You are on page 1of 10

Process Biochemistry 74 (2018) 108–117

Contents lists available at ScienceDirect

Process Biochemistry
journal homepage: www.elsevier.com/locate/procbio

The impact of kinetic parameters on cellulose hydrolysis rates T


a b,1 b,⁎
Jennifer Nill , Nardrapee Karuna , Tina Jeoh
a
Chemical Engineering, University of California, Davis, One Shields Ave., Davis, CA 95616, United States
b
Biological and Agricultural Engineering, University of California, Davis, One Shields Ave., Davis, CA 95616, United States

A R T I C LE I N FO A B S T R A C T

Keywords: Mechanistic kinetic models are vital tools to elucidate rate-limiting interactions and processes hampering
Cellulose hydrolysis complete hydrolysis of cellulose by cellulases. Properly formulated, the models should ultimately simulate ex-
Mechanistic kinetic modeling perimental cellulose hydrolysis time courses given that model parameters are known a priori. Here, we review
Productive cellulase binding capacity reported values of critical parameters in cellulose hydrolysis (catalytic, adsorption/desorption, complexation/
Rate constants
decomplexation rate constants and apparent processivity) and examine how estimation methods may impact the
Complexation/decomplexation
TrCel7A
magnitudes and interpretation of the values. The sensitivity of the extents and rates of cellulose hydrolysis to the
range of reported values were examined by conducting simulations with a general mechanistic model of TrCel7A
hydrolysis of cellulose. When the simulation was conducted at enzyme-limiting conditions, initial bursts in
cellulose hydrolysis rates were evident and highly sensitive to specific hydrolysis rates of productively bound
enzymes and dissociation rates of non-productively bound enzymes. Under substrate limiting conditions, hy-
drolysis rates were significantly reduced and no initial burst was observed, with more pronounced sensitivity of
hydrolysis rates to the range of kinetic values. Enzyme-substrate interaction rate constants reviewed here only
capture short-time hydrolysis rate outcomes and future work to elucidate mechanisms of cellulose hydrolysis
must also incorporate how substrate-limitations evolve over the course of hydrolysis.

1. Introduction cellulases adsorbed or complexed to cellulose at a non-productive binding


site (SNP) without hydrolyzing glycosidic bonds [5], while productively
Cellulose is the most abundant biopolymer on earth, and is com- bound cellulases complexed to cellulose at productive binding sites (SP)
prised of chains of glucose, an excellent feedstock for microbial con- hydrolyze glycosidic bonds. Cellulases may also release from a cellulose
version to biofuels. The economics of this process are highly dependent molecule (decomplexation) or detach from the microfibril surface back
on the price of carbon on which the microorganisms are grown, and into solution (desorption). As previously reviewed, several recent me-
cellulose offers the potential as a cheap and renewable carbon source. chanistic models have focused on resolving several sub-steps within
The main technological impediment to a more widespread utilization of enzyme-substrate interactions [3]. These ‘enzyme-centric’ mechanistic
lignocellulosic biomass is the lack of low-cost technologies to overcome models are highly parameterized and complicated by the need for a
its recalcitrance, the inherent resistance to degradation of the cellulosic priori estimates of many of the kinetic rate constants to validate model
structure. [1] Despite significant gains in understanding of cellulolytic outcomes. Thus, tied closely with efforts to capture elementary me-
enzymes, the complex nature of the heterogeneous reaction kinetics has chanisms of cellulose hydrolysis are efforts to independently measure
limited a full mechanistic understanding of lignocellulose hydrolysis by relevant kinetic constants (e.g. for adsorption/desorption, complexa-
cellulases [2,3]. tion/decomplexation, hydrolysis etc.).
A working hypothesis of cellulose hydrolysis by cellulases follows One substantial challenge in experimentally determining kinetic
the paradigm outlined in Fig. 1 [4], where enzymes in solution adsorb parameters lies in the difficulty of measuring intermediate species in
by the carbohydrate binding module (CBM), complex with an available the reaction. For example, within the total quantity of cellulases bound
cellulose molecule at the active site of the catalytic domain (CD) and to cellulose, experimentally distinguishing modes of cellulase binding
hydrolyze glycosidic bonds that, in the case of exocellulases, release (by CD or CBM, or as productively and non-productively bound) is non-
soluble sugars. It is proposed that cellulases bound to cellulose may trivial [6–8]. Another unresolved challenge is the uncertainty in de-
exist in two general states: non-productively bound enzymes are fining and measuring the substrate in the models. Earlier modeling


Corresponding author.
E-mail address: tjeoh@ucdavis.edu (T. Jeoh).
1
Present address: Biotechnology, Faculty of Engineering and Industrial Technology, Silpakorn University, Nakhon Pathom, Thailand.

https://doi.org/10.1016/j.procbio.2018.07.006
Received 7 June 2018; Received in revised form 9 July 2018; Accepted 10 July 2018
Available online 11 July 2018
1359-5113/ © 2018 Elsevier Ltd. All rights reserved.
J. Nill et al. Process Biochemistry 74 (2018) 108–117

Fig. 1. Proposed mechanism of enzyme hydrolysis. Enzyme adapted from [4].

approaches commonly defined substrate as the mass of cellulose in the 2. The catalytic rate constant, kcat
reaction (examples of models reviewed by [2]). However, this approach
does not capture the inaccessibility of large fractions of cellulose to The catalytic rate constant is the rate at which a complexed cellulase
cellulases due to its supramolecular structuring into microfibrils, fibrils hydrolyzes glycosidic bonds in cellulose (i.e. the number of catalytic
and further into hornified particles such as those of microcrystalline events per time). In homogeneous reactions that are not mass transfer
cellulose. Furthermore, for a reducing-end cellobiohydrolase, defining limited, an enzyme’s catalytic efficiency limits the overall reaction rate.
substrate by mass does not capture the concentration of ‘cellulase-ac- Improving the catalytic rate of cellulases has been the focus of many
cessible reducing-ends’ [9–11], or distinguish between productive and years of research [15,16]. However, because there are several poten-
non-productive binding sites. Cellulases have been shown to adsorb to tially rate limiting steps such as adsorption and complexation that
hydrophobic surfaces of cellulose fibrils by the CBM [12], thus a precede cellulose hydrolysis that are experimentally difficult to de-
measure of ‘hydrophobic binding sites’ may be representative of SNP, couple, estimates of intrinsic catalytic rate constants of cellulases have
but non-productive binding may also occur with a cellulose chain been challenging.
complexed in the active site of the CD (Fig. 1). Thus, to fully understand Early experimental estimation of kcat was obtained for TrCel7A on
cellulose hydrolysis mechanistically, a detailed knowledge of both the amorphous, bacterial and microcrystalline cellulose by fitting an
rate constants and intermediate species are required to validate pro- equation analogous to the mixed type Michaelis-Menten inhibition ex-
posed reaction mechanisms. pression to the initial release of soluble sugars and [3H] labelled re-
In this review, we explore how kinetic parameters for cellulose ducing ends on cellulose [17] (Table 1). More recently, direct mea-
hydrolysis have been defined and estimated, and how the range of re- surements of the rate of unidirectional translation of TrCel7A on the
ported magnitudes can impact our understanding of rate limiting en- surface of a crystalline cellulose Iα fibril by high speed AFM (assumed to
zyme-substrate interactions during cellulose hydrolysis. There is a be due to processive hydrolysis) gave an estimate of a kcat of ∼ 5.3 s−1
wealth of data from many sources employing various methods to when averaging initial measurements of 3.5 s−1 [18], and a follow up
measure kinetic constants of the reducing-end specific exocellulase refinement of 7.1 s−1 [19]. A number of studies utilized the assumption
Cel7A from Trichoderma reesei (Hypocrea jecorina) (TrCel7A). TrCel7A is that during the initial burst phase of the reaction, all TrCel7A are
relatively easily obtainable in pure form [13,14] and releases soluble productively bound and contributing to the measured hydrolysis rate
cellobiose from the reducing ends of cellulose. We review kinetic con- [20], referred to as “occupying the obstacle free path” [8] or as “limited
stant values reported for TrCel7A on cellulose and evaluate how they to performing a single processive run” [21]. Measurements fitting a
influence mechanistic interpretation of cellulose hydrolysis by per- pseudo steady state model incorporating adsorbed, irreversibly bound
forming a sensitivity analysis with a previously published working and complexed enzyme states to hydrolysis data measured by quartz
mechanistic kinetic model. crystal microgravimetry (QCM) produced another estimate of kcat = 4.8

109
J. Nill et al. Process Biochemistry 74 (2018) 108–117

s−1 [22]. Additionally, a computational approach examining the reac-

[18,19,26]

[26,30]
Source
tion pathway for the hydrolysis of a glycosidic bond at the +1/−1

[20]
[24]

[23]
[22]
[17]

Biochemical Measurements – estimated from initial rates of cellobiose formation and the concentration of “active site occupied TrCel7A” [21]
[8]
subsites within the active site of TrCel7A estimated kcat from the cal-
culated energy of activation of the reaction [23].
These disparate approaches applied to cellulosic substrates with
varying processing histories and crystalline allomorphs, including
bacterial cellulose, Avicel and cellulose III, arrived at kcat estimates well
within 2-fold magnitudes (Table 1). The average of kcat tabulated in

Biochemical Measurements – Estimated from initial rates of formation of total reducing groups after 10 s of hydrolysis
Table 1 is 4.3 ± 2.4 s−1, showing remarkable agreement across the
Biochemical Measurements – from applying a first order relationship to initial rates of product (cellobiose) release.

board [4]. Together, the convergence of the kcat estimates by these


Computational – calculated kcat from Ea of reaction coordinate for bond cleavage within the enzyme active site.
groups strongly suggests that TrCel7A that is productively engaged to
cellulose will hydrolyze cellulose at a kcat of ∼ 5 s−1 regardless of type,
crystallinity or morphology of the cellulose. That is, cellulose hydrolysis
High Speed AFM – The velocities of TrCel7A translating on the surface of cellulose were measured.

rates are not explicitly limited by the ‘energy of decrystallization’, the


free energy cost of dissociating a chain of cellulose from the fibril
Biochemical Measurements – Fit to pseudo-steady state hydrolysis rate determined by QCM.

surface to be threaded into the active site. Most mechanistic studies of


Biochemical Measurements – Fit non-linear regression of initial rates of cellobiose release

cellulose hydrolysis agree that the hydrolytic rate is not limiting, yet
there is disagreement regarding whether enzyme attachment or de-
tachment from cellulose truly controls hydrolysis rates [3].
Published measurements of kcat of TrCel7A hydrolysis of cellulose (expanded from [10]). All measurements were obtained at pH 5–5.5 and T = 25–30 °C.

3. Specific rates of adsorption and complexation

A frequently used approach to experimentally estimate the adsorp-


tion and desorption rate constants from cellulose has been to assume a
general mechanism shown in Eq. (1), where free/unbound enzyme (E)
binds to the substrate (S) reversibly. The adsorption rate constant is the
second order rate constant for cellulase binding from solution to solid
cellulose (Eq. (2)).
k on
E + S ⇄ Eb
k off (1)

radsorption = kon [E ][S ] (2)


How S in Eq. (1) is defined and measured can lead to discrepant
conclusions about characteristic binding rates. For example, kon is the
complexation rate constant if S is specifically defined as the con-
centration of productive binding sites on cellulose (μmoles/g) [5].
However, S is defined in many other ways in the literature (e.g. as
Method

cellulose mass or mass concentration). Thus, although adsorption rate


constants of TrCel7A binding to cellulose has been estimated in various
ways by many research groups (Table 2), when comparing results from
3.5 ± 1.1
7.1 ± 3.9
5.3 ± 2.1
5.7 ± 2.4
6.8 ± 3.5

1.5 ± 0.3

2.5 ± 0.3
1.8 ± 0.5
2.8 ± 0.4
2.2 ± 0.5
kcat (s−1)

different sources, it is critical to understand how the measurements are


4±1

conducted and how the substrate S is defined.


4.75
10.8
5.1
2.4

4.8

1.7

The most experimentally convenient definition of S is the mass


concentration (g/L) of cellulose in the reaction. In one recent example,
adsorption rate constants of Hypocrea jecorina Cel7A (HjCel7A) on re-
generated amorphous cellulose (RAC), Avicel and bacterial micro-
Amorphous cellulose (Nata de Coco in LiCl and dimethylacetamide)
Regenerated amorphous cellulose (Sigmacell 20 in phosphoric acid)

Regenerated cellulose film (Avicel in 4-methylmorpholine-N-oxide)

crystalline cellulose (BMCC) were estimated using the total initial cel-
Amorphous cellulose (Whatman filter paper and phosphoric acid)

lulose concentration S0 (g/L) as the substrate [24]. The enzyme on-rate


Regenerated amorphous cellulose (Avicel in phosphoric acid)

was estimated by fitting cellobiose release rates measured within the


first 100 s with a model incorporating Eq. (2), where enzymes directly
Bacterial microcrystalline cellulose from Nata de Coco

C labelled Bacterial Cellulose (Acetobacter xylinum)

form a productive enzyme-substrate complex from solution. Thus here,


Bacterial microcrystalline cellulose (Nata de Coco)

Cellodextrin within the active site of the enzyme

the estimated association rate is an estimate of the complexation rate


constant, kon,P. The authors reported an apparent influence of substrate
concentration on kon,P; at high enzyme loadings, the limited availability
Bacterial cellulose (Acetobacter xylinum)

of productive binding sites led to higher apparent kon,P values, while at


Bacterial cellulose (Nata de Coco)

sufficiently high cellulose concentrations, cellobiose release rates and


complexation rate estimates were independent of substrate concentra-
tion (Table 2). These results underscore the challenge in defining S as
Cladophora cellulose III
Cladophora cellulose Iα

the mass concentration of cellulose when estimating complexation rate


constants because the relationship between the mass of cellulose and
the productive binding site concentration on cellulose are not constant
between celluloses [5].
Cellulose

The enzyme on-rate has also been estimated by defining S in Eq. (2)
Table 1

Avicel

as cellulose surface area (m2/mg), where the rates of adsorption and


14

desorption of cellobiohydrolase Cel7A of Trichoderma longibrachiatum

110
J. Nill et al. Process Biochemistry 74 (2018) 108–117

Table 2
Adsorption/complexation and desorption/decomplexation rate constants in the literature. All rate constants were estimated at pH 5–5.5 and T = 25–30 °C.
a
Enzyme Cellulose Adsorption/complexation rate Desorption/decomplexation rate Ref.

−1
b
‘S’ c
Type Value Units c
Type d
Value (s )

−1 −1
H. jecorina Cel7A RAC (Avicel/H3PO4) Initial mass concentration Complexation 0.035 (g/L) s Decomplexation 0.023 [24]
Avicel (g/L) 0.0092 0.011
BMCC from Nata de Coco 0.016 0.01
T. reesei Cel7A BC (Nata de Coco) Initial molar concentration Adsorption 0.097 μM−1s−1 Desorption 0.067 [28]
(μM) Complexation 0.029 s-1 Decomplexation 0.013
Association sites and attack Adsorption 0.33 μM−1s−1 Desorption 0.19
sites (μM) Complexation 0.18 μM−1s−1 Decomplexation 0.016
μM−1s−1
BC (Acetobacter xylinum) – – – – Decomplexation 0.0007 [8]
Amorphous cellulose 0.0032
(Filter paper/H3PO4) – – – – Decomplexation (Pop. 0.017 [31]
1)
Decomplexation (Pop. 0.0018
2)
Desorption 0.1
Cladophora cellulose Iα – – – – Decomplexation (Pop. 0.023
1)
Decomplexation (Pop. 0.003
2)
Desorption 0.125
Substrate not limiting Complexation 0.28 s−1 Decomplexation 0.03 [27]
Unit area of cellulose fibril General 3.0 × 1010 M−1 μm−2 s−1 Decomplexation 0.12 [26]
surface (μm−2) Desorption 0.86
Cladophora cellulose III 4.5 × 1010 Decomplexation 0.14
Desorption 0.74
– – – – Decomplexation 0.2 [30]
T. longibrachiatum Regenerated Cellulose Specific surface area of General 0.28 ppm−1 h−1 Decomplexation 0.00183 [25]
Cel7A Film (Avicel/4MMO) cellulose (m2/mg)

a
RAC = Regenerated Amorphous Cellulose; BMCC = Bacterial Microcrystalline Cellulose; BC = Bacterial Cellulose.
b
The definition of ‘S’ (Eq. (1)) used to estimate adsorption rate constants.
c
Adorption/desorption, complexation/de-complexation or general (combined) rate constant.
d
Units are s−1 unless otherwise noted.

on regenerated cellulose films were estimated using QCM [25]. The the hydrolysis of cellulose increases the residence time of an enzyme on
adsorption rates, modeled as second order with respect to unbound the substrate. Under the assumption of excess substrate at the low
enzyme concentration (ppm) and unoccupied surface area on the cel- (picomolar) enzyme loading, the complexation rate constant was esti-
lulose film (m2/mg) were reported in units of inverse enzyme con- mated as first order, kon,P = 0.28 s−1. In this case, the substrate was
centration per time (ppm−1 h−1). These units are a departure from the implicitly assumed to not limit the rate of complexation by TrCel7A to
general convention of inverse substrate concentration per time for ad- Cladophora cellulose.
sorption rate constants and thus present a challenge in comparing to Adsorption (kon,NP) and threading (kon,P) rates were estimated by a
values in the literature. In this case, the estimated on-rate constants are model expanding Eq. (1) into two steps; the first describes the asso-
neither specific to non-productive adsorption, nor to complexation. ciation of free enzyme with a binding site to form an associated enzyme
In another study, Shibafuji et al. [26] defined the on-rate constant with an unoccupied binding site, followed by the threading of a cellu-
kon as the number of TrCel7A bound to cellulose in a unit of Cy3- lose strand into the binding tunnel of the catalytic domain, as described
TrCel7A concentration, cellulose length and time (M−1 μm−1s−1) for in Eq. (3):
quantitation and subsequently normalized by the unit area of cellulose
E + S ⇌ Eassoc ⇌ Ethread (3)
(M−1 μm−2s−1). The authors quantified dynamically bound Cy3-la-
beled TrCel7A molecules on the surface of algal (Cladophora) cellulose These on-rates were estimated on the basis of a molar concentration
fibrils by total internal reflection fluorescence microscopy (TIRFM) and of productive binding sites (μM) converted from cellulose mass con-
high-speed AFM (HS-AFM). The distributions of kon were fitted by centration (g/L) using the maximum TrCel7A binding capacity
multiple Gaussians corresponding to the number of microfibrils in the (μmoles/g) from an equilibrium binding isotherm on bacterial cellulose
cellulose fibrils. The authors thus reported kon for single microfibrils as [28]. There is some debate as to whether an equilibrium estimation of
given in Table 2 [26]. Here again, combining surface area of cellulose binding sites underestimates initial concentrations of binding sites,
and bound enzyme (non-productively adsorbed and complexed) as the since it is thought that the initial number of productive binding sites
basis, resulting kon estimates inform on a general binding rate constant rapidly decreases within the first minute of hydrolysis [5]. The time
of enzymes from solution onto the cellulose surface. In a similar ap- resolved amount of bound TrCel7A with a free active site was quanti-
proach, Jung et al. tracked binding lifetimes of individual TrCel7A fied by generating a standard curve by incubating TrCel7A with a re-
molecules on Cladophora cellulose Iα by TIRFM, from which binding porter substrate (4-methylumbelliferyl-β-D-lactobioside, MUL). The loss
rate constants were estimated. In contrast to the previously described in MUL activity when cellulose, MUL, and TrCel7A were combined was
methods, binding rate constants were estimated without assuming a defined as a measure of threaded enzyme. The association and
binding mechanism such as Eq. (1) [27]. Two binding lifetimes of threading rate constants were estimated from numerical analysis of the
TrCel7A on Cladophora cellulose were observed with the shorter life- temporal development of the measured enzyme species to estimate
time assigned to non-productive (kon,NP) binding, and the longer life- association and using the scheme described in Eq. (3). The model was
time assigned to productive binding (kon,P), since it was assumed that further refined by distinguishing association (Sα) and attack (Sβ) sites
by partitioning the total estimated number of binding sites using a ratio

111
J. Nill et al. Process Biochemistry 74 (2018) 108–117

Table 3
Estimates of processivity for TrCel7A. Measurements were made at pH 5–5.5 and T = 25–30 °C.
Cellulose Apparent Method Ref.
processivity

Amorphous cellulose (Filter 20.8 Biochemical - Fluorescently labelled reducing ends reveal Ninit, while soluble reducing groups estimate Ncatal: [8]
paper/H3PO4) napp =
[Soluble groups] + [End Label]
Bacterial Cellulose (Acetobacter 61 [End Label]

xylinum)
Amorphous cellulose (Avicel/SO2- 14 Biochemical - Glucose and cellotriose reveal Ninit and cellobiose estimates Ncatal: [37]
DEA) n app =
[Cellobiose]
BMCC 23 [Glucose] + [Cellotriose]

(Nata de Coco)
Cellulose Iα 20.3 [30]
(Cladophora sp.)
Amorphous cellulose (Avicel/ 15.5
H3PO4)
Cellulose III 34.8
(Cladophora sp.) 23.6 HS-AFM – Processivity is estimated from the time and velocity of observed cellulase movement:
movement half − life * average velocity
napp =
ln(2) * cellobiose length
Bacterial Cellulose (Acetobacter 66 Biochemical - 14C-labelled ends reveal Ninit, while the estimated concentration of enzymes with occupied active [21]
xylinum) [14C − labelled ends]
sites approximates Ncatal: napp =
[Tr Cel7A with occupied active site]
Bacterial Cellulose (Nata de Coco) 88 Biochemical - Fluorescently labelled reducing ends reveal Ninit, while soluble reducing groups estimate Ncatal: [36]
BMCC 42 napp =
[Soluble groups]
(Nata de Coco) [End Label]

Amorphous cellulose 13 Mechanistic - Derived from kinetic analysis of the pre-steady state regime fitting a mechanistic expression to [20]
(Sigmacell 20/H3PO4) hydrolysis curves of amorphous cellulose TrCel7A
Avicel 23–24 Biochemical – Cellotriose concentrations (or production rates) reveal Ninit and cellobiose estimates Ncatal: [38]
[Cellobiose]
n app =
[Cellotriose]

of Sα/Sβ reported by Jalak and Valjamae [29]: respectively. A sensitivity analysis on the effects of changes in rate
constants on hydrolytic activity determined that dethreading (decom-
E + Sα ⇌ Eassoc ⇌ Ethread (4)
plexation) is overall rate limiting for both models.
Eassoc + Sβ ⇌ Ethread (5) In another approach, the decomplexation rates of TrCel7A from
−1 −1 −1 −1 BMCC and amorphous cellulose were approximated from estimates of
Association (kon,NP) rates of 0.097 μM s and 0.33 μM s and
the rate at which the enzyme initiated hydrolysis on the insoluble cel-
threading (kon,P) rates of 0.029 s−1 and 0.18 μM−1s−1 were obtained
lulose [8]. The estimates were predicated on the assumption that de-
from the single binding site and association-attack site models, re-
complexation was ultimately rate limiting and significantly slower than
spectively.
turnover rates. The rate constants, obtained from dividing the estimated
decomplexation rate by the total enzyme concentration in the reaction,
4. Specific rates of desorption and decomplexation were 0.0032 and 0.0007 s−1 on BMCC and amorphous cellulose, re-
spectively. These estimates are considerably lower than the estimates
As is the case with estimating adsorption rate constants, the generic reported by other studies (Table 2). One likely reason for the low values
mechanism in Eq. (1) is generally the basis for estimating desorption is that total enzyme concentration, rather than the bound enzyme
rate constants. The desorption rate constant is a first order rate of concentration, or productively bound enzyme concentration was used in
transferring cellulase from solid surface to fluid phase (Fig. 1) and the estimations. Jalak and Väljamäe, using a soluble in situ reaction
thereby only a function of the bound enzyme concentration: reporter of non-productively bound enzyme, had demonstrated earlier
rdesorption = kd [Eb] (6) that only a fraction of bound enzymes are productively bound [29].
Furthermore, others have reported that only a few percent of bound
The same approaches used to estimate adsorption rate constants TrCel7A appear to be productively bound [10,27]. Accordingly, the
described in the previous section are often used to estimate desorption decomplexation rate constants reported by [8] are likely to be under-
rate constants. Moreover, both on- and off-rate constants are frequently estimated by one or two orders of magnitude.
estimated simultaneously. Thus, to compare reported values of deso- Direct estimations of desorption (koff,NP) and decomplexation (koff,P)
rption rate constants, one must also be cognizant of whether these are rate constants have been made from tracking binding lifetimes of in-
rates of desorption by non-productively bound enzyme, decomplexation dividual cellulases on the surface of cellulose by super-resolution single
or a general rate incorporating both types of desorption (Table 2). molecule imaging. Multi-exponential fits to histograms of binding
Cruys-Bagger et al. estimated decomplexation rates (koff,P) of events over time can reveal populations of bound enzyme with different
HjCel7A from RAC, Avicel, and BMCC from fits to product release rates characteristic binding lifetimes that have been assigned to productively
which were found to be at least two orders of magnitude less than kcat, and non-productively bound enzymes to obtain decomplexation and
implying that a complexed TrCel7A executes several hydrolysis events desorption rate constants, respectively, from the reciprocals of the
before de-complexing [24]. The estimated koff,P value on RAC was binding lifetimes. This approach was reported by Jung et al. [27] and
higher than on crystalline celluloses (Avicel and BMCC), leading the Shibafuji et al. [26], both measuring binding lifetimes of fluorophore-
authors to hypothesize that higher steady-state activities on amorphous labeled TrCel7A on Cladophora cellulose Iα fibrils. While both studies
cellulose could be due to more rapid desorption of the enzyme. Fitting reported two characteristic binding lifetimes within TrCel7A bound to
of temporally evolving concentrations of free, associated and threaded cellulose and assigned the populations to productively and non-pro-
enzymes using the model described in Eqs. (4) and (5) gave dissociation ductively bound enzyme, the reported rate constants were distinctly
rate constants of 0.067 s−1 and 0.19 s−1 and dethreading rate constants different. Jung et al. [27] observed a dominant population of bound
of 0.013 s−1 and 0.016 μM−1s−1 for the one- and two-site models,

112
J. Nill et al. Process Biochemistry 74 (2018) 108–117

TrCel7A with a short binding time of 30 s (koff,NP = 0.03 s−1) assigned mechanistic approach fitting equations describing the pre-steady state
to non-productively bound TrCel7A and a minor population with long regime to hydrolysis profiles of amorphous cellulose by TrCel7A has
characteristic binding time of 173 s (koff,P = 0.006 s−1) assigned to also provided an estimate for napp. For a detailed review of methods of
productively bound TrCel7A. Shibafuji et al. [26] similarly reported a measuring processivity, refer to [39].
dominant population with fast desorption rate and a minor population The values for apparent processivity estimates from the literature
with slow desorption rate assigned to non-productive binding (koff,NP) are summarized in Table 3 showing good agreement by cellulose type
and productive binding (koff,P), respectively. The koff values for TrCel7A despite the variation in methodology. Biochemical [8,30,37] and me-
of koff,P = 0.12 s−1 and koff,NP = 0.86 s−1 from cellulose Iα and koff,P = chanistic [20] estimates for amorphous cellulose yield an average napp
0.14 s−1 and koff,NP = 0.74 s−1 from cellulose III were 10–1000 times of 16 ± 3.5, while various methods of estimating Ninit and Ncatal within
larger than koff values previously reported in the literature. The dif- the biochemical approach estimate napp = 72 ± 14 for bacterial cel-
ferent magnitudes of the rate estimates from other reported values were lulose [8,21,36] and napp = 30 ± 10 for microcrystalline cellulose
attributed to differences in sources of Cel7A and cellulose in the ex- [36–38]. Measurement of napp on the same cellulose III by both bio-
periments because a similar value of koff,P (0.2 s−1) was obtained using chemical and HS-AFM approaches give napp = 29.2 ± 7.9 [30].
HS-AFM to estimate the decomplexation rate from the half-life of pro- While processivity estimation is consistent across estimation
cessive TrCel7A movement on cellulose III [26,30]. methods, there is a clear distinction of napp values by cellulose type. The
Mudinoor et al. [31] also measured the binding lifetimes of TrCel7A degree of polymerization has been shown to have little effect on cel-
by single molecule imaging on cellulose Iα fibrils and amorphous cel- lulose hydrolyzability [10], and it is generally thought that this is due
lulose, but contrary to Jung et al. and Shibafuji et al, three characteristic to the fact that the substrate limits processivity to a number of residues
binding lifetimes were observed, despite the use of the same experi- much less than the total DP of a cellulose strand. However, little is
mental conditions. An automated method of binding lifetime analysis known about processivity from a substrate perspective. For example,
determined a major population with a characteristic short binding time what physical or chemical features of the cellulose surface limit pro-
of 8 and 10 s for Iα and amorphous cellulose, respectively, which due to cessivity, and whether processivity changes throughout hydrolysis re-
the similarity to non-specific binding times on background glass was mains unanswered.
attributed to non-specific binding to the cellulose surface. Two sub-
populations of similar magnitude with medium and long characteristic 6. General mechanistic model of cellulose hydrolysis
lifetimes (44 and 320 s for Iα and 57 and 562 s for amorphous cellulose,
respectively) were ascribed as enzyme populations specifically bound To understand how sensitive cellulose hydrolysis rates are to the
by the active site (i.e., complexed). However, the unclear association of magnitudes of each kinetic parameter, we conducted short (30 s) si-
binding lifetimes with a given cellulase domain and lack of reaction mulations using a general mechanistic kinetic model [5]. The model
progress data prohibited the assignation of the populations as produc- captures the steps described in Fig. 1. Free enzyme in solution (E),
tively or non-productively bound. loaded on the basis of substrate mass concentration (S), binds at pro-
ductive binding sites (SP) to form productively bound enzymes (Eb,P), or
5. Processivity values at non-productive binding sites (SNP) to form non-productively bound
enzymes (Eb,NP). Productively bound cellulases hydrolyze cellulose to
Processive exocellulases such as TrCel7A execute multiple sequen- generate soluble sugars (P) with a catalytic rate constant kcat. The
tial catalytic events (glycosidic bond hydrolysis) while complexed to model also describes the transformation of surface bound enzymes
cellulose [18,32,33]. The processivity of a cellulase, i.e. the number of between productive and non-productively bound states. The association
catalytic events per complexation event, is expected to be limited by the of a non-productively bound enzyme with a productive binding site
length of time a cellulase remains complexed and the catalytic rate of a occurs at the same rate as the association of a free enzyme at a similar
complexed enzyme. Theoretical processivity is defined as [8,24,34,35]: productive binding site (kon,P). The rate of transformation of an enzyme
from a productively bound to a non-productively bound state (kt,NP) is
kcat + koff defined in Eq. (9) as the rate at which an enzyme translates along a
ntheo =
koff (7) fibril (kcat) per average number of hydrolytic cuts (napp) [8,20,24]:

However, a discrepancy between experimentally measured pro- kt , NP = kcat / napp (9)


cessivity and the theoretical proccessivity has been reported and in-
The ODEs describing the rate of change of each species concentra-
terpreted as obstacles on the cellulose surface preventing continued
tion are given in Eqs. (10)–(14):
processive movement by the cellulase [8,21,22,29,32]. Despite this
interpretation, processivity is often viewed solely as a property of the d [Eb, P ] [E]
= kon, P [SP ]−koff , P [Eb, P ] + kon, P [Eb, NP ][SP ]−kt , NP [Eb, P ]
enzyme. Generally, biochemical estimations of apparent processivity, dt [S] (10)
napp, rely on estimating the ratio of the number of processive initiations
(Ninit) to the number of sequential catalytic events (Ncatal): d [SP ] [E ]
= −kon, P [SP ] + koff , P [Eb, P ]
dt [S ] (11)
Ncatal
napp =
Ninit (8) d [SNP ] [E ]
= −kon, NP [SNP ] + koff , NP [Eb, NP ]
dt [S ] (12)
Kipper et al., Kurasin et al., and Jalak et al. estimated Ninit from
concentrations of solubilized labelled end groups and Ncatal from the
concentration of soluble reducing groups [8,36] or estimates for the
d ( ) = −k
[E ]
[S ]
on, P
[E ]
[SP ] + koff , P [Eb, P ]−kon, NP [E ][SNP ] + koff , NP [Eb, NP ]
concentration of TrCel7A with an occupied active site [21]. Another dt [S ]
method assumed that each initiation cut on a molecule of cellulose (13)
produces either glucose or cellotriose, therefore using these con-
d [P ]
centrations as an estimate for the relative Ninit, and cellobiose—the sole = kcat [Eb, P ]
dt (14)
hydrolysis product from processive hydrolysis—as an estimate for Ncatal
[30,37,38]. Direct observation of the processive velocity and movement The model variables are summarized in Table 4.
lifetime of cellulases on crystalline cellulose III by HS-AFM has also This model generally describes cellulose hydrolysis by endo- or exo-
been used to quantify the apparent processivity of TrCel7A. A cellulases, so long as estimates for the kinetic rate constants and

113
J. Nill et al. Process Biochemistry 74 (2018) 108–117

Table 4 the enzyme-limiting scenario models the case where the productive
Definition and units of variables used in the general mechanistic kinetic model. binding site concentration is two orders of magnitude greater than the
Variable Definition Units enzyme concentration ([S]HS∙[SP,T] = 500 μM, [ET] = 5 μM).
There are two immediately noticeable differences in the simulation
Reaction Species outcome between the enzyme- and substrate- limited scenarios (Fig. 2).
E Free Enzyme μM
First, increasing the productive binding site (substrate) concentration
S Substrate Mass Concentration g/L
Eb,P Productively Bound Enzyme μmole/g
from 5 to 500 μM increased the hydrolysis extents and rates by more
Eb,NP Non-Productively Bound Enzyme μmole/g than an order of magnitude. This reaction rate increase indicates that
SP Productive Binding Site μmole/g the low solids case (5 μM productive binding sites) is indeed substrate
SNP Non-Productive Binding Site μmole/g limited. Second, initial bursts in hydrolysis rates occur for the enzyme-
P Soluble sugar μmole/g
limited case but not in the substrate-limited case. The hydrolysis rate
Reaction Parameters
kon,P Productive complexation rate g/μmol/s burst is characterized by a steep rate increase upon the start of the
koff,P Productive decomplexation rate s−1 reaction to a maximum, followed by a rapid rate decrease to a pseudo-
kon,NP Non-productive adsorption rate g/μmol/s steady state. When substrate is limiting, the speed at which the hy-
koff,NP Non-productive desorption rate s−1
drolysis rate reaches and decays from a maximum is considerably more
kt,NP Non-productive transformation rate s−1
kcat Hydrolysis rate s−1
gradual, likely because the reaction rate is controlled by the limited
napp Apparent Processivity unitless number of productive binding sites, rather than the kinetic properties of
the enzyme. When enzyme is limiting, the initial hydrolysis rate burst is
eliminated by either increasing the desorption rates of non-productively
reaction intermediates can be obtained. It is important to note that the bound enzyme (koff,NP) or decreasing the specific hydrolysis rate (kcat).
working model is an ‘enzyme-centric’ kinetic model [3] in that it does These trends suggest that when there is an excess of productive binding
not deplete the substrate and therefore only captures short-time cellu- sites, the initial burst is facilitated by fast processive action of initially
lose hydrolysis behavior. Our analysis is based on numerically solving bound enzymes and brought to a premature slowdown by the presence
the coupled ordinary differential equations (Eqs. (10)–(14), which is of non-productively bound enzymes on the substrate surface. Similar
only one of several approaches [3] towards solving the mechanism of conclusions were previously drawn from observations of hydrolysis rate
cellulose hydrolysis. bursts in the first few seconds of hydrolysis of regenerated cellulose by
TrCel7A at enzyme limiting conditions [20].
7. Sensitivity analysis of kinetic parameters

The sensitivity of cellulose hydrolysis rates and extents to the ranges 7.1. Sensitivity of cellulose hydrolysis rates to non-productive binding
of kinetic values reported in the literature was tested with the high and
low values summarized in Table 1 (kcat), Table 2 (kon and koff) and The rate at which TrCel7A binds non-productively to the cellulose
Table 3 (napp). Each parameter was varied independently while keeping surface (kon,NP), has little to no effect on hydrolysis rates (Figs. 2 and 3).
all other parameters at the base value. The influence of parameter As noted above, however, the rate at which non-productively bound
magnitudes from Table 5 on cellulose hydrolysis rates were examined TrCel7A detach from the cellulose surface (koff,NP) can have an impact
as cellobiose concentrations and release rates in the first 30 s (Fig. 2). on hydrolysis rates under enzyme-limiting conditions. Dissociation
The effect of increasing or decreasing parameter magnitudes 5-fold rates of non-productively bound enzymes primarily affect steady state
from the base values was also examined, comparing cellobiose con- hydrolysis rates. While faster dissociation helps to maintain maximum
centrations at 30 s ([CB]30), maximum rates of cellobiose release (vmax) hydrolysis rates achievable by the enzymes when the substrate is in
and steady state cellobiose release rate (vss) following vmax (Fig. 3). The excess, slow dissociation results in deceleration of hydrolysis rates,
simulations were conducted at substrate-limiting and enzyme-limiting likely due to obstacle created by the non-productively bound enzymes
conditions. While not explicitly modeled here, the kinetic parameters [19,40]. Fig. 3 also shows the influence of non-productive dissociation
used in this model were all obtained at pH 5–5.5 and T = 25–30 °C, and rates on maximum hydrolysis rates, likely because the onset of enzymes
as such, the simulation outcomes are for these reaction conditions. For a hindering the progress of others is earlier with slow non-productive
given productive TrCel7A binding capacity of the cellulosic substrate of desorption and later with faster desorption of the impeding enzymes.
5 μmole/g and total enzyme concentration of 5 μM (Table 5), the sub- When substrate is limiting, no effect of dissociation is observed because
strate-limiting scenario models the case where the concentration of other surface enzymes are no longer the primary cause of rate limita-
productive binding sites (substrate, S) in the reaction is equal to the tions.
enzyme concentration ([S]LS·[SP,T] = 5 μM, [ET] = 5 μM). In contrast,

Table 5
Kinetic parameter values and initial values used in simulations shown in Fig. 2.
Parameter Base Value Base Value Source High/Low Values

−1
kcat (s ) 4.3 Averaged from Table 1 10.8 / 1.5
kon,P (g/μmole/s) 6.43 x 10−4 kon,P = koff,P/Affinity from [5] 2.35 x 10−3 / 1.2 x 10−5
koff,P (s−1) 0.038 Averaged decomplexation values from Table 2 0.14 / 0.0007
kon,NP (g/μmole/s) 6.43 x 10−5 kon,NP = 0.1*kon,P from [5] 2.35 x 10−4 / 1.2 x 10−6
koff,NP (s−1) 0.347 Averaged desorption values from Table 2 0.86 / 0.067
napp 34 Averaged from Table 2 88 / 13

Input Variable Value Explanation


[ET] (μM) 5 Total enzyme concentration in reaction: [ET] = [E] + [Eb,P] + [Eb,NP]
[S] (g/L) [S]HS = 100 Total mass concentration of cellulose, HS = High Solids, LS = Low Solids.
[S]LS = 1
[SP,T] (μmole/g) 5 [5] The productive TrCel7A binding capacity of the substrate = total concentration of productive binding sites, [SP,T] = [SP] + [Eb,P]. Modeled as a
fixed parameter in this analysis.
[SNP,T] (μmole/g) 2 [5] Total concentration of non-productive binding sites.

114
J. Nill et al. Process Biochemistry 74 (2018) 108–117

Fig. 2. Product release curves (left column) and release rates (right column) predicted by our general mechanistic model using high (blue) and low (red) estimates for
kon,P, napp, kcat, koff,P, and kon,NP are shown compared to the base case simulated using values from Table 5 (black). Simulations were repeated for substrate limited
conditions (left panel) and enzyme limiting conditions (right panel) (For interpretation of the references to color in this figure legend, the reader is referred to the web
version of this article).

7.2. Sensitivity of cellulose hydrolysis rates to productive binding in this analysis; one explanation is that increasing decomplexation rates
causes enzymes to stop hydrolyzing prematurely. Slowing down de-
Cellulose hydrolysis rates are sensitive to both complexation (pro- complexation, which could allow enzymes to hydrolyze for longer
ductive binding) and decomplexation rates (kon,P, and koff,P, respec- periods, resulted in a marginal increase in hydrolysis rates and extents
tively) regardless of whether substrate or enzyme is limiting (Fig. 2). (Figs. 2 and 3). This limited increase was most likely observed because
The effect of complexation rates, however, is far greater when substrate this model imposed a baseline apparent processivity of 34 (Table 5),
is limiting. A 5-fold change in kon,P resulted in ∼300–400% changes in preventing further hydrolysis once an enzyme reaches the end of its
the magnitude of cellobiose yields and maximum hydrolysis rates under obstacle free processive run.
substrate limiting conditions, but only ∼40–70% changes under en-
zyme limiting conditions (Fig. 3). These results indicate that regardless
of the abundance of productive binding sites, the faster the enzymes can 7.3. Sensitivity of cellulose hydrolysis rates to processive hydrolysis rates
complex and begin hydrolysis, the greater the overall hydrolysis rates and processivities
and extents. Faster complexation is particularly beneficial when pro-
ductive binding sites are limited. Furthermore, the complexation rates The catalytic rate constant (kcat) affected initial, maximum and
influence both maximum and steady state hydrolysis rates, indicating steady state hydrolysis rates, with the impact more pronounced for the
that faster enzyme complexation is important at the very start of the enzyme-limiting case (Fig. 2). When substrate is in excess, faster hy-
reaction and as the reaction proceeds. drolysis by complexed enzymes gives faster initial hydrolysis rates that
Decomplexation rates also influence maximum and steady state reach a higher maximum rate more rapidly. As seen above, the accu-
hydrolysis rates, and the sensitivity to 5-fold changes in the magnitude mulation of slow dissociating non-productively bound enzymes causes a
of koff,P is similar for substrate and enzyme limiting cases (Fig. 3). In- sharp drop in the hydrolysis rates. Nevertheless, pseudo-steady state
terestingly, increasing decomplexation rates decrease hydrolysis rates rates remain higher for enzymes that hydrolyze faster. When substrate
is limited, a similar trend of faster initial rates achieving higher

115
J. Nill et al. Process Biochemistry 74 (2018) 108–117

Fig. 3. A summary of the percent change in


[CB]30, vss and vmax resulting from a sensitivity
analysis where each parameter was in-
dependently increased (blue solid bars) and
decreased (red striped bars) 5X from its re-
spective base case value (For interpretation of
the references to color in this figure legend, the
reader is referred to the web version of this
article).

maximum rates more rapidly by enzymes with higher catalytic rates model for cellulose hydrolysis demonstrated that the interpretation of
was observed. A distinct peak in maximum hydrolysis rates is absent rate limiting cellulase-cellulose interactions can be influenced by both
because the low abundance of substrate becomes rate limiting in this the parameters and experimental conditions used. Key to this analysis is
case. the definition of substrate as the concentration of productive binding
Notably, the apparent processivity napp had a strong influence on sites to which enzymes can complex and hydrolyze cellulose. For a
cellulose hydrolysis. Both maximum hydrolysis rates (vmax) and steady substrate with a productive binding capacity of 5 μmole/g and an en-
state hydrolysis rates (vss) respond significantly, which together results zyme loading of 5 μM (∼0.3 mg/mL for a 65 kDa cellulase), the sub-
in a strong response in cellobiose concentrations released in the reac- strate-limiting cellulose loading of 1 g/L mimics frequently seen la-
tion (Fig. 3). Higher processivities prolong the initial stage of reaction, boratory reaction conditions, whereas the enzyme-limiting cellulose
allowing hydrolysis rates to increase to a greater maximum. This shows loading of 100 g/L approaches economically feasible conditions. This
that when enzyme processivity is hampered, overall reaction outcomes analysis showed that the sensitivity of instantaneous cellulose hydro-
suffer. While the theoretical (or intrinsic) processivity of an enzyme is lysis rates to the kinetic parameters is strongly influenced by enzyme
solely a function of its catalytic and off- rates, the apparent processivity and substrate loadings in the reaction; cellulose hydrolysis under sub-
value is viewed as an indicator of the length of ‘obstacle free path’ over strate-limiting conditions are considerably more sensitive to the mag-
which complexed cellulases can processively hydrolyze cellulose un- nitudes of the kinetic parameters (Fig. 3).
impeded [8,21,41]. The base case apparent processivity of 34 (Table 5) The sensitivity analysis maintained a constant solids concentration
is at least 3-fold less than the theoretical processivity of 114 calculated and constant productive binding site concentrations within the short
with base case kcat and koff,P values (Table 5 and Eq. (7)), implying that (30 s) simulation. Enzymatic hydrolysis of cellulose necessarily depletes
in this model, a processive run is terminated by substrate features ra- solids and likely alters the concentration of available productive
ther than intrinsic properties of the enzyme. The rate of transformation binding sites. Thus, these simulations only provide insights into in-
of an enzyme from a productively bound to a non-productively bound stantaneous hydrolysis rates impacted by the kinetic parameters. As
state was earlier defined in Eq. (9) as a function of catalytic rate and the research in this field advances, the influence of the rate of productive
apparent processivity. In these simulations, napp was varied according binding site change during enzyme hydrolysis needs to be elucidated
to literature values (Table 5), thus increasing napp inherently decreased and included in mechanistic models. As observed in the analysis, in the
the rate at which productive bound enzyme becomes non-productive in substrate limiting case, hydrolysis rates were limited by the low
the model and can be interpreted as a substrate contribution to rate abundance of productive binding sites and not by the accumulation of
limitations. non-productively bound enzymes as seen when productive binding sites
were not limiting. Thus, potential rapid decrease in the availability of
8. Concluding remarks productive binding sites during hydrolysis can turn an enzyme-limiting
scenario into a substrate-limiting scenario, altering the sensitivity of
Estimates of kinetic parameter values in the literature can vary by reaction rates to kinetic parameters such as desorption rates of non-
orders of magnitude. The sensitivity analysis of a general mechanistic productively bound enzymes.

116
J. Nill et al. Process Biochemistry 74 (2018) 108–117

There is clearly a need for further refinement of both the enzyme- Commun. 9 (1) (2018) 1186.
centric kinetic models that can help define elementary mechanisms and [17] M. Gruno, P. Väljamäe, G. Pettersson, G. Johansson, Inhibition of the Trichoderma
reesei cellulases by cellobiose is strongly dependent on the nature of the substrate,
provide estimates for kinetic rate constants, and substrate-centric ki- Biotechnol. Bioeng. 86 (5) (2004) 503–511.
netic models to clarify the rate limiting properties of the insoluble [18] K. Igarashi, A. Koivula, M. Wada, S. Kimura, M. Penttila, M. Samejima, High speed
cellulosic substrate. We emphasize that as the rate constants reviewed atomic force microscopy visualizes processive movement of Trichoderma reesei
cellobiohydrolase I on crystalline cellulose, J. Biol. Chem. 284 (52) (2009)
here are instantaneous rate constants that do not account for evolution 36186–36190.
of the substrate, understanding the substrate contribution to the var- [19] K. Igarashi, T. Uchihashi, A. Koivula, M. Wada, S. Kimura, T. Okamoto, M. Penttilä,
iation in these parameters by cellulose type, as well as changes in en- T. Ando, M. Samejima, Traffic jams reduce hydrolytic efficiency of cellulase on
cellulose surface, Science 333 (6047) (2011) 1279–1282.
zyme-substrate interactions during long-time hydrolysis arising from [20] N. Cruys-Bagger, J. Elmerdahl, E. Praestgaard, H. Tatsumi, N. Spodsberg, K. Borch,
substrate evolution is key to successful development of an efficient P. Westh, Pre-steady-state kinetics for hydrolysis of insoluble cellulose by cello-
biomass deconstruction scheme. biohydrolase Cel7A, J. Biol. Chem. 287 (22) (2012) 18451–18458.
[21] J. Jalak, M. Kurašhin, H. Teugjas, P. Väljamäe, Endo-exo synergism in cellulose
hydrolysis revisited, J. Biol. Chem. 287 (2012) 28802–28815.
Declaration of interests [22] S.A. Maurer, N.W. Brady, N.P. Fajardo, C.J. Radke, Surface kinetics for cooperative
fungal cellulase digestion of cellulose from quartz crystal microgravimetry, J.
None. Colloid Interface Sci. 394 (2013) 498–508.
[23] B.C. Knott, M. Haddad Momeni, M.F. Crowley, L.F. Mackenzie, A.W. Gotz,
M. Sandgren, S.G. Withers, J. Stahlberg, G.T. Beckham, The mechanism of cellulose
Acknowledgement hydrolysis by a two-step, retaining cellobiohydrolase elucidated by structural and
transition path sampling studies, J. Am. Chem. Soc. 136 (1) (2014) 321–329.
[24] N. Cruys-Bagger, H. Tatsumi, G.R. Ren, K. Borch, P. Westh, Transient kinetics and
JN was supported by the U.S. Department of Energy, Office of rate-limiting steps for the processive cellobiohydrolase Cel7A: effects of substrate
Science, Office of Workforce Development for Teachers and Scientists, structure and carbohydrate binding domain, Biochemistry 52 (49) (2013)
Office of Science Graduate Student Research (SCGSR) program. The 8938–8948.
[25] S.A. Maurer, C.N. Bedbrook, C.J. Radke, Competitive sorption kinetics of inhibited
SCGSR program is administered by the Oak Ridge Institute for Science endo- and exoglucanases on a model cellulose substrate, Langmuir 28 (41) (2012)
and Education for the DOE under contract number DE‐SC0014664. 14598–14608.
[26] Y. Shibafuji, A. Nakamura, T. Uchihashi, N. Sugimoto, S. Fukuda, H. Watanabe,
M. Samejima, T. Ando, H. Noji, A. Koivula, K. Igarashi, R. Iino, Single-molecule
References imaging analysis of elementary reaction steps of Trichoderma reesei
Cellobiohydrolase I (Cel7A) hydrolyzing crystalline cellulose Iα and IIII, J. Biol.
[1] L.R. Lynd, P.J. Weimer, W.H. van Zyl, I.S. Pretorius, Microbial cellulose utilization: Chem. 289 (20) (2014) 14056–14065.
fundamentals and biotechnology, Microbiol. Mol. Biol. Rev. 66 (3) (2002) 506–577. [27] J. Jung, A. Sethi, T. Gaiotto, J.J. Han, T. Jeoh, S. Gnanakaran, P.M. Goodwin,
[2] P. Bansal, M. Hall, M.J. Realff, J.H. Lee, A.S. Bommarius, Modeling cellulase ki- Binding and movement of individual Cel7A cellobiohydrolases on crystalline cel-
netics on lignocellulosic substrates, Biotechnol. Adv. 27 (6) (2009) 833–848. lulose surfaces revealed by single-molecule fluorescence imaging, J. Biol. Chem.
[3] T. Jeoh, M.J. Cardona, N. Karuna, A.R. Mudinoor, J. Nill, Mechanistic kinetic 288 (33) (2013) 24164–24172.
models of enzymatic cellulose hydrolysis—a review, Biotechnol. Bioeng. 114 (7) [28] N. Cruys-Bagger, K. Alasepp, M. Andersen, J. Ottesen, K. Borch, P. Westh, Rate of
(2017) 1369–1385. threading a cellulose chain into the binding tunnel of a cellulase, J. Phys. Chem. B
[4] L. Zhong, J.F. Matthews, M.F. Crowley, T. Rignall, C. Talón, J.M. Cleary, 120 (25) (2016) 5591–5600.
R.C. Walker, G. Chukkapalli, C. McCabe, M.R. Nimlos, C.L. Brooks, M.E. Himmel, [29] J. Jalak, P. Valjamae, Mechanism of initial rapid rate retardation in cellobiohy-
J.W. Brady, Interactions of the complete cellobiohydrolase I from Trichoderma reesei drolase catalyzed cellulose hydrolysis, Biotechnol. Bioeng. 106 (6) (2010) 871–883.
with microcrystalline cellulose Iβ, Cellulose 15 (2) (2008) 261–273. [30] A. Nakamura, H. Watanabe, T. Ishida, T. Uchihashi, M. Wada, T. Ando, K. Igarashi,
[5] N. Karuna, T. Jeoh, The productive cellulase binding capacity of cellulosic sub- M. Samejima, Trade-off between processivity and hydrolytic velocity of cellobio-
strates, Biotechnol. Bioeng. 114 (3) (2017) 533–542. hydrolases at the surface of crystalline cellulose, J. Am. Chem. Soc. 136 (12) (2014)
[6] J. Jalak, P. Väljamäe, Multi-mode binding of cellobiohydrolase Cel7A from 4584–4592.
Trichoderma reesei to cellulose, PLoS One 9 (9) (2014) e108181. [31] A.R. Mudinoor, Surface Interactions of Trichoderma reesei Cel7A and its Variants
[7] T. Jeoh, C.I. Ishizawa, M.F. Davis, M.E. Himmel, W.S. Adney, D.K. Johnson, on Cellulose [PhD Disseration], University of California, Davis, 2018.
Cellulase digestibility of pretreated biomass is limited by cellulose accessibility, [32] G.T. Beckham, J. Ståhlberg, B.C. Knott, M.E. Himmel, M.F. Crowley, M. Sandgren,
Biotechnol. Bioeng. 98 (1) (2007) 112–122. M. Sørlie, C.M. Payne, Towards a molecular-level theory of carbohydrate pro-
[8] M. Kurasin, P. Valjamae, Processivity of cellobiohydrolases is limited by the sub- cessivity in glycoside hydrolases, Curr. Opin. Biotechnol. 27 (2014) 96–106.
strate, J. Biol. Chem. 286 (1) (2011) 169–177. [33] T. Imai, C. Boisset, M. Samejima, K. Igarashi, J. Sugiyama, Unidirectional processive
[9] Y.H.P. Zhang, L.R. Lynd, Determination of the number-average degree of poly- action of cellobiohydrolase Cel7A on Valonia cellulose microcrystals, FEBS Lett. 432
merization of cellodextrins and cellulose with application to enzymatic hydrolysis, (3) (1998) 113–116.
Biomacromolecules 6 (3) (2005) 1510–1515. [34] A.L. Lucius, N.K. Maluf, C.J. Fischer, T.M. Lohman, General methods for analysis of
[10] P.J. O’Dell, A.R. Mudinoor, S.J. Parikh, T. Jeoh, The effect of fibril length and ar- sequential’ n-step’ kinetic mechanisms: application to single turnover kinetics of
chitecture on the accessibility of reducing ends of cellulose Iα to Trichoderma reesei helicase-catalyzed DNA unwinding, Biophys. J. 85 (4) (2003) 2224–2239.
Cel7A, Cellulose 22 (3) (2015) 1697–1713. [35] N. Cruys-Bagger, J. Elmerdahl, E. Praestgaard, K. Borch, P. Westh, A steady-state
[11] Q. Gan, S.J. Allen, G. Taylor, Kinetic dynamics in heterogeneous enzymatic hy- theory for processive cellulases, FEBS J. 280 (16) (2013) 3952–3961.
drolysis of cellulose: an overview, an experimental study and mathematical mod- [36] K. Kipper, P. Valjamae, G. Johansson, Processive action of cellobiohydrolase Cel7A
elling, Process Biochem. 38 (7) (2003) 1003–1018. from Trichoderma reesei is revealed as ‘burst’ kinetics on fluorescent polymeric
[12] J. Lehtio, J. Sugiyama, M. Gustavsson, L. Fransson, M. Linder, T.T. Teeri, The model substrates, Biochem. J. 385 (Pt. 2) (2005) 527–535.
binding specificity of affinity determinants of family 1 and family 3 cellulose [37] I. Von Ossowski, J. Ståhlberg, A. Koivula, K. Piens, D. Becker, H. Boer, R. Harle,
binding modules, Proc. Natl. Acad. Sci. 100 (2) (2003) 484–489. M. Harris, C. Divne, S. Mahdi, Engineering the exo-loop of Trichoderma reesei cel-
[13] W.S. Adney, Y.C. Chou, S.R. Decker, S.Y. Ding, J.O. Baker, G. Kunkel, T.B. Vinzant, lobiohydrolase, Cel7A. A comparison with Phanerochaete chrysosporium Cel7D, J.
M.E. Himmel, Heterologous expression of Trichoderma reesei 1,4-beta-D-glucan Mol. Biol. 333 (4) (2003) 817–829.
Cellobiohydrolase (Cel 7A), in: S.D. Mansfield, J.N. Saddler (Eds.), Applications of [38] J. Kari, J. Olsen, K. Borch, N. Cruys-Bagger, K. Jensen, P. Westh, Kinetics of cel-
Enzymes to Lignocellulosics, American Chemical Society, Washington, D. C, 2003, lobiohydrolase (Cel7A) variants with lowered substrate affinity, J. Biol. Chem. 289
pp. 403–437. (47) (2014) 32459–32468.
[14] J.G. Linger, L.E. Taylor, J.O. Baker, T. Vander Wall, S.E. Hobdey, K. Podkaminer, [39] S.J. Horn, M. Sørlie, K.M. Vårum, P. Väljamäe, V.G.H. Eijsink, Measuring pro-
M.E. Himmel, S.R. Decker, A constitutive expression system for glycosyl hydrolase cessivity, in: J.G. Harry (Ed.), Methods in Enzymology, Academic Press, 2012, pp.
family 7 cellobiohydrolases in Hypocrea jecorina, Biotechnol. Biofuels 8 (2015) 12. 69–95.
[15] J.R. Cherry, A.L. Fidantsef, Directed evolution of industrial enzymes: an update, [40] B.Z. Shang, R. Chang, J.-W. Chu, Systems-level modeling with molecular Resolution
Curr. Opin. Biotechnol. 14 (4) (2003) 438–443. elucidates the rate-limiting mechanisms of cellulose decomposition by cellobiohy-
[16] L.E. Taylor, B.C. Knott, J.O. Baker, P.M. Alahuhta, S.E. Hobdey, J.G. Linger, drolases, J. Biol. Chem. 288 (40) (2013) 29081–29089.
V.V. Lunin, A. Amore, V. Subramanian, K. Podkaminer, Q. Xu, T.A. VanderWall, [41] E. Praestgaard, J. Elmerdahl, L. Murphy, S. Nymand, K. McFarland, K. Borch,
L.A. Schuster, Y.B. Chaudhari, W.S. Adney, M.F. Crowley, M.E. Himmel, P. Westh, A kinetic model for the burst phase of processive cellulases, FEBS J. 278
S.R. Decker, G.T. Beckham, Engineering enhanced cellobiohydrolase activity, Nat. (9) (2011) 1547–1560.

117

You might also like