• Open Access

Occurrence and abundance of soil-specific bacterial membrane lipid markers in the Têt watershed (southern France): Soil-specific BHPs and branched GDGTs



Recently, four bacteriohopanepolyols (BHPs), adenosylhopane, and structurally similar adenosylhopane-type 1, 2-methyl adenosylhopane, and 2-methyl adenosylhopane-type 1, have been suggested to be characteristic of soil microbial communities and therefore can serve as molecular markers for soil organic matter (OM) supply in river, lake, and marine sediments. In this study, we analyzed BHPs in peats and soils collected in the Têt watershed (southern France) and compared them with branched glycerol dialkyl glycerol tetraethers (GDGTs), a more established molecular tracer of soil OM. Adenosylhopane-type I is identified in all of the samples from the study area except one collected near the Têt River mouth with up to three of the related compounds also frequently present, particularly in the surface samples. The concentrations of soil-specific BHPs in peat environments have been shown to increase with lower δ15N values, providing evidence that N2-fixing bacteria are probably a major source of soil-specific BHPs in acidic environments. It seems likely that soil pH is a major factor controlling BHP occurrence based on statistical analysis of environmental parameters and BHP concentration data. The comparison of the soil-specific BHP concentrations with those of branched GDGTs shows no clear relationship in the Têt River system, supporting the concept that these two groups of soil-specific compounds are synthesized by different microbial organisms living in different niches in the soil profile (e.g., oxic top versus anoxic deep).

1. Introduction

Bacteriohopanepolyols (BHPs) are pentacyclic triterpenoids [Rohmer, 1993] biosynthesized as membrane lipids by a wide range of bacteria including cyanobacteria, nitrogen-fixing bacteria, purple nonsulfur bacteria, acetic acid bacteria, methanotrophs, and methylotrophs [e.g., Talbot et al., 2008, and references therein]. Analysis of BHPs in terrestrial samples from various locations around the world showed that four BHPs, adenosylhopane (Ic; see Appendix A), 2-methyladenosylhopane (IIc), and the two related structures with an as yet undetermined terminal group structure termed “adenosylhopane-type 1” (Id) and “2-methyladenosylhopane-type 1” (IId), are common components of soils and peats [Talbot and Farrimond, 2007; Cooke et al., 2008a; Redshaw et al., 2008; Xu et al., 2009; Rethemeyer et al., 2010] but have rarely been observed in lacustrine [Talbot and Farrimond, 2007] and offshore [Cooke et al., 2007] sediments. Another pair of related homologues, with a second, as yet unidentified terminal group, have recently been reported in soils from Svalbard [Rethemeyer et al., 2010]. However, their occurrence seems to be more restricted than those of adenosylhopanes and adenosylhopane-type 1 compounds. Data from surface sediments from the continental shelf off the Rhône River [Cooke et al., 2007] showed a steady decrease in the abundance of soil-specific BHPs with distance from the river mouth. Recent studies of Congo fan sediments [Cooke et al., 2008b; Handley et al., 2010] have shown evidence of the preservation of these compounds in sediments associated with transport of soil OM from the Congo up to ∼100 mbsf and ∼1.2 Ma. These promising initial results indicate the potential of soil-specific BHPs to serve as an indicator for soil OM input from land to the ocean. However, the use of the new soil-specific BHPs has not been widely tested yet as a robust proxy for soil OM input in various environmental settings.

Here, we investigated soils collected in the Têt watershed (southern France) and determined variations in BHP concentrations. For the first time, soil BHP distributions are directly compared to environmental variables including pH, precipitation, and mean annual air temperature (MAT) to determine potential controls on BHP distributions over a range of conditions occurring within a single catchment. BHPs are also compared with a well established proxy of river-borne soil OM input to the ocean, the branched and isoprenoid tetraether (BIT) index [Hopmans et al., 2004; Huguet et al., 2007; Kim et al., 2009]. The BIT index is based on a group of branched glycerol dialkyl glycerol tetraethers (GDGTs) derived from presumably anaerobic bacteria [Weijers et al., 2006a], which occur widely in soils [Weijers et al., 2006b, 2007], and a structurally related isoprenoid GDGT ‘crenarchaeol’, predominantly found in marine planktonic Crenarchaeota [Sinninghe Damsté et al., 2002].

2. Study Area

The Têt watershed, a typical Mediterranean river system, is located in the southern part of France (Figure 1). The uppermost part of the Têt watershed is situated in elevated regions with steep slopes [Garcia-Esteves et al., 2007]. In general, soils are thin in the upstream area of the Têt River. They are mostly cambisols with the vegetation essentially composed of pasture grass. Peats also occur in some places in the upper Têt watershed. The surrounding forest is mainly composed of beech as well as Douglas and Laricio pine trees. Further downstream, intensive agricultural land use becomes dominant in the form of orchards and vineyards.

Figure 1.

Map showing soil sampling sites in the Têt watershed. Open circles indicate the sites where soil depth profiles were studied. Numbers correspond to sample code TESO numbers in Table 1.

3. Material and Methods

3.1. Sample Collection and Preparation

Sampling of Têt soils (TESO) was carried out in June and July 2007 along the Têt River, from the source area in the Pyrenees to the river mouth into the Gulf of Lions (Figure 1 and Table 1). In total, 29 samples including 2 peats (TESO2 and TESO 49) were collected from 15 sampling sites, including 14 topsoils (i.e., upper 10 cm soils) and 3 soil profiles (TESO2, TESO5, and TESO36). All samples were immediately frozen with dry ice in the field and subsequently stored at −40°C. The soil samples were freeze-dried, sieved (<2 mm), and ground with a swing mill to obtain homogeneous material for further geochemical analysis.

Table 1. Sampling Sites, Environmental Data, and Results of Bulk Geochemical, Branched GDGT, and BHP Analyses From the Investigated Têt Soil Samplesa
Sample CodeMaterialSampling Depth (cm)Longitude (E)Latitude (N)Altitude (m)Sampling Date dd/mm/yyyyMATb (°C)Precipitationb (mm)Soil pHbTN (wt. %)δ15N (‰)Branched GDGTb (μg gTOC−1)BIT IndexbTotal BHPs (μg gTOC−1)Soil-Specific BHPs (μg gTOC−1)Soil-Specific BHPs (%)BHPsc
  • a

    MAT indicates mean annual air temperature. Topsoil indicates the upper 10 cm.

  • b

    Data from Kim et al. [2010].

  • c

    Number of BHP structures identified in each sample.

TESO17Soilupper 102.40242.60743712/06/200713.85998.40.46.320.34220474
TESO19Soilupper 102.60642.67415112/06/200714.26337.90.11.540.7408266658
TESO32Soilupper 102.79242.6886013/06/200715.36346.80.36.3171.036557167
TESO35Soilupper 103.03942.714014/06/200715.45778.90.06.750.6353003
TESO39Soilupper 102.77442.7067414/06/200715.26197.50.15.930.611963539
TESO41Soilupper 102.89242.7032514/06/200715.35557.60.14.830.8332194588
TESO47Soilupper 102.30742.55783115/06/200710.55497.

3.2. Bulk Parameter Analyses

Total nitrogen content (TN, Table 1) was analyzed with a Thermo Elemental Analyzer Flash EA 1112 at Royal Netherlands Institute for Sea Research (NIOZ) at least in duplicate. The analytical error is on average 0.01% for TN contents. For total nitrogen stable isotope composition (δ15N, Table 1), bulk soil samples were analyzed using a Flash EA 1112 Elemental Analyzer interfaced with a ThermoFinnigan DeltaPlus mass spectrometer at NIOZ. All samples were analyzed in triplicate. The analytical error is less than ±0.5 ‰.

3.3. BHP Analysis

All the soil samples investigated were extracted at NIOZ. The extraction method is similar to that reported by Talbot et al. [2008], which is based on the Kates modification of the original Bligh and Dyer extraction [Bligh and Dyer, 1959], except for the use of dichloromethane (DCM) instead of chloroform. An aliquot of the total lipid extract was derivatised by heating with acetic anhydride and pyridine (4 ml; 1:1 v/v] at 50°C for 1 h and leaving at room temperature overnight. The derivatised extract was rotary evaporated to near dryness, transferred to a vial using DCM, blown down to dryness under N2, and redissolved in methanol/propan-2-ol (60:40, v/v] for liquid chromatography tandem mass spectrometry (LC-MSn) analysis. Analysis was performed as described previously [Cooke et al., 2008a] using a high-performance liquid chromatography (HPLC) and detection via a Thermo Finnigan LCQ ion trap mass spectrometer equipped with an atmospheric pressure chemical ionization (APCI) source operated in positive ion mode. Structures were assigned from comparison with published spectra where possible [Talbot et al., 2003a, 2003b, 2007a, 2007b, 2008]. A semiquantitative estimate of BHP abundance was calculated from the characteristic base peak areas of individual BHPs in mass chromatograms relative to the m/z 345 ([M + H-CH3COOH]+) base peak area response of the acetylated 5α-pregnane-3β,20β-diol internal standard added prior to derivatisation. Averaged relative response factors (from a suite of five acetylated BHP standards) were used to adjust the BHP peak areas relative to that of the internal standard where BHPs containing one or more N atoms give an averaged response approximately 12 times that of the standard and compounds with no N atoms give a response approximately 8 times that of the standard.

3.4. Statistical Analysis

The relationships between environmental and geochemical parameters were assessed using redundancy analysis (RDA), using the Brodgar v.2.5.2 (www.brodgar.com) software package. Multiple colinearity between environmental variables was examined using variance inflation factors (VIFs). Large VIFs (>50) indicate that a variable is highly correlated with other variables, and thus contributes little information to the ordination. Preliminary ordinations revealed that altitude had a high VIF value. Therefore, the altitude was excluded in the final redundancy analysis. However, statistical analysis including altitude yielded similar results (data not shown).

4. Results

The TOC and TN contents varied between 0.1 and 33 wt. % and between < 0.1 and 1.7 wt. %, respectively, with higher values found in peats (Figure 2a and Table 1). TN contents were only significant for soils near the source of the Têt River (Figure 3). The δ15N values of the soils ranged from 0.7 to 7.6 ‰ with distinctively depleted δ15N values in peats (Figure 2c).

Figure 2.

Comparison of the concentration of the soil-specific BHPs in μg gTOC−1 with (a) TOC in wt. % [Kim et al., 2010], (b) TN in wt. %, (c) δ15N in ‰, (d) total BHPs in μg gTOC−1, and (e) soil-specific BHPs in %. Red diamonds and black circles indicate peat and soil samples, respectively.

Figure 3.

Bulk geochemical (TN in wt. % and δ15N in ‰) and BHP parameters of soil depth profiles collected from the source area and the river mouth of the Têt River: (a) TESO2, (b) TESO5, and (c) TESO36. Red diamonds and black circles indicate peat and soil samples, respectively.

In total, 19 different BHP structures were identified in the investigated peat and soil samples (Table 2). The concentration of total BHPs ranged from 42 to 2289 μg gTOC−1 (Figure 2d). Bacteriohopanetetrol (BHT; Ia) and aminotriol (Ib) were detected in all the samples (Table 3). BHT cyclitol ether (Ih) was also common but not detected in all the samples and was only observed in the surface sample at site TESO36. These source-unspecific BHPs (Table 2) contributed 26 to 92% of total BHPs (Figures 4 and 5).

Figure 4.

Distributions of individual or grouped BHPs relative to the total BHPs in % for topsoils. Cyanobacterial BHPs include 2-methyl BHT (IIa), 2-methyl aminotriol (IIb), unsaturated BHT pentose (IIIg or IVg), BHT pentose (Ig), and 2-methyl BHT pentose (IIg). Soil-specific BHPs include adenosylhopane (Ic), 2-methyl adenosylhopane (IIc), adenosylhopane-type (Id), and 2-methyl adenosylhopane-type (IId). Methanotrophic BHPs are the group of aminotetrol (Ie) and aminopentol (If). Numbers on the x axis correspond to TESO numbers in Table 1.

Figure 5.

Distributions of individual or grouped BHPs relative to the total BHPs in % for soil depth profiles: (a) TESO2, (b) TESO5, and (c) TESO36.

Table 2. BHPs Identified in the Investigated Têt Soil Samplesa
Abbreviated NameStructurebBase Peak m/zKnown Source OrganismsReferencesc
BHTIa655various1–4, 6–8, 10, 11, 17, 19–21, 23–25, 27–29
2-methyl BHTIIa669Cyanobacteria, Rhodopseudomonas palustris2, 23, 29, 32
Unsaturated aminotriolIIIb or IVb712Rhodopseudomonas palustris27
AminotriolIb714various3–5, 7, 12, 13, 15, 16, 18, 20, 22, 25, 27, 28
2-methyl aminotriolIIb728Cyanobacteria31
AdenosylhopaneIc746Purple nonsulfur bacteria, Nitrosomonas europea, Bradyrhizobium japonicum4, 7, 14, 16, 27, 22
2-methyl adenosylhopaneIIc760Bradyrhyzobium japonicum27
Adenosylhopane-type 1Id761Purple nonsulfur bacteria27
AminotetrolIe772Methanotrophs, Desulfovibrio sp.3, 12, 13, 25, 30
2-methyl adenosylhopane-type 1IId775none 
AminopentolIf830Type I methanotrophs5, 13, 25
Unsaturated BHT pentoseIIIg or IVg941Cyanobacteria31
BHT pentoseIg943Cyanobacteria31
2-methyl BHT pentoseIIg957Cyanobacteria31
BHT cyclitol etherIh1002 (C)various6–11, 16, 19, 26–28
BHT glucosamineIi1002 (G)various8, 11, 19
BHpentol cyclitol etherIj1060various10, 26–28
BHhexol cyclitol etherIk1118none 
2-methylBHhexol cyclitol etherIIk1132none 
Table 3. Individual BHP Data From the Têt Watershed in μg gTOC−1a
Sample SiteMaterialSampling Depth (cm)IaIIaIIIb or IVbIbIIbIcIIcIdIIdIeIfIIIg or IVgIgIIgIhIiIjIkIIk
  • a

    MAT indicates mean annual air temperature. Roman numerals refer to BHP structures shown in Appendix A and “-” means not detected.

TESO17Soilupper 1019 -3-16-4-----------
TESO19Soilupper 1011012-16-1301480424---------
TESO32Soilupper 1016019-25-57---813---83----
TESO35Soilupper 1028030-43---------------
TESO39Soilupper 1038--12-46-1163----2-11-
TESO41Soilupper 108414-30-1109601510---------
TESO47Soilupper 105018-14-8087010-----33---

Usually, 1 to 4 of the “soil-marker” BHPs were present in all analyzed samples, except for the topsoil (TESO35) collected near to the river mouth. The concentration of soil-specific BHPs varied between 0 and 589 μg gTOC−1 (Figure 4b). The proportion of soil-specific BHPs relative to total BHPs reached up to 66% (Figure 4e). In general, the concentrations of total and soil-specific BHPs were higher in the upper layers than in the deeper layers down the soil profiles (Figure 3); however, the proportion of soil-specific BHPs relative to the total BHPs did not clearly follow those of the concentrations of total and soil-specific BHPs.

5. Discussion

5.1. Occurrence and Abundance of Soil-Specific BHPs

As observed in previous studies of soils [Cooke et al., 2008a; Xu et al., 2009; Rethemeyer et al., 2010], BHPs in Têt soils are dominated by non-source-specific BHT, aminotriol, and BHT cyclitol ether (Figure 4). In contrast, contributions of BHPs from methanotrophic and likely phototrophic sources, including cyanobacteria and purple nonsulfur bacteria (Figure 4) are relatively minor. Despite their low abundances, cyanobacterial BHPs are commonly found in Têt peats and soils, but absent in deeper soil depths at the river mouth. This is in good agreement with recent studies, showing that the diversity of organisms capable of BHP biosynthesis is greater in terrestrial rather than marine influenced systems [e.g., Pearson et al., 2009] and more specifically that BHP production in marine cyanobacteria seems to be uncommon [Pearson et al., 2007; Talbot et al., 2008].

The methanotrophic bacteria markers aminotetrol (Ie) and aminopentol (If) were only observed at a small number of sites (Table 3) with the most diagnostic structure (aminopentol) [Cvejic et al., 2000a] only present at 4 sites including the upper layers of the peat core (TESO2) and in topsoil TESO32. It was also observed in the deepest layer of site (TESO36) where it likely reflects preservation of a fossilized signature as it indicates an aerobic process which is unlikely to occur only in the deeper and anoxic part of the soil profiles.

The soil-specific BHPs (adenosylhopane, adenosylhopane-type 1, 2-methyl adenosylhopane, and 2-methyl adenosylhopane-type 1) are common in all the samples but the topsoil (TESO35) collected near the river mouth. Strikingly, the concentrations of soil-specific BHPs are highest in peats with depleted δ15N values (Figure 2). The depleted δ15N values in peats therefore suggest that soil-specific BHP production is favored by the growth of N2-fixing bacteria. Adenosylhopane (Ic) and its methylated homologue (IIc) are, for instance, found in the nitrogen-fixing bacterium Bradyrhizobium japonicum [Bravo et al., 2001; Talbot et al., 2007a]. Therefore, the Têt data suggest that B. japonicum and/or yet other unknown N2-fixing bacteria are one of the major sources of soil-marker BHPs in acidic environments.

5.2. Environmental Factors Controlling Soil-Specific BHP Productions

In order to assess which environmental factors determine the concentrations of total and soil-specific BHPs normalized to TOC, the number of BHP compounds detected, and the percentage of the soil-specific BHPs relative to the total BHPs in the Têt watershed, we applied redundancy analysis on the data acquired. The variables pH, MAT, and precipitation (Table 1) explain 41% of the variation in the response (BHP parameters) variables (Table 4). Conditional effects (i.e., increase total sum of eigenvalues after including new variable) (Table 4) indicate that soil pH is the most important environmental factor, influencing the response variables and thus the production and occurrence of total and soil-specific BHPs. This response is in good agreement with recent culture studies showing that BHPs, 2-methyl compounds in particular, were produced in greater concentration by photosynthetic bacteria at lower pH [Doughty et al., 2009; Welander et al., 2009]. Therefore, our results support previous findings that BHPs are important in protecting cells from external stresses, such as pH. The pH and BHP abundance relationship is also similar to that of branched GDGT production [Kim et al., 2010]. In general, the concentrations of total and soil-specific BHPs as well as the diversity of BHP producing organisms are higher in the Têt catchment when soil pH values are lower (Figures 6a–6c). However, there is no apparent relationship between the percentage of the soil-specific BHPs and soil pH values (Figure 6d).

Figure 6.

Scatterplots of soil pH values with (a) the concentration of total BHPs in μg gTOC−1, (b) the concentration of soil-specific BHPs in μg gTOC−1, (c) the numbers of BHP compounds detected, and (d) the percentage of the soil-specific BHPs relative to the total BHPs in % (see also Table 3). Red diamonds and black circles indicate peat and soil samples, respectively.

Table 4. Numerical Output of a Redundancy Analysis Applied to the BHP Dataa and Conditional Effectsb
Axisλλ as Cumulative %λ as Cumulative % of Sum of All Carnonical Eigenvalues
  • a

    The sum of all canonical eigenvalues is 0.42 and the total variance is 1 (the variation explained by the first two axes). Eigenvalue (λ) is the standard deviation of the scores.

  • b

    Eigenvalue indicates the increase in explained variation due to adding an extra explanatory variable. Response variables used are the concentration of total BHPs, the concentration of soil-specific BHPs, the percentage of the soil-specific BHPs relative to the total BHPs, and the numbers of BHP compounds detected. Explanatory variables used are soil pH, mean annual air temperature (MAT), and precipitation. Significance level, p < 0.05. Conditional effects are total sum of eigenvalues after including new explanatory variable.

OrderExplanatory VariableConditional Effectsp Value

5.3. Comparison of Soil-Specific BHPs With Branched GDGTs

The topsoil of TESO35 influenced by seawater contains virtually no soil-specific BHPs. Furthermore, the concentration of soil-specific BHPs from the TESO36 soil profile near the Têt River mouth are significantly lower than those from the TESO2 and TESO5 soil profiles upstream of the Têt River. Accordingly, our data generally support the initial hypothesis that soil-specific BHPs can be a useful tool to identify the transport of soil OM to marine sediments [Talbot and Farrimond, 2007; Cooke et al., 2008b; Rethemeyer et al., 2010].

In order to further explore the potential of soil-specific BHPs as a soil OM tracer, we compared soil-specific BHPs with branched GDGT parameters, which are more established indicators of soil OM input from land to the ocean. The comparison of the soil-specific BHP concentrations with the branched GDGT concentrations (Figure 7a) does not show significant relationships for both peats (r2 = 0.5, p = 0.3) and soils (r2 = 0.06, p = 0.24). The comparison of BIT values with the soil-specific BHP concentrations (r2 = 0.3, p = 0.003, Figure 7b) and the percentage of soil-specific BHPs relative to total BHPs (r2 = 0.2, p = 0.03, Figure 7c) also show no significant correlations. Although more data from different environmental settings and statistical evaluations are clearly needed to validate the absence of relationship observed in the Têt River, our results imply that both biomarkers are produced by different organisms.

Figure 7.

Comparisons of soil-specific BHPs and branched GDGTs: (a) the concentration of soil-specific BHPs in μg gTOC−1 versus the summed concentration of branched GDGTs in μg gTOC−1, (b) the concentration of soil-specific BHPs in μg gTOC−1 versus BIT values, and (c) the percentage of the soil-specific BHPs relative to the total BHPs in % versus BIT values. Red diamonds and black circles indicate peat and soil samples, respectively. The branched GDGT and BIT data were previously reported by Kim et al. [2010].

Up to now, only planctomycetes, performing the anaerobic oxidation of ammonium [Sinninghe Damsté et al., 2004], Geobacter species [Fischer et al., 2005], and sulfate-reducing bacteria of the genus Desulfovibrio [Blumenberg et al., 2006] have been shown to produce BHPs under strictly anoxic conditions. Therefore, it is consistent that both total and soil-specific BHPs are more abundant in the aerated zones than in the deeper anaerobic parts of soil profiles (Figure 3). This vertical distribution pattern is, to some degree, inverse to that of the branched GDGTs, which show higher abundances in deeper parts of the soil profiles [Kim et al., 2010] and peats [Weijers et al., 2007]. This is also in agreement with previous studies, suggesting that branched GDGTs are produced by anaerobic bacteria [Weijers et al., 2006a]. Thus, soil-specific BHPs and branched GDGTs are likely produced by bacteria living in different ecological niches of the soil profile.

6. Conclusions

BHP distributions have been analyzed from a range of soils and peats from 14 sites in the catchment of the River Têt. In all samples but one (TESO35) collected near the Têt River mouth, the previously proposed soil-marker BHPs (adenosylhopane and up to three structurally related compounds) were detected in proportions up to 60% relative to total BHPs, significantly higher than the previously reported global mean for soils of 28% [Cooke et al., 2008b] although higher values have been reported in both Canadian soils [Xu et al., 2009] and soils from Svalbard [Rethemeyer et al., 2010]. Concentrations of soil-specific BHPs in peat environments have been shown to increase with lower δ15N values, suggesting that N2-fixing bacteria are one of the major sources of BHPs including soil-marker BHPs in peats. Soil pH is also a major factor controlling the diversity and concentration of total and soil-marker BHPs similar to previous observations for branched GDGTs and in agreement with laboratory culture studies looking at BHP biosynthesis under changing pH regimes. Further studies are clearly required to determine what, if any other environmental factors, determine BHP diversity, concentration and preservation including e.g., temperature, a wider range of pH values, salinity, and pressure such as pCO2.

The concentrations of soil-specific BHPs from the soil profile near the Têt River mouth are significantly lower than those of the soil profiles upstream of the Têt River. This supports the initial hypothesis that soil-specific BHPs can be useful biomarkers for tracking the input of soil OM from land to aquatic environments [Talbot and Farrimond, 2007]. However, there is no clear correlation between the percentage of soil-specific BHPs relative to total BHPs and the BIT index likely due to the different ecological niches occupied by the major sources organisms (aerobic topsoil versus deeper anaerobic horizons). More work is needed to extend our limited knowledge on production and occurrence of both soil-specific biomarkers in various environmental settings.

Appendix A

BHP structures in the soil samples are investigated in this study (Figure A1). The stereochemistry indicated was previously determined from NMR studies, but other configurations may be possible.

Figure A1.

Structures of BHPs found in soil samples from the Têt watershed.


We thank E. Hopmans, S. Crayford, J. Ossebaar, and M. Kienhuis for analytical support. F. Peterse is acknowledged for her help with the Têt soil sampling. We are grateful to S. Schouten and J. S. Sinninghe Damsté for supporting this study and two anonymous reviewers and the associate editor for their constructive comments. This study was supported by a Marie Curie IntraEuropean Fellowship grant to J.-H. Kim. The Science Research Infrastructure Fund (SRIF) from HEFCE for funding the purchase of the Thermo Electron Finnigan LCQ ion trap mass spectrometer. Thomas Wagner acknowledges the Royal Society–Wolfson Research Merit Award.