Tracing the transformation of labelled [1-13C]phenanthrene in a soil bioreactor

Article abstract of DOI:10.1016/S0269-7491(99)00205-5

[1-¹³C]-labelled phenanthrene was incubated in a closed bioreactor to study the flux and biotransformation of polycyclic aromatic hydrocarbon (PAH) in contaminated soils on a bulk and molecular level. The degradation of extractable phenanthrene was observed by GC-MS measurements and the mineralisation was monitored by ¹³CO₂ production. The transformation of the ¹³C-label into non-extractable soil-bound residues was determined by carbon isotopic measurements. With these data we were able to calculate a carbon budget of the ¹³C-label. Moreover, the chemical structure of non- extractable bound residues was characterised by applying selective chemical degradation reactions to cleave xenobiotic subunits from the macromolecular organic soil matrix. The obtained low molecular weight products yielded ¹³C-labelled compounds which were identified using IRM (isotope ratio monitoring)-GC-MS and structurally characterised with GC-MS. Most of the ¹³C-labelled products obtained by chemical degradation of non-extractable bound residues are well-known metabolites of phenanthrene. Thus, metabolites of [1-¹³C]phenanthrene formed during biodegradation appear to be reactive components which are subsequently involved in the bound residue formation. Hydrolysable amino acids of the soil residues were significantly labelled with ¹³C as confirmed by IRM-GC-MS measurements. Therefore, phenanthrene- derived carbon was transformed by anabolic microbial processes into typical biologically derived compounds. These substances are likely to be incorporated into humic-like material after cell death. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0269-7491(99)00205-5

Environmental Pollution 108 (2000) 91±101
www.elsevier.com/locate/envpol
1
3
Tracing the transformation of labelled [1- C]phenanthrene in a
soil bioreactor
a,
a
b
b
H.H. Richnow *, E. Annweiler , M. Koning , J.-C. Lu
È
R. Stegmann , C. Garms , W. Francke , W. Michaelis
th ,
a
b
c
c
a
Institut fuÈr Biogeochemie und Meereschemie, UniversitaÈt Hamburg, Hamburg, Germany
bAB. Abfallwirtschaft, Technische UniversitaÈt Hamburg, Hamburg, Germany
cInstitut fuÈr Organische Chemie, UniversitaÈt Hamburg, Hamburg, Germany
Received 6 December 1998; accepted 12 July 1999
`Capsule'': Detailed experiments are described in which the transformation of 13C-labelled phenanthrene is tracked in a
soil bioreactor.
`
Abstract
1
3
[1- C]-labelled phenanthrene was incubated in a closed bioreactor to study the ¯ux and biotransformation of polycyclic aromatic
hydrocarbon (PAH) in contaminated soils on a bulk and molecular level. The degradation of extractable phenanthrene was observed
by GC±MS measurements and the mineralisation was monitored by 13CO production. The transformation of the 13C-label into non-
2
extractable soil-bound residues was determined by carbon isotopic measurements. With these data we were able to calculate a carbon
budget of the 13C-label. Moreover, the chemical structure of non-extractable bound residues was characterised by applying selective
chemical degradation reactions to cleave xenobiotic subunits from the macromolecular organic soil matrix. The obtained low mole-
cular weight products yielded 13C-labelled compounds which were identi®ed using IRM (isotope ratio monitoring)±GC±MS and
structurally characterised with GC±MS. Most of the 13C-labelled products obtained by chemical degradation of non-extractable
13
bound residues are well-known metabolites of phenanthrene. Thus, metabolites of [1- C]phenanthrene formed during biodegrada-
tion appear to be reactive components which are subsequently involved in the bound residue formation. Hydrolysable amino acids of
the soil residues were signi®cantly labelled with 13C as con®rmed by IRM±GC±MS measurements. Therefore, phenanthrene-derived
carbon was transformed by anabolic microbial processes into typical biologically derived compounds. These substances are likely to
be incorporated into humic-like material after cell death. # 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Transformation; 13C-label; Phenanthrene; Soil bioreactor
1
. Introduction
(Richnow et al., 1994). Progress in search of the chemi-
cal structure of bound residues has been made applying
1
3C-labelled model compounds to trace the transforma-
The formation of non-extractable bound residues is a
signi®cant sink of organic pollutants in soils (Bollag and
Loll, 1983; Calderbank, 1989; Richnow et al., 1999a). In
the case of hydrophobic pollutants like polycyclic aro-
matic hydrocarbon (PAH) recent advances in bound
residue research could clearly assign microbial processes
to be a key factor catalysing bound residue formation.
A metabolic activation is required to transform rela-
tively inert parent compounds into reactive microbial
metabolites which may subsequently interact with
natural soil organic matter to form bound residues
tion of pollutants in soils (Richnow et al., 1998).
Experiments with 13C-labelled anthracene revealed that
microbial metabolites of PAH are structural subunits
of bound residues partly linked by covalent bonds to
soil organic matter. Enzyme-catalysed oxidative cross-
coupling reactions which form covalent chemical bonds
may be involved in the residue formation reaction
(Berry and Boyde, 1985; Hatcher et al., 1993; Dawel et
al., 1997). Moreover, ageing processes in soils have been
shown to be responsible for limited bioavailability and
biodegradability of PAH (Hatzinger and Alexander,
1
995). Sequestering within non-aqueous phase liquids,
micropores, clay minerals and in the soil organic matter
are considered to reduce the mineralisation of anthro-
pogenic chemicals in soils.
* Corresponding author. Tel.: +49-40-4123-4987; fax.: +49-40-
123-6347.
E-mail address: richnow@geowiss.uni-hamburg.de (H.H. Richnow)
4
0
269-7491/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0269-7491(99)00205-5

9
2
H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
According to the IUPAC de®nition, ``non-extractable
taminated during the destruction of a fuel depot in
World War II and is at least more than 53 years old.
Soil material (40 kg) was sieved to a particle size of
<1 mm to remove gravel material and carefully homo-
genised prior to any further treatment. The pH was
adjusted to 6.3 by adding ®ne Ca(OH)2 powder to
the soil. The water content of the soil was adjusted
with distilled water to 50% of the maximal water-
holding capacity before incubation in the bioreactor
experiments.
residue (sometimes also referred to as bound residues, or
non-extracted residues) in plants and soils are de®ned as
chemical species originating from pesticides, that remain
unextracted by methods which do not signi®cantly
change the chemical nature of these residues. These
non-extractable residues are considered to exclude frag-
ments recycled through metabolic pathways leading to
natural products.'' (Roberts, 1984). The chemical spe-
cies in this context refer either to parent material or
metabolites of it.
To stimulate the microbial activity the soil was sup-
plemented by the addition of compost (5 wt.% of dry
compost). The compost added was produced at a wind-
row composting plant in Hamburg which processed
yard waste and separately collected vegetable kitchen
waste. The mature biocompost had a grain size of less
than 4 mm, a maturity degree of V (German systematic
for highly matured compost) and was at least more than
Considering biological transformation processes in
soils, the former pollutants may be mineralised in cata-
bolic processes or transformed in the microbial anabo-
lism to chemical structures identical to those of natural
origin, such as carbohydrates, amino acids, and fatty
acids. These latter compounds do not belong to the
bound residue fraction, even if parts of their carbon
originate from anthropogenic chemicals. In practice,
bound residues are very dicult to distinguish from
compounds transformed into biological fractions, espe-
cially when radioactively labelled tracer substances are
used to quantify the bound residue fraction without any
structural assignments. In this study we apply 13C-
labelled contaminants to elucidate the transformation of
xenobiotics in soils. We analysed the anabolic conver-
sion of xenobiotic carbon into typical chemical struc-
tures of the microbial biomass. In particular, we discuss
13
6 months old. Prior to spiking the [1- C]phenanthrene,
the soil was mixed with dry compost in a stainless steel
pastry blending machine.
Soil material (4 kg) was loaded with the isotopic tra-
cer by adding dropwise (drop size <10 ml) ethylacetate
13
solutions containing [1- C]phenanthrene (50.03 mg
ml 1). For a batch experiment with non-labelled phe-
nanthrene 3.2 kg soil material was contaminated like-
wise. The soil material was carefully homogenised
during the contamination procedure. Then the solvent
(ethyl acetate) was evaporated at ambient temperature
for 1.5 h during continuous mixing of the soil material.
�
1
3
the transformation of [1- C]phenanthrene in soil on a
bulk and on a molecular level emphasising: (1) miner-
alisation; (2) non-extractable bound residue formation;
and (3) transformation into the biomass.
2.2. Chemicals
Moreover, ageing e€ects were investigated by spiking
a soil with 13C-labelled phenanthrene which was con-
taminated at least 53 years ago and yield signi®cant
amounts of non-aqueous phase liquids (tar). In a soil
bioreactor the biodegradation of labelled and tar oil-
derived phenanthrene was traced to obtain information
about their bioavailability.
Several grams of 13C-labelled phenanthrene were syn-
thesised (Scheme 1). The synthesis of small amounts of
13
[1- C]phenanthrene was described earlier (Soderquist et
al., 1992) but some modi®cations were essential. The
commercially available 1-methylnaphthalene (1, Scheme
1) was deprotonated with butyl-lithium in a mixture of
0
0
diethyl ether and N,N,N ,N -tetramethylethylendiamine
TMEDA). The reaction with racemic 2-(2-bro-
moethoxy)-tetrahydropyran gave compound 2 (Scheme
) (Mallan and Bebb, 1969). In the next step the tetra-
(
2. Materials and methods
1
2
.1. Soil
hydropyran group (THP) was removed with p-toluene-
sulphonic acid in methanol (Marshall, 1969) and then
the obtained alcohol was treated with triphenyl-
phosphoniumdibromide to yield the corresponding
bromide 3 (Scheme 1) according to Wiley et al. (1969).
The bromide 3 (Scheme 1) was converted to its Grignard
The soil material used in this experiment was collected
from a military base in Achim near Bremen (Germany).
The soil was derived from the upper soil column and
consisted of 1.2 wt.% clay, 2.2 wt.% silt, 92.0 wt.%
sand and 5.2 wt.% gravel. The total organic carbon
content was 1.5 wt.%. The soil was heavily con-
taminated by mineral-oil derived aliphatic hydro-
carbons and PAHs. The acidic soil (pH 4.8) was free of
carbonates, and the molecular ratio of C:N:P (100:3:0.2)
revealed nitrogen and phosphorous de®ciencies. The
historical investigation indicates that the soil was con-
13
reagent with [ C]carbon dioxide, generated from
13
[ C]sodium carbonate with concentrated sulphuric acid,
to obtain the labelled carboxylic acid 4 (Scheme 1). Com-
pound 4 was transformed to its acid chloride with oxalyl
chloride and then cyclisised with tin tetrachloride in
benzene to give the ketone (5, Scheme 1) (Bachmann et
al., 1940).

H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
93
Scheme 1. Synthesis of [1-13C]phenanthrene.
After reduction of 5 with lithium-aluminium-tetra-
hydride using a description of Brown and Krishna-
murthy (1979), dehydratisation, and oxidation with 2,3-
dichloro-5,6-dicyano-p-benzoquinone (DDQ) (Fu and
dichloromethane, twice]. For extraction, the soil ma-
terial (10±20 g soil) was mixed with an excess of organic
solvent (50±100 ml). The extraction procedure was
repeated four times using ultrasoni®cation (RK100H,
Sonorex, Bandelin, Berlin, Germany) to enhance the
eciency of the extraction. Aliquots of the extracts were
analysed for the phenanthrene concentration by GC
and GC±MS.
1
3
Harvey, 1978) the [1- C]phenanthrene (6) was obtained.
1
3
[
1- C]Phenanthrene was ®nally puri®ed by column
chromatography with silica gel as stationary phase
Merck, 230±400 mesh) using n-hexane as developer.
(
13
The chemical and isotopic purity of the [1- C]phen-
anthrene (99%) were veri®ed by H-NMR, 13C-NMR
The pre-extracted soil was dried for 15 h under
ꢀ
1
reduced pressure (40 mbar) at 40 C to remove the
and GC±MS.
organic solvent. The extracted soil sample was de®ned
as the non-extractable soil-bound residue fraction and
then subjected to further analyses. The total organic
carbon content of the total soil and bound residue frac-
tion was determined with a LECO CHN 100 elementary
analyser (Leco Instruments, Kirchheim, Germany).
All other chemicals used for synthesis, extraction and
preparation were obtained from Merck (Darmstadt,
Germany). The organic solvents (p.A.-quality) were
distilled before use.
2.3. pH determination
2
.5. Chemical degradation
The pH of the soil was measured according to DIN
8404-65. Brie¯y, 10 g soil was mixed with 75 ml 0.01 M
CaCl and the pH was measured after 24 h.
3
Alkaline hydrolysis was applied to cleave ester-linked
bound residues from the non-extractable macromol-
ecular soil matrix (Richnow et al., 1994). Extracted soil
(1±2 g) was transferred into a glass ampoule and mixed
with 10 ml 2 M NaOH. The ampoule was closed under
2
2.4. Extraction procedures
Prior to the extraction, soil samples were dried in a
N atmosphere and the mixture was hydrolysed for 2 h
2
at 100 C. The reaction mixture was separated by cen-
ꢀ
freeze dryer (Christ Alpha 1-4, Martin Christ Gefrier-
trocknungsanlagen GmbH; Osterode, Germany). The
water content was calculated by weight di€erence of the
soil before and after drying. Soil samples were extracted
with an organic solvent sequence of decreasing polarity
trifugation followed by two extractions with distilled
water. Then the alkaline solution was acidi®ed with 6 M
HCl to pH 1±2 and extracted with dichloromethane.
The dichloromethane extract was dried with anhydrous
Na SO . Organic acids and aromatic alcohols were
[
I, acetone; II, acetone:dichloromethane (1:1, v:v); III,
2
4

9
4
H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
derivatised with diazomethane to methyl esters and
methyl ethers.
2.9. Stable isotope analyses
For the determination of the mineralisation of the
13C-tracer, the 13CO was trapped in 20 ml 1 M NaOH
2.6. Hydrolysable amino acids
2
solution. Dissolved carbonate in the NaOH solution
was precipitated with a BaCl2 solution. The BaCO3
precipitate was washed several times with distilled water
and dried before determination of the weight to calcu-
late the amount of CO2.
Precipitated barium carbonate and pre-extracted soils
were analysed for the composition of stable carbon iso-
topes using a MAT 251 (Finnigan, Bremen, Germany)
mass spectrometer connected to an elemental analyser.
The isotopic composition of carbon is expressed in
standard di€erences to the Pee Dee Belemnite standard
Soil material (1±2 g) was hydrolysed with 5 ml dis-
ꢀ
tilled 6 M HCl for 22 h at 110 C. The acidic supernatant
of the reaction mixture was used for further analysis.
An aliquot of the reaction mixture was used for the
quanti®cation of amino acids using ortho-phthaldialde-
hyde reagent for derivatisation prior to HPLC analysis
Dawson and Liebezeit, 1983). Another aliquot was
derivatised with 2-propanol to obtain isopropyl esters
and tri¯uoroacetic acid to obtain corresponding tri-
(
¯
uoroacetates (Silfer et al., 1991). The amino acids were
1
3
13 12
identi®ed using GC±MS and authentic reference com-
pounds before determination of their isotopic composi-
tion by isotope ratio monitoring±gas chromatography±
mass spectroscopy (IRM±GC±MS).
(PDB) in terms of d values: (d C [%]=([ C/ C)sample]/
13
12
[ C/ C] � 1)Â1000) (Hoefs, 1997). To calibrate the
std
1
3
mass spectrometer for the determination of C-enri-
ched carbon, reference material IAEA-309 UL-Glucose
A 93.9 (% PDB) and IAEA-309 UL-Glucose B 535.3
(% PDB) was purchased from Analytical Quality Con-
trol Service (Vienna, Austria).
2
.7. GC and GC±MS analyses
High-resolution gas chromatographic analyses were
performed on a gas chromatograph (GC-6000 Vega 2,
Carlo Erba Instruments, Milan, Italy) equipped with a
fused silica capillary column (DB-5, 30 mÂ0.25 mm,
J&W Scienti®c, USA). The temperature program for the
analysis of aromatic hydrocarbons and metabolites was:
2.10. Fixed-bed bioreactor with continuous oxygen
supply
The bioreactor consisted of a 5-l glass vessel which
was loaded with 3 kg contaminated soil material (Fig.
1). A continuous ¯ow of air was maintained at a ¯ux of
ꢀ
ꢀ
ꢀ
� 1
80 C, 3 min isothermal; 80±300 C, 4 C min , 20 min
isothermal 300 C; injection mode: on column; carrier
ꢀ
gas: H . GC±MS measurements were performed with a
2
Varian CH7A (Finnigan, Bremen, Germany) mass
spectrometer (ionisation mode: EI; 70 eV ionisation
ꢀ
energy; 250 C source temperature; mass range: m/z 50±
600; resolution 1000) coupled with a Carlo Erba gas
chromatograph (carrier gas: He) using the same condi-
tions as for the GC measurements. Phenanthrene-d10
was used as internal standard for quanti®cation of
PAH. Ion chromatograms of m/z 188 (internal stan-
dard), m/z 178 and m/z 179 were integrated and used to
quantify the concentration of tar oil-related phenan-
1
3
threne and [1- C]phenanthrene, respectively.
2.8. Isotope ratio monitoring GC±MS
The isotopic composition of amino acids and chem-
ical degradation products was measured with IRM±
GC±MS (Finnigan Mat, Bremen, Germany). The
system consisted of a GC-unit, interfaced to a Finnigan
Mat combustion device with a water removal assembly
and coupled to a Finnigan Mat 252 mass spectrometer.
The organic substances in the e‚uent stream from the
GC were oxidised to CO in the combustion interface.
2
The CO produced was transferred continuously to the
2
12
mass spectrometer to determine the 13CO / CO ratios.
2
2
Fig. 1. Fixed-bed bioreactor.
Methodical details are described in Hayes et al. (1990).

H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
95
�
1
� 1
dry soil. The humidity of the air was
2). The mineralisation of 13C-labelled phenanthrene was
determined by the amount and isotopic composition of
the CO fraction. The concentration of organic carbon
1
l h
kg
achieved by leading the gas stream through a gas-
washing ¯ask ®lled with water. A detector array includ-
ing a ¯ow controller was used for permanent registra-
tion of gas ¯ow, CO , volatile organic carbon (VOC).
2
The data set was stored on a computer for further ana-
lysis. CO was measured with an IR-spectrometer (Gas
Analyzer ADC 7000, Analytical Development Co. Ltd,
Hoddesdon, UK), gas ¯ow with a mass ¯ow controller
Bronkhorst, AK Ruurlo, Netherlands) and VOC with
a gas analyser (Hydrocarbon Analyzer Mod. 3002, Ber-
nath Atomic, Wennigsen, Germany). In the case of the
experiment with labelled phenanthrene, the air was
purged to a gas-washing ¯ask (600 ml) ®lled with 4 M
NaOH (400 ml) to remove CO from the air supply. A
2
similar gas-washing ¯ask (600 ml) with 400 ml 2 M
NaOH was used to trap CO before the gas stream was
sent to the detector array. More information about the
bioreactor system is given by Hupe et al. (1996) and
Koning et al. (1998).
2
and its isotopic composition was used to quantify
the immobilisation of the 13C-label within the non-
extractable soil-bound residues according to Richnow et
al. (1999b). Both phenanthrene species in the solvent-
extractable fraction was measured with conventional
GC and GC±MS to quantify residual parent material.
2
13
(
The concentration of extractable [1- C]phenanthrene
�
1
decreased from 354.3 to 1.5 mg g within the ®rst 50
days and then slightly declined until the end of the
experiment (0.11 mg g 1). 13CO was already formed
�
2
after 5 days characterising the short adaptation period
necessary for the micro¯ora to develop the ability to
mineralise phenanthrene. The 13CO evolution was most
2
pronounced between days 0 and 14, indicating min-
eralisation as the main elimination process of phen-
anthrene. Between days 15 and 40, bound residue
formation was an important sink of phenanthrene in the
soil. The residue formation reached the highest con-
2
�
1
centration of 35.5 mg g phenanthrene equivalents at
day 35. Parts of the bound residues were then slowly
degraded in the next 47 days to a ®nal concentration of
3
. Results and discussion
� 1
3
.1. Carbon budget traced with the 13C-labelled
compounds
21.3 mg g phenanthrene equivalents. The bound resi-
due formation takes place during the period of intensive
biodegradation, indicating a clear microbial contribu-
tion in the formation process.
The total recovery of 13C was 83% of the initially
applied amount after 82 days. The carbon budget
revealed a continuous loss of about 17% over time
To elucidate the transformation of PAH in soil during
1
3
microbial degradation [1- C]phenanthrene with a the-
1
3
oretical d- C of 5845% (PDB) was applied as a model
compound. The distribution of the C-label between
1
3
CO , solvent extractable and non-extractable bound
2
which may be the result of CO leakages or incomplete
2
residue fraction was used to calculate a carbon budget
and to obtain information about the fate and transfor-
mation processes of PAH in soils (Fig. 2; Tables 1 and
CO sorption in the NaOH traps of the bioreactor sys-
2
tem. From a number of similar bioreactor experiments
with 14C-labelled substances, we know that volatile
Fig. 2. The fate of [1-13C]phenanthrene during biodegradation in soil. (Phenanthrene (eq)=recovery of the labelled 13C-carbon calculated as phe-
nanthrene equivalents.)

9
6
H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
Table 1
Carbon budget of 13C-phenanthrene in the bioreactora
Time
days)
Soil
(g dry wt)
CO
(mg g soil)
2
CO
(mg g soil)
2
cumulative
� 1
Total extract
� 1
Bound residue
� 1
Budget
� 1
�
1
(
(mg g soil)
(mg g soil)
(mg g soil)
0
5
1
3
5
6
8
2699.1
2613.0
2527.0
2440.2
2353.3
2266.2
2178.1
0.0
112.3
124.5
13.6
3.6
0.0
112.3
236.8
250.4
254.0
254.3
254.3
350.1
186.6
62.7
22.7
12.8
17.2
15.0
0.8
14.4
21.5
35.5
26.9
22.4
21.5
350.9
313.3
321.0
308.7
293.7
294.0
290.8
4
5
0
3
2
0.3
0.0
a
The concentration of 13CO
, total extracts, bound residues and the budget are given in 13C-phenanthrene equivalents. Phenanthrene equivalents
2
are recalculated based on the 13C composition of [1-13C]phenanthrene, assuming that all carbon atoms of the model substance behave similarly in
the transformation process.
Table 2
Concentration and degradation rates of extractable tar oil-borne phenanthrene and [1-13C]phenanthrene in the bioreactor experiment
(mg g 1)
�
Time interval (days)
Degradation (% day
� 1
)
Time (days)
12C
13C
12C
13C
0
88.6
21.3
3.69
1.48
0.23
0.23
354.3
84.8
±
±
±
1
3
5
6
8
4
5
0
3
2
0±14
14±35
35±50
50±63
63±82
5.42
3.94
4.00
6.49
0.03
5.43
4.48
4.74
6.30
3.18
5.09
1.47
0.27
0.11
metabolic products are quantitatively insigni®cant but a
volatilisation of small amounts of the parent phenan-
threne in the gas stream of the oxygen supply cannot be
3.2. Bioavailability of phenanthrene
In order to elucidate the bioavailability of the tar oil-
related phenanthrene in the, at least 53-year-old, soil
contamination, the soil material was spiked with
excluded. Nevertheless, the mean recovery of applied
1
3C was 83% and may be sucient for a carbon budget.
The partition of the 13C-label among various fractions
normalised to the applied amount was ananysed at
day 82. Seventy-three per cent of the 13C-label was
[1- C]phenanthrene and the concentration of both phe-
13
nanthrene species was monitored in the bioreactor. The
tar oil contamination consists of typical tar oil-related
PAH like phenanthrene, anthracene, chrysene, benzo-
(a)anthracene and methylated homologues, whereas
¯uoranthene and pyrene are predominant compounds
found as CO . Four per cent was extractable of which
2
very minor amounts were positively identi®ed as the
1
3
parent [1- C]phenanthrene (0.03%) and 6% was trans-
formed into bound residues. Compared to bound
(Fig. 3). Large amounts of extractable tar oil (15±25 mg
� 1
13
residue formation of [9- C]anthracene (11%) and
1
g ) characterise a signi®cant non-aqueous liquid phase
in this soil contamination. The initial concentration of
4
[9- C]-anthracene (20%) in pristine soils (Richnow et
al., 1999b) and bioremediation of [9- C]phenanthrene
14
� 1
tar oil-borne phenanthrene in the soil was 88.6 mg g
which was spiked with 354.3 mg g
nanthrene (Table 2). During the time course of the
�
1 13
(
(
36%) with white rot fungi in a contaminated soil
Brotkorb and Legge, 1992) the residue formation in
C-labelled phe-
our experiment was signi®cantly lower. The relatively
low content of bound residues might be a good indi-
cation for the ability of the soil micro¯ora to degrade
phenanthrene eciently. In earlier studies we observed a
low tendency to form bound residues in the case of a
high mineralisation capacity of the soil micro¯ora
degradation experiment, phenanthrene species were
�
1
degraded eciently compared to 0.23 mg g unlabelled
�
1
and 0.11 mg g labelled phenanthrene, respectively. The
degradation rates were calculated as the percentage of
decreasing phenanthrene concentrations per day in the
extractable fraction (Table 2). During the ®rst 14
�
1
(
Richnow et al., 1999b). In this experiment a short lag
days, mean degradation rates of about 5.4% day
phase for the phenanthrene degradation and a high
mineralisation capacity point to a micro¯ora already
adapted to degrade the contamination.
were observed for both phenanthrene species (Fig. 4).
Between 35 and 63 days tar oil-related phenanthrene of
the old contamination was degraded in the same order

H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
97
�
1
� 1
of magnitude between 3.9 and 6.5% day as the spiked
1
lower than 0.2 mg g , the tar oil-related phenanthrene
was degraded at a very small rate (0.03% day ) and
remained as a relatively stable species in the soil. The
3
� 1
[
1- C]-phenanthrene. Thus, during the degradation of
the bulk material, both compounds were degraded at
quite similar rates. Only at minor concentrations, e.g.
13
spiked [1- C]phenanthrene was degraded further but
the degradation rates also slowed down at very low
concentrations between day 63 and 82 (Table 2). Thus,
the degradation rates of spiked 13C-phenanthrene and
the tar oil-borne phenanthrene were in the same order
of magnitude in the concentration range between 350
and 0.3 mg g 1. Only at very low concentrations we did
observe a lower degradation rate of tar oil-related phe-
nanthrene. This result was striking as the old tar oil
contamination was aged at least for about 53 years, and
ageing is believed to reduce the bioavailabilty of con-
taminants (Hatzinger and Alexander, 1995). In model
experiments with aged 14C-phenanthrene, a reduced
mineralisation of phenanthrene was observed in the
presence of non-aqueous phase liquids, clay minerals
and natural soil organic matter. The reduced miner-
alisation was interpreted as a slow desorption of the
aged pollutants into the water phase of soil, which is
suggested to be necessary to make the pollutants bioa-
vailable (Efroymson and Alexander, 1995). Compared
to our approach, these authors have used much lower
concentrations of phenanthrene in their experiments. At
higher concentrations, which may be more realistic for
many tar oil-polluted soils, ageing is less pronounced.
Soils may have a limited capacity to sequester hydro-
phobic pollutants, and in the case of high concentrations
free sorption sites are not available. Ageing processes
are likely to be dependent on the concentration of the
pollutant and the soil properties.
�
Our results suggest that bioavailability in heavy tar
oil contaminated soils may be less e€ected by ageing as
previously expected and a limited biodegradation of
pollutants on bioremediation sites is not necessarily a
soil matrix problem. Furthermore, the addition of
Fig. 3. [1-13C]Phenanthrene degradation in a tar oil-contaminated soil
matrix. Gas chromatograms of the solvent-extractable fraction. Ant,
anthracene; Ph, phenanthrene.
Fig. 4. Phenanthrene degradation rates. (Dotted bars=tar oil-borne phenanthrene; white bars=[1-13C]phenanthrene.

9
8
H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
the 13C-labelled substances is a very useful method to
measure bioavailability of contaminants to elucidate the
remediation capacity of a soil.
In our case the micro¯ora in the contaminated soil
was already adapted to degrade the contaminants as the
lag-phase was relatively short compared to other degra-
È
dation experiments with PAH (Kastner et al., 1995;
Richnow et al., 1999b). The supplementation with
compost, aeration, and the increase of the pH from 4.8
to 6.3 have stimulated the microbial capacity to degrade
bulk phenanthrene that was already aged for more than
53 years.
3.3. Chemical structure of bound residues
To investigate the nature of bound residues the pre-
extracted soil was hydrolysed and the products analysed
by IRM±GC±MS to screen the content of 13C-labelled
metabolites. The chemical structure of products was
subsequently investigated by conventional GC±MS
comparing mass spectra, and additionally by co-injec-
tion of authentic reference compounds.
The products of the alkaline hydrolysis of the bound
residue fraction at day 50 were analysed by IRM±GC±
MS (Fig. 5). 13C-labelled products produce a very
strong signal in the chromatogram depicting the mass
1
3
12
ratio m/z 45/44 ( CO / CO ) due to the relatively high
2
2
abundance of 13C in labelled products. Although the
absolute concentration of labelled products may be low,
the mass ratio m/z 45/44 clearly indicates the presence
of labelled metabolites. Major compounds in the
hydrolysis fraction are long-chain carboxylic acids
deriving from the natural soil organic matter without
any enrichment of 13C. Possible precursors are fatty
acids from plants which have been incorporated into
the soil organic matter during humi®cation processes.
Methoxylated derivatives of benzoic acids may be
derived from lignin degradation products which con-
tribute to the aromatic domain of soil humic substances.
A number of 13C-labelled compounds were detected in
the hydrolysates. The chemical structure of most of the
major constituents could be identi®ed. This includes the
2
Fig. 5. m/z 44 shows an ion chromatogram of CO generated by
IRM±GC±MS of the products obtained by alkaline hydrolysis of
bound residues. IS, internal standard, major compounds are long-
chain carboxylic (v) and dicarboxylic acids (o). Numbers indicate the
carbon atoms of the chains.
react with soil organic matter and mayget incorporated
into humic material during natural humi®cation pro-
cesses in soil. Some constituents such as the parent
phenanthrene might be sequestered within micropores
leading to a limited extractability. Functionalized meta-
bolites such as acids and hydroxylated compounds
might be bound to a signi®cant extent via ester bonds to
the macromolecular organic matrix in the soil which has
been demonstrated to be a signi®cant mode of binding
in bound residues formation during biodegradation of
selected PAH in pristine soil (Richnow et al., 1994).
Moreover, aromatic alcohols such as phenanthrenols
and phenanthrenediols might also be linked by oxida-
tive coupling via ether and carbon±carbon bonds and
contribute to bound residues. These linkages are not
amenable to the alkaline degradation technique and the
characterisation of ether and carbon±carbon bonds was
not covered by our analytical approach.
1
3
parent [1- C]phenanthrene, typical primary oxidation
products such as hydroxy- and dihydroxyphenan-
threnes, a number of ring cleavage products including
1,2-dicarboxynaphthalene and hydroxynaphthoic acid.
Moreover, phthalic acid and salicylic acid represent
typical metabolites of the second ring cleavage. Beside
these major metabolites, a signi®cant number of non-
identi®ed labelled metabolites were present and con-
tribute to the total amount of bound residues. The
major metabolites in the hydrolysates of the residues are
well-known metabolites of the aerobic phenanthrene
degradation described in liquid cultures with Pseudo-
monas strains (Evans et al., 1965). Therefore, metabo-
lites of the aerobic phenanthrene degradation might

H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
99
3
.4. Transformation into biological compartments
acids, carbohydrates, glycerols and other cell con-
stituents. Although we have not analysed these cell
constituents, we assume that a signi®cant amount of
contaminant-derived carbon was bound within these
fractions.
The soil microorganisms may use organic xenobiotica
as a source of carbon and transform the xenobiotic
carbon in anabolic processes into typical chemical
structures of the biomass. Up to 80% of the bacterial
biomass consists of protein, which are built up by pro-
teineous amino acids (Gottschalk, 1986). To elucidate
the microbial transformation of the 13C-label of phe-
nanthrene into amino acids in soil, we hydrolysed the
solvent-extracted soil and analysed the isotopic compo-
sition of hydrolysable amino acids with IRM±GC±MS
Our approach did not allow the distinction between
living and dead biomass. The contribution of the bac-
terial-derived living biomass is considered to make up
very small parts of the total organic matter in soils (1±
5%) although bacteria transform major amounts of
organic carbon in soils (Oades, 1995). Therefore, large
parts of bacteria-derived organic substances may belong
to the dead biomass. A good indication that parts of the
amino acids belong to non-living organic matter in soils
is the relatively high concentration of labelled g-amino
butyric acid. g-Amino butyric acid is a non-proteineous
amino acid which is formed during the degradation
of proteinous organic matter in sediments (Lee and
Cronin, 1982; Ittekott et al., 1984). The microbial
decarboxylation of glutamic acid leads to the formation
of g-amino butyric acid (Meister, 1965). Therefore,
labelled g-amino butyric acid may well indicate the
decay of dead cells which had have transformed
phenanthrene related carbon into cell constituents. The
dead biomass may provide a food source for other soil
organisms. During this process, parts of the dead
microbial biomass are incorporated into soil organic
matter during natural humi®cation processes in the
natural carbon cycle.
In the light of the IUPAC de®nition, non-extractable
bound residues of highly metabolised xenobiotic carbon
in a chemical structure identical to biological com-
pounds is not considered as bound residues. In most
classical transformation studies with 14C-labelled
PAH, the structural assignments of non-extractable
bound residues in soils and sediments were not carried
out. Therefore, limited information about the nature
of soil-bound residues was obtained. Our results
clearly indicate that xenobiotic carbon transferred into
(
Table 3). In the background experiment with non-
labelled phenanthrene, the isotopic signature of amino
acids was between � 14 and � 26% (PDB) which is
similar to the isotopic composition of the total organic
carbon (� 23.5% PDB) and the non extractable organic
carbon in the soil (� 25.7% PDB). The hydrolysable
amino acids of the bound residue fraction revealed an
isotopic composition between � 3.1 and � 10.4% (PDB)
13
which implies a signi®cant enrichment in C. The only
source of 13C-enriched carbon in the bioreactor was the
1
3
labelled [1- C]phenanthrene, which was incorporated
into amino acids of the soil biomass via anabolic pro-
cesses. The total amount of 13C derived from 13C-phe-
nanthrene transferred into hydrolysable amino acids
�
1
accounted for 3.1 mg g phenanthrene equivalents in
the extracted soil. This amount is equivalent to 11% of
the total bound residues present in the soil at day 50.
Thus, the soil-bound amino acids receive a signi®cant
amount of 13C. However, we have analysed only
hydrolysable amino acids in the soil. The total con-
tribution of 13C derived from phenanthrene bound
within amino acid derivatives may be larger because
amino acids undergo humi®cation reactions in the
soil forming non-hydrolysable condensation products
which are not amenable to our method. Moreover, our
results imply, that bacterial anabolic metabolism may
incorporate contaminant derived carbon into fatty
Table 3
Isotopic composition of hydrolysable amino acidsa
Amino acid
Background d13C
% PDB)
50 days (SL3)
�
d
13C
Phenanthrene (eq)
� 1
1
(
(mmol g soil)
(% PDB)
(ng g soil)
Alanine
Glycine
� 17.1
� 18.5
� 17.0
� 20.6
� 24.4
� 23.9
� 17.2
� 14.1
� 14.0
� 26.5
� 19.9
2.38
3.22
10.00
1.26
1.65
0.86
1.44
2.84
2.50
0.71
0.35
� 6.3
� 6.0
� 6.7
� 8.0
� 9.2
� 10.4
� 3.3
� 9.3
� 3.1
� 7.2
� 3.3
178.5
186.2
957.9
184.9
350.4
160.8
231.9
125.4
315.2
286.6
81.0
g-Amino butyric acid
Valine
Leucine
Isoleucine
Proline
Aspartic acid
Glutamic acid
Phenylalanine
Lysine
a
The isotopic composition is corrected for the derivative.

1
00
H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
the biomass by anabolic processes in the course of
biodegradation is signi®cant. These types of residues
clearly possess no ecological risk related to the parent
compound of concern and may often mimic bound
residues in 14C-studies.
®nancial support from the Deutsche Forschungsgem-
einschaft in the framework of SFB-188 (Reinigung
È
kontaminierter Boden).
References
Bachmann, W.E., Cole, W., Wilds, A.L., 1940. The total synthesis of
the sex hormone equiline and its stereoisomers. J. Am. Chem. Soc.
62, 824±839.
4. Conclusions
Berry, D.F., Boyd, S.A., 1985. Decontamination of soil through
enhanced formation of bound residues. Environ. Sci. Technol. 19,
1132±1133.
13C-labelled PAH are useful tracers to elucidate the
transformation of PAH in soil on a bulk and on a
molecular level. They may also be applied to investigate
the transformation of PAH in tar oil-contaminated
soils. The carbon budget revealed a relatively fast and
ecient mineralisation of the tracer substance after the
adjustment of the pH and a sucient oxygen supply.
The tendency to form bound residues was lower com-
pared to former studies, which is a result of the high
intrinsic capacity of the soil micro¯ora to degrade the
contaminant.
13C-labelled PAH can be used to assess the ageing
e€ects on the bioavailability of contaminants in old soil
contaminations. We observed no signi®cant di€erences
in the bioavailability of the tar oil-related phenanthrene
in the 53-year-old contamination compared to the
spiked 13C-tracer over a large range of concentration.
Only at minor concentrations were ageing e€ects evi-
dent. During the degradation of the bulk of phenan-
threne, no signi®cant ageing e€ects reducing the
bioavailability of tar oil-related phenanthrene were
observed. In our case, ageing processes sequestering the
parent material in the soil were not relevant for phe-
nanthrene present in a concentration range of environ-
mental concern. The matrix e€ects of tar-related non-
aqueous phase liquids and humic substances were neg-
ligible in our case. Further investigations with di€erent
types of aged real contaminated material will have to
elucidate the e€ects of non-aqueous phase liquids and
macromolecular natural organic matter in soil under
real conditions in more detail.
Bollag, J.M., Loll, M.J., 1983. Incorporation of xenobiotics into soil
humus. Experientia 39, 1221±1231.
Brotkorb, T.S., Legge, R.L., 1992. Enhanced biodegradation of phe-
nanthrene in oil-tar contaminated soils supplemented with Phaner-
ochaete chrysosporium. Appl. Environ. Microbiol. 58, 3117±3121.
Brown, H.C., Krishnamurthy, S., 1979. Forty years of hydride reduc-
tions. Tetrahedron 35, 567±607.
Calderbank, A., 1989. The occurrence and signi®cance of bound pes-
ticide residues in soil. Rev. Environ. Contam. Toxicol. 108, 71±103.
È
È
Dawel, G., Kastner, M., Michels, J., Poppitz, W., Gunther, W.,
Fritsche, W., 1997. Structure of a laccase-mediated product of cou-
pling of 2,4-diamino-6-nitrotoluene to guiacol, a model for coupling
of 2,4,6-trinitrotoluene metabolites to a humic organic soil matrix.
Appl. Environ. Microbiol. 63, 2560±2565.
Dawson, R., Liebezeit, G., 1983. Determination of amino acids and
carbohydrates. In: Grashof, K., Ehrhardt, M., Kremling, K. (Eds.),
Methods of Seawater Analysis, 2nd Edition. Verlag Chemie, Wein-
heim, pp. 319±342.
DIN 38404-65. Deutsche Industrie Norm.
Efroymson, R.A., Alexander, M., 1995. Reduced mineralization of
low concentration of phenanthrene because of sequestering in non-
aqueous-phase lipids. Environ. Sci. Technol. 29, 515±521.
Evans, W.C., Fernley, H.N., Griths, E., 1965. Oxidative metabolism
of phenanthrene and anthracene by soil pseudomonads. Biochem. J.
95, 831±891.
Fu, P.F., Harvey, R.G., 1978. Dehydrogenation of polycyclic hydro-
aromatic compounds. Chem. Rev. 78, 317±361.
Gottschalk, G., 1986. Bacterial Metabolism, 2nd Edition. Springer,
New York.
Hatcher, P.G., Bortiatynski, J.M., Minard, R.D., Dec, J., Bollag,
J.M., 1993. Use of high resolution 13C NMR to examine enzymatic
covalent binding of 13C-labelled 2,4-dichlorophenol to humic sub-
stances. Environ. Sci. Technol. 27, 2098±2103.
Hatzinger, P.B., Alexander, M., 1995. E€ect of aging of chemicals in
soil on their biodegradability and extractability. Environ. Sci.
Technol. 29, 537±545.
Part of the xenobiotic carbon is transformed into
biomass-related compounds, such as amino acids via
anabolic processes. An incorporation of carbon from
the pollution into other cell components is likely. This
part of the non-extractable residues is not considered as
bound residue in the light of the IUPAC de®nition.
Without a precise structural investigation on a mole-
cular level, the bound residue formation may lead to
misleading conclusions.
Hayes, J.M., Freeman, K.H., Popp, B.N., Hoham, C.H., 1990. Com-
pound-speci®c isotope analyses: a novel tool for reconstruction of
ancient biogeochemical processes. Org. Geochem. 16, 1115±1128.
Hoefs, J., 1997. Stable Isotope Geochemistry. Springer, Berlin.
È
Hupe, K., Luth, J.C., Heerenklage, J., Stegmann, R., 1996. Enhance-
ment of the biological degradation of soil contaminated with oil by
the addition of compost. Acta Biotechnol. 16, 19±30.
Ittekott, V., Deuser, W.G., Degens, E.T., 1984. Seasonality in ¯uxes of
sugars, amino acids and amino sugars to the deep ocean: Saragasso
Sea. Deep-Sea Res. 31, 1057±1069.
È
Kastner, M., Lotter, S., Heerenklage, J., Breuer-Jammali, M., Steg-
mann, R., Mahro, B., 1995. Fate of 14C-labeled anthracene and
hexadecane in compost manured soil. Appl. Microbiol. Biotechnol.
Acknowledgements
43, 1128±1135.
Koning, M., Hupe, K., Lu
È
th, J.-C., Cours, J., Quandt, C., Stegman,
Pierre Albrecht and Patrick Wehrung (University of
Strasbourg) generously provided analytical facilities
for IRM±GC±MS measurements. We are grateful for
R., 1998. Comparative investigations into biological degradation of
contaminants in ®xed bed and slurry reactors. In: Contaminated
Soil. Thomas, London, pp. 531±538.

H.H. Richnow et al. / Environmental Pollution 108 (2000) 91±101
101
Lee, C., Cronin, C., 1982. The vertical ¯ux of particulate organic
nitrogen in the sea: decomposition of amino acids in the Peru
upwelling area and equatorial Atlantic. J. Mar. Res. 40, 227±251.
Mallan, J.M., Bebb, R.L., 1969. Metalations by organolithium com-
pounds. Chem. Rev. 69, 693±755.
Organic pollutants associated with macromolecular soil organic
matter and the formation of bound residues. In: Block, J.C.,
Baveye, P., Goncharuk, V.V. (Eds.), Bioavailability of Organic
Xenobiotics in the Environment (NATO ASI Series). Kluwer Aca-
demic Publishers, Netherlands, pp. 297±326.
Marshall, J.A., 1969. Photosensitized ionic additions to cyclohexenes.
Acc. Chem. Res. 2, 33±40.
È
Richnow, H.H., Eschenbach, A., Mahro, B., Kastner, M., Annweiler,
E., Seifert, R., Michaelis, W., 1999b. The formation of non-
extractable soil bound residuesÐa stable isotope approach. Envir-
onm. Sci. Technol. (in press).
Meister, A., 1965. Biochemistry of Amino Acids, 2nd Edition, Aca-
demic Press, New York.
Oades, J.M., 1995. An overview of processes e€ecting the cycling of
organic carbon in soils. In: Zepp, R.G., Sonntag, Ch. (Eds.), Role
of Nonliving Organic Matter in the Earths Carbon Cycle. Wiley,
Chichester, pp. 293±303.
Roberts, T.R., 1984. Non-extractable Pesticide Residues in Soil and
Plants (IUPAC Reports on Pesticides No. 17). Pure Appl. Chem.
56, 945±956.
Silfer, J.A., Engel, M.H., Macko, S.A., Jumeau, E.J., 1991. Stable
carbon isotope analysis of amino acid enantiomers by conventional
isotope ratio mass spectrometry and combined gas chromato-
graphy/isotope ratio mass spectrometry. Anal. Chem. 63, 370±374.
Soderquist, A., Hughes, C.D., Horton, W.J., Facelli, P., Grant, J.C.,
1992. Carbon-chemical shifts tensors in aromatic compounds. 3.
Phenanthrene and triphenylene. J. Am. Chem. Soc. 144, 2826±2833.
Wiley, G.A., Hershkowitz, R.L., Rein, B.M., Chung, B.C., 1969. Studies
in organophosphorous chemistry. I. Conversion of alcohols and phe-
nols by tertiary phosphine dihalides. J. Am. Chem. Soc. 86, 964±965.
È
Richnow, H.H., Seifert, R., Hefter, J., Kastner, M., Mahro, B.,
Michaelis, W., 1994. Metabolites of xenobiotica and mineral oil
constituents linked to macromolecular organic matter in polluted
environments. Org. Geochem. 22, 671±681.
Richnow, H.H., Eschenbach, A., Seifert, R., Wehrung, P., Albrecht,
P., Michaelis, W., 1998. The use of 13-C-labelled polycyclic aro-
matic hydrocarbons for the analysis of their transformation in soils.
Chemosphere 36, 2211±2224.
È
Richnow, H.H., Annweiler, E., Fritsche, W., Kastner, M., 1999a.