ENDOXYLANASES FROM GLYCOSYL HYDROLASE FAMILIES 5 AND 10:
BACTERIAL ENZYMES FOR DEVELOPMENT OF GRAM-POSITIVE BIOCATALYSTS
FOR PRODUCTION OF BIO-BASED PRODUCTS

By

FRANZ JOSEF ST. JOHN

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2006

Copyright 2006

by

Franz Josef St. John

I dedicate this work to the memory of Josefa Florentine St. John

1940-1990

Wife to one, mother of four, five who miss, evermore

ACKNOWLEDGMENTS

I acknowledge the steadfast support of my family and wife who have vicariously enjoyed

graduate school through my continued praise. Without their support this may have been

impossible. Further, I thank my advisor, James. F. Preston III, for taking the time to help me

grow as a scientist and teaching me that my imagination may be the best scientific tool of all.

I would also like to thank my Ph. D. research committee; Doctors A. Edison, L. O. Ingram,

J. Maupin-Furlow and K. T. Shanmugam, for their support and guidance. Their comments and

suggestions have helped guide this work to a fruitful conclusion. I thank them.

TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ......... ................. ...............................................................................4

LIST O F TA B LE S ............................................................................................. ............

LIST OF FIGURES .................................. .. .. .... ..... ................. .8

CHAPTER

1 INTRODUCTION ............... ................. ........... .............................. 12

2 BIOM ASS AND BIOCONVERSION ........................................................ ............. 16

The Complex Composition of Bioenergy Crops and Agricultural Crop Residues ................16
C ellu lo se ...............................................................................................17
Hemicellulose ................................................................ ..... ..... ........ 17
Lignin .................................. ........................... .......... 18
Polymer Interactions Which Create Recalcitrant Tissues ............................................18
Pretreatm ent of Lignocellulose............................................................................................20
Description of Biomass Pretreatm ent M ethods.................................. ............... ......21
Analysis of the Methods to Determine Which Pretreatment Protocols are Most
E effective ...................... ...... .. ... ... ............... ............ ......... 23
Enzyme Systems for Utilization of Glucuronoxylan .............................. .... .............25

3 Family 10 glycosyl hydrolaseS: Structure, Function and Phylogenetic relationaships..........35

Xylanases of Glycosyl Hydrolase Fam ily 10 ........................................ ...... ............... 35
Enzym e Structure and M echanism ...................................................................... ...... 36
Modular Characteristics of GH 10 Xylanases............................................................36
CBM classification by target substrate................................................... ................ 38
CBM modules common in bacterial GH 10 xylanases and their general
architectural arrangem ent .............................................................................. 39
F ungal m odules ......................................................... ................. 46
Other modules and sequences from GH 10 xylanase............... ...............47
Glycosyl Hydrolase Accessory Module Discussion....................... ..................49
Hydrolysis of Substituted Xylans by GH 10 Xylanases .....................................................51
H ydrolysis of M ethylglucuronoxylan......................................... ......................... 52
Hydrolysis of Methylglucuronoarabinoxylan........... ....... ....................... 53
Hydrolysis of Rhodymenan by GH 10 xylanases................................. ................. 55
GH 10 Xylanase Substrate Binding Cleft Studies.............................................. 55
Phylogenetic Relationships of Glycosyl Hydrolase Family 10 Xylanases...........................57
Plant and Related Bacterial GH 10 Xylanase .............. ........................................... 58
Fungal and Streptomyces Association ..................................... .......................... ....... 59
Bacterial GH 10 Xylanases: Tools to Work With..................................................59
C o n c lu sio n ................... .......................................................... ................ 6 0

4 Paenibacillus SPECIES STRAIN JDR-2 AND XynAi: A NOVEL SYSTEM FOR
GLUCURONOXYLAN UTILIZATION ............................................................................. 76

In tro du ctio n ................... ...................7...................6..........
M materials and M methods ...................................... .. ......... ....... ...... 79
R e su lts ................... ...................8...................9..........
D iscu ssio n ................... ...................9...................4..........

5 CHARACTERIZATION OF XynC FROM Bacillus subtilis SUBSPECIES subtilis
STRAIN 168 AND ANALYSIS OF ITS ROLE IN DEPOLYMERIZATION OF
G L U C U R O N O X Y L A N ............................................................................ ..................... 110

In tro du ctio n ................... ...................1.............................0
M materials and M methods ..................................... .......... .. .... ...... .. ........ .... 115
R e su lts ................... ...................1.............................2
D iscu ssio n ................... ...................1.............................6

6 Summary Discussion ................................................................ 142

C current R research D irections....................................................................... ......... ........... 142
G ram -positive B iocataly sts......................................................................... ....................143

L IST O F R E F E R E N C E S .................................................................................... ...................147

B IO G R A PH IC A L SK E T C H ......................................................................... .. ...................... 166

6

LIST OF TABLES

Table page

2-1 Composition of potential biofuel crops and other biomass sources ..............................30

3-1 Distribution by bacterial genus of carbohydrate binding modules and other functional
domains associated with GH 10 xylanases......................................................................63

3-2 Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some
of their properties ............................................................... ... .... ......... 65

4-1 Source and characteristics of sequences used for phylogenetic comparison................. 104

5-1 Relationship of XynC activity to the degree of MeGA substitution on MeGAXn ..........134

5-2 M ALDI-TOF peak assignments ........................................... ................... 136

5-3 Relative transcript quantity measured by Q-RT-PCR for gapA, abnA, xynA and
xy n C g en e s ...................... .. .............. .. ............................... ................ 14 0

LIST OF FIGURES

Figure pe

2-2 Common structural elements and sites of enzymatic hydrolysis which degrade
methylglucuronoxylan and methylglucuronoarabinoxylan. ............................................32

2-3 Generation of hydrolysis products by different families of xylanases. ..........................33

2-4 X ylanase structure and function............................................... .............................. 34

3-1 Common domain arrangements found in GH 10 xylanases. ...........................................62

3-2 Products formed by the hydrolysis of methylglucuronoxylan and
methyglucuronoarabinoxylan by a glycosyl hydrolase family 10 xylanase....................64

3-3 Phylogenetic distribution of catalytic domains of glycosyl hydrolase family 10
xylanases .................................................................................75

4-1 Growth of Paenibacillus sp. strain JDR-2. .......................................... ............... 101

4-2 Genetic map ofxynA1 and surrounding sequence resulting from sequencing of the
Paenibacillus sp. strain JDR-2 genomic DNA insert of pFSJ4................................ 102

4-3 Domain alignment of GH 10 subset B and subset A sequences.................................. 103

4-4 Phylogenetic analysis of a randomly selected set of GH 10 xylanases with respect to
the XynA i CD GH 10B subset........... ................. .......... ............... ............... 105

4-5 Localization of modular XynAi in subcellular fractions. ...........................................106

4-6 Lineweaver Burk kinetic analysis of XynAi CD................................ ..................107

4-7 Kinetic analysis of product formation catalyzed by XynAi CD hydrolysis of SG
M eG A X n. ................................................................................108

4-8 Differential carbohydrate utilization by Paenibacillus sp. strain JDR-2.........................109

5-1 O ptim ization of X ynC activity................................................................ .................... 133

5-2 MALDI-TOF MS analysis of the Filtrate (A) and Retentate (B) resulting from 3 kDa
ultrafiltration of a SG M eGAXn XynC digest.............. .............................................. 135

5-3 1H-NMR of SG MeGAXn 3 kDa filtrate revealing the general action of XynC
hydrolysis of MeGAXn and the predicted limit product of XynC MeGAXn digestion...137

5-4 Identification of products generated by XynA (GH 11) and XynC (GH 5) secreted by
B su b tilis 16 8 ....................................................................... 13 8

5-5 Regulation of expression of xynA and xynC genes in early- to mid-exponential phase
growth cultures of B. subtilis 168 with different sugars as substrate.............................. 139

5-6 Limit aldouronates expected from a SG MeGAXn digestion with a GH 11 xylanase
and a GH 5 xylanase co-secreted in the growth medium of B. subtilis 168 ..................141

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

ENDOXYLANASES FROM GLYCOSYL HYDROLASE FAMILIES 5 AND 10:
BACTERIAL ENZYMES FOR DEVELOPMENT OF GRAM-POSITIVE BIOCATALYSTS
FOR PRODUCTION OF BIO-BASED PRODUCTS

By

Franz Josef St. John

December 2006

Chair: James F. Preston III
Major Department: Microbiology and Cell Science

Fossil fuels are a nonrenewable resource. Their wide geological distribution and relatively

simple acquisition have allowed massive increases in human population and associated energy

expenditure over the last century. The current rate of consumption and the expectation of

reduced fuel supplies predicate the need to develop new energy sources that may be merged with

the current energy infrastructure. As an underutilized renewable resource, lignocellulosic

biomass may, through microbial bioconversion, contribute to the environmentally benign

production of alternative fuels and chemical feedstocks. A major target for this conversion is

methylglucuronoxylan, the predominant structural polysaccharide in the hemicellulose fraction

of hardwood and crop residues. The research described herein has focused on xylanolytic

bacteria and their secreted endoxylanases that are involved in the depolymerization of the

methylglucuronoxylan. In this work, endoxylanases of glycosyl hydrolase families 5 and 10

(GH 5 and GH 10 xylanases) have been characterized with emphasis on the native bacterial host

and utilization of the hydrolysis limit products of methylglucuronoxylan. Recombinant

constructs of the genes encoding these xylanases have made them available for definitive

characterization and for expression in transgenic organisms. XynA1, a multimodular GH 10

xylanase anchored to the cell surface ofPaenibacillus sp. strain JDR-2, generates aldouronates

that are efficiently assimilated and metabolized by this organism. The proximity between XynA1

and the native Paenibacillus host, and the proficient utilization of the resulting hydrolysis

products, identify a process of vectoral assimilation of methylglucuronoxylan-derived products.

A GH 5 xylanase encoded by the ynfF gene in Bacillus subtilis 168 is directed for catalysis by

methylglucuronosyl substitutions on the xylan chain, supporting its application in an accessory

role in the overall depolymerization process. Secretion of this GH 5 and a GH 11 endoxylanase

by the genetically malleable B. subtilis 168, for which the entire genome has been sequenced,

recommends it as a target for introduction of genes encoding the GH 10 endoxylanase, XynAi

and aldouronate-utilization enzymes for efficient depolymerization and metabolism of

methylglucuronoxylan. These discoveries provide insight needed for the development of second-

generation bacterial biocatalysts for the direct conversion of lignocellulosic biomass to

alternative fuels and bio-based products.

CHAPTER 1
INTRODUCTION

The political and social stability of the world is dependent on a plentiful supply of energy.

For the last century, this supply has been in the form of fossil fuels mined or pumped from the

earth. At the current rate of use, it is predicted that approximately half-way through this century,

the supply of easily obtainable fossil fuels will be significantly limited. Three recent studies of

proven world oil supplies show an average estimate of currently obtainable crude oil reserves of

1188 billion barrels (Energy Information Administration, 2006a). In 2003 world oil consumption

was estimated at about 29 billion barrels per year (Energy Information Administration, 2006b).

These values lead to an estimated time of 41 years before the current proven reserves are

exhausted. Although there are newly realized additions to the world crude oil reserve identified

frequently, consumption is increasing steadily and outpacing discovery (Erickson, 2003).

Regardless of the number of years until a severe shortage in crude oil reserves, the inevitable

consequence of dependence on this finite resource is troubling. Increasing population and

industrialization around the world exacerbate this situation. Furthermore, as crude oil supplies

decrease, that which remains becomes more difficult to obtain. These circumstances result in

high demand for a diminishing commodity which can stifle worldwide growth and challenge

international relations. Fossil fuels are also a primary contributor to "greenhouse" gases (GHG)

which could, over the next few decades, have a major effect on global climate. There is no need

to wait until this inevitable energy crisis suffocates the world economy. It is not a requirement

that we utilize every obtainable drop of crude oil. In fact, it would be extremely irresponsible to

do so. Research efforts must focus on increasing energy yield of current alternative energy

sources, on developing novel methods for harvesting energy, and decreasing per capital energy

consumption. Together, these goals may add up to a sustainable energy future for the world.

Further, investment in this new energy infrastructure may spur equivalent growth as that

witnessed during the 20th century in the United States. After all, survival is a strong motive for

reform.

Over the last several decades, methods have been developed by which the constituent

sugars of lignocellulosics can be converted to clean burning ethanol using microorganisms as

biocatalysts (Dien et al., 2003; Ingram et al., 1999; Jeffries, 2005). Large scale application of

this technology to renewable resources, such as woody biomass and crop residues, may help to

balance the demand for fossil fuels by supplementing the supply with ethanol. In the long term,

this could become a major contributor to liquid fuel supplies for transportation needs. Use of

ethanol produced from biomass is carbon neutral, meaning that it does not increase the net

atmospheric CO2 concentration. Therefore, it does not contribute to GHG emissions. The

process by which this conversion takes place can be broken down into two key steps. The first is

a preprocessing step, which is typically required to prepare the complex, recalcitrant

lignocellulosic biomass for conversion. The second utilizes engineered microbial biocatalysts to

convert the simple sugars released by the pretreatment to ethanol. In some systems, conversion

of the sugars in lignocellulose to ethanol has been accomplished to near theoretical yields

(Ingram et al., 1998).

Two food crops which are used for large scale ethanol production are corn grain and

sugarcane. In the US, corn grain is the primary source for ethanol production representing

approximately 2% of liquid fuel use whereas in Brazil, sugarcane has been used to produce

enough ethanol to significantly displace gasoline. Another biofuel, biodiesel, used as a diesel

fuel replacement, can be prepared from triglyceride rich crops such as soybean, rapeseed and

palm. A primary concern frequently raised when considering ethanol or biodiesel production

from food crops is the net energy balance. Recent studies have shown that production of ethanol

from corn grain and biodiesel from soybean are energy positive (Dewulf et al., 2005; Farrell et

al., 2006). When considering these food crops as substrate biomass, the balance is obtained by

considering all inputs, including all costs of farming and biomass preprocessing, and the output

costs including the net yield of fuel and all other valuable components. In a recent study,

bioconversion of corn grain to ethanol yielded only a 25% energy increase over the energy

consumed in the process and biodiesel from soybean yielded a 93% energy increase (Hill et al.,

2006). The lower energy yield for corn conversion is primarily attributed to the greater growth

requirements of corn and the use of gas-fired boilers for ethanol purification. Although soybean

may be a good candidate crop for biodiesel production, analysis of total crop land requirements

for use in growing any of theses crops as major sources for fuel is thought to be very ambitious,

and would still contribute only a small amount of total liquid fuel requirements (Hill et al.,

2006). Large scale food crop cultivation for bioenergy also concerns scientists as it links energy

production directly to food production and also increases demand for valuable agricultural land.

It is generally acknowledged that the greatest benefit for society will come from production of

cellulosic ethanol that results from conversion of lignocellulosics and crop residues (Dewulf et

al., 2005; Farrell et al., 2006; Hammerschlag, 2006; Hill et al., 2006; Perlack et al., 2005).

Idealized substrates for ethanol production will be lignocellulosics such as crop residues like

corn stover, forestry products including pulp and paper mill waste, and renewable energy crops

such as switchgrass and poplar. As considered above, bioenergy crops are expected to compete

for generally useful agricultural lands. However, they will not directly compete with agricultural

crop products. Use of these underutilized renewable biomass resources as substrates for biofuel

production greatly reduces the input requirements therefore increasing energy yield. The United

States Department of Agriculture recently published a report entitled "Biomass as Feedstock for

a Bioenergy and Bioproducts Industry: The Technical Feasibility of a Billion-Ton Annual

Supply" (Perlack et al., 2005). The analysis presented in this report proposes that it is possible to

replace 30% of the US petroleum consumption with biofuels by 2030. Development of

bioenergy crops like poplar (hardwoods) and switchgrass graminaceouss plants) along with food

crop residues, such as corn stover and sugar cane bagasse, will contribute greatly to ethanol

production and accomplishment of the "Billion-Ton Annual Supply" and energy independence.

CHAPTER 2
BIOMASS AND BIOCONVERSION

The Complex Composition of Bioenergy Crops and Agricultural Crop Residues

Energy harvested from the sun through the process of photosynthesis is stored by plants as

fixed carbon in the form of several different, tightly associated complex polymers. The

combined characteristics of these interacting polymers impart to the plant the strength and

resilience that is common in wood products. For this reason, wood has been ubiquitous in the

urbanization of society, allowing for relatively affordable, easily constructed homes and

buildings. When considering bioconversion processes, these characteristics are referred to as

recalcitrance, and are a primary concern for development of processes which liberate the

complex sugar components as simple utilizable sugars. The primary components of wood which

interact to create this strength and recalcitrance are cellulose, hemicellulose and lignin.

Cellulose is the most abundant carbohydrate polymer in the biosphere and is the major structural

component of plant life. In bioenergy crops and crop residues, cellulose can range from 36% to

50% total mass (Table 2-1) (Lynd et al., 1999; Scurlock). The second most abundant polymer in

biomass is the combined hemicellulose fraction, which in hardwoods is composed primarily of

methylglucuronoxylan, and can range from 20% to 35% of total mass (Timell, 1967). This

fraction in graminaceous plants such as switchgrass and the crop residue corn stover represents

up to 35% of mass and is primarily composed of the polymer methylglucuronoarabinoxylan

(Lynd et al., 1999). A secondary hemicellulose from hardwoods, glucomannan, is a minor sugar

component representing no more that 4% of mass, and will not be directly considered in this

discussion. Lignin is a complex non carbohydrate polymer which is directly associated with

recalcitrance of biomass. Renewable biomass sources receiving the most intense scrutiny

contain lignin from between 5.5% for early cut harvest of switchgrass growth up to 24% for

common hardwood sources such as poplar (Lynd et al., 1999; Timell, 1967). Together, these

interacting polymers impart the characteristics for which wood has been cherished, but they

further act to prevent enzymatic accessibility to the individual carbohydrate polymers for direct

use of woody biomass as a renewable resource.

Cellulose

Cellulose is a homopolymer composed of repeating 0-1,4 linked glucose molecules. When

derived from woody biomass it has a degree of polymerization of approximately 10,000 glucose

residues, making it the largest naturally occurring polymer (O'sullivan, 1997). Although the

word cellulose refers to the 0-1,4-linked polymer described above, it also describes the tightly

associated crystalline fibers that result from many individual cellulose strands hydrogen bonding

together to form the chemically and physically recalcitrant cellulose fiber. In general, cellulose

from biomass is composed of many identical molecules which are tightly associated through

hydrogen bonding interactions, intermixed with hemicellulose. This tight association between

cellulose molecules and fibrils makes this specific polymer recalcitrant to chemical and

enzymatic degradation.

Hemicellulose

Hardwood hemicellulose differs from graminaceous sources of hemicellulose primarily by

not having arabinose substitutions. As the name suggests methylglucuronoxylan is a 0-1,4

linked xylose chain randomly substituted (Jacobs et al., 2001) with a-1,2 linked 4-0-

methylglucuronate moieties. Common hardwood sources have methylglucuronate substitutions

on one in every ten xylose residues, but there have been reports of hardwoods with significantly

higher methylglucuronate content (Hurlbert and Preston, 2001; Preston et al., 2003; Puls, 1997).

Detailed 13C-NMR studies of methylglucuronoxylan from the hardwood sweetgum, a candidate

bioenergy hardwood, reflect degrees of substitution as high as 1 in every 6 xylose residues.

Methylglucuronoarabinoxylan from graminaceous biomass resources have fewer

methylglucuronate substitutions, but can contain arabinose at a ratio of one for every six xylose

residues (Puls, 1997; Wilkie, 1979). An additional substitution on both the

methylglucuronoxylan and methylglucuronoarabinoxylan polymers is O-linked acetyl groups

(Puls, 1997; Timell, 1967). Methylglucuronoxylan has this substitution in either the 02 or 03

position on 70% of all the xylose residues and methylglucuronoarabinoxylan contains

approximately half this amount (Puls, 1997; Timell, 1967).

Lignin

Of all the polymers in lignocellulose, lignin is the most complex due to its random

amorphous nature. There is little if any in the primary cell wall, but it forms a major component

in the secondary wall and the middle lamella. It is composed of characteristic phenyl propane

derivatives, randomly linked through carbon-carbon bonds by an enzymatic dehydrogenation

process, shedding light on the reason for the complexity. The phenyl propane derivatives differ

depending on the lignin source, but in general lignin from most plants is composed of guaiacyl,

syringyl and p-hydroxyphenol moieties (Fujita and Harada, 2001).

Polymer Interactions Which Create Recalcitrant Tissues

The secondary cell wall in woody tissue is the primary structural component of biomass,

which imparts rigidity and strength to the plant and is the primary source of stored carbon

composed of cellulose, hemicellulose and lignin. Based on data reviewed by Mellerowicz et al.

(Mellerowicz et al., 2001), 86% to 88% (v/v) of cells in poplar wood have secondary cell walls

consisting of 80% utilizable carbohydrate. This quantifies the value of hardwood as a biomass

resource. This wood type is characterized by its high content of methylglucuronoxylan and

lower content of lignin when compared to softwoods. These properties make this wood type

more amenable to pretreatment processes. Formation of a full secondary cell wall begins with

synthesis of a primary wall. Cellulose microfibrils, which are composed of many individual

cellulose molecules, are synthesized on the cell surface by large complexes of cellulose synthase

enzymes called rosettes (Ding and Himmel, 2006; Doblin et al., 2002). The layers of cellulose

microfibrils that are produced in the primary wall form a net-like cellulose matrix (Fujita and

Harada, 2001). Together with this matrix are the primary wall hemicellulose polymers and

associated polysaccharides and proteins which lend support to and flexibility for cell expansion

(Carpita et al., 2001; Somerville et al., 2004). After the primary wall is synthesized, if the cell is

a xylem fiber type, it begins to synthesize a secondary cell wall. The secondary cell wall is much

thicker than the primary wall and is typically divided into at least three distinct layers. As can be

observed in Figure 2-1, cellulose fiber synthesis in each layer is offset from the next by some

degree so that after synthesis of the complete secondary cell wall, there are cellulose fibers

bracing the wall from most angles (Fujita and Harada, 2001). It is thought that hemicellulose

interacts with and coats the outer layer of cellulose microfibrils to allow for movement, in effect

acting as a lubricant and preventing formation of larger cellulose fibers through hydrogen

bonding (Ding and Himmel, 2006; Fengel, 1971; Reis et al., 1994). Layers of the secondary cell

wall are described as showing a helicoid structure. It has been postulated that it is the intimate

interaction with hemicellulose polymers, which are at a greater concentration between secondary

cell wall layers that control the formation of this helicoid structure (Reis et al., 1994). Although

the mass of cellulose and associated hemicellulose polymers impart strength to the cell wall and

the plant, the wall is further reinforced by cross linking with lignin. This randomly formed

complex, non-carbohydrate polymer forms ester linkages to various moieties on the

hemicellulose chain (Puls, 1997; Reis et al., 1994). The final secondary cell wall structure of

mature xylem cells contains three heavy layers of cellulose (Table 2-1) intertwined with

hemicellulose, which is lignified through covalent bonds to lignin. The middle lamella is fully

lignified, filling in between individual cells (Fromm et al., 2003). The complex matrix formed

by these three associated materials can be compared to reinforced concrete and the outermost

layer of lignin filling the middle lamella acts to glue it all together.

Softwood conifers are composed of the same three major components discussed above.

However, the hemicellulose fraction is not composed of methylglucuronoxylan as in hardwood,

but rather consists of two different polymers. The minor polymer is similar to

methylglucuronoarabinoxylan of the graminaceous plants, but has a greater amount of

methyglucuronate substitution and no acetyl groups. The predominant hemicellulose polymer in

softwood is galactoglucomannan. This polymer is an O-acetylated polymer composed of

terminally branched galactose, glucose and mannose in the ratio 1 : 1 : 3. Generally, softwood is

more heavily lignified than hardwoods and is considered to be less efficient in bioconversion

processes. Further, the pulp and paper and construction industries are major consumers of

softwood supplies. It is possible that future bioconversion endeavors may combine all wood

sources for bioconversion, but current objectives primarily include hardwoods and switchgrass as

biofuel crops and waste agricultural residues, e.g., corn stover and sugar cane bagasse.

Pretreatment of Lignocellulose

Due to the complex nature of lignocellulosics, utilization of the embedded carbohydrates

requires preprocessing, which usually includes a physical and chemical pretreatment. Following

this, all pretreatment methods in development require supplementation with fungal cellulolytic

hydrolase mixtures (e.g., Spezyme CP). The most established pretreatment methods have

recently been reviewed. In this particularly constructive effort, several laboratories worked

together to obtain equivalent, directly comparable data (Kim and Holtzapple, 2005; Kim and

Lee, 2005; Liu and Wyman, 2005; Lloyd and Wyman, 2005; Mosier et al., 2005a; Teymouri et

al., 2005; Wyman et al., 2005b). The seven preprocessing methods studied used NREL standard

methods for data collection and result analysis. Results were reported as total sugars released

after the pretreatment step and also after the subsequent enzyme hydrolysis. Besides the review

of the individual preprocessing methods and general results of each (Wyman et al., 2005a),

further articles published simultaneously (same volume) combined complete data sets of all

seven pretreatments for comparative analysis (Eggeman and Elander, 2005). They also provided

an in-depth economic analysis of the most promising approaches (Mosier et al., 2005b) and

considered the effect of preprocessing on the biomass structure (Lloyd and Wyman, 2005). Here

I briefly consider these processing methods to better understand the requirements for subsequent

bioconversion to ethanol.

Description of Biomass Pretreatment Methods

The pretreatment methods compared include dilute sulfuric acid, flow-through, pH

controlled water, ammonia fiber explosion (AFEX), ammonia recycle percolation (ARP) and

lime pretreatment. Each pretreatment method is reviewed to better understand how each affects

biomass. The first applies mild acid conditions (0.5-3.0% H2S04) and temperatures and

pressures from 1300C to 2000C and 3atm tol5atm, respectively. In this method, the acid,

temperature and pressure function to liberate the hemicellulose in an almost completely utilizable

form. In addition, dilute acid treatment disrupts the normally recalcitrant cellulose for effective

subsequent enzymatic hydrolysis (Liu and Wyman, 2005). The flow-through pretreatment

method was simultaneously compared to an improved method termed partial flow-through

pretreatment (PFP), both of which are similar in principle to the dilute acid pretreatment. Both

of these methods apply temperatures of 190-2000C and pressure of 20-24 atm. Partial

flowthrough pretreatment was superior to water flow-through pretreatment because the final

solubilized xylan was more concentrated by the reduced water elution volumes. This

significantly affects downstream costs associated with product recovery for fermentation.

Furthermore, this method can remove significant lignin content prior to enzyme hydrolysis

(Mosier et al., 2005a). The pH controlled water pretreatment method uses high temperature

(160-190C) and high pressure (6-14 atm) for times between 10 and 30 minutes. The application

of high temperature and neutral pH water has several advantages over the acid hydrolysis

methods. Hemicellulose is solubilized and primarily remains as small oligomers, reducing the

conditional formation of side products such as furfural, which can inhibit fermentations.

Furthermore, this method significantly reduces the lignin content to increase enzymatic

hydrolysis (Teymouri et al., 2005). Ammonia fiber explosion (AFEX) is a promising

pretreatment method which applies ammonia at elevated temperatures (70-90C) and pressures

(15-20 atm) for only short periods of time (<5 minute). The inherent solvent properties and

volatility of ammonia are the characteristics which allow this unique approach to disrupt the

biomass. The explosion occurs with the sudden release of pressure resulting in rapid expansion

of the volatile heated ammonia. The method does not remove any component of the biomass,

but rather disrupts it sufficiently to allow near complete enzyme hydrolysis (Kim et al., 2006).

ARP uses a 10-15% (w/w) ammonia solution at 150-170C and 10-20 atm of pressure for 10-20

minutes. This method is somewhat similar to the PFP described above, but uses an aqueous

ammonia solvent rather than acidic water. As with PFP, this method has the advantage of

obtaining significant lignin removal, but requires downstream separation of solubilized

carbohydrates for maximum realized sugar yield (Kim and Holtzapple, 2005; Kim and

Holtzapple, 2006a). Use of lime (CaOH2) in pretreatment applies opposite chemistry as

compared to acid based pretreatments. Combination of lime and oxygen with lignocellulose

degrades hemicellulose and cellulose by the peeling reaction endwisee reducing end 3-

elimination), releasing glucoisosaccharinate and xylosaccharinate. In general, this pretreatment

method produces a complex mixture of degradation products, and allows removal of a large

percentage of lignin (Wyman et al., 2005a). This treatment, as with all the discussed

pretreatment methods, significantly improves enzymatic hydrolysis of the resulting pretreated

biomass.

Analysis of the Methods to Determine Which Pretreatment Protocols are Most Effective

From the pretreatment studies of corn stover described above, comparisons were

performed by analysis of total sugar yields (Eggeman and Elander, 2005). With a standard

enzyme amount of 15 filter paper units Spezyme CP per gram glucan applied post pretreatment,

all methods yielded no greater than 86% of the total carbohydrate, indicating that they are all

candidates for further refinement and optimization. The lowest yield (86.8%) was obtained with

an optimized lime treatment and three of the seven methods resulted in yields over 90%. These

included dilute acid treatment yielding 92.4%, AFEX yielding 94.4% and flow-through yielding

96.6%. These three pretreatment methods differ greatly in their resulting carbohydrate product

mixtures. The dilute acid treatment converted approximately 83% of the total xylose content to

free xylose and slightly more than 2% to xylooligomers. This achieved near complete removal

of xylan from the cellulose and lignin matrix and provides a likely explanation for the almost

complete hydrolysis of cellulose (-92%) observed with the subsequent cellulase treatment.

AFEX is by far the most interesting case and resulted in the second best sugar yield (94.4%).

The pretreatment did not release any sugars, but resulted in a much altered biomass structure,

which facilitated near complete enzymatic hydrolysis. The best results were obtained with the

flow-through pretreatment method. With this approach, almost complete xylose conversion

occurred, but it was primarily in the form of xylooligomers (-92%). Only a small amount of free

xylose was detected (-4.5%).

Economic analysis of these preprocessing methods showed that no single method had a

clear advantage (Eggeman and Elander, 2005). The dilute acid, AFEX, ARP and lime

pretreatments were each estimated to require a similar fixed capital investment. As an example

of this, in the dilute acid method, the primary cost was associated with equipment requirements

needed to handle the corrosive conditions, and a minor cost was associated with chemical supply

requirements. The opposite case was observed for AFEX pretreatment, which requires costly

pure ammonia, which is less corrosive then acid. Use of pure ammonia is a substantial cost

although AFEX plants are designed to recover most by condensation. Dilute acid, AFEX and

lime pretreatments resulted in the lowest total fixed capital per gallon of annual ethanol

production capacity making these the best current methods for large scale pretreatment. Both

dilute acid and AFEX were at $3.72/gallon and lime was significantly lower at $3.35/gallon.

Except for AFEX, it is clear that all these pretreatment methods act primarily by removing

xylose or lignin, and in some cases, a significant amount of both. Since the hemicellulose and

lignin fractions are thoroughly embedded within the cellulose matrix, it seems likely that

methods which remove either will significantly alter cellulose accessibility and render it

susceptible to enzymatic hydrolysis. In the case of AFEX, no degradation of the treated biomass

is apparent, but based on the subsequent ability for almost complete enzyme hydrolysis it seems

likely that this process readily alters the lignin and xylan association with cellulose (Kim and

Holtzapple, 2006b; Teymouri et al., 2005). Although it is thought that residual lignin in biomass

inhibits enzymes added for hydrolysis, (Berlin et al., 2006) these studies show no indication of

this. Dilute acid preprocessing does not remove lignin, but enzyme hydrolysis resulted in the

third highest yield of carbohydrate. AFEX, as mentioned, retained the majority of its lignin

content and resulted in the second highest yield of carbohydrate. This suggests that 15 FPU

Spezyme CP/g glucose may be a wasteful amount of hydrolytic activity for complete hydrolysis

following some pretreatments.

Enzyme Systems for Utilization of Glucuronoxylan

Although further research into pretreatment methods is required, it seems likely that the

methods reviewed above approach their optimal performance. Limitations with the current

methods for ethanol production from biomass include the high cost of pretreatment and the cost

of commercial enzyme preparations required to obtain maximum yield. Advancements which

lower the cost of bioconversion of lignocellulosics to ethanol are likely to come from the

development of less expensive, more robust enzyme systems with a greater range of enzymatic

activities and development of robust microbial biocatalysts. The latter research direction

includes biocatalyst advances such as: decreasing growth requirements and increasing the

substrate range, development of hydrolytic enzyme secretion systems to reduce commercial

enzyme use, and in general, optimizing a specific biocatalyst for use with a specific

preprocessing and bioconversion method. The ultimate biocatalyst would secrete most, if not all

of the required hydrolytic activity and efficiently transport and ferment hydrolysis limit products

to fuel ethanol. Research in this direction will facilitate advances for low-cost, high yield

bioconversion processes.

Dilute acid pretreatment is currently being developed as a leading pretreatment method.

Other than the energy and chemical requirements discussed above, limitations specific to this

process include formation of acid hydrolysis side products such as furfural and a-1,2-

glucuronoxylose. Furfural forms from the acid and heat catalyzed dehydration of xylose.

Formation of this side product reduces process efficiency in two ways. First, it reduces the net

convertible xylose concentration and secondly, furfural is known to inhibit microbial growth and

fermentation (Zaldivar et al., 1999). The aldouronate, a-1,2-glucuronoxylose, results from acid

hydrolysis of methylglucuronoxylan and methyglucuronoarabinoxylan due to the stability of the

a-1,2 glycosyl linkage, which is thought to form an internal lactone between the carboxylate

moiety on the glucuronic acid and a hydroxyl on the substituted xylose while under acidic

conditions (M. E. Rodriquez, A. Martinez, S. W. York, K. Zuobi-Hasona, L. O. Ingram, K. T.

Shanmugam, J. F. Preston, Abstr. 101st ASM General Meeting, abstr.O-21, 2001) (Jones et al.,

1961). Whereas the arabinose and acetyl linkages are considered acid labile, the stability of the

a-1,2 glucuronosyl linkage allows for the buildup of singly substituted aldouronates, which are

unable to be utilized by any current biocatalyst. Considering that the frequency of substitution of

methylglucuronate is at least 1 for every 10 xylose residues, this suggests at best, a

bioconversion process can only recover 90% of the total xylose fraction. This does not take into

account the significant potential contribution made by free glucuronic acid to the net ethanol

yield.

In the short term, feasible goals can be met by developing enzyme systems which function

efficiently to allow reduced pretreatment of biomass, in effect, lowering the total cost of

preprocessing. Limited pretreatment with dilute acid would allow for reduced energy and/or acid

consumption in the pretreatment process and would also lower the formation of furfural and

other fermentation inhibiting compounds. The resulting pretreated biomass would still have a

significant content of polymerized xylose and may require enzymatic treatment to fully release

fermentable carbohydrate. For this reason, enzymes which degrade hemicellulose are primary

research targets to facilitate utilization of methylglucuronoxylan and methyglucuronarabinoxylan

by biocatalysts.

As detailed above, these two polymers make up the second most abundant carbohydrate in

bioenergy crops and agricultural crop residues and unlike the chemically simple, but physically

recalcitrant cellulose polymer, methylglucuronoxylan and methylglucuronoarabinoxylan are

chemically complex and require a battery of enzymes with a wide range of activities to fully

degrade them to simple sugars. As shown in Figure 2-2 these activities include several different

xylanases, an a-glucuronidase, acetyl esterases, arabinofuranosidases, and lignin esterases (not

shown). Xylanases have a primary role in the degradation of xylan as they reduce the large

linear polymer to small xylooligomers and small substituted xylooligomers. Accessory enzymes

such as the a-1,2-glucuronidase are known to have activity on small substituted hydrolysis

products resulting from xylanase digestion, but not on the intact polymer (Nagy et al., 2002;

Nurizzo et al., 2002). This research will address how xylanases of glycosyl hydrolase (GH)

family 5 and 10 function to hydrolyze polymeric methylglucuronoxylan.

Throughout this dissertation, methylglucuronoxylan is considered an idealized substrate

consisting of a 0-1,4-xylan substituted with a-1,2-linked 4-O-methylglucuronate moieties. Based

on the acid labile character of the less significant substitutions on the methylglucuronxylan and

methylglucuronoarabinoxylan polymers, pretreatment utilizing limited dilute acid conditions

may result in methylglucuronoxylan being the primary retained polymer. Processing of

methylglucuronoxylan by the three major families of xylanase enzymes is depicted in Figure 2-3.

Xylanases of glycosyl hydrolases family 5 (GH 5) are the newest xylanases to be characterized.

This work (Chapter 5) details the current understanding of this novel xylanase family. Although

all indications are that it is specific for the hydrolysis of methylglucuronoxylan, resulting

products are thought to be too large for direct utilization by biocatalysts. The abilities of this

enzyme may well complement the activities of the other two primary xylanase families.

Xylanases from families GH 10 and GH 11 are relatively well characterized. Both of these

xylanase families are known to produce primarily xylobiose and xylotriose as primary neutral

limit products of methylglucuronoxylan. However, while GH 10 xylanases yield the smallest

substituted aldouronate, aldotetrauronate (MeGAX3) (Fig 2-3) which is substituted directly on

the nonreducing terminal xylose with a-1,2 glucuronate, GH 11 xylanases yield aldopentauronate

(MeGAX4) which is substituted penultimate to the nonreducing terminal xylose (Biely et al.,

1997). This slight difference has significance in that substrates for the xylanolytic a-1,2-

glucuronidase accessory enzyme can only hydrolyze methylglucuronate from xylooligomers

when it is substituted directly on the nonreducing terminal xylose (Nagy et al., 2002; Nagy et al.,

2003; Nurizzo et al., 2002). Further, most bacterial a-1,2-glucuronidase enzymes are

intracellular and supporting research has indicated that MeGAX3 is the largest aldouronate which

is readily transported for catabolism (G. Nong, V. Chow, J. D. Rice, F. St. John, J. F. Preston,

Abstr. 105th ASM General Meeting, abstr.O-055, 2005) (Shulami et al., 1999; St. John et al.,

2006). Figure 2-3 depicts the processing of methylglucuronoxylan as an idealized substrate for

utilization by bacterial biocatalysts.

Although GH 10 and GH 11 xylanases share identical hydrolytic mechanisms (as with GH

5) these two families differ in primary protein fold. Catalysis of the 0-1,4 xylan chain proceeds

by a double displacement mechanism with retention of anomeric configuration. Figure 2-4

identifies the structural differences and presents the common mechanism by which these

xylanases function. The different limit aldouronates resulting from GH 10 and GH 11 xylanases

result from steric interactions between the substituted xylan polymer and the binding cleft of

these two xylanases with different protein folds.

For consistency, throughout the following chapters and just below, 4-0-

methylglucuronoxylan will be abbreviated MeGAXn and corresponding substituted

xylooligomers as MeGAXx, where x equals the number of xylose residues. In sections

considering arabinose substitutions the name methylglucuronoarabinoxylan will be used and

arabinose substituted xylooligomers will be denoted as AXx, where x equals the number of

xylose residues.

The following chapters contain the analysis of xylanases from glycosyl hydrolase families

5 and 10 and explore their mode of action and their hydrolysis products on the substrate

MeGAXn. By developing a strong understanding of how these enzymes act to hydrolyze

MeGAXn, how they function to benefit the native bacterial host in MeGAXn utilization and how

they may facilitate enzyme systems for complete hydrolysis and utilization of MeGAXn, we may

better employ these enzymes in development of bacterial bioconversion processes and next-

generation biocatalysts.

Table 2-1. Composition of potential biofuel crops and other biomass sources
Non-carbohydrate
Biomass Resource Carbohydrate Composition Non-carbohydrate
Composition
Hardwood Populus tremuloides 48% cellulose 21% lignin
(Poplar)a 24% glucuronoxylan
Betula papyrifera 42% cellulose 19% lignin
(Paper Birch)a 35% glucuronoxylan

Herbaceous Switchgrass (early 40.7% cellulose 5.5% lignin
cut)b 35.1% hemicellulose

Switchgrass (late 44.9% cellulose 12% lignin
cut)b 31.4% hemicellulose

Crop residue Corn stoverb 36.4% glucan 16.6% lignin
18.0% xylan
Wheat straw 38.2% glucan 23.4% lignin
21.2% xylan

Sugarcane Bagasse' 32-48% cellulose 23-32% lignin
19-24% xylan

Adapted From:
b Timell, T. E. (1967). Recent progress in the chemistry of wood hemicelluloses. Wood Sci
Technol 1, 45-70.
a Lynd, L. R., C. E. Wyman, and T. U. Gerngross (1999). Biocommodity Engineering.
Biotechnol Prog 15, 777-793.
c Scurlock, J. (http://bioenergy.ornl.gov/papers/misc/biochar factsheet.html). Bioenergy
feedstock characteristics (Oak Ridge: Oak Ridge National Laboratory, Department of
Energy).

I

U

Figure 2-1. Pattern of cellulose fiber deposit in different layers of the primary and secondary cell
wall. A "P" designation refers to layers of the primary cell wall while an "S"
designation refers to layers of the secondary cell wall. Glucuronoxylan is thought to
be more concentrated at the interface between secondary cell wall layers. Figure
adapted from, Fujita, M., and H. Harada (2001). Ultrastructure and formation of
wood cell wall, In Wood and Cellulosic Chemistry, D. N.-S. Hon, and N. Shiraishi,
eds. (New York: Marcel Dekker), pp. 1-49.

3-1,4-Xylosidase
non-reducing end

l Amo,

Acetylesterase 2-Glucurodase a-1,3-Arabinofuranos

Figure 2-2. Common structural elements and sites of enzymatic hydrolysis which degrade methylglucuronoxylan and
methylglucuronoarabinoxylan.

idasc

Processing of Glucuronoxylan by Bacterial Enzyme Systems

GH 11 0% GH 10 4r GH 11 and 10
Non Reducing End Reducing End

Primary Aldouronates Generated
GH 10 GH 11 GH 5

S( (n).)t
Penultimate reducing end
Aldotetrauronate Aldopentauronate substimtted prduct
substituted product
> Most characterized a-glucuronidases act only on aldouronates with nonreducing end substitutions
Transporter >Aldotetrauronate has been shown to act as a catabolite activator binding the UxuR represor in Geobacillus
stearothermophilus (Shulami, et al 1999)

Xylosidase
AI

Xylotriose
Sa-Glucuronidase

Value added
products
Glucuronic acid

Figure 2-3. Generation of hydrolysis products by different families of xylanases highlighting the intricate role of GH 10 xylanases in
complete pentosan utilization with respect to the other xylanases.

Glu172

Cellvibrio japonicus GH 10
it' \B 1CLX (Harris et al. 1996)
I I-, a'), barrel

Bacillus circulans GH 11
RCSB 1XNB (Campbell et al. 1994)
P-jellyroll fold

Erwinia chrysanthemi GH 5
RCSB 1NOF (Larson et al. 2003)
(P/a), barrel with attached beta structure

Glycosylation Glu78

Glu172

HOR'

S-0 -
HO -O-H
OH

Glu78
Deglycosylation
Glul'

ROH

Glu78
Retaining reaction mechanism for Bacillus circulans GH 11 endoxylanase

Figure 2-4. Xylanase structure and function. Diverse xylanase structures catalyze identical reactions by the same mechanism.

72

%O

CHAPTER 3
FAMILY 10 GLYCOSYL HYDROLASES: STRUCTURE, FUNCTION AND
PHYLOGENETIC RELATIONSHIPS

A Paenibacillus sp. (strain JDR-2) has been isolated that is capable of efficient utilization

of MeGAXn. Studies of this organism have attributed this ability to the production of a large

multimodular GH 10 xylanase. This 154 kDa secreted protein has eight separate modules which

contribute functions for efficient hydrolysis and utilization of polymeric xylan. Two different

modules are involved with substrate association while another module is involved in cell surface

localization. The proximity between the cell, substrate and hydrolysis products which results

from the combined function of these appended modules are thought to facilitate vectoral or

directional utilization of xylan hydrolysis products (St. John et al., 2006). Analysis of

Paenibacillus sp. strain JDR-2 and characterization of XynA1 is presented in Chapter 4. This

chapter reviews GH 10 xylanases and endeavors to establish functional themes through analysis

of associated modules and application of phylogenetics.

Xylanases of Glycosyl Hydrolase Family 10

Glycosyl hydrolase family 10 (GH 10) xylanases are arguably the best studied and

understood family of xylanolytic enzymes. Their substrate is the ubiquitous P-1,4-linked xylose

backbone of the major xylans of hardwood and crop residues, the primary sources of biomass for

bioconversion to ethanol. It is expected that GH 10 xylanases have a leading role in the

degradation of MeGAXn, allowing for subsequent turnover of this major biomass component.

To date, they have been found in all three domains of life. The Carbohydrate Active Enzymes

database (CAZy) (www.cazy.org/CAZY/) has 175 bacterial GH 10 entries and 110 eukaryote

entries (Davies and Henrissat, 1995; Henrissat, 1991; Henrissat and Bairoch, 1993). As with any

sequence database in this era of genomics, most of the sequences have been deposited in

conjunction with genome sequencing projects. Hence, only a few have been studied for kinetic

properties and fewer have received detailed molecular analysis to understand mechanistic

enzyme substrate interactions.

Enzyme Structure and Mechanism

The primary unit of GH 10 xylanases is the catalytic domain (CD). This module typically

ranges in size from 30 kDa to 40 kDa. Several examples of GH 10 xylanases have been

crystallized and x-ray diffraction for structural analysis revealed a common a/8P protein fold. As

with many endo acting glycosyl hydrolases, GH 10 xylanases have a substrate binding cleft that

appears from crystal structures to run the breadth of the enzyme (Figure 2-3). The binding site of

GH 10 xylanases (as with most glycosyl hydrolases) is composed of a series of subsites that

position and bind individual xylose residues. The nomenclature for describing the organization

of subsites has been reviewed (Davies et al., 1997). It is based on the convention for the naming

of polymeric carbohydrates and the point of hydrolysis within the enzyme. Subsites are

numbered increasing from the point of hydrolysis with negative designation toward the

nonreducing terminus (glycone) (left) and a positive designation in the direction of the reducing

terminus (aglycone) (right). Hydrolysis of xylan occurs through a double displacement

mechanism with retention of anomeric configuration (Davies and Henrissat, 1995; Gebler et al.,

1992; Henrissat et al., 1995) (Fig. 2-4). Two glutamate residues have been identified which

catalyze this hydrolysis, one acting as the primary nucleophile and the other as the proton donor.

Modular Characteristics of GH 10 Xylanases

Often GH 10 xylanases are associated within their translated protein product, with

additional separately folding domains. Although a variety of different functional domains have

been identified, the majority represent carbohydrate binding modules (CBM). These separately

folding modules are P-sheet structures that bind target carbohydrates, not necessarily xylan.

Further, it is common for there to be modular repeats such that there are multiple modules of the

same family. The largest GH 10 xylanase in the CAZy database has eight separate modules and

is 194 kDa in mass. Six of the modules represent three different CBMs and a seventh module, an

additional enzymatic activity.

No direct interaction has been identified between a GH 10 CD and its associated CBM.

The modules are generally connected through linker regions that in some cases have

characteristic protein sequences. It is generally thought that linker regions lack structure, having

the singular task of connecting two functional domains. In only two cases has a CBM been

crystallized together with its cognate CD (Fujimoto et al., 2000; Pell et al., 2004a). In these

studies the linker did not yield a precise electron density map for structure analysis. The

tethering action of the linker sequence between a CD and CBM identifies a simple spatial

relationship required for enhancement in CD function. This contrasts with the concept of a

coordinated interaction in which accessory modules may directly interact with catalytic modules

for enhanced functionality (Akin et al., 2006; Irwin et al., 1998; Sakon et al., 1997). Boraston et

al. have recently reviewed the structure function and classification of CBM modules (Boraston et

al., 2004). Conventional wisdom suggests that these modules help target the CD to the expected

substrate, thereby increasing the localized substrate concentration. In contrast to idealized

kinetic systems with a soluble substrate, the recalcitrant, composite character of lignocellulosic

biomass requires enzymes to be targeted to specific regions of the substrate for effective

hydrolysis. The frequent occurrence of GH 10 xylanases that do not contain an appended

carbohydrate binding module suggest that in many cases CBMs are not necessary for desired

function.

There are 46 families of carbohydrate binding modules in the CAZy database assigned on

the basis of sequence similarity and hydrophobic cluster analysis (Henrissat and Bairoch, 1993;

Tomme et al., 1995). Currently, eleven are found associated with GH 10 xylanases. These

include members of families 1, 2, 3, 4, 5, 6, 9, 10, 13, 15 and 22. The occurrence of these

modules within GH 10 xylanases varies greatly. For instance, of all the different 106 families of

glycosyl hydrolases in the CAZy database, CBM 22 modules are primarily found associated with

bacterial and plant GH 10 xylanases (Table 3-1). CBM 9 modules are only associated with GH

10 xylanase that already have a CBM 22 module, but for these modules, only bacterial

associations are known. CBM's of families 2, 3, 5, 6 and 10 are primarily found in bacterial

enzymes, with only a few from fungal glycosyl hydrolases. Many of these modules are

associated with a variety of enzymatic activities. To exemplify an extreme distribution, CBM

family 5 has only a single module associated with a GH 10 xylanase, but has about 200 entries in

the CAZy database associated mainly with chitinase and cellulase enzymes, while family 15 has

only two entries in the database and both are associated Cellvibrio GH 10 xylanases. Of all the

families, CBM family 13 is probably the most diverse, with representatives in bacterial, fungal,

plant and mammalian proteins. These modules are only common in GH 10 xylanases from

Streptomyces (Table 3-1). CBM 1 modules are common in fungal GH 10 xylanases, but are

found more often in other families of glycosyl hydrolase enzymes of fungal origin.

CBM classification by target substrate

A relatively new classification system for carbohydrate binding modules identifies their

target substrate rather than their protein fold. Type A modules bind crystalline substrates such as

crystalline cellulose, which is not necessarily the target of the associated catalytic module. Type

B modules bind soluble substrates, which are usually the intended substrate for the catalytic

module and Type C modules bind small soluble sugars such as cellobiose. The CBM modules

listed above which are found appended to GH 10 catalytic modules have representatives in each

of these types. Modules 1, 2a (see below), 3, 5, 10 classify as Type A, modules 2b (see below),

4, 6, 15 and 22 classify as Type B and modules 9 and 13 classify as Type C (Boraston et al.,

2004). The following descriptions clarify their carbohydrate binding preferences detailing the

differences between the three types.

CBM modules common in bacterial GH 10 xylanases and their general architectural
arrangement

Modules of CBM family 22 are associated with GH 10 xylanases. There are examples

of this module associated with GH 10 xylanases of bacterial and plant origin. Even with this

diversity, there is only one example of a characterized CBM 22 not from bacterial origin. Early

studies assigned a thermostabilizing function to these domains as removal of the module resulted

in decreased thermal stability of the cognate xylanase CD (Fontes et al., 1995). It was soon

realized that these domains have a primary role in binding carbohydrate polymers. Most CBM

22 modules are located N-terminal to the CD and are often observed as a duplicate or triplicate

set (Ali et al., 2001b; St. John et al., 2006). Xynl0B of Clostridium thermocellum has a unique

CBM 22 configuration and also has been well characterized (Charnock et al., 2000; Dias et al.,

2004; Xie et al., 2001). In this case the CD is flanked on both sides by a single CBM 22 module.

Substrate binding studies showed that while the module on the C-terminal side of the CD has

affinity for xylan, the N-terminal localized CBM 22 has no detectable affinity for tested

substrates. Crystal structure analysis of the functional module revealed a P-sandwich structure

with a small cleft for binding substrate sugars. The tandem N-terminal CBM 22 modules of

XynlOA of Clostridiumjosui expressed together showed similar carbohydrate binding properties

as the C-terminal located CBM 22 of XynlOB from Clostridium thermocellum described above

(Ali et al., 2005a).

XynlOC of Clostridium thermocellum has a single CBM 22 in the N-terminal region. In a

recent report, absorption assays with the recombinantly expressed CBM showed that it bound

best to acid-swollen cellulose and ball milled cellulose but native affinity polyacrylamide gel

electrophoresis (NAPAGE) analysis showed the greatest gel retardation with birch wood xylan.

Although in these studies substrate affinity for this CBM 22 is not definitive, results showed a 4-

fold activity increase between the Xynl0C CD expression product and the native Xynl0C

protein product, confirming the generally accepted role of CBM modules (Ali et al., 2005b). The

nonbacterial contribution to CBM 22 characterizations comes from the ruminal protozoan

Polyplastron mutivesiculatum. XynlOB of this protozoan has a single N-terminal CBM 22 as

described above for XynlOC of Clostridium thermocellum. While it was shown to bind xylan, it

did not function to enhance catalytic activity (Devillard et al., 2003).

In two cases, CBM 22 modules have been shown to bind mixed linkage 0-1,3-1,4 glucan

chains. Work with XynlOB of Clostridium stercorarium did not differentiate between an N-

terminal duplicate of CBM 22 modules, but showed that they only slightly increased activity on

oat spelt xylan with respect to the separate catalytic domain. Unexpectedly, activity on 0-1,3-1,4

glucan, which was very low with the separate catalytic domain, was higher than activity on oat

spelt xylan for the native non truncated enzyme (Araki et al., 2004). Previous work with this

enzyme showed that the CBM modules facilitated binding to cellulose (Ali et al., 2001b). XynA

of Thermotoga maritima has an identical CBM 22 arrangement. Very detailed studies of these

modules identified major differences between the first (CBM 22-1) and second (CBM 22-2) (left

to right) modules. Meissner and colleagues showed by NAPAGE that CBM 22-2 bound 0-1,3-

1,4 glucan, 0-1,3-1,4 xylan, and 0-1,4 xylans while CBM 22-1 failed to show separate affinity

for these potential substrates (Meissner et al., 2000).

Their wide-spread diversity and apparent variety of specificity make CBM 22 modules

interesting platforms to study carbohydrate epitope recognition and binding. Further structural

work may lead to sugar binding cleft engineering for development of tools in biotechnology.

Thorough studies of CBM 22 modules clearly show that sequence based determination of the

presence of these modules cannot confidently be correlated to a specific function, it is clear that

these modules are involved with binding carbohydrate polymers.

CBM 9 modules are frequently associated with CBM 22 modules. These modules are

usually positioned just C-terminal to the CD and their presence is most common in modular GH

10 xylanases which already have a CBM 22 module to the N-terminal side of the CD. In several

cases, modular GH 10 xylanases from thermopiles have tandem CBM 9 modules. Defining

research, characterizing the second CBM 9 module of T. maritima Xynl0A (CBM 9-2) showed

that this module had high affinity for small soluble oligosaccharides, including glucose, xylose,

cellobiose, xylobiose. This attribute classifies these modules as Type C. Binding affinities for

oligomers over two residues did not increase, indicating that the binding epitope recognized no

more that two sugar residues. This type of CBM also displayed affinity toward xylans and

cellulose of all types (Ali et al., 2001a; Clarke et al., 1996; Notenboom et al., 2001). The

Mechanism of binding was characterized using sodium borohydride reduced polymers.

Replacement of the hemiacetyl reducing terminal sugar with sugar alcohols prevented binding of

CBM 9-2. Subsequent crystal structure analysis supported these findings revealing that every

hydroxyl group of the reducing terminal sugar in a cyclic conformation interacted with the

protein via multiple hydrogen bonding interactions (Boraston et al., 2001; Notenboom et al.,

2001).

Analysis of CBM 9-1 of T. maritima Xynl0A failed to identify a functional role for this

very similar module. Modeling of CBM 9-1 using CBM 9-2 as template and a sequence

alignment of many CBM 9 sequences showed that CBM 9-1 as well as all other CBM 9 modules

in the same modular position in a tandem arrangement lacked the structurally characterized

conserved sugar binding amino acids identified in CBM 9-2. Based on the differences between

CBM 9-1 and CBM 9-2, it has been proposed that two subfamilies be designated (CBM 9a and

9b). In all of the CBM 9 tandem arrangements, the first CBM 9 classifies as a CBM 9a and the

second CBM 9 classifies as a CBM 9b (Notenboom et al., 2001).

Many other GH 10 xylanases, including three in the alignment discussed above, have

single CBM 9 modules. The three included in the alignment and the single CBM 9 of

Paenibacillus sp. strain JDR-2, a mesophilic, aggressive xylan utilizing organism, show near

complete conservation of the key residues attributed to sugar binding in CBM 9-2 of T

maritima. This suggests that GH 10 xylanases, in which there is a single CBM 9 module, it may

function similar to that of CBM 9-2 of T. maritima.

The concise studies performed with XynlOA CBM 9b of T. maritima identified a role for

these modules in binding of reducing terminal sugars. Although this module showed affinity for

the reducing ends of xylan and cellulose, it had a much higher association constant with

cellobiose that with xylobiose, possibly indicating a preference for binding of cellulose.

Modules of CBM family 2 can bind crystalline cellulose and xylan. While there are

eleven sequences within the GH 10 family that contain this domain, there are about 200 in the

database from glycosyl hydrolase families of chitinases and various cellulases. This family has

been grouped into two subfamilies designated CBM 2a (Type A) and CBM 2b (Type B). While

CBM 2a binds to crystalline cellulose, CBM 2b has been shown to bind soluble xylan. The

difference in structure that changes substrate specificity is attributed to a single amino acid

switch which reorients a tryptophan for binding to xylan (Simpson et al., 2000). Based on this

analysis, out of the eleven CBM 2 modules found in GH 10 xylanases, only one is classified as a

CBM 2b. The other ten classify as CBM 2a, presumably having specificity for crystalline

cellulose. More examples of CBM 2b modules are associated with GH 11 xylanases. CBM 2a

modules bind crystalline cellulose irreversibly (Creagh et al., 1996) but are thought to be mobile,

allowing for movement on the surface of cellulose crystalline fibers (Jervis et al., 1997). An

example of a CBM 2a module from a GH 6 cellobiohydrolase has been shown to disrupt

crystalline cellulose (Din et al., 1994), revealing a synergism between the CBM 2 module and

the associated CD. Thermodynamic and structural analysis of these modules conclude that

binding of crystalline cellulose occurs through an entropic driven process, probably due to

displacement of water molecules between the cellulose surface and the near planar face of the

carbohydrate binding module (Creagh et al., 1996; McLean et al., 2000).

Family 3 CBM modules bind crystalline cellulose. This family is of notable interest for

cellulases of GH family 9. It can be found in five GH 10 xylanases, three of which have two

separated modules. These modules have been divided into three subfamilies. Although both

CBM 3a and 3b bind to the surface of crystalline cellulose, CBM 3a differs from CBM 3b

primarily in a loop structure which contributes to substrate binding (Jindou et al., 2006). Further,

CBM 3a modules are associated with scafoldin components of the cellulosome (Shimon et al.,

2000) where CBM 3b modules are enzyme localized (Gilad et al., 2003). The last subtype, CBM

3c, is a glycosyl hydrolase family 9 CD fusion domain which is thought to feed or guide the

cellulose chain into the GH 9 catalytic domain. This type of association is attributed to

processive degradation of cellulose (Irwin et al., 1998; Sakon et al., 1997).

CBM modules of family 6 bind soluble polymeric sugar substrates. There are seven

GH 10 xylanases containing this Type B CBM module. They differ from other group B CBM

domains in that the substrate binding location is a ridge rather than the typical cleft of the 3-

sandwich structures. They have been shown to bind a variety of soluble substrate sugars with

similar affinities. A CBM 6 module from Cellvibrio mixtus endoglucanase 5A has revealed two

binding sites, each with unique substrate specificities. Binding ofxylan was specific for cleft A

which could also bind cellooligosaccharides, while cleft B also bound cellooligosaccharides, but

was specific for p-1,3-1,4-glucans (Boraston et al., 2003; Henshaw et al., 2004; Pires et al.,

2004).

Xylanases from Streptomyces spp. have CBM 13 modules. Of the 10 CBM 13 modules

associated with GH 10 xylanases, 7 are in Streptomyces sequences (Table 3-1). These modules

have similarity to the lectin like B-chain of ricin toxin which has specificity for galactose. Each

CBM is composed of a triplicate repeat of approximately 40 amino acids and each repeat is a

separate site for carbohydrate binding. CBM 13 modules are selective for pyranose sugars with

generally low association constants at each site. Upon binding of polymeric xylan there is a

"cooperative and additive" effect (Notenboom et al., 2002), increasing the affinity for this

substrate more than by a simple additive result of the three contributing sugar binding sites

(Boraston et al., 2000; Fujimoto et al., 2002; Notenboom et al., 2002). Studies have indicated

that the three binding sites (a, 0, y) accommodate three different xylooligomers (Scharpf et al.,

2002).

CBM modules of families 4, 5, 10 and 15 are rare in GH 10 xylanases. CBM modules

of family 4 have been identified in about 30 sequences from the CAZy database. Only one of

these sequences is a GH 10 xylanase which has an N-terminal tandem set. All the others which

have been identified are associated with various P-1,4 and P-1,3 glucanases. Structural studies

have characterized this module family as having a P-sandwich jelly roll fold (Johnson et al.,

1996b). Binding of soluble carbohydrates occurs within a binding cleft. The bottom of the cleft

is lined with hydrophobic residues and the walls have hydrophilic residues for hydrogen bonding

interactions with the carbohydrate polymers. Several subfamilies of CBM 4 modules have been

identified. In general, the CBM 4 modules bind substrate for the associated catalytic module

(Simpson et al., 2002; Zverlov et al., 2001). The first studies of this module family were

performed with the N-terminal tandem CBM 4 modules from Cellulomonasfimi CenC. These

modules were specific for soluble P-1,4 glucan and did not associate with xylan (Brun et al.,

2000; Johnson et al., 1996a; Johnson et al., 1996b; Tomme et al., 1996). Xynl0A of

Rhodothermus marinus was found to have a related tandem N-terminal set of modules and

substrate binding studies showed that although they had a low affinity for soluble cellulose they

showed specificity for xylans (Abou Hachem et al., 2000). Structural analysis of the second

module of this system allowed the researchers to postulate differences which dictate substrate

specificity between those from C. fimi CenC that bound soluble cellulose and those from R.

marinus XynlOA that bind xylan (Simpson et al., 2002). The CBM 4 modules from T.

neapolitana Lam 6A, a laminarinase (P-1,3-glucan), do not bind soluble cellulose but are

specific for various P-1,3 linked glucan polymers (Zverlov et al., 2001). The diversity of

carbohydrate binding in this family is similar to that found in family 22 modules. Both are

classified together as Type B CBMs in a larger superfamily (Sunna et al., 2001).

There is only one example of a CBM 5 module in GH 10 xylanases. These modules are

thought to bind cellulose but the primary associated enzymatic activity is a chitinase. The CBM

5 ofErwinia c 1hiyn\1heini endoglucanase Z has been structurally characterized and the authors

correlated it with CBM 5 modules associated with chitinase enzymes (Brun et al., 1995; Brun et

al., 1997). At present, family 15 CBMs have only been found in two enzymes. Both are GH 10

xylanases of Cellvibrio (Table 3-1). With these modules, association constants increase up to

xylohexaose, indicating there are 6 subsites in the binding cleft. Although no natural substituted

polymer such as MeGAXn achieved as high an association constant as observed with

xylohexaose, affinity in the worst case decreased by only one-half, not a significant decrease in

the measure of association. These modules are thought to efficiently bind decorated xylan

because the 02 and 03 hydroxyls (substituted positions in native xylan) of most xylan binding

subsites are solvent exposed (Szabo et al., 2001). Only a single CBM 10 module has been found

associated with GH 10 xylanases. These 45 amino acid modules have a hydrophobic side

involved with cellulose binding. The mechanism of association with crystalline cellulose is

similar to CBM 2 modules with coplanar aromatic amino acid residues stacking on the cellulose

surface (Millward-Sadler et al., 1995; Ponyi et al., 2000; Raghothama et al., 2000).

Fungal modules

Of all the domains associated with GH 10 xylanases, CBM 1 modules are strictly found in

sequences from fungal enzymes. However it is not restricted to xylanases, being primarily found

in fungal cellobiohydrolases and cellulases. These small 36 amino acid structures have four

highly conserved cysteine residues involved in the formation of disulfide bridging (Kraulis et al.,

1989). This module has been shown to facilitate association of cellobiohydrolases and cellulases

with cellulose (Carrard et al., 2000; Gilkes et al., 1991). In GH 10 xylanases these modules are

usually located to the far N or C-terminal region, some distance from the CD. One report

showed that the CBM 1 of a GH 7 reducing terminal cellobiohydrolase (Cbhl) from Penicillium

janthinellum had a disruptive effect on cellulose that enhanced activity (Boraston et al., 2004).

No research has determined possible differences between the CBM 1 modules in cellulose active

fungal enzymes and those in fungal GH 10 xylanases. It may be that similarities among these

small modules are significantly high to discourage such endeavor. If the primary purpose of this

module is to associate the fungal GH 10 CD with cellulose, it serves a similar role as several

bacterial CBM modules.

Other modules and sequences from GH 10 xylanase

Surface Layer Homology (SLH) domains anchor proteins to the cell surface. SLH

domains have several roles in bacterial physiology. With respect to GH 10 xylanase and

glycosyl hydrolase function, these domains are often arranged as C-terminal sets (up to three

separate domains) and are involved in anchoring the associated enzyme to the cell surface. They

are also involved as a primary surface anchoring mechanism for the multicomponent lignolytic

cellulosome complex produced by several Clostridium spp. Bacterial surface binding studies of

several SLH module sets have identified two mechanisms of binding. Several studies have

identified binding to secondary cell wall polysaccharide (SCWP). These binding sites consist of

carbohydrates associated with the peptidoglycan cell wall. Genetic verification for this binding

mechanism was obtained from a csaB gene knockout in Bacillus culll/lt, i% (Mesnage et al.,

2000). Other results indicate that SLH domains bind directly to the cell wall peptidoglycan

layer. Recent work with the SLH C-terminal triplicate of the scafoldin dockerin binding protein

(SdbA) from C. thermocellum found that it bound to the peptidoglycan layer of Escherichia coli.

This was in contrast to the SLH doublets from the xylanases, Xynl0A and XynlOB of C. josui

and C. stercorarium, respectfully. In this case, these SLH domains displayed specificity for the

Clostridia SCWP extract and reduced affinity for hydrofluoric acid extracted secondary cell wall

polymer (Zhao et al., 2006a). This preference for binding native peptidoglycan suggests that the

SLH modules from the SdbA protein may be used in biotechnology applications. Recently,

similar binding selectivity was observed between the two surface layer proteins (Slpl and Slp2)

and the cellulosome anchoring protein (Ancl) of C. thermocellum (Zhao et al., 2006b). Studies

of XynAi and Xyn5, both GH 10 xylanases from different Paenibacillus spp. (Ito et al., 2003; St.

John et al., 2006) have shown that the C-terminal triplicate SLH module anchors the GH 10

xylanase to the cell surface. These triplicate SLH domains as well as the linker region to their N-

terminal have homology to the same region of the SdbA protein discussed above. Based on this

homology, these two xylanases may bind with specificity similar to those examples above which

bind native peptidoglycan with no requirement for SCWP. Currently, there may be enough

characterized examples available to allow for sequence based determination of amino acid

functionality and define the differences between these two similar modules that bind different

polysaccharides on bacterial cell surfaces.

Characteristic linker regions connect modules in some enzymes. Although the ascribed

function of linker regions in glycosyl hydrolases is that they connect together functional

domains, the identification of linkers with unique amino acid sequences has made them an

interesting topic of research. These unique sequences are characterized as having very high

content of specific amino acids. These include the serine rich linker (Sr) (Cellvibrio,

Pseudomonas, and Saccharophagus) (Hall et al., 1989), the asparagine rich linker (Nr)

(Ruminococcus), the proline and threonine rich linker (PTr) (Caldibacillus, Caldicellulosiruptor,

and Cellulomonas), the proline and glutamate rich linker (PEr) (Cohwellia, Pseudomonas, and

Saccharophagus), and the proline and glycine rich linker (PGr) (Thermobifida). The serine rich

linker regions of XynA and XynC (an arabinofuranosidase) of Cellvibriojaponicus have been

characterized (Table 3-1) (Black et al., 1997; Black et al., 1996; Ferreira et al., 1990). Initial

studies with XynA determined that the linker sequence was not required for activity and

substrate binding functions. Removal of this intervening sequence resulted in lower activities.

Although this could be attributed to other functions, it was concluded that it resulted from

reduced flexibility of the CD with respect to the CBM.

A completely novel linker sequence has been identified in XynB of Neocallimastix

patriciarum. It is composed of 12 tandem repeats of the core amino acid sequence TLPG

followed by 45 tandem repeats of the octapeptide XSKTLPGG (X=S, K or N). This linker

region connects a C-terminal family 1 CBM. Research to elucidate this modular system failed to

obtain good expression for functional analysis but showed that the CD sequence coded for a

functional GH 10 xylanase (Black et al., 1994).

As discussed above, CBM 22 modules were originally thought to confer thermostabilizing

properties to GH 10 xylanases. New research shows it is possible that these conclusions resulted

from the presence of the linker sequence. The 18 amino acids connecting the CBM 22 module

with its cognate GH 10 CD has recently been shown to attribute thermo stabilization and

resistance to proteolysis (Dias et al., 2004).

Glycosyl Hydrolase Accessory Module Discussion

The descriptions above regarding the modules appended to GH 10 CDs exemplify the

functional diversity common in glycosyl hydrolase families for lignocellulose degradation.

Whether associated with a xylan or crystalline cellulose binding domains the assumed goal of

these modules is to facilitate interaction with the substrate. Modules like family CBM 2, which

target crystalline cellulose, may have roles in xylan hydrolysis by GH 10 xylanases that cannot

easily be determined. Endeavors to distinguish functionality of these modules with respect to the

GH 10 catalytic core may facilitate development of applications for efficient enzymatic

hydrolysis of lignocellulosics.

Bacterial GH 10 domain architecture. As can be observed in Figure 3-1, common

domain arrangements are evident. Significant modular arrangements include: CBM 22 modules

are localized to the N-terminal region of the GH 10 CD (except for one GH 10 of C.

thermocellum), all CBM 9 modules are localized to the C-terminal side of the CD and all but one

is associated with CDs that also have a CBM 22 module. All sequences which have SLH

modules for possible cell surface anchoring also have both CBM 22 and 9 modules. CBM 3

modules are always immediately flanked by proline and threonine-rich linker regions and are

only found in Caldibacillus and Caldicellulosiruptor. In several cases there are two of these in

the same xylanase. CBM 2 modules are in GH 10 xylanases from Cellvibrio, Sacharophagus,

Cellulomonas, Streptomyces and Thermobifida (Table 3-1). These modules are also often

flanked by a proline and threonine-rich or a serine-rich linker sequence. Predictions of GH 10

xylanase function can be proposed based on common architectural module associations and an

understanding of the function of these carbohydrate binding modules.

Information concerning the method by which the CBM facilitates CD activity can also be

deduced from positional relationships. While many of the CBM modules of bacterial GH 10

xylanases are usually in a specific position with respect to the CD, some modules are not. The

CBM 1 module is only found associated with fungal CDs. Of the fourteen which have this

domain, seven have it toward the N-terminal and seven have it toward the C-terminal. From this,

it seems that the only purpose of this module is to ensure localization to the lignocellulose

substrate.

From Figure 3-1 and the brief description of associated modules above, we can imagine a

mode of action for these GH 10 xylanases based on their respective module assemblages. The

Thermoanaerobacterium saccharolyticum (P36917) and Paenibacillus species JDR-2

(62990090) xylanase would be expected to associate with soluble polymers such as xylan or 3-

1,3-1,4 glucan with their N-terminal CBM 22 domains. The CBM 9 module is expected to bind

the reducing terminus of a cellulose chain fixing the catalytic module in place and the C-terminal

SLH modules should anchor this enzyme system to the cell surface. The combined properties of

these appended modules favor simultaneous substrate and cell surface localization, perhaps

increasing hydrolysis product recovery by the cell through a process of vectoral transport. The

triplicate family 22 CBM in the N-terminal region of the Arabidopsis thaliana GH 10 xylanase

(Q9SM08) is expected to facilitate substrate localization for this CD. How these enzymes

function in A. thaliana is difficult to determine but it can be imagined that they may function in

expansion of the cell wall. The Caldibacillus cellulovorans xylanase (7385020) has a C-terminal

localized double CBM 3 set. These modules would be expected to bind the crystalline surface of

cellulose and the N-terminal CBM 22 would promote association with soluble substrate. A

similar mode of action can be imagined for the xylanase from Cellulomonasfimi (73427793).

The irreversible binding and mobile character of the C-terminal CBM 2 module would allow an

associated CD to translate the surface of cellulose crystals, in search for substrate. The

Streptomyces coelicolor (Q8CJQ1) modular xylanase is the simplest of all the examples. The

CBM 13 module in the C-terminal region is expected to associate with soluble xylan and

increase localized substrate concentration to enhance enzymatic efficiency.

These examples offer a glimpse into the possible mode of action for several modular GH

10 xylanases. Although these descriptions are not absolute, they provide a framework for

development of methods which utilize these enzymes for complex biomass degradation.

Hydrolysis of Substituted Xylans by GH 10 Xylanases

Xylan hydrolysis by GH 10 xylanases primarily result in the limit products xylose,

xylobiose, xylotriose and small substituted xylooligomers. Early studies using GH 10 xylanases

from Cryptococcus albidus and Streptomyces lividans to digest methylglucuronoxylan resulted in

the characterization of aldotetrauronate (MeGAX3) as the smallest substituted

xylooligosaccharide (Fig. 3-2) (Biely et al., 1997). Similar work digesting insoluble wheat

arabinoxylan with the GH 10 xylanase, XylA from Thermoascus aurantiacus resulted in two

small substituted limit products. Arabinofuranose-xylobiose (AX2) with the substitution in the

03 position of the nonreducing xylose of xylobiose and arabinofuranose-xylotriose (AX3) with

the same substitution on the middle xylose of xylotriose resulted as 50% and 30% respectively of

the total arabinofuranose (Araj) substituted products (Fig. 3-2) (Vardakou et al., 2003). These

biochemical methods have recently been validated with detailed structural studies of two GH 10

xylanases together with these limit products (Fujimoto et al., 2004; Pell et al., 2004b).

Xylan has been reported to have a three fold helical symmetry. Binding subsites of GH 10

xylanases and CBM modules specific for xylan accommodate this characteristic. Native xylan is

usually substituted at the 02 or 03 hydroxyl positions (Chapter 2). Substitutions in these

positions along the xylan chain can either be accommodated into the protein structure or exposed

to the solvent so as not to interfere with subsite xylan interaction. Specific interactions can be

understood from the positioning of the 02 and 03 hydroxyl in the subsite bound xylose residue

relative to the protein structure. If a subsite orients the bound xylose moiety such that these

positions of the xylose are sterically confined by protein structure, no substitution can be

accommodated in that position. For a subsite to bind a substituted xylose moiety in the xylan

chain, there can either be a pocket into which the substitution can fit in the protein tertiary

structure, or the 02 and 03 hydroxyl positions can be solvent exposed away from the protein

surface. As will be seen below, substituted hydrolysis products can also result from subsite

flexibility. Resulting substituted hydrolysis limit products reflect subsite accommodation by GH

10 xylanases.

Hydrolysis of Methylglucuronoxylan

Crystal structure analysis of GH 10 xylanases from Streptomyces olivaceoviridis (XynlOA)

and Cellvibrio mixtus (XynlOB) have provided molecular level determination of subsite

interactions of the methylglucuronosyl moiety on the xylan chain (Fujimoto et al., 2004; Pell et

al., 2004b). The limit product, MeGAX3, was cocrystallized with an active site mutant and

structure analysis revealed binding of this hydrolysis product in the -3 through -1 subsites and +1

through +3 subsites. Binding of MeGAX3 reflected enzyme substrate interactions, indicating the

-3 and +1 subsites accommodate methylglucuronosyl substitutions. For C. mixtus XynlOB, the -

3 subsite methylglucuronosyl could not be modeled as electron density was diffuse, but for the

same position in S. olivaceoviridis XynlOA electron density was clear. In this position the 02

hydroxyl is solvent exposed and the substituted methylglucuronate is extended up into solvent.

No interactions between the methylglucuronate and protein were identified to explain the clear

electron density observed for XynlOA. In the +1 subsite, the 02 position points into the protein.

A pocket in this position accommodates 02 substituted methylglucuronosyl moieties. For S.

olivaceoviridis XynlOA, diffuse electron density did not allow modeling, indicating that the

protein has minimal interactions with this carbohydrate residue but is structured to allow

unrestricted access in this position. In the case of C. mixtus XynlOB, clear electron density was

observed for the methylglucuronosyl in this position. In XynlOB, the +1 subsite has more

xylose-binding interactions then other GH 10 xylanases, and while in the methylglucuronosyl

pocket the glucuronate moiety is hydrogen bound to two separate amino acid residues. The

additional stability in this position is used to explain the clear electron density for the

methylglucuronate and significantly increased activity with respect to other xylanases on the

polymeric substrate MeGAXn (Fujimoto et al., 2004; Pell et al., 2004b). Identification of the

methylglucuronosyl pocket within the aglycone +1 subsite suggests that GH 10 xylanases may

have evolved to address this specific 02 hydroxyl substitution.

Hydrolysis of Methylglucuronoarabinoxylan

GH 10 xylanase crystal structure analysis of Araf substituted hydrolysis products, AX2 and

AX3, did not identify conserved Arafprotein interactions. Results for XynlOB of C. mixtus and

XynlOA of S. olivaceoviridis were comparable. XynlOB binding of AX2 and AX3 in the

glycone subsites resulted in clear xylose modeling in subsites -1 through -2 and -1 through -3,

respectively. In both cases the Araf substitution yielded clear electron density. The Arafof AX2

had two alternative conformations. In one, Arafhydrogen bonds with the protein and in the

other, similar to the positioning of Arafin AX3, the 03 hydroxyl hydrogen bonds to the 05

endocyclic oxygen of the xylose in subsite -3 having no direct interaction with the protein.

XynlOA is similar to this, but electron density is not clear for Arafof AX2 in the -1 through -2

subsites. AX3 however resulted in clear modeling of the Arafmoiety. In this case the 03

hydroxyl of Arafhydrogen bonded with two separate positions within XynlOB.

Interactions of AX2 and AX3 in the aglycone subsite region identified possible xylose

subsite binding flexibility. XynlOB of C. mixtus did not have clear electron density data for

AX2, but the xylotriose backbone of AX3 modeled into subsite +1 through +3 as expected. For

both enzymes, no Arafmoiety could be clearly modeled in the aglycone sites. For XynlOC of S.

olivaceoviridis, the oligomers only allowed modeling of xylobiose in subsite +1 through +2 with

the third xylose of AX3 not clear in subsite +3. Based on the modeling for xylose residues in the

+1 and +2 subsites for both oligomers, it is assumed Arafis positioned in these subsites for AX2

and AX3, respectively. In the case of AX3, the +2 subsite xylose was slightly displaced from the

binding subsite suggesting that Arafwas wedging into an awkward position. It is a good

indication that this flexibility in arabinose accommodation is normal as XynlOC was used to

generate AX3 as the major Araf substituted hydrolysis product of wheat arabinoxylan. AX2 was

produced by hydrolysis of the same with XynlOB. Subsite binding of this oligomer into the

expected aglycone subsites did not allow modeling. Further the authors identified restrictions of

the 03 hydroxyl of xylose in the +2 subsite of XynlOB, suggesting that accommodation of AX3

would be more difficult then in XynlOC.

It is apparent that glycone subsites of both enzymes can accommodate 02 linked

glucuronosyl in the -3 and an 03 linked Arafin the -2 subsites. These substitutions occur as the

02 and 03 in these positions are solvent exposed. For aglycone subsites, 02 glucuronosyl

substitutions in the +1 subsite are easily accommodated within a pocket. Araf accommodation in

this area of the catalytic cleft seems to be variable among xylanases. The differences between

these two enzymes can be highlighted by the fact that XynlOB of C. mixtus was used to produce

AX2 and XynlOC of S. olivaceoviridis was used to produce AX3. The latter, as positioned in the

aglycone binding region of XynlOC, revealed a flexibility of xylose binding in the +2 subsite

which helps explain how the Arafin this position is accommodated. XynlOB was suggested not

to have this flexibility for 03 linked Arafin the +2 subsite, but based on hydrolysis product

analysis must accommodate it in the +1 subsite.

Hydrolysis of Rhodymenan by GH 10 xylanases

Only one reported study has considered the hydrolytic products of GH 10 xylanases on

substrates other than P-1,4-linked xylans. Rhodymenan, a 0-1,3-1,4-linked xylan digested with

the two GH 10 xylanases of Cryptococcus albidus and Streptomyces lividans discussed above,

resulted in the hydrolysis limit product xylosyl-P-1,3-xylosyl-P-1,4-xylose (Biely et al., 1997).

GH 10 Xylanase Substrate Binding Cleft Studies

An important expectation has recently been addressed, which changes the way we must

consider synergy of methylglucuronoxylan hydrolysis between different families of xylanases.

This expectation was that the smallest methyglucuronate substituted hydrolysis product released

by a GH 11 xylanase, aldopentauronate (MeGAX4), would be further hydrolyzed by a GH 10

xylanase with release of xylose and generation of aldotetrauronate. MeGAX4 is substituted

penultimate to the nonreducing terminal xylose of xylotetraose and this methylglucuronosyl

substitution would be expected to guide the substrate into the +1 subsite where the

methylglucuronosyl can be accommodated. The additional xylose would then be expected to lie

across the active site residues and hydrolysis would release xylose. In this interesting study, four

different GH 10 xylanases did not have activity on this substrate (Kolenova et al., 2006).

However, hydrolysis of the substrate aldohexauronate, in which the methylglucuronosyl moiety

is substituted on the middle xylose of xylopentaose, resulted in release of xylobiose. This study

may have identified a substrate requirement for GH 10 xylanases. The inability of these GH 10

xylanases to use MeGAX4 as substrate but use MeGAX5, indicates that binding of xylose to the -

1 subsite does not occur with a single xylose (Kolenova et al., 2006). This reflects strong

binding of xylose at the -2 subsite compared to binding at the -1 subsite.

In a similar study, the GH 10 xylanase of T. aurantiacus (Xynl0) was shown to use an 03

Arafsubstitution in the -2 subsite as a substrate specificity determinant (Vardakou et al., 2005).

It was determined that multiple interactions between the Arafmoiety and amino acids in the

enzyme stabilized the interaction with this substituted substrate more than with unsubstituted

substrate. The purpose of this interaction was validated with comparison of activity on

xylotriose and AX3 in which the arabinose was substituted 03 on the nonreducing terminal

xylose. The results showed a four-fold higher activity on the Arafsubstituted substrate.

An interesting comparison between the above result and the previous discussion,

considering hydrolysis of MeGAX4, is that a single xylose extending across the active site from

the glycone region was hydrolyzed, suggesting that the +1 sub site binds xylose without

additional interactions into the +2 subsite. In the work described for GH 10-catalyzed hydrolysis

of MeGAX4, a single xylose residue extending across the active site into the glycone region was

not hydrolyzed, but the larger xylobiose was hydrolyzed. It seems possible that the

methylglucuronosyl substitution in MeGAX4 may limit the binding of xylose in the -1 subsite.

These findings suggest that GH 10 and GH 11 xylanases may not function synergistically.

Rather, GH 11 xylanases may hinder the full potential of a GH 10 xylanase. Identifying how

decorated substrates interact with the catalytic cleft of GH 10 xylanases is important. This

knowledge can be used to develop enzyme mixtures for efficient hydrolysis of target biomass

substrates. Studies to determine the functional properties of GH 10 xylan binding subsites will

also help to develop synthetic xylanases with engineered characteristics.

Phylogenetic Relationships of Glycosyl Hydrolase Family 10 Xylanases

Phylogenetic analysis of 241 GH 10 CD sequences is presented in Figure 3-3. A complete

list of compared sequences and their attributes is presented in Table 3-2. The tree identifies three

major branches of divergence (A, B and C). The first major branch, A, diverges to plants (Ai)

and a bacterial clade (A2). The bacteria in this clade closely associate with the B branch bacterial

clade which contains most members of the phytopathogenic genus Xanthomonas. This large

bacterial clade is made larger, grouping close to the Ci bacterial subgroup which diverges from

the major C branch. The C1 clade contains interesting bacterial genera such as Rhizobium,

Agrobacterium, Synechococcus, Anabaena and Nostoc. The divergent C2a branch leads to most

fungal sequences. It diverges into two major fungal clades one of which splits into a

Streptomyces clade. This association is of notable interest as filamentous prokaryotic

Streptomyces spp. have similar cell structure and morphological stages as some fungi. The third

fungal group is composed entirely of Fusarium which branches separately from other fungi.

Sequences 1-59 are comprised of almost entirely bacterial species. Many of these sequences

cluster by bacterial genus or enzyme characteristics, such as the associated modular architecture.

For instance, sequences 22 through 44 are all highly modular enzymes consisting of similar

modular type and architecture.

Plant and Related Bacterial GH 10 Xylanase.

The phylogenetic tree highlights associations between GH 10 xylanases of plants and

bacteria which can only be attributed to close evolutionary origin or interaction. Bacterial

sequences from 189 through 206 which originate partially from branches B and all of A2 are

closest to plant sequences but represent such a diverse assemblage of genera that it is difficult to

draw conclusions. However, the remaining sequences in branch B (177 through 188) and those

in the closely associated C1 clade contain bacteria with clear similarities or associations to plants.

These represent sequences from the phytopathogenic genera Xanthomonas and Argobacterium

and the well studied plant pathogen Pseudomonas syringae. Also included are three different

genera of cyanobacteria, the nonsulfur purple photosynthetic bacterium Rhodopseudomonas

palustric and two rhizosphere nitrogen-fixing plant endosymbionts.

Similarities between these sequences may arise from common ancestry or common

evolutionary ascendancy defined by prolonged plant-bacterial interaction. It would be expected

that plant GH 10 xylanases have a role in expansion of the cell wall and may have the inherent

capability for hydrolysis of highly substituted xylans. Plant pathogens probably would benefit

from these same properties found in plant GH 10 xylanases as they are expected to perform a

similar task in a similar environment.

Due to these possibilities, GH 10 xylanases from plant pathogens may have interesting

characteristics when compared to the same from saprophytic microorganisms. Although the goal

of each of these enzymes is considered to be the same, the substrate for saprophytes is not

unaltered plant tissue but is rather decaying biomass, occurring through the function of

saprophytic microbial consortia. The combined activities of many hydrolytic enzymes within

this environment may present a significantly altered substrate. These GH 10 xylanases may have

evolved high turnovers on simplified substrate vs. others, having lower rates on complex

substrates.

Fungal and Streptomyces Association

GH 10 xylanases from fungal origin are intriguing in that they seem significantly less

complex then the modularly diverse bacterial xylanases. Of the two major fungal clades, the one

containing sequences 127 through 162 have seven sequences with an appended CBM 1 module,

six having this module in the N-terminal domain. The other clade, containing fungal sequences

97 through 113 (17 sequences) contains seven which have a CBM 1 module in the C-terminal.

From this, it seems that the difference between the two clades including the positioning of the

CBM 1 module is reflected within the sequence of the CD. A clade for the genus Streptomyces

intervenes between the two fungal groupings. Every sequence in this clade has a C-terminal

CBM module. Most have CBM 13 modules (p-trefoil), but there are four CBM 2 modules, two

found with Cellulomonasfimi sequences in this clade. These are the only two which are not

Streptomyces spp.

Bacterial GH 10 Xylanases: Tools to Work With

Sequences 1 through 79 and also 89 through 96 are primarily bacterial. Approximately

52% of these sequences contain accessory modules. Several small clades come directly off the

C2 branch but most branch from the C2b (Fig. 3-3). In this subgroup, sequences 1 through 9, 14

through 21 and 45 through 59 consist of CDs only. Of these, sequences 1 through 9 do not

contain detectable secretion signal sequences and are therefore considered intracellular.

Sequences 10 through 13 and 22 through 44 are all modular, many showing a common modular

architecture consisting of a CBM 22 and CBM 9 appended to the CD. Several of these that

group closely, including Paenibacillus sp. strain JDR-2, also have SLH modules involved with

cell surface anchoring.

Conclusion

This review emphasizes the wide distribution and significant diversity of glycosyl

hydrolases of family 10 xylanases. The array of accessory modules often found associated with

GH 10 xylanases highlights possible functional variability and suggests that directed effort to

develop xylanases to facilitate preprocessing may benefit from inclusion of these modules.

Substrate binding studies of GH 10 xylanases have revealed the details describing the interaction

of the GH 10 catalytic cleft with substitutions on the methylglucuronoxylan chain. For subsites

which bind xylose orienting the 02 or 03 hydroxyls into the protein, substitutions can only be

accommodated by open secondary structure such as the existence of a pocket as in the case of 02

substituted methylglucuronate in the +1 subsite, or by subsite flexibility as suggested for binding

of AX3 in the +2 subsite. Aglycone substrate binding accepts methyglucuronate in the -3 subsite

and 03 substituted Arafin the -2 subsite. These positions are solvent exposed and generally

display little to no interactions between the protein and the substrate appendage. The product

variability found for hydrolysis of methyglucuronoarabinoxylan suggest that minor amino acid

changes within the xylan binding cleft may contribute to to large differences in hydrolysis

product profiles. Even though the catalytic cleft is well conserved, differences in the ability of

GH 10 xylanases can occur as a result of appended accessory modules and variations in the

catalytic cleft.

Phylogenetic tree analysis identified interesting associations between the GH 10 xylanases

from different organisms. Although the role of GH 10 xylanases has not been determined in

plants, the number of available sequences from Arabidopsis and other plant genera suggests that

they may have some function in cell wall alteration. Proximity in the phylogenetic tree identifies

a close similarity of GH 10 xylanases from plants and several plant pathogens and photosynthetic

organisms. Other than the distribution of the CBM 22 modules which are significant in plant

sequences, and the CBM 1 module that is strictly fungal, all other modules and the CBM 22

module are found only in bacterial GH 10 xylanases. This again, highlights the significant

diversity available to accentuate GH 10 catalytic abilities and identifies bacterial GH 10

xylanases as a significant biotechnology resource for bioengineering and development of next-

generation bacterial biocatalysts.

Common Domain Arrangements

Thermoanaerobacterium saccharolyticum (P36917)
alllllGMMMMM llllllll GD OD ~ ~

Paenibacillus sp. strain JDR-2 (27227837)

Arabidopsis thaliana (Q9SM08)

Caldibacillus cellulovorans (7385020)

Cellulomonasfimi (73427793)

Streptomyces coelicolor (Q8CJQ1)

IIII CBM 22 GH GH 10 iE CBM 9

777 CBM 2

CBM 13 SLH Domain

Figure 3-1. Common domain arrangements found in GH 10 xylanases.

f^ CBM3

EOGDED

Table 3-1. Distribution by bacterial genus of carbohydrate binding modules and other functional domains associated with GH 10
xylanases.
Other
Genus Family of Carbohydrate Binding Module Linking Sequence SLH
Domain
2 3 4 5 6 9 10 13 15 22 Nr Sr PEr PTr PGr

Aeromonas
Anaerocellum
Bacillus
Caldibacillus
Caldicellulosiruptor
Cellulomonas
Cellvibrio
Clostridium
Colwellia
Cytophaga
Eubacterium
Fibrobacter
Nonomuraea
Paenibacillus
Prevotella
Pseudomonas
Rhodothermus
Ruminococcus
Saccharophagus

* *0

* *

* *

* *

Streptomyces *g
Thermoanaerobacterium *
Thermobifida *
Thermotoga *
a GH 5 cellulase module and a truncated GH 43 module. b Chitin binding module and a deacetylation module. C Esterase module.
d Cadherin repeat module and Salmonella repeat of unknown function. e GH 11 module. f Chitin binding module and a truncated GH
43 module. g GH 62 module and a chitinase module.

Glycone Hydrolysis Aglycone
non reducing
reducing end
end

MeGAX3
lCbH3 0H13
Glucuronoxylan

Arabinoxylobiose

Arabinoxylan *N,4.T .J
Arabinoxylotriose

+

X2

X3
Figure 3-2. Products formed by the hydrolysis of methylglucuronoxylan and methyglucuronoarabinoxylan by a glycosyl hydrolase
family 10 xylanase. Substituted hydrolysis limit products are determined by the interaction between the substitutions and
binding subsites.

Table 3-2. Glycosyl hydrolase family
Organism

1
2
3
4
5
6
7
8
( 9
10
11
12
13
14
15
16
17
18
19
20

Caldicellulosiruptor saccharolyticus
Caulobacter crescentus CB15
thermophilic anaerobe NA10
Ampullaria crossean
Eucalyptus globulus subsp. globulus
Hordeum vulgare subsp. vulgare
Aeromonas punctata
Bacillus sp. BP-23
uncultured bacterium
Geobacillus stearothermophilus
Geobacillus stearothermophilus
Bacillus alcalophilus
Bacillus sp.
Thermobacillus xylanilyticus
ButyrivibrioJ,.i- .. h ,i,
Thermotoga sp. strain Fj SS3-B.1
Thermotoga maritima
Thermotoga neapolitana
Thermotoga sp. strain Fj SS3-B.1
Clostridium stercorarium
Clostridium stercorarium
Geobacillus stearothermophilus
Bacillus sp. NG-27
Bacillus firms
Bacillus halodurans
Bacillus so.

10 sequences included in phylogenetic studies and some of their properties
Accession Module Architecture
Number
144299 CD
16127272 CD
024820 CD/PTr/CBM3/PTr/Cel5
66474472 CD
88659658 CD
Q8GZB5 CD
61287936 CD
3201483 CD
Q7X3W7 CD
499714 CD
73332107 CD
37694736 CD
662884 CD
069261 CD
48963 CD/tAes
Q9WWJ9 CBM22(2)/CD/CBM9(2)
Q60037 CBM22(2)/CD/CBM9(2)
Q60042 CBM22(2)/CD/CBM9(2)
Q9R6T4 CBM22(2)/CD/CBM9(2)
216419 CD
23304849 CD
P40943 CD
2429332 CD
34978678 CD
56567273 CD
216371 CD

Secreted

Table 3-2. Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some of their properties. Continued.

Organism Accession Module Architecture Secreted
Number

Bacillus halodurans
unidentified
Caldicellulosiruptor saccharolyticus
Caldicellulosiruptor sp. Rt8B.4
Anaerocellum thermophilum
Caldicellulosiruptor saccharolyticus
Caldicellulosiruptor saccharolyticus
Caldicellulosiruptor sp. Tok7B. 1
unidentified
Caldicellulosiruptor sp. Tok7B. 1
Aeromonas punctata
Bacillus sp. BP-23
Cellulomonas fimi
Cellulomonas pachnodae
Clostridium thermocellum

22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47

22597186 CD Yes

39749821
2645425
P40944
1208895
2645417
40646
4836168
39749823
4836167
3810965
3201481
1103639
5880612
144776
Q60046
P36917
Q60043
5360744
23304851
12225048
62990090
27227837
7385020
109701250
P48789
21110686

tCBM22/CD/CBM9t
CBM22(2)/CD
CBM22(2)/CD
CBM22(2)/CD
CBM22(2)/CD
CD/PTr/CBM3/PTr/Cel5
CD/PTr/CBM3/PTr/CBM3(2)/PTr/Cel5
CD
CBM22(2)/CD/PTr/CBM3(2)/PTr/CBM3/PTr/GH43t/CBM6
CBM22/CD/CBM9(2)
CBM22(2)/CD/CBM9(2)
Deac/CBM22/CD/CBM9
CBM22(2)/CD/CBM9/CBM5/ChtBD3
CBM22/CD/CBM9(2)/SLH(3)
CBM22(2)/CD/CBM9(2)/SLH(3)
CBM22(2)/CD/CBM9(2)/SLH(2)
CBM22(2)/CD/CBM9(2)/SLH(3)
CBM22/CD/CBM9
CBM22/CD/CBM9
CBM22/CD/CBM9/tSLH
CBM22(3)/CD/CBM9/SLH(3)
CBM22(2)/CD/CBM9/SLH(2)
CBM22/CD/PTr/CBM3/PTr/CBM3
CD
CD
CD

21

Thermoanaerobacterium ;i, ,. iiil--, ,., ,. ,
Thermoanaerobacterium saccharolyticum
Thermoanaerobacterium sp.strain JW/SL-YS 485
Clostridium stercorarium
Clostridium stercorarium
Clostridium josui
Paenibacillus sp. JDR-2
Paenibacillus sp. W-61
Caldibacillus cellulovorans
Pseudoalteromonas atlantica T6c
Prevotella ruminicola
Xanthomonas axonopodis pv. citri str. 306

Table 3-2. Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some of their properties. Continued.

Organism Accession Module Architecture Secreted
Number

48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74

Xanthomonas campestris pv. vesicatoria str. 85-10
Xanthomonas campestris pv. campestris str. 8004
Xanthomonas campestris pv. campestris str.
Bacteroides ovatus
uncultured bacterium
Flavobacterium sp. MSY2
Saccharophagus degradans 2-40
Cellvibrio japonicus
Cellvibrio mixtus
Caldicellulosiruptor saccharolyticus
Dictyoglomus thermophilum
uncultured bacterium
Clostridium cellulovorans
Clostridium thermocellum
Polyplastron multivesiculatum
Clostridium thermocellum
Ruminococcus flavefaciens
Epidinium caudatum
ButyrivibrioJ, *-h.. a,
Eubacterium ruminantium
Cellvibrio japonicus
Cellvibrio mixtus
Cellvibrio japonicus
Saccharophagus degradans 2-40
Neocallimastix patriciarum
Clostridium thermocellum
Rhodopirellula baltica SH 1

78038341
66575835
21115393
450852
56709936
68525474
89951878
5690438
37962277
2645419
973983
Q8VPE4
47716661
4850306
Q9UOG1
P51584
P29126
28569972
P23551
974180
38323070
757809
45520
89952176
Q02290
P10478
32446690

CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CBM22/CD
CBM22/CD
CBM22/CD
CBM22/CD/CBM22/Est
GH11/Nr/CD
CD/CBM13
CD
CBM22/CD/CBM9
Sr/CBM15/CD
Sr/CBM15/CD
CBM2/Sr/CBM10/Sr/CD
CBM2/Sr(2)/CD
CD/XSKTLPGG(45)/CBM1
Est/CBM6/CD
CD

No
Yes
Yes
Yes
Yes
ND
Yes
Yes
TAT
No
ND
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
No
ND
ND
Yes
Yes
Yes
Yes
No

Table 3-2. Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some of their properties. Continued.

Organism Accession Module Architecture Secreted
Number

75
76
77
78
79
80
81
82
83
84
85
86
87
00 88
89
90
91
92
93
94
95
96
97
98
99
100
101

Thermotoga sp.
Thermotoga neapolitana
Thermotoga maritima
Clavibacter michiganensis subsp. michiganensis
Streptomyces turgidiscabies
Aspergillus nidulans FGSC A4
Fusarium oxysporum
Fusarium oxysporum
Fusarium oxysporum f. sp. lycopersici
Fusarium oxysporum
Fusarium oxysporum
Fusarium oxysporum
Fusarium oxysporum
Fusarium oxysporum
Prevotella ruminicola
Streptomyces avermitilis MA-4680
Thermobifidafusca YX
Streptomyces avermitilis MA-4680
Thermobifida alba
Thermobifidafusca YX
Colwellia psychrerythraea 34H
Cryptococcus adeliensis
Patent 5693518
Aspergillus nidulans FGSC A4
Aspergillus oryzae
Penicillium funiculosum
Talaromyces emersonii

Q60044
Q60041
Q9WXS5
31559721
57338460
40745311
Q8TGC2
Q8TGC3
093976
Q8TGC4
19912845
19912843
21699819
19912853
P72234
29605742
71916922
29608643
P74912
71917054
71145740
013436
3015123
40742582
83775646
53747929
21437253

CD
CD
CD
CD
CD/ChNT
CD
CD
CD
CD
CD
CD
CD
CD
CD
CBM22/tCD
CD
CD
CD/CBM2
CD/PGr/CBM2
CD/PGr/CBM2
CD
CD
CD/CBM1
CD/CBM1
CD
CD/CBM1
CD/CBM1

Table 3-2. Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some of their properties. Continued.

Organism Accession Module Architecture Secreted
Number

102
103
104
105
106
107
108
109
110
111
112
113
114
C 115
116
117
118
119
120
121
122
123
124
125
126
127
128

- fl i,,y ..'..' i,." grisea 70-15
Neurospora crassa OR74A
Neurospora crassa OR74A
J f,, .... ioh.. grisea
Gibberella zeae
Neurospora crassa OR74A
Humicola grisea
i i,,.''.... thi '.. grisea 70-15
Aureobasidium pullulans var. melanigenum
Gibberella zeae
Agaricus bisporus
Phanerochaete chrysosporium
Streptomyces coelicolor
Streptomyces lividans
Streptomyces olivaceoviridis
Streptomyces thermocyaneoviolaceus
Streptomyces thermoviolaceus
Streptomyces avermitilis
Patent 6300114
Nonomuraea flexuosa
Cellulomonas fimi
Cellulomonas fimi
Streptomyces chattanoogensis
Streptomyces coelicolor
Streptomyces halstedii
Alternaria alternate
Cochliobolus carbonum

39973147
32416834
32413873
24496243
50844266
32407695
P79046
39963865
84469404
56555501
060206
Q9HEZ1
Q8CJQ1
P26514
Q7SI98
Q9RMM5
38524461
Q9X584
34606109
Q8GMV6
73427793
144425
Q9X583
Q9RJ91
Q59922
Q9UVP5
49066418

Bpht/CD
CD
CD
CD
CD
CD/CBM1
CD/CBM1
CD/CBM1
CD
CD
CD
CBM1/CD
CD/CBM13
CD/CBM13
CD/CBM13
CD/CBM13
CD/CBM13
CD/CBM13
CD/CBM13t
CD/CBM13
CD/PTr/CBM2
CD/PTr/CBM2
CD/CBM13/GH62
CD/CBM2
CD/CBM2
CBM1/CD
CD

Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
ND
Yes
Yes
Yes
Yes

Table 3-2. Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some of their properties. Continued.

Organism Accession Module Architecture Secreted
Number
129 Claviceps purpurea 074717 CD Yes
130 Fusarium oxysporum P46239 CBM1/CD Yes
131 Fusarium oxysporum f. sp. lycopersici 059937 CBM1/CD Yes
132 Gibberella zeae 50844270 CD Yes
133 i h,.i,lt... h, .i grisea 22415585 CBM1/CD Yes
134 Coniothyrium minitans 11876710 CBM1/CD Yes
135 Fusarium oxysporum f. sp. lycopersici 059938 CD Yes
136 Gibberella zeae 50844272 CD Yes
137 Cryptovalsa sp. BCC 7197 53636303 CD Yes
138 Neurospora crassa OR74A 32410597 CD Yes
139 Hypocreajecorina 6705997 CD Yes
140 il,,lq... di. grisea 70-15 39951799 CD Yes
141 il,,,q ..... icgrisea Q01176 CD Yes
- 142 Agaricus bisporus Q9HGX1 CD No
143 Volvariella volvacea Q7Z948 CBM1/CD Yes
144 Emericella nidulans 95025700 CD Yes
145 Emericella nidulans Q00177 CD Yes
146 Aspergillus oryzae 83772405 CD Yes
147 Thermoascus aurantiacus P23360 CD Yes
148 Aspergillus oryzae 15823785 CD Yes
149 Aspergillus oryzae 83766611 CD Yes
150 Aspergillus sojae Q9P955 CD Yes
151 Aspergillus terreus 68161138 CD Yes
152 Aspergillus oryzae 094163 CD Yes
153 Aspergillus oryzae 83775732 CD Yes
154 Penicillium canescens 55792811 CD Yes
155 Penicillium simplicissimum P56588 CD No

Table 3-2. Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some of their properties. Continued.

Organism Accession Module Architecture Secreted
Number
156 Penicillium purpurogenum Q9P8J1 CD Yes
157 Aspergillus aculeatus 059859 CD Yes
158 Patent 6197564 14480380 CD Yes
159 Aspergillus kawachii P33559 CD Yes
160 Penicillium chrysogenum 46406032 CD Yes
161 Penicillium chrysogenum 83416731 CD Yes
162 Penicillium chrysogenum P29417 CD Yes
163 Rhizobium etli CFN 42 86282913 CD Yes
164 Rhizobium leguminosarum bv. trifolii 88657052 CD Yes
165 Anabaena variabilis ATCC 29413 75701321 CD No
166 Nostoc sp. PCC 7120 Q8YNW3 CD No
167 Pseudomonas syringae pv. phaseolicola 1448A 71555629 CD Yes
168 Pseudomonas syringae pv. syringae B728a 63258442 CD Yes
S 169 Caulobacter crescentus CB15 16127035 CD TAT
170 Synechococcus elongatus PCC 7942 81169090 CD Yes
171 Synechococcus elongatus PCC 6301 56685123 CD Yes
172 Acidobacterium capsulatum 13591553 CD No
173 Agrobacterium tumefaciens str. C58 17740854 CD Yes
174 Agrobacterium tumefaciens str. C58 15157542 CD Yes
175 Bradyrhizobiumjaponicum USDA 110 27377352 CD TAT
176 Rhodopseudomonas palustris BisB18 90104203 CD Yes
177 Xanthomonas oryzae pv. oryzae KACC10331 58428646 CD No
178 Xanthomonas oryzae pv. oryzae MAFF 311018 84369769 CD No
179 Xanthomonas oryzae pv. oryzae Q9AM29 CD No
180 Xanthomonas campestris pv. vesicatoria str. 85-10 78038346 CD Yes
181 Xanthomonas axonopodis pv. citri str. 306 21110692 CD Yes
182 Xanthomonas campestris pv. campestris str. 8004 66575838 CD Yes

Table 3-2. Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some of their properties. Continued.

Organism Accession Module Architecture Secreted
Number

183
184
185
186
187
188
189
190
191
192
193
194
195
196
S 197
) 198
199
200
201
202
203
204
205
206
207
208
209
210

Xanthomonas campestris pv. campestris str.
Xanthomonas campestris pv. vesicatoria str. 85-10
Xanthomonas oryzae pv. oryzae KACC10331
Xanthomonas oryzae pv. oryzae
Xanthomonas axonopodis pv. citri str. 306
Xanthomonas oryzae pv. oryzae MAFF 311018
Rhodothermus marinus
Fibrobacter succinogenes S85
Fibrobacter succinogenes S85
Fibrobacter succinogenes S85
Cellvibrio japonicus
Saccharophagus degradans 2-40
Pseudomonas sp. ND137
Saccharophagus degradans 2-40
Clostridium acetobutylicum ATCC 824
Clostridium acetobutylicum ATCC 824
Cytophaga hutchinsonii ATCC 33406
Cytophaga hutchinsonii ATCC 33406
Colwellia psychrerythraea 34H
Colwellia psychrerythraea 34H
Pseudomonas sp. PE2
Saccharophagus degradans 2-40
Cytophaga hutchinsonii ATCC 33406
Rhodopirellula baltica SH 1
Arabidopsis thaliana
Arabidopsis thaliana
Arabidopsis thaliana
Medicago truncatula

21115397
78038344
58428645
12658424
21110690
84369768
P96988
11526752
9965987
9965986
45524
89952852
57999823
89949430
15004819
15004757
110280325
110281120
71145380
71143508
25137524
89949572
110281182
32446276
081754
081751
081752
92868656

CD
CD
CD
CD
CBM4(2)CD
CD/CBM6
CD/CBM6t
CD/CBM6
CBM2/Sr/CBM6/Sr/CD
GH43t/CBM6/Sr/CBM2/Sr/CBM22/CD
Sr/CD
CD/Sr(2)/ChtBD3
CD
CD
CD/CBM22
CD/CBM9
PEr/CD/Cad/DUF823
CD
PEr/CD
PEr/CD
CBM22/CD
CD
CBM22/CD
CBM22/CD
tCD/CBM22/CD
CBM22/CD

Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
ND
Yes
No
ND
Yes
Yes
Yes
No
No
Yes

Table 3-2. Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some of their properties. Continued.

Organism Accession Module Architecture Secreted
Number

211
212
213
214
215
216
217
218
219
220
221
222
223
S 224
225
226
227
228
229
230
231
232
233
234
235
236

Triticum aestivum
Oryza sativa
Oryza sativa (japonica cultivar-group)
Zea mays
Carica papaya
Arabidopsis thaliana
Arabidopsis thaliana
Arabidopsis thaliana
Arabidopsis thaliana
Thermosynechococcus elongatus BP- 1
Bacillus pumilus
Clostridium thermocellum
Ampullaria crossean
Oryza sativa (japonica cultivar-group)
Nicotiana tabacum
Nicotiana tabacum
Populus tremula x Populus tremuloides
Arabidopsis thaliana
Arabidopsis thaliana
Arabidopsis thaliana
Oryza sativa (japonica cultivar-group)
Oryza sativa (japonica cultivar-group)
Oryza sativa (japonica cultivar-group)
Oryza sativa (japonica cultivar-group)
Hordeum vulgare
Hordeum vulgare subsp. vulgare

40363757
19920133
Q7XFF8
Q9ZTB8
Q8GTJ2
Q9ZVK8
081897
082111
Q9SZP3
22295628
20386142
37651955
Q7Z1V6
28411931
73624749
73624751
60656567
Q9SM08
Q9SYE3
080596
29788834
15528604
55168219
15528602
P93186
71142590

CD
CBM22/CD
CD
CD
CBM22/CD
CD
CD
CD
CD
CD
CD
CBM22/CD
CD
CBM22(4)/CD
CBM22(3)/CD
CBM22(3)/CD
CBM22(3)/CD
CBM22(3)/CD
tCBM22/CD
CBM22(4)/CD
CBM22/tCBM22/CD
CBM22/CD
CD
CD
CD
CBM22/CD

Table 3-2. Glycosyl hydrolase family 10 sequences included in phylogenetic studies and some of their properties. Continued.

Organism Accession Module Architecture Secreted
Number
237 Hordeum vulgare 71142588 CBM22/CD No
238 Triticum aestivum (bread wheat) Q9XGT8 CD Yes
239 Hordeum vulgare P93185 CD No
240 Hordeum vulgare 14861199 CBM22/CD No
241 Hordeum vulgare subsp. vulgare 71142586 CBM22/CD No

CD refers to a GH 10 catalytic module, Aes refers to esterase/lipase module, Cel5 refers to a GH 5 cellulase module, Deac refers to
deacetylase domain, ChtBD3 refers to chitin-binding domain, Est refers to esterase, Cad refers to Cadherin repeat domain, DUF823
refers to Salmonella repeat of unknown function, Nr refers to asparagine rich domain, Sr refers to serine rich domain, PEr refers to
proline glutamate rich domain, PTr refers to proline threonine rich domain, PGr refers to proline glycine rich domain, XSKTLPGG
refers to unique Neocallimastix linker, GH 62 refers to arabinofuranosidase domain, ChNT refers to chitinase N-terminal domain, Bph
refers to bacterial phosphatase, t refers to a predicted truncation and is positioned to the side of the truncation.

Xanthomonas

Figure 3-3. Phylogenetic distribution of catalytic domains of glycosyl hydrolase family 10 xylanases. Numbering corresponds to
Table 3-2. Full xylanase sequences were aligned in MEGA 3.1 and the highly conserved catalytic domain was trimmed
out. These sequences were realigned and used to generate a Neighbor-joining Bootstrap phylogenetic tree.

CHAPTER 4
Paenibacillus SPECIES STRAIN JDR-2 AND XynAi: A NOVEL SYSTEM FOR
GLUCURONOXYLAN UTILIZATION

Introduction

A version of this chapter has previously been published as a peer reviewed manuscript in

the Febuary 2006 issue of Applied and Environmental Microbiology.

Increasing cost and demand of fossil fuels highlights the need to develop efficient

methods to utilize renewable resources for conversion to alternative energy sources such as fuel

ethanol (Aldhous, 2005; Sun and Cheng, 2002). Supplementing the energy infrastructure with

ethanol may help to shift economic dependence from petroleum-based energy. Microbial

biocatalysts, both yeast and bacteria, have been developed for the conversion of glucose derived

from cellulose and pentoses from hemicellulose to ethanol (Dien et al., 2003; Ingram et al., 1999;

Jeffries and Jin, 2004; Jin et al., 2004), and similar approaches with bacteria have been

successfully applied to the formation of value-added products such as optically pure lactic acid

(Dien et al., 2002; Zhou et al., 2003a; Zhou et al., 2003b). Current research efforts are directed

at improving the pretreatment processes to maximize the release of fermentable pentoses as well

as glucose, and also further development of the biocatalysts for specific fermentations of these

sugars. The adoption of microbial strategies for the efficient depolymerization and assimilation

of hemicellulose derived carbohydrates offer promise for maximizing the conversion of the

hemicellulose fraction of lignocellulosic biomass to alternative fuels and biobased products

(Preston et al., 2003).

Hemicellulose represents 20% to 30% of lignocellulosic biomass and

methylglucuronoxylan (MeGAXn) is the predominant form of hemicellulose found in hardwood

and crop residues (Preston et al., 2003; Singh et al., 2003). This polymer consists of P-1,4 linked

xylan in which 10% to 20% of the xylose residues are periodically substituted with a-1,2-4-O-

methylglucuronic acid (MeGA) moieties. Complete enzymatic hydrolysis of MeGAXn requires

the combined action of several families of glycosyl hydrolases, including 3-1,4-endoxylanase, a-

1,2-4-O-methylglucuronidase and 3-1,4-xylosidase (Preston et al., 2003). Secreted microbial

xylanases that catalyze the depolymerization of MeGAXn are primarily represented by two

families of glycosyl hydrolase, GH 10 and GH 11, based on sequence similarity and hydrophobic

cluster analysis (http://afmb.cnrs-mrs.fr/CAZY), (Gilkes et al., 1991; Henrissat and Bairoch,

1993; Henrissat and Davies, 1997).

In bacteria capable of utilizing MeGAXn, the metabolism of the aldouronates generated

by enzyme-catalyzed depolymerization is dependent on their assimilation and cleavage of the

MeGA substitution. Most substrate and structural studies of a-glucuronidases, the enzymes

required to initiate complete degradation of MeGA substituted xylooligosaccharides, have

clearly established that only aldouronates in which MeGA is linked to the nonreducing terminal

xylose are suitable substrates (Nagy et al., 2002; Nurizzo et al., 2002). This distinguishes the

role of GH 10 xylanases from GH 11 xylanases in generating products for direct assimilation and

metabolism. This argument is further supported by evidence that aldotetrauronate acts as a

catabolic signaling molecule for its further metabolism (Shulami et al., 1999). Studies of the

glucuronic acid utilization gene cluster of Geobacillus %e, Itilh' mniV,,/hiihi have identified a

putative MeGAX3 transporter in an operon composed of genes involved with the degradative and

catabolic processing of glucuronoxylan. The uxuR gene product, a DNA binding protein, was

found to be a self-regulating element of this operon that acts to repress transcription. Binding of

MeGAX3 by UxuR alleviates repression. From this, it appears that GH 10 xylanases play a

prominent role, both directly and indirectly, in processing of MeGAXn for its complete

catabolism. There is no evidence to support a similar role for the GH 11 xylanases. It is possible

that GH 11 xylanases act to hydrolyze polymeric xylan primarily into shorter fragments that can

then be further acted upon by GH 10 xylanases and P-xylosidases. (Pell et al., 2004a).

Another factor affecting the efficiency of metabolism is the localization of the xylanase

relative to the cell. The cellulosome, found primarily in anaerobic Clostridium spp. and some

ruminant microorganisms, and the xylanosome from aerobic soil bacteria, often have associated

GH 10 xylanases (Bayer et al., 2004; Doi et al., 2003; Jiang et al., 2004). These extracellular

surface anchored complexes often display a variety of enzymes from several glycosyl hydrolase

families with diverse functions. Clostridium thermocellum has a well described cellulosome

with twenty-six glycosyl hydrolases (62% cellulases, 23% xylanases, 15% other) and several

associated esterase activities that contribute to hydrolysis of the lignocellulose complex (Doi and

Kosugi, 2004). Localization of this complex to the surface of the organism presumably allows

efficient utilization of hydrolysis products, which may provide a competitive advantage in an

anaerobic niche. Extracellular GH 10 xylanases may also occur as large multi-modular surface-

anchored enzymes separate from other glycosyl hydrolases. Representative GH 10 xylanases

from Clostridium, Thermoanaerobacterium, Caldicellulosirupter, Thermotoga,

Promicromonospora, Paenibacillus and several other genera have been shown to have similar

modular architectures.

Here we describe the properties of an extra-cellular multidomain endoxylanase from an

aggressively xylanolytic Paenibacillus sp. (strain JDR-2). The association of XynAi with cell

wall preparations indicates an anchoring role for the SLH domains near the C-terminus of the

155 kDa enzyme. The marked preference of this organism for polymeric MeGAXn as a growth

substrate compared to xylose or the aldouronates generated by the action of the GH 10

endoxylanase, supports a role for this enzyme in the vectoral processing of MeGAXn for

subsequent transport and metabolism.

Materials and Methods

Isolation and identification ofPaenibacillus sp. strain JDR-2. Paenibacillus sp. strain

JDR-2 was isolated from fresh cut discs (5 cm diameter by 2-4 mm thick) of sweetgum stem

wood (Liquidamber styraciflua) incubated about one inch below the soil surface in a sweetgum

stand for approximately three weeks. Discs were suspended in 50 ml sterile deionized water and

sonicated in a 125 Watt Branson Ultrasonic Cleaner water bath for 10 min. The sonicate was

inoculated into 0.2% (w/v) sweetgum (SG) MeGAXn containing the mineral salts of Zucker and

Hankin (Zucker and Hankin, 1970) at pH 7.4 and incubated at 30 OC. The SG MeGAXn was

prepared and characterized by 13C-NMR as described previously (Hurlbert and Preston, 2001;

Jones et al., 1961; Kardosova et al., 1998). Isolated colonies were passed several times in

MeGAX1 broths and agars until pure. A culture growing on 0.2% MeGAXn Z-H medium was

cryostored by mixing 0.5 ml of exponentially growing culture with 0.5 ml 50% (v/v) sterile

glycerol and freezing at -70 OC. The purified isolate was submitted to MIDI Labs

(http://www.midilabs.com) for partial 16s rRNA sequencing. The organism was identified as

Paenibacillus sp. with 96% identity to Paenibacillus granivorans by blastn submission of 530

nucleotides of sequenced 16s rRNA. The organism has been deposited with the Bacillus Genetic

Stock Center (BGSC), (http://www.bgsc.org).

Growth studies. A common protocol was applied in the maintenance and analysis of

Paenibacillus sp. strain JDR-2 cultures. Each time a culture was prepared for study, a sample

from the cryostored stock culture was transferred into 4 ml of 0.5% SG MeGAXn Z-H medium

in 16 x 100 mm test tubes. After 36 to 48 hours of growth the culture was plated on agar

medium containing 1.0% yeast extract (YE) and 0.5% oat spelt xylan in Z-H and grown for 36 to

48 hours until appropriately sized colonies were observed. In a slight deviation, inoculum from

the cryostored culture was plated directly onto the agar medium and grown for 48 to 72 hrs

before picking an isolated colony. All colonies regularly displayed the expected phenotype, i.e. a

clearing zone on the opaque oat spelt xylan background with the expected colony morphology.

For the various growth studies described below a single colony was inoculated into medium

specified for the particular experiment. All growth was performed at 300C.

Growth optimization studies were performed aerobically in 16 x 100 mm test tubes

containing 4 ml volumes of medium and optical densities of cultures were measured at 600 nm

with a Beckman DU500 series spectrophotometer with a 16 x 100 mm test tube holder.

Individual 4 ml cultures for study were inoculated with 200 [pl (5% volume) of an exponentially

growing culture (4 ml medium of 1.0% YE in Z-H). For these test tube cultures, agitation was

achieved by setting a test tube rack in a large flask holder on a New Brunswick G-2 gyrotory

shaker at an angle of approximately 45. Under these conditions, rotation at 200 rpm yielded the

best agitation when compared to simple rotation.

Studies comparing Paenibacillus sp. strain JDR-2 utilization of MeGAXn, with or without

xylose or glucose as co-substrates, were performed in 125 ml baffle flasks with shaking at 150

rpm on a G-2 gyrotory shaker. Cultures were initiated by the addition of 4 ml (8% volume) of Z-

H mineral salts washed cells from an overnight culture (25 ml) of 1.0% YE Z-H medium.

Growth was monitored using an HP Diode Array spectrophotometer at 600 nm in a 1.00 cm

cuvette. For these cultures, sample dilutions were performed to obtain OD 600 nm readings

between 0.2 and 0.8 absorbance units and the resulting value was corrected by the dilution factor.

Culture aliquots were centrifuged, supernatants filtered and carbohydrate utilization was

measured by HPLC using a complete modular Waters chromatography system comprised of a

600 controller, 610 solvent delivery unit, 2410 RI detector and a 710B WISP automated injector.

Carbohydrate separation was achieved with a Bio-Rad HPX-87H column running in 0.01 N

H2SO4 with a flow rate of 0.8 ml/min at 65 C. Data analysis was performed using Waters

Millennium Software.

The differential utilization of MeGAXn and XynAi CD generated products from MeGAXn

as growth substrates by the organism was evaluated by the initiation of 50 ml cultures with 4 ml

(8% volume) of Z-H washed cells from 25 ml overnight cultures in 0.5% SG MeGAXn Z-H

medium. Growth was monitored as described above and aliquots examined by TLC (see

procedure below).

DNA cloning, sequencing and analysis. A genomic library of Paenibacillus sp. strain

JDR-2 DNA, prepared in pUC 18 with gel purified 6-9 kb fragments obtained from a partial

Sau3AI digest, was kindly provided by Ms. Loraine Yomano from the laboratory of Professor

Lonnie Ingram. All cloning and general DNA manipulation methods originate from Molecular

Cloning: A Laboratory Manual (Sambrook et al., 1989). In addition, DNA purification and gel

extraction was performed using kits purchased from Qiagen (Valencia, California). Cloning

analysis, planning and image preparation was performed with Clone Manager 6 and Enhance

(Scientific and Educational Software, Cary, NC). Analysis of sequences for regulatory elements

was conducted using the online tools available through Softberry

(http://www.softberry.com/berry.phtml). The pUC18-based 6-9 kb library was transformed into

E. coli DH5ca and screened for xylanase positive clones by plating transformed cells on Remazol

Brilliant Blue xylan plates and observing agar clearing after 24 hours (Braun and Rodrigues,

1993). Sequencing of cloned DNA was done in house by subcloning the insert into smaller sizes

and using pUC18 M13 priming sites for sequencing of both strands. Primer walking at the ICBR

Genome Sequencing Services Laboratory at the University of Florida filled in gaps and

completed 2x coverage. All sequencing employed the Sanger dideoxy chain termination method.

The final sequence was assembled using the CAP3 sequence assembly program (Huang and

Madan, 1999) located on the P81e Bio-Informatique Lyonnais server (http://pbil.univ-lyonl.fr/).

Sequence analysis was performed with online resources available through the NCBI

(http://www.ncbi.nlm.nih.gov) and BCM (http://searchlauncher.bcm.tmc.edu) websites. The

main tools employed were BLAST and CD-Search of the CDD (Conserved Domain Database)

(Marchler-Bauer et al., 2003) on the NCBI site and the 6 Frame Translation and Readseq utility

at the BCM site.

Phylogenetic Analysis of Paenibacillus sp. strain JDR-2 XynAi. All presented

phylogenetic analyses resulted from sequences that had been trimmed to contain only the highly

conserved catalytic domain from the proton donor (WDVVNE) to the catalytic nucleophile

(ITELDI). These sequences were aligned using Clustalx and phylogenetic trees constructed

using MEGA 2.1 (Molecular Evolutionary Genetics Analysis, Version 2.1.; Kumar et al., 2002).

The domain arrangement of the whole xylanase was determined with CDD (Marchler-Bauer et

al., 2003) at NCBI (http://www.ncbi.nih.gov/Structure/cdd/wrpsb.cgi). Signal sequences were

analyzed by the on-line program Signal-P (Bendtsen et al., 2004)

(http://www.cbs.dtu.dk/services/SignalP). Eighty-four bacterial GH 10 xylanases were

downloaded from the CAZy(ModO) database (http://afmb.cnrs-mrs.fr/CAZY/) and processed as

described above. This processing ensured the strictest comparison between all the bacterial GH

10 xylanases. Four xylanases showed high similarity to Paenibacillus sp. strain JDR-2 XynAi.

These and eleven randomly chosen sequences from the eighty set were presented in figures for

this chapter.

Carbohydrate and protein assays. Total carbohydrate concentrations related to substrate

preparations and enzymatic kinetic analysis were determined by the phenol-sulfuric acid assay

(Dubois et al., 1956). In conjunction with the total carbohydrate assay, measurements to define

the degree of polymerization of substrate and increased reducing terminus levels due to

xylanolytic activities were performed by the method of Nelson (Nelson, 1944). Xylose was used

as the reference for both assays. Protein levels were determined using Bradford assay reagents

(BioRad) with BSA (Fraction V) as the standard (Bradford, 1976).

XynA1 CD cloning, overexpression and purification. The expression vector pET15b+

(Novagen, San Diego, CA) was used to overexpress the catalytic domain of XynA1 independent

of other modules. Primers were designed to delimit the CD based on the modular endpoints

identified by Pfam (Bateman et al., 2004). The forward primer (5'aagcatggctccactcaaa)

included an Ndel site for in-frame fusion with the His-Tag sequence, and the reverse primer

(5'tgtgctcagccggatcaat) contained a Blpl (Bpu1021) site for directional cloning into pET15b+.

This primer selection method added a Gly, Ser, His, Met sequence to the N-terminus just prior to

the beginning of the pfam designated sequence. This additional sequence was derived from the

pET15 expression ORF coding sequence. There was additional sequence corresponding to Ala,

Glu, Gln at the C-terminal end resulting from vector-derived sequence just upstream from the

vector-encoded stop codon. The PCR product was generated using Proof Start high fidelity PCR

(Qiagen, Valencia, CA). The construct was verified by sequencing. For expression, the

pET15XynA1 CD N-terminal His construct was transformed into E. coli Rosetta (DE3)

chemically competent cells, and grown with selection (see below) for about 24 hours. A single

colony was picked and inoculated into 50 ml LB containing 50 [g/ml ampicillin and 34 [g/ml

chloramphenicol, grown at 37 C to an OD 600 nm of about 0.6 to 0.8, and the entire culture

centrifuged for 10 minutes at 5000 x g at 350 C. The pellet was resuspended in 100 ml LB

containing 100 [g/ml ampicillin and 34 [g/ml chloramphenicol and grown for about 1 hr at

370C. Cells were harvested as above, resuspended in 10 ml LB medium and used to inoculate a

1 L LB batch culture (preequilibrated at 370C) containing antibiotics as above. This was grown

at 37 C to an OD 600 nm of 0.6 to 0.8 and overexpression induced by the addition of IPTG to

1.0 mM final concentration, and incubation continued for no more than 3 hours before cells were

harvested. Cell pellets representing the growth of 1 liter of culture were suspended in 35 ml of

20 mM sodium phosphate buffer, pH 7.0 and lysed at 16000 psi with a single pass through a

French pressure cell. The total volume was estimated to be 37.5 ml and 1 M MgCl2 was added

to a final concentration of 1.5 mM to obtain optimal Benzonase (Novagen, San Diego,

California) activity for hydrolysis of nucleic acid. Benzonase was added at 8 units/ml and the

cell lysate was incubated with gentle mixing at room temperature for about 45 minutes. The

crude cell lysate was centrifuged at 40C for 30 minutes at 35000 x g, the supernatant filtered

through a 0.45 um syringe tip filter, and 5.0 M NaCl added to a final concentration of 0.50 M.

Ten ml aliquots of the cell-free extract were affinity-purified using the HiTrap Chelating HP

column procedure (Amersham Biosciences, Piscataway, NJ). Each loaded volume yielded a

single expected band detected with Coomassie Blue (CB) following SDS-PAGE analysis.

Removal of the N-Terminal His-Tag was accomplished using the thrombin cleavage capture kit

available through Novagen. XynAi CD and XynAi CD N-terminal His were stored for short

periods of time at 40C in 50 mM potassium phosphate buffer, pH 6.5. For longer storage, these

stocks were split with equal volumes of glycerol and stored at -20 OC. Enzyme analysis by

activity measurement and protein profiles following SDS-PAGE staining with CB were the same

after 6 months storage at -20 C.

Xylanase activity measurements for enzyme optimization and kinetic analysis. The

temperature optimum for XynAi CD xylanase activity was determined by incubation in 0.25 ml

reaction mixes containing 1.0% SG MeGAXn in 0.1 M potassium phosphate, pH 7.0 for 10 to 30

minutes over a 40 C to 60 OC range. Reactions were halted by the addition of 0.25 ml Nelson's

A:B reagent (25:1) (v:v) and the increase in reducing termini determined (Nelson, 1944). The

resulting temperature optimum (45 C) was subsequently used to determine the optimal activity

for the enzyme over the pH range from 5.5 to 7.0 in reaction mixes containing 1.0% SG

MeGAXn in 0.1 M potassium phosphate. The optimal conditions from these determinations (pH

6.5 at 450C) were used in experiments to examine the reaction kinetics of the enzyme with SG

MeGAXn as substrate. Activity units are described as the amount of enzyme producing 1 [tmole

of reducing termini per minute at 450C. Production rates were linear through 30 minutes and data

obtained are averages of 3 separate experiments performed in triplicate.

Chromatographic resolution and detection of aldouronates and xylooligosaccharides.

Standards were obtained by acid and enzymatic hydrolysis of SG MeGAXn. Aldouronate

oligomers, MeGAX1 through MeGAX5, were prepared by acid hydrolysis of MeGAXn in 0.1 N

H2SO4 at 121 C for 60 min. The acid hydrolysate was neutralized with BaCO3 and the

aldouronates adsorbed onto Bio-Rad AG2-X8 anion exchange resin in the acetate form. Xylose

and xylooligosaccharides were eluted with water, and the aldouronates were then eluted with

20% acetic acid. After concentration by flash evaporation, aldouronates were fractionated with

50 mM formic acid eluent on a 2.5 cm x 160 cm BioGel P-2 column (BioRad, Hercules, CA)

equilibrated in the same buffer. Identities of MeGAX1 and MeGAX2 were confirmed by 13C and

1H-NMR spectrometry (K. Hasona, unpublished data). Identities of MeGAX3, MeGAX4 and

MeGAX5 are based upon the elution profile from the BioGel P-2 column and TLC analysis of

aldouronates resulting from GH 10 and GH 11 catalyzed MeGAXn hydrolysis. Xylobiose and

xylotriose were generated by hydrolysis with a GH 11 xylanase, XynII of Trichoderma

longibrachiatum (Hampton Research, Aliso Viejo, CA), and fractionated using water based

BioGel P-2 column chromatography. These methods allowed the isolation of X2, X3 and the

aldouronates MeGAX1 through MeGAX5.

To follow the depolymerization of MeGAXn catalyzed by XynA1 CD, a 250 [tl reaction

containing 0.5 units of enzyme and 5 mg SG MeGAXn in potassium phosphate buffer, pH 6.5,

was incubated at 30 OC. Samples (5 IL) were removed every 10 min up to 120 min, and spotted

on 20 x 20 cm pre-coated 0.25 mm Silica Gel 60 TLC plates (EM Reagents Darmstadt,

Germany). An additional 0.5 units of XynAi CD was added after the initial 120 minutes and

incubation was continued for an additional 16 h. A 5 [tl sample representing the reaction limit

products was also spotted on the plate. Oligomers were separated by ascension with a solvent

system containing chloroform: glacial acetic acid: water (6:7:1) (v:v:v) two times for 4 hours

each with at least 1 hour of drying time between each solvent presentation. After the second

development the plate was allowed to dry for at least 30 minutes and then sprayed with 6.5 mM

N-(1-naphthyl)ethylenediamine dihydrochloride in methanol containing 3% sulfuric acid with

subsequent heating to detect the carbohydrates (Bounias, 1980).

To compare the ability of Paenibacillus sp. strain JDR-2 to utilize MeGAXn, aldouronates

and xylooligosaccharides, MeGAXn was digested with XynA1 CD to generate primarily X2, X3,

and MeGAX3. For digestions with XynAi CD, 50 ml of substrate containing 30 mg/ml SG

MeGAXn were prepared with 10 mM sodium phosphate buffer, pH 6.5. Digestions were

initiated by addition of 3.5 units XynA1 CD and incubated with rocking at 30 OC for 24 h. An

additional 1 unit was added after 24 hours and incubation was continued for 40 h.

Digests were processed by stir-cell filtration under nitrogen pressure through YM-3

ultrafiltration membranes (3 kDa MWCO) (Millipore, Billerica, MA). The filtrates containing

oligomers with molecular weights less than 3000 Da were concentrated by flash evaporation and

analyzed by the phenol-sulfuric acid assay (Dubois et al., 1956). These concentrated fractions

then served as substrates comparing the growth rates and yield of Paenibacillus sp. strain JDR-2

on MeGAXn. Cultures were incubated at 30 OC in baffle flasks containing 50 ml medium

supplemented with 10 mg/ml of anhydroxylose equivalents (determined by the total

carbohydrate assay). Growth was followed by measuring OD 600 nm. Samples of 250 pll were

removed at selected times, centrifuged at maximum speed in a microfuge and the supernatant

and pellet were separated and saved. The supernatant was incubated at 70 OC for 10 minutes

prior to storage. A volume of 6 pll was used for each sample on the TLC plate. TLC plates were

developed as described above.

Immunolocalization of XynAi. Sodium dodecyl sulfate-polyacrylamide gel

electrophoresis (SDS-PAGE) was performed according to Laemmli (Laemmli, 1970) using a

MINI-PROTEAN 3 electrophoresis cell, a 12% Ready Gel and Precision Plus Dual Color pre

stained molecular weight standards (Bio-Rad Laboratories, Hercules, CA) following described

methods (Mini- PROTEAN 3 Cell Instruction Manual, Bio-Rad Laboratories, Hercules, Ca.).

Immunodetection was performed as previously described (Schmidt et al., 2003). XynAl CD was

purified to homogeneity as judged by SDS-PAGE after staining with CB. Chickens were

inoculated with XynAl CD as antigen. An amount of 100 tg was delivered in a total volume of

1 ml PBS. A volume of 700 pll was injected subcutaneously under the wing and 300 ul was

injected in the footpad. No adjuvant was used. Eggs were collected from before injection and

through the entire process. A boost injection was administered as above about two weeks post

primary injection. The eggs were screened by ELISA in groups of three as crude unprocessed

yolk in PBS. Peak fractions were pooled and chicken IgY polyclonal antibody obtained

following the method of Polson, et al (Polson et al., 1985).

Immunolocalization studies were performed with cell fractions following growth on SG

MeGAXn. Bacillus subtilis 168 was cultured as a negative control and compared with

Paenibacillus sp. strain JDR-2. Colonies of B. subtilis and Paenibacillus sp. strain JDR-2 were

suspended in 2 ml, lx Z-H and vortexed until cells were fully suspended. The complete 2 ml

volume was used to inoculate 50 ml of media in 250 ml baffle flasks containing 0.2% YE, 0.36%

SG MeGAXn in Z-H. Cultures were grown overnight (16 hr) at 300C with shaking at 150 rpm

on a New Brunswick G-2 gyrotory shaker. Cells were harvested at an OD 600 nm of 1.0

(Paenibacillus sp. strain JDR-2) and 0.7 (B. subtilis). Cultures were centrifuged at 5000 x g for

15 minutes at room temperature and the supernatant was recovered. Cell pellets were

resuspended in 50 mM sodium phosphate buffer, pH 6.5, and centrifuged as above. The

procedure was repeated with 50 mM sodium phosphate pH 6.5 containing 0.5 M NaC1. The final

cell pellet was resuspended in 5 ml, 50 mM sodium phosphate, pH 6.5. Some cell lysis of

Paenibacillus sp. strain JDR-2 was apparent, observed as increased viscosity, probably due to

osmotic shock. A volume of 50 pl Promega DNase RQ1 at 1 unit/pl was added with 1/10 the

volume of 10 x DNase RQ1 buffer (0.40 M Tris-HC1, 0.10 M MgCl2, 0.01 M CaC12, pH 8.0) and

the suspension was incubated for 30 minutes at room temperature. Cells were then lysed by two

passes at 16,000 psi through the French pressure cell. Lysates were centrifuged at 30,600 x g for

20 min at 40C. Supernatant was collected as the cell free extract, the pellet was resuspended in 1

ml, 50 mM sodium phosphate pH 6.5 and designated the cell wall suspension. All supernatants

were concentrated using YM-10 Centriprep concentrators (Millipore, Billerica, MA) to volumes

less than 4 ml. Samples of the media supernatant concentrate (MSC), NaCl wash (NaC1), cell

free extract (CFE) and cell wall suspension (CWS) were analyzed by SDS-PAGE. Reactive

antigens were detected on immunoblots using rabbit anti-chicken alkaline phosphatase conjugate

(Sigma, St. Louis, Missouri ) as previously described, and proteins were detected in gels with

CB (Schmidt et al., 2003).

Results

Growth analysis of Paenibacillus sp. strain JDR-2. Based on OD 600 nm

measurements, the initial growth analysis of Paenibacillus sp. strain JDR-2 indicated that the

organism utilized MeGAXn more efficiently compared to glucose or xylose as substrates (Figure

4-1A). More detailed studies (Figure 4-1B-D) using HPLC to follow substrate concentration

showed that MeGAXn is almost completely utilized. Additionally, Paenibacillus sp. strain JDR-

2 preferentially utilized the MeGAXn in the presence of glucose or xylose. Under these

conditions, the concentrations of glucose and xylose in the medium decreased more slowly, and

at a nearly linear rate. Figure 4-1B also shows that xylose accumulates in the medium to a small

extent during growth on MeGAXn, indicating that what is produced during the extracellular

depolymerization may not be directly assimilated.

Identification and sequencing ofxynAi encoding a secreted modular GH10

endoxylanase. Analysis of the Paenibacillus sp. strain JDR-2 chromosomal DNA library in E.

coli for xylanases led to the isolation of four clones. Restriction analysis of these clones

suggested that the inserts were from the same genomic DNA location. Plasmid pFSJ4 was

selected for sequencing which revealed an insert (Figure 4-2) including a large modular xylanase

(xynA1) of 4401 nt (1467 aa). Sequencing of the compete genomic DNA insert identified genes

flanking xynAi. In the 5' direction on the same chain there is a mdep gene encoding a putative

multi-drug efflux permease with 43% amino acid identity to the same in Bacillus halodurans

(gene = BH3482) determined by blastp. In the 3' direction on the opposite strand there is a

putative c-1,6-mannanase gene (amanA) that codes for a protein with 67% identity to Aman6

protein (aman6 gene) from Bacillus circulans. Domain analysis revealed that AmanA has the

exact modular structure of Aman6 with a GH 76 catalytic module followed by triplicate family 6

CBM. In silico sequence analysis identified a probable promoter region and rho-independent

terminator for xynA1, but only a terminator for the mdep gene and a promoter for the gene

encoding AmanA.

Much like many other glycosyl hydrolases, XynAi is a modular protein composed of 8

separate modules (Figure 4-2). The domains include a triplicate N-terminal set of CBM 22

modules which have previously been shown to bind soluble xylan and P-1,3-1,-4-glucan (Dias et

al., 2004; Xie et al., 2001). These modules are followed in sequence by a GH 10 CD and a CBM

9, which has been shown to bind to the reducing end of carbohydrate chains (Boraston et al.,

2001; Notenboom et al., 2001). Following CBM 9 is an undefined sequence with high similarity

to the same region in Xyn5, a GH 10B xylanase of Paenibacillus sp. W-61. This region, as

previously reported, has high identity to the lysine-rich region of the SdbA protein of C.

thermocellum. Xyn5 and XynAi have 36% and 35% amino acid identity, respectively, to this

region of SdbA, and this region in XynAi has 49% identity to the same in Xyn5. Although these

identities to the lysine rich-region of SdbA are relatively high, XynAi and Xyn5 contain only

about 5% and 6.5% lysine, respectively, to the same region of SdbA which has 13% lysine (data

not shown) (Ito et al., 2003; Leibovitz et al., 1997). The C-terminal region includes a triplicate

set of SLH modules which are predicted to function in surface anchoring (Cava et al., 2004;

Kosugi et al., 2002; Mesnage et al., 2000). The xynAi coding sequence has been deposited in

EMBL with the accession number AJ938162.

Phylogenetic analysis of XynAi. Initial phylogenetic analysis revealed that XynA1 and

XynA1 CD amino acid sequences, when subjected to blastp, had high bit scores to the same set

of four modular GH 10 xylanases (Figure 4-3). Comparison of the top 9 blastp hits to XynA1 of

Paenibacillus sp. strain JDR-2 shows the comparative modular structures. Additionally, the bit

scores are represented for the whole sequence blastp and the CD sequence blastp. XynAi and

the top four hits were classified as GH 10B and the lower set as GH 10A based on the number of

amino acids separating the glutamate residue functioning as the catalytic proton donor from the

glutamate functioning as the catalytic nucleophile. Although there are some exceptions, most

catalytic domains of GH 10 xylanases have about 105 amino acids separating the two catalytic

residues. In the case of sequence group GH 10B, the distance separating the catalytic residues is

about 123 amino acids (Table 4-1). For further analysis we reasoned that the catalytic residue

bridge sequence was probably the most highly conserved portion of GH 10 xylanases and

compared this sequence from many xylanases. Figure 4-4 represents a phylogenetic comparison

of the GH 10B subset to eleven randomly selected GH 10A xylanases. Table 4-1 characterizes

the modular structures for the xylanases, indicating significant diversity among those represented

in subset 10A. In Figure 4-4A the Clustal alignment of the CD region used to prepare the

phylogenetic tree revealed three areas in which GH 10B subset differs from GH 10A. The

additional sequences accounted for the extra length between the catalytic proton donor and

nucleophilic glutamate residues. A phylogenetic tree developed with the neighbor-joining

method for the alignment in Figure 4-4A shows sequences within a clade distinct from the others

(Figure 4-4B). It should be noted that this comparison set is biased to the extent that it contains

five very similar GH 10B sequences with eleven other random GH 10A sequences. However,

the presentation of data identifies a relationship that indicates that GH 10B sequences have a

common lineage. Large-scale analysis of 84 bacterial GH 10 xylanases obtained from CAZy

identified GH 10B as a subset and allowed few other subsets to be created with a > 95%

bootstrap value. Many of these sequences did not place with confidence in any subset potentially

allowing for only a few well-defined subgroups of GH 10 xylanases.

XynAi localization. Chicken polyclonal IgY generated against XynAi CD (anti-CD) was

used to examine the localization of XynAi in Paenibacillus sp. strain JDR-2 cell fractions. CB

stained SDS-PAGE bands of both Paenibacillus sp. strain JDR-2 and Bacillus subtilis 168

proteins were primarily greater than 100 kDa for all fractions (Figure 4-5A). However, anti-CD

showed reactivity with Paenibacillus sp. strain JDR-2 CWS protein (approx. 150 kDa) which

was not apparent with the CB stained gel (Figure 4-5B). The antibody reacted well with XynAi

CD with essentially no cross reactivity towards B. subtilis fractions. Size estimation of the

reactive Paenibacillus sp. strain JDR-2 CWS protein at approximately 150 kDa compares

favorably with the MW (154 kDa) obtained from the translated amino acid sequence of the

XynAi modular enzyme. The band identified as XynAi in the immunoblot is not visible in the

CB-stained gel (Figure 4-5A), indicating that XynAi represents a minor component of the

surface protein complement. What is obvious from the CB-stained gel is a band size of

approximately 80 kDa. This undoubtedly is the most prominent protein overshadowing all

others. Observing that XynAi is anchored to the surface supports the possibility that

Paenibacillus sp. strain JDR-2 produces a crystalline surface layer. The size of the prominent

band at 80 kDa is roughly the same size as the Sap and 80K surface layer proteins from Bacillus

amlani i% and Bacillus sphaericus respectively (Bowditch et al., 1989; Etienne-Toumelin et al.,

1995).

Kinetic and product analysis of XynAi CD. XynAi CD was overexpressed in pET 15b+

and affinity purified using an N-terminus His-Tag. Removal of the affinity tag by thrombin

protease treatment resulted in an increased activity against SG MeGAXn of approximately 50%.

Initial characterization showed XynAi CD to have an optimal pH and temperature of 6.5 and

450C, respectively (data not shown). Kinetic analysis with SG MeGAXn as substrate (Figure 4-

6) showed XynAi CD to have Vmax and Km values of 8 units/mg and 1.96 mg/ml, respectively,

and a kcat of 306.8 /min. Analysis of products by TLC (Figure 4-7) showed that XynAi CD is a

typical GH 10 xylanase hydrolyzing MeGAXn primarily to X2 and MeGAX3. Small amounts of

X3 and MeGAX4 were also produced. True limit products of the reaction included xylose, which

built up from the seemingly slow conversion of X3 and MeGAX4 to X2 and MeGAX3. Thirty

minutes after reaction initiation, X2 and MeGAX3 are the predominant products (Figure 4-7).

The small amounts of X3 and GAX4 disappeared by 24 hr.

Paenibacillus sp. strain JDR-2 utilization of aldouronates and xylooligosaccharides in

comparison to MeGAXn. Xylooligosaccharides and aldouronates were generated by hydrolysis

of SG MeGAXn with XynAi CD. A 3 kDa molecular weight cutoff ultrafiltration filtrate product

was used to evaluate growth of Paenibacillus sp. strain JDR-2 on xylanase generated

aldouronates and xylooligosaccharides. Growth was compared for SG MeGAXn and XynAi CD

filtrate. Data presented in Figure 4-8 shows the aldouronates and xylooligosaccharides resolved

by TLC during the growth of Paenibacillus sp. strain JDR-2. Through the time course for

growth on SG MeGAXn, neither aldouronates nor xylooligosaccharides were detected in the

media during exponential growth. In contrast to growth observed on the XynAi CD-generated

products, higher growth rates and yields were observed with MeGAXn as substrate, indicating

preferred utilization of polymeric glucuronoxylan compared to aldouronates and

xylooligosaccharides generated by the in vitro XynAi CD-catalyzed depolymerization of

MeGAXn.

Discussion

Based upon growth and substrate utilization analysis, Paenibacillus sp. strain JDR-2 has

been shown to more efficiently utilize the biomass polymer MeGAXn compared to simple sugars

such as glucose and xylose. In addition, growth on MeGAXn with competing simple sugars does

not seem to affect its utilization of MeGAXn (Figure 4-1). This observation stands in contrast to

a similar xylanolytic system from Paenibacillus sp. W-61 in which the investigators found that

glucose strongly repressed xylanase activity (Viet et al., 1991). Although there appear to be

metabolic differences, Paenibacillus sp. W-61 produces Xyn5, a GH 10B xylanase that is the top

blastp hit of XynAi. With 51% identity the two full sequences are very similar with Xyn5

differing; i.e., having only 2 CMB 22 modules rather than three. Kinetic properties of the two

xylanases are similar but the generation of aldouronates by Paenibacillus sp. W-61 was not

determined, precluding a comparison to XynAi secreted by Paenibacillus sp. strain JDR-2 (Ito et

al., 2003).

Even though Paenibacillus sp. strain JDR-2 utilizes MeGAXn very efficiently, it is

probable that XynAi is the only extracellular xylanase responsible for this ability. Genomic

library screening led to the isolation of four xylanolytic clones with identical restriction profiles,

each containing the same xynA] coding sequence. Only one other xylanase gene has been

identified from this organism during intensive cosmid library screening in E. coli, and this

encodes a 40 kDa GH 10 catalytic domain designated XynA2. The primary amino acid sequence

for XynA2 does not have a detectable secretion signal sequence and is expected to be localized to

the cytosol. The xynA2 gene sequence is located within an operon including aguA, encoding a

GH 67 a-glucuronidase, and encodes a GH 10 xylanase that may be involved in the intracellular

processing of aldouronates and xylooligosaccharides generated by the action of XynAi on the

cell surface (G. Nong, V. Chow, J. D. Rice, F. St. John, J. F. Preston, Abstr. 105th ASM General

Meeting, abstr.O-055, 2005). MeGAX3, the primary aldouronate limit product of GH 10

xylanases (Biely et al., 1997), is presumably efficiently assimilated as it has been identified as an

inducer for genes involved in hydrolysis and catabolism of glucuronoxylan in Geobacillus

stearothermophilus (Shulami et al., 1999).

Phylogenetic characterization of XynAi placed the sequence with a highly similar set of

GH 10 xylanases referred to in this paper as the GH 10B subset (Figure 4-3). This classification

is supported by differences observed in the CD coding sequence. Specifically, the area bridging

the two catalytic glutamate residues contains three areas of additional sequence that are not

observed in other xylanases. Although GH 10B xylanases are modular, there are many similarly

modular xylanases that may not be classified as GH 10B. This suggests a unique mode of action

for GH 10B xylanases. It is interesting to note that this subset includes GH 10 xylanases in

anaerobic Clostridium spp. and aerobic Paenibacillus spp., all of which are found in soil

environments. The common modular architecture (Figures 4-3 and 4-4) that, in this case,

includes anchoring motifs, suggests a positive role in niche development of these bacteria. The

variability in the number of CBM and SLH modules suggests these may be mobile elements that

may be combined from different genes during evolution.

XynAi CD analysis identified XynAi as a typical GH 10 endoxylanase, producing

primarily X2, X3 and MeGAX3 in the early stages of the reaction (Figure 4-7) (Biely et al., 1997;

Preston et al., 2003). Hydrolysis seemed to proceed in two stages. The first stage resulted in

formation of X2 and MeGAX3 with small amounts xylose, X3 and MeGAX4. The second stage

included the slow conversion of the minor products to X2 and MeGAX3 with increased formation

of xylose. The XynA1 CD Km for SG MeGAXn was within a comparable range to other GH 10

xylanases but the rate of catalysis was significantly lower than that found in other reports (Figure

4-6). This may in part be due to the special attention given to xylanases showing high activity.

Wild type purification of Xyn5 from Paenibacillus sp. W-61 yielded an enzyme showing a

similar specific activity as XynAi. (Roy et al., 2000).

Xylan binding subsite analysis revealed that XynA1 contains a CD that has four well-

conserved subsites. Subsites -2 through +2 are highly conserved in GH 10 xylanases and XynA1

is no exception. In addition, by alignment analysis (data not included), XynA1 does not appear

to have a +3 subsite and subsites -3 and +4 do not exist as defined in some other GH 10

xylanases (Charnock et al., 1998; Pell et al., 2004a). Analysis of these subsites as they may

impact the catalysis of MeGAXn seems to support the results of product analysis by TLC. A

xylanase with strong binding -2 through +2 subsites should yield X2 as a primary product of

MeGAXn hydrolysis. Accumulation of xylose and small odd numbered xylooligomers (X, X3)

would only result from processing through odd numbered oligomers such as X5 and X7.

Structural studies have also identified a glucuronic acid pocket in the +1 subsite that would

facilitate hydrolysis of MeGAXn, leaving the MeGA substitution on the nonreducing terminus

(Pell et al., 2004b). Since MeGAX3 is a primary limit product of GH 10 xylanases, it follows

that large xylooligomers containing MeGA as a nonreducing end substituted residue can only be

further processed by positioning into the -3 subsite yielding MeGAX3 (Fujimoto et al., 2004).

XynAi is the largest GH 10 xylanase so far identified from a Paenibacillus sp. The net

modular architecture is similar to other Paenibacillus sp.; however, the triplicate N-terminal

CBM 22 is unique to Paenibacillus sp. strain JDR-2. Although at least one other bacterial GH

10 xylanase has been identified with a triplicate set of N-terminal CBM 22 modules (XynB from

Caldicellulosiruptor sp. Rt69B. 1), it classifies as a GH 10A subset member according to this

analysis (Figure 4-3) (Morris et al., 1999). Published reports that consider the role of

carbohydrate binding modules with respect to the function of the catalytic domain suggest that

these modules accentuate activity by increasing the localized substrate concentration (Boraston

et al., 2004). It is difficult to imagine hydrolysis of MeGAXn by XynAi being potentiated by the

addition of a third CBM 22 module. Two publications concerned with a set of CBM 22 modules

(not necessarily in tandem) from different xylanases have shown that while one seems to

function to bind a potential carbohydrate substrate, the other does not (Charnock et al., 2000;

Meissner et al., 2000). Additionally, in some cases these modules have been shown to have

better binding to P-1,3-1,4-glucan (barley 1-glucan) than to xylan (Araki et al., 2004; Meissner et

al., 2000). Based on these inconsistent findings it is impossible to assume any precise

functionality of these CBM 22 modules. Analysis of this system by expression of each module

separately and the application of native affinity polyacrylamide gel electrophoresis (NAPAGE)

would clarify the role of each CBM 22 as they may function to accentuate CD catalytic ability

(Meissner et al., 2000).

It seems likely that there is a competitive advantage in colonizing a niche for an organism

that utilizes surface anchored enzymes to hydrolyze biomass polymers. The proximity of the

resulting hydrolysis products would decrease diffusion-dependent assimilation rates. This

strategy has been attributed to Clostridium spp. that produce the cell surface-localized

cellulosome (Bayer et al., 2004; Doi and Kosugi, 2004; Doi et al., 2003). However in anaerobic

ecosystems in which Clostridium spp. are the primary utilizers of biomass it has been suggested

that the main product of crystalline cellulose hydrolysis, cellobiose, actually decreases

cellulolytic activity and that excess cellobiose and other free sugars are utilized by other

members of the ecosystem. This relationship is truly communal in that these other bacteria

receive carbon substrates and may return the favor in the form of vitamins or other beneficial

growth factors for the cellulolytic organisms (Bayer et al., 1994). Inclusion of xylanolytic

enzymes in the cellulosome does not establish that these organisms can utilize the products

resulting from hydrolysis of MeGAXn. It's probable that hemicellulolytic activities are

associated with the cellulosome to increase cellulase access to cellulose by removing associated

non-cellulose polymers (Bayer et al., 2004). It has been reported that the mesophilic Clostridium

cellulovorans can ferment xylan, but there is no clear analysis of hydrolysis products and the

extent to which the xylan is utilized (Kosugi et al., 2001; Sleat et al., 1984). Additionally,

although there is increasing evidence of GH 10 xylanases in the Clostridium, there is no evidence

of accessory enzymes such as an a-glucuronidase which is thought to be required for complete

utilization of MeGAXn (Han et al., 2004). All of these characteristics pertaining to the

cellulosomal systems seem to stand in contrast to the MeGAXn hydrolytic system of

Paenibacillus sp. strain JDR-2. This system does not utilize the hydrolytic products of XynAi

CD efficiently and seems to require the activity of XynAi anchored to the cell surface for

efficient utilization of MeGAXn. This would suggest that the XynAi anchoring/ vectoral

transport mechanism has evolved to yield almost complete recovery of hydrolytic products as an

advantage against potential niche competitors. This may be further supported by the fact that

Paenibacillus sp. strain JDR-2 requires no nutritional supplement for growth on MeGAXn, but

growth of C. thermocellum and C. cellulovorans requires medium supplemented with yeast

extract (Bayer et al., 1983; Quinn et al., 1963; Sleat et al., 1984).

Paenibacillus sp. strain JDR-2 XynAi anchoring to the cell surface may be considered

somewhat analogous to the surface anchoring of the cellulosome in the Clostridia (Bayer et al.,

2004). An important distinction between these genera is that Clostridia are strictly anaerobic

organisms and Paenibacillus sp. strain JDR-2 requires oxygen for growth. This physiological

difference spatially separates these two genera in environmental niche development. In addition,

the cell surface associations of cellulases and other enzymes in Clostridium spp. is mediated

through the cellulosome. This complex of enzymes is maintained via interactions between

dockerin modules on individual enzyme proteins and cohesion modules on a scaffoldin protein,

which is then anchored to the cell surface (Doi, Kosuge et al., 2003). Surface anchoring of

biomass degradative enzymes may provide a strategy for efficient hydrolysis and transport of

resulting products, yielding a distinct advantage over organisms with free enzyme lignocellulose

degradative systems. In the example of C. thermocellum cellulose, cellobiose and cellodextrins

(degrees of polymerization < 4) are directly transported by a cellodextrin ABC transporter.

Transport of these oligosaccharides conserves ATP by the action of intracellular phosphorylase

yielding a significant growth advantage (Zhang and Lynd, 2005). Paenibacillus sp. strain JDR-2

may utilize a similar mechanism for the efficient utilization of X2 and MeGAX3. Additionally,

Paenibacillus sp. strain JDR-2 seems to couple the action of substrate hydrolysis to product

uptake. This may be a secondary method for the conservation of energy. Once internalized, the

MeGAX3 is apparently processed by a-glucuronidase (AguA)-mediated hydrolysis to MeGA and

X3, with subsequent hydrolysis of xylotriose by intracellular GH 10 xylanases (XynA2) with

specificity for small xylooligosaccharides and P-xylosidase (Gallardo et al., 2003; Pell et al.,

2004a; Preston et al., 2003). Although this may be the case, it is also possible that Paenibacillus

sp. strain JDR-2 utilizes an intracellular phosphorylase as described for several cellulolytic

organisms (Lou et al., 1996; Lou et al., 1997; Reichenbecher et al., 1997).

Considering the efficient utilization of MeGAXn by Paenibacillus sp. strain JDR-2, this

organism may provide a platform for future biocatalyst development. Under conditions of low

oxygen, Paenibacillus sp. strain JDR-2 produces succinate and acetate as fermentation products

(unpublished data). Alternatively, the genes encoding the cell surface anchored XynAi, as well

as those involved in the assimilation and metabolism of XynAl-generated products, may be used

to engineer other bacterial platforms to efficiently convert MeGAXn to desired fermentation

products. The aggressive utilization of MeGAXn by Paenibacillus sp. strain JDR-2 supports its

further development and genetic exploitation for the conversion of lignocellulosic biomass to

alternative fuels and bio-based products.

	Title Page
	Dedication
	Acknowledgement
	Table of Contents
	List of Tables
	List of Figures
	Abstract
	Introduction
	Biomass and bioconversion
	Family 10 glycosyl hydrolases:...
	Paenibacillus species strain JDR-2...
	Characterization of XynC from Bacillus...
	Summary discussion
	References
	Biographical sketch

Endoxylanases from Glycosyl Hydrolase Families 5 And 10: Bacterial Enzymes for Development of Gram-Positive Biocatalysts...

Material Information

Record Information

Material Information

Record Information

This item has the following downloads: