The in vitro synthesis of cellulose – A mini-review

Carbohydrate Polymers 285 (2022) 119222

Contents lists available at ScienceDirect

Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol

The in vitro synthesis of cellulose – A mini-review
Anna F. Lehrhofer a, 1, Takaaki Goto a, b, 1, Toshinari Kawada c, Thomas Rosenau a, d,
Hubert Hettegger a, *
a

University of Natural Resources and Life Sciences, Vienna (BOKU), Department of Chemistry, Institute of Chemistry of Renewable Resources, Muthgasse 18, A-1190
Vienna, Austria
b
Wood K Plus – Competence Center for Wood Composites and Wood Chemistry, Altenberger Straße 69, A-4040 Linz, Austria
c
Kyoto Prefectural University, Graduate School of Life and Environmental Sciences, Nakaragi-cho 1-5, Shimogamo, Sakyo-ku, Kyoto 606-8522, Japan
d
Johan Gadolin Process Chemistry Centre, Åbo Akademi University, Porthansgatan 3, FI-20500 Åbo/Turku, Finland

A R T I C L E I N F O

A B S T R A C T

Keywords:
Anhydroglucose
Biopolymer
Cellulose
In vitro synthesis
Polysaccharide

Ring-opening polymerization

The implementation of cellulose as a green alternative to classical polymers sparks research on the synthesis of
defined derivatives of this biopolymer for various high-tech applications. Apart from the scientific challenge, the
in vitro synthesis of cellulose using a bottom-up approach provides specimens with absolutely accurate substit­
uent patterns and degrees of polymerization, not accessible from native cellulose. Synthetic cellulose exhibiting a
comparably high degree of polymerization (DP) was obtained starting from cellobiose by biocatalytic synthesis
implementing cellulase. Cationic ring-opening polymerization has been established in the last two decades,
representing an excellent means of precise modification with regards to regio- and stereoselective substitution.
This method rendered isotopically enriched cellulose as well as enantiomers of native cellulose (“L-cellulose”, “D,
L-cellulose”) accessible. In this review, techniques for in vitro cellulose synthesis are summarized and critically
compared – with a special focus on more recent developments. This is complemented by a brief overview of
alternative enzymatic approaches.

1. Introduction and scope
Cellulose, as the essential building block of cell walls, is the most
abundant natural organic polymer. Already in 1838, cellulose was
discovered, isolated, and named as such by Anselme Payen, who also
found the molecular formula of the biopolymer to be C6H10O5 (Klemm
et al., 2005). Its structure was first determined by Hermann Staudinger,
the “father” of polymer chemistry, in 1920, who identified cellulose as a
linear homopolysaccharide consisting of covalently linked β-(1 → 4)Dglucose units (Staudinger, 1920). Since the glycosidic bond is formed by
condensation of two glucose units and is thus accompanied by the loss of
water, the repeating unit is often denoted as anhydroglucose unit (AGU).
For a long time, the dimer of glucose, cellobiose, was given as a
repeating unit and is still denoted as such in literature (Brown, 1996).
However, the elucidation of the three-dimensional structure of the
cellulose-synthesizing enzyme cellulose synthase (CesA) led to the
conclusion that natural synthesis starts from glucose (as the respective
UDP derivative), and thus D-anhydroglucopyranose/AGU is universally

accepted nowadays as the fundamental building block that makes up
cellulose (French, 2017).
Cellulose – in contrast to starch and especially highly branched
amylopectin – is arranged in a characteristic straight chain and can
reach a length of up to several 1000 AGUs (French et al., 2018; Nunes,
2017). Because of the inherent chirality of the repeating unit D-anhy­
droglucopyranose with its five chiral centers (C1-C5), the resulting
polymer cellulose is also a chiral molecule. Due to the high number of
hydroxy groups present in cellulose – three hydroxy groups per AGU – an
exceptionally strong and complex inter- and intramolecular hydrogen
bond network is formed, which is the reason why the cellulose molecule
is insoluble both in water and many organic solvents. According to its
degree of order, these hydrogen bond interactions are responsible for the
formation of higher structures consisting of crystalline (highly ordered)
as well as amorphous (less highly ordered) regions. The H-bonds are also
held accountable for the biopolymer's rigidity and stability, both phys­
ically and chemically. Cellulose is a polysaccharide exhibiting poly­
morphism, with four main allomorphs being known. The major one is

* Corresponding author.
E-mail addresses: (A.F. Lehrhofer), (T. Goto), (T. Kawada),
(T. Rosenau), (H. Hettegger).
1
These authors contributed equally to this work.
/>Received 20 December 2021; Received in revised form 31 January 2022; Accepted 1 February 2022
Available online 7 February 2022
0144-8617/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( />

A.F. Lehrhofer et al.

Carbohydrate Polymers 285 (2022) 119222

cellulose I, which is the native form synthesized by the majority of living
organisms, whereas natural cellulose II is only synthesized by a very
limited number of species, most of which are bacteria (Brown, 1996;
Nunes, 2017). Native cellulose I found as a structural polymer in cell
walls is arranged in rod-shaped structures known as microfibrils. These
microfibril rods make up the stiff, supporting part of the cell wall's
biocomposite, “glued” together by xyloglucans, pectins, proteins, and
lignin for providing additional strength and plasticity (Brown, 1996;
French et al., 2018).
Cellulose I consists of the allomorphs Iα/Iβ, which differ in their
crystal packing, molecular conformation as well as H-bonding, and the
ratio of the latter two forms can significantly vary depending on the
source of cellulose (Atalla & Vanderhart, 1984). In contrast to cellulose
I, which is packed in parallel strands, the chains in cellulose II are ar­
ranged in an antiparallel manner with extensive intersheet H-bonds
resulting from differences in the crystal structure (Brown, 1996; French
et al., 2018; Nunes, 2017). Cellulose II exhibits higher thermodynamic
stability than cellulose I and can be obtained by mercerization, swelling
in aqueous sodium hydroxide, or dissolution/regeneration. The modi­
fication of cellulose I to attain cellulose II typically comes with a sig­
nificant reduction in its molecular weight (Brown, 1996). Cellulose III
can be accessed by reacting cellulose I or II with various amines (e.g.,
liquid ammonia). When cellulose is treated with glycerol at high tem­
peratures, cellulose IV is obtained (French et al., 2018; Nunes, 2017).
While the chemical structure of cellulose is rather simple (only one
monomer involved, strict linearity, strictly regioselective linkages), its
biosynthesis is by far more complex (Brown, 1996). Besides vascular

plants, which produce the majority of all naturally occurring cellulose,
also other organisms like algae, bacteria, tunicates, or oomycetes can
synthesize cellulose in a highly specific and orderly process (Bessueille &
Bulone, 2008; Brown, 1996; Saxena & Brown, 2005). Using a freezefracture technique in 1976, Brown and Montezinos established that
the production of this biopolymer involves a linear, rod-like assembly
located at the growing end of a microfibril, which they called “terminal
complex” (TC) (Brown & Montezinos, 1976). The TCs of different or­
ganisms show various morphologies, with the most frequent one being a
hexagonal rosette-shaped structure found in higher plants and some
Chlorophyta (green algae) as well as various linear structures found in
other algae (Williamson et al., 2002). The dimensions of the synthesized
microfibrils mainly depend on the precise geometry of the TC, more
specifically the number of catalytic centers located there and the relative
arrangement of the latter to one another, which are responsible for
micro-sheet formation as the first stage of crystalline cellulose con­
struction (Brown, 1996; Brown & Montezinos, 1976).
Within the terminal complex, the plant enzyme cellulose synthase
(CesA) or bacterial cellulose synthase (BcsA), both examples of glyco­
syltransferases, catalyze the formation of cellulose by glycosidic bond
formation and chain extension at the non-reducing C4 terminal unit
(Koyama et al., 1997; Olek et al., 2014). Typically, the cellulose syn­
thases are arranged in hexameric clusters called rosettes, which are
located in the plasma membrane of the cell, with each rosette subunit
containing six CesA or BcsA and thus synthesizing a total of 36 cellulose
strands per TC, which then form an elementary fibril by spontaneous
lateral aggregation and bundling (Cosgrove, 2005; Heath, 1974; Olek
et al., 2014). More recent investigations on the microfibrils' exact
structure implementing X-ray diffraction techniques rather suggest an
arrangement containing three CesA enzymes per subunit and thus only
18 cellulose strands forming an elementary fibril, however, this is still a

matter of ongoing discussion (Jarvis, 2013). The formation of β-linkages
in the glucan chains starting from α-UDP-glucopyranose as the substrate
is catalyzed within the active site of CesA, corresponding to the so-called
inverting mechanism. This inversion of the stereogenic center at the C1
position is coupled to the extension of the growing glucan chains from
their non-reducing ends (Bessueille & Bulone, 2008). The catalysis oc­
curs at the cytoplasmic side of the membrane, where the formed cellu­
lose chains are extruded through the plasma to reach the cell wall by a

so-far unidentified mechanism to fulfill their function as a structural
biopolymer (Bessueille & Bulone, 2008). The biosynthesis of cellulose
has been extensively described and reviewed (Bessueille & Bulone,
2008; Brown, 1996; Delmer, 1999; Saxena & Brown, 2005).
The isolation of cellulose from wood or other plant material is
essential for the use of this biopolymer and is conducted by pulping the
raw material. This technical process has been established especially in
the paper industry and has already been described in the literature in
detail (Biermann, 1996; Mboowa, 2021; Ragnar et al., 2014).
Here the question arises why a chemist would synthesize cellulose
when nature provides enormous amounts of this material every year.
Methods for the utilization and purification of natural cellulose are
available on a large scale. After all, the classic paper and pulp industry –
one of the most important monetary pillars in many national economies
– is based on the isolation and utilization of cellulose as a basic material
for paper, textile fibers, and cellulose derivatives. So why prepare cel­
lulose in a complex multi-step synthesis, in small quantities probably
below the gram scale?
The first, general answer is scientific curiosity. Will it be possible to
copy the process that nature has invented and optimized so perfectly for
the production of the – in terms of mass – most prominent and important

natural substance? The second answer is based on the fact that biosyn­
thesis naturally always generates cellulose. This is done with great
perfection, but on the other hand, the possibilities for alterations, e.g., in
isotope content or substituents/functionalities, are almost non-existent.
If such modifications are desired, one must always start from the natural
polymer. Reactions on cellulose are always “polymer analogous”, in the
best cases they are almost regioselective, but just not completely. This
would be different for chemically synthesized cellulose and cellulose
derivatives. By using selectively modified monomers, which, for
example, have different isotope or substitution patterns, celluloses or
cellulose derivatives could be built up with 100% selectivity, which
would otherwise not be accessible in this perfection starting from nat­
ural cellulose. A third question is hoped to be answered with the syn­
thesis of cellulose: which allomorph is obtained in such “non-natural”
syntheses, cellulose I or II, or continuous amorphous material?
Despite the structure of cellulose being determined already in 1920
and numerous attempts of chemists to either enzymatically or chemi­
cally synthesize cellulose, the published examples of actually successful
chemical in vitro synthesis of this biopolymer are rather limited. This can
be attributed to the fact that bottom-up synthesis approaches are all
quite challenging, mostly starting from glucose, the logical and obvious
monomer choice. In 2001, Kobayashi and co-workers stated that a
variable and convenient method for the synthesis of different poly­
saccharides had not yet been established (Kobayashi et al., 2001).
Cellooligomers with a defined DP can be obtained employing a topdown approach. For example, Isogai et al. reported the hydrolysis of
alkali-treated native and regenerated cellulose with dilute acid. The
obtained products exhibited a DP of 35–101 for the major, highmolecular-mass fraction and around 20 for the minor low-molecularmass fraction (Isogai et al., 2008). In 2016, Zweckmair et al. reported
a method for the preparation and isolation of well-defined cellooligo­
saccharides from cellulose. The authors acetolyzed a native cellulose
sample and isolated monodisperse peracetylated cellooligomers with a

DP of up to 20 using preparative normal-phase HPLC (Zweckmair et al.,
2016). However, these top-down approaches do not present an in vitro
cellulose synthesis and are not discussed further here.
In this review, a brief overview of existing methods for the in vitro
synthesis of cellulose using chemical as well as enzymatic methods is
provided. Kobayashi and co-workers already discussed some of the
chemical and especially enzymatic approaches for obtaining synthetic
cellulose and related polysaccharides in 2001, including some mecha­
nistic aspects (Kobayashi, 2005; Kobayashi et al., 2001). Further reviews
also discussing methods for the synthesis of polysaccharides (Kobayashi
& Makino, 2009; Yoshida, 2001) as well as generally cellulose as a
biopolymer (Klemm et al., 2005) have been published in the last
2


A.F. Lehrhofer et al.

Carbohydrate Polymers 285 (2022) 119222

decades. The present article focuses on more recent developments
within the last 20 years, with a special emphasis on the accessibility of
cellulose via chemical synthesis.

widely used one because the imidoyl group can be easily introduced, the
imidate is generally stable enough to outlast purification steps and at the
same time reactive enough to allow high yields in the glycosylation
reactions (Schmidt & Jung, 2000). The chemical structures of frequently
used glycosyl donors are shown in Fig. 2.
It is worth mentioning that the first regiocontrolled glycosylation
was reported by Shapiro et al. in 1969 for the synthesis of lactose from

2,3-di-O-acetyl-1,6-anhydro-β-D-glucopyranose with tetra-O-acetyl-α-Dgalactopyranosyl bromide (acetobromogalactose) (see Fig. 3) (Shapiro
et al., 1969).
Tackles towards the chemical synthesis of cellulose from a bottom-up
approach employing the stepwise addition method had begun with the
synthesis of cellooligosaccharides. To achieve cellulose synthesis, the
glycosylation reaction must regioselectively occur between C1 of a
glycosyl donor and C4-OH of a glycosyl acceptor. Such regiocontrol is
possible by selective protection of all other hydroxy groups except the
glycosyl acceptor's C4-OH group. To obtain cellooligomers or cellulose
and not just the dimer cellobiose, repetitive glycosylation is required to
elongate the chain. After each glycosylation step, the obtained in­
termediates need to be isolated, purified and converted into the corre­
sponding glycosyl donor or acceptor once again. The first arrow using
this stepwise addition method was fired by Freudenberg and Nagai to hit
cellobiose synthesis (DP = 2) already in 1933. Starting from levoglu­
cosan (1,6-anhydro-β-D-glucopyranose) and acetobromoglucose (1bromo-α-D-glucose tetraacetate), they reported the synthesis of octaa­
cetyl cellobiose, however, the yield was still very low (Freudenberg &
Nagai, 1933). Another example of this stepwise addition method is the
synthesis of α-cellotriose (DP = 3) by Hall and Lawler in 1971 by
condensation of selectively protected 2,3,6-tri-O-substituted methyl β-Dglucopyranosides with tetra-O-acetyl-α-D-glucopyranosyl bromide in the
first and hepta-O-acetyl-α-D-cellobiosyl bromide in the second step,
respectively. After substitution of the methyl- by acetyl-groups, α-Dcellotriose hendecaacetate was obtained, which could be efficiently
transformed into the β-anomer (Hall & Lawler, 1971). Ten years later,
Takeo and co-workers reported a similar synthesis of these α- and
β-cellotriose hendecaacetates and an additional pathway to obtain
6,6′ ,6′′ -tri-substituted derivatives of methyl β-cellotrioside (Takeo et al.,
1981). Cellotetraose (DP = 4) was first reported by Schmidt and Michel
in 1982 (Schmidt & Michel, 1982). As a starting material, they used the
α-trichloroacetimidate derivative of 2,3,4,6-tetra-O-acetyl-glucopyr­
anose, which they coupled stepwise with high β-selectivity to obtain

both linear and branched cellooligomers. In this case, three steps were
necessary to convert the obtained intermediates – the suitably protected
mono-, di- and trisaccharide – after successful glycosylation to the next
glycosyl donor (imidate) again before being able to couple it once more.
Takeo and co-workers also reported the synthesis of cellotetraose by
stepwise elongation in 1983. The authors described the selective

2. Chemical synthesis of cellulose
To achieve the chemical synthesis of cellulose, three conditions need
to be fulfilled: a) regioselectivity of the linkage (only 1 → 4 linkages and
no disubstitution, which would cause branching), b) stereoselectivity
(strictly β-configured anomeric carbon), and c) DP control (to allow
precise regulation of the chain length or at least a low dispersity). While
regiocontrol can be achieved by selective protection of the corre­
sponding monosaccharide glucose prior to the glycosylation reaction,
the choice of the reaction type itself is crucial for accomplishing precise
stereocontrol. The formation of the O-glycosidic linkage is usually
conducted by nucleophilic attack of the free hydroxy group of the
glycosyl donor at the anomeric carbon of the glycosyl acceptor carrying
a leaving group. Complete stereocontrol of the conformation at the
anomeric carbon (C1) can thus be achieved employing an SN2 reaction
at an α-configured acceptor, in which the anomeric carbon undergoes
complete inversion. On the contrary, an SN1 pathway via an oxocarbe­
nium cation intermediate (Hosoya et al., 2014; Xiao & Grinstaff, 2017)
permits nucleophilic attack from both sides of the ring and consequently
gives a racemic mixture of both the α- and β-anomers. Even though the
latter one is preferred because of the often-discussed “anomeric effect”
(in short, the adjacent oxygen atom in the pyranose ring exerts a sta­
bilizing effect on the β-anomer due to orbital interactions) (Graczyk &
Mikolajczyk, 1994; Juaristi & Cuevas, 1992), the SN1 pathway is

disadvantageous with regard to anomeric stereocontrol. To ensure ste­
reoregularity, precise selection of the type of leaving group, catalyst,
protecting groups, and the reaction conditions is required to enforce
selective progression of the glycosylation via an SN2 pathway (Fügedi,
2006; Xiao & Grinstaff, 2017; Zhu & Schmidt, 2009).
There are mainly two approaches to chemically synthesize cellulose:
1) the stepwise addition method and 2) the polymerization and poly­
condensation method.
2.1. Stepwise addition method
The stepwise method is classically employed in oligosaccharide
synthesis, where it is used to obtain a substance with defined composi­
tion, linkage pattern, and DP, or at least a very narrow dispersity. In this
method, the glycosylation reaction is repeated until the desired DP is
achieved using a derivative of the repeating unit. The glycosylation re­
action is formally a dehydrative condensation reaction between the
hydroxy group of a glycosyl donor carrying the anomeric carbon of the
new glycosidic bond and the hydroxy group of a glycosyl acceptor. An
important example, directly illustrating the principle, is the acidcatalyzed Fischer glycosylation, which is depicted in Fig. 1.
To control the configuration at the anomeric carbon, several glyco­
sylation methods utilizing glycosyl donors with specific leaving groups
and reactivities and special catalysts have been developed, inter alia the
Kă
onigs-Knorr method (Koenigs & Knorr, 1901), the Schmidt imidate
method (Schmidt & Michel, 1980), the thioglycosylation method
(Fügedi et al., 1987) and the pentenyl method (Mootoo et al., 1988). The
imidate method, mostly relying on trichloroacetimidate, is the most

Fig. 1. Acid-catalyzed Fischer glycosylation of glucose yielding a glucoside (R
= alkyl or aryl, typically methyl).

Fig. 2. Chemical structures of glycosyl donors comprising a) halide, b) tri­
chloroacetimidate, c) thioalkyl, and d) O-pentenyl leaving groups.
3


A.F. Lehrhofer et al.

Carbohydrate Polymers 285 (2022) 119222

Fig. 3. Regio- and stereo-controlled synthesis of a lactose precursor as described by Shapiro and co-workers (Shapiro et al., 1969).

condensation of hepta-O-acetyl-α-D-cellobiosyl bromide with benzyl
2,3,6,2′ ,3′ ,6′ -hexa-O-benzyl-β-D-cellobioside and subsequent depro­
tection to yield the desired product (Takeo et al., 1983).
An additional challenge the stepwise synthesis faced was the high
number of steps involved, such as those that entail byproduct formation
or require purification of the products, which significantly decrease the
overall yield of the synthesis. To overcome this problem, Nakatsubo
et al. proposed a convergent synthetic method (Nakatsubo et al., 1985).
They used a precursor with active X and Y groups at the C1-OH and C4OH, respectively, with the other hydroxy groups being protected (see
Fig. 4). The X and Y groups can be cleaved selectively, and thus the
glycosyl donor can be prepared in only two steps (removal of X group
and introduction of the leaving group Z, see Fig. 4, left), and the glycosyl
acceptor in only one step (removal of Y group, see Fig. 4, right). The
product has the same combination of terminal protecting groups as the
respective precursors (X and Y) and can thus be directly used as a re­
agent for subsequent repetitive coupling using the same chemistry and
conditions.
Using this strategy, cellopentaose to cellooctaose derivatives were
synthesized by Kawada and co-workers in the form of perbenzylated

allyl 4n’-O-p-methoxybenzyl cellooligosaccharides starting from
2,3,6,2′ ,3′ ,6′ -hexa-O-benzyl-4′ -O-(p-methoxybenzyl)-β-D-cellobioside by
repetitive alternating removal of the p-methoxybenzyl group on the one
hand and stereoselective β-glycosylation using the imidate method on
the other hand (Kawada et al., 1990). Further, the authors reported the

first synthesis of cellooctaose acetate starting from allyl 2,3,6-tri-Obenzyl-4-O-(p-methoxybenzyl)-β-D-glucopyranoside (Kawada et al.,
1994).
In this synthesis approach, the “1 + n” convergent method or linear
synthetic method, they employed benzyl protective groups at O-2, O-3,
and O-6, which significantly decreased the reactivity of both the glycon
(i.e., glycosyl donor) and the aglycon (i.e., glycosyl acceptor) with
increasing DP. Thus, an “n + n” convergent method could not be realized
employing per-benzyl protection and the synthesis of cellooligo­
saccharides bearing a DP higher than eight was found to be challenging.
Therefore, Takano and co-workers thoroughly studied the substituent
effects on the glycosylation reactions, not only regarding their reactivity
but also focusing on the stereoselectivity. They investigated various
protective groups and found that an electron-withdrawing pivaloyl
group at the O-2 and O-6 position of the glycon and an electron-donating
benzyl group at the O-3 of the aglycon were the most convenient choice.
This finely tuned protective system made an “n + n” convergent syn­
thesis of cellooligomers possible (Takano et al., 1990, 1994).
Combining the new findings by implementing this convergent
method and the ideal choice of protective groups as suggested by Takano
and co-workers (vide supra), Nishimura and Nakatsubo synthesized a
series of regioselectively substituted oligomers up to celloeicosaose (DP
= 20) starting from cellooctaose by stepwise chain extension using a
monomer derivative (Nishimura & Nakatsubo, 1997). This protective
system proved the possibility of synthesizing cellooligosaccharides with

Fig. 4. Schematic of the glycosylation procedure in the convergent synthetic method reported by Nakatsubo and co-workers (Nakatsubo et al., 1985).
4


A.F. Lehrhofer et al.

Carbohydrate Polymers 285 (2022) 119222

a higher DP, and thus opened up opportunities for later ring-opening
polymerization methods using a similar arrangement.
This stepwise method, of course, still required multiple stages to
prepare the precursors, and individual glycosylation steps are required.
In addition to these time-consuming, laborious chemical steps, also
purification is usually required after each step. Moreover, solubility and
reactivity usually get lower and lower as the DP of this “homologous
series” of the cellooligomers increases. This is especially true in the case
of a linear oligo- or polysaccharide with extensive inter- and intra­
molecular hydrogen bonds, such as cellulose, rendering the practical
aspects of the synthesis even more challenging. However, despite its
challenges, the stepwise method remains the method of choice for the
synthesis of cellooligosaccharides or cellulose with a single, strictly set,
and well-defined DP.

glucose tricarbanilate”) in a mixture of chloroform and dimethyl sulf­
oxide (DMSO) in the presence of phosphorus pentoxide for the removal
of water that is generated during polycondensation. After removal of the
protecting groups, a predominantly β-(1 → 4)-linked, slightly branched
glucan with a DP of 48–63 was obtained (Husemann & Müller, 1966).
Shortly after, Hirano reported a similar synthesis procedure using the

same glucose derivative, 2,3,6-glucose tricarbanilate, DMSO, and
phosphorus pentoxide without any additional solvents to give a similar
polysaccharide. The predominant formation of the β-linkage was
attributed to the steric hindrance of the protective groups and thus the
disfavoring of α-linkages (Hirano, 1973). Since then, no conceptually
different attempts to synthesize cellulose by a polycondensation
approach have been conducted and, to the best of our knowledge to
date, no successful stereo- and regioselective synthesis route towards
cellulose with a sufficiently high DP has been reported in the literature.

2.2. Polycondensation method

2.3. Ring-opening polymerization method

In contrast to the stepwise addition method, no intermediate pro­
tection or deprotection steps are required in the polycondensation
method for the synthesis of cellulose. Only one single type of monomer
precursor containing reactive groups at corresponding carbon atoms is
used. To enable synthesis of cellulose implementing polycondensation, a
monomeric glucose derivative exhibiting an anomeric leaving group at
C1 and a free, unprotected hydroxy group at C4, which can simulta­
neously act as a glycosyl donor as well as an acceptor, is needed (Xiao &
Grinstaff, 2017).
Already in 1941, Schlubach and Lührs reported first attempts to
polycondensate D-glucose by treatment with dry gaseous HCl. The au­
thors stated that these polyglucosans exhibited a spherical or highly
branched structure and a high molecular weight of 12 000 g/mol, which
had not been previously described in the literature for a synthetic
molecule at that time. However, the authors did not state the confor­
mation of the anomeric carbon in the polymer or whether their product

exhibited stereoregularity (Schlubach & Lührs, 1941). The first reported
attempt to synthesize a stereoregular, linear polysaccharide based on
glucose using a polycondensation approach in a controlled manner was
conducted by Haq and Whelan in 1956, which was inspired by the
Kă
onigs-Knorr method for stepwise addition (vide supra) (Haq & Whelan,
1956). Implementing a regioselective polymerization, the authors star­
ted from selectively protected 2,3,4-tri-O-acetyl-α-D-glucopyranosyl
bromide, which polycondensed upon treatment with silver oxide in
chloroform. However, their attempt caused also intramolecular side
reactions with the formation of O-acetylated levoglucosan (which is also
known as a chemical biomass tracer of carbohydrate pyrolysis) and
gentiodextrins, i.e., (1 → 6)-linked β-glucopyranose derivatives (see
Fig. 5) with a low DP of up to only nine repeating units.
A few years later, Husemann and Müller as well as Hirano attempted
polycondensation of a glucose derivative inspired by a nonregioselective polymerization approach of different monosaccharides
previously proposed by Micheel and co-workers (Micheel et al., 1961).
Husemann and Müller reported polycondensation of selectively pro­
tected
2,3,6-tri-O-(N-phenylcarbamoyl)-D-glucopyranose
(“2,3,6-

This polymerization method requires a monomer containing a
strained ring system readily available for ring-opening polymerization.
When synthesizing polysaccharides in a bottom-up approach, ringopening polymerization of anhydrosugars has become the method of
choice. An anhydroglucose derivative that either forms a strained ring
system itself or carries a cyclic protecting group can be used as a pre­
cursor. These bi- or tricyclic ring systems are prone to undergo poly­
merization upon release of the ring strain as the respective driving force
(Xiao & Grinstaff, 2017).

The precursors are typically prepared by vacuum pyrolysis of
monosaccharides and subsequent protection of the free hydroxy groups.
After that, polymerization is started by adding a Lewis acid catalyst at a
low temperature under the exclusion of water (Xiao & Grinstaff, 2017).
This general approach was established by Ruckel and Schuerch in 1966.
The reactivity of the monomers is based on the release of the strain in the
monomers. They reported successful preparation of a stereoregular
linear polysaccharide starting from 1,6-anhydro-2,3,4-tri-O-substituted
β-D-glucopyranose monomers, yielding polymeric compounds with a DP
of up to 300 (Ruckel and Schuerch 1966a and 1966b). Although this
approach – employing levoglucosan (1,6-anhydrosugar) derivatives – is
the most established and thoroughly studied one, obviously 1,4-anhy­
droglucose derivatives are needed as precursors for cellulose synthesis.
Their synthesis is made more challenging, on the one hand, by much
higher ring strain and, on the other hand, by the fact that they can be
regarded as both 1,4-anhydropyranoses as well as 1,5-anhydrofurano­
ses. Thus, the precursors can polymerize according to two different
ring-opening pathways to give either (1 → 4)-pyranosidic or (1 → 5)furanosidic repeating units, both possibly with α- or β-linkages (see
Fig. 6).
Synthesis of cellulose-type polymeric compounds requires selective
1,4-cleavage of a 1,4-anhydropyranose, in most cases via a tri­
alkyloxonium ion intermediate. As the 1,5-anhydro ring oxygen pos­
sesses higher basicity than the 1,4-anhydro ring oxygen (Xiao &
Grinstaff, 2017), the formation of C1-O+-C5 oxonium ions as in­
termediates is preferred. Consequently, (1 → 5)-linked glucofuranosides
are predominantly formed during polymerization. Thus, the precise se­
lection of catalyst, choice of protecting groups, and careful control of the
reaction conditions are required to direct the preferred pathway and
accomplish selective 1,4-scission for the successful synthesis of cellulose
(Uryu, Yamanouchi, Hayashi, et al., 1983; Uryu, Yamanouchi, Kato,

et al., 1983; Xiao & Grinstaff, 2017). The catalyst must selectively
complex the 1,4-anhydro ring oxygen, and at the same time, the 1,4anhydro ring oxygen of the co-reacting monomer must nucleophili­
cally attack C1 from the reverse side of the C1-O4 bond. Due to the
involvement of the oxonium species, the reaction type is specified as
“cationic”, and CROP is often used as an abbreviation for the “cationic
ring-opening polymerization” sequence.
The cationic ring-opening polymerization of 1,4-anhydro-2,3,6-tri-

Fig. 5. Structures of levoglucosan (1,6-anhydro-β-D-glucopyranose) and gen­
tiobiose (1-β-D-glucopyranosyl-6-D-glucopyranose).
5


A.F. Lehrhofer et al.

Carbohydrate Polymers 285 (2022) 119222

Fig. 6. Ring-opening pathways of 1,4-anhydrosugar derivatives (redrawn from (Xiao & Grinstaff, 2017)).

O-benzyl-α-D-glucopyranose was reported by Micheel and Brodde as
well as Uryu et al., but mainly (1 → 5)-glucofuranose repeating units
were obtained due to the preferred glucopyranose-scission pathway.
However, they also attributed this outcome to the steric hindrance of the
bulky benzyloxy-groups, suspecting inhibition of complexation by the
catalyst (Micheel & Brodde, 1974; Uryu et al., 1985).
In 1994, Kochetkov reported the synthesis of a completely stereo­
regular β-(1 → 4)-D-glucopyranoside under high pressure by poly­
condensation of 4,6-di-O-acetyl-3-O-trityl-1‚2-O-cyanoethylidene-α-Dglucose, however, the author did not reveal detailed information about
the conditions or the product (Kochetkov, 1994). Kamitakahara and coworkers described a synthetic approach using 1,4-anhydro derivatives,
such as 1,4-anhydro-3,6-di-O-benzyl-2-O-pivaloyl-α-D-glucopyranose,

which led to a regular 3,6-di-O-benzyl-2-O-pivaloyl-β-(1 → 5)-D-gluco­
furanan and the respective non-natural polysaccharide after depro­
tection (Fig. 7, top) (Kamitakahara et al., 1994). The authors stated that
only the 1,4-anhydro-3,6-di-O-benzyl-2-O-pivaloyl-α-D-glucopyranose,
but not the 1,4-anhydro-2,3-di-O-benzyl-6-O-pivaloyl-α-D-glucopyr­
anose, led to a stereoregular polymer, which they attributed to a
participation of the pivaloyl as the neighboring group at the O-2 during
the polymerization. Zachoval and Schuerch already previously dis­
cussed this phenomenon of neighboring group participation of ester
groups during polymerization of sugars in detail (Zachoval & Schuerch,
1969). Additionally, coordination of the catalyst PF5 is facilitated, as the
benzyl group at the O-6 position increases the electron density of the
1,5-anhydro ring oxygen (Kamitakahara et al., 1994). Shortly after that,
in 1996, Nakatsubo et al. achieved the first successful synthesis of

cellulose by CROP using a derivative of the precursor implemented by
Kamitakahara. Instead of incorporating simple 1,4-anhydro derivatives,
they used a different intermediate in which O-1 and O-4 were separated
by a C1-spacer. This way, sufficient strain – and thus reactivity – was
maintained and the competitive glucopyranose ring scission was pre­
vented. The OH groups 1, 2, and 4 were integrated into an orthopivalate
structure, a system with high ring strain, but different geometry from the
above 1,4-anhydroglucopyranoses. Starting from D-glucose and
involving several protection and deprotection steps, 3,6-di-O-benzylα-D-glucose 1,2,4-orthopivalate was synthesized as the reactive precur­
sor, which underwent highly selective ring-opening catalyzed by trity­
lium tetrafluoroborate (Ph3CBF4) to give a highly stereoregular, β-(1 →
4)-linked cellulose derivative with a DP of about 20 after cleavage of the
protecting groups (Fig. 7, bottom). The pivaloyl group remained at O-2,
while O-6 and O-3 were protected by benzyl groups from the beginning.
A careful selection of the protecting groups was crucial to direct the

mechanism towards selective ring-opening (the desired 1,4-cleavage vs.
the 1,2- and 2,4-alternatives) and thus the formation of selectively
linked β-(1 → 4)-glucopyranose repeating units (Nakatsubo et al., 1996).
Especially the 3-O-benzyl group was found to be essential for cellu­
lose formation when studying the substituent effect of the protecting
group in the O-3 position. According to Kamitakahara and co-workers –
due to the rather sterically demanding axial 3-O-benzyl group – the
polymerization of this orthoester might proceed via the so-called diox­
alenium ion mechanism rather than the trialkyloxonium ion mechanism
proposed for other stereoregular ring-opening polymerizations of
anhydrosugars (Kamitakahara et al., 1996). Hori and co-workers further
studied the influence of the orthoester group, which was implemented
during this synthesis of cellulose. Studying orthopropionate-, orthoace­
tate- as well as orthobenzoate-derivatives and comparing their poly­
merization behavior to the respective orthopivalate derivative, they
found that the latter was crucial for regioregularity, as the other de­
rivatives gave a mixture of β-(1 → 4)-linked and β-(1 → 2)-linked
glucopyranose-derivatives (Hori et al., 1997; Yoshida, 2001).
The authors further adapted their polymerization method and suc­
cessfully synthesized 13C-labeled cellulose with 99% isotopic enrich­
ment, which was otherwise not accessible employing in vivo synthesis
(Adelwă
ohrer et al., 2009) and was used to study dynamic changes of the
hydrogen bond patterns during swelling and dissolution (Rosenau et al.,
2019). Recently, also the enantiomeric mirror image of native cellulose,
i.e., “L-cellulose”, was obtained by this approach (see Fig. 8 for the
chemical structure). The successful synthesis of β-(1 → 4)-L-glucopyr­
anan starting from L-glucose exhibiting an average DP of 32.8 and an
Mw/Mn ratio of 1.97 was reported by Yagura et al. in 2020, and its
enantiomeric identity was confirmed by optical rotation of the respec­

tive peracetate derivative. The synthesized L-cellulose triacetate had a
positive specific optical rotation of +8.3◦ , whereas authentic D-cellulose
triacetate had a negative specific optical rotation of − 23.4◦ (Yagura
et al., 2020).

Fig. 7. First successful chemical synthesis of cellulose by CROP of orthopivalate
precursors as reported by Nakatsubo and co-workers. Reprinted with permis­
sion from Nakatsubo et al., 1996. Copyright © 1996 American Chemi­
cal Society.
6


A.F. Lehrhofer et al.

Carbohydrate Polymers 285 (2022) 119222

Fig. 8. Chemical structure of native “D”-cellulose compared to its artificial enantiomer “L-cellulose” as synthesized by Yagura et al. (Yagura et al., 2020).

Shortly after, the authors also reported the synthesis of optically
inactive cellulose using the same method, starting from a “racemic
mixture” of D- and L-glucose. The equimolar mixture of the corre­
sponding orthopivalate precursors yielded “racemic” 3-O-benzyl-2,6-diO-pivaloyl-β-(1 → 4)-D,L-glucopyranan, which was deprotected and
subsequently derivatized to give an acetylated derivative with an
average DP of 18.6 and a specific optical rotation of +0.01◦ , suggesting
that the underivatized biopolymer consisted of a nearly racemic mixture
of D- and L-anhydroglucose units, namely “D,L-cellulose” (Ikegami et al.,
2021).
Further, CROP was used to synthesize highly regioselectively alky­
lated cellulose derivatives. Kamitakahara and co-workers prepared 2,6di-O-methylcellulose by CROP of the previously reported precursor 3,6di-O-benzyl-α-D-glucose 1,2,4-orthopivalate, removal of the pivaloyl
protecting groups, methylation using MeI in the O-2 and O-6 position,

and subsequent three-step deprotection of O-3 (Kamitakahara et al.,
2008). Moreover, Kamitakahara et al. studied the influence of regiose­
lectively substituted synthetic cellulose bearing methyl and ethyl groups
in the O-2 and O-3 positions on the biopolymer's solubility. The authors
synthesized the cellulose derivatives by polymerization of 1,2,4-ortho­
pivalate-type precursors, partly already carrying ethyl or methyl sub­
stituents in the O-6 position of the monomer and thus demonstrating
derivatization before CROP (Kamitakahara et al., 2009). In 2010,
Kamitakahara and co-workers reported the successful synthesis of
regioselectively substituted 2-O-, 3-O- and 6-O-ethyl celluloses. Starting
from a 1,2,4-orthopivalate cellulose precursor carrying different sub­
stituents in O-2, O-3, and O-6 positions, they prepared the polymeric
compound by CROP and subsequent (multiple) deprotection steps to
study the structure/property relationship of the cellulose derivatives
(Kamitakahara et al., 2010).

microscopy, and were much shorter compared to the typical fibrils
produced by Acetobacter xylinum cellulose synthase in vivo, which might
be attributed to the absence of the TCs in the solubilized, isolated
enzyme (Lin et al., 1985).
A major drawback of cellulose synthesis by the biosynthetic pathway
using glycosyltransferases is the comparatively high cost of the glycosyl
donor in the case of UDP-forming cellulose synthase. Further, these
enzymes are prone to product-inhibition, which significantly diminishes
the yield (Kobayashi et al., 2001). Thus, also non-biosynthetic pathways
are frequently employed. In 1991, the first confirmed enzymatic syn­
thesis of cellulose was accomplished by Kobayashi and co-workers. The
authors reported an entirely non-biosynthetic pathway for cellulose
synthesis implementing an enzyme catalyst, a glycosylase or glycosyl
hydrolase of which the biocatalytic activity was reversed. Glycosylases

are a convenient choice for biocatalysis since they usually have high
glycosylation activity and are comparatively tolerant against organic
solvents. They are, furthermore, readily available and relatively wellstudied (Kobayashi et al., 2001). Cellulase, an example of glyco­
sylases, is originally a degrading (hydrolyzing) enzyme. As using cellu­
lase for cellulose synthesis is the reverse reaction, various ideas, mainly
an equilibrium control and kinetic control, have been developed. Shoda
and co-workers developed the glycosyl fluoride method as a new tech­
nology for glycosylation reactions (Shoda et al., 2016). Kobayashi and
co-workers implemented this technique and used β-D-cellobiosyl fluo­
ride as a starting material, which was polymerized with cellulase as the
enzymatic catalyst in acetonitrile/acetate buffer (pH 5) to obtain the
first-ever example of in vitro synthesized cellulose. To promote the ste­
reoregular formation of the β-(1 → 4)-glycosidic linkage, the β-config­
uration of the starting substrate was crucial. The authors confirmed the
structure of the isolated polymeric compound by comparing the 13C
solid-state NMR and IR spectra to the respective spectra of native cel­
lulose and by conducting a hydrolysis experiment. Further, they acety­
lated the synthetic compound and measured a DP of at least 22 using gel
permeation chromatography, which is still relatively low compared to
native cellulose (Kobayashi et al., 1991). Using an extensively purified
enzyme, Lee and co-workers achieved the first and highly remarkable in
vitro synthesis of the native cellulose I allomorph (Lee et al., 1994). This
technique was further extended to the synthesis of cellulose derivatives
exhibiting, inter alia, alternating methyl groups at C6-OH (Okamoto
et al., 1997) or glucose and N-acetylglucosamine units (Kobayashi et al.,
2006) using fluorinated disaccharide monomers and cellulase catalysis.
A major drawback of this method, however, is the rather complex,
multistep synthesis of the glycosyl fluorides, making the use of different
monomers desirable (Shoda et al., 2016).
Samain et al. successfully synthesized cellooligosaccharides

employing cellodextrin phosphorylase (CDP) for catalysis, together with
cellobiosyl derivatives and α-D-glucopyranosyl 1-phosphate as glycosyl
donors. However, the rather costly glycosyl donor was also a major
drawback in this case and the achieved DP of around eight was
comparatively low (Samain et al., 1995). Starting from glucose and α-Dglucopyranosyl 1-phosphate, Hiraishi and co-workers synthesized
highly ordered cellulose II cellooligosaccharides also using CDP as a
catalyst (Hiraishi et al., 2009). Serizawa et al. were the first ones to
implement C1-modified β-D-glucose derivatives as end-group substrates

3. Enzymatic synthesis of cellulose
As nature is synthesizing cellulose through enzymes far more spe­
cifically and efficiently than any chemist in the laboratory, it is logical
that several enzymatic in vitro approaches towards cellulose have been
proposed. Regioselective and stereoregular formation of the β-(1 → 4)glycosidic linkage is especially challenging by chemical synthesis, while
native enzymes can do this with ease. Enzymatic catalysis for the in vitro
synthesis of cellulose and similar polysaccharides, including mecha­
nistic aspects, has already been reviewed in detail by Kobayashi
(Kobayashi et al., 2001; Kobayashi & Makino, 2009), Kadokawa
(Kadokawa, 2011), and Shoda et al. (Shoda et al., 2016). Thus, only a
brief overview is provided in this review for the sake of completeness.
Using a UDP-forming cellulose synthase enzyme isolated from Ace­
tobacter xylinum, Aloni et al. and Lin et al. synthesized a β-(1 → 4)-Dglucan (Aloni et al., 1982; Lin et al., 1985). Aloni et al. used a PEGsupported enzyme to reach a rate of almost 40% of in vivo cellulose
production. However, they only concluded that the reaction product was
cellulose by using derivatization or digestion with cellulose hydrolyzing
enzymes, but did not provide detailed analytical data or information
about the crystalline and microfibrillar structure (Aloni et al., 1982).
Shortly after, Lin et al. reported an in vitro synthesis of cellulose micro­
fibrils which, however, accumulated in clusters, according to electron
7

A.F. Lehrhofer et al.

Carbohydrate Polymers 285 (2022) 119222

(“primers”) in CDP-catalyzed polycondensation of α-D-glucopyranosyl 1phosphate to allow precise end-group modification of the cellulose chain
at its reducing end. Due to the poor substrate specificity of CDP, the
modified glycosyl acceptors were readily accepted by the enzyme. For
example, it was used as a catalyst in a related synthesis of alkyl β-cel­
lulosides, i.e., O-alkyl reducing-end-modified cellooligomers. By varying
the length of this single alkyl group, the crystallization behavior of the
oligomers could be influenced. While the use of n-butyl or shorter-chain
alkyls promoted the formation of antiparallel cellulose II, n-hexyl or
longer chains led to the parallel cellulose I allomorph, which also
affected the self-assembly of the strands into different tertiary structures
(Serizawa et al., 2021; Yataka et al., 2016). The authors also synthesized
azide-containing cellulose oligomers by the same path. Implementing a
β-glucosyl azide primer as the glycosyl acceptor, Yataka et al. synthe­
sized a cellulose II crystalline allomorph with reactive N3-groups situ­
ated at the reducing ends on the surface prone to post-functionalization
by Cu(I)-catalyzed Huisgen alkyne-azide cycloaddition reactions (Yataka
et al., 2015). The same group recently reported the synthesis of block
copolymers of cellooligosaccharides and oligo(ethylene glycol) using
bifunctional oligomeric primers and CDP catalysis, in which a cellulose
II-like crystalline structure was observed for the products (Sugiura et al.,
2021).
In 2007, Egusa et al. successfully synthesized cellulose exhibiting a
high DP from non-substituted cellobiose, implementing catalysis with a
commercially available cellulase. They employed the well-known N,Ndimethylacetamide (DMAc)/LiCl solvent system: first, to overcome the
solubility issue of cellulose, and second, to prevent partial inactivation

of the cellulase by acetonitrile, which were the main reasons for the low
DP of the product previously obtained by Kobayashi and co-workers.
The authors introduced the enzyme in the form of a cellulase/surfac­
tant complex. Using this method, synthetic cellulose exhibiting a DP
>100 was obtained, whose structure was confirmed by NMR spectros­
copy as well as X-ray diffraction. However, the yield of higher molecular
weight products (DP ≥6) was less than 5%, and more than half of the
starting cellobiose was not consumed, which was attributed to a possible
decreasing effect of aprotic polar solvents such as DMAc or DMSO on
enzymatic activity (Egusa et al., 2007). Later, Egusa et al. further opti­
mized the conditions and were able to achieve higher conversion rates. A
surfactant-enveloped enzyme (SEE) and a protic component (acid) as an
additive in the same aprotic organic solvent system enabled the syn­
thesis of cellulose with a DP of more than 120 at approximately 26%
conversion (see Fig. 9). Furthermore, the protective surfactant in the
SEE, i.e., dioleyl-N-D-glucona-L-glutamate, largely prevented deactiva­
tion of the cellulase in the strongly polar, aprotic organic solvent system
DMAc/LiCl and thus allowed for a more efficient biocatalytic synthesis
(Egusa et al., 2012).

4. Concluding remarks and outlook
Only limited options are available to efficiently synthesize cellulose
in vitro, even though much effort has been put into establishing efficient,
convenient, and practical protocols. The complete regio- and stereo­
control during the synthesis of the polysaccharide is especially chal­
lenging and requires accurate fine-tuning of protecting groups and
reaction conditions. The method of choice strongly depends on the
purpose of the synthesis. If isotopically labeled and completely regio­
selectively substituted cellulose or cellulose with a discrete DP (defined
chain length, no molar mass distribution) are needed, chemical synthesis

might be a suitable choice. Especially the stepwise addition method can
provide cellulose (cellooligosaccharides) with exactly defined DP (or at
least a very low dispersity). However, the high number of synthetic steps
required and the low amounts accessible are significant drawbacks, as is
the limited accessibility of higher DPs. Chemical polymerization, espe­
cially the ring-opening polymerization approach, is also able to provide
pure cellulose. Yet, this method is prone to various quenching mecha­
nisms, which again significantly limits the accessible DP range. With this
method, however, D-cellulose, isotopically labeled celluloses, cellulose
derivatives with complete substitution selectivity as well as the enan­
tiomers “L-cellulose” and optically-inactive “D,L-cellulose” have been
synthesized. In general, celluloses chemically synthesized are identical
to genuine celluloses according to all analytical techniques, and usually
belong to the cellulose II allomorph. In vitro chemical synthesis of cel­
lulose exhibiting a type I crystalline structure may thus also be desirable
and was already achieved under special conditions (Lee et al., 1994). If a
higher molecular weight is required than accessible by chemical syn­
thesis, enzyme-mediated synthesis is a convenient choice, although the
conversion yield of the literature protocols is still very low due to the
deactivation of the implemented cellulases by the required organic (co-)
solvent, and further purification of the product might be more
challenging.
A common challenge in all the discussed methods is the insolubility
of cellulose in water as well as in almost all organic solvents. Especially
with increasing DP, the biopolymer precipitates and renders efficient
elongation of the chain challenging, both by chemical and enzymatic
means. In terms of enzymatic catalysis, the available solvent systems,
such as DMAc/LiCl, are critical for the enzyme's activity. For efficient
polymerization, the careful exclusion of any type of quenching agent is
crucial for obtaining a high DP. Even a solvent-free method might be

accessible in the future, also rendering the synthesis more convenient in
terms of workup and cellulose isolation. Soon, novel bottom-up ap­
proaches to cellulose and cellulose derivative synthesis can be expected,
which make a completely regioselective substitution of the biopolymer
or well-defined isotopomers easier accessible and thus more widely
available for specialized studies.

Fig. 9. Biocatalytic cellulose synthesis using a surfactant-enveloped enzyme (SEE) and a protic component (acid) in an organic medium. Reprinted with permission
from Egusa et al., 2012, Copyright © 2012 American Chemical Society.
8


Carbohydrate Polymers 285 (2022) 119222

A.F. Lehrhofer et al.

CRediT authorship contribution statement

Heath, I. B. (1974). A unified hypothesis for the role of membrane bound enzyme
complexes and mierotubules in plant Cell Wall synthesis. Journal of Theoretical
Biology, 48, 445–449. />Hiraishi, M., Igarashi, K., Kimura, S., et al. (2009). Synthesis of highly ordered cellulose
II in vitro using cellodextrin phosphorylase. Carbohydrate Research, 344, 2468–2473,
10/fc35dj.
Hirano, S. (1973). The preparation of a cellulose-like polymer from 2,3,6-tri-O-(N
-phenylcarbamyl)-D-glucopyranose by the action of phosphorus pentoxide in
dimethyl sulfoxide. Agricultural and Biological Chemistry, 37, 187–189, 10/gmxs5r.
Hori, M., Kamitakahara, H., & Nakatsubo, F. (1997). Substituent effects of the orthoester
group on ring-opening polymerization of α-D-glucopyranose 1,2,4-orthoester
derivatives. Macromolecules, 30, 6. />Hosoya, T., Takano, T., Kosma, P., & Rosenau, T. (2014). Theoretical foundation for the
presence of oxacarbenium ions in chemical glycoside synthesis. The Journal of

Organic Chemistry, 79, 7889–7894, 10/f6gx5h.
Husemann, V. E., & Müller, G. J. M. (1966). Über die Synthese unverzweigter
Polysaccharide. Makromolekulare Chemie, 91, 212–230, 10/bsbcc9.
Ikegami, W., Kamitakahara, H., Teramoto, Y., & Takano, T. (2021). Synthesis of optically
inactive cellulose via cationic ring-opening polymerization. Cellulose, 28,
6125–6132, 10/gj6hq5.
Isogai, T., Yanagisawa, M., & Isogai, A. (2008). Degrees of polymerization (DP) and DP
distribution of dilute acid-hydrolyzed products of alkali-treated native and
regenerated celluloses. Cellulose, 15, 815–823, 10/fbn7jd.
Jarvis, M. C. (2013). Cellulose biosynthesis: Counting the chains. Plant Physiology, 163,
1485–1486, 10/gncwbf.
Juaristi, E., & Cuevas, G. (1992). Recent studies of the anomeric effect. Tetrahedron, 48,
5019–5087, 10/dxtgt6.
Kadokawa, J. (2011). Precision polysaccharide synthesis catalyzed by enzymes. Chemical
Reviews, 111, 4308–4345. />Kamitakahara, H., Funakoshi, T., Nakai, S., et al. (2010). Synthesis and structure/
property relationships of regioselective 2-O-, 3-O- and 6-O-ethyl celluloses.
Macromolecular Bioscience, 10, 638–647, 10/czz73n.
Kamitakahara, H., Funakoshi, T., Takano, T., & Nakatsubo, F. (2009). Syntheses of 2,6-Oalkyl celluloses: Influence of methyl and ethyl groups regioselectively introduced at
O-2 and O-6 positions on their solubility. Cellulose, 16, 1167–1178, 10/fvdfd9.
Kamitakahara, H., Hori, M., & Nakatsubo, F. (1996). In , 29. Substituent effect on ringopening polymerization of regioselectively acylated β-D-glucopyranose 1,2,4-orthopivalate
derivatives (p. 6).
Kamitakahara, H., Koschella, A., Mikawa, Y., et al. (2008). Syntheses and comparison of
2,6-di-O-methyl celluloses from natural and synthetic celluloses. Macromolecular
Bioscience, 8, 690–700, 10/cwtms4.
Kamitakahara, H., Nakatsubo, F., & Murakami, K. (1994). In , 27. Ring-opening
polymerization of 1,4-anhydro-α-D-glucopyranose derivatives having acyl groups and
synthesis of (1→5)-β-D-glucofuranan (p. 6).
Kawada, T., Nakatsubo, F., & Murakami, K. (1990). Synthetic studies of cellulose. VII:
Synthesis of a series of perbenzylated allyl 4n’-O-p-methoxybenzyl cellooligosaccharides up to octamer. Cellulose Chemistry and Technology, 343–350.
Kawada, T., Nakatsubo, F., Umezawa, T., et al. (1994). Synthetic studies of cellulose. XII:

First chemical synthesis of cellooctaose acetate. Mokuzai Gakkaishi, 7, 738–743.
Klemm, D., Heublein, B., Fink, H.-P., & Bohn, A. (2005). Cellulose: Fascinating
biopolymer and sustainable raw material. Angewandte Chemie, International Edition,
44, 3358–3393, 10/bs2fw2.
Kobayashi, S. (2005). Challenge of synthetic cellulose. Journal of Polymer Science, Polymer
Chemistry Edition, 43, 693–710, 10/d2cdrt.
Kobayashi, S., Kashiwa, K., Kawasaki, T., & Shoda, S. (1991). Novel method for
polysaccharide synthesis using an enzyme: The first in vitro synthesis of cellulose via
a nonbiosynthetic path utilizing cellulase as catalyst. Journal of the American
Chemical Society, 113, 3079–3084, 10/cvrs7k.
Kobayashi, S., & Makino, A. (2009). Enzymatic polymer synthesis: An opportunity for
green polymer chemistry. Chemical Reviews, 109, 5288–5353, 10/df2w7c.
Kobayashi, S., Makino, A., Matsumoto, H., et al. (2006). Enzymatic polymerization to
novel polysaccharides having a glucose-N-acetylglucosamine repeating unit, a
cellulose− chitin hybrid polysaccharide. Biomacromolecules, 7, 1644–1656, 10/
cmzn45.
Kobayashi, S., Sakamoto, J., & Kimura, S. (2001). In vitro synthesis of cellulose and
related polysaccharides. Progress in Polymer Science, 26, 1525–1560, 10/fp7 ppf.
Kochetkov, N. K. (1994). Recent developments in the synthesis of polysaccharides and
stereospecificity of glycosylation reactions. In Studies in natural products chemistry
(pp. 201–266). Elsevier.
Koenigs, W., & Knorr, E. (1901). Ueber einige Derivate des Traubenzuckers und der
Galactose. Berichte der Deutschen Chemischen Gesellschaft, 34, 957–981, 10/cf2wv7.
Koyama, M., Helbert, W., Imai, T., et al. (1997). Parallel-up structure evidences the
molecular directionality during biosynthesis of bacterial cellulose. Proceedings of the
National Academy of Sciences, 94, 9091–9095, 10/d894q2.
Lee, J. H., Brown, R. M., Kuga, S., et al. (1994). Assembly of synthetic cellulose I.
Proceedings of the National Academy of Sciences, 91, 7425–7429. />10.1073/pnas.91.16.7425
Lin, F. C., Malcolm Brown, R., Cooper, J. B., & Delmer, D. P. P. (1985). Synthesis of fibrils
in vitro by a solubilized cellulose synthase from azobacter xylinum. Science, 230,

822–825. />Mboowa, D. (2021). A review of the traditional pulping methods and the recent
improvements in the pulping processes. Biomass Conversion and Biorefinery. https://
doi.org/10.1007/s13399-020-01243-6
Micheel, F., & Brodde, O.-E. (1974). Polymerisation von 1,4-Anhydro-2,3,6-tri-O-benzylα-D-glucopyranose. Justus Liebigs Annalen der Chemie, 1974, 702–708, 10/fnv7dh.

Anna F. Lehrhofer: Writing – original draft, Visualization. Takaaki
Goto: Writing – original draft, Visualization. Toshinari Kawada:
Writing – review & editing, Supervision. Thomas Rosenau: Conceptu­
alization, Writing – review & editing, Supervision, Project administra­
tion, Funding acquisition. Hubert Hettegger: Conceptualization,
Writing – review & editing, Visualization, Supervision, Project admin­
istration, Funding acquisition.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgements
We would like to thank the University of Natural Resources and Life
Sciences, Vienna (BOKU), the County of Lower Austria, and Lenzing AG
for their financial support in the framework of the “Austrian Biorefinery
Center Tulln” (ABCT) and the BOKU doctoral school “Advanced Bio­
refineries: Chemistry & Materials” (ABC&M). The financial support by
ărderung
Wood K plus (T.G.) and the GFF Gesellschaft fỹr Forschungsfo
ăsterreich m.b.H. (A.F.L. and H.H., project LSC20-002) is grateư
Niedero
fully acknowledged.
References
Adelwă
ohrer, C., Takano, T., Nakatsubo, F., & Rosenau, T. (2009). Synthesis of 13Cperlabeled cellulose with more than 99% isotopic enrichment by a cationic ringopening polymerization approach. Biomacromolecules, 10, 2817–2822, 10/brw7jb.

Aloni, Y., Delmer, D. P., & Benziman, M. (1982). Achievement of high rates of in vitro
synthesis of 1,4-β-D-glucan: Activation by cooperative interaction of the acetobacter
xylinum enzyme system with GTP, polyethylene glycol, and a protein factor.
Proceedings of the National Academy of Sciences of the United States of America,
6448–6452, 10/cswkgb.
Atalla, R. H., & Vanderhart, D. L. (1984). Native cellulose: A composite of two distinct
crystalline forms. Science, 223, 283–285, 10/d7kjp4.
Bessueille, L., & Bulone, V. (2008). A survey of cellulose biosynthesis in higher plants.
Plant Biotechnology, 25, 315–322, 10/bkh69r.
Biermann, C. J. (1996). Handbook of pulping and papermaking (2nd ed.). Corvallis,
Oregon: Academic Press - An Imprint of Elsevier.
Brown, R. M. (1996). The biosynthesis of cellulose. Journal of Macromolecular Science,
Part A, 33, 1345–1373, 10/bwbd7h.
Brown, R. M., & Montezinos, D. (1976). Cellulose microfibrils: Visualization of
biosynthetic and orienting complexes in association with the plasma membrane.
Proceedings of the National Academy of Sciences, 73, 143–147, 10/fw3kgf.
Cosgrove, D. J. (2005). Growth of the plant cell wall. Nature Reviews. Molecular Cell
Biology, 6, 850–861, 10/bn26z3.
Delmer, D. P. P. (1999). Cellulose biosynthesis: Exciting times for a difficult field of
study. Annual Review of Plant Physiology and Plant Molecular Biology, 50, 245–276,
10/fv4r3q.
Egusa, S., Goto, M., & Kitaoka, T. (2012). One-step synthesis of cellulose from cellobiose
via protic acid-assisted enzymatic dehydration in aprotic organic media.
Biomacromolecules, 13, 2716–2722, 10/f379b4.
Egusa, S., Kitaoka, T., Goto, M., & Wariishi, H. (2007). Synthesis of cellulose in vitro by
using a cellulase/surfactant complex in a nonaqueous medium. Angewandte Chemie,
International Edition, 46, 2063–2065, 10/c5fhw6.
French, A. D. (2017). Glucose, not cellobiose, is the repeating unit of cellulose and why
that is important. Cellulose, 24, 4605–4609, 10/gch6nf.
French, A. D., P´erez, S., & Bulone, V. (2018). Cellulose. In Encyclopedia of polymer science

and technology (1st ed., pp. 1–69). John Wiley & Sons, Inc.
Freudenberg, K., & Nagai, W. (1933). Die Synthese der Cellobiose. Berichte der Deutschen
Chemischen Gesellschaft A/B, 66, 27–29, 10/d2pwqw.
Fügedi, P. (2006). Glycosylation methods. In The Organic Chemistry of Sugars (pp.
89–180). Boca Raton: CRC Press.
Fỹgedi, P., Garegg, P. J., Lă
onn, H., & Norberg, T. (1987). Thioglycosides as glycosylating
agents in oligosaccharide synthesis. Glycoconjugate Journal, 4, 97–108, 10/cs5pdb.
Graczyk, P. P., & Mikolajczyk, M. (1994). Anomeric effect: Origin and consequences. In
Topics in stereochemistry (pp. 159–350). New York, New York: John Wiley & Sons,
Inc.
Hall, D. M., & Lawler, T. E. (1971). New routes to the synthesis of 2,3,6-tri-O-substituted
methyl β-D-glucopyranosides an improved synthesis of α-cellotriose hendecaacetate.
Carbohydrate Research, 1–7. />Haq, S., & Whelan, W. J. (1956). 873. The chemical synthesis of polysaccharides. Part I.
Synthesis of gentiodextrins. Journal of the Chemical Society, 4543–4549, 10/dwk4d7.

9


A.F. Lehrhofer et al.

Carbohydrate Polymers 285 (2022) 119222

Micheel, V. F., Bă
ockmann, A., & Meckstroth, W. (1961). Darstellung und Struktur
synthetischer Polysaccharide. Makromolekulare Chemie, 48, 1–16, 10/cbm8rc.
Mootoo, D. R., Vandana, D., & Bert, F.-R. (1988). N-pentenyl glycosides permit the
chemospecific liberation of the anomeric center. Journal of the American Chemical
Society, 110, 2662–2663, 10/dtb2rz.
Nakatsubo, F., Kamitakahara, H., & Hori, M. (1996). Cationic ring-opening

polymerization of 3,6-Di-O-benzyl-α-D-glucose 1,2,4-orthopivalate and the first
chemical synthesis of cellulose. Journal of the American Chemical Society, 118,
1677–1681, 10/fd6vqw.
Nakatsubo, F., Takano, T., & Kawada, T. (1985). Synthetic studies of cellulose, 1:
Synthetic design and selection of the protective groups. Memoirs of the College of
Agriculture, Kyoto University, 127, 37–47.
Nishimura, T., & Nakatsubo, F. (1997). Chemical synthesis of cellulose derivatives by a
convergent synthetic method and several of their properties. Cellulose, 4, 109–130.
/>Nunes, R. C. R. (2017). Rubber nanocomposites with nanocellulose. In Progress in rubber
nanocomposites (pp. 463–494). Elsevier.
Okamoto, E., Kiyosada, T., Shoda, S.-I., & Kobayashi, S. (1997). Synthesis of alternatingly
6-O-methylated cellulose via enzymatic polymerization of a substituted cellobiosyl
fluoride monomer catalyzed by cellulase. Cellulose, 4, 161–172. />10.1023/A:1018479621509
Olek, A. T., Rayon, C., Makowski, L., et al. (2014). The structure of the catalytic domain
of a plant cellulose synthase and its assembly into dimers. Plant Cell, 26, 2996–3009,
10/f6hj52.
Ragnar, M., Henriksson, G., & Lindstră
om, M. E. (2014). Pulp. In Ullmanns encyclopedia of
industrial chemistry (4th ed., pp. 1–92). Weinheim: Wiley-VCH.
Rosenau, T., Potthast, A., & Hofinger, A. (2019). Toward a better understanding of
cellulose swelling, dissolution, and regeneration on the molecular level. In Cellulose
science and technology. Chemistry, analysis and applications (pp. 99–125). Hoboken,
USA: John Wiley & Sons Inc.
Ruckel, E. R., & Schuerch, C. (1966a). Chemical synthesis of a stereoregular linear
polysaccharide. Journal of the American Chemical Society, 88, 2605–2606, 10/brkj3j.
Ruckel, E. R., & Schuerch, C. (1966b). Preparation of high polymers from 1,6-Anhydro2,3,4-tri-O-substituted β-D-glucopyranose. The Journal of Organic Chemistry, 31,
2233–2239. />Samain, E., Lancelon-Pin, C., F´erigo, F., et al. (1995). Phosphorolytic synthesis of
cellodextrins. Carbohydrate Research, 271, 217–226, 10/bc733f.
Saxena, I. M., & Brown, R. M. (2005). Cellulose biosynthesis: Current views and evolving
concepts. Annals of Botany, 96, 9–21, 10/b9rx3k.

Schlubach, H. H., & Lührs, E. (1941). Einwirkung von Chlorwasserstoff auf Glucose.
Synthese eines Polyglucosans. Justus Liebigs Annalen der Chemie, 547, 73–85, 10/
d423c7.
Schmidt, R. R., & Jung, K.-H. (2000). Trichloroacetimidates. In B. Ernst, G. W. Hart, &
P. Sina (Eds.), Carbohydrates in chemistry and biology (pp. 5–59). Weinheim,
Germany: Wiley-VCH Verlag GmbH.
Schmidt, R. R., & Michel, J. (1980). Facile synthesis of α- and β-O-glycosyl imidates;
preparation of glycosides and disaccharides. Angewandte Chemie (International Ed. in
English), 19, 731–732, 10/b88hfh.
Schmidt, R. R., & Michel, J. (1982). Synthesis of linear and branched cellotetraoses.
Angewandte Chemie (International Ed.), 21, 72–73, 10/c93tdr.
Serizawa, T., Tanaka, S., & Sawada, T. (2021). Control of parallel versus antiparallel
molecular arrangements in crystalline assemblies of alkyl β-cellulosides. Journal of
Colloid and Interface Science, 601, 505–516. />jcis.2021.05.117
Shapiro, D., Rabinsohn, Y., & Diver-Haber, A. (1969). A new micro-scale synthesis of
lactose. Biochemical and Biophysical Research Communications, 37, 28–30, 10/chmpxj.

Shoda, S., Uyama, H., Kadokawa, J., et al. (2016). Enzymes as green catalysts for
precision macromolecular synthesis. Chemical Reviews, 116, 2307–2413. https://doi.
org/10.1021/acs.chemrev.5b00472
Staudinger, H. (1920). Über Polymerisation. Berichte der Deutschen Chemischen
Gesellschaft A/B, 53, 1073–1085, 10/fq7gp3.
Sugiura, K., Sawada, T., Tanaka, H., & Serizawa, T. (2021). Enzyme-catalyzed
propagation of cello-oligosaccharide chains from bifunctional oligomeric primers for
the preparation of block co-oligomers and their crystalline assemblies. Polymer
Journal, 53, 1133–1143, 10/gkq3wt.
Takano, T., Nakatsubo, F., & Murakami, K. (1990). Synthetic studies of cellulose. VI:
Effect of the substituent groups of glycon on β-glycosylation. Cellulose Chemistry and
Technology, 333–341.
Takano, T., Nakatsubo, F., & Murakami, K. (1994). Synthetic studies of cellulose V: Effect

of the 3-O- and 6-O-substituent groups of the aglycon on β-glycosylation. Cellulose
Chemistry and Technology, 135–145.
Takeo, K., Okushio, K., Fukuyama, K., & Kuge, T. (1983). Synthesis of cellobiose,
cellotriose, cellotetraose, and lactose. Carbohydrate Research, 163–173. https://doi.
org/10.1016/0008-6215(83)84014-2
Takeo, K., Yasato, T., & Kuge, T. (1981). Synthesis of α- and β-cellotriose hendecaacetates
and of several 6,6′ ,6′′ -tri-substituted derivatives of methyl β-cellotrioside.
Carbohydrate Research, 148–156. />Uryu, T., Yamaguchi, C., Morikawa, K., et al. (1985). Ring-opening polymerization of
1,4-anhydro-2,3,6-tri-O-benzyl-α-D-glucopyranose and 1,4-anhydro-2,3,6-tri-Obenzyl-β-D-galactopyranose. Macromolecules, 18, 599–605, 10/c8jzgb.
Uryu, T., Yamanouchi, J., Hayashi, S., et al. (1983a). Selective ring-opening
polymerization of 1,4-Anhydro-2,3-di-O-benzyl-α-D-xylopyranose and synthesis of
stereoregular (1→5)-α-D-xylofuranan. Macromolecules, 16, 320–326.
Uryu, T., Yamanouchi, J., Kato, T., et al. (1983b). Selective ring-opening polymerization
of di-O-methylated and di-O-benzylated 1,4-anhydro-α-D-ribopyranoses and
structure proof of synthetic cellulose-type polysaccharide (1→4)-β-D-ribopyranan
and (1→5)-α-D-ribofuranan. Journal of the American Chemical Society, 105,
6865–6871, 10/dtmf37.
Williamson, R. E., Burn, J. E., & Hocart, C. H. (2002). Towards the mechanism of
cellulose synthesis. Trends in Plant Science, 7, 461–467, 10/fpvkwr.
Xiao, R., & Grinstaff, M. W. (2017). Chemical synthesis of polysaccharides and
polysaccharide mimetics. Progress in Polymer Science, 74, 78–116, 10/gch58t.
Yagura, T., Ikegami, W., Kamitakahara, H., & Takano, T. (2020). Synthesis of an
enantiomer of cellulose via cationic ring-opening polymerization. Cellulose, 27,
9755–9766, 10/gmxtb4.
Yataka, Y., Sawada, T., & Serizawa, T. (2015). Enzymatic synthesis and postfunctionalization of two-dimensional crystalline cellulose oligomers with surfacereactive groups. Chemical Communications, 51, 12525–12528, 10/gn4h4q.
Yataka, Y., Sawada, T., & Serizawa, T. (2016). Multidimensional self-assembled
structures of alkylated cellulose oligomers synthesized via in vitro enzymatic
reactions. Langmuir, 32, 10120–10125, 10/f86zfs.
Yoshida, T. (2001). Synthesis of polysaccharides having specifc biological activities.
Progress in Polymer Science, 26, 379–441. />00045-9

Zachoval, J., & Schuerch, C. (1969). Steric control in the polymerization of 1,6-anhydroβ-D-glucopyranose derivatives. Journal of the American Chemical Society, 91,
1165–1169. />Zhu, X., & Schmidt, R. R. (2009). New principles for glycoside-bond formation.
Angewandte Chemie (International Ed.), 48, 19001934, 10/c5fk7w.
Zweckmair, T., Oberlerchner, J. T., Bă
ohmdorfer, S., et al. (2016). Preparation and
analytical characterisation of pure fractions of cellooligosaccharides. Journal of
Chromatography. A, 1431, 47–54, 10/gmxtb5.

10