._
.4 .1. . o. . - <- . . .
7.3.. at... .V... 3.. 2,- ...$1¢Vl... £73.33. .1 I3... ‘0: . ..
a. 53...... UNIV. .. . 3.3. . z}. .0 o L}... 6...: on...
. .V‘.
.. V. V: .
1 1.4930. .o.... C I! ... 7. ..
.,.. V _ V . V ..,H‘J.pl) . 104.9-l- J.
. .v... . L. 93:: « fart...- l..1o¢.. flu. . at)?“
2... :29 .39..- . V z... I: &$CI!ao.¢: o.’...%; no . V
V VV . . V. .9. II .. _ 3.0.9... \O!o.~l V 3"
. ..1p1..)tIt-.2J 0:230! V. ‘9’...
.011... lat—4“.
. o

 

       

'. 9,.- .‘|.o V9:
. § .32 .08. $9.. s? Q 9. Erositlutlsg
. V ‘9'*...’l.

       

'.A... V
to. .2!‘.... .3. av. I

 

     

13¢".-

 

   

 

 

 
      

 

 

  

 

  
 

 

,.!._ ..: .. 3L9... ...!:..t Bias!!! . . . 151 .o 4‘! .
. .7 . a _.' 3...... V. . .k .9. 1.1.1 V... 00.? .. .V...!..’(!c_.....31.r.:s 3‘.” 0.. ...
. .v«?l.?..2..:O-.!P . 3:3} 1.2. o..s.a.0:1~ 91?. u-UEO SA
V. .. ... 1..- 1......‘3. ._ DY-.o~"v)ro‘. . f...
4213'...’ ‘22960‘1fl‘

   

8.2.79

              

- H. V

.-.- . . . . .
"V... V . V. V V. r. . V

. . . . .- ciftV ‘...

. V . .....V 3 .‘3. '-

 

 

. . no . o. v o .
J. 7‘ 00.....05 .
.- . ..o¢..o.v0. V V . . . .
. . r '3
V V V .

           

.31...
V I \‘U! .V‘.‘

. . o I c .
V A}. :0. 1‘ .0
(It... -. 0“...»6
.3. ._ l .. .

I...

 

‘3.
'u...
3!! .ltuJ 2.

.0; civil ,...V V

            

   

.. to... a. _
. 1.3.0.6...
.

 

 

 

9
1

 

   

... . .
3.3.... V V
V... 2.2 .3

.9:

v-I
.
,...... ‘10-...
. V2; .V

  

1-?
.o 3..
2- 4

  

 

0.10
‘I

 

 

 

 

 

      

 

 

 

 

 

 

 

    

 

 

  

 

 

                   

 

'

q

‘.

’V

.. . . .

.2 . 07“..’?...

1. £11. .9

.5

‘-0 V

. v... Pu... 7... 1.
’0. V. o.\v3n.$..nq..:3 i UQu" I‘V‘,

. 3.5., .9. .92... .. I. l .. 131-. 0“.le
.0 .. . o . .
o

 

 

‘9'

 

c .p. Q :1:
.211. ‘3

 

 

1'.

 

to
fix :0 c .

 

               
      

.o.‘~ 7-.

We"?!

IV

 

       

 

 

 

nu: .
"9. .3 ....o:~oc 0.... V
. ,. v}... .V ... .
O.!J...?.t.. _ .V . .. . V V I. :
.2 . ft}? . V. . . . . V .V . V . V V _... V
.V. .V V . V . . V . V . . V V V . O. i.
. V V. V V. . V |o~ .. .t....~. .0. ‘I 391
V ...~V...!:. ,0! V. ~V.Vfi-d.V V . V on
. .91§.!1bbq0| . \5
. . . 3 O .

 

   

...I....I. 9:. is 119%.}? 13.?!)3.
..lo...9....399 . .-V.:9l!'!!7s.¢.} 7... I30¢
9].; .9. 14‘. .1... )t’h.2V u $.33 .JI-
.3._w.¢u Vt; 02't’o‘ IQQI. v. 3"
1 0.! Vow. 7’928.~4 f‘l.) O.

 

."

 

    

 

 

 

 

1.2-3. 0:
V 1 .5 . .

  

   

pugs-bu

. I... . . .V. .V
1 .V

 

     

    

....... . .V .. .VV
:-....X.-.v. 7‘: V. .

 

 

 

. . . V . .
. V . .V . V V . . . . V V . l ..|.
. V .. . V. . . 2... .0.
V - . Co .0. ~VV..V0
..V:..|.‘I

o

.‘u...‘.. V o

..uo.o.3|§m.

 

 

«1999 ms

 

-
o
O
a
o
u
o

.

 

5“”,

   

V0. 5"; V 0.".
l... . 3.1.0....
.‘On . .. 'r.'.

      

    

. o3...

 

  

    

 

   

~14”

0'. 2.1.3.1.. u. .-
. . V n
.-

V .

   

_ V. .. V
.V. 3.9..
. . .. :V.

 

.
....V
O...
. .t.
V. .01..

                

mm

    

                

.

. ’0.

V..............

. .-V V .

l. V ......O...-.lc.
V. .u...

   

 

o.

‘9‘...” ”f.

R734

      

          

 

         

        
         

.
p.
.. . V . V V
. . V . V . V V ..
. . . V. o . 3 . V V .. . V
V V . . . . . . . . . V . V V. V .. V .
. . V V . . . V ..V . . .
V . V V . V V V ... V. . .. _ t V V . . V. -.
a C V V XV . V 39.9“! V.
. Y. o .3. 71.0... n; I a V 39 z! 0 it.§a
I .2... . 0.1 9511.¢Q.I~ 9.. 5... 119]. I)!
.14. V: ‘. o D. I. .- 37-01
1- 09" .V‘QOna'. In I. I .
c2302‘c I«V."-
v.0..cli-

.
u

o
;
.

         

 

.
Q .

. .
. . .
.VV. .V.
a c v to-...¢c.. . . . ~ . .
.. ... . . .. .l...u¢....V-9 073...;
.V V . . . . _. \AVt‘nc. .Vu..-o..‘...‘o.- ‘.
V «9...: .9.

              

 

‘0‘... u 3-...
01.- QES. ‘3. "- \R.
in”?

 

   

 

¢ 3‘. 4 rt.‘ .V.vv.|Pa.-

 

on .
:... c.1139... O!
. 1.7...

V . . V. . V V .V. V . ... .. . .. V. . 4 . . V . V a.
V . . .ul..,u.09

0

fl." '9‘ N “_‘ I wuch-Vo

            

 

 

   

 
 

 

.
n . .
..V V .
a .u.' u 0 at. 3.:-
~ . .u. u. a .
I O -.- .u a t .-
m . . . . . .V . . f
. I b . .o .. V. . .. . V v o
. . u
. I. p n .
H 6 .o .. . a. u. .. u . .. . ..n\. .
IV... . . ...... . . .. .. . .. .1...o‘?.3.|'..
.- n
o) u. . a o . V . . . . . 4.. .
.0 .. a o v V o ‘n. ,- 0n . . ~ 0
o . o. o. o . ..- o v
V . . . . . o .. ‘0 AA 1094
.OIIOQ...I..V-.v ..
..'!‘4’.-0Iai two 0
$01.00;...

 

o
>

“i".c a;
3

O
.
a

 

 

. .V . . V.
. V . V V. .V .. ..A1oc~0.o\.ho..
V Vt. .. .-

...
..-.-.O~oo..o

 

 

V .4.
3... .

 

O3". otnflmu‘“u

..n
.n..

 

 

          

  

 

 

w. . I 'l n I. _ .w. b h A v
u .. .. . .
f .. . . V x. .
- .- —.I-o'.- .07.

‘ . o . .

t .. o . . .
'. 0 O. n. D c ‘00:. ‘Q .4. .0 l4 ..
O. . . . . p. ....o . oon..‘.n.‘!.¢v.cn
. I l - I.-§-'IVIVOI.II..OI‘l-‘I-V
‘ o . . . . . font.- . f.V.Y . ..QV. ..0-.....

. . V... on ..’..6.- uncle-V. .AOcV

"o. .. V V.. . l...n.. .. .I-Vo..‘o.o...o a .Vh..|¢ : u. .

. . . . . V. outgoo.b§7.n. 00......len u¢1.-Q.V ‘Q ' u-
“ o . lo. .04.. ro-Vu.oV’ p...-OIDI...0 OQIO‘ kt.
O - .. . .- .0 .'.V.~uc.o ~V.o I30..... .V

. 0.9-. V...II¢.....oo.O-V1 .. . a.. .
. . .a'. c.~ .- .-. . .IVpuo.v..| V... 0......
.. .....-.. .01... 0.5.... ..VoV- v... I. V... .8... cg
. .‘uu... .vo. .1 .‘.~.V. a .o4 ......... .. .3 A..V~ .o
o‘.- lunc....0v. V .00.. . V ... .Ol?o .
. u . V «... c . .. . ._.V .0 I. -so...._
. . a ...o.. .. .V 0-. . V.
l u . . o
..

       

1 .C Q u
a ..O ...o .

VVVVVVVszVVVVVéiEéVV

_ o . .. .
. V . ..
.HVVV...04.J..—Ip.-.

.. t? 45...... .
l V.

‘ .o... b...

    

2 '..u.‘..

LflT’WYqo-agm.

.. ...,.........,u... .

. . 3
Rab... 1:30....) Flumagf‘

V . . .

an.... .-n>‘Q~AV.6
. o

 

 

. .VVI.“£ o. a... . V .
. f» a... V.3V.....V...113Jvt 1V :0 V. 15! 5.7.2.... , ..V. ~ V V t
.11.}. .2; 2.3;? _."...V...Vx..u.r........h. .x . 23.9.?
.14... .v:\.o¢.oV 3 «I! nfi). :0 .I.v3 2.75.4...0. 33.38 V ‘33.... V‘ 111‘ . «US-630:. d a
. V ........2:l.8 1...... . ‘.¢.LV-.l 3. . o: .
V n. i, V V."!..VES ..‘o.3‘...-oJJ
V V . V V V a V. o

    

    

  
 

 
 

 

\u...0‘!....
. :- .‘ _ 0... ....¢. ‘ .v... . .
o O . . o to

-‘l

 

      

                         

     

                

 

 

 

 

 

I In" l '5‘,

NHL}: II: ‘ ”ate
University

 

 

 

This is to certify that the
thesis entitled

A BROADLY SPECIFIC BENZOATE COENZY ME A LIGASE
IS COUPLED WITH TAXUS ACYLTRANSFERASES IN
VITRO TO BIOSYNTHESIZE PACLITAXEL ANALOGUES

presented by

Sean Austin Sullivan

has been accepted towards fulfillment
of the requirements for the

MS. degree in Chemistry

 

 

mixer/Ax

/ Major Professor’s Signature

MAY ’1 20 IO

 

Date

MSU is an Affirmative Action/Equal Opportunity Employer

-o—A-A—o— -

.O-O-I-l-o-l-o-C-I-O-l-l-l-I-.-O-l-a-

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KIIProleoc8-PrelelRC/DateDuetindd

A BROADLY SPECIFIC BENZOATE COENZYME A LIGASE IS COUPLED WITH
TAXUS ACYLTRANSFERASES IN VITRO TO BIOSYNTHESIZE PACLITAXEL
ANALOGUES
By

Sean Austin Sullivan

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Chemistry

2010

ABSTRACT
A BROADLY SPECIFIC BENZOATE COENZYME A LIGASE IS COUPLED WITH
TAXUS ACYLTRANSFERASES IN VITRO To BIOSYNTHESIZE PACLITAXEL
ANALOGUES
By

Sean Austin Sullivan

Herein an optimized in vitro coupled assay system that couples the activity of a benzoate
coenzyme A ligase and NDTBT is described. Overall product yields are within
experimental error (i 5 %) of those found in comparable NDTBT in vitro assays with the
natural co-substrates. Moreover, it will be demonstrated that the current coupled assay
system can be used to biocatalyze several N—acyl-N-debenzoyl-2'-deoxypaclitaxel
analogues, as well as paclitaxel. As important, in vitro coupled assays with another
T axus 2-0—acyltransferase, designated mTBT, are also demonstrated, including the
effective production of multiple 2-acyl-2-debenzoyl-7,13-diacetylbaccatin III analogues.
Paclitaxel analogues with similar functionality to those produced herein have been
identified as more efficacious than paclitaxel or docetaxel. The advantages to the
coupled assay system include improved substrate availability (compared to in vitro assays
for the described acyltransferases), coenzyme A recycling capabilities, and

counteractivity against benzoyl coenzyme A degradation in solution.

To my mother, Monica Sullivan

iii

ACKNOWLEDGEMENTS

First and foremost, I would like to thank my friends and family for all of their
support and understanding throughout this process. I would also like to thank my fellow
graduate students in Dr. Walker’s lab for experimental support. Danielle Nevarez, Irosha
Nawarathne, Behnaz Shafii, and Yemane Mengistu deserve special thanks for their
continued assistance while working with the enzymes and substrates researched herein.
The undergraduate student Noelle Byrne also deserves thanks for her technical assistance.
Finally, I would like to thank Dr. Kevin Walker for his guidance and advice, as well as
the other members of my committee, Dr. Dana Spence and Dr. J etze Tepe.

This research could not have been completed without the badA cDNA generously
provided by the University of Washington (Seattle, WA). I would also like to thank the

MiChigan State University Mass Spectrometry Facility and the Genomics Core.

iv

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... vi

LIST OF FIGURES ....................................................................................................... vii

LIST OF ABBREVIATIONS .......................................................................................... xi

INTRODUCTION ............................................................................................................ 1
CHAPTER 1

Expression, Activity, and Further Studies with Benzoate Coenzyme A Ligase ............... 7

1.1: Introduction .................................................................................................... 7

1.2: Experimental .................................................................................................. 9

1.3: Results and Discussion ................................................................................ 17

1.4: Conclusions .................................................................................................. 25
CHAPTER 2

Coupled Activity of BadA with the T axus Acyltransferases NDTBT and mTBT ......... 27

2.1 : Introduction .................................................................................................. 27

2.2: Experimental ................................................................................................ 29

2.2.1 NDTBT Experimental .................................................................... 30

2.2.2 mTBT Experimental ....................................................................... 39

2.3: Results and Discussion ................................................................................ 42

2.3.1: NDTBT Results and Discussion ................................................... 42

2.3.1: mTBT Results and Discussion ...................................................... 61

2.4: Conclusions .................................................................................................. 64
CHAPTER 3

Future Aims .................................................................................................................... 66

REFERENCES ........................................................................ ' ....................................... 69

LIST OF TABLES

Table I ............................................................................................................................... 5
The cost of several acyl-coenzyme A substrates, (Sigma-Aldrich).

Table II ............................................................................................................................ 21
Series of assays, including controls, used to determine BadA activity after
purification. Y indicates inclusion in the reaction mixture; N indicates absence from
the reaction mixture; N/D represents “Not Detectable.”

Table III .......................................................................................................................... 24
Retention times for the putative products of the BadA catalyzed reaction when
incubated with the reported acyl-carboxylic acid donor.

Table IV ........................................ 50
Assay compositions used in the optimized NDTBT and BadA coupled assay system.
Y indicates inclusion in the assay composition.

Table V ............................................................................................................................ 52
Retention times (min) for the N—substituted-N-debenzoyl-2'-deoxypaclitaxe1 products
of the NDTBT and BadA coupled reaction when incubated with the reported acyl-
carboxylic acid donor.

vi

LIST OF FIGURES

Figure l .............................................................................................................................. 2
The antineoplastic taxanes paclitaxel (left) and docetaxel (right).

Figure 2 .............................................................................................................................. 8
Enzymatic activation of BzOH with CoASH to yield BzCoA proceeds via the activity

of BadA, which requires both Mg2+ and ATP as cofactors.l

Figure 3 ............................................................................................................................ 15

Calibration curve for BadA product quantitation using A210 Peak Area as a function
of BzCoA Amount, reported in nmol. The 40 ,uL injection volume allows for
conversion of BzCoA amount to concentration. Error bars are reported as the
standard deviation of three separate measurements.

Figure 4 ............................................................................................................................ 18
The nucleotide (5 '-3' top strand) and amino acid (N- to C-terminus translated bottom
strand) sequences obtained from the recombinant badA gene. Sequencing alignment
(not shown) with the cDNA sequence of badA in the source vector pPE204 showed
identical sequences.

Figure 5 ............................................................................................................................ 19
SDS-PAGE analysis of four of the seven separate elutions collected during BadA
purification. Lanes: 1, elution at 100 mM imidazole; 2, elution at 150 mM imidazole;
3, elution at 200 mM imidazole; 4, elution at 250 mM imidazole in Assay Buffer; 5,
Protein ProSieve® ladder. The two bands in the ladder closest to the putative BadA
band are labeled with their respective molecular masses (75 kDa and 50 kDa). The
band in Lane 4 was used to estimate BadA mass at ~58 kDa.

Figure 6 .......................................... ' .................................................................................. 20
Chromatogram obtained following the analysis of a BadA activity assay. Assay
conditions were as follows: 2.5 mM MgC12, 1.0 mM ATP, 0.25 mM CoASH, 0.25
mM BzOH, 100 pg/mL BadA, incubated at 31 °C for 10 min followed by acid
quench.

Figure 7 ............................................................................................................................ 28
A reaction scheme describing the coupled activity of BadA and NDTBT. The N-
debenzoyl-2'deoxypaclitaxel is pictured as it is considered the natural substrate for
NDTBT. Both the kcat for BadAl and NDTBT (reported in this thesis) are listed with
their respective natural substrates. The five separate acyl donors displayed describe
the substrates that have been investigated.

vii

Figure 8 ............................................................................................................................ 29
A reaction scheme describing the coupled activity of BadA and mTBT. 7,13-
Diacetyl-2-debenzoylbaccatin III is pictured, as it is considered to be a surrogate
substrate for mTBT. Both the kcat for BadAl and mTBT2 are listed with their
respective natural substrates. The three separate acyl donors displayed describe the
substrates that have been investigated in this research.

Figure 9 ............................................................................................................................ 43
Reverse phase separation followed by UV detection at 228 nm of an activity assay for
the enzyme NDTBT. Peak identities: 14.0 min, baccatin III; 16.7 min, N-debenzoyl-
2'-deoxypac1itaxel; 18.8 min, 2'-deoxypaclitaxel. The baccatin III was included as a
standard, as will be discussed in later experiments.

Figure 10 .......................................................................................................................... 45
Hanes-Woolf analysis of the data obtained from incubating NDTBT with 500 ,uM
NDBZDT and varying concentrations of BzCoA at a 10 min incubation period. The
plot shows the concentration of BzCoA divided by the experimentally determined v0
(reported as nmols of 2’-deoxypaclitaxel divided by 10 min) vs. BzCoA
concentration. Error bars are reported as the standard deviation of three separate
measurements.

Figure 11 .......................................................................................................................... 47
Hanes-Woolf analysis of the data obtained from incubating NDTBT with 500 ,uM
NDBZDT and varying concentrations of BzCoA at a 10 min incubation period upon
the addition of CoASH as indicated in the legend. The plot shows the concentration
of BzCoA divided by the experimentally determined v0 (reported as nmol of 2'-
deoxypaclitaxel divided by 10 min) vs. BzCoA concentration. Error bars are reported
as the standard deviation of three separate measurements.

Figure 12 .......................................................................................................................... 50
The above chart displays 2'—deoxypaclitaxel concentrations obtained when the assay
compositions listed in Table IV were incubated for the prescribed 2 h time point.
Error bars are reported as the standard deviation of three independent measurements.

Figure 13 .......................................................................................................................... 54
Mass spectrum of the collected de novo peak (at a retention time of 18.6 min) from
the separation of an NDTBT and BadA coupled reaction incubated with the substrates
NDB2DT and BzOH. Presence of the pseudomolecular ion [M+H]+ at 838.5 m/z as
well as the regular fiagmentation pattern of paclitaxel analogues confirms the identity
of the product peak as 2'-deoxypaclitaxel.

viii

Figure 14 .......................................................................................................................... 55
Mass spectrum of the collected de novo peak (at a retention time of 19.4 min) from
the separation of an NDTBT and BadA coupled reaction incubated with the substrates
NDBZDT and 4-methylbenzoic acid. Presence of the pseudomolecular ion [M+H]+ at
852.5 m/z as well as the regular fragmentation pattern of paclitaxel analogues
confirms the identity of the product peak as N-debenzoyl-N—(4-methylbenzoyl)-2'-
deoxypaclitaxel.

Figure 15 .......................................................................................................................... 56
Mass spectrum of the collected de novo peak (at a retention time of 18.4 min) from
the separation of an NDTBT and BadA coupled reaction incubated with the substrates
NDBZDT and thiophene-2-carboxylic acid. Presence of the pseudomolecular ion
[M+H]+ at 844.6 m/z as well as the regular fragmentation pattern of paclitaxel
analogues confirms the identity of the product peak as N-debenzoyl-N—thiophene-Z-
carbonyl-2'-deoxypaclitaxel.

Figure 16 .......................................................................................................................... 57
Mass spectrum of the collected de novo peak (at a retention time of 17.6 min) from
the separation of an NDTBT and BadA coupled reaction incubated with the substrates
NDB2DT and 3-furoic acid. Presence of the pseudomolecular ion [M+H]+ at 828.4
m/z as well as the regular fragmentation pattern of paclitaxel analogues confirms the
identity of the product peak as N-debenzoyl-N-(3-furanoy1)-2'-deoxypaclitaxel.

Figure 17 .......................................................................................................................... 58
Mass spectrum of the collected de novo peak (at a retention time of 19.0 min) from
the separation of an NDTBT and BadA coupled reaction incubated with the substrates
NDBZDT and 2-fluorobenzoic acid. Presence of the pseudomolecular ion [M+H]+ at
856.5 m/z as well as the regular fiagmentation pattern of paclitaxel analogues
confirms the identity of the product peak as N—debenzoyl-N-(2-fluorobenzoyl)-2'-
deoxypaclitaxel.

Figure 18 .......................................................................................................................... 60
Mass spectrum of the fragmented parent ion 854 m/z from the analysis of an NDTBT
and BadA coupled reaction incubated with the substrates N-debenzoylpaclitaxel and
BzOH. Presence of the pseudomolecular ion [M+H]+ at 854.3 m/z as well as the
regular fragmentation pattern of paclitaxel analogues confirms the identity of the
product peak as paclitaxel. The base peak 286.1 m/z represents the protonated N-
benzoylated—phenylisoserinoyl sidechain.

Figure 19 .......................................................................................................................... 61
Product mass spectrum for 7,13-diacetylbaccatin 111 product as a result of an mTBT

activity assay. The base peak at m/z = 671.4 corresponds to the [M+H]+
pseudomolecular ion.

ix

Figure 20 .......................................................................................................................... 63
Product mass spectrum for 7,13-diacetyl-2-debenzoyl-2-(4-methy1benzoy1)-baccatin
111 product as a result of an mTBT and BadA coupled assay. The base peak at m/z =

707.4 corresponds to the [M+Na]+ pseudomolecular ion.

Figure 21 .......................................................................................................................... 64
Product mass spectrum for 7,l3-diacety1-2-debenzoyl-2-(thiophene-2-carbonyl)—
baccatin III product as a result of an mTBT and BadA coupled assay. The base peak

at m/z = 677.4 corresponds to the [M+H]+ pseudomolecular ion.

LIST OF ABBREVIATIONS

Absorption Measurement at 210 nm
Absorption Measurement at 228 nm

Absorption Measurement at 600 nm

Adenosine Triphosphate

Benzoate Coenzyme A Ligase (specific)
Benzoate Coenzyme A Ligase (general)
Benzoate Coenzyme A Ligase Gene Sequence
Benzoic Acid

Benzoyl Coenzyme A

Bristol-Myers Squibb

Bovine Serum Albumin

Centimeter

Collision Induced Dissociation

Coenzyme A

Confer

Electrospray Ionization — Tandem Mass Spectrometry
Environmental Protection Agency

Gram

Gravity

High Performance Liquid Chromatography

Hour

xi

A210
A228

A 600

ATP

BadA
BzCoA Ligase
badA

BzOH
BzCoA
BMS

BSA

cm

CID
CoASH

cf.
ESI-MS/MS

EPA

g

g

HPLC

Immobilized-Metal Affinity Chromatography
Initial Velocity
Isopropyl-B-D—thiogalactopyranoside
Kilogram

KiloDalton

Liquid Chromatography - Electrospray Ionization Mass Spectrometry

Liter
Luria-Bertani

Maximum Velocity

Michaelis Constant

Microgram
Microliter
Micrometer
Micromolar
Milligram
Millimeter
Millimolar
Minute

Molar
Modified TBT
Molecular Weight Cutoff
Nanometer

Nanomole

xii

IMAC

V0

IPTG

kg

kDa
LC-ESI-MS
L

LB

Vmax

,ug
pL

pm

mg
mm

mM

min

mTBT

MWCO

nmol

National Cancer Institute

Normal

N—Debenzoyl-(3’R)-2'-deoxypaclitaxel
N—Debenzoyl-2'-deoxypaclitaxel:N-benzoyltransferase

Nitriloacetic Acid
Optical Density at 600 nm

Plant Cell Culture

Polymerase Chain Reaction
Proton Nuclear Magnetic Resonance

Quadrupole — Time of Flight Mass Spectrometer

Second

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis
Species

Thin Layer Chromatography

Triethanolamine

Trifluoroacetic Acid

Tris(hydroxymethyl)aminomethane

Turnover Number

Ultraviolet
Volt

'-Deoxypac1itaxel
2-0—Deacyltaxane:2-0-benzoyltransferase

10-Deacytelbaccatin III

xiii

NCI

N
NDB2DT
NDTBT

NTA
OD600

PCC

PCR
lH-NMR
Q-ToF

5
SDS-PAGE
spp.

TLC

TEA

TFA

Tris

kcat

UV

V

2DOT

TBT

1 O-DAB

INTRODUCTION
Natural products remain some of the most potent medications used in the

pharmaceutical sector to date. Many diseases are treated with natural products or with
bioactive compounds derived from natural product precursors.3 The National Cancer
Institute (NCI) conducted a large-scale screen of flora in the Pacific Northwest (USA),

out of which paclitaxel (Taxol®) was discovered as an active component in the Pacific

yew tree (T axus brevifolia).4 Paclitaxel was found to be an antimitotic agent that
stabilizes the formation of microtubules during cell division, and thus induced cellular

9

apoptosis.5’ Since its approval for clinical use in 1992, paclitaxel and its synthetically
derived analogue docetaxel (Figure 1) have been used to treat multiple cancer types,
including ovarian, breast, lung, head, and neck carcinomas.8 Paclitaxel has also been
used in the treatment of coronary artery disease as a preventive measure to reduce the

. . . 9 . . . . . . ‘ . .
occurrence of in-stent restmosrs. In addition, there is some mterest 1n usmg paclitaxel 1n

the treatment of Alzheimer’s disease where it is currently being used as a research tool to
counter the effects of the degraded microtubulin cytoskeleton in neuronal cells of patients
suffering from this dementia affliction.10 The broad range of both research and medical
uses for paclitaxel (and its analogue docetaxel) has presented a strong demand for this
successful antimitotic drug, as well as for modified paclitaxel analogues that could prove

more effective in medical applications.

 

Paclitaxel (Taxol) Docetaxel (Taxotere)

Figure l. The antineoplastic taxanes paclitaxel (left) and docetaxol (right).

The original source of paclitaxel for preclinical trials was the natural resource T.

brevifolia, from which only a dearth of paclitaxel could be isolated (1 kg paclitaxel per

6000 kg of T. brevifolia bark, i.e. 0.02 % w/w).11 Harvesting the bark from the plant was

a sacrificial process that depredated the plant from old growth forests.4 The

Environmental Protection Agency (EPA) issued an edict that forced the research
community to look for alternative methods of production of the pharmaceutical that
would be both more efficient and environmentally friendly. The alternative methods for
production that have been researched so far include semisynthesis from naturally

occurring taxane precursors and plant cell cultures derived from T axus plant species.”-14

While the total synthesis of paclitaxel was achieved in the 1990’s by several different

15,16,17,l8,19,20,21,22,23 .
groups, these syntheses presented elegant examples of multi-step

organic chemistry transformations, and were never truly intended to address drug supply
issues. Completing the total synthesis of paclitaxel has inspired the synthesis of many

other natural products, and has also contributed to some of the methods of production

used presently.24

The methods used for paclitaxel production today represent the culmination of
decades of research; however, these processes do not fully address production efficiency
and minimization of environmental impact. The first significant route towards paclitaxel
production was a semisynthetic method, which relied on harvesting the leaves of T axus
saplings and extracting relatively abundant natural products lO—deacetylbaccatin III (10-
DAB) and baccatin III, which are late-occurring metabolites on the paclitaxel
biosynthetic pathway.12 While these precursors can only be isolated in relatively low
yields of 0.1 % (w/w) from the leaves, the semisynthetic method utilizes renewable

leaves as a source of paclitaxel and its precursors without killing the slow-growing tree.25

The original contracted producer of Taxol®, Bristol-Myers Squibb (BMS), has been
influential in continuing the search for more environmentally-responsible and efficient
methods for paclitaxel and its analogues.4 The current method of synthesis employed by
EMS, and other generic companies, is a plant cell culture (PCC) process that is
considered a greener method of production compared to the previously used
semisynthetic routes.14 Overall yields of paclitaxel reported for the originally published

plant cell culture method range from 0.012 — 0.05 %, which is estimated at 0.3 mg of

paclitaxel per gram of dry cell weight per day, which can be maintained for up to 40

,14

days.13 These methods, considered more efficient and green than previous

semisynthetic methods, have allowed the supply of paclitaxel to meet the demand of

. . . . . . 26
medical and research interests, which 18 estimated at one metric ton.

While current PCC methods are addressing the supply issue, there remains a

sustained demand for paclitaxel in basic research and in the development of taxane

analogues that may aid in the continued fight against cancer and other diseases.
Currently, most paclitaxel analogues are only available via the semisynthetic methods

due to the lack of investigating the utility of PCCs to manufacture the non-natural
paclitaxel compounds in viva.4 Structure-activity relationships studying the necessity of

the different functional groups attached to the taxane core have revealed that some

analogues with N—acyl-N-debenzoyl-substitutions may be equally or more efficacious

than the currently used paclitaxel or docetaxel.27 While interest in such analogues is

high,28 supply issues still plague further studies of such compounds. One possible

method for producing such analogues is the engineering of a transgenic host expressing
all the necessary genes on the paclitaxel biosynthetic pathway. Current work on such a

method has been characterized in yeast (Saccharomyces cerevisiae); however, as of yet,

taxadiene is the only diterpenoid that has been successfully isolated from such cultures.29

Several more enzymatic steps are involved in synthesizing paclitaxel from taxadiene, thus
still more research is necessary before such a system can become feasible for industrial
scale-up. Two of the enzymes already identified in the final steps of the natural
paclitaxel biosynthetic pathway are N-debenzoyl-2'-deoxypaclitaxel:N—benzoyltransferase
(NDTBT) and 2-0-deacyltaxane:2-0—benzoyltransferase (TBT) have been suggested as
possible enzymes for such a transgenic host.2’30 Both of these T axus acyltransferases
exhibit broad substrate specificity that could potentially be used to construct paclitaxel
analogues. However, the ability for the biocatalytic production of such analogues relies

on the availability of the acyl-coenzyme A (acyl-CoA) co-substrate in the catalyzed

reaction.

Currently, the in vitro assay procedure described in the analysis of NDTBT and
mTBT requires incubation with acyl-CoA substrates obtained commercially or through

synthetic methods where an acyl acid is coupled to CoASH to form the corresponding
thioester.2’30 The prohibitively high cost of commercially available acyl-CoA substrates

is demonstrated in Table I.

Table I. The cost of several acyl-coenzyme A substrates (Sigma-Aldrich).

 

 

Acyl-Coenzyme A Substrate Cost/5 mg

2-Butenoyl Coenzyme A $ 186.50
Acetyl Coenzyme A $ 57.07
Arachidonoyl Coenzyme A $ 180.00
Benzoyl Coenzyme A $ 184.50
Butyryl Coenzyme A $ 106.00
Decanoyl Coenzyme A $ 102.50
Hexanoyl Coenzyme A $ 119.50

While these acyl-CoA substrates can be synthesized at high reported yields, the general
syntheses require several organic solvents and purification steps.30 One possible

alternative to using these costly substrates in assays with the acyl-CoA-dependent
acyltransferase is to incubate the enzymes in the presence of a benzoate coenzyme A
ligase (BzCoA ligase). Such a coupled assay system would prevent the necessity of
completing the difficult mixed-phase syntheses or purchasing the expensive acyl-CoA
substrates.

Previous research has been conducted wherein an acyl-CoA synthetase was used

31,32,33

to recycle free CoASH in chemical reactions; however, none of the past research

has shown recycling of CoASH with a broadly specific BzCoA ligase. Moreover, there

are no accounts demonstrating such a coupled assay system proposed herein for the
production of paclitaxel analogues. The research described in this thesis aims to test the
feasibility of a coupled enzyme assay system where a BzCoA ligase from
Rhodopseudomonas palustris produces the necessary acyl-CoA substrate for the reaction

catalyzed via NDTBT or mTBT with their respective, diterpene substrates.

CHAPTER 1
Expression, Activity, and Further Studies with Benzoate Coenzyme A Ligase
1.1: Introduction

While plant cell culture methodologies currently dominate the market for
production of paclitaxel)?”14 further studies into heterologous expression hosts could

provide a more efficient and environmentally friendly method of production. A
heterologous expression host would require multiple enzymes in the paclitaxel
biosynthetic route to effectively biocatalyze the complex diterpenoid molecule. Many of

the acyltransferases involved in the paclitaxel biosynthetic route in T axus plants have

been identified and characterized;34 however, many of the other important enzymes

within the pathway have not yet been elucidated.34 One of these important enzymes is a

putative benzoate coenzyme A ligase (BzCoA ligase), which catalyzes the formation of a
thioester bond between benzoic acid (BzOH) and coenzyme A (CoASH), thus yielding
benzoyl coenzyme A (BzCoA). BzCoA is an important substrate for the T axus

acyltransferase enzymes, yet is not described in many of the feasible hosts for transgenic
expression.35 Since the pertinent BzCoA ligase from T axus plants has not been

identified, any heterologous host employed to construct novel paclitaxel or its precursors
would require a novel ligase be introduced into the system through a genetic engineering
methodology.

The necessity to identify a BzCoA ligase for heterologous expression in a suitable
host is clear. More important, the selected BzCoA ligase can offer many added benefits,
including a broad substrate specificity range and a high turnover number. The research

presented here focuses on the utilization of a BzCoA ligase (BadA), isolated from

Rhodopseudomonas palustris (R. palustris), in the production of paclitaxel and baccatin
III analogues by coupling BzCoA production to the activity of two acyltransferases
involved in the paclitaxel biosynthetic pathway. The reaction catalyzed by the reported
BadA is shown in Figure 2. This specific BadA enzyme was chosen because of its

previously reported broad substrate specificity, high turnover number, and

availability.1’36

 

NH2
0 0 ,6 9 u </N I \N
OH + HSkaANM/O‘lr'O—P—O N NJ
H H OH OH m
OH
9 OH
o=P-0H
BadA OH
(Msz*.ATP)
V
NH2
° 9 9 m
5\/\ M ' 0".”‘°"."° J + AMP + PP-
N N OH OH O N N I
o H H OH
9 OH
o=P—0H
OH

Figure 2. Enzymatic activation of BzOH with CoASH to yield BzCoA proceeds via
the activity of BadA, which requires both Mg2+ and ATP as cofactors.1

1.2: Experimental Section

General. An Agilent 1100 high performance liquid chromatography (HPLC)
system (Agilent Technologies, Wilmington, DE) connected to a UV detector was used for
separation and analysis of assays. Solvents of reagent grade were obtained from Sigma-
Aldrich. Restriction endonucleases were purchased from New England Biolabs (Ipswich,
MA). In terms of the performed biochemical work, including microbial and recombinant

techniques, most are described in detail by Sambrook.37

Substrates. Benzoic acid (BzOH) was purchased from Mallinckrodt

(Phillipsburg, NJ). All other acyl-carboxylic acids, as well as BzCoA, adenosine
triphosphate (ATP), and MgClz, were purchased from Sigma-Aldrich. Coenzyme A
(CoASH) was obtained from ARC (St. Louis, MO). Initial studies were completed with
laboratory stocks of synthesized BzCoA via literature reported syntheses.30

Bacterial strains and culture components. All media solutions were prepared
in distilled, deionized water. Luria-Bertani (LB) medium contained tryptone (10 g)
(Bacto Laboratories Pty Ltd, Australia), yeast extract (5 g) (Bacto) and NaCl (10 g)
(Sigma) or 25 g of LB mixture (Neogen, Lansing, MI) per 1 L of distilled, deionized
water. Solid medium was prepared by the addition of 1.5% (w/v) agar (Bacto) to the
medium with appropriate antibiotics. All media was sterilized via autoclave.

The antibiotic kanamycin (Sigma) was added where appropriate to a
concentration of 50 ,ug/mL. Isopropyl-B-D-thioglucopyranoside (IPTG) (Gold
Biotechnology, St. Louis, MO) was added where appropriate to a final concentration of

100 ,uM. Stock solutions of kanamycin and IPTG were prepared in distilled, deionized

water. Antibiotic and IPTG stock solutions were filter-sterilized through 0.22-,um
membrane filters (Millipore, Billerica, MA).

Subcloning badA gene from R. palustris genome. The Harwood group at the
University of Washington (Seattle, WA) generously provided two plasmids pPE202 and

pPE204, which contain fragments of the R. palustris genome that include the badA

gene.36 The badA gene was subcloned and directionally inserted into the pET28a vector

(Novagen), which incorporated an N-terminal His6-tag epitope. This was accomplished

through amplification of the badA gene with the forward primer set (5'-
TATGAATGCAGCCGCGGTC-3'; 5'-TGAATGCAGCCGCGGTCAC-3') and reverse
primer set (5'-GTCAGCCCAACACACCC-3'; 5'-TCGAGTCAGCCCAACACACC-3').
The resultant amplicon contained the badA gene and incorporated NdeI and Xhol cut
sites for directional insertion into the pET28a vector. The recombinant badA plasmid,
henceforth referred to as pBadA, was transformed into Escherichia coli (E. coli) cell lines
DHSa (for plasmid storage) and BL21(DE3) (for overexpression) following supplier’s
protocol (Stratagene, La Jolla, CA).

Plasmid sequencing was carried out by the MSU Genomics Core on an ABI
PRISM® 3730 Genetic Analyzer (Life Technologies Corporation, Carlsbad, CA), and
sequencing comparison was completed using DNASTAR Lasergene software
(DNASTAR, Inc., Madison, WI). The pBadA vector containing the badA gene purified
from the DH5a E. coli cell line with the forward primer 5'-
CTGCTGACGGCGGACGACTACG-3' and the reverse primer 5'-

TTGGTCAGGCCGTGTTCGTCGG-3'.

10

BadA overexpression in E. coli and purification. The overall procedure for

protein overexpression and purification is adapted fi'om previously reported procedures
used in the overexpression and purification of NDTBT,30 with necessary modifications

for optimal expression of badA. The recombinant badA gene was expressed using the
BL21(DE3) E. coli 'cell line transformed with the pBadA vector. As described in the
company’s technical bulletin (Stratagene, La Jolla, CA), the BL21(DE3) cell line
contains a T7 polymerase gene inserted into the bacterial chromosome; this gene is
induced following the introduction of IPTG. Aliquots (5-mL) of LB medium
supplemented with 50 ,ug/mL kanamycin were inoculated with the described E. coli
containing the recombinant badA gene and incubated overnight at 37 °C. The 5-mL

cultures were then used to inoculate l-L cultures of LB medium, also supplemented with
50 pg/mL kanamycin, and grown at 37 °C to an OD600 = 0.7 — 1.0, at which point badA

gene expression was induced with 100 pM IPTG. Cultures were then grown for 18 h at
18 °C. The cells were harvested by centrifugation at 2000g at 4 °C for 20 min; the
supernatant was then discarded. The remaining cell pellet was resuspended in 5-mL of
Lysis Buffer (50 mM sodium phosphate, 300 mM NaCl, and 10 mM imidazole at pH 8.0)
per gram wet weight. Sonication was used to lyse the resuspended cells with a Misonix
XL-2020 sonicator (Misonix Inc., Farrningdale, NY), where the sonicator was set at 50 %
power for six consecutive 20 s bursts with 1 min intervals. Clarification of the cell-lysate
was accomplished through ultracentn'fugation at 149,000g for 1.5 h at 4 °C. The clarified
cell-lysate will henceforth be referred to as the crude lysate. Cell pellets not immediately

used for BadA purification were stored at -20 °C.

11

Purification of the expressed BadA protein from the crude lysate was carried out
using immobilized-metal affinity chromatography (IMAC) in a batch mode. Specifically,
the crude lysate (adjusted to pH 8.0) was incubated with Ni - Nitrilotriacetic acid (NTA)
Agarose resin (QIAGEN, Valencia, CA) for 10 min at 4 °C with constant shaking. The
resin and crude lysate mixture was then transferred to a Poly-Prep® column (BioRad, 0.8
x 4 cm) and the remaining buffer was drained. The resin was then washed with three
column volumes of Lysis Buffer to remove any unbound proteins from the column.
Bound protein was eluted in a stepwise fashion using one column volume of each of the
following buffers: 10 mM, 50 mM, 100 mM, 150 mM, 200 mM, and 250 mM imidazole
in Lysis Buffer.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was

used to identify the eluent aliquots containing the highest percentage of purified BadA
(described by Sambrook).37 Acrylamide gels (12 %) cast in a 0.75 mm x 8.3 cm x 7.2 cm

mold were used, and electrophoresis was completed in a Bio-Rad apparatus (Mini-
PROTEAN Tetra Cell, Bio-Rad) set at 140 V and run until the Coomassie dye reached
the bottom of the gel. Protein samples were prepared by diluting S-pL of sample
(adjusted depending on sample volume) to a total volume of 20-pL with sample buffer,
boiling the solution at 97 °C for 10 min to ensure protein denaturation, followed by
centrifugation to pool the total 20-,uL. SDS-PAGE ladder Protein ProSieve® Markers
(50547; Lonza, Rockland, ME) was used as a protein standard. Gels were then visualized
using Coomassie Blue dye. The aliquots containing purified BadA were then identified
and combined, and imidazole was removed using consecutive dilution and concentration,

accomplished with 30,000 molecular weight cutoff (MWCO) filters (regenerated

12

cellulose membrane, Millipore, Billerica, MA). Centrifugation was achieved at 4,000g at
4 °C for 20 min. Following concentration, the retentate was washed with Assay Buffer
(50 mM sodium phosphate, 5% glycerol, pH 8.0) until imidazole concentration was
reduced to 1.5 pM.

Protein concentration was determined via the Bradford method and comparison to
bovine serum albumin (BSA) standards. A Coomassie (Bradford) Protein Assay Kit
(Pierce, Rockford, IL) was used to prepare all standards and the A600 was used for the
construction of a Bradford assay calibration curve.

After determining final BadA purity and concentration, the enzyme solution was
diluted to 10 mg/mL with Assay Buffer, and 50-yL aliquots, previously flash frozen with
liquid nitrogen, were stored at -80 °C.

BadA activity assay. Following purification using Ni-NTA Agarose resin, the
activity of the BadA enzyme was tested using the following assay conditions, adapted

from previous work by Geissler et al.1 Briefly, a reaction mixture containing 2.5 mM

MgC12, 1.0 mM ATP, 0.25 mM CoASH, 0.25 mM BzOH and 100 pg/mL BadA in a 20

mM triethanolamine—HCI (TEA) buffer at pH 8.0 was prepared on ice. Following
incubation at 31 °C for 10 min, the enzyme activity was quenched with 0.5 N HCl to pH

<1.0.

Because coupled assays with NDTBT were to be completed in the described
Assay Buffer (as this buffer has been used in previous studies of NDTBT),30 the activity

of BadA was compared in both the TEA buffer as well as the Assay Buffer. Results from
these comparisons showed no statistical difference in activity between the two buffers,

hence the Assay Buffer was used in all reported studies on BadA activity.

13

Following the acid quench, an aliquot from each assay was subjected to HPLC

and the eluent was analyzed by ultraviolet (UV) absorbance at 210 nm. Initial assay
analyses were completed using the mobile phase gradient as described by Geissler et al.1
Briefly, separation was achieved via injection with an Agilent 1100 HPLC system
(Agilent) onto a C13 reverse-phase column (AtlantisTM dClg - 3 pm; Waters, Milford,
MA) and elution was completed using a gradient mobile phase composed of solvent

A/solvent B [solvent A: 99.9 % H20 with 0.1 % trifluoroacetic acid (TF A) (v/v); solvent

B: 99.9 % acetonitrile with 0.1 % TFA; gradient: 0 — 40 min at 2 — 4O % (B)].l However,

this method was later altered to allow for more efficient separations without loss of peak
resolution. The altered separation technique was achieved using the same instrument and

column as before, except the following gradient mobile phase composition was used:
[solvent A: 99.9 % H20 with 0.1 % TFA (v/v); solvent B: 99.9 % acetonitrile with 0.1 %
TFA (v/v); gradient: 0 — 5 min at 2 — 17 % (B), 5 — 18 min at 17 — 30 % (B), 18 — 20 min
at 30 — 40 % (B), and 20 - 22 min at 40 - 2 % (B)] with a flow rate of 1.0 mL/min and a
total run time of 22 min. Detection was completed by monitoring the A 210 of the effluent.

Peak identities in a typical Chromatogram were determined through injection of substrate
and product standards.

BzCoA product amounts were quantified through comparison to a calibration
curve constructed using commercially available BzCoA (see Substrates section above).
The calibration curve was constructed using serial dilutions of a 500 ,uM BzCoA stock
solution prepared in distilled, deionized water at pH 5.0 (see Figure 3 on the next page).

An acidic solution was used to stabilize the thioester bond in BzCoA, which is prone to

14

hydrolysis in basic solutions.38 The reported calibration curve had a linear dynamic

range of 0.313 — 10.0 nmol of BzCoA (corresponding to a concentration range of 7.81 —
250 pM BzCoA). At higher concentrations of BzCoA, the analytical signal was no

longer linearly dependent on BzCoA amounts.

 

 

 

 

14000

12000 -
”I:

.2 10000 -
C
D
+3

3 8000 ~
to
9.’
<

.2 6000 ~
to
Q)
Q

3 4000 -
N
<

2000 -

0 I I I r I
0 2 4 6 8 10

BzCoA Amount (nmol)

Figure 3. Calibration curve for BadA product quantitation using A210 Peak Area as
a function of BzCoA Amount, reported in nmol. The 40 ”L injection volume allows
for conversion of BzCoA amount to concentration. Error bars are reported as the
standard deviation of three separate measurements.

While the construction of this calibration curve may seem pedestrian, there were
several initial problems when attempting to quantify product amounts from BadA
reactions. Several attempts at generating a BzCoA calibration curve were made using a

synthesized BzCoA from laboratory stocks. The detailed synthesis of this substrate is

15

reported fully in literature by Nevarez, et al.,30 apparent purity was assigned through

several methods of analysis, including proton nuclear magnetic resonance (lH-NMR),

electrospray ionization — tandem mass spectrometry (ESl-MS/MS) fragmentation, and
thin layer chromatography (TLC) preparation. Issues in this thesis research arose when
reported product amounts of the BadA reaction exceeded possible theoretical yields by up
to 50 %. As the controls run alongside the BadA activity assays showed no residual
BzCoA present in any of the other assay components, including the purified enzyme
fraction, it was hypothesized that the BzCoA used to construct the calibration curve was

not as pure as previously reported(at 95 — 99 %).30 For this reason, 5 mg of BzCoA was

purchased from Sigma to reconstruct a more accurate standard curve. The final
calibration curve reported in Figure 3 was therefore constructed using the commercially
available BzCoA rather than the synthesized version available from laboratory stocks.
BadA substrate specificity. The two T axus acyltransferases used in later studies
exhibit broad substrate specificity with regard to the acyl-CoA substrate. For this reason,
the substrate specificity of BadA was tested using four different acyl-carboxylic acid

substrates, of which the products correspond to four of the reported acyl-CoA substrates
for the T axus acyltransferases.”2 The activity of BadA with the four acyl-carboxylic
acids. 4-methylbenzoic acid, 2-fluorobenzoic acid, 3-furoic acid, and thiophene-2-
carboxylic acid was tested using similar assay conditions to those described for BadA

activity assays. Briefly, 2.5 mM MgC12, 1.0 mM ATP, 0.25 mM CoASH, and 0.25 mM

carboxylic acid were added to a reaction vessel and diluted to final concentrations with
Assay Buffer. The reaction vessel was incubated at 31 °C for 7.5 min, followed by an

acid quench with 0.5 N HCl to pH < 1.0. Controls for these substrate specificity studies

16

were included so that the presence of any de novo peaks could be assigned to putative
product formation; these controls also allowed for identification of acyl-carboxylic acid

peaks in the resultant chromatograms. Separation and analysis was completed using the
non-optimized HPLC method described by Geissler et al.1 Further verification of the

production of the expected acyl-CoAs was completed through the coupled assays with

both NDTBT and mTBT, as discussed in Chapter 2.

1.3: Results and Discussion
Subcloning badA gene from R. palustris genome. The sequence obtained from

the source plasmid pPE20436 matched the extant recombinant badA gene subcloned

herein (Figure 4). These results lend support to the conclusion that the BadA used for

further studies herein is the same enzyme characterized in previous research.

17

atgaatgcagccgcggtCachCgccacccgagaagtttaattttgccgagcaCctgctgcag
M N A A A v T P P P E K P N F A E H L L Q
iccaatch3tgcggCC33acaa3aC33C3ttcgthaCgacatthgtcgctgagcttC3cg
T N R J R P D K T A F V D D I S S L S F A
CaathgaagctCag?CgcgtCagCtcgccgccgccttac3cgcgatcgggitgaaachqaa
Q L E A Q T R Q L A A A L R A I G V K R E
gagcgcgtcctgctgctgaLgCLCgacgng aCc 3gattggCCg '3tggcgtttctcggcgcaatc
E R V L L L M L D G T D W P V A P L G A I
tanccggcatcgtgccggtcgcggtcaata chthtgangng acgactacg CCtacatg
Y A G I V P V A V N T L L T A D D Y A Y M
thgagcattCQCOqjxLCQGJCTITQC ggtcagc ggcgcqctgc aCC cggtgctcaaggca
L E H S R A Q A V L V S G A L H P V L K A
gcgctgaccaagagcgatcacgaggtgcagc gagtgatcgtttcgc3cccagcggcthgctg
A L T K S D H E V Q R V I V S R P A A P L
gagccgggcgaggtcgacttcgct3ag ttchC ggC 3cacatgcgccgcttgag:agccthC
E P G E V D F A E F V G A H A P L E K P A
gctacgcaagcggacgatCng:gttctggctgtattcgtcgggttctaccgggcggccgaag
A T Q A D D P A F W L Y S S G S T G R P K
ggcgthHtg aca ctca~3 gCCE atCthaCtgqaccthqqq tgtacggccgcaacacgctg
G V V H T H A N P Y W T S E L Y G R N T L
CatCtgcgcgaagaC gaC gtctg-C ttttcggcggC caaactgt tt ttcgcttacggcctcggc
H L R E D D V C P S A A K L F F A Y G L G
aacgcgcthcgtttccqeLganqtcggCgcgaccacgctgctgatgggcgagcgaccgacg
N A L T F P M T V G A T T L L M G E R P T
CC33aC3ngtgttcaa3CgctqgctcggcggcgtcggcggtgtgaaaccgaCCgtgttctaC
P D A V F K R W L G G V G G V K P T V F Y
ggcgcgcccaccg3ctaccccggCatgttggccgcgccgaacctgccgtcgcgcgaccaggtg
G A P T G Y A G M L A A P N L P S R D Q V
gcgttgcggctcgcgt C3tcg3 g3gC3aa 3caCtgccggcggagattgggcagcgcttccag
A L R L A S S A G E A L P A E I G Q R F Q
chcatttC33CCtcqacathtcaatggcatcggctcgachagatgctgcacatctttctg
R H F G L D I V D G I G S T E M L H I F L
tcgaacctgccagaccgqgtgcgctacggcaccaccggaCCgccggtgccgggctatcagatc
S N L P D R V R r G T T G W P V P G Y Q I
gagctgchggC3angngngaCngtcgccgacggaqagccgggCgatctctaCattcaC
E L R G D G G G P V A D G E P G D L Y I H
ggcccgtcatcggcgacgatrtattq3ggcaaccgggccaaqagccgcgaca CC ttccagggc
G P S S A T M Y W G N R A K S R D T P Q G
ggctggacczagagccgcgaca aaa atacgtCC3C aacgacgacggctcctacacctatgcgqgc
G W T K S G D F Y V R N D w G S Y T Y A G
cgcaccgacgacathtqaa33tC agcggca atCt CTgTCPQPer .tC cgagatc 3a agcgacg
R T D D V L K J S G I Y V S P F E I E A T
ctggtgcagcatcccggtgtgctha agccgcagtgctcggggtCgccgacgaacacggcctg
L V Q H P G V L E A A V V G V A D E H G L
accaaaCCgaa_i3 t5t3t33t3‘C3 93 MCchgCCagaccctgtcggagaccgagctgdag
T K P K A Y V V P R P G Q T L S E T E L K
accttcatcaaggatC 3aCthC3/Cg aC aaatatccgcgcagcacggtgttcgtcgccgaa
T F I K D R L A P Y K Y P R S T V F V A E
ttgccgaagacqgcgaccggcaagaLtca_cgcttcaagctgcgcgagggt3tg ttggg: tga
L P K T A T G K I Q R F K L R E G V L G -

Figure 4. The nucleotide (5'-3' top strand) and amino acid (N- to C-terminus
translated bottom strand) sequences obtained from the recombinant badA gene.
Sequencing alignment (not shown) with the cDNA sequence of badA in the

source vector pPE20436 showed identical sequences.

18

BadA overexpression in E. coli and purification. SDS-PAGE analysis of the
seven separate elutions, each with different imidazole concentrations, was completed
with comparison to the aforementioned protein ladder (Lonza). The following image

displays a sample SDS-PAGE gel used to determine protein purity and molecular mass.

 

Figure 5. SDS-PAGE analysis of four of the seven separate elutions collected during
BadA purification. Lanes: 1, elution at 100 mM imidazole; 2, elution at 150 mM
imidazole; 3, elution at 200 mM imidazole; 4, elution at 250 mM imidazole in Assay

Buffer; 5, Protein ProSieve® ladder. The two bands in the ladder closest to the
putative BadA band are labeled with their respective molecular masses (75 and 50
kDa). The band in Lane 4 was used to estimate BadA mass at ~58 kDa.

In Figure 5, it can be seen that a protein eluting at a molecular mass of ~58 kDa

was present in several of the eluents from the purification of BadA. This molecular mass
corresponds to that of the previously characterized BadA.1 The eluents corresponding to

Lanes 2, 3, and 4 (150 mM, 200 mM, and 150 mM imidazole in Assay Buffer) were
combined with an estimated purity of 95 %, as determined using the SDS-PAGE analysis

(see Figure 5). Following buffer exchange and imidazole removal, the total amount of
the putative N-terminal His6-epitope of BadA isolated throughout the prescribed

purification process was 6.25 mg BadA per L of harvested culture. While SDS-PAGE

19

analysis does show that a protein at molecular mass ~58 kDa was successfully purified, it
does not conclusively prove the identity of the protein. The following described activity
assays were used to lend further support to the identity of the ~58 kDa protein as the
hypothesized BadA.

BadA activity assay. A typical Chromatogram obtained from the analysis of

BadA activity assays as described above is shown below in Figure 6.

 

 

 

 

 

 

 

 

 

2500
ATP
2000 - 2.0 min
NA
42
5 1500 .
e’
S.
8 1000 -
2 BzCoA
g 11.23 min
G. q
o 500
H
N
< l
0 JM V A
‘500 I T T
0 s 10 15 20

Retention Time (min)

Figure 6. Chromatogram obtained following the analysis of a BadA activity assay.

Assay conditions were as follows: 2.5 mM MgClz, 1.0 mM ATP, 0.25 mM CoASH,

0.25 mM BzOH, 100 pg/mL BadA, incubated at 31 °C for 10 min followed byacid
quench.

In the activity assay described above, BzCoA was produced at a 91.5 % yield. This yield

was calculated based on comparison of available starting material (0.25 mM BzOH and

20

CoASH) with the product amount quantified using the calibration curve displayed in
Figure 3 of the Experimental Section. In the above Chromatogram, a cluster of peaks
near 2.0 min was observed in all activity assays. When compared to controls, these peaks
were found to have the same retention time as solutions containing ATP; hence, this

cluster of peaks is most likely due to ATP as well as AMP, another by-product of the

. 1
BadA reaction.

For any given set of activity assays completed after purification of the putative
BadA, several controls were run to ensure that any observed BzCoA product could be

attributed entirely to the activity of the enzyme (Table II).

Table II. Series of assays, including controls, used to determine BadA activity after
purification. Y indicates inclusion in the reaction mixture; N indicates absence from
the reaction mixture; N/D represents “Not Detectable.”

 

 

 

 

 

 

 

 

 

Component Activity BadA MgClz ATP ”$331,
Concentration Assay Control Control Control an
I Control
2.5 mM MgC12 Y Y N Y N
1.0 mM ATP I Y Y Y N N”
0.25 mM
BzOH Y Y Y Y Y
0.25 mM
CoASH Y Y Y Y Y
100 ,ug/mL
BadA Y N Y Y Y
Observed 0.23 mM N/D 0.16 mM N/D N/D
BzCoA ,

 

 

 

 

 

Not included within the table are controls where BzOH and CoASH were either

separately or simultaneously excluded from the overall reaction mixture. All three of

21

 

these controls gave no detectable BzCoA product. The results of these assays and

controls show that any observed product within a given activity assay is due entirely to

the activity of the putative BadA protein. It should be noted that when MgC12 was not

included in the assay (MgC12 Control), BzCoA was still observed as a product. While
this may be the case, the activity of the enzyme is reduced by 31 % or more without the

presence of externally introduced MgClz. One possible explanation for this observed

activity is that residual Mg2+ could be coordinated with BadA throughout purification
procedures. Because previous studies on BadA showed no inhibition of the enzyme with

MgClz at 2.5 mM,1 this experimental observation was noted but not deemed a cause for

concern. The addition of MgC12 to the assays improved overall turnover, hence it was
routinely used in all subsequent assays of BadA.

The results of both the SDS-PAGE analysis and the activity assays lend strong
support to the argument that the purified protein with molecular mass ~58 kDa does
correspond to the previously reported BadA. Additionally, the sequence alignment of the

badA cDNA in the source plasmid pPE20436 with the recombinant plasmid pBadA show

identical nucleotide sequences, and confirmed the introduction of an N-terminal His6-tag

allowing for Ni — NTA purification.

BadA substrate specificity. The previously published data with regards to the
substrate specificity of BadA showed that BadA catalyzes the formation of the thioester
bond between several carboxylic acids and CoASH. Included in the previous

investigation was the analysis of the relative velocity of BadA with respect to the natural

substrate BzOH of both 2-fluorobenzoic acid and 4—methylbenzoic acid.l These

22

productive substrates were utilized in the present investigation, along with 3-furoic acid
and thiophene-Z-carboxylic acid, which were analyzed for the first time with BadA.

For this reason, the activity of BadA with respect to the four additional substrates
discussed was assessed via the assay conditions described in the Experimental section. It
should be noted that this data analysis did not allow for quantitation of product amounts
as product standards for the acyl-CoA products were not available for the construction of
standard curves. Following HPLC separation and UV analysis, the presence of de novo
peaks was taken as evidence of the production of the putative acyl-CoA product. Further
verification of these products was completed through coupled assay activity, as discussed
in the next chapter. The observed retention times for the putative acyl-CoA products of

the BadA reaction with non-natural carboxylic acid substrate are listed in Table III.

23

Table [11. Retention times for the putative products of the BadA catalyzed reaction

when incubated with the reported acyl-carboxylic acid donor.

 

Acyl-carboxylic Acid
Substrate

Putative Acyl—CoA
Product

Retention Time

 

E O
O
I

Q...

 

 

 

 

 

SCoA 20.2 min
Benzoic Acid Benzoyl-CoA
O O
/©/U\OH fiscm 27.8 min
4-Methylbenzoic Acid 4-Methylbenzoyl-COA
O 0
do“ [XLSCOA 20.5 min
F F
2-Fluorobenzoic Acid 2-Fluorobenzoyl-CoA
O O
Xvi/[kOH oéJkSC“ 15.4 min
/ /
3-Furoic Acid 3-Furanoyl-COA
O 0
S S
\ / 0” \ / SC°A 18.3 min

Thiophene—Z-carboxylic
Acid

 

Thiophene-Z-carbonyl-CoA

 

 

 

As discussed in the Experimental section above, separation and detection of this finite
substrate specificity study was completed using a non-optimized HPLC method. For this
reason, the retention time for the natural product, BzCoA, (Table III) does not correspond
to the same retention time as the Chromatogram displayed in Figure 6 (see page 20),
which was obtained via the optimal HPLC method.

The result of this limited substrate specificity study for BadA confirms the
previously reported broad substrate specificity of the enzyme, and also lends evidence
demonstrating that BadA is active with the two previously untested substrates 3-furoic
acid and thiophene-Z-carboxylic acid. As a result of these assays, it was hypothesized
that coupling the activity of BadA with NDTBT or mTBT using non-natural acyl-
carboxylic acid donors would result in the production of acyl variants of paclitaxel or
baccatin III. The results in Chapter 2 show that BadA must be active with all of the
tested non-natural substrates as the expected, substituted paclitaxel and baccatin III

analogues were observed.

1.4: Conclusions

The first step in the proposed coupled assays depends on the activity and broad
substrate specificity of BadA, thus obtaining a functional and purified BadA was a key
step in furthering this research. The successful construction of the recombinant plasmid
pBadA was confirmed with sequencing alignment, which offers strong evidence that the
BadA expressed via transformed E. coli BL21(DE3) cells matches the amino acid

sequence of the BadA enzyme previously studied by researchers at the University of

Washington."36 The molecular mass of a purified protein obtained through the use of Ni-

25

affinity chromatography at ~58 kDa matched the molecular mass of the protein BadA.
Additionally, activity assays obtained through incubation of the ~58 kDa protein with the
natural substrates of the reported BadA enzyme (BzOH and CoASH) led to the
observation of BzCoA production as a function of the protein solution. These
experimental results lend strong support to the conclusion that the purified, functionally

expressed protein has the same amino acid sequence and enzymatic functionality as the
previously studied BadA."36 However, the most important aspect of BadA activity, that

of production of BzCoA, was confirmed with multiple concentrations of enzyme and
substrates. Only when enzyme concentrations were below 1.0 ,ug/mL in an assay was the
observed product yield affected; typical yields of BzCoA were between 90 —— 95 % at 5
min. These results led to the initial studies with coupled assay activity.

Another important aspect of the proposed coupled assay system lies in the
hypothetical ability to produce substituted paclitaxel or baccatin 111 products. As the
product of the BadA reaction is the acyl-CoA substrate for either of the Taxus
acyltransferases used in later experiments, the possible substituted products depend on
the ability of BadA to produce the necessary acyl-CoA substrate. The limited substrate
specificity study performed in this research shows that, when incubated with an acyl-
carboxylic acid, BadA catalyzes the production of a de novo compound. These novel
compounds are most likely the hypothesized acyl-CoA products, and later evidence will

show this to be the case.

26

CHAPTER 2
Coupled Activity of BadA with the T axus Acyltransferases NDTBT and mTBT
2.1: Introduction

The production of paclitaxel as a natural product in T axus spp. involves, in part,
the activity of five separate acyltransferases, each of which requires a coenzyme A-
activated acyl group as a substrate.34 While these enzymes have been characterized,
there have been few studies exploring their utility as biocatalysts for the production of
paclitaxel or second generation paclitaxel analogues.39 As discussed in the Introduction,
the implementation of such a proposed biocatalytic system depends greatly on the
availability of substrates either in vitro or in vivo. The research presented in this chapter
focuses on coupling the activity of a broadly-specific BzCoA ligase to the activity of two
separate T axus acyltransferases in vitro.

Displayed on the subsequent pages, Figures 7 and 8 are the proposed coupled
reaction schemes for the coupled activity of BadA with NDTBT, and with mTBT. In the
scheme, there are several important experimental observations listed that helped in the
development of the coupled assay system. Firstly, the kcat (turnover) values for both
enzymes are reported. Comparison of the turnover number (i.e. product biosynthesized
per unit time) showed that BadA operates at a much faster rate than does NDTBT,

suggesting that coupled assays conducted herein are limited by the T axus acyltransferase

reaction.

27

0 . _1
)L BadA, kcat = 1570 mm

 

                 

R OH > AMP + PPi
0
CoASH R’lksc o A
AcO 0
R OH kcat‘ 1.6 min"1
02‘ng o a
i O
0““ a i
H0 082 OAc
Substituted paclitaxel product N-debenzoyI-2'-deoxypaclitaxel

 

0‘1 o 0: :1 0

Figure 7. A reaction scheme describing the coupled activity of BadA and NDTBT.
The N-debenzoyl-2'deoxypaclitaxel is pictured as it is considered the natural
substrate for NDTBT. Both the km for BadA1 and NDTBT (reported in this thesis)
are listed with their respective natural substrates. The five separate acyl donors '
displayed describe the substrates that were investigated.

 

Benzoic acid (likely convergent to the paclitaxel pathway in T axus), along with four
surrogate carboxylic acids were tested as substrates in each of the experiments presented
in this document. This information provided a primer for assessing possible analogues

that could be produced via the coupled reaction scheme.

28

 

(,1: Oil;\ Rj: BadA,kcat=1560 min'1

 

 

 

 

OH > AMP + PPi
R2: 011 Mg , ATP 0
“ASH RJkSCoA
R mTBT
. -1

      

a?

“0 5H 6Ac
7,13-diacetyI-2-debenzoylbaccatin Ill

AcO‘“

 

Substituted 7,13-diacetylbaccatin III product

Figure 8. A reaction scheme describing the coupled activity of BadA and mTBT.
7,13-diacetyl-2—debenzoylbaccatin III is pictured, and it is considered to be a

surrogate substrate for mTBT. Both the km for BadA1 and mTBT2 are listed with
their respective natural substrates. The three separate acyl donors displayed
describe the substrates tested with mTBT.

2.2: Experimental

General. Most of the general instruments used in this chapter were identical to
those used in Chapter 1. A Q-ToF Ultima Global electrospray ionization tandem mass
spectrometer (ESI—MS/MS, Waters, Milford, MA) with a Waters CapLC capillary HPLC
was used for mass spectral analysis.

Substrates. The substrates necessary for BadA activity are identical to those
reported in the previous chapter. Baccatin III was purchased from Natland (Research

Triangle Park, NC). The syntheses of laboratory stocks of N—debenzoyl-(3'R)-2'-

29

deoxypaclitaxel (NDB2DT), 2'-deoxypaclitaxel (2DOT), N-debenzoylpaclitaxel, and
7,13-diacetyl-2-debenzoylbaccatin III are described in the literature. 2’30 The former two

substrates were used in NDTBT and NDTBT/BadA coupled assays, while the latter was
used in assays with mTBT.

Bacterial strains and culture components. The same LB media, kanamycin,
and IPTG solution conditions and concentrations were used for NDTBT expression as
those for BadA expression. In addition, the antibiotics ampicillin (Roche) and
chloramphenicol (Sigma) were added to media, where appropriate, to final concentrations
of 50 ,ug/mL and 34 pg/mL, respectively. Stock solutions of ampicillin were prepared in
water while stock solutions of chloramphenicol were prepared in 100 % ethanol.
Antibiotic and IPTG stock solutions were filter-sterilized through a O.22-,um membrane

(Millipore).

2.2.1: NDTBT Experimental

NDTBT subcloning. Ongoing research with the enzyme NDTBT and work

towards obtaining a crystal structure for the enzyme has led to the use of a different
recombinant vector for expressing ndtbt than described previously.30 The cDNA ndtbt
was subcloned into the pET28a vector (Novagen) to incorporate a C-terminal His6-tag,

and designated ndtbt-ct.
NDTBT overexpression in E. coli and purification. Recombinant ndtbt-ct was

expressed in the described bacterial expression system and harvested according to
previous literature, with some modifications.30 Cultures were grown overnight at 37 °C

in lOO-mL of LB medium supplemented with 50 ,ug/mL kanamycin and 34 pg/mL

3O

chloramphenicol. Bacteria transformed with empty vector were processed analogously.
To six-850 mL portions of LB medium supplemented with the appropriate antibiotics

were added 16-mL of the lOO-mL inoculum; the flasks were incubated at 37 0C until the

cells reached OD600 = 0.5-0.7. Gene expression was induced with 100 yM IPTG, and the

culture was incubated at 20 °C for 18 h. The cells were harvested by centrifugation at
4000g for 10 min at 4 °C, the supernatant was discarded, and the pellet was stored at -20
°C.

The pellet was resuspended in 3 mL NDTBT Lysis Buffer (50 mM
tris(hydroxymethyl)aminomethane (Tris) -HCl, 300 mM NaCl, 10 % glycerol, pH 8.0)
per wet weight at 4 °C, and sonicated with a Misonix XL-2020 sonicator (Misonix) set at
60 % power for six consecutive 15 s bursts with 45 5 intervals. The cell lysate was
clarified by ultracentrifugation at 149,000g for l h at 4 °C to provide the crude lysate.

The crude lysate was incubated with 1 mL of HIS-Select Nickel Affinity Gel
(Sigma) per 8 g of wet pellet in batch mode at 4 °C for 1 h. The mixture was poured into
an Econo column (BioRad, 20 cm X 2.5 cm). After the resin settled, the head volume
was drained, the resin was washed with five column volumes of NDTBT Wash Buffer
(50 mM Tris-HCl, 300 mM NaCl, 10 % glycerol, pH 8.0) to remove any unbound protein
from the column, and the bound protein was eluted with 1.5 column volumes of each of
the following buffers: 10 mM, 50 mM, 100 mM, 150 mM, 200 mM, and 250 mM
imidazole in NDTBT Wash Buffer. SDS-PAGE was used to identify the fractions
containing the highest percentage of purified NDTBT.

Samples to be loaded onto the SDS—PAGE gel were prepared by combining 5 ,uL

of enzyme elution with 10 ,uL of SDS-PAGE sample buffer; samples were subsequently

31

boiled, centrifuged, and loaded onto the gel. Usually the recombinant protein eluted in
the 100 mM and 150 mM imidazole eluent aliquots. The aliquots containing purified
NDTBT were then combined, and imidazole was removed using consecutive
dilution/concentration cycles with a 30,000 MWCO filter (regenerated cellulose
membrane, Millipore, Billerica, MA). Centrifugation was completed at 4,000g at 4 °C
for 20 min. The concentrated sample was washed with NDTBT Storage Buffer (50 mM
Tris-HCl, 10 % glycerol, pH 8.0) until the imidazole concentration was 1.5 ,uM.

The protein was diluted into 50 mL of NDTBT Storage Buffer before loading
onto a column packed with 40 mL of Q-Sepharose Fast Flow (Amersham) that was
previously equilibrated with Storage Buffer. Proteins were eluted with a linear NaCl
gradient (0-400 mM; 5 mL/min; 33 min), and the bulk of the target enzyme was found in
fractions containing 90-150 mM NaCl. These fractions were combined and NaCl was
removed by consecutive dilution/concentration cycles as before, using a 30,000 MWCO
filter. The protein solution was diluted in NDTBT Gel Buffer (50 mM Tris-HCl, 150
mM NaCl, 10 % glycerol, pH 8.0) to a final concentration of 2 mg/mL.

At least 10 mg of NDTBT protein was loaded onto a self-packed gel filtration
column (150 mL of Superdex 200 prep grade in a XK16/7O column (GE Healthcare)) and
eluted at 1 mL/min for 2.5 h in Gel Buffer. The enzyme eluted between 75-90 min and
was judged to be ~95 % pure by SDS-PAGE and Coomassie blue staining, followed by
quantitation on Kodak lD Image Analysis Software (Version 3.6.3). Fractions
containing the protein were combined and NaCl was removed by consecutive

dilution/concentration cycles as before, and the protein solution was diluted to 2 mg/mL

32

in NDTBT Storage Buffer, as determined by the Bradford method. This solution was
then separated into 100 ,uL aliquots, flash frozen in liquid nitrogen, and stored at -80 °C.
NDTBT activity assays and analysis. Functional expression of NDTBT was
verified using assays with the following composition: 500 ,ug/mL purified NDTBT, 200
pM BzCoA, and 200 pM NDB2DT, diluted to a final volume of 200 ,uL with Assay
Buffer. The assays were incubated at 31 °C for 20 min, and the reactions were quenched
by the addition of 0.5 N HCl, reducing the pH below 1.0. The acid quench also ensured
that the free primary amine of the NDB2DT substrate was protonated to its ammonium
ion form so that it would partition into the aqueous fraction during extraction. The assay
was then extracted with ethyl acetate (3 x 1 mL), after which the organic fractions were
combined and the solvent was evaporated under vacuum. The remaining residue was
then resuspended in 100 ,uL of acetonitrile. The resuspended organic fraction (50 ,uL)
was then injected onto a reverse phase column (Allsphere ODS-25 pm, 250 mm X 4.6

mm, Alltech, Mentor, OH) and eluted by a gradient mobile phase composition with the

following solvent A/solvent B ratios [solvent A: 99.9 % H20 with 0.1 % TFA (v/v);

solvent B: 99.9 % CH3CN with 0.1 % TFA (v/v); gradient: 0 - 5 min at 80 % (A), 5 - 11
min at 80 - 50% (A), 11 - 21min at 50 —20 % (A), 21 - 23 min at 20 — 0 % (A), 23 - 25
min at 0 % (A), 25 - 33 min at 0 — 80 % (A)] at a flow rate of 1.5 mL/min and A228

monitoring of the effluent on an Agilent 1100 HPLC system (Agilent Technologies,
Wilmington, DE) connected to a UV detector.

Products eluting from the HPLC column corresponding to a de novo UV-
absorbance peak, identified by comparison with assay controls, were collected and

further analyzed by liquid chromatography — electrospray ionization mass spectrometry

33

(LC-ESI-MS), in positive ion mode. A Q-ToF Ultima Global ESI-MS/MS (Waters,
Milford, MA) with a Waters CapLC capillary HPLC was used for separation and mass
spectral analysis. A C18 reverse phase column (Betasil C18, 5 pm, 150 x 2.1 mm,
Therrno Scientific Inc, Waltham, MA) was used during the HPLC separation. Briefly,
from the approximate 250 ,uL collected from the HPLC effluent, 10 ,uL were injected
onto the LC column coupled to the ESI-MS/MS detector. Because the effluent from the

Agilent HPLC separation contained trifluoroacetic acid (TFA), a known ionization
suppressant in ESI,40 TFA was separated from the other analytes of interest using a

gradient mobile phase composition with the following solvent A/solvent B ratios [solvent
A: 99.9 % H20 with 0.1 % formic acid (v/v); solvent B: 99.9 % CH3CN with 0.1 %

formic acid (v/v); gradient: 0 - 5 min at 98 % (A), 5.00 -— 5.01 min at 98 — 50 % (A), 5.01
— 10 min at 50 - 0.0 % (A), 10 — 10.5 min at 0.0 % (A), 10.50 - 10.51 min at 0.0 — 98 %
(A), 10.51 — 13 min at 98 % (A),] at a flow rate of 0.3 mL/min. The effluent was
diverted for the initial 5.0 min to allow for complete removal of TFA from the sample;
from 5.0 — 13 min, the effluent was analyzed using the first-stage mass spectrometer.
The mass spectrum obtained from the most abundant peak in the resultant Chromatogram
had a molecular weight and ionization pattern that were identical to those of authentic 2'-
deoxypaclitaxel when analyzed by similar methods.

Kinetic evaluation of NDTBT with substrate N—debenzoyl—(3'R)—2'-
deoxypaclitaxel (NDB2DT). To establish linearity with respect to time and enzyme
concentration, varying amounts of NDTBT were incubated with BzCoA (the natural acyl
group donor, at 100 ,uM), and NDB2DT was maintained at saturation (500 ,uM) in 1.6 mL

of Assay Buffer (50 mM sodium phosphate, 5 % glycerol, pH 8.0). Aliquots (200 yL)

34

from each assay were collected, quenched with 80 yL of 0.5 N HCl at 5, 10, 20, 30, 40,
and 60 min time points, and extracted with ethyl acetate (3 X 1 mL). Immediately prior
to the acid quench, 4 nmol of baccatin III was added to the reactionmixture as a control
for extraction and injection efficiency during further assay processing. Solvent was

removed in vacuo, and the products in the resultant residue were dissolved in 100 ,uL of
acetonitrile and were analyzed by UV-HPLC with A228 monitoring of the effluent, as

described for NDTBT activity assays. The peak area corresponding to the biosynthetic
2'—deoxypaclitaxel (2DOT) was converted to concentration by comparison to a calibration
curve obtained through serial dilution of authentic 2DOT product standard (linear
dynamic range fi'om 0 to 1.25 mM; concentrations greater than 1.25 mM showed loss of
linear analytical signal dependence). Kinetic parameters were determined under steady
state parameters using 50 pg/mL of protein and a 10 min incubation time. The

concentration of BzCoA was independently varied (0-500 ,uM) in separate assays while
NDB2DT was maintained at apparent saturation (500 ,uM). The initial velocity (v0) was
plotted against substrate concentration for each data set, which exhibited the expected
Michaelis-Menten curve. The Hanes-Woolf method of analysis ([BzCoA]/v0 vs.
[BzCoA]) was used, and the equation of the best-fit line (R2 = 0.97) was determined
(Microsoft Excel 2003, Microsoft Corporation, Redmond, WA) to calculate the kcat and
KM parameters.

The effect of CoASH on the NDTBT activity was determined using an inhibition
study. Assay compositions were identical to those used to determine the kcat and KM

described above, except for the addition of a single concentration of CoASH, which was

35

established between 50 ,uM and 500 ,uM, in separate assays. The assays were processed
by methods identical to those used above, and the Hanes-Woolf method was used to
determine the effect of CoASH on NDTBT activity.

Optimizing coupled activity assays for NDTBT and BadA. Initial coupled
activity assays were completed using assay concentrations similar to those used for the
separate assays of NDTBT and BadA activity. In addition to the proposed coupled assay

system, an activity test for NDTBT was run as a control, where the BzCoA concentration
was set at 0.25 mM. Briefly, a 500 pL reaction mixture containing 2.5 mM MgC12, 1.0

mM ATP, 0.25 mM CoASH, 0.25 mM BzOH, 0.50 mM NDB2DT, 500 pg/mL NDTBT,
and 200 ,ug/mL BadA, prepared on ice, was diluted to final concentrations with Assay
Buffer. The assays were incubated at 31 °C for 18 h, after which, enzyme activity was
quenched by the addition of 0.5 N HCl, reducing the pH below 1.0. Separation and
analysis of these initial coupled assays was carried out according to the procedure used to
analyze NDTBT activity assays.

Results from the kinetic studies completed on NDTBT activity were used to
modify the coupled assay to yield optimal turnover. The following assay conditions were

considered optimal after several variations were completed. A reaction mixture with total
volume 200 ,uL containing 2.5 mM MgC12, 1.0 mM ATP, 0.4 mM BzOH, 0.4 mM

CoASH, 0.5 mM NDB2DT, 200 ,ug/mL NDTBT, and 50 pg/mL BadA, prepared on ice,
was diluted to final concentrations with Assay Buffer. The assays were incubated at 31
°C for 2 h, and all remaining steps were completed as before. A series of assays, as

discussed in the Results and Discussion section, was used to determine the effects of

36

BzCoA production through BadA activity on observed 2'-deoxypaclitaxel product
amounts.

An analogous set of assays were used to determine whether CoASH could be
effectively recycled using the coupled assay system wherein CoASH was added at a

concentration 4- to 8-fold led than in previous assays. A reaction mixture with total
volume 200 ,uL containing 0.5 mM NDB2DT, 2.5 mM MgC12, 1.0 mM ATP, 50 ,uM

CoASH, 500 ,ug/mL NDTBT, and 200 yg/mL BadA was diluted to final concentrations
with Assay Buffer. The assays were incubated for 18 h, followed by an acid quench, as
described previously. In one assay, the BzOH concentration was equimolar to the
CoASH concentration (50 ,uM) while in the second assay, the BzOH concentration was
twice that of CoASH (100 ,uM).

Coupled NDTBT and BadA activity with non-natural carboxylic acid and N-
debenzoylpaclitaxel substrates. Coupled activities with the non-natural carboxylic acid
donor substrates for the BadA reaction discussed in the previous chapter were also
incubated in the coupled enzyme system. These assays were completed to assess whether
the coupled enzyme assay could biosynthesize various N—substituted-Z'-deoxypaclitaxel
analogues. As these assays were completed prior to the optimization of the coupled assay
system, the concentrations of the various reaction components were not identical to those

described above. A reaction mixture with total volume 250 ,uL containing 2.5 mM
MgClz, 1.0 mM ATP, 0.25 mM BzOH, 0.25 mM CoASH, 0.25 mM NDB2DT, 250

pg/mL NDTBT, and 250 pg/mL BadA, prepared on ice, was diluted to final
concentrations with Assay Buffer. Assays were incubated at 31 °C for 18 h, after which

enzyme activity was quenched by the addition of 0.5 N HCl (pH < 1.0). Extraction and

37

initial separation on the Agilent 1100 HPLC (described previously) was performed as
before in the NDTBT activity assays. As before, products eluting from the HPLC
column corresponding to a de novo UV-absorbance peak, identified via comparison with
assay controls, were collected and further by analyzed LC-ESI-MS in positive ion mode.
Peak identity was confirmed through the characterization of MS fragmentation, as
previously observed for authentic 2'-deoxypaclitaxel product standard. In addition, the
mass spectra for the products of this coupled assay system with non-natural carboxylic

acid substrates correspond to the previously characterized products of the NDTBT
enzymatic reaction when incubated with the corresponding acyl-CoA substrate.3O

As shown previously, NDTBT is active with several N-debenzoyl-taxane variants.

The substrate N-debenzoylpaclitaxel is important as the hydroxylation at the 2'-position
on the C13 side chain of paclitaxel is hypothesized to occur prior to N—benzoylation.41
Coupled assays completed with this taxane substrate were completed with the following
compositions: 1.25 mM N-debenzoylpaclitaxel, 2.5 mM MgC12, 1.0 mM ATP, 1.0 mM

BzOH, 1.0 mM CoASH, 250 ,ug/mL NDTBT, and 250 ,ug/mL BadA. Assays were
incubated for 18 h at 31 °C, followed by the same assay processing used for NDTBT
activity assays as well as NDTBT and BadA coupled activity assays. However, assays

were analyzed by direct injection LC-ESI-MS/MS analysis (as described in the

literature”) rather than the HPLC separation and UV analysis described in this document.

The expected product ion [M+H]+ (m/z = 854) was directed to a fiagmentation cell, and

the second mass spectrometer was used to scan the array of fragrrrent ions.

38

2.2.2: mTBT Experimental

Bacterial strains and culture components. The same LB media, kanamycin,
and IPTG solution conditions and concentrations were used for mTBT expression as were
used for both NDTBT and BadA expression. E. coli BL21(DE3) cells previously

transformed with p28PK-TBT were used to overexpress the point-mutant tbt as described
in the literature.2 The two point mutations Q19P and N23K introduced in the tbt gene

isolated from cDNA obtained from T. cuspidata allowed increased soluble expression of
the modified taxane-2-0-benzoyltransferase, and the subsequent protein was renamed
modified taxane-2-O-benzoyltransferase, or mTBT.

Overexpression of mTBT in E. coli and purification. Laboratory stocks of
functional mTBT were used throughout the experiments described in this chapter. These

stocks were prepared according to the previously described protocols, with some
exceptions.2 In brief, E. coli BL21(DE3) transformed with p28K-TBT was used to

inoculate a 100-mL culture containing LB medium supplemented with 50 pg/mL
kanamycin, which was grown overnight at 37 °C. Aliquots (5-mL) from this 100 mL
culture were then separately used to inoculate 1 L of LB media along with the appropriate

antibiotic at 37 °C until an OD600 = 0.95 — 1.05 was reached. At this point, mtbt

expression was induced with 50 pM IPTG and the cultures were incubated at 18 °C for
16 h. As before, cultures were centrifuged at 4000g for 20 min at 4 °C to harvest the
cells. Cell pellets were resuspended in mTBT Lysis Buffer (50 mM sodium phosphate,
pH 8.0) with 3 mL/g wet cell pellet weight. Sonication at 4 °C was used to lyse the
resuspended cells via a Misonix XL-2020 sonicator (Misonix) where the sonicator was

set at 50 % power for five consecutive bursts with 1-min intervals followed by sonication

39

at 70 % power for two consecutive bursts with a 2-min interval. Clarification of the cell
lysate was accomplished through ultracentrifugation at 149,000g for 1 h at 4 °C; the
supernatant from this centrifugation will be referred to as the soluble enzyme fraction.

Purification of mTBT from the soluble enzyme fraction was completed in a two-
step fashion. First, the soluble enzyme fraction was loaded onto a Whatman DE-52 anion
exchange column (2.5 x 6 cm, 15 g resin) to remove any remaining small molecules or
cellular debris. The remaining enzyme was eluted from the column using a 200 mM
NaCl solution (in mTBT Lysis Buffer). The resultant eluted enzyme fraction was then
further purified through incubation with His-Select Nickel Affinity Gel (3 g, Sigma-
Aldrich) at 4 °C in batch mode. Following an hour of incubation, the mixture was poured
into an Econo column (Bio-Rad, 20 x 2.5 cm) and remaining buffer was drained. The
resin was then washed with seven column volumes of mTBT Wash Buffer (300 mM
NaCl, 20 mM imidazole in mTBT Lysis Buffer). Bound protein was eluted in a stepwise
gradient with 1.5 mL of a solution containing 200 mM imidazole (with 300 mM NaCl in
mTBT Lysis Buffer) followed by elution with 1.5 mL of a solution containing 100 mM
imidazole (with 300 mM NaCl in mTBT Lysis Buffer).

As with both BadA and NDTBT, imidazole was removed through consecutive
dilution/concentration cycles via centrifugation (30,000 MWCO, regenerated cellulose
membrane, Millipore). Upon imidazole removal, protein was diluted in mTBT Assay
Buffer (25 mM 3-(N-morpholino)-2-hydroxypropanesulfonic acid, 5 % glycerol (v/v), pH

7.4). As stated in literature,2 SDS-PAGE analysis revealed a 30 % pure protein at

molecular mass ~50 kDa, as determined by comparison to molecular mass standards

40

(Lonza). Total protein concentration was determined by the Bradford method (Pierce,
Rockford, IL).

mTBT activity assays and analysis. mTBT activity was tested using similar
assay conditions to those reported previously with some modifications.2 Briefly, a 500

,uL reaction mixture containing 500 ,uM 7,13-diacety1-2-debenzoylbaccatin III, 100
,ug/mL total protein, and 500 pM BzCoA, diluted to final concentration with mTBT
Assay Buffer, was incubated for 15 h at 31 °C. The assay was then moved to ice and 2 x
1 mL of ethyl acetate were used to quench the enzyme activity and extract the product,
7,13-diacetylbaccatin 111, from the assay. The organic fractions were then combined and
solvent was removed in vacuo. Any remaining residue was subsequently resuspended in
100 pL of acetonitrile, and the resultant assay extract was subjected to further analysis.
Assays were separated by reverse phase chromatography and analyzed using ESI-
MS. A 10 ,uL aliquot of the resuspended assay residue was injected onto a C18 column
(Betasil C18, 5 ,uM, 150 x 2.1 mm, Thermo Fisher Scientific Inc, Waltham, MA) eluted
with a gradient mobile phase composed of the solvent A/solvent B ratios [solvent A: 99.5
% H20 with 0.5 % formic acid (v/v); solvent B: 99.5 % CH3CN with 0.5 % formic acid
(v/v); gradient: 0 - 7 min at 70 — 0.0 % (A), 7 - 9 min at 0.0 % (A), 9 — 10.1 min at 0 — 70
% (A), 10.1 - 11 min at 70 % (A)] at a flow rate of 0.3 mL/min on a capillary HPLC
system (CapLC capillary HPLC, Waters, Milford, MA). The effluent was monitored via
LC-ESI-MS analysis (Q-ToF Ultima Global, Waters, Milford, MA), and the resultant
Chromatogram was used to identify de novo peaks when compared to assay controls.
Coupled activity assays for mTBT and BadA. The coupled activity of BadA

and mTBT was tested in the following manner. A reaction mixture containing 500 ,uM

41

7,13-diacetyl-2-debenzoylbaccatin III, 100 pg/mL total protein(containing mTBT), 100
,ug/mL BadA, 500 ,uM BzOH, 500 ,uM CoASH, 2.5 mM MgC12, and 1.0 mM ATP,

diluted to final concentration with mTBT Assay Buffer, was incubated for 15 h at 31 °C.
Assays were analyzed using the method previously described for analyzing mTBT assays.

Coupled activity assays for mTBT and BadA with non-natural carboxylic
acid substrates. Two of the four non-natural substrates proven to be active in the BadA
reaction were incubated in the coupled assays system to determine its possible use in
synthesizing 0-substituted-7,13-diacetyl-2-debenzoylbaccatin III analogues. The assay
conditions used were identical to those listed in the procedure testing the coupled activity
of mTBT and BadA. However, in place of the 500 ,uM BzOH, the two substrates 4-
methylbenzoic acid and thiophene-2-carboxylic acid were incubated at 500 yM. Assay

analysis was again identical to the procedure used for assessing mTBTactivity alone.

2.3: Results and Discussion
2.3.1: NDTBT Results and Discussion

NDTBT activity assays. Analysis of the activity assays discussed in the
experimental section yielded chromatograms similar to the following displayed in Figure

9 on the next page.

42

 

14.0 min

600 a 18.8 min
’5
3. 400 1
'8
S,
Q
g 16.7 mirfi
200 -

 

 

 

 

 

0 n———Jk————:———-/‘-—"A"’”J

I I I I

 

0 5 10 15 20 25 30

Retention Time (min)

Figure 9. Reverse phase separation followed by UV detection at 228 nm of an
activity assay for the enzyme NDTBT. Peak identities: 14.0 min, baccatin III; 16.7
min, N-debenzoyl-Z'-deoxypaclitaxel; 18.8 min, 2'-deoxypaclitaxel. The baccatin III
was included as a standard, as will be discussed in later experiments.

The sample eluting off the HPLC column at 18.8 min was collected and identified as 2'-
deoxypaclitaxel by comparison to authentic standard (Figure 9). The peak area was
converted to concentration of product using a standard curve constructed using laboratory
stocks of 2'—deoxypaclitaxel.

Kinetic Evaluation of NDTBT with N-Debenzoyl-(3'R)-2'-deoxypaclitaxel

(NDB2DT). The proposed coupled assay system requires incubation of NDTBT with the

43

carboxylic acid and CoASH substrates, along with cofactors, necessary for BadA activity.
The affect of these substrates on NDTBT activity have not been previously studied;
however, an understanding of these effects could potentially lead to optimization of the

coupled assay system. Initial attempts at evaluating these effects were made via

comparison of the kinetic parameters km and KM in NDTBT assays incubated with and
without one of the newly introduced substrates or cofactors (CoASH, BzOH, MgClz, and
ATP). However, many challenges were faced as the kcat and KM values for NDTBT

alone did not match those previously reported.30’42 The differences in kinetic evaluation

were later rectified on the basis of several observations. First, enzyme used in the

original kinetic study of NDTBT was functionally expressed and purified as an N-

terminal His6-tag epitope, whereas the enzyme used in this thesis research was

functionally expressed as a C-terminal His6-tag epitope. Second, the previous studies

utilized soluble protein solutions with an estimated purity of roughly 70 % based on SDS-
PAGE analysis; the additional purification steps included in the Experimental section
above increased protein purity to roughly 95 % based on similar evaluation
methodologies. Finally, the BzCoA used in the original kinetic studies was not as pure as
originally calculated and contained some residual CoASH salts; as discussed in the
previous chapter, actual concentrations of BzCoA were found to be anywhere from 10 -—
15 % of the reported concentrations. All of these reasons led to the necessity for the
reevaluation of NDTBT kinetics with the purer soluble protein solution and with

commercially available BzCoA.

44

As kcat was reached within the first 5 min of incubation when a protein

concentration of 500 ,ug/mL (used previously)30 was used, range-finding experiments

with varying amounts of NDTBT found that steady state kinetics were observed at an
enzyme concentration of 50 ,ug/mL and a 10 min incubation period. Following
incubation of the enzyme with varying benzoyl CoA concentrations (0 — 500 ,uM) and a
constant N-debenzoyl-2'-deoxypaclitaxel concentration (500 pM, saturating), the

following Hanes-Woolf plot was obtained.

 

2500

2000 -

 

l

1500

 

l

1000

[BzCoA]/vo(;1M min nmol-1)

500 -

 

 

 

 

O I I I I I
0 100 200 300 400 500 600

[BzCoA] (I’M)

Figure 10. Hanes-Woolf analysis of the data obtained from incubating NDTBT with
500 ”M NDB2DT and varying concentrations of BzCoA at a 10 min incubation
period. The plot shows the concentration of BzCoA divided by the experimentally
determined v0 (reported as nmols of 2'-deoxypaclitaxel divided by 10 min) vs.
BzCoA concentration. Error bars are reported as the standard deviation of three
separate measurements.

45

From this data, it was found that KM = 116 i 29 ,uM benzoyl CoA and Vmax = 0.33 :t 0.06
nmol 2'-deoxypaclitaxel per minute; taking into consideration the enzyme concentration,

it was found that kcat = 1.7 d: 0.3 min']. When compared to the previously reported
values”42 evaluated for the same enzyme (KM = 410 pM, kcat = 1.5 :t 0.3 $4), it can be
seen that the original reported KM was roughly 3.5 times greater than the newly defined
KM, and that the turnover number kcat was roughly 55 times greater than the newly

defined kw. These significant variations are most likely due to the several reasons listed

above.

Following the determination of the newly defined kinetic parameters describing
the reaction catalyzed via NDTBT activity, experiments to determine the effect of
CoASH on enzyme turnover were completed. Figure 11 displays the results of these
experiments, where the same assay conditions were used to determine original kinetic

parameters, but CoASH was added at a constant concentration within a given set of v0

measurements.

46

 

 

 

 

 

 

 

 

 

2500
A 2000 -
'1' *-

‘5

E

.s 1500 1 3

E

a . i ~-

E’ 1000 —

'3'

Q o [BzCoA]/vo, [CoASH] = 0 ,uM

59- 500 - o [BzCoA]/vo, [CoASH] = 50 ,uM

v [BzCoA]/v0, [CoASH] = 500 ;1M
0 I I I I I
0 100 200 300 400 500 600

[BzCoA] (WI)

Figure 11. Hanes-Woolf analysis of the data obtained from incubating NDTBT with
500 ,uM NDB2DT and varying concentrations of BzCoA at a 10 min incubation
period upon the addition of CoASH as indicated in the legend. The plot shows the

concentration of BzCoA divided by the experimentally determined v0 (reported as
nmol of 2'-deoxypaclitaxel divided by 10 min) vs. BzCoA concentration. Error bars
are reported as the standard deviation of three separate measurements.
The results from these experiments show that coenzyme A has no observable effect on
NDTBT activity with the given conditions. This conclusion can be drawn from the
inhibition data that coincides with the original measurements of NDTBT kinetic
parameters, described earlier in this document. Greater concentrations of coenzyme A

were not included in this study because coenzyme A concentrations never exceeded 500

pM in the coupled assay system.

47

Optimizing coupled activity assays for NDTBT and BadA. A set of assays
were completed to verify that the coupled enzyme assay system was optimized (Table
IV). The evidence provided by these assays is also important in proving several of the
other hypotheses made concerning the activity of the coupled assay system. First, by

comparing Assays A and B (Figure 12), it can be seen that the 2'-deoxypaclitaxel (2DOT)

yield when NDTBT is incubated with Mng and ATP is within experimental error of the

2DOT yield when MgC12 and ATP are not present. This result confirms the results of

several single point inhibition studies completed at various time points with the same
enzyme and substrate concentrations, where overall 2DOT production was not affected.
Comparing Assays A and C also show that the addition of BadA does not hinder NDTBT
turnover. Second, when comparing Assays A and D, product yields are again within
experimental error of one another. As Assay A was simply testing NDTBT production,
while Assay D was testing production levels in the coupled assay system (with equimolar
amounts of BzCoA in Assay A to BzOH and CoASH amounts in Assay D), this result
proves that the coupled assay system is just as effective at producing 2DOT as is the
simple NDTBT assay system. A comparison of Assays D and E, where the CoASH
concentration was halved, again shows product yield within experimental error. This
shows that, as hypothesized, CoASH concentrations do not need to be equimolar to the
BzOH donor in order to achieve maximal turnover. However, as product levels never
exceeded the concentration of CoASH, this series of assays alone does not prove that
CoASH can be effectively recycled in the coupled assay system. Assay G, which

contained both BzOH and BzCoA at 200 ,uM, was also designed to test the hypothesis

48

that CoASH is recycled in the coupled assay system. However, because product levels
did not exceed 200 pM, it is not clear whether or not CoASH was recycled.

Interestingly, Assay F shows the lowest production levels in the optimized assay
system. One possible explanation for the observed product amounts may be that with

CoASH concentrations at 50 ,uM, the level of BzCoA at any time during the reaction

cannot exceed 50 ,uM. The KM for NDTBT with respect to BzCoA was previously

shown to be 116 i 29 ,uM, which implies that at 50 pM BzCoA, the kcat for NDTBT

would not be reached. Hence, with BzCoA levels below 116 ,uM, NDTBT was not
operating at optimal velocity, meaning that a two hour time point may not have been long

enough to observe the largest product levels for the described assay composition.

49

Table IV. Assay compositions used in the optimized NDTBT and BadA coupled
assay system. Y indicates inclusion in the assay composition.

 

Final

A N b
Component ssay um er

 

Concentration
A B E F G
(,uM) C D

 

NDB2DT
(500) Y Y Y Y Y Y Y

 

BzOH (400) Y Y Y

 

BzOH (200) Y

 

BzCoA (400) Y Y Y

 

BzCoA (200)

..<

 

..<

CoASH (400)

 

CoASH (200)

<

 

CoASH (50)

 

MgC12 (2500) Y

 

ATP (1000) Y

 

BadA
(50 #g/mL)

 

 

NDTBT
(200 gngL)

~< ~<~<~<
~< ~<~<~<
~< ~<~<~<
-< ~<~<~<~<
~< ~<~<~<

 

 

 

 

 

 

 

 

 

g 250
8A
£3200“
£51504
=33
5;:1001
5:8
g8 50.4
<1)

C90

24 o.

 

 

   

ABCDEFG

Assay

Figure 12. The above chart displays 2'-deoxypaclitaxel concentrations obtained
when the assay compositions listed in Table IV were incubated for the prescribed 2
h time point. Error bars are reported as the standard deviation of three
independent measurements.

50

 

These assay conditions were considered optimal when taking into account several
previously completed studies where enzyme and substrate concentrations, as well as time
of incubation, were all varied independently of one another. Two hours was chosen as
the appropriate stop time because similar assays at 20 min showed less production, while
assays at time points up to 48 hours did not show a statistically significant increase in
production levels. It should be noted that while similar assays were used to optimize
conditions, it was exceedingly difficult to keep specific variables consistent. For example

(cf. Chapter 1 Results and Discussion), BzCoA laboratory stocks prepared synthetically
via previously described methods30 were found to be 10 — 15 % pure based on

comparison with the BzCoA standard curve constructed with commercially available
BzCoA. Many of the preliminary coupled assays were completed using such stocks, and
thus prompted all subsequent experiments to be completed with a commercially available
stock of BzCoA. Assays conducted with NDTBT alone, with the unknowingly impure
stock of BzCoA and the appropriate taxane co-substrate, had production levels that were
50 % less than similar assays in which the BzCoA was derived from BadA activity in a
coupled enzyme system. Due to the prohibitively high cost of the acyl-CoA substrate, as
discussed in the Introduction, the economics of re-running initial assays were considered
too great to warrant further investigation.

While the assays presented above do not prove that CoASH can be effectively
recycled in the coupled assay system, previous preliminary assays did prove this
hypothesis. The composition of those assays differed from the optimal set of conditions
because these experiments were designed before the optimal conditions were established.

In a series of assays, N—debenzoyl-2'-deoxypaclitaxel was maintained at 500 ,uM when

51

BzOH and CoASH were equimolar at 50 ,uM. 2'-Deoxypaclitaxel was produced at 85 %
yield (7.1 ,ug, 42.5 ,uM) in 16 h. However, when BzOH concentration was doubled to
100 ,uM while CoASH concentration remained constant at 50 ,uM, 2'-deoxypaclitaxel was
produced at 96 % yield (16.5 pg, 98 ,uM) in 16 h. This shows that CoASH was
effectively recycled in the coupled assay system because product yields exceeded
CoASH concentration.

Coupled NDTBT and BadA activity with non-natural carboxylic acid and N-
debenzoylpaclitaxel substrates. The following Table V summarizes the retention times
of de novo peaks eluting from the column after separation and UV analysis as discussed

in the experimental section.

Table V. Retention times (min) for the N-substituted-N-debenzoyl-Z'-
deoxypaclitaxel products of the NDTBT and BadA coupled reaction when incubated
with the reported acyl-carboxylic acid donor.

  
   

 

 

AcO O OH
R‘lng O
O
O .'-.
W ”0 532 0Ac
R: o O O O
o . S .3
9’: 9’5 \ g1 \ / ‘
9‘1 F 0
/
4 M h 1b . Thiophene-Z-
. . ' et y 6112019 2-Fluorobenzoic - - Carboxylic
Benzmc Ac1d Acid Acid 3-Furorc ACld Acid
18.7 min 19.4 min 18.4 min 17.6 min 19.0 min

 

 

 

 

 

 

 

52

The identity of each peak was characterized via ESI-MS analysis, as discussed in
the experimental section. The following set of Figures (13 - 17) display the mass spectral

data obtained as a result of the above peak collection and subsequent analysis. Common
fragment ions of the parent ions [M+H]+, as previously resolved in literature,30 include

loss of the C-13 side chain to yield the positively charged taxane core (m/z = 569), the
positively charged C-l3 side chain (m/z dependent on substrate), and subsequent loss of

an acetyl group from the taxane core (m/z = 509). Another commonly observed ion was
that of [M+formic acid+acetonitrile+H]+, which yields an [was]+ ion. This adduct has
been reported in previous literature;43 its presence is not surprising as the mobile phase

used to separate the sample before mass spectral analysis contained both acetonitrile and

formic acid.

53

 

 

 

 

 

 

 

 

925.6
100 ‘ [M+HCOOH+CH3CN+H]+
838.5
A 80 " +
25. [M+H]
0)
U
C
8
g 60 -
D 552.6
<
0)
.2
g 40 - 569.4
'03
(I
20 , 270.2 494.5
742.8
0 _ . LU--LlLdL. 1.111 In. L' -
200 400 600 800 1000 1200
m/z

Exact Mass = 837.3
[C-l3 Side Chain + 111* m/z = 270.2

 

Figure 13. Mass spectrum of the collected de novo peak (at a retention time of 18.6
min) from the separation of an NDTBT and BadA coupled reaction incubated with

the substrates NDBZDT and BzOH. Presence of the pseudomolecular ion [M+H]+ at
838.5 m/z as well as the regular fragmentation pattern of paclitaxel analogues
confirms the identity of the product peak as 2'-deoxypaclitaxel.

54

 

 

 

 

 

 

100 . 852.5
iM+H1“
,0. 80 .
o\
‘6 939.7
U
5 [M+HCOOH+CH3CN+H]+
g 60 -
3
.D
<
0)
>
'g 40 -
E
a:
20 _ 569.4
284.2
509.3
0 -LLL n Lnanh- n-.L n TJLALJEUAAFI ll. 1' A I I
200 400 600 800 1000 1200 1400
m/z

Exact Mass = 851.4
[C-13 Side Chain + 111* m/z = 284.2

 

Figure 14. Mass spectrum of the collected de novo peak (at a retention time of 19.4
min) from the separation of an NDTBT and BadA coupled reaction incubated with
the substrates NDB2DT and 4-methylbenzoic acid. Presence of the
pseudomolecular ion [M+H]+ at 852.5 m/z as well as the regular fragmentation
pattern of paclitaxel analogues confirms the identity of the product peak as N-
debenzoyl—N—(4-methylbenzoyl)-2'-deoxypaclitaxel.

55

 

 

 

 

 

 

 

 

100 _ 844.6
[M +Hl+ 931.7
[M+HCOOH+CH3CN+H]+

g 80 ‘
8
C
45’
g 50 ' 569.4
.0
<
0)
.2
E 40 d
31:: 509.5

20 , 276.2

0 ' r ‘ JFL L I L I I
200 400 600 800 1000 1200 1400
m/z
S /
Exact Mass = 843.3
0 NH

 

[C-13 Side Chain + 111* m/z = 276.2

Figure 15. Mass spectrum of the collected de novo peak (at a retention time of 18.4
min) from the separation of an NDTBT and BadA coupled reaction incubated with
the substrates NDB2DT and thiophene-Z-carboxylic acid. Presence of the
pseudomolecular ion [M+H]+ at 844.6 m/z as well as the regular fragmentation
pattern of paclitaxel analogues confirms the identity of the product peak as N-
debenzoyI-N-thiophene-2-carbonyl-2'—deoxypaclitaxel.

56

 

 

 

 

 

 

 

100 - 915.6
[M+HCOOH+CH3CN+H]+

4? 80 —
2\_’
OJ
U
C
8
c 60 -
3
.Q
<
3 8284
33 40 d . +
:2 569.3 [MW]

20 -

260.1
509.3
0 PAL .l.1'_nntixi..lnr - LlLllian 111' - n
200 400 600 800 1000 1200
m/z

Exact Mass = 827.3
[C-l3 Side Chain + 111* m/z = 260.1

 

Figure 16. Mass Spectrum of the collected de novo peak (at a retention time of 17.6
min) from the separation of an NDTBT and BadA coupled reaction incubated with
the substrates N DBZDT and 3-furoic acid. Presence of the pseudomolecular ion
[M+H]+ at 828.4 m/z as well as the regular fragmentation pattern of paclitaxel
analogues confirms the identity of the product peak as N-debenzoyl-N-(3—furanoyl)-
2'-deoxypaclitaxel.

57

 

 

 

 

 

 

 

 

100 . 943.6
8565+ [M+HCOOH+CH3CN+H]*
[M+Hl

? 80 A 569.3
5:.

Q)

U

C
8

c 60 -

3
.Q
<

0)
.2
E 40 r
:2 288.2

20 - 509.3
0" All I“ J‘TLl-LlTll l‘ l I A I
200 400 600 800 1000 1200
m/z
Exact Mass = 855.3
0 NH

[C-l3 Side Chain + in“ m/z = 288.2

 

Figure 17. Mass spectrum of the collected de novo peak (at a retention time of 19.0
min) from the separation of an NDTBT and BadA coupled reaction incubated with
the substrates NDBZDT and 2-fluorobenzoic acid. Presence of the pseudomolecular
ion |M+H]+ at 856.5 m/z as well as the regular fragmentation pattern of paclitaxel
analogues confirms the identity of the product peak as N-debenzoyl-N-(Z-
fluorobenzoyl)-2'-deoxypaclitaxel.

58

All of the displayed mass spectra indicate that the expected N-substituted-N-debenzoyl-
2'-deoxypaclitaxel product was biosynthesized as a direct result of the coupled assay
activity. These are important findings in that they conclusively prove the activity of
BadA with the acyl-carboxylic acid substrate to produce the expected acyl-CoA product.
This data also proves that the coupled enzyme assay system can be used to produce 2'-
deoxypaclitaxel analogues, thus decreasing the reliance on the necessary acyl-CoA
substrate in an uncoupled NDTBT assay.

When the coupled assay system was incubated with the taxane substrate N-

debenzoylpaclitaxel, as discussed in the Experimental section, ESI-MS/MS fi'agmentation
was used for analysis as in previously reported literature.3O Tandem mass spectrometry

was used for analysis because the product yields of the biocatalytic reactions were
previously reported as below detection limits using the described UV-HPLC system. The
following mass Spectrum (Figure 18) was obtained from fragmentation of the parent

paclitaxel ion ([M+H]+ = 854 m/z). The fragmentation pattern is similar to that of the 2'-

deoxypaclitaxel ions discussed above; however, because fragmentation was induced via
collision induced dissociation (CID), the abundances of the fragment ions are much

greater than the parent ion peak.

59

 

 

 

 

 

 

 

 

286.1

100 -

80 -
d)
U
C
to
'O
S 60 -
.D
<
0
.2
('0
7, 40 -
a: 509.2

2

0 i 569.2 854.3
[M+Hl+
0 _[ 7 1 L 'l a d m 1. l i ‘I L h l u
200 400 600 800 1 000
m/z
O Exact Mass = 853.3

[C-13 Side Chain + H]+ m/z = 286.1

 

Figure 18. Mass spectrum of the fragmented parent ion 854 m/z from the analysis
of an NDTBT and BadA coupled reaction incubated with the substrates N-
debenzoylpaclitaxel and BzOH. Presence of the pseudomolecular ion [M+H]+ at
854.3 m/z as well as the regular fragmentation pattern of paclitaxel analogues
confirms the identity of the product peak as paclitaxel. The base peak 286.1 m/z
represents the protonated N-benzoylated-phenylisoserinoyl sidechain.

60

2.3.1: mTBT Results and Discussion

Coupled and un-coupled activity assays for mTBT and BadA. Activity assays
completed for mTBT produced results similar to those reported in previous literature.2
The following Figure 19 displays the mass spectrum obtained by integrating the mass

spectra collected during elution of the peak corresponding to the product 7,13-

diacetylbaccatin III, as identified by an identical retention time to product standard.

 

 

 

 

 

[M+H]+
? 80 ‘ 693.4
a. Aco" HO , ; +
8 0Ac [M+Na]
C
m C)
1; 60 .
D
.D
<
0)
'5 40 ‘ Exact Mass - 670 3
‘7'; ' 652.4
I
20 _ 611.4
0 A.“ 1"; LLl‘nLlleJ‘LL 1111' n. “1.1-!

 

 

 

200 400 600 800 1000

m/z

Figure 19. Product mass spectrum for 7,13-diacetylbaccatin [11 product as a result

of an mTBT activity assay. The base peak at m/z = 671.4 corresponds to the [M+H]+
pseudomolecular ion.

In Figure 19, the base peak has an m/z corresponding to the [M+H]+ pseudomolecular

ion. The next most abundant peak at m/z = 693.4 corresponds to the Na+ adduct of the

61

molecular ion. Loss of an acetyl group from both the [M+H]+ ion and the [M+Na]+ ion

yield the peaks at m/z = 611.4 and 652.4, respectively.

Following confirmation of mTBT activity, the coupled assays (as described in the
experimental section) were used to determine whether the activity of BadA could be
coupled in vitra to the activity of mTBT. The resultant mass spectrum matched that of
the expected product, as displayed above in Figure 19. This important result shows that
the activity of BadA can be coupled to at least two of the T axus acyltransferases involved
in the biosynthetic pathway for the production of paclitaxel.

Results of the mTBT and BadA coupled assays incubated with the non-natural
acyl-carboxylic acid substrates 4-methylbenzoic acid and thiophene-Z-carboxylic acid are

displayed on subsequent pages in Figures 20 — 21.

62

 

 

 

 

 

 

 

100 ~ 707.4
[M+Na]+
.... 80 -
3\:
a)
8 o
m
‘g 60 -
3
.D
<
d)
1% 40 -
+3 Exact Mass = 684.3 685.4
c“ +
[M+H]
20 -
0 ALJLAL- [‘41 l n- 1 LA. I
200 400 600 800 1000

m/z

Figure 20. Product mass spectrum for 7,13-diacetyl-2-debenzoyl-2-(4-
methylbenzoyl)-baccatin III product as a result of an mTBT and BadA coupled

assay. The base peak at m/z = 707.4 corresponds to the [M+Na]+ pseudomolecular
ion.

63

 

 

 

 

 

 

 

 

 

100 i 677.4+
[M+H]
A 80 -
3g
8
g 60 a 699.4
C d +
g / S [Md-N3]
a, /
>
'43 40 5
Fr, Exact Mass = 576.2 658.4
20 4
599.3
0 41.1. I L m nthLl'LLl Ainn ill. L L411!
200 400 600 800 1000
m/z

Figure 21. Product mass spectrum for 7,13-diacetyl-2-debenzoyl-2—(thiophene-2-
carbonyl)-baccatin III product as a result of an mTBT and BadA coupled assay.

The base peak at m/z = 677.4 corresponds to the [M+H]+ pseudomolecular ion.
These results indicate that the mTBT and BadA coupled activity assays are feasible in the

production of 2-0—substituted 7,13-diacetyl-Z-debenzoylbaccatin III analogues.

2.4: Conclusions

The results of the coupled assay system, where BadA activity is coupled with
NDTBT as well as mTBT, are promising in many respects. First, when NDTBT is
incubated with the natural substrates BzCoA and N-debenzoyl-(3'R)-2'-deoxypaclitaxel,
the amount of observed product 2'-deoxypaclitaxel is statistically identical to that

observed when NDTBT is incubated with BadA. These results are important because

64

they indicate that the activity of NDTBT is unaffected in the coupled assay system.
Furthermore, these results prove that the production of 2'-deoxypaclitaxel via the
biocatalyst NDTBT can be achieved without the need for synthesizing or purchasing the
expensive BzCoA substrate necessary for NDTBT activity. Additionally, the coupled
activity with non-natural carboxylic acid substrates further reduces the necessity for
synthesized or purchased acyl-CoA substrates. Finally, the results Show that a reduced
amount of CoASH can be used in the coupled assay system to yield identical product
amounts to uncoupled NDTBT assays. This result is especially important because it
reduces the need for CoASH, which is an expensive co-substrate (at $ 1,590.00 per g
(Sigma)). While CoASH may not be as expensive as the respective acyl-CoA substrates,
reducing the necessary amounts within the coupled assay system offers further
advantages over the NDTBT assay alone. The studies completed with coupled mTBT
and BadA activity indicate similar benefits for work with other T axus acyltransferases.
While previous work has shown both acyltransferases considered in this research to have
broad acyl-CoA substrate specificity, all of the necessary acyl-CoA substrates for such
studies had to be synthesized. The use of the coupled assay system circumvents this
necessity, allowing for a possibly more efficient method of producing identical

substituted paclitaxel analogues without the same high labor and monetary costs.

65

CHAPTER 3

Future Aims
While the results presented within this thesis research Show some promising
capabilities of the coupled enzyme assay system discussed, still more research is
necessary to further elucidate the complicated nature of such a proposed system. The

following brief list outlines some areas of future interest.

1. Optimize the BadA, mTBT coupled in ‘vitra assay.
2. Compare coupled assay turnover of non-natural carboxylic acid substrates to that
of acyltransferase activity with the respective acyl-CoA substrate.

3. Apply coupled assay system with in viva experiments.

The first future aim requires several independent experiments. As with the
coupled BadA, NDTBT assay, the first steps would involve elucidating the effects of
BadA substrates and cofactors on the reaction catalyzed by mTBT. While this may at
first consideration seem mundane, it would require the reevaluation of mTBT kinetic
parameters, as was completed with NDTBT. This would be necessary due to the
previous discoveries made concerning the purity of the BzCoA substrates utilized in
determining both NDTBT and mTBT kinetic parametersz’30 With those results in hand,
an optimized coupled assay system with BadA and mTBT could be devised, leading to
the maximal production of baccatin III analogues possible.

As one of the benefits of the coupled assay system discussed throughout this

thesis has been the possible production of paclitaxel and baccatin III analogues due to

66

respective broad substrate specificities, the second future aim is an important one for the
viability of such a coupled system. These necessary comparisons would truly
demonstrate the possible advantages of the proposed system. However, there are many
challenges in completing such research. As an example, recall the drastic changes
observed when the NDTBT activity assay, incubated with synthesized BzCoA, Showed
much less turnover than the analogous coupled assay. These observations were later
attributed to the fact that the amount of benzoate donor in the NDTBT activity assay was
not equimolar to that of the coupled assay due to impurities. From the discovery of
impurities in the synthesized BzCoA, it can be inferred that similar, previously

undetected impurities are present in the other acyl-CoA substrates synthesized via the
methods presented in previous work.2 Obviously, if such substrates were used as the acyl

donors for NDTBT activity assays, and such assays were compared to coupled assay
activity, true comparisons would not be achieved. Thus, the research of aim two is
hindered by substrate availability and purity, and would require either newly synthesized
or purchased acyl-CoA substrates. This would be an expensive and time consuming
endeavor.

While more research is necessary within the in vitra coupled assay system, the
results of this thesis research have proven that BadA can produce acyl-CoA substrates at
a high enough level to be used by two T axus acyltransferases. The third future aim, while
perhaps still far off, would be an important one for presenting a complete consideration
of the coupled assay system. As one of the arguments in support of this research lies in
the possibility of producing non-natural acyl-CoA substrates in a transgenic host cell for

use as substrates with T axus acyltransferases, the co-expression of BadA and either

67

NDTBT or mTBT in viva would allow for a complete consideration of such a coupled
enzyme system. Incubation of such a transgenic host with the appropriate N-debenzoyl-
2'-deoxypaclitaxel or 7,13-diacetyl-2-debenzoylbaccatin III substrate, followed by
analysis regarding the possible products, would discern the capabilities of such a
transgenic host. One main difficulty of such experiments lies in the ability of E. coli (or
other host cells) to allow the appropriate taxane substrate to permeate the cell membrane.
The complex nature of any transgenic host could also prove difficult in determining
experimental results. Nevertheless, in viva considerations of the coupled assay system
would allow for further implications in the use of transgenic host cells to produce

possible second generation paclitaxel analogues.

68

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

REFERENCES

Geissler, J. F .; Harwood, C. 8.; Gibson, J. Purification and Properties of
Benzoate-Coenzyme a Ligase, a Rhaa'opseudomonasPalustris Enzyme Involved
in the Anaerobic Degradation of Benzoate. Journal of Bacteriology 1988, 1 70,
1709-1714. ‘

Nawarathne, I. N.; Walker, K. D. Point Mutations (Q19P and N23K) Increase the
Operational Solubility of a 2 alpha-O-Benzoyltransferase that Conveys Various

Acyl Groups from CoA to a Taxane Acceptor. Journal of Natural Products 2010,
73, 151-159.

Cutler, S. J.; Cutler, H. G. Biologically Active Natural Products: Pharmaceuticals;
CRC Press LLC: Boca Raton, FL, 2000.

Suffness, M. Taxol®: Science and Applications. In Pharmacology and
Toxicology: Basic and Clinical Aspects; Hollinger, M. A. Ed.; CRC Press, Inc.:
Boca Raton, 1995.

Schiff, P. B.; F ant, J.; Horwitz, S. B. Promotion of Microtubule Assembly Invitro
by Taxol. Nature 1979, 2 77, 665-667.

Manfredi, J. J.; Horwitz, S. B. Taxol - an Antimitotic Agent with a New
Mechanism of Action. Pharmacology & Therapeutics 1984, 25, 83-125.

Horwitz, S. B. Mechanism of Action of Taxol. Trends in Pharmacological
Sciences 1992, 13, 134-136.

Eisenhauer, E. A.; Vermorken, J. B. The taxoids - Comparative clinical
pharmacology and therapeutic potential. Drugs 1998, 55, 5-30.

Liistro, F.; Stankovic, G.; Di Mario, C.; Takagi, T.; Chieffo, A. et al. First clinical
experience with a paclitaxel derivate-eluting polymer stent system implantation

for in-stent restenosis - Immediate and long-term clinical and angiographic
outcome. Circulation 2002, 105, 1883-1886.

Weissmann, C.; Reyher, H. J.; Gauthier, A.; Steinhoff, H. J.; Junge, W. et al.
Microtubule Binding and Trapping at the Tip of Neurites Regulate Tau Motion in
Living Neurons. T raffic 2009, 10, 1655-1668.

Hoke, S. H.; Wood, J. M.; Cooks, R. G.; Li, X. H.; Chang, C. J. Rapid Screening

for Taxanes by Tandem Mass-Spectrometry. Analytical Chemistry 1992, 64,
2313-2315.

69

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

Ojima, 1.; Habus, 1.; Zhao, M.; Zucco, M.; Park, Y. H. et al. New and Efficient
Approaches to the Semisynthesis of Taxol and Its C-13 Side-Chain Analogs by
Means of Beta-Lactam Synthon Method. Tetrahedron 1992, 48, 6985-7012.

Kingston, D. G. I. Taxol, a molecule for all seasons. Chemical Communications
2001, 867-880.

Seki, M.; thora, C.; Takeda, M.; Furusaki, S. Taxol (Paclitaxel) production
using free and immobilized cells of Taxus cuspidata. Biotechnology and
Bioengineering 1997, 53, 214-219.

Holton, R. A.; Somoza, C.; Kim, H. B.; Liang, F .; Biediger, R. J. et al. First total

synthesis of taxol. 1. Functionalization of the B ring. Journal of the American
Chemical Society 1994, I 16, 1597-1598.

Holton, R. A.; Kim, H. B.; Somoza, C.; Liang, F .; Biediger, R. J. et al. First Total
Synthesis of Taxol .2. Completion of the C-Ring and D-Ring. Journal of the
American Chemical Society 1994, 116, 1599-1600.

Nicolaou, K. C.; Yang, 2.; Liu, J. J .; Ueno, H.; Nanterrnet, P. G. et al. Total
Synthesis of Taxol. Nature 1994, 367, 630-634.

Danishefsky, S. J .; Masters, J. J .; Young, W. 8.; Link, J. T.; Snyder, L. B. et al.
Total synthesis of baccatin III and taxol. Journal of the American Chemical
Society 1996, 118, 2843-2859.

Wender, P. A.; Badham, N. F .; Conway, S. P.; Floreancig, P. E.; Glass, T. E. et al.
The Pinene Path to Taxanes. 5. Stereocontrolled Synthesis of a Versatile Taxane
Precursor. Journal of the American Chemical Society 1997, 119, 2755-2756.

Wender, P. A.; Badham, N. F.; Conway, S. P.; F loreancig, P. E.; Glass, T. E. et al.
The Pinene Path to Taxanes. 6. A Concise Stereocontrolled Synthesis of Taxol.
Journal of the American Chemical Society 1997, 119, 2757-2758.

Morihira, K.; Hara, R.; Kawahara, S.; Nishimori, T.; Nakamura, N. et al.
Enantioselective total synthesis of taxol. Journal of the American Chemical
Society 1998, 120, 12980-12981.

Mukaiyama, T.; Shiina, 1.; Iwadare, H.; Saitoh, M.; Nishimura, T. et al.
Asymmetric total synthesis of Taxol®. Chemistry-a European Journal 1999, 5,
121-161.

Kusama, H.; Hara, R.; Kawahara, S.; Nishimori, T.; Kashima, H. et al.
Enantioselective total synthesis of (-)-taxol. Journal of the American Chemical
Society 2000, 122, 3811-3820.

70

(24)

(25)

(26)

(27)

(28)

(29)

(30)

(31)

(32)

(33)

(34)

Wender, P. A.; Miller, B. L. Synthesis at the molecular frontier. Nature 2009,
460, 197-201.

Denis, J. N.; Greene, A. E.; Guenard, D.; Guerittevoegelein, F .; Mangatal, L. et al.
A Highly Efficient, Practical Approach to Natural Taxol. Journal of the American
Chemical Society 1988, 110, 5917-5919.

Walji, A. M.; MacMillan, D. W. C. Strategies to bypass the taxol problem.
Enantioselective cascade catalysis, a new approach for the efficient construction
of molecular complexity. Synlett 2007, 1477-1489.

Fang, W. S.; Liang, X. T. Recent progress in structure activity relationship and
mechanistic studies of taxol analogues. Mini-Reviews in Medicinal Chemistry
2005, 5, 1-12.

Fang, W. S.; Wang, S. R. Structural studies of taxol analogues for drug discovery.
Expert Opinion on Drug Discovery 2008, 3, 1109-1122.

Engels, B.; Dahm, P.; Jennewein, S. Metabolic engineering of taxadiene
biosynthesis in yeast as a first step towards Taxol (Paclitaxel) production.
Metabolic Engineering 2008, 10, 201 -206.

Nevarez, D. M.; Mengistu, Y. A.; Nawarathne, I. N.; Walker, K. D. An N-
Aroyltransferase of the BAHD Superfamily Has Broad Aroyl CoA Specificity in

Vitro with Analogues of N-Dearoylpaclitaxel. Journal of the American Chemical
Society 2009, 131, 5994-6002.

Ouyang, T.; Walt, D. R. A New Chemical Method for Synthesizing and
Recycling Acyl Coenzyme a Thioesters. Journal of Organic Chemistry 1991, 56,
3752-3755.

Patel, S. S.; Conlon, H. D.; Walt, D. R. Enzyme-Catalyzed Synthesis of L-
Acetylcamitine and Citric-Acid Using Acetyl Coenzyme-a Recycling. Journal of
Organic Chemistry 1986, 51, 2842-2844.

Billhardt, U. M.; Stein, P.; Whitesides, G. M. Enzymatic Methods for the
Preparation of Acetyl-Coa and Analogs. Bioorganic Chemistry 1989, I 7, 1-12.

Jennewein, S.; Wildung, M. R.; Chau, M.; Walker, K.; Croteau, R. Random
sequencing of an induced T axus cell cDNA library for identification of clones

involved in Taxol biosynthesis. Proceedings of the National Academy of Sciences
of the United States of America 2004, 101, 9149-9154.

71

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

(43)

Babbitt, P. C.; Kenyon, G. L.; Martin, B. M.; Charest, H.; Slyvestre, M. et al.
Ancestry of the 4-Chlorobenzoate Dehalogenase - Analysis of Amino-Acid-
Sequence Identities among Families of Acyl-Adenyl Ligases, Enoyl-Coa

Hydratases Isomerases, and Acyl-Coa Thioesterases. Biochemistry 1992, 31,
5594-5604.

Egland, P. G.; Gibson, J.; Harwood, C. S. Benzoate-coenzyme A ligase, encoded
by badA, is one of three ligases able to catalyze benzoyl-coenzyme A formation
during anaerobic growth of Rhodopseudomonas palustris on benzoate. Journal of
Bacteriology 1995, 1 77, 6545-6551.

Sambrook, J .; Russel, D. W. Molecular Cloning: A Laboratory Manual; 3 ed.;
Cold Spring Harbor Laboratory Press: Cold Spring Harbor, New York, 2001.

Sharma, P. K.; Sangwan, N. S.; Sangwan, R. S. Trichloroacetic acid-facilitated
coenzyme A-SH-stabilization-based end-point spectrophotometric 5,5'-dithiobis-
2-nitrobenzoic acid assay for plant terpene alcohol acetyltransferases. Analytical
Biochemistry 2005, 346, 176-178.

Loncaric, C.; Merriweather, E.; Walker, K. D. Profiling a taxol pathway 10 beta-
acetyltransferase: Assessment of the specificity and the production of baccatin III
by in vivo acetylation in E-coli. Chemistry & Biology 2006, 13, 309-317.

Annesley, T. M. Ion suppression in mass spectrometry. Clinical Chemistry 2003,
49,1041-1044.

Long, R. M.; Croteau, R. Preliminary assessment of the C13-side chain 2'-
hydroxylase involved in Taxol biosynthesis. Biochemical and Biophysical
Research Communications 2005, 338, 410-417.

Walker, K.; Long, R.; Croteau, R. The final acylation step in Taxol biosynthesis:
Cloning of the taxoid C13-side-chain N-benzoyltransferase from Taxus.
Proceedings of the National Academy of Sciences of the United States of America
2002, 99, 9166-9171.

Tong, H.; Bell, D.; Tabei, K.; Siegel, M. M. Automated data massaging,
interpretation, and E-mailing modules for high throughput open access mass
spectrometry. Journal of the American Society for Mass Spectrometry 1999, 10,
1174-1187.

'72