2.5K:

0‘. «I
1.4.! km
méfi .. .a

. “3m... .... fi
. ‘1’ .

.42}... an:

:3. r ‘ u.
(.1: 1: $9
in... ”am

I.

A

 

 

 

 

. f... .15.
.5; .4..-

.. flan.-.

 

k. ‘ . K £4§3$§
. . u , ‘ . .. ‘ «aha; 1.1.x. .
flflwéfiia, . ,. , ,A a . .V , 7, L ‘

 

A LIBRARIES
MICHIGAN STATE UNIVERSITY
EAST LANSING, MICH 48824-1048

This is to certify that the
dissertation entitled

STUDIES OF THE ACTIVATION AND DEACTIVATION OF
PYRUVATE FORMATE-LYASE FROM ESCHERICHIA COL!

presented by

Mbako Ronald Nnyepi

has been accepted towards fulfillment
of the requirements for the

 

 

 

 

PhD. degree in Chemistry
JP pf z/O ’ 4—,“
/ Major Professor s Signature
_D__;_(_jx<fj(‘6/C /é/ ZOOV
Date

MSU is an Affirmative Action/Equal Opportunity Institution

o- nvo» ~I-.—u~‘-n-onn‘. -

-. ... .-.u.-.-u-—-.-—o--—.—-—-—.-- -‘-.—-—-—-—.-.-—.._

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJClFlC/DateDuepSS—p. 15

STUDIES OF THE ACTIVATION AND DEACTIVATION OF PYRUVATE
FORMATE-LYASE FROM ESCHERICHIA COLI
By

Mbako Ronald Nnyepi

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

2004

ABSTRACT
STUDIES OF THE ACTIVATION AND DEACTIVATION OF PYRUVATE
FORMATE-LYASE FROM ESCHERICHIA COLI
By

Mbako Ronald Nnyepi

This study set out to investigate the conditions necessary for deactivation of
pyruvate formate-lyase from'Escherichia coli by alcohol dehydrogenase (AdhE) from the
same species. One of the objectives was to gain some insight into the mechanism by
which AdhE achieves this reported deactivation of PF L, and particularly what role each
of the putative cofactors implicated in this activity, plays in the mechanism.

To that end, the adhE gene was successfully engineered and cloned onto a
commercial vector using standard molecular biological techniques. Recombinant plasmid
DNA engineered as such was then transformed into a DNA amplification Escherichia
coli strain, purified, fingerprinted, and subsequently introduced to a protein expression
strain for overproduction of AdhE. AdhE was chromatographically purified to
homogeneity and characterized by gel electrophoresis as well as by enzymatic assays
with respect to its known activities. Pyruvate formate-lyase and pyruvate formate-lyase
activating enzyme (PFL-AE) were purified according to procedures optimized in Dr. Joan

Broderick’s research laboratory.

A new method for activating PF L was developed whereby the activation mix was
allowed to react at room temperature and ambient light. This was different from the
methd previously used in our lab in which the activation mix was intensely illuminated in
ice. PF L activated by the new method was characterized both by electron paramagnetic
resonance (EPR) and by a coupled enzymatic method.

Systematic deactivation studies of PFL using various combinations of AdhE and
the putative cofactors COA, Fe2+and NAD+ revealed that the presence of AdhE did not
make a difference to the rate and extent of PF L deactivation. Our results, however,
Showed that both chemical and biological reductants, including Fe2+ and, to a lesser
extent COA, deactivate PFL in a significant and systematic manner. Investigation of
thiols also revealed that their rate of PF L deactivation is generally inversely related to
their Size, thus suggesting diffusion-limited access to the buried radical of active PF L. It
was Significant to observe that after deactivation by thiols, PF L can be fully reactivated.

This study also revealed that NADH is a relatively potent deactivator of PFL and
that the presence of AdhE stabilizes active PF L towards deactivation by this coenzyme.
The presence of AdhE was also found to confer a greater degree of reactivatability to
PFL, compared to PFL that is deactivated in the absence of AdhE. Our findings have thus
revealed that AdhE, rather than deactivate PFL, actually provides some kind of protection
for active PF L against deactivation. This role is consistent with tha fact that PF L, which
is activated under anaerobic conditions, is extremely sensitive to oxygen and would
therefore need to be protected from oxygenolytic degradation under microaerophillic
conditions during the transition from anaerobiosis to aerobiosis. Since it is expressed

under similar conditions, AdhE would be a logical candidate for such a protective role.

I dedicate this work in part to my mother Lechedzani Saina Nyepi for her sacrifice and
commitment to the education of all her children, inspite of enormously daunting
economic circumstances. I also dedicate it to my wife Maria for her helping to keep us
both sane in the middle of our bewildering responsibilities as parenting graduate students,

and to our children Mbano, Titi and Bubuya for putting up with less than a fair share of
their father’s time.

iv

ACKNOWLEDGEMENTS

I would like to express Sincere and heartfelt appreciation to my advisor Dr. Joan
Broderick. For the years that I have worked on this project, I have been privileged to
receive from her unreserved academic guidance, personal and sometimes much needed
financial support. It has been a truly special privilege to have such a renouned and
capable scientist as my advisor. From Joan I have come to appreciate that it is possible to
be both a highly successful scientist and still maintain one’s humanity. Thank you for
everything and my best wishes to you in all that you aspire for.

I also thank my guidance committee which comprised Dr. Milton Smith, 111.,
Dr. James Geiger and Dr. Babak Borhan. It is my hope that I will have an opportunity to
meet you again one day, maybe in Botswana!

I would like to thank Dr. Sheila Smith for showing me how to carry out enzymatic
assays. Tim Henshaw was tremendously helpfiIl with the molecular biological techniques
of the project. My thanks also go to Dr. Jennifer Cheek for EPR measurements. I have
also been very fortunate to have a great group of laboratory mates over the years, each of
whom, in their own way, has made my stay at MSU (and particularly in the lab) an
enriching experience. I thank Dan for his occasional assistance with Chemme and (not
less significantly) for keeping the lab environment warm and friendly with his baking,
and for recognizing the birthdays of each group member. Thanks to Jeff for the
stimulating discussions of current events and for his friendship and to Cliff for interesting
perspectives on research topics. It has been a pleasure to have you as a friend.

V

I appreciate Magdalena for always being ready to assist those who need help in
the lab. I am grateful for her input during my preparation for my defence and her
friendliness. I wish all the best to her. I thank and acknowledge Shujuan for sharing her
PFL with me for the last set of experiments when I desperately needed it. Thanks for
being a good friend. I could always count on Meng for her computer skills, whether it be
for editing a figure or preparing for a presentation. I will remember you with much
appreciation. I also acknowledge Sofia, Yi, Liton and Washington for their friendship. I
will always remember you with a smile. My thanks also go to Egis and Thalia for
sharing so much of yourselves with me. My interaction with you has left me culturally
enriched and broadened. Thank you all for being such a great bunch collectively and
Special friends individually.

Lastly, I would like to thank all those who have contributed one way or another to
the advancement of my educational career. I particularly thank Mr. Tafilani “Tat”
Machacha at Botswana Geological Surveys for his encouragement and support when I
joined the department straight from college. I also thank Sjaak Meulenberg for giving me

the opportunity to switch from government work to academia.

vi

TABLE OF CONTENTS

 

 

LIST OF TABLES X1
LIST OF SCHEMES XII
LIST OF FIGURES .......................................................................... VII

CHAPTER I LITERATURE OF STUDIES ON PYRUVATE FORMATE-
LYASE: DISCOVERY, ACTIVATION-CATALYTIC MECHANISMS

 

 

 

 

 

 

 

 

 

 

 

AND DEACTIVATION 1
1.1. INTRODUCTION 2
1.2. EARLY WORK ON PYRUVATE FORMATE-LYASE 3

1.2.1. Discovery 4

1.2.2. The Formate-Pyruvate Exchange System is Oxygen-Sensitive ................... 5

1.2.3. Activation Studies 5

1.2.4. “Enzyme III” and “Fraction IV” are Essential For PFL Activity ............... 6

1.2.5. Enzyme 11 Requires Fe (11) for Activity 7

1.2.6. The Identity of “Enzyme 1” and “Enzyme 11” 9

1.2.7. Comparison of Reducing Systems 10
1.3. STRUCTURE OF PFL 14
1.4.THE RADICAL COFACTOR OF PFL 18

1.5. ACTIVATION OF PFL REQUIRES THE PRESENCE OF PYRUVATE

OR A STRUCTURAL ANALOGUE 19

1.6. THE MECHANISM OF PFL RADICAL GENERATION 21

 

1.7. INVESTIGATIONS OF THE CATALYTIC MECHANISMS OF PFL ....... 24

 

1.8. THEORETICAL STUDIES 4]

REFERENCES 44

 

CHAPTER 11 PFL: PURIFICATION , ACTIVATION AND STABILITY ...... 51

 

ABSTRACT 52

11.1. INTRODUCTION 53

 

vii

11.2. MATERIALS AND METHODS

56

 

11.2.1. Chemicals, bacterial cells and plasmids -

56

 

11.2.2. Bacterial production of PFL

56

 

58

 

11.2.3. Bacterial synthesis of PFL-AE
11.2.4. Purification of PFL

59

 

11.2.5. Determination of PFL concentration

63

 

11.2.6. Activation of PF].

67

 

11.3. RESULTS

71

 

11.3.1. Expression

7]

 

11.3.2. Purification of PFL

71

 

11.3.3. Activation and stability of PFL

72

 

11.4. CONCLUSIONS

73

 

REFERENCES

74

 

CHAPTER 111 STUDIES OF ALCOHOL DEHYDROGENASE FROM
ESCHERICHIA COLI

75

 

ABSTRACT

76

 

111.1. INTRODUCTION AND BACKGROUND

77

 

111.2. MATERIALS AND METHODS

84

 

111.2.1. Cloning of Hiso-tagged AdhE

84

 

111.2.2. Extraction and analysis of PET-21' (+)-ath.

85

 

111.2.3.Tranformation of BL21(DE3)pLysS with PET-21a(+)-adhE ..............

87

 

111.2.4. Streaking of the glycerol stocks to obtain colonies
111.2.5. Growth, Harvesting and Lysis

88
90

 

111.2.6. Purification of His,- AdhE

91

 

91

 

[11.2.6.1. Purification Principle
[11.2.6.2. Purification and Analysis of Hisg-tagged AdhE

92

 

111.2.7. Cloning, Growth and Purification of untagged AdhE

96

 

96

 

111.2.7.]. PCR and ligation

111.2.7.2. Transformation, plating, DNA extraction and fingerprinting .......

111.2.7.3. Growth of Cell Cultures and Purification

 

111.2.7.4. Characterization of untagged AdhE

 

 

111.2.7.5. Characterization of the His.;-tagged AdhE

viii

.100

104
110
111

111.3. RESULTS 115

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

111.3.1. Over-expression and Characterization of Untagged AdhE .................. 115
[11.3.1.]. Overexpression of untagged AdhE 115
[11.3.1.2. Characterization of AdhE 116

111.3.2. Over-expression and Purification of His,- AdhE 117
111.3.2.]. Test growths 117
111.3.2.2. Purification ...................................................................... 118
111.3.2.3. Alcohol Dehydrogenase Activities 119
III.3.2.4. Iron Content of AdhE 123
[11.3.2.5. Comparison of “as isolated” and iron-stripped AdhE activities ..... 124
[11.3.2.6. Effect of Fe2+ on the alcohol dehydrogenase activity of AdhE ........ 126
111.3.2.7. Investigation of PFL deactivation 128

111.4. DISCUSSION 139
111.5. CONCLUSIONS 142
REFERENCES 143
CHAPTER 1V DEACTIVATION OF PFL BY SMALL MOLECULES .......... 144
ABSTRACT 145
1V.1. INTRODUCTION 146
1V.2. MATERIALS AND METHODS 147

1V.2.1. Materials 147

1V.2.2. The PFL deactivation experiment 148
1V.3. RESULTS 150

1V.3.1.Deactivation by DTT and fl-MF 150

1V.3.2. Effects of Illumination on the deactivation of PFL 152

1V.3.3.Deactivation of PFL by other thiols and small molecules 156

1V.3.4. Investigation of PFL deactivation by other small molecules ................. 157

1V.3.5. Comparison of methionine and cysteine 160

1V.3.6. Other non-thiol reducing agents 161

1V.3.7. Deactivation of PFL by B-ME is reversible 162

1V.3.8. Effect of the NADH/NAD+ Ratio on PFL Deactivation by ADHE ....... 166

1V.4. DISCUSSION AND CONCLUSIONS - 176

 

REFERENCES 179

 

ix

CHAPTER V: MOTIVATION FOR THIS STUDY, SUMMARY OF

 

 

 

FINDINGS AND FURTHER INVESTIGATIONS 180
V.1. MOTIVATION FOR THIS STUDY 181
V.2. ACTIVATION AND STABILITY OF PFL 183

V.2.]. Some Philosophical and Practical Considerations on Activation ........... 183

V.2.2. Activation Without Intense Illumination is Advantageous 184

 

V.2.3. In vivo PFL Activation Employs A light-independent Reductant ........ 184

 

 

 

V.3. STUDIES OF THE DEACTIVATION OF PFL BY ADHE - 186
V.4. CONTEXT, RELEVANCE AND OTHER STUDIES 189
V.5. DEACTIVATION BY THIOLS AND OTHER REDUCTANTS ................ 196
V.6. SUGGESTED EXPERIMENTS FOR FURTHER STUDIES 199

V.6.1. Mechanistic Studies 199

 

V.6.2. In vivo studies 199

 

List of Tables

Table 1.1. Minimum required conditions for PFL activation ............................................ 8
Table 1.2. Catalytic Efficiency of PF L-activase in Various Reducing Systems ........... 11
Table 1.3. Minimum required conditions for PFL activation ................................. 19
Table 11.1. Mixes for the Bradford assay ......................................................................... 64
Table 11.2. Typical series of standards used for calibration ............................................. 65
Table 11.3. The typical protein concentration caculated fi'om the calibration curve ........ 66
Table 111.1. PCR Reaction Components ......................................................................... 98
Table 111.2. Mixes for experimental and control PETBlue-adhE fingerprinting. .......... 102
Table 111.3. Activities of crude extracts from Tuner(DE3)pLacI-adhE ........................ 106
Table 111.4. Typical composition of mixes used for ADH activity of AdhE ................. 110

Table 111.5. Typical composition of mixes for alcohol dehydrogenase activity assay .. 112
Table 111.6. Typical composition of mixes used for acetaldehyde assay ....................... 113
Table 111.7. The components of the mix used to investigate PFL deactivation ............. 130

Table 111.8. Reaction mixes to investigate deactivation under anaerobic conditions 133

Table 111.9. Isolating the species responsible for PFL deactivation. ........................... 1354
Table 111.10. Effect of Fe(II) on Extent of PF L Activation ........................................... 137
Table IV.1. Residual PFL activities of PFL with small molecules ......................... 159

Table 1V.2. Mixes to study effect of NADH/ NAD+ on PFL deactivase of AdhE. ...... 168

Table 1V.3. Mixes for investigation ofNADH and AdhE effect on PFL deactivation . 174

xi

List of Schemes

Scheme 1.1. The first committed step in pyruvate catabolism following g1ycolysis........3
Scheme 1.2. The two steps of the anaerobic pyruvate “dissimilation ................................ 4
Scheme 1.3. The Structural Formulae of Pyruvate and Oxarnate. .................................... 19
Scheme 1.4.The overall reaction catalyzed by PFL .......................................................... 25
Scheme 1.5. The Ping-Pong Mechanism proposed by Knappe and co-workers .............. 26

Scheme 1.6. The mechanism of PF L action as proposed by Kozarich and co-workers. . .28
Scheme 1.7. Formation of the acety-PF L radical intermediate ....................................... 30
Scheme 1.8. Hypophosphite and acetyl-PF L; acetylphosphinate and active PF L . .33
Scheme 1.9. The salt bridge between oxarnate and Arg 435 in the active site of PFL ..... 35
Scheme 1.10. Chemical structures of pyruvate and alpha-oxobutyrate ............................ 36
Scheme 1.11. Comparison of residues found at the active sites of Tch and PFL. . . . . ....37

Scheme 1.12. Proposed mechanism of transfer of the radical and acetyl group of PF L ...38

Scheme 1.13. The mechanism proposed by Knappe and co-workers. ...................... 39
Scheme 1.14. Mechanism Proposed by Himo and Siegahn ............................................ 43
Scheme 11.1. A structural depiction of the activation of PFL ......................................... 53
Scheme 111.1. A Schematic of glycolysis .......................................................................... 78

Scheme 111.2. TCA: acetyl—COA oxidation, production of the NADH and FADHZ. ....... 79

Scheme 111.3. Reduction of pyruvate to ethanol during alcoholic fermentation .............. 81
Scheme 111.4. Two reations catalyzed by AdhE. ............................................................. 82
Scheme 111.5. The reaction catalyzed by PFL .................................................................. 82
Scheme 111.6. The proposed role of AdhE as a PF L deactivase ...................................... 83
Scheme 111.7. The ADH/acetaldehyde reductase reaction catalyzed by AdhE . ...l 12

xii

Scheme 111.8. The coupled enzymatic assay for the PFL activity .................... 129

Scheme IV.1. Enzymatic activity assay used to monitor residual activity of PFL . ...... 149

xiii

List of Figures

(Images in this dissertation are presented in color
Figure 1.1. Plot of redox potentials necessary for PFL activation .................................... )10

Figure 1.2.Time course of the activation of PF L with variously treated PFL-activase 12

Figure 1.3. Anaerobic gel filtration chromatograph of activation mixture. ..................... 13
Figure 1.4. Cartoons of the partial crystal structure of PF L ............................................. 16
Figure 11.1. A Chromatogram of PF L purification by ion exchange. .............................. 60
Figure 11.2. SDS-PAGE of PFL fractions obtained from the first. .................................. 61
Figure 11.3. A Chromatogram from the second purification of PF L ................................ 62
Figure 11.4. Calibration curve used to calculate the concentration of PFL. ..................... 66
Figure 11.5. The EPR spectrum of the glycyl radical on activated PF L .......................... 68
Figure 11.6. A time course activation of PF L under ambient light and 23°C ................... 69
Figure 111.1.The primers used to amplify the AdhE gene of E. coli by PCR .................. 84
Figure 111.2. Fingerprints of DNA digested with Xho I and BamH-I ............................ 86

Figure 111.3. Streaking a glycerol stock onto an agar plate for overnight incubation. 89
Figure 111.4. Typical appearance of colonies after overnight incubation ........................ 89
Figure 111.5. Ni is coordinated to a nitrilotriacetic acid (NTA) support. ........................ 91

Figure 111.6. Interactions between NTA-bound nickel and histidines from the H156. ..... 92

Figure 111.7. Effect of protein load on purification quality. ............................................ 94
Figure 111.8. The oligonucleotide primers used to amplify untagged AdhE. .................. 97
Figure 111.9. The adhE PCR product for the untagged AdhE ......................................... 99
Figure 111.10. Agarose gel of PETBlue—adhE plasmid digested with Scal. .................. 103
Figure 111.11. DNA finger-prints of plasmids extracted from colonies 2, 3 and 4 ........ 104
Figure 111.12. Overexpression of untagged AdhE by colonies 1 and 2 ......................... 105

xiv

Figure 111.13. Whole cell SDS-PAGE of Tuner (DE3) pLacI cells .............................. 115
Figure 111.14. AdhE lacking a His6 tag was also partially purified by gel filtration ..... 116
Figure 111.15. SDS-PAGE of BL21(DB3) cells transformed with PETZIaM-adhE. 117
Figure 111.16. The level of purity achieved for Hisé-tagged AdhE using a Ni-NTA 119

Figure 111.17. Acetaldehyde reductase/alcohol dehydrogenase activities of by AdhE. 121

Figure 111.18. Activity was measured by monitoring AA340 of NADH ........................ 122
Figure 111.19. Correlation between the amount of AdhE and Fe ................................... 123
Figure 111.20. Comparison of specific activity Of apo-AdhE and holo-AdhE .............. 125
Figure 111.21. Effect of Fe“ on the activity of holo- and apo-AdhE ............................. 126

Figure 111.22. Time course of PFL deactivation by AdhE, NAD+, Fe2+ and CoA... 13129
Figure 111.23. Time course of PF L deactivation by AdhE, NAD+, Fe2+ and CoA ..... 1321

Figure 111.24. PFL deactivation using various combinations of cofactors and AdhE..1332

Figure 111.25. Isolation the component responsible for PFL deactivation ................... 1353
Figure 111.26. Effect of Fe (II) in the reactivation of PFL ...................................... 135
Figure 111.27. Effect of Fe (II) on the reactivation of PFL ................................. 137
Figure 1V.l. Small molecules used in deactivation studies ................................. 147
Figure 1V.2. The effect of DTT and B-ME on the activity of PFL ................................ 151

Figure 1V.3.The effect of DTT on the activity of PFL over extended period of time... 152
Figure IV.4. Effect of DTT and B-ME on the activity of PF L under ambient light. ..... 153
Figure IV.5. Deactivation of previously reactivated PFL ............................................. 155
Figure [V.6. PFL deactivation properties of ethanethiol, DTT, and B-ME ................... 156
Figure 1V.7.Comparison of deactivation rates of various thiols and small molecules. 158

Figure 1V.8.Comparison of methionine and cysteine as PFL deactivators. .................. 160

XV

Figure 1V .9. Effect of non-thiolic reductants on the activity of PFL ............................. 162
Figure IV .10. PFL deactivation by B-ME. ..................................................................... 164

Figure 1V.11. Effect of different NADH/NAD+ ratios and AdhE on PFL activity 169

Figure 1V.12. The residual PFL activities after 9 hours of incubation .......................... 171
Figure 1V.13. PF L activation-deactivation cycles in the reaction mixes. ...................... 172
Figure 1V.14. The effect of NADH and AdhE on the activity of PFL. ......................... 175

xvi

CHAPTER I

LITERATURE OF STUDIES ON PYRUVATE FORMATE-LYASE: DISCOVERY,
ACTIVATION-CATALYTIC MECHANISMS AND DEACTIVATION

1.1 . INTRODUCTION

As a facultative anaerobe, Escherichia coli (E.coli) is capable of growing under
both aerobic and anaerobic conditions. In the presence of oxygen, this organism
metabolizes glucose via a pathway that employs pyruvate dehydrogenase to convert
pyruvate to acetyl-COA. The next step in this pathway is the consumption of acetyl-COA
in the Kreb’s cycle, wherefrom oxidative phosphorylation leads to the production of 38
molecules of ATP per glucose molecule. However when conditions are anaerobic, an
alternative metabolic route is followed, whereby pyruvate formate-lyase (PFL) catalyses
the reversible and non-oxidative cleavage of pyruvate to acetyl-COA and formate.l The
acetyl-COA is then fed into an alternative metabolic pathway that generates only 4
molecules of ATP per molecule of glucose. Thus, pyruvate dehydrogenase and pyruvate
formate-lyase catalyze the first committed aerobic and anaerobic metabolic steps,
respectively, in the catabolism of the pyruvate generated by glycolysis. The two
reactions are summarized in Scheme 1.1, where A represents the anaerobic reaction and B
represents the reaction under aerobic conditions. Though adaptable to both aerobic and
anaerobic conditions, Escherichia coli (E.coli) grows more efficiently under aerobic
conditions, as more ATP is produced per molecule of glucose, reverting to the anaerobic

catabolic route only when necessary.

 

 

 

 

 

CoA-SH H0P032' 4 ADP
O i )1 )OK
I PTA acetate kinase
0- PFL H c OPO 2' -
H3C)Y H3C S-CoA 3 3 H3C O
O >
HJLO-
8
C02
CoA-SH
O . .
_. Oxrdatuve
join/D pyruvate k /l|\ Krebs NADH
H30 : HsC SCOA — Cycle FADH —— Phos horylation
O dehydrogenase 2 p
/\ /\
38ADP 02
CO2
38ATP H20

Scheme 1.1. The first committed step in pyruvate catabolism following
glycolysis under aerobic conditions (A) and anaerobic conditions (B).
PF L, Pyruvate formate-lyase; PTA, phosphotransacetylase.

1.2. EARLY WORK ON PYRUVATE FORMATE-LYASE
3

1.2.1. Discovery

The anaerobic “dissimilation” of pyruvate was first discovered in 1943 by
Kalnitsky and Werkrnan.2 Lipmann and Tuttle later found that cell extracts of E.coli
required prolonged incubation under assay conditions in order to induce the dissimilation
activity.3 From the work of Chantrenne and Lipmann and that of Stadtman the
“dissimilation” reaction was recognized as proceeding in two steps as shown in Scheme

1.2. 4‘5

 

 

1. pyruvate + CoA acetyl-GOA + formate

 

2. acetyl-CoA + phosphate acetylphosphate + COA

 

 

 

Overall: pyruvate + phosphate acetyl phosphate +formate

 

 

Scheme 1.2. The anaerobic pyruvate “dissimilation” reaction was found to
occur in two steps

Other characteristics of the PF L dissimilation activity from various Species were
investigated by several other workers including Asnis and co-workers, who observed that
the dissimilation activity in E. coli was stimulated by addition of boiled extracts of yeast,
whereas for both E. coli and Micrococcus lactilyticus, stimulation could be achieved by
adding extracts of the respective organisms heated to 65°C. 6‘7’8 These observations
suggest the involvement of non-protein co-factors in the reactions. The intriguing
cleavage of the C-C bond of pyruvate in the first step of pyruvate dissimilation, catalyzed

by the enzyme now known as pyruvate formate-lyase (PFL), led several workers to
4

investigate the mechanism by which this cleavage occurs, and this reaction has continued

to attract the interest of researchers from a broad spectrum of specializations.

1.2.2. The F ormate-Pyruvate Exchange System is Oxygen-Sensitive

During some of their early work on the activity of PF L, Knappe and co-workers
observed that during the dissimilation reaction there was a time lag of several minutes
before formate was produced, after which it was produced at a constant rate. They
attributed this to poisoning of at least one of the essential reaction components by
residual oxygen. In order to test this hypothesis, they added several oxygen-removing
reagents to the substrate mixture of the assay system. Since the enzyme mix was already
supplemented with Fe(11), it was assumed to be already oxygen-fi'ee and was therefore
not treated with any of these oxygen-removing reagents. Interestingly, the previously
observed time lag was abolished by addition of these reagents, thus supporting the
hypothesis that the delay in pyruvate production had been due to poisoning of the PFL
activity by oxygen. It is now known that PFL must be post-translationally activated, with
the activation requiring strictly anaerobic conditions due to the PFL activating enzyme
(PFL-AB), and that the activated PFL is also very oxygen sensitive.9’lo Essentially all

reactions of PFL are therefore carried out under strictly anaerobic conditions.

I. 2. 3. Activation Studies
In 1965 Knappe and co-workers demonstrated that the pyruvate dissimilation
activity required the presence of two (then unidentified) protein factors and S-adenosyl-

L-methionine (SAM).ll Wood had previously used isotopic labeling to Show that the

5

exchange reaction between formate and the C-2 carbon of pyruvate proceeds at a
physiologically significant rate, and therefore Chase and Rabinowitz were able to use
this exchange reaction to measure catalytic activity by mixing l4C-labeled formate with
unlabelled pyruvate in the presence of PFL.“l3 Since both reactions shown in Scheme
1.2 are reversible, incorporation of 14C was used as evidence that reaction (1) had taken
place and therefore that the pyruvate “dissimilation” activity was present. They referred
to the activity as pyruvate formate-lyase and to the enzyme responsible for this exchange
simply as “the formate-pyruvate exchange system”.13 From this study, these workers
were able to establish that the pyruvate formate-lyase activity was stimulated by the
presence of both pyruvate and SAM and that the effect of the two components was
additive.l3 The work of Knappe and co-workers using sedimentation and anaerobic
sucrose gradient techniques revealed for the first time that there were at least four
components involved in the pyruvate formate-lyase reaction, which they referred to
simply as enzymes 1, 11, H1 and fraction IV.14 Based on subsequent investigations using
primarily ultraviolet-visible spectroscopy and gradient centrifugation techniques, this
team was able to identify enzyme III as a flavoprotein containing 1 molecule of flavin
mononucleotide (FMN). Fraction IV was not identified, but it was recognized to be a

protein. 1 5

1.2. 4. “Enzyme III ” and “Fraction IV" are Essential For PFL Activity
Further investigation showed that exclusion of either one of “enzyme III” or
“fraction IV” abolished PFL activity, whereas substitution of Co(II) / thiol for both

afforded PFL activity.14 Consequently, these investigators suggested that “enzyme III”

and “fraction 1V” only play an auxiliary role in the PFL reaction.14 Based on the fact that
CO2+ and thiols form strongly reducing complexes, these researchers suggested that the
role of enzyme III and fraction 1V in the PFL reaction is to provide an appropriate redox
potential via pyruvate as an electron donor.16 For the first time this group demonstrated
that, in the presence of either one of the auxiliary systems (Co(II)/thiol or enzyme
III/enzyme IV), the only other condition necessary for pyruvate formate-lyase activity
was the interaction of enzyme 1, enzyme 11 and SAM.16 The fact that these two pairs of
components were essential but interchangeable suggested that they played a vital role in
the activity of the system but did not interaction with the “pyruvate-formate exchange

system” in a specific manner.

1.2. 5. Enzyme 11 Requires Fe (II) for Activity

Contrary to the previously held view that the primary role of iron was to scavenge
for oxygen, subsequent studies had shown that cobalt (II) competitively interfered with
Fe (11) to inhibit pyruvate dissimilation in the “pyruvate-formate exchange system”, thus
leading Knappe and co-workers to conclude that iron actually binds to enzyme 11 as a
prosthetic group.""15 Furthermore, incubation of F e-reconstituted enzyme 11 with a
chelating agent afforded retention of a significant amount of activity, thus suggesting that
Fe was indeed tightly and Specifically bound by the enzyme.14 Having shown that Fe (II)
was essential for the “pyruvate-formate exchange reaction”, it became necessary to
identify the exact component of the system whose activity required Fe (H). In pursuit of
this, a series of reactions were performed in which various components of the exchange

reaction system were pre-incubated with Fe (11) before the reaction was initiated. The

extent of reaction was then determined by measuring the total amount of formate
produced when the components were mixed.

The various pre-incubations investigated are listed in Table 1.1, in which the
amounts of formate produced upon mixing the components together (right column) were
used as an indicator of the activity of the system. The results demonstrated
unequivocally that pre-incubation with Fe (II) was essential for the activity of “enzyme
H” but had no effect on “enzyme 1”. Dithiothreitol (DTT) was also proved necessary for
significant activity of the “exchange system”. However, based on this data it was not

possible to determine which of “enzyme 1” or “enzyme 11” required DTT.

Table 1.1. Initial Turnover Results Identifying Minimal Required Conditions for PFL

 

 

activityl4
Components preincubated with Fe(II) and DTT umoles of formate formed
Enzyme 1 plus enzyme H 1.70
Enzyme 1 0
Enzyme H 1.71
Enzyme 1 plus enzyme H without Fe” 0
Enzymeljlus enzyme H without dithiothreitol 0.1

 

 

The data in Table 1.1 clearly identified enzyme H to be the one specifically
activated by pre-incubation with Fe (11), while Co (11) was not able to activate enzyme 11,
and was indeed found to be inhibitory to activation.14 The findings from this work led to
the first suggestion that the pyruvate formate-lyase reaction is catalyzed by the interaction
of “enzyme 1” and “enzyme 11”. The kinetics of the PFL reaction suggested that the
nature of the interaction of “enzyme 11” with enzyme 1 during the PFL reaction was

catalytic, since the amount of formate produced was found to be dependent only on the
8

amount of “enzyme 1”.14 These kinetics were observed when Fe (II) was used alone, or

when it was present in excess over Co (11). However, when Co (11) was present in excess,
the rate of formate production depended on both the concentration of both “enzyme 1”
and “enzyme H”, indicating that enzyme 11 was activated by binding Fe (II), and that Co
(11) competed with Fe (II) for the binding site on the enzyme but was itself not a viable
co-factor. ‘4

1.2. 6. The Identity of “Enzyme I" and “Enzyme II”

The work of Knappe and co-workers shed some more light into the identity of the
components required for pyruvate formate-lyase activation and also corroborated a
previous investigation in which it was established that “enzyme 1”, now known as
pyruvate formate-lyase (PF L), was a polypeptide of molecular weight of about 140 kD,
which could be converted from an inactive to an active form upon interaction with
“enzyme 11”, now known as the pyruvate formate-lyase activating enzyme (PF L-AE).14’15
PFL-AE was confirmed to have a molecular weight of about 30 kD as previously
reported and was shown to require pre-incubation with iron and a thiol in order to
convert PFL to the catalytically active form.15 The essential role of S-Adenosyl-L-
methionine (SAM) was again recognized but was at this point believed to be only
allosteric. It was of interest, however, that in addition to the requirement for the absence
of oxygen in the reaction mix, PFL activation was found to be more efficient under
highly reducing conditions.15

At potentials above -390 mV (more oxidizing), no activation was detected,
whereas optimum activation was achieved at -440 mV (Figure 1.1). Once activated and
bound to its substrate (pyruvate), however, PFL was observed to become dramatically

9

more stable, converting to its inactive form only when conditions were more oxidizing
than -170 mV (Figure 1.1 B).16 This turned out to be of practical advantage as it suggested

a means to trap activated PFL by adding pyruvate, as demonstrated by Chase and

Rabinowitz. ' 3

1A 3 120-
3b 100 .
.2 80 ~
0
_§ 60 .
E 40-
51 20 ~

 

 

1 T 1 U fi

-500 ~300 -100 100
Eh [mV]

Figure 1.1. Plot of redox potentials necessary for PFL activation (A) and
stability of the PFL-pyruvate complex. In the presence of the substrate
(pyruvate) active PFL was Shown to withstand relatively highly oxidizing
conditions, only breaking down when conditions became more oxidizing than

about —170 mV (3).“3

1.2. 7. Comparison of Reducing Systems

A number of reducing systems were investigated in an effort to identify one with
the most suitable potential for the activation reaction.l6 Although all of the experiments
were carried out in vitro, they shed some light on the broader aspects regarding the nature
of the reductant components involved in the activation of PF L.16 At least four reducing

systems, including 3 mM Fe2+/ 12 mM 2,3-dimercaptopropanol and 4 mM COZI/ 12 mM

10

DTT, were shown to be suitable for achieving the potential required for PF L activation.1°
However, using photo-reduced flavodoxin, it was observed that the PFL-activase had a
higher catalytic efficiency as shown in Table I.2.'° This reducing system was therefore

employed for the subsequent studies of PF L activation.

'Table 1.2. Catalytic Efficiency of PF L-activase in Various Reducing Systems

 

 

Reducing System PFL,-“active (ug) PF L-activase (ug) PF Lsctivcflmitsf
CoZfiDTT 16 50 1.8
Fe2+-dimercaptopropanol 16 50 1.2
Photoreduced Flavodoxin 16 50 8.6

800 50 395

 

 

 

 

.1 unit corresponds to l umole of formate formed per hour at 28°C. OAdapted from reference; Metabolic
Interconversion of Enzymes, Springer, Berlin Heidelberg New York, pp319-329, Ed.: Wieland, O;
Helmreich, E; Holzer, H.

Further support for the suggestion that ferrous iron actually binds to PFL-AE
came from two important experiments. In the first experiment, the activation reaction
was carried out using PFL-AE that had been previously incubated with ferrous iron and
DTT, leading to successful activation of PFL.1° The other experiment was done similarly,
except that PFL-activase was incubated in 10 mM Trion (a strong iron-chelating agent)
for 10 minutes.16 The activation of PFL with activase from this incubation was
significantly reduced (Figure 1.2).1° The third data set shows that in the absence of
adenosylmethionine, essentially no activity is observed, therefore indicating that it is
likely an essential component in the reaction rather than an allosteric effector as

previously postulated.

11

25-

 

 

1, I I I
s 20 — I
E
'5 I ,
"_', 15 ~ I O O OAE w/Tnon
& l . o I+Fe(l|) +DTT
"6 1° “ Aw/o AdoMet
3 I
- , e
5 5

o t t it . a. .

o 20 4o 60 80

Time (minutes)

Figure 1.2.Time course of the activation of PF L (32ug) with variously treated PFL-activase
(217ug). The reducing agent used was Coy/DTT.1°

Kinetic studies showed that the rate of PFL activation is a function of both
inactive PFL and PFL-activase, thus indicating that the activation process involves direct
interaction between the two proteins.15 Chromatographic studies (Figure 1.3), however,
showed that activated PF L elutes out of a gel filtration column at the same buffer volume
as inactive PFL. In addition, no PFL-activase was detected in the peak of active PFL.
Previously, sedimentation studies by Knappe and co-workers had shown that the
molecular dimensions of PF L do not change upon activation.14 Taken together, these
findings strongly suggested that though both proteins interact with each other during the

activation reaction, they do not bind to each other to form a complex.

12

Activity (Ulml)

450 -
400 -
350 -

to

O

O
1

N N
O 8"
O O
1 1

J

Cl!

0
1

1004

501

 

 

o 1 T 1 1 1 1 1
40 50 60 70 80 90 100 110

Elution Volume (ml)

Figure 1.3. Anaerobic gel filtration chromatograph of a PFL activation mixture
on Sephadex G-100. The peaks at 63 m1 and 90 ml correspond to active PFL and
PFL-activase respectively.16

13

1.3. STRUCTURE OF PFL

PF L is a homodimeric enzyme consisting of 85,000 dalton monomeric subunits,
thus making PFL a 170,000-da1ton enzyme. Goldman and co-workers, following on
earlier work by Knappe and others, reported the first well resolved crystal structure of a
PFL fragment.“18 The crystal structure solved by Knappe and co-workers“ earlier had
been poorly resolved because of the difficulties encountered obtaining isomorphous
heavy metal derivatives, “due to the pair of the adjacent cysteines (C418 and C419).”3°
Goldman and co-workers circumvented this challenge by using a combination of both
multiple isomorphous replacement with anomalous scattering (MIRAS) and multiple
wavelength anomalous diffraction (MAD). They were able to obtain a well-resolved
electron density map of residues 4-614 (Figure 1.4.). The first three and last ten residues
were found to be disordered and were therefore we excluded from the structure analysis.
The structure was fine-tuned by truncating to alanine those surface loops and residues
that showed weak density. The refined crystal structure was resolved to 2.8 ' with an R
factor and Rm, of 22.8 and 25.2 respectively.

Structurally, the protein was found to comprise two domains consisting of parallel
barrels with large excursions (Figure 1.4). The overall structure was found to resemble a
right hand with fingers slightly curved in and the thumb folded from the right across the
palm. Significantly, the catalytically essential C418 and C419 are located on the thumb.
Viewed from the top of the barrels, the structure resembles the capital letter G written

backward, with the thumb located on the inside and a B-finger on the outside curvature of

the letter (Figure 1.4 B). These workers thus referred to the barrel as the “G-barrel”.l7

14

The excursions wrap around the G-barrel and create the finger domain. Goldman
and co-workers established that there are a total of five “up” and three “down” strands all
forming part of the 01/13 barrel. C419 and C418 reside on a unique B-finger that links the
two sets of B-Sheets. One of the significant findings from this work is that PFL has
structural homology to anaerobic ribonucleotide reductase III (RNR-III), an enzyme that,
in its active form, also harbors a glycine-centered radica1.2° Both PFL and RNR—HI also
require SAM and an activating enzyme for the generation of the glycyl radical co-
factor.°°‘2° These two enzymes have also been shown by several workers to have
extensive structural and mechanistic similarities.4°‘°‘l° Subsequently, it was discovered
that the activating enzymes for both belong to the same superfarnily of SAM—dependant

radical enzymes referred to as the “Radical SAM Superfamily.”5 ‘

15

 

Figure 1.4. MOLSCRIPT and Raster3D Cartoons of the partial crystal structure

of PFL as solved by Goldman and co-workers (adapted from reference 17).
Helices are shown as spirals and strands as ribbons. (A) The overall fold of PFL
in stereo, shown from the front to emphasize the right-hand nature of the
structure. A right hand is drawn to the side for comparison. The three large
excursions are labeled El, E2 and E3. (B) The left-handed G-barrel of PFL in
stereo, viewed from the ‘top’ where substrate enters. The secondary structure
elements are labeled. '7

In a later study, Becker et al. reported the crystal structure of 11111 length but
inactive PFL co-crystallized with oxarnate, as well as the crystal structure of the double
mutant C418A, C419A.°° Whereas the previous workers had concluded that there were

three “down” strands,20 these workers described the structure as comprising a “10-

16

stranded (it/B barrel assembled in an anti-parallel manner from two parallel five-stranded
B-Sheets.”5° However, the Becker structure also confirmed the earlier observation by
Goldman and co-workers that PF L bears homology to RNR—III.17

It is noteworthy that due to the difficulty of obtaining isomorphous heavy-metal
derivatives for PFL, Becker et a1. capitalized on the fact that the reflection intensities of
the double mutant (C418A/C419A) crystals have very good correlation to those of the
data of wild type PFL co-crystallized with oxamate. They therefore used a difference
Fourier technique to elucidate the wild-type structure and its complex with oxamate.5°
From this work it was established that G734 and C419 are located at the tips of two
opposing loops, again Showing consistency with the findings of Goldman and co—
workers.17 Furthermore, this work revealed that the two residues meet at the center of the
barrel structure. The catalytically essential C418 was also confirmed to be in
juxtaposition with C419, and oxamate was found to be lodged in a tight cleft with C2
symmetry at a distance of only 3.3A from the sulfur of C418. The sulfur atoms of C418

and C419 were found to be 3.7A apart.

l7

1.4. THE RADICAL COFACTOR OF PFL
Active PF L exists as a radical.2° Knappe and co—workers, through 013 and N-15
isotopic labeling techniques and EPR spectroscopy first identified the exact location of

l. 22 The splitting of the EPR signal observed as a result of coupling between the

the radica
spins of the glycyl radical and the isotopically labeled atoms was used as the main
analytical tool in this investigation. This study also revealed that the activation process
involves the abstraction of the pro-S hydrogen of the Gly-734 residue by a 5’-
deoxyadenosyl radical, as proven by incorporation of the hydrogen isotopic label into 5’-
deoxyadenosine.22 In addition, when active PF L was exposed to oxygen, it underwent
rapid oxidative cleavage leading to the formation of two fragments of 82 and 3 kDa
respectively. N-terminal block on the 3 kDa fragment and analysis by mass spectrometry
identified the small fragment as an oxylyl residue derived from G734, thus assigning the
position of the glycyl radical to G734. Following this assignment, the vast majority of
work done on PF L was centered on the questions 1) what is the mechanism by which the
glycyl radical of activated PFL is generated? and, 2) how does active PFL proceed to
mediate homolysis of the C-C bond during catalytic pyruvate turnover. The first of these
questions involves the participation of another enzyme, namely, pyruvate formate-lyase
activating enzyme (PFL-AE). Both of these questions have continued to receive a great

. ~ ‘4
deal of attention fi'om several researchers.9 33 5

18

1.5. ACTIVATION OF PFL REQUIRES THE PRESENCE OF PYRUVATE OR A
STRUCTURAL ANALOGUE

The observation that PF L activation required the presence of the substrate,
pyruvate, raised the question as to whether the activation of PF L and its subsequent
catalysis of pyruvate homolysis were intimately related, or whether the two processes
were separate and distinct from one another. To address this question, PFL was activated
in the presence of pyruvate and its structural analogue oxamate, which does not act as a
substrate, in separate reactions (Scheme 1.3). As a negative control, another activation
reaction was carried out in the absence of both pyruvate and oxamate and a fourth was
done in the absence of SAM. The level of PF L activity was measured and the results are
shown in Table 1.3. Oxamate was found to have the same effect on the activation of PFL
as pyruvate both qualitatively and quantitatively. There was no detectable difference in

the extent of PF L activation when using oxamate instead of pyruvate as indicated by the

data on Table 1.3.
O O
\\ //° \\ //°
/C—C\ /C—C\
0 CH3 0 NH2
pyruvate oxamate

Scheme 1.3. The Structural Formulae of Pyruvate and Oxamate.

Table 1. 3. Minimum required conditions for PFL activation1°

 

 

 

Activation reaction ACtiVC PFL units
Complete including pyruvate 6.2
Without pyruvate 0.1
Pyruvate replaced by oxamate 6.5
Without S-Adenosyl-L-adenosyl methionine O

 

l9

It was significant that oxamate proved to be non-competitive to pyruvate for
activated PFL in the dissimilation reaction, in spite of its facility to replace pyruvate in
the activation reaction.16 This observation led to the suggestion that oxamate (and
therefore pyruvate) probably has a regulatory role, whereby it binds to the activating
enzyme, as was also suggested for SAM.l3 More recently, however, Knappe and co-
workers solved the crystal structure of the double mutant of PFL (C4188, C4198 ) co-
crystallized with both pyruvate and oxamate, from which they were able to Show that
both these molecules do bind to PF L.2° Interestingly, and perhaps uniquely, pyruvate and
oxamate are not just allosteric effectors; they actually bind to the active site of PFL. The
binding of pyruvate to PFL, therefore, presumably results in an essential conformational

change on the enzyme, which in turn, affords the subsequent steps leading to activation.

20

1.6. THE MECHANISM OF PFL RADICAL GENERATION

The work of Conradt et a1. revealed that PF L is translated in an inactive form
(PFL,) and is subsequently converted to the active form (PFL,) by a sophisticated post-
translational mechanism involving an activating enzyme (PFL-activase) and a number of
co-factors.19 Owing to the work of Knappe and co—workers, it is now well accepted that
the active form of PFL comprises a glycyl radical that is produced by abstraction of a
hydrogen atom at Gly-734. 2°

Based on EPR spectroscopy and isotopic labeling techniques, active PFL has been
shown to consist of a free radical co-factor located on the protein backbone. 193° Knappe
and co-workers further probed the mechanism by which the radical at G734 is
generated.22 They activated PF L by illuminating a mix containing PFL, PFL-AE,
riboflavin as a photoreductant, oxamate as an allosteric effector, 5’-deoxyadenosyl-L-
methionine, and DTT, all in MOPS buffer, pH 7.2. By using deuterium labeling at G734
and nuclear magnetic resonance (NMR) and mass spectrometric (MS) analyses, they
observed that 2H was stoichiometrically incorporated at the methyl group of 5’-
deoxyadenosine, a byproduct of the activation reaction.22 In the same study, this team
used synthetic short chain peptides to model the region containing G734. The objective
was to gain insight into the features of the active site that are responsible for recognition
between PFL and PFL-AB. The approach entailed identification of peptides that are able
to produce l4C-labeled 5’-deoxyadenosine from l4C-labeled SAM under activation
conditions. Significantly, and as predicted, the models that showed activity in this
respect also proved to be competitive inhibitors to PF L during activation, that is, to the
generation of the G734 radical.22 A small peptide of sequence Arg-Aal-Ser-Gly—Tyr-Ala-

21

Val, corresponding to the sequence of residues 731-737 of PFL, was used as the model to
investigate the interaction between PF L and PFL-AE. The results of this investigation
showed that substitution of glycine for L-alanine preserved the affinity of the peptide for
PFL-AE, whereas replacement with D-alanine abolished the affinity, thus proving that
interaction between PFL and its activating enzyme is stereospecific with respect to
Gly734. Taken together with the observed incorporation of 2H into 5’-deoxyadenosine at
the methyl carbon (as determined by isotopic labeling and mass spectrometry), two
conclusions were reached:

1) The abstraction of the hydrogen at C-2 of G734 is stereospecifically pro-S

and

2) The abstractng chemical species is a 5’-deoxyadenosyl radical.
At this point, however, the involvement of the 5’-deoxyadenosyl radical was only based
on strong circumstantial evidence, rather than direct observation, and the PFL radical was
surmised to reside on a B-turn segment of PFL.22 At this point, the nature of and the
mechanism operational at the active site of PF L were not well understood.

Based on work done previously by various other workers, the solvent
exchangeability of the G734 hydrogen had been unequivocally established through
isotopic labeling and EPR techniques. It was now known, for instance, that activation of
PF L in D20 resulted in a singlet EPR signal as opposed to the doublet (due to spin-
coupling to the Iii-hydrogen) observed when the activation was carried out in H20.23 It
was also known at this point, primarily due to the work of Gautney and Miyagawa 24, that
the glycyl radical undergoes hydrogen exchange both intra- and inter-molecularly, with
the former being too rapid to be observed by EPR.23 Indeed it was on the basis of these

22

findings that Knappe and co-workers suggested that the (it-hydrogen of G734 must have
an uncharacteristically low pKa.2°

In order to push the fi'ontiers of understanding of the mechanistic features of the
exchange reaction further, Kozarich and co-workers carried out site-directed mutagenesis
and mechanism-based inactivation studies on PF L.23 They discovered that when wild type
PFL, or the mutants C4188 and C4198 were activated in H20, EPR recorded doublet
Signal. The doublet arose from magnetic coupling between the radical on Gly734 and
hydrogen on the tit-hydrogen. When activated wild type PFL is incubated in D20, an
exchange takes place between the a-H and D. This exchange between the magnetically
active a-hydrogen with deuterium from D20 results in the EPR doublet collapsing into a
Singlet. When Kozarich and co-workers activated wild type PF L and the mutants C4188
and C4198 in H20, followed by equilibration with excess D20, the glycyl radical EPR
Signal of the C4188 mutant and wild type PFL collapsed from the characteristic doublet
to a singlet, whereas the signal remained a doublet in the case of C4198. This finding
provided evidence that the a-hydrogen in the C4188 mutant exchanges with deuterium
from the solvent and that this exchange does not involve C418. 0n the other hand, the
fact that the glycyl radical EPR Signal of C4198 remained as a doublet, even after
incubating with D20 over a relatively extended period of time, proved that C419 is
1 directly involved in the mechanism of (it-hydrogen exchange between G734 and solvent.
This finding would prove to be one of the fundamental considerations in subsequent
proposals of the mechanism by which PFL accomplishes the homolytic cleavage of

pyruvate during catalytic turnover.

23

1.7. INVESTIGATIONS OF THE CATALYTIC MECHANISMS OF PFL

One of the significant findings in the investigation of the catalytic mechanism of
PFL was the discovery that active PF L can be acetylated with preservation of the radical
cofactor. The acetyl-PFL intermediate was generated by incubating active PFL with
pyruvate in the absence of COA and formate.23 Exclusion of COA eliminated the
conditions necessary for tumover, as this coenzyme is the co-substrate for PFL. Formate
was excluded from the reaction in order to ensure that the position of the equilibrium was
favorable for the production of detectable amounts of acetylated PF L, as formate is one
of the products of the acetylation reaction. The effect of PFL acetylation on hydrogen
exchange was helpful in shedding some light on the catalytic mechanism of PFL.”
Consistent with the findings of Unkrig and co-workers”, the EPR signal of acetylated
PFL was found to be a doublet identical to that of free PFL. When acetylated PF L was
incubated in D20, the doublet collapsed into a singlet, once again indicating that the a-
hydrogen is still solvent exchangeable. This observation, taken together with the
observation that mutation of C419 abolishes hydrogen exchange between G734 and
solvent”, provided indirect evidence that acetylation of active PF L takes place at C418.

After the aforementioned studies and the data obtained therefrom, the conditions
required for activation of PF L were reasonably well established. Another intriguing
question that needed to be addressed, however, involved the mechanistic details by which
activated PFL mediates the non-redox homolytic cleavage of the C-C bond of pyruvate to
transfer the acetyl group to COA and produce formate in the process. Knappe and co-
workers probed this mechanism by using elegant inhibition kinetics experiments.15 When

they added formate to the PFL reaction mix and measured the rate of COA consumption,

24

they observed competitive inhibition. Given that the overall PFL-catalyzed reaction can
be represented as in Scheme 1.4, competitive inhibition of the COA-pyruvate reaction by
formate suggested that the reaction had to involve a second step in the mechanism.
However, when the experiment was repeated by varying pyruvate instead, inhibition was
found to be non-competitive, suggesting that formate and pyruvate do not compete for a
common reactant.15 These Observations led these workers to propose the “ping-pong”
mechanism depicted in Scheme 1.5.15 According to this proposed mechanism, the acetyl
group from pyruvate is transiently transferred to the active PFL to form the acetylated
active PFL intermediate. Since reaction in the first step (a) is reversible, addition of
formate shifts this reaction to the left, thereby decreasing the amount of PFL-acetyl

available to react with COA to yield

 

 

o PFL o
\\ //° 3 // CR
/C-—C + COA COA—C\ + C—H
'0 CH3 CH3 0’

Scheme 1.4.The overall reaction catalyzed by PF L

25

active PFL + pyruvate <-> PFL-acetyl + formate (a)

PFL-acetyl + COA <—> active PF L + acetyl-COA (b)

 

Overall: pyruvate + CoA (—> acetyl-COA + formate

 

Scheme 1.5. The Ping-Pong Mechanism proposed by Knappe and co-workers

aCetyl-COA. Consistent with the observed kinetics, COA and formate do not interact with
each other directly, nor do they have a common reactant.

At this point the available evidence to prove that the intermediate species was
indeed acetylated PF L was largely circumstantial. Direct and definitive proof was
provided by radioactive labeling experiments in which pyruvate was labeled at C-2 with
14C and then reacted with active PF L in the absence of COA.15 If the mechanism proposed
in Scheme 1.5 was correct, then 1) the acetylated PFL would be produced and would be
trapped Since the co-substrate (COA) was not available to participate in the next step
leading to a turnover, and 2) the radioactive label attached to active PFL would be
detectable. The results showed that indeed radioactivity was incorporated into the protein
only when the radiolabel was at C-2 of pyruvate but not at C-1. .In addition, only
activated PF L was able to incorporate the radioactivity at C-2 of pyruvate, whereas
inactive PFL and all the other enzymes involved in the activation process were not.
These observations provided positive proof that during homolysis of pyruvate, mediated
by PFL, the acetyl group of pyruvate is transiently attached to PFL before being
transferred to COA to form the acetyl-COA.15

Though considerable progress has been made toward understanding of the

mechanism of the PF L reaction, so far there is no consensus on one mechanism for the
26

processing of pyruvate by PFL. Based on the sum total of the findings of various
workers as discussed above, a number of mechanisms have been proposed, all of which
involve the participation of the catalytic triad comprising G734, C418 and C419. 1523’”
Based on previous work by several workers and their own extensive studies on PF L,

Kozarich and co-workers proposed the mechanism shown in Scheme 1.6.23

27

 

o
H
O
HN H HN H 020 CH3
H H % SH
i 1’ l l
4190 C418 C419 C413
0
Cows/1km,3
COA-SH i
HN 0
III
HN O H
VI HN H .
HN . H i 020%
O\ S CH3 SH
SH \\ I
/ I
419C C418
4190 T
C418
HCOz-d é VI
HN o
. -
HN’JflJ-Hr O F CC)2
\ 0H3 HN H 0\
Hco
’ 2 /S 3 CH3
HN S H \ 3 SH
H C418 I I
4190 ‘ 0419 C418

Scheme 1.6. The mechanism of PFL action as proposed by Kozarich and
co-workers.23

28

In this mechanism, it is proposed that the radical at G734 on activated PFL is
transferred to C419. The thiyl radical at C419 then attacks the C-2 carbonyl group of
pyruvate, whereupon a tetrahedral intermediate is formed in which the radical resides on
the C-2 oxygen of the pyruvate moiety. This intermediate then collapses to a more stable
thioester with concomitant elimination of a formate radical anion. The radical anion then
abstracts a hydrogen atom from G734 and is then released as formate. Simultaneously,
the acetyl group at C419 is transferred to C418. In the last step the acetyl group is
postulated to transfer from C418 to COA by a transesterification process, thereby
completing the catalytic cycle.

It is intriguing to note that this proposed mechanism involves an acetyl group
transfer from C419 to C418 immediately prior to transesterification of COA. The above
mechanism was proposed based on data from several experiments including studies by
Knappe’s groups in which they demonstrated that the glycyl radical EPR signal (centered
at g = 2.0037) is a doublet with a Splitting of 15 G resulting from hyperfine coupling with
the (II-hydrogen of G734.2°25 The studies of the Knappe laboratory had also shown that
the a-hydrogen of G734 rapidly exchanges with solvent, as demonstrated by the collapse
of the doublet EPR spectrum into a singlet when active PF L is incubated in D20, leading
them to suggest that the rapid exchangeability of the (II-hydrogen with solvent was due to
an unusually low K3.25 However, Mulliez et al. later demonstrated that though structurally
very similar to the PF L glycyl radical, the glycyl radical of RNR-HI of E. coli does not
exchange with solvent. 2° Because the active site of PFL differs from that of RNR-IH in
having the two residues C418 and C419, Kozarich and coworkers used Site-directed

mutagenesis to investigate the role that these residues play in the H-exchange process. 23

29

Their results showed that whereas the exchange reaction in the C4188 mutant is identical
to that of the wild type enzyme, such exchange is totally abolished in the C4198 mutant.
Contrary to the conclusions of Wagner and co-workers, these findings proved that
hydrogen exchange is not a spontaneous process inherent in the properties of the G734 a-
hydrogen but that it requires the participation of C419. 2°

Further implication of C419 in the exchange mechanism arose from investigation
of the (II-hydrogen exchange in acetylated PFL.” Knappe and co-workers had earlier
Shown that in the absence of COA and formate, stable acetylated PF L could be isolated
with the G734 radical intact.” Thus, the reaction can be stopped after the first half
reaction when one of the reactants needed for the second reaction (COA) is not available.
The exclusion of formate minimizes the Shifi of equilibrium back to the reactants.

Scheme 1.7 represents the reaction involved.

G734 ’ (3734 {If
\H 0 CH3 \H
+ \\c—c/ CH HCO —
C419 H -0/ \\ == C419 H 3 + 2
C418 SH 0413 / \0
Active PFL Acetylated PFL radical

Scheme 1.7. The formation of the acety-PF L radical intermediate by reaction of
activated PF L with pyruvate in the absence of COA and formate. Only the
reactants and products are shown.

Following upon the work of Unkrig et a1.25 , Kozarich‘s group acetylated active PFL
by reacting it with pyruvate in the absence of COA and formate.23 Knappe and co-workers
had previously reported the equilibrium constant for PFL acetylation to be 50, and based

on this equilibrium constant the concentration of acetylated PF L was calculated to be

30

10,000-fold that of free PFL.18 In all practicality, therefore, active PF L was completely
converted to the acetylated PFL radical intermediate. The EPR spectrum of the acetylated
PFL was indistinguishable from that of activated wild type PFL, Showing the doublet
centered at g = 2.0037.23 Upon incubation in D20, however, the doublet EPR signal
collapsed to a singlet identical to that observed in the deuterated C4188 mutant and wild
type PF L, indicating that acetylated active PFL is capable of III-hydrogen exchange with
solvent and therefore, on the basis of the mutagenesis experiments, provided evidence
that C419 is free to facilitate the exchange.23 These findings suggested very strongly that
C418 is favored overwhelmingly over C419 as the site of acetylation during pyruvate
turnover by PFL. Based on the foregoing discussion, two conclusions followed.

1. The (It-hydrogen of the glycyl radical in activated PFL undergoes rapid exchange
with solvent by a mechanism involving active and necessary participation of
C419.

2. Acetylation of active PFL during the homolytic cleavage of pyruvate in the first
of the two-step “ping pong” mechanism of pyruvate catabolism in E. coli takes
place with overwhelming preference for C418 over C419.

The results Obtained from PF L inactivation by hypophosphate, however,
introduced an additional complication to the mechanism by which reaction at the active
site of PFL proceeds during pyruvate turnover, particularly the role that each of the
cysteines plays in the transfer of the acetyl group to its position on PFL following
homolytic lysis of pyruvate.23 Knappe and co-workers and Brush had previously
demonstrated that hypophosphite (H2P02') is a mechanism-based inactivator of acetyl-
PFL?"Z7 Unkrig et al. then later demonstrated that when either pyruvate or acetyl-COA is

31

present, hypophosphite reacts with active PF L to produce a new and stable radical
characterized by a distinct and highly split EPR signal.25 These workers concluded that

the radical on this intermediate was localized at C418, based on the observations that:

1. Perdeuteriation of the acetyl group of pyruvate leads to perturbation of the EPR
spectrum of this radical, and
2. A peptide modified by attachment of a (hydroxyethyl) phosphonate moiety at

C418 was isolated.

An additional perspective was provided by the work of Ulissi-Demario et a1. 28
who discovered that acetylphosphinate also inactivates PF L and generates a radical
Species with identical features to that generated by the reaction of hypophosphite with
acetyl-PFL. With the benefit of this knowledge, Kozarich and co-workers carried out
experiments aimed at establishing which of C418 and C419 is responsible for the transfer
of the radical to the inactivator, as this might shed light on the extent to which each is
involved in the acetyl group transfer during pyruvate turnover.23 To address this question,
acetylphosphinate was added to the glycyl radical in the PFL mutants C4188 and C4198.
The experiment was done under anaerobic conditions to avoid destruction of the radical
by oxygen. For the C4198 mutant, no radical intermediate was observed. Instead, the
glycyl radical was found to persist unchanged for up to 15 minutes in the presence of
acetylphosphinate. The C4188 mutant, however, reacted in less than 30 seconds to
generate a radical intermediate stable enough to be detected by EPR.23 It was observed
that the splitting pattern and splitting constants for the C4188 radical intermediate were
slightly different from those of the radical formed with wild type PF L. These differences

32

were interpreted to be a result of the change from a sulfur-phosphorus bond in the wild
type to an oxygen-phosphorus bond in C4188.23 Theoretical calculations are also
consistent with these findings.29

The foregoing findings led to the proposal by Ulissi-Demario et a1., and later
corroborated by Knappe and co-workers, that inactivation of acetyl-PFL by

hypophosphite and inactivation of PF L by acetylphosphinate mechanistically converge to

 

 

 

 

a common radical intermediate as summarized by the Scheme 1.8. ”’22
r—fi
MCH SH SH
HCO '
2 6734 0419 C418 H2P02'
0 O
\
fic—c’f
H30 0. OH
O\\ )—CH3
/P\ -
WWW('3H SH SH WCHz SH 8 0

II

6734 C419 C418 6734 C419 C418

on

//O
O
I
u

:1:

\

/
O

OH

0 )‘CH

MCH SH OH “p\ 3
¥ / O.

l l l T 7 MCstH
c734 0419 0413

G734 C419 C418

 

Scheme 1.8. Reactions of hypophosphite with acetyl-PFL and acetylphosphinate with
active PFL lead to the same product.

At this point, data obtained from site-directed mutagenesis experiments had
led to the following conclusions;
a) C418 is exclusively acetylated when activated PFL is reacted with

pyruvate in the absence of COA and formate;
33

b) C419, but not C418, is absolutely required for (Ii-hydrogen exchange with
solvent;

0) trapping of the G734 radical by acetylphosphinate requires C419; and
(1) following acetylphosphination of both wild type and the C4188 mutant,
the (hydroxyethyl)phosphonate moiety, covalently attaches to residue 418.
The Observation that C418 is necessary for thioester transfer during the CoA-acetyl-COA
exchange reaction, proved that it is necessary in the initial homolytic cleavage of the C1-
C2 bond in pyruvate.23 The reaction of PFL with acetylphosphinate, on the other hand,
suggested that C418 is not necessary for the transfer of the radical to acetylphosphinate.23
Instead, this transfer was found to require C419.23 This apparently divergent findings
introduced an unwelcome complication to the mechanistic investigations of the PFL-
pyruvate reaction. To reconcile these findings, Kozarich and coworkers proposed that
C419 is the “direct mediator of the glycyl radical action” while C418 mediates the
transfer of the acetyl moiety to COA after homolysis of the pyruvate C1-C2 bond.23 It was
on this basis that they modified the mechanism of Brush et al. by proposing a step
involving transesterification of the acetyl group between both residues.”27 Thus, it is
proposed that upon cleavage of the pyruvate C1-C2 bond, the acetyl group is first
transferred to C419 and subsequently to C418 from where it is transferred to COA in the
second step of the reaction.23
Knappe and co-workers proposed an alternative mechanism based on the crystal
structure of wild type PF L, both in its free state and as a complex with oxamate.3° In the
same study, they solved the crystal structure of the C4188, C4198 double mutant at 2.6A
resolution and found that oxamate and pyruvate bind specifically and uniquely at the

34

active site without affecting the overall structure of the protein. An interesting, and
perhaps unique, finding was that pyruvate and oxamate are not simply allosteric effectors,
but that they bind to the active site of PF L.3 0 Since PFL activation is known to require the
presence of either one of the two”, it can be inferred that the positioning of either one of
these two at the active site facilitates H-abstraction from G734. The nature of this
facilitation is not yet understood and is therefore a subject for on-going investigation.

The mechanism that Knappe and coworkers proposed was based primarily on
structural information obtained from crystal crystallographic studies, with the PF L-
oxamate co-crystal providing significant insight into the special relation between the PFL
active site residues and the substrate pyruvate.30 The structure revealed both electrostatic
and hydrophobic interactions between PF L and oxamate, in which the carboxylate
oxygen atoms form a salt bridge with the guanidino group of Arg 435 as illustrated in
Scheme 1.9. The delocalized negative charge on the carboxylate oxygen atoms leads to

the formation of a salt bridge with the delocalized positive charge on the N8 of Arg.

,5 0
RC// NHZ ------ O

H—c—CH —NH-C G) _ /
I ( 3)3 \ 9/0 C\
H3C NH2 ------ O NH2

Scheme 1.9. The salt bridge formed between oxamate and Arg 435 in the active
site of PFL

The flat moiety is also sandwiched between two aromatic and hydrophobic sided chains,
Phe 432 and Trp 333, the catalytically essential C418 being Situated with its 8 atom at a

distance of 3.3A from the C1 and C2 of oxamate, and inserted between the planes of Trp
35

333 and oxamate. Of particular interest was the Observation that the O-C-S angle at C2 is
~103°, which is very close to the optimal angle of 109° optimally required for
nucleophilic attack at an Sp2 carbon of the carbonyl group.

Using the crystallographic distances observed in the PF L—oxamate complex,
Knappe and coworkers deduced that if pyruvate, which is structurally homologous to
oxamate (CH3 in pyruvate instead of NH2 in oxamate), were to be substituted for oxamate
in this structure, its methyl group would be within 3.4A of the CH3 group of Ala 272.
Ala 272, in turn, is hydrogen-bonded to the OH group of Tyr 323. This inferred binding
mode for pyruvate at the active site of PFL was corroborated by the Situation in a-
oxobutyrate formate-lyase (Tch). 4° This enzyme is homologous and bears 82%
sequence identity to PFL.4° It carries out an analogous reaction to PFL but acts on a-

oxobutyrate (Scheme 1.10) as substrate.

0 O
\\C <0 \\C_ //
_ . /
'0/ CH3 0 CHch3
pyruvate alpha-oxobutyrate

Scheme 1.10. Chemical structures of pyruvate and alpha-oxobutyrate

At analogous positions to Ala 272 and Tyr 323 in PFL, Tch has glycine and
phenylalanine. These are both smaller residue residues than the corresponding alanine
and tyrosine in PF L (Scheme 1.11) and therefore affording more room for the larger

substrate a-oxobutyrate.

36

/
HO—C/
\C/H
/
O
//
CH3 OH
OH
alanine tyrosine
A
O
//
HO-C
\C/H
/ \
O H C
2 _ §
\
H OH
glycine phenylalanine
8

Scheme 1.11. ComparisOn of residues found at the active sites of Tch (A)
and PFL (B)

Relying, in part, upon previous biochemical studies, which showed that C419 is
situated within 3.7A of the G734 a-C and C418 and the data obtained in this work,
Knappe’s group proposed the catalytic mechanism shown in Scheme 1.11, involving the
cooperation of the catalytic triad during the acetylation of PFL in the first step of the

“ping pong” mechanism. 3°

37

 

 

_l
o E
I f g HN H
a O
HN H
3 )L—CHS
3 5 SH
.9
HS H a I I
| .3 C419 C418
C419 C418 g
g .3
HN

1 EO
acetyl-PFL resting state
;0
<
o

O
H HN/\H \\\
HN H o i . / 0H,
2 A s T
0 c .s COA'SH
T“ T CA CH” (L419 C418
C419 C418

Scheme 1.12. Proposed mechanism for the transfer Of the radical and acetyl
group within the amino acid triad at the active site of PFL. Adapted from ref. 30.

The conversion of 1 to 11 and III constitutes the initiation step. The formation of the
acetyl-PFL intermediate involves the reaction of 111 with either pyruvate or acetyl-COA,
with concomitant production of formate or COA respectively. The processing of pyruvate
by PFL is proposed to begin with PFL 111 above and proceed as illustrated by Scheme

1.13.

38

Step 1: “ping”

 

 

 

III VII VIII IX
CH3 CH3 O\
0\0— 0\ C‘W “30 o\
// \0 // |\ O// \ /O C—H H3C\
0 O C/ // éo
K. s __ H / o 0
SH S SH J S S ——- I /
I 0 18 | S 3
C419 0419 0 18 I Ci
0 19 c 18 C419 18
Step 2: “pong”
IX x XI
COA—f? H3C\ 0 /CH3
H 05 COA—S ”30 \
/ 1° \0‘5’0 COA—S/C\0
T S v—_______—~_ V V ‘
SH 0
0419 0418 | J ISH 3
C419 C 18 C419 C418

Scheme 1.13. The mechanism proposed by Knappe and co-workers. 3°

The proposals put forward by Knappe and co-workers are fundamentally similar
to those of Kozarich, both starting with the transfer of the radical from G734 to C419.3°
However, whereas Kozarich and co-workers propose that the thiyl radical at C419
directly attacks the C2 carbon of pyruvate to form the tetrahedral hemiketal
intermediate”, Knappe’s group suggested that the thiyl radical is first transferred to C418
and that the tetrahedral intermediate forms at this residue.30 Since it is now known that

the acetyl group in acetyl-PFL is located at C418, the mechanism of Kozarich and others
39

includes the transfer of the acetyl group from C419 to C418 immediately prior to the
acetyl group transfer to COA to form the thioester linkage in acetyl-COA. Kozarich’s
group used as part of their rationale for the transesterification step between C418 and
C419, the fact that mutagenesis experiments showed that radical trapping by
acetylphosphinate takes place only when C419 is in place, whereas C418 is not required
for this process.23 However, C418 was shown to be essential for the transfer of the acetyl
group to COA.3O

In their latest proposal, on the other hand, Knappe and co—workers preclude the
involvement of C419 in acetyl group transfer to C418 and instead suggest that acetylation
takes place directly at C418. Notably, in Kozarich’s proposal, the G734 radical is
recovered after every turnover as would be expected in a catalytic cycle, while Knappe’s
proposal involves multiple turnovers “without regeneration of the glycyl radical.”3°
Clearly, the mechanism by which PFL processes pyruvate remains a controversial and

on-going area of investigation.

40

1.8. THEORETICAL STUDIES

Owing to the lack of consensus regarding the mechanisms by which pyruvate is
processed by PFL, a number of workers have carried out theoretical studies aimed at
testing the feasibility of some of the steps proposed in the various mechanisms.”33 The
stability of the glycyl radical has been proposed to be due to the captodative effect, a
chemical phenomenon whereby an electron-withdrawing group and electron-donating
group cooperate to enhance the resonance stability of a radical located between them.”37
The glycyl radical in PFL is located between an electron donating amino group and an
electron withdrawing amide carbonyl group and was therefore hypothesized to benefit
from stabilization by the captodative effect.52

The other phenomenon that affects the stability of amino acid radicals is the non-
bonding interaction between the side chain and the carbonyl group. It is known that this
interaction has the effect of destabilizing a radical at the alpha carbon. Since the glycyl
radical is located on a planar (sz-hybridized) carbon, this interaction is absent, thereby
rendering the radical more stable than other corresponding Cor-amino acid radicals. This
phenomenon has been studied extensively by a several workers.”44 The stability of the
glycyl radical was confirmed by calculating the bond dissociation energies (BDE) of the
Cor-H bond in a series of models wherein it was discovered that whereas the substitution
of —NH2 and —CHO in methane lower the Cor-H bond energy by 13.1 kcal/mol and 10.7
kcal/mol, respectively; the combined effect of the two groups in NH2-CH2-CH0
decreases the BDE by 33.5 kcal/mol.39 When the chain length was increased with
additional peptide bonds to both the —CHO and the —NH2 groups, it was found that there

was a smaller decrease in BDE. This was consistent with the expected lowering of the

41

electron—donating and withdrawing effects of the —NH2 and the —CH0 groups
respectively, and suggested that the captodative effect on glycyl radical situated along the
backbone of a polypeptide will be lower than was observed in NH2-CH2-CH0.
Calculations have also shown that the Cor-H and 8-H BDES are within 2 kcal/mol of each
other, thereby affording reversibility in the transfer of the radical between G734 and
either one of C418 and C419.

The mechanism proposed by Kozarich and co-workers was further subjected to
analysis by DFT calculations, which proved that all of the steps subsequent to the transfer
of the radical from G734 to the C419 are energetically plausible.4° Although the
regeneration of the glycyl radical by abstraction of a hydrogen atom by the formyl radical
was also found to be energetically plausible, Himo and Eriksson proposed that the
hydrogen is abstracted fiom C419.4° This subsequently facilitates the transfer of the
acetyl group to COA. Taking into consideration the proposed mechanisms of both the
groups of Kozarich and the Knappe, Himo and Siegbahn used the results of the
calculations to propose the presently accepted and slightly different mechanism depicted

in Scheme 1.1.4.45

42

G734
0419 SH
SH
C418
0H2 "
so 3
C
COA-S/ \CH3

CH2

0
n”
o
‘9
a

O
‘0

a
a
-C

C
CH2 HSC/ \002' CH2
H
S >3 . SH $05
C 00
S' ’
" S \CH3
CH2 CH2
HS-CoA SH .002-
SC
.. 9
I O ' S’C\
S \CH3 HC 2 CH3

Scheme 1.14 Mechanism Proposed by Himo and Siegahn. Adapted from ref. 45.

However, Ramos and co-workers recently published results from a

theoretical study in which they modeled pyruvate as an anionic species, since,

given its pKa of 2.5, it is expected to be deprotonated at physiological pH. Based

on the interpretation of their DFT data, these workers conclude that thiyl addition

to pyruvate occurs via a quasi-planar structure and not a tetrahedral transition

state as hitherto proposed by several workers. These workers also contend that

rather than the C-C bond cleavage, it is the H transfer between G734 and C419

that is the rate limiting step. They therefore propose yet another mechanism for

the processing of pyruvate by PFL. Clearly, much work remains to be done

before consensus on the PFL mechanism is reached.

43

1.9. THE ROLE OF ALCOHOL DEHYDROGENASE (AdhE)

1.9.1. Reported role of AdhE as a PFL deactivase

The discovery of pyruvate formate-lyase activating activating enzyme (PF L-AE)
14‘ 15 led researchers working on the PFL-AB system to think of the possibility of a second
activity that would play an opposite role to the activating enzyme, namely to
enzymatically converting PFL from its active radical form to its inactive form. From the
standpoint of metabolic economy the existence of such an activity seemed rational since
it would ensure that PF L was deactivated conservatively, thus leaving it intact and well
poised for reactivation when conditions required its deployment.

In 1992 Kessler and Knappe53 reported that they had discovered the PFL
deactivase activity to reside in the multifimctional alcohol dehydrogenase (AdhE). This
report therefore provided the activity that achieved the reverse reaction to PF L-AE,
thereby completing the interconversion cycle between between active and inactive states
in response to the growth conditions of Escherichia coli. It is with respect to this reported
activity that this study was conducted, with the aim of illucidating the nature and

mechanistic role of the purported cofactors, namely Fe”, co-enzyme A and NAD+'

1.9.2. Other physiological characteristics of AdhE

In addition to its latest reported role as a PFL deactivase, AdhE is known to play a
major role in the anaerobic metabolism of Escherichia coli. It catalyzes the reduction of
acetyl-COA by NADH to ethanol via a two-step mechanism involving acetaldehyde as an

intermediate.54 In this physiological role, AdhE is very similar to the alcohol

44

dehydrogenase from the obligate anaerobe and pathogenic protozoan Entamoeba
histolytica to which it also bears striking structural resemblance. 55 Both enzymes are also
NAD+ and Fe2+-dependent form helical assemblies of protomers called spirozomes.
Clark and co-workers56 exploited these structural and functional similarities and
successfully cloned the E. histolytica alcohol dehydrogenase (AhADHZ) into E. coli and
proceeded to use the E. coli/EhADHZ clones to screen various compounds for their
ability to inhibit anaerobic growth of the clone by interfering with EhADHZ. This
provided a rapid assay for identifying possible phamaceauticals against E. histolytica,

with the aim of finding efficatious drugs for amebic dysentery.

1.9.3. Regulation of AdhE

In addition to being anaerobically induced Clark discovered what appeared to be a
positive correlation between induction of the adhE gene and the levels of NADH in the
growth medium.57 In order to determine the role of this co-enzyme in the induction of the
adhE gene, Clark and co-workers carried out a systematic study in which they grew E.
coli on various carbon sources whose metabolism is known to produce various relative
amounts of NADH. They also used treatments that are known to change the NAD+ :
NADH ratio or alter the NAD+ and NADH pool. Some of the treatments included the use
of chemical inhibitors and mutations to block the electron transport chain and thus
artificially induce the adhE gene under aerobic conditions. These workers also
introduced mutational blocks into the biosynthetic pathway for nicotinic acid in order to
limit the synthesis of NAD“, while in some experiments they introduced exogenous
NAD+ and nicotinic acid. The results obtained from these studies led to the conclusion

45

that “the enzymatic activity of of the AdhE protein modulates the level of NADH under
anaerobic conditions, and thus indirectly regulating its own expression”. In another
study”, it was also found that the expression of the adhE gene is transcriptionally
regulated by Cra and post-transcriptionally by RNase G. Thus, the physiological roles of
AdhE are complex and very interesting, while its regulation under different conditions
remains intriguing. In this study we hope to gain some insight into the role that this
enzyme plays in the deactivation of the anaerobically indispensable pyruvate formate-

lyase of Escherichia coli.

46

REFERENCES

1. Knappe, J and Sawers, G. 1990, FEMS Microbiol. Rev. 75, 383-398
2. Kalnitsky, G.; Werkman, CH. 1943, Arc. Biochem. 2, 113-124.

3. Lipmann, F.; Tuttle, LC. 1945, J. Biol. Chem. 158,505-519

4. Chantrenne, H.; Lipmann, F. 1950, J. Biol. Chem. 187, 757

5. Stadtman, ER. 1951, Biol. Chem. 191 365

6. GD. Novelli, et al. 1954, Bacterial. Proc. p.97;

7. Novelli, GD. 1955, Biophys. Acta 18, 594-596.

8. McCormick, N. et. al. 1962, J. Bacterial. 83: 899-906.

9. Broderick, J .B.; Dederstardt, R.E.; Fernandez, D.C.; Wojtuszewski, K.; Henshaw,
T.F.; Johnson, M.K.J. J. Am. Chem. Soc. 1997, 119, 7396-7397.

10.Knappe, J .; Sawers, G. 1990, FEMS Microbial. Rev. 75, 383-398

11.Knappe, J .; Bohnet, E.; Brummer, W. 1965, Biachim. Biophys. Acta. 107, 603-
605.

12. Wood, N.P. Methods in Enzymology, 1966, 69, 718.
13. Chase, T.Jr.; Rabinowitz, J. J.Bacterial. 1968, 1065-1078.

14.’Knappe, J .; Schacht, J .; Mockel, W.; Hopner, Th.; Vetter, H. Jr.; Edenharder, R.
Eur. .1. Biochem. 1969, 11, 316-327.

15.Knappe, J .; Blaschkowski, H.P.; Schmitt, T. Eur. J. Biochem. 1974, 50, 253-263.

16.Knappe, J .; Blaschkwoski, H-P.; Edenharder, R. Metabolic Interconversion of
' Enzymes, Springer, 1972, 319-329.

17. Leppanen, V. M.; Merckel, M. C.; Ollis, D. L.; Wong, K. K.; Kozarich, J. W.;
Goldman, A. 1999. Structure, 7, 733-744.

18.Knappe, J .; Wagner, A.F.V. 1995, Methods Enzymal 258, 343-362.

47

19. Conradt, H.; Hohmann-Berger, M.; Hohmann, H.P.; Blaschkowski, H.P.;
Knappe, J. 1984, Arch Biochem. Biophys. 228, 133-1 42.

20. Wagner, A.F.V., Frey, M., Neugebauer, F.A., Schafer, W., and Knappe, J. 1992,
Proc. Natl. Acad. Sci.U.S.A 89, 996-1000.

21. Knappe, J .; Neugebauer, F .A.; Blachskowski, H.P.; Ganzler, M. 1984, Proc. Natl.
Acad. Sci. USA, 81, 1332-1335.

22. Frey, M.; Rothe, M.; Wagner, A.F.V.; Knappe, J. 1994, J. Biol.Chem. 269, 17,
12432-12437.

23. Parast, C.V.; Wong, K.K.; Lewisch, S.A.; Kozarich, J .W.; Peisach, J .; Magliozzo,
RS. 1995, Biochem. 34 (8), 2394-2399.

24.Gautney, L.; Miyagawa, I. 1975, Radiat. Res. 62, 12-17.
25. Unkrig, V., Neugebauer, F.A., Knappe, J .(1989) Eur. J. Biochem. 184, 723-728.

26. Mulliez, E.; Fontecave, M.; Gaillaid, J .; Riechard, P. 1993,.1. Biol.Chem., 268,
2296-2299.

27. Brush, E.J., Lipsett, K.A., Kozarich, J .W. (1988) Biochemistry, 27, 2217-2220.

28. Ulissi-Demario, L.; Brush, E. J .; Kozarich, J .W. 1991, J. Am. Chem. Soc. 113,
4341-4342.

29. Reed, A.E.; Schleyer, P. 1990, J. Am. Chem. Soc. 112, 1434.

30. Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, J .; Schultz, 8.; Wagner, A.F.V.
1999, Nature: Struct. Biol. 6 (10), 969-975.

31. Knappe, J .; Albert, S.; Frey, M.; Wagner, A.F.V. 1993, Biochem. Soc. Trans. 21,
731-734.

32. Himo, F.; Eriksson, LA. .1. Am. Chem. Soc. 1998, 120, 11449.
33. Gauld, J .W.; Eriksson, L. A. 2000, J. Am. Chem. Soc.,122, 2035.
34. Wong, K. K.; Kozarich, J. W. In Metal Ions in Biological Systems,

Metalloenzymes Involving Amino Acid-Residue and Related Radicals, Eds. Si gel,
H.; Sigel A.; Marcel Dekker: New York, 1994; Vol. 30, p 279.

48

35. Vieche, H. G.; Merenyi, R.; Stella, L.; Zanousek, Z. 1979, Angew. Chem, Int.
Ed. Eng]. 18, 917.

36. Vieche, H. G.; Zanousek, Z.; Merenyi, R.; Stella, L. 1985,Acc. Chem. Res., 18,
148.

37. Stella, L.; Harvey, J. N. In Radicals in Organic Synthesis, Volume I: Basic
Principles, pp 360-380, Renaud, P., Sibi, M. P., Eds.; Wiley-VCH: New York,
2001.

38. Armstrong, D. A.; Yu, D.; Rauk, A. 1996, Can. J. Chem. 74, 1192.

39. Rauk, A.; Yu, D.; Taylor, J .; Shustov, G. V.; Block, D. A.; Armstrong, D. A.
1999,Biachem. 38, 9089.

40. Himo, F. 2000, Chem. Phys. Lett. 328, 270.
41. Himo, F.; Eriksson, L. A. 1998, J. Chem. Soc., Perkin Trans. 2, 2, 305.

42.Barone, V.; Adamo, C.; Grand, A.; Brunel, Y.; Subra, R. 1995, J. Am. Chem. Soc.
11 7, 1083.

43. Barone, V.; Adamo, C.; Grand, A.; Subra, R. 1995, Chem. Phys. Lett. 242, 351.

44. Barone, V.; Adamo, C.; Grand, A.; Jolibois, F.; Brunel, Y.; Subra, R. 1995, J.
Am. Chem. Soc. 117, 12618.

45. Himo, F.; Siegbahn, P.E.M. 2003, Chem. Rev. 103, 2421-2456.

46.HeBlinger, C.; Fairhurst, S.A.; Sawers, G. 1998, Mol. Microbiol. 27, 477-492.

47. Reichard, P., and Ehrenberg, A. 1983, Science 221, 514-519.

48. Sun, X.; Ollagnier, S.; Schmidt,P. P.; Atta, M.; Mulliez, E.; Lepape, L.; Eliasson,
R.; Graslund, A.; Fontecave, M.; Reichard, P.; Sjoberg, B. M. 1996, J. Biol.
Chem. 271, 6827-6831.

49. Logan, D. T.; Andersson, J .; Sjoberg, B.-M.; Nordlund, P. 1999, Science 283,
1499-1504.

50. Becker, A.; Kabsch, W. 2002, J. Biol. Chem. 277, 42, 40036-40042.

51. Sofia, H.J.; Chen, G.; Hetzler, B.G.; Reyes-Spindola, J. F.; Miller, N. 2001,
Nucleic Acids Research, 29, 5, 1097-1 106.

49

52. Sustman, R., Korth, H-G. 1990. Adv. Phys. Org. Chem. 26:131-78.

53. Kessler, D and Knappe, J. 1992. J. Biol. Chem. 267, 25, 18073-18079.

54.'Goodlove, RE. 1989. Gene. 85, 209-214.

55. Espinosa, A.; Yan, L.; Zhang, Z.; Foster. L.; Clark, D.; Li, B; Stanley, S.L., Jr.
2001. J. Biol. Chem. 276, 20136-20143.

56. Yong, T.; Li, B; Clark, D.P.; Stanley, S.L., Jr.

57. Clark, DP. 1989. FEMS Bicrobiol. Rev. 63, 223-234.

58. Leonardo, M.R.; Dailly, Y.; Clark, DP. .1. Bacteriol. 1996. 178, 6013-6018.

50

CHAPTER II

PFL: PURIFICATION, ACTIVATION AND STABILITY

51

ABSTRACT

PFL was overexpressed in BL21(DE3)pLysS cells that were previously
transformed with the recombinant vector pKK-pfl. The recombinant vector was
engineered by ligating the pfl gene (obtained fiom the polymerase chain reaction) onto
the commercial vector pKK. Purified pyruvate formate-lyase (PFL) was then purified
and activated according to the standard protocol in which all the activation components,
including co-substrate S-adenosylmethionine (SAM), the photoreductant deazariboflavin
and pyruvate formate-lyase activating enzyme (PFL-AB) were mixed and then subjected
to an intense halogen lamp for 45 minutes. PFL thus activated was used to carry out
deactivation studies using AdhE and various small molecules (Chapters HI and IV). In
the process of these deactivation studies, which were carried out under ambient light, we
observed that the control reactions containing only PFL and buffer, or those containing
components that proved unreactive to the glycyl radical of PFL, always showed a steady
increase of PFL activity over time. This led us to the hypothesis that PFL can be
activated under ambient light, if all of the essential components of the activation mix are
available, thereby circumventing the use of a high intensity halogen lamp. This
hypothesis was confirmed. By activating PFL under ambient light, we were able to
achieve PF L activity that was comparable to the highest levels we have obtained using
the illumination method. The maximum activity was achieved within a period of two
hours. This discovery enabled us to carry out all subsequent PFL activations at ambient
light and temperature, thus simplifying the procedure significantly. In addition, the
glycyl radical in activated PF L proved to be extremely stable, being detectable for well

over 24 hours at room temperature.

52

11.]. INTRODUCTION

Pyruvate formate-lyase (PF L), the enzyme which, in the presence of co-enzyme A
(CoA) and under anaerobic conditions, converts pyruvate to acetyl-COA and formate, is
translated in an inactive form inside the cell. When conditions become anaerobic, it is
converted to its active form by pyruvate formate-lyase activating enzyme (PFL-AB),
which generates a radical on a glycine residue located at position 734 of the PFL
polypeptide chain.1 In vitro, this activation involves the participation of several cofactors
and co-substrates and requires strictly anaerobic conditions.2 The activation reaction is

summarized as shown by Scheme 11.].

.S O
_.= \CH 2
(3625

OH
OH

 

ll
O—I

PFL-AE, reduced flavodoxin % \fg’

’NH,

 

Scheme 11.1. A structural depiction of the activation of PFL showing the
essential participants in the activation process. Following reductive cleavage of
SAM ,into methionine and a 5'-deoxyadenosyl radical, the transient 5'-
deoxyadenosyl radical generates a glycyl radical on the PFL backbone by
abstracting the pro-S hydrogen atom at Gly 734. PFL thus activated proceeds to
catalyze homolysis of the C2-C3 bond of pyruvate followed by transacetylation of
CoA and formation of formate.

53

The scheme shows only glycine 734 before and after hydrogen atom abstraction,
representing the inactive and active states of the enzyme, respectively. The role of the
photoreductant deazariboflavin is to reduce an iron-sulfur cluster on the PF L-AE to the
catalytically active [4F e-4S]+ form. The reduced activating enzyme facilitates the
abstraction of the hydrogen atom at glycine 734 via a process that requires the
participation of S-adenosylmethionine (SAM) as described in Chapter I. In the process of
the PFL activation reaction, S-adenosylmethionine is reductively cleaved to methionine
and 5'-deoxyadenosine.3

Studies of the activation system have shown that the hydrogen atom abstracted
from glycine 734 of PFL ends up at the 5’-carbon of 5’-deoxyadenosine, suggesting that
the radical generated at glycine 734 is transferred to that position from the 5’-carbon of
5’-deoxyadenosine.4 That, in turn, led to the conclusion that the interaction between PFL-
AE and SAM results in the generation of a 5’-deoxyadenosyl radical. The nature of this
interaction is a subject of continuing investigation in our laboratory.

Activation of PFL in vitra is usually performed by mixing the activation
components under anaerobic conditions, followed by illumination with a high intensity
lamp.5 Maximum activity is generally achieved within 30 to 45 minutes. Because of the
large amount of heat generated by the lamp, it is necessary to keep the activation mix
cool by carrying out the process in an ice-water bath. This introduces technical
difficulties in sample handling and conceivably introduces a further deviation from the
natural activation conditions within the cell. We have now discovered that under
conditions of ambient light and room temperature (~23°C), PF L can be activated to levels

comparable to, and sometimes higher than, those previously achieved by illumination in

54

an ice-water bath as described above. We have also discovered that active PF L, as
determined by the enzymatic activity assay, is extremely stable under the same
conditions. These findings are significant because they afford simpler practical
manipulations both in the preparation for activation as well as after the activation of PF L.

Additionally, the activation conditions are a closer approximation of the situation in vivo.

55

11.2. MATERIALS AND METHODS
[1.2.]. Chemicals, bacterial cells and plasmids

The plasmid pKK-PFL used to transform BL21(DE3)pLysS was from a stock
solution obtained by amplification of pKK-PF L previously used in our laboratory.6
Plasmid pCAL-n-AE was also amplified from a stock solution previously engineered in
our laboratory6 using pCAL-n-EK expression vector obtained from StratagenTM and the
pMG-AE plasmid generously donated by Dr. John Kozarich (Merck). 5-Deazariboflavin
was synthesized by Megan Kibbey and Dr. W. Broderick following published
procedures.7'9
11. 2.2. Bacterial production of PFL

Escherichia coli (E. coli) BL21(DE3) pLysS was transformed with pKK-PFL
according to the supplier’s protocol. Single transformant colonies were inoculated into 50
mL LB media containing a total of 50 ug ampicillin per mL of LB liquid medium. The
cultures so prepared were allowed to incubate for approximately 16 hours to reach
maximum cell density. In order to check overexpression in each of the flasks, small
pellets were routinely prepared from each growth and denatured in sodium dodecyl
sulfate (SDS) denaturing buffer containing 3.5% 2-mercaptoethanol. Sodium dodecyl
sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was then carried out using a
12% Tris-HCl commercial gel from BioRadTM to check overexpression of the protein. A
sample that showed good overexpression (in BL21(DE3)PlysS, PF L overexpresses even
without induction) was used to inoculate 5 mL into each of 2.8 L Fembach flasks
containing 1 L of liquid LB medium impregnated with ampicillin to a final concentration

of 50 ug/mL. Alternatively 50 mL of the saturated growth was inoculated into a bench
56

top fermentor (New Brunswick) containing 10 L of LB medium containing 50 pg
ampicillin per mL of medium. When Fernbach flasks were used, shaking the flasks
vigorously at 250/min ensured good aeration. In the fermentor, the growing cells were
aerated by continuously bubbling a stream of filtered air into the medium while
simultaneously agitating vigorously with an electric stirrer. In either case, the cells were
grown at a temperature of 37°C. When the optical density of the cell suspension reached
~O.7, PFL overexpression was induced by adding a sterile-filtered solution of isopropyl-
B-D—thiogalactopyranoside (IPTG) to a final concentration of lmM. Although the cells
overexpressed even without induction, overexpression improved somewhat when H’TG
was added. The cells were incubated for ~2 hours after IPTG induction, and then were
harvested by centrifugation at 800 rpm using a Sorvall GS3 rotor. The supernatant was
discarded and the cell pellet was weighed and then either lysed immediately or flash-

frozen with liquid nitrogen and stored at -80°C to be used later.

57

11. 2.3. Bacterial synthesis of PFL-AE

PCAL-n-AE, engineered in our lab according to published procedures 6, was used
to transform E. coli BL21(DE3) pLysS. Cell suspensions were prepared by inoculating
half of a colony into each of a number of flasks containing 50 mL of autoclaved LB/Amp
media. When the optical density of the cell suspension reached approximately 0.5
enough sterile-filtered IPTG was added to give a final concentration of 1 mM. The cells
were grown for two hours post-induction and cell pellets were prepared for SDS-PAGE
analysis. When overexpressing colonies were identified, the remaining half of the colony
was used to grow a cell suspension that was then used to prepare glycerol stock cells.
Glycerol stocks were prepared by mixing, in 1.5mL micro-tubes, equal volumes of
autoclaved glycerol with a cell suspension grown to saturation overnight. These were
flash-frozen and stored at -80°C to be used whenever needed.

A typical growth of PFL-AE began with the plating of the glycerol stock cells on
LB-agar media containing ~50 ug ampicillin per mL of LB-agar and incubating overnight
at 37°C. To prepare an overnight cell suspension for a growth, a single colony was
inoculated in LB/Amp medium and incubated overnight. The overnight suspension was
used to inoculate 10 L of LB/Amp using a bench-top fermentor (New Brunswick)
equipped with a continuous airflow inlet and an agitator. 50 mL of the overnight culture
was inoculated into 10 L of LB/Amp and grown at 37°C. At an optical density (ODGOO) of

0.6, IPTG was added to a final concentration of lmM.

58

11.2. 4. Purification of PFL

Using a cell pellet from freshly harvested cells or one from a -80°C freezer,
enough lysis buffer was added to make a suspension of approximately 2 mL of lysis
buffer per gram of cells. The lysis buffer comprised 20 mM HEPES containing 1 mM
DTT and adjusted to pH 7.2, 1% Triton X-100, 5%w/v glycerol, 10 mM MgC12, 8 mg
lysozyme, 1 mM phenylmethylsulfonylsulfate (PMSF) and trace quantities (~0.1 mg) of
each of RNAse and DNAse. The cell suspension was thoroughly mixed and then
homogenized continuously with a magnetic stirrer at 4°C for 2 hours. The lysed cells
were centrifuged at 15, 000 rpm using a Sorvall SS34 rotor for 30 minutes at 4°C. In the
first step 50 mL of the crude lysate was injected onto AP-5 Superdex ion-exchange
column using a low ionic strength buffer (Buffer A) consisting of 20 mM HEPES - pH
7.2, 1 mM DTT. Purification was performed automatically using the BioLogicTM
program from BioRadTM. The program was written based an optimization carried out
under the same conditions used for the experiment.

Immediately following sample injection, the column was flushed with the low
ionic strength buffer (buffer A) for 60 minutes at a rate of 3 ml/min in order to wash out
the proteins that were not bound specifically by the column. Following the initial wash,
the column was eluted with a steadily increasing ionic strength starting from 100% buffer
A to 100% buffer B but with a total combined flow rate of 3 ml/min. Buffer B was of
identical composition to buffer A except that it also contained 500 mM NaCl. When the
eluent composition reached 100% buffer B, it was kept isocratic for 60 minutes before
dropping down to 100% buffer A. Under these conditions, PFL started to elute out of the

column at approximately 375mM NaCl and was collected in 10mL fractions. The

59

fractions were analyzed using SDS-PAGE in order to determine which ones contained
PFL and the relative amounts of it in each. Based on the SDS-PAGE results, fractions
containing 2 75% PFL were pooled together and, whenever necessary, concentrated
down to ~ 50 mL. The pooled fractions were then dialyzed for 2 hours in enough buffer
A to dilute the salt 40-fold (e.g. 50 mL dialyzed in 2L of buffer A). Afler 2 hours, the
protein was dialyzed in fresh buffer overnight. When dialysis was complete, the protein
was again concentrated down to less than 10 mL using an AmiconTM concentrator. Figure
III shows a typical Chromatogram obtained from the first purification using ion
exchange and an SDS-PAGE gel electrophoresis for fourteen fractions spanning the PFL

peak is shown in Figure 11.2.

Fractions

 

 

LUO/SLU 'AuAuonpuog

 

Absorbance, 280 nm

 

 

 

 

 

‘ v V v v v v v v v v '

00:00:00 02-5000 """" 55:00:00

HrMin:Sec

Figure 11.1 A Chromatogram of PFL purification by ion exchange. PFL is in
the peak eluting from approximately 3.5-4.5 hours.
60

In order to identify the fractions containing PFL and the relative concentration of PFL in
each, SDS-PAGE was performed on the fractions spanning the PFL peak. Figure [1.2.

shows a typical SDS-PAGE gel of fractions from the ion-exchange column.

Std 2 8 13 18 21 2530354045 50556053
200 kDa\
115 kDa\:
97 kDa-"'7 O
65 kDa ‘: ~
45 kDa/-

b

  

Figure 11.2. SDS-PAGE of PFL fractions obtained from the first run on the
Superdex ion exchange column. The fraction numbers are indicated at the top of
the gel.

The fractions with 3~90% purity, as judged by the relative intensity of the band at 85
kDa, were pooled together, and dialyzed into 20 mM HEPES (pH 7.2) containing 1 mM
DTT. The fractions of lesser purity were also pooled together and stored to be run
together with the next crude lysate to be purified.

Afier dialysis, the pooled fractions from the ion exchange column were
concentrated down to 5 5mL, dialyzed against 40mM HEPES (pH 7.2) containing lmM
DTT and 1M ammonium sulfate. The dialyzed protein sample was then loaded onto a
Hi-Load phenylsepharose (Pharmacia) hydrophobic interaction column (1.6 cm x 10 cm)
that had previously been equilibrated with a high ionic strength (high salt) buffer. The
sample was injected using the high ionic strength buffer (buffer B) consisting of 40 mM

HEPES at pH 7.2, containing 1 mM DTT and 1M ammonium sulfate. The purification
61

protocol was programmed to begin with a pre-wash, using the high ionic strength buffer
in order to remove all the non-specifically associated species from the column. This was
followed by elution with buffer of progressively decreasing ionic strength from 100%
buffer B to 100% buffer A. The overall flow rate was set at 0.3 mL/min. Buffer A was
compositionally similar to buffer B, except that it contained no ammonium sulfate.
Fractions were collected in volumes of 2 mL each and analyzed using SDS-PAGE. Only
fractions containing _>_95% pure PFL were pooled and concentrated. After removing
about 10 uL for analysis, the concentrated protein was routinely flash frozen in O-ring-
sealed micro-tubes and stored at -80°C. A typical chromatograph for this step of the

purification is shown in Figure 11.3.

1 8 17

m .............. .. 0
-600
.

 

 

 

1.50

100 .. 800

05" ~13]

/ - 50.0
b 0953 .

00mm 10000 20000 311100 «000 90000

 

Ill! 03; '00.“.‘1'
l

\
cm. mSIcm

 

 

ms»

Figure 11.3. A chromatogram from the second step in purification PFL using a
phenylsepharose hydrophobic interaction column.

62

Typically, because a smaller volume of protein was used for the phenyl sepharose
column, all protein eluted out of the column within a smaller buffer volume. The narrow
bandwidth of the sample also proved that there was good interaction between the column
and the sample. The fractions from this column therefore tended to be more concentrated

as is attested to by the peak being off scale.

11. 2.5. Determination of PFL concentration

The concentration of PF L was determined by using the method of Bradford5 with
bovine serum albumin (BSA) as a standard. BSA was obtained from Sigma as a powder
and the standard was prepared by preparing stock solutions of 1.19 mg/mL BSA. The
working BSA solutionwas prepared by diluting the stock 10x. The procedure involved
preparation of a protein-dye-water mixture to a final volume of 1000 ul, of which 200 pl
was dye and the remaining 800 pl was divided between water and protein. A standard
curve was prepared by pipetting different volumes of the standard into 1.5mL micro
tubes, incubating for 30 minutes to allow the blue color of the protein-dye complex to
reach maximum intensity. The absorbance of the standards was measured at 595 nm and
used to plot a calibration curve. The sample was prepared similarly using appropriately

diluted sample as illustrated with Table 11.1

63

Table II. 1. An example of the mixes used for the standard curve and sample during the
determination of PF L concentration using the Bradford assay

 

 

 

 

 

Sample/Std # Dye Protein H20
volume transferred (pl)

1 200 0 800

2 200 1 0 790

3 200 20 780

4 200 30 770

 

 

 

 

A typical analysis is illustrated by the data presented in Table 11.2, Table 11.3 and Figure
11.4. The concentration of the standard was 0.119 mg/mL and the PFL was diluted 200-
500 x before preparing sample for analysis. The concentration of the original
concentrated PFL was calculated from the equation of the standard curve, taking into
account the dilution factor. In determining the concentration of protein, samples were
routinely prepared in duplicate or triplicate per given volume, and various volumes were
used. This was done in order to test for both random error arising from the analyst, and
any systematic error that might arise as a function of the amount of protein used.
Analytically, only the absorbance values lying within the range of the standard curve can
be used to calculate the unknown PFL concentration. In the example given in Table 11.3,
the sample containing 50 pl of diluted PFL could not be included in the calculation of
PF L concentration because its absorbance was above the highest absorbance (0.885)
measured for 7.14 pg of standard protein. The example used here also shows illustrate
another tendency that was typically observed in the analysis of PFL, namely, that the
higher the amount of PFL in the sample, the lower the estimate of the concentration of

PFL in the original concentrated sample. Further investigation of this phenomenon

64

revealed that whenever the concentration of PF L in the sample was higher than the
maximum possible for a linear response, the Bradford analysis tended to underestimate
the concentration of the original protein solution, even if the sample absorbance may be
within the linear range of the calibration curve. Thus it is important to ensure not only
that the concentrations of the standard used fall within a linear response range, according
to the Beer-Lambert law, but also that the amounts of PFL used in the samples also fall
within linear response range for PFL. For all subsequent PFL concentration
measurements using this method several PFL samples were measured and plotted to
check for response linearity before subjecting them to analysis based on the BSA

standard curve.

Table 11.2 Typical series of standards used for calibration

 

0.119 mg/mL BSA standard

 

 

Volume (pl) pg Protein Absorbance

0 0 0.475

5 0.595 0.502

10 1.19 0.551

15 1.785 0.592
20 2.38 0.633
40 4.76 0.762

60 7.14 0.885

 

 

 

65

Table 11.3. The typical raw data and calculated protein concentration
calculated from the calibration curve.

 

 

 

l/500x diluted PFL
Volume (pl) Absorbance pg Protein Original Avg.
conc.(mg/ml) conc.(mg/ml)
20 0.673 3.37 84.36
20 0.690 3.67 91.77 88.1
40 0.851 6.48 80.95
40 0.840 6.28 78.55
40 0.843 6.34 79.20 79.6
50 0.913 7.56 75.56
50 0.898 7.29 72.94
50 0.933 7.90 79.04 75.8

 

 

 

 

 

 
   

 

0.95 —.
E J y = 0.0579x + 0.4809
g 0-85 i R2 = 0.9954
G)
‘0. 0.75 7'
8 .'
5 0.651
5
g 0.55
(O
0.45 1 I I 1 I T T
0 2 4 6 8
ug ptrotein

Figure 11.4. An example of a calibration curve used used to calculate the
concentration of protein analyzed by the Bradford assay.

66

11. 2. 5. Activation of PFL

The standard protocol followed in our lab to activate PF L entailed mixing all the
species involved in a specific order.6 A typical PFL activation mix consisted of 100-150
mM Tris-Cl buffer (pH 7.6), 0.1 M KCl, 10 mM oxamate, 8 mM DTT, 0.03 pM PFL-AE,
2 pM PFL, 0.2 mM AdoMet and 50 pM 5-deazariboflavin, all added to the reaction
vessel in the given order with the final concentration of each sample being as indicated.
After mixing in all the reagents inside an anaerobic chamber, the mix was then
illuminated with a high intensity (300W) halogen lamp. The temperature was maintained
between 16 and 23°C using an ice bath. Illumination was carried out for 30 to 45 minutes
at a time and the PF L activity was measured by using a coupled enzymatic assay
containing 100-150 mM Tris-Cl (pH 8.5), 3 mM NAD+, 55 pM CoA, 0.1 mg BSA, 10
mM pyruvate, 10 mM malate, 6 units of citrate synthase and 14 units of malic
dehydrogenase. All reagents used for PFL activation and for the assay mix were pre-
weighed into a Falcon tube and then pumped into an MBRAUNTM glove box. Tris-Cl and
water were made anaerobic in advance using a Schlenk line, and then stored inside the
anaerobic chamber. To 698 pL of the assay mix was added 2 pL of the mix containing
active PFL. To prevent reaction, the active PFL mix was placed on the inside cuvette
wall well above the surface of the assay mix. Because of the small volume of the active
PFL delivered to the cuvette wall, surface tension was strong enough to keep it from
flowing down into the assay mix, thus keeping the two separate until the desired moment.
After sealing it with an air-tight screw-on rubber stopper, the cuvette was carefully placed
upright inside the mini-chamber of the anaerobic box. Meanwhile, the computer-

controlled spectrophotometer was pre-set to measure the rate of NADH production
67

indirectly through the rate of increase of absorbance at 340 nm, using a coupled

enzymatic assay whereby the rate of production of NADH is directly related to the

concentration of active PFL (Figure 111.8.). The cuvette was brought out of the box and

the active PF L and the assay mix were brought into contact with two rapid shakes. The

cuvette was then placed in the instrument to measure the rate of NADH production in

units of nmoles per minute. The activity of PFL was periodically confirmed by using

EPR to directly detect and quantify the glycyl radical on PFL. The characteristic signal

for the glycyl radical is shown in Figure 11.5.

[10]

200

100

-100

p— | 1

P— .— "' I. '

L

p—

 

3000 3400 3800

Figure 11.5. The EPR spectrum of the glycyl radical on activated PFL

During the studies of activated PF L, we discovered that previously activated PF L,

upon prolonged exposure to ambient light, increased in activity, as measured by the rate

of NADH production during the enzymatic assay. This led us to explore the possibility

that PFL could be activated without need for high intensity illumination, but simply by

68

exposing the activation mix to ambient light. To investigate this, the reaction mix was
prepared as usual and incubated inside the anaerobic chamber under ambient light and
temperature of ~23°C. PFL activity was monitored over time by periodically removing 2
pl of the activation mix and adding it to 698 pl of the assay mix. The change in PFL
activity over time is shown in Figure II.6. The activity showed a dramatic increase from
0 to 116 nmolNADH/min in just less than 2 hours, reaching a maximum after

approximately seven hours.

180.0 —

160.0 —

140.0 -

120.0 -

100.0 4
80.0 -
60.0 J
40.0 4
20.0 4

0.0 m 1 , . . 1
0.0 2.0 4.0 6.0 8.0 10.0

Time (hours)

 

PFL activity (nmolNADHH/min

 

Figure 11.6. A time course activation of PFL under ambient light and 23°C. The
level of activation was found comparable to that achieved by high intensity
illumination. As predicted, PFL activated by this method, once it reached a
threshold activity, did not show any sign of reactivation during subsequent
experiments. This was in contrast to PFL that had been previously activated by
illumination, which showed evidence of additional activation upon exposure to
ambient light.

These findings have led us to propose that high intensity illumination of PF L is
not necessary for activation. Indeed, it may be detrimental as cooling with ice water

69

leads to rather sharp changes in temperature. We periodically observed some
precipitation of PFL during activation by intense illumination. However, in that case, the
temperature of the activation mix was always maintained below 23°C by using ice. We
hereby suggest that such precipitation, though relatively uncommon during activation,
presented a problem when it occurred because it effectively reduced the total PFL
concentration in the activation mix. Consequently, the maximum achievable PFL
activation would be decreased. It is also possible that activated PF L is less stable to
denaturation than unactivated PFL, thereby aggravating the problem. That precipitation
has been observed even at temperatures less than 23°C may be due to the rapid
temperature changes resulting from local effects within the bulk solution, particularly
when ice is added to cool the activation mix. This suggestion is also supported by the
observation that PFL activated without illumination was found to be exceedingly stable,
showing activities of over 50% of the original even after standing for' 2 days at 23°C
inside the chamber. It is also significant that when activation is done without need for
introduction of ice into the anaerobic chamber, it is easier to maintain the oxygen level
within the box at less than 1 ppm. It may also be argued that the latter activation
conditions approximate more closely those that probably prevail inside a living organism.
More quantitative studies are needed in order to determine the extent of activation during
each set of conditions. Our preliminary findings strongly suggest that illumination offers

no advantage over “dark” activation.

70

11.3. RESULTS

11. 3.1. Expression

BL21(DE3)plysS E. coli strain transformed with the plasmid pKK-pfl, expressed
PFL well. It was of interest to note that even though the plasmid used to clone the pfl
gene has an IPTG-inducible promoter, the transformed cells were able to overexpress
PFL extremely well, even without induction. There was only a marginal difference in the
PFL yield between the induced and uninduced cells. Sometimes the cells were only

grown to the stationary phase and harvested without induction.

11. 3.2. Purification of PFL

We routinely purified PFL using ion-exchange chromatography. The fractions
from this column generally contained PFL with at least 90% purity but still contained a
significant amount of impurities. The main chromatographic peak corresponding to PFL
in this purification appeared as essentially a single peak, thereby indicating that the
impurities found in this purification had similar PIs as PFL. Some of the impurities were
probably trapped in the PF L structure, just because of the high concentration of PFL
relative to the other proteins. As a routine protocol in our lab, all fi'actions containing
~90% or more were pooled together, concentrated and run through a hydrophobic
interaction column to improve purity. Based on SDS-PAGE analysis of the fractions,
only those containing ~95% PFL were pooled together and concentrated and used for the

investigations.

71

11. 3.3. Activation and stability of PFL

Initially, activation was done by using high intensity illumination as described in
the Materials and Methods section. However, our discovery that PF L could be activated
under ambient light afforded us a way to routinely activate this enzyme by merely
exposing it to regular fluorescence light in the presence of all the activation components.
To our surprise, PFL could be activated to the same extent using this method. This
method was preferred because it was practically far less cumbersome than activation by
high intensity illumination. Ambient light also approximated more closely the natural
conditions under which PFL would have to be activated.

During the course of experimental manipulations of PFL we also discovered that
this enzyme is extremely stable both in its active and inactive forms. PFL left at room
temperature to up to 10 days was found to be still activatable, for example. Notably, and
quite amazingly, there was also no sign of precipitation. Presumably, some of the protein
had denatured, but this was not immediately observable. That the enzyme could still be
activatable after such prolonged exposure at room temperature was quite remarkable.
Prompted in part by this observation, and during the course of other studies of PFL, we
also observed that activated PFL also preserved its activity for well over 24 hours under
anaerobic conditions, even at temperatures as high as 23°C. We therefore believe that in
addition to the inherent stability enjoyed by the glycyl radical due to the captodative
effect, the overall stability of the PFL protein plays some role in the unusual stability

observed for the PFL radical.

72

[1.4. Conclusion

PF L was successfully overproduced in an E. coli protein expression strain
(BL21(DE3)plysS) that had been transformed with a plasmid carrying an IPTG-inducible
promoter. Interestingly, this strain overexpressed PFL even without induction by IPTG.
This property of the bacteria proved advantageous, as the cells could be grown to the
stationary phase and harvested without need for induction. PFL was purified to
homogeneity using ion-exchange chromatography followed by hydrophobic interaction.
During the course of our experiments, we discovered that PF L can be activated by using
ambient light, as opposed to the high intensity light that we previously used to illuminate
it, consistent with published procedures. This proved advantageous because 1) it
simplified the experimental manipulations significantly and 2) although the general
conditions of activation remained artificial, they more closely approximated the natural
conditions under which PFL activation takes place.

PFL is remarkably stable, both in its active and inactive forms. Unactivated PF L
can be activated even after sitting at room temperature for over a week. Activated PF L
was found to be capable of retaining essentially all its activity for 24 hours at 23°C. We
suggest that this unusually high stability of activated PFL may be due to the inherent
stability of the PFL protein itself as well as the local inductive effects that stabilize the
radical captodatively. We expect that the discovery of this remarkable stability of PFL

will prove useful in any future studies conducted on this enzyme.

73

REFERENCES

1 Wagner, A.F.V.; Frey, M.; Neugebauer, F.A.; Schafer, W.; Knappe, J. 1992,
Proc. Natl. Acad. Sci.U.S.A 89, 996—1000.

2 Knappe, J .; Sawers, G. 1990, FEMS Microbiol. Rev. 75, 383-398

3 Knappe, J .; Elbert, S.; Frey, M.; Wagner, AF. 1993. Biochem. Soc. Trans. 21,
731 -734.

4 Conradt, H.; Hohmann-Berger, M.; Hohmann, H.P.; Blaschkowski, H.P.; Knappe,
J. 1984, Arch Biochem. Biophys. 228, 133-142.

5 Knappe, J .; Blaschkowski, H.P.; Schmitt, T. Eur. J. Biochem. 1974, 50, 253-263.

6 Broderick, J .B.; Henshaw, T.F.; Cheek, J .; Wojtuszewski, K.; Smith, S.R.; Trojan,
M.R.; McGhan, R.M.; Kopf, A.; Kibbey, M.; and Broderick, W.E. 2000.
Biochem. Biophys. Res. Commun. 269: 451-456

7 Janda, M.; and Hemmerich, P. 1976, Angew. Chem. Int. Ed. Engl. 15,7

8 Ashton. W.T., Brown, R. D. and Tolman, R. L. 1978, J. Heterolytic Chem. 15,
489-491.

9 Smit, P. 1986, Recl. Trav. Chim. Pays-Bays. 105, 538-544.

74

CHAPTER III

STUDIES OF ALCOHOL DEHYDROGENASE FROM ESCHERICHIA COLI

75

ABSTRACT

Alcohol dehydrogenase from E. coli (AdhE) is known to harbor both the alcohol
dehydrogenase and acetaldehyde reductase activities. Both of these activities utilize the
NADVNADH redox pair during processing of substrate. In the last decade this enzyme
has also been implicated as a pyruvate formate-lyase (PFL) deactivase, facilitating the
conversion of PFL from the active radical state to the inactive non-radical state by
quenching the radical that is characteristic of the active form of PF L. We were unable to
confirm this activity for AdhE. We did, however, observe clear deactivation of PFL by
other assay components in the assay mix. Significantly, some of these components were
also included in the deactivation studies from which AdhE was implicated as a PF L
deactivase. In this chapter we report the cloning, over-production and purification of
AdhE and its characterization with respect to both the alcohol dehydrogenase and
acetaldehyde reductase activities. Enzymatic assays were used to investigate further the
purported role of this multienzyme as a PF L deactivase. Our findings suggest that this
enzyme does not deactivate PF L. We propose that the deactivation observed in previous
reports was due to other components in the assay mix rather than AdhE. This proposal is
supported by the observation that several thiols that we investigated have proven to be
effective deactivators of PFL under conditions similar to those reported in the literature

(Chapter IV).

76

111.1 INTRODUCTION AND BACKGROUND

Metabolic breakdown of glucose in Escherichia coli (E.coli) takes place via
glycolysis, a series of chemically coupled phosphorylation reactions involving multiple
enzymes that ultimately lead to the production of pyruvate as the final product.
Glycolysis occurs under both aerobic and anaerobic growth, and results in a net reduction
of two NAD+ molecules to two NADH molecules (Scheme III.1). During glycolysis, a
net of 2 molecules of ATP are produced by substrate-level phosphorylation, supplying
sufficient energy to enable E. coli to survive during anaerobiosis. During aerobic growth,
pyruvate produced fiom glycolysis is oxidized to acetyl-COA by NAD+ under the
mediation of the pyruvate dehydrogenase enzyme complex (PDHc), with concomitant
production of NADH and C02, whereas in the absence of oxygen the conversion of
pyruvate to acetyl-COA is catalyzed by pyruvate formate-lyase (PF L), a homodimeric
enzyme whose active form harbors a radical co-factor. The PF L-catalyzed conversion of
pyruvate to acetyl-COA also produces formate as a by-product.

Subsequent processing of acetyl-COA in E. coli leads to two possible fates. If
catabolism occurs under aerobic conditions, acetyl-COA enters the tricarboxylic acid
(TCA) cycle wherein the acetyl group is oxidized to two C02 molecules (Scheme 1112).
A molecule of ATP is also produced for every complete TCA cycle, that is, two
molecules of ATP are generated for every molecule of glucose catabolized. The free
energy released from this oxidation is stored in the reduced co-enzymes NADH and
FADHz and is passed along the electron transport chain and ultimately used to reduce
oxygen to water. Concomitant with the recovery of the oxidized co-enzymes during the

electron transport, ATP is generated from ADP and phosphate in a process referred to as

77

2P1 + 2NAD+ 2NADH + 2H+

 

 

 

 

TIM
GAP DHAP _ 7 2(1,3-BPG)
GAPDH 2ADP
aldolase pGK
F-1 ,6-BP 2“" 3PG
ADP
PFK PGM
ATP
F6P 2PG
PGI enolase
36" PEP
11/ ADP 2ADP
HK pyruvate kinase
K ATP 2ATP
H
OH O
o H -
II
HO O
H H
Glucose Pyruvate

Scheme 111.1. A Schematic of glycolysis showing the steps involved during the breakdown of
glucose to pyruvate. The enzyme involved in each step is indicated on the side of the arrows in
abbreviation. Two molecules of ATP are consumed during the first step and four are produced in
the second part, thus yielding a net of two ATP molecules per molecule of glucose.

78

Glucose

i

Glucose-6-phosphate

i
i

2 x Gcheraldehyde-3-phosphate

2NAo+

NADH

2 x 1,3-Biphosphoglycerate

i

2 x Pyruvate
2 x AcetyI-CoA
2NADH
2 x Oxaloacetate 2 x citrate
2NAD+
2 x Malate 2 x lsocitrate
2NAD‘
TRICARBOXYLIC
2 X F marate ACID (TCA) CYCLE
2 x alpha-ketoglutarate
2FAD

2 x succinate

\\\_. 2NAo*
2 x SuccinyI-CoA

2NADH

Scheme 111.2. During aerobic growth, acetyl-COA is fed into the TCA cycle
wherein its oxidation is coupled to production of the reduced co-enzymes NADH
and FADHz.

79

oxidative phosphorylation. During aerobic respiration, 38 ATP molecules are produced
for every glucose molecule catabolyzed, and this mode of energy production is very
efficient.

During anaerobic growth, however, glycolysis provides the only source of ATP
for E. coli. One of the crucial steps in this catabolic process is the conversion of the
isomers glyceraldehyde-3-phosphate (GAP) and dihydroxyacetone phosphate (DHAP) to
two molecules of 1,3-biphosphoglycerate under the mediation of glyceraldehyde-B-
phosphate dehydrogenase. This substrate-level phosphorylation step involves the
consumption of two phosphates and reduction of two NAD+ molecules to NADH. The
NAD+ necessary for this step is supplied by the reduction of pyruvate via a series of
reactions starting with the decarboxylation of pyruvate to produce acetaldehyde (Scheme
111.3). Acetaldehyde produced by pyruvate decarboxylase is reduced to ethanol by
NADH under the catalysis of AdhE, and in the process, NAD+ is cycled for the
conversion of the GAP/DHAP pair to 1,3-BPG under the mediation of GAPDH. 1,3-
BPG produced in this step, in the presence of ADP, is then acted upon by
phosphoglycerate kinase (PGK) to generate more ATP and pyruvate. The alcoholic
fermentation process is known to involve two steps as summarized in Scheme HI.3. The
first step (A) is the decarboxylation of pyruvate catalyzed by pyruvate decarboxylase,
producing acetaldehyde as an intermediate. The second and final step is the reduction of
acetaldehyde by NADH to ethanol, and this reaction is catalyzed by alcohol

dehydrogenase (AdhE).

80

co2

o O J o
\\ // T , H— C//

 

 

/C- C\ + \ A
'0 CH3 pyruvate decarboxylase CH3
NADH NAo+
O
.0 \\\\_g’,// B
H—C + CH3CH2OH

\CHa alcohol dehydrogenase (AdhE)

Scheme HI.3. The two-step reduction of pyruvate to ethanol during alcoholic
fermentation

During anaerobic grth and in the absence of exogenous electron acceptors, E.
coli can produce any of a number of fermentation products, depending on the carbon
source on which it is growing. Some of the possible products include ethanol (Scheme
111.3), glycerol, formate, acetate, D-lactate, succinate, C02 and H2.1 Under these
conditions, AdhE is also able to reduce acetyl coenzyme A (acetyl-COA) in a two-step
process whereby two molecules of the reduced co-enzyme NADH are oxidized to two

molecules of NAD+ per turnover (Scheme 111.4).2

81

CH3CHO + NADH + H+

Acetyl-CoA + 2NADH + 2H+ = CoA + 2NAD” + CH3CH2OH C

Scheme 111.4. The reactions catalyzed by AdhE. The two steps (A and B) lead to

the overall (C)

This fermentation route is of interest in relation to pyruvate forrnate-lyase (PF L),
which, like AdhE, is an anaerobically induced and activated enzyme. In the presence of
CoA, PFL mediates the homolytic cleavage of pyruvate producing acetyl-CoA and
formate (Scheme 111.5). Based on the two reactions above, it is notable that during
anaerobiosis under conditions favoring the reduction of acetyl-CoA as depicted in
Scheme 111.4, PFL-mediated homolysis of pyruvate supplies the substrate for the
fermentation reaction catalyzed by AdhE. There is therefore interesting cooperation
between these two enzymes under anaerobic respiration, whereby the reduction of

pyruvate to ethanol begins with employment of PFL but is finally mediated by AdhE.

 

O O Q
\ /0 PFL / \
>c—c\/ + CoA ~ 8 “ CoA— C\/ + >C—H
'0 CH3 CH3 '0

Scheme 111.5. The reaction catalyzed by PFL.

In addition to the alcohol dehydrogenase and acetaldehyde reductase activities,

AdhE has been implicated in the deactivation of PFL, as shown by Scheme 111.6.3’4

82

H H, H

I AdhE /<
c 2
/ \ CoA, NADI, Fe2+

Gly 734

 

Gly 734

Scheme 111.6. The proposed role of AdhE as a PFL deactivase

This purported PFL deactivation activity is intriguing because it has to be invoked when
conditions change from anaerobic to aerobic, whereas both enzymes are known to
function under anaerobic conditions. PFL is translated in an inactive state and is then
activated post-translationally under anaerobic conditions by an activating enzyme called
PF L-activating enzyme (PF L-AE). 5 Its active state comprises a free radical located on at
Gly734, °‘° which has been demonstrated to be extremely sensitive to oxygen, undergoing
oxygenolytic fragmentation within a few seconds of exposure to air.7 In order to be an
effective deactivator of PFL, AdhE would have to do so very rapidly. It would also have
to be able to sense small amounts of oxygen before it can destroy the very sensitive PFL
radical. The PFL deactivase activity of this enzyme is, in part, the subject of this
investigation. Recently, a novel antioxidant role of AdhE was discovered under aerobic
conditions.9 If verified to be real, this activity adds valuable perspective towards
elucidation of the roles that AdhE plays in both E. coli as well as other medically
important facultative anaerobes, particularly those belonging to the family

E nterabacteriaceae.

83

111.2. MATERIALS AND METHODS
111. 2. I Cloning of His6-tagged AdhE

The gene sequence for adhE was obtained from the literature.10 The adhE gene
was amplified by the standard polymerase chain reaction (PCR) technique using E.cali
BL21(DE3)pLYS chromosomal DNA as a template. The synthetic oligonucleotide
primers had the sequences shown in Figure 111.1. The bold letters indicate the restriction
enzyme recognition sequences, with the corresponding restriction enzyme used at each

site indicated in italics above each restriction site sequence.

 

Xho
antisense strand: 5’ CGC CGC GGA TCC ATG GCT GTT ACT AAT GTC..3’

BamH-I

sense strand: 5’ CGC CGC CTC GAG AGC GGA ['11 1'11 CGC III..3’

 

 

 

Figure 111.1. The primers used to amplify the AdhE gene of E. coli by PCR. The
restriction sites are shown in bold and were included to improve cloning
efficiency.

The PCR product was digested with the restriction enzymes BamH-l and Xho and
then cloned into a PET-21a(+) expression vector (Novagen), also digested with the same
enzymes. The expression vector contains a His-tag sequence downstream of the Xho site.
This tag attaches a chain of 6 histidines onto the C-terrninus of the overexpressed protein.
The recombinant DNA obtained from the cloning of the adhE gene into PET-21a(+) was
first transformed into competent Epicurian XLl Blue E. coli cells obtained fiom
Novagen. The transformation mix was plated on LB-agar impregnated with SOpg/ml

84

ampicillin and incubated ovemight at 37°C. Twelve colonies were identified and marked
by circling over each on the cover plate of the petri dish. Using autoclaved toothpicks,
about a quarter of each colony was used to inoculate a Falcon tube containing 5 ml of
liquid LB medium to which was added 250 pg ampicillin. All twelve tubes were
incubated overnight in a shaker at 250 rpm. The temperature was maintained at 37°C and
each tube was covered so as to allow free diffusion of air into the cell suspension. Afier
incubating overnight, all twelve tubes showed good growth, as indicated by the turbidity

of the cell suspensions.

111. 2.2 Extraction and analysis of PE T -21 a ( +)-adhE

From each of the overnight cultures, a total 1.5 ml was centrifuged in order to
obtain a cell pellet. Plasmid DNA was then extracted from each cell pellet using a
commercial DNA extraction kit (Stratagenm). The purity of the extracted DNA was
determined by measuring the ratio of the absorbance due to DNA (A260nm) and
absorbance due to protein (Azgonm). An A260nm / A230nm of 1.6 or greater was considered to
indicate an acceptable level of DNA purity. In order to ascertain its genetic identity, the
recombinant plasmid DNA purified as above was digested with Xho and BAM-HI ,
followed by electrophoretic analysis on 0.6% agarose gel impregnated with 10 pg
ethidiurn bromide per mL. A lkb ladder commercial DNA standard was used as a marker
to establish the size of the fragments resulting from the restriction digest. Figure 111.2

shows the fingerprints of DNA obtained from each of the 12 colonies.

85

Std C1 C2 C3 C4 C5 C6 C7 CS C9 C10 C11C12

6108 i | H

5090
4072
3054
2035

    

Figure 111.2. DNA fingerprints of colonies 1 to 12 after digestion with Xho I

and BamH—I. The digested sample (10 pL) was thoroughly mixed with

approximately the same volume of loading buffer before loading into an agarose

gel containing 10 pg ml'I ethidium bromide. Lane 1 contains the DNA molecular

weight marker and the molecular weights of the bands are shown on the side.

The expected size of the adhE gene is 2650 bp in size. The rest of the PET-
21a(+)-adhE recombinant plasmid constitutes 5443 bp. Figure 111.2 shows that DNA from
colonies 1, 4, 5, 6 and 10 is cut into two fragments that correspond to the expected
molecular weights. DNA from colonies 2 and 7 was not detectable, probably because of
misleading of the wells. The band patterns from the other colonies also make sense.
Colonies 8 and 12 were partially digested as evidenced by the appearance of two bands of
identical masses to those seen with colonies 1, 4, 5, 6 and 10. The bands at ~4500 bp and
~8000 bp are probably due to supercoiled and nicked DNA respectively. Similarly, the
bands observed for colonies 9 and 11 were interpreted as being due to nicked and super
coiled DNA. DNA from colony 3 was the most difficult to explain. A lone band of
molecular weight corresponding to a PET21°(+) was observed for this colony. However,

the fragment corresponding to the adhE PCR product is not observed, therefore

suggesting that the band at ~5400 did not result from a double digest of PET-21a(+)-
86

adhE. Based on this interpretation, it was hypothesized that colonies 1, 4, 5, 6 and 10 and
possibly 8, 9, 11 and 12 were successful transformants and would therefore over-express

AdhE but colony 3 would not.

111. 2. 3. T ranfarmatian of BL21(DE3)pLysS with PE T -21 a( +)-adhE

The hypothesis proposed in section 111.3.2 was tested by transforming
recombinant plasmid DNA from colonies 3, 4, 5, 6 and 10 into BL21(DE3)pLysS and
using the transformants to test for over-expression of the adhE gene product. Competent
BL21(DE3)pLysS cells were obtained from Stratagenm. Following the prescribed
protocol, 50 pL of competent cells was transformed using 2 pL of the test recombinant
DNA. The transformed cells were then plated on LB/agar medium containing 10 pg of
ampicillin per milliliter of medium and incubated at 37°C overnight. For each of the cell
lines, the colonies obtained from the overnight culture plates were used to grow an
overnight liquid culture of the transformants in 5mL of Luria-Bertini (LB) medium
containing 10 pg of ampicillin per milliliter of medium. For each cell line, 50p] of the
overnight cultures were used to inoculate two 50 ml LB cultures and the optical density
of each culture was monitored at 600 nm (ODOOO). At 0D600 = 0.5, only one of each pair
of growths per cell line was induced by adding 1 M isopropyl B-D-thiogalactopyranoside
(IPTG) to a final concentration of 1 mM. Both the induced and uninduced cells were
allowed to grow for a further 2 hours before harvesting cell pellets fi'om 1 mL of each of
the cell suspensions. To test over-expression of AdhE, the cell pellets were denatured by
boiling for 10 minutes in a sodium dodecyl sulfate (SDS) denaturing buffer (Maniatis).

The denatured cells were then centrifuged at 12,000 rpm for 5 minutes and the

87

supernatant was loaded onto a 12% Tris-HCI electrophoresis gel purchased from
BioRadTM. Electrophoresis was run at a voltage of 250 mV and the protein bands were
exposed with Coomasie stain.
111. 2. 4. Streaking of the glycerol stocks to obtain colonies

The process of growing the cells from the glycerol stocks involved lightly
scraping the surface of a frozen stock of cells with a bacterial streaking needle and lightly
streaking onto an arnpicillin-impregnated agar culture medium contained in a bacterial
culture plate. At all times the glycerol stock was prevented from thawing by keeping it in
a Dewar flask chilled with liquid nitrogen and then putting it back into the fieezer
immediately after plating was completed. The streaked plate would then be incubated at
37°C overnight. At all times the working environment was kept sterile by applying
ethanol to the workbench and then performing the streaking under a flame. Figure 111.3
illustrates the streaking technique used. The streaking needle was sterilized in a flame
and then cooled by touching it to the sides of the chilled micro tube containing the
glycerol stock. The needle was lightly scraped on the surface of the frozen glycerol stock
and then lightly streaked in different regions as illustrated in steps 1 to 3. The darkest
streak (with an arrow) is the last one in each step and represents a progressively

decreasing concentration of cells as streaking proceeds.

88

 

Figure 111.3. Steps followed in streaking a glycerol stock onto an agar plate for
overnight incubation.

Typically, several colonies were observed on the plate after 16 hours as illustrated in
Figure 111.4, and tended to be more isolated with the later streaks as shown in the boxed

regions. Regions 1, 2 and 3 correspond to the streaking steps in Figure 111.3.

 

 

 

Figure 111.4. The typical appearance of colonies on an agar plate after overnight
incubation. Regions 1,2 and 3 (boxed) correspond to the streak steps in Figure
111.4 and show stepwise dilution of the colonies with each subsequent streak step.

Since all the cells in the suspension used to prepare the glycerol stocks can be traced back
to a single transformant cell, all of the colonies on the streaked overnight plate could
reasonably assumed to be isogenic. The choice of colony for the to be used for the
overnight grth was usually based on the appearance of the colony. Only colonies that
appeared healthy (not overgrown and having no satellite colonies around them) were used

for overnight growths.

89

111. 2. 5. Growth, Harvesting and Lysis

Using a colony from a streaked plate prepared as in section 111.3.4, a single colony
was inoculated into a 50 mL of LB/Amp medium and incubated overnight in a
shaker/incubator set at 37°C and 250 rpm. To a set of six 2.8 L Fembach flasks
containing ampicillin at a concentration of 50 pg/mL, 1 mL of the overnight culture was
inoculated. The growth of the cells was monitored by measuring the optical density of
each culture at 600 nm (0D600). When 0D600 reached 0.5, the cells were induced by
adding 1M isopropyl B-D-thiogalactopyranoside (IPTG) to a final concentration of 1mM
and then allowed to grow for a further 2 hours.

After 2 hours, the cells were harvested in pre-weighed bottles by high-speed
centrifugation using a Sorvall GS3 rotor at a speed of 8000 rpm for 5 minutes. The
supernatant was then decanted and the wet cell pellet was drained by inverting the
centrifuge bottles on absorptive paper for approximately 10 minutes. The bottles plus the
drained cells were weighed and the mass of the cell pellets determined by difference.
The cells were either lysed directly or stored at -80°C for later use. The lysis buffer
consisted of 50 mM NaH2P04, 10 mM imidazole and 300 mM NaCl, the whole mix
adjusted to pH 8. To an average of 6 g of cells per harvest bottle, 20 mL of lysis buffer
containing 8 mg lysozyme, 1 mM phenyhnethylsulfonyl-sulfate (PMSF) and trace
quantities (~0.1 mg) of each of RNase and DNase were added. The cell suspension was
thoroughly mixed and then homogenized continuously with a magnetic stirrer at 4°C for
approximately 2 hours. The lysed cell paste was centrifuged at 15,000 rpm using a

Sorvall SS34 rotor for 30 minutes at 4°C. The supernatant containing the over expressed

90

AdhE was carefully pipetted into a 50 mL FalconTM tube, flash-frozen in liquid nitrogen

and stored at -80°C or used directly to purify AdhE.

111. 2.6. Purification of His6- AdhE

111. 2. 6. 1. Purification Principle

The protein was purified using a commercial nickel nitrilotriacetic acid (Ni-NTA)
immobilized metal affinity chromatography (IMAC) resin from QIAGEN®. This column
consists of Ni2+ immobilized by a tetradentate nitriloacetic acid chelate in highly cross-
linked 6% agarose beads on a superflow support. Figure 111.5 is a schematic to show how

Ni is immobilized by coordination to nitrilotriacetic acid (N TA.

1.

,Co—CH CH2
\ 2/
.2+. ..-“““N\

' Cvaw»
\ /

0*CO

4‘

Q\.........o

I

N

0\
011111111111101211110

I
N

Figure [11.5. Illustration of how immobilized Ni is coordinated to a
nitrilotriacetic acid (NT A) support.

Upon loading the column, His-tagged AdhE protein displaces the water molecules
and binds to the two remaining Ni2+ coordination sites via the nitrogen atoms on the
imidazole rings of the histidine tag. The interactions between the NTA-immobilized Ni
ion and two histidines of the tag are illustrated in Figure 111.6. Once AdhE is bound as

shown, all other proteins that do not specifically bind to the resin were washed out of the
91

column using a buffer containing mild concentrations (20 mM) of imidazole before
eluting AdhE out with similarly constituted buffer containing a higher concentration (250

mM) of imidazole.

 

c—o N \\

C//CH2—<\ fl /C\CH2 H H

\ N“ (:3 N/ H H2 H2 CZ ’0:

NH \0. he Va” C.“
,x‘ \NJ 2 2 H2 HO

O

. 1.1
Z \
L

2.

0\

é o’

\

0%

Figure 111.6. Interactions between NTA-bound nickel and the two histidines
from the His, affinity tag. The Niz’i ion is held in place by nitrilotriacetic acid via
tetradentate chelation resulting in a very stable complex. Under conditions of
near neutral pH and prior to the introduction of imidazole or the histidines-rich
tagged AdhE, two of the six coordination sites on Ni2+ are occupied by water
molecules.

111.2. 6.2 Purification and Analysis of His6-tagged AdhE

The purification protocol entailed three major steps. 1) The crude lysate
containing all the water-soluble proteins expressed by the bacterium was loaded onto the
column in low-imidazole-concentration (lysis) buffer. 2) The loaded column was washed
with medium-imidazole-concentration (wash) buffer to remove all components not
specifically bound to the column. 3) The Ni-bound AdhE was eluted out of the column
using a high-imidazole-concentration (elution) buffer. The lysis, wash and elution

buffers all contained 50 mM NaH2P04 and 300 mM NaCl and were adjusted to pH 8.
92

They only differed in the amount of imidazole they contained, that being 10, 20 and 250
mM for the lysis, wash and elution buffers respectively. The purification protocol was
developed by first determining the column capacity with regard to protein load and the
optimum pH for the elution buffer. The optimum buffer pH was established by eluting
the protein with buffer of pH 6.9, 7.5, 7.6 and 8.0 and comparing the quality of elution in
terms of both amount and purity of protein obtained. Using the buffer at pH 6.9 and 7.5,
satisfactory and indistinguishable levels of enzyme purity were achieved (Figure 111.7).
However, in order to minimize Ni2+ leaching of the column due to protonation of NTA
under acidic conditions, and thereby rapidly compromising the column’s binding affinity,
pH 7.5 was chosen for all subsequent purifications. After several trials were canied out,
it was concluded that the optimal protein load of the fully regenerated and equilibrated
column was maximally ~3 g of total protein per mL of resin. Each trial involved loading
a known amount of total protein onto the column followed by several washes with pre-
determined volumes of wash buffer. Fractions of equal volumes of wash buffer were
collected and saved for analysis using SDS-PAGE. This was followed by eluting the
protein with elution buffer in like manner and saving all elution fractions for subsequent
analysis by SDS-PAGE. Figure 111.8 shows a sample data obtained from such an analysis
and the basis upon which the optimal conditions used to purify Hisb-tagged AdhE were

determined.

93

 

Figure 111.7. Comparison of the effect of protein load on purification quality.
The concentration of crude lysate was 6 mg/ml. In panel A, 2 ml of clear lysate
were loaded onto 1.5 ml resin. Lane legend: 1: molecular weight marker (kDa),
2: crude lysate, 3: flow-through from loading, 4, 5, 6: first, second and third
washes, 7, 8, 9, 10: elutions 1, 2, 3 and 4. In panel B lml clear lysate was loaded
onto 1.5 ml resin. Panels B and C show data from purifications using elution
buffers at pH 6.9 and 7.5 respectively. For both panels lane 1 is the molecular
weight standard, lane 2 is the flow-through, lanes 3 and 4 are washes 1 and 2
respectively and lanes 5, 6 and 7 are elutions l, 2 and 3 respectively. The lysis,
wash and elution buffers all contained 50 mM NaH2P04 and 300 mM NaCl and
were adjusted to pH 8. They differed by containing 10, 20 and 250 mM
imidazole respectively. The volume of the FT was 2 for panel A and 1 ml for
panels B and C. All washes and elutions were done with 1 ml of buffer per
fraction.

In panel A, lane 2 shows that AdhE was both very well over-expressed and
soluble in the lysis buffer as evidenced by the very broad band at molecular weight 97
kDa. From the composition of the flow-through and first wash (lanes A3 and A4
respectively) it is evident that the majority of the proteins in the lysate did not interact
with the column specifically, hence they passed right through the column with the
loading (lysis) buffer (lane A3) and the remainder were washed out in the first wash (lane
A4). Lanes A5 and A6 show that washes 2 and 3 contain AdhE almost exclusively,
indicating that AdhE was overloaded to the column but was able to interact with the
column sufficiently strongly to slow down its migration significantly compared to the

94

other proteins. It should be noted that after the fifth wash (not shown), virtually no AdhE
was detectable in the wash buffer. When the concentration of imidazole was increased to
250 mM, essentially all AdhE was washed out of the column afier 4 elutions (lanes A7 to
A10). Panel A also shows a relatively significant co-expression of a protein of molecular
weight of approximately 45 kDa. Most of this protein washes out with the rest of the
impurities but a small amount of it co-elutes with AdhE but constitutes a relatively
insignificant amount of impurity. Panels B and C show that loading 4 mg total protein
per mL of resin resulted in improved retention of AdhE by the column as opposed to 8
mg total protein per mL of resin used in panel A. No AdhE was detectable in the flow
through and first wash. It was not until the second wash that some AdhE was detected,
thus showing improved AdhE retention by the column at a lower protein load.
Comparison of lane 4, 5 and 6 in both panels B and C shows that a relatively large
amount of AdhE washed out of the column after wash 2 (lane 4), following which
relatively little AdhE came out in the first elution (lane 5), and then a large amount once
again came out with the second and third elutions (lanes 6 and 7). Wash 2 (lane 4)
probably contains AdhE that was specifically but weakly bound to the column due to
saturation effects, resulting in dynamic exchanges between bound and unbound AdhE
molecules. This would account for the delay of the AdhE detected in the second wash.
However, when the “excess” AdhE molecules were washed out, the remaining AdhE
molecules were more resistant to elution (lane 5) until a critical concentration of the
competing imidazole was established (lanes 6 and 7).

After determining the optimum purification conditions, large-scale purification of
the protein was carried out by applying 30 ml of 50% Ni-NTA/ethanol slurry to a 50 ml

95

Flex Columnm. The column was equilibrated with 50 ml of lysis buffer at a rate of 2
ml/min, and 7ml of clear lysate (6mg/ml protein) were then applied. (The manufacturer
recommends mass of protein loaded was within the range 5-10 mg protein / ml of resin).
The level of AdhE purity achieved using the Ni-NTA IMAC purification technique was

very good a will be shown in the “Results” section.

111. 2. 7. Cloning, Growth and Purification of untagged AdhE
111. 2. 7. 1. PCR and ligation

In an effort to investigate the effect, if any, of the His6-tag on the catalytic activity
of AdhE, the adhE gene was cloned into a commercial vector that would allow
expression of untagged AdhE. Primers were designed as described previously for the
His6-tagged AdhE. The oligonucleotide primers were ordered from Integrated DNA
Technologies, INC, Coralville IA (Figure 111.8) The antisense primer was 45 bases long,
had a molecular weight of 13,837g/mol, and contained 65300 ng of DNA. The sense
primer was also 45 bases long but had a molecular weight of 13,757 g/mol and contained

83100 ng of DNA.

96

 

Xho I
antisense strand : 5’-GCG CGC GGA TCC ATG
GCT GT1 ACT AAT GTC
GCT GAA CI'I' AAC GCA ...... 3’

BamH-l
sense strand: 5’-GCG CGC CTC GAG TTA
AGC GGA TTT TIT CGC
'ITT TIT CT C AGC TTT ......... 3’

 

 

 

Figure 111.8. The oligonucleotide primers used to amplify the AdhE gene for
expression of the untagged enzyme. The bold sequences are the restriction sites
and the restriction enzymes are shown in italics, each above its corresponding
recognition site.

Based on the formula Tm (°C) = 2(NA +NT) + 4(NG + NC) obtained from the Stratagen Pfu
DNA Polymerase Manual (2002), the melting points for both primers were calculated to
be 90°C. For good yield of PCR product and to ensure high fidelity, between 25 and 30
cycles are generally used. The PCR thermocycler was therefore programmed to run 25
cycles before maintaining the reaction mix at 4°C. In addition, based on the guidelines
from Stratagen (page 3 of Stratagen Pfu DNA Polymerase Instruction Manual (2000)),
the extension time should be 1 minute per kb for vector targets up to 10 kb and genomic
targets up to 6 kb. Since adhE is 2.65 kb long, an extension time of 3 minutes was used.
Genomic DNA was obtained from BL21 E. coli cells grown overnight at 37°C in
LB medium, with shaking at 250/min to ensure good aeration. DNA was extracted using
the Wizard® genomic DNA extraction kit from Promega. The genomic DNA obtained
was analyzed both qualitatively and quantitatively by UV-visible spectroscopy. A 50x
diluted sample was measured to simultaneously determine A260 (absorbance by DNA)

and A280 (absorbance by protein impurities). The ratio A260/A280 was measured in

97

duplicate and averaged 1.7. The absorbance at 260 nm (A260) was 0.24. Using the
relationship that 1 AU corresponds to 50pg/m1 of DNA, and taking into account the
dilution factor of 50, the concentration of the genomic DNA was calculated to be 600
ng/pl.

For the PCR reaction, the template DNA was diluted 6x to a final concentration of

~100 ng/ pl. The mixes prepared for the PCR reaction are summarized Table 111.1.

Table 111.1. PCR Reaction Components

 

 

 

 

 

 

Cocktail 1

Component Vol,(pl) Amount in reaction
H20 20 -
25mM dNTP 0.4 0.2 mM
70 ng/ pl Atisense primer 1.9 133 ng
80 ng/ pl Sense primer 1-7 136 118
100 ng/ pl Genomic DNA (template) 1 100 118

Cocktail 2
H20 20 -
10x Cloned Pfu DNA polymerase buffer 5 1x
2.5 U/ pl Pfu Turbo DNA polymerase I 2.5U

 

Cocktails 1 and 2 were mixed together immediately prior to being placed in the
thermocycler. The program parameters of the thermocycler program were set as
follows: Melting temperature, 90°C; Annealing temperature, 55°C;
Polymerization temperature, 72°C; Extension time, 3 minutes; Number of cycles,
25.

Four reactions were carried out altogether. In the design of the primers, 8 Bam
H1 restriction site was engineered immediately upstream of the start codon with a 6-base

GC overhang immediately upstream of it. Correspondingly, a Xho restriction site was

98

inserted immediately downstream of the stop codon, followed by a 6-base GC overhang.
Although blunt-end ligation was used for the untagged adhE, the restriction sites were
engineered into the primers in order to provide an option for sticky-end ligation in case it
became necessary. The adhE gene has 2650 base pairs. Taking into account the GC
overhangs and the restriction sites, the total number of bases in the PCR product was
expected to be 2650 + 24 = 2674. Upon running an agarose gel, a band was therefore
expected between the 3054 and 2036 base pair bands on the standard marker. The PCR

band fell exactly where it was expected as shown in Figure 111.9.

Std adhE PCR product

3054

2036
1636

 

Figure 111.9. The adhE PCR product for the untagged AdhE. The PCR
product lies exactly where it is expected (2674 bp).

The adhE gene was ligated onto PETBlue, a blunt-end vector from NovagenT'“,
according to the protocol prescribed. A positive control was included using an insert

supplied in the commercial kit. The ligation procedure involved two steps.

99

1) End—conversion, in which the PCR product was reacted with a dephosphorylating
agent in order to convert it to a form that will allow it to ligate through the loose
triphosphate chain on the commercial vector

2) Ligation to the vector, whereby the end-modified PCR product was reacted with
the blunt-end commercial vector PETBlue (N ovagen).

For the first step, the mix was incubated for 2 hours at 220C and the reaction was
quenched by heating at 750C for 5 minutes. The reaction mix was then cooled in ice for
2 minutes and then centrifuged for 5 minutes to collect the condensate. In the second
step, lpl each of the commercial blunt-end vector and T4 DNA ligase were added. For
the positive control, a control insert supplied with the kit was used. The mixes were then
incubated overnight at 16°C. It should be noted that blunt-end ligation does not require
the inclusion of restriction sites on either end of the gene product as was done in this
work. This was done as a back-up measure, providing an option for sticky-end ligation in

the event that the blunt-end ligation was not successful.

111.2. 7. 2. Transformation, plating, DNA extraction and fingerprinting

111. 2. 7. 2. 1 Transformation

The overnight ligation mixes were transformed into NovaBlue cells according to
the following protocol. Five tubes, previously stored at -800C, each containing 25 pl of
commercial (Novagen) competent cells, were thawed in ice for 5 minutes. To two of
them were added 1 and 2 pl of control insert ligation mix respectively. To the remaining
three were added 1, 2 and 5p] of the experimental ligation mix respectively. After

mixing well and leaving in ice for 5 minutes, the cells were heat-shocked for exactly 30
100

seconds at 42°C and placed in ice for two minutes. To each tube 250p] of SOC medium

was added, followed by incubation for 30 minutes.

111.2. 7.2.2. Plating

The ampicillin-impregnated medium to be used for plating had previously been
treated to allow for white/blue screening of the bacterial colonies. To each plate, 20 pl of
100mM IPTG and 35 pl of 50 mg/ml X-galactose (X-gal) were added, spread and
allowed to soak into the medium for 30 minutes. 50 pl of each of the transformation
mixes were plated and incubated at 37°C overnight.

The principle behind the screening process is the following. Beta-galactosidase is
one of the enzymes encoded by the LacZ operon of the PETBlue-l plasmid vector
palylinker. Beta-galactosidase metabolizes X-galactose to produce a blue product. When
ampicillin is included in the growth medium, only cells that are resistant to the antibiotic
(i.e., those containing the vector containing the ampicillin resistance gene) can grow. Of
these, those without insert (containing an intact palylinker) will be blue due to breakdown
of X-galactose. The cells containing an insert are not able to metabolize X-gal because
the insert disrupts the expression of the LacZ operon thereby abolishing the expression of

beta-galactosidase. The positive colonies are therefore white.

111.2. 7. 2. 3. DNA Extraction
Both the positive control plates had over 1000 colonies and about 35% were
white. Of the experimental plates, all had at least ~200 colonies but only about 3% were

white. The positive colonies fi'om the three experimental plates were numbered 1 to 16.
101

After labeling 15 ml Falcon tubes correspondingly and filling each with 5 ml of
autoclaved LB medium containing 50 pg ampicillin / ml, an autoclaved tooth pick was
used to transfer some cells from each colony into the medium. The cells were then grown
overnight at 37°C and shaken at 250 rpm. Plasmid DNA was extracted from each of the
overnight cultures. To recover a cell pellet of the putative PETBlue-l-adhE cells (#1 to
#17), 2 m1 of cell suspension were centrifuged. The DNA was extracted following the

protocol in the commercial Plasmid DNA extraction kit fiom Promega.

111. 2. 7. 2. 3. Fingerprinting

The plasmid DNA extracted from each of the colonies was digested with Sea], for
which only one restriction site is known to occur on the vector, but not on the insert.
Table 111.2 shows an example of the mixes used for an experimental and control digest
for each of the DNA extracts investigated.

Table 111.2. Typical mixes for experimental and control restriction digest mixes for
PETBlue-adhE fingerprinting.

 

Plasmid (pl) 10x Buffer 6(pl) H2O (pl) Sca 1(pl)

 

Experimental 6 1 l .5 1.5
Control 6 1 3 0

 

 

 

 

 

Afier digesting for three hours at 37°C, the samples were analyzed by
electrophoresis on 0.6% agarose gel containing 0.5 pg/ml ethidium bromide using a
commercial DNA standard as a molecular weight marker. An exhaustive and systematic
analysis of all 17 extracts was done. DNA from only one of the colonies showed bands
consistent with PETBlue (3476 base pairs) with an adhE insert (2650 base pairs). A
single cut with Sea] is expected to result in a band corresponding to linearized DNA of

molecular weight in the region of 6126 base pairs. In addition, the control was expected

102

to migrate farther since it consisted of circular supercoiled DNA as demonstrated in

Figure 111.10.

 

Figure 111.10. Agarose gel of PETBlue-adhE recombinant plasmid digested with
Scal. Lane a is the molecular weight marker, lanes b&c, d&e, f&g, h&1, j&k,
l&m show digested and undigested DNA extracted from colonies 2, 3, 4, 5, 6 and
17 respectively. Only DNA from colony 3 (d&e) showed bands that are
consistent with the bands expected of singly digested and undigested PETBlue-
adhE. The singly digested plasmid (lane d) shows a band at the expected
molecular weight of ~6100bp. The two bands observed for the undigested
plasmid (lane e) were probably due to supercoiled DNA (lined with ~4000bp
marker) and nicked DNA (at ~6000bp).

In order to confirm the identity of the plasmid from colony 3, the DNA finger-
printing was repeated, but this time with 3 pl of plasmid DNA instead of 6p]. DNA from
colonies 2 and 4 was also included in the analysis to confirm the negative results from the
first experiment. In addition, the PCR product was run together with the digested
plasmid DNA in order to verify its molecular weight. The results previously obtained
were reproduced and thus confirmed that only colony 3 contained PETBlue-adhE. The

identity adhE PCR product was also confirmed (Figure 111.11).

103

std123 4 55 adhE

 

Figure 111.11.Confirmatory DNA fingerprints of plasmids extracted from
colonies 2, 3 and 4. Lanes 1&2, 3&4, 5&6 contain undigested and digested
plasmid DNA from colonies 2, 3 and 4 respectively. Lane 3 (digested plasmid
from colony 3) lines up with the 6000bp band of the molecular weight marker.
The undigested plasmid from colony 3 migrated at the same rate the 4000bp
marker. This is consistent with undigested and super-coiled DNA. The adhE
PCR product corresponded with the expected molecular weight of 2674 bp.

111.2. 7.3. Growth of Cell Cultures and Purification

111.2. 7.3.]. Test Grawths

Afier confirming the positive recombinant DNA from colony 3, 2 p1 of it was
transformed into 50 pl of Tuner (DE3) pLacI cells according to the transformation
protocol described previously. The cell suspension (100pl) was plated in duplicate and
incubated overnight at 370C. Afier 20 hours, two colonies were observed on one of the
plates. The colonies were given reference numbers labeled 1 and 2 respectively.
Overnight cultures of both colonies were started in LB/Amp. The overnight cell
suspensions (lmL) from each of colonies 1 and 2 were inoculated into each of two 250
mL flasks containing LB-Amp liquid medium. The cell suspension of each of colonies l
and 2 was divided into two equal volumes that were grown in parallel. The optical
density measured at 600nm (0D600) was monitored over time and when it reached 0.66

and 0.65 for colonies 2 and 1 respectively, 50 pl of a 1M solution of isopropyl-B-D-
104

thiogalactopyranoside (IPTG) was added to only one half of the growth to induce
overexpression. All cell suspensions were allowed to grow for two hours post-induction.
Pellets from each of the four growths were made by centrifuging 3 ml of each
suspension. The pellets were lysed according to standard protocols using SDS denaturing
buffer containing 3.5% B-mercaptoethanol, followed by centrifirgation for 5 minutes. For
each sample, 10 and 15p] were loaded into 15p1 wells of a 4-15% Tris-HCl gradient gel.
Both colonies showed good over-expression in growths without IPTG but significantly
decreased expression of AdhE in the grth to which IPTG was added, as can be seen in
Figure 111.12. Glycerol stocks were made by mixing the cell suspensions with autoclaved
glycerol to 30% in a total volume of 1000 pl. Curiously, repeated test growths using the
glycerol stocks from colony 2 consistently showed no AdhE overexpression when
growths were repeated. All subsequent growths of this cell line were therefore done with

stocks from colony 1.

97 kDa— fill an
H , ; .“1
H ‘
H .» ~M Ni

“—-

Figure 111.12. Overexpression of untagged AdhE by colonies l and 2. The first
lane is the molecular weight marker, the next four lanes are samples from colony
1 and the last four from colony 2. For each, the first pair of lanes are a duplicate
from the uninduced grth and the second pair are from the induced growth.

In order to provide firrther proof that IPTG suppressed overexpression of
AdhE in the Tuner (DE3) pLacI-adhE transformants, 5mL of an overnight culture cell

105

suspension were inoculated into two 2.8 L flasks each containing 1L of LB medium
impregnated with ampicillin to a final concentration of 50 mg/ml. Both flasks were
incubated at 37°C in a shaker incubator going at 250 shakes per minute. When the cell
suspension reached an optical density (0D600) of ~0.7 1 mL of 1M IPTG was added to
flask 1. Both cell suspensions were allowed to incubate for a further two hours and then
harvested. The wet cell masses were 6.07g and 5.85 g for flask 1 and flask 2 respectively.
Both were lysed with equal volumes of lysis buffer and the crude extract recovered by
careful pipetting. Alcohol dehydrogenase activity of the crude extract was measured as
an indirect means of determining the effect of IPTG on the expression of AdhE in Tuner

(DE3) pLacI-adhE transformants. The results obtained are shown in Table 111. 3.

Table 111.3. Alcohol dehydrogenase activities of crude extracts from induced and
uninduced Tuner (DE3) pLacI-adhE transformants

Measurement # ADH, Flask 1 (nmolNADH/min) ADH, Flask 2 (nmolNADH/min)

 

 

 

 

1 21.4 80.4
2 22.7 84.2
3 24.6 83.6

 

As shown in Table 111.3, the AdhE suppression observed in the whole cell pellet (Figure
111.12) was corroborated by a corresponding decrease in alcohol dehydrogenase activity
of the crude lysate obtained fi'om cells that had been “induced” by addition of IPTG
during growth. The uninduced cells showed close to four times the activity, per gram of
cells, than the induced cells. Based on these findings, all subsequent growth of these
cells were done without addition of IPTG, and AdhE overexpressed well when the cells

were grown up to the stationary phase.

106

111. 2. 7.3.2. Large Scale Growths

To prepare for a growth, glycerol stocks were routinely streaked on an LB-agar
medium containing 50pg/ml ampicillin and incubated overnight at 37°C. An isolated
colony was selected and transferred to LB liquid medium containing 50pg/ml ampicillin
(LB-Amp) and again grown overnight at 37°C. For a large—scale growth, 5 ml of the
overnight grth was transferred to 2.8 L flasks each containing 1.5 L of LB-Amp
medium. The flasks were incubated in a shaker going at 250 rpm and maintained at 37°C
until the optical density at 600 nm (ODOOO) was 2.5. This tended to be the maximum
attainable cell density, and it was reached within approximately 6 hours. Because
addition of IPTG was found to suppress AdhE overproduction in Tuner (DE3) pLacI cells
(Figure 111.12), these cells were routinely grown to the stationary phase and then
harvested without adding IPTG. Each growth was routinely checked for over-expression
by spinning out a small pellet of cells, denaturing it by boiling in SDS denaturing buffer

before loading the supernatant into a Tris-Cl polyacrylamide gel.

111.2. 7.3.3. Cell Harvesting and Lysis

The cells were harvested by centrifugation at 8000rpm for 5 minutes in 500 ml
pre-weighed polyethylene centrifirge bottles and alter discarding the supernatant, the cell
pellet was drained by inverting the bottles on tissue paper for about 10 minutes. The
bottle plus the cells was then weighed and the difference from that of the empty bottle
was recorded as the mass of the cells. In preparation for AdhE purification, the Tuner
(DE3) pLacI-adhE cell pellets were lysed in a pH 7.5 buffer consisting of 50 mM Tris-

sulfate ( pH 7.5), 200mM NaCl, 1% TritonX-IOO, 5% glycerol, 10 mM Mng, 1 mM
107

DTT, 1 mM PMSF, 8 mg of lysozyme and trace amounts (~0.1 mg) of each of DNase
and RNase and incubated at 4°C for 2.5 to 3 hours. For each lysis batch, 1.5 mL of lysis
buffer was used per gram of cell pellet. An average grth yielded a total of ~25 g of
cell pellets. The cell paste resulting from enzymatic lysis was then centrifuged in a
SorvallTM super centrifuge equipped with an SS 34 rotor spinning at 18,000 rpm for 30
minutes. A The supernatant was carefully decanted into 50 mL Falcon tubes and either

flash fiozen or directly used to purify AdhE.

111.2. 7.3.4. Purification of untagged AdhE

The untagged AdhE protein was purified by size exclusion chromatography using
a Superdex 75TM gel filtration column 5cm x 60cm). Typically approximately 40 to 50
mL of crude protein extract obtained as described in section 111.3.7.3.3, were loaded onto
the column at a rate of 3 mL per minute using Tris buffer. The optimization of the
purification method was achieved by monitoring the change in absorbance at 280 nm as
protein eluted out of the column. Fractions corresponding to the peaks observed were
analyzed by SDS-PAGE to determine their composition by molecular weight. After
identifying the peak corresponding to AdhE under the experimental conditions, the
procedure was automated by writing a program that allowed fractions to be collected
automatically in the chromatographic region spanning the AdhE peak. SDS-PAGE was
routinely used to determine the composition of each of the fractions and to decide which
fractions to pool for further purification. The purity of the protein was firrther improved
by loading it onto an ion exchange column followed by gradient elution from low to high

salt concentration. The low salt buffer consisted of 20mM HEPES containing 1 mM

108

DTT and adjusted to a pH of 7.2. The high salt buffer was similar, but also contained
200mM NaCl.

The overexpression of untagged AdhE in the Tuner (DE3) pLacI cells was poorer
than that of Hisé-tagged AdhE expressed in BL21 (DE3)plysS, as seen by comparison of
Figures 111.13 and 111.15. The untagged AdhE also proved to be more difficult to purify to
homogeneity, even even after repeated trials. After carrying out SDS-PAGE on the
fractions corresponding to the AdhE peak, those containing 3 ~90% wild type AdhE or 3
~95% for the Hisé-tagged AdhE were pooled together and concentrated using an

AmiconTM concentrator equipped with a 50 kDa molecular weight cut-off (MWCO) point
membrane purchased from Milliporem. The concentration of the protein was determined
by the Bradford method. To prepare the standards, a series of microtubes containing
various known volumes of a 0.119 mg/ml solution of bovine serum albumin. The
volumes of the protein standard solution ranged from 0 to 60 pl. To each of the tubes,
enough sterilized water was added to make up the volume to 800 pl. 200 pl of the
Bradford dye (obtained from BioRadT”) were added to each of a series of microtubes to

make up the volume to a total of 1000 pl and to develop the spectrophotometrically
active analytical complex. The tubes were then allowed to stand for 30 minutes to allow
development of color. In parallel to the standards, a series of samples containing various
volumes of diluted AdhE were prepared and incubated. The absorbance of the protein
complex in each was measured at 595 nm and a calibration curve was plotted and used to

calculate the concentration of AdhE in the samples.

109

111. 2. 7.4. Characterization of untagged AdhE

In order to ascertain that the observed over-expressed protein was indeed AdhE, a
serni-quantitative analysis was carried out on AdhE fractions obtained fiom a partial
purification of the untagged AdhE using gel filtration. The objective was to determine
the correlation between the band at 97 kDa and the alcohol dehydrogenase activity. The
fractions were collected such that they spanned the peak corresponding to the band at 97
kDa (as determined by SDS-PAGE). Crude extract from overexpressing cells was loaded
onto a Superdex 75 gel filtration column, and fractions were collected for analysis by the
alcohol dehydrogenase assay and by SDS-PAGE. Each fraction was assayed for alcohol
dehydrogenase and the measured activity for each fraction was then compared to the
corresponding height of the A230 peak. In a typical assay for each fraction, all
components shown in Table 111.4 (previously degassed and kept inside an anaerobic
chamber) were mixed, except the sample from the fraction to be assayed. To a solution
containing 300 mM K2C03 (pH 10), 0.25 mM NAD+ and 170 mM ethanol and an
appropriate volume of purified (MilliQ) water, was added enough fraction volume to
bring the total volume to lmL. After rapid mixing, the rate of NADH production was
measured spectrophotometrically by monitoring the rate of increase of absorbanve at 340
nm. The rate of NADH production was measured as a function of both the fraction

volume and the substrate ethanol.

Table 111. 4. Typical composition of mixes used for ADH activity of AdhE

 

17 M EtOH (pl) SOOmM ch0.. pH 25 mMNAD+(pl) Fraction (pl) H20
10(111)

 

10 600 10 30 350

 

 

 

 

 

110

The results (Figure 111.16) showed a direct correlation between the alcohol

dehydrogenase activity (nmolNADH/min) and the intensity of the band at 97 kDa.

111. 2. 7. 5. Characterization of the [lists-tagged AdhE

111. 2. 7. 5. 1 . Identification of AdhE by mass spectrometry

Hisb-AdhE, purified as outlined in section 111.362, was concentrated using an
Amiconmconcentrator equipped with a 30 kDa molecular weight cut off membrane from
Milliporel". The concentration of the protein was measured by using the Bradford
method, using bovine serum albumin (Sigma) as a standard and a commercial protein dye
from BioRadl”. The protein was dialyzed into water (previously adjusted to pH 7.2) to
remove all elements of buffer and salts. A sample of the protein (in water) was then
submitted to the Macromolecular Facility in the biochemistry department at Michigan
State University for analysis by matrix assisted laser desorption ionization mass
spectrometry (MS-MALDI). The molecular ion of 97 kDa was positively identified with

greater than 90% confidence to be AdhE.

111.2. 7. 5.2. Alcohol dehydrogenase assay
The alcohol dehydrogenase activity of AdhE is the reverse of the reaction
represented by Scheme 111.7. The activity was measured by monitoring the rate of

change of NADH absorbance at 340 nm. The assay was carried out as described in

section 111.2.7.4.

lll

AdhE +
CH3CHO + NADH + H+ = NAD + CH3CH2OH

Scheme 111.7. The alcohol dehydrogenase/acetaldehyde reductase reaction
catalyzed by AdhE.

First the concentration of AdhE was held constant while that of EtOH was varied and
vice versa (Table 111.5) In either case, the species under study was added last. All
components were first degassed and kept inside an anaerobic chamber before mixing
them in a 1mL quartz cuvette equipped with a screw-on air tight rubber septum. The
component under investigation (AdhE or EtOH) was added last, using a Hamilton
syringe. After two quick inversions to mix the components, the automated kinetic
program was immediately started to monitor absorbance at 340 nm using a Model 8453

Hewlett Parckard spectrophotometer.

Table [11.5 Typical composition of mixes for alcohol dehydrogenase activity assay

 

 

17 M 131011 (pl) 500mM cho, (pl) 25 mM NAD" 98 21M AdhE H20

10 600 10 o 380
10 600 10 10 370
10 600 10 20 360
5 600 10 40 345
10 600 10 40 340
15 600 10 40 335

 

 

 

 

 

111.2. 7.5.3. Acetaldehyde reductase assay

This activiy is responsible for the forward reaction of Scheme 111.7 as written. In
this case, the activity was measured as the rate of NADH consumption (as measured by
the rate of decrease of A340). Similarly to the case of the alcohol dehydrogenase activity,
the rate was measured with respect to the concentration of each of the substrate

(acetaldehyde) and the enzyme (AdhE) in turn. Table 111.6 shows a typical mix reaction
112

mix used for this assay. To a cuvette containing 50 mM Tris-HCl (pH 7.6), 0.256pM
NADH, 100mM acetaldehyde and an appropriate volume of purified water (MilliQ) was
added varying amounts of AdhE, ranging from 0 to approximately 4pM. The total
volume of the assay mix was always fixed at 1mL. After quick mixing, the rate of
consumption of NADH was measured by automatic measurement of the rate of decrease
of absorbance at 340nm. Measurements were also made with the concentration of AdhE

held constant and varying that of acetaldehyde.

Table III. 6.Typical composition of mixes used for acetaldehyde assay

 

 

1M Tris-HCl, pH 7.6 (pl) 98 pM AdhE (pl) 10M CH3CHO 256pM NADH H20
50 0 10 10 930
50 20 10 10 910
50 40 10 10 890
50 30 20 10 890
50 3O 30 10 880
50 30 40 10 870

 

 

 

 

 

111. 2. 7.5.4. Iran analysis

In order to calculate the mole ratio of iron in AdhE, fractions were collected
during elution fiom the Ni-NTA column. After washing the loaded lysate with
approximately 50mL of wash buffer, the samples were eluted with elution buffer
containing 250 mM imidazole. Each fraction was analysed for AdhE and Fe, in order to
determine if there was correlation between Fe and AdhE. Protein was determined using
the Bradford method as described in Section 11142.2 Fe was determined
spectrophotometrically using bathophenanthroline as the complexing agent. The basic
procedure involved preparation of a series of standards using a commercial solution of

Fe2+ dissolved in sulfuric acid. The volume of standard used ranged fi'om 0 to 200 pl.

113

The volume was then made up to 1000 pl with M0 water. 500 pl of a solution
comprising 4.5% w/v Kmn04 and 12N HCl was then added to each tube. The purpose of
this step is to denature the protein and oxidize iron to Fe3+ in order to facicilate its release
from the protein matrix. After heating for 2 hours at 65°C, 100 pl of a solution consisting
of 9.7 g ammonium acetate, 8.8 g ascorbic acid, 80 g of each of neocuprine and ferrozine,
all dissolved in a total of 45 mL of filtered (MilliQ) water. After standing for 30 minutes
to allow for complete complexation, the absorbance of the standards and samples were
read at 562 nm. A calibration curve was plotted and used to calculate the concentration
of Fe in the samples.

111. 2. 7.5.5. Calculation of activity from kinetic data

The kinetics program used to measure the rate of change in NADH absorbance as
a function of active PFL measures the change in absorbance per second. However, the
program also makes provision for multiplication of the measured rate by any constant and
the choice of units to be used for the output. Most of the PFL and AdhE activities are
measured based on the rate of consumption of NAD+, and are reported in units of nano
moles of NADH produced per minute. We therefore conformed to this norm and took
advantage of the program to command the computer to convert absorbance units/s to
nmolNADH/min. We accomplished this by converting absorbance units to nmoles of
NADH though the Beer-Lambert law, A =8bC, wheres, b and c are the molar absorptivity
of NADH, the cuvette path length and concentration of NADH, respectively. Using the
cuvette path length of lcm and the molar absorptivity of 6270 M'lcm'l, coversion factor
of 9569.4 was calculated and used calibrate the program to report the rates in

nmolNADH/min.
1 14

111.3. RESULTS
111. 3. 1. Over-expression and Characterization of Untagged AdhE

111.3. 1.1. Overexpression of untagged AdhE

A typical whole cell SDS-PAGE analytical gel for Tuner (DE3) pLacI-
adhE is shown in Figure 111.15. Expression of untagged AdhE was suppressed by
addition of IPTG and therefore all growths were done without induction.
Comparison of Figures 111.15 to Figure 111.13 also reveals that the crude lysate of
BL21 (DE3)plysS-PET21a+-adhE cells was cleaner than that of Tuner(DE3)Lacl-
PetBlue-adhE cells. Consequenlty, it was easier to purify the tagged AdhE to
homogeneity than it was the untagged protein (compare Figures 111.14 and 111.16).

Std PETBlue-adhE whole cells
200——-
116. .. a

97"—
66—.

   

'AdhE

Figure 111.13. Results from whole cell SDS-PAGE of Tuner (DE3) pLacI-adhE cells grown to
stationary phase in LB/Amp. The cells were grown to an ODwo of ~2.5 and then a cell pellet was
prepared by centrifuging 1 mL of cell suspension in a 1.5 mL micro tube. The pellet was
suspended in equal volumes of SDS denaturing buffer containing 3.5% 2-mercaptoethanol and
water to make a total volume of 60p]. The cell suspension was then boiled for 10 minutes and
then centrifuged at 12,000 rpm to separate out particulate matter from the mixture of denatured
protein. 10p] of supernatant containing denature dissolved proteins was loaded into a commercial
(BioRadTM) 4-15% Tris-Cl polyacrylamide gel.

115

[11.3.1.2. Characterization of untagged AdhE

1113.12. 1. Correspondence of untagged AdhE activity to the 97 kDa band

The results presented in Figure 111.16 demonstrate a strong correlation between
AdhE activity and the protein band at approximately 97 kDa, thus proving that we have

successfully overproduced native AdhE.

 

 

   

 

 

.540:J
a 35 . .11 14
a 301 .10 .
E
E 25 .8
5 20 .
g 151 .7 .18
g 101
g 51 .6
3 0 l ..4 . T ‘ I .27]
0 5 10 15 20 25 30
Fractionnumber
97 1.. .f"i"““‘*"‘"""’ "" ~
22 man-i
45 h. :3 11% ’ i
33 1‘ . _. ..8 a d

 

 

Std4678101114l827|

 

Figure 111.14. AdhE lacking a His6 tag was partially purified by gel filtration
using an AP-5 Superdex 75 gel filtration column (5cm x 60 cm). Fractions were
assayed for alcohol dehydrogenase activity. The results show a direct correlation
between the intensity of the band at 97 kDa and alcohol dehydrogenase activity.
This provided additional evidence that the overexpressed protein is AdhE.

116

111.3.2. Over-expression and Purification of His6- AdhE

111.3.2.]. Test growths

Based on DNA analysis (Figure 111.2) four small growths from different colonies
were of cells transformed with PET21a1-adhE were grown in LB-Amp medium. The
results (Figure 111.13) were consistent with the DNA fingerprints, showing good
overexpression of the Hisé-AdhE. As shown in Figure 111.15, colony number 3 showed
no over-expression in either the induced or uninduced growth. In contrast, colonies 5, 6
and 10 show a dramatic difference between the induced and uninduced growths.

Std 3 3i 5 5i 6 6i 1010i
F“ ~ '5’” "1.9-r ‘-
200 a '

116.

66%.

 

Figure 111.15. Sodium dodecyl sulfate polyacrylamide gel electrophoresis
(SDS-PAGE) of cell pellets of colonies 3, 5, 6 and 10 from a plate of
BL21(DE3)pLysS transformed with PET21a("-adhE encoding Hisé—tagged
AdhE. The colonies were inoculated and grown in duplicates whereby X and X,
denote uniduced and induced colony X respectively. Whereas colony 3 did not
show any AdhE overexpression both in the absence and in the presence of IPTG,
colonies 5, 6 and 10 showed a marked response to IPTG, over-expressing AdhE
almost exclusively when IPTG was added.

Whereas the uninduced growths for these colonies are indistinguishable from colony 3,
the induced growths show a very intense band at molecular weight of ~97 kDa,
corresponding to the molecular weight expected for AdhE tagged with six histidine

residues. Notably, the intensity of the band appearing below 45 kDa was observed to be

117

lower in the induced growths than in the uninduced ones for each of the colonies. This
was consistent with the expectation that upon induction with IPTG, the energy of the
cells will almost exclusively be directed towards the production of AdhE. The uninduced
cell suspension fiom colony 10 was used to prepare glycerol stocks by mixing equal
volumes of autoclaved glycerol and the cell suspension in capped micro tubes. The 1:1
glycerol stocks were kept at -80°C and were used to grow BL21(DE3)pLysS-PET—21“(+)-
adhE bacterial colonies on ampicillin-impregnated (50 ug/mL) agar plates whenever
needed. Large-scale growths fi'om these glycerol stocks showed similar levels of

overexpression as the test growths illustrated in Figure 111.13.

111.3.2.2. Purification

As outlined in the previous section, the Hisé-tagged AdhE was purified using a
Ni-NTA affinity column. The crude extract was obtained by lysing the cells in phosphate
buffer containing imidazole, lysozyme, DNAse, RNAse and PMSF. The crude extract
was then loaded onto the column and then washed with low imidazole buffer, followed
by elution with the same buffer, but containing a higher concentration of imidazole. The
flow-through, the wash and the elution fractions were all collected and analyzed for
protein using SDS-PAGE. Figure 111.14 shows gels of fractions collected during

purification using the Ni-NTA affinity column.

118

 

Figure 111.16. Gels showing the level of purity achieved for Risa-tagged AdhE
by using a Ni-NTA affinity column to purify the protein. For gel A lane 1
contains the standard molecular weight marker, lane 2 contains the flow through
from loading the Ni-NTA column and the subsequent lanes contain washes l to
11. Lanes 14 and 15 were not used. For gel B, lane 1 was not used. Lanes 2 and
3 contain the first eluted fraction the standard molecular weight marker
respectively. The rest contain elution fractions 2 to 13.

Hiss-tagged AdhE showed very specific affinity for the Ni-NTA column and was
therefore purified to homogeneity. Virtually all impurities were removed by washing
with buffer containing 20 mM Imidazole (Figure 111.14). Typical yields of purified AdhE

were between 50 and 60 mg of purified protein per liter of bacterial culture.

111. 3. 2. 3. Alcohol Dehydrogenase Activities

Alcohol dehydrogenase activity was measured as described in sections III.2.7.4
and III.2.7.5. In one experiment, the concentration of the substrate (ethanol) was held
constant and that of AdhE varied. The activity was measured by observing the increase
in NADH absorbance at 340 nm. In the second experiment, the concentration of AdhE
was held constant and the increase in NADH absorbance at 340 nm observed as a
function of the change in the amount of ethanol. The total volume of the reaction mix

119

was 1000 pl. A typical mix was as a shown in Table 111.4. The reactions were carried
out anaerobically using degassed reagents kept inside an anaerobic chamber. All
components except enzyme or substrate were mixed together into a quartz cuvette inside
the anaerobic chamber. The cuvette was then sealed with a screw-on rubber stopper such
that it was airtight. The appropriate volume of substrate or enzyme was then transferred
into and held inside a Hamilton syringe. Both the syringe and the cuvette were then
brought out of the chamber and the cuvette placed into the UV-vis instrument. The
enzyme (or substrate) was then released into the reaction mix through the rubber stopper
and quickly mixed by inverting the cuvette twice. For each of the forward and reverse
reactions (Scheme 111.7), the reaction was first carried out with the enzyme held constant
and then with the substrate held constant. The results of the assays are shown in Figures

111.17 (reverse reaction) and 111.18 (forward reaction).

120

140 —

120 l
5 .
§ 100 1 A mg EtOH, AdhE
g 80 ~ constant
§ 60 « I mg AdhE, EtOH
g constant
C

 

 

0 T l l
0 20 40 60

mg EtOH, mg AdhE x 100

 

Figure III.17. Alcohol dehydrogenase catalyzed by AdhE. The rate of
production of NADH was measured as a fimction of the amount of AdhE or
EtOH, respectively, during the alcohol dehydrogenase reaction. The constant
[EtOH] assays were done at 7.7 mg/ml. The constant [AdhE] assays were done at
0.38 mg/ml. Reactions were monitored spectrophotometrically at 340 nm. The
data shown was obtained using Hisb-tagged AdhE. The untagged AdhE similarly
showed alcohol dehydrogenase activity.

121

N
(J1
__J

 

  
    
 

 

20 .
.E
E_ 15 y A mg AdhE, CH3CHO
E i constant
<2: ‘ I mg CH3CHO, AdhE
5 10 4 constant
E .
c

 

 

O 10 20 30 40
mg AdhEx100, mgCH3CHO

Figure 111.18. Acetaldehyde reductase activity of AdhE was measured by
monitoring the decrease in NADH absorbance at 340 nm. The graph shows the
rate of decrease of absorbance plotted as a function of the amount of AdhE or the
substrate acetaldehyde, respectively, during the acetaldehyde reductase reaction.
The constant [AdhE] assays were done at 0.29 mg/ml. The constant [CH3CHO]
assays were done at 7.7 mg/ml.

These results demonstrated that the Hiso-tagged AdhE indeed harbors the
expected activities. Based on the enzyme concentration of 0.3 uM (Figure 111.17) and the
activities observed, the calculated specific alcohol dehydrogenase activity for AdhE 63
nmolNADH/min/mgAdhE. The specific acetaldehyde reductase activity was calculated
as 26 nmolNADH/min/mgAdhE. This compares very favorably with values published in

the literature1 " 12'

122

111. 3.2.4. Iron Content of AdhE

The data presented in Figure 111.19 shows a correlation between Fe concentrations
and the AdhE peak. Each fraction was analyzed with respect to the concentration of both
AdhE and Fe. The concentration of each was plotted as a fimction of the volume of
eluent passed over the column. Figure 111.19 shows the plot of concentration as a
function of fraction number for both Fe and AdhE. The peaks for both Fe and AdhE
coincide with each other, suggesting that the Fe is associated with AdhE in a specific
fashion. The calculated Fe: AdhE mole ratio ranged from 0.25 to 1.25. The iron of the
purified enzyme averaged approximately 0.6 Fe/AdhE, again confirming the identity of

AdhE .

0.141
0.12 A

0.1 ~

  
 

0.08 i +AdhE x 2

0.06 J ' Fe

umollul

 

O l l l 1 l 1
0 5 1 0 1 5 20 25 30

Fraction number

 

Figure 111.19. Correlation between the amount of AdhE and Fe in successive
fiactions collected during purification of Hiso-tagged AdhE using a Ni-NTA
affinity column. .

123

The observations in Figure 111.19 are consistent with those previously reported in the
literature in which it was found that there is 1 Fe ion per molecule of AdhE.3 In our

hands, the FezAdhE of purified His-tagged AdhE ranges from ~ 0.2 to ~06.

[11. 3.2.5. Comparison of “as isolated ” and iron-stripped AdhE activities

A comparison of the specific activities of apo- and holo- enzyme was made by
calculating the change in the rate of production of NADH during the alcohol
dehydrogenase assay as a function of the concentration of enzyme used in the assay. The
holo- enzyme was the His6-tagged enzyme as-isolated from the Ni-NTA column, and
contained 0.6 moles Fe per mole of AdhE. The apo- form of the enzyme was generated
by incubating the “as-isolated” enzyme for 45 minutes in 1mM EDTA to complex out
Fe”. At the end of the incubation period, the protein was passed through a PD-lO gel
filtration column to separate the protein from the F e-EDTA complex. Iron content of the
apo-AdhE proved to be below the detection limit for the complexometric method.
Alcohol dehydrogenase activity was then measured as a function of protein volume. The
specific activity was calculated as the slope of the plot of rate of NADH production (in
nanomoles per minute) vs mg of protein in the assay. Figure 111.20 is a graphical
depiction of the results obtained. The slopes of the graphs give the specific activities of
the respective forms of AdhE, with the specific activity of holo-AdhE being 28
nmolNADH/min/mg compared to 12 nmolNADH/min/mg for apo-AdhE. The results
indicate that the specific activity of the as isolated enzyme was more than twice that of
the Fe-depleted form. This is consistent with the previous reports that the alcohol

dehydrogenase activity is dependent on the presence of Fey. 1

124

nmolNADH/min
o '5’ ‘5 8 93 8

120 — y =28.806x+ 9.1101
R2 = 0.9842

y = 12.062x- 7.9907

/ R2=0.913
1 r 1

l 2 3 4
mg AdhE

 

 

O

O Apo AdhE I Holo AdhE

Figure III.20. Comparison of specific activity of EDTA-treated (apo-) AdhE
and as-isolated (holo-) AdhE . A decrease in specific activity was observed when
AdhE was incubated in EDTA and gel filtered, suggesting that the Fe in the
enzyme is important for the alcohol dehydrogenase activity. The activity assays
were carried out aerobically.

125

111. 3.2.6. Effect of F e2 + on the alcohol dehydrogenase activity of AdhE

In order to further investigate the effect of Fe on the alcohol dehydrogenase
activity of AdhE, an approximately stoichiometric amount of F e2+ was added to each of
the EDTA-treated and the “as-isolated” AdhE. The alcohol dehydrogenase activity was
then determined as a function of protein concentration. For each of the EDTA-treated

and the “as-isolated” samples, enzyme assays were done on both before and after iron

 

 

 

addition.
250 1
' y = 49.638x + 20.457
._. 200 l .
§ 1
an: 150 1
< 1 y=28.013x+ 3.7917
30 100 l
E I — =
c 50 l y 11.7 x+36.7l8
0 41— y : 11.852x - 4.9463
0 1 2 3 4
mg AdhE

O Apo AdhE I Apo AdhE + Fe A Holo AdhE a": Holo AdhE + Fe

Figure 111.21. Effect of Fe2+ on the alcohol dehydrogenase activity of holo- and
apo-AdhE. Overall, holo-AdhE has a higher specific activity than apo-AdhE as
determined by the slopes of the linear fit of the data points. The calculated
specific activities are 12 nmolNADH/min/mg AdhE for both apo- and apo+Fe,
and 28 and 50 nmolNADH/min/mgAdhE for holo- and holo + Fe respectively.

The results in Figure 111.21 show that the “as-isolated” enzyme is more active
than the EDTA-treated enzyme, consistently with the results obtained in Figure 111.20

above. Interestingly, the specific activity of the Fe2+-supplemented and unsupplemented

126

apo-AdhE is essentially identical. On the other hand, alcohol dehydrogenase activity of
the Fe-supplemented “as-isolated” enzyme increased dramatically over that of the

unsupplemented enzyme.

The specific activities observed here are within the same order of magnitude as
those in the literature.'0 It is noteworthy that the activity of holo-AdhE without added iron
was found to have similar activity to the “as-isolated” AdhE in Figure 111.20. Our data
also shows that when Fe was removed from AdhE the specific activity was not restored
to that of the “as isolated” enzyme, remaining practically the same both with and without
iron supplementation. The data also suggests that once Fe is stripped out of the enzyme, it
becomes more difficult to re-incorporate back into the enzyme than into AdhE that
already has occupied sites. This apparent lack of sensitivity of the apo-AdhE to iron
supplementation may be a result of failure of the enzyme to reconstitute the iron active
site due to conformational changes resulting from EDTA treatment. For the “as isolated”
enzyme, however, there was a dramatic increase in the specific activity after incubating
with Fe2+. This suggested that the “as isolated” AdhE still had the capacity to incorporate
more iron, presumably into vacant active sites. That there could have been some vacant
F e-coordinating sites is plausible since the measured AdhE: Fe ratio for purified Hiso-
AdhE was determined to be ~0.6. If such incorporation of iron does indeed occur, it
would explain the dramatic increase in the specific activity of the enzyme. It would
therefore appear that the presence of some Fe in AdhE primes it for further incorporation
of Fe. It is also plausible, however, that the observed decrease response to EDTA by

AdhE results from denaturation of AdhE by EDTA.

127

111.3. 2. 7. Investigation of PFL deactivation

AdhE has also been reported to harbor PFL deactivase as a third activity in
addition to ADH and acetaldehyde reductase.3’ 4’ 12 One of the reports further asserts that
in order to carry out deactivation of PFL, AdhE required NAD+, Fe2+ and CoA as
cofactors.3 In this study the putative PF L deactivase activity was investigated by
preparing a series of reactions in which activated PFL was incubated in the presence of
different combinations of AdhE and cofactors. All the reactions were carried out inside
an anaerobic chamber. After incubating activated PF L with a given combination of
cofactors and AdhE, the residual activity of PFL was followed periodically using the
coupled enzymatic assay shown in Scheme 111.8, in which the PF L-catalyzed formation
of acetyl-CoA is coupled to the reduction of NAD+, which can be monitored
spectrophotometrically. The first set of PFL deactivation experiments was done using
AdhE and various combinations of the other purported cofactors. Although no
correlation could be seen between the rate of PFL deactivation and AdhE, there appeared

to be a correlation with CoA, as illustrated in Figure 111.22.

128

PFL activity (nmolNADH/min)

 

activated PFL
2‘1

A
‘7

k.

 

 

 

CH3
NAD+ NADH CoA
"13'3“ K j = oxaloacetate / = citrate
malic dehydrogenase citrate synthase

Scheme 111.8. The coupled enzymatic assay used to measure the activity

 

 

 

of PFL.
60 ~
so J,
‘ A
40 — .PFL only
0 + CoA
30 — l
. ‘+AdhE
A
20 7 . l-I-CoA + AdhE
10 1
y I.
0 + r r ,
0 5 10 15

Time (hours)

Figure 111.22. Some correlation was observed between CoA and PFL
deactivation

129

Following on this observation, the investigation of the cofactors was carried out
further to try to ascertain the effect, if any, of the cofactors by themselves and in the
presence of AdhE, on the activity of PFL. Table 111.7 shows an example of reaction
mixes used to carry out this investigation. The first set consisted of activated PFL
incubated with AdhE and all three putative cofactors. In the second set, active PFL was
incubated with the cofactors only and in the third set of reactions activated PF L was
incubated with “as isolated” AdhE only. The reactions were carried out under strictly

anaerobic conditions.

Table 111.7. The components of the mix used to investigate PFL deactivation

 

 

Reaction React 1 (ul) React 2 (ul) React 3 (ul)
component
1M Tris, pH 7.5 100 100 100
20 mM NAD+ 3O 3O 0
10 mM CoA 30 30 0
160 nM Fe“ 30 30 0
694 M PFL 100 100 100
100 M AdhE 10 0 10
H20 60 70 150

 

 

 

 

Representative results are shown in Figure 111.23. These results demonstrate
clearly that AdhE by itself does not deactivate PFL. Furthermore, the mixture of
cofactors by themselves steadily and systematically deactivated PFL over several hours
of incubation. The reaction mix containing AdhE in the presence of the cofactors showed
an identical effect to the one containing the cofactors alone. These results therefore
indicate that either one or a combination of the cofactors NAD+, Fe2+ and CoA

deactivates PFL. Also, the observation that when AdhE was introduced to the reaction

130

mix, exactly the same rate of PF L deactivation was observed proved that AdhE plays no

role in the deactivation process in this experiment.

 

 

90?

8O ‘ A
a l A ‘
E70 4 ga
i o
060 7 ..
< as OPFL+NAD+Fe+CoA+AdhE
E50 4 §3PFL+NAD+Fe+CoA
o APFL+AdhE
E40 4 e"!
330 .
-.-= 1
220 Q“

101I

Oi‘**‘i*i“%“‘lii“ii

O 2 4 6 8
Time(hours)

Figure 111.23. AdhE, the cofactors NAD1, Fe2+ and CoA and a combination of
both AdhE and the cofactors. All the reaction mixes were of a total volume of
360 111. The final PFL concentration was 193 1.1M. Where applicable the final
concentrations of each reactant were; AdhE: 3 11M, NAD+: 0.2 mM, CoA: 0.1
mM, Fe”: 1 mM, respectively.

The experiment illustrated in Figure 111.23 was repeated but this time the
amount of AdhE in the mix was doubled. The trend (Figure HI.24) was found to
be identical to that observed previously, again indicating that the observed
deactivation was due to the effect of at least one of the components on activated
PFL, and that AdhE was not that component. Moreover, it was yet again shown

that AdhE does not make any difference to the rate of deactivation of PFL.

131

160 7

 

 

.140 - an A
.E A
§120 —
g . O" OPFL+CoA+NAD+Fe+
<100 . N AdhE
% gaPFL + CoA + NAD + Fe
E 80 7 9,,
f? 5‘ APFL + AdhE
i 6° 1
4o 4
E 9.:
20 l “33
o . T a
0 2 4 6 8

Time (hours)

Figure III.24. A time course of PFL deactivation in reaction mixes containing
AdhE, the cofactors NADi, Fe”, CoA, and a combination of both AdhE and the
cofactors.

All the reaction mixes were identical to those in Figure 111.23, except that
the concentration of AdhE was doubled to 6 uM. Further experiments were
carried out in an attempt to isolate the cofactor(s) responsible for the observed
deactivation, and the mixes for this series of experiments were prepared as shown
in Table 111.8. The total concentrations of NAD+, CoA and Fe in the reaction
mixes were 3, 3 and 4 mM respectively. All the reaction mixes were incubated
inside an anaerobic chamber at a temperature of approximately 24°C, and the
residual activity was determined over time.

All of the experimental mixes containing various combinations of AdhE, CoA,
NAD+ and Fe” exhibited virtually identical behavior (Figure 111.25). All showed a

higher rate of PFL deactivation than the control and did so at essentially the same rate.

132

These results were interpreted as suggesting that at least two of the putative AdhE
cofactors are able to deactivate PFL, since the absence of all (PFL only) resulted in a
somewhat decreased rate of PFL deactivation, whereas systematic exclusion of one at a
time results in essentially identical rates of PF L deactivation, as when all components are

present.

Table 111.8. Compositions of the reaction mixes used to investigate PFL deactivation
under anaerobic conditions

 

 

 

 

 

 

 

 

 

 

 

Component React 1 React 2 React 3 React 4 React 5 React 6
(#1) (#1) (#1) (Ill) (#1) (#1)

1M Tris-C1 100 100 100 100 100 100
0.03 M NAD1 30 30 O 30 30 0
0.03 M CoA 30 30 30 0 30 0
0.4 M Fe” 30 30 30 30 0 0
55 M PFLa 100 100 100 100 100 100
139 M AdhE 20 0 20 20 20 0
H20 0 20 30 30 30 110

A 180 —

E 160 — ‘3 A

E 140 J 1 oCoA+Fe+NAD+AdhE

g 1201? 23 . lCoA+Fe+NAD

'2 100< ’ 1 . ACoA+Fe+AdhE

5 80 4 XFe+NAD+AdhE

5; 60 _ :3 ESCOA + NAD + AdhE

E 40 q ‘ . PFL only

a; 20 4

0 T 1 1
0.0 5.0 10.0 15.0
Time (hours)

Figure 111.25. Graph of PFL deactivation using various combinations of
cofactors and AdhE. The control showed relatively slowbut steady deactivation
over an extended period of time, making analysis of deactivation more difficult.

133

To fiirther pursue this question, more deactivation studies were done using more
cofactor combinations. Each cofactor was also tested individually in order to determine
which of the cofactors was sufficient to effect PF L deactivation by itself. The reaction
mixes investigated are listed in Table 111.9. All the components were added to Tris buffer
in a reaction vial inside an anaerobic chamber and were well mixed. Active PFL was
then quickly added to each of the vials before sealing the vials with screw-on cap
equipped with an airtight rubber septum. The time at which PFL was added to the
reaction mix was noted and recorded. The residual PFL activity in each mix was
monitored by periodically removing 5 ul of the reaction mix and placing it carefully on
the inside wall of a cuvette containing 995 pl of the PFL activity assay mix (see page
132, Scheme 111.8). The cuvette was then sealed with an air-tight screw-on rubber
septum, brought out of the chamber, quickly shaken to wash down the reaction mix
containing active PF L and immediately placed in the UV-vis instrument. The rate of
production of NADH was then monitored over a period of 60-120 seconds. The rate of
NADH production is directly correlated to the concentration of active PF L in the
reaction. The instrument is programmed to record the rate of NADH production in units
of Au/s. The conversion factor from Au/s to nmol/min was calculated fi'om the Beer-
Lambert Law using the cell path length of 1 cm and the molar absorptivity of NADH as

6270 M'lcm'l. The results obtained are presented in Figure 111.26.

134

Table 111.9. Composition of mixes used to isolate the species responsible for PFL
deactivation. The concentrations of the stock solutions used to prepare the mixes are

 

 

 

 

 

 

 

 

 

 

 

indicated.
Component React 1 React 2 React 3 React 4 React 5 React 6
(#1) (#1) 1131) (#1) (£1) 1131)
IM Tris-Cl 100 100 100 100 100 100
46 mM NAD+ 0 0 0 30 0 30
42 mM CoA 0 O O 0 0 30
170 mM Fe” 0 30 0 0 30 30
694 M PFLa 100 100 100 100 100 100
139 M AdhE 0 0 3o 30 30 30
H20 120 90 90 6O 60 0
A 180 4
c
E 160 4 f aPFL only
g 140 4 _ "
g 120 - ‘ ‘ ""x - PFL + Fe(ll)
a % '
g 100 A PFL + AdhE
. I. 80 _
3 60 ‘ A xPFL+AdhE+NAD
33, 4o 4 -
E 20 _ .4: .9 A PFL + AdhE + Fe(ll)
0 ' ‘ "C F -PFL+AdhE+CoA+
0.00 10.00 20.00 30.00 40.00 NAD + Fe

Time (hours)

Figure 111.26. The effect of various combinations of components used to isolate
the component responsible for PF L deactivation.

The data falls into two distinct groups. The first group comprises three reactions

showing an increase in the activity of PF L over time, whereas the other group shows a

steady decrease in PF L activity over the same period. The only common distinguishing

feature between these two groups is the presence or absence of F e(11). All of the reactions

135

that contain F e(II) also show systematic deactivation of PF L, while those containing no
Fe (11) show no deactivation. Notably, the reaction mix in which active PF L was
incubated with only F e(11) exhibited deactivation kinetics that are indistinguishable from
those containing F e(11) and other components such as AdhE. These results are consistent
with our previous findings suggesting that AdhE has no effect on PFL deactivation.
However, intriguingly, Fe (11) appears to deactivate PF L. In fact, the reactions without Fe
(11) showed an increase in PF L activity over the course of the experiment. The observed
increase in the activity of PF L in the mixes without Fe (11) suggested that PF L could be
activated fiirther by exposure to ambient light in the presence of traces of the initial
activation components. To test this hypothesis, activated PFL mixes containing different
combinations of the putative cofactors were prepared as given in Table 111.10, except
SAM and AE were excluded. Iron was included in reactions 2 and 3 but not in the other
three, which included the control. After incubating for approximately 18 hours, PF L-AE
and SAM (both of which are essential components for the activation of PF L), were added
as indicated in Table 10. The activity of PFL in each of the reaction mixes was then
monitored periodically using the coupled enzymatic assay. The experiment was canied
out under strictly anaerobic conditions. The results are obtained are summarized in

Figure 111.27.

136

Table 111.10 . Effect of F e(11) on Extent of PF L Activation

 

 

 

 

 

 

 

 

 

 

Component React 1(pl) React 2 (pl) React 3 (pl) React 4 (p1) React 5 (pl)
1M Tris-Cl 100 96 ' 100 100 100
46 mM NAD+ 0 30 3O 30 30
42 mM CoA 0 30 30 30 30
170 mM Fe” 0 30 30 0 0
694 pM PFL, 100 100 100 100 100
139 pM AdhE 0 30 0 30 0
H20 116 0 26 26 56
700 M AB 2 2 2 2 2
10 mM SAM 2 2 2 2 2

‘E 180 4

g 160 _1 .23 .PFL only

I ..

3 14° ‘ .x X‘ x I +AdhE. CoA,

Z 120 4 ..1 NAD, Fe(ll)

g 100 ”is A+CoA,NAD.Fe(II)

5 80 4 ‘

3‘ x +CoA,NAD.AdhE

'; 60 ~ I I

.3 .

8 4° ‘ ‘ A £3+COA.NAD

.1 I

IL 20 ”A

a O i T T Y t 1

0.0 2.0 4.0 6.0 8.0 10.0

Time (hours)

Figure 111.27. Effect of Fe (II) on the reactivation of PF L

These results showed a very interesting and consistent behaviour when compared

to Figure 111.26. As expected based on the results from Figure 111.26, the initial PFL

activities in the mixes containing Fe (11) were both low (~20 nmolNADH/min) compared

to those without Fe (II) (~100 nmolNADH/min). This was consistent with the previously

observed PF L deactivation by Fe (11) in Figure 111.26. Whereas the reactions without Fe

showed a steady increase in activity to approximately 160 nmolNADH/min, those to

137

which Fe was added showed marginal activation to ~45 nmolNADH/min in ~8 hours.
The monitoring of activity was not continued to saturation under the experimental
conditions. However, these results prove that Fe (11) does deactivate active PF L in a

systematic way.

138

111.4. DISCUSSION

The multiplicity of enzymatic functions harbored by AdhE is now commonly
accepted within the fields of biochemistry and bioinorganic chemistry. Specifically, this
enzyme is believed to be a triple function enzyme, mediating alcohol dehydrogenase,
acetyl-CoA reductase and pyruvate formate-lyase (PFL) deactivase. Indeed recently it
was reported that AdhE also exhibits antioxidant activity 9. Our lab has been involved
with the investigation of the interaction between the activating enzyme of PF L (PFL-AB)
with both the co-substrate S-adenosyl-L-methionine and the substrate PFL for some
time.14 One of the objectives has been to understand the nature of the interaction among
these three by employing various physical methods in order to elucidate the mechanism
by which AE achieves the generation of a glycyl radical on PFL.

In this study, our interest in AdhE was with regard to its purported PFL deactivase
activity. Taken together with the understanding of the chemistry of the activation process
mediated by PF L-AE, elucidation of the chemical mechanism of the deactivation process
catalyzed by AdhE would shed light on the bigger picture regarding the PF L activation-
deactivation system. An understanding of the interconversion of PFL between the radical
and non-radical forms could be seminal since there are several other enzymes that are
similar to PFL in that they bear a glycyl radical co-factor that is similarly generated by an
activating enzyme. PFL is up-regulated under anaerobic conditions when pyruvate has to
be turned over anaerobically. Activated PF L, however, is rapidly cleaved by exposure to
oxygen in vitro, suggesting that PFL is either inactivated or deactivated as conditions
become more aerobic. Under anaerobic conditions, AdhE is also expressed and used to

convert acetyl-CoA to ethanol and CoA by alcoholic fermentation. Notably, the acetyl-

139

CoA is a by-product of the PFL reaction. That AdhE should also function to deactivate
PFL was a very intriguing proposal as it suggested that AdhE somehow acts as both a
microaerophillic sensor and a deactivator of PF L, and thus protects it from oxygenolytic
degradation. The fact that AdhE is known to have a mononuclear iron (11) center led us
to hypothesize that it is probably this iron that provides the electron necessary to quench
the glycyl radical. We therefore set out to investigate the role of this iron in the PFL
deactivation reaction and to seek to understand the mechanism by which the deactivation
process proceeds.

To that end, the AdhE gene was successfully cloned onto a PET21(a)+ vector and
the protein was subsequently overexpressed in BL21 (DE3) pLysS and purified to
homogeneity using an Ni-NTA affinity chromatographic column. Enzymatic assays
confirmed the presence of both the alcohol dehydrogenase and acetyl-CoA reductase
activities in AdhE and this corroborated the positive identification of the enzyme using
gel electrophoresis. However, repeated attempts to reproduce the AdhE-catalyzed PFL
deactivation reported in the literature all led to negative results. Whereas it was easily
demonstrable that AdhE cloned and isolated in our laboratory harbors both of the other
known activities, no association could be made between PFL deactivation and the
presence of AdhE. Indeed the rate of PF L deactivation in the reactions containing AdhE
was consistently similar to that of the control experiment in which only buffer and PF L
were present. On the other hand, of the reactions containing the purported cofactors
NAD“: CoA and Fe”, those containing Fe” showed a consistent enhancement in
deactivation relative to the control. The presence of NAD+ alone showed no discernible

effect on the rate of PFL deactivation. Based on the observations made in this study,

140

Fe”, and to some measurable degree CoA, was found to be capable of deactivating PF L
either by itself or in the presence of other co-factors. The presence or absence of AdhE
did not have any effect in this deactivation, thus causing us to put to question its role in
the previously reported PF L deactivation, particularly considering that the reactions were
done in the presence of Fe”. Other experiments (data not shown) also showed that CoA
deactivates PFL at a significant rate relative to background deactivation observed in
activated PFL alone.

It is noteworthy that the deactivation observed in all of these reactions did not
show enzymatic behavior, occurring over a period of many hours. The deactivation
kinetics are consistent with behavior that one would expect from the quenching of the
PFL radical by a small non-enzymatic molecule. On that basis, the deactivation of PFL
in these reactions was attributed to the action of Fe” or CoA. However, it was clear that
this same explanation could not be immediately used to account for the deactivation
reported in the literature because in that case the deactivation was complete within 30
minutes,3 which is much faster than we have observed. We also noted that the
deactivation mixes used by Knappe and Kessler in all their experiments contained DTT
in large excess over PFL.3 Having tentatively ascribed the observed deactivation by CoA
to the thiol group, we postulated that DTT, also being a thiol, would likewise deactivate
PFL by quenching of the glycyl radical at Gly734. We further postulated that the
deactivation previously ascribed to AdhE was probably due to the combined
contributions of Fe” and DTT. The tentative implication of DTT was in part predicated
upon previous reports of glycyl radical quenching by thiols.l3 The effect of thiols on the
activity of PFL is further pursued in Chapter 4.

141

111.5. CONCLUSIONS

AdhE from Escherichia coli was successfully cloned, overexpressed and purified
to homogeneity. Analysis by sodium dodecyl sulfate polyacrylamide gel electrophoresis
(SDS-PAGE), enzymatic assays for alcohol dehydrogenase, acetaldehyde reductase,
DNA fingerprinting using restriction enzymes, partial sequencing of the adhE PCR gene
product, and MS-MALDI of the AdhE gene product were all used to confirm the identity
of the AdhE. Iron analysis using complexometric photometry corroborated previous
reports that AdhE harbors one Fe atom per molecule of the enzyme.

Several attempts to reproduce the reported PF L deactivase activity, which has been
reported as one of the activities harbored by AdhE, consistently showed that AdhE does
not have any role in PF L deactivation. On the other hand, the reactions containing Fe”
(and to a lesser extent CoA) always showed an increased rate of deactivation relative to
the control. However, this deactivation was clearly not enzymatic as it occurred over
several hours. These observations have therefore led us to conclude that AdhE is not a
PF L deactivase as previously reported, but that the deactivation that was observed was
likely due to one or any combination of Fe”, CoA and DTT, all of which constituted part

of the assay mix used in the deactivation studies.

142

REFERENCES

l.

2.

10.

11

12.

13.

Clark, DP. 1989, FEMS Microbiol. Lett. 63, 223-234.

Goodlove, RE. 1989, Gene 85, 209-214

Kessler, D and Knappe, J. 1992, J. Biol. Chem. 267, 25, 18073-18079.
Kessler, D.; Leibrecht, 1.; Knappe, J. 1991. FEBS Lett. 281, 59-63

Frey, M.; Rothe, M.; Wagner, A.F.V.; Knappe, J. 1994, J. Biol.Chem. 269, 17,
12432-12437.

Knappe, J .; Neugebauer, F.A.; Blachskowski, H.P.; Ganzler, M. 1984, Proc. Natl.
Acad. Sci. USA, 81, 1332-1335.

Wagner, A.F.V.; Frey, M.; Neugebauer, F.A.; Schafer, W.; Knappe, J. 1992,
Proc. Natl. Acad. Sci.U.S.A 89, 996-1000.

Unkrig, V.; Neugebauer, F.A.; Knappe, J. 1989, Eur. J. Biochem. 184, 723- 728.

Echave, F.; Tamarit, Jordi.; Cabiscol, E.; Ros, J. 2003, J. Biol. Chem. 278, 30193-
30198.

Holland-Stanley, C.A.; Lee, K.; Clark, D.P.; Cunningham, PR. 2000, J.Bacteriol.
182, 6049-6054.

. Membrillo-Hemandez, J.; Echave, P.; Cabiscol, E.; Tamarit, Jordi.; Ros, J .; Lin,

E.C.C. 2000. J. Biol. Chem. 275, 33869-33875.
Knappe, J.; Sawers, G. 1990. FEMS. Microbiol. Rev. 75, 383-398.

Armstrong, D.A.; Yu, D.; Rauk, A. 1998. J.Am. Chem. Soc. 120, 8848-8855.

143

CHAPTER IV

DEACTIVATION OF PFL BY SMALL MOLECULES

144

ABSTRACT

Several small molecules were studied with respect to their ability to deactivate
PFL. Based on preliminary findings indicating that the presence of CoA resulted in a
measurable and reproducible increase in the rate of PF L deactivation relative to the
control, we carried out a systematic investigation of several small molecules. Based on
the hypothesis that the thiol group was responsible for PF L deactivation, the initial set of
molecules studied were all thiols. All showed a similar effect as CoA, with the rate of
deactivation generally proving to be more rapid for the smaller molecules. Whereas
DTT, 2-mercaptoethanol and ethanethiol all deactivated PF L in a systematic manner,
they all did so at a markedly higher rate than CoA. Among these compounds, the rate of
PFL deactivation showed a correlation to the size of the thiol, with the smaller thiols
showing a generally higher rate of PFL glycyl radical quenching. In order to determine
whether the observed PFL deactivation by thiols was an inherent property of the thiol
functionality or whether thiols were exerting their property as reducing agents in general,
other biological reducing agents were also studied to determine their effect on active

PF L. All the reducing agents studied proved to be capable of deactivating PFL.

145

IV.1. INTRODUCTION

Upon observing that CoA deactivates PF L, we were intrigued to further
investigate the effect of thiols on the activity of PF L. Our hypothesis was that the glycyl
radical of activated PFL was quenched by the thiol group of CoA, as that was the most
plausible functional group by which CoA could supply the hydrogen atom required to
quench the glycyl radical. Thiols have been shown to quench glycyl radicals]. However,
in previous studies done on PFL, as indeed is the case with most other enzymes studied in
research laboratories, small molecule thiols were used routinely as components of the
buffers. Dithiothreitol (DTT), for example, is the most common thiolic reductant used in
enzymatic assays and is routinely used as a component of lysis buffers and buffers used
during protein purifications. Following the observation that CoA slowly inactivates PFL,
a number of thiols were systematically studied, with the aim to investigate the effect of
thiol size and structure on the rate of deactivation. Other small molecule reducing agents
relevant to biological systems were also studied in order to determine whether the
observed activity of the compounds containing a thiol group is specific to the thiol group,
or whether it can be extended to reducing agents in general. Because the glycyl radical of
active PFL is known to be buried within the structure of the protein, it was expected that
the rate of PF L deactivation would be sensitive to the size of the reducing species. This
would be so particularly if there was direct hydrogen transfer from the reductant to the
glycyl radical of active PFL. In this chapter, the effect of several small molecules on
active PFL was investigated systematically in order to shed some more light on the

properties of active PFL, particularly with regard to its stability.

146

1V.2. MATERIALS AND METHODS
IV. 2.1. Materials

All the chemicals used in these experiments were of the highest quality
commercially available. Dithiothreitol was obtained from American Bioanalytical, and
2-mercaptoethanol, ethanethiol, cysteine, homocysteine, glutathione and methionine were
obtained from Sigma. The enzymes and reagents used in the coupling assay were
purchased from Sigma. PFL was purified and activated as described in Chapter 2. The

structures of the compounds investigated in the study are shown in Figure 1V.1.

Ho
CH3CH28H \
Ethanethiol O//C—CH(NH2)CH2$H
cysteine
HO
CH2(SH)CH20H
C—CH NH CH CH SH
2-mercaptoethanol O// ( 2) 2 2
homocysteine
“‘1
HSCH2(OH)CHCH(OH)CHZSH //C—CH(NH2)CH26HZSCH3
DTT 0 methionine

Figure IV.1. Structures of the small molecules used in PFL deactivation
studies

147

l V. 2.2. The PFL deactivation experiment

For each deactivation carried out, the active PFL in 300 mM Tris-HCl buffer, pH
8.5, was incubated in an anaerobic chamber (Mbraun) with a known concentration of the
reagent under investigation. The residual PF L activity was determined periodically by
transferring 2 pl of the deactivation mix into 698 pl of the coupling reaction mix as
previously described in Chapter II. To avoid reaction while transporting the reaction mix
from the anaerobic chamber to the spectrophotometer (HP Model 8543), the 2 p1 of the
PFL deactivation mix was carefiilly deposited on the inside wall of the cuvette. Care was
taken to ensure that there was no contact between the deactivation mix and the assay mix.
The cuvette containing the activity mix was made airtight by sealing it with a screw-on
septum prior to bringing it out of the anaerobic chamber. The deactivation mix and the
assay mix were then quickly mixed by rapidly and vigorously shaking the cuvette so that
the 2 p1 of deactivation mix (containing residual active PFL) was washed into the bulk of
the assay mix. The cuvette was rapidly placed in the instrument and the residual PFL
activity was measured by monitoring the production of NADH as a fimction of time. The

activity assay is summarized in Scheme IV.1.

148

 

COA <- -.
m l
8. :
.4. f 1
a :
Q :
s \ a
1‘ H .
0%0' 5
V ;
O I
CoAS /
CH3

NAD’ NADH

\J, j:

oxaloacetate

malic dehydrogenase citrate synthase 7 citrate

 

Scheme IV.1. Enzymatic activity assay used to monitor residual activity

of PF L in the deactivation reactions.

As summarized in Scheme IV .1, in the presence of CoA, active PFL mediates
homolysis of pyruvate between Cl and C2 and transfer of the acetyl group to CoA to
produce acetyl-CoA. In the presence of oxaloacetate, citrate synthase then converts
acetyl-CoA back to CoA and in the process produces citrate. CoA produced in this step
is fed back to the pyruvate-PFL reaction and produces more acetyl-CoA. Meanwhile,
oxaloacetate is replenished by the conversion of malate (in the presence of NAD+) by
malic dehydrogenase. Provided NAD+ and malate are in excess, the concentration of
oxaloacetate is essentially constant. The rate of consumption of oxaloacetate (and
indirectly NAD”) under such conditions is directly proportional to the activity of PFL,
provided a that pyruvate is also present in excess. The rate at which the reactions in the
coupled assay proceed can therefore be measured by the rate of production of NADH.

This is the principle upon which the coupled enzymatic assay is based.

149

1V.3. RESULTS
IV. 3.1. Deactivation by DTT and ,B-ME

Based on the observation that CoA showed measurable deactivation of PFL, and
the hypothesis that this deactivation was due to thiol moiety of CoA, we decided to
investigate the effect of small molecule thiols on PFL activity. Since dithiothreitol (DTT)
was routinely used as a reductant in the lysis and purification buffers for PFL, it was one
of the first small molecules investigated. For comparison, 2-mercaptoethanol (B-ME)
was also tested. B-ME is structurally similar to DTT and is also a thiol. Like DTT, this
compound is routinely added to buffers used during protein purifications and handling.
With regard to PFL, it is important to establish whether the presence of B-ME in solution
had any effect on the lifetime of the glycyl radical. The experiment was carried out by
incubating activated PFL (0.14 mM) with 0.14 M B-ME and 0.10 M DTT and measuring
the residual PFL activity over time, using the enzymatic assay represented in Scheme
1V.l. The control consisted of replacement of the thiol with water followed by similar
treatment as the experimental. The preparation and the incubation of the deactivation
mixes were done inside an anaerobic chamber, with oxygen concentration maintained at
less than 2 ppm. Each deactivation mix was prepared by transferring 90 pL of activated
PFL into a reaction vial equipped with a cap bearing an air tight rubber septum. The
volume was made up to 100 pL with the thiol under study, or water in the case of the
control, mixed thoroughly, and then aliquots were taken periodically to check for residual
activity. Representative results are shown in Figure 1V.2. The activity of PF L in the
control remained constant over the entire duration of the experiment, whereas the activity
of PFL incubated with DTT showed a slight downward trend over the time of the

150

experiment, but was only marginally different from that observed for the control. B-ME,

on the other hand, completely deactivated PFL within 1 hour.

 

 

250 4
‘E
E 200 4 o O
I A g A. o
2 l ‘ PFL I
z 150- o my
3 [+0.14MBME
E
5 100 i A+0.10M on
Q .
.2 50 I
‘6
“ , I

O i l 1 J l

0 0.5 1 1.5

Time (hours)

Figure IV.2. The effect of DTT and B-ME on the activity of PFL as ,
determined by the enzymatic assay shown in Scheme IV .3.1. Conditions:
[PFL] =125 11M; [DTT] = 100 mM; [B-ME]=140mM; buffer is 300mM
Tris, pH 8.5.

The results in Figure IV.2 demonstrate that PFL by itself remained active and
stable over the entire period of the experiment. It is also evident that B-ME rapidly
deactivates PFL. Although the PFL activity in the presence of DTT showed a downward
trend, the activity was too close to that of the control to allow any concrete conclusion to
be made. In order to investigate this further, the deactivation experiment was repeated,
but carried out for a longer time. The concentration of DTT was maintained at 1mM as
in the first experiment, in order to make a direct comparison of the results obtained from
the two experiments. The results in Figure IV.2 illustrate the effect of DTT on PFL

151

activity over a longer period. The effect of DTT on the activity of PF L in the first hour in
Figure 1V.3 is identical to that shown in Figure 1V.2. After about 7.5 hours, the effect of
DTT became more apparent in that the PF L activity in the experimental mix was virtually
all gone, whereas the control showed activity at a similar level to those at the beginning
of the experiment. Remarkably whereas no PFL activity was detectable in the
experimental mix after 24 hours, the control retained 90% of the highest measured

activity within that period, thus attesting to the stability of the PFL radical.

200
180 I

140 t

120 ‘ o PFL
I PFL +1 mM DTT

a
I

 

 

 

 

Residue PFL activity
(nmolNADH/min)
N -i> O) 8 8

O
;.I

 

00

f 1” 1T fl

4 6 8 10
Time (hours)

 

O
O
N

Figure 1V.3. The effect of DTT on the activity of PFL over an extended period
of time. After seven hours of incubation, the control shows PFL activity at
similar levels as at the beginning of the experiment, whereas the activity of the
reaction to which DTT was added to a concentration of 1 mM showed virtually
no activity after the same period of time. Conditions were as in Figure 1V.2.

1 V. 3.2. Effects of Illumination on the deactivation of PFL

In some of our previous experiments it was observed that when the

deactivation was carried out under ambient light, the activity of PFL in the control
152

reaction steadily increased over time (see, for example, Figure 111.26). In the

current experiments similar observations were made (Figure 1V.3). The activity

of PF L increases both in the control and in the presence of DTT, although the

activity of the reaction containing DTT increased at a slower rate than the control.

In the presence of B-ME, the activity still decreased over time. Whereas B-ME

deactivated PF L rather rapidly (about 80% in 20 minutes), the activity in the

presence of DTT steadily increased similarly to, but at a slightly slower rate than

 

 

that of the control.
g 400 4
2 300 4 M
% 250 ‘ O ESPFLonIy
E 200 4 >13 . OPFL+1mMDTT
1; 150 — A APFL+1mMBME
- ‘f s,
§ 100 4‘1. 9
d 50 1 ‘
h 0 —1H 4 A 4 4
0 4 6

Time (hours)

Figure IV.4. Effect of DTT and B—ME on the activity of PFL when exposed to
ambient light. Conditions: [PFL] = 98 pM; [DTT] = 10 mM; 113-ME ] = lOmM;
buffer was 300 mM Tris, pH 8.5. The reactions also contained quantities of PFL-
AE (4.5 pM), SAM (0.7 mM) and deazariboflavin (10 pM) left over from the

activation of PFL.

The observation of steady increase in activity of the control (and to a lesser extent

the reaction containing 1 mM DTT), suggested that the PFL was not fully activated prior

to initiating the deactivation experiments, and thus, that PFL activation was competing

with deactivation.

The on-going activation was presumably occurring due to the

153

presence of small quantities of PF L—AE (2 pM), AdoMet (700 pM) and deazariboflavin
(140 pM), which were not removed from the activated PFL prior to initiating
deactivation studies. Since deazariboflavin is a photoreductant used as the electron donor
in activating PFL, it appeared that the ambient lighting was sufficient to drive further
PFL activation. This finding was surprising, given that previously PFL activations were
carried out using an intense 300W halogen lamp.2

In order to test our hypothesis that activation by ambient light was competing
with deactivation, the stock solution of previously activated PFL was incubated in
ambient light for about five hours and then used to carry out the deactivation studies.
The PFL thus treated exhibited significantly higher activity, and subsequent deactivation
experiments showed identical results, regardless of whether the experiment was carried
out in the dark or under ambient light (Figure 1V.5). No increase in control PFL activity
was observed even under ambient light, showing that after reactivation of PFL by
ambient light, PF L had been activated to the maximum extent. It is particularly
noteworthy that activated PFL proved exceedingly stable, maintaining its activity over an
extended period at average temperatures ranging from 22-23°C; this proved to be of great
experimental advantage for these studies. The data illustrate again that B-ME causes
rapid deactivation of active PF L, reducing the activity by ~80% within 6 seconds, while
for the DTT-treated sample no detectable decrease was observed until after ~2 hours of
incubation. Notably, the initial activity of PF L was more than 120% greater after
reactivation than before reactivation (~230 vs ~100 nmol NADH/min) and remained
virtually constant in the control over the six hours of incubation. PFL activity in the

reaction mix containing DTT decreased relatively slowly (by ~ 60% of the initial) during

154

the same period. This was consistent with previous observations that showed that at the
given relative concentration, DTT deactivates PF L rather slowly (hours). The activity in
the mix containing a similar concentration of B-ME on the other hand was completely

quenched within 25 minutes.

300

r 1

$250 5”

D O .

gzool ’ ,4, ”

B lPFLonly
E 150—

E . OPFL+1mM
g 1004 o .

*5

(U

.1

Li.

D.

 

 

O 2 4 6 8
Time (hours)

Figure IV.S. Deactivation of PFL fully activated by exposure to ambient light for about 6
hours before deactivating with 1 mM of each of B-ME and DTT. Experimental conditions
were as outlined under Figure IV.4.

155

IV. 3. 3. Deactivation of PFL by other thiols and small molecules

Other small molecules and thiols were also investigated in order to determine
whether the observed deactivation was particularly a thiol effect, or whether it was due to
some other more general property of these small molecules. Specifically, it was of
interest whether the observed deactivation by DTT and B-ME was a result of the presence
of the thiol functional group in these molecules, or whether it was their capacity as
reducing agents that conferred on them the capacity to deactivate PFL. Figure 1V.6
shows data obtained fi'om experiments conducted to compare DTT, B-ME and ethanethiol

(C2H5SH), all of which showed systematic PFL deactivation.

 

 

g 250
D
< 200 e ..
z a a
'6
g 150
S a o +1.4 mM BME
g 100 4 O 0+1.4mM DTT
8 A +1.4 mM 02115311
7' 50 4 '
a EPFL only
22 ° . . 0
a: 0 44 . W

0 5 10 15

Time (hours)

Figure 1V.6. Comparison of the PFL deactivation properties of ethanethiol, DTT, and [3-
ME. Ethanethiol showed similar deactivation kinetics to DTT and B-ME. Conditions:
[PFL] = 98 pM; [thiol] = as shown above; buffer is 300 mM Tris, pH 8.5.

156

IV. 3.4. Investigation of PFL deactivation by other small molecules

In order to gain a broader view of the nature of the reaction responsible for PF L
deactivation by the small molecules, we expanded the investigation to include
glutathione, methionine, cysteine and homocysteine. There were three basic reasons for
choosing this particular set of small molecules. Firstly, these molecules are very
common in biological systems and are therefore physiologically relevant molecules.
Methionine and cysteine are amino acids and are synthesized regularly by bacteria in
order to be incorporated into various types of polypeptides, including enzymes and
structural proteins. Cysteine is also credited with the ability to strengthen the antioxidant A
capability within the living cell, particularly by mitigating the harmful effects of heavy
metals.”4 Homocysteine is generally associated with the risk of heart disease5 in humans,
and is therefore of general biological interest. It was included because of its structural
and chemical similarity to cysteine. Glutathione is most commonly known for its
involvement in detoxification, transforming toxins such as heavy metals, solvents, and
pesticides into a form that can be excreted in urine or bile.6 Glutathione is also an
important antioxidant and is now known to protect against toxic chemicals in the brain,
hence an increased interest in it by researchers in Alzheimer’s disease.7

Secondly, we wanted to determine whether there was a correlation between the
size of the thiol and the rate of PFL deactivation. 1f deactivation required direct
interaction between the small molecule and the active site radical of PFL, then the rate of
deactivation would show a direct correlation with the size of the small molecule, such

that the smaller molecules would deactivate PFL more efficiently than the larger ones.

157

The third reason for choosing these molecules was to determine whether or not the
deactivation of PF L previously observed during investigation of Fe”, DTT, CoA and [3-
ME was due to the specific involvement of the thiol moiety, or whether it was due to the
general property of these molecules as reducing agents. If the deactivation was a specific
function of the thiol group, methionine would not deactivate PF L. Each compound was
prepared such that its effective concentration in the deactivation mix was 4 mM. The

results of this investigation are presented in Figure IV.4.7.

 

 

180 -
c .
g 160 4 d o "
E I‘ o o
z 9 £3 O+BME
g 120 d “a: o I+Cysteine
1: ll .
v A+Homoc ste ne
3‘ 100 4 I A . o a . “Y I
:; A . <> Methionine
r}: 80 “ ES .011"
d 60 4 I O +GSH
IL
_ $3 I aControl
g 40 ‘1 I."
.2
m 20 —
I? ; e
0 l . ’ t.
0 2 4 6

Time (hours)

Figure IV.7. Comparison of deactivation rates of various thiols and small
molecules. The concentration of each of the components was 4 mM. Conditions:
[PFL] = 370 pM; [small molecule] = 4 mM; [Deazaroboflavin] =10 pM; [AB] =

35 pM; Buffer is 300 mM Tris, pH 8.5.

158

Among the molecules shown in Figure 1V. 7, B-ME showed the most rapid
deactivation, achieving half of the original activity within two hours. DTT and
cysteine showed comparable rates of deactivation, decreasing the activity of PF L
by half within ~3.5 hours. Glutathione showed rather marginal deactivation
compared to the control, bringing the activity of PFL to half its maximum in
approximately 6 hour. Within the same period, the control retained ~65% of its
original activity. Methionine, the only molecule without a thiol group among
those investigated here, showed virtually identical kinetics to the control,
suggesting that it was not capable of deactivating PFL. The PFL activities

measured in the mixes containing each of the species afier ~ 6 hours are listed

 

 

Table IV. 1.
Table IV.1 Residual PFL activities of PFL in deactivation
mixes containing some small molecules
Species studied Residual PFL activity
Control 101
Homocysteine 88
Methionine 98
Glutathione 8 1
Cysteine 47
DTT 37
j-ME 0

 

 

Relative to the control, methionine proved virtually ineffective with regard to PF L
deactivation. Homocysteine and glutathione were comparable in their degree of
PFL deactivation in 6 hours. Cysteine proved to be relatively effective,
deactivating ~80% of PFL within the same period. As observed before, DTT and

B-ME deactivated PFL relatively rapidly, with B-ME doing so a lot faster.
159

IV. 3.5. Comparison of methionine and cysteine

Following on the results presented in Figure 1V.7, the deactivation
experiment was repeated using only methionine and cysteine, in order to ascertain
that indeed methionine, as hypothesized based on its lack of a thiol functional
group, does not deactivate PFL, whereas cysteine does. The concentration of
each amino acid was increased to 4 mM for ease of comparison. The results of
this investigation (Figure 1V.8) provide further and more convincing support for

the deactivation of PFL by cysteine.

100

80 I
70 . O PFL + Cysteine
50 4 I PFL + methionine
40 4 9 e I
30 -
20 —
1O - Q

0 1 . 1 .

A PFL only

 

 

PFL activity (nmolNADH/min)

—<
.4

Time (hours)

Figure IV.8. Comparison of methionine and cysteine as PF L deactivators.
Conditions: As in Figure IV .7.

Within the time frame of the experiment, methionine proved to be a much less
effective quencher of the glycyl radical of PF L than cysteine, consistent with our

previous observations.

160

IV. 3. 6. Other non-thiol reducing agents

Previous studies (Chapter 111) had shown that Fe (11) deactivates active
PFL in a systematic manner, and within a similar time frame as was observed
with some of the thiols investigated in this study, therefore suggesting that the
quenching of the glycyl radical in active PFL might be a function of the presence
of any species capable of providing reducing equivalents. Following that line of
reasoning, ascorbic acid and dithionite were the chosen candidates of study in
order to address this question. Figure 1V.9 gives the results obtained from the
investigation. Although the initial activity of the PFL used in this experiment was
lower than in the previous ones, the activity of the control was steady for the
entire 5 hours for which the experiment was run. At the end of the experiment, all
of the experimental runs showed markedly reduced PFL activities compared to
the control. The experimental reaction containing dithionite contained no
detectable PFL activity after the same incubation period as that containing
ascorbic acid.

Based on the hypothesis that the observed deactivation of PF L was due to
the participation of the thiol fimctionality in the small molecules studied, we had
predicted that ascorbic acid and dithionite might not deactivate PFL. That these
molecules also proved to be capable of deactivating PFL was interesting, and led
us to propose that the capacity to deactivate PFL, exhibited by the molecules
investigated in this study, was a property of reductants in general. This proposal
is supported by the observation that dithionite, a strong reducing agent without a
thiol group, actually deactivated PFL more ascorbic acid. This was also

161

consistent with the observation that Fe(II) also proved to be an effective

deactivator of PF L (Figure 111.25).

 

 

35
33301 A
g 25 A
§ 1 o J ‘ OPFL+asoorbicacid
-5 I PFL + dithionite
s 154 I ’
5 APFLonly
£10“ I
"3 5‘
fl
.1 0 . 4 I 1
LI.
0- 0 2 4 6

Time (hours)

Figure 1V.9. Effect of non-thiolic reductants on the activity of PFL. A 1 pL
aliquot of an 400 mM solution of each was transferred to a final solution of 90
pL of active PF L and the volume made up to 100 pL with ultra pure water.
Reaction conditions: As in Figures IV .7. and 1V.8.

I V. 3. 7. Deactivation of PFL by ,B-ME is reversible

A crucial question arising from all of the fore-going observations was whether the
observed loss of PFL activity was due to deactivation or inactivation. In other words,
was active PFL being converted to its native non-radical state by the quenching of the
radical at glycine 734, such that it could once again be activated under the right
conditions (‘deactivation”), or was the loss of activity associated with a permanent
chemical or conformational change that rendered PFL unactivatable (“inactivation”)?

To address this question, active PFL was reacted with B-ME as described

previously. Activated PFL was diluted 2x with ultra pure and degassed water that had
162

been kept inside an MBraunTM anaerobic chamber. To 990 pL of active PF L, 30 pL of a
1.4M B-ME stock solution was added. The solution was mixed well to ensure
homogeneity. A control was prepared by using water in the place of B-ME. The residual
PFL activity was then monitored as a function of time using the coupled enzymatic assay
shown in Scheme 1V . 1. The results in Figure 1V.10 show that after 3 hours the activity of
PF L in the control had increased slightly upon incubation in ambient light (~120 to 160
nmol NADH/min), the average temperature being 23°C. In the next 7 hours the activity
decreased by less than 10% (~160 to ~150 nmol NADH/min). During the same period (7
hours), the activity of PF L containing B-ME decreased by over 99%. These results were

consistent with those made previously.

At 10 hours, 500 pg of PFL-AE and 385 pg deazariboflavin were added to each
of the control and experimental samples (double arrow) and illuminated with a 300-watt
halogen lamp for 30 minutes. During illumination, both samples were kept in ice water
in order to keep the temperature of the reactions within the stability range of the proteins.
The activity of PF L in both samples was then monitored as before over a time range of

spanning 30 hours.

163

200 4
180 4
160 4,,,,,. .. )at‘ .
140 —== :3

O
120 W

 

1004 .
80~
501 e

40- Q
‘204 ,

Activity (nmolNADH/min)

O O

O TTT‘fifiiiiiliTlliilfljiIi]

0.0 10.0 20.0 30.0 40.0 50.0
Time (hours)

 

 

e PFL + BME a: PFL only

Figure 1V.10. PFL deactivation by B-ME. After 10 hours, only 12% of PFL
activity remained in the experimental sample, whereas the control sample
still had 93% of PFL activity remaining. Conditions: [PFL] =125 pM; [B-
ME]=140mM; buffer is 300mM Tris, pH 8.5. Also present from original PFL
activation: AB, 20 pM; SAM, 1 mM; deazariboflavin, 10 pM. At 10 hours
deazariboflavin was added to a final concentration of 20 pM. The reaction was
then illuminated for 30 minutes and PFL activity monitored over time.

Following addition of PF L-AE and deazariboflavin, and illuminating for 30
minutes, the activity of PF L in the “+B-ME” reaction increased more than 10-fold from

~15 to ~160 nmol NADH/min within 2 hours. The activity of PFL in the control

164

increased by only 19% from ~148 to ~176 nmol NADH/min and remained stable at about
the same level for the entire duration of the experiment. Notably, the maximum PFL
activity achieved after additional activation of PF L in the control reaction is essentially
the same as that attained afier reactivation of PF L in the experimental reaction.

Our interpretation of these results is that under the initial activation conditions,
PFL was activated to about 75% of the maximum possible (~130 nmolNADH/min). In
the control sample, additional activation to ~90% of maximal occurred due to ambient
lighting. This additional activation of previously activated PF L was also observed in
previous experiments. The addition of 41 mM of B-ME resulted in the deactivation of
almost 90% of the active PFL within 10 hours while essentially no deactivation was
observed in the absence of B-ME (control). Upon addition of the PF L activating enzyme
and deazariboflavin to both reactions, the deactivated PF L in the experimental reaction
was reactivated to slightly over its original level. The activity of PFL in the control,
however, had already been close to the maximum attainable under the conditions of the
experiment and therefore increased only slightly upon illumination.

Another way to interpret these results is that the addition of PFL-AB and
illumination simply activates PFL that had not been previously activated. However, in
this case we would expect to see additional activation in the control as well, unless the
observed activities are artificially limited by the experimental conditions. Using the
coupled PFL activity assay described previously under similar conditions, we have
consistently measured PFL activities that are well over 500 nmol NADH/min. The
possibility that the observed maximal activity measured in this experiment is artificially

limited by the experimental conditions can therefore be ruled out. So, had the dramatic

165

increase in PFL activity observed for the test reaction been due only to previously
unactivated PFL, the same increase in activity would also have been observed in the
control reaction, giving NADH production rates in the region of 350 nmol/min for the
control. Taken together, these results suggest that the increase in activity of PF L
observed in the experimental reaction resulted from the B-ME-deactivated PFL, and thus
we conclude that deactivation of PFL by B-ME is reversible.

It is also of interest to note that shortly after reaching the maximum activity of
~160 nmol NADH/min, the activity of the reactivated PF L begins to decrease relatively
rapidly in the reaction mix containing B-ME (Figure 1V.10). The kinetics of this second
deactivation phase are noticeably different from those of the first. Whereas the first
deactivation phase showed the typical decreasing rate of deactivation (concave curve),
the second phase shows a steadily increasing rate of deactivation. The significance of

this difference is not clear at present, but it is noteworthy.

IV. 3.8. Effect of the NADH/NAD+ Ratio on PFL Deactivation by ADHE

Previous studies have shown that overexpression of AdhE is directly correlated
with the NADH/NAD+ ratio.8 Since the NADH/NAD+ ratio is a function of the redox
status of the cell, it is expected that the expression of redox-sensitive or redox-responsive
enzymes will be correlated to this ratio, and therefore the observed correlation makes
sense. The ratio of NAD+ to NADH in E. coli is sensitive to the growth conditions. The
efficiency of reoxidation of NADH produced by glycolysis varies depending on whether

it is carried out by electron acceptor via the electron transport chain (aerobic growth), or

166

through fermentation (anaerobic growth). Previous studies have also shown that AdhE is
translated under both aerobic and anaerobic conditions, but that its translation under
aerobic conditions is considerably less than under anaerobic conditions. As a PF L
deactivase, AdhE would express this activity during the transition from anaerobic to
aerobic growth. However, during anaerobic growth the two enzymes coexist, as they are
both required for mixed acid fermentation. In order to change its role from alcohol
fermentation to PFL deactivation, and to prevent PFL deactivation under conditions
where active PFL is needed, AdhE would have to switch to the appropriate mode in
response to cellular signals.

This intriguing possibility was investigated by carrying out deactivation studies
aimed at determining the effect of the NADH/NAD" ratio on the putative PFL deactivase
activity of AdhE. Different ratios of NADH to NAD+ were studied and each of the
experimental runs was controlled by running a parallel reaction in which AdhE was
excluded. The concentrations of NADH used in the reactions were 0.0128 pmol/pL for
Reaction 1 and 0.128 pmol/pL for Reactions 2 and 3. For NAD+, the concentration of
the stock solution was 1 pmol/pL. The mixes used in the experiment are listed in Table
1V.2. All the mixes were prepared under strictly anaerobic conditions, inside an

anaerobic chamber.

167

Table IV.2. The mixes prepared to investigate whether the NADI-I/ NADJr ratio (x)
affects the putative PFL deactivase of AdhE.

 

 

 

 

NADH NAD+(1M) MOPS PFL. AdhE 3
(500 mM) (75m) (123 pM) f
Volumes transferred (pL)

Rxn 1 Mix 1 6.25 3.92 34.8 50 5 .02
Rm 2 Mix 2 5.2 3.33 36.5 50 5 .2
Rxn 3 Mix 3 13.4 2.29 29.3 50 5 .75
Ctrl 1 Mix 4 0 O 45 50 5 -
Ctrl 2 Mix 5 0 O 50 50 0 -
Ctrl 3 Mix 6 13.4 2.29 34.3 50 0 .75

 

 

 

 

 

 

 

 

' Activated PF L was the last component to be added to each of the mixes, the total time
needed for addition and mixing being ~ 10 seconds. The reaction vials (all covered with
air tight rubber septa) were incubated inside the chamber and the residual activity of PFL
was measured as a fimction of time, as described in the preceding sections. The results of

this experiment are shown in Figure 1V.11.

168

Residual PFL activity

Residual PFL activity (nmolNADH/min)

(nmolNADH/min)

700 —

600 3

500 —

400 1

300 4

200 ~

100 ~

 

'23

u NADH/NAD = 0.02
e NADH/NAD = 0.2
A NADH/NAD = 0.75
a PFL only

 

900 -
800 —
700 -
600 ~
500 ~
400 ~
300 —
200 1
100 1

0

 

Time (hours)

_d

O PFL only
IPFL + AdhE

APFL + NADH/NAD
(0.75)

A

 

0.0

l I 1

2.0 4.0 6.0
Time (hours)

V

8.0

10.0

Figure IV.11. Comparison of the effect of different NADH/NAD+ ratios and
AdhE on the activity of PFL. PanelA shows the effect in the presence of AdhE.

Panel B shows a set of controls as outlined in Table 1V.2. Conditions are as
outlined in Table IV.2

169

The data (Figure IV.11) shows that the rate of PF L deactivation is directly correlated with
the NADH/NAD+(A). All of the samples in Figure IV.11A, except the control, contained
AdhE, together with NAD+ and NADH in the indicated ratios. The data shows that the
PFL activity remaining in the mixes containing AdhE plus NADH and NAD+decreased
with increasing NADH/NAD+ ratios. From the data on 1V.10A, alone it was not possible
to determine which one of AdhE, NADH, or NAD+, or combination of the components
was responsible for the observed PFL deactivation. The answer to this question was
provided, in part, by the data from the controls (Figure 1V.11B). These showed that even
in the absence of AdhE, PFL activity decreased more rapidly in the reaction containing
the NADH/NAD+ mixture than in the control. Interestingly, it was also observed that in
the reaction containing PFL and AdhE only, the activity of PFL remained at about the
initial level for up to 5 hours after the beginning of the experiment, even though the
activity in the “PFL only” control decreased somewhat. After nine hours, the residual
PFL activity in the respective control mixes were; ~l6% for the reaction containing
NADH and NAD+, ~31% for the reaction containing PF L only and ~77% in the reaction
containing PFL and AdhE only. These results showed that in. the presence of NADH and
NAD+ the deactivation of PF L took place more rapidly. Intriguingly, active PFL
appeared to be more stable when AdhE was present. This finding suggested that AdhE

stabilizes active PFL. Another view of these results is provided in Figure 1V.12.

170

500 —
450 1
400 3
350 -
300 -
250 n
200 ~
150 -
100 4

50 -

Residual PFL activity after 9 hours

 

         

1 2 3 4 5 6
Reaction mix number

Figure 1V.12. The residual PFL activity in each of the reaction mixes (Table
IV .2) after 9 hours of incubation. Mix numbers 1, 2 and 3 correspond to
experimental mixes containing AdhE in the presence of NADH/NAD+ ratios of
0.02, 0.2 and 0.75, respectively, as given in Table IV.1. Mix numbers 4, 5 and 6
correspond to the controls reactions containing PFL+AdhE only, PFL alone and
PF L+NADH/NAD 1 (= 0.75), respectively, as shown in the same table.

The increase in PFL deactivation could not be conclusively correlated to the
increase in the ratio NADH/NAD+ per se, as the increase in the ratio was directly
proportional to an increase of NADH in solution. However, the results from controls 1, 2
and 3, respectively, show that AdhE appears to confer greater stability to PF L. The
residual PFL activity of the mix containing PFL and AdhE was approximately 3x higher
than that of PFL alone (Figure IV.12). The results from control 3 also showed that the
deactivations observed in reactions 1 to 3 were due to the effect of at least one of the
components of the NADH-NAD+ mix, but independent of AdhE. In order to investigate
their behaviour further, the reactions were left overnight inside the anaerobic chamber,

and then the activity of PFL in each mix was measured. Though the activity in all of the

171

mixes had decreased considerably, the relative magnitudes were similar to those
measured after nine hours of incubation (Figure 1V.13). To each of the reaction mixes, l
pL of each of SAM, PFL-AB and deazariboflavin was added and incubated inside the
anaerobic chamber to bring the total concentrations to 0.5, 0.02 and 0.1 mM respectively.
The PFL activity in each was measured after 4 and then again after 6 hours of incubation
in ambient light. The mixes were then lefi overnight for the second time. The activities
were again measured and reactivation repeated by adding SAM, PFL-AE and

deazariboflavin. The activity profiles of the mixes are summarized in Figure IV .13.

       
   

 

 

 

 

 

500 -
g 450 ‘ \ Iovemight 1
'5 400 -
5;? 350 - \ is: El ~4 hrs post
:3 8 5:": ' '
8 300 _ . § reactlvatlon1
—' 8 8 a: I ~6 hrs post
u. 250 r 8 .8 8 . .
n. 8 8 reactlvatlon1
200 8 ' 8 8
a, 8 8 . 8, I ovemlght 2
1. 8 8 8
a 150 4
w 8 8 ' 8
8 100 — 81.4248 8 . .... .~1'5 hrs DOSt
E 50 - 8 1’8 8 reactivation2
o - 8 8 8

 

 

1 2 3 4 5 6
Reaction mix number

Figure 1V.13. PF L activation-deactivation cycles in the reaction mixes.
The composition of the mixes is as indicated in Table 1V.10.

The first three reaction mixes show that the higher the proportion of NADH in the
mix, the greater the extent of deactivation, as shown by the activities measured afier both

overnight incubations. The data also shows that the degree of PFL reactivation decreased

172

when the NADH/NAD+ ratio was higher, as shown by both “reactivation l” and
“reactivation 2”. The data from reactions 4 and 5 (controls 1 and 2, respectively),
demonstrate that when AdhE is present, the rate of PFL deactivation is significantly
reduced. AdhE also appears to increase the degree of PFL reactivation as shown by
comparison of data from “post reactivation 2” for these reactions. Comparison of data
for reaction 3 (mix #3) and control 3 (mix #6) also supports this observation. Both of
these reactions contain NADH and NAD+ in a ratio of 0.75, but reaction 3 also contains
6 uM of AdhE. The residual PFL activity measured after reactivation and afler the second
overnight incubation for reaction 3 (+AdhE), were all higher than the corresponding
measurements for control 3 (-AdhE). This strongly suggested that AdhE plays a role in
both the enhancement of PFL activation and the stabilization of the PFL radical once it is

generated.

Chemical intuition, our experience with other reductants, and these observations,
together led to the hypothesis that the deactivation seen in these experiments was due to
the action of NADH, and that NAD+ does not play an active role in the process. We
proceeded to test this hypothesis by preparing the series of reaction mixes (Table 1V.3)
inside an anaerobic chamber and monitored their PFL activity over time. Reactions 1 and
4 together were used to test if NADH does indeed deactivate PF L, whereas 1 and 2 were
used to address the question about the effect of AdhE on the rate of such PFL

deactivation by NADH, as would be expected based on previous observations.

173

Table 1V.3. The reaction mixes used to investigate the effect of NADH
and AdhE during PFL deactivation

 

 

 

 

 

 

0.0128 M 500 mM 75 uM PFLa 123 uM
Mix # NADH MOPS AdhE
Volumes transferred (pl)
1 8 14.5 25 2.5
2 8 17 25 O
3 0 25 25 O
4 0 22.5 25 2.5

 

 

 

 

 

Reaction 3 was the control, in which only PFL and buffer were incubated. Reactions 3
and 4 together were use to investigate whether AdhE confers greater stability to active
PF L. The results of this investigation are shown in Figure 1V.14. These results confirm
that the deactivation observed in the previous experiments was indeed due to the action of
NADH, rather than the combined action of NADH and NAD+. This was demonstrated by
the fact that reactions 1 and 2 consistently showed lower activities than reactions 3 and 4.
Based on reactions 1 and 2, it is concluded that AdhE counteracts the deactivation of
active PFL by NADH, as demonstrated by the consistently higher activity of PFL in
reaction 2. The samples containing NADH (1 and 2) lost 290% of PFL activity within 7
hours. On the other hand, PFL activity in the samples without NADH was virtually
unchanged within same period (191 vs 180 nmol/min for reaction 3 and 183 vs 194
nmol/min for reaction 4). Whereas the PFL activity in mixes 1 and 2 decreased within
the first three hours, the activity increased in reactions 3 and 4 within the same period.
AdhE also appeared to preserve PF L activity in the absence of NADH, as shown by the
data for reactions 3 and 4 over time. The compositions of the two reaction mixes were

identical, except that AdhE was added to reaction mix 4 but not to reaction mix 3. At this

174

point the main observations made were that l) the amount of residual PFL activity at any

given time tended to be higher in the mix containing AdhE, and 2) the extent of PFL

reactivatability, tended to be higher in the mix containing AdhE.

450 —
400 -

A 350 - _

C

E 300 -

 

measured PFL activity
nmolNADH
..I. B N
8 o 8

 

 

   

. . .
. . .

. . .
g.;.;.
A'I‘I'
.-.-.'
-.:.;.
1.-.|l
n..."
»=-.v.
u'l‘u'
. .
.'2'.‘
l'l'l'
b'l‘lI
I'o'i.
.I‘I'I
. . .

V. .I.‘

1 2 3

reaction mix number

 

 

 

I ~25 minutes

~60 minutes

E] ~180 minutes

I~7 hours

I overnight

I ~2.5 hrs post
reactivation

Figure 1V.14. The effect of NADH and AdhE on the activity of PFL. The mixes
were prepared as in Table IV.2 and the residual PFL activiy measured at various

time intervals as indicated in the legend.

175

IV.4. DISCUSSION AND CONCLUSIONS

Several small molecule thiols were investigated with the aim to establish their role
in the deactivation of the active (radical) form of PFL. All of the thiols studied showed a
measurable deactivation of active PFL, with the capacity of such deactivation (measured
by the extent of deactivation within a given time), showing a general correlation to the
size of the thiol. In order to determine whether this capacity to deactivate PFL was a
function of the thiol functionality, or whether it was exhibited as a property of reducing
agents in general, other small molecule reductants were similarly investigated. The
results of these investigations revealed that non-thiol biological reducing agents also
deactivate PFL at a significant rate.

Investigation of the effect of the NADH/NAD+ ratio and NADH by itself revealed
that this ubiquitous biological reductant also deactivates PFL with fast kinetics relative to
the background deactivation observed in the control. The alcohol dehydrogenase
multifunction enzyme (AdhE), whose overexpression is known to be up-regulated
similarly to PFL during anaerobic growth, was found to stabilize active PF L. In the
presence of AdhE, the activity of PFL was greater than in the absence of AdhE. In
addition, deactivation of PFL by NADH was measurably slower in the presence of AdhE
than when AdhE was absent. Considering the findings of other investigators which
showed that PFL and AdhE expression both respond to an increase in the ratio
NADH/NAD+, this apparent protective role of AdhE is intriguing. It is when E. coli is
growing anaerobically that it deploys PFL for post-glycolytic catabolism of pyruvate.
PFL so deployed has to be in the active (radical) form in order to act on its substrate

pyruvate, in the presence of the co-substrate coenzyme A. Under the same conditions,

176

NADH is present at concentrations comparable to those of the other products of
glycolysis, having been a product of one of the steps in glycolysis. The observation that
NADH is effective at deactivating PFL suggests that under conditions in which both exist
simultaneously, a mechanism must be present by which active PFL is protected against
deactivation by NADH. Based on our observations herein, we propose that the
multifunction enzyme AdhE provides such protection.

Regarding the role of reducing agents in general, it is significant that the thiols,
particularly B-ME and DTT, have proven to be relatively potent deactivators of PFL.
Both of these reductants are very routinely used during enzyme purification in order to
provide the reducing conditions necessary to avoid denaturation of the proteins under
study. It would appear that it is necessary to only use minimal concentrations of each of
these reductants to prevent protein denaturation, without significantly counteracting the
activation of PF L What that concentration should be can only be established by
systematic study. Interestingly, DTT is required in small concentrations during PF L
activation. Determination of the optimum DTT concentration necessary to facilitate PFL
activation without compromising the stability of the same would provide valuable
information for any research involving PFL activation. Several theoretical studies on
protein radicals have demonstrated that the activation enthalpy for the transfer of a
hydrogen atom from a thiol to a C-based radical is in the region of ~12 kJ mol'1.9‘10 The
observation of deactivation by thiols is therefore consistent with predictions made from
theoretical studies.

We also find it fascinating that cysteine deactivates PFL, because there are two

cysteine residues (C418 and C419) at the active site of PFL. Given the high effective

177

concentration of cysteine in the vicinity of the PFL glycyl radical, and based on our
findings, it is conceivable that the glycyl radical would be rapidly deactivated shortly
after formation. However, this is not found to be the case. One of the active site
cysteines does, however, “quench” the glycyl radical during the PFL-catalyzed reaction,

as the glycyl radical is transferred to Cys419.11

178

REFERENCES

10.

11.

Rabek, J. F. 1987, Mechanisms of Photophysical Processes and
Photochemical Reactionsin Polymers. Theory and Applications, Chichester,
New York, Brisbane, Toronto, Singapore: Wiley.

Broderick, J.B.; Henshaw, T.F.; Cheek, J.; Wojtuszewski, K.; Smith, S.R.;
Trojan, M.R.; McGhan, R.M.; Kopf, A.; Kibbey, M.; Broderick, W.E. 2000.
Biochem. Biophys. Res. Commun. 269, 451 -456.

Droge W. 1993. Pharmacol. 46(2):61-5.

Quig D. 1998. Altern Med Rev. 3(4):262- 70.

Alfthan, G.; Aro, A.; Gey, KR 1997. Lancet 349: 397.

Dorval, J .; Hontela, A. 2003. Toxicol. Appl. Pharmacol. 192(2):191-200.

Li, Y.J; Oliveira, S.A.; Xu, P.; Martin, E.R.; Stenger, J.B.; Scherzer, C.R.;
Hauser, M.A.; Scott, W.K.; Small, G.W.; Nance, M.A.; Watts, R.L.; Hubble,
J.P.; Koller, W.C.; Pahwa, R.; Stem, M.B.; Hiner, B.C.; Jankovic. J.; Goetz,
C.G.; Mastaglia, F.; Middleton, L.T.; Roses. A.D.; Saunders, A.M.;
Schmechel, D.E.; Gullans, S.R.; Haines, J.L.; Gilbert, J.R.; Vance, J.M.;
Pericak-Vance, M.A. 2003. Hum Mol Genet. Dec 15: 12(24):3259-6 7.

Leonardo, L.R.; Dailly, Y.; Clark, DP. 1996, .1. Bacteriol, 178, 20, 6013-
6018.
Armstrong, D.A.; Yu, D.; Rauk, A. 1998. J.Am. Chem. Soc. 120, 8848-8855

Reid, D.L.; Shustov, D.A.; Armstrong, A.; Rauk, A.; Schuchmann, M.N.;
Akhlaq, M.S.; von Sonntag, C. Phys. Chem. Chem. Phys.

Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, J.; Schultz, S.; Wagner,
A.F.V. 1999, Nature: Struct. Biol. 6 (10), 969-975.

179

CHAPTER V

MOTIVATION FOR THIS STUDY, SUMMARY OF FINDINGS AND FURTHER
INVESTIGATIONS

180

V.1. MOTIVATION FOR THIS STUDY

At the onset of this research project, our aim was to study the Fe(II) alcohol
dehydrogenase fi'om E. coli (AdhE) with respect to its purported deactivase activity on
pyruvate formate-lyase. This would include the characterization of the Fe (II) center in
terms of the identity of the ligands directly coordinated to Fe (11), using a number of
techniques, including Mossbauer, XAS and EXAF S spectroscopies. The second part of
the project would then be to investigate the mechanism by which AdhE accomplishes the
deactivation of PFL. The activation of PF L, which involves the generation of a glycyl
radical on a glycine residue at position 734 of the PFL polypeptide, has received a great
deal of attention from numerous researchers."3

Based on a wealth of data collected over the years, the basic mechanism by which
the activation of PFL is accomplished, as well as the co-factors involved in the process
are now known.“ Indeed the knowledge base for the activation reaction is now at a point
where the focus is now on trying to understand how the co-factors and enzymes involved
in the activation interact with one another in order to achieve the generation of the radical
co-factor on the PF L backbone. The correct determination of the mechanism by which
the activation of PFL proceeds derives importance from the fact that many vital bacterial
enzymes are now known that are characterized by the generation of a radical co-factor
mediated by another enzyme, analogously to PFL-AB on PF L.7’8 Also, it is now known
that PFL-AE belongs to a super family of enzymes collectively known as the radical
SAM super family, all of which use SAM to generate a radical co-factor on another

enzyme or at another site on the same enzyme.9 It is therefore conceivable that the

mechanism of generation of the radical co-factors in these enzymes would be similar.

181

Whereas the activation of PF L and the components involved therein have
received considerable attention, the reverse reaction, which is purported to be mediated
by AdhE, has never been investigated further since the publication of 2 papers by one.
research group over a decade ago.””11 This is in spite of the fact that virtually every
publication on AdhE since around 1992 attributes its reference to AdhE as a PFL
deactivase to the same.”16 The fact that the PFL deactivase activity of AdhE was
reported by only one research group, and given the potential applicability of the
deactivation as a model reaction for at least some of the radical SAM enzymes, we sought
to better understand this fascinating reaction. However, before embarking on a study of
the metal center of AdhE and the mechanism of PFL deactivation by AdhE, we
considered it logical to first establish whether AdhE indeed contained deactivase activity,

hence the deactivation studies.

182

V.2. ACTIVATION AND STABILITY OF PFL
V.2. I. Some Philosophical and Practical Considerations on Activation

The classical challenge facing any researchers in the areas of biochemistry and
bioinorganic chemistry is the difficulty of creating experimental conditions in vitro, that
approximate those found within the living cells closely enough, such that the
observations made therein and conclusions drawn therefrom can be applicable to the real
life situation. The disparity between the conditions under which PF L has been activated
in all previous studies, and those that actually obtain within the cell, dictates that the
extrapolation of any findings made in vitro to what happens within the cell, be considered
with caution. Any modification of the experimental conditions that reduces the disparity
between the in vitro and in vivo situation with regard to the identity and nature of the
components involved, however, renders the conclusions drawn from the biochemical
study concerned more applicable to the real situation.

It is against such a background that the activation of PF L by illumination with a
high intensity halogen lamp has always been of concern to us. It known that in vivo, the
activation reaction employs flavodoxin/flavodoxin reductase as the reductant.
Understandably, this system does not require light, as many of the anaerobic
environments in which E. coli is found, such as in the stomach, are actually dark,
therefore light activation is artificial. Also, it is extremely difficult to maintain constant
temperature during activation while illuminating the sample with a high intensity lamp.
The addition of ice to the cooling water tends to cause a drastic drop in temperature

initially, followed by a steady increase between ice additions. The kinetics of PFL

183

activation under these conditions, therefore, can be difficult to control and to analyze, and

are certainly not comparable to those inside the cell.

V.2.2. Activation Without Intense Illumination is Advantageous

Our discovery that PFL can be activated without the need for high intensity
illumination enabled us to perform the activation at a constant temperature of
approximately 23°C. This temperature much more closely approximates temperatures of
the natural habitats of E. coli, relative to the wide range fiom < 10°C to ~23°C observed
during intense illumination experiments. Another advantage offered by activation
without illumination is that sample handling becomes greatly simplified, as it is no longer
necessary to try to control the temperature of the activation reaction by periodically
adding ice to the water bath. This also minimizes fluctuations of the oxygen
concentration inside the anaerobic chamber, which tended to be common when ice was
brought into the chamber, sometimes in spite of care having been taken to thoroughly

evacuate the ice prior to bringing it into the chamber.

V.2. 3. In vivo PFL Activation Employs A light-independent Reductant

As indicated, the (anaerobic) conditions under which E. coli requires the use of
active PFL tend to be dark, and yet E. coli is able to metabolize glucose. PF L-AE is able
to activate PFL under such conditions because the flavodoxin/flavodoxin reductase
system responsible for providing the required reducing equivalents is not light induced.
We therefore set out to attempt to carry out PFL activation under conditions that more

closely approximate those prevailing in an E. coli cell carrying out anaerobic metabolism.

184

As a first step, PF L was activated under ambient light and temperature. The activation
mix was prepared as described previously but, instead of illuminating with a high
intensity lamp and cooling the activation mix with ice-water, the mix was incubated
inside the anaerobic chamber maintained at ~23°C. All the light in the box was from a
fluorescent bulb inside the box and from those in those in the lab. As was shown in
Chapter II, PFL activity in the activation mix increased steadily up to a threshold level
comparable to (and sometimes higher than) that obtained when using the usual
illumination method. In effect, therefore, the variable that was made to approximate the
in vivo situation more closely was the temperature, but not light, since the photo-
reductant was still exposed to light, only at lower intensity.

A closer approximation of the situation in vivo for E. coli would be achieved by
using a biological reductant that is capable of generating active PFL under anaerobic
conditions, room temperature, and in the absence of light. Within the cells, such a
reductant would more closely model the flavodoxin/flaxodoxin reductase system
involved in the PFL/PFL-AE activation mechanism. Successful activation of PFL
without intense illumination, achieved in this study, was only a step towards this goal

because the photoreductant used was exposed to light, albeit at lower intensity.

185

V.3. STUDIES OF THE DEACTIVATION OF PFL BY ADHE

Kessler et al. have published work in which they indicate that AdhE is a
deactivase for PF L, namely, that it mediates enzymatic and non-destructive conversion of
PFL from the radical fi'om to the non-radical formm‘ll These workers therefore assert
that AdhE catalyzes the reverse reaction to that catalyzed by PFL-AB, thus completing an
interconversion cycle between active and inactive PFL. We found this report very
intriguing, as it proposed an elegant mechanism by which E. coli apparently avoids
oxidative destruction of PFL upon environmental change from anaerobic to aerobic. This
made sense, since by conservatively converting PF L from the active to the inactive form,
this bacterium would save a considerable amount of energy required synthesizing new
PFL when conditions again become anaerobic. According to these workers, AdhE
achieves the radical quenching in cooperation with NAD": Fe2+ and CoA as cofactors.

We therefore set out to first confirm these reported findings by repeating the
deactivation experiments under similar conditions as those described by Kessler et a1.11
Interestingly, our findings were at variance with the reported data. Based on our study,
the rate of deactivation of active PFL samples to which known amounts of DTT, F e2+ and
CoA (all components of the AdhE assays of Knappe and Kessler”) were added, was
found to be systematic and to proceed at rates significantly higher than that of the control
containing PFL alone. Surprisingly, AdhE was found to protect active PFL fiom
inactivation in the presence of a deactivator and to confer PFL with greater stability
compared to a negative control in which AdhE was excluded. From their study, Kessler
and co-workers noted that during the deactivation of PF L by AdhE, NAD+ was not
replaceable by non-reducible analogs, and on that basis they reasoned that NAD+

186

therefore directly participates as an oxidant.ll They also noted that there was a net
increase in the amount of NADH in solution during PF L deactivation and that addition of
NADH inhibited PF L deactivation.

From a chemical standpoint, it is not immediately clear how NADJr would be
involved in the quenching of the glycyl radical of active PFL. The quenching reaction
effectively involves the addition of a hydrogen atom at Gly-734 of active PFL, which is a
reduction reaction. It is therefore not conceivable that NAD+ would be able to achieve
this reaction, directly or indirectly. Their observation of an increase in NADH during
PFL deactivation may be explained by a side reaction between NAD+ and another
reductant in solution, particularly since the reaction was carried out under reducing
conditions. These workers also noted that the production of NADH was concomitant
with and correlated to the amount of PFL deactivated, and that the background amount of
NADH was essentially constant when inactive PF L was used. They therefore argued that
this provided further proof that NAD+ was being reduced in the process of PF L
deactivation. It should be noted, however, that the activated PFL used by these workers
contained several components, including reducing agents such as Fe2+ and DTT. Under
such reducing conditions, NAD+ could be reduced to NADH detected in the mix. The
more activated PFL transferred, therefore, the larger the amount of reducing equivalents
and so the amount of NADI reduced. Unactivated PF L, on the other hand, contained
much lower concentrations of the reducing agents found in the active PFL mix,
particularly F e2+.

As additional proof that NAD+ was directly involved as an oxidant, these workers

”ll

report that it was “not replaceable by non-reducible analogs . In the absence of the data

187

based upon which this conclusion was drawn, it is difficult to comment on what might
have led to the said observations. Whereas these workers report that NADH proved to be
a competitive inhibitor, our findings show that NADH deactivated PFL quite effectively,
relative to a negative control. This is also amenable with the chemical properties of both
NADH and active PFL. NADH is a ubiquitous biological reductant and readily reduces
biological oxidants and is converted to NAD+ in the process. The quenching of active
PFL effectively entails addition of a proton and an electron at the radical-bearing carbon
of glycine-734. The conversion of NADH to NAD+ can be visualized as
NADH —> H+ + 2e' + NAD”

Therefore, in NADH-mediated deactivation of PF L, the hydrogen atom would be
transferred to the radical center and the electron transferred to an oxidant in solution. We
presently do not have sufficient data to provide more details regarding the mechanism by
which this deactivation proceeds. However, in a subsequent section we propose some
experiments that can be carried out in order to preliminarily test if there is direct
interaction between active PFL and NADH.

It is also intriguing that even though Knappe and co-workers reported that CoA
was required for the deactivation of PFL, they also report that it did not undergo any
detectable chemical modification.11 However, they observed that CoA was probably
directly involved in the deactivation through the thiol group because the desulfo analog
of CoA competitively inhibited deactivation. Our findings are consistent with these

reported observations, except that we find PF L deactivation to be independent of AdhE.

188

V. 4. CONTEXT, RELEVANCE AND OTHER STUDIES

In order to develop a perspective of the role of AdhE within the context of the
metabolic machinery of the organism, particularly E. coli, a brief discussion of some
findings from the literature will first be presented. E. coli is able to survive and thrive
under very adverse conditions because it has evolved mechanisms to allow it to cope with
a wide range of environmental conditions. Such conditions may vary widely in terms of
factors such as pH and oxygen availability. The adaptability of this organism can be
traced to its ability to express its genes selectively depending on environmental
demands.l7 This organism is a facultative anaerobe and controls gene expression by using
a combination of transcription factors that respond to the overall oxygen status of the
growth environment.” Fumarate and nitrate reduction regulator (FNR) is one of the most
important transcription regulators responsible for turning on the expression of most of the
enzymes needed for anaerobic growth.17 The sensing mechanism of FNR under anaerobic
conditions involves the acquisition of an oxygen-labile [4Fe—48] cluster, which triggers
the dimerization of FNR, thereby enhancing site-specific DNA binding.18'20

The ple gene encoding PFL is FNR-regulated. Recently Green and co-workers
discovered that another FNR-regulated enzyme, YfiD, bears very close sequence
homology (77% homology over a 64 amino acid overlap) to the C-terminal region of
pyruvate formate-lyase (PF L).21 Based on its strong homology to pyruvate formate-lyase
(PFL), these workers predicted that YfiD would also be a radical-containing enzyme.
They not only confirmed that to be so, but also found that PFL-AB is able to introduce
the radical on YfiD, similarly to PFL. Based on the PF L deactivase activity of AdhE
reported in the literature, it was expected that AdhE would quench the radical generated

189

on YfiD during its activation by PFL-AE. All attempts to identify the reverse activity,
under the mediation of AdhE, however, gave negative results.21

The genes for aerobic metabolism are regulated by the Oz-responsive Arc system,
comprising two enzymes ArcA and Ach.”’23 Ach is a membrane protein and primarily
plays a sensory role, while ArcA is the response regulator. The expression of PFL and
AdhE under conditions ranging from aerobic through microaerobic to anaerobic have
been studied quite extensively. Many of these studies have ranged in detail from 0;-

24'26 to the genes targeted by those regulators,27 the

responsive transcriptional regulators,
promoter structures involved,28 and the mode by which the system senses oxygen.29

In their study of the effect of oxygen tension on the expression of anaerobic
enzymes and their metabolic products, Becker and co-workers found that AdhE
expression was more sensitive than PF L to oxygen tension.30 These workers defined
p005 as the oxygen tension required to bring down protein or metabolite synthesis to half
the maximum observed under anaerobic conditions. Using this operational definition,
they measured pOo_5 for the fermentation products of PF L and AdhE, as well as for the
enzymes themselves. They observed that the general trend was that the p005 for the
synthesis of the fermentation products acetate, ethanol, lactate and succinate, fell within
the narrow range from 0.2-0.4 mbar. The exception to this was formate, which was
found to have 3 p005 value of 1 mbar. What these findings suggested was that during the
change from aerobic to anaerobic metabolism, formate is formed earlier than the other
products. Also, for PFL and AdhE, p005 was measured to be 5 and 0.8 mbar 0;,
respectively, while the corresponding p005 for the formate and ethanol were 1 and 0.4
mbar 0;, In metabolic terms, these findings indicate that during the transition from

190

aerobiosis (high p02) to anaerobiosis (lower pOz), PF L is deployed earlier than AdhE,
hence the detection of its metabolic product, formate, earlier than the metabolic product
of AdhE, ethanol. The interpretation offered by these workers is that the earlier
deployment of PFL could be because it has to be post-translationally modified (activated)
prior to catalysis. The limitation of this interpretation, however, is that both the
activating enzyme (PF L-AE) and activated PFL are highly sensitive to oxygen in vito and
would therefore be expected to operate only at lower oxygen tensions, which is later in
the transition fiom aerobiosis to anaerobiosis.

Notably, deletion mutations at the ubiquinone synthase and quinoloxidase loci,
both of which are members of the electron transport chain, resulted in a shift of both
formate and ethanol production to higher p02 values, suggesting that both PFL and AdhE
become operational earlier during the shift from aerobic to anaerobic metabolism.16
Removal of some of the members of the electron transport chain is expected to lead to
accumulation of NADH, since the mechanism of disposal of reducing equivalents is no
longer available. Previous studies had shown that accumulation of NADH inhibits
pyruvate dehydrogenase (PDH). 31 It is therefore conceivable that under those conditions
pyruvate metabolism would shift towards PFL, since it is the other of two systems
available to metabolize pyruvate to acetyl-CoA. Notably, accumulation of NADH also
stimulates NADH reoxidation by alcoholic fermentation, a reaction that requires AdhE.32
Based on the fore-going analysis, it is therefore sensible that both PFL and AdhE and
their corresponding metabolic products would shift towards higher p02 values when the
electron transport chain enzymes are either removed or inactivated. However, this has to

be reconciled with the fact that PFL and its activating enzyme are very sensitive to

191

oxygen in vitro. Based on our observation that AdhE appears to confer greater stability
to active PFL, we propose that the observed shift of PFL and AdhE expression towards
more aerobic conditions when the electron transport activities are not present, is induced
by metabolic necessity and made possible by the protective role of AdhE for active PFL.
More recently, Teixeira de Mattos and co-workers investigated the metabolic
response of E. coli to varying oxygen availability and found that PFL was active at low
oxygen concentrations.32 Their measurements show that even under microaerobic
conditions, when respiration is at half its maximal rate, the processing of as much as half
of pyruvate is due to PFL, showing that both the pyruvate dehydrogenase system and the
PFL system are operational under these conditions. They further proposed that PF L is
stabilized by cytochrome bd through rapid consumption of oxygen via uncoupled
respiration.32 In their review, Poole and Hill note that cytochrome bd has been discovered
to be the terminal oxidase responsible for protecting oxygen-labile nitrogenase during
nitrogen fixation in Azotobacter vinelanciii.33 The discovery of this protective role of
cytochrome bd for nitrogenase removed the paradox of aerobic nitrogen fixation in
prokaryotes, a process mediated by a very air-sensitive enzyme. The implication of
cytochrome bd as a protector of PFL during microaerobic respiration was previously
suggested by Becker and co-workers, who found cytochrome bd to be maximally
expressed under microaerobic conditions, similarly to PF L.33 Since PFL is known to be
extremely sensitive to oxygen, and yet it is found to be active under conditions of
relatively high oxygen tension, this is strongly suggestive of an agent that protects it
against oxygenolytic destruction, a fate that it suffers almost instantly in vitro when
exposed to air.33 Our finding that AdhE appears to stabilize active PFL, and that of

192

Echave et al.”, who found AdhE to have some antioxidant properties has lead us to
suggest that the presence of AdhE allows PFL to be able to operate under
microaerophillic conditions.

The co-expression of AdhE and PFL, observed by several researchers,‘u4'33is
also consistent with the fact that both enzymes are sequentially involved in the processing
of pyruvate to ethanol under anaerobic growth. The suppression of pyruvate
dehydrogenase resulting from the accumulation of NADH would necessitate the
deployment of PFL, which would, in turn, require activation to convert it to the catalytic
state. Since the acetyl-CoA produced by the PFL reaction with pyruvate and CoA is a
substrate for AdhE, it also makes sense that AdhE should be co-expressed with PFL.
Furthermore, alcoholic fermentation mediated by AdhE consumes NADH, which
accumulates under microoxic conditions. Under this set of conditions, the
multifunctional role of AdhE as an alcohol dehydrogenase participating in the
fermentation reaction, while at the same time protecting active PFL against oxygen,
makes sense.

The PF L protective role proposed for AdhE is not only consistent with the
precedence of the apparent PF L protection by cytochrome bd mentioned above,32 but also
with the reported antioxidant properties of AdhE.12 The latter results showed that E. coli
cells with an AdhE knockout were not able to grow aerobically in minimum medium. It
was also noted that these cells were particularly prone to oxidative stress, but
reintroduction of the adhE gene into the mutants restored the cells’ ability to grow
aerobically.12 AdhE purified from the adhE+ cells was found to undergo metal catalyzed
oxidation (mco) very rapidly when exposed to H202 leading to inactivation of the

193

enzyme.12 Upon substitution of Zn for Fe in the AdhE, no metal catalyzed oxidation was

observed. The Zn-substituted AdhE also led to reduced cell viability compared to the Fe

AdhE. These observations led to the conclusion that AdhE was acting as an antioxidant

and that Fe2+ was essential for the antioxidant activity.

In conclusion, the following observations have been made in the studies of AdhE

and related enzymes carried out by a number of workers:

1.

Although it is generally known to be an anaerobically active enzyme,
AdhE is expressed under aerobic conditions as well.

AdhE and PFL are expressed to comparable levels during
microaerobiosis, under which conditions PFL is found to be active,
despite its high sensitivity to oxygen.

Cytochrome bd, a terminal oxidase found to be effective in protecting
nitrogenase against oxidative degradation, is also found to be present
in higher concentrations under microoxic growth conditions.

AdhE was discovered to protect growing bacterial cells against
hydrogen peroxide.

In our work, AdhE was found to confer greater stability to active PF L
and to protect it against deactivation.

In our work, NADH, which accumulates both under microoxic
conditions and in mutants devoid of terminal quinone oxidases, was

found to deactivate PFL.

These findings, among others (Chapter III) put to question the assertion that AdhE

is a PF L deactivase. Deactivase properties would not be consistent with the finding that

194

under microoxic conditions AdhE and PF L are co-expressed and that PF L is active under
the same conditions. This is inconsistent because
i) active PFL would either be destructively inactivated by oxygen or
ii) active PF L would be deactivated by AdhE
With what we now know of the effect of NADH on active PFL, its accumulation under
microaerobic growth would also contribute towards PFL deactivation and thus render
fermentative metabolism impossible. The case of YfiD is also instructive. Whereas this
enzyme shows 77% sequence homology to PF L, and whereas it is under the same regulon
as PFL and is also activated by PFL-AB by introduction of a glycyl radical, attempts to
deactivate it with AdhE proved ineffective.2| Based on our observations, and the findings
of others cited above, we propose that AdhE is co-expressed with PFL primarily for two
reasons:

1. Both enzymes are needed for anaerobic metabolism, PFL being responsible for
the first step consisting of non-redox conversion of pyruvate to formate and
acetyl-CoA, followed by a second step comprising AdhE-mediated reduction of
acetyl-CoA to ethanol, with concomitant recycling NAD+ to feed the glycolytic
cycle.

2. AdhE provides protection for the highly sensitive active PF L against destructive
inactivation by traces of oxygen or any biological reductant capable of

inactivating PFL.

195

V.5. DEACTIVATION BY THIOLS AND OTHER REDUCTANTS

The deactivation studies carried out in this project were done as an offshoot to the
test for PF L deactivation by AdhE. As it turned out, however, thiols proved to be
effective deactivators of PFL. This did not come as a great surprise per se, because one-
electron oxidation to produce thiyl radical is a well-established reaction.34 These radicals
participate in a number of reactions including electron transfer, hydrogen abstraction and
addition reactions with several biological constituents and xenobiotics, and can be
detected by optical spectroscopy or by electron spin resonance (ESR) spectroscopy.34
Bithiols are known to be particularly good antioxidants, providing cells with protection
against oxidative damage by free radicals.35 In the process, thiols undergo one-electron
oxidation and form thiyl radicals (R8). The efficacy of thiols as antioxidants stems from
the fact that under physiological conditions, thiyl radicals can react with thiolate anion,
yielding disulfide radical anion (RSSRf as an intermediate.” The disulfide radical is then
passed on to oxygen, yielding disulfide and superoxide radical anion (02'), which is then
inactivated in a reaction catalyzed by superoxide dismutase (SOD).35 Thiyl radicals are
also known to undergo reaction with reductants such as ascorbate, with concomitant
formation of low-activity ascorbyl radical, which subsequently undergoes a
disproportionation reaction.35 Thus biological systems have devised an efficacious means
of using thiols to neutralize potentially harmful radicals by converting them into forms
that can be disposed of enzymatically or species that are inherently self-destructive.

The deactivations that were observed in our work therefore demonstrated that
when carrying out investigations on active PFL, it should be taken into account that any

thiols present in the solution will have the potential to deactivate PFL. It is therefore

196

 

important to take this into consideration, so that control experiments can be designed to
filter out any background deactivation that may be taking place as a result of the
participation of the thiol(s). In their work on PFL, Kessler and co-workers used DTT in
the assay mix at concentrations that were several-fold higher than that of active PF L.ll
Given the well-established reactivity of thiols towards protein-based radicals, we submit
that the presence of such high concentrations of thiol relative to the radical compromised
the rigor with which any radical quenching study by other components were pursued.
Our study demonstrates that the deactivation observed by Knappe and Kesslerll cannot
be unequivocally attributed to AdhE. Indeed, we have demonstrated that AdhE, rather
than deactivate PFL, actually stabilizes the radical form of PFL. This result is in direct

10.11

opposition to those reported by Knappe and Kessler . Based on the strength of these

findings and consistently with those of other workers, it is hereby proposed that AdhE is
not a PF L deactivase but rather, protects PFL against deactivation.3°'33

Clearly, more work needs to be done in order to firlly understand the role(s)
played by AdhE during anaerobic catabolism of pyruvate. If indeed AdhE does protect
active PFL, it is expected to specifically interact with it. Such interaction, if present,
should be demonstrable. One of the first studies to be done in this regard would be to
carry out binding studies to show that indeed there is a specific interaction between these
two enzymes. It would also be interesting to investigate the interaction between AdhE
and both activated and inactive PFL to determine whether there is a discernible
difference. The role of Fe(II) in this process is also intriguing. The antioxidant role of
AdhE reported by Echave et al. was found to be dependant on the presence of Fe(II) 12,

thus suggesting a direct role by this metal. It is curious what the role of this metal might

197

play, both in the putative interaction between AdhE and PF L, as well as in the apparent
PFL protection that AdhE affords. The next question to answer would be whether the
interaction between AdhE and PFL (active or inactive) has any dependence on the
presence or absence of Fe (II). This can be done by using site-directed mutagenesis to
replace the iron-binding histidine residues with phenylalanine, in order to overexpress
iron-free AdhE. Based on the findings of these studies, the investigation can be carried
further towards elucidation of the mechanism by which the putative PFL protection is
accomplished by AdhE. The questions regarding the cooperation between PFL and
AdhE are very interesting, particularly since it has become evident that this area may not
be as well understood as first thought. The route towards answering them promises to be

an exciting one.

198

V.6. SUGGESTED EXPERIMENTS FOR FURTHER STUDIES

V.6. 1. Mechanistic Studies

Part of the initial motivation of this project was the desire to elucidate the
mechanism by which the glycyl radical is quenched during PF L deactivation by AdhE.
Although AdhE was found to be inactive with regard to PFL deactivation, several
reductants, some of which are biologically relevant, were found to be effective
deactivators of PFL. That deactivation rather than inactivation occurred was
demonstrated in the case of 2-mercaptoethanol and ethanethiol. This was accomplished
through reactivation experiments and SDS-PAGE. However, based on the data obtained
by these techniques, it is not possible to gain insight into the molecular events leading to
the quenching of the PFL radical. Time course deactivation studies using EPR may
provide useful information about both the progress of deactivation and the nature of the
radical species formed in the process. More insight would also be gained by using rapid

freeze-quench to trap any transient species that may not be amenable to regular EPR.

V.6.2. In vivo studies

All studies done in this work were done outside the cell. Attempts were made to
reproduce conditions that are as reasonably close to physiological as possible. However
close the experimental conditions may be to physiological, they are always an
approximation. The difference between the experimental and physiological conditions is
particularly important when we work in the “test tube” to study changes that take place

within a physiological cell. The pflB gene encoding PFL is FNR-regulated. Any

199

conclusions that we draw from such in vitro studies are subject to inaccuracies imposed
by the difference between the realities of the situation in vivo and in vitro. Whenever
possible, therefore, it is better to use a real cell to study changes that take place in the
cell. It is fair to aclmowlegde, however, that for practical reasons, sometimes it is not
possible to do in vivo studies. Considering the foresaid, we believe that useful
information would be gained by carrying out both whole cell and in vitro in the study of

the PFL activation and deactivation studies.

200

 

REFERENCES

1.

10.

11.

12.

13.

14.

15.

Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, J .; Schultz, S.; Wagner,
A.F.V. 1999, Nature: Struct. Biol. 6 (10), 969-975. '

Ulissi-Demario, L.; Brush, E. J.; Kozarich, J.W. 1991, J. Am. Chem. Soc.
113, 4341-4342.

Mulliez, E.; Fontecave, M.; Gaillaid, J .; Riechard, P. 1993, J. Biol.Chem.,
268, 2296-2299.

Gauld, J .W.; Eriksson, L. A. 2000, J. Am. Chem. Soc., 122, 2035

Wong, K. K.; Kozarich, J. W. In Metal Ions in Biological Systems,
Metalloenzymes Involving Amino Acid-Residue and Related Radicals, Eds.
Sigel, H.; Sigel A.; Marcel Dekker: New York, 1994; Vol. 30, p 279.

Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, J .; Schultz, S.; Wagner,
A.F.V. 1999, Nature: Struct. Biol. 6 (10), 969-975.

Wagner, A.F.V.; Frey, M.; Neugebauer, F.A.; Schafer, W.; and Knappe, J.
1992, Proc. Natl. Acad. Sci.U.S.A 89, 996-1000.

Himo, F.; Eriksson, L.A. J. Am. Chem. Soc. 1998, 120, 11449.

Sofia, H.J.; Chen, G.; Hetzler, B.G.; Reyes-Spindola, J. F.; Miller, N. 2001,
Nucleic Acids Research, 29,5, 1097-1106.

Kessler, D.; Leibrecht, 1.; Knape, J. 1991, FEBS Lett. 281, 1-2, 59—63.

Kessler, D.; Herth, W.; Knappe, J. 1992, J. Biol. Chem. 267, 25, 18073-
18079.

Echave, P.; Tamarit, J.; Cabiscol, E.; Ros, J. 2003. J. Biol. Chem. 278, 32,
30193-30198.

Asanuma, N.; Yoshii, T.; Hino, T. 2004, Arch. Biochem. 181, 2, 122-128.

Melchiorsen, C.R.; Jokumsen, K.V.; Villadsen, J .; Johnsen, M.G.; Israelsen,
H.; Arnau, J. 2000, J. Bacteriol. 182, 17, 4 783-4 788.

Holland-Staley, C.A.; Lee, K.; Clark, D.P.; Cunningham, PR. 2000, J.
Bacteriol. 182, 21, 6049-6054.

Mikulskis, A.; Aristarkhov, A.; Lin, E.C.C. 1997, J. Bacteriol. 179, 22,
7129-7134.

201

 

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Kiley, P. J .; Beinert, H. 1999. FEMS Microbiol. Rev. 22, 341-352.

Lazzazera, B. A.; Beinert, H.; Khoroshilova, N.; Kennedy, M. C.; Kiley, P.
J. 1996. J. Biol. Chem. 271, 2762-2768.

Jordan, P. A.; Thomson, A. J .; Ralph, E. T.; Guest, J. R.; Green, J. 1997.
FEBS Lett. 416, 349-352.

Popescu, C. V.; Bates, D. M.; Beinert, H.; Munck, E.; Kiley, P. J. 1998.
Proc. Natl. Acad. Sci. USA, 95, 13431 -13435.

Wybom, N.R.; Messenger, S.L.; Henderson, R.A.; Sawers, G.; Roberts R.E.,
Attwood, M.M.; Green, J. 2002. Microbial. 148,1015-26.

Iuchi, S.; Chepuri, V.; Fu, H.; Gennis, R.B.; Lin, E.C.C. 1990. J. Bacteriol.
172, 6020-6025.

Iuchi, 8.; Lin, E.C.C. 1988. Proc. Natl.Acad. Sci. USA, 85, 1888-1892.
Rice, C.W.; Hempfling, WP. 1978. J. Bacteriol. 134, 115-124.

Tseng, C.P.; Albrecht, J.; Gunsalus, RP. 1996. J. Bacteriol. 178, 1094-
1098.

Govantes, F.; Orjalo, A.V.; Gunsalus, RP. 2000, Mol. Microbial. 38,
5,1061-1073.

Michel-Angelo Sciotti, M-A.; Chanfon, A.; Hennecke, H.; Fischer, H-M.
2003, J. Bacteriol. 185, 18, 5639-5642.

Govantes, F .; Albrecht, J. A.; Gunsalus, RP. 2000. Mol. Microbiol. 37, 6,
1456-69.

Becker, 8.; Holighaus, G.; Gabrielczyk, T.; Unden, G. 1996. J. Bacteriol.
1 78, 4515-4521.

Becker, S.; Vlad, D.; Schuster, S.; Pfeiffer, P.; Unden, G. 1997. Arch.
Microbiol. 168, 290-296.

Hansen, R.G.; Henning, U. 1966. Biochim. Biophys. Acta 122, 355-358.

Alexeeva, S.; de Kort, B.; Sawers, G.; Hellingwerf, K.J.; Teixeira de Mattos,
M. J. 2000. J. Bacterial. 182, 17, 4934-4940.

202

 

 

32. Poole, R. K.; Hill, S. 1997. Biosci. Rep. 17, 303-317.
33. Kalyanaraman, B. Biochem. Soc. Symp., Portland Press, Volume 61.

34. Wodek, L. 2002, Pol. J. Pharmacol, 54, 215-223.

203

 

   

nijl'llflflijiljjljljjllilijj‘u