Volume 13, Issue 1 p. 40-52
Comparative and Intercalibration Studies
Free Access

Toxic effects of lab-grade butyl rubber stoppers on aerobic methane oxidation

Helge Niemann

Corresponding Author

Helge Niemann

Department of Environmental Sciences, University of Basel, Switzerland

Corresspondence: [email protected]Search for more papers by this author
Lea Steinle

Lea Steinle

Department of Environmental Sciences, University of Basel, Switzerland

GEOMAR Helmholtz Centre for Ocean Research Kiel, Department of Marine Biogeochemistry, Germany

Search for more papers by this author
Jan Blees

Jan Blees

Department of Environmental Sciences, University of Basel, Switzerland

Search for more papers by this author
Ingeborg Bussmann

Ingeborg Bussmann

Alfred Wegenener Institute for Polar and Marine Research, Marine Station Helgoland, Germany

Search for more papers by this author
Tina Treude

Tina Treude

GEOMAR Helmholtz Centre for Ocean Research Kiel, Department of Marine Biogeochemistry, Germany

Search for more papers by this author
Stefan Krause

Stefan Krause

GEOMAR Helmholtz Centre for Ocean Research Kiel, Department of Marine Biogeochemistry, Germany

Search for more papers by this author
Marcus Elvert

Marcus Elvert

MARUM Center for Marine Environmental Sciences and Department of Geosciences, University of Bremen, Germany

Search for more papers by this author
Moritz F. Lehmann

Moritz F. Lehmann

Department of Environmental Sciences, University of Basel, Switzerland

Search for more papers by this author
First published: 28 January 2015
Citations: 32

Abstract

Methods for measuring aerobic methane oxidation (MOx) rates in aquatic environments are often based on the incubation of water samples, during which the consumption of methane (CH4) is monitored. Typically, incubation vessels are sealed with butyl rubber because these elastomers are essentially impermeable for gases. We report on the potential toxicity of five different commercially available, lab-grade butyl stoppers on MOx activity in samples from marine and lacustrine environments. MOx rates in incubations sealed with non-halogenated butyl were > 50% lower compared to parallel incubations with halogenated butyl rubber stoppers, suggesting toxic effects associated with the use of the non-halogenated butyl type. Aqueous extracts of non-halogenated butyl rubber were contaminated with high amounts of various organic compounds including potential bactericides such as benzyltoluenes and phenylalkanes. Comparably small amounts of organic contaminants were liberated from the halogenated butyl rubber stoppers but only two halogenated stopper types were found that did not seem to leach any organics into the incubation medium. Furthermore, the non-halogenated and two types of the halogenated butyl elastomers additionally leached comparably high amounts of zinc. While the source of the apparent toxicity with the use of the non-halogenated rubber stoppers remains elusive, our results indicate that leaching of contaminants from some butyl rubber stoppers can severely interfere with the activity of MOx communities, highlighting the importance of testing rubber stoppers for their respective contamination potential. The impact of leachates from butyl rubber on the assessment of biogeochemical reaction rates other than MOx seems likely but needs to be verified.

Elastomers are (semi) synthetic polymers with numerous applications in all aspects of modern life and are used in many types of sampling equipment. The lids of most Niskin Bottle-type water samplers, for instance, are sealed with rubber O-rings. While the use of synthetic polymers has innumerable (and very obvious) advantages, direct or indirect contact of these materials with living biological samples can be problematic. The release of potentially toxic substances from medical hardware including rubber (e.g., instruments, tubing, gloves) for example, has been a major concern in medical research (Ball et al. 2012). In the wider field of environmental/aquatic sciences, recent studies typically focused on aspects related to environmental pollution, for example, the leaching of toxic components from tires and their liberation and accumulation in aquatic or terrestrial ecosystems (Stephensen et al. 2003; Sheehan et al. 2006; Wik 2007). Systematic investigations of the potential toxicity (i.e., harmfulness to organisms with respect to metabolic activity, growth or life) of rubber used in experimental setups are scarce. An early report by DeWitt (1957) showed elevated mortality of aquatic insects and fish in an experimental setup, in which water was circulated through latex tubing. Similarly, Hubschman and Engel (1965) found that waters in contact with black rubber stoppers (not further specified) were toxic to crustaceans. Analogous effects were also observed in more recent studies, for instance the inhibition of plankton productivity related to utilization of rubber parts in a Niskin Bottle (Williams and Robertson 1989), or toxic effects of latex-, tygon- (Price et al. 1986), and rubber (Galbraith and Burns 1997) tubing on plankton and bacteria. The nature of the observed toxicity remained mostly unknown, although many studies speculated that leaching of organic and inorganic compounds could be responsible for the observed inhibiting/lethal effects.

Aqueous samples used for the analysis of volatile solutes and/or incubation experiments often require a gas-tight closure that allows further manipulation of the closed/sealed sample, for instance injection of fixatives or metabolic substrates. Microbial rates of aerobic methane oxidation (MOx) in aquatic environments, for example, are best determined by measuring the turnover of radio-labeled CH4-tracer or the decrease of CH4 concentrations during ex situ incubations, which are typically carried out in crimp-top sealed vials (Fig. 1a) (Reeburgh 2007; Mau et al. 2013; Blees et al. 2014a).

Details are in the caption following the image
(a) A typical setup for measuring rates of aerobic methanotrophy in the water column: water samples are crimp-top sealed with butyl rubber stoppers in glass vials. Trace amounts of radio-labeled CH4 (C3H4; 14CH4) are subsequently injected into the vial and the rate of methanotrophy is estimated from the fractional turnover of labeled substrate. Commercially available butyl rubber stoppers tested in this study were (b) black, non-halogenated butyl, (c) black bromobutyl, (d) blue chlorobutyl, (e) gray chlorobutyl (with and without PTFE coating), and (f) gray bromobutyl rubber. The texture of halogenated butyl is more rigid compared to regular butyl stoppers, which made injections and general handling of the thick blue chloro- and black bromobutyl variants cumbersome.

Butyl rubber, a copolymer of 2-methyl-1-propene and 2-methyl-1,3-butadiene (isoprene) with different levels of curing (comprising non-halogenated and halogenated butyl matrices), has a low permeability for gases (Lundberg et al. 1969) and is comparably flexible. These characteristics make butyl rubber the commonly used material for manufacturing gas-tight seals and stoppers used for laboratory and field sampling as well as for incubation experiments. However, during our recent research projects on CH4 oxidation in marine and lacustrine settings, we observed that MOx seemed to be inhibited when using certain commercially available, lab-grade butyl rubber stoppers. Toxic effects of the butyl seals would undoubtedly compromise MOx rate measurements. Biased rate estimates are futile for reliable budgets, which are important for determining the contribution of different environments to the atmospheric pool where CH4 acts as a potent greenhouse gas.

The aim of this study was (1) to test the suitability of different lab-grade butyl rubber stoppers for measuring MOx rates in aquatic environments and (2) to identify leachables from the elastomers that are potentially toxic for microbes mediating MOx. Our findings raise important concerns regarding the pervasiveness of organic and inorganic compound-leaching from lab-grade rubber stoppers, as well as the extent to which toxic effects of butyl rubber stoppers may have affected MOx rate estimates in previous studies.

Material and Procedures

Butyl rubber stoppers

We tested five types of commercially available, lab-grade butyl rubber stoppers (Fig. 1b–f): black, non-halogenated butyl (Rubber B. V., The Netherlands, part No.: 7395; Fig. 1b), black bromobutyl (Glasgerätebau Ochs, Germany, part No.: 102049; Fig. 1c), blue chlorobutyl (Glasgerätebau Ochs, Germany, no part No. available; Fig. 1d), polytetrafluoroethylene (PTFE) coated gray chlorobutyl (Wheaton, USA, part No.: 224100-175; Fig. 1e) and gray bromobutyl rubber (Helvoet Pharma, Belgium (now Dätwyler Holding), part No.: V9230; Fig. 1f). The PTFE coating of the gray chlorobutyl stoppers was pierced during injections facilitating contact of the incubation medium with the rubber matrix. Hence, we removed the coating from one set of stoppers and treated them as an additional stopper type. Prior to ex situ incubations, all stoppers were washed > 5 times in a dishwasher with laboratory dishwasher detergent and subsequently rinsed with deionized (DI) water.

Many labs subject stoppers to additional pretreatments (e.g., boiling in base or acid) to further clean them prior to usage. We tested the effect of several pretreatments on ex situ MOx activity by comparing rates from incubations sealed with untreated and pretreated stoppers. We applied the common methods of soaking stoppers (12 h) in aqueous oxalic acid solution (ox. A; 3%, w/v) followed by 2 × autoclaving and boiling stoppers (2 × 2 h) in and aqueous sodium hydroxide solution (NaOH; 20%, w/v). Furthermore, we also tested the effects of pretreating stoppers with an organic solvent by boiling stoppers (2 × 2 h) in aqueous methanol solution (MeOH; 20%, v/v).

Sampling

Water samples were retrieved from three marine and one lacustrine setting (Table 1). All environments were sampled with 5 or 10 L Niskin Bottles, and aliquots were taken directly from the Niskin bottle through silicon tubing. Samples were collected in 20 mL glass serum vials, which were crimp-top sealed bubble free with the different butyl rubber stoppers (n ≥ 4; subsamples per stopper type). The Eckernförde Bay (E-Bay) was sampled during five cruises (autumn 2012, spring 2013, summer 2013, spring 2014, and summer 2014) (Table 2). Ex situ incubations were setup with non-halogenated (n = 20) and gray bromobutyl stoppers (n = 20) during all sampling campaigns as well as with gray chlorobutyl stoppers (±PTFE) in summer 2013 and 2014 (n = 8) and blue chlorobutyl stoppers in spring and summer 2014 (n = 8). The harbor of Helgoland was sampled in spring 2013 (gray bromobutyl stoppers, n = 4) and in summer 2014 (gray- and black bromobutyl stoppers, n = 7). During an expedition in summer 2012, the water column above the cold seeps West of Svalbard was sampled three times (with a temporal gap of six days between sampling events) and ex situ incubations were setup with non-halogenated (n = 12) and gray bromobutyl stoppers (n = 12). Lake Lugano was sampled once in autumn 2012 and water samples were incubated with gray bromobutyl (n = 4), blue chlorobutyl (n = 4), and non-halogenated butyl stoppers (n = 4).

Table 1. Environments sampled for comparison of MOx rates. Eckernförde Bay (E-Bay) is a shallow coastal inlet of the Baltic Sea in northern Germany and is characterized by seasonal hypoxia/anoxia of bottom waters. The sea floor at the E-Bay sampling site features gassy sediments from which CH4 is transported by diffusion or advection (gas bubbles) into the water column (Bange et al. 2010). The harbor of the central North Sea Island Helgoland contains relatively high CH4 contents in the water column. The Svalbard sampling site is located West of the Svalbard Archipelago at the shelf break where numerous gas flares emanate from the sea floor leading to elevated CH4 concentrations in the water column (Berndt et al. 2014). Lake Lugano is a deep, narrow and meromictic south-alpine lake with high contents of CH4 at the redox interface (Blees et al. 2014a).
Environment Lat (N) Lon (E) Water depth (m) Sampling depth (m) Habitat characteristics in situ CH4 (nM)
Eckernförde Bay 54.53° 10.05° 28 15 hypoxic, marine bay waters 70–120
Helgoland Harbor 54.18° 7.88° 4 1 oxic, marine habour water ∼100
Svalbard 78.56° 9.48° 385 350 oxic, marine water column 50–325
Lake Lugano 46.01° 9.02° 288 140 hypoxic lake waters at redox cline ∼ 2000
Table 2. Sampling campaigns for MOx rate measurements with untreated and pretreated butyl stoppers. Sampling dates are indicated as season (spring–winter; 1–4) and year (2012–2014; 12–14) (e.g., summer 2013 = 2–13).
Environment
Stopper type (pretreatment) E-Bay H. Harbor Svalbard L. Lugano
gray Br-butyl 3–12, 1–13, 2–13, 1–14, 2–14 1–13, 2–14 2–12 4–12
gray Br-butyl (NaOH) 2–14
gray Br-butyl (MeOH) 2–14
gray Br-butyl (ox. A) 2–14
gray Cl-butyl (± PTFE) 2–13, 2–14
gray Clr-butyl (± PTFE, NaOH) 2–14
gray Cl-butyl (± PTFE, MeOH) 2–14
gray Cl-butyl (± PTFE, ox. acid) 2–14
blue Cl-butyl 1–14, 2–14 4–12
blue Clr-butyl (NaOH) 2–14
blue Cl-butyl (MeOH) 2–14
blue Cl-butyl (ox. A) 2–14
black Br-butyl 2–14
black Br-butyl (NaOH) 2–14
black Br-butyl (MeOH) 2–14
black Br-butyl (ox. A) 2–13 1–13, 2–14
non-halo. butyl 3–12, 1–13, 2–13, 1–14, 2–14 2–12 4–12
non-halo. butyl (NaOH) 1–13, 2–14
non-halo. butyl (MeOH) 1–13, 2–14
non-halo. butyl (ox. A) 2–14

The performance of pretreated stoppers was tested in E-Bay in spring 2013 (NaOH-, MeOH-pretreatment, non-halogenated butyl stoppers, n = 4), summer 2013 (oxalic acid pretreatment, black bromobutyl stoppers, n = 4), and summer 2014 (all pretreatments, all stopper types [except black bromobutyl stoppers], n = 4) (Table 2). In the harbor of Helgoland, we performed tests on pretreated stoppers in spring 2013 (oxalic acid pretreatment, black bromobutyl stoppers n = 7) and summer 2014 (all pretreatments, black bromobutyl stoppers).

General handling of the non-halogenated butyl stoppers (Fig. 1b) was most convenient because the elastomer is relatively soft allowing for easy and tight sealing of the incubation vial as well as easy injection. Conversely, the halogenated butyl elastomers are rather stiff, which made handling comparably cumbersome, particularly injections through the thick black bromobutyl (Fig 1c) and blue chlorobutyl stoppers (Fig. 1d).

Rate measurements

MOx rates were determined with radiotracer assays (Reeburgh 2007) as specified in previous publications with 3H- (Berndt et al. 2014) or 14C-labelled CH4-tracer (Mau et al. 2013; Blees et al. 2014a,b). In short, 10 μL of tracer was added to lacustrine and marine samples, respectively, by injection through the rubber stoppers, and the samples were incubated ex situ (at in situ T in the dark) for one (lake samples) or three days (marine samples). MOx rates (rMOx) were calculated assuming first-order kinetics (Reeburgh 2007):
(1)
where k is the first-order rate constant (determined from the fractional turnover of labeled CH4 per unit time and corrected for tracer turnover in killed controls) and [CH4] is the concentration of CH4 at the beginning of the incubation.

We measured MOx rates with the gray bromobutyl stoppers at all study sites, while tests with other stoppers were only performed in selected environments. To account for variations in MOx rates between the different environments, and for the purpose of comparison between different stopper types, we normalized all MOx rates to those from incubations sealed with the gray bromobutyl stoppers. For this approach, rates obtained with a given stopper from repeated sampling campaigns in the same environment were pooled. MOx rate measured with gray bromobutyl stoppers were normalized (i.e., transformed to a value of 1). Using the same conversion factor, mean rates obtained from parallel incubations with other stopper types were transformed accordingly and differences of the normalized rates were analyzed statistically. Differences were considered significant at p-values ≤ 0.05. With a two-tailed t-test, we tested for differences between rates obtained with the gray bromobutyl seals and those obtained with the other stopper types from the same environment (for statistics, NaOH, MeOH, and oxalic acid pretreatments were considered as additional stopper types). Furthermore, we tested for environmental effects on the performance of the butyl stoppers with a one way analysis of variance (ANOVA) by comparing rates obtained with a given stopper type from different environments.

Extraction and analyses of organic contaminants

We investigated leaching of organic contaminants from the butyl stoppers into the incubation medium by gas chromatographic and mass spectrometric analyses of aqueous extracts of the different elastomers. The extracts were prepared by adding 22 of the thick black and blue stoppers, or 32 of the thin, gray stoppers for 2 h into 1 L of boiling, deionised (DI) water (high-T/short-t extracts). The wetted surface area of 22 thick, or 32 thin, stoppers is about 255 cm2 L−1, which is roughly fivefold higher compared to the wetted stopper surface area per volume of media (1.13 cm2 20 mL−1) during incubations for MOx rate measurements. We also investigated leaching of organic compounds from the non-halogenated butyl elastomer at temperatures and contact time periods (elastomer–incubation medium) that are similar to the actual conditions during incubation experiments by extracting stoppers for 48 h at ∼ 20°C (low-T/long-t extracts). Finally, we tested if stopper pretreatments with NaOH or MeOH reduced leaching of organics. After boiling in NaOH or MeOH solution, the pretreated stoppers were rinsed with DI water and extracted as described above (high-T/short-t extraction). After removal of the stoppers from the aqueous solutions, organic compounds in this solution were extracted three times with 25 mL of dichloromethane (DCM) - MeOH (9 : 1; v/v). The organic phases were combined, excess solvent was evaporated, and further separation and derivatization was carried out with slight modifications to our previous works (Elvert et al. 2000; Niemann et al. 2005, 2006). Briefly, the extract was saponified and a neutral fraction was extracted with hexane. The neutral fraction was then further separated into apolar and polar fractions over a silica column (∼ 0.5 mm particle size, ∼ 60 Å pore size, 0.5 g packing) with 5 mL n-hexane - DCM (95 : 5; v/v) and 10 mL DCM - acetone (9 : 1; v/v), respectively. Prior to analyses, we derivatized alcohols in the polar fraction with bis(trimethylsilyl)trifluoroacetamide in pyridine (Klebee et al. 1966; Niemann et al. 2005). Chromatographic and mass spectrometric analyses were performed as described elsewhere (Blees et al. 2014a). Identities of acquired mass spectra were compared to data published in the scientific literature and the mass spectral library of the National Institute of Standards and Technology (NIST).

Heavy metal measurements

The concentrations of heavy metals such as zinc (Zn) were determined from 10 mL aliquots of the aqueous stopper extracts (high-T/short-time and low-T/long-t) with inductively coupled plasma optical emission spectrometry (ICP-OES; Spectro Ciros Vision) according to standard methods (Rüdel et al. 2007).

Assessment

MOx rates

Rates of aerobic MOx observed in incubations with halogenated butyl rubber stoppers were more than twofold and up to 35-fold higher compared to incubations with non-halogenated butyl rubber (Fig. 2a, see statistical values in Table 3). This discrepancy provides clear evidence for an inhibiting and possibly toxic effect of the non-halogenated butyl rubber seals on aerobic methanotrophs. However, the rather variable extent of MOx activity reduction, which was greatest in Lake Lugano incubations and substantially lower in incubations with marine waters (Fig. 2a), suggests differential impacts of the stopper material on MOx communities in different environmental settings. Theoretically, fertilization of MOx communities by the halogenated butyl elastomers could serve as an alternative explanation for the higher MOx rates in incubations sealed with halogenated butyl stoppers. Yet, results from our leaching experiment (showing low exudation of organic and inorganic compounds from halogenated butyl when compared to non-halogenated butyl rubber, see below) argue against this. Furthermore, MOx rates in incubations with PTFE-coated, gray chlorobutyl stoppers were similar to rates in incubations sealed with non-coated halobutyl stoppers further arguing against the release of compounds from halobutyl elastomers, which may act as MOx-fertilizers.

Table 3. Statistical differences of MOx rates related to stopper type (MOx rates measured in a given environment with gray bromobutyl stoppers in comparison to parallel incubations with other stopper types) and the tested environment (comparison of MOx rates measured with a given stopper type in different environments). Statistical parameters are indicated in parentheses (statistical test: p-value, statistical power at α = 0.05). Insignificant differences (p > 0.05) are indicated (—). nt = not tested. env. = environment.
Stopper type Difference to gray Br-butyl Environmental effect
gray Cl-butyl ± PTFE >(E-Bay; t-test: <0.001, >0.94) nt
blue Cl-butyl
black Br-butyl
black non-halo. butyl <<(all env.; t-test: <0.001, >0.99) +(ANOVA: <0.001, 0.97)
Details are in the caption following the image
Normalized MOx rates during (a) ex situ incubations of various marine and lacustrine water column samples sealed with different types of halogenated and one non-halogenated butyl rubber stoppers: gray bromobutyl, gray chlorobutyl (with and without PTFE coating), blue chlorobutyl, black bromobutyl and non halogenated, black butyl. Notably, rates in incubations capped with non-halogenated butyl rubber were > 50% lower compared to parallel incubations sealed with halogenated butyl. (b) Pretreatment of the butyl stoppers by boiling/soaking stoppers in NaOH or methanol, or soaking in oxalic acid did not lead to a substantial increase in MOx during ex situ incubations.

MOx activity in incubations sealed with halobutyl stoppers were comparably similar but we found small, yet significant, differences (Table 3). The mean MOx rate in incubations with gray chlorobutyl stoppers was ∼ 20% higher compared to rates obtained with gray bromobutyl stoppers. We tested the gray chlorobutyl stoppers only with E-bay water samples, so that a direct comparison to rates obtained with the gray bromobutyl stoppers in the other environments was not possible. However, the overall performance of both, gray bromobutyl and gray chlorobutyl seems quite similar. The removal of the PTFE coating of the gray chlorobutyl stoppers (i.e., facilitating direct contact of incubation medium and the elastomer) did not seem to have any effect on the measured MOx rates.

Incubations sealed with pretreated non-halogenated butyl stoppers (boiling stoppers in NaOH/MeOH or soaking them in oxalic acid) did not yield higher MOx rates when compared to incubations with untreated non-halogenated (Fig. 2b, Table 4). On the contrary, our measurements rather indicate slightly lower rates after NaOH/MeOH treatment (t-test, p = <0.016; power = >0.73). Similarly, the use of pretreated halobutyl stoppers seemed, if at all, to cause a reduction in MOx (Fig. 2b, Table 4). Particularly, incubations with oxalic acid pretreated gray chlorobutyl stoppers (±PTFE) showed substantially lower rates when compared to incubations with the untreated stoppers (31–37% lower rates). Only pretreatments of gray bromobutyl stoppers with NaOH and black bromobutyl stoppers with MeOH showed slightly positive effects on MOx rates (17% and 15 % higher rates, respectively). Nevertheless, with the exception of incubations with oxalic acid pretreated chlorobutyl stoppers (and to a lesser degree the NaOH and MeOH pretreated gray chlorobutyl- and black bromobutyl stoppers, respectively), MOx rates in incubations with pretreated halobutyl stoppers seemed quite similar when compared to incubations with untreated stoppers.

Table 4. Statistical differences of MOx rates in incubations sealed with pretreated stoppers (boiled in NaOH, methanol or soaked in oxalic acid) in comparison to their untreated counterpart. Statistical differences were determined with a t-test and statistical parameters are indicated in parentheses (p-value, statistical power at α = 0.05). Insignificant differences (p > 0.05) are indicated (—).
Stopper type (pretreatment) Difference to untreated
gray Br-butyl (NaOH) >(0.001, 0.95)
gray Br-butyl (MeOH)
gray Br-butyl (ox. A)
gray Cl-butyl+PTFE (NaOH)
gray Br-butyl+PTFE (MeOH) <(0.003, 0.96)
gray Br-butyl+PTFE (ox. A) <(0.001, 1.00)
gray Cl-butyl-PTFE (NaOH)
gray Br-butyl-PTFE (MeOH) <(0.02, 0.71)
gray Br-butyl-PTFE (ox. A) <(0.001, 0.95)
blue Cl-butyl (NaOH)
blue Cl-butyl (MeOH) <(0.001, 0.97)
blue Cl-butyl (ox. A)
black Br-butyl (NaOH)
black Br-butyl (MeOH) >(0.015, 0.75)
black Br-butyl (ox. A)
black non-halo. butyl (NaOH) <(0.001, 0.99)
black non-halo. butyl (MeOH) <(0.001, 0.99)
black non-halo. butyl (ox. A.)

In conclusion, our results provide clear evidence that the non-halogenated black butyl rubber stoppers were inferior compared to the tested halobutyl elastomers and are not suitable for assessing ex situ MOx rates in aqueous samples. Albeit less practical with respect to experimental handling, halobutyl elastomer should generally be the preferred stopper material. At this point, however, it is challenging to clearly identify the most suitable halogenated stopper type for measuring MOx activity, because mean differences of normalized rates were usually within one standard deviation (Fig. 2).

Leaching of organic and inorganic contaminants from the butyl elastomer into the incubation medium

Organic contaminants

A possible cause for the apparent reduction in MOx activity (Fig. 2) is the leaching of toxic organic substances from the butyl elastomer matrix into the incubation medium where they possibly hamper microbial activity. Our chromatographic analyses of the aqueous extracts of the different stopper types revealed clear differences in the amount and identity of extractable compounds (Fig. 3a–j). Both, the apolar (Fig. 3a) and polar fraction (Fig. 3f) of non-halogenated butyl rubber stopper extracts contained high amounts of organic contaminants. In the apolar fraction, we could identify three structural isomers of benzyltoluene (with ortho-, meta- or para-methyl substituted phenyl groups). In addition, we were able to detect several phenylalkane isomers, which comprised four isomers with a C12-, five isomers with a C13-, and six isomers with a C14-alkyl side chain. Each isomer was characterized by a specific fragment ion of either m/z 105, 119, 133, 147, 161, or 175 indicating that the structures were 2-, 3, 4, 5, or 6 phenylalkanes, respectively (King 1965). The accompanied unresolved complex mixture (UCM) also appeared to comprise phenylalkanes, as indicated by the high abundance of fragment ions at m/z 91 and 105 (King 1965; McLafferty and Tureček 1993). In the polar fraction (Fig. 3f), we found that one of the organic contaminants was a benzothiazole (2-(ethylamino)benzothiazole) but the associated mass spectra of compounds i–iii and a cluster of isomers (iv) were rather inconclusive (when comparing them to literature data and matching them against the NIST library; see Fig. 4a–d).

Details are in the caption following the image
Partial gas chromatograms of apolar (a–e) and polar fractions (f–j) of aqueous extracts (short (2h) high temperature (100°C) extractions; see material and procedure section) obtained from butyl elastomer stoppers (black non-halogenated butyl: a, f; black bromobutyl: b, g; blue chlorobutyl: c, h; gray chlorobutyl d, i; gray bromobutyl e, j). Concentrations of organic contaminants were highest in the non-halogenated and lowest in the gray chloro- and gray bromobutyl elastomers. Peak areas of the internal standards α-cholestane and n-nonadecanol (C19 : 0-ol) correspond to 5 μg L−1 extract; concentrations of the last-eluting benzyltoluene (RT ≈ 25.5 min.; panel a) and the most concentrated phenylalkane (RT ≈ 34.8 min.; panel a) are thus 0.9 μg  L−1 and 9.7 μg  L−1, respectively. Note that contaminant concentrations in microbial incubations were > fivefold lower. Identified compounds are indicated by names and/or structural formulas, and unknown compounds by roman numerals (see mass spectra of unknown compounds in Fig. 4). Several of the identified compounds are known biocides.
Details are in the caption following the image
Mass spectra of organic contaminants found in the polar fractions of aqueous extracts from black non-halogenated (a–d), black bromobutyl (e–g) and blue chlorobutyl stoppers (h) (short (2h) high temperature (100°C) extractions; see material and procedure section). The mass spectra of compounds i (a), ii (b) and iii (c) from black non-halogenated butyl stoppers are similar, comprising major fragment ions at m/z 72 (iii), or 72/74 (i, ii), 86 (iii), or 86/88 (i, ii), which are often found in compounds with pyrimidine or morpholine moieties (however, other substructural compositions seem possible as well). (d) A cluster of compounds (iv, Fig. 3f) was additionally found in extracts from non-halogenated butyl rubber stoppers. The mass spectra of these compounds are almost identical strongly suggesting that these constitute an isomeric series. The fragment ions at m/z 77, 91, 105, and 167 tentatively suggests that the isomers contain a biphenyl moiety. Compounds v–vii are probably aromatic compounds containing hetero atoms (probably nitrogen and sulfur) as indicated by abundant fragment ions at m/z 65 and 109 but no further assignment was possible due to missing standards or accurate identification via spectral library searches. Similar to cluster iv, the abundant fragment ions at m/z 77 in the mass spectra of compounds vi (f) and vii (g) suggests to a phenyl moiety in these molecules.

Although the amount of organic contaminants was overall lower (about sevenfold) in low-T/long-t extracts of untreated non-halogenated butyl stoppers, we detected all compounds (including the UCM), which were present in the high-T/short-t extracts (data not shown). Our results thus provide evidence that the non-halogenated stoppers leached potentially toxic compounds from the rubber matrix into the incubation medium (where they are bioavailable) at conditions relevant for microbial incubations. Extracts from non-halogenated stoppers pretreated with NaOH/MeOH contained the same spectrum of organic contaminants with quantities comparable to low-T/long-t extracts (data not shown). The applied pretreatments are consequently insufficient for reducing contamination from butyl seals.

In contrast to the non-halogenated butyl stoppers, leaching of organic compounds from halobutyl rubber was much lower (Fig. 3b–e,g–j). Remarkably, the extracts of the gray chlorobutyl (Fig. 3d,i) and gray bromobutyl stoppers (Fig. 3e,j) did not contain any detectable amounts of organic contaminants. However, we found considerable concentrations of fluoranthene and pyrene, n-alkanes (C16–C34) (Fig. 3b) as well as three benzothiazoles (3-methylbenzothiazol-2-one, 2-(ethylamino)benzothiazole, 3-methylbenzothiazole-2-thione) and three unidentified contaminants (compounds v-vii; Fig. 3g, see mass spectra in Fig. 4e–g) in the extract of the black bromobutyl stoppers. The blue chlorobutyl stoppers also leached some n-alkanes (Fig. 3c) and another unidentified contaminant (compound viii; Fig. 3h; see mass spectrum in Fig. 4h).

Inorganic contaminants

Similar to organic contaminants, inorganic substances such as heavy metals leaching from rubber (Wik 2007) can be toxic for microbes reducing their activity in environmental samples (Giller et al. 1998; Bong et al. 2010; Ytreberg et al. 2010). Our ICP-OES measurements revealed that leaching from non-halogenated butyl-, black bromobutyl- and blue chlorobutyl stoppers increased the Zn concentration in the aqueous medium to micromolar concentrations independent of the extraction method (high-T/short-t, low-T/long-t; Table 5). In contrast, the gray bromo- and chlorobutyl stoppers released comparably little amounts of Zn into the medium. In this study, we focused on Zn, as Zn-containing compounds are widely used for manufacturing rubber, for example, as a cure activator and gelling agent (Bhowmick et al. 1994). Moreover, we did not observe elevated levels (> 0.02 μM) of heavy metals other than Zn in the leachates (data not shown).

Table 5. Zinc concentrations in aqueous stopper extracts. Note that Zn concentrations in microbial incubations are > fivefold lower; see material and procedure section). For comparison: marine and lacustrine waters typically contain submicromolar Zn concentrations.
Zn (μM)
Stopper type high-T/short-t low-T/long-t
black non-halo. butyl 22.03 35.67
black Br-butyl 8.26 6.93
blue Cl-butyl 4.28 1.00
gray Cl-butyl <0.02 0.15
gray Br-butyl 0.31 0.21

Potential origin and toxicity of organic and inorganic contaminants

Organic contaminants

Synthetic polymers may leach a multitude of organic molecules ranging from monomer building blocks, plasticizers, and additives (e.g., vulcanization activators, accelerators, and fillers or reinforcing agents) to incomplete reaction (side) products and dyes (Airaudo et al. 1989; Oshima and Nakamura 1994; Wik 2007). All identified compounds detected in the aqueous stopper extracts, fall into one of these categories. Benzyltoluene is a constituent of heat transfer fluids and could thus have been used during fabrication of the non-halogenated black butyl rubber. Phenylalkanes and n-alkanes are common petrochemicals that are often used as building blocks or rubber antiaging agents during rubber production (Bhowmick et al. 1994). Fluoranthene/pyrene is used as reinforcing agents in rubber, and benzothiazoles constitute an important class of vulcanization accelerators (Bhowmick et al. 1994; Karásek and Sumita 1996).

Several of the detected compounds are potentially toxic, but toxicity tests, if available, were mostly performed on eukaryotic organisms. Tests on prokaryotes, in contrast, are rather scarce, and test-matrices used as reference material are highly enriched with microbes. The half maximal effective concentration expressed in the EC50 value (OECD 2010), for instance, is tested on activated sludge, which contains ∼ 1012 cells mL−1 (Foladori et al. 2010), while natural, aquatic environments typically contain about 105 cells mL−1 (marine) to 106 cells mL−1 (lacustrine) (Whitman et al. 1998). Available toxicity estimates may thus not be representative for applications in biogeochemical research. To the best of our knowledge, none of the identified contaminants (with the exception of fluoranthene/pyrene; Rockne et al. 1998) has yet been tested with respect to their potential effect on MOx.

Many of the detected contaminants contain an aromaticity, making them potentially toxic to microorganisms (Heipieper and Martínez 2010; Segura et al. 2010). However, the toxicity of aromatic compounds on microorganisms is variable. Toxicity tests of benzyltoluene, for example, revealed an EC50 value of < 1 g L−1, while Pseudomonas putida (tested on cultures with 106 cells mL−1) seems to tolerate much higher contaminant levels (see data base of the European Chemicals Agency, www.echa.europa.eu). Furthermore, several microbes (Bhatia and Singh 1996; Widdel and Rabus 2001; Herter et al. 2012), including MOx communities (Colby et al. 1977; Burrows et al. 1984; Smith and Murell 2010), were found to utilize aromatic hydrocarbons, such as benzene, ethylbenzene, styrene. On the other hand, the degradation of phenylalkanes with even and odd-numbered alkyl side chains yields intermediates (phenylacetic and benzoic acid, respectively) that are known bacteriotoxins (Goshorn et al. 1938; Russell 1991), in particular at low pH (Goshorn et al. 1938; Heipieper and Martínez 2010). However, a strain of the type II MOx bacterium Methylocystis sp. was found to degrade benzoic acid (Uchiyama et al. 1989). The potential toxicity of the aromatic hydrocarbons (which were released to a great extent by the non-halogenated butyl stoppers, Fig. 3a) to MOx communities thus remains unclear. Exudation of benzothiazoles was greatest in incubations with black bromobutyl stoppers. Toxicity and physiological effects of these heterocyclic aromatic compounds on microbes as well as microbial consumption of benzothiazoles are well documented (De Wever and Verachtert 1997, 2001; Haroune et al. 2002). However, MOx activity in incubations with black bromobutyl stoppers did not seem to be suppressed (Fig. 2), suggesting that the detected benzothiazoles are nontoxic to the MOx communities in the studied environments.

In general, a relatively high susceptibility of MOx communities to organic compound interference can be expected because the substrate specificity of one of the key enzymes of the MOx pathway, methane mono-oxygenase (MMO), is relatively low. In addition to CH4, the soluble form of the MMO (sMMO) co-oxidizes short-chain n-, branched-, cyclic-, and halogenated alkanes as well as alkenes and aromatic compounds even though the metabolic products from oxidizing C-compounds other than CH4 cannot support cell growth (see review by Smith and Murell 2010). The particulate MMO (pMMO) appears to be more specific, co-oxidizing only short chain n- and chlorinated alkanes as well as alkenes and ammonia. Substrate competition for the active sides of the MMOs might thus explain some of the reduction in methane turnover observed in our experiments. The polyaromatic hydrocarbons such as fluoranthene/pyrene found in black bromobutyl stoppers, for example, may be used as a carbon substrate by some marine methanotrophs (Rockne et al. 1998). Similarly, long chain n-alkanes, of which high amounts were released from the black bromobutyl and (to a lesser degree) from blue chlorobutyl stoppers, also occur naturally in the environment and the ability to metabolize n-alkanes is widespread among microorganisms (Rojo 2010) including MOx communities (Colby et al. 1977; Hou et al. 1979; Burrows et al. 1984).

In conclusion, we found putative evidence for direct links between the presence of organic contaminants in butyl rubber stoppers, in particular the non-halogenated type, and the inhibition of MOx during ex situ incubations. However, we can only speculate about the potential toxicity of the detected compounds (identified and unidentified). In fact, some of the organic contaminants may be co-oxidized by MOx communities and may interfere with MOx.

Potential Zn interference

Zn is a common cofactor in many metalloenzymes and is thus essential for growth of many autotrophic and heterotrophic microorganisms, but an excess of Zn was shown to cause toxicity and inhibition of microbial processes (Giller et al. 1998; Bong et al. 2010; Ytreberg et al. 2010). Sea and lake water typically contains Zn concentrations in the submicromolar range (Baeyens et al. 1987; Xue et al. 1997; Bruland et al. 2006). Zn concentrations in our extracts, particularly from the non-halogenated butyl stoppers as well as the black bromobutyl and the blue chlorobutyl stoppers strongly exceed this range (Table 5) and can consequently be regarded as heavily Zn contaminated. Zn concentrations of 10 μM, for example, were found to almost completely inhibit extracellular enzyme activities (aminopeptidase) of a mixed marine microbial community (Bong et al. 2010) and a second study reported an EC50 value of 30.6 μM Zn for the bacterium Vibrio fishery (bioluminescence inhibition) (Ytreberg et al. 2010). Nevertheless, the potentially harmful effects of Zn have, to the best of our knowledge, not been tested for methanotrophs. The relatively high Zn concentrations in the black bromobutyl and blue chlorobutyl stopper extracts (equivalent to about 0.2–1.4 μM in the incubation media), concomitant with seemingly unsuppressed MOx rates suggest a comparably high tolerance of MOx communities to Zn. This might be related to the fact that one of the three metal centers of pMMO is occupied by Zn (Lieberman and Rosenzweig 2005). Indeed, investigations into the expression of pMMO in recombinant Rhodococcus erythropolis revealed optimum growth conditions at 1.8 μM Zn (Gou et al. 2006). Whether the much higher Zn concentrations in incubations with non-halogenated butyl stoppers (equivalent to about 7.1 μM) are partially responsible for the observed suppression of MOx thus remains unclear.

Discussion

A widely applied technique to measure MOx rates in aquatic environments is based on the ex situ incubation of water samples. For this approach, a sample aliquot is filled into an incubation chamber, typically a glass vial, and sealed with a butyl rubber stopper. The MOx rate is then determined by monitoring the consumption of CH4, for example, by measuring the change of CH4 concentration over time or by measuring the turnover of isotopically labeled substrate (Reeburgh 2007; Mau et al. 2013; Blees et al. 2014a). Our data show a clear effect of the rubber seal material on MOx rates, and suggest that non-halogenated butyl rubber strongly inhibits CH4 turnover (Fig. 2). Our results furthermore show that the non-halogenated butyl stoppers leach a multitude of organic compounds (Fig. 3a,f) as well as comparably high amounts of Zn (Table 5) into the incubation medium. Although the exact cause for MOx suppression remains uncertain, we argue that leaching of potentially toxic substances is responsible for the observed inhibition of MOx activity. Our study showed that only two of the tested halobutyl seals (gray bromo- and chlorobutyl stoppers) seemed not to leach detectable amounts of contaminants. This could be related to their use as a pharmaceutical product with potentially stricter “low-bleed” requirements, at least for butyl seals. The other halogenated butyl elastomers tested here, in contrast, were found to exude substantial amounts of organic substances and Zn. Manufactures usually do not provide details on the chemical composition and leaching behavior of elastomers, so that specifications in the scientific literature are equally sparse. It thus remains speculative how extensively compromising elastomer types have been used in previous work for measuring MOx rates, potentially resulting in biased reports of CH4 turnover.

The relevance of this study goes beyond applications of butyl rubber seals for assessing MOx in aquatic environments and likely has implications for other incubation/enrichment studies, and other elastomer seals. The non-halogenated butyl elastomer used in this study, for example, appears to inhibit the activity of the Bacterium Candidatus Methylomirabilis oxyfera, which mediates the anaerobic oxidation of CH4 with nitrite (Ettwig et al. 2010). Possibly, the toxic effects are similar as for MOx communities because M. oxyfera utilizes most of the enzymatic machinery found in MOx communities. The black bromobutyl stoppers (which seem rather unproblematic for MOx rate measurements), may hamper the metabolic activity of fungi or nitrifying bacteria because these organisms are susceptible to benzothiazoles (De Wever and Verachtert 1997), which are released from this stopper type. Elastomer types other than butyl rubber may also leach potentially toxic contaminants. Certain types of latex and Tygon were found to suppress aquatic microorganisms (Price et al. 1986), and, although tested on mouse cells, cytotoxicity was confirmed for some types of pharmaceutical nitrile rubber and polyvinyl chloride (Lönnroth 2005). Converesely, the contaminants may also constitute an additional source of carbon, which might be utilized by the organism under investigation, potentially compromising activity/growth estimates. Alternatively, other microorganisms may benefit from the additional substrate, which could lead to a shift in the community structure in the incubation vessel.

Our study demonstrated that an experimental setup involving non-halogenated butyl elastomer can yield unreliably low rates of MOx activity, or may even lead to its oversight in the environment. Contaminants released from butyl rubber may also compromise the assessment of biogeochemical reaction rates other than MOx. With respect to the measurements of MOx activity, leaching behavior and ease of general handling, it is plausible to recommend using the gray bromo- or chlorobutyl stoppers reported in this study. For any other stopper types and/or manufactures, and independent of the studied biological processes, we recommend testing the elastomer's potential for leaching of organic and inorganic contaminants.

Acknowledgments

We are indebted to the captain, crew, and scientific research parties of the research vessels Littorina, Polarfuchs, Alkor (cruise 410) and Maria S. Merian (cruise MSM 21/4). Marco Simona is thanked for his excellent support during fieldwork on Lake Lugano and Judith Kobler Waldis for ICP-OES measurements. The reviewer John Kessler and David Valentine are thanked for their constructive comments/criticism. Funding for this study was provided by the Swiss National Science Foundation and German Research Foundation (NSF-DFG DACH grant 200021L_138057 and SNF grant 200021_121861) as well as COST Action PERGAMON (STSM ES0902-14336 and STSM ES0902-10695). M.E. was supported by the DFG through the Research Center/Excellence Cluster MARUM-Center for Marine Environmental Sciences.