by anaerobic O-demethylation under sub-
strate-limited conditions (Fig. 1B and Fig. 2).
This model is consistent with the geologic
context of these materials. The least-mature
Belchatow lignites underwent early coalifi-
cation in a shallow setting unfavorable for
establishing strongly reducing conditions
(<250 m of clastic overburden in a tectonic
graben; materials and methods, section S1),
which would have inhibited anaerobic metab-
olisms. The onset of microbial CBM produc-
tion tends to coincide with a thermal maturity
of 0.3% Ro—just beyond that of these lignites
[0.26 to 0.29% Ro( 24 )] (Fig. 2). Thus, Belcha-
tow lignites likely record only initial O-
demethylation, which probably occurred
during peat formation within a few meters
of the surface and approximately tens of
thousands of years after deposition, consistent
with previous findings ( 7 ). In more-mature,
deeply buried coals of approximately the same
age (early Miocene), methoxyl abundances
drop by another 10 to 100×, which suggests a
later proliferation of anaerobic O-demethylation.
This extends the conditions over which micro-
organisms directly modify humic materials
to depths of 2 km, temperatures of 50°C in
subbituminous coals, and therefore on time
scales of tens of millions of years.
At low methoxyl concentrations (0.05 to
2% remaining),d^13 Cmethoxylvary by 25‰.Itis
unlikely that this scatter is primarily the re-
sult of source variability. Woodd^13 Cmethoxylvary
by at most 10‰(table S6). Leafd^13 Cmethoxylare
up to 45‰lower than wood; however, most
leaf methoxyl groups are in labile compounds
unlikely to survive peatification, wood con-
tains approximately five times more meth-
oxyl groups by dry mass, and woody tissues
are the main feedstock for these coals ( 7 , 25 )
(materials and methods, section S1, and fig. S11).
Instead, we speculate that methoxyl isotopic
variability results from differences in the phys-
ical encapsulation of lignin within the coal
matrix (e.g., armored in intact macerals ver-
sus disseminated in groundmass), which var-
iably restricted the ability of extracellular
enzymes to access the substrate, and is cap-
tured in our simple model as a 5× range inBn.
Combined with evidence for methyl fer-
mentation and methanogenesis in the same
Shimokita coals ( 13 , 26 ), our findings suggest
a complete microbial reaction pathway from
coal methoxyl groups to methane. Temper-
ature is a key variable for deep biosphere life,
influencing community size, activity, and com-
position ( 13 , 14 ). However, the observation of
pervasive biodegradation with minimal iso-
topic fractionation in situ in subbituminous
coals suggests that substrate availability ulti-
mately limited O-demethylation in these units.
Models of the Shimokita coal community also
suggest a carbon-starved, substrate-limited exis-
tence ( 27 ). Considering that initial activation
of coal moieties may be the rate-limiting step
of microbial methanogenesis ( 28 ), our results
indicate that, here, coal bioavailability limited
microbial CBM production. If these results are
generalizable, the abundance of methoxyl
groups that persist until gas can accumulate
may be an important control on CBM yield
(Fig. 2 and materials and methods, section S5).
The existence of a widespread, labile car-
bon pool that is anomalously^13 C-enriched
bears on the interpretation of methane origins
in coals. Stable isotopic fingerprinting princi-
ples developed in conventional hydrocarbon
reservoirs ( 15 , 29 ) are routinely violated in
unconventional coal-bed systems: Coal gases
of probable microbial origin (based on thermal
maturity and methane/ethane abundance
ratio) often have methaned^13 C values above
− 50 ‰, which is the nominal upper limit for
primary microbial methane ( 30 , 31 ) (Fig. 3).
Existing explanations include the following:
Bulk organic carbon is more^13 C enriched in
coals versus conventional sources ( 30 ); a
greater fraction of microbial CBM is formed
through acetoclastic versus CO 2 reductive path-
ways ( 30 , 31 ); and methanogenic substrates
from coal are partly diverted to other fates,
leaving residues^13 Cenriched( 31 ). We suggest
an alternative explanation: Near-quantitative
coal methylotrophy generates methane with a(^13) Ccontentsimilartothis (^13) C-enriched substrate
(Fig. 2). Specifically, initial O-demethylation,
probably occurring during peatification, iso-
topically fractionates lignin methoxyl groups.
Remaining^13 C-enriched methoxyl groups are
anaerobically biodegraded to C1 intermedi-
ates ( 11 ) that can feed methanogenesis (mate-
rials and methods, section S5). Because this
anaerobic O-demethylation is near quantita-
tive, the cumulatived^13 C of the C1 intermediate
matches the initial coal methoxyl pool. If
methanogenesis is also quantitative or sub-
strate limited, thed^13 C of accumulated meth-
ane will match the initial as well. Taking the
methoxyl abundance and^13 C content of the
Belchatow lignites as representative of the pool
delivered to the deep subsurface, this scheme
suggests that the biodegradation of humic
coals should generate up to 2 to 3 mmol/g of
cryptic microbial methane with unusually high
d^13 C values (−20 to− 10 ‰; Fig. 2). Consid-
ering that total CBM potentials are ~8 mmol/g
(materials and methods, section S5), contri-
butions from methoxyl-derived methane could
appreciably raise the bulkd^13 C value of mi-
crobial CBM (Fig. 3). Given the paucity of
methoxyl groups in conventional hydrocarbon
deposits, this likely explains the enigmatic
d^13 C offset between microbial methane from
coal and conventional sources.
REFERENCES AND NOTES
- World Energy Resources 2016 (World Energy Council, 2016);
http://www.worldenergy.org/publications/entry/world-energy-
resources-2016.
2. D. Ritteret al.,Int. J. Coal Geol. 146 , 28–41 (2015).
3. M. Saunoiset al.,Earth Syst. Sci. Data 12 , 1561– 1623
(2020).
4. A. J. Turner, C. Frankenberg, E. A. Kort,Proc. Natl. Acad.
Sci. U.S.A. 116 , 2805–2813 (2019).
5. D. Rice,AAPG Stud. Geol. 38 , 159–184 (1993).
6. D. Strąpoćet al.,Annu. Rev. Earth Planet. Sci. 39 , 617– 656
(2011).
7. P. G. Hatcher, D. J. Clifford,Org. Geochem. 27 , 251–274 (1997).
8. C. Durand-Souron, R. Boulet, B. Durand,Geochim. Cosmochim.
Acta 46 , 1193–1202 (1982).
9. I. Blom, L. Edelhausen, D. W. van Krevelen,Fuel 36 , 135– 153
(1957).
10. D. F. Payne, P. J. Ortoleva,Org. Geochem. 32 , 1073– 1085
(2001).
11. J. Heider, G. Fuchs,Eur. J. Biochem. 243 , 577–596 (1997).
12. D. Mayumiet al.,Science 354 , 222–225 (2016).
13. F. Inagakiet al.,Science 349 , 420–424 (2015).
14. V. B. Heueret al.,Science 370 , 1230–1234 (2020).
15. M. J. Whiticar,Chem. Geol. 161 , 291–314 (1999).
16. C. Clayton,Mar. Pet. Geol. 8 , 232–240 (1991).
17. Y. Tang, J. K. Perry, P. D. Jenden, M. Schoell,Geochim.
Cosmochim. Acta 64 , 2673–2687 (2000).
18. K. L. Londry, K. G. Dawson, H. D. Grover, R. E. Summons,
A. S. Bradley,Org. Geochem. 39 , 608–621 (2008).
19. J. B. Best,J. Cell. Comp. Physiol. 46 ,1–27 (1955).
20.T.N.P.Bosma,P.J.M.Middeldorp,G.Schraa,
A. J. B. Zehnder,Environ. Sci. Technol. 31 , 248– 252
(1996).
21. M. Thullner, M. Kampara, H. H. Richnow, H. Harms, L. Y. Wick,
Environ. Sci. Technol. 42 , 6544–6551 (2008).
22. S. J. B. Mallinsonet al.,Nat. Commun. 9 , 2487 (2018).
23. D. Naidu, S. W. Ragsdale,J. Bacteriol. 183 , 3276– 3281
(2001).
24. L. Gao, M. Mastalerz, A. Schimmelmann, inCoal Bed Methane:
From Prospect to Pipeline, P. Thakur, S. Schatzel, K. Aminian,
Eds. (Elsevier, 2014), pp. 7–29.
25. F. Keppler, R. M. Kalin, D. B. Harper, W. C. Mcroberts,
J. T. G. Hamilton,Biogeosciences 1 , 123–131 (2004).
26. E. Trembath-Reichertet al.,Proc. Natl. Acad. Sci. U.S.A. 114 ,
E9206–E9215 (2017).
27. S. A. Bowdenet al.,Basin Res. 32 , 804–829 (2020).
28. B. Wawriket al.,FEMS Microbiol. Ecol. 81 , 26–42 (2012).
29. M. Schoell,AAPG Bull. 67 , 2225–2238 (1983).
30. M. J. Whiticar,Int. J. Coal Geol. 32 , 191–215 (1996).
31. D. S. Vinsonet al.,Chem. Geol. 453 , 128–145 (2017).
32. H. Lee, X. Feng, M. Mastalerz, S. J. Feakins,Org. Geochem.
136 , 103894 (2019).
33. O. A. Sherwood, S. Schwietzke, V. A. Arling, G. Etiope,Earth
Syst. Sci. Data 9 , 639–656 (2017).
ACKNOWLEDGMENTS
We thank K. Snell for use of analytical facilities and M. Formolo and
S. Petsch for sharing materials. Shimokita samples were provided
by the International Ocean Discovery program. H. L. O. McClelland,
D. Stolper, A. Schimmelmann, C. Ponton, and H. Lee provided helpful
suggestions at various stages of the project.Funding:This work
was supported by NSF EAR 1322058 to J.M.E.; a NASA postdoctoral
fellowship with the NASA Astrobiology Institute to E.T.-R.; the
L’Oréal USA For Women in Science fellowship to E.T.-R.; the
L'Oréal-UNESCO FWIS International Rising Talent Endowment to
E.T.-R.; and the ACS award PRF-53747-ND2 to S.J.F., which
initiated their work on this project.Author contributions:
Conceptualization: M.K.L. and E.T.-R.; Methodology: M.K.L., E.T.-R.,
S.J.F., A.L.S., and J.M.E.; Investigation: M.K.L., E.T.-R., and K.S.D.;
Resources: E.T.-R., S.J.F., M.M., V.J.O., A.L.S., and J.M.E.; Writing–
original draft: M.K.L.; Writing–review and editing: M.K.L., E.T.-R.,
S.J.F., M.M., V.J.O., A.L.S., and J.M.E.; Supervision and funding
acquisition: J.M.E.Competing interests:The authors have no
competing interests on this project.Data and materials
availability:All data are available in the main text or the
supplementary materials.SUPPLEMENTARY MATERIALS
science.org/doi/10.1126/science.abg0241
Materials and Methods
Figs. S1 to S14
Tables S1 to S6
References ( 34 – 88 )
MDAR Reproducibility Checklist
8 December 2020; accepted 13 September 2021
10.1126/science.abg0241SCIENCEscience.org 12 NOVEMBER 2021•VOL 374 ISSUE 6569 897
RESEARCH | REPORTS