Supplementary Information for

0 downloads 0 Views 2MB Size Report
... the growth of aerobic, heterotrophic bacteria that produce only a subset of ..... the bacterial taxa present in Lake Lugano showed a strong correlation with the ..... McMurdie PJ, Holmes S (2013) Phyloseq: An R Package for Reproducible ...
Supplementary Information for Redox-dependent niche differentiation provides evidence for multiple bacterial sources of glycerol tetraether lipids in lakes Yuki Weber*, Jaap S. Sinninghe Damsté, Jakob Zopfi, Cindy De Jonge, Adrian Gilli, Carsten J. Schubert, Fabio Lepori, Moritz F. Lehmann, Helge Niemann * To whom correspondence should be addressed. Email: [email protected] This PDF file includes: Supplementary text Supplementary Materials and Methods Figures S1 to S13 Tables S1 and S2 References for SI citations Other supplementary materials for this manuscript include the following: Dataset S1: Bacterial 16S-rDNA abundances from the water column of Lake Lugano Dataset S2: Branched GDGT abundances from Lake Lugano and other Alpine lakes R code: https://github.com/yukiweber/phylo.lipids

1 www.pnas.org/cgi/doi/10.1073/pnas.1805186115

Supplementary Information Text In vitro production of brGDGTs under oxic vs. anoxic conditions. In order to test for the potential of brGDGT synthesis in freshwater habitats, we incubated SPM collected from the epilimnion and monimolimnion (10 m and 260 m water depths, respectively) with additions of 99% 13C-labeled organic substrates (glucose and algal biomass) under quasi in situ conditions of the epilimnion (i.e., oxic/20°C/light) and the deeper hypolimnion (anoxic/4°C/dark). After six weeks of incubation, de novo brGDGT synthesis was evidenced by high levels of 13C incorporation into brGDGTs (20–26 % 13C) (Fig. S6B,C). Remarkably, only five out of the nine brGDGTs initially present in 10 m SPM were synthesized (i.e., Ia, Ib, and the C6-methylated brGDGTs IIa’ IIb’, and IIIa’; see Fig. S6A), but the concentrations of specific brGDGTs in the epilimnetic biomass increased by a factor of ~100 (Figs. S6A,B), corresponding to a production rate of 0.62 µg g(Corg)-1 d-1. Moreover, the production rates we observed were ~100 times higher than those previously reported from a peat bog (6.3 µg g(Corg)-1 d-1; ref. 1), where incubation experiments were conducted without addition of organic C substrates and under hypoxic conditions. Together this strongly suggests that the addition of organic electron donors (glucose and algal biomass) supported the growth of aerobic, heterotrophic bacteria that produce only a subset of the naturally occurring brGDGTs. In contrast to the aerobic experiment, we did not observe considerable brGDGT production in the anoxic incubation of SPM collected from below the RTZ at 260 m depth. More precisely, both concentrations and 13C contents were identical to the t0 values within analytical error (i.e., ±0.5% 13C for isotope analysis, and ±10% of t0 concentrations). Instead, we found substantial 13C incorporation (up to 33 % 13C at t1) into GDGT-0 (caldarchaeol), an isoprenoid GDGT derived from archaea, providing evidence for microbial substrate turnover under the given experimental conditions. The lack of measurable brGDGT synthesis in this anaerobic experiment therefore suggests that the brGDGT-producing bacteria at the RTZ are either (1) micro-aerobic and require at least trace amounts of O2, or (2) use alternative electron acceptors (e.g. nitrate) that were not present in the deeper anoxic waters from 260 m.

Intact polar lipid content of SPM. Intact polar lipids (IPLs) are thought to derive from (recently) living biomass and IPL-brGDGTs can therefore be used to trace the bacterial sources of brGDGTs in the environment (2, 3). Our recent work on the extractability of brGDGTs from SPM in Lake Lugano indicated that up to 76 % of IPL-brGDGTs are not recovered by the most commonly applied protocol for IPL lipids – the ‘modified Bligh-Dyer’ method (4). We also demonstrated in the above study that a large portion (38–75 %) of brGDGTs resists extraction by non-aqueous solvent mixtures (i.e., methanol and dichloromethane mixtures) but was readily released after acid hydrolysis of the extracted sample residue. Remarkably, this resilient fraction of brGDGTs was not recovered by the phosphate-buffered Bligh-Dyer method, suggesting that previous studies have vastly underestimated the total amount or brGDGTs in the water column of lakes. Further experimental investigation then revealed that the ‘resilient’ fraction could in fact be solubilized by 5 % trichloroaceitic acid extraction, and contained acid-hydrolysable head group moieties of unknown type. Here, we therefore assume that the non-solvent extractable fraction of brGDGTs released by acid hydrolysis of the extracted sample residue consist of IPLs. This inference is further supported by our 13C labeling experiment (Fig. S6C), showing the production of highly 13Cenriched brGDGTs that were dominantly non-solvent extractable.

Potential production of brGDGTs within lake sediments. In many lakes the oxycline is located within the top centimeters or millimeters of the sediment, so that both the aerobic and the anaerobic (methanotroph-associated) brGDGT-producers may be active within oxic and anoxic sediment layers, respectively. In Lake Lugano, however, the flux-weighted and Corg-normalized annual average abundance of brGDGTs was almost two times higher in settling particles collected in the deeper water (129 µg g[Corg]-1 in the sediment trap at 176 m) than in the surface sediment (68 µg g[Corg]-1; Fig. S5). The lower concentration in the sediment may be caused by further decomposition of settling organic matter by

2

anaerobic microbes between 176 m water depth and the sediment surface at 286 m, and/or may be due to dilution by terrigenous organic matter (with a lower relative brGDGT content) that has been transported laterally by bottom water currents and/or turbidites. Moreover, the C6-methylated brGDGT IIIa′ is relatively much less abundant in the sediment (45 %) than in the deepest sediment trap (74 %), whereas the relative distribution of all other brGDGTs remains largely unchanged (Fig. S5, Table S2). This suggests that IIIa′ is preferentially degraded during passage of the deeper anoxic water body, probably because it originates from the upper oxic water column and cells containing IIIa′ have been subject to biodegradation for a longer period of time. Together this strongly suggests that, at least in Lake Lugano, sedimentary brGDGTs derive predominantly from the water column and no substantial amounts are produced in the sediments.

Supplementary Materials and Methods 13

C incorporation experiments with SPM from Lake Lugano. SPM was collected by in situ filtration of 140–185 L of lake water using glass fiber filters (0.7 µm; Whatman GF/F) in August 2013. Immediately after retrieval, filters were placed in sterilized flasks together with 800 ml of in situ lake water of the respective depth, and 100 mg 13C6 Glucose (99%, Euriso-Top®, France) as well as 50 mg freeze-dried Spirulina sp. cells (99% 13C, Sigma) were added. The flasks containing SPM from 10 m depth were closed with cellulose stoppers to facilitate gas exchange (oxic) and incubated at 20°C under artificial daylight (12 h day-1). Flasks containing SPM from 260 m were flushed with N2, sealed airtight and left at 4°C. Both treatments were agitated at 30 rpm on orbital shakers, and all materials were autoclaved (121°C) prior to the experiment. After 180 days (t1), all solids were collected by filtration (0.7 µm GF/F), freeze-dried, and processed as described for SPM samples (i.e., solvent extraction followed by acid hydrolysis of the residue). The hydrolyzed extracts and the residue hydrolysates were then analyzed on the same HPLC-MS system as the regular SPM samples (5). The remaining polar fraction was subjected to ether cleavage, and the brGDGT-derived alkanes were analyzed by GC-MS and GC/C/IRMS. However, due the massive presence of sterol-derived hydrocarbons (released from the 13Clabelled Spirulina biomass), we were unable to resolve 13C incorporation below 0.5% (as determined by co-injection of standards). Initial brGDGT abundances (t0) were measured in the lipid extract of unamended SPM. Assessment of 13C label incorporation by UHPLC-APCI-MS. For the analysis of 13C-labelled brGDGTs, we monitored all single ions between m/z 1018 and m/z 1077. We then assigned brGDGTs to each peak in the base peak chromatograms (BPC; Fig. S6A), based on relative retention times (with respect to the internal standard) and comparison with unlabeled samples. For each peak, all monitored ions (m/z 1018 to m/z 1077) were integrated using the corresponding extracted mass chromatograms (EIC). We then weighted the peak areas by the number of 13C substitutions, e.g. 0×AREA(m/z 1022) + 1×AREA(m/z 1023) + 2×AREA(m/z 1024) + (…) + 23×AREA(m/z 1055) in case of brGDGT Ia (m/z 1022). The sum of the weighted peak areas was then divided by the sum of the un-weighted peak areas, yielding the average number of 13C substitutions in each brGDGT. The relative 13C content was obtained through division by the number of C atoms per brGDGT molecule. We note that co-eluting, minor brGDGTs may have contributed to the calculated peak areas in the EICs, though, we consider their contribution to be small (likely ≲5 %).

3

Processing of 16S-rDNA data. Bioinformatical processing of the raw sequencing data (a total of 1,741,418×2 paired raw reads from 15 water column samples) was performed with MacQIIME V.1.9.1 (6) and Cutadapt (7). Briefly, paired reads were stitched (min. overlap: 50 bp), quality-filtered (Phred≥25), cleaned from PCR primers, and de-replicated. Sequences were then subjected to both de novo- and reference-based chimera detection, and only those passing either filter were retained (UCHIME; [8]). The remaining sequences were, clustered into operational taxonomic units (OTUs) at 97% sequence identity (USAERCH 6.1 ; [9]). Taxonomy was assigned with the RDP classifier (10) (confidence threshold: 0.6), using a set of QIIME-formatted and pre-clustered (97 %) reference OTUs (SILVA 128 SSU). Downstream phylogenetic analyses were performed in ‘R’ (11), using the package ‘Phyloseq’ (12). OTUs occurring in less than five samples, and those with less than 40 reads across all samples, were discarded before correlation analysis.

Comparison of 16S-rDNA data with brGDGT profiles in Lake Lugano. In order to compare vertical trends in brGDGT concentrations vs. that of OTUs, we approximated OTU-specific DNA concentrations in the water column of Lake Lugano. For this approach we accounted for the contribution of eukaryotic 16S-rDNA from chloroplast and mitochondria, because otherwise bacterial taxa would be overestimated. To this end, the raw OUT abundances (including chloroplast- and mitochondria) were divided by the total number of acquired reads, and thereafter multiplied with total DNA concentrations in the water column determined by absorption spectrophotometry of the bulk DNA extract. Finally, these ‘OTU-specific DNA concentrations’ and the concentrations of brGDGTs were subjected to bivariate correlation analysis, and the taxonomic affiliations of the highly correlated OTUs (r>0.75) were summarized. The R code used for lipid–DNA correlation and summarizing phylogenetic affiliations is available at github.com/yukiweber/phylo.lipids (release: 0.0.0900, commit: 513f5ba).

Gas chromatographic analysis of brGDGT-derived alkanes. Hydrocarbons released by ether

cleavage were dissolved in n-hexane and injected at 70°C on a 50 m low-bleed Rxi®-5ms GC column (0.2 mm iD, 0.33 µm df, Restek, USA). After 2 min hold time, the temperature was rapidly increased to 220 °C at 20 °C min-1. Temperature was held for 45 min to improve separation of the target compounds from co-eluting steranes, and subsequently increased at 4 °C min-1 to 320 °C (held for 35 min). The carbon isotopic composition of the analytes was calibrated by separate analysis of an alkane mixture with known isotopic compositions (n-alkane mixture B3 provided by A. Schimmelmann, Indiana University, USA). GC/C/IRMS performance was monitored by regular analysis of this standard mixture.

Calculation of MAT and brGDGT-based proxy indices. As air temperatures were not directly measured in this study, we approximated MAT at each of the lake sites using a linear lapse rate model derived from instrumental data obtained from the Federal Office of Meteorology and Climatology MeteoSwiss: MAT(°C)=11.76–0.00494×h (R2=0.91 ; p=10-15 ; n=204), where h is the lake surface elevation in meters above sea level. Previously published brGDGT proxy indices were calculated from the LCMS peak areas as follows: MBT’5me= (Ia+Ib+Ic) / (Ia+Ib+Ic+IIa+IIb+IIc+IIIa) ; (Ref. 13), MBT’6me= (Ia+Ib+Ic) / (Ia+Ib+Ic+IIa′+IIb′+IIc′+IIIa′) ; (Ref. 14), MBT56 = (Ia+Ib+Ic+IIa′) / (Ia+Ib+Ic+IIa+IIb+IIc+IIIa+IIIa′) (Ref. 15).

4

Supplementary Figures and Tables

Ia IIIa Ib IIIb

Ic

IIIc IIa IIIa′ IIb IIc

IIIb′

IIa′

IIIc′ IIb′ IIc′

IIIa′′

Fig. S1. Structures of all known brGDGTs, including C5 methylated isomers (16), C6 methylated isomers (indicated with a prime; refs. 13, 17), and the recently discovered C5/6 methylated isomer (IIIa′′; ref. 18). Note that brGDGTs IIIa, IIIa’ and IIIa’’ each comprise a symmetric and an asymmetric isomer, which cannot be separated by HPLC-based chromatography. Structures of the hexa- and penta-methylated brGDGTs with cyclopentyl moiety(ies) IIb′, IIc’, IIIb′ and IIIc′ are tentative.

5

15

0.0 0.3 0.6 0.9

10

5

0.000 0.002 0.004 0.006 0.008 0.0 0.5 1.0 1.5 2.0 0.3 0.00 0.01 0.02 0.03 0.04 Ic

0.03

0.2

0.1

0.0

Ib

0.01

Ia

IIc

0.02

0.02

IIb'

2

1

0

IIa

0.015 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.00

0.010

0.005

0.000

IIIb

IIIc'

IIb

0.00 0.25 0.50 0.75 IIc'

0 50 100 150 200 250

0.01

0.00

0.000 0.025 0.050 0.075

IIIc

IIa'

0 50 100 150 200 250

IIIa'

0 1 2 3 4 0

10

0

IIIa''

IIIb'

0 50 100 150 200 250

depth (m)

5

IIIa

0 50 100 150 200 250

concentration (ng L-1) RES

IPL−ex

CL

Fig. S2. Cumulative concentration of individual brGDGTs in suspended particles from Lake Lugano obtained by ultrasonic extraction with methanol and dichloromethane (CL and IPL-ex), and acid hydrolysis of the extracted residual material (RES). Key: CL, core brGDGTs analyzed in the total lipid extract; IPL-ex, intact polar brGDGTs present in the total lipid extract, determined by acid hydrolysis of the extract (ref. 19); RES, intact polar brGDGTs released from the pre-extracted sample residues by acid hydrolysis (ref. 4).

6

IIIa

0

IIIa''

IIIa'

IIIb

50 100

IIIb'

0

IIIc

IIIc'

30

20

10

1500

1000

500

0 10 20 30 40 50

200

150

100

200

50

150

IIa

50 100

IIb

IIb'

75

50

25

0.0 0.5 1.0 1.5 2.0 2.5

1.2

0.8

0.4

0.0

9

IIa'

0

IIc

50 100

0

IIc'

Ia

Ib

3

2

1

10 20 30 40 50

30

20

10

200

150

200

100

150

50

depth (m)

6

200

3

150

Ic

50 100

0

PM

7.5

10.0

5.0

2.5

40

30

20

10

120

80

40

200

1 2 3 4 5 6

150

POC

50 100

3e+08

2e+08

1e+08

2e+09

200

1e+09

150

average flux (ng m−2 d−1)

Fig. S3. Average fluxes of brGDGTs as well as total particulate matter (PM) and particulate organic carbon (POC) in sediment traps deployed at three discrete depths in Lake Lugano throughout an annual cycle (Mar/1990 – Mar/1991). Error bars show the standard deviation.

7

Fig. S4. GC-combustion mass chromatogram of m/z 44 (CO2), showing partial co-elution of brGDGTderived alkanes b and c, and d and f, respectively. Therefore, δ13C values were determined by combined integration of the ‘double peaks’.

8

Fig. S5: Organic carbon (Corg) -normalized concentrations and relative distributions of brGDGTs in samples from the catchment, water column, and sediments of Lake Lugano. Sediment trap values represent flux-weighted annual averages. Labels on the left refer to water depth in case of lake samples, and sampling location (see Fig. S11) for catchment samples. The lower brGDGT abundance in the sedimentary organic matter compared to settling (traps) and suspended particles (SPM) indicates that in Lake Lugano, no substantial amounts of brGDGTs are produced in the sediments. The factional abundance of brGDGTs in traps vs sediment is further detailed in Table S2.

9

A

response

0.03

IS

744 1050 1036

.02 .01

t0

IIIa′

m/z 1034 1022 1020

IIa′ IIa

Ia Ib

IIb′ IIb

IIIa

Ic

0

IIIa′

IIa′

1.6

IIb′

Ia

1060

Ib

t1

0.5 0 1

IS

IIa′

0.5

IIIa′ BPC 1

IIb′ Ia Ib

?

0

1.2 1.4 relative retention time

1.6

t1 30

25

25

20

20

15

15

10

10

5

5

0

Ia

Ib

*

*

*

Ic

IIa

IIb

* IIIa IIa′ IIb′ IIIa′

-1

30

concentration (ng L )

C

% 13C

1080

1020

1040

1060

m/z 1080

1040

1020

m/z 1080

1040

1080

m/z 1020

1040

1020

1080

1060

1

1040

1020

abundance response

m/z

m/z

1060

B

1.2 1.4 relative retention time

1060

1

13

C content extract

residue

concentration residue extract

0

Fig. S6. Abundance and 13C content of brGDGTs before (A) and after incubation with 13C-labeled substrates (B+C), respectively, under oxic conditions. Cumulative mass spectra (cf. Supplementary Materials Methods) show an average mass shift of ~+15 amu due to incorporation of the 13C-label (B), corresponding to an average 13C content of ~25% for all brGDGTs (C). Note the presence of C5methylated brGDGTs (IIIa, IIa, and IIb) in the unamended SPM (t0 ; A), which were not biosynthesized under the laboratory conditions and thus were below detection limit (*) after incubation (t1; B,C). brGDGTs were analyzed in hydrolyzed total lipid extracts (TLE; blue bars and diamonds), as well as in the fraction released by acid hydrolysis of the post-extraction residues (gray bars and diamonds). Experiments were performed in duplicate (light and dark colors in panel C). Error bars indicate the error of quantification (±10 %; 2 SD). Mass spectrometer response in panels A and B is shown relative to 0.1 µg of internal standard (IS, m/z 744). The concentration of brGDGTs in SPM from 10 m prior to incubation was ~0.8 ng L-1 of lake water.

10

acidobacteria (n=48)

Σ(bG) Ia

other bacteria (n=1533)

IIa' IIa IIIa'' IIIa' IIIa -1

-0.5

0

0.5

1

correlation (r) Fig. S7. Correlation coefficients (r) between OTU-specific bacterial 16S-rDNA abundances and brGDGT concentrations in the water column of Lake Lugano. Several members of the Acidobacteria (red filled circles) as well as other bacterial phyla (gray open circles) correlate well with each of the major brGDGTs, suggesting that the biological sources of brGDGTs are among these taxa. In contrast, none of the bacterial taxa present in Lake Lugano showed a strong correlation with the summed concentration of all brGDGTs (Σ(bG)) (r≲0.5; dashed line), in support of the interpretation that individual brGDGTs have different (consortia) of source organisms. Positive correlations >0.5 are highlighted in yellow. Negative r values indicate anti-correlation.

11

SD6

0

SD3

SD8

50 100 150 200

SD4

0

0.0 2.5 5.0 7.5 10.0

10

5

0

30

20

10

0

250

SD21

SD5

depth

50 100 150 200

0

SD22

1.2 0.0 0.2 0.4 0.6 0.8

0.8

0.4

0.0

1.0

0.5

0.0

250

SD15

SD17

50 100 150 200

0.06

0.04

0.02

0.00 0.03 0.06 0.09 0.12 0.000 0.025 0.050 0.075 0.100 0.00

250

estimated concentraion (ng L-1)

Fig. S8. Taxon-specific 16S-rDNA concentration estimates of OTUs affiliated with different Subdivisions (SDs) of the Acidobacteria. Note the differential preference of certain groups for the oxic (above 90 m) and anoxic parts (below 90 m) of the water column, resembling the differential vertical trends of brGDGTs IIIa′ and IIIa′′ (cf. Fig. 2C). Panels are arranged in the order of decreasing abundance (upper left to bottom right).

12

0

DIC DOC

depth (m)

50 100 150 200 250 −30

−25

−20

−15

−10

−5

δ13C (‰) Fig. S9. Stable C isotope composition of dissolved organic (DOC) and dissolved inorganic (DIC) carbon pools in Lake Lugano (01 Nov. 2014).

13

20 m

85 m

176 m 20 15

0.4

10

0.2

T (°C)

MBT

0.6

5

MBT′5me

MBT′6me

MBT56

Jan

Oct

Jul

Apr

Jan

Oct

Jul

Apr

Jan

Oct

Jul

Apr

0.0

water T

Fig. S10. Comparison of in situ water T with three different brGDGT temperature proxies (i.e., Methylation Indices of brGDGTs – MBT) in settling particles of Lake Lugano collected at three depths in 1990/1991. Despite large temporal variations in the MBT values near the surface (upper trap at 20 m), the proxy values in settling particles further below (i.e., at 85 m and 176 m) were almost invariant throughout the year, consistent with the constantly low T conditions in the hypolimnion. Locally weighted scatterplot smoothing (LOESS) was applied to the proxy data, and 95 % confident intervals are shown as shaded areas.

14

δ13Ca,b,c > -36‰

0.5

δ13Ca,b,c < -36‰

r = 0.53 p = 0.14

0.0 0.5

r = 0.87 p = 0.001

0.2 0.1

r = -0.33 p = 0.32

0.0 2

4

6

8

10

r = 0.89 p = 0.04 2

4

6

8

10

r = 0.81 p = 0.004

r = -0.24 p = 0.48

r = 0.28 p = 0.65

0.4 0.2

6

8

10

r = -0.59 p = 0.29

0.6 0.4 0.2

2

4

6

8

10

2

4

6

8

2

4

6

8

depth > 40 m

r = 0.78 p = 0.01

0.6

4

r = -0.55 p = 0.08

0.8

depth < 40 m

0.2

2

r = 0.69 p = 0.03

δ13Ca,b,c < -36‰

0.4

0.0

r = 0.14 p = 0.71

MAT (°C)

0.6

0.0

0.0

0.0

MAT (°C) δ13Ca,b,c > -36‰

0.2

10

MAT (°C)

Fig. S11. Lapse rate-based mean annual air T (MAT) plotted against three recently proposed brGDGT temperature proxy indices (MBT′5me, ref. 13; MBT′6me, ref. 14; and MBT′56, ref. 15) in lake surface sediments from the European Alps (n=35). Sites are binned by maximum water depth and the average stable C isotope composition of brGDGTs (δ13Ca,b,c). Note that all three proxies show highest positive responses to MAT in shallow lakes characterized by a low 13C content of brGDGTs. Correlation coefficients (r) as well as p-values of the slope coefficients are given for each bin.

15

10

depth > 40 m

0.3

depth > 40 m

0.4

0.4

MBT′6me

0.1

0.6

depth < 40 m

0.2

depth < 40 m

0.3

MBT56

δ13Ca,b,c < -36‰

0.8

0.4

MBT′5me

δ13Ca,b,c > -36‰

8.966°E

9.000°E

9.033°E

o

Cu

zz

cci

Re

o

8.933°E

(275 m)

46.033°N

#2

#1 20 10

ate

0

IT

sar Cas

0

CH 0

-17 m

(275 m)

0

La

ke

Lu

n ga

o

N

Basi h t or

n

46.000°N

46.000°N

Lugano

DE

FR 45.966°N

Soils#1 (670 –780 m)

IT

2 km

0

8.933°E

AT

CH

8.966°E

Soils#2 (340 –390 m)

45.966°N

10 0

20

46.033°N

Porlezza

9.000°E

River sediments

9.033°E

Lake sampling

Fig. S12: Sampling sites at Lake Lugano north basin (Southern Switzerland). Lake sampling was performed at 46.01077° N; 9.02021° E, close to the deepest location. Note that catchment soil samples from Lake Lugano were collected at two distinct elevation levels.

16

7

8

9

47.5 47

*

-30

-1

µg g(Corg)

CH

IIIa′′ (%) IIIa′′ traces

5 4 2

*

6 2

AU

**

11

-34

3

*

3 5

46

-32

-36

46.5

lat. (°N)

4

11

D

brGDGTs: 45 180

10

-38 3

-40 2

F

3

5

IT

δ13C(a,b,c) (‰)

6

-42 -44

6

7

8

9

10

11

lon. (°E)

Fig. S13: Location of lake sites investigated in this study, as well as bulk sedimentary brGDGT concentration normalized to bulk organic carbon (Corg) (size of circles), weighted average 13C content of brGDGT–derived alkyl chains (a,b,c) (color scale), and the percent contribution of the uncommon isomer IIIa′′ to the sedimentary brGDGT pool (numbers).

17

river sed.

catchment soils

? an an ? ? ? ? ox ? ? an an ? ? ? an ? ox‡‡ ox ? ox an an ? ox ? ox ? an ox an ox ox ox ox an

? eu eu ? ? ? ? olig ? ? eu eu ? ? ? eu ? olig‡‡ olig ? mes mes mes ? mes ? olig ? eu olig eu¶¶ mes## mes olig mes eu

y y y y y y y y (y) (y) y y (y) y y (y) y n n n n n n n n n n n n n n n n n n n n n n

TOC ¶

trophic state†

40 66 21 16 10 6 9 45 24 34 10 32 14 23 85 12 4 24 46 47 68 286 13 (30)‡ 152 56 85 69 16 29 25 78 215 7 112 135

13

IIIa''#

bottom water*

m (asl)

ID lake name AZ Alzasca, Lago d' 1857 BA Baldeggersee 464 CD Cadagno, Lago di 1926 CM Cama, Lago di 1266 CN Canovasee 777 DS Dosso, Lago 1656 ED Endine, Lago di 335 EN Engstlensee 1852 FA Fälensee 1455 GA Garlate, Lake 198 GE Gerzensee 602 GF Greifensee 452 GH Ghirla, Lago di 443 GL Glattalpsee 1848 GS Grimselsee 1909 HB Hinterburgsee 1523 HS Hinterer Schwendisee 1160 IF Iffigsee 2067 KT Klöntalersee 847 LD Ledro, Lago di 654 LN Lungerensee 693 LL Lake Lugano 271 LZ Lauerzersee 447 MG Mognola, Lago di 2005 NB Lac de Neuchâtel 434 OE Oeschinensee 1580 PS Poschiavo, Lago di 963 RI Ritom, Lago 1855 RO Rotsee 428 SI Silsersee 1800 SO Soppensee 598 SV Silvaplana, Lago di 1791 TH Thunersee 559 TR Trüebsee 1767 VW Vierwaldstättersee 599 ZH Zürichsee 406 AZ Alzasca, Lago d' 1857 BA Baldeggersee 464 CD Cadagno, Lago di 1926 CM Cama, Lago di 1266 CN Canovasee 777 DS Dosso, Lago 1656 FA Fälensee 1455 GE Gerzensee 602 GH Ghirla, Lago di 443 GL Glattalpsee 1848 HB Hinterburgsee 1523 HS Hinterer Schwendisee 1160 IF Iffigsee 2067 MG Mognola, Lago di 2005 RO Rotsee 428 SI Silsersee 1800 SO Soppensee 598 TR Trüebsee 1767 LL Lake Lugano 670–780 LL Lake Lugano 340–390 LL Lake Lugano 318 LL Lake Lugano 271

max. depth (m)

surface sediments

sample type

Table S1: Lake sites investigated in this study

-31.5 -31.9 -30.3 -27.1 -31.1 -30.6 -34.6 -26.8 -30.7 -28.5 n.a -35.2 -30.8 -27.2 -26.6 -35.7 -33.3 -27.6 -28.8 -32.1 -28.7 -30.8 -31.3 -27.9 -30.8 -27.5 -28.8 -26.5 -32.0 -28.4 -34.0 n.a -27.1 -24.9 n.a -34.6 -26.5 -29.0 -26.0 -26.7 -27.2 -25.2 -27.2 -28.5 -27.6 -25.4 -26.3‡ -27.9 -26.3 -26.4 -28.5 -27.6 -28.8 -26.6 -26.6 -26.8 -27.7 -26.2

-34.3 (±0.4) -41.1 (±0.2)

(d+f) avg. (STD) -43.1 (±0.6) -36.6 (±0.6) -34.8 (±0.7) -32.1 (±1.2) -37.3 (±0.5) -35.8 (±0.3) -40.0 (±0.4) -29.4 (±0.5) b.d -33.1 -41.7 (±0.8) -43.3 -36.4 (±1.0) -27.3 (±0.7) -29.2 (±0.0) -49.7 -42.3 (±1.4) b.d -29.6 (±1.0) -35.9 (±1.2) -29.8 (±1.0) -39.1 (±0.8) -33.6 (±0.4) -29.7 (±0.5) -35.1 -29.3 (±0.7) -34.5 (±0.8) -27.6 -37.5 (±0.7) -30.5 (±0.8) -40.7 (±0.9) -33.6 (±0.7) -32.1 (±1.4) b.d -35.6 (±0.2) -34.3 (±0.3) -41.7 (±0.1) b.d b.d

-26.7

-26.5

avg. -42.7 -36.0 -34.2 -28.9 -37.4 -34.4 -40.6 -28.8 -38.3 -34.9 -41.5 -42.3 -35.2 -27.9 -30.3 -42.8 -44.8 -28.5 -29.9 -39.7 -30.6 -36.9 -34.5 -29.3 -33.7 -27.3 -30.3 -29.5 -37.1 -30.3 -40.2 -32.1 -33.2

(a) (STD) (±0.01) (±0.3) (±0.4) (±0.5) (±0.6) (±0.4) (±0.2) (±0.2) (±0.2)

C (‰)

(±0.9) (±0.3) (±0.8) (±0.1) (±0.03) (±0.5) (±0.6) (±0.8) (±0.0) (±0.7) (±0.3) (±0.5) (±0.3) (±0.1) (±0.2) (±0.8) (±0.2) (±0.2) (±0.2) (±0.4)

(b+c) avg. (STD) -41.6 (±0.1) -39.5 (±0.4) -34.9 (±0.6) -30.9 (±0.2) -37.6 (±0.3) -36.3 (±0.5) -40.1 (±0.2) -30.6 (±0.2) -37.9 (±0.8) -34.8 -42.5 (±1.0) -43.1 (±0.1) -37.4 (±1.1) -29.3 (±0.7) -28.2 (±0.2) -44.5 -43.4 (±0.4) -33.7 (±0.2) -32.3 (±0.5) -41.3 (±0.4) -31.0 (±0.4) -36.9 (±0.4) -36.0 (±0.2) -30.2 (±0.4) -34.8 -29.5 (±0.7) -33.9 (±0.4) -28.1 -38.0 (±0.4) -30.4 (±0.5) -42.2 (±0.1) -34.0 (±0.0) -34.7 (±0.2)

(e) avg. (STD) -37.2 (±1.4) -34.5 (±0.8) -33.3 (±0.7) -29.8 (±0.8) -36.8 -32.7 (±1.1) -36.3 (±1.4) -27.9 (±0.4) b.d -27.8 -38.6 -36.4 -35.5 (±0.4) -26.3 (±1.0) b.d -42.8 -36.7 (±0.3) -34.7 -29.3 (±3.1) -36.0 (±0.9) -30.4 (±0.4) -34.8 (±0.3) -32.7 (±0.9) -29.9 (±1.1) -32.4 -30.3 (±0.4) -27.4 (±0.9) -26.3 -33.9 (±2.0) -29.2 (±1.1) -37.2 (±0.1) -33.8 (±1.2) -32.5 (±0.4) b.d -33.9 b.d

-27.4

-24.5 (±0.1) -24.3 (±0.04) -26.7 (±0.1) -27.8 (±0.5) -26.9‡ (±1.5) -27.9‡ (±1.5) -26.6 (±0.3) -25.2 (±0.1) -24.5 (±0.8) -26.8 (±0.2) -26.3 (±0.3) -26.7 (±0.2) -28.1 -28.0 -29.8 -26.0

(±0.4) (±0.4) (±0.7) (±0.3)

-26.7 -26.7 -29.7 -27.0

(±0.5) (±0.2) (±0.7) (±0.1)

Table S1. Lake sites investigated in this study, as well as stable C isotopic composition of brGDGTderived alkanes (a–f) and total organic carbon (TOC). Abbreviations and key — *bottom water oxygenation. an: seasonally or permanently anoxic; ox: year-round oxic — †: trophic state based on epilimnetic total N and/or total P concentrations (20); eu: eutrophic; mes: mesotrophic: olig: oligotrophic; water chemistry data was provided by the Federal Office of the Environment, or taken from literature -— #: occurrence of isomer IIIa′′; parentheses indicate trace amounts — ¶: average of triplicates (STD