United States
Environmental Protection
Agency
Environmental Research
Laboratory
Athens GA30613
Research and Development
EPA/600/S3-89/080 May 1990
&EPA Project Summary
Structure-Activity
Relationships and Estimation
Techniques for
Biodegradation of Xenobiotics
Susan A. Moore, John D. Pope, John T. Barnett, Jr., and Luis A. Suarez.
The current status of structure-
activity relationships for the
biodegradation of xenobiotics is
reviewed. Results are presented of a
pilot study on biodegradation
constants obtained from computer
databases. New analyses for a
relatively large number of anilines
and phenols are presented in which
the kinetic constants for
biodegradation successfully
correlate with the pKa's of the
ionizing groups. The use of molecular
connectivity indices is reviewed.
These indices have broad application
over a wide array of chemical classes
in structure-biodegradability
relationships, and they have the
benefit of being purely calculated
parameters. It is proposed that the
biodegradability of complex
molecules and polymers containing
labile R-X-R' linkages may be
accurately estimated based on the
biodegradability of their component
parts, where X is one or more
heteroatom. Estimation techniques
are reviewed with respect to the
kinetic processes that are associated
with biodegradation.
This Project Summary was
developed by EPA's Environmental
Research Laboratory, Athens, GA, to
announce key findings of the research
project that is fully documented in a
separate report (see Project Report
ordering information at back).
Introduction
The EPA reviews more than 2000
premanufacture notices each year to
identify ecological and human health
effects and exposure. The consideration
of biodegradability of these chemicals is
important to the exposure assessment.
Under the Toxic Substances Control Act
(TSCA), the EPA also must determine the
risk from exposure to the thousands of
chemicals already in commerce. These
and other EPA activities create a need for
structure-activity correlations that
describe and predict the biodegradability
of chemicals in natural environments
(Boethling and Sabljic, 1989).
In drug design research, quantitative
structure-activity relationships (QSARs)
have undergone their greatest advances
among biological systems. For example,
the QSAR for antimalarial drugs in whole
mice is shown in eqn 1 (Hansch, 1987). C
is the moles of drug per kilogram of body
weight needed to cure malaria in mice.
Equation 1 is based on 646 compounds,
including ring-substituted 2,6-
diphenylpyridines, 2-phenylquinolines,
and phenanthrene carbinols (n = 646; r2
= 0.806; s = 0.309).
log 1/C= 0.56So + 0.17Sir + 0.17logP
-0.019(logP)2 + 2.69
-0.17(c-side) + 0.32CNR2
-0.14AB - 0.8<3-cures
+ 0.28(MR-4'-Q) (1)
+ 0.25(Me-6,8-Q)
+ 0.081 (2-Pip)
+ 0.17NBrPy-0.67Q2P378
+ 0.27Py
The basic QSAR of eqn 1 includes five
terms, where o refers to the Hammett
constant for electronic character, IT refers
to the hydrophobicity of the substituents,
logP refers to the hydrophobic interaction
-------
of the whole molecule, and 2 indicates
that the ir and o terms are summed for
multiple substituents. The QSAR is then
modified beyond these five terms by
another ten terms that are associated
with secondary or side-chain substituents
for specific parent compounds. This
analysis shows the types of terms that
are needed to obtain a broadly based but
successful QSAR.
The QSAR is a mathematical
representation of the physical structure of
the active site of a protein (or a series of
proteins) that is involved in the rate-
determining step(s) of a pathway. The
process of applying the QSAR analysis to
biodegradation studies is relatively new.
Such equations have been dubbed
QSBRs (quantitative structure
biodegradability relationships).
Computer Databases
A list of computer databases that are
available to the EPA for structure-activity
analyses is presented. The best
databases on biodegradation appear to
be CHEMFATE and BIODEG (Syracuse
Research Institute, Syracuse, N.Y.).
Summaries of 301 literature articles from
the CHEMFATE and BIODEG databases
were reviewed for 19 compounds.
Enough information was found in 217 of
these to allow calculation of the half-time
for biodegradation (t50) by means of eqns
2 and 3.
tso = LAG + t1/2
tia-- 0.693
(2)
(3)
Equation 2 assumes the presence or
absence of a lag followed by first order
kinetics of disappearance. Equation 3
gives the ti/a for the first order portion
based on. the fraction of xenobiotic
remaining (f0ts) at a given time (t0bs). An
abbreviated summary of the mean and
standard deviations of the half-times thus
obtained is given in Table 1. Conclusions
from this preliminary analysis are as
follows.
In general, the mean half-time
decreases per environment as: SEWAGE
** ACTIVATED SLUDGE < FRESH
WATER < SOIL < MARINE WATER
SEDIMENT < MARINE WATER,
although the order varies somewhat
depending on the chemical. Surprisingly,
the mean half-time for biodegradation in
aerobic fresh water is typically within 2-
fold of that measured in activated sludge.
This difference is smaller than commonly
assumed. Biodegradation has not been
measured across multiple environments
for a significant fraction (11 out of 19) of
the compounds in the initial survey.
However, for most of the compounds for
which it could be calculated (7/9), the
half-time quantitated over multiple
environments (called t50 (QUME)) was
within 5-fold identical to that for one of
the fastest biodegrading environments,
acclimated fresh water.
The variation in half-times among
literature articles was 5-fold on the
average in going from plus one to minus
one standard deviation of the mean. The
largest variation was 15-fold for aniline in
fresh water over 9 literature articles. This
variation is much smaller than commonly
assumed and suggests that mean half-
times determined from computer
databases may be useful in establishing
QSBRs.
QSBR to Date
Table 2 shows an abbreviated list of
QSBR equations that were found in the
literature or established by the authors.
Parameters such as the rate constant
(KOH) or half-time (t50OH) for alkaline
hydrolysis, as well as the van der Waal
radius (Yvdw). octanol/water partition
coefficient (Kow). and Hammett constant
(o) have been correlated with
biodegradability, but are either difficult
and expensive to measure, or may be
applicable only within a relatively small
series of compounds. It is best for EPA
purposes to base QSBRs on easily
measured or purely calculated
parameters, if possible. Some of these
parameters are becoming easier to
determine. For example, KOw has been
calculated from retention times in high
pressure liquid chromatography.
Infrared (IR) spectra contain a wealth of
substructural information in the form of
peak frequencies and intensities. Aspects
of this information have been found to
correlate with rate constants for
biodegradation (Steen and Collette,
1989). An advantage of using IR in
QSBRs is that the spectral data can be
measured in a rapid, precise, and
inexpensive manner.
Correlation of biodegradation constants
with pKa is limited by the fact that it can
be applied only to classes of xenobiotics
that contain dissociating protons.
However, it has the advantage that pKa
values do not need to be measured.
Existing tables of pKa values are
extensive, and computer software
programs have been developed that allow
accurate calculation of pKa. Figure 1 was
established in this study. It shows the
correlation of the biodegradation data of
Pitter (1976,1984) with pKa for anilines
and phenols. Kinetic constants associated
with the initial rate of transformation of
anilines to acetanilide by a Rhodotorula
glutinis yeast isolated from river water
was also found by the authors to
correlate with pKa (Figure 2).
Molecular connectivity indices are
calculated parameters based on the
molecular structure of a molecule. These
parameters have been successfully
correlated with the biodegradation
constants of a number of classes of
chemical compounds including esters,
carbamates, ethers, alcohols and acids
(Boethling, 1986). Molecular connectivity
indices may currently be the best
parameters to correlate with
biodegradation constants for EPA
purposes because the indices are highly
sensitive to even small variations in
structure, widely applicable, and
calculated rather than measured.
Semi-Quantitative Relationships
Algorithms have been devised that
semi-quantitatively predict whether a
compound is biodegradable (e.g. tso <
15 days) or persistent (e.g. t50> 15
days), based on calculated parameters
such as molecular connectivity indices,
and using biodegradation data on 250 to
357 chemicals (Niemi et al., 1987;
Enslein et al., 1984). Impressively, the
algorithms typically displayed ^90%
accuracy in back calculating the
biodegradability of the compounds that
were used to derive the algorithm.
Approximately 10% of the compounds
were not correctly assessed.
Such algorithms are based on the
frequencies at which substructural
fragments occur in the chemicals of the
database. In order to accurately predict
the biodegradability of new chemicals,
these frequencies must be accurate for
each of the substructural fragments in the
new chemicals of interest. This may be
unlikely because such algorithms can
change when as little as 2% of the
compounds are removed from the
database (T. Collette, personal
communication). Such an analysis points
out the difference between obtaining a
successful correlation, and the more
difficult task of developing an accurate
predictor of biodegradability. In
quantitative structure activity relationships
(QSBRs), predictions are limited to
compounds of similar structure. However,
the requirement for similarity appears to
increase the accuracy of the predictions.
-------
Table 1. Examples of Biodegradation Half-Times (Days) Calculated from Computer Databases
Chemical
Acetanilide
Acrylonitrile
Aniline
(Bis) 2-
Chloroethyl
Ether
Soil
-
—
157 ± 49
>60
Sewage
8 ± 2
431"*
16 + 13*
4.6 + 2.5
"
Act. Sludge
6+5
18 ± 1"*
2A
4.6 ± 3.1
>40
Fresh Water
21 ± 9
22 ± 10UA
4+1*
8+7
120
Fresh Marine
Water/Sed. Marine Water Water/Sed.
..
345U*
..
p-Cresol
0.5
• 4.6 + 0.1
13 ± 2
-------
E + S
kcaU
E-S *"""> E + P
Jki (7)
E (inactive)
Krtvln(St/S0) + S,-S0-
kcat'C° (e-w-t-1) = 0
ki
(8)
So/Km
10-
SolKm
Figure 4. Linear and semilog plots for
substrate disappearance by
theoretical Michaelis-Menten
kinetics.
[S] < < Km, the observed rate constant
(kobs) is equal to Vmax/Km, and k0bs is
independent of substrate concentration.
However, these conclusions can be, and
have been, made incorrectly when the
reaction is determined to be first order
from partial, rather than complete, time-
course kinetics. Figure 4B shows at So
> Km, the kinetics of substrate
disappearance bend downward on the
semi-log plot, and the kobs determined
from the slope at early times differs from
that determined at late times. Figure 5
shows the deviation from the true
Vm/Km, of k0tjs obtained from partial
time-course kinetics. The presence of a
toxicity term in the Michaelis-Menten
equation can also skew the data to look
first order, where it can be wrongly
concluded that kobs equals Vmax/Km
and is independent of substrate
concentration.
The observation of (apparent) lags in
the kinetics of biodegradation is, common
but difficult to predict. A lag can be due
to the time required for induction of
enzyme, or growth of biodegrading cells.
Care must be taken before concluding
that a lag is in fact present. Figure 4A
shows theoretical curves for Michaelis-
Menten kinetics on a linear plot of [S]
versus time where no lag is present.
These same curves are plotted in semi-
log fashion in Figure 4B. At high [S], it
appears that a lag occurs because the
log scale decreases the slopes at high
compared to low [S].
When cells use a xenobiotic as a
nutrient to support cell growth, Monod
kinetics apply (eqn 9), where p. is the
specific growth rate of cells in units of
reciprocal time. Biodegradation in this
case often involves complete oxidation of
the hazardous chemical to CO2 and toxic
intermediates rarely accumulate. The
disappearance of substrate by Monod
kinetics (Simkins and Alexander, 1984)
and the corresponding increase in cell
number are described by equations 10
and 11, respectively (see next page).
= Umax[S]/(Ks + [S])
(9)
Figure 6 depicts one set of theoretical
curves for substrate disappearance and
cell number increase where Monod
Growth kinetics apply (qC0 < < S0) and
where S0 < Ks (Ks = 100 mM, S0 = 10
mM, C0 = 10 cells/mL, 9 = 0.001
mM/(cells/mL), and p^, = 0.1 min-i). Cmax
is the maximum concentration of cells
reached when all of the substrate has
been consumed, and 9 is the amount of
substrate required to generate one new
cell. It is shown that the time at which the
cell concentration reaches one-half Cmax
is identical to the time at which the
substrate concentration reaches one-half
S0 (Figure ,6). It is also shown that the cell
number doubling time at early times
(9obs) 's identical to the half-life of
substrate disappearance at late times
(t-i/2) (Figure 6). These kinetic identities
are proposed to be useful diagnostic
tools for the demonstration of
biodegradation by Monod kinetics where
it is not feasible to directly demonstrate
14CO2 production using i4Olabeled
xenobiotic.
References
1. Boethling, R.S., and Sabljic, A.,
Environ. Sci. Technol., 23, 672-679,
(1989).
2. Boethling, R.S., Environ. Toxicol.
Chern., 5, 797-806, (1986).
3. Enslein, K., Tomb, M.E., and Lander,
T.R., in Quantitative Structure Activity
Relationships in Environmental
Toxicology, (Kaiser, K.L.E.,ed.), D.
Reidel Publ., Dordrecht, Netherlands,
pp. 89-109, (1984).
4. Hansch, C., in "Molecular Structure
and Energetics: Biophysical
Aspects,"Vol. 4, Liebman, J.F., and
Greenberg, A., eds., VCH, New York,
p. 341-379, (1987).
5. Kol|ig, H.P., Toxicol. Environ. Chem.,
77,287-311, (1988).
6. Niemi, G.J., Veith, G.D., Regal, R.D.,
and Vaishnav, D.D., Environ. Toxicol.
Chem., 6, 515-527, (1987).
7. Pitter, P., Collect. Czech. Chem.
Commun., 49 2891-2896, (1984).
8. Pitter, P., Water Res., 10, 231-235
(1976).
9. Simkins, S., and Alexander, M., Appl.
Environ. Microbiol., 47, 1299-1306,
(1984).
10. Steen, W.C., and Collette, T.W., Appl.
Environ. Microbiol., 55, 2545-2549,
(1989).
-------
t =
(Ks + S0 + qC0)ln[(S0 + qC0 - St)/gC0] + Ks-ln(So/St)
+ qC0)
(10)
(Ks + S0 + qC0)ln(Ct/C0) + Ks-ln[S0/(S0 + qC0 -qCt)]
Hmax(S0 + qC0)
(11)
Mobs) Baser! on:
3 Half-Lives
\(87.5% Reaction)
1 Half-Life
(50% Reaction)
0.01
0 20 40 60 80 100,
So/Km
Figure 5. Variation in the observed rate
constant relative to the true
Vmax/Km at various values of
So/Km.
10'
1/2 Cmax = 1/2 So
I
400 800
Time (mm)
1200
10
Figure 6. Monod kinetics of substrate
disappearance (dashed line) and
cell number increase (solid line).
-------
The EPA authors, Susan A. Moore (also the EPA Project Officer), John D. Pope,
John T. Barnett, Jr., and Luis A. Suarez are with the Environmental Research
Laboratory, Athens, GA 30613.
The complete report, entitled "Structure-Activity Relationships and Estimation
Techniques for Biodegradation of Xenobiotics," (Order No. PB 90-149 357/AS;
Cost: $23.00, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Research Laboratory
U.S. Environmental Protection Agency
Athens, GA 30613
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S3-89/080
------- |