(56f) Extractives of Douglas-fir and their Implications for Fermentation Based Biofuel Production

Douglas fir (Pseudotsuga
menziesii) has been investigated as a feedstock for biofuel processes due
to its abundance in western North American timberland.  This plentiful feedstock has the
potential to be a large component of a forestry residue based biofuel
industry.  Residues from timber
harvest and forest thinnings (sometimes called "slash") are comprised of
branches, broken or defective tree parts, tops, and trees not meeting grade
specifications. This material is often considered unmerchantable and so it is
collected, piled, and burned (or left to decay) after forestry operations to
reduce fire hazards and improve forest stands.  Routing Douglas fir residues into
biofuel production offers the potential for a sustainable low carbon fuel.

To design economical biofuel facilities using forestry
residues, a better understanding of how chemicals components in the feedstock
affect biofuel process steps is needed. 
Diverse extractive compounds make up 5% to 25% of the dry weight for
different tissues of Douglas fir [1,2], but are rarely accounted for in biofuel
studies.  Despite the
well-documented role of specific extractive components in the pulp and paper
literature, little detailed attention has been focused on extractive molecules
in biofuel processing. Instead, it is typical to define a lumped species with a
single property set, a psuedo-extractive,
that is tracked in process simulations. Our studies provide a starting
point for understanding the effects of molecularly distinct extractives in
fermentation-based biofuel production.

We create a molecularly representative model of Douglas fir
logging slash by drawing from a wide range of literature.  Our feedstock model extractive
components consist of 3.3% (w/w) polymeric polyphenols, 2.4% flavonoids, 1.6%
fats and waxes, 0.4% diterpenes, 0.2% phytosterols, and 0.1% monoterpenes on a
dry (no water) basis.  See Figure 1
for representative molecules of these extractive classes [2-6].  ASPEN simulations are used to track how
extractives move through a biofuel process based on sulfite/bisulfite
pretreatment, saccharification, and fermentation.  Steps such as pretreatment, pressing to
remove spent sulfite liquor, enzymatic hydrolysis, and evaporation are
simulated and molecules of interest are identified in key streams. Unlike
process simulations that lump all extractives into a single psuedo-species,
we find different classes of extractive molecules preferentially partitioning
between the vapor, product, and waste streams.

Our simulation results indicate that the saccharification
step may be particularly susceptible to inhibition from extractives.
Saccharification is inhibited by noncellulose
components binding to cellulose as well as binding to the hydrolyzing enzymes.
Astringent molecules known to bind to cellulases such
as the condensed tannins are very abundant in feedstocks containing bark and
our simulation indicates they appear in significant quantities in the feed to
saccharification.  As enzymes are
still one of the most expensive steps in biofuel processing, ways to use lower
titers of enzymes or bioengineer more effective enzymes need to be
investigated.  Knowledge of the
extractive inhibitors can help guide this investigation and subsequent
optimization. 

Figure
1: The six most abundant classes of extractives in Douglas fir and molecules
representative of their class.

Molecular dynamics simulations of cellulase-extractives
systems are conducted using GROMACS. 
These simulations identify which extractives bind to different cellulase
active sites, and the stability of extractive-enzyme complexes.  The cellulases
are modeled in aqueous media and include a cellobiohydrolase
from Hypocrea Jecorina, cellobioase from Aspergillus Niger,
and an endoglucanase from Acidothermus Cellulolyticus. Both the catalytic
domain and carbohydrate-binding module are simulated for the cellobiohydrolase and endoglucanase.
The extractives used include condensed tannin, dihydroquercetin, a fatty acid,
palustric acid, sitosterol, and α-pinene. Figure
2 shows dihydroquercetin binding onto the carbohydrate-binding module of the cellobiohydrolase.

Figure
2: The extractive dihydroquercetin (green) is shown binding to the
carbohydrate-binding module of the cellobiohydrolase
from Hypocrea jecorina (colored
by "index", based on amino acid location in the peptide sequence).

This work identifies the common extractives in Douglas fir,
models how they propagate in important process streams, and assesses their
affect on saccharification. These studies significantly improve our knowledge
of the effect extractives have on biofuel processes as well as lay the
groundwork for further studies of Douglas fir forestry residue.

References

[1]       Kaar WE,
Brink DL. Summative Analysis of 9 Common North American Woods. J Wood Chem
Technol 1991;11(4):479-94.

[2]       Kurth EF.
Chemicals from Douglas-Fir Bark. TAPPI 1953;36(7):119A-22A.

[3]       Dellus V,
Mila I, Scalbert A, Menard C, Michon V, duPenhoat C. Douglas-fir Polyphenols
and Heartwood Formation. Phytochem 1997;45(8):1573-8.

[4]       Foster DO,
Zinkel DF, Conner AH. Tall Oil Percursors of Douglas fir. TAPPI
1980;63(12):103-5.

[5]       Erdtman H,
Kimland B, Norin T, Daniels PJL. The Constituents of the "Pocket
Resin" from Douglas Fir Pseudotsuga menziesii (Mirb.) Franco. Acta Chem
Scand 1968;22(3):938-42.

[6]       Fischer F,
Koch H, Borchers B, Hontsch R, Pruzina KD. Preparation and Use of Phytosterols
from Wood. Pharm 1981;36(7):456-62.
ADDIN EN.REFLIST