Citation
Improving Klebsiella oxytoca for Ethanol Production from Lignocellulosic Biomass

Material Information

Title:
Improving Klebsiella oxytoca for Ethanol Production from Lignocellulosic Biomass
Creator:
WOOD, BRENT E. ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Biomass ( jstor )
Carbon ( jstor )
Enzymes ( jstor )
Ethanol ( jstor )
Ethanol production ( jstor )
Fermentation ( jstor )
Ions ( jstor )
Klebsiella oxytoca ( jstor )
Plasmids ( jstor )
Sugars ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Brent E. Wood. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
2/28/2006
Resource Identifier:
495636469 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

IMPROVI NG Klebsiella oxytoca FOR ETHANOL PRODUCTI ON FROM LI GNOCEL LUL OSIC B IOMA SS By BRENT E. WOOD A DISSERTATI ON PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVE RSITY OF FL ORIDA I N PARTIAL FULFI LL MENT OF THE REQUIREMENTS FO R THE DEGREE OF DOCTOR OF PHIL OSOPHY UNIVERSI TY OF FLORI DA 2005

PAGE 2

Copyright 2005 by Brent E. Wood

PAGE 3

To my wife L isa and my children Jordan and Etha n. Without their love, support, and patience this work would not have been possible.

PAGE 4

iv ACKNOWLEDGMENTS I thank all those who encoura ged, counseled, and cajoled me in this endea vor . I especia lly w ould like to thank the curren t and for mer mana gement of BC I nternat ional, LL C for thei r indulg ence in a llowing m e to pursu e this deg ree. Mos tly, I wou ld like to thank Prof. Lonnie I ngram whose guidance a nd friendship brought my academic and professional career this far.

PAGE 5

v TABLE OF CONT ENTS page ACKNOWLEDGMENTS ................................................ iv LI ST OF TABLES ..................................................... vii LI ST OF FIGURES .................................................... viii ABSTRACT ........................................................... x CHAPTER 1 BACKGROUND AND GENERAL INTRODUCTI ON ...................... 1 What is L ignoce llulosic B iomass? ....................................... 1 Why L ignocellulose? ................................................. 2 Why E thanol? ....................................................... 4 Lignocellulosic Biomass Proce ssing ..................................... 5 What is the I deal Bio cataly st? ......................................... 11 Wh y Klebsiella oxytoca M5a1? ........................................ 11 Economic Factor s Affec tin Ethano l Producti on from L ignoce llulose ........... 16 Scope of This Work ................................................. 17 2 DEVELOPMENT OF A UREABASED MEDIUM AND EL IMI NATION OF TH E 2,3-BUTANEDI OL FERMENTATI ON PATHWAY TO IMPROVE ETHANOL PRODUCTI ON BY ET HANOL OGENI C Klebsiella oxytoca ................. 18 Introduction ....................................................... 18 Materials and Methods ............................................... 18 Results and Discussion .............................................. 26 Conclusions ....................................................... 38 3 OVE R E XP RE SS ION O F TH E MU TAN T GLO BAL R EGU LATO RY P ROT EIN CRP(IN) I MPROVES PENTOSE USE IN MIXED SUGAR FERMENTATIONS 43 Introduction ....................................................... 43 Materials and Methods ............................................... 45 Results and Discussion .............................................. 49 Conclusions ....................................................... 65

PAGE 6

vi 4 CONSTRUCTION OF ENDOGLUCANASE PRODUCING, ETHANOLOGENIC STRAINS OF Klebsiella oxytoca LACKING GENES FOR 2,3-BUTANEDIOL PRODUCTION .....................................................67 Introduction ....................................................... 67 Materials and Methods ............................................... 67 Results and Discussion .............................................. 71 Conclusions ....................................................... 82 5 GENERAL CONCLUSIONS ..........................................87 REFERENCES ........................................................92 BIOGRAPHICAL SKETCH .............................................105

PAGE 7

vii LI ST OF TABLES Table page 1-1. Composition of various lignocellulosic residues ........................... 3 2-1. Strains and plasmids used in media development and construction of budAB strains of K. oxytoca . ..................................................... 19 2-2. Composition of media .............................................. 23 2-3. Effect of budAB deletion on ethano l product ion and by -produc ts ............. 30 2-4. Estimated cost of me dia compo nents ................................... 42 3-1. Strains and plasmids used in the study and construction of Crp( in) over-expressing K. oxytoca ........................................................ 46 3-2. Su mm ar y of fe rm en ta ti on pr od uc ts an d c ar bo n b al an ce by K. oxytoca SZ21 .... 53 3-3. Ef fe ct of Cr p( in ) o ve r e xp re ss io n b y K. oxytoca SZ21 ..................... 59 3-4. Effect of chromosomal Crp(in) expression and budAB deletion on product formatio n after 96 hours f rom 30 g/ L gl ucose plu s 60 g/L xylos e ............. 64 4-1. Strains and plasmids used for the study and c onstruction of budAB , endoglucanase producing strains of K. oxytoca ....................................... 68 4-2. Product formation and carbon balance afte r 72 h from 90 g/L g lucose in OUM1 by budAB , ethanologenic, endoglucana se producing strains of K. oxytoca . ........ 73 4-3. Theoretical y ields from SSF and SSCF using 50 : L ce llulase pe r g Sig macell. .. 81

PAGE 8

viii LI ST OF FIGURES Figures page 1-1. Fuel ca rbon dioxide (CO 2 ) cycle . ....................................... 6 1-2. Enzy matic hy droly sis of ce llulose by the major constitue nts of ce llulase ........ 8 1-3. Glycoly sis and fermentation pathway s in K. oxytoca ....................... 13 2-1. Co ns tr uc ti on of pl as mi ds us ed to de le te th e 2 ,3 -b ut an ed io l f er me nt at io n p at hw ay. 21 2-2. Clustal W ali gnment o f selec ted reg ions of the genes encoding acetola ctate decarboxylase ( budA ) and acetolactate sy nthase ( budB ) .................... 27 2-3 Phylogene tic tree derived from budAB’ DNA se quences ................... 28 2-4. Fermentation of K. oxytoca P2 in vario us media .......................... 29 2-5. Development of optimized medium for K. oxytoca P2 ..................... 31 2-6. Distributi on of car bon in the f ermenta tion of 90 g /L g lucose in OUM1 ........ 34 2-7. Color palate of Voges-Proskour assay for acetoin production ................ 35 2-8. Effects of ) budAB on growth and ethanol production by ethanologenic strains of K. oxytoca ........................................................ 36 2-9. Co mp ar is on of et ha no l yi el d a nd pr od uc ti vi ty b y K. oxytoca strains ........... 37 2-10. Ef fe ct of pH on th e f er me nt at io n o f 9 0 g /L x ylo se by K. oxytoca BW21 ....... 39 3-1. Transcriptional regulation by Crp ..................................... 44 3-2. Scheme used for chromosomal replacement of wild-ty pe crp with crp(in) . ..... 48 3-3. Clustal W alignment of cloned DNA sequence containing c oding regions for crp from K. oxytoca M5a1, E. coli B and crp(in) from E. coli ET25 .............. 50 3-4. Clustal W ali gnment o f Crps fr om K. oxytoca M5a1, E. coli B, and Crp(in) from E. coli ET25 ...................................................... 51

PAGE 9

ix 3-5. Gr ow th an d e th an ol pr od uc ti on by K. oxytoca SZ21 from glucose, xy lose, and arabino se ......................................................... 52 3-6. Et ha no l p ro du ct io n b y K. oxytoca SZ21 from sugar mixtures ................ 55 3-7. Ef fe ct of Cr p( in ) o ve r e xp re ss io n o n e th an ol pr od uc ti on by K. oxytoca SZ21 from 30 g/L glucose and 60 g /L xy lose ...................................... 56 3-8. Ef fe ct of Cr p( in ) o ve r e xp re ss io n o n e th an ol pr od uc ti on by K. oxytoca SZ21 from 30 g/L glucose , 30 g/L xylos e, and 30 g/L arabino se ....................... 58 3-9. Ef fe ct of Cr p( in ) o ve r e xp re ss io n o n e th an ol pr od uc ti on by E. coli KO11 from 30 g/L glucose , 30 g/L xylos e, and 30 g/L arabino se ......................... 60 3-10. Et ha no l p ro du ct io n b y K. oxytoca BW23 fro m 30 g/L glucose and 60 g /L xy lose 62 3-11. Et ha no l p ro du ct io n b y budAB strains of K. oxytoca from 30 g/L g lucose and 60 g/L xylos e ..................................................... 63 4-1. Gr ow th an d e th an ol pr od uc ti on fr om 90 g/ L gl uc os e i n O UM 1 b y K. oxytoca strains BW21, BW34, BW35, and BW35 pCPP2006 ............................ 72 4-2. Ethanol pr oduction i n SSF of 100 g/L Sigmace ll .......................... 75 4-3. Residual s ugars ( 144h) in SSF of Sigma cell ............................. 77 4-4. Product f ormation i n SSF from 100 g/L Sigmace ll ........................ 78 4-5. Ethanol pr oduction a nd xyl ose consu mptionin SSCF of 40 g/L xylos e and 45 g /L Sigmace ll ........................................................ 80 4-6. Product formation by ethanolog enic, budAB strains of K. oxytoca in SSCF ..... 83 4-7. Residual s ugar in SSCF of 40 g /L xy lose and 4 5 g/L Sigmace ll .............. 84 4-8. Ethanol production, from cellulose, per unit enzy me in SSF and SSCF by K. oxytoca BW34 .................................................. 86

PAGE 10

x Abstract of Dissertation Presented to the Graduate School of the University of F lorida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy IMPROVI NG Klebsiella oxytoca FOR ETHANOL PRODUCTI ON FROM LI GNOCEL LUL OSIC B IOMA SS By Brent E. Wood August 2005 Chair: Lonnie O. I ngram Major Department: Microbiology and Cell Science Using e xis ting i nfrast ructu res, t he ex pande d use o f ethan ol cou ld pro duce l arge amounts of liquid fue l for use in transp ortation. Carbohy drates f ound in lig nocellul ose (cellulose and hemicellulose) are an abunda nt and unused source of raw material (sug ar) for prod ucing c ommodity chemica ls such as ethanol. Hemicel lulosic sug ars are easily hydroly zed with dilute acids; however, the hy drolysis of cellulose re quires more severe conditions, resulting in sugar degra dation and the production of biologically inhibitory compound s. Hemice llulose sug ars (pa rticular ly p entose su gars) are esp ecially sensitive to acid d egradat ion, and th us mu st be s eparat ed from cellu lose b efore c onti nued p rocess ing, adding to equipme nt costs. A lternati vely , enzy mes (ce llulases) may be used to depolymerize cellulose without the problems a ssociated with severe acid hy drolysis. The simultaneous saccharification and fermentation (SSF) of c ellulose has previously bee n shown to greatly reduce the amount of cellulase needed f or conversion to ethanol. Even so the costs of cellulases have been a major bar rier to their large-scale use . The

PAGE 11

xi fermen tation of h emicellu lose comb ined with SS F (SSCF) provides further process simplification. Ethanologenic strains of Klebsiella oxytoca M5a1 have been shown to be exceptionally well suited for SSF. Previous studies have use d expensive laboratory media, impractical for large sca le use. Other work showed that in the presence of glucose , the use o f the pen tose sug ars xy lose and a rabinose were re pressed . This observation was attributed to the transcriptional regulation of the g enes for xylose and/or arabinose metabolism by the c yclic-AMP rec eptor protein (CRP). This study used K. oxytoca M5a1's natural ability to use ure a as a nit rogen source t o develop an inexpensive medium (containing corn steep liquor) to produce ethanol from lignocellulose. In g lucose (9%) fermentation, using this new medium, elevate d levels of products from the 2,3-butanediol pathway were produced. Deleting two g enes exclusive to th is pa th wa y ( budAB ) grea tly i mproved b oth ethan ol produc tivity (30%) a nd y ield (12%). In para llel work, it was found that over expression of a mutant CRP, CRP(in), insensitive to the normal intracellular signals, improved the metabolism of pentose suga rs in the presence of glucose. F inally, ge nes encoding endoglucana se activities from Erwinia chrysanthemi (previo usly shown to be benefic ial in SSF) were c ombined w ith other improvements to construct strains further improved for ethanol produc tion from cellulose . The bes t strain ( BW34) pr oduced 3 8 g/L ethanol f rom 100 g /L ce llulose (i n SSF) compared to 28 g/L in stra in P2, a 36% improvement. Ethanol yields wer e increa sed from 0.38 to 0.4 9 g etha nol per g cellulose in SSCF, whi ch conta ined both xylose and cellulose (using 50% less fungal cellulase than in SSF).

PAGE 12

1 CHAPTER 1 BACKGROUND AND GENERAL INTRODUCTI ON Energy from biomass and particularly fuels from biomass can provide tremendous economic, environmental, and energy security bene fits. As energy prices rise and become i ncreas ingly volatile, and as ev idence o f globa l warming mounts, the case for clean, renewable, domestic sourc es of energy has never been clearer ( Greene et al. 2004 p. 1 ). Ethanol as an automotive fuel provides a clean-burning and renewable alternative to petrole um-base d fuels ( Arntzen and Dale, 1999 ). Current technology for ethanol production is based on edible crops such as sugar c ane juice (molasses) and corn starch ( Zaldivar et al. 2001 ). These feed stocks may contribute as much as 40% of the producti on costs. L ignoce llulosic bi omass, the inedible fractio n of plant materia l, is availab le at cos ts competi tive with p etroleu m ( Wyman, 2003 ). The continued development of improved microorganisms for the conve rsion of lignocellulosic sugars offers the potential to decrease depe ndance on petroleum and create ne w manufacturing opportuni ties from existing p lant mate rials. What is Ligno cellulosic Biomass? As descr ibed in The Fuels Security Act of 2005 , lignocellulosic biomass is the inedible portion of plant material (available on a renewa ble or recurring basis) including wood and wood residues, plants, grasses, ag ricultural residues, fibers, waste paper, a nd municipal yard wa ste. In gene ral, lignocellulosic biomass is the mixture of structural components of plant matter as opposed to the sugar storag e compounds, starch and sucrose. Lignoc ellulosic biomass is a mix of complex organic polymers. The primar y structural component (cellulose) is a $ -1,4 linke d poly mer of g lucose, f ound in hig hly

PAGE 13

2 ordere d cry stalline b undles. Ce llulose bu ndles ar e held tog ether by covalen t linkag es to the polysaccha ride, hemicellulose. Hemicellulose is a (highly branched) heterogene ous poly mer comp osed of th e pentose sugars xylos e and ar abinose a s well as t he hexose su ga rs gl uc os e, ma nn os e, an d g al ac to se . F ur th er cr os s l in ki ng is pr ov id ed by a pheny l-prope ne poly mer, lig nin. Propo rtions of t he compon ent poly mers in lignocellulose are dependant on the source (Table 1-1). Why Lignocellulose? Availability. A recent report by the U.S. Department of Agriculture and Departm ent of Ene rgy estimate d that more than 1.3 b illion dry tons (590, 000 metric tons) of fore stry and ag ricultur al waste materia l could be availab le with onl y mo dest cha nges to current harvesting practice ( Perlack et al. 2005 ). Another review of global ava ilability of agric ultural w aste (no t including forestr y r esidues) sugge sted more than 1.5 b illion metr ic tons are potentially ava ilable ( Kim and Dale, 2004 ). Crop residues such as sugar cane bagas se, gra in hulls, a nd suga r-beet pulp, and also for estry residue s are g eneral ly collecte d in the fie lds and se parate d as they are pro cessed. Centrali zed colle ction of th ese residues may pr ovide the best near-term opportunity for the use of lignocellulose as a fee d stock. Better land use. According to the National Commission on Energy Policy (2004) , Cultivation of crops (dedicated to energy production) generally requires less fertilizer and energy inputs compared to food crops. This allows land of marginal quality (for food production) to be used, thus eliminating the competition for land between food and energ y c rops ( Wyman, 1999 ). When fo od-crop residue s are use d, the add ed value from

PAGE 14

3 Hemicellulose CelluloseXylanArabinanMannanGalactanTotalLigninOtheraBagasse38.620.41.70.30.623.023.115.3 Corn Stover34.618.32.50.40.922.117.725.6 Corn Fiber17.721.414.21.24.741.54.036.8 Wheat Straw32.619.22.30.30.722.516.828.1 Hybrid Poplar39.213.10.91.80.916.716.727.4 Monterey Pine41.75.91.510.72.420.525.911.9a Includes ash, extractives, free sugar, starch, oils, and protein. Data compiled from The Biomass Feedstock Composition and Properties Database (2005) . Table 1-1. Composition of various lignocellulosic residues (% dry weight)

PAGE 15

4 the lignocellulosic crop offsets some of the fertilizer and ene rgy inputs, improving land us e e ff ic ie nc y. Solid waste dis posal. Many of the c rop residues separated in processing c an produce large volumes of waste mate rial. While some of these may be either sold as animal feed or burned to provide processing energy , there is typically more waste material than can be absorbed by either of these uses. Other residues (such as sug ar cane leaves and rice straw) h ave histo rically been bur ned in the field. Th is practi ce is re stricted in many areas, because of envir onmental concer ns about th e larg e amount o f partic ulate matter it produces. Furthermore, household and offic e lignocellulosic waste (paper, y ard waste, e tc.) re present s at least 40% of the volume of materia l curre ntly being p laced in land fills ( Bergeron and Hinman, 1991 ; Bulls, 1991 ; Kerstetter and Ly ons, 1991 ). Cost. As a waste material most lignocellulose is available at little or no cast. Even if lignocelluose costs $40.00 per dry ton, it would equal the value of petroleum at $13.00 per bar rel ( Wyman, 2003 ). When municipal waste is used, tipping fees create a dditional revenue. Why Ethanol? Reduced CO 2 gas emi ssions. All biomass-based ethanol, including starch and sugarbased et hanol, ha s the pote ntial to re duce g reenhou se-ga s product ion. Acc ording t o the Renewable Fuels Association (2003) , ethanol is a potent fuel oxyge nate and as such, when blended with gasoline, improves engine c ombustion (resulting in reduced tail-pipe emissions) . In a ddition to r educing CO 2 , ethanol also reduces the amount of ground-le vel ozone for ming com pounds (SO x , NO x , and CO). Bergeron (1996) , estimate d that a light-d uty vehicle running on biobase d ethanol (pure) would pro duce only 7% of the CO 2

PAGE 16

5 produce d by the same vehicle using re formulat ed gas oline. Ad ditional CO 2 reductions are re alized whe n the CO 2 cy cle is co nsidere d (Fig ure 1-1) . The only sources of net CO 2 producti on are f ossil-fue l based. T he CO 2 from ethanol is recy cled into the growth of the next crop. Inc reased blending of ethan ol into motor fuel pro portiona lly r educes C O 2 emissions. Reduced fertilizer and energy inputs for cultivating dedicated energ y crops could fur ther re duce CO 2 emissions. A lignoc elluose e thanol pla nt would us e the lig nin re si du es fo r s te am an d e le ct ri ci ty ( Wyman, 1999 ). Such a facility w ould consume no fossil fue ls and cou ld result i n net CO 2 consumption ( Greene et al. 2004 ). Energy security. Increasing demand for petroleum has resulted in world oil prices topping $60.00 per barrel. The United States consumes 25% of the oil produced worldwide, most of which is imported ( Greene et al. 2004 ). Of this, 70% is used in the transportation sector (97% going to fue l) ( Wyman, 2003 ; Greene et al. 2004 ). Growth of developing countries such as China and I ndia will further limit the supply of petroleum ( Wyman, 1999 ; NCEP, 2004 ). Recent military ac tions and political instability in the middle east may furthe r contribute to price volatility. I ncreased use of ethanol as a motor fuel cou ld grea tly r educe d ependa nce on fo reign s ources o f oil ( Greene et al. 2004 ; NCEP, 2004 ). Lignocellul osic Biom ass Pro cessing The Biorefinery. Th e c ur re nt mo de l f or a b io re fi ne ry ( Kamm and Kamm, 2004 ) using lig nocellul ose as a r esource is the sug ar platf orm model ( U.S. DOE, 2005 ). In the sugar platform model, carbohy drate fractions of lignocellulose (ce llulose plus hemicell ulose) a re conve rted to its constitue nt sugar s. These sugars are the n availa ble to produce an array of valuable chemicals and products in addition to biofuels (ethanol)

PAGE 17

6 Plant CropStarch Sugar Cellulosic Biomass Fossil Fuel DerivedEnergy Fuel Fertilizer Fuel Blending and Use Harvesting Processing CO2From Fossil Fuel CO2From Ethanol CO2From Fossil Fuel Plant Growth CO2From Ethanol Production Conversion to Ethanol Plant CropStarch Sugar Cellulosic Biomass Fossil Fuel DerivedEnergy Fuel Fertilizer Fuel Blending and Use Harvesting Processing CO2From Fossil Fuel CO2From Ethanol CO2From Fossil Fuel Plant Growth CO2From Ethanol Production Conversion to Ethanol Figure 1-1. Fuel carbon dioxide (CO2) cycle.

PAGE 18

7 needed for the transportation sector. Conceptually , the biorefinery mimics current practices in both the petrochemical and the wet cor n milling industries. Those industries essentially conver t all of the raw material to products. Biorefining consists of an initial thermochemical pretreatment (usually with dilute acids) that solubilizes (and depolymerizes) the hemice llulose and some of the lignin. The cellulose-containing residue must then be further processed to liberate f ree glucose. While this may also be done thermochemically with acid, the cry stalline nature of cellulose requires severe proces s cond itio ns for depol ymerizat ion: high t empera ture ( 150 t o 250° C) and /or hi gh acid con centra tions (one to 25%). T his gene rally create s sugar -degr adation p roducts, inhibitory to biolog ical pro cesses ( Hahn-Hägerdal et al. 1991 ; Taherzadeh, 1999 ; Zaldivar and Ingra m, 1999 ; Zaldivar et al. 2001 ; Klinke et al. 2004 ), and usually re duces sugar re co ve ry ( Lee et al. 1999 ; Torget et al. 2000 ). When high acid concentrations are used, economic s requir es reco very and reus e of the a cid cata lys t ( Sivers and Zacchi, 1995 ). Therefore the preferr ed route for hy drolysis of cellulose is to use enzy mes (cellulases). Enzymatic hydrolysis of cellulose. The use of cellulase enzy mes requires significantly lowe r temperatures (30 to 60°C). I n the absence of sugar deg radation, cellulas es can g reatly increa se suga r y ields ( Wyman, 1999 ), and can limit the production of toxic pro ducts. Ce llulase a ctivity is the sy nergis tic actio n of thre e broad e nzym e activitie s (Figu re 1-2) . Endogl ucanase s (EGs) h ydr oly ze bonds wi thin long er gluc ose chains, producing free ends for cellobiohydrolases ( CBHs) to act, releasing cellobiose. Ce ll ob io se is th en ac te d o n b y $ -glucosidases (BGs) relea sing glucose monomers. Each of these categories contains seve ral independent proteins, each with differing specificities and acti vities. I n one of th e best ce llulaseproducin g org anisms use d in comme rcial

PAGE 19

8CH2O O O CH2O O O O CH2O O O O CH2O O O O CH2O O CH2O O CH2O O O O CH2O O O O CH2O O O O CH2O O CH2O O O CH2O O CH2O O CH2O O O O CH2O O O CH2O O CH2O O O O CH2O O O CH2O O O CH2O O O O CH2O O O CH2O O CH2O O O CH2O O CH2O O O CH2O O O CH2O O CH2O O CH2O O CH2O O CH2O O CH2O O Endoglucanase CBH CBH -glucosidaseCH2O O O CH2O O O O CH2O O O O CH2O O O O CH2O O CH2O O CH2O O O O CH2O O O O CH2O O O O CH2O O CH2O O O CH2O O CH2O O CH2O O CH2O O O CH2O O O O CH2O O O O CH2O O O O CH2O O CH2O O CH2O O O O CH2O O O O CH2O O O O CH2O O CH2O O O CH2O O CH2O O CH2O O O O CH2O O O CH2O O CH2O O O O CH2O O O CH2O O O CH2O O O O CH2O O O CH2O O CH2O O O CH2O O CH2O O O CH2O O O CH2O O CH2O O CH2O O CH2O O CH2O O CH2O O Endoglucanase CBH CBH -glucosidaseFigure 1-2. Enzymatic hydrolysis of cellulose by the major constituents of cellulase.

PAGE 20

9 cellulase production ( Trichoderma reesei ) two CBHs, five EGs, and two BGs have bee n identified ( Kubicek and Penttila, 1998 ; Ly nd et al. 2002 ). Trichoderma does not produce large quantities of BG activity . Because of the strong inhibition of CBHs by cellobiose, BG must be suppleme nted for m other so urces ( Kubicek and Penttila, 1998 ). BG ac tivity itself is inhibited (to a lesser degree) by glucose. Another way to overcome the end-product inhibition of cellulases, in converting lignocellulose to ethanol was developed during the oil cr isis in the 1970s. The simultane ous sacc harific ation and fermen tation (SSF ) proce ss combine d the cel lulosic substrate with cell ulases a nd the y east Saccharomyces cerevisiae ( Gauss et al. 1976 ). In th e S SF pr oc es s s ub -i nh ib it or y gl uc os e c on ce nt ra ti on s w er e m ai nt ai ne d b y S. cerevisiae consuming the liberated glucose. This reduc ed the need for supplemental BG. Howeve r, the low levels of BG in T. reesei broths an d the stro ng inhibi tion of CB H, by cellobio se still required some BG supplementation (Brooks and Ing ram, 1995 ). What about hemicellulose? While SSF is good for converting cellulose, pretreated lignocellulose contains solubilized hemicellulose that must be separated f rom the cellu lose res idue. This require s extensive washing , resultin g in dilut e suga r stream s or the use of expensive counter-current washing systems. Efficient conve rsion of hemic ellul ose su gars is vital to th e econo mics o f bior efini ng (Hinman et al. 1992 ; Hettenhaus, 1998 ; Wyman, 1999 ; Saha, 2003 ). Use of y easts suc h as S. cerevisiae works well on the cellulose fraction of lignoce llulose, b ut the hem icellulos e frac tion (par ticularl y pe ntose sug ars) is no t well used. Pen tose-usin g y easts (su ch as Pichia stip itis ) have an absolute requirement for an external electron acceptor (g lucose co-metabolism and/or carefully controlled O 2

PAGE 21

10 concen trations) to use pen tose sug ars ( Skoog and Hahn-Häger dal, 1990 ). The careful process controls needed for P. stipitis make it impractical for large-sc ale production. To simplify the processing of lignocellulose, a modification of the original SSF was proposed. In the simultaneous saccha rification and co-fermentation (SSCF) the bulk of the pretreated lignocellulose (ce llulose + hemicellulose) is combined with cellulases and a biocatalyst, thus eliminating the need for extensive liquid-solid separation. Because of the simplif ied unit op eration s, SSCF or SSF is prefe rred to a separa te enzy me hy droly sis and fermentation ( Wright et al. 1988 ). In either c ase, a bi ocataly st is need ed that c an use all of the sugars found in lignocellulose. The met abolic engi neering so lution. While no na tive y east or b acteria is complete ly s atisfac tory for conv erting pentose s ugars t o ethanol ( Chandrakant and Bisaria, 1998 ), the advent of genetic eng ineering techniques provided a way to extend the spectrum of sugars that could be converte d to ethanol. Several research g roups have used these techniques to broaden the substrate rang e of natural ethanol-producing org anisms. In S. cerevisiae the genes for using xy lose ( Ho et al. 1998 ) and arabinose ( Sedlak and Ho, 2001 ) have be en adde d. The et hanol pro ducing Gram neg ative ba cterium Zymomonas mobilis also has b een impr oved to use pentose s ugars ( Zhang et al. 1995 ; Deanda et al. 1996 ). Using another approach with a strain of Escheric hia coli B (ATCC#11103) Ohta et al. (1991b) chromosomally integ rated the genes encoding pyruvate de carboxylase ( pdc ) and alcohol dehy drogenase ( adhB ) from Z. mobilis . Escheric hia coli B is natur ally able to use a wide range of sug ars including those found in lignocellulose. I ncorporating the Z. mobilis ethanol p athway create d an org anism able to produc e ethano l, at hig h yi elds,

PAGE 22

11 from a wi de rang e of sug ars ( Beall et al. 1991 ). The sa me appro ach was later ta ken with the other Gram-ne gative organ isms Klebsiella oxytoca M5a1 ( Wood and Ingram, 1992 ) and Erwinia chrysanthemi ( Beall and Ing ram, 1993 ) as well as the Gram-positive bacter ium Bacillus s ubtilis ( Barbosa and I ngram, 1994 ). What is the Ideal Biocatalyst? Several traits have previously been identified for a biocataly st in the conversion of lignoce llulose to e thanol ( Hettenhaus, 1998 ; Zaldivar et al. 2001 ; Dien et al. 2003 ). Included are; • High et hanol y ields. • Broad substrate utilization range. • Tolerance to low pH. • Minimal nut rient re quireme nts. • Minimal By-produc t formation. • Produce enzymes for the depolymerization of lignoc ellulose. Why Klebsiella oxytoca M5a1? Klebsiella oxytoca M5a1 (here after K. oxytoca ) is consid ered non pathog enic, ha s a history of safe use, and is nearly incapab le of per sisting in humans ( Parker a nd Abbot, 2003 ). Orig inally it was isola ted as Aerobacter aerogenes from a 2,3-butanediol fermen tation, la ter rec lassifie d as Klebsiella pneumoniae M5a1 ( Mahl et al. 1965 ). It is co mm on ly f ou nd in th e e ff lu en ts fr om th e p ul p a nd pa pe r i nd us tr y ( Grimont et al. 1991 ). Klebsiella oxytoca has several of the above traits desirable for a biocatalyst conve rting lignocellulose. Broad substrate range. Klebsiella oxytoca was previously shown to use a wide range of monomeric suga rs derived from lignocellulose ( Ohta et al. 1991a ; Wood and Ingram, 1992 ; Bothast et al. 1994 ). Unlike analogous strains of E. coli ( Ohta et al. 1991b ;

PAGE 23

12 Yomano et al. 1998 ), K. oxytoca has the na tive abili ty to metabolize other sol uble products from lignocellulosic biomass including cellobiose, cellotriose, xy lobiose, xylotr iose, and arabino sides ( Wood and Ingram, 1992 ; Burchardt and I ngram, 1992 ; Qian et al. 2003 ). Additionally, K. oxytoca has the ability to use many of the products from starch d egrad ation suc h as maltod extrins and cy clodextrin s ( Feederle et al. 1996 ; Fiedler et al. 1996 ). High yield and t olerance to low pH. Ethanologenic strains of K. oxytoca have been shown to produce ethanol efficiently from a variety of sugars, at y ields approaching theoretical maxima ( Ohta et al. 1991a ; Wood and Ingram, 1992 ; Bothast et al. 1994 ). At pH 5 to 5.5, near the optimal fo r commer cial ce llulases ( Nieves et al. 1998 ), the ethanologenic strain K. oxytoca P2 ( Wood and Ingram, 1992 ) was ab le to eff iciently metaboliz e incomplet ely hydroly zed product s from lignoc ellulose i n SSF ( Doran and Ingram, 1993 ; Doran et al. 1994 ; Brooks and Ing ram, 1995 ; Wood, 1997 ; Wood et al. 1997a ). This trait provides added advantage during SSF processes. Klebsiella oxytoca P2 re qu ir ed le ss th an ha lf of th e f un ga l e nz yme s r eq ui re d b y S. cerevisiae to achieve equivale nt ferme ntation ra tes and y ields ( Brooks and Ing ram, 1995 ). Minimal by-product format ion and nutrit ional requir ement s. Native K. oxytoca , has four fermen tative pa thway s ( Ørskov, 1984 ; Böck and Sawers, 1996 ; Figure 1-3) • The pyruva te formate-ly ase pathway (producing formate, H 2 + CO 2 , acetate, and ethanol). • Th e l ac ti c a ci d p at hw ay. • Th e s uc ci na te pa th wa y. • Th e b ut an ed io l p at hw ay. The distribution of the fermentation products, in Klebsiella , is dependant on the culture

PAGE 24

13Glucose Pyruvate Lactate Acetolactate Acetoin2,3-Butanediol Acetyl-CoA Acetate Ethanol Acetaldehyde Oxaloacetate Citrate Succinate 2-Oxoglutarate FormateldhA pflB budB budA budC adhE pta ackA frdABCD ATP ADPDihydroxyacetonePhosphate Glyceraldehyde-3-Phosphate NADH NAD+ NADH NAD+ NAD+NADH ADP ATP ADP ATP NAD+NADH ADP ATP NAD+NADH PEP H2+ CO2FHLPyruvate BiosynthesisNADH NAD+ adhE A B C DGlucose Pyruvate Lactate Acetolactate Acetoin2,3-Butanediol Acetyl-CoA Acetate Ethanol Acetaldehyde Oxaloacetate Citrate Succinate 2-Oxoglutarate FormateldhA pflB budB budA budC adhE pta ackA frdABCD ATP ADPDihydroxyacetonePhosphate Glyceraldehyde-3-Phosphate NADH NAD+ NADH NAD+ NAD+NADH ADP ATP ADP ATP NAD+NADH ADP ATP NAD+NADH PEP H2+ CO2FHLPyruvate BiosynthesisNADH NAD+ adhE A B C DFigure 1-3. Glycolysis and fermentation pathways in K. oxytoca . (A) The pyruvate formate-lyase pathway; (B) The lactate pathway; (C) The succinate pathway; (D) The 2,3-butanediol pathway.

PAGE 25

14 conditions such as nutrient limitation, carbon source, and pH ( Johansen et al. 1975 ; Teixeira de Mattos and Neijssel, 1997 ; Tolan and Finn, 1987 ). In the fermenta tion of xylose (rich media, pH 6.0) by ethanologenic K. oxytoca , acidic product formation was sightly eleva ted, but 2,3-butanediol production was negligible ( Ohta et al. 1991a ). However, the high ethanol y ields which they observe d suggested that by -product formation was minimal. Findings were similar in an ethanol producing strain of Klebsiella planticola ( Tolan and Finn, 1987 ). In an SSF, wh ere the rate limit ing step is the release of sugars, sug ar would be limiting for the growth of K. oxytoca . Under t his scenario by -product formation would be expected to be low ( Teixeira de Mattos and Tempest, 1983 ). While all of the prior studies with ethanologenic K. oxytoca used the rich lab oratory medium L uria Br oth, work with the r elated s trains Klebsiella aerogenes ( Teixeira de Mattos and Tempest, 1983 ), K. plantico la ( Tolan and Finn, 1987 ), and K. oxytoca strain ATCC 8724 ( Jansen et al. 1984 ) used minimal-salts media, indicating the low nutritional requirements of the genera. Production of polysaccharases . Native K. oxytoca produces the starch-degra ding enzyme pullulanase (Pul). The Pul sy stem has been well studied and is one of the paradig ms for ty pe I I sec retion sy stems ( Wandersman, 1996 ). To produce enzy mes for the depolymerization of ce llulosic substrates, initial studies focused on the co-production of enzymes exogenously , to then be used in cellulose fermentations. I n 1992, Wood and Ingram showed the efficacy of producing the endoglucana ses CelA, CelB, CelC, and CelD, fr om Clostridium thermocellum , as a co-product in ethanol fermentations by K. oxytoca P2. The most effective of these (CelD) was shown to partially replace the commercial cellulases in fermenting Solka-F loc. Later work used the ethanologenic E.

PAGE 26

15 coli strain KO11 ( Ohta et al. 1992b ) to co-p roduce e ndogluc anase Ce lZ from E. chrysanthemi ( Beall, 1995 ; Wood, 1997 ; Wood et al. 1997b ), CelA, f rom B. subtilis ; and CenA, Cen B, CenC, a nd CenD, f rom Cellulomonas fimi ( Wood, 1997 ; Wood et al. 1997b ). The la tter study showed th at at the p H used fo r SSF (5.2 ) with fun gal ce llulases, only the E. chrysanthemi Ce lZ re ta in ed si gn if ic an t a ct iv it y. The eff icacy of CelZ , at pH 5.2 , led to fu rther sim plificat ion of the SSF proce ss usin g K. oxytoca . In 1999, Zhou an d Ing ram integrated the gene celZ , encoding the E. chrysanthemi endoglucanase (CelZ) on the chromosome of K. oxytoca P2. When combined with the E. chrysanthemi specific secret ion gen es ( out ), the be st of these (strain SZ6) pr oduce d more than 2 4,000 IU per L endo glucan ase act ivit y while ret ainin g high ethanol productivity. Zhou and Ing ram (2000) then foun d that com bining Ce lZ with another endogl ucanase from E. chrysanthemi (CelY ) resul ted in a high d egree of synergy betwee n the two e nzym es. The g enes for both enzy mes ( celY and celZ ) were then placed on the chromosome of K. oxytoca P2 ( Zhou et al. 2001 ). This strain (SZ21) expressing the out secret ion gen es produc ed more t han 20,00 0 IU o f extra-c ellular e ndogluc anase activity and produce d more than 20% more ethanol, in SSF with fungal cellulases, than did the unmodified strain P2. Their study showed that, in conjunction with commer cial enzy mes, CelY was the mo st benef icial. Str ain SZ21 plus out , was abl e to dire ctly convert acid-swollen cellulose to ethanol without commercial cellulase supplementation ( Zhou and Ing ram, 2001 ).

PAGE 27

16 Economi c Fact ors Affect ing Ethanol P roduction from Ligno cellulose Enzyme cost. The purc hase pri ce of co mmercia l cellula ses was p reviousl y estimated at $0.30 to 0.50 per gallon of ethanol produced ( Hettenhaus and Glasner, 1997 ; Himmel et al. 1999 ). This represents a significant frac tion of the current ethanol sales price of $1.22, as reported in The Kingsman Ethanol Report, 2005 . As part of the sugar platform model, the U.S. Dept. of Energy subcontracted with the cellulase manufacturer s Genencor Inte rnational, Inc. and Novozy mes A/S to reduce the cost of cellulase for the production of ethanol from lignocellulose by at least ten fold ( U.S. DOE, 2005 ). Both companie s have sin ce repo rted 30 f old cost r eduction s ( Genencor Inte rnational, Inc., 2004 ; Novozymes A/S, 2004 ). Howev er, firm price sc hedules f or the re duced c ost enzymes are not y et available (personal communication). Assuming that these price-reduction claims are accurate, further cost saving s may be possibl e throug h other imp rovemen ts. There is linear relation ship betwe en enzy me cost and enz yme loading in SSC F or SSF ( Himmel et al. 1999 ). The amount of comme rcial cellu lase r equir ed for i s dire ctly relat ed to s everal factor s incl udin g; • Efficacy of the biocatalyst in SSF (SSCF). • Co-producing additional cellulases can reduc e the need to add external cellulase. • Pretreatment of lignocellulose can signific antly affe ct the digestibility of lignoce llulosic su bstrates ( Ly nd et al. 2002 ; Mosier et al. 2005 ). Improvements in any of these areas can further r educe the cost of producing ethanol fr om lignocellulose. Media cos t and product yields. An inherent advantage of produc ing ethanol from grain a nd molasse s is that mos t of the nu trients r equired for cel l growth is contain ed in the substrate. Conversely , the hydroly sates of lignocellulose are nutrient poor ( Kadam

PAGE 28

17 and Newman, 1997 ). Thus inexpensive industrial media needed to be developed that retain h igh etha nol produc tivity and y ield from lignoce llulosic f eedstoc ks ( Ing ram et al . 1998 ; Zhang and Grea sham, 1999 ). The ma intenanc e of hig h ethanol yie lds nece ssarily equates to low by-pr oduct formation. In a previous sensitivity analysis, Ly nd et al. (1996) showed that improving ethanol y ields has a potential cost savings of 3.6 to10.0¢ per gallon e thanol. Scope of This Work This pres ent study address es three areas i n the impro vement of ethanolo genic strains of K. oxytoca . All focused toward reducing the cost of e thanol production from lignoce llulosic bi omass. • Develop a low-cost media for ethanol production. • Reduce b yproduct f ormation b y ma king spe cific g ene dele tions. • Improve pentose metabolism in the presenc e of glucose. • Incorporate E. chrysanthemi endogl ucanase s into impro ved stra ins.

PAGE 29

18 CHAPTER 2 DEVELOPMENT OF A UREABASED MEDIUM AND EL IMI NATION OF TH E 2,3-BUTANEDI OL FERMENTATI ON PATHWAY TO IMPROVE ETHANOL PRODUCTI ON BY ET HANOL OGENI C Klebsiella oxytoca Introduction Klebsiella oxytoca strain P2 contains chromosomally integ rated genes encoding pyruvate de carboxylase ( pdc ) and alcohol dehy drogenase ( adhB ) from Zymomonas mobilis. All previous work with K. oxytoca P2 has use d complex g rowth med ia containi ng labo ratory nutrients such as y east extra ct and Dif co Try ptone, cl early impractical for commodity c hemicals. In this study , we have developed a new medium that contains low levels of corn steep liquor, mineral salts, urea as a sole nitrogen source ( Ørskov, 1984 ), and fermentable sugar. With this medium, further g enetic engineering of the bioca taly st was re quired to eliminate byproducts and reta in high e thanol pr oductivity and yield. Materi als and Met hods Strains and plasmids. Table 21 lists the o rganis ms and pla smids used in media developm ent and by -produc t reduct ion. Ethan ologen ic strain s were ma intained on Lur ia agar ( Ausubel et al. 1996 ) contai ning 2% g lucose a nd chlora mphenico l (40 or 60 0 mg/L on alternate day s) under argon. Other strains wer e maintained on Luria ag ar plates lacking added sugar with appr opriate antibiotics. Unless otherwise noted; ampicillin (50 mg/L), kanamy cin (50 mg/L), apr amycin (50 mg /L), and tetracy cline (12.5 mg/L) we re used for selection. Strains harboring plasmids with temperature c onditional replicons were grown at 30°C. All other strains were maintained at 37°C, except where noted.

PAGE 30

19 Strain or PlasmidTraitsSource/Reference E.coli TOP10F’ hsdR mcrA lacZ)Tj/T1_0 1 Tf0.0052 Tc 0 9.96 -9.96 0 195.45 302.85 Tm(M15 endA recA ; F’ tet lacI Invitrogen DH5)Tj/T1_1 1 Tf0.0005 Tc 0 9.96 -9.96 0 208.41 232.2901 Tm[(lacZ)Tj/T1_0 1 Tf0.0043 Tc 0 9.96 -9.96 0 208.41 256.1701 Tm(M15 recA Bethesda Res Lab K.oxytoca M5a1 prototroph Ohta et al. 1991a P2 pflB ::( Zm pdc , adhB ) cat Wood and Ingram, 1992 BW15 )Tj/T1_1 1 Tf0.0054 Tc 0 9.96 -9.96 0 262.41 240.9301 Tm(budAB ::FRTtet -FRT This study BW19 P2 transductant from BW15, )Tj/T1_1 1 Tf0.0054 Tc 0 9.96 -9.96 0 275.97 356.13 Tm(budAB ::FRTtet -FRT This study BW21 BW19 Tcs, )Tj/T1_1 1 Tf0.0054 Tc 0 9.96 -9.96 0 289.41 284.8501 Tm(budAB ::FRT This study pCR2.1-TOPOTOPO T/A PCR cloning vector bla kan Invitrogen pLOI2065 FRTtet -FRT Underwood et al. 2002b pFT-K FLP Recombinase kan Posfai et al. 1997 pKD46 Red recombinase, bla Datsenko and Wanner, 2000 pHP45aac Blondelet-Rouault et al. 1997 pLOI2745temperature cond itional vector, pSC101ts, kan This study pLOI3301pCR2.1 budAB’ This study pLOI3310pLOI3301 budA’ FRTtet -FRT ’budB’ This study pLOI3313pLOI2745 budA’ FRTtet -FRT ’budB’ This study pLOI3420 1.8 kbp Sma I DNA frag.containing aac from pHP45W Xmn I site of pKD46 This studyTable 2-1. Strains and plasmids used in media development and construction of budAB strains.

PAGE 31

20 Plasmid preparations were stored at -20°C. Stock cultures were stored in 40% gly cerol at -75°C. Geneti c met hods. Stan dard m ethod s were u sed fo r PCR -based gene clo ning, plasmid co nstructio ns, and g enetic a naly ses ( Miller, 1992 ; Ausubel et al. 1987 ; Sambrook and Russel, 2001 ). Methods for integration, chromosomal deletions, and the use of removable antibiotic resistance genes ha ve been previously described ( Datsenko and Wanner, 2000 ; Martinez-Morales et al. 1999 ; Zhou et al. 2003 ; Causey et al. 2004 ). Escheric hia coli DH5 " was used for constructions. Bacteriophag e P1 (an E. coli phage) was used to transduce the budAB chromosomal deletion constructed in K. oxytoca M5a1 in to st ra in P2 ac co rd in g t o t he me th od s d es cr ib ed by Silhavay et al. (1984) . Although the budAB genes have not been previously described in K. oxytoca , homologous genes are known fr om two rel ated org anisms: Enterobacter aerogenes (fr. Klebsiella mobilis ) and Roultella terrigena (fr. Klebsiella terrigena ) ( Blomqvist et al. 1993 ; Mayer e t al. 1995 ). Based on these sequences and the pa rtial (unannotated) genome of K. pneumoniae ( Washington University Genome Project, 2003 ), primer s were d esigne d for the PCR ampli ficat ion o f a DNA f ragment conta inin g budAB' (Figure 2-1). These g enes encode " -acetolactate decarboxy lase and " -acetolactate sy nthase, respectively . A DNA fragment conta inin g " -acetolactate decarboxy lase ( budA ) and the 5' end of " -aceto lactate syn thase ( budB ) was amp lified by PCR using g enomic DN A from K. oxytoca as a template in Taq PCR Master Mix (Qiagen). After an initial denaturation at 94°C for 3 min, DNA was amplified for 30 cy cles; denaturation at 94°C for 30 s, annealing at 55°C for 30 s a nd extension at 72°C for 70 s. A final elongation step at 72°C for 10 min was also included. The budAB' DNA fragment was amplified (for ward primer,

PAGE 32

21 1000 2000 3000 4000 budR Fnr budA budB budC Blp I Sac II Fnr budA budB' pLOI33015113 bps Blp I Sac II Eco RI Eco RI budA budB' f1 ori kan bla 'budA budB'' pLOI33105877 bps Apa I Hin dIII budA' FRT tet 'budB' kan bla pUC ori Hin dIII, Apa I Hin dIII, Apa I Gel purify 2.1 kbp fragment Sma I+ pCR2.1-TOPO Blp ISac II, Blunt w/T4 pol GCTGAATCGGGTCAACATTT TTTCGGTTTGTCCAGGTAGTGel purify 1.7kbp FRTtet -FRT fragment low pH BudR A BSac II Blp I pLOI33135502 bps Bsr BI Bsr BI Bsr BI Bsr BI Bsr BI kan budA' tet 'budB'pLOI27453515 bps Apa I Hin dIII kan pLOI20654356 bps Sma I Eco RI Sma I Eco RI tet bla pSC101tspSC101tspUC ori f1 ori FRT FRT FRT FRT FRT 1000 2000 3000 4000 budR Fnr budA budB budC Blp I Sac II Fnr budA budB' pLOI33015113 bps Blp I Sac II Eco RI Eco RI budA budB' f1 ori kan bla 'budA budB'' pLOI33105877 bps Apa I Hin dIII budA' FRT tet 'budB' kan bla pUC ori Hin dIII, Apa I Hin dIII, Apa I Gel purify 2.1 kbp fragment Sma I+ pCR2.1-TOPO Blp ISac II, Blunt w/T4 pol GCTGAATCGGGTCAACATTT TTTCGGTTTGTCCAGGTAGTGel purify 1.7kbp FRTtet -FRT fragment low pH BudR A BSac II Blp I pLOI33135502 bps Bsr BI Bsr BI Bsr BI Bsr BI Bsr BI kan budA' tet 'budB'pLOI27453515 bps Apa I Hin dIII kan pLOI20654356 bps Sma I Eco RI Sma I Eco RI tet bla pSC101tspSC101tspUC ori f1 ori FRT FRT FRT FRT FRTFigure 2-1. Construction of plasmids used to delete the 2,3-butanediol fermentation pathway. (A) Operon and transcriptional regulation. As indicated, expression of the budAB operon is increased by low pH, BudR, and Fnr. (B) PCR primers used to clone budAB’ and plasmids used in construction.

PAGE 33

22 5'-GCTGAATCGGGTCAACATTT-3'; re verse primer, 5'-TGATGGACCTGTTTGGCTTT-3' ) and cloned into pCR2.1-TOPO (Figure 2-1). The recovered frag ment was sequenced and has been deposited in Ge nBank (Accession No. AY722056). As shown in Figure 2-1 (after se quencing), the unique restriction endonuclea se sites Blp I and Sac II were identified. These sites were used to re move a 0.9 kbp fragment within the coding regions of budAB . The single stranded overhang s on the remaining 4.2 kbp fragment of pL OI3201 were made blunt using T4 DNA polymer ase. A 1.7 kbp SmaI fragment from pL OI2065, containing F RTtet -FRT, was gel purified and liga ted to the blunt 4.2 kbp fragment of pLO I3201 to make pL OI3210. A 2.1 kbp Hin dIIIApa I fragment from pL OI3210 was ge l purified and ligated into the Hin dIIIApa I site s in pLOI 2745 (a vector that contains a temperature c onditional pSC101 replicon) to make pLOI 3213. Plasmid pLOI3213 wa s cut with Bsr BI (to minimize background) and a 1.8 kbp budA ’-FRTtet -FRT-‘ budB ’ PCR product was amplified from the 2.1 kbp Bsr BI fragm ent from p LOI 3313, usin g the sa me primer s and rea ction con ditions use d in cloning. Flanking FRT ( F lp recombinase R ecognition T arget ) sites we re inclu ded to facilitate marker removal after c hromosomal integration ( Posfai et al. 1997 ; Martinez-Morales et al. 1999 ; Underwood et al. 2002b ). Media. Components of media used in our study are summarized in Table 2-2. Each was used with either 50 g/L (278 mM) glucose or 90 g/L (500 mM) glucose. Components were purchased from either Fisher Scientific Company or Sigma Che mical Company. I norganic salts were reag ent grade. Urea wa s technical grade. M9 medium was prepared as previously described ( Neidhardt et al. 1974 ) and further supplemented

PAGE 34

23 Media Composition (mM) Componenta LBbM9(+Fe)cU-M9(+Fe)d1% CSL+MeU-1%CSL+MfOUM1gKH2PO4 22227.47.410.7 K2HPO4 2.92.9 Na2HPO4 42421.3 Total PO4646410.310.312 NaCl 85.699 CaCl2 0.10.1111 MgSO4 11221 FeCl3 0.0740.0740.0740.0740.0740 NiCl2 0.00680.00680.0068 NH4Cl 19 (NH4)2SO4 23.5 NH2CONH21023.510 Total Nitrogenh1920474720 Tryptone (g/L)10 Yeast extract (g/L)5 CSL (g dry wt./L)555a Degree of hydration is omitted for simplicity. b Luria Broth (Ausubel et al.1987). c M9 medium (Neidhardt et al. 1974). d M9 with NH4Cl replaced with equivalent urea nitrogen.e 0.5% CSL+M (Martinez et al. 1999). f 0.5% CSL+M media except that (NH4)2SO4 was replaced with equivalent urea nitrogen. g OUM1, optimized urea media number 1. h Total mmoles of available nitrogen (excluding nitrogen from complex media components).Table 2-2. Composition of media (excluding fermentable sugar).

PAGE 35

24 with 0.07 mM FeCl 3 to ens ure ade quate level s for t he iro n-requ iring Z. mobilis alcohol dehydrog enase ( Scopes, 1983 ). Corn ste ep liquor (CSL) medium for ethanolo genic E. coli , 0.5% CSL+ M, has been previously described ( Martinez et al. 1999 ). Both M 9 and 0.5% CSL+M media were use d as starting points to optimize a medium for ethanologenic derivatives of K. oxytoca M5a1. When urea was used as the nitrogen source, 0.0068 mM N iCl 2 was added for urease activity . Note that CSL levels are expressed on a dry w eight basis. Stock solutions of CSL were prepar ed and sterilized as previously descr ibed ( Underwood et al. 2002a ). Media opti mizat ion. K. oxytoca P2 was use d in all med ia optimiza tion studie s. Isolated colonies from fre shly grow n plates (24 h) were resuspended in 1 ml of deionized H 2 0 and used to inoculate 125 ml flasks (~50 : L inoculum) containing 75 ml of medium (50 g/L glucose ). Growt h and etha nol produc tion were monitored after 2 4 and 48 ho urs. Ferm entation. Seed cultures (150 ml in 250 ml flasks) were grown for 16 h at 35°C (120 rpm) in media containing 50 g/L glucose. Cells were harvested by centrifugation (5000 x g, 5 min) and used as inocula to provide a n initial concentration of 33 mg/L dry cell weight (OD 550nm =0.1). Respective media used for fermenta tions were also used for seed growth but with a lower conc entration of glucose (50 g/L ). Fermentation vessels were previously described ( Beall et al. 1991 ) and contained an initial volume of 350 ml (90 g/L glucose) . Cultures were incubated at 35°C (150 rpm). Broth was maintained at pH 5.2 (except where noted) by automatic addition of 2N KOH . Screening for the butanediol pathway. Strains we re scre ened for acetoin production using a modification of the Vogues-Proskaur (VP) agar method described by Blomqvist et al. (1993) that used microtiter plates instead of petri plates, increasing the

PAGE 36

25 sensitivity by limiting diffusion of the colored product. Each well was filled with 1 ml of the medium (per liter: 2.5 g Difco bacto peptone, 1.0 g Difco yea st extract, 10 g glucose, 1.0 g sodium pyruva te, and 25 g agar), a nd inoculated. After 24 hours, 200 : L of 5% " -napthol solution in 2.5 N NaOH was added to each well. Color development was monitored for one h at room temperature. The abse nce of red color confirmed the lack of ace to in (pr odu ct o f Bu dA a cti vi ty). Add it io nal con fir mat io n wa s pr ovi ded by HP LC analy sis of fer mentation products . Analytica l metho ds. Cell mass was estimated by deter mining OD 550nm with a Bausch & L omb Spectronic 70 spectrophotometer. With this instrument, 1 OD 550nm corres ponds to a c ell densit y of 0.33 mg ( dry wt.)/L . Note tha t measure ment of c ell density for K. oxytoca has a lar ge err or due to t he clumpi ng natu re of the biocata lys t. Ethanol was measured by gas chromatography using a Varian model 3400 X as previously descr ibed ( Ohta et al. 1991b ). Other fermentation products were dete rmined by high-pe rformance liquid chromatography (HPLC) using a Hewle tt-Packard model 1090 seri es II chromat ograp h and a B io-Rad Am inex HPX-87 H ion par tition colu mn (45°C; 4 mM H 2 SO 4 ; 0.4 ml/min.; 10 : L injection volume) with dual detectors (re fractive index and UV 210nm ) a s d es cr ib ed by Underwood et al. (2002a) . Carbon balances were calculated as previously described ( Zabriskie et al. 1984 ; Causey et al. 2004 ). W hen LB was used as the fe rmentati on medium, cell mass was assu med to be p roduced exclusive ly from the complex media components and was not included in calculations of carbon balance.

PAGE 37

26 Results and Discussion Cloni ng and s equen cing o f budAB' . After PCR cloning of budAB' from K. oxytoca , the sequence was compared to closely related organisms (Figur e 2-2). Phylogene tic comparison of these genes was found to be c onsistent with other genes previously compare d ( Martinez et al. 2004 ), with the K. oxytoca genes being mo st closely related to those of R. terrigena (Figure 2-3). Media development for K. oxytoca P2. As a starting point, M9(+Fe) and 0.5% CSL+M media (a mmonia as t he sole ni trogen source) were te sted at pH 5.2 using strain P2 (Figure 2-4A and 2-4B) . LB medium at pH 5.2 was included a s a control to provide a benchmark for performance . As expected ( Martinez et al. 1999 ), LB medium supported the highest cell y ield and the most rapid ethanol production. Equivalent levels of ethanol were produced in 0.5%CSL+ M and LB. Replace ment of ammonia with urea resulted in a small dec rease i n ethanol producti on (Fig ure 4C an d 4D). I n gene ral, eth anol prod uctivity and cell yields incr eased with the richness of the media (L B>0.5%CSL+M>M9+F e), regardless of nitrogen sour ce (Table 2-3). Maximum cell densities were quite similar for M9(+Fe) and 0.5% CSL+ M media. This observation suggested that the lower le vels of nitrogen (19.0 mM in M9) and phosphate (10.3 mM in 0.5% CSL+M) in these respective media may be a dequate (Table 2-2). Based on the compositions of M9(+Fe) and 0.5% CSL+M , flask e xperiments were de signed t o evalua te diffe rent lev els of nut rients (Figure 2-5): phosphate (12-72 mM), mag nesium (0.25-1.0 mM), CSL (0-15 g/L ), and urea nitrogen (2.5-15 mM). Over the r anges tested, only CSL and urea had clea r optima, 5 g dry wt./L and 10 mM, r especti vely . Similar c oncentr ations of ethanol ( 13.9 ± 1.3 g /L were produced after 48 h with all levels of othe r components. Although it is possible that

PAGE 38

27CLUSTAL W (1.7) multiple sequence alignment PCR Primer GCTGAATCGGGTCAACATTT budAK. oxy -----------------------GCTGAATCGGGTCAACATTTATTTAACCTTTCTGATATTCGTTGAACGAGGA AGTGGGCAATGAACC... K. pne TGGTTTCTATATTGGAACTGTGAGCTGAATCGGGTCAACATTTATTTAACCTTTCTTATATTTGTTGAACGAGGA AGTGGTATATGAATC... R. ter TGGATTCTATATTGGAACTGAGAGCTGAATCGGGTCAATATTTATTTAATCTTTCTTATATTTGTTGAACGAGGA AGTGGATTGTGAATC... E. aer TGGATTCTATATTGGAACTCTCTGCTGAATCGGGTCAACATTTATTAAACCTTTATAAATAAAGTTGAA-GAGGA CGGGCATGATGATGC... *************** ******* ** **** * * ****** ***** * * *** * budBK. oxy ...CGGATCTGCATCCAGACAATCTTGATGCCGCCATTCGCTCAGTCGAAAAC TAA--GGAG ATTCTCG TGGATAATCAACATCAA... K. pne ...CTAATCTGCATCCCGATAATCTCGATGCCGCCATCCGTTCCGTAGAAAGT TAA--GGGG GT-CACA TGGACAAACAGTATCCG... R. ter ...CCGACCTGCATCCTGACAATCTCGATGCCGCTATTCGTGCGGTAGAAAAC TAA--GGAG CTTCAGA TGGACAAACCGCGTCAC... E. aer ...CCAACCTGCACCCCAGCAACCTTGATGCGGCTATCCGCGCCGTCGAAAAC TAACAGGAG AACTACCGTGAACAGTGAGAAACAG... * * ***** ** ** ** ***** ** ** ** * ** **** *** ** * ** * * PCR Primer TTTCGGTTTGTCCAGGTAGT K. oxy GGCGACCCGGTGGTGGCGCTGGGCG-CGCGGTAAAACGCGCCGATAAAGCCAAACAGGTCCATCA-------------------------K. pne GGCGACCCGGTGGTGGCCCTGGGCGGCGCGGTAAAACGCGCCGATAAAGCCAAACAGGTCCATCAGAGTATGGATACGGTGGCGATGTTCA R. ter GGCGACCCGGTGGTGGCGCTGGGCGGCGCGGTGAAGCGCGCGGATAAGGCCAAGCTGGTTCACCAAAGCATGGACACCGTGGCGATGTTCA E. aer ------------------------------------------------------------------------------------------Figure 2-2. Clustal W alignment of selected regions of the gene s encoding acetolact ate decarboxylase ( budA ) and acetolactate synthase ( budB ). PCR primers used in cloning are also shown. Start and stop codons are i ndicated in bold, ribosomal binding sites are underlined. Ov erall DNA sequence identity to K. oxytoca were 84, 84, and 80% for the sequences published for K. pneumoniae , R. terrigena and E. aerogenes . Amino acid sequences were 92, 91, and 88 % identical.

PAGE 39

28K. oxytoca (AY722056) R. terrigena (L04507) K. pneumoniae (NC 002941) E. aerogenes (L04506) 0.02K. oxytoca (AY722056) R. terrigena (L04507) K. pneumoniae (NC 002941) E. aerogenes (L04506) 0.02Figure 2-3. Phylogenetic tree (Kimura 2 parameters) derived fro m budAB’ DNA sequences. Accession numbers used are shown in parentheses.

PAGE 40

29 0 24 48 72 96 0 1 2 3 4 5 6Time (h)O.D.550nm 0 24 48 72 96 0 10 20 30 40 0 250 500 750 1000Time (h)Ethanol (g/L)Ethanol (mM)C D 0 24 48 72 96 0 2 4 6 8 10 12Time (h)O.D.550nm 0 24 48 72 96 0 10 20 30 40 0 250 500 750 1000Time (h)Ethanol (g/L)Ethanol (mM)B AFigure 2-4. Fermentation of K. oxytoca P2 in various media (90 g/L glucose). A and B. LB and media with ammonia nitrogen. Symbols: , LB; , M9 + Fe; , 0.5% CSL+M. C and D. Media with urea nitrogen. Symbols: , U-M9 + Fe; , U-0.5% CSL + M; , OUM1. Standard errors are included for data with n3.

PAGE 41

30 Strain Medium pH nEthanol EthanolFormate Lactate Succinate AcetateAcetoin +Carbon (mM) Yieldb (mM) (mM) (mM) (mM) 2,3 Butanediol Recoveryc,d (%) (mM) (% total) P2LB5.2284885<1258<131110 P2 M9+NH 4 5.2 2 761 76 <1 22 18 <1 24 96 P2 0.5%CSL+NH 4 5.2 2 831 83 <1 13 8 <1 23 101 P2M9+Urea5.2270871<13312<11889 P20.5%CSL+Urea5.2277678<11520<11995 P2OUM15.210825(65)83<110(5)13(5)9(3)72(20)101(6) BW21OUM15.24926(17)93<14(1)13(3)5(1)2(1)100(2) P2LB6.029981001137122539111 P2LB6.82979986628113438112 P2OUM16.8280681443017471696aValues are corrected for dilution by added base. Standard de viations are shown in parentheses for n values of 3 or more. bPercentage of theoretical yield based on total glucose (90 g/L). cIncludes unmetabolized glucose remaining after 72 hours. dWhere LB was used, cell mass was assumed to be produced exclusively from the complex media components as was not included in carbon recovery calculations.Table 2-3. Effect of budAB deletion on ethanol production and by-products (90 g/L glucose, 72 h)a.

PAGE 42

31 0.0 1.2 5.0 7.5 0 1 2 3 4 5 6 0 5 10 15 20Corn Steep Liquor (g dw/L)OD550nmEthanol (g/L) 2.5 5.0 10.0 15.0 0 1 2 3 4 5 6 0 5 10 15 20Urea (mM)OD550nmEthanol (g/L)A BFigure 2-5. Development of optimized medium (OUM1) for K. oxytoca P2 (flask-cultures, 48 h, 50 g/L glucose). Data are shown for growth (hashed bars) and ethanol (solid bars) in media with varying corn steep liquor (A) and urea (B). Standard errors are included for data with n> 3.

PAGE 43

32 lower co ncentra tions may be adeq uate, 12 m M PO 4 -3 , 1 mM MgSO 4 , and 1 mM Ca Cl 2 were selected for the optimized urea medium (OUM1) a s summarized in Table 2-2. Fermenta tion of gl ucose in OU M1 medi um. OUM1 (ur ea) wa s tested in pH-cont rolled fe rmentati ons at pH 5 .2 with 90 g /L g lucose. E thanol pr oduction i n this medium was slightly hig her than urea-containing formulations of M9(+ Fe) and CSL+M media (Figure 2-4C and 2-4D), c onfirming that higher levels of nitroge n, phosphate, and CSL were not necessar y. Ethanol titers were equivalent to those obtained with LB at pH 5.2 with strain P2. However, ethanol titers with all media were lower at pH 5.2 tha n previously repor ted ( Ohta et al. 1991a ; Wood and Ingram, 1992 ), albei t in rich me dia (LB) at more ne utral pH. This detrimental effect of low pH on ethanol production wa s confirmed for P2 fermentations at pH 6.8 with both L B and OUM1 media (Table 2-3). Althoug h by -produc ts were m ade by strain P2 in all media , the hig h level of products from the 2,3-butanediol pathway with OUM1 medium at pH 5.2 was unexpected and appear s relate d both to OU M1 composi tion and lo w pH. At pH 6.8, buta nediol+a cetoin levels were reduced while ace tate (and formate) levels increase d. In native strains of Klebsiella , enzyme activities conce rned with the production of these by -products are known to increase in response to low pH ( Blomqvist et al. 1993 ; Mayer e t al. 1995 ; Yang et al. 2000 ). At more neutral pH, acidogenic a ctivities such as pyruvate formate-lya se, acetat e kinase , and lac tate deh ydr ogena se produc e more a cidic pr oducts ( Tolan and Finn, 1987 ). A similar regulation appears to remain ac tive in strain P2 (Table 2-3; Figure 2-6). In both L B medium and OUM1 medium, fermentations at pH 5.2 contained a hig her pro por ti on o f ne ut ral by-p rod uct s (a cet oi n an d 2, 3-b ut ane di ol ) th an a t p H 6. 8. In LB medium, the levels of neutral byproducts remaine d relativ ely constant while ac idic

PAGE 44

33 byproducts decline d. In OUM1 medi um, the le vels of ne utral by -produc ts were 4 .5 fold higher at pH 5.2 than at pH 6.8 (Fig ure 2-6 A and B). Elimination of the budAB operon reduces by-products . After integration of the PCR product containing the budAB deletion into strain K. oxytoca strain M5a1 by electroporation in the presence of Red re combinase (pLOI 3421), ten clones were grown in OUM1 containing 5% glucose and scre ened for the absence of ac etoin and 2,3-butanediol. Deletion of budAB was also confirmed by PCR analysis. One deletion clone (B W15) was se lected a nd used a s a donor f or trans duction ( P1 phage ) into K. oxytoca P2. Ten resulting transductants were scree ned for acetoin and butanediol production. One was selected for further study, strain BW19 (Fig ure 2-7). The FRT-flanked tet gene w as subseq uently removed using F LP re combinas e (pFTK) to produce strain BW21. Const ruct ion of ) budAB increase d ethanol yie lds . In O UM1 at pH 5 .2 (90 g /L glucose), unwanted by -products from the 2,3-butanediol pathway (2 mol pyruvate pe r mol product) were produced from appr oximately 14% of the glucose available for ethanol producti on by strain P2 ( Table 23). Dele tion of g enes enc oding a cetolac tate sy nthase and acetolactate decarboxy lase (strain BW21) eliminated both by -products. Lactate a nd acetate levels were also lower in stra in BW21 than in P2, Figure 2-6. I n OUM1 medium at pH5.2, the decr ease in b yproducts using B W21 was ac companie d by a 12% inc rease i n ethanol titer and yield in compa rison to strain P2 (Table 2-3; Figure 2-8). Althoug h the growth of BW21 a nd P2 were essentia lly t he same, maximum and a verag e volumet ric rates of ethanol production were consistently higher for strain BW21 (Figur e 2-9).

PAGE 45

34 Ethanol 2CO Formate Lactate Succinate Acetate Acetoin + 2,3-Butanediol Cell Mass 0 1 2 3 4 5 6 7 8 9 10 11 30 50 70 0.0 0.5 1.0 1.5AFermentation Products (% Total Carbon)Cell Mass (% Total Carbon) Ethanol 2CO Formate Lactate Succinate Acetate Acetoin + 2,3-Butanediol Cell Mass 0 1 2 3 4 5 6 7 8 9 10 11 30 50 70 0.0 0.5 1.0 1.5BFermentation Products (% Total Carbon)Cell Mass (% Total Carbon) Ethanol 2CO Formate Lactate Succinate Acetate Acetoin + 2,3-Butanediol CellMass 0 1 2 3 4 5 6 7 8 9 10 11 30 50 70 0.0 0.5 1.0 1.5CFermentation Products (% Total Carbon)Cell Mass (% Total Carbon) Ethanol 2CO Formate Lactate Succinate Acetate Acetoin + 2,3-Butanediol Cell Mass 0 1 2 3 4 5 6 7 8 9 10 11 30 50 70 0.0 0.5 1.0 1.5AFermentation Products (% Total Carbon)Cell Mass (% Total Carbon) Ethanol 2CO Formate Lactate Succinate Acetate Acetoin + 2,3-Butanediol Cell Mass 0 1 2 3 4 5 6 7 8 9 10 11 30 50 70 0.0 0.5 1.0 1.5BFermentation Products (% Total Carbon)Cell Mass (% Total Carbon) Ethanol 2CO Formate Lactate Succinate Acetate Acetoin + 2,3-Butanediol CellMass 0 1 2 3 4 5 6 7 8 9 10 11 30 50 70 0.0 0.5 1.0 1.5CFermentation Products (% Total Carbon)Cell Mass (% Total Carbon) Figure 2-6. Distribution of carbon in the fe rmentation of 90 g/L glucose in OUM1. A) K. oxytoca P2,pH 6.8; B) K. oxytoca P2, pH 5.2; C) K. oxytoca BW21, pH 5.2.

PAGE 46

35 M5a1 BW15 P2 BW19 M5a1 BW15 P2 BW19 Figure 2-7. Color palate of Voges-Pr oskour assay for acetoin production. Parent strains, K. oxytoca M5a1 and P2 and their respective budAB derivatives are indicated.

PAGE 47

36 0 24 48 72 96 0 1 2 3 4 5 6Time (h)OD550nm 0 24 48 72 96 0 10 20 30 40 0 250 500 750 1000Time (h)Ethanol (g/L)Ethanol (mM)A BFigure 2-8. Effects of )Tj/T1_2 1 Tf-0.0005 Tc 12.018 0 0 12.0042 279.0267 164.5483 Tm(budAB on growth and ethanol production by ethanologenic strains of K. oxytoca (90 g/L glucose). A) Growth; B) Ethanol produc tion. Symbols for A and B: )Tj/T1_0 1 Tf-0.0002 Tc 0.0002 Tw 12.018 0 0 12.0042 208.4814 122.0533 Tm[(, strain BW21 ()Tj/T1_2 1 Tf-0.0003 Tc 12.018 0 0 12.0042 290.8042 122.0533 Tm(budAB ); , strain P2 (parent). Standard errors are in cluded for data with n)Tj/T1_0 1 Tf12.018 0 0 12.0042 413.9883 107.8883 Tm(3.

PAGE 48

37 P2, U-M9 P2, U-0.5% C SL+ M P2, O U M 1 B W 21, O U M 1 0 5 10 15 20 25 30 35yield maximum volumetric average volumetric 0.0 0.1 0.2 0.3 0.4 0.5Productivity (mMethanol/h)Yield [g ethanol/(g glucose)] P2, U-M9 P2, U-0.5% C SL+ M P2, O U M 1 B W 21, O U M 1 0 5 10 15 20 25 30 35yield maximum volumetric average volumetric 0.0 0.1 0.2 0.3 0.4 0.5Productivity (mMethanol/h)Yield [g ethanol/(g glucose)]Figure 2-9. Comparison of ethanol yield and productivity by K. oxytoca strains. Maximum volumetric productivity o ccurred early in fermentation, between 8 h and 24 h. Average pr oductivities were calculated for the initial 72 h. Ethanol yields were calculated after 72 h.

PAGE 49

38 Does pH affect PFL activity? In c onjunctio n with the i ncreas e in ace tolactat e synthase to redire ct carbon into the butanediol pathway , a reduction in cell mass (33%), was observed in strain P2 and was similar to strain BW21. Reductions in formate and acetate accumulation were also obser ved in both strains (Figure 2-6). The lac k of measured formate may be due to formate hy drogen ly ase activity (F HL). I n E. coli , FHL activity is known to be increa sed at low pH ( Böck and Sowers, 1996 ). A red uction in c ell mass and a cetate producti on would be also see n if py ruvate f ormate ly ase (PF L) or acetat e kinase (ACK) activities were reduced a t low pH. Consistent with a reduction in PFL (or ACK ) ac ti vi ti es w ere th e ob ser ved res ul ts in th e fe rme nt ati on o f x ylos e (F igu re 2 -10 ). It has been previously demonstrated, in E. coli , that functional PFL and ACK activities are required for anaerobic g rowth on xylose ( Hasona et al. 2004 ). In ethanologe nic E. coli strain KO11 (using 0.5% CSL + M, 9% xy lose, pH 6.5) it was demonstrated by Underwood et al. (2002a) a similar decline in cell mass was due to a limitation in carbon skeletons (from TCA cy cle intermediates) available for biosy nthesis. While similar limitations appear to occur in ethanologenic K. oxytoca , the apparent pH dependance of this limitation suggest a different mechanism. I t is well established that in many Gram po si ti ve (l ac ti c a ci d p ro du ci ng ) b ac te ri a t ha t P FL a ct iv it y is re gu la te d ( tr an sc ri pt io na ll y, translationally, and post-transla tionally) by pH ( Melchiorsen et al. 2002 ; Asanuma and Hino, 2000 ). Similar control in K. oxytoca may occur, b ut involve ment of AC K activit y cannot be ruled out. Conclusions The product yie lds and costs associated with the media are important factors in the economics of commodity che micals such as ethanol. Klebsiella oxytoca has the native

PAGE 50

39 0 24 48 72 96 0 1 2 3 4 5 6 7ATime (h)OD550nm 0 24 48 72 96 0 10 20 30 40 BTime (h)Ethanol (g/L)Figure 2-10. Effect of pH on the fermentation of 90 g/L xylose in OUM1 by K. oxytoca BW21. A) Cell mass and B) Ethanol produc tion. Closed circles ()Tj/T1_0 1 Tf-0.0008 Tc 0.0008 Tw 11.9955 0 0 11.9831 231.3638 97.7517 Tm()pH 6.0; Open circless () pH 5.2.

PAGE 51

40 ability to use ure a as a nit rogen source. On an equ ivalent n itroge n basis, ur ea is ty pically half the cost of ammonium salts. A comparison of the cost of the media used in our study (T ab le 34) sh ow s t he us e o f u re a a s a so ur ce of ni tr og en re du ce s m ed ia co st s b y $ 10%. The new medium developed in our study , OUM1, offers further potential savings fr om the low co ncentra tions of oth er salts a nd corn st eep liquo r. On a dr y w eight ba sis, OUM1 medium co nsists of 0. 5% CSL , 0.06% ur ea, and 0.2% inor ganic s alts plus f ermenta ble sugar. With this new medium co mbined wit h the incr eased e thanol y ields obta ined with K. oxytoca BW21 (42.6 g/L max.), the media costs for e thanol production would be 8.5¢ per gallon. The use of urea a s a nitrogen source has additional benefits. Unlike ammonium salts, the metabolism of urea does not contribute to the acidification of the medium ( Teixeira de Mattos and Neijssel, 1997 ) and thus reduce s the amou nt of base required for pH control. The low pH used in these fermentations is particularly appropriate for lignoce llulosic f eedstoc ks. Fung al cellul ases and xyla nases ty pically exhibit optim a around pH 5 ( Nieves et al. 1998 ). However, the pathway for butanediol (and acetoin) production is activated by low pH ( Johansen et al. 1975 ) leadin g to an in crease in by-products and de cline in ethanol yields. De le ti on s i n t he tw o g en es un iq ue ly i nv ol ve d i n t hi s p at hw ay ( budAB ) resulte d in an improv ed strain , BW21, tha t achiev ed 12% hig her etha nol y ield than the pare nt, strain P2. Ethano l product ion by BW21 at pH 5.2 in OUM1 (90 g/L glucose ) was ess entially complete after 4 8 h and exc eeded th at of the parent ( strain P2) in LB medium. Th ese improvements were likely due to the improved ability to balance NAD + /NADH in glyc olysis.

PAGE 52

41 The differences in K m for py ruvate b y th e compet ing enzy mes py ruvate de ca rb ox la se an d a ce to la ct at e s ynt ha se (0 .4 an d 8 mM ) r es pe ct iv el y ( Bringer-Mey er et al.1986 ; Yang et al. 2000 ) and the accumul ation ac etoin (a nd butane diol) sug gest relatively hig h intracellular py ruvate levels. The fermentation results from xy lose (at pH 6.0) are consiste nt with inc reased pyr uvate poo ls ( Underwood et al. 2002a ).

PAGE 53

42 Table 2-4. Estim ated cos t of media componen ts ($/m 3 ). Component LB M9(+Fe) U-M9(+Fe) 1% CSL+M U-1%CSL+M OUM1 Nonnitrog en Salts a 1.00 1.93 1.93 0.36 0.36 0.38 Nitrogen b 0.00 0.32 0.11 0.45 0.25 0.11 CSL c 0.70 0.70 0.70 Other d 90.00 Total 91.00 2.24 2.04 1.51 1.32 1.19 a For simplicity , non-nitrog en salts are assume d to be $200.00 pre 1000kg . b Nitrog en costs are as of April 18, 2 005 ( Chemical Mark et Reporter ). c CSL is assum ed to be $70.00 p er 1000kg. d LB com ponents Trypto ne and Yeast e xtract.

PAGE 54

43 CHAPTER 3 OVE R E XP RE SS ION O F TH E MU TAN T GLO BAL R EGU LATO RY P ROT EIN CRP(IN) I MPROVES PENTOSE USE IN MIXED SUGAR FERMENTATIONS Introduction Previous studies with ethanologenic K. oxytoca had shown that in fermentations containing mixtures of glucose, xylose, a nd/or arabinose, pentose use was slow, and incomple te in some c ases ( Bothast et al. 1994 ; Moniruzzaman et al. 1996 ). This trait was attributed to Crp-mediated catabolite repression ( Moniruzzaman et al. 1996 ). In enteric bacter ia, such a s K. oxytoca , the use of sugar is trancriptionally regulated by the cyclic-A MP receptor protein Crp ( Busby and K olb, 1996 ; Saier et al. 1996 ). The active form of Crp is a homodimer requiring cy clic-AMP (cAMP) for functionalization. As illustrated in Figure 3-1, cAMP is produced by adenylate cyclase (Cya). Cy a is activated by th e ph osp hor ylat ed f orm of e nz yme II A (EII A ~P) of the phosphotransterase sy stem (PTS). This occurs when sugars metabolized using the PTS, such as g lucose, are not present . As a re sult, cata bolite re pression occurs i n the pre sence of PTS sugar s. That is, the use of alternative sugars (ca rbon sources) is repressed ( Saier, 1998 ). Additional regulation is typic ally provided by sugar specific reg ulatory proteins (e .g., Xy lR, AraC). Many previous ef forts at relieving catabolite repre ssion have focused on mutations of the PTS ( Lindsay et al. 1995 ; Hernandez-Montalvo et al. 2001 ; Nichols et al. 2001 ). Many of these muta nts do coutilize gl ucose an d xylo se, but thi s comes a t the expen se of gl ucose utilization r ates ( Hernandez-Montalvo et al. 2001 ). There is a class of mutant Crps which do not require cAMP for functional dimerization. Many of these cAMP insensitive

PAGE 55

44 cAMP ATP EIIA~P Hpr ( ptsH ) EIIA( crr ) Hpr~P EI~P EI ( ptsI ) PEP Pyruvate EIIBC( ptsG ) Glc Glc6P cAMP Crp ( crp) Crp Transcriptional Activation Cya ( cya ) cAMP ATP EIIA~P Hpr ( ptsH ) EIIA( crr ) Hpr~P EI~P EI ( ptsI ) PEP Pyruvate EIIBC( ptsG ) Glc Glc6P cAMP Crp ( crp) Crp Transcriptional Activation Cya ( cya ) cAMP ATP EIIA~P Hpr ( ptsH ) EIIA( crr ) Hpr~P EI~P EI ( ptsI ) PEP Pyruvate EIIBC( ptsG ) Glc Glc6P cAMP Crp ( crp) Crp Transcriptional Activation Cya ( cya ) Figure 3-1. Transcriptional regulation by Crp.

PAGE 56

45 mutations, crp(in) , ( crp* ) have been previously characterized ( Melton et al. 1981 ; Harman et al. 1986 ; Moore, 1993 ). In all cases the mutations are within the region of the protein k nown to be involved i n functio nal dimer ization with cAMP ( Harman e t al. 1986 ). Our wo rk examine d the eff ects of th e over e xpression o f one of t hese pre viously charac terized Cr p(in) pr oteins (Eppler and Boos, 1999 ) on pentose use, in mixtures of sugar, by the ethanologenic, endogluca nase producing strain K. oxytoca SZ21. Materi als and Met hods Strains and pl asmids. Table 31 lists the o rganis ms and pla smids used for this work. Ethanologenic strains were maintained on L uria agar containing 2% g lucose and chloramphenicol (40 or 600 mg/L on alternate day s) under argon. Other strains wer e maintained on Luria ag ar plates without added sugar. When necessa ry for se lection or plasmid maintenance, 50 mg/L ka namycin, 50 mg /L apramy cin, or 12.5 mg/L tetrac ycline was added. Geneti c met hods. Stan dard m ethod s were u sed fo r PCR -based cloni ng, transformation, and genetic analy sis. Mutant and wild type crp s were c loned usin g PCR from K. oxytoca M5a1 and E. coli strains B and ET25. Primers were designe d to include th e e nt ir e c od in g r eg io n a s w el l a s t he na ti ve pr om ot er s, pu rc ha se d f ro m S ig ma Ge no sys (The Woodlands, TX), and were as follows; forwar d primer5'-TCC ACTGC GTCAA TTTTC -3', rev erse pr imer5'TCTT CTCC CTCG TTCC TG-3'. The PCR reaction had an initial melting at 94°C for 3 min followed by 30 cycles of melting at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 60 s. A final exte nsion at 72 °C for 10 mi n was also included . PCR produc ts were i mmediate ly recovered in pCR2.1-TOPO (I nvitrogen, Carlsbad, CA) according to the manufacturers

PAGE 57

46 Strain/PlasmidRelevant traits Source/Reference E. coli TOP10F’ hsdR mcrA lacZ)Tj/T1_0 1 Tf0.0052 Tc 0 9.96 -9.96 0 198.93 325.53 Tm(M15 endA recA ; F’ tet lacI Invitrogen E. coli DH5lacZ)Tj/T1_0 1 Tf0.0043 Tc 0 9.96 -9.96 0 211.53 278.8501 Tm(M15 recA Bathesda Res. Lab E. coli S17 pir Lab Stocks E. coli BW.T.Lab Stocks E. coli KO11 ( Z . m. pdc , adhB ), )Tj/T1_2 1 Tf-0.0022 Tc 0 9.96 -9.96 0 248.73 331.41 Tm(frd , cat Ohta et al. 1991b E. coli ET25 crp(in) Eppler and Boos, 1999 E. coli CA8445)Tj/T1_2 1 Tf0.0018 Tc 0 9.96 -9.96 0 273.33 261.0901 Tm[(crp)Tj/T1_2 1 Tf0.0018 Tc 0 9.96 -9.96 0 273.33 280.5301 Tm(cya E. coli genetic stock center K. oxytoca M5a1W.T.Lab Stocks K. oxytoca SZ21 ( Z. m pdc , adhB) , cat , ( E.c. celYcelZ ) Zhou et al. 2001 K. oxytoca BW17 M5a1 )Tj/T1_2 1 Tf0.0018 Tc 0 9.96 -9.96 0 309.93 287.0101 Tm[(crp , tet This study K. oxytoca BW18BW17 crp(in) , aac This study K. oxytoca BW23SZ21 crp(in) , aac This study K. oxytoca BW21 ( Z. m pdc , adhB) , cat , )Tj/T1_2 1 Tf0.0054 Tc 0 9.96 -9.96 0 346.53 346.7701 Tm(budBA This study (Chapter 2) K. oxytoca BW24BW21 crp(in), aac This study pCR2.1-TOPOTA PCR Cloning Vector, kan , bla Invitrogen pUC18 bla pLOI2065 FRTtet -FRT Underwood et al. 2002b pLOI2745SC101ts ori , kan This study (Chapter 2) pLOI2785R6K ori , kan This study pLOI3219 crp(in) PCR product from E. coli ET25 in pCR2.1-TOPOThis study pLOI3244 Eco RI crp(in) frag.from pLOI3219 in pUC18 Eco RI,This study pLOI3266 Sma I FRTtet -FRT frag. from pLOI2065 in pLOI3244 Ale IEco RV This study pLOI3268 Sma IBsr BI frag, from pLOI3266 in pLOI2745 Eco RVThis study pLOI3270 Sma I FRTaac -FRT form pLOI3421 in pLOI3244 Eco RVThis study pLOI3275 Pvu II crp(in) -FRTaac -FRT frag. From pLOI3270 in pLOI2785 Pme IThis study pLOI3420FRTaac -FRTThis study pLOI3421 red recombinase, aac This study (Chapter 2)Table 3-1. Strains and plasmids used in the study and construction of Crp(in) over-expressing K. oxytoca .

PAGE 58

47 instructions. Recovered plasmids were transformed into E. coli CA8445 and screened on Mackonkey a gar containing 1% lactose. Scr eening of wild-ty pe Crps required the inclusion of 1 mM cAMP in the media. Plasmids conferring a bright red c olony phenotype we re selected for further study . Gene replacements were initially constructed in K. oxytoca M5a1 using the strategy illustrated in Figure 3-2 and using methods de sc ri be d p re vi ou sl y ( Martinez-Moralles et al. 1999 ; Datsenko and Wanner, 2000 ). Chromosomally expressed crp(in) was tr ansdu ced to ethan ologen ic str ains b y P1 phage transduction ( Silhavay et al. 1984 ). Ethanol Fe rment ation. All fermentations were performed in 350 ml pH-stats as previously descr ibed ( Beall et al. 1991 ). Lu ria broth with 90 g /L tota l sugar was used in all ferm entation s. Mixtures c ontained either 3 0 g/L glucose plus 60 g /L xy lose or 30 g/L glucose, 30 g/L xylose, and 30 g/L arabinose. Automatic addition of 2N KOH was used to maintain pH at 6.0. Inocula (150 ml in a 250 ml erlenmey er flask) was grown for 16 hours at 35°C with mild shaking (120 rpm) in Luria broth conta ining 50 g/L g lucose. Cells were harvested by centrifugation (5000xg, 5 min.) to provide an initial concentration of 33 mg/L dry cell we ight (OD 550nm =0.1). Analytica l Methods. Cell mass was estimated by deter mining OD 550nm usin g a Bausch and Lomb Spectr onic 70 spectrophotometer where 1 OD 550nm corres ponds to a c ell de ns it y of 0. 33 g ( dr y we ig ht ) p er li te r. Et ha no l w as me as ur ed by g as ch ro ma to gr ap hy, aspreviously desc ribed ( Ohta et al. 1991b ), using a Varian model 3400X. Sugar a nd other fermentation products were determined by high-performance liquid chromatog raphy using a Hewlett-Packard model1090 series I I chromatogr aph and a Bio-Rad Aminex HPX-87H i on partiti on column ( 45°C’s; 4 mM H 2 SO 4 ; 0.4 ml/min.; 10 µl injection

PAGE 59

48 Ale I Eco RV crp(in) yhfK' crp' FRT tet FRT yhfK' Ale I crp(in) FRT aac FRT yhfK'Eco RI frag.from pLOI3244 Ale IEco RV + Sma I frag. from pLOI2065 Eco RV + Sma I frag. from pLOI3421 Red Recombinase Integration in M5a1Select for Tetrand lack of growth on glycerol Designated strain BW17 Red Recombinase Integration in BW17Select for Aprr,Tetsand restored growth on glycerol Designated strain BW18 P1 phage transduction to SZ21, BW21Strains BW23,BW24 Ale I Eco RV crp(in) yhfK' Ale I Eco RV crp(in) yhfK' crp' FRT tet FRT yhfK' crp' FRT tet FRT yhfK' Ale I crp(in) FRT aac FRT yhfK' Ale I crp(in) FRT aac FRT yhfK'Eco RI frag.from pLOI3244 Ale IEco RV + Sma I frag. from pLOI2065 Eco RV + Sma I frag. from pLOI3421 Red Recombinase Integration in M5a1Select for Tetrand lack of growth on glycerol Designated strain BW17 Red Recombinase Integration in BW17Select for Aprr,Tetsand restored growth on glycerol Designated strain BW18 P1 phage transduction to SZ21, BW21Strains BW23,BW24Figure 3-2. Scheme used for chromosomal replacement of wild-type crp with crp(in) .

PAGE 60

49 volume) equipped with dual refractive index and UV 210nm detecto rs, as pr eviously described ( Underwood et al. 2002a ). Carbo n balanc es were calcula ted as pr eviously described ( Zabriskie et al. 1984 ; Causey et al. 2004 ). Cell ma ss was ass umed to co me exclusively from the complex media components and wa s not included in calculations of carbon balance. Results and Discussion Sequence c omparis on. Comparing the DNA s equenc es clone d from K. oxytoca M5a1, E. coli B, and E. coli ET25 (Figure 3-3) reve aled a high degree of identity among all three strains ( 99%). All of the dif ferenc es betwe en M5a1 a nd ET25 we re in the 3' half of the coding region, allowing the strategy for chromosomal integration described previously. One hundre d percent identity wa s also found between K. oxytoca M5a1 and the sequence for E. coli K12 (not shown). When the de duced a mino acid sequenc es were compare d (Fig ure 3-4) the only chang es obser ved wer e in the Cr p(in) fr om E. coli ET25. The amino acid change A144T, alone and in combination with T127I, has previously been shown to confer a Crp(in) phenotype ( Harman et al. 1986 ). The identity of the a mino acid sequences between K. oxytoca M5a1 and E. coli B (as we ll as E. coli K12, not shown) is indicative of the conserved function of Crp in enteric bacter ia. Ferm entation o f sugar mi xtures. As previo usly observe d with etha nologe nic K. oxytoca ( Bothast et al. 1994 ), 90 g/ L gl uc os e, xyl os e a nd ar ab in os e w er e, in di vi du al ly, well utilized by strain SZ21 (Fig ure 3-5). As previously seen (Chapter 2) with K. oxytoca P2 yields appeare d to be reduced by significant by -product formation (Table 3-2). Not unexpectedly pentose use inc reased acetate production, a likely response to decreased net

PAGE 61

50CLUSTAL W (1.7) multiple sequence alignment (DNA) M5a1 crp 1 CTCCACTGCGTCAATTTTCCTGACAGAGTACGCGTACTAACCAAATCGCGCAACGGAAGGCGACCTGGGTCATGCTGAAGCGAGACACCAGGAGACACAAAGCGAAAGCT Ec B crp CTCCACTGCGTCAATTTTCCTGACAGAGTACGCGTACTAACCAAATCGCGCAACGGAAGGCGACCTGGGTCATGCTGAAGCGAGACACCAGGAGACACAAAGCGAAAGCT ET25 crp(in) CTCCACTGCGTCAATTTTCCTGACAGAGTACGCGTACTAACCAAATCGCGCAACGGAAGGCGACCTGGGTCATGCTGAAGCGAGACACCAGGAGACACAAAGCGAAAGCT ************************************************************************************************************** M5a1 crp 111 ATGCTAAAACAGTCAGGATGCTACAGTAATACATTGATGTACTGCATGTATGCAAAGGACGTCACATTACCGTGCAGTACAGTTGATAGCCCCTTCCCAGGTAGCGGGAA Ec B crp ATGCTAAAACAGTCAGGATGCTACAGTAATACATTGATGTACTGCATGTATGCAAAGGACGTCACATTACCGTGCAGTACAGTTGATAGCCCCTTCCCAGGTAGCGGGAA ET25 crp(in) ATGCTAAAACAGTCAGGATGCTACAGTAATACATTGATGTACTGCATGTATGCAAAGGACGTCACATTACCGTGCAGTACAGTTGATAGCCCCTTCCCAGGTAGCGGGAA ************************************************************************************************************** M5a1 crp 221 GCATATTTCGGCAATCCAGAGACAGCGGCGTTATCTGGCTCTGGAGAAAGCTTA TAACAGAGGA TAACCGCGCATGGTGCTTGGCAAACCGCAAACAGACCCGACTCTCG Ec B crp GCATATTTCGGCAATCCAGAGACAGCGGCGTTATCTGGCTCTGGAGAAAGCTTG TAACAGAGGA TAACCGCGCATGGTGCTTGGCAAACCGCAAACAGACCCGACTCTCG ET25 crp(in) GCATATTTCGGCAATCCAGAGACAGCGGCGTTATCTGGCTCTGGAGAAAGCTTA TAACAGAGGA TAACCGCGCATGGTGCTTGGCAAACCGCAAACAGACCCGACTCTCG ***************************************************** ******************************************************** M5a1 crp 331 AATGGTTCTTGTCTCATTGCCACATTCATAAGTACCCATCCAAGAGCACGCTTATTCACCAGGGTGAAAAAGCGGAAACGCTGTACTACATCGTTAAAGGCTCTGTGGCA Ec B crp AATGGTTCTTGTCTCATTGCCACATTCATAAGTACCCATCCAAGAGCACGCTTATTCACCAGGGTGAAAAAGCGGAAACGCTGTACTACATCGTTAAAGGCTCTGTGGCA ET25 crp(in) AATGGTTCTTGTCTCATTGCCACATTCATAAGTACCCATCCAAGAGCACGCTTATTCACCAGGGTGAAAAAGCGGAAACGCTGTACTACATCGTTAAAGGCTCTGTGGCA ************************************************************************************************************** M5a1 crp 441 GTGCTGATCAAAGACGAAGAGGGTAAAGAAATGATCCTCTCCTATCTGAATCAGGGTGATTTTATTGGCGAACTGGGCCTGTTTGAAGAGGGCCAGGAACGTAGCGCATG Ec B crp GTGCTGATCAAAGACGAAGAGGGTAAAGAAATGATCCTCTCCTATCTGAATCAGGGTGATTTTATTGGCGAACTGGGCCTGTTTGAAGAGGGCCAGGAACGTAGCGCATG ET25 crp(in) GTGCTGATCAAAGACGAAGAGGGTAAAGAAATGATCCTCTCCTATCTGAATCAGGGTGATTTTATTGGCGAACTGGGCCTGTTTGAAGAGGGCCAGGAACGTAGCGCATG ************************************************************************************************************** M5a1 crp 551 GGTACGTGCGAAAACCGCCTGTGAAGTGGCTGAAATTTCGTACAAAAAATTTCGCCAATTGATTCAGGTAAACCCGGAC A TTCTGATGCGTT TGTCTGCACAGATGGCGC Ec B crp GGTACGTGCGAAAACCGCCTGTGAAGTGGCTGAAATTTCGTACAAAAAATTTCGCCAATTGATTCAGGTAAACCCGGAC A TTCTGATGCGTC TGTCTGCACAGATGGCGC ET25 crp(in) GGTACGTGCGAAAACCGCCTGTGAAGTGGCTGAAATTTCGTACAAAAAATTTCGCCAATTGATTCAGGTAAACCCGGAC C TTCTGATGCGTT TGTCTGCACAGATGGCGC ******************************************************************************* *********** ****************** M5a1 crp 661 GTCGTCTGCAAGTCAC TTCAGAGAAAGTGGGCAACCTGGCGTTCCTCGACGTGACGGGCCGCATT G CACAGACTCTGCTGAATCTGGCAAAACAACCAGACGCTATGACT Ec B crp GTCGTCTGCAAGTCA C TTCAGAGAAAGTGGGCAACCTGGCGTTCCTCGACGTGACGGGCCGCATT G CACAGACTCTGCTGAATCTGGCAAAACAACCAGACGCTATGACT ET25 crp(in) GTCGTCTGCAAGTCAT TTCAGAGAAAGTGGGCAACCTGGCGTTCCTCGACGTGACGGGCCGCATT A CACAGACTCTGCTGAATCTGGCAAAACAACCAGACGCTATGACT *************** ************************************************* ******************************************** M5a1 crp 771 CACCCGGACGGTATGCAAATCAAAATTACCCGTCAGGAAATT GGTCAGATTGTCGGCTGTTCTCGTGAAACCGTGGGACGCATTCTGAAGATGCTG GAAGATCAGAACCT Ec B crp CACCCGGACGGTATGCAAATCAAAATTACCCGTCAGGAAATC GGTCAGATTGTCGGCTGTTCTCGTGAAACCGTGGGACGCATTCTGAAGATGCTT GAAGATCAGAACCT ET25 crp(in) CACCCGGACGGTATGCAAATCAAAATTACCCGTCAGGAAATT GGTCAGATTGTCGGCTGTTCTCGTGAAACCGTGGGACGCATTCTGAAGATGCTG GAAGATCAGAACCT **************************************** ****************************************************** ************** M5a1 crp 881 CACGGTAAAA GATCTCCGCACCATCGT CGTTTACGGCACTCGTTAATCCCGTCGGAGTGGCGCGTTACCTGGTAGCGCGCCATTTTGTTTCCCCCGATGTGGCGCAGACT Ec B crp CACGGTAAAA GATCTCCGCACCATCGT CGTTTACGGCACTCGTTAATCCCGTCGGAGTGGCGCGTTACCTGGTAGCGCGCCATTTTGTTTCCCCCGATGTGGCGCAGACT ET25 crp(in) CACGGTAAAG GATCTCCGCACCATCGC CGTTTACGGCACTCGTTAATCCCGTCGGAGTGGCGCGTTACCTGGTAGCGCGCCATTTTGTTTCCCCCGATGTGGCGCAGACT ********* **************** *********************************************************************************** M5a1 crp 991 GATTTATCACCCCGATATCAACTATGCACTTCGACAAACGCTGGTGCTATGTTTGCCCGTGGCCGTTGGGTTAATGCTTGGCGAATTACGATTCGGTCTGCTCTTCTCCC Ec B crp GATTTATCACCCCGATATCAACTATGCACTTCGACAAACGCTGGTGCTATGTTTGCCCGTGGCCGTTGGGTTAATGCTTGGCGAATTACGATTCGGTCTGCTCTTCTCCC ET25 crp(in) GATTTATCACCCCGATATCAACTATGCACTTCGACAAACGCTGGTGCTATGTTTGCCCGTGGCCGTTGGGTTAATGCTTGGCGAATTACGATTCGGTCTGCTCTTCTCCC ************************************************************************************************************** M5a1 crp 1101 TCGTTCCTG Ec B crp TCGTTCCTG ET25 crp(in) TCGTTCCTGFigure 3-3. Clustal W alignment of clone d DNA sequence containing coding regions for crp from K. oxytoca M5a1, E. coli B and crp(in) from E. coli ET25. Nucleotides differing in E. coli B are underlined. Nucleotides differing in E. coli ET25 are bold. Start and stop codons are ita licized. Ribosomal binding sites are double underlined.

PAGE 62

51CLUSTAL W (1.7) multiple sequence alignment M5a1 Crp 1 MVLGKPQTDP TLEWFLSHCH IHKYPSKSTL IHQGEKAETL YYIVKGSVAV LIKDEEGKEM Ec B Crp MVLGKPQTDP TLEWFLSHCH IHKYPSKSTL IHQGEKAETL YYIVKGSVAV LIKDEEGKEM ET25 Crp(in) MVLGKPQTDP TLEWFLSHCH IHKYPSKSTL IHQGEKAETL YYIVKGSVAV LIKDEEGKEM ********** ********** ********** ********** ********** ********** M5a1 Crp 61 ILSYLNQGDF IGELGLFEEG QERSAWVRAK TACEVAEISY KKFRQLIQVN PDILMRLSAQ Ec B Crp ILSYLNQGDF IGELGLFEEG QERSAWVRAK TACEVAEISY KKFRQLIQVN PDILMRLSAQ ET25 Crp(in) ILSYLNQGDF IGELGLFEEG QERSAWVRAK TACEVAEISY KKFRQLIQVN PDLLMRLSAQ ********** ********** ********** ********** ********** ** ******* M5a1 Crp 121 MARRLQVTSE KVGNLAFLDV TGRIAQTLLN LAKQPDAMTH PDGMQIKITR QEIGQIVGCS Ec B Crp MARRLQVTSE KVGNLAFLDV TGRIAQTLLN LAKQPDAMTH PDGMQIKITR QEIGQIVGCS ET25 Crp(in) MARRLQVISE KVGNLAFLDV TGRITQTLLN LAKQPDAMTH PDGMQIKITR QEIGQIVGCS ******* ** ********** **** ***** ********** ********** ********** M5a1 Crp 181 RETVGRILKM LEDQNLISAH GKTIVVYGTR Ec B Crp RETVGRILKM LEDQNLISAH GKTIVVYGTR ET25 Crp(in) RETVGRILKM LEDQNLISAH GKAIAVYGTR ********** ********** ** * ***** Figure 3-4. Clustal W alignment of Crps from K. oxytoca M5a1, E. coli B, and Crp(in) from E. coli ET25. Differing amino acids are indicated in bold.

PAGE 63

52 0 24 48 72 96 0 5 10 15Time (h)OD550nm 0 24 48 72 96 0 10 20 30 40Time (h)Ethanol (g/L)A BFigure 3-5. Growth (A) and ethanol production by K. oxytoca SZ21 from glucose ()Tj/T1_0 1 Tf-0.0005 Tc 0.0005 Tw 12 0 0 12 303.21 86.8501 Tm[(), xylose ()Tj/T1_0 1 Tf-0.0014 Tc 0.0014 Tw 12 0 0 12 362.25 86.8501 Tm(), and arabinose ().

PAGE 64

53 Sugar aEthanolYield bLactateSuccinateAcetateAcetoin+Butanediol% Carbon (mM)(mM)(mM)(mM)(mM)Recovery cGlucose815.979.622.023.620.439.693.6 Xylose 760.7 80.0 5.4 44.9 47.6 26.5 95.6 Arabinose746.876.811.928.440.167.698.8 Results are average values for duplicate fermentations. a 90 g/L total suar added. b % of theoretical. c % of starting sugar accounted for at e nd of fermentation including products and residual sugar. Cell mass is not included and is assumed to come solely from complex mediaTable 3-2. Summary of fermentation products and carbon balance by K. oxytoca SZ21.

PAGE 65

54 ATP production in glyc olysis. Overall, the distribution of the by -products appeared dependant on the sugar metabolized. I n fermentations containing glucose (30 g /L) and xyl os e ( 60 g/ L), gl uc os e i s c on su me d r ap id ly ( # 16 h). Xy lose was t hen slowly used with half of th e initial xy lose rema ining unu sed afte r 72 h (F igure 3 -6A). Whe n arabin ose is included (30 g/L each sug ar), ag ain gluc ose was d epleted within 16 ho urs, ara binose follows c losely and is go ne within 2 4 hours. X ylo se appea red to be co-meta bolized wit h arabinose albeit more slowly (Figure 3-6B). This phenomenon w as likely due to the transpor t of xy lose by the ara binose pr oton sy mport pro tein Ara E, known to be in K. oxytoca ( Shatwell et al. 1995 ), as previously de monstrated in E. coli ( Hasona e t al. 2004 ). Although K. oxytoca may not conta in the ana logous xy lose tran sporter Xy lE ( McClelland et al. 2000 ; Laikova et al. 2001 ), Hasona et al. (2004) showed it was not an effective transporter in anaer obic cultures of E. coli . Overexpr ession of Cr p(in) im proves sug ar consum ption. To see the effects of plasmid-based expression of Crp(in), fermentation of identical sugar mixtures was used, strains harboring the cloning vector pCR2.1-TOPO wa s used as a control. Considering the relatively g ood results seen with three sugar mixtures (Figure 3-6B) it was not surprisin g that th e improve ment from Crp(in) e xpression w as most pro nounced in mixtures of glucose and xylose (F igure 3-7). I n two sugar mixtures, expression of Crp (in ) re sul ted in mo re r api d x ylos e us e an d ha lf a s m uch xyl ose rem ain in g at 72 h . In Crp(in) e xpressing strains, i t appear s that xy lose cons umption co mmenced immediate ly after g lucose is deplete d. In control f ermenta tions, af ter glu cose dep letion, the re was a lag of about eight hours before xylose consumption began. Of pa rticular note was the increa se in xy lose cons umption, e xceeding that obse rved in st rain SZ 21 without p lasmids.

PAGE 66

55 0 24 48 72 0 10 20 30 40 50 60 0 10 20 30 40 Time (h)Sugar (g/L)Ethanol (g/L) 0 24 48 72 0 10 20 30 0 10 20 30 40 Time (h)Sugar (g/L)Ethanol (g/L)A BFigure 3-6. Ethanol production ()Tj/T1_0 1 Tf-0.0015 Tc 0.0015 Tw 12 0 0 12 312.45 105.5701 Tm[() by K. oxytoca SZ21 from sugar mixtures. A) 30 g/L glucose () and 60 g/L xylose (). B) 30 g/L glucose (), 30 g/L xylose (), and 30 g/L arabinose ()Tj/T1_0 1 Tf12 0 0 12 452.61 77.2501 Tm()

PAGE 67

56 0 24 48 72 0 10 20 30 40 50 60 0 10 20 30 40Time (h)Sugar (g/L)Ethanol (g/L) 0 24 48 72 0 10 20 30 40 50 60 0 10 20 30 40Time (h)Sugar g/LEthanol (g/L)A BFigure 3-7. Effect of Crp(in) over expression on ethanol production by K. oxytoca SZ21 from 30 g/L glucose and 60 g/L xylose. A) K. oxytoca SZ21 pCR2.1-TOPO. B) K. oxytoca SZ21 pLOI3219. Ethanol ()Tj/T1_0 1 Tf-0.0011 Tc 0.0011 Tw 12.0142 0 0 12.0037 293.9595 82.1961 Tm(); Glucose (); Xylose (X).

PAGE 68

57 In three suga r mixtures (Figure 3-8) the maintenance of the high-copy cloning vector p CR2.1-TOPO signific antly reduce d sugar consumpti on rates compare d to strain SZ21 alo ne (Fig ure 3-6B ). Howev er, the r ates of c onsumption were inc reased for all three sugars in Crp(in) expressing strains, when c ompared to strains with plasmid vector alone . In both cases e thano l prod uctiv ity and yield s mir ror su gar cons umpt ion, altho ugh ethanol yields wer e further reduced by the high levels of by -products formed (Table 3-3). In pa rt the re duced e thanol pr oduction r esulted f rom the me tabolic b urden of protein expression from high copy plasmids but, the levels of products from the 2,3-butanediol pathway we re of particular note. Redirection of 93 to 148 mM acetoin plus 2,3-buta nediol to e thanol co uld resul t in 183 to 29 6 mM (8 to 11 g/L ) addition al ethan ol. In similar experiments, using three sug ar mixtures, with the analogous ethanologen E. coli KO11 a dif ferent pattern of suga r use wa s observe d (Fig ure 3-9) . Again xylos e and arabinose appeared to be cometabolized. In strain KO11 pCR2.1-TOPO, xy lose was used more rapidly a nd more completely c ompared to K. oxytoca SZ2 1 pC R2 .1TOP O. In the presence of Crp(in), overall pentose use was similar in both organisms, suggesting a n additional rate-limiting step elsewhere in the pentose pathwa ys. However , in E. coli KO11 pL OI32 19, the inc reased rates in xy lose meta bolism app eared t o be at the expense of arabinose use. In K. oxytoca SZ21 pLOI 3219, the improvements appear in large par t to have be en due to i mproved u se of ar abinose. The hig her etha nol conce ntrations in E. coli KO11 was a direct result of the 2,3-butanediol pathway , absent in E. coli . A preference for arabinose by K. oxytoca . A priority in pentose utilization has been previously de scribed in E. coli K12 ( Kang et al. 1998 ) and als o in ethan ologen ic E. coli KO11 ( Moniruzzaman and Ingram, 1998 ). In the former study it was shown that

PAGE 69

58 0 24 48 72 0 10 20 30 0 10 20 30 40 50Time (h)Sugar (g/L)Ethanol (g/L) 0 24 48 72 0 10 20 30 0 10 20 30 40 50Time (h)Sugar (g/L)Ethanol (g/L)A BFigure 3-8. Effect of Cr p(in) over expression on ethanol production by K. oxytoca SZ21 from 30 g/L glucose, 30 g/L xylose, and 30 g/L arabinose. A) K. oxytoca SZ21 pCR2.1-TOPO. B) K. oxytoca SZ21 pLOI3219. Ethanol ()Tj/T1_0 1 Tf-0.0011 Tc 0.0011 Tw 12 0 0 12 221.73 71.8501 Tm(); Glucose (); Xylose (X ); Arabinose ()Tj/T1_0 1 Tf-0.003 Tc 12 0 0 12 423.57 71.8501 Tm().

PAGE 70

59 SugaraPlasmidnEthanolYieldbLactateSuccinateAcetateAcetoin+Butanediol% Carbon (mM)(mM)(mM)(mM)(mM)RecoverycGlc,XylpCR2.1-TOPO4319.6(25.3)32.7(2.3)14.5( 2.6)19.6(3.5)43.4(4.5)93.1(15.0)102.6(4.5) pLOI32194534.2(9.5)53.5(9.6)47.4(3.2) 25.0(3.3)40.3(2.7)130.0(20.5)105.3(6.2) Glc,Xyl,ArapCR2.1-TOPO2646.168.764.922.764.2147.9105.7 pLOI32194698.0(34.2)70.1(3.6)29.6(6.6) 24.2(6.2)59.3(16.1)111.3(6.2)100.1(2.3) Where applicable, standard devia tions are shown in parentheses. aSugar concentrations total 90 g/L. Either 30 g/L glucose + 60 g/L xylose or 30 g/L glucose + 30 g/L xyose + 30 g/L arabinose. bPercentage of theoretical yield based on total added sugar (90 g/L). cIncludes unmetabolized sugar at 72 hours.Table 3-3. Effect of Crp(in) over expression by K. oxytoca SZ21 (90 g/L sugar, 72 h).

PAGE 71

60 0 24 48 72 0 10 20 30 0 10 20 30 40 50Time (h)Sugar (g/L)Ethanol (g/L) 0 24 48 72 0 10 20 30 0 10 20 30 40 50Time (h)Sugar (g/L)Ethanol (g/L)A BFigure 3-9. Effect of Crp(in) over expression on ethanol production by E. coli KO11 from 30 g/L glucose, 30 g/L xylose, and 30 g/L arabinose. A) E. coli KO11 pCR2.1-TOPO B) E. coli KO11 pLOI3219. Ethanol ()Tj/T1_0 1 Tf-0.0011 Tc 0.0011 Tw 12 0 0 12 355.77 85.5701 Tm(); Glucose (); Xylose (X ); Arabinose ()Tj/T1_0 1 Tf-0.003 Tc 12 0 0 12 279.69 71.4101 Tm().

PAGE 72

61 arabinose is used preferentially to other pentoses (xylose and ribose) . The fermentation of th re e s ug ar mi xt ur es by K. oxytoca SZ21 revealed a similar priority , a strong preference for ar abinose was seen in all cases, further amplified by the expression of Crp(in). Xylose use w as likely due to nonspecific transport by the ar abinose transport system. Chromosomally expressed Crp(in) im proves xylose use in strain SZ21 but not BW21. Since the use of xy lose appe ared to b e more pr oblematic than ara binose in K. oxytoca , mixtures of glucose (30 g/L) and xylose (60 g/L ) was used to examine chromosomal integrants of Crp(in) . As shown in Figu re 3-6A , ferme ntations u sing str ain SZ21 had ~30 g/L xy lose remaining after 72 hours. I n fermentations with Crp(in) expresse d on a plas mid (pL OI32 19), the r esidual xy lose was r educed to ~15 g/ L (Figure 3-8). When the crp(in) chromoso mal integ rant K. oxytoca BW23 was used, residual xylose concentra tions are further reduced to ~10 g/L (Figure 3-10). By -product formation by strain B W23 was similar to SZ21 (Table 3-4). Relief of plasmid burden resulted in a significant reduction (~40%) in ac etoin plus 2,3-butanediol and proportional increases in ethanol production, compared to plasmid-based e xpression. To further reduce acetoin plus 2,3-butanediol for mation, crp(in) was transduced into the budAB strain BW21. Both strain BW21 and BW24 had improved ethanol producti on and re duced by -produc t formati on compar ed to stra ins SZ21 a nd BW23 Ta ble 3-4). The addition of crp(in) to strain B W21 did not r esult in fu rther imp rovemen t in strain BW24 (Table 3-4; Figure 3-11).

PAGE 73

62 0 24 48 72 0 10 20 30 40 50 60 0 10 20 30 40 Time (h)Sugar (g/L)Ethanol (g/L)Figure 3-10. Ethanol production by K. oxytoca BW23 from 30 g/L glucose and 60 g/L xylose. Ethanol ()Tj/T1_0 1 Tf-0.0013 Tc 0.0013 Tw 12 0 0 12 243.33 171.8101 Tm(); glucose (); xylose().

PAGE 74

63 0 24 48 72 96 120 0 10 20 30 40 50 60 0 10 20 30 40 Time (h)Sugar (g/L)Ethanol (g/L) 0 24 48 72 96 120 0 10 20 30 40 50 60 0 10 20 30 40 Time (h)Sugar (g/L)Ethanol (g/L)A BFigure 3-11. Ethanol production by budAB strains of K. oxytoca from 30 g/L glucose and 60 g/L xylose. A) strain BW21 and B) Strain BW24. Ethanol ()Tj/T1_33 1 Tf-0.0013 Tc 0.0013 Tw 12 0 0 12 280.89 83.8501 Tm[(); glucose (); xylose ().

PAGE 75

64 StrainnEthanolYieldaFormateLactateSuccinateAcetateAcetoin +Butanediol Carbon Balanceb(mM)(%)(mM)(mM)(mM)(mM)(mM)% SZ214524.3(67.3)52.6(6.7)24.0(4.8)12.2(7.1)28.5(9.7)50.4(12.7)72.0(8.3)103.9(10.6) BW23 2 619.7 62.1 23.2 19.1 31.7 47.9 79.0 90.7 BW213702.6(77.5)70.4(7.8)-c4.1(0.0)13.7(2.1)15.8(2.6)8.9(0.7)107.2(4.2) BW243710.5(27.5)71.2(2.8)-4.1(0.6)13.7(0.4)19.5(2.1)7.6(0.4)107.3(3.3) Where applicable, standard deviations are shown in parentheses. aPercentage of theoretical yield based on total added sugar (90 g/L). bIncludes unmetabolized sugar at 72 hours. cNot detected. Table 3-4. Effect of chromosomal Crp(in) expression and budAB deletion on product formation after 96 hours from 30 g/L glucose plus 60 g/L xylose.

PAGE 76

65 Conclusions The lack of improved xylose use, in budAB strai ns was somew hat su rpris ing. However, in E. coli , it is known that many of the inte rmediate s and prod ucts of g lyc oly sis can hav e regu latory effec ts on cent ral car bon metab olism ( Suzuki, 1969 ; Frankel, 1996 ). As described in Chapter 2, in Klebsiella , acetate play s a role in the regulation of the butanediol pathway a s a transcriptional co-activator as well as an a llosteric affecter of en zym e a ct iv it y. Additiona lly, the pheno menon of i nducer e xclusion ha s been pr eviously demonstrated in E. coli . Inducer exclusion occurs whe n the unphosphorylated for m of enz yme IIA glc , e nc od ed by ptsG , interacts with the non-PTS sugar permeases spec ific for lac to se, mal to se, mel li bi ose , an d ra ffi nos e, i nhi bi ti ng s uga r up tak e. E nz yme IIA glc has also bee n shown to b ind gly cerol ki nase, thu s inhibiting gly cerol me tabolism ( Brückner and Titgemey er, 2002 ). In this study it was fou nd (in mixtur es of sug ars) pe ntose consumption generally did not commence until glucose was depleted, even in strains expressing Crp(in). This combined with the lack of further improvements in budAB strains suggests additional regulation of pe ntose metabolism. The ratio of ADP/ATP has also bee n shown to a ffect th e rate o f gluco se uptake and gly coly sis ( Koebman n et al. 2002 ). Pentose metabolism would increase ADP/ATP ratios. Additionally , the intrace llular re dox state, in the form o f NADH le vels, neg atively affec ts both phosphotransacety lase ( Suzuki, 1969 ) and citrate sy nthase ( Underwood et al. 2002a ). With the impr oved gl ucose fe rmentati on (gly coly sis) rate s we saw w ith strain BW21 in Cha pter 2 of our study , the leve ls and ra tios of the se interm ediates would like ly be altered. Although these ty pes of regulation have not been identified in the xy lose

PAGE 77

66 metabolism of Klebsiella , it is an area that has not been extensively studied beyond the transcr iptional r egulat ion by CRP and Xy lR. The ad ditional su bstrates availab le to K. oxytoca make other regulatory circuits possible. As the use of hemicellulosic feed st oc ks in cr ea se , t hi s m ay r ep re se nt a f ut ur e a re a o f s tu dy.

PAGE 78

67 CHAPTER 4 CON ST RUC TION OF E NDO GLUC ANA SE PR ODU CING , ET HAN OLOG ENIC STRAINS OF Klebsiella oxytoca LACKI NG GENES FOR 2,3-BUTANEDI OL PRODUCTION Introduction The utility of co-produc tion of bacterial endoglucanases in re ducing the requirement for commercial cellulases has be en previously demonstra ted ( Wood and Ingram, 1992 ; Wood, 1997 ; Wood et al. 1997b ). The most effective of these were found to be those from Erwinia chrysanthemi ( Wood et al. 1997b , Zhou et al. 2001 ). The combination of CelY and CelZ, working sy nergistically , allowed for the direct conversion of amorp hous cell ulose to e thanol ( Zhou and Ing ram, 2001 ). CelY a lone, wa s shown to provide t he gre atest be nefit whe n used in c onjunctio n with comm ercial c ellulase prepar ations ( Zhou et al. 2001 ) in the simultaneous saccharification and fermentation (SSF) of cry stalline c ellulose. Our work aimed to c ombine the producti on of thes e endoglucanases with the improvements observed in the e limination of the 2,3-butanediol pathway (Cha pter 2). Materi als and Met hods Strains, plasmids, and media. Strains a nd plasmid s used in ou r study are liste d in Table 4-1. Strains were maintained on L uria agar plates without added sug ar. Antibiotics wer e ad ded as r equ ire d fo r se lec ti on: amp ici ll in , 50 mg/ L; ap ram ycin , 50 mg/ L; kanamycin, 50 mg /L; tetracy cline, 12.5 mg/L; and spectinomy cin, 50 mg/L. Ethanol producin g strai ns were maintaine d on Lu ria ag ar plate s which a lso conta ined 2% g lucose and 40 or 600 mg/L chlora mphenicol (alternating daily , under argon). Aga r plates (used

PAGE 79

68 Strain/PlasmidRelevant traits Source/Reference E. coli DH5)Tj/T1_1 1 Tf0.0005 Tc 0 9.96 -9.96 0 232.29 201.4501 Tm[(lacZ)Tj/T1_0 1 Tf0.0043 Tc 0 9.96 -9.96 0 232.29 225.3301 Tm(M15 recABathesda Research Labs E. coli S17 pir Lab Stocks K. oxytoca SZ21 ( Z.m pdc , adhB) , cat , ( Er.c. celYcelZ ) Zhou et al. 2001 K. oxytoca SZ22 SZ21 celZ , aac Zhou et al. 2001 K. oxytoca BW15 K. oxytoca M5a1 budAB ::FRTtet -FRTThis study (Chapter 2) K. oxytoca BW21 ( Z.m pdc , adhB) , cat , budAB This study (Chapter 2) K. oxytoca BW32BW21, celY , kan This study K. oxytoca BW33 SZ21 budAB , tet This study K. oxytoca BW34BW21 , celY This study K. oxytoca BW35 SZ21 budAB This study pCPP2006 spm , ~40 kbp frag. Containing out genes from E. chrysanthemi He et al. 1991 pFT-A bla , FLP recombinase Posafi et al. 1997 pFT-K kan , FLP recombinase Posafi et al. 1997 pLOI2224 R6K ori , kan , FRT flanked MCS ( Asc I) Martinez-Morales et al. 1999 pLOI2348 celY from E. chrysanthemi ; 3 kbp M5a1 chromosomal fragment Zhou et al. 2001 pLOI3420 Red Recombinase, RepA ori , aac This study (Chapter 2) pLOI3421FRTaac -FRTThis study (Chapter 3) pLOI3290 Dra I aac frag. from pLOI3421 in pFT-A Cla IThis study pLOI3293 celY , M5a1 glgP frag. ( Asc I) from pLOI2348 in pLOI2224 Asc IThis study Table 4-1. Strains and plasmids used for the study and construction of budAB , endoglucanase producing strains of K. oxytoca .

PAGE 80

69 in the scr eening of endog lucanas e produc tion also c ontained 3 g/L low-visc osity carboxymethy lcellulose (CMC). CMC plates were stained with Congo red ( Wood et al. 1988 ) after overnight grow th at 37°C. Strains containing pLOI 3420 were incubated at 30°C, all o thers we re mainta ined at 37 °C. Geneti c proce dures. Stan dard m ethod s were u sed fo r PCR -based gene clo ning, plasmid co nstructio ns, and g enetic a naly ses ( Miller, 1992 ; Ausubel et al. 1987 ; Sambrook and Russel, 2001 ). Methods for integration, chromosomal deletions, integra tion, and the use of removable antibiotic resistance ge nes have been previously described ( Datsenko and Wanner, 2000 ; Martinez-Morales et al. 1999 ; Zhou et al. 2003 ; Causey et al. 2004 ). Escheric hia coli DH5 " was used for most constructions. Escheric hia coli S17 was used for the construction of plasmids containing an R6K origin of re plication. Phage P1 was us ed fo r g en er al iz ed tr an sd uc ti on ac co rd in g t o t he me th od s d es cr ib ed by Silhavay et al. (1984) . Plasm id and strain c onstruct ion. The plasmid used for direct integration of the gene celY encoding the endog lucanas e CelY fr om E. chrysanthemi was constructed by removing a 5.7 kbp Asc I fragment from p LOI2348 con taini ng celY , behind a surrog ate promoter , and a c hromosoma l frag ment from K. oxytoca M5a1 to target integration (previously deter mined to contain glgP , a gene in the gly cogen sy nthesis pathway) ( Zhou et al. 2001 ) followed by lig ation in the Asc I site of pL OI2224 ( MartinezMorales et al. 1999 ) to create pLOI 3293. Integra tion of pLOI 3293 in the budAB strain K. oxytoca BW21 was facilitated by the expression of 8 Red Recombinase from pLOI 3420. Resultant integrants were cured of pL OI3420 by outgrowth at 39°C, selected for kanamycin re sistance and subsequently screened for endogluca nase production and

PAGE 81

70 retention of resistance to high-level chlora mphenicol resistance (600 mg/L ). Several isolates w ere fur ther test ed for e thanol pr oduction i n pH contr olled fe rmentati ons in OUM1 (Chapter 2) containing 90 g/L glucose, all were found to be similar. One wa s selected for further study and designated strain BW32. Strains c ontaining two endog lucanas e gene s from E. chrysanthemi , celYcelZ , and budAB wa s c on st ru ct ed es se nt ia ll y as de sc ri be d i n C ha pt er 2 o f t hi s s tu dy. Br ie fl y, K. oxytoca strain BW15 was used as a donor strain for the P1 phage transduction of a deletion in the butanediol operon ( budA ’-FRTtet -FRT-‘ budB ’) to strain SZ21 (Zhou et al. 2001 ). Isolation and screening of transductants were identical to those used above except tetracyc line was used for selection. Strain BW33 was carried on f or further work. To facilitate the removal of kan from strain BW32, pFT-A (FL P recombinase) was modified by the addition of an a pramycin re sistance gene, aac . Plasmid pFT-A was linearized by restric tion digestion with Cla I, followed by treatment with the Klenow fragment of E. coli DNA po lymerase. A 1.6 k bp fragm ent co ntain ing aac from pLOI 3420 was ligated to the linearized pFT-A cre ating pLOI 3290. Removal of tet from strain BW33 used the previously de scribed pFT-K ( Posafi et al. 1997 ). Kan s and Tet s strains were rescreened a s above for endoglucanase pr oduction, chloramphenicol resistance, and ethanol production, resulting strains wer e designated BW34 and BW35 respec tively . Inte gratio ns were also conf irmed usin g PCR. Ferm entation. Seed cultures (150 ml in 250 ml flasks) were grown for 16 h at 35°C (120 rpm) in the same media used in pH controlled fermentations but contained 50 g/L glucose . Cells were harvested by centrifugation (5000 x g, 5 min) and used as inocula to provide an initial concentration of 33 mg/L dry cell we ight (OD 550nm =0.1).

PAGE 82

71 Fermentation vessels were previously described ( Beall et al. 1991 ) and contained an initial vol ume of 350 ml. Glucos e ferme ntations c ontained 90 g/L sugar a nd were in OUM1 (de scribed in Chapter 2 of this wo rk). SSFs containe d 100 g/L Sigmace ll 50 in Luria broth and SSCFs contained 45 g /L Sigmacell 50 plus 40 g/L xylose in OUM1. Cultures were incubated at 35°C (150 rpm). I n fermentations with strain BW35 pCPP2006, spectinomycin was added for pla smid maintenance. All other fermentations contained no added antibiotics. Broth of glucose f ermentations and SSFs were maintaine d at pH 5.2 by t he automa tic additi on of 2N K OH. For improved xylos e fermentation, SSCFs were maintained at pH 5.8. Results and Discussion Glucose ferme ntation. To compare the fermentation performanc e of strains expressing recombinant endoglucanase (s), final strains were used in the fermentation of 90 g/L glucose in OUM1. T he CelY pr oducing strain B W34 was eq uivalent to strain BW21 in bot h growt h and etha nol produc tion (Fig ure 4-1) . The stra in express ing CelY and CelZ , BW35, required the co-expression of the E. chrysanthemi out genes from pCPP2006 for improved growth. Strain BW35 pCPP2006 initially had ethanol productivity similar to both BW21 and BW34. Final ethanol c oncentrations were reduced in strain BW35, possibly due to instability of pCPP2006. Maximum ethanol yields from 90 g/L glucose were 92.7, 91.4, and 80.5% of theoretical by strains BW21, BW34, and BW35 pCPP2006, respectively. When by -products were analy zed (Table 4-2) strains BW21 and B W34 were again s imilar whi le strain BW35 had a slight inc rease i n lactate production. The increased lactate production wa s retained with BW35 pCPP2006. The

PAGE 83

72 0 24 48 72 96 0 1 2 3 4 5 6ATime (h)OD550nm 0 24 48 72 96 0 10 20 30 40 BTime (h)Ethanol (g/L)Figure 4-1. Growth (A) and ethanol production (B) from 90 g/L glucose in OUM1 by K. oxytoca strains BW21 ()Tj/T1_0 1 Tf-0.0006 Tc 0.0006 Tw 12 0 0 12 441.09 98.0101 Tm[(), BW34 (), BW35 (), and BW35 pCPP2006 ().

PAGE 84

73 StrainGlucoseLactateSuccina teAcetateAcetoin + Butane diolEthanolCarbon Balance (mM)(mM)(mM)(mM)(mM)(mM)(%) BW2142921411335121926171002 BW344251072141712091732911 BW35394121781220011833241003 BW35 pCPP2006388261831512030178846981 Table 4-2. Product formation and carbon balance after 72 h from 90 g/L glucose in OUM1 by budAB , ethanologenic, endoglucanase producing strains of K. oxytoca .

PAGE 85

74 producti on of ace tate is co mmon in stra ins expres sing hig h levels o f recom binant pr otein ( Aristidou et al. 1995 ; Farmer and L iao, 1997 ). Cellulose ferme ntation. Two commercial cellulase preparations fr om Genencor Inte rnationa l were e valuate d for the ir use in c ombination with CelY ( and CelZ ). Both Spezyme CE (no longer available commercially ) and GC220 are standardized based on their ac tivities on carboxy methy l cellulos e and ar e repor ted to be b lends of h ydr olase activitie s. To elimin ate any potentia l effec ts by the media , Lur ia broth w as used in evaluating each enzy me blend. Not surprisingly , the highest ethanol concentrations of 38 g/L (Fig ure 4-2C) were obtained with the highest conc entrations of Spezyme CE (100 : L/g Sigmace ll). This c orrespo nded to an ethanol y ield of 67 % of theo retical . In c ontrast to previous work with Sigmacell and Spezy me CE, there was no apparent benefit from CelY (BW3 4 vs. BW21) or CelZ (BW35 pCPP2 006 vs. BW2 1) in comb ination wi th Spezyme CE ( Zhou et al. 2001 ). At lowe r enzy me loadin gs, the r esults cle arly indicate that Spezyme GC220 is the superior to Spezy me CE (Figure 4-2A&C). At equiva lent enzyme conce ntrations (50 : L/g cellulose) the initial rate s and final ethanol concentrations were increased by 26% and 21% respectively (with strain BW34) when Spezy me GC220 is used, co mpared t o Spezy me CE. Etha nol produc tion using Spezy me GC220 at an enzyme loading of 50 : L/g Sigmace ll was nea rly equivale nt to that w ith Spezyme CE at 100 : L/g Sigmace ll. With Spezyme GC220 (50 : L/g), the expression of CelY (in strain BW34) appeared to have the most benefit. SSFs (with strain B W34) produced 15% higher final ethanol conc entrations than did strain BW21 and 70% higher than strain P2.

PAGE 86

75 0 24 48 72 96 120 144 168 0 10 20 30 40 ATime (h)Ethanol (g/L) 0 24 48 72 96 120 144 168 0 10 20 30 40 BTime (h)Ethanol (g/L) 0 24 48 72 96 120 144 168 0 10 20 30 40 CTime (h)Ethanol (g/L) 0 24 48 72 96 120 144 168 0 10 20 30 40 ATime (h)Ethanol (g/L) 0 24 48 72 96 120 144 168 0 10 20 30 40 BTime (h)Ethanol (g/L) 0 24 48 72 96 120 144 168 0 10 20 30 40 CTime (h)Ethanol (g/L)Figure 4-2. Ethanol production in SSF of 100 g/L Sigmacell using (A) 50 )Tj/T1_3 1 Tf-0.0006 Tc 0.0006 Tw 0 12 -12 0 413.01 442.17 Tm(L Spezyme GC220 per g cellulose; (B) 50 )Tj/T1_3 1 Tf0 12 -12 0 413.01 654.5701 Tm(L Spezyme CE per g cellulose; and (C) 100 )Tj/T1_3 1 Tf-0.0008 Tc 0.0008 Tw 0 12 -12 0 427.29 321.57 Tm(L Spezyme CE per g cellulose by ethanologenic K. oxytoca strains P2 ()Tj/T1_3 1 Tf-0.003 Tc 0 12 -12 0 427.29 673.29 Tm(); SZ22 (); BW21 (); BW34 (); and BW35 pCPP2006 ()Tj/T1_3 1 Tf-0.003 Tc 0 12 -12 0 441.57 408.33 Tm().

PAGE 87

76 Because the commercial enzy mes are blends of activities, it is likely that they are periodically re formulated to meet specific customer requirements and thus a re not constant over sev eral y ears. A lso consis tent with a chang e in the f ormulatio n of Spezy me CE was the relatively poor performance of the previously developed strain P2 (Doran and Ingram, 1993 ). The previously de veloped strain SZ22, expressing CelY ( Zhou et al. 2001 ) was also included for comparison. Using either e nzyme (at any loading) the amount of cellulose degradation products (sug ar) present at the end of fermenta tion (Figure 4-3) appea red to correlate to the production of ethanol and may reflect the metabolic activity of the or ganism relative to the effectiveness of the cellulase used. While Sigmacell is reported to be a purified cellulose product, the dete ction of xylose sugge sts a small xylan component. Alternatively , the HPLC peak identified as xylos e may be anothe r metabo lic produ ct from c omponents (prese rvative s, stabilize rs, etc.) present in commercial cellulases. Be cause of this uncertainty , residual xylose is not included in yield calc ulations. In SSFs where the hy drolysis is more rapid (50 : L Spezyme GC220 and 100 : L Spezy me CE per g cellu lose) the stability of pCPP2006 in strain BW35 may ag ain cause a reduction in final ethanol concentr ations. In gener al, the st ra in s l ac ki ng ge ne s o f t he 2, 3bu ta ne di ol fe rm en ta ti on pa th wa y ( budAB ) outperformed those which retain those genes. Consistent with observations from Chapter 2 of our study the elimination of budAB also res ulted in the reducti on of othe r by -produc ts as well (Figu re 4-4) . An excep tion to this w as in stra in BW35 pCPP2 006, as se en in glu cose fermentations (Table 4-2) both acetate a nd lactate production were elevated. Products of the 2,3-butanediol fermentation pathway , in strains P2 and SZ22, appeared to be

PAGE 88

77 P 2 S Z2 2 B W35 pCP P 20 0 6 B W 2 1 BW 3 4 0 10 20 30 40 50Sugar (mM) P 2 S Z2 2 B W 3 5 p CP P2 0 0 6 BW2 1 B W3 4 0 10 20 30 40 50Sugar (mM) P2 SZ 2 2 B W 35 pCP P20 06 B W 21 B W 3 4 0 10 20 30 40 50Cellobiose Glucose Xylose Sugar (mM)aA B CFigure 4-3. Residual Sugars (144h) in SSF of Sigmacell using: (A) 50)Tj/T1_2 1 Tf-0.0009 Tc 0.0009 Tw 12 0 0 12 279.57 110.7301 Tm[(L Spezyme GC220 per g Sigamcell, (B) 50)Tj/T1_2 1 Tf-0.001 Tc 0.001 Tw 12 0 0 12 227.73 96.5701 Tm[(L Spezyme CE per g Sigmacell, and (C) 100)Tj/T1_2 1 Tf-0.002 Tc 0.002 Tw 12 0 0 12 448.41 96.5701 Tm[(L per g Sigmacell. aTentatively identified as xylose

PAGE 89

78 Figure 4-4. Product formation in SSF from 100 g/L Sigmacell (144h) with (A) 50 : L Spezy me GC220 per g cellulose; (B) 50 : L Spezyme CE per g c ellulose; and (C) 100 : L Spezy me CE per g cellulose.

PAGE 90

79 correlated to the rate of cellulose hy drolysis, i.e., glucose limitation (50 : L CE > 50 : L GC220 > 100 : L CE per g ce llulose), contrary to studies with K. aerogenes in gluco se limiting c hemostat c ultures ( Teixeira de Mattos and Tempest, 1983 ). With the exception of strain P2, which performed relatively poorly, succinate levels were consistent among all strains regardless of enzy me type or conc entration. Simultaneous saccharification and co-fermentation (SSCF). The elimination of need for extensive liquid-sol id separ ation ca n simplify process ing of li gnoce llulose to ethanol ( Wright et al. 1988 ). The same enzyme s (Spezyme CE and Spezy me GC220) used in SSF were again evaluated f or their use in SSCFs containing 45 g/L Sigmacell and 40 g/L xy lose. An enzyme loading of 50 : L/g cellulose was use d for each enzy me. To fully evaluate the potential for commercial application, all SSCFs used OUM1, developed in Chapter 2 of our study, only the budAB strains BW21, BW34, and BW35 pCPP2006 were tested. In SSCF it appea rs that the reduced hy drolysis rates by Spezyme CE are actually benefic ial for th e rapid a nd comple te use of xylos e (Fig ure 4-5) . With Spezy me GC220, g lucose le vels altho ugh low ( ~1mM at 24 h ), may initially exceed a threshol d where xylose metabolism becomes r epressed as reported in Chapter 3 of our study . Further evidence of this is found in the reduced r ate of xylose consumption in the endoglucanase producing strains BW34 and BW35 pCPP2006, presumably a re sult of the improved hydroly sis seen in SSFs with Spezyme GC220. The requirement for increased PFL( or ACK) a ctivity (eleva ted pH) f or improv ed xy lose cons umption wa s previou sly shown in Figure 2-7. I n OUM1 at pH 6.0, strain BW21 produced 44.3 g/L ethanol from 90 g/L xy lose, 96.4% of theoretical y ields. A slightly more conse rvative 95% was assumed for xylose conver sion, in cellulose yield calcula tions, in SSCF at pH 5.8

PAGE 91

80 0 24 48 72 96 120 144 0 10 20 30 40 Time (h)Ethanol, Xylose (g/L) 0 24 48 72 96 120 144 0 10 20 30 40 Time (h)Ethanol, Xylose (g/L)A BFigure 4-5. Ethanol production (so lid lines) and xylose consumption (dotted lines) in SSCF of 40 g/L xylose and 45 g/L Sigmacell with 50 )Tj/T1_0 1 Tf-0.0011 Tc 0.0011 Tw 11.9849 0 0 11.9818 240.629 111.0419 Tm[(L Spezyme CE per g Sigmacell (A) and 50 )Tj/T1_0 1 Tf11.9849 0 0 11.9818 455.6403 111.0419 Tm(L Spezyme GC220 per g Sigmacell (B) by K. oxytoca strains BW21 ()Tj/T1_0 1 Tf0.0005 Tc -0.0005 Tw 11.9849 0 0 11.9818 240.5092 82.7646 Tm[(), BW35 pCPP2006 (), and BW34 ().

PAGE 92

81 SSCF Yield (%)a,bSSF Yield (%)a,dEnzyme Strain Xylose+ Cellulose CellulosecSpezyme CE BW2181.871.245.8 BW35 pCPP200682.572.443.6 BW3486.679.847.0 BW2175.058.951.8 Spezyme GC220BW35 pCPP200675.359.537.9 BW3480.970.158.8a % of theoretical based on 0.51 g ethanol per g xylose and/or 0.568 g ethanol per g cellulose.b SSCFs were in OUM1 medium, pH 5.8 and yields ignore residual xylose. c Assumes 95% of theoretical ethanol yield ( 19.4 g/L) from added xylose. d SSFs were in Luria broth, pH 5.2. Table 4-3. Theoretical yields from SSF and SSCF using 50 L cellulase per g Sigmacell.

PAGE 93

82 (Table 4 -3). As in SSF, the by -produc ts formed were sim ilar betw een stra ins in SSCF (Fi gur e 46) a nd m ost ly co nsi st ent wit h t hos e se en i n S SF. The in cre ase d AC K(P FL) activity at pH 5.8 d id result i n an incr ease in a cetate producti on relat ive to tha t seen in glucose or in SSF. The inclusion of xylose in SSCF resulted in improved cellulose conver sion. Table 4-3 compares ethanol y ields from SSF and SSCF (at enzyme loading s of 50 : L/g Si gm ac el l) by budAB strains. T he most sig nifican t increa ses were seen with Spezy me CE wher e estimat ed cellu lose conv ersions w ere >55 % highe r in SSCF tha n in SSF alone . Althoug h not as dr amatic, u sing Spe zyme GC220 in SSCF also res ulted in increa sed etha nol y ields fro m cellulos e (14 to 19 %). With bot h enzy mes, the b est strai n was BW34 in either SSF or SSCF. The improved metabolism of cellulose degrada tion products, in SSCF, was reflected by the low levels of sugar detected at the e nd of fermen tation (F igure 4 -7). When Spezy me GC220 w as used, t he reduc ed rate of xy lose use (and overall ethanol productivity ) by strains BW34 and B W35 pCPP2006, translated to an increase in residual sugar de tected. Conclusions It is evident that the choice of commer cial cellulase can have significa nt impact on the extent o f cellulo se hy droly sis. The in clusion of xylos e, or poss ibly other fr ee sug ars, increased the effectiveness of c ellulases used in SSCF. This was likely due to incre ased biocatalyst conce ntrations, which were better able to maintain sub-inhibitory concentrations of cellobiose and glucose. When incr eased biocataly st concentrations were c ombined w ith the pro duction of additiona l endog lucanas e activit y, c ellulose convers ion was e ven gr eater, even at f ermenta tion condi tions less t hat optima l for fun gal

PAGE 94

83 Figure 4-6. Product formation by ethanologenic, budAB strains of K. oxytoca in SSCF of 40 g/L xy lose and 45 g/L Sigmac ell (144h) using: (A) 50 : L Spezy me CE per g Sigmacell and (B) 50 : L Spezy me GC220 per g Sigmacell.

PAGE 95

84 BW 3 5 pCP P2 006 B W21 B W 34 0 10 20 30 40 50Sugar (mM) BW 3 5 p CP P 2 00 6 BW 2 1 BW 3 4 0 10 20 30 40 50Cellobiose Glucose Xylose Sugar (mM)A BaFigure 4-7. Residual sugar in SSCF of 40 g/ L xylose and 45 g/L Sigmacell using: (A) 50)Tj/T1_0 1 Tf-0.0011 Tc 0.0011 Tw 12 0 0 12 151.41 98.0101 Tm[(L Spezyme CE per g Sigmacell and (B) 50 )Tj/T1_0 1 Tf-0.0017 Tc 0.0017 Tw 12 0 0 12 365.97 98.0101 Tm[(L Spezyme GC220 per g Sigmacell. aTentatively identified as xylose.

PAGE 96

85 cellulas e activit y. T he incre mental imp rovemen ts in ethan ol produc tion from c ellulose are illustrated in Figure 4-8. The a mount of ethanol per unit enzyme has a direct impact on the cos t of etha nol produc tion from c ellulose u sing ce llulase. I n SSF, as Sp ezy me CE loading decreased 50%, the enzy me cost per gram ethanol would decre ase an equivalent 50%. Assu ming Spezy me GC220 ha s an equa l to the co st with Spezy me CE, but th e use of Spezyme GC220 would reduce enzyme costs by an additional 21%. In SSCF the enzyme cost of ethanol produc tion from cellulose could be reduced further by using Spezyme GC220 (14%) or Spezy me CE (60%) at equivalent enzy me loadings (50 : L/g Sigmacell) u sed in SS F. In this work, no attempt was made to optimize the conditions for SSCF and previous SSF optimizations with ethanologenic K. oxytoca used strains that did not contain r ecombina nt endog lucanas es ( Doran and Ing ram, 1993 ). Future optimization co ul d f ur th er im pr ov e t he co nv er si on of li gn oc el lu os e t o e th an ol by K. oxytoca . Additional improvements may be possible throug h the reduction of succinate production by removal of g enes encoding fumarate re ductase ( frdABCD ).

PAGE 97

86 0.0 2.5 5.0 7.5 10.0SSF SSCF100 L CE/g Sigmacell 50 L CE/g Sigmacell 50 L GC220/g Sigmacell g EtOH/ ml enzymeFigure 4-8. Ethanol production, from cellulose, per unit enzyme in SSF (100 g/L Sigmacell) and SSCF (40 g/L xylose and 45 g/L Sigmacell) by K. oxytoca BW34. In SSCF 95% of theoretical yi eld (19.4 g/L) from added xylose was assumed. Spezyme type and enzyme loadings are indicated.

PAGE 98

87 CHAPTER 5 GENERAL CONCL USIONS The expan ded use o f ethano l can pro vide an e nvironme ntally friendly alterna tive to petroleum-based products. The use of ethanol derive d from lignocellulose may have even great er implic ations. B y pr oviding a domestic, renewa ble sourc e of tra nsportat ion fuel, ethanol f rom lign ocellulo sic biomas s can re duce de pendanc e on fore ign sour ces of oi l, and the military/political implications therein. The costs associated with processing, nutrients, and enzy mes have been major barriers to realizing the potential of lignocellulosic ethanol. Etha nologenic strains of K. oxytoca have already been shown to require significantly less fungal enzy me than ethanol producing y east strain, in the conversion of cellulose to ethanol. Our study directly addr esses these areas to further improve etha nologenic K. oxytoca for the lowcost production of ethanol from lignocellulosic biomass. The strains developed in our study can aid i n overco ming man y of these co st issues. The use of inexpensive medium is essential for the economic production of commodity chemicals from lig nocelluose. It was found that a simple ure a, salts, and 0.5% corn steep liquor medium could support high ethanol y ields. In this medium at pHs used in combination with fungal cellulases in the simultaneous saccharifica tion and fermen tation of c ellulose, a signif icant amo unt of gl ucose ca rbon was diverte d to by-products of the 2,3-butanediol pathway . Deletion of two genes, budAB , specif ic to this pathway restore d high ethanol y ields from glucose. At the ethanol concentration achieved with strain BW21(42.6 g/L ) the new medium (OUM1) costs 8.5¢ per gallon.

PAGE 99

88 Corn steep liquor, a by-pr oduct of the corn processing industry , is rich in many essential minerals added as part of OUM1, including ca lcium, iron, magnesium, and phosphorous. It is li kely that many of these could be e xcluded wh ich would further reduce the costs to ~5.8¢ per gallon of ethanol. An appare nt, pH dependant, reduction in either pyruvateformate lya se or acetate kinase activity necessitated slightly higher fermentation pH for the efficie nt conver sion of xy lose to eth anol. The ability to utilize all of the sugars de rived from lignocellulose rapidly and efficiently impac ts product yields. Ethanolog enic strains of K. oxytoca have been previously shown to use these sug ars, including soluble oligimers, independently . Howeve r in prev ious studie s it was sho wn in mixture s of suga rs, wher e gluco se was a major constituent, other sugars (i.e., xy lose and arabinose) were consumed slowly and incom plete ly. This pheno menon was at tribu ted to catabo lite repres sion throu gh transcriptional regulation by the global regulator of ca rbon metabolism Crp. In our study it was found to be a particular problem in mixtures of glucose and xylose. The presence of arabinose was actually found to aid in the metabolism of xylose, likely due to activation of the a rabinose transport sy stem and its nonspecific transport of xylose. The plasmid-based over e xpression of the cAMP insensitive Crp mutant Crp(in), in K. oxytoca SZ21, re sulted in im proved su gar up take and ethanol p roductio n, in all mixtures te sted, but m ost signi ficantly in gluco se + xy lose mixture s. The impr ovements in xylose utilization observed in plasmid-based over expression of Crp(in) were retained when crp(in) was chromosomally expressed in stra in BW23. However, when the budAB stain BW21 was modified with crp(in) , strain BW24, no further improvements were seen.

PAGE 100

89 This result suggested an additional reg ulatory circ uit possibly involving an intermediate, cofact or, or pr oduct of g lyc oly sis. The spar ing of xy lose, in g lucose + xylos e mixtures, may be a par ticular p roblem in batch cultivation. At large scale, a fe d-batch type pr ocess would likely be implemented. It is known that under glucose limitation catabolite repression does not operate in E. coli culture s and are able to imm ediately use a var iety of alter native su gars ( Lendenmann and Egli, 1995 ). In our study , the ability of K. oxytoca to utilize xylose in the presence of limiting glucose was subsequently confirmed in the simultaneous saccharification and co-fermentation of xylose + cellulose. The bene fits of co -produc tion of re combinan t endog lucanas es, in com bination w ith commercially a vailable fungal cellulases, have bee n previously shown. I n our work, as well as in previous studies, the extent of the benefit from endoglucanase s from E. chrysanthemi (CelY and/or CelZ) was dependant on the f ormulation of the commercial enzy mes. This w as expect ed as com mercial cellulas es are b lends of c elluloly tic activ ities, possibly from differe nt sources. The complete cellulose hy drolysis requires thre e broad categ ories of activitie s, endog lucanas e activit y be ing only one. Within the endog lucanas e category there are several differ ent specificities and activities, those commercial blends (deficient in the activities provided by CelY or CelZ) would benefit from the co-production of these enzy mes. The highest ethanol concentration (38 g /L), from cellulose, was achieved with 100 : L of Spezy me CE per g cellulose using the budAB , CelY producing strain BW34. Compared to strain P2 this was a 36% improvement. However, at both 100 : L and 50 : L Spezy me CE per g cellu lose, BW3 4 and its non-endoglucanase producing parent strain (BW21) were similar. Using a nother blend,

PAGE 101

90 Spezyme GC220 (at 50 : L enzy me per g cellulose), strain BW34 produced 15% higher ethanol concentration than was produced by strain BW21 and 70% higher than strain P2. At similar enzyme loading (50 : L/g), strain BW34 produc ed 25% more ethanol when using Spezyme GC220 as opposed to Spezy me CE. In gener al, the strains lacking the 2,3-b utane diol pathw ay out perf ormed thos e cont ainin g budAB in SSF using either cellulase blend. In c ontrast to SSF (at re duced e nzym e loading ) the SSCF o f the mixtur e of xy lose and cellulose, the use of Spezyme CE produced ethanol more rapidly and slightly hig her (3%) fin al ethan ol conce ntrations . This is like ly d ue to the i mproved c ellulose h ydr oly sis by Spezy me GC220, seen in SSFs with similar enzyme loading. The increased rate of sugar r elease probably induces r egulat ory circuits , inhibiting xylos e metabo lism. With strain BW34, using 50 : L Spezy me CE per g cellulose, the total ethanol y ields (including xylose and cellulose) exceede d 85% of theoretical. Total y ields with Spezyme GC220 were 81% of theoretical. With each enzy me blend, strain BW34 was superior to strains BW21 and BW35 pCPP2006. Using either enzyme ble nd in SSCF, with strain BW34, estimated ethanol yields fr om cellulose was improved by 70% , with Spezyme CE, and 20%, with S pezy me GC220 w hen compa red to y ield obta ined in SSF at similar enzy me loading. While no a ttempt was made in our s tudy to optimi ze the condi tions for S SCF, chang es in the f ermenta tion condi tions (i.e ., pH, tem peratur e) or by reducin g the en zyme loading or timing of addition could further improve the results. The diff erence in optimal conditions between fungal cellulases and metabolism of the bioca talyst has long be en thought t o be a pro blem in the enzy matic con version o f lignoc elluose to ethanol. This

PAGE 102

91 incongruity is one of the reasons why tolerance to low pH is listed as a desirable trait. The results from SSCF suggest that as long as suff icient free sugar (xy lose) is present, for cell growth and metabolic activity , high yields fr om cellulose can still be obtained even at sub-optimal conditions for cellulase activity. I n both SSF and SSCF succinate production was two fold higher than was observed in g lucose fermentations with budAB strains, providing an additional opportunity f or future improvement through the elimination of genes encoding fumara te reductase ( frdABCD ).

PAGE 103

92 REFERENCES Arist idou , A.A. , K. S an, an d G.N. Bennet t. 19 95. Me tabol ic engi neerin g of Escheric hia coli to enhance recombinant protein production through a cetate reduction. Biotechnol. Prog. 11:475-478. Arntzen, C.E., and B.E. Dale (co-chairs).1999 . Biobased industrial products, priorities for res earch a nd commer cializatio n. Nation al Acad emy Press, Was hington, D.C. Assanuma, N., and T. Hino. 2000. Effects of pH and energ y supply on activity and amount of pyruvate formate-lya se in Streptoc occus bo vis . Appl. Environ. Microbiol. 9:3773-3777. Ausubell, F.M., R. Brent, R.E. Kingston, D.D. More, J.G. Seidman. J.A. Smith, and K. Struhl. 1996. Current protocols in molecular biology . Green Publishing and Wiley Interscience, N ew York. Barbosa, M. de F.S., and L .O. Ingra m. 1994. Expression of the Zymomona s mobilis alc oho l d ehyd rog ena se II ( adhB ) and pyruva te decarboxylase ( pdc ) gene s in Bacillus. Cur. Microbiol. 28:279-282. Beall, D.S. 1995. Genetic engineering of Erwinia for the c onversio n of bioma ss to ethanol. Ph.D. Thesis. University of F lorida. Gainesville Florida USA. Beall, D.S., and L.O. I ngram. 1993. Genetic engineering of soft-rot bacteria for ethanol production from lignocellulose. J. Ind. Microbiol. 11:151-155. Beall, D.S., K. Ohta, and L.O . Ingram. 1991. Parametric studies of ethanol production from xy lose and o ther sug ars by recombi nant Escheric hia coli . Biot echno l. Bio eng. 38:296-303. Bergeron, P. 1996. Environme ntal impac ts of bioe thanol, pp . 163-178 .In Wy man CH (ed.), H andbook o n bioetha nol: produ ction and utilization. Tay lor & Fr ancis, Washingt on DC. Bergeron, P., and N.D. Hinman. 1991. F uel ethanol usage and environmental ca rbon dioxide production. In Klass , D.L. (ed ), Ene rgy from bi omass waste s XIV. Ch icago Institute of Gas Technolog y. Chicag o.

PAGE 104

93 Blomqvist , K., M. Nikkola, P. Lehtovaara, M. Suihko, U. Airaksinen, K .B. Stråby, J.K. Knowles, and M.E. Penttila. 1993. Characterization of the 2,3-butanediol operons from Klebsiella terrigena and Enterobacter aerogenes . J. Bacte riol. 175:1392-1404. Blondel et-Rouau lt , M.H., J. Weiser, A. Lebrihi, P. Branny , and J.L. Pernodet. 1997. Antibiotic resistance gene cassette der ived from the S interposon for use in E. coli and Streptomyces . Gene. 190:315-317. Böck, A., and G. Sawers. 1996. Fermentation, p. 262-282. In F. C. Nei dhardt, R . Curtiss II I, J. L . Ing raham, E . C. C. L in, K. B. Low, B. Mag asanik, W. S. Reznikof f, M. Riley, M. Schaechter , and H. E. Umbarger (ed.), Escheric hia coli and Salmonel la : ce ll ul ar an d m ol ec ul ar bi ol og y, 2 nd ed . A me ri ca n S oc ie ty f or Mi cr ob io lo gy, Washingt on, D.C. Bothast, R.J., B. Saha, A.V. Flosenzier, and L.O. I ngram. 1994. Fermentation of L-a rabinose , D-xy lose and D -gluco se by ethanolo genic r ecombina nt Klebsiella oxytoca strain P2. Biotechnol. Lett. 16:401-406. Bringer-Mey er , S., K.L. Schimz, and H. Sahm. 1986. Pyr uvate decarboxyla se from Zymomona s mobilis . Iso lation an d partia l chara cterizat ion. Arc h. Microb iol. 146:105-110. Brooks, T.A., and L.O. I ngram. 1995. Conversion of mixed waste office paper to ethanol by ge netically eng ineered Klebsiella oxytoca strai n P2. Biote chnol . Prog. 11:619-625. Brückner, R., and F. Titemey er. 2002. Carbon catabolite repression in bacteria: choice of the car bon sourc e and au toregu latory limitation o f sugar utilization. FEMS Microbiol. Lett. 209:141-148. Bulls, S.R. 1991. The U.S . De par tm ent of E ner gy Bio fue ls res ear ch p rog ram , pp . 114. In Klass, D.L. (ed), Ener gy f rom biomass wastes XIV. Chicag o Institute of Gas Technology . Chicago. Burchardt, G., and L .O. Ingra m. 1992. Conversio n of xy lan to eth anol by ethanolo genic strains of Escheric hia coli and Klebsiella oxytoca . Appl. Env iron. Mic robiol. 58:1128-1133. Busby, S., and A. Kolb. 1996. The CAP modulon, p. 255-280. In E.C.C. L in and A.S. Ly nch (ed.), Regulation of gene expression in Escheric hia coli . R.G. Landes Co., Austi n Tex.

PAGE 105

94 Causey, T. B ., K.T. Shanmugam, L.P. Yomano, and L .O. Ingra m. 2004. Engineering Escheric hia coli for Efficient Conversion of Glucose to Py ruvate. Proc. Natl. Acad. Sci. 101, 2235-2240. Chandrakant, P., and V.S. Bisaria. 1998. Simultaneous bioconversion if cellulose and hemicellulose to ethanol. Crit. Rev. Biotechnol. 18:295-331. Chemical Market Reporter. 2005 . Prices and people. Chem. Market Rep. 267(12):16-18. Datsenko, K.A., and B.L . Wanner. 2000. One-ste p inactiv ation of c hromosoma l genes in Escheric hia coli K-12 using PCR products. Proc. Natl. Acad. Sci. 97:6640-6645. Deanda, K., M. Zhang, C. Eddy , and S. Picataggio. 1996. Development of an arabin ose-fe rment ing Zymomona s mobilis strai n by metab olic p athwa y engineeri ng. Appl. Environ. Microbiol. 62:4465-4470. Dien, B.S., M.A. Cotta, and T.W. Jeffries. 2003. Bacteria enginee red for fuel ethanol production: Current status. Appl. Microbiol. Biotechnol. 63:258-266. Doran J.B., H.C Aldrich, and L.O. I ngram. 1994. Saccharification and fermentation of su ga rca ne ba ga ss e b y Klebsiella oxytoca P2 containing chromosomally integ rated genes encoding the Zymomona s mobilis ethan ol pat hway. Biot echno l. Bio eng. 44:240-247. Doran, J.B., and L.O. I ngram. 1993. Fermentation of cry stalline cellulose to ethanol by Klebsiella oxytoca containing chromosomally integrated Zymomona s mobilis genes . Biotechnol. Prog. 9:533-538. Eppler, T., and W. Boos. 1999 . Gly cerol3-phosph ate media ted repr ession of malT in Esc her ich ia c ol i d oes not req ui re m eta bol is m, dep end s on enz yme II A Glc and is mediated by cA mp levels. Mol. Microbiol. 33:1221-1231. Farmer, W.R., and J.C. Liao. 1997. Reduction of aerobic acetate production by Escheric hia coli . Appl. Environ. Microbiol. 63:3205-3210. Feederle, R., M. Pajatsch, E. Kremmer, and A. B öck. 1996 . Metabolism of cyclodextrins by Klebsiella oxytoca M5a1: pur ificatio n and cha racter ization of a cy toplasmic ally located cy clodextrinase. Arch. Microbiol. 165:206-12. Fiedler G., M. Pajatsch, and A. Böck.1996. Genetics of a Novel Starch Utilization Pathway Present in Klebsiella oxytoca . J. Mol. Biol . 256:279-291.

PAGE 106

95 Frankel, D.G. 1996. Glycoly sis, p. 189-198. In F. C . Ne id har dt , R . C urt is s III, J . L. Ing raham, E . C. C. L in, K. B. Low, B. Mag asanik, W. S. Reznikof f, M. Riley , M. Schaechter, and H. E. Umbarge r (ed.), Escheric hia coli and Salmonel la : cellular and molecular biology , 2nd ed. American Society for Microbiology , Washington, D.C. Fuels Security A ct of 2005. 2005. S650. 109 th Cong. Gauss, W.F., S. Suzuki, and M. Takagi. 1976. Manufac ture of a lcohol fr om cellulo sic material using plural ferments. U.S. Patent No. 3,990,944. Genencor Inte rnational, Inc. 2004. Genencor Celebrates Major Progr ess in the Conversion of Biomass to Ethanol. World Wide Web. URL http://www.genencor.com/wt/gcor/pr_1098313606 , accessed May 2005. Greene, N ., F.E. Ce lik, B. Da le, M. Jack son, K. Jay awardh ana, H. Jin , E.D. L arson, M. Laser, L . Ly nd, D. MacKenzie, J. Mark, J. McBride, A. McLaug hlin, and D. Saccar di. 2004. G rowing energ y: H ow biofue ls can he lp end Ame rica’s o il dependa nce. Na tural Res ources D efense Council, Wa shington , D.C. Grimont, G., P.A.D. Grimont, and C. Richard. 1991. The Genu s Klebsiella , p.2775-2796. In A. Balows, H.G. Truper, M. Dworkin, W. Harder, a nd K.H. Schliefer (ed.), The Prokaryotes, 2 nd ed.. Spriner-Verlag, New Y ork. Hahn-Hägerdal, B., L .T Senac, and T.K. Skoog. 1991. Ethanolic fermentation of pentoses in lignocellulosic hydroly sate. Appl. Biochem. Biotechnol. 28/29:131-144. Harman, J.G., K McKenney , and A. Peterofsky . 1986 . Structure-function analy sis of three cAMP-ind epende nt forms of the cAMP r eceptor protein* . J. Biol. Che m. 261:16332-16339. Hasona A, Y. Kim, F.G. Healy , L.O. I ngram, and K.T. Shanmugam. 2004. Pyr uvate Formate Ly ase and Acetate Kinase Are E ssential for Anaerobic Growth of Escheric hia coli on Xylose. J. Bacter iol. 186:7593-7600. He, S.Y., M. Lindeberg , A.K. Chatterjee, and A. Collmer. 1991 . Cloned Erwinia chrysant hemi out genes enable Escheric hia coli to selec tively secret e a diver se family of heter ologous proteins to its milieu. Proc. Natl. Acad. Sci. USA. 88:10791083. Hernandez-Montalvo, V., A. Martinez, G. Gernandez-Chavez, F. B olivar, F. Valle, and G. Gosset. 2003 . Expression of galP and glk in a Escheric hia coli PTS mutant restore s glucos e transp ort and in crease s gly coly tic flux to f ermenta tion produ cts. Biotechnol. Bioeng. 83:687-694.

PAGE 107

96 Hernandez-Montalvo, V., F. Valle, F. Bolivar , and G. Gosset. 2001 .Characterization of sugar mixtures utilization by an Escheric hia coli mutant devoid of the phosphotransferase sy stem. Appl. Microbiol. Biotechnol. 57:186-191. Hettenhaus, J.R. 1998 . Ethanol fermentation strains: Present and future requirements for biom ass to ethan ol com merci aliz ation , a repo rt to t he U.S . DOE O ffice o f Energy Efficie ncy and Rene wable En ergy Ethanol Pr ogram a nd Nation al Renew able En er gy La bo ra to ry. Hettenhaus, J.R., and D. Glassner. 1997. Cellulase assessment for biomass hydr olysis, a report for the DOE Office of Ene rgy Efficiency and Renewable Energy Ethanol Pro gra m, on W orl d W id e W eb U RL http://www .ceassi st.com/as sessment. htm . Himmel, M.E., M.F. Ruth, and C.E. Wyman. 1999. Cellulase for commo dity products from cellulosic biomass. Curr. Opin. Biotechnol. 10:358-364. Hinman, N.D., D.J. Schell, C.J. Riley, P.W. Bergeron, and P.J. Walter. 1992 Preliminary estimate of the co st of etha nol produc tion for SSF technolo gy . Appl. B iochem. Biotechnol. 34/35:639-649. Ho, N.W.Y., Z.D. Chen, and A.P. Brainard. 1998. Genetically e ngineered Saccharomyces yeast ca pable of effective coferme ntation of glucose and xylose. A ppl. Environ. Microbiol. 64:1852-1859 Ingram, L .O., P.F. Gomez, X. Lai, M. Moniruzzaman, B.E. Wood, L .P. Yomano, and S.W. York. 1998. Metabolic engineering of ba cteria for ethanol production. Biotechnol. Bioeng. 58:204-214. Jansen, N.B., M.C. Flickinger, and G.T. Tsao. 1984. Production of 2,3-butanediol from Dxyl os e b y Klebsiella oxytoca ATCC 8724. Biotechnol. Bioeng. 26:362-369. Johansen, L.,K. Bry n, F.C. Størmer. 1975. Physical and Bioc hemical Role of the Butanediol Pathway in Aerobacter ( Enterobacter ) aerogenes . J. Bacte riol. 123:1124-1130. Kadam, K.L., and M.M. Newma n. 1997. Development of a low-cost fermentation medium for ethanol p roductio n from bio mass. App l. Microb iol. Biote chnol. 47:625-629. Kamm, B., and M. Kamm. 2004. Principle s of biore finerie s. Appl. Mic robiol. B iotechno l. 64:137-145. Kang, H., S. Song, and C. Park. 1998. Priority of pentose utilization at the level of transcription: arabinose, xylose, and r ibose operons. Mol. Cells. 8:318-323.

PAGE 108

97 Kerstetter, J.D., and J.K. Ly ons. 1991. Mixed waste office paper to ethanol f uel. Washington State Energy Office. Oly mpia WA. Kim, S. and B.E. Dale. 2004. Global potential bioethanol production from wasted crops and crop residues. Biomass and Bioener gy. 26:361-375. Kingsman Ethanol Report . May 25, 2005. Klinke, H.B., A.B. Thomsen, and B.K. Ahring . 2004. Inhibition of ethanol-producing yeast and ba cteria by de gradation products produced during pre-treatment of biomass. Appl. Microbiol. Biotechnol. 66:10-26. Koebmann, B.J., H.V. Westerhoff, J.L. Snoep, D. Nilsson, and P. R. Jensen. 2002. The Glycoly tic Flux in Escheric hia coli Is Con trolled by the Dema nd for AT P. J. Bacteriol. 184: 3909–3916. Kubicek, C.P., and M.E. Penttila. 1998. Regulation of production of plant poly saccharide de gr ad in g e nz yme s b y Trichoderma , p. 49-72. In G.E. Harmon and C.P. Kubicek (ed.) , Trichoderma and Gliocladium , vol. 2. Taylor and F rances Ltd., L ondon, U.K. Laikova, O.N., A.A. Mironov, and M.S. Gelfa nd. 2001. Computational analysis of the transcriptional regulation of pentose utilization sy stems in the gamma subdivision of Proteobacteria. FEMS Microbiol. L ett. 205:315-322. Lee, Y.Y., P. I yer, and R.W. Torg et. 1999. Dilute-a cid hy droly sis of lig nocellul osic biomass. Adv. Biochem. Engin. Technol. 65:99-115. Lendenmann, U., and T, Eg li. 1995. Is Escheric hia coli growing in glucose-limited chemostat culture able to utilize other sugars without lag? Microbiol. 141:71-78. Lindsay , S.E., R.J. Bothast, and L.O. Ing ram. 1995. Improved strains of rec ombinant Escheric hia coli for etha nol produc tion from s ugar mi xtures. App l. Microb iol. Biotechnol. 43:70-75. Ly nd, L.R., R.T. Elander, and C.E. Wy man. 1996. Likely features and costs of mature biomass ethanol technology . Appl. Biochem. Biotech. 57/58:741-761. Ly nd, L.R., P.J. Weimer, W.H. van Zy l, and I.K. Pretorius. 2002. Microbia l cellulos e utilization: Fundamentals and biotechnology. Mic robiol. Mol. Biol. Rev. 66:506577. Mahl, M.C., P.W. Wilson, M.A. Fife, and W.H. Ewing. 1965 . Nitrogen fixation by members of the tribe Klebsielleae . J. Bacte riol. 89:1482-1487.

PAGE 109

98 Martinez, A., S.W. York, L.P. Yomano, V.L. Pineda , F.C. Davis, J.C. Shelton, and L.O. Ingram. 1999. B iosynthetic burden and pla smid burden limit expression of chromoso mally integr ated het erolog ous gen es ( pdc , adhB ) in Escheric hia coli . Biotechnol Prog. 15:891-897. Martinez, J., L. Martinez, M. Rosenblueth, J. Silva, and E. Martinez-Romero. 2004. How are gene sequence analyses modify ing bacterial taxonomy? The case of Klebsiella . Int. Microbiol. :261-268. MartinezMorales , F., A.C. Borge s, A. Mar tinez, K.T . Shanmug am, and L .O. I ngram. 1999. Chromosomal integration of heterologous DNA in Escheric hia coli with precise removal of marke rs and re plicons us ed during construc tion. J. Bac teriol. 181:7143-7148. Mayer, D ., V. Schlensog, and A. Böck. 1995. I dentification of the transcriptional activator controlling the butanediol fermentation pathway in Klebsiella terrigena . J. Bacteriol. 177:5261-5269. Melton, T., L.L . Snow, C.S. Freitag, and W.J Dobroosz. 1981. Isolation and characterization of cAMP suppressor mutants in Escheric hia coli K12. Mol. Gen. Genet. 182:480-489. McClelland, M., L. Florea, K. Sande rson, S.W. Clifton. J. Parkhill, C. Churcger, G. Douglan, R.K. Wilson, and W. Miller. 2000 . Comparison of the Escheric hia coli K-12 genome with sampled genomes of Klebsiella pneumoniae and three Salmonella enterica serovar s, Ty phimurium, Typ hi and Par aty phi. Nucl . Acids Re s. 28:4974-4986. Melchiorsen C.R., K.V. Jokumesen, J. Villadsen, H. Israelsen, and J. Arnau. 2002. The level of pyr uvate-f ormate ly ase cont rols the sh ift from h omolactic to mixed-ac id product formation in Lactococ cus lactis . Appl. Microbiol. Biotechnol. 8:338-344. Meyer, M., P. Dimroth, and M. B ott. 2001. Catabolit e repre ssion of th e citrat e fermentation genes in Klebsiella pneumoniae : evidence for involvement of the cyclic AMP re ceptor protein. J. Bacteriol. 183:5248-5256. Miller, J.H. 1992. A Short Course in Bacterial Genetics :a L aboratory Manual and Handbook for Escheric hia coli and Related Bacteria. Cold Spring Har bor Laboratory Press, Plainview, N.Y. Mon iru zzam an, M., B.S . Di en, B. F err er, R.B . He spe ll, B .E. Dal e, L .O. In gr am, and R.J. Bothast. 1996 . Ethanol production from AFEX pretreated cor n fiber by recombinant bacteria. Biotechnol. L ett. 18:985-990.

PAGE 110

99 Moniruzzaman, M., and L.O. I ngram. 1998. Ethanol pr oduction f rom dilute acid hydroly sate of rice hulls using genetically engineered Escheric hia coli . . Lett. 20:943-947. Moore, J.L.1993. The mechanism of activation of the Escheric hia coli cy clic AMP re ce pt or pr ot ei n b y cyc li c A MP . P h. D. Th es is . T he Ge or ge Wa sh in gt on Un iv er si ty. Washingt on D.C. Mosier, N., C. Wyman, B. Dale , R. Elander, Y.Y. Lee, M. Holtzapple, and M. L adisch. 2005. Fe atures o f promisin g techn ologies for pre treatme nt of lig nocellul osic biomass. Bioresource Technol. 96:673-686. Neidhardt, F.C., P.L. Bloch, a nd D.F. Smith. 1974. Cult ure med ia f or e nte rob act eri a. J. Bacteriol. 119:736-747. Nichols, N.N., B.S. Dien, and R.J. Bothast. 2001. Use of catabolite repression mutants for fermen tation of s ugar mi xtures to et hanol. Ap pl. Micro biol. Biot echnol. 56:120-125. Nieves, R.A., C.I. Ehrman, W.S. Adney , R.T. Elander, and M.E. Himmel. 1998 . Technic al Communic ation: Sur vey and Anal ysi s of Commer cial Cellu lase Prepara tions Suitab le for B iomass Conv ersion to Ethanol. Wo rld J. Micro biol. Biotechnol. 14:301-304. Novozy ms A/S . 2005. Novozymes and NREL reduce enzy me cost. World Wide Web. URL http://www.novozymes.com/cgi-bin/bvisapi.dll/press/press.jsp?id=32730&lang=e n accessed May 2005. Ohta, K., D.S. Beall, J.P. Mejia, K.T. Shanmugam, and L.O . Ingram. 1991a . Metabol ic engineering of Klebsiella oxytoca M5A1 for ethanol production from xylose a nd glucose. Appl. Environ. Microbiol. 57:2810-2815. Ohta, K., D.S. Beall, K.T. Shanmugam, and L .O. Ingra m. 1991b . Genetic improvement of Escheric hia coli for ethanol production: chromosomal integration of Zymomonas mobilis gen es e nco di ng p yruv ate dec arb ox ylas e an d al coh ol deh ydro gen ase II. Appl. Environ. Microbiol. 57:893-900. Ørskov, I. 1984.Genus V Klebsiella , p. 461-465. In Kreig , N.R. ; Hol t J. G. (ed) , Bergey's Manual of Systematic B acteriology , Vol.1, ; Williams and Wilkins, Baltimore, MD. Parker, A.S. and P.J. Abbot. 2003. alpha-c yc lodextrin, I n: Safety Evaluati on of Cer tain Food Add itives and Contamina nts. WHO fo od additiv es serie s # 48. World Health Organization, Geneva.

PAGE 111

100 Perlack, R.D., L.L . Wright , A. Turhollow, R.L. Graham, B. Stokes, and D .C. Erbach. 2005. Biomass as Feedstock for a bioener gy a nd bioproducts industry: The technic al feas ibility of a billio n-ton ann ual supply , A joint stu dy b y Th e U.S. De pt. of Agriculture and The U.S. Dept. of Ene rgy . Oak Ridge National La boratory, Oa k Ridge, TN. Pósfai, G ., M.D. Koob, H.A. Kirkpatrick, and F.R. Blattner. 1997. Ve rsatile insertion plasmids for targeted genome manipulations in bac teria: isolation, deletion, and rescue of the pathogenicity island LEE of the Escheric hia coli O157:H7 genome. J. Bacteriol. 179:4426–4428. Qian Y., L.P. Yoman o, J. F. Pres ton, H.C. A ldric h, and L.O. Ingram. 2003. Clon ing, characterization, and functional expression of the Klebsiella oxytoca xylod extrin utilization operon ( xynTB ) in Escheric hia coli . Appl. Env iron. Mic robiol. 69:5957-5967. Renewable Fuels Association. 2003. Building a secure energ y future: Ethanol industry outlook. Re newable Fuels As sociation , Washing ton, D.C. Saha, B.C., 2003. Hemicel lulose bio convers ion. J. I nd. Micro biol. Biot echnol. 30:279-291. Saier, M.H., 1998. Multiple mechanisms controlling carbon metabolism in bacteria. Biotechnol. Bioeng. 58:170-174. Saier, M.H., T.M. Ramsier, and J. Reizer. 1996. Regulation of Carbon Utilization, p. 1325-1943. In F. C. Neidhard t, R. Curt iss III, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Mag asanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H . E. Umbarger (ed.), Escheric hia coli and Salmonel la : c el lu la r a nd mo le cu la r b io lo gy, 2nd ed. Am erican Society for Micr obiology , Washing ton, D.C. Sambrook, J., and D.W. Russell. 2001. Molecula r Cloning : a La borator y Ma nual. Cold Spring Harbor La boratory Press, Cold Spring H arbor, NY. Scopes, R.K. 1983. An iron-activated alcohol dehy drogenase. FEBS L ett. 156:303-306. Sedlak, M., and N. W. Ho. 2001. Expression of E. coli araBAD operon encoding enzymes for metabolizing L -arabinose in Saccharomyces cerevisiae . Enzy me Microb. Technol. 28:16-24. Shatwell, K.P., B.M. Charalambous , T.P . McDo nald, and P. J. He nders on. 19 95. C loni ng, sequencing, and expression of the araE gene of Klebsiella oxytoca 8017, which encodes arabinose-H + symport activity . J. Bacteriol. 177:5379-5380.

PAGE 112

101 Silhavay, T.J., M. L. B erman , and L .W. Enquist. 1984. Exper iments with gene f usions. Cold Springs Harbor Labora tory Press. Cold Springs Har bor, NY. Sivers, M.V., and G. Zacchi. 1995. A techno-economical comparison of three pr ocesses for the production of ethanol from pine. Bioresour. Te chnol. 51:43-52. Skoog K., and B. Hahn-Häg erdal. 1990. Effect of oxyg enation on xylose fermentation by Pichia stip itis. Appl Environ Microbiol. 56:3389–3394. Suzuki, T. 1969. Phosphotransacetylase of Escheric hia coli B, activation by py ruvate and inhibition by NADH and cer tain nucleotides. Biochemica et Biophy sica ACTA. 191:559-569. Taherzadeh, M.J. 1999. Ethanol from lignocellulose: Physiolog ical effects of inhibitors an d f er me nt at io n s tr at eg ie s. Ph .D . T he si s. Ch al me rs Un iv er si ty o f T ec hn ol og y. Teixeira de Mattos, M.J. and O.M. Neijssel. 1997. Bioenergetic consequenc es of microbial adaptation to low-nutrient environments. J. Biotechnol. 59:117-126. Teixeira de Mattos, M.J. and D.W. Tempest. 1983. Metabolic and energetic aspec ts of the growth of Klebsiella aerog enes NCTC 418 on glucose in anaerobic che mostat culture. Arch. Microbiol. 134:80-85. The Nati onal Commis sion on Ene rgy Policy . 2004. End ing the e nergy stalemat e: A bipartisan strategy to meet America’s energy challenges. The National Commission on Energ y Pol icy , Washing ton, D.C. Tolan, J.S., and R.K. Finn. 1987. Fermen tation of D -xylo se to etha nol by geneti cally modified Klebsiella planticola . Appl. Environ. Microbiol. 53:2039-2044. Torget, R.W., J.S. Kin, and Y.Y. Lee. 2000. Fundame ntal aspe cts of dilu te acid hyd roly sis/frac tionation k inetics of hardwoo d carboh ydr ates. 1. c ellulose hydroly sis. Ind. Eng. Chem. Re s. 39:2817-2825. Underwood, S.A., M.L. Buszko, K.T. Shanmuga m, and L.O. I ngram. 2002a. Flux through citrate sy nthase limits the growth of ethanologenic Escheric hia coli KO11 during xylose fer mentation. Appl. Environ. Microbiol. 68:1071-1081. Underwood, S.A. , S. Zhou, T.B. Causey , L.P. Yomano, K.T. Shanmugam, and L .O. Ingram. 2002b. Ge netic Changes to Optimize Carbon Partitioning Between Ethanol and Biosynthesis in Ethanolog enic Escheric hia coli . Appl. Env iron. Mic robiol. 268:1715-1727.

PAGE 113

102 U.S. Department of Energy , Office of Energy Efficiency and Renewable Energy . 2005. Biomass program, information resource s, Biomass feedstock composition and propert y da tabase . World Wide We b. URL http://www .eere. energ y.g ov/biomas s/feeds tock_dat abases. html , accessed May 2005 U.S. Department of Energy , Office of Energy Efficiency and Renewable Energy . 2005. Biomass program, technologies, sugar platform . World Wide Web. URL http://www .eere. energ y.g ov/biomas s/cellula se_cost. html , accessed May , 2005 U.S. Department of Energy , Office of Energy Efficiency and Renewable Energy . 2005. Biomass program, technologies, sug ar platform . World Wide Web. URL http://www .eere. energ y.g ov/biomas s/sugar _platfor m.html , a cc es se d M ay, 2005 Wandersman, C. 1996. Secretion across the outer bacterial membra ne, p. 955-966. In F. C. Neidhardt, R. Curtiss II I, J. L. I ngraham, E. C. C. Lin, K. B . Low, B. Mag asanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umba rger (ed.), Escheric hia coli and Salmonel la : cellular and molecular biology , 2nd ed. American Society for Microbio logy , Washing ton, D.C. Washingt on Univer sity in St. L ouis, Gen ome Seque ncing Ce nter, B acteria l Projects . Klebsiella pneumoniae . World Wide We b. URL http://genome.wustl.edu/projects/bacterial/kpneumoniae/ , accessed Nov., 2003. Wood, B.E. 1997. Improvements to the simultaneous sacchar ification and fermentation of li gn oc el lu lo se by Klebsiella oxytoca strain P2 using ultrasound and heterologous endogl ucanase s expresse d in the et hanolog enic bac terium Escheric hia coli strain KO11. M.S. Thesis. University of Flor ida. Gainesville Florida. Wood, B.E., H.C. Aldrich, and L.O. I ngram. 1997a . Ultrasound stimulates ethanol production during the simultaneous saccharification and f ermentation of mixed waste office paper. Biotechnol. Prog . 13:232-237. Wood, B.E., Beall, D.S. and L.O. I ngram. 1997b . Production of recombinant bacterial endoglucanase as a co-pr oduct with ethanol during fermentation using derivatives of Escheric hia coli KO11. Biotech. Bioeng. 55:547-555. Wood, B.E. and L.O. I ngram, 1992. Ethanol production from cellobiose, amorphous cellulose and cry stalline c ellulose b y r ecombina nt Klebsiella oxytoca containing chromosomally integ rated Zymomona s mobilis genes for ethanol production and plasmids e xpressing thermosta ble cellu lase ge nes from Clostridium thermocellum . Appl. Environ. Microbiol. 58:2103-2110.

PAGE 114

103 Wood, P.J., J.D. Erfle, and R.M. Teather. 1988. Use of complex formation between Congo re d and poly saccha rides in d etection and assa y of poly saccha ride hy droly sis. Methods Enzymol. 160:59-74. Wright, J.D. 1988. Ethanol from lignocellulosics : an overview. Energy Prog. 84:71-80. Wyman, C.E. 1999. Biomass ethanol: technical progress, opportunities, and commer cial challenges. Annu. Rev. Energ y Environ. 24:189-226. Wyman, C.E. 2003. Potential s yne rgies a nd challe nges in r efining cellulosi c biomass to fuels, chemicals, and power. Biotechnol. Prog . 19:254-262. Yang, Y ., M. Peredelchuk, G.N. Bennet, and K. San. 2000. Eff ect of Variation of Klebsiella pneumoniae Acetolactate Sy nthase Expression on Metabolic Flux Redistribution in Escheric hia coli . Biotechnol. Bioengin. 69:150-159. Yomano, L.P., S.W. York, and L.O . Ingram. 1998. Isolation and charac terization of eth ano l tol era nt mu tan ts o f Es che ric hia col i KO 11 f or f uel eth ano l pr odu cti on. J. Industrial Microbiol. Biotechnol. 20: 132-138. Zaldivar, J. and L.O. I ngram. 1999. Effect of organic ac ids on the growth and fermentation of ethanologenic Escheric hia coli LY01. Bi otech nol. Bioeng. 66:203-210. Zaldivar, J., A. Martinez and L.O. I ngram. 2000. Effect of alcohol components of hemicell ulose hy droly sate on th e grow th and fe rmentati on of etha nologe nic Escheric hia coli LY01. Biotechnol. Bioeng . 68:524-530. Zaldivar, J., J. Nielsen, L. Olson. 2001. Fuel eth anol prod uction fr om lignoc ellulose: a challen ge for metabolic engine ering a nd proce ss integ ration. A ppl. Micr obiol. Biotechnol. 56:17-34. Zab riski e, D. W. , W.B. Arm iger, D .H. Phillips , and P.A. Albano.1 984. Tra ders gu ide to fermentation media formulation. Traders Protein. Memphis, TN. Zhang, J.and R. Greasham. 1999. Chemically defined me dia for commercial fermentations. Appl. Microbiol. Biotechnol. 51:407-421. Zhang, M., C. Eddy , K . Deanda, M. Finkelstein, and S. Picatagg io. 1995. Metabolic engineering of a pe ntose metabolism pathway in ethanolog enic Zymomonas mobilis . Science 267:240-243. Zhou, S., T.B. Causey , A. Hasona, K.T. Shanmugam, and L .O. Ingra m. 2003. Production of optically pure D -lactic acid in mineral salts medium by metabolically eng ineered Escheric hia coli W3110. Appl Environ Microbiol. 69:399-407.

PAGE 115

104 Zhou, S., F.C. Davis, and L.O. I ngram. 2001. Gene integration and expression and extracellular secretion of Erwinia chrysanthemi endoglucanase CelY ( celY ) and CelZ ( celZ ) in ethanologenic Klebsiella oxytoca P2. Appl. E nviron. Mi crobiol. 67:6-14. Zhou, S. and L.O. I ngram. 1999. Engineering endog lucanase-secreting strains of ethanologenic Klebsiella oxytoca P2. J. Ind. Microbiol. Biotechnol. 22:600-607. Zhou, S. and L.O. I ngram. 2000. Synerg istic hydroly sis of carboxymethy l cellulose and acid-sw ollen ce llulose by two endog lucanas es (EGY a nd EGZ) from Erwinia chrysant hemi. J. Bacteriol. 182:5676-5682. Zhou, S. and L.O. I ngram. 2001. Simultaneous saccharification and fermentation of amorphou s cellulos e to etha nol by recombi nant Klebsiella oxytoca SZ21 without supplemental cellulase. Biotechnol. L ett. 23:1455-1462.

PAGE 116

105 BIO GRAPHI CAL SK ETCH Brent E . Wood atte nded the U niversity of Flori da as an u nderg raduate beginn ing in 1986. He received his B.S. in microbiology and cell science in December 1991. He then worked in industry until Dece mber 1994 when he enrolled in the Graduate School of the University of Flor ida beginning in January on 1995. He then received a Master of Science degree from the D epartment of Microbiology and Cell Science in May 1997. After another four y ears in industry, he r ejoined the Graduate School at the University of Florida to pursue his Ph.D. H e will re join his for mer emplo ye r after completio n of his degree.