sample lipid extraction (Sweeting et al., 2006), samplestorage and preservation appears to be a crucial point.There is a general agreement that oven drying, freezing
Journal of Experimental Marine Biology an1. Introduction
Analysis of tissue stable isotope ratios in nitrogen(15N/14N or 15N vs. AIR) and carbon (13C/12C or 13Cvs. PDB) is an essential part of modern food web studiesin marine benthic communities (e.g. Fry, 1988; Hobsonand Welch, 1992; Herman et al., 2000; Jennings et al.,2002). The underlying principle is the enrichment of theheavier isotope 15N and 13C with each assimilation stepin the food chain owing to the selective metabolic loss of
the lighter isotopes during food assimilation and growth.On average, fractionation per trophic level amounts to2.55 in 15N and to 01 in 13C (De Niro andEpstein, 1978, 1981; Minagawa and Wada, 1984; Post,2002).
There are many natural sources of variability instable isotope ratios, but bias owing to sample treatmenthas drawn a considerable amount of attention recently.Besides sample acidification (Bosley and Wainright,1999; Jacob et al., 2005; Carabel et al., 2006) andMacrobenthic infaunal sampling campaigns that aim at unpreserved fresh animals for later biochemical analysis follow one oftwo strategies: either samples are sieved and sorted on board and individual specimens are frozen, or grab/corer samples are frozencompletely and animals are separated from sediments later in the laboratory. This study demonstrates that deep freezing ofcomplete grab samples causes a significant bias in stable isotope ratios of nitrogen (15N/14N or 15N vs. AIR) and carbon (13C/12Cor 13C vs. PDB) as well as in the C/N ratio of nine infaunal species from the German Bight. On average, molar C/N ratio increasesby 0.93, 13C decreases by 1.87, and 15N decreases by 1.01. Mechanical cell destruction and subsequent loss of cytosol aswell as metabolic degradation by free enzymes and by microorganisms are discussed as major causes for the observed effects. Werecommend to abstain from using bulk frozen grab samples for the analysis of C/N ratio or stable isotope ratios. 2007 Elsevier B.V. All rights reserved.
Keywords: Freezing storage; Infaunal macrozoobenthos; Sample processing; Stable isotopesAbstractDoes sample bulk freezinginfaunal mac
Jennifer Dannheim a,, Ula Alfred Wegener Institute for Polar and Marine Reb Museum of Natural History Berlin, Humboldt U
Received 19 SeptemberCorresponding author. Tel.: +49 471 48311718; fax: +49 47148311724.
E-mail address: Jennifer.Dannheim@awi.de (J. Dannheim).
0022-0981/$ - see front matter 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.jembe.2007.06.001ect stable isotope ratios ofzoobenthos?
Struck b, Thomas Brey a
h, PO Box 120161, 27568 Bremerhaven, Germanysity, Invalidenstrasse 43, 10115 Berlin, Germany
; accepted 5 June 2007
d Ecology 351 (2007) 3741www.elsevier.com/locate/jembeand freeze drying show little or no effects, whereasformalin and ethanol fixation cause significant bias
Fig. 1. Plots of (a) C/N ratio, (b) 13C and (c) 15N ratio in bulktreatment (BF) versus individual treatment (IF) samples. Horizontaland vertical bars represent standard deviation of x and y variable forEchinocardium cordatum (Ec), Euspira pulchella (Ep), Laniceconchilega (Lc), Nephtys sp. (Ns), Poecilochaetus serpens (Ps), Sco-lelepis bonnieri (Sb), Tellimya ferruginosa (Mf), Tellina fabula (Tf).Line indicates 1:1 ratio. : significant difference between IF and BF inthis species as indicated by post-hoc test (pb0.05).
l Marine Biology and Ecology 351 (2007) 3741(Bosley and Wainright, 1999; Ponsard and Amlou,1999; Kaehler and Pakhomov, 2001; Feuchtmayr andGrey, 2003; Lorrain et al., 2003; Sweeting et al., 2004).
In studies on infaunal macrozoobenthic species, thebulk freezing of complete grab/corer samples directlyafter collection is an appealing approach (e.g. Lovvornet al., 2005). Separating tiny animals from the sedimentand identifying them is a time consuming and laborioustask, which can be handled much better in the laboratorythan on board of a research vessel with tight restraints ontime and space. Here we evaluate whether bulk freezingof complete grab/corer samples and defrosting themagain in the laboratory causes bias in stable isotope ratiodeterminations in animals collected from this sample,and whether such bias is species specific.
2. Materials and methods
Infaunal macrozoobenthos was collected with a0.1 m2 van Veen grab at one site in the German Bight(5403.835400.96N and 00627.30006 43.13E,autumn 2004, RV Heincke). Thirteen samples werefrozen directly after collection, i.e. the bulk content ofthe grab was transferred into a polyethylene box andstored at 20 C (bulk freezing, BF). From another 25samples, we handpicked 91 individuals belonging tonine dominant infaunal species (Echinocardium corda-tum, Euspira pulchella, Lanice conchilega, Nephtyssp., Poecilochaetus serpens, Scolelepis bonnieri, Telli-mya ferruginosa, Tellina fabula and Urothoe poseido-nis). Each individual was stored separately in a plasticjar at 20 C (individual freezing, IF). After 13 monthsstorage, BF samples were slowly defrosted at 8 C, andthe sediment was removed layer-by-layer following thethawing horizon down into the sample. 77 infaunalspecimens were handpicked from the sediment andtransferred to cooled seawater (34 C) where theyremained for a maximum of 20 min before further pro-cessing. Subsequently, both BF and IF treated sampleswere lyophilised for 24 h (Finn-Aqua Lyovac GT2E),homogenized in a ball mill, treated with 1 mol l1
hydrochloric acid (HCL) to eliminate calcium carbo-nates following the procedure of Jacob et al. (2005),redried at 55 C and ground to powder in a mortar.
Stable isotope ratios 13C and 15N were determinedby an isotope-ratio mass spectrometer (Thermo/Finni-gan Delta Plus at GeoBioCenterLMU, University ofMunich). Gaseous standards (N2 and CO2 respectively)were calibrated against atmospheric nitrogen (AIR) fornitrogen and the PeeDee Belemite standard (PDB) forcarbon. An internal standard (Peptone with known
38 J. Dannheim et al. / Journal of Experimentaisotope composition) was used for every 6th sample.
l MarExperimental precision was 0.2. Both isotopesratios were expressed in notation:
dX x Rsample=Rstandard 1 103
where X is 13C or 15N, and R is the 13C/12C or 15N/14Nratio.
All samples with b5 g nitrogen content wereexcluded from further analysis because of insufficientmeasurement precision (10 individuals). Multivariateoutliers in the sample space [C/N, 15N, 13C] wereidentified by Mahalanobis distances (Barnett and Lewis,1994) and excluded, too (1 individual). Delta13C datawere multiplied by 1 to obtain positive values, and allthree data series (C/N, 15N, 13C) were BoxCoxtransformed (Sokal and Rohlf, 1995) in order to achievenormality and equality of variances. Full interaction 2-way analysis of variance (ANOVA) followed by post-hoc Tukey HSD test on differences between means wasapplied to check for effects of treatment (IF, BF) and ofspecies on the three parameters.
Compared to individual freezing (IF), bulk freezing(BF) caused a significant (pb0.001) increase in C/N (onaverage 0.93, Fig. 1a), and a decrease in both 13C (onaverage 1.87, Fig. 1b) and 15N (on average 1.01,Fig. 1c). In C/N ratio and 15N ANOVA detectedsignificant interactions between treatment and species(Table 1). On the species level, BF treatment effects were
Table 12-way analysis of variance testing the effects of treatment (IF, BF) and
Source df 13C
Treatment 1 1.98e-4 309.34
Species 8 2.71e-5 42.36
Interaction 8 7.47e-7 1.17
df: degrees of freedom, MS: mean square; significant effect with pb
J. Dannheim et al. / Journal of Experimentasignificant in seven (C/N, Fig. 1a), nine (13C, Fig. 1b),and three species (15N, Fig. 1c), respectively.
Freezing by itself, i.e. the exposure of organic matter tobelow zero temperature, has no effect on C/N ratio andstable isotope ratios (Bosley and Wainright, 1999;Ponsard and Amlou, 1999; Kaehler and Pakhomov,2001; Sweeting et al., 2004). There are, however, severaldistinct differences between the IF and the BF treatmentprocedure which may affect tissue condition andbiochemical properties: (i) The freezing process of abulk sample takes much longer owing to its larger massand smaller surface/volume ratio compared to singleanimals. Slower freezing allows for the formation ofmoreice crystals inside tissues which can destroy cells mechan-ically and lead to cell content leakage (Salonen andSarvala, 1980; Feuchtmayr and Grey, 2003). (ii) Tissuesmay be damaged further by mechanical stress caused bythe antagonistic forces of liquid volume expansion duringfreezing against sample mass. (iii) Tissue componentsmay be metabolized by free enzymes or by microorgan-ism inhabiting the sediment (e.g. Ponsard and Amlou,1999), in particular during freezing and thawing of thesample and following mechanical tissue destruction. (iv)Certain macromolecules, in particular fatty acids can beoxidized chemically at temperatures as low as 30 C (E.Brodte, pers. comm.).
The observed increase in C/N ratio may be due toeither an increase in C content or a decrease in N content.Salonen and Sarvala (1980) report a significant leakageof carbon during freezing but do not provide informationon the corresponding loss in nitrogen. We presume thatmechanical cell destruction during BF treatment leads toa loss of cytosol when thawing. Cytosol has a higherprotein content and thus nitrogen content thanorganelles and membranes which are comparativelyricher in carbohydrates and in fatty acids (Alberts, 2002).
May cytosol loss foll