilization.
Specific comments
-Line 34. I suggest changing “requisite” to “required”.
-Line 72. Stn 17 is not in the shelf. Please revise.
-Line 73. Please explain all the abbreviations used in the text (T, S).
-Line 131. This must be an error, 25 mL is not enough for DNA extraction.
-Lines 169-171. Please clarify for what samples 15 mL sample incubations were used.
-Lines 188-192. The author should be cautious using a too reduced dataset for correlations among many variables. I suggest using only correlations with the entire dataset.
-Lines 207-208. I suggest including a figure with O2 profiles. The location of O2 minimum is relevant for the discussion of the results.
-Libe 291. The authors could discuss the high abundance of these bacteria in the stn 20 where the BP was highest.
-Line 299. Did the differences were statistically significant?
-Line 383. Please revise as there are two figure 6.
-Lines 392-398. I would remove these correlation analyses with only part of the dataset, as in some cases, the number of samples may be too small to obtain reliable correlations.
-Lines 415-417. The authors should exploit much better their dataset to contribute to this issue.
-Line 509. What do the authors mean here by “selfish”?
Citation: https://doi.org/10.5194/egusphere-2024-615-RC2 -
AC2: 'Reply on RC2', Chad Lloyd, 29 Jun 2024
We would like to thank the reviewer for the thoughtful comments. We have copied their comments and posted our replies in bold below each comment.
REVIEWER 2
The authors present a very interesting study about the spatial and vertical variability of carbohydrate composition and utilization in the ocean using new analytical approaches, which is of extremely high value. The study also includes information about bacterial community composition, which the authors aim to relate with inventories and rates of utilization of such compounds. The dataset is of high quality, the manuscript is overall very well written, and the information is novel and relevant, yet there are some major issues that should be addressed. Particularly interesting is the observation of the disconnection between the composition of monomers and that of complex carbohydrates. A major flaw of the study is the insufficient effort to link chemical and biological information. The correlation between bacterial community composition and carbohydrates composition and enzymatic rates is surprisingly missing. The authors should include multivariate analysis statistically linking bot datasets (chemical a taxonomic) and include a detailed discussion on the links between bacterial composition and the carbohydrate composition and utilization rates. On the other side, the authors should also consider using an appropriate transformation of compositional data to avoid the associated biases (e.g. using the clr transformation). Another issue is related with the primers used for microbial taxonomic composition, the authors should mention that they are missing an important part of the microbial community, the archaea, which may numerically dominate in meso- and bathypelagic waters and may also play a relevant role in carbohydrate utilization.
As mentioned above, our primers do not include archaea in the analyses; while it is possible that they play an important role in carbohydrate degradation and utilization, we are focusing solely on the bacterial community for this dataset.
As commented in response to Reviewer #1, other than for ‘selfish bacteria’ (Cuskin et al. 2015; Human gut Bacteroidetes can utilize yeast mannan through a selfish mechanism. Nature 517, 165-173; Reintjes et al. 2017 An alternative polysaccharide uptake mechanism of marine bacteria. The ISME Journal 11, 1640-1650), it is currently not possible to link specific organisms to specific hydrolytic activities – this is beyond the state of the art currently. We can make links between groups (Gammaproteobacteria, Bacteroidetes, etc.) that often show up in blooms and genetically/genomically/experimentally have many enzymes, but the enzyme complement differs considerably even in closely related species (i.e., Avci et al., 2020; Polysaccharide niche partitioning of distinct Polaribacter clades during North Sea spring algal blooms. The ISME J doi.org/10.1038/s41396-020-0601-y). Multivariate analyses are not helpful in this respect, since they could simply correlate higher abundances of specific bacteria to higher enzymatic activities, even in cases where they are not able to produce the enzymes necessary for carbohydrate degradation. We do not think that statistical correlations will be informative in linking specific bacteria to measurements of enzyme activities.
Specific comments
-Line 34. I suggest changing “requisite” to “required”.
We will change the wording, as suggested.
-Line 72. Stn 17 is not in the shelf. Please revise.
Thank you for catching this. We will change this to continental slope.
-Line 73. Please explain all the abbreviations used in the text (T, S).
We will define all of the abbreviations before using them.
-Line 131. This must be an error, 25 mL is not enough for DNA extraction.
As we noted in response to Reviewer #1, 25 ml is sufficient for DNA extraction, and is in fact a larger volume than is used in some studies of bacterial community composition in seawater. The DNeasy PowerWater Kit we used is designed for the extraction of low amounts of DNA from aquatic, especially marine, environmental aquatic samples. With 25 mL sample volume, we extract DNA from a maximum of 4x108 cells in surface waters of station 19 and 2.25 x105 cells in bottom water of station 20 – a sufficient number of bacteria to cover the full community. See further replied to the line-by-line comments. Additionally, as part of the bioinformatics processing, we created rarefaction curves to ensure that we had an adequate sampling of the microbial community; if we had too few reads, we would have removed them from the analysis, but there were not. As commented above, these samples were analyzed in the Molecular Ecology Department at the Max Planck Institute for Marine Microbiology, where we have decades of experience in this methodology. Reintjes et al., (2017) used only 10 mL in their DNA extractions and obtained sufficient DNA to adequately analyze the bacterial community (doi: 10.1038/ismej.2017.26).
-Lines 169-171. Please clarify for what samples 15 mL sample incubations were used.
The polysaccharide incubation (FLAPS) samples were done in 15-mL tubes, we will note this specifically in the manuscript.
-Lines 188-192. The author should be cautious using a too reduced dataset for correlations among many variables. I suggest using only correlations with the entire dataset.
We have presented both the overall dataset (Fig. 7), as well as reduced datasets (Fig. 8) in the manuscript. As stated, and seen in Fig. 8, by looking at correlations within a station, we see a very different picture than what is seen with the overall dataset, highlighting the point that station-related differences in enzyme activities can be lost when combining all of the data. We understand the pitfalls of using a reduced dataset to draw correlations between specific activities, but in this case, we want to highlight the point that there are differences between stations that emerge from the corrplot. These data illustrate the point that station and depth are significant factors for many of the differences we measured.
-Lines 207-208. I suggest including a figure with O2 profiles. The location of O2 minimum is relevant for the discussion of the results.
We will add a plot to the supplemental information.
-Libe 291. The authors could discuss the high abundance of these bacteria in the stn 20 where the BP was highest.
We will discuss this observation at greater length, as suggested.
-Line 299. Did the differences were statistically significant?
We did not test for statistical relevance, since these are descriptive comparisons of few community members standing out in abundance, based on individual measurements from few sampling sites. However, based on comments from Reviewer 1, we plan to add the statistical p-values to support our statements.
-Line 383. Please revise as there are two figure 6.
We will fix this issue; thank you for catching this!
-Lines 392-398. I would remove these correlation analyses with only part of the dataset, as in some cases, the number of samples may be too small to obtain reliable correlations.
We have presented both the overall dataset (Fig. 7), as well as reduced datasets (Fig. 8) in the manuscript. As stated above and seen in Fig. 8, by looking at correlations by depth, we see a very different picture than what is evident with the overall dataset. Depth-related differences make sense, given depth-related distinctions in microbial community members and therefore likely also in their metabolic capabilities. These differences are highlighted by the differences in enzymatic activities measured.
-Lines 415-417. The authors should exploit much better their dataset to contribute to this issue.
We will revise the discussion to more thoroughly discuss this point.
-Line 509. What do the authors mean here by “selfish”?
‘Selfish’ described a specific mode of bacterial substrate utilization, in which selfish organisms use surface-associated enzymes to bind and partially degrade HMW substrate with little to no loss of substrate to the external environment (Cuskin et al. 2015; Human gut Bacteroidetes can utilize yeast mannan through a selfish mechanism. Nature 517, 165-173; Reintjes et al. 2017 An alternative polysaccharide uptake mechanism of marine bacteria. The ISME Journal 11, 1640-1650; Reintjes et al. 2019; https://doi.org/10.1038/s41396-018-0326-3). We will explain the context more thoroughly.
Citation: https://doi.org/10.5194/egusphere-2024-615-AC2
-
AC2: 'Reply on RC2', Chad Lloyd, 29 Jun 2024
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 396 | 143 | 50 | 589 | 32 | 18 | 20 |
- HTML: 396
- PDF: 143
- XML: 50
- Total: 589
- Supplement: 32
- BibTeX: 18
- EndNote: 20
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1