! " # 
  

#  
#

$ %
& 

 ' 

()(*(+),


&  

   -
# ' 
*.(/
 0


1

2


(
  
0
1

"
3
4   

*
  0
1

5
2 

3

1 0
**
.(
*.(6
   

 


   


 !  "!   



  

 



   

    
 



 



  

 



 


# $%  "  &
'            "



  
   
 

 




































!" 
#

##$



% " 

!
$  " 

 
   

































!" 
##$



& 
% ' "  


 
   




















!" 
###$


(       "!     

(

# )*  

*+ 
* 
%+ 

,-
  


.  +-+- /
!" 
#0$



"  " 
* 
  " 
)1  

%  *  



 " 



&.


& '"
1' 

2  3 - 
 
















































































!" 
0$



.   


 
  

 
*




"  "
!" 
#0$

4
4
 " 

 
!
$%  " 3
)1
.  +-+- / 


!" 
0#$



 
* 
%+ 

!
$.  +-+- /


!" 
0#$


) $

(
* +  


, - 



'"1*  

 

''
'' 
:&!=$
:-  
& "  != , $
&5
& +  
: 
: +   ,6 

 

 "
 

   

     

:.9
, * 78
*  " 

, "",

3  
2>      
&
&3   =&8
=+ & -'/+


, '
  
&5
& +  
=<=
=+ /?, , 
9&
 "
=.=
=+ /?, , 
6&
6& +  
=+
=+ 
6.:
6  " 
0'6

 



  
;8
5
8  
@ 
@ '-
5&
5 - & +  

 "
 
#9&
#9 "

    
<&
<"& +  
@ 7
@ '-

 "
 
.2
./ 

    
8==
 8'/+, +

,+ /, , 

 A  
  
%#=3)/ B

1'- 
'"1*  
)1
3  

%#  
  



C  * 

" 1' - 

   

&
%, " 

7"+ 

 
  


# 

-


 * 
D  
"
3 
# 
 
*+
  
* 
= "
-   
6 - " 

= * "
  *
 ' 
 

-
 

'"+-
9 " 

 
 
*+

 -

  )
9 " 
!*+"" $
&
 )"3 

 "  
  
7"+ 



 * 
D  
*
'   B

6 " 

*+ 
'  
*
 

 " '  
=' 

%'3
 

% ', " 
* 
E3  

! $  " 
 
!$
 "  
'  
F

 
% '  * 
 
  - 
  B


% " 

!
$  " 
  
' "
%
 

.F - "  

 
=   B
*7

: * 

 " '  
& '" B
)1
!
$
"
 - 3"

C 
  
' 


 )"3 

% " 
 
!
$  " 


 - *  
G  - 
  
 
% ' "   
   
' )
-
 
' 


%+ 

      
 
-

"  "
#
  
"  "
  
.  
  B
 -

)
 
 " '  
& '"
  
  
E3  
 
-


  



!
$  " 

 
   
' )B
' -  '
 

% " 
  
!
$
#
 
 
-  A 


% ' "  
  
*+
 
  - 
.

 " 

5  
"F 
* "1)
')-
E "*
3" 
* 
*
" B


'  
' 

! $  " 
 

6



7 , 
 )1
*  
 .B



')-
+

&.& '"

 )

###

=<=' 3  
*+
 

6  "
#
  

7 

%+ 

      
.  +-+- /
 
' )
 )
 
 " ' 
&.& '"
* 1-"
)) 
 .
' -  'B


'
 
 "


)

&.    
.  

! $  " 

 
  
      
! $ .  +-+- /
  
  
 .
' )
-


%+ 

" ,    
!
$.  +-+- /
-
  '
"  "





C  * 

" 1' - 

 
/!   0! 1! $$   $   &
'    $     

! "# !$B
2B
4H4
&2#I
J%44

. 
  KB
<+
%" KB
5)
D
=B
. 
F  B
G


7- B
E)
- 

&  
" 
' -  '

/, B
9  



!
$  " 
 
 
    
 )  
9F '    
  
*  

    

 
/! 0     $   &
'    $     

! "#
!4$B
3B
(
&2#I
J%44(4

. 
  KB
<+
%" KB
5)
D
=B
. 
F  B
G


7- B
E)
- 

&
 " 

!
$  " 

 
   


  
" 
' -  ' 
 ' 



 )'

5  " 

 -

" 
C  
' -  ' 

  
!  %  
 * B


G   

" 
C  

 -

,3)  *
*
'  
E   

 
% " 


% "  

*  "  
3   

 -

&3   
 - B

* 
 ' 
6  "


!
$  " 
 
F 
&
 "  
 
0 "1* 

:  

 

' )
  

  "        $        $   &
'    $     

$ % 
!$B
(B
44
&2#I
J) '4

<+
%" KB
. 
  KB
. 
F  B
G

7- B
E)

- 

#
  
" 


G  - 

 
% ' "  

!
$  " 
 
 
   
' -  ' 
.
)

.      


+

 

 
'   
# '", /
  
-  
3   
 )* B

)1

   

8'/+  ,,

% ' 

" 
C  
-
 
E   


E -  

"* ,,3)  *
-

= " 
*  

K
 
  - )



C  * 

" 1' - 

 .           !       0$   ! 

"& "&
!$B
B
44
&2#I
J-'-4

<+
%" B
8 
=  B D 
&-"  ))B
' 
:'B
 "D

D B
 
E *B
. 
F  B
G

7- 

# 
", / 


   
%  *  
  

7  
)1
= * "
' F 
&

  
" 
  
'  
 
 "

 
 
F - 

%+ 
 
B
! $.  +-+- /B

* 
%  *  
&
# )*  
   
 .  

)

' 


0   

&.& '"
  
*  
&
  
  
.  

 
"  "

 -

) 

-   
.  

8 % *
 )
 
"  "
  
' -  ' 

 . 4 $  ! $  -+-0 ! 0$!  $   ! 00 

!&  !$B
B
4
&2#I
JL' -  4

6'
% )) . '  B
8
0-" B
 
 B
 "
< B
  "

.B
' 
:'B
<+
%" B



7  B
 "D
D B
= 

7    B
. 
F  B
G

7- 

& 
9  
' -  '

'

<   - - B
% "    - 

')- 
+


###
 

=<=' 3  
*+
 

6  "
#
.
)

)  )  
% M *  
  
% M * 6 "C   3
  
& -
 

 )  
0 3 
)1

7"+

! $  " 
 


.F - " 

% M * 6 ", 


 )) 

*+ 

' 
' " 
" 3

 . "!  $ &
5(
' ( !! 6 !  !   

(&
!$B
&2#I
JL 

<+
%" B
8 
=  B
< "
9*B
G

7- 

&
%+ 

!
B
$.  +-+- /
) 

0   
 
 "

   * 
,3)  *
* 
 "
 
 .
 .
' -  '

  
 
- 
 "

 
-  A 
  


"  "

 


5** + 
' )
  
  
 
* 
%+ 

! B
$&   
  - 

4



*+-
%+ 

 


!  "!   

 
  


3  

7   
 
7  -

< ' 
3  

 
  

5  '  
)1
=  B
*
 
% " , 



 B


 B
 

- 
 )  B

 -
7"+ B


*+ B
F 


-
" 


: 
3 )

% "  3 '

" 
 

:  
B


-  
E"  

 - ) 
 )   
5A
  
) 


 


= * B
6 -  B
 - 

. * 4
%
   
 

(
N

 
" ) 
= * "

D 
4
) " 
!''
$

 7 40 $   -    8  #9) + 05   / :  4   4    :$$ ( / 0        $  4  ) "   0  *  .  $  4   4    :$$   ,

E
 

7 ,  
!''
$
- -
B
 

-
)

   
E")) B

* 
*
" 
&  
 


-
' 1 B

' 
 
 -  -
E"  
 )  
"F 

7 , 


 - )

! $. - 
&
! $ 


 '*
' 
.'


"
6B
3  

!
$ 
" 
E" 
*  
#
  
6
*  

 * 
'  
,  

9- 
" 
  
 "  B
L -

 
- 
' 
 
% '* 

6
7  
 ) "" 
  
 



8  %" 


 
D  
&
- -
5 -

E")) 
  
 

%- )

7     
 ' (
&

' )


  -
%-  - ) 1' " 
) 
 B
 


 
%-   
-  ' B
&
 
* 

: '  

6 '  
)1 
* 
 ) -" 

E" 

 -  - 
  
E3  

!
$ 
3- -
 
 - ) 
 ) B


! $ 

'+- 3  
E"  
7  

  
6

- B

- 

   
    

)1

  


. *

= *




*+-
%+ 

 

 #
:        

"F 
  
 )
- - B
   
 )
*+- 
E 
  
  
7 

- - 
   
" 

 
- 
. "+ B

9   B
#   B B

= ,   B

= ,   B
* 
*


- 
G ' 
*
- )) 

  
.F - " 

- - 
   


  "
8 

 

8'3  
 
 
- - 
5 - 
&
- -
   
) 
3 )
-
7   

7 *
 )
 ,  

& -"
E   
  
 
9 " 
)

 - 





*+-
%+ 

 

 #7 #7 !  ;   :       : :    < <        ' ' :! :! 5 5 4' =   6 4' =   6 &=>'&=>'5 5 /' /' ;   ; &; 3'   3' 3 !    3 !  &3:?' & 3: ?'       

 (
   

=+ /?, , !=<=$' 3 
*+ B


  ,,

 

=+ /? , , !=<=$' 3 
*+ B 

  ,,

 
 
 )
 
"* ,
)   
 ' 
  ' 
"F 
"F  

 

 
 
- - 
5 -
  

 
 

 
  "
 * 
!" -
! "   -
9- ,  B
9- ,  B
''
''
$
$  
 
 
"F 
"F  
  
 ,-  

  
!+ -
%+  B
''
7$
E3  
' 

" - 
9- , 
  -

!+ -
%+  B
''
7$ E3  
' 

" - 
9- , 
  -

/
 ' 


N
 - 
  
"B

' 

+ - 
%+ 

/
 ' 

 N
 - 
  
" B

' 

 +  - 
% +  


N
 ' 
F - 
  N
 ' 
 F - 

 (7 ; !0    <       ;0 &'    !  "!  &4'  $ 

7 
9 " 
) 
 
= = 77. -  4
& 
. -  
"

* 
' " 
  
  
!''
4$4B4
#

 
' "


  ,,

 
1' 
-  
#  
 )

8)"
=<=
1'  B
 - 


7 

=+ /?, , 
!=.=$
   
5 - * 


 
 
)   *
& 
 
   
*
7 

9 "


  
 B
 - 
 

"+- 
<+


8)"
=<=
'  



*+-
%+ 

 

O' 

7 

 &   


"+-
<+
)   *


/ 
 
 
 

=<=
  
&
'"

=

P   ))

 
)1 
* 
7 

-  
#   
&  
  
   
-

 ) 
 
=
* 

&
-  A 
9 "

 
E  "1
'  


' "


6  * 

  
 


 

=.=
#

*  
' "


"* ,
1' 

 
9  )

#   

 
" ,    

  

=<=
 
   
& 
8)"
=<=
 
3  

9 "

 "  "  B4B44
 )  

,
  
=+  

8)"
=<=
"

 
 
< 

#  
-

'"

=
-

& "  

9 + 
",  
  

 )7 =    )5)# 3        0  5      0     5  ;     

#
 -
9-  

9 "
'3 )



5 - - 
' 3  
E3  
' 

" - 
9- , 
)

= "' 
' * 
B

 
' 


(



*+-
%+ 

 

+ - 
%+ 
 
- 

6 4
 )  
  
-
'

-  
.F - "  

5 - - 
*
-  ' 
  
"

O' - 

 
  *
  B


5 - - 

9-  

= "' 
*
-  ' 
 ' 
 
L -
 )
-  
  B

-
  
O' - 
" 
% ' '  

*+

   
'

' 3 


-   
  
  
.F - "  I
E


&
  B
"

5 - - 
-

*
 
& +  
!&5$ 4 B
 
=+  & -'/+
!=&8$ 4
 
 
<"& +  
!<&$ 4(B4
- ' 
  
 ' 


8= "
=+ 
 

9 "
) B

* 
0 -  ' 

5 - - 
* 
= " 
)1 
&
&5
 

' 
=+ 


B
 -
=+ 
 

9 "
) 


 - * 


 


9 "
 
&
=&8
 *
=+ 
*
-  +

  )) /
B
3  

<&
<"
 
=+ 
 
C
' -  

L - B

  
  
*+ B


5 - 
 
6& +  
!5&
 
6&$
' F 
  B


8)"
- 
9 -+-  " 
*
    
#
#,,+

 B

"

5 - - 
-

61-  " 

  
- 
- ' 
   B

 ( 
   #

 ' 
6  " 

=<=' 3  
*+

 


6  " 
#

#0
*
) 4B4B4

G   
 

% M *   
)1 
* 
7 

 - 
G   ,, 
#0# 

 
.F - " 
 
*
))  *  

 
  
% ', "  4
& 
 '
-

  

PB
Q

 
!$
P 
 * 
% '
B


8'/+) "

P=
* 
  ,,
 )  
Q 
 
"* ,  
% ' B

  

  ))
*- 

8'/+

  ,,
' 
&
* 
  
  ))
"
'
*


'  
 


G   ,,

Q 

"F 

% '
 )

8'/+  ,,
*-  

  
7   
  
 



% ' "   
2' 

" 
8'/+  ,,

% '
' F B




<
 
P= +  +

 -




 ' 
" 
C  
*
' 

*  *  B(
&
"
C  


   
 
 
"  

 
A 
7  -
 ) ' 
E3  

" 
7  -
/
=*
)1

.  +
 
 +  ,,
 ) B
"


A 
7  -
 

8'/+  ,,



 -

= +

P= +  +
'  
   



*+-
%+ 

 

 ( #
  )  

- B

  


<

% '
 
8'/+  ,,
*  * B
 
 
*+


 * 
D  


6 " 

 )-  

 


! $  "


% 
R

' -  '  4(B
D -

D -


D 



# )*  

 
5  M 
D 



# )*  

 
5  M  *
 
! $  " 
  
 ' 
 

-


 
 
  
!0'6 $ 
:-
-   
.  

" 
C   B



 )"3 

% " 
-
="
R
D  
%  
  

% "  

! $  " 
 
 ) "3

' -   ' 
!''
$B4

 *7 "  & '        3  =       ?  4  5 %   / $  -+- 0     $% 

! $  "
 

   B
 
. 
L 

   B
 
. 
L  
3   
 
3   
* 
" 
C  
'  
&
"
C  

-
 


# ")3-
*- 
'  
.  
&
% "  
' 3  

,  
G   

" 
C  


" 


A
7   - 
& 
8)"
=< =
 
"  
'  

"+- 
<+

G
-

+ -
%+ 
 

! $

 -

!
$ 
 

*   

' )
!
$  "
 
) - 
7

 
  -
* 
."  
' -  ' B



<
 
!
$
 *   4B
D -
 
" 
#) 
1' 
 
5  M *
' "
=      0  -
* 
G" 

,3)  *

! $
*
!
$
-   
  
 

  
)  - 
7 , 
 

%   
  
' -  '  (
& -

)1  
 
= " 
" 

  " 3
)1
46 , +- 
 " 
  
&  
%-
' *
-
L -
 
 )
  
% 'B
 3  

  " 3
)1

  
% '
' '  
  
#
D 


F 
  #
-

  !
% 

# )*  


,  
!
$  " 
 


6  "
#0

=<=' 3  
*+  B
&
# )*  

% M * 


 -
F - 
 -  






*+-
%+ 

 

*  "  
3   
!78$

&3   
!&$
F - 

 )) 
  
5  M * 
0
  


" 

F-

# ! 
/, 

 
!
$%  " 3
'   
   
2' 
 
 
!
$

 -
! $ 
' -  ' 
 B
)  
-
 
% " ) 
)1
!
$  "
 

)-
O'  

% " 

! $  " 
 
 )

!
$  " 
*+

- 
F - B

' 
 
 -  - 
6    ,, 

=<=' 3  
*+
 F 
%  
1' 

% ', " 

!
$
  B
 

' 

! $  " 
B
- )1 

* 
  
 
. 

" 
C  
 *  
&
% ', " 

! $

 -
!
$  " 


' - 3"
 )
% '

 
"  

 
A 
% '  
G
% '

* 
FA  
9  
*  * B
) 
*  -
=     0  -
 

! $

 -

!
$
 BB
 

' "  
7 , 

 ' 

%+ 

% , B4
& -
 
=   B
) 

 
-   
  B
"
 

   

% ,B
 
&'   " B
* 
  
7 

 
=   
 
L -
 )

   
* 1-" )) 
  B


" 
% " ) 

!
$
'
&  
. 
 
 )
5  

% " 

78
 
!  , - 
>. )*
!, '8 I
&$
 B
 -
-


 - 
6    ,,


!
$
' ) 
G
* 
  
 -
, -  
!
$
*   B

L -

   
- 
  -  
   
"F 
 
)1

 )'

!
$
 ) 
D -
1 

- 

E"- " 
, - B


-
  -


. 
 B
 - 
 
 
  
*+

  
% ', " 
 
  
 )
5 
  
-


 )"3 

% " 

!
$
  
% 



  
' 
' -  '  
 ' 
  
* 
  
 "  

!
$  " 
 
 
+! & !!! 
   
' 3 
  
, '* B
7 
5 ,, 
  -  
' )


% ' "   
E3  
%+ 
  #
 

# ' 
- )1    B
'  * 
<+""
  
.       
7 
5 ,, 
" 
*
-   
 ' 
<+""
  #
 )*  

 

!  , B
 - 
)1


% ' "  
-

#
5 *
*
' -  ' 
%+ 
  # B


 

!  ,
*
 

# '", /
) 
 B

)1


% ' "  
-
*
 



 )"3 

% " 

!
$

 

% ' "  

#
$%  "  &
'            "

# 
         &   ' &
  

   
   
)1



% "  )"3 

!
$
 
 )
5  

/, B

 ' 
-

9  

  
" 3
))  
61)
*+
  
  3 B

 
'"  

 -
,*-
 
 ,3   
 ' 
 

-


!
$  " 
  
 
   B
    B
+ !& B
-&    

'!!   

)1)
+*
  

# ! 
1'  /, 

  
&
9  
) 
 
 
:-" :%3 
 
C )  

8  
 

*+ 
:-
 
* 

 
-  A  
"*   
  
-  

&
.
  
 

' " 
. " 
% " ' 

G 3
5 )
!=)
&
E
- $
- )1 





+ !&
  
-
 
 
  
7 


&
.
  
- )1  
F " , - 
+ 
*  
-
L - B


&))"' 
  

- 
  '
 
G

F -
,-"  


*
 - B


:  
"1*
  
&
G *'" 
  
0 ' 
 
 
 
   
  - 

% "  
)1
  
 

78
 
!  , - 
>.
!, '8 I
&$
  B
 -
'  
' 

=   

!
$  " 

 
&! , - 
)  -
  
  
&
 
   
*   B

 

:

 -

8  
) /' 
 

- 


   
% " 

= 
    
!''
$
& 
8  
 
 ' 
L -
- 

3 
* B

-
 


* 
9  
' ) 




 )"3 

% " 

!
$

 

% ' "  

 ,7 :           0       %    0 A= &0/  (3' %  3  =       ?  4    5 %   / $  -+- 0     $%  3 4  @ 5    5    =    ;   

 
  
     
 
-  

3  

:   
*
)  
/,   

"1* 


+ !&
*   B


5 *
* 
 "1* 
B


O'  /, 

F- 
= 
"
' 3-  
  
/, ,   

.   -  B
 ,    
' 

/, B
0   
-   
# ""*  
 
# "* , " 


0   

8  ,  
)1  
*
"  
'   
 '  
G   
" 

 
 
   
!,6 $

 
,,  ) " )3A

-, )) 
-
 - 
. 
' 
4;8
&
F " , - 
G  -  


& ,*   
  

&
.
  
- )1 

 )F 

B
S
"
 - 
  
  
- )1 

 )F 

B
S
"
 - 
  

# #
"  &
'         & ' &
"+-
 
' 

!
$  " 

 
   


  
%
'  
 
 


 
7
=+, , " 

L  

3  
-  A-
  B

 
 

3  

. 
-


  
  
 
7 

)   
3  
 

-
L  



3  

8  


 
'  
3  

:   

&
 3
% " 



   
, 

+,- 
. "

6 

*+ 

6  "
#0

=<= ' 3  
*+
 

. 
"
  
  


" 
&3 B

A
&3


 
#  3 ,
!''
$


5 *)3-
'  
. 
' ) 
-

"



 )"3 

% " 

!
$

 

% ' "   

 

% ' "  

C  B
 - 
-

#  3 ,
'  *
 
7 
. 
 
3   
* 
" 
C  
'  

 B7 "  &
'         3 =     4  ; <   3  3 %  =     4  ;    3  C 3 %   /!  >  $% 3 %   /!  >  $% 3 / $  -+-  0    3 / $  -+-  0      $% $%

#
0   -
*
   
0   

6  "
#0


5  " 

!
$
 
3 -
*
78

&
G - 
 
  -

,  ) 

 
 
P /

:  
!" 
##B
'' 
$
7 
  
 
P /
 

-



7  - B

' 
  
/,   




+ !&
"1*
  B


'  
M 3
*
 - 
2' 
" 
+ , ' 
7  -
 )

   
% " 

 

   

"1* 
:  
- -
 B

 
 /
- 
)1

F-
/, 

!
$  " 
  
#
  
C   

 

*
"  B

'  -  
   
- 
- 

*
3



9F  "  
 
 
 

   
0   
 ' 
6  "
"F 

 
 
 )

5  " 

1-  
*+
' B
L -
"F 
 - B


  
6


:   
P /B
-  
% "    
)  
& 
8)"
=<=
' ) 
-

" 
C  
*- 

"  


A 
&3 
C 
-  
C 3

=<=
 


% " 
" '
  

=<=
"  


"+-
<+
'  

   
 '  

 "

 
 )*  
<  
&
   

=<=

-
) 
1' 

0 * 

3    
4B
 B
 B
 4B
# 

 (
"    

= ,   ,,

8)" B
3  
< 4


= , 1-" 

5+

= 

=+ / 
"   
E   


5 
*
3  B
 - 

# ,


%-"))


4


 )"3 

% " 

!
$

 

% ' "  

=+ / 
 


)1

 - )

=<=

 "  "
-


,   
F 
  B


"
C  

!
$  " 
 
 
 
"  

 
A 
7  -
 ) ' 

& -
!  /,  
 
 
 
  - 
 *
 


= 
 
 ..

 "

=<=


L   
  

P= +  +
-"
 )  

- B


=+ /


= ,   ,,

-" 
=<= "
 



<

=<=


 " 
1'   B
 

 
 
T 3

! 
   
 
  
! 
"

"
C  B

,  B


" 


A
7  -
  
  
#
L 
!  0  -
 

.  +  ,,


"  
7  - B
'  
 
0B
= 

# 4B
'  
& 
-
= +
 
' )
-
L 
.


A 
7  - B
'  
 
KB
+(B
 KB
04(

,(
!K
 
)1
3   
 

  
. $
2' 
 

 "
 
=<=


!
$ 

P= +  +B

 -

 "


! $ 
 
 
-"
  B
"
 

!
$  " 3
*  
  
<  -

 "


!
$ 


, B


'"

=
-

"+-
<+
F -

& 
0   -

" 
C  

!
$

 

78

&
*   B

  -

5 

 - 
-
  '

=+ / 
' ) 



"+-
<+
"   
 B
3  
-

  
3   
 -  
& 
 )'

" 
C  
  
3  



78



&B


A
7  -
-
 -
, -  
9 
 

!" 
##B
'' 
($

# (
3 "        & ' % 
)1
! $ 

 -
)1
-
3   



% ' "  
'  
' -  '  B(
& -
)3   

3    

" 
C  


   

-   
% ' 

 -  - 
% '  

) "  
5 ,, 



 ' 
7  -

*+
F - 
 -
' 

!
$
 
   



% ' "  
-

G * 

P= +  +

 -


' "
&
 )"3 

% " 

!
$  " 

 
   
*   B


 

!  ,


6 /'3
 ) 
)3   
  
3    
"F 
-
)1





 )"3 

% " 

!
$

 

% ' "  

% ' "  
 
G
  
 )-  
1' 

9

 
*
 B
  
 
.      -  

+ 
- )1 
.
)

5'- B
 
' " 
# '

  B( B


   

8'/+  ,,
-
  
 
 
  - 
  

 
' 3 B


   
* 

5'- 

" 3


 
   
'
* 
#" 3
"B



 ' 
7 

5'- 

=<=
' 1 
 
 

-  A  
5'-    B
F " , - 
G  -  


  
  

&
.
  
 

' " 
)1
. " 
% " ' 


G 3
5 )
!=)
&
E
- $
- )1 
&
% " 

'   
# '
 

 
 )F 

B
S
F
G - 
* 
5  " 


 
   
 
# '
 
 ' 
  -

!  , 
E3  

. 



!  ,

. 
7

-  
)
 B
 

. 
7


!  ,

. 


)) 
)

!" 
###B
'' 
$
E3  


))  
)

8'/+  ,,

# '=<= "
 B
/ 
 

"" 
 -  


-   
)
!" 
###B
'' 
$
&
8'/+  ,,

 
1' 
* 
E  "1 B
 
-

9 
K

+(


* 
-

 B
"  
!" 
###B
'' 
4$
&
 "
&*

 
* 
8'/+  ,,

 8'/+, +,+ /, , 
! 8==$
' 3 
B(
S
% 
*

" , - 
/,   
 

 
*
 

  

- -
  
" 3

-   
"* , B
 


 -
 
8'/+  ,, B
  - 
 ' 
*  
- B


0   -
* 
E +,*+

. 
 
  -

9 " 3


N

=+ 
' A
' )


" 3

= 

"* ,
' 
,6 
 
 " B
L -


0   -

=+ 
*
= 
 
= 
' * 
!" 
###B
'' 
$
 )  

" , - 
 '
  

9 
K

+(
3 
  - 
.
)

&.& '"
!
'- 

 
' -  ' $
  

3   
< -

= +
)1

=
K

= +
)1

=
+(

  
.  
 )* 

-  A 
G  - 

   
*   B

 -
&,, 

& )-   


=
 B
' 3  

" , - 
" 
!''
($
& 
9 

' )
-  
)1

   

% ' 




 )"3 

% " 

!
$

 

% ' "  

 D7 ;   %  =        3  %  =    $  E!0 %

E3  

*  
+(=
 -
 
" 3  
 -  B
' ' 

=3)  * 
- -

"* , 
 - 
#
5 *
*
   
-

.  
=

<


 .  
 -

  
 
" 
" 3  
 )B
*  

 
E -  

"* , ,3)  *

&,, 

& )-   
)1  
*
 
)
3  
#" 3

  
& 
3   
,
' 


))  
)

!  ,

%*'1-"
* 
 
 
!" 
###B
'' 
$
G

% )"*
  
3    
)1

" 3

*+
*
  - B
  
.  
  
3 

 -  -
0   

*  
3   ,, 
!'- B
  B
  B
- B
*   $
- )1 

*  
- B


%*'1-"
- 
  
)1

" 3




& -

 )"3 

% " 

!
$  " 

 
   


 
  
.      
"

   
%- 

0 3 
  
*+

7 *
 )
 
% ''  B
  * 

%  , *)3
- 
  
 
  



C 
* 
 
=   
' 

7   
-   
6    
 -  

' - 3"
-
- 
 
 
 )
    B
 -
A
G - 
* 
 - 
% " 
 )  
"F 



  

 


%+ 

      
 

(
      "!      :- 
 
 
% "  

! $

 -

!
$  " 
 
' "


 -
 

% ' "  
 
  - 
  B


  
  



%+ 

      
 
*  
  
#

, * - 
# 
) 
 
 
*+"-

- *+"-


   
%  *  
 
   B
L -

 
   
 
-

  )   
 
*  

 

%    " 3
' 

  

% ' 

A 

"  
% '  B


+,- 
 )'
  
.  ' 
P= +  +
, -  4B
&  
'  
 )

7 

-   
% '  


, -  
7  - 
& -

, *)-
7 

"  
'*

A 
% '  

% ' 


" 
'*
A
7  -

 B


, *)-
2   

" 
C  
*  B
 -
*

7 
 
 
)1 
&
6  
  
  
2   

" 
C  
"
*
%  " 3   
)1  
0
 
' 
*+"- 
% ' 
))
 
* B


)
- 

"- 
G   

 
A 

"  
% '  
  B


' 
+,- 
.  ' 



6

O' 
%  ,3)  * 


- -
   
%  *  

-
- 
 
' "
* 
7 , 
)1

    "
%+ 

 
  

  

6
8 /
!8 /
 %-   B
8 /
#-B
G%$
' -  '  ((
= 
  #
*   
*7

  "
%+ 

!7 * * +$  +,,  4
-

 (4
&
' 
' -  '  
%+  

 '   *+"- 
      
 
*
7 
  
' 
'    
 ,3- -
% '

 
A 
% '   (B(B(4B(
#
9  
  
&"' 
  - 
-

G * 
 
*+"- 
 '   
% ' 

 
"  
"+ '  
 )1
 
.  +-+- /

.  '
3 
!''
$

(


  

 


%+ 

      
 

 F7 "!    (= !! 6 3 +3:G?3:"!  < 5   ?    ;  $  -    

0
" * 
  
* 
  
= '" 
F)) - B

-

"  
*+"- 
 '   
% ' 
' )  (B(
E3  
9-  
  #
-

-   
"  
*+"- 
 B

8+- /B B
.  +-+- /
 
B4& +  ,   !$
' ) ( B
  -  
.
  #
' )
.  +-+- /(

( 
$  !  "!   0   ( = !! 6 & .' C 
-  
 B
 
! $

 -
!
$%  " B
  
 )

G * 

- - 
.  +-+- /
  - 
 ' 
 
-

 

  
!0'6 $

  
 
D -
 
 
 
" 
%  " 3
1' 

.  + '  
 )B


- -
% '
" 
  *
  

 
G  - 

&     *   * 
*  


  " 3
)1

)1  

  ,, 
&  

-  
F -
 )  

'  
3  
-   
, *)- 
7 

% ' 

" 
C  B
   ) 
-

6  

 
A 

"  
% '  
E   
"
)   
  B


  ,,

 -  
!
$=3)  *
 )1 
 B


5 *
* 

 
! $%  " 3

*+
 4(BB
&
G-
 )1
"
 )

+ - 
 )' B

  

"  
.  + '  B
* 1-" )1 
  
C 
%+ 

      
.  +-+- /

 


%  " 3


.  +  ,,

 -


  ,,
) - 
G

)  
%  " 3


.  +  ,,
*
",   B


)1
 
*+"- (

-  *+"- ((B(

' -  ' B

*+""
  
  
&
 
 
  

 ' 

 
"  "

 



  

 


%+ 

      
 

 "
"' 
  B


%  " 3

.  +  ,,
*
3   
!''
$

 97       & ; 3'    &'  "!      (= !! 6 3 +3:G?3:"!  <    ?    ;  $  -    

(  
     $   "  
.F - " 
    "

*   B


0   

"  "  (B((
#

 * 
D  
) 
 
"  " B
'  
*+"" B
  


   
-   
-  
7   
& 
0 

'
   
*+
 

E )
*  )3  
#   

9   

#   B


C  ,

9 "

')
-
-
* 
E   


F -
'

/-
#  
 "

' B
- /-
= "
*  * B(

#
0   -
*
- - 
%+   
  
*+"" 
-  
0 
 )
*+
'  
  

3 
%+ B
' 
, + - 
,E  B
 ,   

, 3   -" 4B
&
 
7"+ 
3 -
9 "'   
' * B


'  

" 

*+ 
)
)-  B B


' 
- - 
.  ) +  

6

E   
' * 
*+


8 

9   " 3B


- - 
0 )  
-
)1   

%- *  ,, 
'*
"   
'  
,  
 1*
  
 
 

-   
0 

L -


%    " 3

*+ 
& -

'
-   
*+
"F 
  
%  *  
  "
 )1 
  
&
0   

*+ 

   * 
=3)  *
F - 
 A  

C 
*
 
0 * 

-   
= " 
-


  '
*+""  
"  " 
 
-

  B
   B
, 

*+"-
') 
   
!''
$((B




  

 


%+ 

      
 

 7 ; !0   F#

#

  
"


% '
1' 
  
#  
*
 
= "
  *
#

   
"
) 
 ' 

 - 
9 "

8)" 9 -+-
 

8= " )  


7 , 
 )1


=+  & -'/+ %+  4


<"& +   J5 - & +   9 -+-  %+  4( 
#
G - 
* 
   
"
) 


,  
"
* 
9 " 
B

1' 

8)"9 -+-
",, 
 B

 
'  
= "

E 
 
7 

  
0
 

-

*+"-
" 

7 , 
 )1

+-
" -
9- ,  

& -    
: 
 

 

9-  

* 
#  
  *B

 
""   
- B
3  

#  
 
* 
% '
 

7 , 
 )1


  

.2
* 
   
,-
  
-  
  
&
/"
  

"  " 


L   
9 "" 

- 
 
F - B

* 
7 , 

 
  
"

  
"
)1

8)" 9 -+-
  
 
" 
0 -  
*+" 
"F 

  
"' 
  

*+"


 * 
C 
3" 
 ) "" 
' 
*+"" 
  B


 "  BB(B

(  #
   " 
F 
*

2/  " B


    "
9 "

PJU  3  
 

 + 
"+   
 ' 
) 
* 3- 

+  ) 

:&!=$
* 
"  
'   
6.:
B
 - 
 * 
 
#

*  
  "


+ 
 )

PJU  3 

1'   
&

 "
 

' -  ' 



 

: 
V2
> 
*+ V
!2>$
' "
#

: 
  
 " 
 ,3- -

& /)*   B
*7

:  

: +-  B
 


7   

. ' 

% "  3 '  (
 " 


=)* B
 ) 




  

 


%+ 

      
 

7"  B

* 
7 , 

    
88
4B
  
#
  
7"  
  
 
 -" 
!@ B
@ 7

: $
 )* B

, -

*+"-


 +

 -
*+"-
#

8 
 * 
# 
  

"  " 
"
' )
 
  
*

.F - " 

  " 
   

= " 

'
*
* 
%  *  B
"

' 
" 
* 
  

 " 


#  

61

%+ 

      
.  +-+- /
 

 "
2>
 
!!& ! ! 
3 
& 
 "
 


N
! $%  " 3
)1

%+ 

.  +-+- /
 
.  +-+- /  
 )
#

"  "

   

 
  

   
 "
2>
"


!
B $.  +-+- /

 
&    O' - 

4
N
  
  
!''
$
& 
 
N E 

 ' 
 )


%  " 3


' 
  ,, ) 
* 1-"* )1  
& -
=   

)

&.& '"


%  " 3

 


) "
'  
  

( #
(3=   3   > %    & .' 
'
-  
')-
E "* B

) 
"F 
  
.  
)1
' 
% '
 
   
 *  
 
  
) 

&.B

 )
% "   
' 
& '" 
%
'   
& 

-   
T  B


% " ) B
% M *  B
<  "" 
'

*
.     
 

<  
 - 
& 
&.   
 

% "  B


 - 
% "  "
 
'  
0 B
*7

=<=' 3  
*+
 

6  "
#B
 )  

% "  
  
 

*
 
A 
% "   
 
& 


V " V9 


) /' 
7  -
  
#
"
' ) 
-

% "    
'*
3    B
 
=

 
*+ 
"   
 B
3  

) /' 
7  -
*7
,
*
) 
 B

-
 )
5 
  
 
6 /'3
- 
  
 
&
-

"
' )  
= 
  
-

&.& '"

   
%
  

3  

  
=

*+

- 
3 

  
% M *  
"
& -
 '
 




  

 


%+ 

      
 

& '"
' )

0   -

-   
=  

*+ 

 
 
% M * 3 
!'

N$
*  

#
 
% , )

&.& '"B
*7

=<=' 3  
*+
 

6  "
#B
  
% ') 
  
& 
% ') 
  
% "  B



# 3
*  
 )  
 
  
*  
% "  
  
7<%% - 
- )1 

' )
= 

*+
 
% " ) 
*
   


L   
% ') 
*
*   
&
" 5FA
 
% , )

' 3 

  
 "   
0 )
D
-   

% "  

   
*+
 B


"  


" 
&
' -  ' 
% , )

=<=' 3  
*+
 

6  "
#

* 
G  - 

 
 *
  
D -
-"

" 9 
- 

3  

" 
C  
'
&
 


-
' 1 B


0 * 

9 " B
   B
   B
& -'/+   B
-B
-

*+
  
% , )
"+ 
  

-
* )

" 
C  
"
 -  
"F 
G


G  - 
*
F - B
 

% '& '"
 
& 
2   3 -
 &.& '"
  
 ,3- -
% M * 


###
 
%
" 

3   

" 
C  

"
' -"
   
 

  
& 


& '"
" 
3     

*  
= B
"   
.  

* 
$
)1
% ', *)3 
 
  
%*'1-" 

&3   * 
* 
%'  

= 

)
- 
"   B

-  
' 

3  , 

 -
9  
  ) 
G

6 "

*+
' * '  B
 
' 
3  , 
 
  
  
&
"   
. + : * "
'
1' 
-
= 
 " )

 )* 
"   
.  
 

* 3*-
 
G  - 

 " 
%


"   
. +

   )3  
- 
   "*

  
& '"
7 "
0    

3  


: 
"F 
 )-  
1' 
' B
'

' 
. 

 
'  
=
  

 
- 


3 


: 

  
=

 )B
"F


 
 )

F -
6  ,' "
 
#"   

*+
 
3     
'*
"   
3  
F - 

   

  
$ 
.
)
  
$


F -
 
' " 
*+
*7

F - 
% ', " 
*
*   
!" 
0B
' 
$
7 



  

 


%+ 

      
 

3     

- 
*+ 
"F 
 )

' 
" 3
  
 ' 
 

-
L -
 

 
 
C 3*-
"F 
  
6" 

"+- 
 - ) 
 
*+
' )  
.
)

&.& '"
"F 
.  
 )* 
  B

-
- 
 )

C )   
1* B
 
 )
' "
0    

3  


: 
& -
"F 
-  
.  
'  

0) 
  -  
  B

-


: 
 3
- 
-  * 
" 
 )  
  
-
"F 
"   B
' 
)) " 
7' " 
'*
*  
.  
)1
, *)-
6    
 


%-   )
 * 
  B

( (
=               & .' #
- )1  
.    


 " 


 
  
!, '8 I
4T$


/  
 B
 
=<=

.  +-+- /
'   B
 

3   
< -

 )* 
&  
3   
"F
) 
 )

  " 3
' 
.
)

&.& "


 
 )
 
   
2  3 - 
 % , ) B
 
 
F -
.  

  
=
' 
  
% - 
3  
  
 
  
3 ) " 

=<=' 3  
*+ 


6  "
#
  3 
:- 

.  
F-

# ! 
1'  /, 

 

  
 
 B
*   
  -

0 
0'6
< #

0'6
< 0

1-  
9  

'-"
 )
" 3

%  " 3
  


*+""


 "
2>
"


0
0'6
< 0

&    1' - 


N
)1

!
B $ .  +-+- /
 - 
  
!''
$
#
5 *
* 
E +,


'   
0
< 0B
*  

. 
< #

G" 


,3)  *
#

"  "
"

! B $.  +-+- /

 
9  


N
  
  
7
*
  
C , "

G  -  


' 
) 

'  
&    
"


,3,
7"+

-  A  
:.9 G  -  B
- )1 

D
&-"  ))
 

' " 
+-
7- 

G 3
5 )
!=)
&

E *$B
"3


' 
) 

'  
= " 

4


  

 


%+ 

      
 

#

"  "
" 
 -  -

( )
    &
'" % $ (= !! 6#  & .' &
%+ 

* 
  
&   
!!
B
$

! B
$$
) 

 "

3*- 
  " 3
)1
.  +-+- / 
!''
$

 #7 "!   &
5(
'  & 5(
' (= !! 6       &
'    & ; 3'    &'



  

 


%+ 

      
 

E3  
)1
 
% '
 
 " 
)1

! $

 -
)1

!
$ 
' "
 B

 
- 

6
)1
.  +-+- / 

' -  ' 
E +,*+
  

! $%  " 3
 ) BB
<  -
7  "
  #
' -  ' 

 " . 
 
!    
!>ML.$B
 -

!
$  " 3
 )  (
& 
0  - B
 )
5  
  
.  B

,3)  *

 
  
' )
 ,  
 " B
@ 
 
    
88
4B
*
3 B
   

F- 
= 
 -
- )1 
/, ,   
  
.  
-   
 
  
5 
  

' -  '  
&,,   
!8+,J#  

8+,J4,$


E +,5 

 "
>ML.
 )1 

F -  


C 
*
 
!
$  " 
 "
)1

% '
.  +-+- / 
&
&,,  
>ML.
8+,J#  
 

'  
F-
/, 


. 
>ML.
8+,J4,
 )B
 -

   
0  -

>ML.
8+,J#  
- )1 
  
 )

9  
  
 "
 
*-  B

G  -  

6*,-"
  #
*   B


 "
>ML.
"
" 3  
-

)1  


 )  

"-
9  

>ML.
.  
3
,*, 
F -
   B B
L -
'  
5** '"+ 

0 
 
   
8)"9 -+- + 
E   
 
" 
  
) 

# ! 
. ' 
 )

9  

   
!
$.  +-+- /
' '-  

"  " 
  
 , - 

  
C  B


0
>ML.
8+,J#  
'    B
- )1 

( *
  "!    (&
'= !! 6 & .'  

* 
- )1  
G  -  

< .  

0'6

' "B


.  

!
B
$&   
1' 

! B
$&   
' * 
&
F- 
%    " 3
)1

!
B
$&   
 
 ' 

0'6
< #  . 
 )
#

"  "

>ML.
8+,J#  
"

  
.  B
 

* '  
   
*+

 -

- /,  
*+


5** '"+ B

&      " 3


N
)1
!
B
$ .  +-+- /
 - 
  
&
5** '"+

- /,  
 "


 '

, 

* 3*- 
.  

' -  
' 

9  


G
L -
 
G*
*
 - B


C '

<&J5&
* 
5** '"+
) - 
E3  

  "

 "
' )B




  

 


%+ 

      
 

""   
-

#  
.  +-+- /

 
- 
  
* 
= "
.  +-+- /
  *
& -

* 
C '

<&J5&
!*
7 

9 "

-

$
"
-
(

 
G*
 - 
  

&
%+ 

! B
$&   
 
 
-
-   
G

 
*
) B


! B
$&   
' * B
  
' 
 

.  +-+- /


 " 


 
  
-"
& 
  
= B
 


A 
!($B

 -


"  
7  -
!= 

0$
  
 )* B


%+ 

! B
$ .  +-+- /
' )  
"F 
.
)

2  3 - 
 
&.& '"
 

3    

  
 
= 


: 
 
+ 
& ) 
  
4
.  
" B

# ! 
/, 

  - 


.  
*  
' 
G*

- - 
.  +-+- /

=3)  *

! B
$&   
1' 

!
B
$&   
' )
 

C '




C  ) 

)
$

&
#  

 



 * 
D  
"

 ' 
 

6" 

 
 )

 )"3 

% "  

 B
' 
 -
 )
 
  


%+ 

 
", /  
 
 
#
  
' 
' )
-
-


% "  )"3 

!
$  " 
 
 
   



%+ 

      
.  +-+- /



!
$  "

 
   
*
"  

 
% " 

 
 )F 

B
S
*
F 
& 
'

,  
 )'

" 
C  

!
$  " 

 
' 3 
E   
 


% ' "  
  
!
$
  - 
&  
) 
-

 


# '
5'- 

-
.      
 ' 
 

 

 )* B
 - 
*  

 
 

 
+B
)1

   

8'/+  ,,


'*
=+ 


A 
7  -

" 
C  
-

& -
 
" 
"

0 3 
1' 
!
$  "
 
  
  

'  
 


5  
)1
* "1)
2,   

   
  
*+


   

 
C 
%+ 

      
.  +-+- /
 

! $  "

 
  
!0'6 $
  
E3  

  " 3 
)1
 
% '
' "
 B
 


* 3- 

)

&.& '"
)1

  "
  

.  +-+- /
, 
  

,, 

   
.  
0'6
< #

0'6
< 0

-   
 " 
)1 
* 
)  - 
%+ 

 

F - 
 
&    

 
9   

'
*

N

 
G3* 
'
*

N
%
"

)  -
  "
%+ 
 
*+"- 


 

 
%  *  
-

*+""
*  
  
& 
 

.F - "  

"  " 

 

 " 
* 
%+ 

", / 
      
 

(


<   * - 

*
+  


 ,,  + B
%W
)-"B
DW
7  B
W
' B
.
%# '# "& #
FF#5
/0 B
44( 4

% B


 #
#
FB5
12 B
4

 B
#
#W

7 "B

W
%)) B


.
6W
&    B
7W
9 L B
#
.
8
.
' # + # $
#
#9#5
02 B

4
F  B
.B
7- B
G

#
   ! &3
.+B
2B
5F B
B
& *B
E
 W
 
E +08I
E  B
#9# W
(

 B
EW
6-  B
.W
6 - B
8
%W
< -  B
W
. B
6
5W
= *B
&W
9L ,B
W
% B
D
W
%B
9
8W
%B

# !
# #
#9(5
*. B


F  B
.
7B
G

"& ""&
#99F5
*B
4

."B
.W
="B
%W
% B
D%
 # '! # !& #
#9#5
41 B

(
 B
W
% )) . '  B
6W
F  B
.
" # # "& #  #
#9)5
*4 B
(

9 B
DW
7 - B
7W
%+ "B
8
' 5
#9#5
6B


7    B
=W
 ' B
.
%W
7 '+B
8
#
" & "& 3
8 B

.W
>B

 W
  I
  B
#9# W
4

 ) B
=W
<9B
DW
D  B
D
%W
M B
:W
: B
EW
E  +B
D
.
!& # #
#95
*78 B
44

,IJJ, -+ -  -J,   B



. -  W
5 
5  W
 +

  W
=  
9  W
%- 3)   B
.
' !&     # ,& !&  &     (5  # W
E - )-
0    - )I
%  B
#99D W
0

4
.B
&
8W
= B
9W
 B
8
DW
.B
8W
5B
5W
6 B
.
&W
8B
9
6W
8B
9
E
!  (&  !  #
FF#5
*6 B
44

"B
D
.W
EB

D
:
# "& 9 :# # #99)5 1/9 4(B
# "& #
#99)5
**2 B
(

 1B
=

5 # & !&#
#9995
*B


7 B
G
# &# " #
#99#5
/11 B

(
% B
5
6
# %# & ! #
F,95
*0 B


#B
W
 B
W
% * "B
W
2 B
W
B
W
# B
>W
> - B
>W
 B

!!
#995
/6. B
4

% , B

&W
7  B
8
D
EW
6 B
7
D
!& # & ! #
#9995
04 B
4( 4

7 B
.W
&- B
W
'- B
W
 B
7W
A  B
.W
%1 B
9W
C "B

# "& 9 :# # #99)5 1/9.88)8619 # "& #
#99)5
**2 B
((4

: B

8W
% *+B
.
# &# " #
#995
/06 B
(

%B
9
:W
8  B
%
7
:# # "& # !# # #
FD*5
61 B
44
4
7 - *B
W
- B
0W
7- B
G

!     (!&  W
E + 08I
E  B
#99* W

 ) 

E  B
%W
# B
%W
. B
W
 B
G
; "& #
#95
*/ B
4

56 X *B
0W
7 B
W
5B
0
# &# " #
#99,5
/18 B
(


9L"B
6W
%  B
9

(&
#99)5
27 B

(
7"   B
6W
&- B
W
 B
7W
<  B
E
%#  # "& # <"& )=
FFB5
//4 B
((4

 / B
.W
 B
W
&B
.
DW
.  B
.W
  B
:
D
# "& #9 :# #
#99#5
1* B
(B
# "&
#99# B
**1 B


8B
9W
 / B
.W
&B
.
DW
56 X *B
0W
 ,  +B
8
W
  B
:
D
"& "&
#99*5
2B




<   * - 


. " B
W
% * "B
.W
%  B
%W
B
W
>  B
W
: B

!#9 !& #9 !& #
#95
.0 B
((

. " B
W
B
W
>  B
W
B
:W
>B
W
: B

 # '! # !& #
#9(5
4. B
(((

5 B
.W
.1 B
W
E  B
9W
F  B
.
%# ' # " # >  #
#9)5
**7 B
 
4
# B
:W
>- B
8W
 B

%# ' # " # >  #
#9995
*7 B
(

' B
.
DW
0X*M *6 B
W
E B
:
7W
. B
D
8W
7 B

%
# "& #
#9#5
*61 B
444

8  B
=W
. B
=

"& #
!#
#995
*B
444

 B
8
=W
8  B
E
E
!& 
F,)5
/B
(4
(
8  B
E
E
!& # &# !
F,(5
2. B
4

8  B
E
E
!& # &# !
F,(5
2. B
(
4
8  B
E
E
!& # &# !
F,(5
2. B
((
4
B
W
- B
D

 !&
#99)5
./ B
(4
4
D B
D
:
" # #  !#  #
FFD5
8B

4
&  B

8
!# + # !# !#
F,,5
00 B

44
&  B

8
#
#   #
 #  ' #  # W
D 
E +
R
%B
#-B
#99, W
4
4
% B
D%W
B
75
!& # #
FFD5
27 B
44
4
*  "B
&W
<  B
#W
8+B
&W
5 '*B
5
.W
9** B
&W
 B
E
# "& #
#99D5
*67 B
44(
4
F  B
.W
1 B
%W
9'B
W
7- B
G

"& "&
#99D5
4B

4(
% B
D%W
B
75
!& 
FFF5
20 B

4
*  "B
&W
<  B
#W
8+B
&W
9** B
&W
 B
E
# &# " #
#99D5
/07 B


. B
6
5W
% B
DW
 ' B
W
 B
E
"& ""&
#95
/B


%  B
8
W
D +B
D
.W
.  ))B

8W
. B
D
8W
B
%W
D B
E
9W
8' -"B
D
8W
 '' B
W
6 *B
6
DW
7 B
DW
&  B
=
:W
 B
5
EW
 B
5
D
!!
#995
/64 B


= - B
9W
= -- B

'"  #
#99F5
*7 B
 
7 " - B
6W
7   B
W
 B
W
B

W
6 +B
=
W
& B
8
<
!# + # !# !# ?# # #
#99)5
*7* B
((
4
< , B
7
EW
9 *-"B
6
DW
6 +B
=
W
9 B
&
!# + # !# !# ?# # #
#99*5
*76 B
((4

 B
7>W
8 B
7 W
> B
W
  B
W
B
75
%# ' # " # >  #
#99*5
/. B
4

%-   B
.W
< " B
:W
71  B
7W
%-  B

# &# " #
#9(5
/00 B


% )) . '  B
6W
0-" B
8W
 B
W
%- 3* B
%W
 - B
W
%   
 ' B
.W
< B
W
7    B
=W
7- B
G
W
F  B
.
"& ""&
#9(5
0B

(
% )) . '  B
6W
0-" B
8W
 B
W
%- 3* B
%W
.   B
W
%   
 ' B
.W
< B
W
7    B
=W
7- B
G
W
F  B
.
"& ""&
#9(5
0B
4

% B
D%W
B
75
%# # "& #
#99#5
2. B
(4((

% B
D%W
B
75
!& # #
FFB5
00 B
4((

% B
D%W
> B
W
D B
DEW
="B
#W
B
75
 # '! # !& #
#99(5
2* B
44

8 B
7 W
="B
>W
% B
DW
B
DW
 B
DW
< B
7%W
B
75
!& # #
#99D5
44 B
(4

="B

W
D B
B
=&7& '"  I
)"> /+?:
#9 




<   * - 

4
 ' B
.
%W
8L B

W
Y"B
.W
'  B
>
9W
E B
7W
' B
0W
6  B
DW
7    B
=W
< B
&

$ %#
#9#5
6.4 B


. )B

%W
B
9W
B
%W
-*B
.
DW
.-8 B
W
5  B
&
W
.+B
W
8  B
D
%W
 B
.W
0'B
W
. B
DW
5  LB
%W
E B
D
E
 # #  #
#9(5
62 B


9 - B
8W
< -  B
W
%-  ) B
W
%" B

  !# $ !#  #
#9(5
8* B
4(

#"B
W
> B
>W
*"B
:W
#" "B
>W
 B
D
 # '! # !& #
#99,5
24 B
4
(
%   B
.W
7 '+B
8W
< '  B
<W
7    B
=
"& ""&
#995
6B
(

F  B
.
%B
%W
D- B
W
9'B
W
7- B
G

+# "& #  #
#995
2B
((

7 B

%
+ "& 
#995
2B
4

%- 3* B
%W
% )) . '  B
6W
 B
W
F  B
.W
9'B
W
7- B
G

# &# " #
#95
/0/ B
444

="B
%W
& B
D>W
% B
D%
 # '! # !& #
#9)5
48 B


:'B
W
% )) . '  B
6W
 B
W
 B
#W
%- *" B
8W
F  B
.W
7- B
G

"& ""&
#9*5
.B

4
& B


# "& #9 :# #
#95
07 B
4B
# "& # #9 B
*6/ B
( 

%+ B
8W
. * B
9
DW
9-  B
:W
# , B
.
:W
 B

8W
< -  B
D
W
E B
D
.
$ %#
#9)5
68* B
4

Z+""B
W
5 ' B
8W
% "  B
5W
%- 1B
.W
%- 'B
W
5 ' B
W
%  B

, +
#9)5
4B
(

 B
W
B
5W
0  B
5
  !# $ !#  #
#99#5
1. B
4
(
 B
W
5B
.W
2"B
W
.+ B
#
(& "& !
!
#99*5
0B


'+ B
W
.+*B
%W
  B
W
. B
W
 B

(& ,#
FB,5
*. B
4
(
9 B
9
9
!& 
FBB5
*2 B
444
(
<B
DW
  +B

9W
>B
DW
7  B
5
<W
5 B
7
W
B

.W
B
.
.W
,B
W
< B
CW
%B

9W
 ,,B
.
&
# ,#
#9)5
*2 B

(
8'B
6
<B
# 
= ,,-I
E2
J44

#9# (
8 B
W
<  B
#W
5B
0
"& # " #
#9(5
14 B
((
(4
=  B
CW
E B
D
EW
 B

8W
= - , "&  -B

<
W
8" B

D
# ,#
#9)5
*2 B
((
(
%B
W
. B
6
5W
5- "B
7W
 ) B
%
6W
6'B
E
.
6W
5 B
5W
 B
E
# &# " #
#9(5
/00 B
(
(
.B
&W
6- B
.
8W
8B
.W
= B
6W
5B
6
5W
7 B
W
9 B
%
"& ""&
#9*5
.B

(
9-  B
:W
%B
9
8W
< -  B
W
 B
EW
E B
D
.W
 B

8
#   # "& #
#9*5
*/ B
((4((
((
% B
W
 B

8W
E B
D
.W
E  B
9W

<  B
W
2)) B
W
E , B
9W
= B
.W
9 B
&
# "& #9 :# #
#9(5
06 B
B
# "& # #9( B
*60 B
4(
(
 B
W
 B
.W
&-"  ))B
DW
E *B
W
8[  B
W
7    B
=W
7   "B
W
7- B
G
W
F  B
.
# &# " #
#9*5
/0. B
(((4

' 
B
W
%*B
:W
7 *B
6
%# !& #
#9(5
*28 B


. -B
%W
<+ B
5
DW
7 *B
6
%# !& #
#95
*00 B


9--B
W
7 - B
7W
%-   B
D

# &# " #
#95
/0/ B


. - B
DW
=  B
8W
2'   B
:W
.   -B
.
&W
7- B
G
W
9 ))B
6
"& # " #
#9*5
0* B
((



<   * - 

4
2*5  B
#W
5-\D - B

" # # "& #  #
#9(5
*. B
4

%B
9
8W
9-  B
:W
7 B
W
 B
E
" " #
#9)5
1B
4

  B
0W
  B
:
D
"& # " #
#9*5
0* B
444

7 "B
W
%-  B
9W
 'B

# !
# #
#99(5
.B
4
(
 B
EW
9 , B
.
" " #
#9)5
1B
((

2'   B
:W
=  B
8W
. - B
DW
 B
.W
%AB
%W
7+ B
W
%- )B
=W
#M'B
:W
9 ))B
6W
.   -B
.
&W
7- B
G

"& ""&
#9(5
0B
4(

 B
9W
9 *B
.

"& # " #
#9(5
14 B
4

5B
.W
  B
.W
=  B
8W
=  B
#
0W
D-*+"B
=W
51 B
=
=W
7  B
%W
%- , B
W
5B
W
7-" B
.W
5 A B
W
< +B
D
=W
 B
D.W
 B
5W
=B
5
DW
- B
EW
7- B
G

# "& #9 :# #
#9)5
0/ B
444B
# "& #
#9)5
*62 B
44

  B

%W
5  B
D
.W
%- B
:
%
"& ""&
#995
6B
(4

 B
9
.W
. +B
0
%#  # "& #
FFD5
6./ B

4
, B
=
W
6/B

.W
. +B
0
$  %#
FF*5
4B
(

2'  ) B
5W
% "  B
5W
% -" B
8W
6' B
W
5 ' B

"& ""&
#95
/B


7   ,B
8W
% B
DW
7  B
GW
 ' B
9W
.-  /B
=W
%-  B
W
8  B

 ! 
#995
4B
44

E'  B
2W
8 B
E
+ !& 
F(#5
67 B
((((
(
EB
9
W
7 - B
:
8
'! #
#99#5
*18 B
4

=  B
8W
F*- B
9W
 B
.W
%" B
<W
9 ))B
6W
.   -B
.
&W
7- B
G

"& ""&
#9)5
2B


, B
9
W
D B
DW

7 " B
E
D
W
< ) "B
:
5
W
9L B
W
#B
W

C  B
6W
D- B
W
7- B
GW
0  B
5W
.

%B
0

=W
%- ,B
=
D
  !# $ !#  #
#995
.8 B


% )) . '  B
6W
0-" B
8W
 B
W
< B
W
.B
W
:'B
W
%" B
<W


7  B
W
D B
DW
7    B
=W
F  B
.W
7- B
G

!& # #
#9*5
// B
4

 B
9W
D- B
W
7- B
%W
, ]B
9W
= B
%
W
5B
.W
7F- B
&W
D ]B
DW
7- B
G

"& "&
#995
** B
4

:'B
W
5B
.
5W
=  B
#
0W
 ,B
.
<W
%-  B
.W
7- B
G

$ %#
#9(5
687 B
(4
4
6 - B
8
%W
 )B
.W
.  B
.W
%B
9
8W
7  B
W
9 "B
D
:
W
9  B
EW
C , -"B
6W
 B
E
# &# " #
#9)5
/02 B


8+B
&W
E" B
8W
X B
5W
6' B

!& # ,
#9)5
/2 B


% B
.
W
7 L B
%W
% B
>W
%-  B
.W
7- B
G
W
. B
&
%
"& ""&
#9#5
1B
44

6 B
>W
8  B
W
E  7*B
&
!& # #
#9(5
**7 B

(
7  " B
&
DW
 B
%W
  ' B
W
9 *B
.

# &# " #
#99F5
/0* B
( 

6*,-"B

7W
   B
%W
* B
W
8  B
W
 B
:W
.-  /B
=
 5   #
#99)5
/2 B
(

* B
W
6*,-"B

7W
E" B
8W
%B
DW
7  " B
5
=W
.-  /B
=W
8  B

%#  # "& #
#99*5
687 B
4

' *B

%W
. +B
0
%#  # "& #
FB,5
60* B





= '"

,
- 

#
9  
  
' 
F)) - 
= '" 

 

L   
  I
  /!   0! 1! $$   $   &
'    $      ! "# !$B
2B
4H4
&2#I
J%44

.
  KB
<
%"  KB
5D
=B
.
F  B
G
7- B
E
- 

.
  I
= "B
& %  B
=*   

    B
0 ) 

&" 

. ", W
<
%" I
= /, 

9   B
0 ) 

&" 

. ", W
5D
=I
G 1* 
' 


&   W
.
F  B
G
7- 

E
- I
 " 

&" 

. ", 

  /! 0     $   &
'    $      ! "#
!4$B
3B
(
&2#I
J%44(4

.
  KB
<
%"  KB
5D
=B
.
F  B
G
7- B
E
- 

.
  I
= "B
& % 

=*    B
% " F  B

  

+

% " B
0 ) 

&" 

. ", W
<
%" I
.    B
= /, 

9   B
!  0  - B
0 ) 

&" 

. ", W
5D
=I
G 1* 
' 

&   W
.
F  B
G
7- 

E
- I
 " 

&" 

. ", 

  "        $        $   &
'    $      $ % 
!$B
(B
44
&2#I
J) '4

<
%"  KB
.
  KB
.
F  B
G
7- B
E
- 

<
%" I
.    B
= /, 

9   B
7  

*+" 3 B
0 ) 

&" 

. ", W
.
  I
= "B
 

# 'B
& % 

=*    B
% " F  B

   
+

% " B
0 ) 

&" 

. ", W
G
7- 

E
- I
#  

=L " W

  
 


&" 

& 

 " 

. ", 
'   

K
 
  - )



= '"

 .
          !       0$   ! 

"& "&
!$BB
44
&2#I
J-'-4

<
%"  B
8
=  B
D
&-"  ))B

:'B
D
D B

E *B
.
F  B

G

7- 

<
%" I
/, B
9  

" 3  

 B
'  B
5    

.  B
& - )1  

7"+ 

'  

+"B
0 ) 

&" 

. ",W
8
=  I
/, B
9  

" 3  

 " B
&" 

. ",W
D
&-"  ))I
:.9 + B
&" 

. ",W

:'I
5    
*  
.  B
&" 

. ",W
D
D I
'  

&.& '"W

E *B
.
F 

G

7- I
 " 

&" 

. ", 

 .
4 $  ! $  -+-0 ! 0$!  $   ! 00 

!&  !$B
B
4
&2#I
JL' -  4

6
% )) . '  B
8
0-" B

 B

< B

.B

:'B
<
%"  B




7  B
D
D B
=
7    B
.
F  B
G

7- 

.
F 

G

7- 
  

9  B
6
% )) . ' 

.
F 
) 

"* , 
&  B
6
% )) . ' 

8
0-" 
"   

& 

<  +

% ' " 

*  

.
F 
 

 
+

% , )
- )1 
6
% )) . '  B
8
0-" B

 B

< B

.B

:'

<
%" 
+  
<  

&.  
*  
G   ,, B
  
$

 

%-  ' 

G ", 
'   



7 

D
D 
'  

&.& '"

-  ' 

" ,    
, 
.
F 
 

9 " -  0 
.
)

8
0-" 

G

7- 
 

 ,' '  

6
% )) . ' 

.
F 
- )1 
6
% )) . '  B
8
0-" B
=
7    B
.
F 

G

7- 
)  

9  

4


= '"

 . "!  $ &
5(
' ( !! 6 !  !   

(&
!$B
&2#I
JL 

<
%"  B
8
=  B
<
9*B
G

7- 

<
%" I
/, B
9  

" 3  

 B
/, 

" 3  

 " B
!  0  - B
5    

.  B
& - )1  

"

'  

+"B
0 ) 

&" 

. ", W
8
=  I
/, B
9  

" 3  

 " B
5    

.  B
&" 

. ", W
<
9*I
5    

.  W
G

7- I
 " 

&" 

. ", 




















































































































7  





































































































=  



Artikel I crystallization communications

Acta Crystallographica Section F Structural Biology Crystallization and preliminary X-ray diffraction and Crystallization studies of the ( R)-selective amine transaminase from Communications Aspergillus fumigatus ISSN 1744-3091

Maren Thomsen,‡ Lilly Skalden,‡ The ( R)-selective amine transaminase from Aspergillus fumigatus was expressed Gottfried J. Palm, Matthias in Escherichia coli and purified to homogeneity. Bright yellow crystals appeared Ho¨hne, Uwe T. Bornscheuer and while storing the concentrated solution in the refrigerator and belonged to space ˚ Winfried Hinrichs* group C222 1. X-ray diffraction data were collected to 1.27 A resolution, as well as an anomalous data set to 1.84 A ˚ resolution that was suitable for S-SAD phasing. Institut fu¨r Biochemie, Universita¨t Greifswald, Felix-Hausdorff-Strasse 4, D-17489 Greifswald, Germany

‡ These authors contributed equally to this work. 1. Introduction Transaminases belong to the pyridoxal-5 0-phosphate (PLP)-dependent Correspondence e-mail: enzymes and catalyze the reversible transfer of an amino group to an [email protected] -keto acid, ketone or aldehyde (Hayashi, 1995). The PLP and the catalytic lysine side chain are the key elements in this reaction (Eliot Received 4 October 2013 & Kirsch, 2004). Accepted 11 November 2013 Transaminases are of biotechnological significance because of their ability to produce enantiopure amines from prochiral precursors. These amines are applied as ingredients or synthons in medicine, agrochemistry, pharmacy and chemistry (Merck, 2001; Deng et al. , 1995; Martens et al. , 1986; Ho ¨hne & Bornscheuer, 2009). Based on their substrate range, transaminases can be divided into -transaminases, !-transaminases and amine transaminases. Whereas the substrates of -transaminases require a carboxylate in the position, the substrates of !-transaminases have up to five extra C atoms between the terminal amino function and the carboxylate. The substrates of amine transaminases can lack the carboxyl group completely (Ho ¨hne & Bornscheuer, 2012; Mani Tripathi & Rama- chandran, 2006). Amine transaminases often show excellent enantio- selectivity and can be grouped into two classes. ( R)-Amines are generated by ( R)-selective amine transaminases when the quinoid intermediate of the reaction is protonated from the catalytic lysine at the si -site (Hanson, 1966). Alternatively, the ( S)-amine is produced by an ( S)-amine transaminase when the protonation occurs at the re -site. Aspergillus fumigatus is a mildew which can cause respiratory allergy. It is a thermophilic saprophytic fungus with a worldwide distribution (Latge´, 1999). The sequence of an ( R)-selective amine transaminase from A. fumigatus was identified by an in silico search (Ho ¨hne et al. , 2010) and is available online at NCBI (NCBI Reference Sequence XP_748821.1). Several structures of -transaminases have been described and these enzymes have been studied in detail (Schwarzenbacher et al. , 2004; Han et al. , 2006). Recently, a few crystal structures of non- homologous ( S)-selective amine transaminases have been published (Steffen-Munsberg et al. , 2013; Humble et al. , 2012). In contrast, only a homology model of an ( R)-selective amine transaminase from an Arthrobacter species based on a d-amino-acid aminotransferase (PDB entry 3daa) has been published (Savile et al. , 2010). Here, we describe expression, purification, crystallization and initial crystallo- # 2013 International Union of Crystallography graphic results to elucidate the structure of the ( R)-selective amine All rights reserved transaminase (AspFum) from A. fumigatus .

Acta Cryst. (2013). F 69 , 1415–1417 doi:10.1107/S1744309113030923 1415 crystallization communications

2. Materials and methods Table 1 Data-collection and processing statistics. 2.1. Protein expression and purification Values in parentheses are for the outermost resolution shell. The gene for the amine transaminase was expressed in Escherichia Data set Native Anomalous coli BL21 (DE3) cells containing the expression vector pET-22b, which encodes the sequence of the amine transaminase including an Beamline 14.1, BESSY II 14.1, BESSY II additional C-terminal His tag (SGSHHHHHH; Ho ¨hne et al. , 2010). Detector Pilatus 6M Pilatus 6M 6 Wavelength (A ˚ ) 0.91841 1.77122 The recombinant protein consists of 332 amino-acid residues with a Temperature (K) 100 100 molecular weight of 37.16 kDa. The cells were grown at 310 K in Orthorhombic space group C222 1 C222 1 À1 Unit-cell parameters (A ˚ ) a = 102.2, b = 120.9, a = 102.2, b = 120.9, 400 ml LB medium containing 0.1 mg ml ampicillin until an OD 600 c = 135.4 c = 135.4 of 0.4 was reached. The temperature was then reduced to the Resolution range (A ˚ ) 50.0–1.27 (1.35–1.27) 50.0–1.84 (1.95–1.84) expression temperature of 293 K and the cells were further incubated No. of unique reflections 426722 (68273) 135117 (17260) Multiplicity 3.38 (3.3) 5.6 (2.6) until they reached an OD of 0.7. Expression of the protein was 600 Rmerge (%) 6.3 (60.3) 3.9 (9.5) induced by the addition of 1 m M IPTG. The cells were harvested 20 h Mean I/(I) 13.2 (2.0) 30.29 (8.34) after induction (Ho ¨hne et al. , 2010). CC 1/2† (%) 99.9 (73.0) 99.9 (98.8) Completeness (%) 99.1 (97.9) 95.9 (75.8) The cell pellet was resuspended in 50 m M sodium phosphate buffer Overall B factor from Wilson plot (A ˚ 2) 17.4 18.8 pH 7.5, 300 m M sodium chloride (buffer A) containing an additional Total rotation, increment ( ) 180, 0.1 360, 0.1 0.1 m M PLP and 30 m M imidazole. Cell disruption was performed † CC 1/2 is the percentage correlation between intensities from random half data sets by two passages through a French press at 10.3 MPa. The resulting (Karplus & Diederichs, 2012). suspension was centrifuged for 45 min at 10 000 g. The filtrated supernatant was applied onto a nickel–NTA column (GE Health- The crystals could easily be seen by eye (>1 mm) and had a very care). After washing with three column volumes of buffer A bright yellow colour, suggesting bound PLP (Fig. 1). Only the small containing 60 m M imidazole at a flow rate of 5 ml min À1, the protein crystals (<0.4 mm) present at the bottom of the tube diffracted to was eluted with buffer A containing 300 m M imidazole. The amine high resolution. The mechanical stress on the large crystals attached transaminase-containing fractions were identified using an aceto- to the wall of the reaction tube while fishing and upon cooling led to phenone assay (Scha¨tzle et al. , 2009), collected and pooled. The loss of diffraction quality. pooled protein was then desalted by gel chromatography against 20 m M tricine buffer pH 7.5, 10 mM PLP at a flow rate of 2 ml min À1 (Ho ¨hne et al. , 2010). The desired concentration of AspFum was 2.3. Data collection and X-ray crystallographic analysis achieved by ultrafiltration with Vivaspin 6 columns (molecular-weight For cryoprotection, a solution consisting of 35%( v/v) glycerol, cutoff 10 kDa; Sartorius Stedim). 20 m M tricine pH 7.5, 10 mM PLP was used. X-ray diffraction data were collected at 100 K on beamline 14.1 at the BESSY II synchro- 2.2. Crystallization tron source, Berlin, Germany (Mueller et al. , 2012). Two data sets were collected from one crystal. The first was collected at a wave- Initial crystallization hits were obtained with a variety of PEG- length of 0.9184 A ˚ using the highest intensity and the second was based conditions (JBScreen Classic 1–10, Jena Bioscience) within 4 d. collected at 1.77 A ˚ to obtain a large anomalous signal from the S However, all diffraction images of these crystals were not indexable. atoms present in the protein. The resolution range of the anomalous Suitable crystals of AspFum appeared after six months in an data set was limited by the detector geometry. All diffraction images Eppendorf reaction tube containing concentrated protein were processed with XDS (Kabsch, 2010) using the graphical user À1 M mM (10.7 mg ml ) and 20 m tricine pH 7.5 with 10 PLP at 277 K. interface XDSapp (Krug et al. , 2012). The rotation function was calculated using MOLREP (Vagin & Teplyakov, 2010; Winn et al. , 2011) with a resolution range of 30–3 A ˚ and a radius of integration of 30 A ˚ . Data-collection and processing statistics are given in Table 1.

3. Results and discussion The ( R)-selective amine transaminase from A. fumigatus was successful expressed, purified and crystallized and X-ray diffraction data collection was performed. The calculation of the Matthews ˚ 3 À1 coefficient VM (Matthews, 1968) as 2.9 A Da with a corresponding solvent content of 58% for two monomers offers the most probable solution. The self-rotation function (Fig. 2) shows an independent noncrystallographic twofold axis. Based on the self-rotation function and the Matthews coefficient, we deduced the presence of a dimer in the asymmetric unit. The structure could be solved directly at the beamline using the SAS protocol of the automated crystal structure- determination platform Auto-Rickshaw (Panjikar et al. , 2005), which incorporates SHELXC (Sheldrick, 2001), SHELXD (Schneider & Sheldrick, 2002), ABS (Hao, 2004), SHELXE (Sheldrick, 2002) and Figure 1 Crystals of the ( R)-selective amine transaminase from A. fumigatus grown in an DM (Cowtan, 1994). Automatic tracing using ARP /wARP (Perrakis Eppendorf reaction tube. et al. , 1999) yielded 97% of the polypeptide model and indeed shows

1416 Thomsen et al.  (R)-Selective amine transaminase Acta Cryst. (2013). F 69 , 1415–1417 crystallization communications

References Cowtan, K. (1994). Jnt CCP4/ESF–EACBM Newsl. Protein Crystallogr. 31 , 34–38. Deng, L., Mikusova´, K., Robuck, K. G., Scherman, M., Brennan, P. J. & McNeil, M. R. (1995). Antimicrob. Agents Chemother. 39 , 694–701. Eliot, A. C. & Kirsch, J. F. (2004). Annu. Rev. Biochem. 73 , 383–415. Han, Q., Robinson, H., Gao, Y. G., Vogelaar, N., Wilson, S. R., Rizzi, M. & Li, J. (2006). J. Biol. Chem. 281 , 37175–37182. Hanson, K. R. (1966). J. Am. Chem. Soc. 88 , 2731–2742. Hao, Q. (2004). J. Appl. Cryst. 37 , 498–499. Hayashi, H. (1995). J. Biochem. 118 , 463–473. Ho ¨hne, M. & Bornscheuer, U. T. (2009). ChemCatChem , 1, 42–51. Ho ¨hne, M. & Bornscheuer, U. T. (2012). Enzymes in Organic Synthesis , edited by W. Drauz, H. Gro ¨ger & O. May, pp. 779–820. Weinheim: Wiley-VCH. Ho ¨hne, M., Scha¨tzle, S., Jochens, H., Robins, K. & Bornscheuer, U. T. (2010). Nature Chem. Biol. 6, 807–813. Humble, M. S., Cassimjee, K. E., Ha˚kansson, M., Kimbung, Y. R., Walse, B., Abedi, V., Federsel, H. J., Berglund, P. & Logan, D. T. (2012). FEBS J. 279 , 779–792. Kabsch, W. (2010). Acta Cryst. D66 , 125–132. Karplus, P. A. & Diederichs, K. (2012). Science , 336 , 1030–1033. Krug, M., Weiss, M. S., Heinemann, U. & Mueller, U. (2012). J. Appl. Cryst. 45 , 568–572. Latge´, J.-P. (1999). Clin. Microbiol. Rev. 12 , 310–350. Mani Tripathi, S. & Ramachandran, R. (2006). J. Mol. Biol. 362 , 877–886. Martens, J., Gu ¨nther, K. & Schickedanz, M. (1986). Arch. Pharm. 319 , 461–465. Matthews, B. W. (1968). J. Mol. Biol. 33 , 491–497. Merck (2001). The Merck Index: An Encyclopaedia of Chemicals, Drugs and Figure 2  Biologicals. Whitehouse Station: Merck Manuals. The self-rotation function at = 180 for the diffraction data of ( R)-selective amine Mueller, U., Darowski, N., Fuchs, M. R., Fo ¨rster, R., Hellmig, M., Paithankar, transaminase from A. fumigatus in space group C222 1 reveals one independent K. S., Pu ¨hringer, S., Steffien, M., Zocher, G. & Weiss, M. S. (2012). J. twofold axis with noncrystallographic symmetry. In the orthorhombic space group Synchrotron Rad. 19 , 442–449. the dyad-related monomers and their rotational symmetry mates display 16 Panjikar, S., Parthasarathy, V., Lamzin, V. S., Weiss, M. S. & Tucker, P. A. noncrystallographic relationships including eight twofold axes. In the packing (2005). Acta Cryst. D61 , 449–457. arrangement these axes coincide pairwise, causing four peaks (60% of the origin) in 6 the self-rotation function at = 180  with ! = 15 or 75  and ’ = 0 or 180 . Perrakis, A., Morris, R. & Lamzin, V. S. (1999). Nature Struct. Biol. , 458– 463. Savile, C. K., Janey, J. M., Mundorff, E. C., Moore, J. C., Tam, S., Jarvis, W. R., a dimer in the asymmetric unit. Currently, manual completion of the Colbeck, J. C., Krebber, A., Fleitz, F. J., Brands, J., Devine, P. N., Huisman, model and refinement against the high-resolution data is in progress. G. W. & Hughes, G. J. (2010). Science , 329 , 305–309. Similarly to this amine transaminase, we have crystallized another Scha¨tzle, S., Ho ¨hne, M., Redestad, E., Robins, K. & Bornscheuer, U. T. (2009). Anal. Chem. 81 , 8244–8248. (R)-selective amine transaminase from Neosartorya fischeri (96% Schneider, T. R. & Sheldrick, G. M. (2002). Acta Cryst. D58 , 1772–1779. sequence identity) from a concentrated protein solution without Schwarzenbacher, R. et al. (2004). Proteins , 55 , 759–763. adding a specific precipitant. Sheldrick, G. M. (2002). Z. Kristallogr. 217 , 644–650. Sheldrick, G. M., Hauptman, H. A., Weeks, C. M., Miller, R. & Uso ´n, I. (2001). MT thanks the Landesgraduiertenkolleg Mecklenburg-Vorpommern International Tables for Crystallography , Vol. F, edited by M. G. Rossmann for financial support. We thank the European Union (KBBE-2011-5, & E. Arnold, pp. 333–351. Dordrecht: Kluwer Academic Publishers. grant No. 289350) for financial support within the European Union Steffen-Munsberg, F., Vickers, C., Thontowi, A., Scha¨tzle, S., Tumlirsch, T., Humble, M. S., Land, H., Berglund, P., Bornscheuer, U. T. & Ho ¨hne, M. Seventh Framework Programme. Diffraction data were collected on (2013). ChemCatChem , 5, 150–153. BL14.1 operated by the Helmholtz-Zentrum Berlin (HZB) at the Vagin, A. & Teplyakov, A. (2010). Acta Cryst. D66 , 22–25. BESSY II electron-storage ring (Berlin-Adlershof, Germany). Winn, M. D. et al. (2011). Acta Cryst. D67 , 235–242.

Acta Cryst. (2013). F 69 , 1415–1417 Thomsen et al.  (R)-Selective amine transaminase 1417

Artikel II research papers

Acta Crystallographica Section D Biological Crystallographic characterization of the Crystallography (R)-selective amine transaminase from Aspergillus ISSN 1399-0047 fumigatus

Maren Thomsen,‡ Lilly Skalden,‡ The importance of amine transaminases for producing Received 19 December 2013 Gottfried J. Palm, Matthias optically pure chiral precursors for pharmaceuticals and Accepted 15 January 2014 Ho¨hne, Uwe T. Bornscheuer and chemicals has substantially increased in recent years. The Winfried Hinrichs* X-ray crystal structure of the ( R)-selective amine transami- PDB reference: (R)-selective nase from the fungus Aspergillus fumigatus was solved by amine transaminase, 4chi S-SAD phasing to 1.84 A ˚ resolution. The refined structure at Institute of Biochemistry, University of 1.27 A ˚ resolution provides detailed knowledge about the Greifswald, Felix-Hausdorff-Strasse 4, molecular basis of substrate recognition and conversion to 17489 Greifswald, Germany facilitate protein-engineering approaches. The protein forms a homodimer and belongs to fold class IV of the pyridoxal- ‡ These authors contributed equally to this 50-phosphate-dependent enzymes. Both subunits contribute work. residues to form two active sites. The structure of the holoenzyme shows the catalytically important cofactor 0 Correspondence e-mail: pyridoxal-5 -phosphate bound as an internal aldimine with [email protected] the catalytically responsible amino-acid residue Lys179, as well as in its free form. A long N-terminal helix is an important feature for the stability of this fungal ( R)-selective amine transaminase, but is missing in branched-chain amino-acid aminotransferases and d-amino-acid aminotransferases.

1. Introduction During the last decade, interest in transaminases has increased strongly (Koszelewski et al. , 2010; Kroutil et al. , 2013; Malik et al. , 2012; Mathew & Yun, 2012; Rudat et al. , 2012; Tufvesson et al. , 2011). Many new transaminases have been discovered and applied in organic syntheses to obtain optically pure amines and non-natural amino acids for chemical and pharmaceutical applications (Ho ¨hne & Bornscheuer, 2012). This includes oxazolone derivatives used for the treatment of diabetes (Sutin et al. , 2007), rivastigmine serving in the treatment of Alzheimer’s disease (Fuchs et al. , 2010; Ro ¨sler et al. , 1999), a protected kedarcidine aglycon useful as an antitumour anti- biotic (Ogawa et al. , 2009), mexiletine for the treatment of cardiac arrhythmia (Koszelewski, Clay et al. , 2009; Kosze- lewski, Pressnitz et al. , 2009) and imagabalin, which has been suggested for the treatment of generalized anxiety disorder (Midelfort et al. , 2013). Transaminases belong to the pyridoxal-5 0-phosphate (PLP)- dependent enzymes. Besides transamination, the cofactor PLP facilitates a broad variety of other enzymatic reactions such as racemization, decarboxylation and elimination, where it serves as an electron sink to stabilize carbanion intermediates (Christen & Mehta, 2001). The reaction catalyzed by trans- aminases is the reversible conversion of -keto acids, ketones and aldehydes to the corresponding amino acids or amines (Hayashi, 1995). The catalysis itself is divided into two half- # 2014 International Union of Crystallography reactions. During the first half-reaction the amino group of a

1086 doi:10.1107/S1399004714001084 Acta Cryst. (2014). D 70 , 1086–1093 research papers suitable amino donor is transferred to PLP to yield pyridox- Table 1 amine-5 0-phosphate (PMP) with the simultaneous release of Data-collection and processing statistics. the co-product, the deaminated donor. In the second half- Values in parentheses are for the outermost resolution shell. reaction the amino acceptor is converted to the corresponding Data set Native Anomalous amine and PLP is thus regenerated (Eliot & Kirsch, 2004; Jansonius, 1998). Transaminases can be used in the kinetic Beamline BL14.1, BESSY II BL14.1, BESSY II Detector Pilatus 6M Pilatus 6M resolution of racemic amines and amino acids with a maximum Wavelength (A ˚ ) 0.91841 1.77122 yield of one enantiomer or in asymmetric synthesis starting Temperature (K) 100 100 from prochiral ketones to yield the corresponding optically Space group C222 1 C222 1 Unit-cell parameters (A ˚ ) a = 102.2, b = 120.9, a = 102.2, b = 120.9, pure amine at up to 100% yield, if a suitable method to shift c = 135.4 c = 135.4 the equilibrium to amine formation is employed. In particular, Resolution range (A ˚ ) 50.0–1.27 (1.35–1.27) 50.0–1.84 (1.95–1.84) the latter method makes them very useful in the production of No. of unique reflections 426722 (68273) 135117 (17260) Multiplicity 3.38 (3.3) 5.6 (2.6) building blocks for pharmaceuticals (Martens & Schickedanz, Rmerge† (%) 6.3 (60.3) 3.9 (9.5) 1986; Blaser, 2002). Mean I/(I) 13.2 (2.0) 30.29 (8.34) Transaminases can be divided into -transaminases, CC 1/2‡ (%) 99.9 (73.0) 99.9 (98.8) Completeness (%) 99.1 (97.9) 95.9 (75.8) !-transaminases and amine transaminases based on their Overall B factor from 17.4 18.8 substrate scope. Whereas the substrates of -transaminases Wilson plot (A ˚ 2)  require a carboxylate group in the -position to the carbonyl Total rotation/increment ( ) 180/0.1 360/0.1 function, !-transaminases also accept substrates with several † Rmerge = Phkl Pi jIiðhkl Þ À h Iðhkl Þij =Phkl Pi Iiðhkl Þ, where Ii(hkl ) is the observed C atoms (Schrewe et al. , 2013) between the carbonyl and the intensity and hI(hkl )i is the average intensity of multiple measurements. ‡ CC 1/2 is the percentage correlation between intensities from random half data sets (Karplus & carboxylic acid function and, typically, the ketone or aldehyde Diederichs, 2012). function is at the (sub-)terminal C atom of the substrate. Amine transaminases convert ketones to amines and do not 2006) and also of a few ( S)-selective amine transaminases have require a carboxylate group in the substrate (Ho ¨hne & been published and investigated (Humble et al. , 2012; Sayer et Bornscheuer, 2012). al. , 2013; Steffen-Munsberg et al. , 2013), but a structural Seven fold classes of PLP-dependent enzymes are currently analysis of an ( R)-selective amine transaminase has not been known, and transaminases have been identified in classes I and published to date. Presently, a homology model of an amine IV (Eliot & Kirsch, 2004; Jansonius, 1998). All of the members transaminase from Arthrobacter sp. is the only existing of these fold classes share the characteristic that the smallest toehold (Savile et al. , 2010). catalytic unit is a homodimer (Eliot & Kirsch, 2004). The In this paper, we present the crystal structure analysis of the monomer can be divided into a large and a small domain. The (R)-selective amine transaminase from the fungus Aspergillus two active sites lie at the interface between the domains, and fumigatus . amino-acid residues of each monomer contribute to the catalytic centre. The active sites of fold classes I and IV can be regarded as mirror images. Whereas ( S)-selective amine transaminases occur in fold class I, ( R)-selective amine 2. Materials and methods transaminases belong to fold class IV (Jansonius, 1998; Eliot & 2.1. Expression and purification Kirsch, 2004). This assignment also matches observations during protonation in the catalytic mechanism. In ( R)-selec- The expression, purification and crystallization of the tive amine transaminases the si -site (Hanson, 1966) of the (R)-selective amine transaminase from A. fumigatus were generated quinoid intermediate is solvent-facing, whereas in performed as reported previously (Thomsen et al. , 2013). the ( S)-selective amine transaminases it is the re -site. To enable the production of enantiopure compounds, amine transaminases with both enantiopreferences are required. In 2.2. Crystallization and diffraction data collection 2010, Ho ¨hne and coworkers discovered 17 ( R)-selective amine For cryoprotection, a solution consisting of 35% glycerol, transaminases using an in silico search (Ho ¨hne et al. , 2010). 20 m M tricine pH 7.5, 10 mM PLP was used. X-ray diffraction To find these putative ( R)-selective amine transaminases data were collected at 100 K on beamline 14.1 at the BESSY II sequences, the in silico search was based on the determination synchrotron, Berlin, Germany. Two data sets were collected of specific sequence motifs which characterize either d-amino- from one crystal. The first data set at a wavelength of 0.918 A ˚ acid aminotransferases ( d-ATAs) or branched-chain amino- was obtained using the highest intensity of the storage ring acid aminotransferases (BCATs) to filter out motifs for ( R)- and the second was collected at a wavelength of 1.771 A ˚ to selective amine transaminases. Based on these criteria, the obtain the highest anomalous signal of the S atoms present sequences of BCATs and d-ATAs could be excluded and the in the protein. The resolution of the anomalous data set was remaining sequences (approximately 0.4% of all investigated limited by the detector size. All diffraction images were sequences) were experimentally confirmed to be ( R)-selective processed with XDS (Kabsch, 2010) using the graphical user amine transaminases (Ho ¨hne et al. , 2010). Structures of interface XDSapp (Krug et al. , 2012). Data-collection and -transaminases (Schwarzenbacher et al. , 2004; Han et al. , processing statistics are summarized in Table 1.

Acta Cryst. (2014). D 70 , 1086–1093 Thomsen et al.  (R)-selective amine transaminase 1087 research papers

2.3. Structure determination of the holoenzyme by S-SAD Table 2 phasing Summary of SAD phasing. The crystal structure of the ( R)-selective amine transami- SHELXD CC (all) 27.11 nase from A. fumigatus was determined by single-wavelength CC (weak) 16.93 anomalous dispersion sulfur (S-SAD) phasing using the PATFOM 3.68 ‘native crystals SAS’ protocol of the automated crystal struc- SHELXE CC between Eobs and Ecalc 20.47 ture determination platform Auto-Rickshaw (Panjikar et al. , CC for partial structure against native data 47.86 2005). SAD was preferred over MAD because higher CC (all/ FOM 0.725 weak) parameters were obtained in SHELXD . The automated MapCC 0.901 No. of residues built by ARP /wARP 626 SAS protocol incorporates SHELXC (Sheldrick et al. , 2001) for data preparation as well as SHELXD (Schneider & Sheldrick, 2002) to find heavy-atom positions. With two Table 3 Refinement statistics. cysteine and ten methionine residues per monomer, we sear- ched for 24 S-atom positions per asymmetric unit. The reso- Resolution (A ˚ ) 50.0–1.27 Working/test reflections 208810/10991 lution limit for substructure determination and initial phasing R/Rfree† (%) 10.3/12.7 was set to 2.5 A ˚ . The best solution obtained resulted in CC No. of protein residues 639 (all/weak) of 27.11/16.93 and a PATFOM of 3.68. The program No. of water/glycerol molecules 994/2 No. of ions (Cl À/K +/Na +) 4/4/2 ABS (Hao, 2004) determined the correct hand of the R.m.s.d. from ideality substructure, which was subsequently used by SHELXE Bond lengths (A ˚ ) 0.014 (Sheldrick, 2002) for initial phasing. SAD phasing statistics are Bond angles ( ) 1.842 Average B factors (A ˚ 2) listed in Table 2. Density modification was performed with Protein (5910 atoms) 14.7 DM (Cowtan, 1994). Automatic tracing using ARP /wARP Water (1031 atoms) 32.1 (Perrakis et al. , 1999) yielded 97% of the polypeptide model Others (96 atoms) 15.8 ˚ Ramachandran statistics‡ (%) at 1.84 A resolution. Manual completion of the model was Most favored regions 97.64 carried out with Coot (Emsley & Cowtan, 2004). Final Outliers 0 refinement with anisotropic B factors was carried out with PDB code 4chi ˚ data extending to 1.27 A resolution using REFMAC 5 † R = jF j À j F j = jF j, where F and F are the observed and Phkl obs calc Phkl obs obs calc (Murshudov et al. , 2011). The quality of the refined protein calculated structure factors, respectively. Rfree is analogous to the R factor for 5% of the diffraction data excluded from refinement. ‡ Categories were defined by model was validated using MolProbity (Chen et al. , 2010). MolProbity . Refinement statistics are listed in Table 3. All molecular graphics were prepared using PyMOL (Delano, 2002). model. The enantiomeric PLP adducts of ( R)- and ( S)- - methylbenzylamine [( R)- - and ( S)- -MBA] were generated 2.4. Docking studies in YASARA and energy minimization was performed to the lowest energy conformation. The completely flexible ligands The docking studies were performed with YASARA were then alternatively docked into the active site. The chosen (Krieger et al. , 2002) with default parameters using the poly- simulation cell was defined to be 18 Â 17 Â 18 A ˚ around the peptide chains of the homodimer of our crystallographic catalytic residue Lys179. All residues of the active site and the active-site loop were included. H atoms were added in riding positions. The correct solution of the docking analysis was distinguished by the orientation of the cofactor PLP. The comparison of the docked enantiomeric PLP adducts with the PLP in the solved crystal structure led to the final assignment of the correct enantiomer.

3. Results 3.1. Structure analysis The phasing contributions of the chloride ions and the S atoms of Cys and Met are shown in Fig. 1. Interestingly, the highest occupancy is observed for two chloride ions, but not for the other possible elements (S, P or K), even taking into account that these atoms show clear signals in the final Figure 1 anomalous electron-density map. The initial phasing based on Occupancies of the heavy-atom sites found by SHELXD (Schneider & ˚ Sheldrick, 2002). Meaningful heavy atoms are labelled by locating their the anomalous diffraction at 1.84 A resolution was sufficient positions in the refined model. for automatic tracing. Refinement using the high-resolution

1088 Thomsen et al.  (R)-selective amine transaminase Acta Cryst. (2014). D 70 , 1086–1093 research papers

data converged to an R and Rfree of 10.3 and 12.7%, respec- those of BCATs and d-ATAs, with the best fit to the BCAT tively. from Thermus thermophilus [PDB entry 1wrv; root-mean- The final model contains 639 amino-acid residues of two square difference on C atoms (r.m.s.d.) of 1.8 A ˚ , fitting 297 polypeptide chains ( A and B), two PLP molecules, four residues; RIKEN Structural Genomics/Proteomics Initiative, potassium ions, four chloride ions, two glycerol molecules and unpublished work] and the d-ATA from Bacillus sp. YM-1 994 water molecules. Ions are assigned according to electron density and meaningful chemical terms and refinement conditions. Both polypeptide chains are well defined by the electron-density maps (Supplementary Fig. S2 1) and the final model is consistent with the anomalous map (Supplementary Fig. S3). Some residues with poor electron density at the N- and C-termini (monomer A, Met1 and Ser322–His332; monomer B, Met1–Ala2 and Ser322–His332) were excluded from the structural model. The cofactor PLP was modelled into the active site using Fobs À Fcalc difference maps and refined with a summed occupancy of 0.8. The occupancies of the two PLP states were assigned so that the B factors were consistent with those of neighbouring residues. The remaining occupancy of 0.2 was filled with a single phosphate in the same position as the phosphate group of the cofactor. Additional positive difference electron density was observed in the substrate-binding site within covalent bond Figure 2 distance of the cofactor PLP in each monomer (Supplemen- Overall structure of the ( R)-selective amine transaminase from tary Fig. S4). Owing to the low occupancy of the ligand, we A. fumigatus viewed normal to the molecular dyad. The subunits of the homodimer are shown in blue and red, respectively. The monomer is could not conclusively model this density. All compounds used divided into colour-coded domains: the small domain (blue, red) with the in purification and crystallization and common metabolites active-site loop (yellow) and the large domain (cyan, orange). The active- were ruled out; also, GC-MS-analysis of acid-denatured and site loop derived from the left subunit is shown in yellow (with Arg126 as a stick model). The cofactor PLP is bound to Lys179 at the domain heat-denatured enzyme did not uncover the identity of this interface (shown as a stick model in pink). ligand. Besides the tricine molecule in the buffer, no amine or carbonyl compounds were added after cell disruption. Nevertheless, d-amino acids were also tested as possible ligands. In this case, the -carboxyl group could not be modelled into the small binding pocket. Alternative conformations were modelled for 138 amino- acid side chains out of 639 residues ( 20%). Some peptide backbone O atoms could also be modelled in alternative conformations. Differences in the main-chain conformation could be detected for residues Thr204–Gly206. The final refinement and validation statistics are shown in Table 3.

3.2. Overall fold The ( R)-selective amine transaminase crystallized in space group C222 1 with two monomers in the asymmetric unit forming a homodimer (Fig. 2). Each polypeptide chain is constituted of 332 amino-acid residues with a molecular weight of 37.1 kDa. The tertiary structure of one subunit consists of the typical fold of enzymes belonging to the fold class IV of PLP-dependent enzymes, as first described for d-ATA from Bacillus sp. (Sugio et al. , 1995). The subunit divides into a small domain (N-terminus to Pro144) with an / -structure, an inter-domain loop (Tyr145–Met149) and a Figure 3 large domain (Ala150 to the C-terminus) with a pseudo-barrel Comparison of the overall monomer fold between the ( R)-selective amine transaminase from A. fumigatus (green), d-amino-acid amino- structure (Fig. 2). The enzyme belongs to fold class IV of PLP- transferase (PDB entry 3lqs; r.m.s.d. 2.0 A ˚ , cyan) and branched-chain dependent enzymes, and the overall structure is very similar to amino-acid aminotransferase (PDB entry 1wrv; r.m.s.d. 1.8 A˚ , violet) distinctly shows the unique long N-terminal helix found in the ( R)- 1 Supporting information has been deposited in the IUCr electronic archive selective amine transaminase. The cofactor PLP is shown as a stick model (Reference: DZ5319). in yellow.

Acta Cryst. (2014). D 70 , 1086–1093 Thomsen et al.  (R)-selective amine transaminase 1089 research papers

(PDB entry 3lqs; r.m.s.d. of 2.0 A ˚ , fitting 280 residues; Lepore (Fig. 4 a), whereas the other state (occupancies of 0.3 and 0.4 et al. , 2010). However, the ( R)-selective amine transaminase in monomers A and B, respectively) represents an adduct with from A. fumigatus has an additional long N-terminal -helix an unidentified ligand (see x3.1) and a free lysine (Fig. 4 b). The (Met4–Arg20) which has a significant effect on protein stabi- covalently bound PLP shows the typical distorted aldimine lity, as discussed below (Fig. 3). Other differences in the bond of PLP-dependent enzymes. The bond angles deviate backbone folding compared with BCATs and d-ATAs are only from the ideal 120  and the internal aldimine bond is out of the visible for loop regions on the surface. Residues from each plane of the pyridoxyl ring by about 11.5 and 14  in monomers domain as well as residues from the other subunit of the dimer A and B, respectively. This typical geometry has been found in participate in forming the active site. The active-site loop various crystal structures of other PLP-dependent enzymes (Gly121*–Asn135*; residues labelled with an asterisk belong and it is supposed that the release of strain on breaking to the other subunit) limits access to the active site and is the internal aldimine bond enhances the catalytic ability contributed by the other subunit. (Dubnovitsky et al. , 2005; Hayashi et al. , 1998). In the free state Lys179 has a distinct alternative conformation and is involved 3.3. Cofactor binding in a hydrogen-bonding network with Arg77 and the phosphate group of PLP via a water molecule. The pyridoxyl ring shows Well defined electron density is observed for the cofactor two distinct orientations, and the phosphate group is tightly PLP, which is located at the bottom of the active site between bound and is involved in several hydrogen-bond interactions the small and the large domains (Fig. 2). There are two distinct (with His74, Arg77, Thr273, Thr274, Ile237 and Thr238) as states observed for the cofactor. One state (occupancies of 0.5 an anchor for the cofactor. Residues Ile237 and Thr238 are and 0.4 in monomers A and B, respectively) is covalently located at the N-terminus of helix 7 (according to Sugio et al. , bound to the active-site residue Lys179 of the large domain 1995) such that the dipole moment of the helix additionally facilitates the coordination of the phosphate group. The pyridoxyl ring is sandwiched between Leu234 and the peptide bond of Gly215 to Phe216. The ion pair formed by the highly conserved Glu212 and the N atom of the pyridoxyl ring (N1) provides an electron sink during the reaction mechanism. This glutamate is in turn coordinated by the conserved Arg168. Anchoring of PLP by the phosphate and N1 coordination

Figure 5 Comparison of the docking studies of ( R)- -MBA-PLP (violet) and ( S)- -MBA-PLP (cyan) into the active site of the ( R)-selective amine transaminase from A. fumigatus . On the basis of the docking experiments Figure 4 the active site can be divided into a small binding pocket hosting the (a) Schematic drawing of the coordination of the cofactor PLP (distances methyl group and a large binding pocket which is responsible for ˚ are in A ) and ( b) 2 Fobs À Fcalc map contoured at 1 . Note the covalently coordinating the aromatic ring. Only for the ( R)-enantiomer is the bound (green C atoms) and unbound (cyan C atoms) states with Lys179. catalytic lysine residue (yellow) at a reasonable distance (2.8 A ˚ , dashed The aldehyde O atom O4A of the free PLP, or alternatively the N atom line) from the H atom (white) of -MBA-PLP to initiate deamination. N4A of PMP, could not be identified in the electron-density map and is Residues defining the small binding pocket are coloured green, while therefore omitted in all figures. residues of the large binding pocket are shown in blue.

1090 Thomsen et al.  (R)-selective amine transaminase Acta Cryst. (2014). D 70 , 1086–1093 research papers limits movement of the pyridoxyl ring to a rotation around the transferase (PDB entry 3daa; Peisach et al. , 1998) with a C5—C5A bond by about 19  (Fig. 4 b). flexible N-terminus, an N-terminal deletion of 22 amino-acid residues was introduced to improve the crystallization quality. 3.4. Deletion of the N-terminal helix Whereas the wild-type amine transaminase could be over- expressed as soluble protein, this N-terminal deletion resulted Previously, we performed crystallization studies on the ( R)- in insoluble protein. The insolubility of this amine transami- amine transaminase from Neosartorya fischeri (96% identity nase variant could not be prevented by changing the expres- to the amine transaminase from A. fumigatus ). Presently, the sion temperature, varying the inducer concentration, altering diffraction images of the obtained crystals are not indexable. the induction time or the co-expression of chaperones. d Based on a homology model built from -amino-acid amino- In the solved X-ray structure of ( R)-amine transaminase from A. fumigatus , the N-terminus forms a long helix corre- sponding to the sequence that was deleted in the ( R)-amine transaminase mutant from N. fischeri . It is obvious that the N-terminal helix is important for the soluble expression of fungal amine transaminases, but remarkable hydrophobic patches on the surface of the modelled truncated enzyme are not observed.

3.5. Structural design of the active site Via docking studies performed with the program YASARA , it could be demonstrated that the active site of the (R)-selective amine transaminase from A. fumigatus is divided into a small and a large binding pocket (Fig. 5). Docking was performed with the substrate adducts of ( R)- -methylbenzyl- amine [( R)- -MBA] and ( S)- -MBA to PLP, which starts the first deamination cycle. The pyridoxyl rings and the phosphate group of the modelled substrate adducts superposed very well with the free PLP state (r.m.s.d. of 0.14 A ˚ ) of the X-ray structure, indicating good quality of the docking results. In every docking run the methyl group was bound in the small binding pocket formed by the residues Val60, Phe113 and Ile146 (Fig. 6). The aromatic ring was coordinated in the large Figure 6 binding pocket which is built by the residues His53*, Tyr58, Active-site architecture. The residues forming the small binding pocket Arg126*, Val148 and Trp183. Although ( S)- -MBA-PLP are shown in green and the amino-acid residues responsible for forming the large binding pocket are shown in blue (residues which originate from could be docked without clashes, the enantioselectivity can the other subunit are shown in dark blue and are marked with asterisks; be explained by the orientation of ( R)- -MBA-PLP and ( S)- PLP-binding residues are coloured violet). -MBA-PLP to the catalytically active lysine residue. This residue initiates the deamination reaction by deprotonation and is only at a reasonable distance (2.8 A ˚ ) for abstraction of the proton from ( R)- -MBA-PLP which points directly towards the Lys179 N " atom. In contrast, the proton of ( S)- -MBA-PLP points in the opposite direction and cannot be abstracted by Lys179.

3.6. Active-site comparison Whereas the binding of the cofactor and the backbone are conserved, a comparison of the active-site residues responsible for substrate recognition of the ( R)-selective amine transaminases ( R-ATAs) with the Figure 7 active sites of BCATs and d-ATAs shows Stereo representation of the active-site comparison between the ( R)-selective amine clearly that no amino-acid residues other transaminase from A. fumigatus (green), d-amino acid aminotransferase (PDB entry 3lqs; cyan) and branched-chain amino-acid aminotransferase (PDB entry 1wrv; violet). The inter- than the catalytic Lys179 and Glu212 are domain loop is omitted for clarity. conserved (Fig. 7). When considering

Acta Cryst. (2014). D 70 , 1086–1093 Thomsen et al.  (R)-selective amine transaminase 1091 research papers

Figure 8 Schematic drawing of the active-site arrangement of enzymes of fold class IV in comparison to the active site of the ( R)-selective amine transaminase from A. fumigatus [( R)-ATA], the ( S)-branched-chain amino-acid aminotransferase [( S)-BCAT; PDB entry 1iye] and d-amino-acid aminotransferase [( R)-DATA; PDB entry 3daa]. Grey spheres indicate the space-filling requirements of residues Val60, Gly38 and Val33, respectively. substrates to be converted by ( R)-selective amine transami- substrate in the same pocket of the active site. This assumption nases, the substitution of the carboxyl group by a methyl group needs to be investigated further via mutagenesis. inverts the priority according to the Cahn–Ingold–Prelog rule. Hence, the active-site architecture, based on the definition of the small and large pockets, was postulated to be more similar 4. Conclusion to the BCATs than to the d-ATAs. This can now be verified by In the era of rational protein design, crystal structure or NMR the crystal structure with the docking analysis (see x3.5) as well analyses at atomic resolution are the most valuable tools for as the sequence motif Y/(F)V Z (with Z preferably being protein-engineering experiments. The crystal structure of glutamate) postulated to be an important feature in the the ( R)-selective amine transaminase elucidated here for the structural design of the active site of ( R)-selective amine enzyme from A. fumigatus provides essential information and transaminases (Ho ¨hne et al. , 2010). Whereas the large pocket insights into understanding how substrate recognition occurs of the d-ATAs is mostly built by small hydrophobic residues, in ( R)-selective amine transaminases and distinguishes them the pocket volume is reduced in BCATs and in the ( R)- from other enzymes of fold class IV. For further understanding selective amine transaminase by bulky amino-acid residues of substrate binding and enantioselectivity, soaking and co- (Fig. 8). A search of the DALI secondary-structure database crystallization experiments are in progress. (Holm & Rosenstro ¨m, 2010) revealed the inter-domain loop Note added in proof . A crystal structure analysis of the (R)- as an active-site limiting feature. The loop in question is two selective !-transaminase from Aspergillus terreus has recently amino-acid residues longer than the equivalent loop found in been published (Lyskowski et al. , 2014). d-ATAs and therefore contributes to the small binding pocket MT thanks the ‘Landesgraduiertenkolleg Mecklenburg- of the active site. Whereas the small pocket of the d-ATAs Vorpommern’ for financial support. We thank the European harbours predominantly positively charged residues to coor- Union (KBBE-2011-5, grant No. 289350) for financial support dinate the carboxylate group, in BCATs and R-ATAs aromatic within the European Union Seventh Framework Programme. side chains form a hydrophobic environment for the mostly Diffraction data were collected on BL14.1 operated by the hydrophobic substituents that are accepted. As mentioned Helmholtz-Zentrum Berlin (HZB) at the BESSY II electron- above, the entrance of the active site of the R-ATAs is limited storage ring (Berlin-Adlershof, Germany). We also thank by the active-site loop. Interestingly, similar to dual substrate Professor M. Lalk (University Greifswald, Germany) for the recognition by the ( S)-amine transaminase (Steffen-Munsberg GC-MS analysis. et al. , 2013), the active-site loop of the R-ATA also has a highly flexible Arg126 (slightly increased B factors and two alter- native conformations of residues Arg126, Gly127 and Ser128). By flipping in and out of the active site, Arg126 could facilitate the coordination of the negatively charged carboxylate of the References amino acceptor pyruvate as well as the binding of uncharged Blaser, H.-U. (2002). Adv. Synth. Catal. 344 , 17.

1092 Thomsen et al.  (R)-selective amine transaminase Acta Cryst. (2014). D 70 , 1086–1093 research papers

Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, Martens, J. G. K. & Schickedanz, M. (1986). Arch. Pharm. 319 , R. M., Kapral, G. J., Murray, L. W., Richardson, J. S. & Richardson, 461–465. D. C. (2010). Acta Cryst. D66 , 12–21. Mathew, S. & Yun, H. (2012). ACS Catal. 2, 993–1001. Christen, P. & Mehta, P. K. (2001). Chem. Rec. 1, 436–447. Midelfort, K. S., Kumar, R., Han, S., Karmilowicz, M. J., McConnell, Cowtan, K. (1994). Jnt CCP4/ESF–EACBM Newsl. Protein Crystal- K., Gehlhaar, D. K., Mistry, A., Chang, J. S., Anderson, M., logr. 31 , 34–38. Villalobos, A., Minshull, J., Govindarajan, S. & Wong, J. W. (2013). Delano, W. L. (2002). PyMOL . http://www.pymol.org. Protein Eng. Des. Sel. 26 , 25–33. Dubnovitsky, A. P., Ravelli, R. B. G., Popov, A. N. & Papageorgiou, Murshudov, G. N., Skuba´k, P., Lebedev, A. A., Pannu, N. S., Steiner, A. C. (2005). Protein Sci. 14 , 1498–1507. R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011). Eliot, A. C. & Kirsch, J. F. (2004). Annu. Rev. Biochem. 73 , 383–415. Acta Cryst. D67 , 355–367. Emsley, P. & Cowtan, K. (2004). Acta Cryst. D60 , 2126–2132. Ogawa, K., Koyama, Y., Ohashi, I., Sato, I. & Hirama, M. (2009). Fuchs, M., Koszelewski, D., Tauber, K., Kroutil, W. & Faber, K. Angew. Chem. Int. Ed. 48 , 1110–1113. (2010). Chem. Commun. 46 , 5500. Panjikar, S., Parthasarathy, V., Lamzin, V. S., Weiss, M. S. & Tucker, Han, Q., Robinson, H., Gao, Y. G., Vogelaar, N., Wilson, S. R., Rizzi, P. A. (2005). Acta Cryst. D61 , 449–457. M. & Li, J. (2006). J. Biol. Chem. 281 , 37175–37182. Peisach, D., Chipman, D. M., Van Ophem, P. W., Manning, J. M. & Hanson, K. R. (1966). J. Am. Chem. Soc. 88 , 2731–2742. Ringe, D. (1998). Biochemistry , 37 , 4958–4967. Hao, Q. (2004). J. Appl. Cryst. 37 , 498–499. Perrakis, A., Morris, R. & Lamzin, V. S. (1999). Nature Struct. Mol. Hayashi, H. (1995). J. Biochem. 118 , 463–473. Biol. 6, 5. Hayashi, H., Mizuguchi, H. & Kagamiyama, H. (1998). Biochemistry , Ro ¨sler, M., Anand, R., Cicin-Sain, A., Gauthier, S., Agid, Y., Dal- 37 , 15076–15085. Bianco, P., Sta¨helin, H. B., Hartman, R. & Gharabawi, M. (1999). Ho ¨hne, M. & Bornscheuer, U. T. (2012). Enzymes in Organic BMJ , 318 , 633–638. Synthesis , edited by O. May, H. Gro ¨ger & W. Drauz, pp. 779–820. Rudat, J., Brucher, B. R. & Syldatk, C. (2012). AMB Express , 2, 11. Weinheim: Wiley-VCH. Savile, C. K., Janey, J. M., Mundorff, E. C., Moore, J. C., Tam, S., Jarvis, Ho ¨hne, M., Scha¨tzle, S., Jochens, H., Robins, K. & Bornscheuer, U. T. W. R., Colbeck, J. C., Krebber, A., Fleitz, F. J., Brands, J., Devine, (2010). Nature Chem. Biol. 6, 807–813. P. N., Huisman, G. W. & Hughes, G. J. (2010). Science , 329 , 305– Holm, L. & Rosenstro ¨m, P. (2010). Nucleic Acids Res. 38 , W545– 309. W549. Sayer, C., Isupov, M. N., Westlake, A. & Littlechild, J. A. (2013). Acta Humble, M. S., Cassimjee, K. E., Ha˚kansson, M., Kimbung, Y. R., Cryst. D69 , 564–576. Walse, B., Abedi, V., Federsel, H. J., Berglund, P. & Logan, D. T. Schneider, T. R. & Sheldrick, G. M. (2002). Acta Cryst. D58 , 1772– (2012). FEBS J. 279 , 779–792. 1779. Jansonius, J. N. (1998). Curr. Opin. Struct. Biol. 8, 759–769. Schrewe, M., Ladkau, N., Bu ¨hler, B. & Schmid, A. (2013). Adv. Synth. Kabsch, W. (2010). Acta Cryst. D66 , 125–132. Catal. 355 , 1693–1697. Karplus, P. A. & Diederichs, K. (2012). Science , 336 , 1030–1033. Schwarzenbacher, R. et al. (2004). Proteins , 55 , 759–763. Koszelewski, D., Clay, D., Rozzell, D. & Kroutil, W. (2009). Eur. J. Sheldrick, G. M. (2002). Z. Kristallogr. 217 , 644–650. Org. Chem. 2009 , 2289–2292. Sheldrick, G. M., Hauptman, H. A., Weeks, C. M., Miller, R. & Uson, Koszelewski, D., Pressnitz, D., Clay, D. & Kroutil, W. (2009). Org. I. (2001). International Tables for Crystallography , Vol. F, edited by Lett. 11 , 4810–4812. M. G. Rossmann & E. Arnold, pp. 333–351. Dordrecht: Kluwer Koszelewski, D., Tauber, K., Faber, K. & Kroutil, W. (2010). Trends Academic Publishers. Biotechnol. 28 , 324–332. Steffen-Munsberg, F., Vickers, C., Thontowi, A., Scha¨tzle, S., Krieger, E., Koraimann, G. & Vriend, G. (2002). Proteins , 47 , Tumlirsch, T., Svedendahl Humble, M., Land, H., Berglund, P., 393–402. Bornscheuer, U. T. & Ho ¨hne, M. (2013). ChemCatChem , 5, Kroutil, W., Fischereder, E. M., Fuchs, C. S., Lechner, H., Mutti, F. G., 150–153. Pressnitz, D., Rajagopalan, A., Sattler, J. H., Simon, R. C. & Siirola, Sugio, S., Petsko, G. A., Manning, J. M., Soda, K. & Ringe, D. (1995). E. (2013). Org. Process Res. Dev. 17 , 751–759. Biochemistry , 34 , 9661–9669. Krug, M., Weiss, M. S., Heinemann, U. & Mueller, U. (2012). J. Appl. Sutin, L., Andersson, S., Bergquist, L., Castro, V. M., Danielsson, E., Cryst. 45 , 568–572. James, S., Henriksson, M., Johansson, L., Kaiser, C., Flyre´n, K. & Lepore, B. W., Liu, D., Peng, Y., Fu, M., Yasuda, C., Manning, J. M., Williams, M. (2007). Bioorg. Med. Chem. Lett. 17 , 4837–4840. Silverman, R. B. & Ringe, D. (2010). Biochemistry , 49 , 3138–3147. Thomsen, M., Skalden, L., Palm, G. J., Ho ¨hne, M., Bornscheuer, U. T. Lyskowski, A., Gruber, C., Steinkellner, G., Schurmann, M., Schwab, & Hinrichs, W. (2013). Acta Cryst. F69 , 1415–1417. H., Gruber, K. & Steiner, K. (2014). PLoS One , 9. Tufvesson, P., Lima-Ramos, J., Jensen, J. S., Al-Haque, N., Neto, W. & Malik, M. S., Park, E.-S. & Shin, J.-S. (2012). Appl. Microbiol. Woodley, J. M. (2011). Biotechnol. Bioeng. 108 , 1479– Biotechnol. 94 , 1163–1171. 1493.

Acta Cryst. (2014). D 70 , 1086–1093 Thomsen et al.  (R)-selective amine transaminase 1093
%]]CVIVJ
`7 :CCQ$`:].1H H.:`:H V`1 : 1QJ Q` .V ^_R VCVH 10V :I1JV `:J :I1J: V ``QI
]V`$1CC% `%I1$: %
:`VJ
.QI VJ 5
1CC7
@:CRVJ 5
Q ``1VR
8
:CI5
: .1:
V.JV5
1V
8
Q`J H.V%V`
:JR
1J``1VR
1J`1H.


.V V
:% .Q`
HQJ `1G% VR
V_%:CC7
Q
.1
1Q`@8

%J 1 % V
Q`
1QH.VI1 `75
J10V` 1 7
Q`
`V1` 1:CR5
'VC16R:% RQ```R `8
5
 
`V1` 1:CR5
V`I:J7

/Q``V ]QJRVJHV
VRI:1C7
11J``1VR8.1J`1H. %J1R$`V1` 1:CR8RV


CQJ1J$ Q` .V :I1JV `:J :I1J: V I% :J Q` VQ:` Q`7: `1H.V`1 Q
RVCV V

:I1JQR:H1R
`V 1R%V
:
.V
5R V`I1J%
:J
:RR1 1QJ:C

RV %
`V `1H 1QJ
1 V
1:
1J V` VR
1J Q
.V
V_%VJHV
Q`
.V
:I1JV
`:J :I1J: V
Q`

VQ:` Q`7: `1H.V`1 :` V`
.V
VHQJR
:I1JQ
:H1R
:C:J1JV8
.V
1J V` 1QJ
1:
I:RV
G7
:
6%1@/.:J$V
/7
11 .
.V
6%1@
1$. J1J$
%C 1
8J<7IV
@1
^;$1CVJ
VH.JQCQ$1V _8
.V
6%1@/.:J$V
1:
RQJV
:
RV H`1GVR
1J
.V
I:J%:C
11 .
.V
`Q`1:`R
]`1IV`
//;;/;//;;///;;;;8
.V
`QCCQ11J$
/7
]`Q$`:I
1:
% VR7
QCR
/
`Q`

I1J
GV`Q`V

H7HCV
11 .
/
`Q`

VH5
/
`Q`

VH
:JR
/
`Q`

I1J
:` VR8
'1J:CC7
/
1:
.QCR
:RR1 1QJ:C
`Q`

I1J8
;` V`
.V
6%1@/.:J$V
/7
:
]J %
R1$V 1QJ
``QI
.V
:IV
@1
`QCCQ1VR8

-C
]J %
1:
I16VR
11 .

-C
/7
`V:$VJ
:JR
1JH%G: VR
`Q`

I1J
:
 /
GV`Q`V
.V
VJ<7IV
1:
1J:H 10: VR
G7
 /
`Q`

I1J8
.V
`V %C 1J$
]C: I1R
1:
0V`1`1VR
.`Q%$.
V_%VJH1J$8
.V
`1J:C
RVCV 1QJ
Q`
.V

:I1JQ
:H1R
1:
RQJV
.`Q%$.
:
`V `1H 1QJ
R1$V 1QJ
11 .

RV %
VJ<7IV
^5V1
8J$C:JR
1QC:G _8
-C
]C: I1R
1:
I16VR
11 .

-C
58 RG%``V`
5

-C

RV %
VJ<7IV
:JR

-C
R1 8
1: V`8
.V
R1$V 1QJ
1:
I:RV
:
 /
`Q`
.
GV`Q`V
.V
VJ<7IV
1:
1J:H 10: VR
`Q`

I1J
:
 /8
;
/7
]%`1`1H: 1QJ
11 .
.V
1$.R%`V
/7
/CV:J%]
1H`Q
@1
^7QH.V_
`QCCQ1VR8
.V
]%`1`1H: 1QJ
1:
RQJV
:
RV H`1GVR
1J
.V
I:J%:C8
'1J:CC7
.V
C1$: 1QJ
Q`
.V
]C: I1R
1:
H:``1VR
Q%
11 .
.V
RC1$: V
^.V`IQ
H1VJ 1`1H_8
8
-C
]%`1`1VR
/7
]`QR%H
1:
I16VR
11 .
8
-C
RC1$: V
G%``V`5
8
-C
RC1$: V
:JR

-C
R1 8
1: V`8
.V
`QCCQ11J$
]`Q$`:I
1:
% VR7
QCR
 /
`Q`
.5
.QCR
 /
`Q`
.5
.QCR
 /
`Q`
.5
.QCR
 /
`Q`
.5
.QCR
 /
`Q`
.
:JR
.QCR
 /
`Q`

I1J8
.V
`V %C 1J$
]C: I1R
1:
0V`1`1VR
G7
V_%VJH1J$8

6]`V 1QJ Q` .V 11CR 7]V :JR .V I% :J Q` VQ:` Q`7: `1H.V`1 .V
V6]`V 1QJ
Q`
.V
:I1JV
`:J :I1J: V
I% :J
:JR
11CR
7]V
``QI

VQ:` Q`7: `1H.V`1
1:
H:``1VR
Q%
C1@V
.V
V6]`V 1QJ
Q`
.V
:I1JV
`:J :I1J: V
``QI
]V`$1CC% `%I1$: %
:
RV H`1GVR
^V.JV V :C8 5
_8
Q
Q] 1I1

1V`V
:C Q
V VR
11 .
:J
1JR%H 1QJ
CD 
Q`
8
:JR
8
;C Q
R1``V`VJ
1JR%HV`
HQJHVJ `: 1QJ
Q`
8
I
:JR
8
I
%
1V`V
1J0V 1$: VR8
.V
V6]`V 1QJ
Q] 1I1<: 1QJ
1:
H:``1VR
Q%
11 .
H.:]V`QJV
]`V VJ
1J
.V
:@:`:
H.:]V`QJV
@1
^:@:`:
1Q
%5/8_8
.V
]C: I1R
H:``71J$
.V
H.:]V`QJV
1V`V
HQR `:J `Q`IVR
11 .
.V
:I1JV
`:J :I1J: V
$VJV
H:``71J$
]C: I1R8
D%`1J$
:CC

H%C 10: 1QJ
:JR
V6]`V 1QJ
LCD 
:I]CV
1V`V
:@VJ8

`:J `Q`I: 1QJ :JR HQR `:J `Q`I: 1QJ Q` H.:]V`QJV /QI]V VJ
8 HQC1
C
HVCC
1V`V
1JH%G: VR
`Q`

I1J
QJ
1HV
11 .

-C
Q`
.V
]C: I1R8
;` V`1:`R
:
.V:
.QH@
:
 /
`Q`

VH
`QCCQ1VR8
;` V`

I1J
1JH%G: 1QJ
QJ
1HV5
 
-C
 RQH
1:
:RRVR
:JR
1JH%G: VR
:
 /
`Q`
.8
'1J:CC75
 
-C
H%C %`V
1:
]C: VR
QJ
:$:`R]C: V
11 .
]VH1`1H
:J 1G1Q 1H
`V 1 :JHV
^:I]1H1CC1J
`Q`
11CR
7]V
:JR
I% :J
Q`
.V
:I1JV
`:J :I1J: V
Q`

VQ:` Q`7: `1H.V`1 _8
.V
HQR `:J `Q`I: 1QJ
1:
RQJV
11 .
1J
V:H.
H: V

-C
Q`
.V
]C: I1R
H:``71J$
.V
H.:]V`QJV
:JR
.V
:I1JV
`:J :I1J: V
VJHQR1J$
$VJV 8
.V
`V %C 1J$
H%C %`V
1:
]C: VR
QJ
:$:`R]C: V
11 .
:I]1H1CC1J
^
-$LIC_
:JR
H.CQ`:I].VJ1HQC
^ 
-$LIC_8



R !" R .V
LCD 
:I]CV
1V`V
R1 `%] VR
1Q
1IV
G7
':
`V]
11 .

I
`Q`

VH8
;` V`
HVJ `1`%$: 1QJ
:

W
$
.V
%]V`J: :J
1:
V]:`: VR
:JR
.V
1J QC%GCV
``:H 1QJ
1:
1: .VR
11 .
QR1%I
].Q ].: V
G%``V`
^ 
I5
]
8 5
:R=% VR
11 .
HC_8
'1J:CC75
.V
1J QC%GCV
``:H 1QJ
1:
`VR % ]VJRVR
1J
QR1%I
].Q ].: V
G%``V`
^ 
I5
]
8 5
:R=% VR
11 .
HC_8

DR;8
11 .
:
Q
V]:`: 1QJ
$VC
:JR
Q
HQCCVH 1J$
$VC
1:
% VR8

-C
JQ`I:C1

:_



G_
#1$%`V
7
DR;8
Q`
.V
:I1JV
`:J :I1J: V
I% :J
``QI

VQ:` Q`7: `1H.V`1 8
:JV
:`V
C:GVCCVR
:HHQ`R1J$
Q
.V
1JR%H 1QJ
1IV

^.Q%`_8
:_
QC%GCV
``:H 1QJ
G_
1J QC%GCV
``:H 1QJ8

#1$%`V
7
 QG T  H:CH
I:]
Q`
:
`V]`V VJ : 10V
`V$1QJ
Q`
.V
VJ<7IV
HVJ `VR
:
.V
 
H:CH%C: VR
:
8
K
`V QC% 1QJ
HQJ Q%`VR
:
5
CV0VC8
.V
`1$%`V
1:
]`V]:`VR
11 .
7
^DVC:JQ5
_ .

#1$%`V
7
;JQI:CQ%
I:]
HQJ Q%`VR
:
58
%C].%`
: QI
Q`
V  5
V  5
/7  5
V  5
:JR
V 
:`V
1RVJ 1`1VR8
.V
`1$%`V
1:
]`V]:`VR
11 .
7
^DVC:JQ5
_ .

#1$%`V
7
 V`VQ
01V1
Q`
:RR1 1QJ:C
]Q 1 10V
R1``V`VJHV
VCVH `QJ
RVJ 1 7
^ QG
T
H:CH
I:]5
5
CV0VC_
1J
.V
:H 10V
1 V
JV6
Q
.V
8
.V
`1$%`V
1:
]`V]:`VR
11 .
7
^DVC:JQ5
_8

#1$%`V
7
;JQI:CQ%
I:]
HQJ Q%`VR
:
5
:`Q%JR
%CV
:JR
/7 5
1QJ
:`V
.Q1J
:
].V`V
^]Q : 1%I7
01QCV 6
H.CQ`1RV7
$`VVJ_8
.V
`1$%`V
1:
]`V]:`VR
11 .
7
^DVC:JQ5
_8

References

Delano, W. L. (2002). http://www.pymol.org . Höhne, M., Schätzle, S., Jochens, H., Robins, K. & Bornscheuer, U. T. (2010). Nat. Chem. Biol. 6, 807-813.

Artikel III Structural and biochemical characterization of the dual substrate recognition of the ( R)-selective amine transaminase from Aspergillus fumigatus Lilly Skalden, Maren Thomsen, Matthias H ohne,€ Uwe T. Bornscheuer and Winfried Hinrichs

Institut fur€ Biochemie, Universit at€ Greifswald, Germany

Keywords Chiral amines are important precursors for the pharmaceutical and fine- amine transaminase; dual substrate chemical industries. Because of this, the demand for enantiopure amines is 0 recognition; gabaculine; pyridoxal-5 - currently increasing. Amine transaminases can produce a large spectrum of phosphate; X-ray structure chiral amines in the ( R)- or ( S)-configuration, depending on their substrate Correspondence scope and stereo-preference, by converting a prochiral ketone into the chi- W. Hinrichs, Institut f ur€ Biochemie, ral amine while using alanine as the amine donor producing pyruvate as an Universitat€ Greifswald, Felix-Hausdorff- a-keto acid product. In order to guide the protein engineering of transam- Straße 4, 17489 Greifswald, Germany inases to improve substrate specificity and enantioselectivity, we carried out Fax: +49 0 3834 864373 a crystal structure analysis at 1.6 A resolution of the ( R)-amine transami- Tel.: +49 0 3834 864356 nase from Aspergillus fumigatus with the bound inhibitor gabaculine. This E-mail: [email protected] revealed that Arg126 has an important role in the dual substrate recogni- U. T. Bornscheuer, Institut f ur€ Biochemie, Universitat€ Greifswald, Felix-Hausdorff- tion of this enzyme because mutating this residue to alanine reduced sub- Straße 4, 17489 Greifswald, Germany stantially the ability of the enzyme to use pyruvate as an amino acceptor. Fax: +49 0 3834 864373 Tel.: +49 0 3834 864367 Database E-mail: [email protected] Coordinates and structure factors have been deposited with the Protein Data Bank under accession code 4UUG. Lilly Skalden and Maren Thomsen contributed equally to this work

(Received 5 September 2014, revised 23 October 2014, accepted 13 November 2014) doi:10.1111/febs.13149

Introduction Amines are important building blocks for a range of shown for the large-scale synthesis of Sitagliptin [4]. pharmaceuticals, fine-chemicals and agrochemicals These enzymes are the most suitable for making chi- [1]. Although several chemical methods to produce ral amines because they can form optically pure chiral amines have been developed, biocatalytic amines from a racemic mixture at kinetic resolutions routes have become increasingly important. In the or, in the asymmetric synthesis mode, from ketones past few years, many amine transaminases (ATA) usually using alanine/pyruvate as the donor/acceptor with complementary enantio-preferences have been pair. Various improvements in asymmetric synthesis discovered [1 –3] and protein engineering has enabled have been made to increase the yield of amine for- substantial alteration of the substrate scope, as mation [5,6].

Abbreviations ATA, amine transaminase; mCPP, m-carboxyphenylpyridoxamine phosphate; PDB, Protein Data Bank; PLP, pyridoxal-50-phosphate.

FEBS Journal 282 (2015) 407–415 ª 2014 FEBS 407 Dual substrate recognition of an amine transaminase L. Skalden et al.

Depending upon their substrate scope, transaminases as D-alanine as an amino donor, similarly to ( S)-selec- can be divided into a-, x- and amine transaminases [7]. tive ATAs [15,19]. a-Transaminases convert substrates with the amino To investigate this dual substrate recognition in group in an alpha position to the carboxylate. x-Trans- detail for (R)-ATA, a thorough knowledge of the aminases may have several carbon atoms between the protein structure is necessary. Whereas structures of carboxylate and the terminal amino group [8]. ATAs a-transaminases and (S)-ATA are well documented in are able to convert ketones directly, so a carboxylate is the literature and the protein structure database not required. All transaminases are pyridoxal-50-phos- [18,20,21], until recently, the structures of ( R)-ATA phate (PLP)-dependent enzymes that transfer an amino were missing. Homology models [4] and substrate group from an amino donor to an amino receptor. In scope studies [15] provided a hint of the overall fold the seven different PLP-depending fold classes, trans- and the active site, but not a reliable explanation of aminases are located in fold classes I and IV [9 –12]. dual substrate recognition. Transamination takes place in two half reactions. In Recently, we reported the crystal structure of the the first half reaction, the amino group is transferred (R)-ATA from Aspergillus fumigatus (PDB code: from an amino donor to the cofactor PLP via several 4CHI) [22]. Two additional structures have been intermediates. The corresponding ketone or a-keto acid published in the meantime for the ( R)-ATAs from of the amino donor is released and pyridoxamine- Aspergillus terreus (PDB code: 4CE5) [23] and Nec- 50-phosphate is formed. In the second half-reaction, the tria haematococca (PDB codes: 4CMD and 4CMF) amino group is transferred from pyridoxamine-5 0-phos- [24]. Detailed analysis of the protein structures –espe- phate via the same types of intermediates in the reverse cially to understand dual substrate recognition – order to the amino acceptor. This is converted into the requires inhibitors such as gabaculine to be bound to amine or amino acid and PLP is regenerated [9,13]. the active site. Gabaculine (5-amino-1,3-cyclohexadie- a-Transaminases commonly use amino acids like nylcarboxylic acid) is a neurotoxic natural product aspartate or alanine as an amino donor [14]. Even if ala- from Streptomyces toyocaenis [25] and is known to be nine is not the preferred amino donor, x-transaminases, a covalent inhibitor for transaminases. During the as well as ATAs, can use alanine as the donor [15,16]. transamination of gabaculine, a cyclohexatrienyl sys- Dual recognition is the ability of the transaminase to tem, which is bound to the cofactor PLP, is formed. use the same substrate-binding pocket by accepting After spontaneous aromatization, m-carboxyphenyl- hydrophobic (e.g. phenyl group) as well as hydrophilic pyridoxamine phosphate (mCPP, Scheme 1) is formed (e.g. carboxylic group) substituents [17]. This phenome- and this finally inhibits the enzyme irreversibly [26]. non has been described for different types of transamin- To date, the substrate recognition and selectivity of ases and has been reported for branched-chain amino (R)-ATAs have been discussed based on the structural acid transaminases and (S)-selective ATAs. Branched- determination of a gabaculine complex (PDB code: chain amino acid transaminases, which belong to fold 4CMF) [24], but only in comparison with structural class IV, use hydrophilic residues to coordinate the car- and mutational data for a related enzyme (PDB code: boxylate at the border of the large binding pocket, but 4CE5) [23]. These studies resulted in different models the main hydrophobic character of this pocket is main- of substrate recognition. tained [2,17]. Dual substrate recognitions have also been reported in fold class I, where ( S)-ATAs and aromatic amino acid aminotransferases are grouped. For the aro- matic amino acid aminotransferase from Paracoc- cus denitrificans, loop rearrangements with conformational changes of side chains in the active site caused by the type of substrate are known [17]. In the (S)-ATA from Silicibacte pomeroyi [Protein Data Bank (PDB) code: 3HMU] movement of the amino acid resi- due Arg417 has been described [18]. Depending on the substrate, either L-alanine or a-methylbenzylamine ( a- MBA) is bound, and Arg417 flips in or out of the active site to constitute hydrogen bonding with the carboxylate of L-alanine. (R)-ATAs, which belong to fold class IV, also show Scheme 1. Structure of m-carboxyphenylpyridoxamine phosphate dual substrate recognition, and can use a-MBA as well (mCPP).

408 FEBS Journal 282 (2015) 407–415 ª 2014 FEBS L. Skalden et al. Dual substrate recognition of an amine transaminase

Here, we report the crystal structure of the ( R)-ATA The inhibitor is clearly bound in both monomers from A. fumigatus with the bound inhibitor gabacu- with full occupancy in monomer B and an occupancy line, as well as mutational studies of the same enzyme of 0.8 for the benzoic acid moiety in monomer A, to clarify the dual substrate recognition phenomenon owing to minor conformational disorder. However, for (R)-ATAs. they display the same overall orientation in the active site (rmsd on all atoms: 0.09 A). These two states dif- Results fer only in the orientation of the carboxylate (C7), which rotates in the open form (Fig. 3). Both states are in agreement with the previously reported classifi- Inhibitor binding to the active site cation of the small and the large binding pockets [22]. Crystal structure analysis of the ATA from A. fumiga- The C-atoms C11–C12 of the benzoic acid moiety are tus with bound gabaculine was carried out to obtain surrounded by the small binding pocket which is detailed insight into substrate recognition. Crystals of formed by the hydrophobic residues Val60, Phe113 the native enzyme crystallized in the trigonal space and Leu146, whereas the C-atoms C8, C9, C13 and group P3121 and were soaked with gabaculine to form the steric-demanding carboxylate are coordinated by the covalent inhibitor-adduct mCPP. The resulting His53*, Arg126 * (*both of the adjacent monomer) and crystal structure could be solved and refined to a reso- Tyr58; Val148 and Trp183 compose the large binding lution of 1.6 A. The asymmetric unit contains two pocket. The carboxylate of mCPP is coordinated by monomers forming a homodimer (Fig. 1) displaying a residues His53* and Tyr58 of the large binding pocket fold typical of class IV PLP-dependent enzymes. Each via a water molecule (Fig. 4). The proposed responsi- monomer contributes to the active site of the other by ble residue Arg126* for the dual substrate recognition providing an active-site loop (Gly121 –Asn135). Mono- has a direct distance of 3.8 A to the carboxylate of mer B with the active-site loop of monomer A forms a mCPP. However, here, coordination is also facilitated closed active-site loop conformation, whereas mono- via a water molecule (Fig. 4). mer A with the loop contributed by monomer B is pre- dominantly in an open conformation (occupancy of 0.7). However, some residues (besides Gly127–Asp132) of the active-site loop of monomer B can be traced in the electron-density maps with a low occupancy of 0.3 in the closed conformation. Superposition of these dif- ferent loop conformations of the monomers showed that the main difference in the backbone folding is restricted to the active-site loop only (rmsd on C a: 1.5 A; Fig. 2).

Fig. 2. Comparison of the open and closed active-site loop conformation by superposition of monomers A and B. In the closed loop conformation (green) R126 forms a water-mediated salt bridge with the carboxylate of mCPP (shown as stick model in yellow). The open loop conformation (orange) is stabilized by an alternative Fig. 1. Overall presentation of the ( R)-ATA from salt bridge of Arg126 with the carboxylate of Asp132 (R126-N g1– Aspergillus fumigatus with mCPP atoms shown as van der Waal’s D132-Od2 3.1 A). The Ca position of Arg126 is shifted by ~ 11.2 A. spheres (color code: carbon, yellow; oxygen, red; nitrogen, blue). Hydrogen bonds are indicated by stippled lines.

FEBS Journal 282 (2015) 407–415 ª 2014 FEBS 409 Dual substrate recognition of an amine transaminase L. Skalden et al.

A B

Fig. 3. The 2Fo À Fc electron-density maps contoured at 1 r of the mCPP. (A) mCPP in the open active-site loop conformation. Note alternative conformations of the carboxylate on top. (B) mCPP in the closed active-site loop conformation with the view rotated about 90 °.

Fig. 5. Comparison of the relative activity of the ( R)-ATA from Aspergillus fumigatius with different amino acceptors. All data are given relative to the wild-type activity with pyruvate as amino acceptor (100% = 1105 mU Ámg –1 protein).

Fig. 4. Presentation of the active site with bound mCPP (yellow) using the acetophenone assay [27] for its ability to showing the coordination of the carboxylate mediated via water accept pyruvate, succinic semialdehyde, 2-butanone, molecules to the residues Arg126 *, His53 * and Y58. Distances are displayed in A. Residues defining the small binding pocket are 2-pentanone or pentanal. A comparison of the spe- shown in green and those defining the large binding pocket in light cific activities of the wild-type and the Arg126Ala blue. The catalytic Lys179 is shown in orange. variant for the five amino acceptors revealed that the activity of the Arg126Ala for pyruvate is signifi- cantly reduced to only residual activity (3%) com- Influence of the Arg126Ala mutation on the pared with the wild-type. Also, the activity against substrate recognition pentanal is reduced; nevertheless, it is higher than To obtain deeper insight into the dual substrate rec- the activity for pyruvate. The low activity towards ognition of the ( R)-ATA from A. fumigatus , the the other substrates remained almost unaffected Arg126Ala variant was generated to investigate the (Fig. 5). influence of this residue on the conversion of differ- ent compounds because this arginine might have a Metal-binding sites similar function as Arg417 reported previously for the (S)-ATA from Silicibacter pomeroyi [18]. The At the surface of monomer B, a metal ion was obser- mutant was generated by a MegaWhop-PCR, over- ved with an octahedral coordination sphere established expressed in Escherichia coli, purified and analyzed by three water molecules (2.37 –2.40 A), two carbonyl

410 FEBS Journal 282 (2015) 407–415 ª 2014 FEBS L. Skalden et al. Dual substrate recognition of an amine transaminase

O-atoms of Leu133 and Asn136 (2.20 –2.33 A), and the side chain of Asn137 (2.42 A). Refinement could not discriminate between the isoelectronic ions of Na + or Mg 2+. The distances of the coordination sphere are suitable for Mg 2+, but Na + was defined because of the high concentration for crystallization. This coordination site is only possible with the open loop conformation, which is stabilized by crystal contacts. Thus, enzymatic activity is not depending on this putative metal-binding site.

Discussion Recently, an inhibitor structure of the ( R)-ATA from Nectria haematococca (PDB code: 4CMF) was pub- lished by the Littlechild group [25]. The sequence iden- tity of this transaminase to the here-reported enzyme is 69.1%. In the crystal structure of 4CMF, the Arg126 of the active-site loop – whose guanidinium group is within 5.5 A of the carboxylate of the mCPP – is not important for the binding of the carboxylate [24]. In the here-reported inhibitor complex of ( R)—ATA Fig. 6. Superposition of the inhibitor structures of ( R)-selective from A. fumigatus (PDB code: 4UUG), the mCPP amine transaminase from Aspergillus fumigatus and Nectria haematococca showing two different conformations of mCPP. conformation differs significantly compared with – – – Yellow C atoms: mCPP in chain B of the Aspergillus fumigatus 4CMF. The dihedral angles C4 C4A N9 C10 enzyme; green C atoms: mCPP in chain A of the Nectria (Scheme 1) of these two conformations differ by ~ 17 °, haematococca enzyme (rmsd on all atoms 0.85 A). Note the and the carboxylate C-atom C7 is displaced by alternative positions of Phe113. N and O atoms in both structures  ~ 2.7 A. In the Nectria haematococca enzyme the are blue and red, respectively. mCPP and the pyridoxine are almost in plane to each other, but these planes are at an angle of ~ 140° in the mCPP of the A. fumigatus enzyme (Fig. 6). This bend tion. In the case of ( S)-selective ATAs and aromatic in the inhibitor complex of ( R)-ATA from A. fumiga- amino acid transaminases, which both belong to fold tus results in a relaxed orientation in the active site class I, positional changes of the arginine are observed. without clashes with amino acid side chains. By con- Arginine flips into the active site to facilitate coordina- trast, some distances that are too short are observed in tion to carboxylated substrates, but is not involved in 4CMF, most probably due to some disorder caused by binding of hydrophobic substrates [28,29]. Comparison alternative conformations (e.g. Phe113 in Fig. 6). of our results with ( S)-selective ATAs and the aro- In addition, compared with 4CMF, the side chain of matic amino acid transaminase is difficult, because the Arg126 of 4UUG is closer to the carboxylate of these structures belong to fold class I, whereas the mCPP by 1.7 A and forms a weak salt bridge (N e–O2 ATA from A. fumigatus belongs to fold class IV. Nev- 3.8 A, Fig. 6). This interaction is supported by bridg- ertheless, a loop movement is observed in both fold ing hydrogen bonds via water molecules to the argi- classes [20]. In the ATA from A. fumigatus the open nine side chain and the carboxylate of mCPP (Fig. 4). active-site loop broadens the entrance to the active site The closed active-site loop allows only one distinct significantly, stabilized by a hydrogen bond between conformation of the carboxylate, whereas in the open Arg126 and Asp132. This is probably an artificial state form at least one additional conformation is observed induced and stabilized by crystal contacts, but it might due to more space and weaker alternative interactions be essential for the open conformation of the active- (Fig. 3). site loop. This assumption needs further investigation The conformation of the open active-site loop to prove the dual substrate recognition of ( R)-selective (Fig. 2) is stabilized by contacts to symmetry mates. ATAs with hydrophobic substrates. The closed active-site loop is not influenced by any The distinct importance of Arg126 of the active-site packing contacts, assuming a distinct active-site func- loop for binding carboxylated substrates is confirmed

FEBS Journal 282 (2015) 407–415 ª 2014 FEBS 411 Dual substrate recognition of an amine transaminase L. Skalden et al. by the Arg126Ala variant. This enzyme variant shows primers were used: specific primer for the mutant Arg126 a significant decrease to a residual activity of 3% Ala: CTTCCGGTTTAGAACCCGCAACACCGGTCAG towards pyruvate compared with the wild-type enzyme. and the T7 forward primer: TAATACGACTCACTA The Arg128Ala variant of the ATA from A. terreus , as TAGGG. After digestion of the wild-type plasmid, the reported by the Steiner group [23], also supports our mutated plasmid was transformed into competent E. coli observations. Mutations of the flipping arginine of ( S)- Top10 cells. The cells were plated on Luria –Bertani–agar Á –1 selective ATAs [29] have no significant effect on the plates with 0.1 mg mL ampicillin. After overnight culture ° activity towards nonpolar amino acceptors, whereas at 37 C colonies were picked and the plasmid was isolated the substrate recognition of nonpolar amino acceptors using the innuPREP Plasmid Mini Kit from Analytik Jena (Jena, Germany). Sequencing was performed by Eurofins in ( R)-selective ATAs requires additional interactions, MWG GmbH (Ebersberg, Germany). Finally, the plasmid as indicated by the loss of activity against pentanal. was transformed into E. coli BL21-competent cells. Interestingly, whereas Arg126 seems to play no important role in the fixed carboxylate-binding posi- tion in Nectria haematococca, our crystal structure Expression, purification and desalting analysis of the inhibitor complex, our kinetic data and Expression, purification and desalting of the Arg126Ala the results from the Steiner group [23] support the variant were performed as described for the wild-type importance of Arg126 of the active-site loop in deter- enzyme [30]. The protein content was determined by the mining a carboxylate-binding position in the ( R)-selec- BCA assay using a Varian Cary 50 Bio spectrophotometer. tive ATA from A. fumigatus.

Activity test Conclusion The activity of the wild-type and the Arg126Ala variant The crystal structure of the ATA from A. fumigatus was determined by the acetophenone assay [27]. The ability was solved with the bound inhibitor gabaculine to a to process succinic semialdehyde, pentanal, 2-butanone or  resolution of 1.6 A. The orientation and binding of 2-pentanone as an amino acceptor was investigated by the carboxylate of pyruvate are facilitated by Arg126 measuring the activity of the variant and the wild-type with via water molecules. Furthermore, mutagenesis of this a-MBA as the amino donor at 254 nm. Amino acceptors arginine present in the active-site loop to alanine and a-MBA were used at a concentration of 2.5 m M. results in substantially reduced activity with pyruvate as the amino acceptor. Hence, we conclude that Crystallization and inhibitor soaking Arg126 is required to finalize substrate recognition and is important for dual substrate recognition in the ( R)- Crystals were obtained by the hanging-drop method and ATA from A. fumigatus. with a reservoir solution containing 0.1 M sodium acetate,

Materials and methods Table 1. Data collection and processing statistic. Values in parentheses are for the highest resolution shell. Chemicals and materials X-ray source BL14.1, BESSY II All chemicals were purchased from Fluka (Buchs, Switzer- Detector Pilatus 6M  land), Sigma (Steinheim, Germany), Merck (Darmstadt, Wavelength (A) 0.91841 Temperature (K) 100 Germany), VWR (Hannover, Germany) or Carl Roth (Kar- Space group P3 21 lsruhe, Germany) and were used without further purifica- 1 a = b/c (A) 144.4/96.1 tion, unless otherwise specified. Polymerases were obtained Resolution range (A) 47.28 –1.6 (1.63–1.60) from New England Biolabs GmbH (NEB, Frankfurt am Unique reflections 151504 (7496) Main, Germany) and primers were ordered from Invitrogen Redundancy 4.5 (4.5)

(Life Technologies GmbH, Darmstadt, Germany). Rmerge (%) 5.8 (66.8)

Rmeas (%) 6.6 (75.7)

Rpim (%) 3.1 (35.3) Cloning and mutagenesis 16.1 (2.3) CC ½ (%) 99.9 (75.6) Mutagenesis was performed using the gene encoding the Completeness (%) 99.9 (99.8) ATA from A. fumigatus , which has a C-terminal His tag and Wilson B factor (A²) 26.7 was cloned into a pET22b plasmid [2]. The mutant Arg126 Total rotation/increment (°) 80/0.2 Ala was generated by a MegaWhop-PCR. The following

412 FEBS Journal 282 (2015) 407–415 ª 2014 FEBS L. Skalden et al. Dual substrate recognition of an amine transaminase

Table 2. Refinement statistic. Acknowledgements Data set (R)-ATA/gabaculine MT thanks the Landesgraduiertenkolleg Mecklenburg- Resolution range (A) 47.25 –1.6 Vorpommern for financial support. We thank the Working/test reflections 143 859/7626 European Union (KBBE-2011-5, grant no. 289350) for a R/Rfree (%) 13.4/14.8 financial support within the European Union Seventh Protein residues/water molecules 641/794 Framework Programme. Diffraction data were col- No. ligands 10 lected on BL14.1 operated by the Helmholtz-Zentrum rmsd from ideality Berlin (HZB) at the BESSY II electron-storage ring Bond lengths ( A) 0.0140 Bond angles (°) 1.6979 (Berlin-Adlershof, Germany). Average B factor (A²) 21 Ramachandran statistics Author contributions Most favored regions (%) 97.0 Outliers (%) 0.3 LS cloned and purified the protein, and performed PDB entry code 4UUG kinetic assays. MT crystallized, soaked, solved and a R = Σ|| Fo À |Fc||/Σ|Fo|, where Fo and Fc are the observed and cal- analyzed the protein structure. UTB and WH initiated culated structure factors, respectively. Rfree, Analogous to R-factor the joint project. All authors were involved in discuss- except the summation is over 5% of reflections not included in ing data and preparing the manuscript. refinement.

References pH 4.6, and 2.0 M sodium formate. The crystallization drop contained an equal volume of reservoir solution 1 H ohne€ M & Bornscheuer UT (2009) Biocatalytic and protein solution. The concentration of the protein routes to optically active amines. ChemCatChem 1, – solution used was 12.2 mg ÁmL 1. The obtained crystals 42 –51. had only a slight yellow color, and because of this they 2 H ohne€ M, Sch atzle€ S, Jochens H, Robins K & were first transferred to a solution containing 10% Bornscheuer UT (2010) Rational assignment of key 2-methyl-2,4-pentanediol, 0.1 M sodium acetate pH 4.6, motifs for function guides in silico enzyme 2.0 M sodium formate and 0.1 mM PLP to ensure full identification. Nat Chem Biol 6, 807 –813. occupation of the cofactor. After 15 min, the crystals 3 Kohls H, Steffen-Munsberg F & H ohne€ M (2014) obtained a much brighter yellow color and were trans- Recent achievements in developing the biocatalytic ferred to a solution containing 10% 2-methyl-2,4-pentane- toolbox for chiral amine synthesis. Curr Opin Chem Biol diol, 0.1 M sodium acetate pH 4.6, 2.0 M sodium formate 19 , 180 –192. and 0.1 mM gabaculine. This solution had cryo-protectant 4 Savile CK, Janey JM, Mundorff EC, Moore JC, Tam properties. S, Jarvis WR, Colbeck JC, Krebber A, Fleitz FJ, Brands J et al. (2010) Biocatalytic asymmetric synthesis of chiral amines from ketones applied to sitagliptin Data collection and structure determination manufacture. Science 329, 305 –309. X-Ray diffraction data were collected at 100 K on beam- 5 H ohne€ M, K uhl€ S, Robins K & Bornscheuer UT (2008) line 14.1 at the BESSY II synchrotron (Berlin, Germany). Efficient asymmetric synthesis of chiral amines by All diffraction images were indexed and integrated with XDS combining transaminase and pyruvate decarboxylase. [31], scaling and calculation of the structure factors was ChemBioChem 9, 363 –365. performed with AIMLESS [32]. Data collection and processing 6 Matcham G, Bhatia M, Lang W, Lewis C, Nelson R, statistics are summarized in Table 1. Wang A & Wu W (1999) Enzyme and reaction The structure was solved with Molecular Replacement engineering in biocatalysis: synthesis of S- using the program PHASER [33]. The structure of the apoen- methoxyisopropylamine (= S-1-methoxypropan-2- zyme from A. fumigatus (PDB code: 4CHI) served as a amine). Chimia 53 , 584 –589. template [22]. The ligand was built manually into the 7 H ohne€ M & Bornscheuer UT (2012) Application of

Fo À Fc electron-density map with COOT [34]. Occupancies transaminases in organic synthesis. In Enzymes in of the mCPP were determined by adapting the B factors to Organic Synthesis (May O, Gr oger€ H & Drauz W, eds), neighboring amino acid residues. Refinement was carried pp. 779–820. Wiley-VCH, Weinheim. out with REFMAC5 including TLS segments [35]. The quality 8 Schrewe M, Ladkau N, B uhler€ B & Schmid A (2013) of the refined protein model was validated using MOLPROBITY Direct terminal alkylamino-functionalization via [36]. Refinement statistics are listed in Table 2. All molecu- multistep biocatalysis in one recombinant whole-cell lar graphics were prepared using PYMOL [37]. catalyst. Adv Synth Catal 355, 1693 –1697.

FEBS Journal 282 (2015) 407–415 ª 2014 FEBS 413 Dual substrate recognition of an amine transaminase L. Skalden et al.

9 Eliot A & Kirsch J (2004) Pyridoxal phosphate 22 Thomsen M, Skalden L, Palm GJ, H ohne€ M, enzymes: mechanistic, structural, and evolutionary Bornscheuer UT & Hinrichs W (2014) Crystallographic considerations. Annu Rev Biochem 73 , 383 –415. characterization of the ( R)-selective amine transaminase 10 Percudani R & Peracchi A (2009) The B6 database: a from Aspergillus fumigatus. Acta Crystallogr D 70 , tool for the description and classification of vitamin B6- 1086–1093. dependent enzymatic activities and of the corresponding 23 Lyskowski A, Gruber C, Steinkellner G, Sch urmann€ M, protein families. BMC Bioinformatics 10 , 273 –280. Schwab H, Gruber K & Steiner K (2014) Crystal 11 Berkovitch F, Behshad E, Tang K-H, Enns EA, Frey structure of an R-selective x-transaminase from PA & Drennan CL (2004) A locking mechanism Aspergillus terreus. PLoS One 9, e87350. preventing radical damage in the absence of substrate, 24 Sayer C, Martinez-Torres RJ, Richter N, Isupov MN, as revealed by the x-ray structure of lysine 5,6- Hailes HC, Littlechild JA & Ward JM (2014) aminomutase. Proc Natl Acad Sci USA 101, 15870 – The substrate specificity, enantioselectivity and 15875. structure of the ( R)-selective amine : pyruvate 12 Lepore BW, Ruzicka FJ, Frey PA & Ringe D (2005) transaminase from Nectria haematococca. FEBS J 281, The x-ray crystal structure of lysine-2,3-aminomutase 2240–2253. from Clostridium subterminale. Proc Natl Acad Sci 25 Kobayashi K, Miyazawa S, Terahara A, Mishima H & USA 102, 13819 –13824. Kurihara H (1976) Gabaculine: c-aminobutyrate 13 Jansonius JN (1998) Structure, evolution and action of aminotransferase inhibitor of microbial origin. vitamin B6-dependent enzymes. Curr Opin Struct Biol Tetrahedron Lett 17 , 537 –540. 8, 759 –769. 26 Rando RR (1977) Mechanism of the irreversible 14 Cronin CN & Kirsch JF (1988) Role of arginine-292 in inhibition of c-aminobutyric acid-a-ketoglutaric acid the substrate specificity of aspartate aminotransferase as transaminase by the neurotoxin gabaculine. examined by site-directed mutagenesis. Biochemistry 27 , Biochemistry 16 , 4604 –4610. 4572–4579. 27 Sch atzle€ S, H ohne€ M, Redestad E, Robins K & 15 Park E-S, Dong J-Y & Shin J-S (2014) Active site Bornscheuer UT (2009) Rapid and sensitive kinetic model of ( R)-selective x-transaminase and its assay for characterization of x-transaminases. Anal application to the production of d-amino acids. Appl Chem 81 , 8244 –8248. Microbiol Biotechnol 98 , 651 –660. 28 Okamoto A, Nakai Y, Hayashi H, Hirotsu K & 16 Shin J-S & Kim B-G (2001) Comparison of the omega- Kagamiyama H (1998) Crystal structures of transaminases from different microorganisms and Paracoccus denitrificans aromatic amino acid application to production of chiral amines. Biosci aminotransferase: a substrate recognition site Biotechnol Biochem 65 , 1782 –1788. constructed by rearrangement of hydrogen bond 17 Hirotsu K, Goto M, Okamoto A & Miyahara I (2005) network. J Mol Biol 280, 443 –461. Dual substrate recognition of aminotransferases. Chem 29 Steffen-Munsberg F, Vickers C, Thontowi A, Sch atzle€ Rec 5, 160 –172. S, Meinhardt T, Svedendahl Humble M, Land H, 18 Steffen-Munsberg F, Vickers C, Thontowi A, Sch atzle€ Berglund P, Bornscheuer UT & H ohne€ M (2013) S, Tumlirsch T, Svedendahl Humble M, Land H, Revealing the structural basis of promiscuous Berglund P, Bornscheuer UT & H ohne€ M (2013) amine transaminase activity. ChemCatChem 5, Connecting unexplored protein crystal structures to 154–157. enzymatic function. ChemCatChem 5, 150 –153. 30 Thomsen M, Skalden L, Palm GJ, H ohne€ M, 19 Koszelewski D, G oritzer€ M, Clay D, Seisser B & Bornscheuer UT & Hinrichs W (2013) Crystallization Kroutil W (2010) Synthesis of optically active amines and preliminary X-ray diffraction studies of the ( R)- employing recombinant x-transaminases in E. coli cells. selective amine transaminase from Aspergillus ChemCatChem 2, 73 –77. fumigatus. Acta Crystallogr Sect F Struct Biol Cryst 20 Humble MS, Cassimjee KE, H akansson M, Kimbung Commun 69 , 1415 –1417. YR, Walse B, Abedi V, Federsel H-J, Berglund P & 31 Kabsch W (2010) XDS. Acta Crystallogr D 66 , Logan DT (2012) Crystal structures of the 125–132. Chromobacterium violaceum x-transaminase reveal 32 Collaborative Computational Project Number 4 (1994) major structural rearrangements upon binding of The CCP4 suite: programs for protein crystallography. coenzyme PLP. FEBS J 279, 779 –792. Acta Crystallogr D 50 , 760 –763. 21 Ford GC, Eichele G & Jansonius JN (1980) Three- 33 McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn dimensional structure of a pyridoxal-phosphate- MD, Storoni LC & Read RJ (2007) Phaser dependent enzyme, mitochondrial aspartate crystallographic software. J Appl Crystallogr 40 , aminotransferase. Proc Natl Acad Sci 77 , 2559 –2563. 658–674.

414 FEBS Journal 282 (2015) 407–415 ª 2014 FEBS L. Skalden et al. Dual substrate recognition of an amine transaminase

34 Emsley P & Cowtan K (2004) Coot: model-building 36 Chen VB, Arendall WB III, Headd JJ, Keedy DA, tools for molecular graphics. Acta Crystallogr D 60 , Immormino RM, Kapral GJ, Murray LW, 2126–2132. Richardson JS & Richardson DC (2010) 35 Murshudov GN, Skubak P, Lebedev AA, Pannu NS, MolProbity: all-atom structure validation for Steiner RA, Nicholls RA, Winn MD, Long F & Vagin macromolecular crystallography. Acta Crystallogr D AA (2011) REFMAC5 for the refinement of 66 , 12 –21. macromolecular crystal structures. Acta Crystallogr D 37 Delano WL (2002) PyMOL in http://www.pymol.org. 67 , 355 –367.

FEBS Journal 282 (2015) 407–415 ª 2014 FEBS 415

Artikel IV DOI: 10.1002/cbic.201500074 Communications

Two Subtle Amino Acid Changes in a Transaminase Substantially Enhance or Invert Enantiopreference in Cascade Syntheses Lilly Skalden, [a] Christin Peters,[a] Jonathan Dickerhoff, [b] Alberto Nobili, [a] Henk-Jan Joosten, [c] Klaus Weisz, [b] Matthias Hçhne, [d] and Uwe T. Bornscheuer* [a]

Amine transaminases (ATAs) are powerful enzymes for the ste- cade reaction. This biomimetic approach represents an exciting reospecific production of chiral amines. However, the synthesis recent development in “white biotechnology”.[3] In a cascade, of amines incorporating more than one stereocenter is still a intermediate downstream and purification steps are unneces- challenge. We developed a cascade synthesis to access optical- sary, and so cascades also represent a very promising approach ly active 3-alkyl-substituted chiral amines by combining two from a green chemistry point of view. For example, Sehl et al. asymmetric synthesis steps catalyzed by an enoate reductase used the combination of a thiamine diphosphate (ThDP)-de- and ATAs. The ATA wild type from Vibrio fluvialis showed only pendent acetohydroxy acid synthase and amine transaminases modest enantioselectivity (14 % de ) in the amination of (S)-3- to produce the 1,2-amino alcohols pseudo- and norpseudo- methylcyclohexanone, the product of the enoate-reductase- ephedrine.[4] Several other cascades with up to 13 different en- catalyzed reaction step. However, by protein engineering we zymes have been reported recently, thus showing the general created two variants with substantially improved diastereo- applicability of this strategy.[5] A great advantage is the modu- selectivities: variant Leu56Val exhibited a higher R selectivity larity of this approach, because different combinations of en- (66% de ) whereas the Leu56Ile substitution caused a switch in zymes with distinct enantiopreferences result in a spectrum of enantiopreference to furnish the S-configured diastereomer products with different diastereomeric configurations. (70% de ). Addition of 30% DMSO further improved the selec- We were interested in the synthesis of ring-substituted exo- tivity and facilitated the synthesis of (1 R,3 S)-1-amino-3-methyl- cyclic amines because they are employed as building blocks cyclohexane with 89% de at 87 % conversion. for the production of pharmaceuticals and fine chemicals.[6] Typically, in reactions employing ATAs, acyclic ketones contain- ing one large and one small substituent apiece have been The enantioselective synthesis of chiral primary amines has ad- used as substrates,[7] and the molecular mechanisms governing vanced remarkably during the last decade,[1] and some enzy- substrate recognition and enantioselectivity are fairly well un- matic processes can now outperform well-established chemical derstood: the active sites of an R- and an S-selective ATA differ processes. One benchmark example is the manufacture of Sita- in the localization of the small and the large binding pocket, gliptin, the active pharmaceutical ingredient (API) of Januvia, thus forcing the ketone into one specific binding mode, lead- which demonstrates the potential of amine transaminase tech- ing to the formation of a distinct amine enantiomer. nology.[2] In contrast, only a few reports on the amination of (substi- Many APIs contain multiple stereocenters, and their prepara- tuted) cyclic ketones have been published.[6b,8] Structurally, tion is therefore challenging. An elegant solution with regard cyclic ketones are challenging with respect to activity and ste- to atom economy is the combination of different (bio)catalytic reoselectivity[9] because they do not possess the well-defined asymmetric synthesis reactions facilitating the stepwise crea- large and small substituents that are usually important for ach- tion of several stereogenic centers. If reaction conditions are ieving highly selective transamination. A few cyclic substrates compatible, the synthesis can be carried out as a one-pot cas- with large substituents (e.g., 2-(1H-benzoimidazol-2-yl)-1-meth- ylpiperidin-4-one,[6b] 2-[2-(3,4-dimethoxyphenyl)ethoxy]cyclo- [6a] [8] [a] L. Skalden, C. Peters, A. Nobili, Prof. Dr. U. T. Bornscheuer hexanone, or methyl (3-oxocyclohexyl)acetate ) were inves- Biotechnology and Enzyme Catalysis, Institute of Biochemistry tigated recently, but cyclic compounds only bearing single Felix-Hausdorff-Strasse 4, 17487 Greifswald (Germany) small alkyl substituents have not been explored. Therefore, we E-mail: [email protected] envisioned the synthesis of 1-amino-3-methylcyclohexanes (3) [b] J. Dickerhoff, Prof. Dr. K. Weisz Analytical Biochemistry, Institute of Biochemistry as model compounds by combining two asymmetric synthesis Felix-Hausdorff-Strasse 4, 17487 Greifswald (Germany) steps based on 1) an enoate reductase (ERED), and 2) an amine [c] Dr. H.-J. Joosten transaminase and starting from 3-methylcyclohex-2-enone (1, Bio-Prodict Scheme 1). Nieuwe Marktstraat 54E, 6511 AA Nijmegen (The Netherlands) An efficient cascade requires: 1) a highly enantioselective [d] Prof. Dr. M. Hçhne ERED, 2) that the precursor 1 should not be an ATA substrate Protein Biochemistry, Institute of Biochemistry Felix-Hausdorff-Strasse 4, 17487 Greifswald (Germany) (or otherwise, 1 must be fully consumed in the first step), to Supporting information for this article is available on the WWW under avoid the formation of byproducts, and 3) that the ATA should http://dx.doi.org/10.1002/cbic.201500074. introduce the amino group with the desired enantiopreference

ChemBioChem 2015, 16 , 1041 – 1045 1041  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Communications

after 18 h. The observed compo- sition of 30:26:40:4 for the four amine diastereomers 3a –3d clearly shows a low enantioselec- tivity in the amination reaction. Remarkably, the ratios of the (1 R)- to the (1 S)-amines of 30:26 and 40:4 indicate that ATA-Vfl in- troduces the amino group with R enantiopreference (see below; Figure S4A and B in the Sup- porting Information). This is in Scheme 1. The combination of an ERED with an ATA in a cascade reaction enables access to optically pure contrast with the well-docu- [16] 1-amino-3-methylcyclohexanes. Glucose dehydrogenase (GDH) can be used for cofactor recycling in the ERED-cat- mented S enantiopreference alyzed reaction, whereas lactate dehydrogenase (LDH) and GDH allow a shift of the equilibrium in the ATA-cata- of ATA-Vfl and the previously re- lyzed reaction. Note that different combinations of EREDs and ATAs with complementary enantiopreferences are ported high S selectivity in the required to obtain all four diastereomers. The boxed diastereomers were obtained in this study. conversion of the related six- membered cyclic ketone methyl and with high enantioselectivity, preferentially regardless of (3-oxocyclohexyl)acetate.[8] Obviously, the type and size of sub- the configuration of the 3-methyl group. Enoate reductases, stituents at the b-position in a cyclic ketone such as 2 strongly which are flavin-mononucleotide-containing (FMN-containing) influence enantioselectivity and -preference in the amination NAD(P)H-dependent oxidoreductases, are well known to con- reaction. vert a/b-unsaturated ketones or aldehydes into the corre- To identify amino acid positions governing enantioselectivity, sponding enantiopure compounds.[10] ATAs introduce amine we modeled the internal aldimine intermediate of 2 into the groups, often with excellent enantioselectivities,[7b, 11] into a structure of ATA-Vfl by using YASARA[17] (Figure 1A). From wide range of ketones, but enantioselectivity drops drastically visual inspection, we chose Leu56 as target for mutagenesis if the size difference between the two substituents on either because its side chain would be in contact with the methyl side of the carbonyl group is too small or if the ketones are substituent in the PLP intermediates of 3b and 3c , so muta- small cyclic molecules such as our target ketone 2. If (dia)ste- tions of Leu56 might induce a change in selectivity. Analysis of reoselectivity is not sufficient for a ketone bearing a chiral the amino acid distribution at position 56 by use of the 3DM center in the a- or b-position, this can in principle be improved database[14] identified six amino acids (Ile, Val, Ala, Trp, Tyr, Phe) by protein engineering. One example was shown by Limanto besides leucine as most frequently occurring in PLP fold type I et al., who used ATA to convert 2-[2-(3,4-dimethoxyphenyl)- proteins (Figure 1B). In a subset containing only subfamilies ethoxy]cyclohexanone; [6a] after three rounds of evolution selec- harboring S-selective ATA sequences, leucine is found in 90% tive enzymes for both chiral centers were found. of the sequences. These six mutants were generated by Mega- The aim of this work was to demonstrate the applicability of Whop PCR, overexpressed in E. coli Bl21 (DE3), and purified by the envisioned cascade and, in this context, to identify enan- His-tag affinity chromatography (Figures S1–S3). Only the tiocomplementary ATAs, preferably with high stereoselectivity Leu56Ile, Leu56Val, and Leu56Ala mutants showed sufficient towards 2. Identification of suitable ATAs can be achieved by activity. Because of its lower activity, in combination with enan- methods of protein discovery[7b,12] and protein engineer- tioselectivities that were similar to those of the Leu56Val ing,[2a,13] which can be guided by bioinformatics tools. One of mutant, the Leu56Ala mutant was also excluded. these tools is the 3DM software package, which connects GC analysis revealed that variant Leu56Ile showed an im- a high-quality, structure-guided sequence alignment of a pro- proved 1 R enantioselectivity in the amination of the R enantio- tein superfamily[14] with bioinformatics analysis tools. 3DM inte- mer of rac-2 (Table 1, entry 2). Interestingly, the enzyme ami- grates data from various sources, including sequences, struc- nated (S)- 2 with inverted enantiopreference. The Leu56Val mu- tural information, protein ligand contacts, and mutational data tation improved the R enantioselectivity in the amination of from the literature, using a unified amino acid residue number- (S)- 2, but had no effect on the R enantiomer of 2 (Table 1, ing system for all proteins. We have previously used this plat- entry 3, Figure S4 B). In the case of 2, the carbonyl group is form to guide the degree of randomization for the creation of part of a flexible aliphatic ring: 2 is an almost symmetric com- “small, but smart” libraries in multiple-site saturation mutagen- pound because only a small methyl group is present in the b- esis experiments.[15] In this work, we have used information position. Thus, the typical steric constraints originating from from 3DM to guide the identification of variants of the ATA the interaction of a substrate bearing well-defined small and from Vibrio fluvialis (ATA-Vfl) with suitable stereoselectivity and large substituents with the enzyme’s active site[7a,16b] do not enantiopreference for the synthesis of different diastereomers apply for 2. This difference could lead to the switch of the of 3, because an initial screening had identified this ATA as commonly observed stereoselectivity of the ATA from V. fluvia- having sufficient activity. Unfortunately, ATA-Vfl exhibited only lis resulting in the R configuration here. We were thus able to moderate selectivity, with rac-2 having been fully converted show that subtle changes in the active site geometry produced

ChemBioChem 2015, 16 , 1041 – 1045 www.chembiochem.org 1042  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Communications

Figure 1. A) View of the active site of ATA-Vfl together with the modeled external aldimine of 2. The PLP intermediate and Leu56 are shown as magenta and green sticks, respectively. The cyan spheres indicate which hydrogen atom is replaced by a methyl group if one of the amines 3a –3d is bound to PLP. Note that the side-chain Arg415 can also adopt a conformation flipping outside the active site.[20] B) Amino acid distribution of PLP fold type I proteins at position 56 of ATA-Vfl. In the 3DM database this position corresponds to number 46.

periments. In accordance with Table 1. Diastereomeric purities of the products of the transamination of substrate 2 and of the simultaneous the results described above, ap- cascade reactions starting from 1 with employment of OYE as the ERED and ATA-Vfl or its mutants in the ab- plication of the Leu56Val variant sence or presence of the organic solvents DMF or DMSO. increased the diastereomeric Diastereomeric composition purity of (1 R,3 S)- 3a from 14% de Enzyme Substrate DMSO DMF t 3a 3b 3c 3d de Conversion (WT) to 65% de (mutant). In con- [a] [%] [%] [h] (1 R,3 S) (1S,3 S) (1R,3 R) (1S,3 R) [%] [%] trast, the mutant Leu56Ile gave 1 ATA-Vfl rac-2 0 1 18 30 26 40 4 n.a. 99 the (1 S,3 S)-amine 3b with 70% 2 ATA Leu56Ile rac-2 0 1 18 7 47 46 n.d. n.a. 99 de (Table 1, entries 8 and 9, Fig- 3 ATA Leu56Val rac-2 0 1 18 40 16 39 5 n.a. 99 4 OYE +ATA-Vfl 1 1 0 48 56 44 n.d. n.d. 12 99 ures S5–S7). 5 OYE +ATA Leu56Ile 1 1 0 48 14 86 n.d. n.d. 71 97 Because the ERED reaction ex- 6 OYE +ATA Leu56Val 1 1 0 48 83 17 n.d. n.d. 66 99 clusively produced the S enan- 7 OYE +ATA-Vfl 1 0 1 48 57 43 n.d. n.d. 14 97 tiomer of 2, it was possible to 8 OYE +ATA Leu56Ile 1 0 1 48 15 85 n.d. n.d. 70 95 9 OYE +ATA Leu56Val 1 0 1 48 82 18 n.d. n.d. 65 98 quantify the selectivity of the 10 OYE+ATA-Vfl 1 30 0 72 88 12 n.d. n.d. 76 88 amine transaminase by deter- 11 OYE+ATA Leu56Ile 1 30 0 72 59 41 n.d. n.d. 18 81 mining the ratio of cis and trans 12 OYE+ATA Leu56Val 1 30 0 72 94 6 n.d. n.d. 89 87 products 3a and 3b . All 1H NMR n.d.: not detected. [a] The enantioselectivity of the ERED was >99% ee for the S enantiomer. resonances of both diastereo- mers in the mixture were as- signed through DQF-COSY ex- by small side chain variations at position 56 (located at the periments. The downfield-shifted H1 protons of the cis and the boundary of the two binding pockets) significantly affected trans diastereomer significantly differ in their chemical shifts, diastereoselectivity. by nearly 0.5 ppm, and were subsequently used for a stereo- To provide access to highly optically pure diastereomers, we chemical assignment with a one-dimensional NOESY version then investigated the one-pot reaction starting with the reduc- employing selective refocusing (Figure 2). Thus, on excitation tion of the a/b-unsaturated ketone 1 by using the enoate of the H1 resonance at 2.62 ppm in the selective 1D NOESY ex- reductases from Saccharomyces carlsbergensis (OYE) or from periment, NOEs develop for H3 and H5a as well as for the two Pseudomonas putida (XenB). This afforded optically pure (S)- 2 directly neighboring H2a and H6a protons (Figure 2B). This with > 99% ee [18] (Figure S8). Because OYE showed quantitative places H1 and H3 on the same side of the cyclohexane ring conversion, this enzyme was used for further investigations. To system in close proximity, thus confirming the cis form with shift the equilibrium of the amination reaction, the well-estab- both substituents being in an equatorial position. On the other lished lactate dehydrogenase/glucose dehydrogenase (LDH/ hand, upon excitation of the other H1 resonance at 3.08 ppm GDH) system was used;[16a,19] it also served to recycle the a noticeable NOE develops for the methyl group at position 7 NADPH required for the ERED. Interestingly, 1 was not convert- in addition to the two H2a and H6a neighboring protons (Fig- ed by ATAs, so the process could be performed as a cascade ure 2C). These NOEs, indicative of short interproton distances, reaction in a simultaneous mode (both enzymes being present reveal a trans configuration with the methyl group oriented on from the beginning). This setup was used in all subsequent ex- the same side as H1. Based on the unambiguous assignment

ChemBioChem 2015, 16 , 1041 – 1045 www.chembiochem.org 1043  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Communications

of the two stereogenic centers to yield particular diastereo- mers of the cyclic, alkyl-substituted amine 3. Thus, we consider this cascade to represent a valuable synthetic strategy for the production of compounds of this class. Mutations leading to very small variations in the active site can strongly influence the diastereoselectivity of the amine transaminase from V. flu- vialis, and this is further improved by use of co-solvents.

Acknowledgements

We thank the European Union (KBBE-2011-5, grant no. 289350) and the DFG (grant no. Bo1862/6-1) for financial support. We are grateful to Prof. Jon D. Stewart (University of Florida, USA) for the gene encoding OYE1 and Prof. Byung-Gee Kim (Seoul Nation- al University, Korea) for the gene encoding the ATA from V. fluvia- lis.

Keywords: amine transaminases · cascade synthesis · Figure 2. NMR spectra of the commercial 3-methylcyclohexylamine standard. A) Normal 1D spectrum. B), C) 1D NOESY spectra of B) the cis, and C) the enantiopreference · enzyme catalysis · protein engineering trans diastereomer with selective excitation of the corresponding H1 protons at 2.62 and 3.08 ppm, respectively. The two antiphase multiplets at high [1] a) M. Hçhne, U. T. Bornscheuer, ChemCatChem 2009, 1, 42–51; b) H. field in (B) are artifacts arising from polarization transfer between H1 and Kohls, F. Steffen-Munsberg, M. Hçhne, Curr. Opin. Chem. Biol. 2014, 19 , the two scalar coupled H2b and H6b protons. 180– 192; c) J. Rudat, B. Brucher, C. Syldatk, AMB Express 2012, 2, 11; d) M. Malik, E.-S. Park, J.-S. Shin, Appl. Microbiol. Biotechnol. 2012, 94 , 1163–1171; e) P. Tufvesson, J. Lima-Ramos, J. S. Jensen, N. Al-Haque, W. of the H1 protons of the two diastereomers, a preparative- Neto, J. M. Woodley, Biotechnol. Bioeng. 2011, 108, 1479 –1493. [2] a) C. K. Savile, J. M. Janey, E. C. Mundorff, J. C. Moore, S. Tam, W. R. scale cascade reaction (77 % isolated yield) with the ERED and Jarvis, J. C. Colbeck, A. Krebber, F. J. Fleitz, J. Brands, P. N. Devine, G. W. the ATA-Leu56Ile was found to result in a preference for the Huisman, G. J. Hughes, Science 2010, 329, 305– 309; b) A. A. Desai, trans product 3b over the cis product 3a (Figure S9). Angew. Chem. Int. Ed. 2011, 50 , 1974 –1976; Angew. Chem. 2011, 123, The dependence of enantioselectivity on the type of solvent 2018 –2020. [3] a) J. Muschiol, C. Peters, N. Oberleitner, M. Mihovilovic, U. T. Bornscheu- [21] is well known in traditional chemical synthesis, but has also er, F. Rudroff, Chem. Commun. 2014, DOI: 10.1039/C4CC08752F; b) E. been described for enzyme reactions, including for ATAs.[22] Ricca, B. Brucher, J. H. Schrittwieser, Adv. Synth. Catal. 2011, 353, 2239 – Therefore we sought to identify suitable co-solvents that 2262; c) R. C. Simon, N. Richter, E. Busto, W. Kroutil, ACS Catal. 2014, 4, might further improve the enantioselectivity of the ATA reac- 129– 143; d) I. Oroz-Guinea, E. García-Junceda, Curr. Opin. Chem. Biol. 2013, 17 , 236–249; e) V. Kçhler, N. J. Turner, Chem. Commun. 2015, 51 , tion. Indeed, addition of DMSO or DMF during the transamina- 450– 464. tion of rac-2 (with ATA-Vfl and its mutants) results in increased [4] T. Sehl, H. C. Hailes, J. M. Ward, R. Wardenga, E. von Lieres, H. Offer- formation of the cis products (Table 1, Figure S10). mann, R. Westphal, M. Pohl, D. Rother, Angew. Chem. Int. Ed. 2013, 52 , The cascade reactions (Scheme 1) were then also investigat- 6772 –6775; Angew. Chem. 2013, 125, 6904 –6908. [5] I. Ardao, A.-P. Zeng, Chem. Eng. Sci. 2013, 87 , 183 –193. ed with the co-solvents DMSO and DMF. Fortunately, all of the [6] a) J. Limanto, E. R. Ashley, J. Yin, G. L. Beutner, B. T. Grau, A. M. Kassim, enzymes (ATA, OYE, LDH, and GDH) are active in the presence M. M. Kim, A. Klapars, Z. Liu, H. R. Strotman, M. D. Truppo, Org. Lett. of DMSO. Although OYE tolerated 30% DMSO, as reported 2014, 16 , 2716 – 2719; b) Z. Peng, J. W. Wong, E. C. Hansen, A. L. A. earlier,[23] the enzyme was inactive in 30% DMF. High levels of Puchlopek-Dermenci, H. J. Clarke, Org. Lett. 2014, 16 , 860 –863. [7] a) J.-S. Shin, B.-G. Kim, J. Org. Chem. 2002, 67 , 2848 –2853; b) F. Steffen- DMSO increased the preference for the cis product 3a in all Munsberg, C. Vickers, A. Thontowi, S. Schätzle, T. Tumlirsch, M. Sveden- three cases. We were pleased to find that, by adding 30% dahl Humble, H. Land, P. Berglund, U. T. Bornscheuer, M. Hçhne, Chem- DMSO, the diastereomeric excess of the cis product (1 R,3 S)- 3a CatChem 2013, 5, 150– 153. could be substantially improved to 76 and 89 % de in the pres- [8] E. Siirola, F. G. Mutti, B. Grischek, S. F. Hoefler, W. M. F. Fabian, G. Grogan, W. Kroutil, Adv. Synth. Catal. 2013, 355, 1703 –1708. ence of ATA-Vfl or the mutant Leu56Val, respectively. Because [9] M. Hçhne, K. Robins, U. T. Bornscheuer, Adv. Synth. Catal. 2008, 350, mutant Leu56Ile forms the trans product 3b , DMSO addition 807– 812. had no useful effect for preparative purposes, but we observed [10] H. S. Toogood, J. M. Gardiner, N. S. Scrutton, ChemCatChem 2010, 2, a switch in diastereoselectivity towards a slight preference for 892– 914. [11] D. Koszelewski, I. Lavandera, D. Clay, G. M. Guebitz, D. Rozzell, W. Kroutil, the cis product 3a (Table 1, entry 11). As described in other Angew. Chem. Int. Ed. 2008, 47 , 9337– 9340; Angew. Chem. 2008, 120, studies, such as the hydrolysis of a,a-disubstituted malonate 9477 –9480. esters,[24] the mechanisms of these solvent effects are currently [12] a) M. Hçhne, S. Schätzle, H. Jochens, K. Robins, U. T. Bornscheuer, Nat. not understood. Chem. Biol. 2010, 6, 807 –813; b) J.-S. Shin, B.-G. Kim, Biotechnol. Bioeng. 1997, 55 , 348– 358. Our results demonstrate the advantage of a cascade based [13] A. Nobili, F. Steffen-Munsberg, H. Kohls, I. Trentin, C. Schulzke, M. on the use of EREDs and ATAs for the successive introduction Hçhne, U. T. Bornscheuer, ChemCatChem 2015, 7, 757 –760.

ChemBioChem 2015, 16 , 1041 – 1045 www.chembiochem.org 1044  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Communications

[14] R. K. Kuipers, H.-J. Joosten, W. J. H. van Berkel, N. G. H. Leferink, E. Rooij- Bornscheuer, Biotechnol. Adv. 2015; DOI: 10.1016/j.biotechadv. en, E. Ittmann, F. van Zimmeren, H. Jochens, U. Bornscheuer, G. Vriend, 2014.12.012; b) F. Steffen-Munsberg, C. Vickers, A. Thontowi, S. Schätzle, V. A. P. Martins dos Santos, P. J. Schaap, Proteins Struct. Funct. Bioinf. T. Meinhardt, M. Svedendahl Humble, H. Land, P. Berglund, U. T. Born- 2010, 78 , 2101 –2113. scheuer, M. Hçhne, ChemCatChem 2013, 5, 154 –157. [15] R. Kourist, H. Jochens, S. Bartsch, R. Kuipers, S. K. Padhi, M. Gall, D. [21] a) S. Arseniyadis, A. Valleix, A. Wagner, C. Mioskowski, Angew. Chem. Int. Bçttcher, H.-J. Joosten, U. T. Bornscheuer, ChemBioChem 2010, 11 , Ed. 2004, 43 , 3314– 3317; Angew. Chem. 2004, 116, 3376 –3379; b) M. 1635 –1643. Node, D. Hashimoto, T. Katoh, S. Ochi, M. Ozeki, T. Watanabe, T. Kaji- [16] a) J.-S. Shin, B.-G. Kim, Biotechnol. Bioeng. 1999, 65 , 206– 211; b) K. S. moto, Org. Lett. 2008, 10 , 2653 –2656. Midelfort, R. Kumar, S. Han, M. J. Karmilowicz, K. McConnell, D. K. Gehl- [22] C. S. Fuchs, M. Hollauf, M. Meissner, R. C. Simon, T. Besset, J. N. H. Reek, haar, A. Mistry, J. S. Chang, M. Anderson, A. Villalobos, J. Minshull, S. Go- W. Riethorst, F. Zepeck, W. Kroutil, Adv. Synth. Catal. 2014, 356, 2257 – vindarajan, J. W. Wong, Protein Eng. Des. Sel. 2013, 26 , 25– 33; c) B.-K. 2265. Cho, H.-Y. Park, J.-H. Seo, J. Kim, T.-J. Kang, B.-S. Lee, B.-G. Kim, Biotech- [23] D. Clay, C. Winkler, G. Tasnµdi, K. Faber, Biotechnol. Lett. 2014, 36 , 1329 – nol. Bioeng. 2008, 99 , 275 –284. 1333. [17] E. Krieger, G. Koraimann, G. Vriend, Proteins Struct. Funct. Bioinf. 2002, [24] M. E. Smith, S. Banerjee, Y. Shi, M. Schmidt, U. T. Bornscheuer, D. S. Mas- 47 , 393– 402. terson, ChemCatChem 2012, 4, 472 –475. [18] C. Peters, R. Kçlzsch, M. Kadow, L. Skalden, F. Rudroff, M. D. Mihovilovic, U. T. Bornscheuer, ChemCatChem 2014, 6, 1021 –1027. [19] N. Nakajima, K. Tanizawa, H. Tanaka, K. Soda, J. Biotechnol. 1988, 8, 243– 248. [20] a) F. Steffen-Munsberg, C. Vickers, H. Kohls, H. Land, H. Mallin, A. Nobili, Received: February 10, 2015 L. Skalden, T. van den Bergh, H.-J. Joosten, P. Berglund, M. Hçhne, U. T. Published online on March 20, 2015

ChemBioChem 2015, 16 , 1041 – 1045 www.chembiochem.org 1045  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Supporting Information

Two Subtle Amino Acid Changes in a Transaminase Substantially Enhance or Invert Enantiopreference in Cascade Syntheses Lilly Skalden, [a] Christin Peters, [a] Jonathan Dickerhoff, [b] Alberto Nobili, [a] Henk-Jan Joosten, [c] Klaus Weisz, [b] Matthias Hçhne, [d] and Uwe T. Bornscheuer* [a] cbic_201500074_sm_miscellaneous_information.pdf !"##$%&'()*'(+$%,-&'$(* * Material and Methods Materials. All chemicals were purchased from Fluka (Buchs, Switzerland), Sigma (Steinheim, Germany), Merck (Darmstadt, Germany), VWR (Hannover, Germany), or Carl Roth (Karlsruhe, Germany) and were used without further purification unless otherwise specified. Polymerases were obtained from New England Biolabs GmbH (NEB, Beverly, MA, USA) and primers were ordered from Invitrogen (Life Technologies GmbH, Darmstadt, Germany). The HisTrap FF, the HiTrap Desalting and the GST Trap columns were ordered from GE Healthcare (Buckinghamshire, GB).

Bacterial strains and plasmids. E. coli TOP10 [F’lacIq, Tn10(TetR) mcrA D(mrr-hsdRMS-mcrBC) F80 LacZDM15 DlacX74 recA1 araD139 D(ara leu)7697 galU galK rpsL (StrR) endA1 nupG] was obtained from Invitrogen (Carlsbad, CA, USA). E. coli BL21 (DE3) [fhuA2 [lon] ompT gal (l DE3) [dcm] DhsdS] was purchased from New England Biolabs (Beverly, MA, USA). The plasmid pET24b bearing the gene encoding the amine transaminase from Vibrio fluvialis (UniProt: F2XBU9) was kindly provided by Prof. Byung Gee Kim (Seoul National University, South-Korea) and the plasmid pET26b bearing the gene encoding the enoate reductase ERED (UniProt: Q02899) was kindly supplied by Prof. John D. Stewart (University of Florida, USA).

Modeling. The modeling was performed with YASARA [1] using the default settings of the program. The polypeptide chains of the homodimer of the crystal structure of the ATA from Vibrio fluvialis (pdb- code: 4E3Q) were used. The external aldimine of 1-amino-3-methylcyclohexane was modeled into the crystal structure.

Cloning. All mutants were generated via MegaWhop PCR with specific primers for each mutant: Leu56Ile: 5’-CTCGGGCATTTGGAACATGGTCGCGGGC-3’ Leu56Val: 5’-CTCGGGCGTGTGGAACATGGTCGCGGGC-3’ Leu56Ala: 5’-CTCGGGCGCGTGGAACATGGTCGCGGGC-3’ Leu56Trp: 5’-CTCGGGCTGGTGGAACATGGTCGCGGGC-3’ Leu56Tyr: 5’-CTCGGGCTATTGGAACATGGTCGCGGGC-3’ Leu56Phe: 5’-CAACTCGGGCTTTTGGAACATG-3’. 4 µl of the plasmid was mixed with 5 µl Pfu+ buffer, 1.5 µl dNTPs, 0.5 µl Pfu+ polymerase, 1 µl specific primer for the specific mutant, 2 µl T7 forward primer and 35.5 µl dist. water. The PCR program included the following temperature steps: hold at 95°C for 5 min, afterwards 30 cycles of the following: hold at 95°C for 45 sec, 54.8°C for 45 sec, 72°C hold for 6 min. Finally hold at 72°C for 10 min. 8 µl of the PCR-product were mixed with 5 µl 10x Pfu+ buffer, 2 µl plasmid, 1 µl dNTPs, 1 µl Pfu+ polymerase and 33 µl dist . water. The following program was used: hold 68°C for 5 min, hold 95°C for 1 min, 10 cycles of the following 3 steps: 95°C for 30 sec, 53°C for 30 sec and 68°C for 7 min. Afterwards 10 cycles of the following 3 steps were made: 95°C hold for 30 sec, 55°C hold for 30 sec and 68°C hold for 11 min. To digest the wild-type (WT) plasmid, 25 µl of the MegaWhop-PCR product was mixed with 0.5 µl Dpn-I enzyme. After incubation at 37°C for 2h the enzyme was deactivated at 80°C for 10 min. 5 µl of the Dpn-I digested mutated vectors were transformed into E. coli Top10, sequenced by Eurofins Genomics (Ebersberg, Germany), and finally transformed into E. coli BL21 (DE3). The WT and all constructs contain a C-terminal His-tag to facilitate purification.

1 Cultivation and expression conditions. Bacteria were incubated on agar plates in a Heraeus Instruments FunctionLine incubator under air. All materials and biotransformation media were sterilized by autoclaving at 121°C for 20 min. Aqueous stock solutions were sterilized by filtration through 0.20 mm syringe filters. Agar plates were prepared with LB medium supplemented by 1.5% (w/v) agar. The genes encoding the amine transaminase from Vibrio fluvialis and its mutants were expressed in Escherichia coli BL21 (DE3) containing the expression vector pET24b, which encodes the sequence of the amine transaminase concluding an additional C-terminal His 6-tag. The cells were grown in 400 ml LB medium containing 0.05 mg ml -1 kanamycin in baffled Erlenmeyer flasks in orbital shakers

(InforsHT Multitron 2 Standard) at 37°C and 200 rpm until an OD 600 of 0.4 was reached. After that the temperature was reduced to the expression temperature of 20°C and the cells were further incubated until they reached an OD 600 of 0.7. Protein expression was induced with the addition of 0.5 mM IPTG. 20 hours after the induction the cells were harvested. The gene of the enoate reductase (ERED) was expressed in Escherichia coli BL21 (DE3) containing the expression vector pET26b, which encodes the sequence of the ERED including an additional N- terminal GST-tag. The cells were grown in 400 ml LB medium containing 0.05 mg ml -1 kanamycin in baffled Erlenmeyer flasks in orbital shakers (InforsHT Multitron 2 Standard) at 37°C and 200 rpm until an OD 600 of 0.7 was reached. After that the temperature was reduced to the expression temperature of 30°C. Protein expression was induced with the addition of 0.5 mM IPTG. The cells were harvested 8 hours after the induction. The overexpression of the generated mutants was analyzed via SDS-PAGE (10% acrylamide/bis-acrylamide as resolving gel and 4% acrylamide/bis-acrylamide as stacking gel). The samples were standardized based on the optical density.

200kDa 1 2 3 4 5 6 7 8 119 kDa

66 kDa

43 kDa

29 kDa

20 kDa

14 kDa Figure S1: SDS-PAGE of the expression of the mutant ATA-Leu56Ile, which is representative for all produced mutants. All mutants were overexpressed as soluble protein. Lines 1-4: samples at different time points of the crude extract (0, 1, 6 and 22 h after induction). Lines 5-6: samples of different time points of the soluble protein (0, 1, 6 and 22 h after induction).

Purification and desalting. The cell pellets containing the amine transaminases were resuspended in buffer A (50 mM sodium phosphate-buffer pH 7.5, 300 mM sodium chloride), which contained additional 0.1 mM PLP and 30 mM imidazole. Cell disruption was performed with a French Press at 1500 psi in two cycles. The resulting cell suspension was centrifuged for 45 min at 10,000 g. The filtrated supernatant was applied to a Nickel-NTA column (GE Healthcare). After washing the column with a triple volume of buffer A containing 0.1 mM PLP and 60 mM imidazole at a flow rate of 5 ml min -1, the protein was eluted with buffer A containing 0.1 mM PLP and 300 mM imidazole. The amine transaminase containing fractions were collected and pooled. The pooled protein was afterwards desalted by gel chromatography against 20 mM tricine buffer pH 7.5 and 0.01 mM PLP at a flow rate of 2 ml min -1.

2 The cell pellet containing the ERED was resuspended in purification buffer A (sodium phosphate buffer, 50 mM, pH 7.5), containing 1 mM DTT and 150 mM sodium chloride. Cell disruption was performed with a French Press at 2000 psi. The cell suspension was centrifuged for 45 min at 10,000 g. The filtrated supernatant was applied to a GST His-trap column (Health Care) with a flow of 0.5 ml min -1. After washing the column with 5 CV with buffer A by a flow of 4 ml min -1, the protein was eluted with Tris-HCl elution buffer (50 mM, pH 8) containing 1 mM DTT and 15 mM reduced glutathione. The protein containing fractions were pooled. The successful purification of all enzymes was analyzed via SDS-PAGE (10% acrylamide/bis-acrylamide as resolving gel and 4% acrylamide/bis-acrylamide as stacking gel).

200 kDa 119 kDa 1 2 3 4 5 66 kDa 6

43 kDa

29 kDa

20 kDa

14 kDa

Figure S2: SDS-PAGE of the purified amine transaminase from Vibrio fluvialis and its mutants ATA-Leu56Ile and ATA- Leu56Val. Line 1, 3, and 5: crude extract of the WT (1) and its mutants Leu56Ile (3) and Leu56Val (5). Lines 2, 4, and 6: purified enzyme of the WT (2) and its mutants ATA-Leu56Ile (4) and ATA-Leu56Val (6).

200 kDa 1 2 3 4 5 6 119 kDa 66 kDa

43 kDa

29 kDa

20 kDa

14 kDa

Figure S3: SDS-PAGE of the purification of the OYE via GST-tag . Line 1: crude extract, line 2: flow through, line 3: washing step, line 4-6: elution fractions.

Activity measurements. The activity of the amine transaminase from Vibrio fluvialis and its mutants was determined by the acetophenone-assay. [2] To 95 µl sodium phosphate buffer (50 mM, pH 7.5) and 100 µl acetophenone reaction solution (2.5 mM pyruvate and 2.5 mM ( S)-phenylethylamine in sodium phosphate buffer 50 mM, pH 7.5), 5 µl enzyme was added. The slope of the absorption was determined at 254 nm. The ERED activity was determined spectrophotometrically by the consumption of NADPH measured at 340 nm for 3 min. 968 µl sodium phosphate buffer (50 mM, pH 7.5) were mixed with 1µl N-ethylmaleimide, 1 µl NADPH and 30µl enzyme. [3]

Biocatalyses. Purified and desalted enzyme solution was used for all biocatalyses. The LDH and GDH were ordered from Sigma Aldrich (Germany). Different values of units were utilized: 0.5 U of the amine transaminases, 0.2 U of ERED, 0.9 U of GDH and 0.15 U of LDH. All reactions were incubated at 30°C at 750 rpm.

3 Biocatalysis with amine transaminases and the racemic substrate 2 The total volume was 1.0 ml. The biocatalyses contained the aforementioned enzymes, excluding the ERED, 200 mM L-alanine, 1 mM glucose, 10 mM substrate solution and 1.1 mM NADH dissolved in sodium phosphate buffer (50 mM, pH 7.5) including 0.1 mM PLP. The substrate solutions were prepared as 1 M stock solution in DMF or DMSO. To investigate the effect of organic co-solvents, the buffer was replaced with the respective solvent concentration.

Biocatalysis with the ERED The total volume was 0.5 ml. Biocatalysis was performed in sodium phosphate buffer (50 mM, pH 7.5) with 4 mM substrate, 1 mM glucose, 1.5 mM NADPH, ERED and GDH. The substrate solution was prepared as a 1 M stock solution in DMF or DMSO. The reaction was incubated at 30°C at 750 rpm.

Cascade reactions All aforementioned enzymes were used in the cascade reaction. The total volume of all cascade reactions was 1.5 ml. The biocatalyses contained in addition to the mentioned enzymes 200 mM L-alanine, 1 mM glucose, 10 mM substrate solution, 1.1 mM NADH and 1.5 mM NADPH dissolved in sodium phosphate buffer (50 mM, pH 7.5, 0.1 mM PLP). The substrate solutions were prepared as 1 M stock solutions in DMF or DMSO. From all biocatalyses, 300 µl samples were taken at different times and extracted with 300 µl ethylacetate and 250 µl hexane. The combined organic phases were dried with anhydrous sodium sulfate and analyzed via GC.

Preparative biocatalysis A preparative biocatalysis was performed for the cascade reaction with the mutant Leu56Ile and the ERED (Old Yellow Enzyme) with 1% DMF at a substrate concentration of 10 mM. The reaction conditions were the same as given above in a 50 ml reaction volume. At the end of the reaction, first concentrated HCl was added to adjust the pH to 2. Afterwards the aqueous phase was extracted three times with 1.5 volumes of dichloromethane. Then, pH was adjusted to 12 with NaOH solution followed by three times extraction with 1.5 volumes of dichloromethane. The combined organic phases were washed twice with 50 ml brine. Finally, the organic phase was dried with anhydrous sodium sulfate and residual solvent was removed by a rotary evaporator (isolated yield 77%).

GC-analysis. Two different GC-methods were developed. The composition of the produced diastereomers of the conversion of the racemic substrate 2 and the produced diastereomers of the cascade reaction were determinated by GC-analysis using method 1 after derivatisation. For this, 5 µl trifluoroacetic acid anhydride was added to the organic solvent, washed with 200 µl dist . water, dried with anhydrous sodium sulfate and evaporated with a nitrogen flow. Finally, the compound was dissolved in 50 µl dichloromethane and measured. Method 1: Macherey-Nagel γ-TBDAc (50 m, 0.25 mm); GC program parameters: injector 220°C, FID 220°C, pressure 94.2 kPa, flow 1.06 ml/min; 130°C/hold 5 min, 180°C/rate 1°C per min, hold 180°C for 5 min.

4 WT

"# "$% "&! "# !$% "&! "# !$% !&! "# "$% !&!

a) b)

ATA-Leu56Ile ATA-Leu56Val

c) d) Figure S4: GC-Chromatograms of the derivatisation of the product 1-amino-3-methylcyclohexane. Investigated were the product of the reaction of b) the wild-type and its the mutants c) ATA-Leu56Ile and d) ATA-Leu56Val with the racemic substrate 2 on the chiral column γ-TBDAc with GC method 1. The commercial standard of 1-amino-3-methylcyclohexane 3 (a) is shown too.

NH2 NH2 NH2 NH2

a) b) Figure S5: Chromatograms of the derivatisation of the cascade reaction with the enoate reductase OYE and the mutant (a) ATA-Leu56Ile and (b) ATA-Leu56Val . ! To simplify the measurement of the enantiomeric and diastereomeric excess of the produced diastereomers by the cascade reaction the GC methods 2 without derivatisation was used. The produced diastereomers were group into the cis - and trans -products on the BPX5 column. This separation allowed the measurement of the diastereomeric excess of the cascade reaction on the BPX5 column without derivatisation. Method 2: SGE BPX5 (25 m, 0.25 mm); GC program parameters: injector 250°C, FID 220°C, pressure 50.8 kPa, flow 0.81 ml/min; 60°C/hold 10 min, 180°C/rate 10°C per min, hold 180°C for 5 min. !

5 NH 2 NH2 + O O external standard NH2 NH2 +

!! Figure S6: GC-chromatogram of the product standard 1-amino-3-methylcyclohexane with the achiral column BPX5.

'(! )*+,-./*! )*+,-01/!

NH NH2 NH2 NH2 NH2 NH2 2

a) b) c) !!!!! Figure S7: Close-up of the GC-chromatograms of the cascade reaction of the OYE with (a) the wild type enzyme and its mutants (b) ATA-Leu56Ile and (c) ATA-Leu56Val.

NMR analysis. To determine the absolute configuration of the product, the samples were dissolved in 1 CDCl 3 with TMS added as an internal standard. H-NMR measurements were performed on a Bruker Avance 600 MHz spectrometer equipped with an inverse 1H/ 13 C/ 15 N/ 31 P quadruple resonance cryoprobe head and z-field gradients. DQF-COSY and one-dimensional NOESY experiments with selective refocusing were acquired at 293 K. For the latter, either the H1 resonance at 2.62 or 3.08 ppm was excited using a mixing time of 300 ms. 1 NMR spectroscopic data for the cis -product. H NMR (600 MHz, CDCl 3): δ (ppm) = 0.70 (q, J = 11.8 Hz; H2b), 0.77 (qd, J = 13.0 Hz, J = 3.9 Hz; H4b), 0.90 (d, J = 6.6 Hz; H7), 0.94 (m; H6b), 1.27 (qt, J = 13.3 Hz, J = 3.5 Hz; H5a), 1.41 (m; H3), 1.61 (m; H4a), 1.72 (dtt, J = 13.5 Hz, J = 6.6 Hz, J = 3.4 Hz; H5b), 1.80 (m; H2a), 1.81 (m; H6a), 2.62 ppm (tt, J = 11.2 Hz, J = 3.9 Hz; H1) 1 NMR spectroscopic data for the trans -product. H NMR (600 MHz, CDCl 3): δ (ppm) = 0.9 (d, J = 6.8 Hz; H7), 1.05 (m; H4a), 1.33 (m; H2a), 1.34 (m; H6b), 1.46 (m; H2b), 1.53 (m; H5a, H5b), 1.57 (m; H4b), 1.58 (m; H6a), 1.81 (m;H3), 3.08 ppm (m; H1) Conformational equilibria. The large vicinal coupling of >10 Hz for H1 indicates its axial orientation in the cis -product in line with the two substituents strongly favoring equatorial positions. In contrast, smaller H1 couplings for the trans -product suggest a rapid equilibrium between the two chair conformations with either the amino or the methyl substituent occupying an equatorial position.

6 Determination of the absolute configuration of the diastereomers. The enoate reductase OYE produces the same enantiomer as the enoate reductase XenB. Consequently the produced compound 2 of the cascade reaction is the ( S)-enantiomer. [4] Because of the (S)-configuration of the methyl group of 3-methylcyclohexanone, the ratio between the cis- and the trans-product of 1-amino-3- methylcyclohexane can be connected with the absolute configuration of the amino function.

O O

Figure S8: Comparison of the produced 3-methylcyclohexanones through the enoate reductase XenB (blue) and the enoate reductase OYE (magenta). The racemic mixture of product 3-methylcyclohexanone is coloured in black.

By performing the cascade reaction with the OYE and the ATA-Leu56Ile the absolute configuration of the amino function was solved. With an (S)-configuration at the methyl group the trans -product is the (1 S,3 S)-diastereomer whereas the cis -product is the (1 R,3 S)-diastereomer. The trans -product was preferred over the cis -product in the cascade reaction with the OYE and the ATA-Leu56Ile.

#$%&'( )*'(

Figure S9: Portion of the 1H-NMR product spectrum for the cascade reaction with substrate 1 and the OYE and the mutant Leu56Ile.

7 Solvent dependency of the diastereomeric composition. The addition of co-solvent shifted the ratio of the produced products. More co-solvent resulted in a stronger preference for the cis -products.

Figure S10: Influence of organic solvent on the cis /trans -ratio in the amination of rac -2 catalyzed by ATA-Vfl and the two mutants ATA-Leu56Ile and ATA-Leu56Val .

References

[1] E. Krieger, G. Koraimann, G. Vriend, Proteins: Struct. Funct. Bioinf. 2002 , 47 , 393-402. [2] S. Schätzle, M. Höhne, E. Redestad, K. Robins, U. Bornscheuer, Anal. Chem. 2009 , 81 , 8244 - 8248. [3] J. F. Chaparro-Riggers, T. A. Rogers, E. Vazquez-Figueroa, K. M. Polizzi, A. S. Bommarius, Adv. Synth. Catal. 2007 , 349 , 1521-1531. [4] C. Peters, R. Kölzsch, M. Kadow, L. Skalden, F. Rudroff, M. D. Mihovilovic, U. T. Bornscheuer, ChemCatChem 2014 , 6, 1021-1027.

!

8

Artikel V Biotechnology Advances 33 (2015) 566 –604

Contents lists available at ScienceDirect

Biotechnology Advances

journal homepage: www.elsevier.com/locate/biotechadv

Research review paper Bioinformatic analysis of a PLP-dependent enzyme superfamily suitable for biocatalytic applications

Fabian Steffen-Munsberg a,c, Clare Vickers a, Hannes Kohls a,b, Henrik Land c, Hendrik Mallin a, Alberto Nobili a, Lilly Skalden a, Tom van den Bergh d, Henk-Jan Joosten d, Per Berglund c, Matthias Höhne b,⁎, Uwe T. Bornscheuer a,⁎⁎ a Dept. of Biotechnology & Enzyme Catalysis, Institute of Biochemistry, Greifswald University, Felix-Hausdorff-Str. 4, 17487 Greifswald, Germany b Protein Biochemistry, Institute of Biochemistry, Greifswald University, Felix-Hausdorff-Str. 4, 17487 Greifswald, Germany c KTH Royal Institute of Technology, School of Biotechnology, Division of Industrial Biotechnology, AlbaNova University Center, SE-106 91 Stockholm, Sweden d Bio-Prodict, Nieuwe Marktstraat 54E, 6511 AA Nijmegen, The Netherlands article info abstract

Available online 7 January 2015 In this review we analyse structure/sequence –function relationships for the superfamily of PLP-dependent enzymes with special emphasis on class III transaminases. Amine transaminases are highly important for appli- Keywords: cations in biocatalysis in the synthesis of chiral amines. In addition, other enzyme activities such as racemases or Protein function decarboxylases are also discussed. The substrate scope and the ability to accept chemically different types of Annotation substrates are shown to be re flected in conserved patterns of amino acids around the active site. These findings PLP-dependent enzymes are condensed in a sequence –function matrix, which facilitates annotation and identi fication of biocatalytically Bioinformatics Biocatalysis relevant enzymes and protein engineering thereof. Enzyme discovery © 2015 Elsevier Inc. All rights reserved. Transaminase

Contents

1. Introduction ...... 567 1.1. Motivation and learning objectives ...... 567 1.2. How the review is structured and where do I findwhat?...... 567 1.3. The protein environment of PLP-dependent enzymes diversi fies reaction speci ficity ...... 568 1.4. PLP-dependent biocatalysts as a short cut for multistep chemical syntheses ...... 570 1.5. The ‘predicting function from sequence ’-problem: how analysis of sequence fingerprints of active site residues can provide functional insights 572 2. Analysing sequence –function relationships of PLP-dependent enzymes using 3DM — high quality alignments meet powerful analysis tools . . . . . 574 2.1. The PLP fold type I and ornithine transaminase-like (OrnTL) 3DM databases ...... 574 2.2. Special features of the ornithine TA-like family exemplify the structural flexibility of PLP-fold type I ...... 575 2.3. Reaction and substrate speci ficity determining residues revealed by correlated mutations analysis (CMA) ...... 576 2.4. The sequence –function matrix ...... 576 3. Activities represented in the ornithine TA-like database ...... 577 3.1. ω-Amino acid: α-ketoglutarate transaminases ...... 578 3.1.1. Dual substrate recognition: the glutamate switch ...... 581

Abbreviations: AA, amino acid; AAA, amino acid amide; 3AcOc, 3-acetyloctanal; AcOrn, N-acetylornithine; DAIB, D-aminoisobutyrate; ATA, amine transaminase; CoA βAA, coenzyme A

β-amino acid thioester; DABA, α,γ-diaminobutyrate; DAPA, 7,8-diaminopelargonic acid; DGD, 2,2-dialkylglycine decarboxylase; DTS, dethiobiotin synthase; fumB 1, fumonisin B 1; GABA,

γ-aminobutyrate; glyox, glyoxylate; GSAM, glutamate-1-semialdehyde aminomutase; Lys ε, lysine ε-amino group; HfumB1, hydrolysed fumonisin B 1; Orn, ornithine; KAPA, 7-keto-8- aminopelargonic acid; αKG, α-ketoglutarate; OrnTL DB, ornithine transaminase-like database; βAla, β-alanine; βPhe, β-phenylalanine; PLP, pyridoxal 5'-phosphate; PUT, putrescine; pyr, pyruvate; SAM, S-adenosylmethionine; SuOrn, N-succinylornithine; TA, transaminase; tau, taurine ⁎ Corresponding author. Tel: +49 3834 8622832; fax: +49 3834 86 794391. ⁎⁎ Corresponding author. Tel.: +49 3834 864367; fax: +49 3834 86 794367. E-mail addresses: [email protected] (M. Höhne), [email protected] (U.T. Bornscheuer).

http://dx.doi.org/10.1016/j.biotechadv.2014.12.012 0734-9750/© 2015 Elsevier Inc. All rights reserved. F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 567

3.1.2. Ornithine, acetylornithine and succinylornithine: α-ketoglutarate TAs ...... 583 3.1.3. Lysine- ε:α-ketoglutarate TAs ...... 583 3.1.4. γ-Aminobutyrate: α-ketoglutarate TAs ...... 584 3.1.5. Putrescine and cadaverine: α-ketoglutarate TAs ...... 586 3.1.6. 3-Acetyloctanal transaminase (PigE) ...... 586 3.1.7. 2-amino-4-oxobutyrate transaminases (diaminobutyrate TAs) ...... 587 3.2. ω-Amino acid:pyruvate transaminases ...... 587 3.2.1. Dual substrate recognition: the flipping arginine ...... 588 3.2.2. Natural function of amine transaminases ...... 588 3.2.3. Discriminating high and low activity amine transaminases and βAla:pyrTAs...... 588 3.2.4. Cadaverine/putrescine:pyruvate TAs ...... 590 3.2.5. Taurine:pyruvate TAs ...... 590 3.3. ω-Transaminases with unusual acceptor spectrum ...... 590 3.3.1. Dual substrate recognition ...... 591 3.3.2. β-Phenylalanine aminotransferases ...... 591 3.3.3. Acyl-CoA- β-TAs...... 592 3.3.4. D-p-hydroxyphenylglycine: αKGTAs ...... 592 3.3.5. Diamino pelargonic acid transaminases ...... 592 3.3.6. Alanine:glyoxylate transaminase 2 ...... 593 3.4. Glutamate-1-semialdehyde transaminases (2,1-amino mutases) ...... 594 3.5. Decarboxylation dependent TAs: the 2,2-dialkylglycine decarboxylases ...... 595 3.6. α-H-amino acid amide/ α-amino- ε-caprolactam racemases ...... 596 3.7. Isoleucine 2-epimerase ...... 597 3.8. Enzymes with unclear substrate recognition ...... 597 3.8.1. Neamine TAs, 2 ′-deamino-2 ′-hydroxyneamine and neomycin C TAs ...... 597

3.8.2. (Hydrolysed) fumonisin B 1 TAs...... 598 3.8.3. Phospholyases ...... 598 3.8.4. Multi-domain or non-enzymes ...... 598 4. Challenges for fingerprint-based sequence –function predictions ...... 599 4.1. Limitations of the active site amino acid fingerprint-based approach ...... 599 4.2. 3DM database related issues ...... 599 4.3. The literature mining problem ...... 599 4.4. The challenge to identify unknown speci ficities...... 600 5. Conclusion ...... 600 Author contributions ...... 600 Acknowledgements ...... 600 Appendix A. Supplementary data ...... 601 References ...... 601

1. Introduction 1.2. How the review is structured and where do I find what?

1.1. Motivation and learning objectives Some basic introduction about the diversity of PLP chemistry, PLP- dependent enzyme classi fication and the biotechnological relevance of What's the function of a certain gene or protein? Answering this ques- transaminases is given in the introductory sections 1.3 and 1.4. The sec- tion precisely is still a challenging, but very important task. An over- tion 1.5 introduces the active site fingerprint concept, which forms the whelming number of potentially interesting enzymes for biocatalysis are basis of our structure –function relationship analysis. The most impor- available in public protein databases. However, this resource is only par- tant terms and concepts of sections 1.3, 1.4 and 1.5, which are used tially useful, because often the function and properties of enzymes cannot throughout the review, are summarised in Boxes 1 and 2 . Section 2 con- be predicted reliably. This review exempli fies how structural knowledge denses all information from literature and our bioinformatic analysis: of enzymes and bioinformatics tools can be integrated to increase the pre- first, in section 2.1 we provide a brief description of the algorithms be- cision of function prediction. As an example, we analysed enzymes of the hind 3DM, the bioinformatics platform used for our analyses. General PLP fold type I superfamily with special focus on class III transaminases. structure and sequence features of the class III transaminase family With the help of the review, the reader should be able to: and speci ficity determining residues are analysed in sections 2.2 and 2.3. Section 2.4 presents the sequence –activity matrix, the central part • understand the fascinating mechanisms and features that govern re- of our analysis. It shows a correlation of the function of different pro- action and substrate speci ficity of PLP-fold type I enzymes, teins with amino acid patterns of a few active site residues ( fingerprint). • understand how bioinformatics tools and structural knowledge can be The most important structural details behind these analyses are pre- combined to study structure –function relationships, sented in section 3. In this section we aim to illustrate the artful mech- • understand how the enzymes' activities are re flected in small amino anisms and active site adaptations that facilitated the development of acid sequence fingerprints, 28 different enzyme activities. On the one hand, speci ficity is created • take a class III transaminase amino acid sequence and easily assign the by providing a binding pocket that is complementary to the substrate most probable function (out of 28 different known functions), in shape and polarity and provides electrostatic interactions. On the • apply this knowledge to guide experiments for the discovery of novel other hand, different mechanisms render the active site very flexible enzymes, and allow two or more chemically different substrates to bind in the • apply the guidelines and tools covered in this review to analyse other same pocket (so called dual substrate recognition). An overview of sec- enzyme superfamilies tion 3 is given by Table 3 , which contains structures of substrates and 568 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 products of all presented enzymes. To make understanding easy, we 1.3. The protein environment of PLP-dependent enzymes diversi fies reac- provide a PyMOL (Version 1.6.0.0) session file containing the aligned tion speci ficity structures shown in all figures as Supplementary PyMOL session to this review. This allows the reader to rapidly inspect all presented fig- Pyridoxal 5'-phosphate (PLP) is by far the most versatile cofactor ures in more detail and in comparison to the others. Section 4 contains enabling enzymes to catalyse an outstanding array of reactions includ- a detailed discussion about challenges and possible limitations of the ing transamination, decarboxylation, racemisation, elimination, substi- presented approach. tution and ring opening ( Eliot and Kirsch, 2004 ). The electron sink

Box 1 Important concepts.

Transaminase classification

Besides from the classification of PLP-dependent enzymes based on fold type, transaminases are additionally divided into six classes based on common structural features and sequence similarity ( Grishin et al., 1995 ) (for a summary of members within the classes see Table 3 in section 3). Amine transaminases belong to the class III transaminase family, which is also referred to as ornithine TA-like family. A suitable tool for de- termining a protein's family membership is InterPro ( Hunter et al., 2012 ), which combines several family and domain databases and is also ap- plied by the UniProtKB ( Magrane and UniProt Consortium, 2011 ). For determining the ‘aminotransferase class-III ’ family (IPR005814), InterPro combines the signatures of PANTHER ( Mi et al., 2013 ), Pfam ( Punta et al., 2012 ), PIRSF ( Nikolskaya et al., 2006 ) and PROSITE ( Sigrist et al., 2010 ). We define the class III transaminase family according to this family in the InterPro database. Note that this — initially PROSITE-based — terminology of class III transaminases differs from an earlier attempt for transaminase classification that referred to the ornithine TA-like family as class II ( Mehta et al., 1993 ).

Active site terminology

The active site architecture in PLP enzymes is often described relative to the cofactor. We will apply the re and si -face terminology introduced for transaminases to indicate the face relative to the cofactor's plane ( Soda et al., 2001 ). This term is derived from the protonation or deprotonation step of the C4 ’ of PLP relative to the cofactor's plane ( Fig. in Box 1 B ). In all PLP fold type I enzymes, the active site entrance is located at the re - face of the cofactor. To indicate the position of active site residues relative to the cofactor, the terminology introduced by Wybenga et al. (2012) will be applied. The side where the 3 ’-O of PLP is located is named the O-side and the other is termed the P-side, owing to the location of the phosphate group of PLP ( Fig. in Box 1 A ). A) B)

Figure in Box 1. Model of the quinonoid intermediate of alanine bound to the Vibrio fluvialis ATA (PDB ID: 4E3Q) to exemplify active site nomen- clature. The intermediate is shown in orange and the catalytic lysine in green. A) The P-side is coloured yellow and the O-side is coloured red. B) The protonation of the quinonoid intermediate is the chirality-introducing step in transamination. In this example the ( S)-enantiomer of alanine will be formed through protonation by the catalytic lysine, which is located at the si -face of the cofactor (relative to its C4 ’).

Active site fingerprint

We identified a set of 13 amino acids lining the active site that play an important role for determining substrate and reaction specificity in the different enzymes. We use the terms ‘active site fingerprint ’ and ‘active site pattern ’ throughout this review to refer to subsets of these residues that were found to be the most important for a certain specificity. Fig. 10 shows their location, Supplementary data Table S4 summarises all fingerprints and search results when applying those and the sequence –function matrix ( Table 2 ) enables a comparison of these sequence pat- terns between different enzymes. F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 569

Box 2 Definitions used in the 3DM-based creation and analysis of the superfamily database.

3DM-database 3DM databases comprise sequence alignments based on a superfamily wide structure alignment and thus offer a way of reliably comparing all sequences independent from low sequence similarity. For this review, a large PLP-fold type I database and a small ornithine TA-like database (OrnTL DB) were built. The latter offers a larger core (for details see Table 1 , and section 2.1).

Subfamily By the alignment of all available sequences to the structures of the initial structural alignment, so-called subfamilies are formed as the smallest building blocks of the superfamily alignment. Each subfamily is formed using one structure of the structural alignment as ‘template structure ’ and all sequences that could be aligned to it within a certain cut-off ( section 2.1, steps 4 –7). For the OrnTL DB the template structure's PDB code was applied to name each subfamily. Note: one subfamily can contain several proteins for which a structure is available, but these are se- quentially so similar (see cut-off step 4, section 2.1) that they are not used as a template for a new subfamily; it is also possible that a subfamily comprises enzymes with different activities. As long as there is a structure in the initial structure alignment with sufficient similarity to a se- quence of interest, this sequence can be compared to the whole database.

Core and variable regions To allow for facilitated comparisons and statistics, 3DM introduces the concepts of core and variable regions. The initial structure alignment is evaluated to find the common structural ‘core ’ that is conserved in all structures available in the database. Positions in these regions are called core positions while all structurally non-aligned areas are referred to as variable regions or positions. The larger the database (i.e. the more struc- tural variation), the smaller the core, as local differences in the structures caused by mutations, insertions or deletions prevent a meaningful alignment.

3D number A unified numbering scheme, called 3D numbers, is created for all sequences and structures in the 3DM database by renumbering them accord- ing to the core positions: all structural equivalent residues get the same number. This enables easy comparison of the amino acid distribution of all proteins in the database. All residue numbers used in this review are the 3D numbers of the OrnTL DB if not stated otherwise. Residue num- bers within variable regions are given in italics in combination with the accession number of the enzyme to which the numbers refer.

Correlated mutation analysis (CMA ) Structural or functionally important amino acid positions are often not conserved, but mutations at these positions are made possible by addi- tional mutation(s) within the protein. Therefore several positions, often randomly scattered over a protein's sequence, are commonly mutated together during evolution. Compared to conserved residues, these networks of correlated mutations are much more difficult to detect in a mul- tiple sequence alignment by visual inspection, but they can be identified and visualised by CMA. Hence, the CMA network analysis tool integrat- ed in the 3DM suite, called CorNet (publication in progress, free web based version available from www.3dm.bio-prodict.nl/Comulator ), is a powerful tool to reveal structural or functional important residues. The CMA algorithm, called Comulator, behind the CorNet tool was published previously ( Kuipers et al., 2009 ).

nature of the PLP cofactor allows for this vast variety of chemistry. It is of the reaction, the enzyme provides catalytic functional groups the enzyme scaffold, however, which elegantly determines which possessing a certain positional flexibility, which promotes the desired reaction pathway is followed. Toney (2011) recently reviewed the sub- bond formation(s). Enzymes achieve the above-mentioned tasks by an ject of reaction speci ficity and summarised three main themes. artful design of their active sites. For a deeper understanding of these 1) Stereoelectronic control: the enzyme forces the substrate/aldimine fascinating details we highly recommend to read the review by Toney intermediate to adopt a certain conformation. The bond to be cleaved (2011) . has to be aligned parallel to the π-orbitals (perpendicular to the PLP Besides controlling reaction speci ficity, a second important plane) for a π–σ-orbital overlap, facilitating bond cleavage and reso- issue is controlling substrate speci ficity and enantioselectivity. nance stabilisation of the developing charge by the conjugated For many PLP-dependent enzymes, this is a complex task, as π-electron system of PLP. This principle that explains the preference chemically different substrates have to be accepted (e.g. acidic for cleavage of one of the C α substituents over another was initially and aromatic amino acids). This phenomenon of multiple sub- published by Dunathan (1966) and is therefore known as Dunathan strate recognition is discussed in detail in sections 3.1.1, 3.2.1 principle (see Fig. 1 ). 2) The electrophilic strength of the Schiff base — and 3.3.1. pyridine ring π-electron system: this property governs the capability Altogether, the majority of PLP-dependent enzymes catalysing this of negative charge delocalisation. Different reactions, i.e. racemisation plethora of reactions have been evolved in a very small number of dif- or transamination, require a lower or higher degree of negative charge ferent tertiary structures: only seven different fold types of PLP- stabilisation, respectively. Thus, different enzymes affect the electro- dependent enzymes were discovered until now (Supplementary data philic strength of the aldimine intermediate by controlling the proton- Table S1) ( Percudani and Peracchi, 2009 ). ation state of the pyridoxyl N atom. Its protonation increases the The fold type I, also referred to as the ‘aspartate aminotransferase capability of resonance stabilisation and a quinonoid intermediate can superfamily ’, combines the highest quantity and diversity of mem- be formed. 3) Catalytic side chain placements: to control the outcome bers, compared to the other fold types ( Schneider et al., 2000 ). It 570 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604

Until today, a total of 236 chemical reactions are known to be

catalysed by PLP-dependent enzymes according to the B 6-database (Percudani and Peracchi, 2009 ). This highlights nature's flexibility to tailor enzymes towards a distinct substrate speci ficity. In addition to the already known activities, there are probably a number of additional un- known enzyme functions waiting to be discovered. This great diversity of enzyme reactions, the known and the unknown, is re flected in millions of amino acid sequences of proteins stored in protein databases. There- fore, protein databases represent a gold mine for biotechnologists. To identify the biocatalytically interesting among this diversity, however, Fig. 1. Stereoelectronic control of reaction speci ficity in PLP-dependent enzymes exempli- tools for sequence-based function prediction are required. fied by the alanine external aldimine of a ( S)-selective transaminase. By aligning the C α–H bond σ-orbital with the p-orbitals of the conjugated π-system, this bond is selectively weakened and transamination or racemisation is favoured over decarboxylation ( Toney, 1.4. PLP-dependent biocatalysts as a short cut for multistep chemical 2011). For the transamination dependent decarboxylases (section 3.5) a three subsites syntheses model of the active site has been proposed (A, B and C, highlighted in grey font). If the pro- ton is placed in subsite A, transamination will occur and if the carboxylate is placed in this subsite, the reaction will proceed via decarboxylation. The capability of catalysing various bond breaking and bond making steps in a coordinated and selective manner renders PLP-dependent en- zymes superior to chemical synthesis routes. Various PLP-dependent enzymes have been utilised in industrial applications, such as decarboxylases, racemases and more commonly, transaminases. For in- fi comprises enzymes belonging to all but one of the six classes de ned stance, L-aspartate-4-decarboxylase in whole Pseudomonas dacunhae – by the enzyme commission numbers (i.e. EC 1 5; Webb and cells was used for the production of L-alanine from L-aspartate and if International Union of Biochemistry and Molecular Biology, 1992 ) combined with aspartase, the production of L-alanine from fumarate is ‘ that are acting on a vast variety of substrates. Fold type II or trypto- possible in two biocatalytic steps ( Liese et al., 2006b ). ’ phan superfamily comprises alkyltransferases, ammonia lyases and The α-amino-ε-caprolactam (ACL) racemase from Achromobacter ‘ ’ some racemases while fold type III or alanine racemase superfamily obae together with an enantioselective L-lysine-1,6-lactam hydrolase contains amino acid racemases and decarboxylases. The second fold from Cryptococcus laurentii have been applied in a one pot whole cell ‘ class containing transaminases is fold type IV or D-alanine transam- biotransformation for L-lysine production from racemic α-amino- ε- ’ inase family which also includes a lyase. Fold types V, VI and VII caprolactam ( Fig. 3 , see also section 3.6). This biocatalytic step in combi- fi contain one reaction speci city each i.e. glycogen phosphorylases, nation with four chemical steps allowed for L-lysine production from cy- D-lysine-5,6-aminomutases and L-lysine-2,3-aminomutases, respec- clohexene (Liese et al., 2006a ). In another example, an Ala-racemase tively. This brief list of the various activities illustrates the versatility was used in combination with a D-amino acid transaminase (DATA) to of PLP chemistry that is enabled through the different protein con fig- enable the synthesis of D-amino acids. The D-Ala required for the DATA urations. A summary of the simpli fied mechanisms is given in Fig. 2 . was obtained from L-Ala using the Ala-racemase ( Soda and Esaki, 1994 ). For a computer animation of a simpli fied modelled amine transami- Further biocatalytic applications of PLP-dependent enzymes in- nase reaction see the Supplementary video. volved lyases. Tyrosine synthesis could be achieved from phenol,

Fig. 2. PLP-dependent enzymes catalyse a variety of chemical reactions by stabilising carbanionic intermediates, after the substrate formed a covalent aldimine intermediate with PLP. In many, but not all reactions, a quinonoid intermediate is formed during the reaction. In the centre of the figure, the most important intermediates observed during a transaminase reaction are shown. At each intermediate, there is a range of many possible reactions, as different bonds can be broken or formed, leading to distinct enzymatic activities. Possible bonds to be bro- ken are shown in different colours, bonds to be formed with differently coloured electron arrows. One enzyme activity is given as an example for each reaction, but in nature a much larger collection of activities exist. For a more detailed version of this figure, please see Supplementary data Figure S8. F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 571

Fig. 3. Biocatalytic process for L-lysine production from racemic α-amino- ε-caprolactam. As both enzymes have a comparable pH optimum, it was possible to run the whole cell process in one reactor.

Fig. 4. Transaminases are useful for the manufacture of amines. A) TA allow for amine synthesis in a single step, compared to traditional metal-catalysed chemical procedures like B) enamide reduction or C) imine reduction ( Nugent and El-Shazly, 2010 ).

pyruvate and ammonia with tyrosine phenol lyases from various organ- demonstrated for the large scale synthesis of the antidiabetic drug isms ( Lütke-Eversloh et al., 2007 ). If a tyrosine phenol lyase is combined Sitagliptin, where Merck & Co (USA) initially used an asymmetric trans- with the PLP-dependent tyrosine decarboxylase in a second step, dopa- fer hydrogenation process catalysed by a rhodium-complex, which was mine could be synthesised from catechol, pyruvate and ammonia ( Lee later replaced by a transaminase-catalysed route, in which simply et al., 1999 ). Furthermore, threonine aldolases have been applied in isopropylamine could be used as amino donor. The highly active and asymmetric aldol reactions to form α,α-dialkyl- α-amino acids from stable ( R)-selective amine transaminase was developed in collaboration Ala, Cys and Ser and various acceptor aldehydes ( Fesko et al., 2010 ). between Merck & Co with the company Codexis (USA) as published by Currently, the most useful and widely applied PLP-dependent en- Savile et al. (2010) . This improved process substantially reduced the zymes are transaminases (TA) as they can be used in biocatalytic asym- waste and E-factor of the process as summarised in a highlight article metric synthesis of amino acids and amines. Compared to organo- or (Desai, 2011). Both companies were awarded the ‘Presidential Green metallo-catalysis, transaminases are superior in step ef ficiency as they Chemistry Award USA ’ for this green chemistry route. For details catalyse several steps in a one-pot reaction: 1) reaction of the ketone about the application of transaminases in biocatalysis, readers are re- with a N-source (pyridoxamine 5'-phosphate, PMP) to form the imine, ferred to recent reviews ( Berglund et al., 2012; Höhne and 2) reduction to yield a protected amine, and 3) liberation of the free Bornscheuer, 2009, 2012; Kohls et al., 2014; Kroutil et al., 2013; Rudat amine by cleavage of the N-protecting group (PLP) (see Fig. 4 for a com- et al., 2012 ). parison to the chemical routes). 1 The enzymatic reaction substantially In nature, two types of PLP-dependent transaminases have been helps to increase process ef ficiency as intermediate work-up proce- discovered, according to the type of substrate that is converted: dures as well as toxic heavy metals can be avoided. This has been nicely α-transaminases ( α-TAs) and ω-transaminases ( ω-TAs). Whereas α-TAs (the majority of TAs) exclusively convert α-amino and α-keto acids, ω-TA also accept substrates having a distal carboxylic acid group 1 Interestingly, the intermediates and steps in the enzymatic reaction resemble those of instead of a carboxylate function in α-position. The term ω-TA is used the chemical synthesis, but most transaminase research overlooks these details and sim- ply considers this as amino group transfer — matching the classification in the enzyme to summarise a very heterogeneous group of activities (see sections commission class (EC) 2, transferases. 3.1, 3.2 and 3.3). Two subgroups of ω-TAs studied for biocatalytic 572 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 approaches are β-TAs (Rudat et al., 2012 ) and amine TAs (ATAs). The latter became particularly popular during the last decade (see Fig. 5 and legend for details).

1.5. The ‘predicting function from sequence ’-problem: how analysis of se- quence fingerprints of active site residues can provide functional insights

The ability to predict an enzyme's function based on its amino acid sequence is central for a variety of scienti fic disciplines:

• Physiology and system biology aim to understand an organism's (or habitat's) metabolic capability based on (meta)genomic data. With this capability they could predict ecophysiological functions or discover novel metabolic pathways. • For biotechnological applications, researchers aim to identify and use speci fic enzymes to catalyse a given reaction. Besides the targeted identi fication of enzymes, discovering novel enzymes is important to expand the enzymatic toolbox for biocatalysis. • Protein engineers wish to apply their understanding of structure – fi function relationships to the reverse direction: which (also by modi - Fig. 5. Common nomenclature of transaminases based on the distance of the transferred cation of an) amino acid sequence will generate the desired activity? amino group from the carboxylic function. A) α-Transaminases ( α-TA) catalyse the conversion of α-amino acids to the corresponding α-keto acid and vice versa. Note that An important aspect complicating sequence –function prediction is (S)- and ( R)-selective α-TA occur in nature: typical examples are aspartate transaminases (Asp: α-ketoglutarate TA) and D-amino acid transaminases (DATA). B) ω-Transaminases promiscuity, which is a consequence of evolution. In the case of PLP- (ω-TA) transfer amino groups that are more distant from a carboxylic group (e.g. in γ, δ dependent enzymes, Christen and Mehta (2001) proposed that PLP or ε position). Note that β-TAs are a subgroup of ω-TAs: these enzymes transaminate binding developed first, while evolution of reaction speci ficity preceded β-amino groups with respect to the acid function (n = 0). A typical example for β-TAs substrate speci ficity. There are often less genes encoding PLP- are β-phenylalanine TAs. Enzymes converting the ω-amino group of ω-amino acids dependent enzymes than PLP-catalysed reactions needed in an (e.g. γ-aminobutyrate (GABA), n = 1, R = H) and α,ω-diamino acids (e.g. Lys, n = 3, R = NH 2) are both referred to as ω-TAs. A subgroup of ω-TAs, the amine transaminases organism's metabolism ( Percudani and Peracchi, 2003 ). Therefore, it is (ATAs), also allow for the conversion of chiral amines independently from the presence not surprising that both reaction and especially substrate promiscuity of carboxylic groups in the substrate as exempli fied in C) these ATAs are very useful for (Bornscheuer and Kazlauskas, 2004; Hult and Berglund, 2007; Humble biocatalysis as they can be applied for asymmetric chiral amine synthesis from the corre- and Berglund, 2011) occur across all fold types of PLP-dependent en- sponding prochiral ketones if applied in reverse direction. Owing to their ability to convert -amino acids as well, the term ‘ -TA ’ has been used equally to the term ‘ATA ’ in biocatal- zymes. The fact that enzymes of very low sequence similarity can have ω ω ysis focussing publications. This terminology is misleading because there are several fi similar speci cities, while closely related enzymes will not accept the ω-TAs known that do not convert any amine substrate. As the activity towards amines is same substrates, makes reliable prediction of function with respect to the biocatalytically most relevant one, we prefer to term these enzymes ATAs to empha- uncharacterised enzymes challenging ( Percudani and Peracchi, 2003 ). sise their independency of carboxylic groups in the substrate and not to confuse with These dif ficulties result in error prone functional annotations of se- other ωAA converting enzymes, throughout this review. Both, ( R)- and ( S)-selective ATA have been found in nature. quences of PLP-dependent enzymes. While reaction speci ficity predic- tion is achieved with a relatively high success rate, the prediction of the detailed function at the substrate speci ficity level is complicated. For example, ATAs belong to a subfamily of PLP fold type I that is referred to as class III transaminase family ( Rausch et al., 2013 ) (see additional sequence –function connections that are gained by these stud- Box 1 ). This enzyme class harbours approximately 28 different enzyme ies can then be applied for further predictions and annotations. activities (see Table 3 ). If one was to BLASTP ( Altschul et al., 1997 ) the Even though computational protein function prediction is steadily sequence of the ATA from Ruegeria pomeroyi (PDB ID: 3HMU) advancing with algorithms, including literature mining and machine (Steffen-Munsberg et al., 2013b ) against the non-redundant protein se- learning, there is still a need for improvements: in the Critical Assess- quences database with pre-set parameters, one will obtain results that ment of Functional Annotation (CAFA) experiment, several methods do not allow any conclusion on its substrate speci ficity. Within the have recently been evaluated ( Radivojac et al., 2013 ). Unfortunately, first 100 results (all have 61 –100% sequence identity to 3HMU) four many of these methods are not yet available for standard large-scale an- sequences are annotated as ‘class III aminotransferases ’ (as is 3HMU it- notation projects. Radivojac et al. (2013) conclude that there is a need self), three are termed ‘adenosylmethionine-8-amino-7-oxononanoate for improving the availability of stand-alone tools to allow the predic- aminotransferase ’ and the remaining 93 are referred to as ‘aminotrans- tion of an enzyme's function independently from the slow updating ferase’, from which it is impossible to predict 3HMU's substrate scope rate of sequence databases according to the recent advances in annota- for small amino acids, such as pyruvate and γ-aminobutyrate or amines. tion technology. This example highlights that sequence similarity might not be suf ficient In our past research, we have focussed on investigating active site for a protein's function prediction. design/amino acid composition for analysis and prediction of enzyme The challenge of closing the gap between sequence and function in- function ( Gand et al., 2014; Höhne et al., 2010; Steffen-Munsberg formation has been subject to extensive research over recent years, and et al., 2013b ). The advantage we see in this approach is that predictions is addressed by bioinformatic annotation solutions ( Radivojac et al., are guided from hypotheses; it can be easily performed by all re- 2013 ), solving protein structures ( Jaskolski et al., 2014 ) and biochemical searchers and it deepens the understanding of structure –function rela- characterisation of single enzymes and combinations thereof. For in- tionships of a given superfamily of proteins. If structural information for stance, the Enzyme Function Initiative characterises proteins of un- the enzymes of interest is available, the structural alignment of only ac- known function structurally and functionally to systematically close tive site residues provides a powerful tool for sequence independent knowledge gaps to enable further predictions ( Gerlt et al., 2011 ), while function prediction in evolutionary distant enzymes. For instance ene- other groups predicted function by docking of metabolites to homology reductases that have completely different sequences and structural models and the evaluation of genetic contexts ( Zhao et al., 2013 ). The folds, but similar active site geometry, were recently shown to possess F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 573 comparable activity ( Steinkellner et al., 2014 ). Sequence independent fold types I –IV ( Catazaro et al., 2014 ). However, these approaches re- active site alignments can also be applied for functional comparisons quire detailed structural data for the target enzymes, which is often of structurally unrelated superfamilies and thereby reveal reaction not available. specificity determining features as recently demonstrated for the PLP Another, in many cases more applicable, strategy is to investigate se- quence –function relationships employing structural based sequence alignments. This approach was recently applied to identify selectivity

determining positions in P 450 monooxygenases and thiamine diphos- phate dependent decarboxylases, thereby enabling targeted mutagene- sis to improve the enzymes' selectivities ( Pleiss, 2014). Furthermore, A) Substrate 1 bound by X 100 , Y 200 , Z 300 discrimination between dehydrogenase and oxidase reaction speci ficity was shown to be determined by the presence of Ala or Gly in a single po- sition ( Leferink et al., 2009 ) and oxaloacetate hydrolyase activity within Z 400 the lyase/PEP mutase enzyme superfamily can be predicted based on the existence of a single Ser in the active site ( Joosten et al., 2008 ). A100 The importance of tools to predict an enzyme's function if only the sequences have been deposited in public databases but no biochemical

C300 characterisation or evidence of their function is available can be exem- pli fied by an example from our research: until 2010 ( R)-selective ATA

B200 activity was only found in two wild type strains, but the sequences of the responsible enzymes had been unknown. By analysing the determi- nants for substrate and reaction speci ficity in PLP fold type IV enzymes, we identi fied sequence motifs that allowed for the rapid annotation B) Substrate 1 bound by X 150 , Y 250 , Z 300 of all known enzymes of this fold type (i.e. branched chain amino acid transaminases (BCAT), D-amino acid transaminases (DATA) and Z400 4-amino-4-deoxychorismate lyases). Then, we predicted key mutations that should facilitate amine conversion in the scaffold of a BCAT to achieve patterns for ( R)-ATA prediction ( Höhne et al., 2010 ). Through A fi fi 150 C these ngerprints we identi ed 17 ( R)-ATAs that had been deposited 300 in the sequence databases, but with misleading annotations. Further ex- periments revealed that these new ( R)-ATAs have a great potential in

B250 asymmetric amine synthesis ( Schätzle et al., 2011 ). This example indi- cates how understanding of substrate and reaction speci ficity- determining residues can result in the identi fication of new versatile biocatalysts. C) Substrate 2 bound by X 100 , Y 250 , Z 300 Whereas information about ( R)-ATA (found in PLP fold type IV) was scarce until our study, ( S)-selective ATAs (belonging to fold type I) have

Z400 been explored for biocatalytic applications for more than 15 years ( Shin and Kim, 1997 ), the first sequence being described in 2003 ( Shin et al., D 2003). However, the methods for ( S)-selective amine transaminases 100 discovery had been restricted to enrichment cultures ( Shin et al.,

C300 2003 ) and sequence homology searches ( E.S Park et al., 2010 ) until 2011 when Park et al. proposed substrate speci ficity determining resi- dues based on homology models ( E.S. Park et al., 2011 ). In the following years, solved crystal structures ( Humble et al., 2012; Midelfort et al., E 250 2013; Rausch et al., 2013; Sayer et al., 2013 ) displayed the spatial ar- rangement of these residues in the active site. Interestingly, four crystal structures of ( S)-ATAs were deposited in the database since 2009, but D) Sequence alignment guides annotation not recognised as ATAs because of the lacking experimental data for these enzymes. Their detailed characterisation combined with a muta- genesis study ( Steffen-Munsberg et al., 2013a ) unravelled the detailed mechanism of dual substrate speci ficity and factors affecting catalytic ef ficiency towards amines. All these investigations strongly focused on (S)-ATAs without discussing the important residues in related enzymes,

Fig. 6. Active site patterns can be used to predict enzyme function. Substrate and reaction specificity of enzymes are governed by the presence of key residues in the active site, which can be detected in a multiple sequence alignment. A) & B) The same substrate 1 can be converted in enzymes by placing similar residues, but at different positions in the amino acid sequence. Because of the flexibility of the amino acid side chains, important functional groups might be in a similar geometric position. Therefore, important residues are not conserved in all cases. C) Chemically different residues realise conversion of a dif- ferent substrate 2. D) In a multiple sequence alignment, sequences matching the pattern of Enzyme A) or B) can be identi fied. Sequences 6 –8 have different pattern, and thus might fi Specificity determining Catalytic Probable have different substrate speci cities. Sequence 9 differs in the catalytic residue. Therefore amino acids residue it is either not catalytically competent, or catalyses a different reaction. Conserved amino Substrate acids are shown in different colours. 574 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604

Table 1 with an analysis of respective multiple sequence alignments, we show Comparison of the size of the PLP fold type I database and the OrnTL database. that activities can indeed correlate with a few active site residues, Amount of PLP fold type I OrnTL which enables a more detailed annotation compared to standard tools which are available online. Furthermore, we elucidate, how conserved Crystal structures found initially 717 170 Crystal structures in final alignment 406 170 these sequence motifs are, and which fraction of sequences cannot be Subfamilies 94 21 annotated by this approach at the moment and hence might be interest- Sequences found initially 120,870 31,000 ing for further research. Sequences aligned 42,080 12,956 Core residues 290 379 2. Analysing sequence –function relationships of PLP-dependent enzymes using 3DM — high quality alignments meet powerful which would be required to broaden the insight into mechanisms of analysis tools substrate recognition and catalysis within the whole PLP fold type I. From these results we hypothesise that reaction speci ficity as well as 2.1. The PLP fold type I and ornithine transaminase-like (OrnTL) 3DM substrate speci ficity is mainly re flected by the presence of certain active databases site residues, which interact with the substrate or the cofactor during the reaction. This pattern of active site residues also referred to as ‘active To identify and compare residues that govern substrate and reaction site fingerprint’ can be used to assign a function from a simple align- speci ficity in an enzyme superfamily, multiple structure and sequence ment, if the sequence matches the known pattern ( Fig. 6 ). It is impor- alignments have to be computed to generate an overall alignment. As tant to keep in mind that the situation in nature is more complex: they are the basis of all further analysis steps, their quality is of extraor- there might exist more than one solution to realise the same substrate dinary importance. speci ficity (as shown in Fig. 6 A, B and D, sequences 1 –4), and especially The commercial software 3DM, a protein superfamily analysis suite, PLP-dependent enzymes bind more than one substrate in the same ac- employs highly sophisticated algorithms to ensure that the alignment is tive site (for dual substrate recognition see sections 3.1.1, 3.2.1 and reliable, and at the same time integrates tools that allow one to generate 3.3.1). hypothesis of relevant structure/sequence –function relationships. In the following, we analyse important residues of well-described Many different data types are collected for all the proteins of a super- enzymes of the class III transaminases of PLP-fold type I. Together family in a 3DM system by extensive data collection tools (i.e. all

AcOrn: αKG TAs 1.0

Orn: αKG TAs

DGD

Lys ε:αKG TAs DAPA TAs

GABA: αKG TAs

Not characterised

ATAs βAla:pyr TAs GABA:pyr TAs

βPhe: αKG/pyr TAs Tau:pyr TAs GSAM αAAA racemases

Ala:glyox TAs 2

Fig. 7. Phylogenetic tree comprising all 12,956 sequences of the OrnTL DB. Colouring highlights the characterised enzymes. For each group of substrate/reaction speci ficity the whole in- duced network is highlighted. Colouring: grey: not characterised; yellow: GABA: αKG TAs; red: Lys ε:αKG TAs; light orange: Orn: αKG TAs; dark orange: AcOrn/SuOrn: αKG TAs; dark green: DGD; light blue: DAPA TAs; pink: ATAs, βAla:pyr TAs and GABA:pyr TAs; black: Tau:pyr TAs; brown: Ala:glyox TAs 2; dark blue: αAAA racemases; light green: GSAM; violet: βPhe: αKG/pyr TAs. Abbreviations are explained in Table 3 . The unrooted tree was calculated based on the core alignment of the OrnTL DB using FastTree 2.1.3 ( Price et al., 2010 ) and formatted using Dendroscope (Version 3.2.10) ( Huson and Scornavacca, 2012 ). F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 575 available structures are examined and full-text articles are text-mined PLP fold type I database due to structural diversity in the superfamily. to extract mutation studies from the literature; Kuipers et al., 2010a ). This protein fold was extremely adaptive during evolution, thus Besides the description of the 3DM database on PLP-dependent en- allowing a large quantity of reactions to be catalysed. This database cov- zymes and the sub-database on ornithine TA like (OrnTL) enzymes in ering a large fraction of PLP fold type I enzymes might be employed for this section, the most important concepts of 3DM relevant for this re- fold type overarching questions, but more detailed investigations like view are brie fly summarised in Box 2 . substrate speci ficity require a smaller ‘sub-database ’ comprising The PLP fold type I 3DM database was generated with the human structurally more related enzymes. We therefore created a smaller ornithine: α-ketoglutarate transaminase (Orn: αKG TA) crystal structure 3DM database, containing only sequences that belong to the class III as the starting template for the structural alignment. This enzyme was transaminases, also referred to as the ‘ornithine transaminase-like ’ fam- chosen as the template because the biocatalytically most relevant en- ily (see Box 1 , compare Table 3 ). This yielded the ornithine TA-like data- zymes of this fold type (( S)-ATAs) belong to the class III transaminase base (OrnTL DB), comprising 21 subfamilies with a much larger core family, which are also referred to as ornithine transaminase-like family compared to the large PLP fold type I database (see Table 1 for a compar- (see Box 1 ). The method of the generation of 3DM databases is, in eight ison of the two databases, see Fig. 7 for a phylogenetic tree of the whole steps, brie fly described below. Further details were published else- OrnTL DB). Almost all active site residues of class III TAs were covered in where ( Kourist et al., 2010; Kuipers et al., 2010b ). the core and 81% of all residues in the ( S)-ATA from Chromobacterium violaceum (Cvi-ATA) belong to core regions in the OrnTL DB compared 1) 3DM collects and superimposes all structures that share a common to 62% in the larger PLP fold type I database (see Supplementary data structural fold with the starting template structure (717 structures Figure S2 and S3 for more details). Besides the subfamily selection, in the PLP fold type I superfamily). Structures that are dif ficult to su- this database was generated with the same settings as the full PLP fold perimpose on the template — therefore resulting in only a small type I database. number of superimposed residues — are discarded. The default 3DM was able to generate a high quality alignment that consists of cut-off is at least 60 residues that are within a sphere of 2.5 Å from 12,956 protein sequences from the 31,000 sequences collected by the the equivalent template residues leaving 406 structures in the PLP initial BLAST searches. The discrepancy between the amount of collect- fold type I database. ed class III TA sequences and the sequences present in the OrnTL DB 2) A structural alignment is generated for each combination of two re flects the portion of sequence space with insuf ficient structural structures in an all-to-all comparison. The resulting structure align- information. This demonstrates the need of further research to enhance ments are merged into one large alignment that represents the the structural coverage of the class III TA family. structurally conserved core of the superfamily. Using this 3DM OrnTL DB we aim to extend recent studies, which 3) The positions in this core alignment are numbered (called 3D either used sequence similarity to classify ATA related enzymes and numbers) such that all structural equivalent residues in the 3DM only investigated ATAs' functional residues in detail ( Rausch et al., system have the same number. 2013 ) or focused only on a limited number of residues in the first 4) 3DM generates subfamilies by grouping all structures that are within shell of the active site and was therefore limited to enzymes with struc- a de fined sequence identity cut-off (in the PLP fold type I database tural information ( Catazaro et al., 2014 ). In this review we summarise this cut-off was set to 50% resulting in 94 distinct subfamilies). recent literature and our findings concerning substrate and reaction 5) From each subfamily the structure that structurally matches best speci ficity determinants within PLP fold type I enzymes with special with the starting template, thereby maximising the number of core focus on the class III transaminase family. residues in each subfamily, is chosen as the representative subfamily template (see Supplementary data Table S2 for a list of enzyme ac- 2.2. Special features of the ornithine TA-like family exemplify the structural tivities represented by the subfamilies' template structures). flexibility of PLP-fold type I 6) The subfamily templates are used in a UniProt BLAST search to col- lect protein sequences for which no structures are available The PLP-fold type I is an interesting example how very distantly (120,870 proteins were collected for the PLP fold type I database). related sequences still form similar tertiary structures. The diversity of 7) For each subfamily template a pro file based iterative alignment is enzyme activities within PLP fold type I is re flected in extensive adapta- performed (this is beyond the scope of this article and published tions at the amino acid sequence level: for example, aromatic amino elsewhere; Kuipers et al., 2010a; Kuipers et al., 2010b ). The inclusion acid transaminase (PDB ID: 1AY4) and ornithine aminotransferase cut-off for aligned sequences is 30% sequence identity compared to (PDB ID: 2OAT) share only 7% sequence identity. Within the OrnTL da- the last pro file used in this alignment procedure, which is approxi- tabase low sequence identities down to 13% 2 are observed. Interesting- mately 20% identical to the starting subfamily template. Aligning se- ly, the backbone arrangement of the majority of the active site residues quences of this low sequence similarity with high quality is dif ficult, is still very similar. From an inspection of the structure alignments, we but a high quality alignment is established by a quality control identi fied several features that are conserved on the structure or se- mechanism described by Kuipers et al. (2010b) . To ensure high qual- quence level in the OrnTL family, but differ in other fold type I enzymes. ity alignments, the sequences for which no evident solution can be Four regions are especially conserved among class III transaminases determined are simply deleted from the alignment. Although (see Fig. 8 for secondary structure elements numbering): many aligned sequences are removed this way, this method still al- lows for the generation of very large superfamily alignments. In the – case of the PLP fold type I database, the final alignment generated 1) A small antiparallel β-sheet (residues 23 36, comprising strands β1, ‘ ’ from the 120,870 sequences collected in step 6 contains 42,080 se- β2 and β3) close to the N-terminus on top of the small domain is quences. conserved in most structures of the OrnTL family. This region is lo- 8) The 3D numbering scheme is applied to all aligned sequences, which cated at the domain interface and we speculate that this sheet con- connects all sequences, all structures, and all alignments from the tributes to the suppressed domain movements during substrate different subfamilies to each other. binding, which is, in contrast to other fold type I transaminases (McPhalen et al., 1992 ), commonly observed in this family ( Cha The size of the conserved core is determined by the structural diver- sity in the superfamily: the more slightly different enzymes are includ- ed, the smaller are the structurally conserved regions that make up the 2 Pairwise identity within core regions of the sequences with UniProt IDs G9N4G9 and core. Not all active site residues are covered in the conserved core of the B8MF32 belonging to subfamilies 4A0G and 4AO9, respectively. 576 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604

Table S3). Of particular interest is that the catalytic lysine is present in this list: in the majority of sequences, which are not in the OrnTL family, the catalytic lysine is found two positions later at position 244 (for a dis- cussion see section 4.1). Most other positions in this list are not directly explained by their functional role, which is not surprising, as the OrnTL DB comprises a variety of substrate and reaction speci ficities that de- mand different recognition mechanisms. The only amino acid that is conserved over the whole fold type I database is D213, which is the res- idue responsible for protonating the pyridine nitrogen of the cofactor.

2.3. Reaction and substrate speci ficity determining residues revealed by correlated mutations analysis (CMA)

Evaluating sequence conservation is a fast way to identify positions that are probably relevant for catalysis or protein architecture. However, this strategy will only display a fraction of important residues. Often, more than one solution (amino acid composition) exists to realise a function in a protein. For example, in the case of a salt bridge at a do- main interface that is required for protein stabilisation, it is probably not important on which of the two domains the acidic or basic residue is localised. Thus, these two positions might not be conserved but mu- tated simultaneously. The degree to which mutations occur in a corre- lated fashion depends on the strength of the selection pressure, which fi Fig. 8. Topology plot of the class III transaminases exempli ed on chain A of the human is related to the functional relevance of these positions. Halabi et al. Orn: αKG TA structure (PDB ID: 2OAT). The regions with speci fically conserved secondary structure in class III TAs are highlighted red (see main text for more information). Con- (2009) demonstrated the power of such analyses using proteases: served residues among the whole family are highlighted in grey: the aspartate ‘on top ’ clusters of amino acids (called protein sectors) relevant for substrate of the left-handed helix α2 (D41), the aspartate ‘below ’ PLP, coordinating its pyridine ni- speci ficity, thermostability and catalysis could be discovered from cor- trogen (D213) and the catalytic lysine (K242). related mutations without looking at crystal structures. Thus, correlated mutations analysis (CMA) is a powerful statistical evaluation tool (Kuipers et al., 2009; Kuipers et al., 2010b ). The informative value of a CMA depends on the size, composition and quality of the multiple se- quence alignment. The OrnTL DB mainly contains amino acid sequences of transaminases with different substrate speci ficities. As ex- et al., 2014; Käck et al., 1999; Liu et al., 2004; Wybenga et al., 2012 ). pected, the network that resulted from a CMA on the OrnTL DB ( Fig. 9 ) 2) Another region that might prevent OrnTL enzymes' domain closure includes 10 active site positions that are important for substrate recog- is located ‘below ’ the small domain at the domain interface (residues nition (we found 13 key amino acids during our literature research and 218 –234, including the helix α10). structure inspection, see section 3; for a picture of the human Orn: αKG 3) All active sites (of available holo WT structures) comprise one turn of TA highlighting the active site residues, see Fig. 10 ). a left-handed helix ( α2) at the si -face of PLP (residues 44 –47), as was Residues 132, 185, 216, and 353 and residues 47 and 346 are part of previously described for Orn: αKG TAs, γ-aminobutyrate:αKG TAs the CMA network; they are key positions determining amino acceptor (GABA:αKG TAs), and S-adenosylmethionine:7-keto-8-amino- speci ficity in the transaminases ( α-ketoglutarate versus pyruvate). pelargonic acid TAs (SAM:KAPA TAs) (Novotny and Kleywegt, The majority of the sequences of the OrnTL DB are transaminases with 2005 ). The in fluence of this exceptional secondary structure element at least 28 signi ficantly different substrate specificities. As more than for substrate recognition was shown for the Escherichia coli SAM: one amino acid residue is necessary to create a distinct substrate speci- KAPA TA. The R221A mutation inverted this region to a right handed ficity, it makes sense that these residues mutated simultaneously. This α-helix which strongly declined the catalytic ef ficiency for SAM clearly demonstrates the potential of CMA. The CMA also detected resi- (Sandmark et al., 2004 ). An arginine coordinating the left-handed dues involved in cofactor binding, at the dimer interface, but also resi- helix at the O-side is conserved in 80.1% of the sequences in the dues whose relevance is not yet known. OrnTL family (R221 might be substituted by a lysine or R220 e.g. in 2GSA or K243 e.g. in 1OHV). The helix is additionally stabilised by a 2.4. The sequence –function matrix highly conserved D41 (97.9% of all sequences in the OrnTL DB, see Supplementary data Table S3), which caps the N-terminus of the From the 3DM database statistics, structure inspection and literature helix. These two conserved residues and the left-handed helix to- research concerning each substrate and reaction speci ficity, we con- gether are speci fic for the OrnTL family and are not found in other structed a sequence –function matrix. This is the centrepiece of this re- fold type I enzymes. view: the matrix summarises 13 active site residues within the OrnTL 4) A second OrnTL speci fic region in the active site is located at the family that determine reaction or substrate speci ficity ( Table 2 ). As cer- O-side (residues 267 –272). This loop contains the conserved T/S271 tain patterns are unique for each enzyme activity, we suggest that these that is involved in PLP binding ( Humble et al., 2012 ) and might addi- active site fingerprints can be used to predict the main activity of a given tionally be involved in catalysis by hydrogen bonding the catalytic sequence belonging to the OrnTL family. Not all 13 positions are equally K242 as it was shown for a conserved cysteine in a comparable posi- relevant for each enzyme. The most important residues for each tion in ornithine decarboxylases ( Oliveira et al., 2011 ). This region speci ficity are shown in bold. In the speci ficity-dedicated subsections, substantially differs in the whole fold type I database and therefore we describe the details of these sequence –function relationships at a belongs to the variable region there, whereas it could be properly molecular level. aligned in the proteins of the OrnTL DB. We constructed the matrix in the following way: for each activity, Several amino acid positions were identi fied as conserved and we built small alignments containing only enzymes with experimental- unique within proteins of the OrnTL family (see supplementary data ly con firmed activities. From crystal structures and literature, we F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 577

M PLP stacking

c M M c c c c M M M c c c c dimer M M interface

Fig. 9. CMA of the OrnTL database detects a network of key residues responsible for substrate speci ficity. A 0.8 cut-off for the correlation score was applied (see Kuipers et al. (2009) for an explanation of this score). The strength of the correlation of two positions is indicated by the colour of the connecting lines (red: 1, yellow: 0.8). Residues, which we collected from literature as key residues for determining substrate and reaction speci ficity are marked with an M (as they are part of the sequence –function matrix Table 2 ). To indicate that correlating residues are in contact, the lines are labelled with ‘c’.

extracted the relevant residues and then identi fied all sequences of the the information stored in the matrix to address the following OrnTL DB carrying these minimal sequence patterns. Finally, we used questions: the larger alignment derived from the sequence pattern search to ex- tract the most frequently observed amino acids at the other matrix po- Question Solution sitions for each speci ficity. fi 1. What are the main In an alignment, compare the active site residues of the Owing to the lack of crystal structures with suf cient sequence sim- substrates of a given query sequence to the fingerprints of the matrix. In case ilarity, several class III transaminases have not been included in the protein belonging to of a match, it is likely that also the activities are matching. OrnTL DB (see Table 3 entries with a minus ( −) sign in the ‘subfamily ’ this superfamily? For convenience, a multiple sequence alignment of all column). Most of these enzymes could be aligned to their closest struc- subfamilies' parental structures of the OrnTL DB is available (Supplementary data Figure S1) that helps to ture and sequence in the 3DM database manually and thereby also be match the 3DM numbers of this review with the original compared to the other class III TAs. The results from these manual align- numbering of the crystal structures and therefore also to ments are discussed in the corresponding sections and are also the query sequence. displayed in the sequence –function matrix (see Table 2 ). 2. How can I identify Usually, a BLAST with a query sequence is conducted to Further research is needed to investigate to which extent the se- novel enzymes identify enzymes with similar/equal function. We – having the desired recommend to align available sequences and check, quence function relationships can be generalised. We suggest using activity present in whether they carry the specificity determining residues the superfamily? highlighted in the matrix. This helps especially to evaluate distantly related sequences which otherwise would often not be chosen as candidate because the outcome would be too uncertain. 3. Does the superfamily Enzymes whose active site residues do not fit the pattern contain novel enzyme of the matrix have either an unknown activity, or they activities, which are might contain an alternative active site design to confer not yet known? a known speci ficity. Orn: KG TA, 2OAT 41 16

346 3. Activities represented in the ornithine TA-like database 73 45 347 352 The class III transaminase family represents only a small fraction of known sequences within the PLP fold class I. Nevertheless, it contains 271 353 268 270 348 351 at least 28 distinct enzymatic activities (not all have a separate EC num- 46 349 ber assigned, yet). This section aims to summarise structural details that 47 govern these substrate and reaction speci ficities. 242 A table of contents of this section's subsections is given in Table 3 , which provides an overview of enzyme activities by providing struc- 132 185 tures of their substrates and products, as well as a list of abbreviations. 216 We will discuss these enzymes in the order mentioned in Table 3 . The 215 entries in the sequence –function matrix and the scenes in the Supple-

129 mentary PyMOL session are in the same order to facilitate a convenient comparison of active site fingerprints and their underlying structural features. Enzymes displaying transaminase activities will be described fi 213 rst, followed by a smaller group of enzymes with other reaction spec- ificities, namely decarboxylation dependent transaminases, 1,2-amino mutases and α-amino acid amide ( αAAA) racemases. Besides these Fig. 10. Active site residues and 3DM numbering of the Orn TA-like database highlighted in well-investigated enzymes, we mention fumonisin B TAs, aminosugar the human Orn: αKG TA (PDB ID: 2OAT). Core regions are shown as grey, variable regions 1 as yellow cartoon. The PLP bound ornithine mimicking analogue is shown in orange. This TAs, phospholyases and multi-domain enzymes grouped together at and all other crystal structure figures were created using PyMOL (Version 1.6.0.0) . the end: due to a lack of structural information, the substrate or reaction 578 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 speci ficity determining residues of these enzymes cannot be predicted 3.2) and the third is converting either both αKG and pyr or other amino from the sequence alignments at the moment. A list of sequences, acceptors (section 3.3). As all ω-TAs convert structurally and chemically which have been experimentally characterised, is given in Supplemen- different amino donors and acceptors in the same active site, they all tary data Table S5. require a mechanism for dual substrate recognition. Within the OrnTL The majority of characterised enzymes found in the OrnTL database database three different solutions for this task have been found, can be classi fied as ω-transaminases. Whereas α-transaminases transfer which are each discussed in the beginning of the sections dedicated the amino group at the C α position with respect to the carboxylate to the three amino acceptors (sections 3.1.1, 3.2.1 and 3.3.1). The group, ω-transaminases can accept amino acids where the position of alanine:glyoxylate transaminases 2 (Ala:glyox TA 2) are the only the amino group with respect to the carboxylate group varies (see characterised transaminases within the OrnTL DB that do not require Fig. 5 ). These enzymes can further be clustered into three groups a dual substrate recognition because only α- and βAA are converted based on the preferred amino acceptor. The first group is accepting α- (section 3.3.6). The second known enzyme that only converts αAA, ketoglutarate ( αKG) (section 3.1), the second pyruvate (pyr) (section D-p-hydroxyphenylglycine: αKG TA, however, requires dual sub- strate recognition because its substrate and product have inversed fi D L Table 2 absolute con guration ( -p-hydroxyphenylglycine and -Glu, sec- Sequence –function matrix: overview of the function determining positions in correlation tion 3.3.4). to reaction and substrate speci ficities. Residues in bold are the fingerprint residues that de- Among the ω-transaminases, a few have been found to additionally termine reaction and substrate specificity (as retrieved from mutagenesis or crystallo- convert amines independently from the presence of carboxylic groups fi fi graphic studies). If not suf cient information was available to assign a ngerprint for a in the substrate. Those enzymes, referred to as amine transaminases certain speci ficity, no residue is bold. The colour code indicates the physicochemical prop- erties of the residues. The degree of conservation is indicated by the following notations: (ATAs), are the biocatalytically most interesting enzymes within this capital letters — conserved residues (more than 70% of the subset); lowercase — residues family ( Rausch et al., 2013 ), and are described in sections 3.2.2 and 3.2.3. with a conservation between 30% and 70%, up to 3 amino acids are listed per position with To facilitate the understanding of the most important features of a descending conservation; lowercase italic letters indicate that none of the amino acids at a given subfamily discussed below and to provide a quick summary, we given position occurs with more than 30%, the three most frequent amino acids are given. fi fi A minus ( −) indicates that there is no residue that can be aligned to this position. A special rst present the active site ngerprint containing only amino acids, character is introduced for sequences that do not belong to the OrnTL DB and were aligned which are of high relevance for substrate recognition or catalysis for this manually to their closest homologs within the database. At positions where these manual enzyme followed by a brief summary of the essentials of each subsection. alignments were ambiguous a question mark is shown. For details on substrates, products and abbreviations of the enzymes corresponding to this matrix, see Table 3 . 3.1. ω-Amino acid: α-ketoglutarate transaminases

Fingerprint ωAA: αKG TAs: R132, D/E185, Q216, R353. Summary: the speci ficity for ω-amino acid: α-ketoglutarate ( ωAA: αKG) transamination can be reliably predicted because the four fingerprint residues, involved in dual substrate recognition, are found in all characterised enzymes. However, amino donor speci ficity prediction is more delicate as most ωAA: αKG TAs show a relatively broad substrate scope. Since broad substrate spectra are achieved by unspeci fic binding of different substrates, predictions by use of active site fingerprints are often not reliable. Nevertheless, the ωAA: αKG TAs with narrow substrate scopes

Notes to Table 2: aOnly based on one sequence ( Flavobacterium lutescens Lys ε:αKG TA (UniProt ID: Q9EVJ7 ) that was aligned to its closest homolog A9MMF7 in the OrnTL DB to determine the 3D numbers. bOnly based on one sequence (the E. coli YgjG enzyme (UniProt ID: P42588 ; PDB ID: 4UOX )) cOnly based on one sequence (the Serratia sp. 3-acetyloctanal TA (UniProt ID: Q5W267 ), crystal structure (PDB ID: 4PPM )) dOnly based on nine characterised enzymes (see Supplementary data Table S5 entries 76 –85) eOnly based on two sequences (the Pseudomonas aeruginosa spuC (UniProt ID: Q9I6J2) and its homolog from P. putida (UniProt ID: Q88CJ8 )) fOnly based on four characterised enzymes (see Supplementary data Table S5 entries 137 –140) (UniProt ID: Q6JE91 is different at several matrix positions) gOnly based on one characterised enzyme (from Candidatus cloacamonas acidaminovorans (UniProt ID: B0VH76 )) hOnly based on two sequences (the Pseudomonas stutzeri (UniProt ID: Q6VY99 ) and the P. putida enzyme (GenBank ID: AX467211 )) the structure of the P. stutzeri enzyme (PDB ID: 2CY8 ) is an unpublished apo structure iOnly based on five characterised mammalian enzymes (see Supplementary data, Table S5, entries 162 –166) jOnly based on one sequence (the Lactobacillus buchneri Ile-2-epimerase (UniProt ID: F4FWH4 )) kOnly based on two characterised sequences (the human O-phosphoethanolamine phospholyase (UniProt ID: Q8TBG4 ) and 5-phosphohydroxy- L-lysine phospholyase (UniProt ID: Q8IUZ5 )); the only difference among these residues is 185 where Cys is re- placed by a Val in the latter enzyme lOnly based on two characterised sequences (the enzymes from Sphingopyxis macrogoltabida (UniProt ID: D2D3B2 ) and bacterium ATCC 55552 (UniProt ID: E2E0Q4 )) mOnly based on one characterised sequence (Seq. ID 56 in patent WO2004085624) nOnly based on one characterised sequence (the Mycosubtilin synthase subunit A from Bacillus subtilis (UniProt ID: Q9R9J1 )) F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 579

Table 3 Overview of the extended transaminase classi fication after Grishin et al. (1995) showing substrates converted and products formed with special focus on class III transaminases within the OrnTL DB. For TA classes I, II, IV, V and VI only representative examples are shown. All transaminase speci ficities are described by their natural amino donor and acceptor (if known) where pyruvate (pyr), α-ketoglutarate ( αKG) and amino acids are three letter abbreviated e.g. Asp: αKG TA for aspartate transaminase. The substrates and products are drawn in the ori- entation they bind to the PLP (if this is known) when looking from the re -face, with the P-side on the left (as in Fig. 10 ). In cases with more than one amino group per molecule, the transaminated one is highlighted in grey (if known). In cases where the products undergo spontaneous further reactions (as e.g. in Lys ε:αKG TA), the products of these are shown. The common terms amino donor and amino acceptor for TA substrates might be misleading in a reversible reaction as e.g. γ-aminobutyrate: αKG TAs (GABA: αKG TAs) may also be referred to as Glu:succinate semialdehyde TA if regarded from the reverse direction. We followed the physiological function (if known) or the reaction equilibrium to specify the donor and acceptor in transamination reactions and to term each enzyme.

α

α

α

α

α

α

α

α

α

α

(continued on next page) 580 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604

Table 3 (continued )

α

α F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 581

Table 3 (continued )

α

α

α

α

a RS: small residue, R L: large residue. bThe full coenzyme A structure is shown in Supplementary data Figure S5 A. cFor six of these the amino donor is unknown, the sequences, however, suggest SAM:KAPA TA. dAdditional substrates are shown in Supplementary data Figure S4 A. eSubstrates and products are shown in Supplementary data Figure S4 B –E. fSubstrate structures are shown in Supplementary data Figure S5 B and C. gTransaminase class VI is equal to InterPro's DegT/DnrJ/EryC1 family. developed mechanisms to distinguish the different ωAAs, thereby enabling known enzymes was tested ( Seong Gyu et al., 2001; Tripathi and plausible predictions of their speci ficity . Ramachandran, 2006 ). Therefore it is often not possible to unambigu- ω-Transaminases, which are selective for αKG as the amino accep- ously classify certain ωAA:αKG TAs and we focused on enzymes that tor, but differ in their amino donor speci ficity, include 72 characterised were tested for more than one substrate to compare active site features enzymes of the 12,956 sequences in the OrnTL DB. The above shown that are important for amino donor discrimination. fingerprint residues are characteristic for these αKG speci fic enzymes. Already small substitutions of additional active site residues vary the ac- 3.1.1. Dual substrate recognition: the glutamate switch ceptance of amino donors (mainly ω-amino acids of different length The dual substrate recognition mechanism is conserved among the and substitution pattern). It is of special interest to understand how ωAA: αKG TAs as all substrate pairs demand for the same requirements. these different enzymes learned to favour one of the very similar sub- Given the fact that the L-enantiomer of glutamate is formed in each strates over another and to distinguish between GABA, Orn, AcOrn reaction, the O-side (for active site nomenclature, see Box 1 ) has to ac- and Lys. The majority of these enzymes, however, are not absolutely commodate αKG's 1-carboxylate. This is achieved by a highly conserved speci fic and convert more than one of those substrates with reasonable arginine (R353, see Fig. 11 B) ( Hirotsu et al., 2005 ). activity. Enzyme redundancy and substrate promiscuity seems to be a In the ω-amino acid converting half reaction the O-side only accom- common feature of many class III transaminases ( Lal et al., 2014; modates a proton (instead of a carboxyl group) from the terminal car- Schneider and Reitzer, 2012 ). Unfortunately, the substrate scope of bon of the substrate and therefore the pocket needs to be otherwise several enzymes has not been investigated systematically and often ‘filled ’. For this purpose, the side chain of E185 switches into the active only the substrate pair suggested from sequence identity to other site and forms a salt bridge with R353 to neutralise its positive charge. 582 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604

Due to this movement, the dual substrate recognition in these enzymes α-amino group is prevented ( Markova et al., 2005 ). The importance of is often referred to as ‘glutamate switch ’ (Hirotsu et al., 2005 ). The R132 for substrate recognition, which is probably the reason for its con- ‘switched-in ’ position of the side chain of E185 is further stabilised by servation in ωAA: αKG TAs, becomes evident in the human inborn a hydrogen bond from ‘below ’ by Q216 (see Fig. 11 A). GABA: αKG TA de ficiency, which may result from a R132K mutation in

On the P-side another highly conserved arginine residue (R132) these enzymes that reduces the v max by 25% ( Medina-Kauwe et al., forms a salt bridge with the 1-carboxylate of the ωAA substrate in the 1999 ). A second con firmation of its in fluence on αKG recognition first half reaction (Fig. 11 A) and in most ωAA:αKG TAs also with the was provided in a mutagenesis study on a dialkylglycine decarbo- 5-carboxylate of αKG in the second half reaction ( Fig. 12 B) ( Newman xylase (see section 3.5) where the M132R mutation conferred the et al., 2013 ). One exception is the AcOrn: αKG TA from Thermus ability to convert L-glutamate ( Fogle and Toney, 2010 ). thermophilus , ( Fig. 11 B) where R132 is only involved in weak electronic The functional importance of the interplay of the three residues interactions to αKG ( Hirotsu et al., 2005 ). E185 Q216 and R353 is also re flected by the CMA: these positions R132 also determines the speci ficity towards transamination of the are highly correlated in the OrnTL DB and mainly occur together (see ω-amino group in α,ω-diamino acids: R132 speci fically interacts with CMA network in Fig. 9 ). Therefore the presence of these three residues the substrate's 1-carboxylate, thereby keeping the substrate in the together might be regarded as a strong indication for ωAA: αKG activity right orientation for ω-transamination. Additionally, R132 would repel of an uncharacterised enzyme. the ω-amino group when the di-amino acid would be oriented in the The conserved R353 is additionally found in other enzymes that alterative binding mode at the P-side; thereby the conversion of the accept α-amino acids such as the 2,2-dialkylglycine decarboxylase

A) AcOrn: αKG TA, 1WKG B) AcOrn: αKG TA, 1WKH 16

46 271 353 242

216 132 215 185

129

C) AcOrn: αKG TA, 1WKG D) Orn: αKG TA, 2OAT

15 82 113

Fig. 11. The dual substrate recognition in AcOrn: αKG TAs is achieved by the E185 ‘switch ’ (A & B) and Y16 side chain conformation determines AcOrn: αKG TAs from Orn: αKG TAs (C & D). The substrate (analogue) PLP adducts are shown in orange, variable region residues are shown in yellow. A) E185 neutralises R353 when AcOrn is bound (PDB ID: 1WKG). B) E185 ‘switched ’ out of the active site to allow for the coordination of αKG's 1-carboxylic group by R353 (PDB ID: 1WKH) C) In AcOrn: αKG TAs Y16 points out of the active site to allow for AcOrn binding (Q82 is 1WKG numbering, between core positions 73 & 74) R132 is omitted for clarity reasons. D) In Orn: αKG TAs Y16 points towards the active site to coordinate Orn's α-amino group. This side chain conformation is caused by a second water molecule (compared to one in 1WKG), which is held in place by a hydrogen-bonding network involving the other water, N15 and R113 (2OAT numbering, between core positions 73 & 74). R132 is omitted for clarity reasons. F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 583 subfamily 1D7V and the α-TAs found in PLP fold type I, where it is also (2OAT numbering) on AcOrn/Orn activity has probably already been in- coordinating the substrate's carboxylate ( Sun et al., 1998 ). vestigated as the structure of the N79R mutant of the Salmonella typhimurium enzyme has been deposited in the PDB recently (PDB IDs: 3.1.2. Ornithine, acetylornithine and succinylornithine: α-ketoglutarate TAs 4JF0 & 4JEZ) but unfortunately the mutant is misfolded at the P-side in Fingerprint Orn:αKG TA: Y16, Y46, R132, E185, Q216, R353, R113 both structures ( Bisht et al., 2014 ; unpublished crystal structures). (2OAT numbering, between core positions 73 & 74) Searching for fingerprint Y16, Y46, R132, E185, Q216, R353 in the Fingerprint Orn/AcOrn/SuOrn: αKG TA: Y16, R132, E185, Q216, R353 OrnTL DB resulted in 816 sequences (97% of which are found in the Summary: AcOrn/SuOrn: αKG TAs have a broad substrate speci ficity and 2OAT subfamily). 786 of those sequences have R113 (2OAT numbering, convert both AcOrn, SuOrn, and at higher pH values also Orn. The lack of between core positions 73 & 74) that are therefore predicted to encode speci fic interactions allows for this broad substrate spectrum. On the con- for enzymes highly speci fic for Orn: αKG transamination. trary, Orn: αKG TAs are highly speci fic because the free α-amino group is The fact that AcOrn: αKG and SuOrn: αKG TAs prefer the acylated or- placed between the two conserved Y16 and Y46. The presence of these nithine over the free di-amino acid at neutral pH values cannot be ex- two tyrosines, however, is not suf ficient for creating the high speci ficity. plained by any speci fic interaction with the substrate. This preference, R113, an arginine residue of the variable region between position 73 and which is not found at higher pH values ( Heimberg et al., 1990 ), is prob- 74 is important for the correct positioning of the side chain of Y46. ably established by steric and desolvation effects ( Newman et al., 2013 ). Ornithine aminotransferases (EC 2.6.1.13, Orn: αKG TA, found in sub- Newman et al. (2013) concluded that these enzymes possess a broader family 2OAT), acetylornithine aminotransferases (EC 2.6.1.11, AcOrn: αKG substrate scope or some degree of substrate promiscuity compared to TA, subfam. 2ORD, 1VEF, 3NX3 and 2EO5) and succinylornithine amino- Orn: αKG TAs due to non-speci fic interactions which enables the bind- transferases (EC 2.6.1.81, SuOrn: αKG TA, subfam 2ORD) are found in ing of different substrates in a similar orientation. Speci fic interactions Archaea, Bacteria and Eukaryota , where they are involved in ornithine (e.g. hydrogen bonds) to the substrate would not allow for a broad spec- homeostasis and proline biosynthesis ( Jortzik et al., 2010 ) or the arginine trum because substrates not satisfying these interactions would suffer and lysine biosynthetic pathways ( Xu et al., 2007 ). energetic penalties. Due to the narrow substrate scopes of Orn: αKG TAs, which are Unspeci fic binding of AcOrn is probably also responsible for limited to biotechnologically rather uninteresting compounds, there AcOrn: αKG TA activity that was detected in ‘broad spectrum ’ GABA: αKG are not many biocatalytic or biotechnological applications described. TAs ( Lal et al., 2014; Voellym and Leisinger, 1976 ) which are discussed in However, they were utilised for equilibrium displacement in other section 3.1.4. Additionally, two enzymes from thermophiles that had, αKG forming transaminations as the formed aldehyde product based on sequence similarity, initially been annotated as GABA: αKG TAs (glutamate-5-semialdehyde) is instantly removed from the equilibrium turned out to be AcOrn: αKG TAs ( Koma et al., 2006 ). These are also by spontaneous cyclisation ( Tufvesson et al., 2011 ). discussed in the GABA: αKG TA dedicated section 3.1.4. In general AcOrn: αKG, SuOrn: αKG and often also N-succinyl- L,L- The inhomogeneity of enzymes possessing AcOrn TA activity and the diaminopimelate: αKG activity is found in the same, broad substrate unspeci fic binding of the substrate makes the fingerprint based discrim- scope, enzymes. These AcOrn: αKG TAs have a preference for the acylat- ination of these enzymes from other ωAA: αKG TAs impossible. To some ed ornithine species over ornithine itself at neutral pH ( Heimberg et al., extent, however, it is possible to distinguish the AcOrn: αKG TAs that 1990; Ledwidge and Blanchard, 1999; Newman et al., 2013 ), whereas share sequence similarity to Orn: αKG TAs from the other class III trans- the Orn: αKG TAs do not convert AcOrn ( Heimberg et al., 1990 ) and aminases. Most Orn: αKG TA similar AcOrn: αKG and SuOrn: αKG have GABA is only converted with very low activity ( Markova et al., 2005 ). Y16 on the P-side but not Y46 like the Orn: αKG TAs (1WKG is an excep- The substrate preference of Orn: αKG TAs can be explained by specif- tion). Even though Y16 does not coordinate AcOrn or SuOrn, its conser- ic interactions of Y16 with the free (non-converted) α-amino group vation implies that it is structurally or functionally important for the (Markova et al., 2005 ). The Y16A/G mutations drastically reduced transaminases converting both free and acylated ornithine (see Orn:αKG activity in the human enzyme (PDB ID: 2OAT). Additionally, sequence –function matrix Table 2 ). the role of the highly conserved Y46 in determining Orn: αKG TA speci- ficity was shown by the Y46I mutation in the human enzyme, which switched the substrate speci ficity towards GABA: αKG transamination. 3.1.3. Lysine- ε:α-ketoglutarate TAs Y16 and Y46 might therefore, together with the dual substrate re- Fingerprint Lys ε:αKG TAs similar to 2JJG: R132, E185, Q216, N/S269, cognition residues (i.e. R132, E185, Q216, R353, see section 3.1.1) be R353 employed to distinguish Orn: αKG TAs from other class III transami- Summary: three different kinds of Lys ε:αKG TAs are known: 1) enzymes nases. Nevertheless, these features alone are not suf ficient to discrimi- similar to 2JJG with a broad substrate spectrum that match the fingerprint; nate Orn: αKG TAs from all AcOrn: αKG TAs, as the AcOrn and AcLys 2) enzymes similar to the Flavobacterium lutescens enzyme, that match the converting enzyme from T. thermophilus also has these two tyrosines fingerprint of broad spectrum GABA: αKG TAs, but do not convert GABA; (PDB ID: 1WKG, see Fig. 11 ) ( Miyazaki et al., 2001; Rajaram et al., 3) enzymes which are highly speci fic for Lys and do not convert Orn (se- 2008 ). Rajaram et al. (2008) proposed that the difference in substrate quence unknown). speci ficity between these enzymes arises from side chain conformation- L-Lysine- ε-transaminases (Lys ε:αKG TA), which are found in the al changes of Y16, rather than from differing active site residues. In 2JJG subfamily, catalyse the first step in a bacterial biosynthetic pathway Orn: αKG TAs the side chain is pointing towards the active side, thereby towards β-lactam, the building block of several antibiotic families such preventing the productive binding of AcOrn, whereas this tyrosine as penicillins and cephalosporins ( Tobin et al., 1991 ). This enzyme has points out of the active site in AcOrn: αKG TAs (see Fig. 11 C & D) to cre- been used to convert Nα-protected- L-lysine into precursors of ACE ate additional space resulting in a more relaxed substrate scope. The (Angiotensin Converting Enzyme) inhibitors, which are used as antihy- comparison of the 2OAT and 1WKG structures implies that this subtle pertensive drugs ( Patel et al., 1999 ). Other examples include the use of difference is mediated by the presence of an additional water molecule Lys ε:αKG TA from Sphingomonas paucimobilis to synthesise a precursor behind Y16 in 2OAT. We propose that N15 and R113 (2OAT numbering, of the vasopeptidase inhibitor Omapatrilat ( Patel et al., 2000 ) as well as between core positions 73 and 74), that are highly conserved in 5-hydroxy- L-proline and some protected variants thereof ( Hanson et al., Orn: αKG TAs, but not in AcOrn: αKG TAs are responsible for the correct 2011). The application of these enzymes (similar to that of Orn: αKG positioning of the two water molecules ( Fig. 11 C & D). R113 (2OAT TAs, described in section 3.1.2) to shift the equilibrium of other αKG numbering) in Orn: αKG TAs corresponds to a conserved Q82 (1WKG forming reactions is possible because the aldehyde product is numbering) or N79 (2PB0 numbering) in AcOrn: αKG TAs, while posi- instable ( Tufvesson et al., 2011 ): Lys ε:αKG transamination forms tion 15 is not conserved in these enzymes. The in fluence of the R113 α-aminoadipate- δ-semialdehyde and Glu, while α-aminoadipate- δ- 584 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 semialdehyde undergoes spontaneous intramolecular imine formation N/S269, R353 resulted in 112 sequences of which 88% belong to the to form 1-piperideine-6-carboxylic acid ( Soda et al., 1968 ) (see Table 3 ). 2JJG subfamily. Unfortunately the only two characterised enzymes in the OrnTL DB The Lys ε:αKG TAs of the second type are not found in the OrnTL DB, have been tested for very few substrates ( Romero et al., 1997; Tripathi but the enzyme from F. lutescens was aligned to the closest homolog in and Ramachandran, 2006 ), which makes general conclusions about se- the database to compare the active site residues (see sequence –function quence –function relations in Lys ε:αKG TA dif ficult. We suggest that matrix Table 2 ). It was found that it also possesses the common dual sub- there are at least three types of Lys ε:αKG TAs: the first one is found in strate recognition residues but a G269 and an I46 instead of N/S269 and the 2JJG subfamily and accepts lysine, ornithine, αKG and to a lower ex- V/F46 in the 2JJG subfamily. Interestingly, this enzyme accepts lysine tent also pyr and oxaloacetate as found for the Streptomyces clavuligerus and ornithine, but not GABA ( Yagi et al., 1991 ) even though it shares enzyme (UniProt ID: Q01767 ) ( Romero et al., 1997 ). The second type of the common active site residues with the GABA: αKG TAs as described Lys ε:αKG TAs with known sequence and substrate scope is represented in the next section. Regions that could not be aligned properly must there- by the enzyme from F. lutescens (UniProt D: Q9EVJ7 ) ( Fujii et al., 2000; fore additionally determine substrate speci ficity in this enzyme. Yagi et al., 1991 ), which is, however, not aligned to the OrnTL DB as it shares too little sequence identity with the 2JJG enzyme or any other en- 3.1.4. γ-Aminobutyrate: α-ketoglutarate TAs zyme with known structure. This enzyme is also found to convert orni- Fingerprint ‘narrow spectrum ’ GABA: αKG TAs: I46, R132, E185, thine in addition to the favoured lysine ( Yagi et al., 1991 ). The third type F269, Q216 and R353 of Lys ε:αKG TAs, which is in contrast to the two other mentioned types Summary: I46 is important for GABA: αKG TA activity but no unique fea- by not converting ornithine at all, was found in Candida utilis (Hammer ture, as I46 is also found in some other enzymes within the class III TAs. The and Bode, 1992 ) but its sequence is unknown. Even though it would be active site of eukaryotic GABA: αKG TAs is narrowed by F269 and there- interesting to investigate how this third type achieves the lysine – fore only GABA, LAIB and βAla are accepted as amino donors. Subtle ornithine discrimination, sequence –function relationships can only be amino acid exchanges not covered by the sequence –function matrix investigated for the first two types. can substantially shift the preference from GABA towards βAla. The dual substrate recognition in the enzymes with known sequence is performed as described for all ωAA: αKG TAs in section 3.1.1 and Fingerprint ‘broad spectrum ’ GABA: αKG TAs: (NOT Y16), I46, R132, R132, E185, Q216 and R353 are therefore conserved ( Tripathi and E185, G269, Q216 and R353 Ramachandran, 2006 ). Summary: bacterial GABA: αKG TAs have, due to G269, more space in The creation of a sequence fingerprint for this substrate speci ficity the active site and therefore have a pH dependent relaxed substrate has proven to be challenging owing to its similarity with the other scope. AcOrn is also converted at neutral pH. At higher pH values substrates ωAA: αKG TAs, which all convert very similar substrates (e.g. ornithine, with additional free amino groups like Orn, Lys and putrescine (PUT) are acetylornithine and γ-aminobutyrate). A closer structural comparison accepted as well. Unidenti fied differences to these enzymes (as found in of the 2JJG enzyme with the GABA: αKG TAs and Orn/AcOrn/SuOrn: αKG thermophiles) result in transaminases prefering AcOrn over GABA. TAs showed very similar active sites and it is therefore likely that GABA γ-Aminobutyrate (GABA) is a neurotransmitter and catabolic and AcOrn is converted by these enzymes as well. There is only one GABA: αKG TAs (EC 2.6.1.19) are involved in neurological disorders position (N/S269) in the active site which is fairly unique for the two and therefore are targets for the treatment of e.g. epilepsy. For instance characterised Lys ε:αKG TAs in the OrnTL DB (Supplementary data the suicide inhibitor Vigabatrin is active for epilepsy treatment ( Grant Table S5 entry 42 & 43) and the whole subfamily 2JJG. This residue, and Heel, 1991 ), as it irreversibly inhibits mammalian GABA: αKG however, is not involved in lysine or αKG coordination (see Fig. 12 ) TAs by covalently linking PLP and the catalytic lysine and thereby and therefore the function of this additional hydrogen-bonding donor preventing further catalysis. Interestingly, Orn: αKG TAs are only weakly at the P-side cannot be rationalised, yet. Nevertheless, we suggest and reversibly inhibited by this highly selective inhibitor ( Lee et al., including it in a sequence fingerprint to identify enzymes with compa- 2014 ). The differences of mammalian Orn: αKG and GABA: αKG TAs rable substrate speci ficities but its function should be investigated by from a pharmaceutical point of view have been recently reviewed by mutagenesis studies. The fingerprint search for R132, E185, Q216, Lee et al. (2014) .

A) Lys ε:αKG TA, 2CJD B) Lys ε:αKG TA, 2CJH

46

269 47 242 353

215 216

132 185

129

Fig. 12. Dual substrate recognition in Lys ε:αKG TAs exempli fied for the enzyme from Mycobacterium tuberculosis . A) Lysine's carboxylate is coordinated by R132 and E185 is in contact with Q216 and R353 (PDB ID: 2CJD) B) α-Ketoglutarate is coordinated by R132 on the P-side and R353 and Q216 at the O-side, while E185 ‘switched ’ out of the active site (PDB ID: 2CJH). F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 585

From sequence similarity and substrate scope we suggest to Ustilago maydis . This explains why both, C98 and C101, are only distinguish the GABA: αKG TAs in two groups: eukaryotic enzymes conserved in the mammal enzymes within the 1OHV subfamily. that possess a very narrow substrate scope and are only accepting In bacterial GABA: αKG TAs (found in subfamily 3Q8N) F269 is in GABA and β-alanine (βAla), and the bacterial enzymes which have a most cases replaced by a glycine (found in 91% of sequences in subfam- broader substrate spectrum. ily 3Q8N), thereby creating space and allowing for a more relaxed sub- Due to the very narrow substrate scope the options for biotechno- strate spectrum (compare Fig. 13 A & B). E. coli, for instance, has two logical applications of eukaryotic enzymes is limited, but also for those such ‘broad spectrum ’ GABA: αKG TAs: gabT (PDB ID: 1SFF) is constantly with bacterial origin only few biocatalytic applications have been de- expressed and puuE (UniProt ID: P50457 ) is induced by putrescine. scribed. Bacterial GABA: αKG TAs can be applied as effective biocatalysts These two enzymes have recently been shown to convert AcOrn for the production of GABA or Glu similar substances as demonstrated in vitro and in vivo (Lal et al., 2014 ). Lal et al. (2014) for the first time in- for the herbicide L-phosphinotricin by the E. coli gabT enzyme (PDB vestigated the function of all E. coli class III transaminases systematically ID: 1SFF) ( Schulz et al., 1990 ). and demonstrated that four of its transaminases possess partly redun- All described GABA: αKG TAs share the common residues for dual dant substrate spectra. The gabT enzyme that was further characterised, substrate recognition with the other ωAA: αKG TAs (i.e. R132, E185, was additionally shown to convert L-aspartate ( Liu et al., 2005 ) but Q216, R353, see section 3.1.1 and Fig. 13 A & B) and I46, which has interestingly, in contrast to the eukaryotic enzymes, not βAla ( Park been found to enhance GABA: αKG TA activity in Orn: αKG TAs (see et al., 1993 ). The GABA: αKG TA from Pseudomonas aeruginosa (UniProt section 3.1.2) ( Markova et al., 2005 ). However, residue I46 is not ID: Q9I6M4 ) also ef ficiently converts AcOrn at physiological pH and ad- speci fic for GABA: αKG TAs among the class III transaminases, as sev- ditionally Orn, putrescine (PUT) and Lys with a higher pH optimum eral AcOrn or SuOrn: αKG TAs also have a I46. The discrimination of compared to the GABA: αKG reaction ( Voellym and Leisinger, 1976 ). enzymes with preference for GABA from the AcOrn/SuOrn: αKG TAs We suggest that other class III transaminases, also matching the that are similar to the Orn: αKG TAs, however, is possible if se- active site fingerprint (NOT Y16), I46, R132, E185, Q216, G269 and quences with Y16 are excluded. This residue is highly conserved in R353 could also be able to convert AcOrn at physiological and free the AcOrn/SuOrn: αKG TAs (see section 3.1.2), but not in the known α,ω-diamino acids like Orn and Lys at higher pH values. GABA: αKG TAs. The prediction of a GABA: αKG TAs substrate scope based on the two Structural features limiting the substrate scope of mammalian proposed active site fingerprints is unfortunately not possible in all GABA: αKG TAs (e.g. from pig (UniProt ID: P80147 ) and mouse (UniProt cases. There are three examples described where very subtle amino ID: P61922 ) in subfamily 1OHV) to GABA, L-3-aminoisobutyrate ( LAIB) acid substitutions in the active sites or even in their entrances changed and βAla (Buzenet et al., 1978; Schousboe et al., 1973; Tamaki et al., the substrate spectra. 2000) have been investigated in detail ( Markova et al., 2005; Storici The first example is from the yeast Lachancea kluyveri , which has et al., 1999 ). The F269, which is highly conserved in the eukaryotic en- two ‘narrow spectrum ’ GABA: αKG TAs (UniProt IDs: A5H0J5 and zymes, narrows the active site at the ‘top’ of PLP, thereby preventing A5H0J6, 57% sequence identity), the first one favours βAla three fold the binding of larger substrates. We propose that most of the 233 se- over GABA and the second is selective for only GABA ( Andersen et al., quences in the OrnTL DB that match the fingerprint I46, R132, E185, 2007 ). Homology modelling of these two enzymes with the pig enzyme Q216, F269, R353 encode GABA: αKG TAs with a narrow substrate scope. structure (PDB ID: 1OHV) as template revealed that the only active site A unique structural feature is found in the pig enzyme, where posi- differences are found relatively far away from the cofactor at the P-side tions C98 and C101 of both subunits form a [2Fe –2S]-cluster at the of the active site entrance (substitutions from the βAla converting en- dimer interface, but its function is unknown ( Storici et al., 2004; Sung zyme to the GABA αKG TA: P266A, F349Y (A5H0J5 numbering, between and Kim, 2000 ) (see Fig. 13 B). It is suggested to be involved in an activa- core positions 266 and 267) and C108D (A5H0J5 numbering, between tion mechanism, especially in higher organisms, because this [2Fe –2S]- core positions 73 and 74)) (data not shown). How these mutations cluster was not found in bacteria or yeast except for the basidiomycete are able to effect substrate recognition in such a drastic manner without

A) GABA: αKG TA, 4ATQ B) GABA: αKG TA, 1OHY

16

271 46 271 46 242 242 98 269 353 353 98 215 216 215 216 101 185 185 101 132 132 129 129

Fig. 13. Coordination of GABA (or γ-ethynyl-GABA) by GABA: αKG TA from A) Arthrobacter aurescens (PDB ID: 4ATQ) ( Bruce et al., 2012 ) and B) from pig (PDB ID: 1OHY) ( Storici et al., 2004). The substrate/inhibitor-PLP adducts are shown in orange, which are coordinated by R132. R353, responsible for αKG's 1-carboxylate recognition in the other half reaction, is neutralised by E185 at the O-side. F269 in the pig enzyme narrows the P-side and forces R132 and the GABA analogue in a ‘lower ’ position, thereby causing the narrow substrate scope of this enzyme. In the A. aurescens enzyme, due to G269, the P-side provides more space, which allows more freedom for R132 and therefore longer chain ωAA like Orn and Lys are accepted as well. The iron in the [2Fe –2S] cluster of the pig enzyme is coloured brown. 586 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 reaching the substrate –cofactor complexes remains puzzling without et al., 1977 ) that accept PUT as the amino donor. Transamination of structural information. non-acylated PUT leads to free γ-aminobutyraldehyde, which sponta- The second example of a few active site substitutions in fluencing neously forms the cyclic imine 1-pyrolline ( Shaibe et al., 1985 ). GABA: αKG TA relevance in vivo was discovered in Pseudomonas To date only three cadaverine or PUT: αKG TAs (EC 2.6.1.82) have syringae, which has three gabT variants (gabT1, UniProt ID: Q48QA9; been described: the YgjG enzyme from E. coli (UniProt ID: P42588 ), gabT2, UniProt ID: Q88AT5 and gabT3, UniProt ID: Q885E5 ), which pre- which has been extensively studied ( Schneider and Reitzer, 2012 ), a sumably have different functions ( D.H. Park et al., 2010 ). Knockout of all cadaverine: αKG-converting enzyme involved in the lysine catabolism three and separate complementation through recombinant expression in Streptomyces ambofaciens (Untrau et al., 1992 ) and a partially investi- indicated that only gabT2 is able to restore the ability to grow on gated enzyme found in a methanogenic coculture ( Roeder and Schink, GABA as the sole carbon and nitrogen source. The in vivo function of 2009 ). Unfortunately, only the sequence of the E. coli enzyme is gabT1 and gabT3 remains unclear. Homology modelling with the known, which was found to accept PUT and cadaverine as good amino E. coli enzyme structure (PDB ID: 1SFF) as the template revealed that donors while GABA and Orn showed only low activity and Lys was not the only active site differences that might have a direct in fluence on converted at all ( Kim, 1964; Samsonova et al., 2003 ). Both αKG and GABA: αKG TA activity due to involvement in coordination from gabT2 pyr were converted, while αKG was a ten times better amino acceptor. to gabT1 are K142M and E185D, and from gabT2 to gabT3 only is The structure of the crystallised E. coli enzyme (PDB ID: 4UOX) was pub- Q80N (Q88AT5 numbering, between core positions 73 & 74) (data not lished recently by Cha et al. (2014) after the OrnTL DB was created and shown). Even though the exchange of glutamate to aspartate at position is therefore not included in the database. Even though its sequence is 185 seems very subtle at a first glance, especially as it is not directly in- not found in the OrnTL DB, a protein with 97% identity (UniProt ID: volved in substrate binding, a different substrate or reaction speci ficity A8APX8 ) is found in the 1VEF subfamily, which, together with the of gabT1 seems reasonable with the knowledge that it might prevent structure, allowed for aligning the characterised PUT: αKG TA to the dual substrate recognition and that an aspartate at position 185 other sequences in the database. was (to our knowledge) only found in the racemases among the A special feature of this enzyme compared to other class III TAs was class III transaminase family (see sections 3.6 and 3.7). The case of revealed by its structure: the extended N-terminus folds to an addition- gabT3, however, is not explicable without further experimental data. If al helix (residues 9–23 in P42588 numbering), that interacts with the this enzyme lacks GABA: αKG TA activity, Q80 (Q88AT5 numbering, other subunit. By this additional interaction, the dimer and therefore equal to Q79 in the E. coli gabT (PDB ID: 1SFF)) might have a greater im- temperature stability is increased ( Cha et al., 2014 ). pact on substrate recognition in these enzymes than anticipated. In the Even though there is no structure with bound αKG available, its Putrescine: αKG TA from E. coli this residue was found to coordinate binding at the O-side is clear: E185, Q216 and R353 are present to putrescine's non-reacting amino group (section 3.1.5). achieve the dual substrate recognition like in the ωAA:αKG TAs (see The third example is a report of two enzymes from thermophiles section 3.1.1) ( Cha et al., 2014 ). The αKG coordination at the P-side is that match the ‘broad spectrum ’ GABA: αKG TA fingerprint, but prefer probably achieved by K132 (instead of R132 in the other αKG AcOrn over GABA (Koma et al., 2006 ) (the exact values, however, re- converting enzymes) and maybe additionally by T269 (see Fig. 14 ). main unfortunately unknown). A comparison of all hypothetical active Cha et al. (2014) state that speci fic PUT recognition is realised by site residues of these enzymes to those of characterised GABA: αKG Q119 (P42588 numbering, between OrnTL DB core positions 73 and TAs revealed either a S269 or S267 at the P-side as the only difference 74), because it is found to hydrogen bond PUT's non-reacting amino that can be seen from the alignment. However, the N-terminus may group in chain B of the structure. However, in the two other PUT con- not be aligned properly and its contribution to the active site remains taining active sites (with K242 from chain A and C) PUT adopts orienta- unclear. These two enzymes' preference for AcOrn over GABA can there- tions that do not allow this H-bond ( Fig. 14 shows the active site of fore not be rationalised without additional structural information. chain B, a scene highlighting PUT's orientation in chain A is provided in the Supplementary PyMOL session). A mutagenesis study would 3.1.5. Putrescine and cadaverine: α-ketoglutarate TAs therefore be required to investigate Q119's (P42588 numbering) role Fingerprint PUT: αKG TAs: F46, K132, E185, Q216, R353, Q119 in PUT binding in more detail. This residue is also commonly found (P42588 numbering, between OrnTL DB core positions 73 and 74) among the ‘broad spectrum ’ GABA: αKG TAs (section 3.1.4), which Summary: the dual substrate recognition is probably achieved as in might explain their ability to convert PUT as well. The reduced activity ωAA: αKG TAs with the exception that R132 is replaced by a lysine. of the PUT: αKG TA towards GABA and AcOrn may be explained by the Putrescine's second amino group is coordinated by Q119 (P42588 number- narrow, hydrophobic active site of the enzyme. Residues F46, F145, ing), which is located at the O-side in the variable region between core F327 (P42588 numbering between core positions 266 and 267) and positions 73 & 74. The substrate scope is additionally determined by L419 (P42588 numbering between core positions 348 and 350), proba- bulky hydrophobic residues narrowing the active site. An additional bly hamper their binding. The insertion of amino acids between posi- N-terminal helix provides increased stability by interactions with the tions 348 and 350 and their orientation (protruding into the active other subunit. site; L419 in particular) is characteristic for this enzyme and not found Putrescine (PUT) and cadaverine are biogenic diamines that are in other class III TA structures ( Cha et al., 2014 ). Additionally, K132 is found in almost all living organisms and are known for modulating not as ef ficient for the binding of ωAA's carboxylate as found for the translation and transcription ( Schneider and Wendisch, 2011 ). Owing human Orn: αKG TA (see section 3.1.1). to their drastic in fluences on metabolism, their cellular levels need to A fingerprint identifying PUT: αKG TAs should therefore contain F46, be carefully tuned. Three different pathways for PUT degradation have K132, E185, Q216, R353 and Q119 (P42588 numbering, between OrnTL been proposed ( Kurihara et al., 2005; Lu et al., 2002; Shaibe et al., DB core positions 73 and 74), which matches only 15 sequences in the 1985 ). These catabolic routes proceed via acetylation, glutamylation or OrnTL DB. However, as most active site residues (except the dual sub- direct oxidation and all yield GABA. The oxidation from amine to alde- strate recognition residues and Q119 ) determine the substrate speci fic- hyde is realised by amine oxidases or transaminases, the second oxida- ity by unspeci fic hydrophobic interactions or by narrowing the active tion to GABA is achieved by dehydrogenases. Organisms that apply site, several other combinations of hydrophobic active site residues transamination for the first oxidation step have been shown to either may probably achieve a comparable substrate scope. use αKG ( Kim, 1964) or pyr ( Yorifuji et al., 1997 ) as amino acceptor. For the discussion of PUT:pyr TAs see section 3.2.4. An additional option 3.1.6. 3-Acetyloctanal transaminase (PigE) for PUT transamination can be ‘broad substrate spectrum ’ GABA:αKG Summary: only one enzyme with this activity is known. Structural in- TAs (Voellym and Leisinger, 1976) or ω-amino acid:pyr TAs (Yonaha formation (PDB ID: 4PPM ) became available after the revision of this F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 587 article. The substrate scope is unknown; we propose that it prefers Glu as additional enzyme and can be achieved by ‘broad spectrum ’ GABA: αKG donor and has a broad spectrum for aldehydes as acceptor. TAs, as found in Pseudomonas species ( Brohn and Tchen, 1971 ), but The biosynthesis of the red tripyrrole pigment prodigiosin in Serratia many halotolerant organisms were shown to have a separate enzyme sp. was found to involve the class III transaminase pigE (UniProt ID: for DABA synthesis. The amino donor in the reaction from L-2-amino- Q5W267 ) for aminating the aldehyde of the precursor 3-acetyloctanal 4-oxo-butyrate to L-2,4-diaminobutyrate might be either glutamate (3AcOc) ( Williamson et al., 2005 ). This enzyme has unfortunately only in Glu:2-amino-4-oxobutyrate TAs ( Vandenende et al., 2004 ) (EC been investigated for 3AcOc conversion and amino donor and acceptor 2.6.1.76) or alanine in Ala:2-amino-4-oxobutyrate TAs ( Rao et al., spectra remain unknown. The crystal structure of this enzyme has been 1969) (EC 2.6.1.46), but for most characterised DABA TAs the amino solved recently ( Lou et al., 2014 ), but its structure (PDB ID: 4PPM) was donor speci ficity was not investigated, hence we refer to all of them as not available in the PDB until the revision of this article was completed. DABA TAs. The sequence alignment to its closest homolog in the OrnTL DB, howev- Unfortunately most DABA TAs are not similar enough to any enzyme er, due to the standard dual substrate recognition residues at the O-side with solved crystal structure and are therefore not aligned to the OrnTL (E185, Q216, R353, see section 3.1.1) and K132 at the P-side, suggests DB. The only aligned sequence is ectB from Virgibacillus pantothenticus that Glu is the preferred donor. 3AcOc recognition cannot be predicted (UniProt ID: Q6PR32 ), which belongs to subfamily 1VEF. To be able to without structural information, but is probably achieved by unspeci fic investigate sequence conservation among these enzymes, a manual interactions. This enzyme must, however, be able to somehow discrim- alignment of 20 characterised and predicted ( Reshetnikov et al., 2006 ) inate the aldehyde from the keto function in 3AcOc. We therefore pro- sequences has been created (for details see Supplementary data pose that pigE might possess a broad substrate scope for aldehydes. In Table S5 entries 76 –96). From this alignment in comparison to the case this enzyme is enantioselective at C3, it might be a valuable tool OrnTL DB, the probable substrate coordination enabling residues could for biocatalytic kinetic resolutions of aldehydes. be derived. The substrate coordination in the DABA TAs is comparable to that in 3.1.7. 2-amino-4-oxobutyrate transaminases (diaminobutyrate TAs) the other ωAA: αKG TAs only at the O-side. The common αKG dual sub- Fingerprint for DABA TAs unknown — P-side substrate recognition strate recognition residues E185 and Q216 are found there (see section remains unclear 3.1.1). Only R353 is replaced by K353, which is probably able to substi- Summary: substrate coordination at the O-side is achieved as in tute R353's role in the substrate's α-carboxyl coordination. The R353K ωAA: αKG TAs (E185 & Q216), only R353 is replaced by a lysine. Substrate replacement in the E. coli Asp: αKG TA, for instance, was shown to retain binding at the P-side and the molecular basis for α/γ-amino group discrim- the enzyme's activity ( Cánovas et al., 1998 ). K353 might additionally be ination in DABA cannot be predicted. coordinated from the ‘top ’ by E346, which is also conserved in this small Diaminobutyrate transaminases (DABA TAs) catalyse the first step in alignment but is found relatively seldom in the OrnTL DB (only 134 the biosynthesis of the compatible solute ectoine and are therefore sequences). mainly found in halophile or halotolerant species ( Schwibbert et al., The main difference to other ωAA: αKG TAs is located at the P-side, 2011 ). DABA: αKG transamination in some species does not need an where position 132 is not conserved and contains mainly non- polar amino acids instead of an arginine. The coordination of the α-carboxylic group of DABA and the 5-carboxylate of αKG need there- fore to be achieved in a different way. When comparing the conserva- PUT: αKG TA, 4UOX tion of all residues at the P-side to other families in the OrnTL DB, mainly three residues attracted attention: Y16, H72 and R273. These 119 three amino acids might be involved in the substrates' carboxylate coordination. Furthermore, position 269 contains a suitable hydrogen bond donor (N/T/S269) in all DABA TA sequences (except A269 in the P. aeruginosa enzyme, UniProt ID: A3KUH7 ) and might also be involved in substrate recognition. However, how these enzymes achieve the discrimination of DABA's α- and γ-amino group remains unclear. The 353 327 46 elucidation of this interesting feature and the amino donor coordination 269 at the P-side will require a solved crystal structure. 271 242 3.2. ω-Amino acid:pyruvate transaminases

419 Fingerprint ωAA:pyr TAs: R346 and (NOT D/E132) 216 Summary: amine transaminases (ATAs) are valuable catalysts for asymmetric amine synthesis, but not all ωAA:pyr TAs possess high ATA ac- 132 185 215 tivity. Based on their biocatalytic usefulness, ωAA:pyr TAs can be grouped in ‘high activity ’ ATAs and ‘low activity’ ATAs. Note that stereoselectivity is 145 — 129 usually excellent within this enzyme class and thus activity and substrate scope — is the main property of interest when interrogating the protein se- quence for novel useful enzymes. In addition to the ω-TAs, which accept α-ketoglutarate as an amino Fig. 14. Substrate recognition in the PUT: αKG TA from E. coli (PDB ID: 4UOX). The sub- acceptor, the class III transaminases of PLP fold type I also include ω-TAs strate-PLP adduct is shown in orange, which is coordinated by Q119 (P42588 numbering which accept pyruvate as an amino acceptor ( ωAA:pyr TAs). This group between core positions 73 and 74) in chain B of the structure (see scene ‘Fig14_chainA ’ in the Supplementary PyMOL session for the PUT orientation in chain A). Residues of the of transaminases comprises GABA:pyr TAs, Taurine:pyr TAs, βAla:pyr variable regions are shown in yellow ( Q119 , F327 and L416 (P42588 numbering)). The hy- TAs, vanillylamine:pyr TAs and also includes examples, which catalyse drophobic residues F46, F145, F327 and L416 form a narrow and hydrophobic active site the transamination of substrates lacking carboxylic acid moieties (entrance), which is supposed to prevent binding of bulkier and more polar substrates. (amine transaminases, amine:pyr TAs, ATAs, see Fig. 5 ). ATAs are of The ‘glutamate switch ’ (E185, Q216, R353) is most probably responsible for dual substrate great biotechnological interest as they can be utilised for asymmetric recognition and together with K132 binds αKG. The loop including L416 narrowing the ac- tive site entrance is a major difference of the PUT: αKG TA compared to other enzymes in amine synthesis. These enzymes proved to be able to compete with the OrnTL DB and therefore belongs to the variable regions. established chemical methods for industrial amine production ( Kohls 588 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 et al., 2014 ). The broad applicability of these enzymes in biocatalysis 3.3.6). These enzymes are, however, not comparable as they do not have been reviewed recently ( Berglund et al., 2012; Höhne and need a dual substrate recognition at the O-side as only α- and β-amino Bornscheuer, 2009, 2012; Kroutil et al., 2013; Rudat et al., 2012 ). acids are converted. A notable difference is found in the vanillylamine: The different amino donor speci ficities within this group are often pyr TAs from chilli pepper and a GABA:pyr TA from tomato (see Sup- not clearly distinguishable, as most of the characterised enzymes have plementary data Table S5 entries 125, 126 and 131) that have a S346. a rather relaxed substrate speci ficity, but we attempted to group them In this case pyruvate is probably only coordinated by W47, which is by their usefulness for biocatalytic amine synthesis. Originally, ωAA: also involved in the ‘high activity ’ ATAs (see Fig. 15 A), as there is pyr TAs were described to be identical to βAla:pyr TAs (EC 2.6.1.18) no other residue at the O-side of this enzyme to replace R346. (Yun et al., 2004 ), but it turned out that not all βAla:pyr TAs could Whether an asparagine at position 353, which is present in these also convert a variety of different ω-amino acids and amines and that three enzymes, is additionally involved in this coordination remains others convert several amines and ω-amino acids, but not β-alanine unclear. (Sayer et al., 2013 ). Among the ωAA:pyr TAs, it is not possible to conclude an enzyme's ability to convert amines from only regarding 3.2.2. Natural function of amine transaminases sequence similarity to known ATAs as the group of ωAA:pyr TAs is rel- The natural function of enzymes with ATA activity is not known in atively diverse and are found in six different subfamilies of the OrnTL most cases. There are βAla:pyr TAs ( Yonaha et al., 1977 ), GABA:pyr TAs DB (4E3Q, 3HMU, 3I5T, 3GJU, 3A8U and 3N5M). In each subfamily (Steffen-Munsberg et al., 2013a ) and vanillylamine:pyr TAs ( Weber amine converting enzymes have been described. By experimentally et al., 2014 ) described to accept amines, but several enzymes have only characterising four TAs belonging to subfamilies 3HMU, 3I5T and 3GJU been investigated for amine production without testing their in vivo func- (PDB IDs: 3HMU, 3I5T, 3FCR and 3GJU), we found that only two tion. We hypothesised that the conversion of amines in many cases is a of them (3HMU and 3I5T) possessed relatively high activity towards ‘substrate promiscuous ’ activity and their natural function might be the standard ATA substrates such as 1-phenylethylamine (around 0.5 U/ conversion of small ω-amino acids (e.g. βAla or GABA) with pyruvate or mg or higher). The ‘low activity ’ ATAs showed at least 20 times less ac- glyoxylate as the acceptor ( Steffen-Munsberg et al., 2013a ). Rausch et al. tivity towards those amines. Interestingly, the GABA:pyr activity (mea- (2013) together with our study further strengthened this hypothesis, as sured as reverse reaction) was comparable in all of them ( Steffen- the known ATAs share high sequence identity with characterised βAla: Munsberg et al., 2013b ). We therefore hypothesised that the natural pyr TAs (Supplementary data Table S5 entries 109 –116), GABA:pyr TAs function of these enzymes is GABA:pyr transamination, while amine (Supplementary data Table S5 entries 127 –136) and vanillylamine:pyr conversion is based on substrate promiscuity, which is pronounced dif- TAs (Supplementary data Table S5 entries 125 & 126). Another enzyme ferently among the four enzymes. By comparing their active sites, it was with high similarity to known ATAs is spuC from P. aeruginosa that was found that the mechanism for dual substrate recognition was the same described as putrescine:pyr (PUT:pyr) TA, some ATAs, however, were in all four (see section 3.2.1), while there were main differences at the found to not accept PUT as substrate (section 3.2.4). All these enzymes O-side between the ‘high activity ’ and the ‘low activity ’ ATAs (see sec- are found in the same OrnTL DB subfamilies and in most cases have tion 3.2.3). the same active site residues. Their substrate spectra are relatively relaxed and it is therefore likely that many of them possess more 3.2.1. Dual substrate recognition: the flipping arginine than one natural function (i.e. are involved in more than one As most known ωAA:pyr TAs transaminate a variety of amino donors metabolic pathway). However, the conversion of amines is in most with pyruvate, their substrate recognition has necessarily to be quite flex- cases (except for vanillylamine:pyr or PUT:pyr TAs) not likely to be ible. While the P-side only needs to accommodate small alkyl groups their in vivo purpose. As it is impossible to clearly de fine an ωAA: (ethyl or methyl) or only a proton, the O-side must on the one hand be pyr TA's function, we focused on their biocatalytic usefulness and able to coordinate the carboxylate of pyruvate and ω-amino acids, but only attempted to distinguish between enzymes with high and low on the other hand also accommodate the bulky hydrophobic substituent ATA activity and enzymes that show a preference for small β-amino of the amine substrates. The dual substrate recognition that enables car- acids ( βAla:pyr TAs, for enzymes converting bulkier β-amino acids see boxylate and hydrophobic group binding is achieved by a flexible, so- section 3.3.2 and section 3.3.3 for enzymes converting thioesters of β- called ‘flipping ’ arginine R346 ( Steffen-Munsberg et al., 2013a ). This argi- amino acids). nine is highly conserved in the ωAA:pyr TA containing subfamilies (ex- cept 3N5M) and it is most probable that the same dual substrate 3.2.3. Discriminating high and low activity amine transaminases and βAla: recognition mechanism is utilised in all these enzymes: the flexible argi- pyr TAs nine might form a salt bridge with the 1-carboxylate of the amino accep- Fingerprint high activity ATAs: W47, A185, R346, NOT D/E132 tor (e.g. pyruvate) when it is in its ‘flipped in ’ conformation (see Fig. 15 A). Fingerprint low activity ATAs: Y47, S/T185, R346, NOT D/E132 By ‘flipping upwards ’ and thereby out of the O-side, it allows for the ac- Fingerprint βAla:pyr TAs: W47, S185, R346, NOT D/E132 commodation of large hydrophobic groups of the amine (e.g. the phenyl Summary: subtle amino acid exchanges at position 47 and 185 deter- group of ( S)-1-phenylethylamine, see Fig. 15 B). Furthermore, the coordi- mine ωAA:pyr TAs' ability to convert amines and β-alanine. nation of ω-amino acids' 1-carboxyl group is also achieved by R346 (for Comparisons of the ‘high activity ’ ATAs to other class III transaminases an animation of the dual substrate recognition involving R346 in ATAs, revealed that only those with W47 and A185 turned out to possess a high see Supplementary video). Its side chain is able to adopt sufficient confor- ‘substrate promiscuous ’ ATA activity ( Rausch et al., 2013; Steffen- mations that are required to bind ω-amino acids of different length (see Munsberg et al., 2013a ). Enzymes with hydrogen bond donors at these Fig. 15 C & D and section 3.2.3 for a more detailed discussion of the sub- positions (e.g. Y47 and T/S185) showed less pronounced activity for strate recognition). The central role of this arginine for the ‘dual ’ substrate amines ( Steffen-Munsberg et al., 2013a ). Site directed mutagenesis recognition was proven by the R346A mutation, which drastically de- proved that the residues W47 and A185, among all active site differences creased the activity towards keto acids (pyruvate and succinic semialde- to ‘low activity ’ ATAs, are the most important ones for high ATA activity. hyde), whereas amine transamination was hardly effected ( Steffen- The ‘low activity ’ ATAs from Reugeria sp. (PDB ID: 3FCR) and Munsberg et al., 2013a ). We therefore suggest the fingerprint R346 and Mesorhizobium loti (PDB ID: 3GJU) showed substantially increased ATA NOT D/E132 (to discriminate from the DAPA TAs, which also have R346 activity when the Y47W or the T185A single mutations had been for DAPA coordination; see section 3.3.5) to identify enzymes with introduced. ωAA:pyr TA activity among the class III transaminase family. There are, We therefore suggest using W47, A185, R346 and NOT D/E132 however, also enzymes with a substrate speci ficity for pyr or Ala that do (to discriminate from the DAPA TAs (see section 3.3.5) to identify not match this fingerprint, such as Ala:glyoxylate TAs 2 (see section ATAs with high activity within the class III transaminase family. F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 589

A) “high activity” ATA, 4E3Q B) “high activity” ATA, 4E3Q

332

346

46 47

271 242

216

185 215

129

C) “low activity” ATA, 3GJU D) βAla:pyr TA, 3A8U

Fig. 15. Substrate recognition of amines and α-, β- and γ-amino acids in ωAA:pyr TAs. The modelled quinonoid intermediates of A) alanine in 4E3Q, B) 1-phenylethylamine in 4E3Q, C) γ-aminobutyrate in 3GJU and D) β-alanine in 3A8U are shown in orange. R346 is suf ficiently flexible to coordinate the carboxylate of α-, β- and γ-amino acids, but can also ‘flip out ’ of the active site to create space for e.g. the phenyl ring of PEA (B). In particular positions 47 and 185 are different in enzymes with high ATA activity (A & B) compared to those with low ATA activity (C) and those with preference for β-amino acids (D) as described in Discriminating high and low activity amine transaminases and βAla:pyr TAs section. The inter- mediates were modelled with YASARA Structure (Version 13.6.16) as described elsewhere ( Steffen-Munsberg et al., 2013a ).

For ωAA:pyr TAs with low ATA activity we suggest the fingerprint citreus (Garcia et al., 2006 ) and Bacillus megaterium (Hanson et al., Y47, S/T185, R346 and NOT D/E132. Additionally, L46 might have 2008 ) (98% sequence identity), both of which are not included in any se- an important role for ωAA:pyr TA activity as it is highly conserved quence database (see Supplementary data section 5 for their sequences). (91% of all sequences that match the fingerprint R346, NOT D/ Furthermore, low ATA activity was found in additional class III transami- E132). Its function, however, is not yet known. nases that are comparably evolutionary distant from ωAA:pyr TAs. Two of In addition to the well described enzymes above, there are some high- these enzymes were found in M. loti (UniProt IDs: Q98AI1 and Q98NJ9) ly active amine converting enzymes known that do not match the sug- that both converted several amines with pyr ( Seo et al., 2012 ), while gested patterns, e.g., the two very similar enzymes from Arthrobacter not matching the ωAA:pyr TA fingerprint. These two enzymes, however, 590 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 only possess very low ATA activity (80 times lower activity towards Nocardioides simplex (Kaneoke et al., 1994 ) and was puri fied from benzylamine compared to the V fl-ATA) ( Kwon et al., 2010 ) and are there- Arthrobacter sp. TMP-1 ( Yorifuji et al., 1997 ), both without sequence fore not discussed here in detail. Two enzymes with high ATA activity information of the responsible enzymes. In P. aeruginosa the enzyme were found among the βPhe TAs, which are discussed with their related spuC (UniProt ID: Q9I6J2 ) has been suggested to catalyse the PUT:pyr enzymes in section 3.3.2. These findings indicate that there might be transamination due to its PUT dependent up regulation, the disruption many more enzymes having a promiscuous ATA activity, but most en- of the organism's ability to utilise PUT as sole N- or C-source by knock- zymes are not characterised systematically at the moment. out mutations and PUT:pyr activity measurements in crude extracts Still puzzling are the structural features that allow for β-alanine (Chou et al., 2008; Lu et al., 2002 ). An analogue enzyme (UniProt ID: transamination as some amine accepting enzymes convert β- Q88CJ8 ) was induced by PUT in P. putida as well ( Bandounas et al., alanine while others do not ( Sayer et al., 2013 ). The first described 2011 ). Furthermore, spuC homologs have also been induced by PUT in amine-converting transaminase (from Pseudomonas putida , PDB coastal bacterioplankton as shown by a metatranscriptomic analysis ID: 3A8U ( Watanabe et al., 1989 )) for instance prefers βAla over (Mou et al., 2011 ). carboxylate free substrates ( Yonaha et al., 1977 ), while several The spuC enzyme from P. aeruginosa and its closest homolog in later described ‘high activity ’ ATAs do not convert βAla (e.g. the en- P. putida , that are found in the 3HMU subfamily, share 59 and 58% iden- zymes from Vibrio fluvialis (V fl-ATA, PDB ID: 4E3Q) ( Shin et al., tity with the amine transaminase 3HMU, respectively and all three have 2003 ) and C. violaceum (Cvi-ATA, PDB ID: 4A6T) ( Kaulmann et al., identical active site residues (see sequence–function matrix Table 2). 2007 )). Sayer et al. (2013) reasoned that a difference in active However, when 3HMU was tested with different PUT derivatives, it site flexibility is important for amine conversion, but prevents the turned out to convert glutamyl-PUT and acetyl-PUT but not PUT (Lea conversion of βAla in the broad spectrum ATAs: the left-handed Kennel, unpublished results). The connection of spuC and its homologs helix ( α2, see section 2.2) in Cvi-ATA was inverted in the to the putrescine catabolism will therefore need to be investigated in gabaculine (a common GABA homolog TA inhibitor) bound crystal more detail, as free putrescine conversion by spuC seems to be unlikely. structure (PDB ID: 4BA5), compared to the gabaculine structure of Unfortunately, P. aeruginosa was only tested for PUT:pyr TA activity in the βAla:pyr TA from P. aeruginosa (PDB ID: 4B98), where the left- crude extracts and spuC was never puri fied. The involvement of addi- handed helix stayed intact. The observed inversion of the left- tional enzymes in the detected activity can therefore not be ruled out. handed helix in Cvi-TA might, however, not be necessary for amine binding, but instead be forced by the inhibitor's rigidity: 3.2.5. Taurine:pyruvate TAs the inhibitor would clash with L46 and W47 if the helix α2 was Summary: sulfonate/carboxylate coordination is different from all other not inverted. Normal ATA substrates are less rigid and therefore ωAA:pyr TAs. The left-handed helix α2 is probably also oriented differently should be able to bind without α2 inversion. because tau:pyr TAs have a glycine inserted before position 46 . However, when comparing the active site residues of ATAs with As taurine (tau) is one of the most abundant small organic solutes in preference for βAla (see Supplementary data Table S5 entries several animals, many bacteria have developed ways to utilise it as a S-, 109 –116 and sequence –function matrix Table 2 ) with those of C- or N-source, where tau:pyr transamination (EC 2.6.1.77) is involved broader spectrum ATAs or GABA:pyr TAs that do not accept βAla in most cases ( Laue and Cook, 2000 ). Unfortunately only four sequences (see Supplementary data Table S5 entries 99 –105), the previously of transaminases with proven tau:pyr activity are known and no struc- mentioned positions 47 and 185 are found to be speci fically con- tural information is available (three in 3N5M and one in 3HMU subfam- served. The occurrence of W47 and S185 among the βAla:pyr TAs ily, see Supplementary data Table S5 entries 137 –140). The enzyme might be necessary to fix βAla's carboxylate in the proper position from Bilophila wadsworthia (UniProt ID: Q9APM5 ) is the only example for catalysis, as found when modelling the quinonoid intermediate that has been characterised for its substrate scope ( Laue and Cook, of βAla in the structure 3A8U (see Fig. 15 D). Other ATAs have A185 2000 ): small α- and β-amino acids are accepted. Hypotaurine, taurine or Y47 instead of S185 or W47 ( Fig. 15 A–C), indicating that they are and β-alanine are the best amino donors (in that order), while pyruvate not able to fix the carboxylate in the right position for βAA conver- and 2-ketobutyrate can be employed as the amino acceptor. sion, whereas αAA and γAA are bound in a productive way. 3 Fur- The three enzymes in the 3N5M subfamily share the majority of ac- ther experiments are needed to clarify, whether these are the tive site residues but the enzyme from Rhodococcus opacus (UniProt ID: main factors that facilitate βAla conversion. Q6JE91 ) ( Denger et al., 2004 ) is completely different. All these enzymes However, class III transaminase enzymes developed several employ a different mechanism for substrate recognition compared to ways for βAla binding, which is found in other pyr and βAla converting other ωAA:pyr TAs, as they do not have R346 or other basic residues enzymes like tau:pyr TAs (section 3.2.5), Ala:glyox TAs 2 (section 3.3.6) at the O-side (see sequence –function matrix Table 2 ). Within the and βPhe-TAs (section 3.3.2). Owing to the different requirements for three enzymes in 3N5M subfamily the sulfonate coordination might the binding of their natural substrates these enzymes are equipped with be realised with R145 pointing towards the active site from the entrance different solutions for substrate recognition, but all allow for βAla ‘bottom ’ and R414 ( Q9APM5 numbering, first of seven amino acids be- coordination. tween core positions 349 and 350), which might also point towards the active site if this loop is folded like in the 3N5M structure. W47, which is found in these three enzymes, might additionally be involved 3.2.4. Cadaverine/putrescine:pyruvate TAs in carboxylate coordination, but the region of the left-handed helix α2 Summary: the only two known sequences have not been tested for PUT: has a glycine insertion (before core position 46) that might completely pyr activity in puri fied form and share exactly the same active site residues change the tryptophan's orientation compared to ATAs. It is not possible like the ‘high activity ’ ATA 3HMU. 3HMU, however, does not convert free to elucidate these enzymes' substrate recognition without structural in- PUT but its acylated derivatives. formation and we are therefore not able to suggest a sequence finger- For an introduction to putrescine and cadaverine, see section print for this speci ficity. 3.1.5 about PUT: αKG TAs. The knowledge of PUT:pyr TAs is more limited compared to the αKG 3.3. ω-Transaminases with unusual acceptor spectrum accepting enzymes. This activity has been found in crude extracts of Even though most characterised enzymes within the group of class 3 Exception: the enzyme from Rhodobacter sphaeroides (PDB ID: 3I5T) has W47 and III transaminases are specific for transaminations with αKG or pyr as S185. A possible explanation for its lack of βAla:pyr TA activity ( Steffen-Munsberg et al., 2013b) might be its R142 that is pointing in the active site from the ‘entrance bottom’ acceptor, there are several exceptions known as well. In this section and thus probably disrupts the proper coordination (Supplementary data Figure S5). we summarise enzymes with an ‘unusual’ substrate scope that either F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 591 employ a different mechanism for substrate recognition and therefore et al., 2005; Kim et al., 2007 ) but structural information were obtained accept both αKG and pyr or convert completely different substrates. only recently ( Crismaru et al., 2013; Wybenga et al., 2012 ) (PDB ID: 2YKY and 4AO9, respectively). As the amino donor they prefer 3.3.1. Dual substrate recognition β-amino acids such as βPhe and βAla. Owing to the instability of the cor- Enzymes that are not speci fic for αKG or pyr as amino acceptors and responding β-keto acids, these enzymes are not suitable for ef ficient accept both, in most cases established other dual substrate recognition asymmetric β-amino acid synthesis. Unfortunately the βPhe TAs do mechanisms. The βPhe TAs, which accept βAA, αKG and pyr as sub- not convert the β-keto acid's ester derivatives that would not undergo strates, apply a ‘flipping ’ arginine in the O-side (R348) for dual substrate spontaneous decarboxylation and could therefore be suitable synthesis recognition, which is similar to the mechanism in ATAs (R346, compare substrates ( Wybenga et al., 2012 ). When the β-keto acid was generated section 3.2.1). It may as well ‘flip ’ out of the active site to create space for in situ from the corresponding ester utilising a lipase, only yields up to βPhe's phenyl ring (see Fig. 16 A) and when ‘flipped in ’, coordinate 50% of βPhe could be obtained ( Bea et al., 2011 ). The biocatalytic impact αKG's 1-carboxylate in the second half reaction (see Fig. 16 B). This of these enzymes is therefore relatively limited. Recently, however, two mechanism is applied by βPhe TAs (section 3.3.2) and PhGly: αKG TAs similar enzymes have been discovered that showed high activity (section 3.3.4). towards amines ( Bea et al., 2011; Shon et al., 2014 ) in contrast to the Other class III transaminases with unusual substrate scope employ first mentioned enzyme ( Kim et al., 2007 ). Active site differences of established dual substrate recognition mechanisms for a completely these enzymes, which might explain their stronger preference for different purpose, such as DAPA-TAs, that utilise a flexible R346 for amine substrates, are not obvious from the alignment as most residues KAPA coordination but do not convert α-amino acids at all (section are conserved in all four. The only amino acid in the active site that is 3.3.5). different is Y46, which is F46 in the in the enzymes with higher ATA Some enzymes within the class III TA family do not need a dual sub- activity. Even though this exchange seems small at a first glance, it strate recognition solution at all, as they only convert α- or β-amino might in fluence ATA activity by increasing the active site's hydro- acids like alanine:glyoxylate TA 2 (Ala:glyox TA 2, see section 3.3.6). phobicity. This might also be the reason for higher activity towards Owing to the diversity of mechanisms for substrate recognition βPhe. Additionally, F46 is not able to bind water molecules as among the ‘unusual’ substrate scope TAs, further details are discussed found in the 2YKY structure, thereby gaining enhanced flexibility. in the corresponding sections. We therefore suggest that βPhe TAs with F46 might be more inter- esting for amine synthesis. 3.3.2. β-Phenylalanine aminotransferases Another TA from M. loti (UniProt ID: Q98NJ9 ) ( Kwon et al., 2010 ) Fingerprint β-Phe: αKG/pyr TAs: E45, R348 showed activity towards β-phenylalanine but this enzyme is very Summary: βPhe TAs accept both pyr and αKG due to their different different in its active site residues and belongs to the subfamily mechanism of substrate recognition. Their biocatalytic potential in asym- 3N5M (a discussion of this subfamily can be found in section 4.4). metric synthesis of β-amino acids is limited, as the corresponding keto This enzyme is rather unspeci fic and also converts diverse amines acids are spontaneously decarboxylated. We suggest that the high ATA ac- (Seo et al., 2012 ) and it is therefore not considered to be a tivity of some βPhe TAs might be caused by F46. βPhe: αKG/pyr TA. β-Phenylalanine aminotransferases ( βPhe: αKG/pyr TAs), which ac- βPhe: αKG/pyr TA's dual substrate recognition at the O-side was cept both αKG and pyr as amino acceptors, are found in subfamily found to be similar to that in the ωAA:pyr TAs (compare section 3.2.1) 4AO9. The most prominent enzymes with this speci ficity from but the ‘flipping ’ arginine is located in position 348 instead of Mesorhizobium sp. (UniProt ID: A3EYF7 ) and Variovorax paradoxus 346 (Crismaru et al., 2013; Wybenga et al., 2012 ). The recognition of (UniProt ID: H8WR05 ) had been known for several years ( Banerjee the phenyl ring and the 1-carboxylate is accomplished through a

A) βPhe: αKG/pyr TA, 2YKY B) βPhe: αKG/pyr TA, 2YKX

348 45

R54 46 47 271

242 185 216

215

129

Fig. 16. Dual substrate recognition in β-Phe: αKG/pyr TAs exempli fied by the enzyme from Mesorhizobium sp. The cofactor and substrates are coloured orange, residues outside the core (here R54 in A3EYF7 numbering) are coloured yellow. The cartoon of loops 53 –59 and 302 –311 (A3EYF7 numbering) are shown transparent for clarity reasons. A) β-Phenylalanine's carboxylate is coordinated by R54 , which is positioned by E45, while R348 is ‘flipped ’ out of the active site (PDB ID: 2YKY) B) α-Ketoglutarate is coordinated by R54 on the P-side and R348 at the O-side (PDB ID: 2YKX). 592 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 movement of the flexible R348 (compare Fig. 16 A and B). On the P-side, fashion compared to other L-α-transaminases ( Jomrit et al., 2011 ). The coordination of the 1-carboxylate of ( S)- β-phenylalanine and the 5- phenyl ring is accommodated at the O-side, while the carboxylic func- carboxylate of α-ketoglutarate is realised via salt-bridge formation tion is bound at the P-side. Until now only the two Pseudomonas species with R54 (A3EYF7 numbering, between core positions 17 and 18, see stutzeri (Wiyakrutta and Meevootisom, 1997 ) and putida (Müller et al., Fig. 16 ). An E45 controls the position of R54 's side chain ( A3EYF7 num- 2006; Townsend et al., 2002 ) have been described to possess enzymes bering) and allows for the conversion of pyruvate as it permanently with this activity. Unfortunately the attempt to solve the crystal struc- neutralises the arginine via salt-bridge formation. ture ( Kongsaeree et al., 2003 ) of the P. stutzeri enzyme was only partially As these residues are important for βPhe: αKG/pyr TA activity and all successful with important parts missing at the P-side in the final (apo) four characterised enzymes have them, we suggest the fingerprint E45 structure (PDB ID: 2CY8, Fig. 17 ), which still remains unpublished. and R348 for the identi fication of βPhe:αKG/pyr TAs among class III Owing to the missing regions in this structure it was not included in transaminases. The combination of E45 and R348 is unique for these the OrnTL DB, but it and the other sequence (see Supplementary data enzymes and the suggested pattern separates subfamily 4AO9 from all Table S5, entries 146 and 147) have been manually aligned to the closest other in the OrnTL DB. homolog structures 2GSA and 4AO9 to compare it to the OrnTL DB. The O-side of the P. stutzeri enzyme apo structure ( Fig. 17 ) is mainly folded as in the other class III TAs and is therefore assumed to be in the 3.3.3. Acyl-CoA- β-TAs active conformation. From this structure in combination with the Summary: this enzyme might be interesting for asymmetric β-amino knowledge of substrate coordination in βPhe: αKG/pyr TAs (see section acid production, if, in addition to the tested N-acetylcysteamine thioesters, 3.3.2), it can be concluded that the recognition of the 1-carboxylic group simple esters are accepted as well. The carboxylate coordination is of Glu is achieved by R348. This arginine may ‘switch ’ out of the active (in contrast to the majority of all class III transaminases) probably achieved site (as found in the apo structure) to leave a hydrophobic pocket com- by R216. prising several Phe and His residues. Unfortunately the P-side is not Recently, a metagenomic approach within a wastewater treatment present in the crystal structure and the dual substrate recognition of plant has discovered a new substrate speci ficity within the class III Glu's 5-carboxylate and D-PhGly's 1-carboxylate therefore remain un- transaminases ( Perret et al., 2011 ). The enzyme with UniProt ID clear. An E45 in combination with an arginine, which was found in β- B0VH76, whose closest homolog in the OrnTL DB is that with UniProt Phe: αKG/pyr TAs is missing in these enzymes but the uncommon ID D8F1V2 (66% identity) in the 2GSA subfamily, was shown to Q269 (only 6 sequences in the OrnTL DB) might be involved in the rec- transaminate coenzyme A (CoA) thioesters of β-amino acids with pref- ognition here. The fact that the P. putida enzyme accepts pyruvate erably αKG but also pyr and it was therefore termed Acyl-CoA β TA (Townsend et al., 2002 ) as acceptor whereas the P. stutzeri enzyme (CoAβAA TA). Its physiological role is supposed to be the conversion does not ( Wiyakrutta and Meevootisom, 1997 ) cannot be explained of β-aminobutyryl-CoA: αKG in an alternative Lys catabolic pathway in from the sequence alignment. Structural information preferably with the anaerobic lysine digester Candidatus cloacamonas acidaminovorans . bound inhibitors is highly desirable to finally understand the coordina- Substrate scope investigations showed that not the whole CoA moiety tion at the P-side. is needed for substrate recognition, but N-acetylcysteamine thioesters of β-aminobutyryl, β-homoleucine and β-phenylalanine were also ac- 3.3.5. Diamino pelargonic acid transaminases cepted. Unfortunately no other esters have been investigated so far, Fingerprint SAM:KAPA TAs: Y129, D/E132, R346, Y353 but if the enzyme would also convert e.g. ethyl esters ef ficiently, it Fingerprint Lys:KAPA TAs: Y129, D/E132, R346, not Y353 could be valuable for asymmetric β-amino acid synthesis. Owing to Summary: DAPA TAs are highly speci fic due to fine tuned substrate rec- the instability of the free β-keto acids and the low activity of known ognition. The discrimination between 7- and 8-amino group of DAPA and β-Phe: αKG TAs towards the corresponding esters ( Wybenga et al., between (8R)- and (8S)-DAPA is achieved by a hydrogen bonding 2012 ) (see section 3.3.2), this acyl-CoA- β-TA could be a bene ficial alter- network of Y129, D/E132 and in some cases Y16. Y353 is proposed to deter- native for biocatalysis. mine SAM coordination and is therefore suggested to discriminate Lys: By aligning the sequence (UniProt ID: B0VH76 ) to its closest homo- KAPA TAs from SAM:KAPA TAs. log in the OrnTL DB some of its active site residues could be predicted Diamino pelargonic acid transaminases (DAPA TAs) convert 7-keto- (see sequence–function matrix Table 2) but the substrate recognition 8-aminoperlargonic acid (KAPA) to 7,8-diamino pelargonic acid (DAPA) in this enzyme cannot be elucidated without structural information as employing either S-adenosyl- L-methionine (SAM) or L-lysine (for sub- it is too different from other class III TAs. The only part that we dare to strate and product structures, see Table 3 ) as the amino donor and hypothesise is the carboxylate coordination at the O-side by R216, thereby play an important role in the biotin biosynthesis ( Mann and which is the only basic residue there. Ploux, 2011 ). Owing to the limitation of biotin anabolic pathways to only plants and bacteria, DAPA-TAs are potential antibiotic targets. 3.3.4. D-p-hydroxyphenylglycine: αKG TAs Therefore several studies focusing on selective inhibition of these Summary: PhGly: αKG TAs are the only class III transaminases accepting enzymes have been conducted ( Mann et al., 2009 ). solely α-amino acids. A dual substrate recognition mechanism is, however, Most known DAPA-TAs convert SAM and KAPA to S-adenosyl-2- required as both substrates are bound in inverted orientation. The thereby oxo-4-methylthiobutyric acid (which undergoes β-elimination to result gained enantioselectivity for one (S)- and one (R)-amino acid is virtually in 5-methylthioadenosine and 2-oxo-3,4-butenoic acid) and DAPA unique among transaminases. Substrate recognition at the O-side is (Stoner and Eisenberg, 1975 ). These SAM:KAPA-TAs (EC 2.6.1.62) achieved like in β-Phe TAs. have been discovered in several bacteria and plants (Supplementary D-Phenylglycine ( D-PhGly) and its p-hydroxy derivative are building data Table S5, entries 148 –158, 160 and 161). The only example blocks for the commonly used β-lactam antibiotics ampicillin and where a DAPA transaminase did not utilise SAM as amino donor was amoxicillin, respectively. Biocatalytic approaches for their production found in Bacillus subtilis (PDB ID: 3DU4) ( Van Arsdell et al., 2005 ). usually start from the carbamoyl, involving D-selective carbamoylases This enzyme is closely related to SAM:KAPA-TAs but employs L-lysine and hydantoinases ( Wegman et al., 2001 ). A second option is the appli- as the amino donor and will therefore be referred to as Lys:KAPA-TA cation of D-p-hydroxyphenylglycine: αKG TAs (PhGly: αKG TA, EC (EC 2.6.1.105). Unfortunately, it remains unclear whether this enzyme 2.1.6.72) in asymmetric syntheses from the corresponding keto acids transaminates the α- or ε-amino group of lysine. The transfer of the (Müller et al., 2006 ). These class III transaminases differ from most ε-amino function, however, seems more likely owing to the resulting other TAs by converting substrates with inversed enantio preference. irreversibility of the overall reaction. As described for Lys ε:αKG TAs, Recently it was proven that these enzymes bind D-PhGly in an inverted the formed product, Δ1 piperideine-6-carboxylic acid, would undergo F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 593

D-PhGly: αKG TA, 2CY8 to ( S)-con figuration, results in the amino function pointing away from PLP. Even though unlikely, the formation of the external aldimine might still be possible in this orientation but the deprotonation of the 43 substrate by the catalytic K242 to form the quinonoid intermediate (see Fig. 2 ) is de finitely not. Since there is no base available on the re - side of the cofactor that could do the deprotonation instead, the carbox- ylic group of SAM and sinefungin would have to be located at the O-side 46 353 to enable transamination. The second reason to doubt SAM binding being represented in the 3LV2 structure is an unexplained mutation at 348 the active site entrance (H268R), which substantially reshapes the 325 P-side, while thirdly, the substrate analogue differs from SAM in an im- portant feature. An amino group replaces the methyl group at the sulfo- 216 nium ion in SAM and the sulphur itself is replaced by a carbon, which 186 modifies the electronic properties of the analogue and therefore most 215 probably also its binding. The importance of the sulfonium ion for 185 substrate recognition was proven by the absence of activity for S-adenosylhomocysteine, which is a SAM analogue lacking the methyl group at the sulphur ( Dey et al., 2010 ). Owing to the high conservation of Y353 in all SAM:KAPA-TAs, it was hypothesised that the sulfonium ion is coordinated by the hydroxyl group of this tyrosine ( Dey et al., 2010 ). This proposal can be strengthened by the fact that the only characterised DAPA-TA lacking the tyrosine in that position is the Lys: Fig. 17. O-side of the D-PhGly: αKG TA from Pseudomonas stutzeri apo structure (PDB ID: 2CY8) . The PLP, taken from the aligned β-phenylalanine: αKG TA structure (PDB ID: KAPA-TA from B. subtilis . One could, therefore, hypothesise that the 4AO9), is shown transparently in orange for orientation reasons. presence of Y353 might be employed to distinguish the SAM and Lys converting DAPA-TAs.

3.3.6. Alanine:glyoxylate transaminase 2 internal Schiff-base formation and therefore virtually disable the back Fingerprint Ala:glyox TAs2: Y/W145, F269, R353, NOT D/E185 reaction ( Soda et al., 1968; Van Arsdell et al., 2005 ). Summary: Ala:glyox TA 2, in contrast to most other class III transami- All characterised DAPA-TAs have a very narrow substrate scope and nases, does not require a dual substrate recognition mechanism at the O- only convert substrate analogues that are very similar to the natural side because only αAA and βAA are converted. The broad substrate scope ones ( Cobessi et al., 2012; Izumi et al., 1975; Izumi et al., 1981; Mann includes small αAA such as Ala, βAA, such as βAla and 3- and Ploux, 2006; Stoner and Eisenberg, 1975; Van Arsdell et al., 2005 ). aminoisobutyrate (AIB) but also bulky αAA such as N G,N G- Owing to the narrow substrate scope and to their low activity (e.g. a dimethylarginine. For the conversion of βAA a (R)-enantioselectivity at −1 steady-state k cat of 0.013 s for the E. coli enzyme (PDB ID: 1MLZ) the α-position was found (selectivity for DAIB). (Eliot et al., 2002 )) these enzymes are of low value for biocatalytic The mammalian mitochondrial Ala:glyox TA 2 (AGXT2, EC 2.6.1.44) applications. But this perfect selectivity combined with the fact of is one of the class III transaminases with the broadest substrate scope being absent in humans makes them valuable targets for potential but is limited to α- and β-amino acids and therefore requires no dual antibiotic substances. substrate recognition at the O-side. Owing to the speci ficity for several Compared to other class III transaminases, DAPA-TAs selectively con- key metabolites, its annotation proved to be dif ficult: the rat Ala:glyox vert extraordinary substrates and therefore need characteristic patterns TA 2 (UniProt ID: Q64565 ) had previously also been described as to bind these. In particular, the recognition of KAPA needs to be finely D-3-aminoisobutyrate:pyr TA (EC 2.6.1.40), Ala:4,5-dioxopentanoate tuned to allow the discrimination between the 7-keto and 8-amino TA (EC 2.6.1.43), β-Ala:pyr TA 2 (EC 2.6.1.18), 2-aminobutyrate:pyru- group. By selectively coordinating the 8-amino group and keeping it vate TA and NG,NG-dimethylarginine:pyr TA ( Kontani et al., 1993; away from the C4 ′ atom of PLP, the enzymes are able to prevent side Tamaki et al., 2000 ). This enzyme is supposed to be the mitochondrial reactions at this position. The positioning is achieved by Y129 and addi- counterpart of the peroxisomal Ala:glyox TA 1 (AGXT1, EC 2.6.1.44) tionally either Y16 or the carbonyl oxygen of residue 269 that bind the which keeps physiological glyoxylate levels low and is involved in the 8-amino group (see Fig. 18 ). A highly conserved aspartate at position hyperoxaluria disease ( Baker et al., 2004 ). Although both enzymes 132, coordinates the tyrosine(s) and thereby keeps them in the right share the ability to catalyse the Ala:glyox conversion, their sequences place ( Käck et al., 1999 ). This finely tuned recognition even allows the and substrate spectra differ substantially. AGXT1, also referred to as enzyme to discriminate between the ( R)- and ( S)-con figuration at C8 Ser:pyr TA (EC 2.6.1.51), which only shares 12% identity with AGXT2 of KAPA ( Mann et al., 2009 ). On the O-side of the active site, the carboxyl (UniProt IDs: P21549 and Q9BYV1 in humans, respectively), belongs group of KAPA forms a salt bridge to R346 (same as the ‘flipping ’ arginine to a different class of transaminases (class V) and its substrate scope fa- in ωAA:pyr TAs, section 3.2.1), which is also highly conserved (Eliot et al., vours small α-amino or keto acids (i.e. Ser, pyr, glyox), but Phe Arg and 2002; Wybenga et al., 2012 ). Since these features are unique among the Glu are also converted by the human Ser:pyr TA ( Cellini et al., 2007 ). class III transaminases, KAPA converting enzymes can be selectively These enzymes are of special interest regarding the connection between identi fied by the presence of Y129, D132 and R346. sequence and function as Ala:glyox TAs 2 and Ser:pyr TAs both convert Even though SAM binding has been studied extensively, its exact ori- Ala:glyox transamination despite their low sequence similarity, where- entation in the active site still remains unclear. Dey et al. (2010) have as a close homolog to Ala:glyox TA 2 in Arabidopsis thaliana differs in its been able to solve a structure of the Mycobacterium tuberculosis enzyme substrate spectrum and additionally catalyses the Glu:glyox conversion containing the SAM analogue sinefungin. Unfortunately this structure (Liepman and Olsen, 2003 ). (PDB ID: 3LV2) was not sufficient to unravel the binding of SAM. The The Ala:glyox TAs 2, which are found in the 3N5M subfamily, also orientation of sinefungin in this structure must represent an unproduc- convert β-amino acids like βAla and D-3-aminoisobutyrate ( DAIB) tive binding mode for the following three reasons. The first and most where they showed strict ( R)-selectivity for the α-methyl group. important reason is motivated by the enantioselectivity of the reaction. Its DAIB:pyr TA activity together with the enantiocomplemen- The carboxylic group of sinefungin is located at the P-side, which, owing tary LAIB: αKG TA (EC 2.6.1.22, which is identical to the above described 594 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604

A) SAM:KAPA TA, 1QJ3 B) Lys:KAPA TA, 3DU4

346

46 47 353 271 16 242

216

185 132

129

Fig. 18. Coordination of KAPA in A) SAM:KAPA TA (enzyme from E. coli PDB ID: 1QJ3) and B) Lys:KAPA TA (Enzyme from B. subtilis, PDB ID: 3DU4). PLP and KAPA are shown in orange.

GABA: αKG TA EC 2.6.1.19, see section 3.1.4), interconvert both enantio- Glutamate-1-semialdehyde 2,1-aminomutase (GSAM, EC 5.4.3.8) is mers of AIB in mammals ( Tamaki et al., 2000 ). In contrast to the a unique enzyme within the class III transaminase family, owing to the GABA: αKG TAs, the Ala:glyox TAs 2 do not convert amino groups fact that it does not catalyse the inter conversion between an amino- more distant from the carboxylic function than in β-position ( Tamaki and a keto function of two molecules. GSAM instead catalyses the et al., 2000 ). intra conversion between an amino- and a keto function within the The conversion of these various substrates raises the question how same molecule i.e. the conversion of glutamate-1-semialdehyde to substrate recognition is realised in these enzymes. There is unfortunate- 5-aminolevulinate ( Hoober et al., 1988 ). Another unique feature of ly no structural information for Ala:glyox TAs 2 available and therefore GSAM is that the resting state of this enzyme is in the PMP-form while only hypotheses based on conserved positions within these enzymes the resting state of the majority of the class III transaminases is in the can be developed. The coordination of the carboxylic group of α- PLP-form ( Smith et al., 1991 ). This feature might be induced by the amino acids is probably undertaken by the conserved R353 like in the highly conserved N185 (see Fig. 19 ), which is supposed to modify the ω-amino acid: αKG TAs (see section 3.1.1), but in contrast to those the cofactor's electron sink properties by coordinating its 3 ′-hydroxyl ‘switching ’ glutamate in position 185 is replaced by Val or Ile. This group ( Orriss et al., 2010 ). To which extent the coordination by N185 might also contribute to the fact that the common dual substrate recog- is different from that of Q216, as present in the ωAA: αKG TAs nition is not possible and no ω-amino acids (except for β-amino acids) where the internal aldimine is the resting state (see section 3.1.1), can- are converted. The unique feature of these enzymes within the class III not be predicted without mutagenesis studies. TA family, namely the conversion of huge α-amino acids like GSAM belongs to the subfamily 2GSA, which consists of 2048 se- dimethylarginine and the enantioselectivity at the α-position when quences. Out of these entries, 16 enzymes were experimentally converting the β-amino acid DAIB, might be explained by the unique characterised and several crystal structures have been solved (Supple- combination of residues Y145 and F269 (only 21 sequences in the mentary data Table S5 entries 170 –185) ( Ge et al., 2010; Hennig et al., OrnTL DB). The aromatic ring of F269 at the P-side might be involved 1997; Schulze et al., 2006; Stetefeld et al., 2006 ). Most of these struc- in cation –π interactions with the guanidinium group of tures showed a third unique feature of GSAM when compared to the dimethylarginine, whereas Y145 that points from the entrance towards other class III TAs, an active site gating loop (residues between core the active site and might be responsible for the recognition of the 4- positions 131 & 155, which cannot be aligned to the OrnTL DB) which carboxylic group of aspartate. This hypothesis is strengthened by the opens and closes the entrance to the active site during catalysis fact that glutamate, due to its additional carbon, is not accepted as sub- (Contestabile et al., 2000; Stetefeld et al., 2006 ). The movement of the strate but when Y145 is exchanged to a tryptophan like in two Ala:glyox gating loop is controlled by allosteric communication between the TA 2 homologous enzymes from Arabidopsis , glutamate is converted as two active sites of the homodimeric enzyme. The ability to close the ac- well ( Liepman and Olsen, 2003; Tamaki et al., 2000 ). tive site gives GSAM the ability to control substrate entry and product release. The intermediate 4,5-diaminovalerate is kept in the active site to avoid its release as unwanted product. The general opinion about 3.4. Glutamate-1-semialdehyde transaminases (2,1-amino mutases) the gating loop is that it is controlled through negative cooperativity: when the substrate binds to one subunit the af finity for the substrate Fingerprint GSAM: N185, Y267 decreases in the other subunit ( Hennig et al., 1997; Stetefeld et al., Summary: GSAM transaminates glutamate-1-semialdehyde intramo- 2006). This leads to the enzyme converting the substrate one subunit lecularly to form 5-aminolevulinate. A gating loop closes the active site en- at the time. This negative cooperativity has been shown to be controlled trance after substrate binding to prevent the intermediate diamine's by the communication of a network of the amino acids S98, E101, R108, release. After product formation the entrance is opened again. GSAM is H153, D155 (Q31QJ2 (from Synechococcus elongatus ) numbering the only known class III TA that uses PMP in its resting state in contrast to between core positions 131 and 155) and R111 that connects the active PLP (internal aldimine) in other enzymes. sites ( Stetefeld et al., 2006 ). This theory is, however, challenged by F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 595 a published structure where both monomers are in the PMP-form catalyses the decarboxylation and transamination of dialkylglycine sub- (Ge et al., 2010 ). strates with an α-keto acid (e.g. pyr) as the amino acceptor ( Sun et al., GSAM is, like many other PLP-dependent enzymes, inhibited by 1998 ). This enzyme is unique within PLP fold type I because it can cleave gabaculine ( Grimm et al., 1991 ). This inhibition has, however, been both the substrate's C α–CO 2 bond and the C α–H bond in the same shown to decrease signi ficantly when an active site methionine is mu- active site. Furthermore, it catalyses an oxidative decarboxylation tated to an isoleucine (M216I) ( Ferradini et al., 2011; Grimm et al., (forming α-keto product), which is in contrast to most other amino 1991; Orriss et al., 2010 ). Since GSAM is not present in animals it is a acid decarboxylases where non-oxidative decarboxylation is favoured promising target for herbicides and antibiotics and this mutation also (forming an amino product) ( Li et al., 2012 ). By protonating the C4 ′ of makes selective growth via the addition of gabaculine possible the quinonoid intermediate after decarboxylation instead of the C α as (Hennig et al., 1997 ). in non-oxidative decarboxylases, the ketimine is formed, which leads As many residues of GSAM that are involved in substrate binding be- to transamination and release of the keto product from PMP. PMP's long to the variable regions, establishment of an active site fingerprint amino group is then transferred to an amino acceptor in the second, only based on core positions was not trivial. We suggest applying the transamination half reaction restoring the internal aldimine (see Fig. 1 ). two amino acids N185 and Y267 to identify enzymes performing the in- The selectivity over the site of protonation (C α over C4 ′ or vice tramolecular transamination of glutamate-1-semialdehyde to form versa) has been investigated in both non-oxidative and oxidative 5-aminolevulinate. N185 modulates the cofactors electronic properties decarboxylases ( Bertoldi et al., 2008; Jackson et al., 2000; Sun et al., (Orriss et al., 2010 ) and Y267 is involved in a hydrogen bonding 1998 ). Sun et al. (1998) proposed an active site model for the DGD network with the substrate and with S163 (Q31QJ2 numbering, in the from Pseudomonas cepacia , whereby the site of protonation by the gating loop) (see Fig. 19 ) ( Stetefeld et al., 2006 ). Additional hints for catalytic lysine is controlled by the tilt of the cofactor, which in turn, is GSAM activity could be R108, which is highly conserved in the subunit in fluenced by the identity of the substrate. The model consists according communication network ( Stetefeld et al., 2006 ) involved in inter- to the Dunathan principle (see section 1.3) of three binding subsites: subunit signalling and the presence of a gating loop between core posi- (A) the activated position for bond cleavage (may bind Cα–H or C α– tions 131 –155 that cannot be aligned to the other known class III TAs. CO 2), which is perpendicular to the plane of the cofactor and located at the si -face; (B) a second carboxylate binding subsite at the O-side of 3.5. Decarboxylation dependent TAs: the 2,2-dialkylglycine decarboxylases the cofactor and (C) a hydrophobic subsite at the P-side of the cofactor (see Figs. 1 and 20 ). For α-H- α-amino acids, for which both reactions Fingerprint DGD: Q46, (W129), R353 are possible, the accommodation of the substrate's carboxylate in the Summary: DGD is the only known PLP fold type I enzyme that catalyses A subsite would lead to decarboxylation, whereas binding in the B sub- both, transamination and oxidative decarboxylation to a comparable ex- site would place the C α–H in the A pocket and would lead to deproton- tent. The reaction speci ficity is determined by the orientation of the ation and therefore to transamination. However, dialkylglycine substrate's substituents relatively to the cofactor's plane: if the C α-H is substrates cannot be transaminated due to their lack of a C α–H. There- pointing towards the si-face, transamination is preferred, if C α–CO 2 is fore binding of their carboxylates in the A subsite leads to decarboxyl- pointing towards the si-face, decarboxylation is favoured. ation, while binding in the B subsite would represent an unreactive In addition to the above-mentioned transaminases, the family of binding mode. Structural investigations with phosphonate substrate an- class III TAs also comprises the decarboxylation dependent transami- alogues, however, revealed that the position of the scissile carboxylate is nase, 2,2-dialkylglycine decarboxylase (DGD, EC 4.1.1.64). This enzyme probably not perfectly perpendicular, as the orbital overlap with the

A) GSAM, 2HP1 B) GSAM, 2HP2

R26

46 S23

271 E400

242

S157 185

267 216

215

129

Fig. 19. Substrate and product binding in GSAM. The substrate-cofactor intermediates are shown in orange and residues that belong to the variable regions ( S29 , R32, S163 and E406 (Q31QJ2 numbering)) are shown in yellow. The cartoon is shown transparent for clarity reasons. A) After binding glutamate-1-semialdehyde, the gating loop is closed to prevent the re- lease of the intermediate 4,5-diaminovalerate (PDB ID: 2HP1). Y267 indirectly coordinates the substrate's carboxylate and the gating loop's S163 (Q31QJ2 numbering) B) The gating loop opens when 4,5-diaminovalerate is bound at its 4-amino group, opening up the active site entrance. Thereafter the product 5-aminolevulinate is formed through water attack and may be released from the active site (PDB ID: 2HP2). 596 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 cofactors π–electron system would suggest (see Fig. 1 ), but is tilted to- sequences found in the 1D7V subfamily. The selectivity of the finger- wards the O-side and placed between the subsites A and B (see print for the 1D7V subfamily was further improved to 100% by including Fig. 20 B) ( Liu et al., 2002 ). the substrate speci ficity determining residue W129. Studies of the crystal structure of the DGD from P. cepacia (PDB ID: Sequences possessing the functionally important Q46 but not the 1D7V) led to the identi fication of a glutamine residue (Q46) within substrate scope limiting W129 (i.e. Y or F129), could have the same van der Waals distance of the substrate's carboxylate, which could reaction speci ficity but differ in their substrate speci ficity. maintain the carboxylate in the stereoelectronically activated position Sequences in the 1D7V subfamily, which do not possess the func- via hydrogen bonding (see Fig. 20 ) ( Fogle et al., 2005 ). This binding tionally important Q46 (58 sequences), could have a different reaction model was supported by mutagenesis studies which showed that sub- speci ficity. However, these sequences have a highly conserved N47 stitution of Q46 with amino acids incapable of forming stabilising and K353. One could imagine that these residues function in the same H-bonds resulted in a decreased rate of decarboxylation ( Fogle et al., way as Q46 and R353 in the DGD from P. cepacia , especially as the 2005 ). Further investigation of the substrate scope, in combination R353K mutant of this enzyme retained decarboxylase activity ( Fogle with mutagenesis experiments, identi fied two residues (M132 and and Toney, 2010 ). Therefore, it is possible that this subset includes W129), as being important for substrate speci ficity. Mutation of W129 decarboxylases, which have either a different substrate scope or follow to the smaller phenylalanine allowed conversion of α-keto acids with a different reaction speci ficity after decarboxylation (compare Fig. 2 ). longer side chains, while mutation of M132 to a positively charged argi- nine conferred the ability to decarboxylate L-glutamate (this can be compared to αKG coordination in ωAA: αKG TAs as described in section 3.6. α-H-amino acid amide/ α-amino- ε-caprolactam racemases 3.1.1). Additionally, a proposed non-reactive binding mode, whereby the carboxylate of the dialkylglycine substrate is placed in pocket B Fingerprint αAAA racemases: D185, K216, (E353) (on the O side of the PLP), and forms a salt bridge with R353, was also Summary: αAAA racemases racemise cyclic and noncyclic amino acid investigated by mutagenesis studies (the R353M and R353K mutants amides. The mechanism proceeds, in contrast to PLP fold type III racemases, have been investigated). Quite interestingly, rather than increasing the via a quinonoid intermediate. How the (de)protonation at the re-face is rate of decarboxylation by disfavouring this non-reactive binding achieved remains unknown. mode by introducing a nonpolar M353, these studies demonstrated The subfamily 2ZUK contains two enzymes that were found to that a positively charged residue at position 353 was required for decar- possess α-H-amino acid amide ( αAAA) racemase activity. The enzyme boxylation ( Fogle and Toney, 2010 ). from A. obae (UniProt ID: Q7M181 ) was initially characterised as As only two sequences of DGDs have been characterised, it could be α-amino- ε-caprolactam (ACL) racemase (EC 5.1.1.15) and was industri- dangerous to make presumptions about functionally important ally applied for L-lysine production ( Fig. 3) ( Okazaki et al., 2009 ). The residues, however, as these enzyme have been extensively and elegant- enzyme from Ochrobactrum anthropi (UniProt ID: Q06K28 ) was patent- ly investigated by various research groups, some conclusions and ed for the application in amino acid amide racemisation ( Boesten et al., predictions can be made ( Fogle and Toney, 2010; Fogle et al., 2005; 2003 ). The third known sequence of an αAAA racemase, which was Hohenester et al., 1994; Keller et al., 1990; Liu et al., 2002; described in the same patent, is not found in the sequence databases Malashkevich et al., 1999; Sun et al., 1998; Toney, 2001; Toney et al., but could be aligned to the 2ZUK subfamily manually. As the ACL 1993; Toney et al., 1995 ). racemase accepts both lactams and amides ( Asano and Yamaguchi, Residues essential for decarboxylation dependent transamination 2005), the structural requirement for being a substrate seems to be a are Q46, R353 and the catalytic K242. Residue Q46 is fairly unique to free amino group adjacent to an amide bond ( Ahmed et al., 1984; the DGDs, and a motif search using this residue was 91.9% selective for Asano and Yamaguchi, 2005 )

A) DGD, 1D7V B) DGD, 1M0O

46

47 353 269 271 242

216

185 215

129

Fig. 20. Different substrate coordination mode in DGD determines reaction speci ficity towards transamination or decarboxylation. A) By the coordination of the substrate's carboxylate in the subsite B, transamination of α-H-α-amino acids or unproductive binding of 2,2-dialkylglycine substrates (such as the displayed 2,2-dimethylglycine) is achieved (PDB ID: 1D7V). B) By the coordination of the carboxylate in subsite A, decarboxylation is favoured. The displayed structure (PDB ID: 1M0O) contains a phosphonate analogue of 2-methyl-2- ethylglycine with the phosphonate group located between subsites A & B. The same orientation of a substrate's carboxylate would lead to a productive binding for decarboxylation. F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 597

Racemases are unique among the other types of PLP-dependent en- recognition is most probably achieved, as in GABA: αKG TAs, with zymes, as they are capable of deprotonating and protonating on both R353. The extent to which the relatively uncommon D185 and N216 the si - and the re -face. For the ACL racemase two mechanisms have (only present in 30 sequences in the whole OrnTL DB) or A46 and S47 been proposed, a two base mechanism (where two acid –base groups (found in only 109 sequences in the whole OrnTL DB) are involved can- are situated on either side of the substrate –PLP complex), and a single not be predicted without further experimental data. Notably, position base mechanism (with a single base capable of accessing both faces). 185 seems to be important for racemase activity within the class III Ahmed et al. (1986) found evidence for a single base mechanism, transaminase family because the αAAA racemases and the isoleucine which is at odds with the two base mechanism described by Okazaki 2-epimerase share the uncommon D185 (present in 234 of the 12,956 et al. (2009) which was proposed based on structural information. By sequences in the OrnTL DB). Additionally, position 216 is harbouring a analogy to alanine racemases in PLP fold type III, Okazaki et al. (2009) very uncommon K216 or N216 in the two racemases (only 28 se- proposed Y129 at the re -face and K242 at the si -face to enable quences have K216 and only 30 have N216 in the OrnTL DB). racemisation in the ACL racemase (see Fig. 21 ). The mechanism of ACL racemases, in contrast to alanine racemases, is believed to proceed via 3.8. Enzymes with unclear substrate recognition a quinonoid intermediate because the cofactor's pyridine nitrogen is kept protonated by the D213 ‘below ’ it ( Okazaki et al., 2009 ), whereas In this section we grouped recently discovered class III transaminase it is deprotonated by an arginine residue in PLP fold type III racemases. enzymes, for which not suf ficient structural or functional data is Fold type III racemases do not require the quinonoid formation for available to suggest speci ficity determining residues. In particular, stabilising the carbanionic intermediate: the negative charge is mainly the aminosugar TAs (section 3.8.1) and the multi-domain enzymes stabilised at the protonated Schiff base ( Griswold and Toney, 2011 ). (section) share too low sequence similarity with any known structure The above-mentioned role of Y129 at the re -face for protonation in to reliably align their sequences to the OrnTL DB. These examples, the racemisation mechanism has not been con firmed by mutagenesis above all, highlight the lack of structural information within the class studies and the fact that Y129 is found in 58% of the sequences in the III transaminase family. OrnTL family (e.g. in ATAs, DAPA TAs and βPhe TAs), places doubt on its suitability as a candidate as a 2nd base (if indeed a two base mecha- 3.8.1. Neamine TAs, 2 ′-deamino-2 ′-hydroxyneamine and neomycin C TAs nism is followed by these enzymes). Since the mechanism in the αAAA Summary: Glu:6 ′-oxoglucos(amin)yl TAs are the only known racemases is not yet fully elucidated, no residues determining the reac- aminosugar converting TAs among the class III transaminases. They are in- tion speci ficity may be included in the fingerprint and we therefore volved in aminoglycoside biosynthesis, by selectively aminating 6 ′- focus on the substrate speci ficity determining residues. oxoglucos(amin)yl moieties. Okazaki et al. (2009) proposed that amide nitrogen recognition in Aminosugar converting transaminases are commonly found to be- this enzyme is achieved by D185 (see section 3.7 for a discussion of long to the class VI transaminases (degT/dnrJ/eryC1 family in InterPro), this residue) and that the carbonyl O is coordinated by K216. In the ε- which were recently reviewed by Romo and Liu (2011) . The biosynthe- caprolactam bound internal aldimine structure (PDB ID: 2ZUK), K216 sis of several aminoglycoside antibiotics, such as neomycin, butirosin is kept in place by the coordination of E353 (see Fig. 21 ). The two resi- and kanamycin however, was found to involve class III transaminases dues D185 and K216 that are believed to be essential for substrate rec- to aminate the C6 atoms of the glucos(amin)yl substituents with ognition in the characterised enzymes (and optionally E353), are glutamate as amino donor. The three different substrate speci ficities therefore suggested for the active site pattern identifying αAAA Glu:6 ′-dehydroparomamine TA (EC 2.6.1.93, forming neamine), racemases. As all 18 sequences in the OrnTL DB with K216 and D185 Glu:2 ′-Deamino-2 ′-hydroxy-6 ′dehydroparomamine TA (EC 2.6.1.94, also have E353, including this glutamate in the fingerprint is not forming 2 ′-deamino-2 ′-hydroxyneamine) and Glu:6 ‴-deamino-6 ‴- necessary. oxoneomycin C TA (EC 2.6.1.95, forming neomycin C), that may be summarised as Glu:6 ′-oxoglucos(amin)yl TAs, have been found in 3.7. Isoleucine 2-epimerase

Summary: Ile-2-racemisation probably, like in the αAAA racemases, also proceeds via a quinonoid intermediate. The racemisation mechanism ACL racemase, 2ZUK is also unknown but we suggest that D185 is important for racemase activity in the class III transaminase family because it was found in all so far characterised enzymes. 46 The recent discovery of an isoleucine-2-epimerase (Ile-2-epimerase, 271 353 from Lactobacillus buchneri (UniProt ID: F4FWH4 )) which racemises the C2 (=C α) in aliphatic α-amino acids, further provided insight to the versatility of the family of class III transaminases ( Mutaguchi et al., 242 2013 ). This enzyme is especially interesting because it shares relatively 216 high sequence similarity with GABA: αKG TAs (e.g. 41% sequence iden- tity to B. subtilis GABA:αKG TA (UniProt ID: P94427)) and is found in the 3Q8N subfamily but most of the active site residues that are impor- 185 tant for substrate recognition differ from the GABA:αKG TAs (see se- quence –function matrix Table 2 ) and therefore GABA and αKG are not 215 transaminated by this enzyme. Owing to the lack of structural informa- 129 tion the mechanism of racemisation cannot be elucidated and from the alignment it is not obvious how the protonation/deprotonation of the substrate at the re -face is achieved (the possible involvement of Y129 therein is discussed for the αAAA racemases in section 3.6). The cataly- 213 sis of this enzyme, however, seems to be comparable to the αAAA racemases because it also has D213, which presumably leads to PLP's Fig. 21. Substrate recognition in ACL racemase exempli fied by the ε-caprolactam bound pyridine nitrogen protonation and therefore a mechanism that pro- internal aldimine structure (PDB ID: 2ZUK). The cofactor and ε-caprolactam are show in ceeds via a quinonoid intermediate. The substrate's 1-carboxylate orange. 598 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 different aminoglycoside producing organisms (see Supplementary Q8IUZ5 ) possess no TA activity but are a O-phosphoethanolamine data Figure S4 for substrate and product structures). As the EC 2.6.1.93 phospholyase (EC 4.2.3.2) and a 5-phosphohydroxy- L-lysine phos- and EC 2.6.1.95 reactions are catalysed by the same enzymes (e.g. pholyase (EC 4.2.3.134), respectively. Both enzymes do not belong to NeoB, BtrB and LivB, see Supplementary data Table S5 entries the OrnTL DB, but enzymes with high sequence identity ( N68%) are 192 –194; Clausnitzer et al., 2011; Huang et al., 2007 ), and the EC found in the 2ZUK subfamily, which allowed for the elucidation of sev- 2.9.1.93 also occurred in the kacL enzyme that was originally eral of their active site residues by aligning them to the 3DM database characterised for EC 2.6.1.94 speci ficity (see Supplementary data (see sequence –function matrix Table 2 ). The active site of these en- Table S5 entry 19; J.W. Park et al., 2011 ), we suggest that all three zymes must substantially differ from the other class III transaminases Glu:6 ′-oxoglucos(amin)yl transaminations (EC 2.6.1.93 –95) are because it lacks two amino acids in the common left-handed helix α2 catalysed by one class III transaminase enzyme. Interestingly, the C6 ′ (positions 45 and 46). At least one of the polar side chains of N/S44 is deaminated with αKG as the acceptor in the pseudodisaccharide and N47 (found in both enzymes) are presumably pointing towards neamine and pseudotrisaccharide kanamycin A ( J.W. Park et al., the active site. Whether this deletion and the polar amino acids at posi- 2011 ), but not when incorporated in neomycin C, where the C6 ‴ is ex- tions 44 and 47 are required for phospholyase activity remains unclear clusively deaminated (compare Supplementary data Figure S4) without structural information. The PLP coordination and the catalytic (Huang et al., 2007 ). machinery is the same as in the transaminases of the OrnTL DB, proba- Unfortunately these enzymes share low sequence identity with the bly because the β-elimination follows the same mechanism until the enzymes in the OrnTL DB and therefore the alignments to compare ac- quinonoid intermediate (see Fig. 2 ) and therefore has comparable re- tive site residues for substrate recognition hypothesis gave unsatisfacto- quirements. Owing to the high similarity between these two reaction ry results. Structural information of Glu:6 ′-oxoglucos(amin)yl TAs is speci ficities and (given a good leaving group) the facility of this reaction, highly desired to explain the uncommon substrate scope that allows β-elimination is commonly found among transaminase families ( Eliot for the discrimination of C6 ′ and C6 ‴ in neomycin C. and Kirsch, 2004 ). β-Elimination was also found to occur in class III transaminases when certain inhibitors, especially those with good leav-

3.8.2. (Hydrolysed) fumonisin B 1 TAs ing groups were applied as substrates (e.g. Ala:glyox TA 2 catalyses Summary: (H)fum B 1 TAs were applied to detoxify the poly-hydroxy β-elimination of β-chloro- β-alanine and halogenated alkene –cysteine amine fumonisin B 1. Further substrate scope investigations might be worth- conjugates; Cooper et al., 2003 ). Several known transaminase inhibitors while, as these enzymes could have potential for amino-alcohol synthesis. inactivate the enzymes by forming reactive products through

The carcinogenic mycotoxin fumonisin B 1 (Supplementary data β-elimination that can covalently bind to the cofactor or active site res- Figure S6 B) is a common contaminant of maize produced in warm idues ( Eliot and Kirsch, 2004 ). climate areas ( Hartinger et al., 2011 ). The studies of its degradation in Given the fact that phosphate is an excellent leaving group, the oc- bacteria has led to the discovery of class III transaminases responsible currence of β-elimination with β-phosphate substrates in a class III TA for the deamination of fumonisin B 1 (fumB 1) ( Leslie et al., 2004 ), or is not surprising. The selectivity for β-elimination of these enzymes, after hydrolysis, to 2-amino-12,16-dimethylicosane-3,5,10,14,15- caused by the complete lack of TA activity, however, is of special pentol (hydrolysed fumonisin B 1, HfumB 1, Supplementary data interest. As the whole catalytic TA machinery is available, other factors Figure S6 C) ( Hartinger et al., 2011; Heinl et al., 2011 ) with αKG or in these enzymes must prevent transamination. Since these two pyr as the acceptor, respectively. These enzymes have been investigated phospholyases are the only characterised enzymes of the class III trans- for the application as food additives to detoxify fumB 1 biocatalytically aminase family that lack the common left-handed helix at positions (Leslie et al., 2004; Moll et al., 2010 ). Unfortunately, the scope of 44 –47, it might be hypothesised that it is essential for TA activity amino donors has not been examined and it is unknown, whether sub- among this group of enzymes. The characterisation of additional en- strates other than (hydrolysed) fumonisins are converted. A further in- zymes with deletions in this region, however, is required to strengthen vestigation of their substrate speci ficity might be bene ficial as they this hypothesis. could probably be applied for asymmetric amino alcohol synthesis. A varied carboxylate recognition is probably not the reason for the Even though they share high sequence similarities, the substrate scope differing reaction speci ficity because the phospholyases have K353 of the fumB 1:αKG TA and HfumB 1:pyr TAs differ substantially as the lat- and Q216 that should be able to position αAA properly in the O-side ter only applies pyr as the acceptor and does not accept the non- as achieved by R353 and Q216 in Ala:glyox TA 2 (section 3.3.6). The hydrolysed fumB 1 (Hartinger et al., 2011 ), whereas the fumB 1:αKG TA phosphate recognition in the lyases probably also involves K355 and converted fumB 1 with αKG ( Leslie et al., 2004 ). These differences can- S346, the presence of E267 and D348, however, is puzzling because not be explained based on the active site residues identi fied by the these residues should repel the phosphate and the carboxylic acid moi- alignment to the OrnTL DB, as they are almost all identical among the ety on both sides of the active site entrance. Sequence fingerprints that enzymes with these two speci ficities (see sequence –function matrix strictly determine the different substrate speci ficity of these two en-

Table 2 ). (H)fumB 1 conversion seem to demand a very specialised active zymes cannot be found without structural information or the character- site compared to the other enzymes in the OrnTL DB, as no sequence in isation of similar enzymes. The common active site residues present in this database has the combination of E185 and R216 found in all the sequence–function matrix (Table 2) did not show obvious differ-

(H)fumB 1 TAs. ences compared to the transaminases except for the deletion at posi- tions 45 and 46 and the polar residues N/S44 and N47. 3.8.3. Phospholyases Summary: the two characterised phospholyases among the class III 3.8.4. Multi-domain or non-enzymes transaminase family were found to lack transaminase activity. Factors de- Summary: PLP fold type I and in particular, the class III transaminase termining this switch in reaction speci ficity are not known as the catalytic family's common fold are versatile building blocks for multi-domain en- machinery is the same as in the related transaminases. A substantial differ- zymes or transcriptional factors. ence, however, is the deletion of two amino acids in the left-handed helix The versatility of the common PLP fold type I is expressly highlighted α2, which could reshape the active site and thereby might shift the reaction by recent studies that discovered non-enzymatic or multi-domain ex- speci ficity. amples of this fold. The MocR-like transcriptional factors belong to the By discovering two phospholyases, Veiga-da-Cunha et al. (2012) fur- class I transaminase family (also belonging to PLP fold type I, see ther broadened the spectrum of known reaction speci ficities among the Table 3 ) attached to a helix –turn –helix domain for DNA binding class III transaminases. They found that the two human Ala:glyox TA 2 (Bramucci et al., 2011 ). Recent structural investigations of gabR, the homologs AGXT2L1 (UniProt ID: Q8TBG4) and AGXT2L2 (UniProt ID: transcriptional activator of the GABA: αKG TA in B. subtilis (PDB ID: F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 599

4N0B), revealed that PLP binding is achieved as in other fold type I en- 2) Some interactions of amino acids to the substrate might not be con- zymes, GABA, however, is bound differently and not transaminated by served. Complementary electrostatic interactions or hydrogen bond- this transcription factor ( Edayathumangalam et al., 2013 ). ing are easily analysed, but if the contacts are mainly realised by Another notable example is the presence of PLP fold type I and espe- hydrophobic interactions or if water molecules mediate them, it is cially class III transaminase domains in multimodular nonribosomal more dif ficult to formulate a clear pattern. Additionally, chemically peptide synthetase or polyketide synthase assembly lines ( Milano partially equivalent residues might result in comparable substrate et al., 2013 ). The class III domains required some structural variations scopes. The suggested fingerprints in this review therefore need to (e.g. insertions of ~12 amino acids between core positions 266 and be applied with caution, always considering that similar amino 267), compared to the stand-alone enzymes to enable integration in acids at the suggested positions might be found in enzymes with the multi-domain complexes. However most active site residues and related speci ficities. Furthermore, not all sequences that match a cer- therefore probably also substrate scopes of these enzymes differ tain fingerprint need to possess a related speci ficity as demonstrated substantially from those in all other characterised enzymes and predic- for a GABA: αKG TA that has been transformed into a DGD. In the first tions are therefore not possible. One characterised example illustrates attempt Liu et al. (2005) exchanged all active site residues of the these differences: the class III TA domain from MycA (UniProt ID: E. coli GABA: αKG TA (PDB ID: 1SFF) that were different to those Q9R9J1 , from B. subtilis , involved in cyclic lipopeptide mycosubtilin syn- found in DGD. The I46Q, E185S, V215A, and G269Y quadruple mu- thesis), which has been investigated for substrate scope independently tant showed, however, no increased decarboxylase activity, which of the whole multi-domain complex ( Aron et al., 2005 ). This domain could be explained by slightly different spatial positions of the C α prefers glutamine as the amino donor and converts acyl carrier protein atoms of these residues in the GABA: αKG TA compared to the DGD (ACP) thioesters of β-ketobutyrate. Active site residues, identified by (Liu et al., 2005 ). A final mutant, that had switched reaction speci fic- aligning this sequence to its closest homolog in the OrnTL DB, are not ity, only contained one of the residues (S185) that were predicted by sufficient to suggest substrate recognition as they differ substantially those found in DGD. from those in all other characterised class III transaminases (see se- 3) Proteins including the active sites have various degrees of flexibility, quence –function matrix Table 2 ). which cannot be seen from sequence patterns. Further information of potential flexibility is only accessible by substrate or product 4. Challenges for fingerprint-based sequence –function predictions bound crystal structures. 4) The superfamily to be analysed needs a spatially conserved back- During the process of discovering and evaluating sequence –function bone to allow for a proper alignment. If the active site is mainly com- relationships within the OrnTL DB, we encountered some challenges posed of variable loops, the position of relevant amino acids might regarding this approach. Some resulted from intrinsic limitations of differ signi ficantly, although they are the same in the sequence structure-based sequence alignments; others are challenges for the alignment. This fact highlights the class III transaminase fold's versa- approach of predicting function based on short sequence fingerprints. tility as all active site structural elements (except the N-terminus In this section we attempt to highlight the current challenges or limita- and the loop between core positions 73 and 74, see Fig. 10 ) are con- tions for a general evaluation of this method thereby deepening its served and allow for such a variety of catalysed reactions and accept- understanding and to motivate future approaches to solve current ed substrates. problems. 4.2. 3DM database related issues 4.1. Limitations of the active site amino acid fingerprint-based approach Structural based sequence alignments and their organisation in crys- Enzyme function prediction by a few residues with known impor- tal structure derived subfamilies also have a few issues to be aware of. tance for catalysis or substrate recognition, as introduced in section Superfamily diversity is a major criterion deciding the database's 1.5 and summarised in Fig. 6 , is a powerful approach as highlighted by content of information. For instance the database created for the PLP this account. It is, however, important to keep in mind that the situation fold type I contained too diverse crystal structures, which resulted in a in nature is often more complex than this. relatively small structural core. Even though subgroups within this alignment (e.g. the class III transaminases) share more structural fea- 1) There often exists more than one solution to realise the same sub- tures, these regions cannot be analysed in the large database because strate speci ficity, which leads to a general limitation of sequence they belong to the variable positions. Creating the OrnTL DB, which alignments: amino acid side chains of different C α (and alignment) only contains structurally more related enzymes — thereby increasing positions might fill the same space in the active site (exempli fied in the core, but also reducing the reaction diversity in the database — Fig. 6 A, B and D, sequences 1 –4). PLP fold type I provides a perfect overcame this problem. The database size will in most cases be a example for this case: the catalytic K242 is fully conserved in the trade-off between contained reaction or substrate speci ficities and OrnTL DB, whereas only 20% of the full PLP fold type I database covered positions. have a lysine there. Interestingly the catalytic lysine is located two positions later (K244) in many of the other sequences of that data- 4.3. The literature mining problem base (40% of the whole database have K244, the additional 40% could not be aligned properly at this position). The structural super- A general dif ficulty for connecting sequence and function is gathering position of the human Orn: αKG TA (PDB ID: 1OAT, has K242) and all available literature that describes enzymes' sequence or speci ficity the 2-aminoethylphosphonate:pyr TA from Salmonella enterica characterisations. Attempts are made to connect publications regarding (PDB ID: 1M32, has K244) revealed that both lysine ε-amino groups function and sequence in databases like BRENDA ( Schomburg et al., are located close to each other, while their C α positions differ 2013). These are constantly growing, but unfortunately still far from (Supplementary data, Figure S4). Structural alignments therefore al- complete. Additional hints for sequences with available experimental ways allow further and more accurate conclusions than sequence data can be retrieved from the sequences' ‘evidence on protein existence ’ alignments, but for many activities, crystal structures are not yet annotations in the UniProtKB ( Magrane and UniProt Consortium, 2011 ). available. This lack may lead to erroneous sequence alignments, for 3DM integrates the literature mining software Mutator ( Kuipers instance if the catalytic lysine is aligned at the same common et al., 2010a ) and the PDF reader Utopia Documents ( Attwood et al., position (242) before the crystal structure reveals that it actually 2010 ) to provide a link between 3DM positions in the alignment and ar- occupies another position (244). ticles where these positions were mentioned in a certain context, such 600 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 as mutagenesis studies to increase speci ficity or stability. This new show phospholyase and/or transaminase activity to investigate the feature supports literature research regarding mutagenesis studies. mechanisms responsible for the different speci ficities. However, the current versions of Mutator and Utopia are made for ex- For the discovery of unknown functions or other mechanisms of al- traction of mutation data from literature and they do not include protein ready known ones, the 3N5M subfamily is of special interest. Not only is characterisation studies making manual literature studies still indis- the template structure's function unknown (PDB ID: 3N5M), but also pensable. Although these tools are optimised for finding mutation most of the entries within this subfamily cannot be assigned with a pu- data from literature, only approximately one third of the characterised tative speci ficity by the active site fingerprints. This subfamily that only enzyme list (Supplementary data, Table S5) were not found by the au- contains six characterised enzymes (tau:pyr TAs, Ala:glyox TAs 2 and a tomated methods in combination with characterisation data retrieved βAla:pyr TA that is also able to convert amines, Supplementary data, from BRENDA and UniProt. The development of such tools is a step in Table 5) is very heterogeneous regarding the sequence –function matrix the right direction as the gap between sequence information and relat- positions (Table 2). Several sequences have R346, while several have ed literature data is a main obstacle and requires substantial reduction. R353 instead, and others have neither. The positions 185 and 269, how- ever, are relatively conserved and also correlated, as revealed by CMA 4.4. The challenge to identify unknown speci ficities (35% sequences of the subfamily have A185 and G269, while 41% have G185 and I/V269). Regardless of whether these enzymes possess so The attempts to identify proteins with unknown function, as far unknown speci ficities or established other ways to achieve known described in section 2.4, are limited to enzymes that switched reaction activities, the characterisation of such enzymes with unknown function or substrate speci ficity by only mutating a few residues compared to will be worthwhile as either new biocatalysts will be obtained or the the known enzymes. As the structure-based sequence alignment data- knowledge of functional determinants within the class III transaminase bases only contain sequences that share a certain identity to a known family will be extended. structure, and crystal structures are usually solved for already function- ally characterised enzymes, the identi fication of speci ficities that re- 5. Conclusion quired several mutations to develop are probably not easy to recognise in such databases. To allow for a wider identi fication of bio- In this review we show that the combination of structural informa- technologically interesting enzymes within the databases increasing tion and analysis of multiple sequence alignments as exempli fied for the amount of structural information for low identity sequences the class III transaminase family allowed to extract active site amino (which otherwise cannot be included in such a database) is of para- acid fingerprints that correlate with the different enzymatic activities. mount importance. Therefore, projects like the Protein Structure Initia- The different active site designs identi fied allowed covering 28 known tive ( Berman et al., 2009 ) or the Enzyme Function Initiative ( Gerlt et al., reaction and substrate speci ficities of enzymes within the ornithine 2011 ) are of substantial value for the biotechnological community. transaminase like family. This analysis should be regarded as a hint for Most notably, the recently discovered new substrate and reaction a qualitative prediction of the substrate scope rather than a fixed rule specificities (see section 3.8) showed that the potential of the family because, besides the amino acid distribution of the active site, additional of class III transaminases had been underestimated for several years. factors also affect the catalytic properties to a smaller or greater extend. This common structural fold allows for a variety of reaction and sub- Nevertheless, we encourage to apply these patterns in annotation strat- strate specificities that have not yet been explored or exploited suf fi- egies and to apply this methodology also for other superfamilies, which ciently. We therefore attempted to exemplify how the knowledge are amenable to such analyses. A critical mass of crystal structures, gained by this work can be applied to discover new enzyme activities however, is necessary to build up high quality sequence alignments, or new active site designs for achieving already known speci ficities (as which include a possibly large fraction of the available sequences of discussed in section 2.4). The subfamilies of a 3DM database make this the superfamily. The fingerprint approach allows not only to connect task relatively easy because these groups can be searched as an evolu- known enzyme activities to given sequences, but also to discover en- tionary related set to identify sequences that differ from known activi- zymes with yet unknown speci ficities and to suggest key mutations to ties by not matching the common fingerprints found in each verify hypotheses derived from the bioinformatics guided in-depth subfamily. The investigation of such a group separately from the analysis of an enzyme superfamily. These kinds of systematic investiga- whole superfamily provides a faster overview of present activities and tions will expand the number of useful enzymes and thereby provide potentially new ones because some of the active site residues within the community with potentially interesting biocatalysts as well as it each group are still conserved and differences can be more easily com- substantially helps to improve our understanding of sequence – and pared and evaluated when not all positions are varied at the same time. structure –function relationships of enzymes. For instance the 2ZUK subfamily contains 18 (putative) αAAA racemases (2 characterised and 16 additional sequences match the fin- Author contributions gerprint D185/K216) and 190 sequences that do not match any of the active site patterns for known activities (summarised in Supplementary MH and UTB initiated the review, FS and MH devised the conceptual data, Table S4). Further analysis showed that many of those 190 se- design, FS and CV coordinated data and literature analysis of the subsec- quences have R/K353, but no E185. Therefore, they probably convert tions and together with MH performed the in-depth analysis of the αAA but not ωAA because the usual mechanism for dual substrate superfamily. FS, CV, HK, HL, HM, AN and LS analysed literature and recognition for this speci ficity is not present (see section 3.1.1). Further- 3DM alignments of subgroups, created fingerprints and were involved more, these might be clustered by certain patterns found within the se- in writing subchapters. TvdB and H-JJ established the 3DM database quence –function matrix ( Table 2 ) positions. A few of these, for instance, and wrote the corresponding chapter. MH created the reaction mecha- contain D185 and N216 (like the Ile-2-epimerase, section 3.7) and nism video. With the help of CV and UTB, FS and MH did the main might therefore also be able to catalyse amino acid racemisation. A editing. FS, CV, PB, MH and UTB finalised the review. larger fraction contains a modi fied left-handed helix (N44, N47 and a deletion at positions 45 and 46) as found in characterised Acknowledgements phospholyases (section 3.8.3). It is likely, however, that many of these enzymes have a different substrate scope than the two known We thank Maika Genz for kindly providing Fig. 8 , Sebastian Wenske phospholyases because they do not have E267 and E348 and further- for his support in the creation of the reference list and Lea Kennel for more, several have N185 instead of the unpolar residues found in the providing activity data of the ATA 3HMU towards acylated and non- known enzymes. It would be interesting to determine if these enzymes acylated putrescine. FS thanks the Fonds der Chemischen Industrie F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 601

(Chemiefonds-Stipendium) and HK thanks the Deutsche Bundesstiftung Chou HT, Kwon DH, Hegazy M, Lu CD. Transcriptome analysis of agmatine and putrescine – fi catabolism in Pseudomonas aeruginosa PAO1. J Bacteriol 2008;190:1966 75. Umwelt (Grant No. AZ29937) for nancial support. PB and UB are grateful Christen P, Mehta PK. From cofactor to enzymes. The molecular evolution of pyridoxal-5 ′- for support by the COST Action (CM1303 Systems Biocatalysis). We espe- phosphate-dependent enzymes. Chem Rev 2001;1:436 –47. cially thank the European Union (KBBE-2011-5, grant No. 289350) for fi- Clausnitzer D, Piepersberg W, Wehmeier UF. The oxidoreductases LivQ and NeoQ are responsible for the different 6 ′-modifications in the aminoglycosides lividomycin nancial support within the European Union Seventh Framework and neomycin. J Appl Microbiol 2011;111:642 –51. Programme. Cobessi D, Dumas R, Pautre V, Meinguet C, Ferrer JL, Alban C. Biochemical and structural characterization of the Arabidopsis bifunctional enzyme dethiobiotin synthetase – Appendix A. Supplementary data diaminopelargonic acid aminotransferase: evidence for substrate channeling in biotin synthesis. Plant Cell 2012;24:1608 –25. Contestabile R, Angelaccio S, Maytum R, Bossa F, John RA. The contribution of a Supplementary data to this article can be found online at http://dx. conformationally mobile, active site loop to the reaction catalyzed by glutamate – doi.org/10.1016/j.biotechadv.2014.12.012 . semialdehyde aminomutase. J Biol Chem 2000;275:3879 86. Cooper AJL, Krasnikov BF, Okuno E, Jeitner TM. L-Alanine-glyoxylate aminotransferase II of rat kidney and liver mitochondria possesses cysteine S-conjugate β-lyase activity: a References contributing factor to the nephrotoxicity/hepatotoxicity of halogenated alkenes? Biochem J 2003;376:169 –78. Ahmed SA, Esaki N, Tanaka H, Soda K. L-α-Amino- β-thio- ε-caprolactam, a new sulfur- Crismaru CG, Wybenga GG, Szymanski W, Wijma HJ, Wu B, Bartsch S, et al. Biochemical containing substrate for α-amino- ε-caprolactam racemase. FEBS Lett 1984;174:76 –9. properties and crystal structure of a β-phenylalanine aminotransferase from Ahmed SA, Esaki N, Tanaka H, Soda K. Mechanism of α-amino- ε-caprolactam racemase Variovorax paradoxus . Appl Environ Microbiol 2013;79:185 –95. reaction. Biochemistry 1986;25:385 –8. Denger K, Ruff J, Schleheck D, Cook AM. Rhodococcus opacus expresses the xsc gene to Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W, et al. Gapped BLAST and utilize taurine as a carbon source or as a nitrogen source but not as a sulfur source. PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res Microbiology 2004;150:1859 –67. 1997;25:3389 –402. Desai AA. Sitagliptin manufacture: a compelling tale of green chemistry, process intensi- Andersen G, Andersen B, Dobritzsch D, Schnackerz KD, Pi škur J. A gene duplication led to fication, and industrial asymmetric catalysis. Angew Chem Int Ed 2011;50:1974 –6. specialized γ-aminobutyrate and β-alanine aminotransferase in yeast. FEBS J 2007; Dey S, Lane JM, Lee RE, Rubin EJ, Sacchettini JC. Structural characterization of the 274:1804 –17. Mycobacterium tuberculosis biotin biosynthesis enzymes 7,8-diaminopelargonic acid Aron ZD, Dorrestein PC, Blackhall JR, Kelleher NL, Walsh CT. Characterization of a new synthase and dethiobiotin synthetase. Biochemistry 2010;49:6746 –60. tailoring domain in polyketide biogenesis: the amine transferase domain of MycA Dunathan HC. Conformation and reaction speci ficity in pyridoxal phosphate enzymes. in the mycosubtilin gene cluster. J Am Chem Soc 2005;127:14986 –7. Proc Natl Acad Sci U S A 1966;55:712 –6. Asano Y, Yamaguchi S. Discovery of amino acid amides as new substrates for α-amino- ɛ- Edayathumangalam R, Wu R, Garcia R, Wang Y, Wang W, Kreinbring CA, et al. Crystal caprolactam racemase from Achromobacter obae . J Mol Catal B: Enzym 2005;36:22 –9. structure of Bacillus subtilis GabR, an autorepressor and transcriptional activator of Attwood TK, Kell DB, McDermott P, Marsh J, Pettifer SR, Thorne D. Utopia documents: gabT . Proc Natl Acad Sci U S A 2013;110:17820 –5. linking scholarly literature with research data. Bioinformatics 2010;26:i568 –74. Eliot AC, Kirsch JF. Pyridoxal phosphate enzymes: mechanistic, structural, and evolution- Baker PRS, Cramer SD, Kennedy M, Assimos DG, Holmes RP. Glycolate and glyoxylate ary considerations. Annu Rev Biochem 2004;73:383 –415. metabolism in HepG2 cells. Am J Physiol Cell Physiol 2004;287:C1359 –65. Eliot AC, Sandmark J, Schneider G, Kirsch JF. The dual-speci fic active site of 7,8- Bandounas L, Ballerstedt H, de Winde JH, Ruijssenaars HJ. Redundancy in putrescine diaminopelargonic acid synthase and the effect of the R391A mutation. Biochemistry catabolism in solvent tolerant Pseudomonas putida S12. J Biotechnol 2011;154:1 –10. 2002;41:12582 –9. Banerjee A, Chase M, Clayton RA, Landis B. Methods for the stereoselective synthesis and Ferradini N, Nicolia A, Capomaccio S, Veronesi F, Rosellini D. A point mutation in the enantiomeric enrichment of β-amino acids. WO International Patent Application Medicago sativa GSA gene provides a novel, ef ficient, selectable marker for plant WO2005005633, 2005. genetic engineering. J Biotechnol 2011;156:147 –52. Bea H-S, Park H-J, Lee S-H, Yun H. Kinetic resolution of aromatic β-amino acids by Fesko K, Uhl M, Steinreiber J, Gruber K, Griengl H. Biocatalytic access to α, α-dialkyl- α- ω-transaminase. Chem Commun 2011;47:5894 –6. amino acids by a mechanism-based approach. Angew Chem Int Ed 2010;49:121 –4. Berglund P, Humble MS, Branneby C. C–X Bond formation: transaminases as chiral Fogle EJ, Toney MD. Mutational analysis of substrate interactions with the active site of catalysts: mechanism, engineering, and applications. In: Carreira EM, Yamamoto H, dialkylglycine decarboxylase. Biochemistry 2010;49:6485 –93. editors. Comprehensive chirality. Amsterdam: Elsevier; 2012. p. 390 –401. Fogle EJ, Liu W, Woon ST, Keller JW, Toney MD. Role of Q52 in catalysis of decarboxylation Berman HM, Westbrook JD, Gabanyi MJ, Tao W, Shah R, Kouranov A, et al. The protein and transamination in dialkylglycine decarboxylase. Biochemistry 2005;44:16392 –404. structure initiative structural genomics knowledgebase. Nucleic Acids Res 2009;37: Fujii T, Narita T, Agematu H, Agata N, Isshiki K. Characterization of L-lysine 6- D365 –8. aminotransferase and its structural gene from Flavobacterium lutescens IFO3084. J Bertoldi M, Cellini B, Montioli R, Borri Voltattorni C. Insights into the mechanism of oxida- Biochem 2000;128:391 –7. tive deamination catalyzed by DOPA decarboxylase. Biochemistry 2008;47:7187 –95. Gand M, Müller H, Wardenga R, Höhne M. Characterization of three novel enzymes with Bisht S, Bharath SR, Murthy MRN. Conformational transitions, ligand speci ficity and catal- imine reductase activity. J Mol Catal B: Enzym 2014;110:126 –32. ysis in N-acetylornithine aminotransferase: implications on drug designing and ratio- Garcia ARM, Kamat SV, Pannuri S. Thermostable ω-transaminases. WO International nal enzyme engineering in omega aminotransferases. Unpublished crystal structures Patent Application WO2006063336, 2006. released in the Protein Data Bank; 2014. Ge H, Lv X, Fan J, Gao Y, Teng M, Niu L. Crystal structure of glutamate1-semialdehyde Boesten WHJ, Raemakers-Franken PC, Sonke T, Eu-Verink GJW, Grijpstra P. Polypeptides aminotransferase from Bacillus subtilis with bound pyridoxamine-5 ′-phosphate. having α-H-α-amino acid amide racemase activity and nucleic acids encoding the Biochem Biophys Res Commun 2010;402:356 –60. same. WO International Patent Application WO2003106691, 2003. Gerlt JA, Allen KN, Almo SC, Armstrong RN, Babbitt PC, Cronan JE, et al. The enzyme Bornscheuer UT, Kazlauskas RJ. Catalytic promiscuity in biocatalysis: using old enzymes to function initiative. Biochemistry 2011;50:9950 –62. form new bonds and follow new pathways. Angew Chem Int Ed 2004;43:6032 –40. Grant SM, Heel RC. Vigabatrin. Drugs 1991;41:889 –926. Bramucci E, Milano T, Pascarella S. Genomic distribution and heterogeneity of MocR-like Grimm B, Smith AJ, Kannangara CG, Smith M. Gabaculine-resistant glutamate 1- transcriptional factors containing a domain belonging to the superfamily of the semialdehyde aminotransferase of Synechococcus . Deletion of a tripeptide close to the pyridoxal-5 ′-phosphate dependent enzymes of fold type I. Biochem Biophys Res NH2 terminus and internal amino acid substitution. J Biol Chem 1991;266:12495 –501. Commun 2011;415:88 –93. Grishin NV, Phillips MA, Goldsmith EJ. Modeling of the spatial structure of eukaryotic Brohn F, Tchen TT. A single transaminase for 1,4-diaminobutane and 4-aminobutyrate in ornithine decarboxylases. Protein Sci 1995;4:1291 –304. a Pseudomonas species. Biochem Biophys Res Commun 1971;45:578 –82. Griswold WR, Toney MD. Role of the pyridine nitrogen in pyridoxal 5 ′-phosphate Bruce H, Tuan AN, Sanchez JM, Leese C, Hopwood J, Hyde R, et al. Structures of a catalysis: activity of three classes of PLP enzymes reconstituted with deazapyridoxal γ-aminobutyrate (GABA) transaminase from the s-triazine-degrading organism 5′-phosphate. J Am Chem Soc 2011;133:14823 –30. Arthrobacter aurescens TC1 in complex with PLP and with its external aldimine Halabi N, Rivoire O, Leibler S, Ranganathan R. Protein sectors: evolutionary units of PLP-GABA adduct. Acta Crystallogr F Struct Biol Commun 2012;68:1175 –80. three-dimensional structure. Cell 2009;138:774 –86. Buzenet AM, Fages C, Bloch-Tardy M, Gonnard P. Puri fication and properties of Hammer T, Bode R. Puri fication and characterization of an inducible L-lysine: 2- 4-aminobutyrate 2-ketoglutarate aminotransferase from pig liver. Biochim Biophys oxoglutarate 6-aminotransferase from Candida utilis . J Basic Microbiol 1992;32:21 –7. Acta 1978;522:400 –11. Hanson RL, Davis BL, Chen Y, Goldberg SL, Parker WL, Tully TP, et al. Preparation of Cánovas D, Vargas C, Calderón MI, Ventosa A, Nieto JJ. Characterization of the genes for (R)-amines from racemic amines with an ( S)-amine transaminase from Bacillus the biosynthesis of the compatible solute ectoine in the moderately halophilic bacte- megaterium . Adv Synth Catal 2008;350:1367 –75. rium Halomonas elongata DSM 3043. Syst Appl Microbiol 1998;21:487 –97. Hanson RL, Johnston RM, Goldberg SL, Parker WL, Patel RN. Enzymatic preparation of Catazaro J, Caprez A, Guru A, Swanson D, Powers R. Functional evolution of 5-hydroxy- L-proline, N-Cbz-5-hydroxy- L-proline, and N-boc-5-hydroxy- L-proline PLP-dependent enzymes based on active-site structural similarities. Proteins 2014; from ( α-N-protected)- L-ornithine using a transaminase or an amine oxidase. Enzyme 82:2597 –608. Microb Technol 2011;48:445 –53. Cellini B, Bertoldi M, Montioli R, Paiardini A, Borri Voltattorni C. Human wild-type Hartinger D, Schwartz H, Hametner C, Schatzmayr G, Haltrich D, Moll WD. Enzyme alanine:glyoxylate aminotransferase and its naturally occurring G82E variant: characteristics of aminotransferase FumI of Sphingopyxis sp. MTA144 for deamination functional properties and physiological implications. Biochem J 2007;408:39 –50. of hydrolyzed fumonisin B1. Appl Microbiol Biotechnol 2011;91:757 –68. Cha HJ, Jeong J-H, Rojviriya C, Kim Y-G. Structure of putrescine aminotransferase from Heimberg H, Boyen A, Crabeel M, Glansdorff N. Escherichia coli and Saccharomyces Escherichia coli provides insights into the substrate speci ficity among class III amino- cerevisiae acetylornithine aminotransferase: evolutionary relationship with ornithine transferases. PLoS One 2014;9:e113212. aminotransferase. Gene 1990;90:69 –78. 602 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604

Heinl S, Hartinger D, Thamhesl M, Schatzmayr G, Moll WD, Grabherr R. An aminotransfer- Kroutil W, Fischereder EM, Fuchs CS, Lechner H, Mutti FG, Pressnitz D, et al. Asymmetric ase from bacterium ATCC 55552 deaminates hydrolyzed fumonisin B1. Biodegrada- preparation of prim -, sec -, and tert -amines employing selected biocatalysts. Org tion 2011;22:25 –30. Process Res Dev 2013;17:751 –9. Hennig M, Grimm B, Contestabile R, John RA, Jansonius JN. Crystal structure of glutamate- Kuipers RKP, Joosten HJ, Verwiel E, Paans S, Akerboom J, van der Oost J, et al. Correlated 1-semialdehyde aminomutase: An α2-dimeric vitamin B6-dependent enzyme with mutation analyses on super-family alignments reveal functionally important asymmetry in structure and active site reactivity. Proc Natl Acad Sci U S A 1997;94: residues. Proteins 2009;76:608 –16. 4866 –71. Kuipers RKP, van den Bergh T, Joosten H-J, Lekanne dit Deprez RH, Mannens MMAM, Hirotsu K, Goto M, Okamoto A, Miyahara I. Dual substrate recognition of aminotransferases. Schaap PJ. Novel tools for extraction and validation of disease-related mutations ap- Chem Rec 2005;5:160 –72. plied to fabry disease. Hum Mutat 2010a;31:1026 –32. Hohenester E, Keller JW, Jansonius JN. An alkali metal ion size-dependent switch in the Kuipers RKP, Joosten HJ, van Berkel WJH, Leferink NGH, Rooijen E, Ittmann E, et al. 3DM: active site structure of dialkylglycine decarboxylase. Biochemistry 1994;33: systematic analysis of heterogeneous superfamily data to discover protein function- 13561 –5570. alities. Proteins 2010b;78:2101 –13. Höhne M, Bornscheuer UT. Biocatalytic routes to optically active amines. ChemCatChem Kurihara S, Oda S, Kato K, Kim HG, Koyanagi T, Kumagai H, et al. A novel putrescine 2009;1:42 –51. utilization pathway involves γ-glutamylated intermediates of Escherichia coli K-12. J Höhne M, Bornscheuer UT. Application of transaminases in organic synthesis. In: May O, Biol Chem 2005;280:4602 –8. Gröger H, Drauz W, editors. Enzymes in organic synthesis. Weinheim: Wiley-VCH; Kwon YC, Lee KH, Kim HC, Han K, Seo JH, Kim B-G, et al. Cloning-independent expression 2012. p. 779 –820. and analysis of ω-transaminases by use of a cell-free protein synthesis system. Appl Höhne M, Schätzle S, Jochens H, Robins K, Bornscheuer UT. Rational assignment of key Environ Microbiol 2010;76:6295 –8. motifs for function guides in silico enzyme identi fication. Nat Chem Biol 2010;6: Lal PB, Schneider BL, Vu K, Reitzer L. The redundant aminotransferases in lysine and 807 –13. arginine synthesis and the extent of aminotransferase redundancy in Escherichia Hoober JK, Kahn A, Ash DE, Gough S, Kannangara CG. Biosynthesis of Δ-aminolevulinate coli . Mol Microbiol 2014;94:843 –56. in greening barley leaves. IX. structure of the substrate, mode of gabaculine inhibi- Laue H, Cook AM. Biochemical and molecular characterization of taurine:pyruvate amino- tion, and the catalytic mechanism of glutamate 1-semialdehyde aminotransferase. transferase from the anaerobe Bilophila wadsworthia . Eur J Biochem 2000;267: Carlsberg Res Commun 1988;53:11 –25. 6841 –8. Huang F, Spiteller D, Koorbanally NA, Li Y, Llewellyn NM, Spencer JB. Elaboration of Ledwidge R, Blanchard JS. The dual biosynthetic capability of N-acetylornithine amino- neosamine rings in the biosynthesis of neomycin and butirosin. ChemBioChem transferase in arginine and lysine biosynthesis. Biochemistry 1999;38:3019 –24. 2007;8:283 –8. Lee S-G, Hong S-P, Sung M-H. Development of an enzymatic system for the production of Hult K, Berglund P. Enzyme promiscuity: mechanism and applications. Trends Biotechnol dopamine from catechol, pyruvate, and ammonia. Enzyme Microb Technol 1999;25: 2007;25:231 –8. 298 –302. Humble MS, Berglund P. Biocatalytic promiscuity. Eur J Org Chem 2011;2011:3391 –401. Lee H, Juncosa JI, Silverman RB. Ornithine aminotransferase versus GABA aminotransfer- Humble MS, Cassimjee KE, Hakansson M, Kimbung YR, Walse B, Abedi V, et al. Crystal ase: implications for the design of new anticancer drugs. Med Res Rev 2014;35: structures of the Chromobacterium violaceum ω-transaminase reveal major structural 286 –305. rearrangements upon binding of coenzyme PLP. FEBS J 2012;279:779 –92. Leferink NGH, Fraaije MW, Joosten HJ, Schaap PJ, Mattevi A, van Berkel WJH. Identi fication Hunter S, Jones P, Mitchell A, Apweiler R, Attwood TK, Bateman A, et al. InterPro in 2011: of a gatekeeper residue that prevents dehydrogenases from acting as oxidases. J Biol new developments in the family and domain prediction database. Nucleic Acids Res Chem 2009;284:4392 –7. 2012;40:D306 –12. Leslie H, Paul WD, Lishan Z. Transaminases, deaminases and aminomutases and compo- Huson DH, Scornavacca C. Dendroscope 3: an interactive tool for rooted phylogenetic sitions and methods for enzymatic detoxi fication. WO International Patent Applica- trees and networks. Syst Biol 2012;61:1061 –7. tion WO2004085624, 2004. Izumi Y, Sato K, Tani Y, Ogata K. 7,8-Diaminopelargonic acid aminotransferase, an Li T, Huo L, Pulley C, Liu A. Decarboxylation mechanisms in biological system. Bioorg enzyme involved in biotin biosynthesis by microorganisms. Agric Biol Chem 1975; Chem 2012;43:2 –14. 39:175 –81. Liepman AH, Olsen LJ. Alanine aminotransferase homologs catalyze the glutamate: Izumi Y, Kano Y, Inagaki K, Kawase N, Tani Y, Yamada H. Characterization of biotin biosyn- glyoxylate aminotransferase reaction in peroxisomes of arabidopsis. Plant Physiol thetic enzymes of Bacillus sphaericus : a dethiobiotin producing bacterium. Agric Biol 2003;131:215 –27. Chem 1981;45:1983 –9. Liese A, Seelbach K, Buchholz A, Haberland J. Processes: hydrolases EC 3 — EC 3.4.21.4 to Jackson LK, Brooks HB, Osterman AL, Goldsmith EJ, Phillips MA. Altering the reaction EC 3.8.X.X. In: Liese A, Seelbach K, Wandrey C, editors. Industrial biotransformations. speci ficity of eukaryotic ornithine decarboxylase. Biochemistry 2000;39:11247 –57. Wiley-VCH; 2006a. p. 350 –446. Jaskolski M, Dauter Z, Wlodawer A. A brief history of macromolecular crystallography, Liese A, Seelbach K, Buchholz A, Haberland J. Processes: lyases EC 4. In: Liese A, Seelbach illustrated by a family tree and its Nobel fruits. FEBS J 2014;281:3985 –4009. K, Wandrey C, editors. Industrial biotransformations. Wiley-VCH; 2006b. p. 447 –503. Jomrit J, Summpunn P, Meevootisom V, Wiyakrutta S. Sensitive non-radioactive determi- Liu W, Rogers CJ, Fisher AJ, Toney MD. Aminophosphonate inhibitors of dialkylglycine decar- nation of aminotransferase stereospeci ficity for C-4 ′ hydrogen transfer on the coen- boxylase: structural basis for slow binding inhibition. Biochemistry 2002;41:12320 –8. zyme. Biochem Biophys Res Commun 2011;405:626 –31. Liu W, Peterson PE, Carter RJ, Zhou X, Langston JA, Fisher AJ, et al. Crystal structures of un- Joosten HJ, Han Y, Niu W, Vervoort J, Dunaway-Mariano D, Schaap PJ. Identification of bound and aminooxyacetate-bound Escherichia coli γ-aminobutyrate aminotransfer- fungal oxaloacetate hydrolyase within the isocitrate lyase/PEP mutase enzyme ase. Biochemistry 2004;43:10896 –905. superfamily using a sequence marker-based method. Proteins 2008;70:157 –66. Liu W, Peterson PE, Langston JA, Jin X, Zhou X, Fisher AJ, et al. Kinetic and crystallographic Jortzik E, Fritz-Wolf K, Sturm N, Hipp M, Rahlfs S, Becker K. Redox regulation of analysis of active site mutants of Escherichia coli γ-aminobutyrate aminotransferase. Plasmodium falciparum ornithine δ-aminotransferase. J Mol Biol 2010;402:445 –59. Biochemistry 2005;44:2982 –92. Käck H, Sandmark J, Gibson K, Schneider G, Lindqvist Y. Crystal structure of Lou X,RanT,HanN,GaoY,HeJ,TangL, etal. Crystal structure of the catalytic domain diaminopelargonic acid synthase: evolutionary relationships between pyridoxal-5 ′- of PigE: a transaminase involved in the biosynthesis of 2-methyl-3- n-amyl-pyr- phosphate-dependent enzymes. J Mol Biol 1999;291:857 –76. role (MAP) from Serratia sp. FS14. Biochem Biophys Res Commun 2014;447: Kaneoke M, Shimizu E, Yorifuji T. Metabolism of L-arginine, agmatine, and related 178 –83. compounds in Nocardioides simplex . Biosci Biotechnol Biochem 1994;58:244 –9. Lu CD, Itoh Y, Nakada Y, Jiang Y. Functional analysis and regulation of the divergent Kaulmann U, Smithies K, Smith MEB, Hailes HC, Ward JM. Substrate spectrum of spuABCDEFGH-spuI operons for polyamine uptake and utilization in Pseudomonas ω-transaminase from Chromobacterium violaceum DSM30191 and its potential for aeruginosa PAO1. J Bacteriol 2002;184:3765 –73. biocatalysis. Enzyme Microb Technol 2007;41:628 –37. Lütke-Eversloh T, Santos C, Stephanopoulos G. Perspectives of biotechnological produc- Keller JW, Baurick KB, Rutt GC, O'Malley MV, Sonafrank NL, Reynolds RA, et al. Pseudomonas tion of L-tyrosine and its applications. Appl Microbiol Biotechnol 2007;77:751 –62. cepacia 2,2-dialkylglycine decarboxylase. sequence and expression in Escherichia coli of Magrane M, UniProt Consortium. UniProt knowledgebase: a hub of integrated protein structural and repressor genes. J Biol Chem 1990;265:5531 –9. data. Database 2011;2011. Kim KH. Puri fication and properties of a mine α-ketoglutarate transaminase from Malashkevich VN, Strop P, Keller JW, Jansonius JN, Toney MD. Crystal structures of Escherichia coli . J Biol Chem 1964;239:783 –6. dialkylglycine decarboxylase inhibitor complexes. J Mol Biol 1999;294:193 –200. Kim J, Kyung D, Yun H, Cho BK, Seo JH, Cha M, et al. Cloning and characterization of a novel Mann S, Ploux O. Pyridoxal-5 ′-phosphate-dependent enzymes involved in biotin biosyn- β-transaminase from Mesorhizobium sp. strain LUK: a new biocatalyst for the synthesis thesis: structure, reaction mechanism and inhibition. Biochim Biophys Acta 2011; of enantiomerically pure β-amino acids. Appl Environ Microbiol 2007;73:1772 –82. 1814:1459 –66. Kohls H, Steffen-Munsberg F, Höhne M. Recent achievements in developing the biocata- Mann S, Ploux O. 7,8-Diaminoperlargonic acid aminotransferase from Mycobacterium lytic toolbox for chiral amine synthesis. Curr Opin Chem Biol 2014;19:180 –92. tuberculosis , a potential therapeutic target. FEBS J 2006;273:4778 –89. Koma D, Sawai T, Harayama S, Kino K. Overexpression of the genes from thermophiles in Mann S, Colliandre L, Labesse G, Ploux O. Inhibition of 7,8-diaminopelargonic acid Escherichia coli by high-temperature cultivation. Appl Microbiol Biotechnol 2006;73: aminotransferase from Mycobacterium tuberculosis by chiral and achiral anologs of 172 –80. its substrate: biological implications. Biochimie 2009;91:826 –34. Kongsaeree P, Samanchart C, Laowanapiban P, Wiyakrutta S, Meevootisom V. Crystallization Markova M, Peneff C, Hewlins MJ, Schirmer T, John RA. Determinants of substrate and preliminary X-ray crystallographic analysis of D-phenylglycine aminotransferase speci ficity in ω-aminotransferases. J Biol Chem 2005;280:36409 –16. from Pseudomonas stutzeri ST201. Acta Crystallogr D Biol Crystallogr 2003;59:953 –4. McPhalen CA, Vincent MG, Picot D, Jansonius JN, Lesk AM, Chothia C. Domain closure in Kontani Y, Kaneko M, Kikugawa M, Fujimoto S, Tamaki N. Identity of D-3-amino- mitochondrial aspartate aminotransferase. J Mol Biol 1992;227:197 –213. isobutyrate-pyruvate aminotransferase with alanine-glyoxylate aminotransferase 2. Medina-Kauwe LK, Tobin AJ, De Meirleir L, Jaeken J, Jakobs C, Nyhan WL, et al. Biochim Biophys Acta 1993;1156:161 –6. 4-Aminobutyrate aminotransferase (GABA-transaminase) de ficiency. J Inherit Kourist R, Jochens H, Bartsch S, Kuipers RKP, Padhi SK, Gall M, et al. The α/β-hydrolase fold Metab Dis 1999;22:414 –27. 3DM database (ABHDB) as a tool for protein engineering. ChemBioChem 2010;11: Mehta PK, Hale TI, Christen P. Aminotransferases: demonstration of homology and divi- 1635 –43. sion into evolutionary subgroups. Eur J Biochem 1993;214:549 –61. F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604 603

Mi H, Muruganujan A, Thomas PD. PANTHER in 2013: modeling the evolution of gene Rausch C, Lerchner A, Schiefner A, Skerra A. Crystal structure of the ω- function, and other gene attributes, in the context of phylogenetic trees. Nucleic aminotransferase from Paracoccus denitri ficans and its phylogenetic relationship Acids Res 2013;41:D377 –86. with other class III amino-transferases that have biotechnological potential. Pro- Midelfort KS, Kumar R, Han S, Karmilowicz MJ, McConnell K, Gehlhaar DK, et al. teins 2013;81:774 –87. Redesigning and characterizing the substrate speci ficity and activity of Vibrio fluvialis Reshetnikov AS, Khmelenina VN, Trotsenko YA. Characterization of the ectoine biosynthe- aminotransferase for the synthesis of imagabalin. Protein Eng Des Sel 2013;26:25 –33. sis genes of haloalkalotolerant obligate methanotroph “Methylomicrobium Milano T, Paiardini A, Grgurina I, Pascarella S. Type I pyridoxal 5 ′-phosphate dependent alcaliphilum 20Z ”. Arch Microbiol 2006;184:286 –97. enzymatic domains embedded within multimodular nonribosomal peptide synthe- Roeder J, Schink B. Syntrophic degradation of cadaverine by a de fined methanogenic tase and polyketide synthase assembly lines. BMC Struct Biol 2013;13:26. coculture. Appl Environ Microbiol 2009;75:4821 –8. Miyazaki J, Kobashi N, Nishiyama M, Yamane H. Functional and evolutionary relationship Romero J, Martin JF, Liras P, Demain AL, Rius N. Partial puri fication, characterization and between arginine biosynthesis and prokaryotic lysine biosynthesis through nitrogen regulation of the lysine ε-aminotransferase of Streptomyces clavuligerus . J α-aminoadipate. J Bacteriol 2001;183:5067 –73. Ind Microbiol Biotechnol 1997;18:241 –6. Moll WD, Hartinger D, Grie βler K, Binder EM, Schatzmayr G. Method for the production Romo AJ, Liu HW. Mechanisms and structures of vitamin B6-dependent en- of an additive for the enzymatic decomposition of mycotoxins, additive, and use zymes involved in deoxy sugar biosynthesis. Biochim Biophys Acta 2011;1814: thereof. WO International Patent Application WO2010031101, 2010. 1534 –47. Mou X, Vila-Costa M, Sun S, Zhao W, Sharma S, Moran MA. Metatranscriptomic signature Rudat J, Brucher BR, Syldatk C. Transaminases for the synthesis of enantiopure beta- of exogenous polyamine utilization by coastal bacterioplankton. Environ Microbiol amino acids. AMB Express 2012;2:11. Rep 2011;3:798 –806. Samsonova NN, Smirnov SV, Altman IB, Ptitsyn LR. Molecular cloning and characterization Müller U, van Assema F, Gunsior M, Orf S, Kremer S, Schipper D, et al. Metabolic engineer- of Escherichia coli K12 ygjG gene. BMC Microbiol 2003;3:2. ing of the E. coli L-phenylalanine pathway for the production of D-phenylglycine Sandmark J, Eliot AC, Famm K, Schneider G, Kirsch JF. Conserved and nonconserved (D-Phg). Metab Eng 2006;8:196 –208. residues in the substrate binding site of 7,8-diaminopelargonic acid synthase from Mutaguchi Y, Ohmori T, Wakamatsu T, Doi K, Ohshima T. Identi fication, puri fication, and Escherichia coli are essential for catalysis. Biochemistry 2004;43:1213 –22. characterization of a novel amino acid racemase, isoleucine 2-epimerase, from Lacto- Savile CK, Janey JM, Mundorff EC, Moore JC, Tam S, Jarvis WR, et al. Biocatalytic asymmet- bacillus species. J Bacteriol 2013;195:5207 –15. ric synthesis of chiral amines from ketones applied to sitagliptin manufacture. Newman J, Seabrook S, Surjadi R, Williams CC, Lucent D, Wilding M, et al. Determination Science 2010;329:305 –9. of the structure of the catabolic N-succinylornithine transaminase (AstC) from Sayer C, Isupov MN, Westlake A, Littlechild JA. Structural studies of Pseudomonas and Escherichia coli . PLoS One 2013;8:e58298. Chromobacterium ω-aminotransferases provide insights into their differing substrate Nikolskaya AN, Arighi CN, Huang H, Barker WC, Wu CH. PIRSF family classi fication system speci ficity. Acta Crystallogr D Biol Crystallogr 2013;69:564 –76. for protein functional and evolutionary analysis. Evol Bioinform Online 2006;2: Schätzle S, Steffen-Munsberg F, Thontowi A, Höhne M, Robins K, Bornscheuer UT. 197 –209. Enzymatic asymmetric synthesis of enantiomerically pure aliphatic, aromatic and Novotny M, Kleywegt GJ. A survey of left-handed helices in protein structures. J Mol Biol arylaliphatic amines with ( R)-selective amine transaminases. Adv Synth Catal 2011; 2005;347:231 –41. 353:2439 –45. Nugent TC, El-Shazly M. Chiral amine synthesis — recent developments and trends for Schneider BL, Reitzer L. Pathway and enzyme redundancy in putrescine catabolism in enamide reduction, reductive amination, and imine reduction. Adv Synth Catal Escherichia coli . J Bacteriol 2012;194:4080 –8. 2010;352:753 –819. Schneider J, Wendisch VF. Biotechnological production of polyamines by bacteria: recent Okazaki S, Suzuki A, Mizushima T, Kawano T, Komeda H, Asano Y, et al. The novel struc- achievements and future perspectives. Appl Environ Microbiol 2011;91:17 –30. ture of a pyridoxal 5 ′-phosphate-dependent fold-type I racemase, α-amino- ε- Schneider G, Käck H, Lindqvist Y. The manifold of vitamin B6 dependent enzymes. caprolactam racemase from Achromobacter obae . Biochemistry 2009;48:941 –50. Structure 2000;8:R1 –6. Oliveira EF, Cerqueira NM, Fernandes PA, Ramos MJ. Mechanism of formation of the inter- Schomburg I, Chang A, Placzek S, Söhngen C, Rother Ml, Lang M, et al. BRENDA in nal aldimine in pyridoxal 5 ′-phosphate-dependent enzymes. J Am Chem Soc 2011; 2013: integrated reactions, kinetic data, enzyme function data, improved disease 133:15496 –505. classi fication: new options and contents in BRENDA. Nucleic Acids Res 2013;41: Orriss GL, Patel TR, Sorensen J, Stetefeld J. Absence of a catalytic water confers resistance D764 –72. to the neurotoxin gabaculine. FASEB J 2010;24:404 –14. Schousboe A, Wu JY, Roberts E. Puri fication and characterization of the 4-aminobutyrate- Park J, Osei YD, Churchich JE. Isolation and characterization of recombinant mitochondrial 2-ketoglutarate transaminase from mouse brain. Biochemistry 1973;12:2868 –73. 4-aminobutyrate aminotransferase. J Biol Chem 1993;268:7636 –9. Schulz A, Taggeselle P, Tripier D, Bartsch K. Stereospecific production of the herbicide Park DH, Mirabella R, Bronstein PA, Preston GM, Haring MA, Lim CK, et al. Mutations in phosphinothricin (glufosinate) by transamination: isolation and characterization of γ-aminobutyric acid (GABA) transaminase genes in plants or Pseudomonas syringae a phosphinothricin-speci fic transaminase from Escherichia coli . Appl Environ reduce bacterial virulence. Plant J 2010a;64:318 –30. Microbiol 1990;56:1 –6. Park ES, Kim M, Shin J-S. One-pot conversion of L-threonine into L-homoalanine: biocata- Schulze JO, Schubert WD, Moser J, Jahn D, Heinz DW. Evolutionary relationship between lytic production of an unnatural amino acid from a natural one. Adv Synth Catal initial enzymes of tetrapyrrole biosynthesis. J Mol Biol 2006;358:1212 –20. 2010b;352:3391 –8. Schwibbert K, Marin-Sanguino A, Bagyan I, Heidrich G, Lentzen G, Seitz H, et al. A blue- Park ES, Kim M, Shin J-S. Molecular determinants for substrate selectivity of print of ectoine metabolism from the genome of the industrial producer Halomonas ω-transaminases. Appl Microbiol Biotechnol 2011a;93:2425 –35. elongata DSM 2581T. Environ Microbiol 2011;13:1973 –94. Park JW, Park SR, Nepal KK, Han AR, Ban YH, Yoo YJ, et al. Discovery of parallel pathways Seo J-H, Hwang J-Y, Seo S-H, Kang H, Hwang B-Y, Kim B-G. Computational selection, of kanamycin biosynthesis allows antibiotic manipulation. Nat Chem Biol 2011b;7: identi fication and structural analysis of ω-aminotransferases with various substrate 843 –52. specificities from the genome sequence of Mesorhizobium loti MAFF303099. Biosci Patel RN, Banerjee A, Hanson RL, Brzozowski DB, Parker LW, Szarka LJ. Oxidation of Biotechnol Biochem 2012;76:1308 –14. Nα-protected- L-lysine by Rhodotorula graminis to produce novel chiral compounds. Seong Gyu J, Jae Hoon B, Joong Sik J, Sang Ho J, Byung Ryong L, Kil Soo L, et al. Molecular Tetrahedron Asymmetry 1999;10:31 –6. cloning and functional expression of bovine brain GABA transaminase. Mol Cells Patel RN, Banerjee A, Nanduri VB, Goldberg SL, Johnston RM, Hanson RL, et al. Biocatalytic 2001;12:91 –6. preparation of a chiral synthon for a vasopeptidase inhibitor: enzymatic conversion Shaibe E, Metzer E, Halpern YS. Metabolic pathway for the utilization of L-arginine, L-or- 2 of N -[N-Phenylmethoxy)carbonyl] L-homocysteinyl]- L-lysine (1 → 1′)-disul fide nithine, agmatine, and putrescine as nitrogen sources in Escherichia coli K-12. J to [4S-(4I,7I,10aJ)] 1-octahydro-5-oxo-4-[phenylmethoxy)carbonyl]amino]-7H- Bacteriol 1985;163:933 –7. pyrido-[2,1-b] [1,3]thiazepine-7-carboxylic acid methyl ester by a novel L-lysine Shin J-S, Kim B-G. Kinetic resolution of α-methylbenzylamine with ω-transaminase ε-aminotransferase. Enzyme Microb Technol 2000;27:376 –89. screened from soil microorganisms: application of a biphasic system to overcome Percudani R, Peracchi A. A genomic overview of pyridoxal-phosphate-dependent product inhibition. Biotechnol Bioeng 1997;55:348 –58. enzymes. EMBO Rep 2003;4:850 –4. Shin J-S, Yun H, Jang JW, Park I, Kim B-G. Purification, characterization, and molecular Percudani R, Peracchi A. The B6 database: a tool for the description and classi fication of cloning of a novel amine:pyruvate transaminase from Vibrio fluvialis JS17. Appl vitamin B6-dependent enzymatic activities and of the corresponding protein families. Microbiol Biotechnol 2003;61:463 –71. BMC Bioinform 2009;10:273. Shon M, Shanmugavel R, Shin G, Mathew S, Lee SH, Yun H. Enzymatic synthesis of chiral Perret A, Lechaplais C, Tricot S, Perchat N, Vergne C, Pelle C, et al. A novel acyl-CoA beta- γ-amino acids using ω-transaminase. Chem Commun 2014;50:12680 –3. transaminase characterized from a metagenome. PLoS One 2011;6:e22918. Sigrist CJA, Cerutti L, de Castro E, Langendijk-Genevaux PS, Bulliard V, Bairoch A, et al. Pleiss J. Systematic analysis of large enzyme families: identi fication of speci ficity- and PROSITE, a protein domain database for functional characterization and annotation. selectivity-determining hotspots. ChemCatChem 2014;6:944 –50. Nucleic Acids Res 2010;38:D161 –6. Price MN, Dehal PS, Arkin AP. FastTree 2 — approximately maximum-likelihood trees for Smith MA, Kannangara CG, Grimm B, von Wettstein D. Characterization of glutamate-1- large alignments. PLoS One 2010;5:e9490. semialdehyde aminotransferase of Synechococcus . Steady-state kinetic analysis. Eur Punta M, Coggill PC, Eberhardt RY, Mistry J, Tate J, Boursnell C, et al. The Pfam protein J Biochem 1991;202:749 –57. families database. Nucleic Acids Res 2012;40:D290 –301. Soda K, Esaki N. Pyridoxal enzymes acting on D-amino acids. Pure Appl Chem 1994;66: Radivojac P, Clark WT, Oron TR, Schnoes AM, Wittkop T, Sokolov A, et al. A large-scale 709 –14. evaluation of computational protein function prediction. Nat Methods 2013;10: Soda K, Misono H, Yamamoto T. L-Lysine- α-ketoglutarate aminotransferase. I. identi fica- 221 –7. tion of a product, Δ1-piperideine-6-carboxylic acid. Biochemistry 1968;7:4102 –9. Rajaram V, Prasuna PR, Savithri HS, Murthy MRN. Structure of biosynthetic Soda K, Yoshimura T, Esaki N. Stereospecificity for the hydrogen transfer of pyridoxal N-acetylornithine aminotransferase from Salmonella typhimurium : studies on enzyme reactions. Chem Rec 2001;1:373 –84. substrate speci ficity and inhibitor binding. Proteins 2008;70:429 –41. Steffen-Munsberg F, Vickers C, Thontowi A, Schätzle S, Meinhardt T, Svedendahl Humble Rao DR, Hariharan K, Vijayalakshmi KR. A study of the metabolism of L-αγ - M, et al. Revealing the structural basis of promiscuous amine transaminase activity. diaminobutyric acid in a Xanthomonas species. Biochem J 1969;114:107 –15. ChemCatChem 2013a;5:154 –7. 604 F. Steffen-Munsberg et al. / Biotechnology Advances 33 (2015) 566 –604

Steffen-Munsberg F, Vickers C, Thontowi A, Schätzle S, Tumlirsch T, Svedendahl Humble rather than S-adenosylmethionine as the amino donor in the KAPA-to-DAPA reaction. M, et al. Connecting unexplored protein crystal structures to enzymatic function. Biotechnol Bioeng 2005;91:75 –83. ChemCatChem 2013b;5:150 –3. Vandenende CS, Vlasschaert M, Seah SYK. Functional characterization of an aminotrans- Steinkellner G, Gruber CC, Pavkov-Keller T, Binter A, Steiner K, Winkler C, et al. Identi fication ferase required for pyoverdine siderophore biosynthesis in Pseudomonas aeruginosa of promiscuous ene-reductase activity by mining structural databases using active site PAO1. J Bacteriol 2004;186:5596 –602. constellations. Nat Commun 2014;5. Veiga-da-Cunha M, Hadi F, Balligand T, Stroobant V, Van Schaftingen E. Molecular Stetefeld J, Jenny M, Burkhard P. Intersubunit signaling in glutamate-1-semialdehyde- identi fication of hydroxylysine kinase and of ammoniophospholyases acting on aminomutase. Proc Natl Acad Sci U S A 2006;103:13688 –93. 5-phosphohydroxy- L-lysine and phosphoethanolamine. J Biol Chem 2012;287: Stoner GL, Eisenberg MA. Puri fication and properties of 7,8-diaminopelargonic acid 7246 –55. aminotransferase. J Biol Chem 1975;250:4029 –36. Voellym R, Leisinger T. Role of 4-aminobutyrate aminotransferase in the arginine metab- Storici P, Capitani G, De Biase D, Moser M, John RA, Jansonius JN, et al. Crystal structure of olism of Pseudomonas aeruginosa . J Bacteriol 1976;128:722 –9. GABA-aminotransferase, a target for antiepileptic drug therapy. Biochemistry 1999; Watanabe N, Sakabe K, Sakabe N, Higashi T, Sasaki K, Aibara S, et al. Crystal structure anal- 38:8628 –34. ysis of ω-amino acid:pyruvate aminotransferase with a newly developed Weissenberg Storici P, De Biase D, Bossa F, Bruno S, Mozzarelli A, Peneff C, et al. Structures of camera and an imaging plate using synchrotron radiation. J Biochem 1989;105:1 –3. γ-aminobutyric acid (GABA) aminotransferase, a pyridoxal 5 ′-phosphate, and Webb EC, International Union of Biochemistry and Molecular Biology. Enzyme nomencla- [2Fe –2S] cluster-containing enzyme, complexed with γ-ethynyl-GABA and with the ture 1992. Recommendations of the Nomenclature Committee of the International antiepilepsy drug Vigabatrin. J Biol Chem 2004;279:363 –73. Union of Biochemistry and Molecular Biology on the Nomenclature and Classi fication Sun S, Zabinski RF, Toney MD. Reactions of alternate substrates demonstrate of Enzymes. San Diego, California: Academic Press; 1992. stereoelectronic control of reactivity in dialkylglycine decarboxylase. Biochemistry Weber N, Ismail A, Gorwa-Grauslund M, Carlquist M. Biocatalytic potential of vanillin 1998;37:3865 –75. aminotransferase from Capsicum chinense . BMC Biotechnol 2014;14:25. Sung BK, Kim YT. Structural arrangement for functional requirements of brain recombi- Wegman MA, Janssen MHA, van Rantwijk F, Sheldon RA. Towards biocatalytic synthesis nant 4-aminobutyrate aminotransferase. J Biochem Mol Biol 2000;33:43 –8. of β-lactam antibiotics. Adv Synth Catal 2001;343:559 –76. Tamaki N, Sakata SF, Matsuda K. [35] Puri fication, properties, and sequencing of Williamson NR, Simonsen HT, Ahmed RAA, Goldet G, Slater H, Woodley L, et al. Biosyn- aminoisobutyrate aminotransferases from rat liver. In: Sokatch JR, Harris RA, editors. thesis of the red antibiotic, prodigiosin, in Serratia : identi fication of a novel Methods in enzymology. San Diego, CA: Academic Press; 2000. p. 376 –89. 2-methyl-3-n-amyl-pyrrole (MAP) assembly pathway, de finition of the terminal The PyMOL Molecular Graphics System, Version 1.6.0.0 Schrödinger, LCC, 2013. condensing enzyme, and implications for undecylprodigiosin biosynthesis in Tobin MB, Kovacevic S, Madduri K, Hoskins JA, Skatrud PL, Vining LC, et al. Localization of Streptomyces . Mol Microbiol 2005;56:971 –89. the lysine ε-aminotransferase ( lat ) and δ-( L-α-aminoadipyl)- L-cysteinyl- D-valine Wiyakrutta S, Meevootisom V. A stereo-inverting D-phenylglycine aminotransferase from synthetase (pcbAB) genes from Streptomyces clavuligerus and production of lysine Pseudomonas stutzeri ST-201: puri fication, characterization and application for ε-aminotransferase activity in Escherichia coli . J Bacteriol 1991;173:6223 –9. D-phenylglycine synthesis. J Biotechnol 1997;55:193 –203. Toney MD. Computational studies on nonenzymatic and enzymatic pyridoxal phosphate Wybenga GG, Crismaru CG, Janssen DB, Dijkstra BW. Structural determinants of the catalyzed decarboxylations of 2-aminoisobutyrate. Biochemistry 2001;40:1378 –84. β-selectivity of a bacterial aminotransferase. J Biol Chem 2012;287:28495 –502. Toney MD. Controlling reaction speci ficity in pyridoxal phosphate enzymes. Biochim Xu Y, Labedan B, Glansdorff N. Surprising arginine biosynthesis: a reappraisal of the Biophys Acta 2011;1814:1407 –18. enzymology and evolution of the pathway in microorganisms. Microbiol Mol Biol Toney MD, Hohenester E, Cowan SW, Jansonius JN. Dialkylglycine decarboxylase struc- Rev 2007;71:36 –47. ture: bifunctional active site and alkali metal sites. Science 1993;261:756 –9. Yagi T, Misono H, Tanizawa K, Yoshimura T, Soda K. Characterization of the half and Toney MD, Hohenester E, Keller JW, Jansonius JN. Structural and mechanistic analysis of overall reactions catalyzed by L-lysine:2-oxoglutarate 6-aminotransferase. J Biochem two re fined crystal structures of the pyridoxal phosphate-dependent enzyme 1991;109:61 –5. dialkylglycine decarboxylase. J Mol Biol 1995;245:151 –79. YASARA Structure, Version 13.6.16 YASARA Biosciences, 2013. Townsend C, Gunsior M, Muller U, Van FA, Sonke T. Fermentative production of Yonaha K, Toyama S, Yasuda M, Soda K. Properties of crystalline ω-amino acid: pyruvate D-hydroxyphenylglycine and D-phenylglycine. WO International Patent Application aminotransferase of Pseudomonas sp. F-126. Agric Biol Chem 1977;41:1701 –6. WO2002034921, 2002. Yorifuji T, Ishihara T, Naka T, Kondo S, Shimizu E. Purification and characterization of Tripathi SM, Ramachandran R. Direct evidence for a glutamate switch necessary for polyamine aminotransferase of Arthrobacter sp. TMP-1. J Biochem 1997;122:537 –43. substrate recognition: crystal structures of lysine ε-aminotransferase ( Rv3290c ) Yun H, Lim S, Cho BK, Kim B-G. ω-Amino acid:pyruvate transaminase from Alcaligenes from Mycobacterium tuberculosis H37Rv . J Mol Biol 2006;362:877 –86. denitri ficans Y2k-2: a new catalyst for kinetic resolution of β-amino acids and amines. Tufvesson P, Lima-Ramos J, Jensen JS, Al-Haque N, Neto W, Woodley JM. Process consid- Appl Environ Microbiol 2004;70:2529 –34. erations for the asymmetric synthesis of chiral amines using transaminases. Zhao S, Kumar R, Sakai A, Vetting MW, Wood BM, Brown S, et al. Discovery of new Biotechnol Bioeng 2011;108:1479 –93. enzymes and metabolic pathways by using structure and genome context. Nature Untrau S, Lebrihi A, Germain P, Lefebvre G. Lysine catabolism in Streptomyces ambofaciens 2013;502:698 –702. producer of macrolide antibiotic, spiramycin. Curr Microbiol 1992;25:313 –8. Van Arsdell SW, Perkins JB, Yocum RR, Luan L, Howitt CL, Prasad Chatterjee N, et al. Removing a bottleneck in the Bacillus subtilis biotin pathway: BioA utilizes lysine ! 722(")"064@,!6,

R ")4'5,61,6%",2418&!"!,@,5"55&10, 2)!D)DD398!H%D!'C)%E)(!FD381!P!!)CD398!pWuWiWiepfW!23D!D90EH%C)!'%8!&)!9&E%38)(!0C97!'2C@(381)C!B8'W!c09C! 0C))!09C!DEF()8EDd!9C!'97A36)(!0C97!3ED!9A)8!D9FC')!'9()W!G8!38EC9(F'E398!E9!3ED!FD%1)!'%8!&)!09F8(!38!E2)!P! "353!c 2EEAV``HHHWAP796H353W9C1` dW! L2%38!G!90!E2)!DECF'EFC)D!2%G)!&))8!%6318)(!H3E2!E2)!')%6318!'977%8(W!N)'%FD)!%66!DECF'EFC)D!%C)!%6318)(!E9!E2)! 2F7%8!C8VSQ!G!DECF'EFC)!qGU!E2)P!7312E!)%D36P!&)!'97A%C)(W!2)!03CDE!9&4)'E!Xq9%EbrXb8F7Y!'98E%38D!rX! 8F7&)CD!()C3G)(!0C97!E2)!rX!8F7&)C381!D'2)7)!90!E2)!C8!XNU!%AA63)(!E9!E2)!DECF'EFC)!90!E2)!2F7%8!C8VSQ! G!DECF'EFC)!qGU!H236)!E2)!9&4)'E!!Xq9%Eb6%&)6Y!'98E%38D!E23D!8F7&)CD!%D!6%&)6D!09C!E2)!LSD!90!)%'2!%7389!%'3(W!G66! 9E2)C!9&4)'ED!'98E%38!E2)!9C3138%6!8F7&)C381!0C97!E2)!'CPDE%6!DECF'EFC)DW!8!E2)!6)0E!90!E2)!P!H38(9HU!E2)! C)%()C!H366!038(!D')8)!&FEE98D!E2%E!C)'%66!E2)!031FC)D!90!E2)!7%38!E)IE!H2)8!'63'5)(W!B0!A)C09C7%8')!3D!%8!3DDF)U! C)(F'381!E2)!X(3DA6%P!BF%63EPY!H366!2)6A!c38!E2)!7)8F!X3DA6%P`F%63EPdW!

S ")4'5,61,6%",2418&!"!,!7(,575646",4" 1$0&6&10,)" %0&5),)18&", 2)!79G3)!G3DF%63D)D!E2)!(F%6!DF&DEC%E)!C)'9183E398!90!%738)!EC%8D%738%D)D!%8(!AC)D)8ED!%!D37A6303)(!7)'2%83D7!90! E2)!EC%8D%738%E398!C)%'E398W!B8!E2)!03CDE!2%60!C)%'E398U!E2)!38E)C8%6!%6(3738)!90!E2)!%738)!EC%8D%738%D)!cXN!BXV! r] d!3D!'98G)CE)(!E9!E2)!!09C7!H3E2!E2)!'98'973E%8E!'98G)CD398!90! a%6%838)!E9!APCFG%E)W!!rsu!A938ED! E9H%C(D!E2)!'90%'E9C!E9!'99C(38%E)!%6%838)YD!'%C&9IP6%E)W!2)!79G)7)8E!90!rsuYD!D3()!'2%38!9FE!90!E2)!%'E3G)!D3E)! 'C)%E)D!DA%')!E9!0%'363E%E)!%')E9A2)898)!&38(381!%8(!'98G)CD398!E9!c !dapaA2)8P6)E2P6%738)!38!E2)!D)'98(!2%60! C)%'E398W! 2)!79G3)!H%D!AC)A%C)(!FD381!E2)!#GGG eqf!%8(!P epf!AC91C%7DW!r] !H%D!FD)(!%D!E2)!DE%CE381!DECF'EFC)W!2)! C)%'E398!38E)C7)(3%E)D!H)C)!79()66)(!FD381!#GGG!%AA6P381!(9'5381!%8(!796)'F6%C!(P8%73'D!D37F6%E398!7)E29(DW! 2)!AFCA9D)!90!E2)!79G3)!3D!7%386P!E9!G3DF%63Q)!(F%6!DF&DEC%E)!C)'9183E398!%8(!E2)!7%38!DE)AD!(FC381!E2)!03CDE!2%60! C)%'E398W!2)C)09C)U!E2)!38E)C7)(3%E)D!%8(!)DA)'3%66P!E2)!EC%8D3E398!DE%E)D!(9!89E!C)AC)D)8E!DECF'EFC)D!H3E2!E2)! 69H)DE!A9DD3&6)!)8)C13)D!%D!98)!'9F6(!9&E%38!&P!FD381!!D37F6%E398DW!97)!c()dAC9E98%E398!DE)AD!%C)!973EE)(U! )DA)'3%66P!38!E2)!6%E)!DE%1)D!90E!E2)!03CDE!2%60!C)%'E398!c'%C&3896!%738)!09C7%E398!%8(!2P(C96PD3DdU!%D!E2)!()E%36D!%C)!89E! P)E!589H8W!N%D)(!98!E2)!38E)C7)(3%E)!DECF'EFC)D!1)8)C%E)(!H3E2!#GGGU!E2)!%837%E398D!&)EH))8!E2)!C)%'E398! 38E)C7)(3%E)D!H)C)!'C)%E)(!&P!FD381!PYD!79CA2381!C9FE38)W!2)!D)'98(!2%60!C)%'E398!H%D!79()66)(!986P!%D!%! D29CE)8)(!G)CD398W!

!!

! p! ! T 722(")"064@,!6,#&$74"5, 3D Number 1 17 18 25 26 32 1D7VA 1 lndDATFWRNARQHLVRYGG...... t...... FEPMIIER..AK GSFVY 1MLZB 1 mTTDDLAFDQRHILHP.....ytsmtspl....PVYPV..s.AE GCELI 1OHVC 1 fdydgplmktevpgprSRELMKQLNIIQNA...... eav...... HFFCNYeeSR GNYLV 1VEFA 1 WRALLEAEKTLDSGVYN...... KHDLLIVR..GQ GARVW 2EO5A 1 mlSRKIIEESDIYLATSTR.....dpe...... LFPLVIDH..GE GVWIY 2GSAB 1 fktikSDEIFAAAQKLMPG...gvsspvrafksvgg .QPIVFDR..VKDAYAW 2JJGA 1 ttPDRVHEVLGRSMLV...... d...... GLDIVLDLtrSG GSYLV 2OATA 1 gpptSDDIFEREYKYGAHNYH...... PLPVALER..GK GIYLW 2ORDA 1 ki...HHHHHHMYLMNTYS...... RFPATFVY..GK GSWIY 2ZUKB 1 KALYDRDGAAIGNLQ...... klr...... FFPLAISG..GR GARLI 3A8UX 1 aslasql...... KLDAHWMPYT...anrnflr.....DPRLIVA..AE GSWLV 3DU4A 1 mTHDLIEKSKKHLWLPFT...qmkdyden.....PLIIES..GT GIKVK 3GJUA 1 gmlnqsne...LNAWDRDHFFHPST.hmgthargesp....TRIMAG..GE GVTVW 3HMUB 1 itnhmpTAELQALDAAHHLHPFS..annalgeeg.....TRVITR..AR GVWLN 3I5TB 1 a...VGAAMRDHILLPAQ..emaklgksa.....QPVLTH..AE GIYVH 3N5MA 1 snamktkqTDELLAKDEQYVWHGMR....pfspn...... STTVGAK..AE GCWVE 3NX3B 1 k...... RFDIVLEK..GQ GVYLF 3Q8ND 1 tltqerrlvtaipgpiSQELQARKQSAVAAGV...... gv...... TLPVYVVA..AG GGVLA 4A0GA 277 vfkalketmvlanlerlerln..GMAKLAGEVFWWP...ftqhklvhqet....VTVIDS..RC GENFS 4AO9B 2 haaidqaladayrrftdanpaSQRQFEAQARYMP....gansrsvlfyapfp ... LTIAR..GE GAALW 4E3QA 1 nkpq.SWEARAETYSLYG....ftdmpslhqrg....TVVVTH..GE GPYIV

3D Number 33 36 45 46 66 67 73 1D7VA 37 DAD...... GRAIL DFT SG QMSAVL..... GHCH PEIVSVIGEYA...... GK LDHLF.. 1MLZB 38 LSD...... GRRLV DGM SSWWAAIH..... GYNH PQLNA AMKS QI...... DAMSHVM.. 1OHVC 49 DVD...... GNRML DLY SQISSIPI..... GYSH PALVKLVQ...... qpqnvstfi ...... NRPA.. 1VEFA 33 DAE...... GNEYI DCVG GYGVANL..... GHGN PEVVE AVKR QA...... ET LMAMP.. 2EO5A 38 DVD...... GNKYL DFT SG IGVNNL.gwp ... SH PEVIKIGIE QM...... QK LAHAA.. 2GSAB 48 DVD...... GNRYI DYVGTWGPAIC..... GHAH PEVIE ALKVAM...... EKGTSFG.. 2JJGA 35 DA...it ..G RRYL DMFTFVASSAL..... GMNP PA...... lvddrefhaelmqaa ..L NKPS.. 2OATA 37 DVE...... GRKYF DFL SSYSAVNQ..... GHCH PKIVN ALKS QV...... DK LTLTS.. 2ORDA 32 DEK...... GNAYL DFT SG IAVNVL..... GHSH PRLVE AIKD QA...... EK LIHCS.. 2ZUKB 34 EEN...... GRELI DLSGAWGAASL..... GYGH PAIVA AVSAAA...... an 3A8UX 39 DDK...... GRKVY DSL SG LWTCGA..... GHTRKEIQE AVAK QL...... ST LDYSP.. 3DU4A 40 DIN...... GKEYY DGF SSVWLNVH..... GHRKKELDD AIKK QL...... GKIAHST.. 3GJUA 47 DNN...... GRKSI DAFA GLYCVNV..... GYGRQKIAD AIAT QA...... KN LAYYH.. 3HMUB 46 DSE...... GEEIL DAMA GLWCVNI..... GYGRDELAEVAAR QM...... RE LPYYN.. 3I5TB 38 TED...... GRRLI DGPA GMWCAQV..... GYGRREIVD AMAH QA...... MV LPYAS.. 3N5MA 45 DIQ...... GKRYL DGM SG LWCVNS..... GYGRKELAE AAYK QL...... QT LSYFP.. 3NX3B 17 DDK...... AKKYL DFS SG IGVCAL..... GYNHAKFNAKIKA QV...... DK LLHTS.. 3Q8ND 50 DAD...... GNQLI DFG SG IAVTTV..... GNSA PAVVD AVTQ QV...... AAFTHTC.. 4A0GA 335 IYKasdnssLSQQF DACASWWTQGPdptfq .... AELAREMGYTA...... ARFGHVM.. 4AO9B 62 DAD..g.... HRYA DFIAEYTAGVY..... GHSA PEIRD AVIEAM...... QGGINLT.. 4E3QA 42 DVN...... GRRYL DAN SG LWNMVA..... GFDHKGLID AAKA QY...... ERFPGYH..

3D Number 74 84 90 91 118 1D7VA 78 ..sgm....LSRPVVD LA T.R LANIT Ppgl.DRALLLST GAESN EAAIRM AKLVT...... 1MLZB 79 ..fggi...THAPAIE LCR.K LVAMT Pqpl.EC VF LAD SG SV AVEVAMKMALQYWQA KG...... 1OHVC 93 ..lgil...PPENFVEKLReSLLSVA Pkgm.SQLITMAC GSCSN ENAFKTIFMWYRS KErgqsafskee 1VEFA 74 ..qtl....PTPMRGEFYR.T LTAIL Ppel.NR VF PVN SG TEA NEAALKFARAHT...... 2EO5A 80 ..andf...YNIPQLE LA K.K LVTYS PgnfqKK VFF SN SG TEA IEASI KVVKNTG...... 2GSAB 89 ...a.....PCALENV LAE. MVNDAV P. si.EM VRFVN SG TEA CMAVLRLMRAYT...... 2JJGA 81 ..nsdv...YSVAMARFV E. TFARVLGdpalPHL FF VEG GAL AVENALKAAFDWKSRHN...... qah 2OATA 78 ..raf....YNNVLGEYE E. YITKLFN.y..HK VLPMNT GVEA GETACKLARKWGYTVK...... g 2ORDA 73 ..nlf....WNRPQME LAE. LLSKNTF.g..GK VFF ANT GTEA NEAAIKIARKYGKK KS...... 2ZUKB 70 pagatilsaSNAPAVT LAE. RLLASF PgegtHKIW FGH SG SD ANEAAYRAIVKAT...... 3A8UX 80 ..gfqy...GHPLSFQ LAE. KITDLT Pgnl.NH VFF TD SG SECALT AVKMVRAYWRL KG...... 3DU4A 81 ..llgm...TNVPATQ LAE. TLIDIS Pkkl.TR VF YSD SG AEA MEIALKMAFQYWKNI G...... 3GJUA 88 .ayvgh...GTEASIT LA K.MIIDRA Pkgm.SR VYFGL SG SD ANETNI KLIWYYNNVL G...... 3HMUB 87 .tffkt...THVPAIA LA Q.K LAELA Pgdl.NH VFF AGG GSEA NDTNIRMVRTYWQN KG...... 3I5TB 79 ..pwym...ATSPAAR LAE. KIATLT Pgdl.NRI FF TTG GST AVDS ALRFSEFYNNVL G...... 3N5MA 86 ..msq....SHEPAIK LAE. KLNEWL.gge.YVI FF SN SG SEA NETAFKIARQYYAQ KG...... 3NX3B 58 ..nly....YNENIAAA AK.N LAKASA.l..ER VFF TN SG TESI EGAMKTARKYAFN KG...... 3Q8ND 91 ..fmvt...PYEGYVKV AE. HLNRLT PgdheKRTALFN SG AEA VENAVKIARAYT...... 4A0GA 383 .fpenvy.... EPALKC AE. LLLDGVGkgwaSR VYFSDN GST AIEIALKMAFRKFCV...... d 4AO9B 103 ...g.....HNLLEGR LA R.LICERF P. qi.EQLR FTN SG TEA NLM ALTA ALHFTG...... 4E3QA 83 . affgr...MSDQTVM LSE. KLVEVS Pfds.GR VF YTN SG SEA NDTMV KMLWFLHAAE G......

! q! ! 3D Number 119 142 143 149 150 154 1D7VA 125 .g...... KYEIVGFAQSWHG MTGAAA SATYS.a...GRKGV...... G PAAV...... 1MLZB 131 .e...... RQRFLTFRNG YHG DTFGAM SVCDP...... dnsmhslw KGYLP...... 1OHVC 156 letc minqapgcp DYSILSFMGAFHG RTMGCLAT TH..s kAIHKIDI...... ps 1VEFA 121 .g...... RKKFVAAMRGFSGRTMGSL SVTWE...PKYREP F...... LPLVE...... 2EO5A 129 ...... RKYIIAFLGGFHG RTFGSI SLTA..s kAVQRSIV...... G PFMP...... 2GSAB 133 .g...... RDKIIKFEGC YHG...... a 2JJGA 137 gidp al...... GTQVLHLRGAFHG RSGYTL SLTNT.k.PTITAR F...... p k 2OATA 128 iqky...... KAKIVFAAGNFWGRTLSAI SSSTD...PTSYDG F...... GPFMP...... 2ORDA 122 .ek...... KYRILSAHNSFHG RTLGSLTA TGQ...PKYQKP F...... EPLVP...... 2ZUKB 124 .g...... RSGVIAFAGA YHG CTVGSMAFSGH...... a 3A8UX 132 qat...... KTKMIGRARG YHG VNIAGT SLGGV...NGNRKL F...... GQPMQ...... 3DU4A 133 kpe...... KQKFIAMKNG YHG DTIGAV SVGSI...ELFHHVY...... G PLMF...... 3GJUA 141 rpe...... KKKIISRWRG YHG SGVMTG SLTGL...DLFHNA F...... DLPRA...... 3HMUB 140 qpe...... KTVIISRKNA YHG STVASSALGGM...AGMHAQS...... GLIP...... 3I5TB 131 rpq...... KKRIIVRYDG YHG STALTAAC TGR...TGNWPN F...... DIAQD...... 3N5MA 136 eph...... RYKFMSRYRG YHG NTMATMAA TGQ...AQRRYQY...... E PFAS...... 3NX3B 107 .vk...... GGQFIAFKHSFHG RTLGAL SLTAN...EKYQKP F...... KPLIS...... 3Q8ND 140 .r...... RQAVVVFDHA YHG RTNLTMAM TAK nqp..YKHG F...... GPFAN...... 4A0GA 435 hnfi...... VVKVIALRGS YHG DTLGAMEA...... qapspyt gf 4AO9B 148 ...... RRKIVVFSGG YHG...... gvlgfg ar 4E3QA 136 kpq...... KRKILTRWNA YHG VTAVSA SMTGK pyn....SV F...... GLPLP......

3D Number 155 162 163 172 1D7VA 161 ...... GSFAI PAP....ftyrprferngayd...YLAELDYAFD...... 1MLZB 169 ...... ENLFA PAP...... qsrmgewder.....DMVGFARLMA...... 1OHVC 203 f...... DWPIA PFP.rlkypleefvkenqqeear.CLEEVEDLIV...... 1VEFA 158 ...... PVEFI PYN...... DVEALKRAVD...... 2EO5A 166 ...... GVIHV PYPnpyrnpwhingyenpselvnrVIEFIED...... 2GSAB 148 ...... NTLTT PYN...... DLEAVKALFA...... 2JJGA 177 f...... DWPRIDA P..ymrpgldepamaaleae..ALRQARAAFE...... 2OATA 168 ...... GFDII PYN...... DLPALERALQ...... 2ORDA 160 ...... GFEYFEFN...... NVEDLRRKMS...... 2ZUKB 150 d...... GLILL PYP....dpyrpyrndptgda...ILTLLTEKLA...... 3A8UX 171 ...... DVDHL PHT..llasnaysrgmpkeggia.LADELLKLIE...... 3DU4A 172 ...... ESYKA PIP...yvyrsesgdpdecrdq..ZLRELAQLLE...... 3GJUA 180 ...... PVLHTEA P..yyfrrtdrsmseeqfsqh.CADKLEEMIL...... 3HMUB 178 ...... DVHHINQ P..nwwaeggdmdpeefgla..RARELEEAIL...... 3I5TB 170 ...... RISFLSS P...nprhagnrsqeafldd..LVQEFEDRIE...... 3N5MA 175 ...... GFLHVTP P..dcyrmpgiereniydve..CVKEVDRVMT...... 3NX3B 145 ...... GVKFAKYN...... DISSVEKLVN...... 3Q8ND 178 ...... EVYRV PTS.....ypfrdgetdgaa....AAAHALDLIN...... 4A0GA 469 lqqpwytgRGLFLDP P...... tvflsngswnislpesfseiap 4AO9B 169 pspttvpfDFLVL PYN...... DAQTARAQIE...... 4E3QA 174 ...... GFVHLTC P..hywrygeegeteeqfvar.LARELEETIQ......

3D Number 173 182 187 200 1D7VA 193 ..lidrqssg...... NL AA FIA EP.I LSSG...... G IIELPDGYMAALK 1MLZB 197 ....ahrh...... EI AA VII EPi VQ GAG...... G MRMYHPEWLKRIR 1OHVC 241 ..kyrkkkk...... TV AGIIV EP.I QSEG...... G DNHASDDFFRKLR 1VEFA 176 .....e...... ET AA VIL EP. VQ GEG...... G VRPATPEFLRAAR 2EO5A 202 .yifvnlvppe...... EV AGIFF EP.I QGEG...... G YVIPPKNFFAELQ 2GSAB 166 ....enpg...... EI AGVIL EP.I VGNS...... GFIVPDAGFLEGLR 2JJGA 213 ....trph...... DI ACFVA EP.I QGEG...... G DRHFRPEFFAAMR 2OATA 186 .....dp...... NV AA FMV EP.I QGEA...... GVVVPDPGYLMGVR 2ORDA 178 .....e...... DVC AVFL EP.I QGES...... GIVPATKEFLEEAR 2ZUKB 183 ...avpag...... SIG AAFI EP.I QSDG...... G LIVPPDGFLRKFA 3A8UX 207 ...lhdas...... NI AA VFV EP. LA GSA...... GVLVPPEGYLKRNR 3DU4A 206 ....ehhe...... EI AA LSI ES...... mvqgas GMIVMPEGYLAGVR 3GJUA 216 ...aegpe...... TI AA FIG EP.I LGTG...... G IVPPPAGYWEKIQ 3HMUB 213 ...elgen...... RV AA FIA EP. VQ GAG...... G VIVAPDSYWPEIQ 3I5TB 204 ...slgpd...... TI AA FLA EP.I LASG...... G VIIPPAGYHARFK 3N5MA 210 ...welse...... TI AA FIM EP.I ITGG...... G ILMAPQDYMKAVH 3NX3B 163 .....e...... KTC AIIL ES.VQ GEG...... G INPANKDFYKALR 3Q8ND 208 ...kqvgad...... NV AA VVI EP. VH GEG...... G FVVPAPGFLGALQ 4A0GA 507 eygtftsrdeifdksrdastlariysaylskhlaHVG ALII EPvI HGAG...... G MHMVDPLFQRVLV 4AO9B 195 ....rhgp...... EI AVVLV EP. MQ GAS...... GCIPGQPDFLQALR 4E3QA 210 ...regad...... TI AGFFA EP. VM GAG...... G VIPPAKGYFQAIL

! r! ! 3D Number 201 206 207 223 224 229 235 247 248 1D7VA 229 RK CEAR...... GM LLI LDE AQTG VGRTG..... TM FA C.QRD GVT... PD IL TLS KTLGA G...... 1MLZB 230 ...... kicdreGI LLI ADE IATG FGRTG..... KL FA C.EHAEIA... PD ILCLG KALT GG...... T 1OHVC 276 DISRKH...... GCAFLV DEV QTG GGSTG..... KFW AH.EHW GL. ddp ADVM TFS KKMM...... 1VEFA 205 EITQEK...... GA LLI LDE IQTG MGRTG..... KR FA F.EHF GIV... PD IL TLA KALG GG...... 2EO5A 240 KLAKKY...... GI LL VD DEV QMGLGRTG..... KL FA I.ENFNTV... PD VI TLA KALG GG...... I 2GSAB 198 EITLEH...... DA LL VF DEV MTG FRIA...... YGGVQ.EKF GVT... PD LT TLG KIIG G...... G 2JJGA 245 EL CDEF...... DA LLI FDEV QTG CGLTG..... TAW AY.QQLDVA... PD IVAFG KKTQ...... 2OATA 216 EL CTRH...... QV LFIADE IQTG LA RTG..... RWL AV.DYENVR... PD IVLLG KALS GG...... L 2ORDA 207 KL CDEY...... DA LL VF DEV QCGMGRTG..... KL FA Y.QKY GVV... PD VL TTA KGLG GG...... 2ZUKB 216 DI CRAH...... GI LVVC DEV KVGLA RSG..... RLHCF.EHE GFV... PD ILVLG KGLG GG...... 3A8UX 240 EI CNQH...... NI LL VF DEV ITG FGRTG..... SM FGA.DSF GVT... PD LMCIA KQVTN G...... A 3DU4A 239 EL CTTY...... DV LMIVDEV ATG FGRTG..... KM FA C.EHENVQ... PD LMAAG KGIT GG...... Y 3GJUA 249 AVLKKY...... DV LL VA DEV VTG FGR LG..... TM FGS.DHY GIK... PD LI TIAZ.....kgltsaY 3HMUB 246 RI CDKY...... DI LLI ADEV ICGFGRTG..... NW FGT.QTM GIR... PHIM TIA KGLSS G...... Y 3I5TB 237 AI CEKH...... DI LYISDEV VTG FGR CG..... EW FA SeKVF GVV... PD II TFA KGVTS G...... Y 3N5MA 243 ET CQKH...... GA LLI SDEV ICGFGRTG..... KA FGF.MNYDVK... PD II TMA KGITSA...... Y 3NX3B 192 KL CDEK...... DI LLI ADE IQCGMGR SG..... KF FA Y.EHAQIL... PD IM TSA KALGC G...... 3Q8ND 242 KW CTDN...... GAVFVA DEV QTG FA RTG..... AL FA C.EHENVV... PD LIVTA KGIA GG...... 4A0GA 570 NE CRNR...... KIPV IFDEV FTG FW RLG..... VETTT.ELL GCK... PD IACFA KLLT GG...... M 4AO9B 227 ESATQV...... GA LL VF DEV MT...... srlap .HGLA.NKL GIR...S DLT TLG KYIG GG...... 4E3QA 243 PILRKY...... DIPV ISDEV ICGFGRTG..... NTWGC.VTYDFT... PD AIISS KNLTA G...... F

3D Number 249 258 259 266 267 271 284 1D7VA 276 L PLA AIVTSA...... AIEERAH...... elgy...... LFYT...... THVSD PLPA AVGLR 1MLZB 278 MTLS ATLTTR...... EVAET...... isngeagcf....MHGP...... TFM GNP LA CA AANA 1OHVC 323 ..TGGFFHKE..efr...... PNA P...... y...... RIFN...... TWL GDPSKNLLLAE 1VEFA 252 V PLGVAVMRE...... EVARSMPK...... g...... GHGT...... TFG GNP LAM AAGVA 2EO5A 288 M PIG ATIFRK...... DLDFK...... tfg...... GN ALA CA IGSK 2GSAB 244 L PVG AYGGKR...... EIMQLVA...... pagpm...... YQAG...... TLS GNP LAMTAGIK 2JJGA 290 ..VCGVMAGRrvdevadNVFA...... vps...... RLNS...... TWG GN LTDMVRARR 2OATA 264 Y PVS AVLCDD...... DIMLTIK P...... g...... EHGS...... TYG GNP LG CRVAIA 2ORDA 254 V PIG AVIVNE..ran...... VLE P...... g...... DHGT...... TFG GNP LA CRAGVT 2ZUKB 263 L PLS AVIAPA...... EILDCA...... sa...... FAMQ...... TLH GNP ISA AAGLA 3A8UX 288 I PMG AVIAST...... EIYQT....fmnqptpeyavef..PHGY...... TYSAH PVA CA AGLA 3DU4A 287 L PIAVTFATE...... DIYKAFY....ddyenlktf....FHGH...... SYT GN QLG CA VALE 3GJUA 298 A PLSGVIVAD...... RVWQVLV....qgsdklgsl....GHGW...... TYSAH PICV AAGVA 3HMUB 294 A PIGGSIVCD...... EVAHV...... igkdef.....NHGY...... TYS GHPVAA AVALE 3I5TB 286 V PLGGLAISE...... AVLARI.....sgenakgswf...TNGY...... TYSNQ PVA CA AALA 3N5MA 291 L PLS ATAVKR...... EIYEAFK.....gkgeyeff....RHIN...... TFG GNP AA CA LALK 3NX3B 239 LSVG AFVINQ...... KVASNSL...... eagdhgstyg ...GNP LV CA GVNA 3Q8ND 289 L PLS AVTGRA...... EIMDGPQS...... g...... GLGG...... TYG GNP LA CA AALA 4A0GA 618 V PLAVTLATD...... AVFDSFS.....gdsklkal....LHGH...... SYSAHAMG CA TAAK 4AO9B 272 MSFG AFGGRA...... DVMALF...... dprtgpl.....AHSG...... TFNN NVMTM AAGYA 4E3QA 291 F PMG AVILGP...... elskrletaieaieefPHGF...... TAS GHPVG CA IALK

3D Number 285 291 292 299 311 312 317 1D7VA 315 V LDVVQR.d..GLVARAN.VM GDR LRRG LLD L...... merfdc...... IGDV RG 1MLZB 320 S LAILES.g..DWQQQVA.DIEVQ LREQ LAPA...... rdaem...... VADV RV 1OHVC 357 VINIIKR.e..DLLSNAA.HA GKV LLTG LLD L...... qarypqf...... ISRV RG 1VEFA 289 AIRYLER.t..RLWERAA.EL GPWFMEK LRAI...... pspk...... IREV RG 2EO5A 317 VIDIVKD.....LLPHVN.EI GKIFAEE LQG L...... ADDV RG 2GSAB 284 T LELLRQ.p..GTYEYLD.QITKR LSDG LLAI...... aqetgh...... AACGGQ 2JJGA 330 I LEVIEA.e..GLFERAV.QH GKY LRAR LDE L...... aadfpav...... VLDP RG 2OATA 301 A LEVLEE.e..NLAENAD.KL GII LRNE LMK L...... psdv...... VTAV RG 2ORDA 290 VIKELTK.e..GFLEEVE.EK GNY LMKK LQEM...... keeydv...... VADV RG 2ZUKB 299 V LETIDR.d..DLPAMAE.RK GRL LRDG LSE L...... akrhpl...... IGDI RG 3A8UX 334 A LCLLQK.e..NLVQSVA.EVAPHFEKA LHGI...... kgakn...... VIDI RN 3DU4A 331 N LALFES.e..NIVEQVA.EKSKK LHFL LQD L...... halph...... VGDI RQ 3GJUA 342 N LELIDE.m..DLVTNAG.ET GAYFRAE LAKA...... vgghkn...... VGEV RG 3HMUB 333 N LRILEE.e..NILDHVR nVAAPY LKEKWEA L...... tdhpl...... VGEAKI 3I5TB 330 NIELMER.e..GIVDQAR.EMADYFAAA LAS L...... rdlpg...... VAET RS 3N5MA 334 N LEIIEN.e..NLIERSA.QM GSL LLEQ LKEE...... igehpl...... VGDI RG 3NX3B 277 VFEIFKE.e..KILENVN.KLTPY LEQS LDE L...... inefdf...... CKKRK G 3Q8ND 326 VIDTIER.e..NLVARAR.AI GETMLSR LGA L...... aaadpr...... IGEV RG 4A0GA 661 AIQWFKD...... petnhnitsqgktlrelwdeelvqqisshsaVQRVVV 4AO9B 313 G LTK...lftpEAAGALA.ER GEA LRAR LNA L...... canegvam...... QFT G 4E3QA 335 AIDVVMN.e..GLAENVR.RLAPRFEER LKHI...... aerpn...... IGEY RG

! s! ! 3D Number 318 328 329 341 349 350 361 1D7VA 355 R GLLLGVEIVK....drrtkepad....GLGAKITRECMN.L GLSMNIVQ.lpgmg.GVFRIA PPL TVS 1MLZB 359 L GAIGVVETTH...... p...... VNMAALQKFFVE.Q GVWIRPFG...... KLIYLM PP YIIL 1OHVC 398 R GTFCSFDTPD...... e...... SIRNKLISIARN.K GVMLGGCG...d...KSIRFR PTLVFR 1VEFA 327 M GLMVGLELKE...... KAAPYIARLEK eHRVLALQAG...p...TVIRFL PPL VIE 2EO5A 349 I GLAWGLEYNE...... k...... KVRDRIIGESFK.R GLLLLPAG...r...SAIRVI PPL VIS 2GSAB 324 VSGMFGFFFT..egpvhnyedakksdl.QKFSRFHRGMLE.Q GIYLAPSQ...fe.....AGFTS LAHT 2JJGA 371 R GLMCAFSLPT...... t...... ADRDELIRQLWQ.RAVIVLPAG...a...DTVRFR PPL TVS 2OATA 339 K GLLNAIVIKE...... tkd...... WDAWKVCLRLRD.N GLLAKPTH...g...DIIRFA PPL VIK 2ORDA 330 M GLMIGIQFRE...... e...... VSNREVATKCFE.NKLLVVPAG...n...NTIRFL PPL TVE 2ZUKB 339 R GLACGMELVC....drqsrepar....AETAKLIYRAYQ.L GLVVYYVG..mng..NVLEFT PPL TIT 3A8UX 373 F GLAGAIQIAP...... rdgdai.....VRPFEAGMALWK.A GFYVRFGG...... DTLQFG PTFNSK 3DU4A 370 L GFMCGAELVR...sketkepypadr..RIGYKVSLKMRE.L GMLTRPLG...... DVIAFL PPL AST 3GJUA 382 D GMLAAVEFVA...dkddrvffdasq..KIGPQVATALAA.S GVIGRAMP...qg..DILGFA PPL CLT 3HMUB 373 V GMMASIALTP..nkasrakfasepg..TIGYICRERCFA.NNLIMRHVG...... DRMIIS PPL VIT 3I5TB 369 V GLVGCVQCL...... lgtaedk.....AFTLKIDERCFE.L GLIVRPLG...... DLCVIS PPL IIS 3N5MA 374 K GLLVGIELVN....dketkepidn...DKIASVVNACKE.K GLIIGRNGmttagynNILTLA PPL VIS 3NX3B 317 L GFMQGLSLDK...... s...... VKVAKVIQKCQE.NALLLISCG...e...NDLRFL PPL ILQ 3Q8ND 366 R GAMIAVELVK.....pgttepda....DLTKRVAAAAHA.Q GLVVLTCG..tyg..NVLRFL PPL SMP 4A0GA 705 I GTLFALELKS...... l...... YAKSLLIMLRE.D GIFTRPLG...... NVIYLMCGPCTS 4AO9B 353 I GSLMNAHF..vqgdvrssedlaavdgr .LRQLLFFHLLN.EDIYSSPR...... GFVVLS LPLT 4E3QA 374 I GFMWALEAVK...dkasktpfdgnl..SVSERIANTCTD.L GLICRPLG...... QSVVLC PP FILT

3D Number 362 379 1D7VA 413 EDEI DLGLSLLGQAIERAl 1MLZB 404 PQQLQRLTAAVNRAVQDEtffc 1OHVC 444 DHHAHLFLNIFSDILADF 1VEFA 372 KEDLERVVEAVRAVLA 2EO5A 395 EEEAKQGLDILKKVIKVV 2GSAB 381 EEDI DATLAAARTVMSAL 2JJGA 417 TAEI DAAIAAVRSALPVVt 2OATA 387 EDELRESIEIINKTILSF 2ORDA 376 YGEI DLAVETLKKVLQGI 2ZUKB 395 ETDIHKALDLLDRAFSELsavsneeiaqfagw 3A8UX 423 PQDL DRLFDAVGEVLNKLl 3DU4A 426 AEELSEMVAIMKQAIHEVtsled 3GJUA 440 REQA DIVVSKTADAVKSVfa 3HMUB 430 PAEI DEMFVRIRKSLDEAqaeiekqglmkse 3I5TB 419 RAQI DEMVAIMRQAITEVsaahgl 3N5MA 435 SEEIAFVIGTLKTAMERI 3NX3B 363 KEHI DEMSEKLRKALKSF 3Q8ND 421 DHLL DEGLDILAAVFAEVk 4A0GA 749 PEICRRLLTKLYKRLGEFnrt 4AO9B 408 DADI DRYVAAIGSFIGGHgallpran 4E3QA 430 EAQM DEMFDKLEKALDKVfaeva

@&$74",RD,647 674(,5"!,5"37"0 ",(&$0)"06,1#,((,40,DE,57#)&(@,6")2(6",5647 674"5E !2)!rX!8F7&)CD!%C)!13G)8!38!E2)!03CDE!C9HU! D)BF)8')!8F7&)CD!'9CC)DA98(!E9!E2)!C)D3(F)D!AC)D)8E!38!E2)!%6318)(!'2%38!c38(3'%E)(!&P!E2)!6%DE!6)EE)C!90!E2)!)8QP7)DY!8%7)DdW!L9C)!C)1398D! %C)!D29H8!38!'%A3E%6!6)EE)CD!%8(!C)D3(F)D!38!G%C3%&6)!C)1398D!%C)!89E!%6318)(W!2)!D)BF)8')a0F8'E398!7%EC3I!c%&6)!qd!C)D3(F)D!%C)!23126312E)(! C)(!%8(!E2)!'%E%6PE3'!6PD38)!3D!23126312E)(!P)669HW,

! t! KP, EP,

@&$74",SD,G14",0!,84&(",4"$&105,1#,6%",9%1(",#1(!,6@2",I,TD,!65",IKP,0!,6%",40,DE,IEP,!&52(@"!,&0,6%", 6&8",5&6",4"5&!7"5,1#, 6%",%7)0,40DAO,K,IDED,SKPE, L9C)!C)1398D!%C)!'969FC)(!38!1C)P!cC)D3(F)D!&6F)d!%8(!G%C3%&6)!C)1398D!%C)!'969FC)(!38!P)669HW!G66!%'E3G)! D3E)!'98EC3&FE381!%7389!%'3(D!%8(!E2)!a9C83E238)!7373'5381!3823&3E9C!%C)!D29H8!%D!DE3'5DW!eDA)'3%66P!E2)!aD3()!90!E2)!%'E3G)!D3E)!3D!89E! '98D)CG)(!H3E238!E2)!H296)!096(!EPA)!BU!&FE!'9F6(!&)!%6318)(!38!E2)!C8!!XNW!

u! KP,

EP,

@&$74",TD,647 674(,(&$0)"06,1#,((,40,KH(&'",!65",57#)&(&"5G,F6")2(6"G,5647 674"5 W!L9C)!C)1398D!%C)!D29H8!38!1C)P!%8(!G%C3%&6)! C)1398D!%C)!D29H8!38!P)669HW!L90%'E9CD!%C)!D29H8!%D!9C%81)!DE3'5DW!2)!X!(97%38!90!E2)! "#$%&'()*&* !GVGG!G!cXN!BXV!sGiQd!H%D! 973EE)(!09C!'6%C3EP!C)%D98DW! KP !%8(! EP !'98E%38!E2)!D%7)!%63187)8E!C9E%E)(!%C9F8(!E2)!G)CE3'%6!$a%I3DW!

v! ! (A) O O OH OH OH OH

NH 2 NH 2 NH 2 O NH 2 O alanine L-2-aminobutyrate β-alanine D-3-aminoisobutyrate

NH O O OH N N OH H NH 2 NH 2 O NG,NG-dimethylarginine 5-aminolevulinate

(B) O Glu αKG H2N O O HO HO HO H2N HO H2N H2N H N O 2 O HO NH 2 HO NH 2 OH OH 0"#$%+,$-*1&-*'&'()% neamine

(C) O Glu αKG H2N O O HO HO HO H2N HO H2N HO HO O O HO NH 2 HO NH 2 OH OH !"#/%&'()*#!"#+,$-*.,#0"#$%+,$-*1&-*'&'()% !"#$%&'()*#!"#+,$-*.,)%&'()%

(D) O Glu αKG H2N O O HO HO HO H2N HO H2N HO HO O O HO NH 2 HO NH 2 O O O O OH OH

HO HO NH NH OH 2 OH 2 6-'deamino-6'-oxokanamycin A kanamycin A

(E) H2N Glu αKG H2N O O HO HO HO H2N HO H2N H N H N 2 O 2 O HO O NH 2 HO O NH 2 O OH O OH

O O O OH O OH NH 2 NH 2 O H2N

HO HO HO HO neomycin C 0""" -$%&'()*#0"""#*.*)%*',2()34 !

@&$74",UD,K!!&6&10(,575646",0!,241!7 6,5647 674"5,61, 1)2(")"06,(",TE,,IKP !G((3E398%6!DF&DEC%E)D!%'')AE)(!&P!G6%V16P9I!GD!q!cEPHITP, )%'E398D!'%E%6PD)(!&P!Q6FVuYa9I916F'9Dc%738dP6!GDW!2)!EC%8D%738%E)(!9IP1)8!%8(!83EC91)8!%C)!23126312E)(!38!1C)PW ,9E)!E2%E!E2)!)8QP7)! 986P!()%738%E)D!LuYYY!38!8)97P'38!L!ITPC!&FE!EC%8D%738%E)D!LuY!38!E2)!9E2)C!8)%738)!()C3G)(!DF&DEC%E)D! IEPaIDPW,

!

! w! (A) HO OH O O OH OH N N HS O P O P O N N O N H H O O NH 2 OH N coenzyme A

HO O O O (B) OH OH O OH

NH 2 OH O O O

HO O OH

fumonisin B1

(C) OH OH OH

NH 2 OH OH

hydrolysed fumonisin B1

@&$74",VD,K!!&6&10(,575646",0!,241!7 6,5647 674"5,61, 1)2(")"06,(",TE !cKd!ECF'EFC)!90!'9)8QP7)!G!cL9Gd!c Ed!ECF'EFC)!90!0F7983D3D8! Np!cGd!ECF'EFC)!90!2P(C96PD)(!0F7983D38!N pW,

!

!

!

@&$74",WD,D&##"4"06,(&$0)"06,215&6&105,1#,6%", 6(@6& ,(@5&0", 0,56&((,4"57(6,&0,,5&)&(4,215&6&10&0$,1#,&65,BH)&01,$4172E, 2)!DECF'EFC)D!90! E2)!2F7%8!C8VSQ!G!cXN!pGU!2%D!qsqd!%8(!E2)!qa%7389)E2P6A29DA298%E)VAPC!G!0C97! !$+,(-.++$/.-0.#&1$ !cXN!prqU!2%D!qssd!2%G)! &))8!DFA)C37A9D)(!E9!'97A%C)!E2)!'%E%6PE3'!6PD38)!A9D3E398DW!2)!'90%'E9C!%8(!E2)!'%E%6PE3'!6PD38)!%C)!D29H8!%D!DE3'5DW! x! !

!!!!rsu! !!!!!!su!

!!!!!!sv!

!qsq!

!!!!!!pwt!

!!psq!

!

@&$74",XD,K 6&8",5&6",1#,6%",KK,#41), !"#$#%&'()*+,-"&)*#.$), ,IDE,IDD,TIVPE,psq!3D!AC9&%&6P!(3DCFAE381!TG6%!&38(381!&)EH))8!"svU!pwt! %8(!rtr!%8(!E2)C)09C)!rBt!A9DD)DD)D!986P!69H!TG6%VAPC!%'E3G3EPW!2)!D3()!'2%38D!90!C)D3(F)D!pwt!%8(!psq!D))7!E9!&)!C)6%E3G)6P!06)I3&6)! &)'%FD)!90!E2)!)6)'EC98!()8D3EP!90!E2)!'CPDE%6!DECF'EFC)!%8(!E2)C)09C)!%6D9!E2)!79()6!()A9D3E)(!38!E2)!XN!38(3'%E)D!79C)!E2)8!98)! 9C3)8E%E398W!)D3(F)D!%C)!D29H8!38!1C))8!%8(!!38!9C%81)!DE3'5DW!

pi! Ring opening 1-Aminocyclopropane- 1-carboxylate-deaminase R Enzym β-Elimination + Reduction - OH Serine dehydratase CDP-4-dehydro-6-deoxyglucose reductase + H2O β-Lysine-5,6-aminomutase CO 2H CO 2H β-Elimination - SR Cystathionine- β-lyase O CH 3 Lysine-2,3-aminomutase - SH Cysteine desulfhydrase O NH 2 - NH 3 COOH O - NH 3 2,3-Diaminopropionate- H ammonialyase H Retro Claisen Glycine C-acetyltransferase O OH R cleavage - Cl 3-Chlor-D-alanin- O CDP CO H + 2 dehydrochlorinase NAD H O CO H 2 R Enzym NH 2 CO H 2 - Indol Tryptophane-indol-lyase O 2 NADH PMP - S0 NifS-Protein – R– - Phenol Tyrosine-phenol-lyase NH 2 0 Selenocystein- NH 2 - Se CoA-SH Acetyl-S-CoA O Tryptophan synthase CH 3 Lyase H O (serine + indol) - CO 2 Aspartat- β- H Retro Aldol Serine hydroxy- HO CO 2H O decarboxylase R cleavage methyltransferase -acetylserine-sulfhydrylase OH β-Substitution O CDP - CO 2R Kynureninase R (Acetylserin+H2S=Cys+ac) NH 2 CO 2H COOH Selenocysteine synthase CO 2H Z HO (O-Phosphoserin-tRNA + β-Elimination NH 2 β-proton abstraction NH R=H: –CH 2-THF NH 2 3- 2 SePO 3 + H2O) COOH (radical mechanism) R=Me: –CH 3CHO

NH 2 + HZ – R– – H – R–

R H2O H H Lys H H COO COO COO NH N + R R R 3 H + + + + N H H N H+ N H2O H H H H O COO O OP i COO R + PiO O O O + R OP i OP i OP i NH 3 N O N H H N N N H H H Internal Aldimine External Aldimine Quinonoid-Intermediate Ketimine PMP

H + H+ H – CO 2 –HY Y COOH Phosphorylase N+ Cyclisation Racemisation/ γ-Elimination H Decarboxylation R NH 2 various amino acid PLP decarboxylases Epimerisation 1-Aminocyclopropan- +HZ + O2 1-carboxylate-Synthase R COO - H2O2 + γ-Addition + β-Addition - NH Cystathionine- 3 Ade S NH Threonine Decarboxylation 2 γ-synthase synthase + Oxidation R O DOPA-Decarboxylase COOH various Z aa-racemases/ NH 2 Z COOH COOH O epimerases Decarboxylation R + Ester condensation X NH 2 NH 2 NH 2 O O Acyl-CoA CoA COOH -succinylhomoserin + Cys -phosphohomoserine + H2O = threonine + phosphate COOH Serine-C-palmitoyltransferase NH 2 δ-Aminolevulinate synthase S COOH H2N + Succinat NH 2 !

!$!'(4)'+&#'!$!'3'#'#'%!"#'''8981 #'('"'& '>''#!'?@"0' ""! !

0 "#"* &-*"*-"*/!- 1*/-")'-4$ $%--"- *;"-!;;"!-#--"- "-<=<--"&!(--

$0' ! !12!3!)44%6!!)(!! 45!)5(!2! 8!32%14%0)! 481!!

G@GS@G! C5@A5@ HD!7% 02! 543!!

G@P@P! C5@@ HD!7% 02! 543!! H@G@H! CG5D!2)3"!23!! H@I@G! ;8'!42)3"!23!! H@Q@G! ;'&8'A28'!42)3"!23!! H@R@G! ;(%)0!42)3"!23!! H@T@G! 5'"52!42)3"!23!! ;31244!!(%)042)3"!23!! H@U@G! !'!)0!42)3"!23!! :! 351!2"(%'8! I@S@G! @8 20'3!!C5B5D! P@G@G! 52078'!42)3"!23!! P@G@H! ;' !$8 !!'83!! P@G@UU! 5B5!83!! P@H@G! @8 20!'83!! P@I@G! ;((0)%!'83!! P@P@G! 5B!83!! Q@G@G! !(3!! Q@P@I! :30(!23!! H@Q@G! ;'&8'A28'!42)3"!23!! I@Q@UU! 5B!@8 20!'83!! P@H@G! 5B!@8 20!'83!! 281401$)!38)4$3!! ::! P@H@I! 5B!83!!C1$031$4!3D! 351!2"(%'8! P@I@G! ;((0)%!'83!! P@P@G! 5B!83!! Q@G@G! ;(%)0!% !2!(3!! P@G@G! 52078!'83!! ;')%)!!2!(3!!! P@G@H! ;' !$8 !!'83!! :::! 351!2"(%'8! P@I@G! ;((0)%!'83!! Q@G@G! ;(%)0!% !2!(3!!

8B')%)!!(%)042)3"!23!! H@R@G! ;(%)0!42)3"!23!! :! 351!2"(%'8! P@G@I! 70!% !'83!! S'80#!)!1$031$028'3!! ! H@P@G! S'8038'42)3"!23!3! 351!2"(%'8! 8B83%)!BQ9RB(%)0(543!! :! 351!2"(%'8! Q@P@I! :30(!23!!

83%)!BH9IB(%)0(543!! ::! 351!2"(%'8! Q@P@I! :30(!23!!

! ! ! -

! GH! ! !"1%-VF-)D%-4&-9(%-VU-8@"&!2)1)%8-4&-9(%-73-AB-P949!1-3@2"%7-4&-8%6@%3#%8-)3-9()8-$!9!"!8%F-UVEcY`QG--

@"&!2)1C- 5%#)&)#)9C-4&-9(%-8@"&!2)1CH8-9%251!9%- @"&!2)1C- 8)D%- 897@#9@7%- C;G!
! C-/VWXS!G;!
R4$! Y! ;9C-/VWXS!G;! C,8! Y! F8!?;9C-/VWXS!G;ZB! F[!
 ! ;9C-/VWXS!G;! 4C!
! F8! \\S!
Y! Q).]VWXS!G;!
C@R!  ! S;^;VWXS!G;! _F!
Y! S;^;VWXS!G;! @`a! Y! bP1TP!09(121()c!;G;! 4_! ! bP1TP!09(121()c!;G;! :G!

! bP1TP!09(121()c!;G;! S\a! YY! b&%d!09(121()c!;G;! ;a!  ! e;&0V*)-!G;!
`Qf!
! M;`VX;g;!G;! ;YS! ! M;`VX;g;!G;U!8GM! 8a!  ! Q).VX;g;!G;! ;C ! ! egP+VWXSA*)-!G;! SM;! Y! SM;`!
8R!

! 8S8! faX! Y! ;5Q!-09+30.+! F`!  ! F8! F8V!/%(!'+(+-31/+'!

! -

!
! ! !"1%-WF-(%-2489-#438%7A%$-548)9)438-)3-9(%-73-AB-!3$-9(%-#477%8543$)3'-#438%7A!9)43-)3-9(%-&@11--&41$-9C5%--$!9!"!8%G-(%- &4114B)3'-548)9)438-!55%!7-94-"%-#(!7!#9%7)89)#-&47-9(%-73-NI1)0%-&!2)1CF-AXUE-QUWTE-SUWUE-TUbTE-SUbaE-TVUXE-VWYE-VXV-!3$-SWUcG--

!5%/.+-20(1%/!1/! !5%/.+-20(1%/!1/! 81<<+-+/9+!1/! ;31/%!091'! C-/GQ!8^! gQg!<%&'!()*+!:!8^! 9%/.+-20(1%/ 0!! 8
! !!! ! ! S!
YY! Y! Y! S !
YY! ! ! @
Y! ! ! ! S

!
YY! Y! Y! 4
Y! ! Y!  ! S
! ! ! Y! 8
!
YY!
YY! Y! 4
! ! ! Y! 0 S
! ! F5 ! O! 0 S!
YY! F5 ! O! g! ! ! ! 8! ! 
! ! X!
YY! ! ! 7 S! ! F5 ! O! 7 G
! ! F5 ! O! S
! ! ! Y! S! Y! ! ! 0! GP+!9%/.+-20(1%/!1/!(P+!gQg!<%&'!()*+!:!'0(070.+!d0.!.67(-09(+'!<-%3!(P+!9%/.+-20(1%/!1/!(P+!C-/GQ!8^=! 7!g%.1(1%/!'%+.!/%(!7+&%/T!(%!(P+!9%-+!-+T1%/.!1/!(P+!dP%&+!<%&'!()*+!:!'0(070.+ !

! -

!
! ! !"1%-XF-9!9)89)#8-#43#%73)3'-9(%-!#9)A%-8)9%-&)3'%757)39-8%!7#(%8G-31C-85%#)&)#)9)%8-B(%7%-!-&)3'%757)39-#4@1$-"%-!88)'3%$-!7%-)3#1@$%$G-

49- !9#(- @"&!2G- 34B3- 2!9#(- 5%#)&)#)9C- @"&!2)1)%8 !- X)3'%757)39- &)3'%7I 8)D% "- &@3#9)43 #- &)3'%7I 57)39 $- 57)39 %-

i
U!iU!,
U!4
U!_
U! C-/VWXS!G;! C;G!
! 
! !
Y
! ,U! bbd$ ?C;G!/637+-1/TB!

C;GU!C,8U! C-/A;9C-/AM6C-/VWXS!G;!
R4$U!F[U! Y! 
! i
U!,
U!4
U!_
U!,! !
Y! 4C!

Q).]VWXS!G;! \\S!
Y! ! ,
U4
UFAM U,!

! !

S;^;VWXS!G;!?/0--%d!
C@R!  !
! :U,
U4
U$ U_
U,! ! 
! .67.(-0(+!.9%*+B!

S;^;VWXS!G;!?7-%0'! ?FCG!i
BU:U,
U4
U! _FU!
C@R!
!
!
! ! .67.(-0(+!.9%*+B! S U_
U,! e;&0V*)-!G;! ;a!  ! ! jU?FCG!8A4
BUM
U,! Y! !

;G;!?&%d!09(121()B! S\a! YY! ! iU?FCG!8A4
BUMAG
U,!
!
 !

@`aU!4_U! ;G;!?P1TP!09(121()B!
!
! jU?FCG!8A4
BU;
U,! YY!
! :GU!;a! egP+!G;! ;C ! ! ! 4U,! ! !

`Qf!8a! ! M;`VX;g;!G;!! 
! i
 U!8A4
U!,Ui!
!
Y! ;YS! ?
B

Q).VX;g;!G;! 8a!  !
! i
 U8A4
U,U?FCG!iB! ! !

SM;`! SM;! Y!
! F
Ui!
!


!

8S8!
8R!

! ! _U,!

!  !

W;;;!-09+30.+! faX! Y! ! 8
UX
!
!
Y! 0V!M67<031&)!/03+.!%

!
! ! !"1%-YF-[(!7!#9%7)D%$-%3DC2%8-!243'-9(%-#1!88--97!38!2)3!8%8-8479%$-"C-8@"897!9%-!3$-7%!#9)43-85%#)&)#)9CG-

T397C- N#9)A)9C- @"&!2)1C- 3)749-A !- 5%#)%8- [422%39- %&G "-

! C-/VWXS!G;! C;G! _,G Y! 8'E9A05G9$QA1855$$ g8^!:8V!
f8! mn! ! C-/VWXS!G;! C;G! gY

! 0A9A$E'B51@E$ g8^!:8V!C;G! mOn! ! C-/VWXS!G;! C;G! gY
! 3')588GE$EG(F585E$$ ! mn! ! C-/VWXS!G;! C;G! gY ! F'B4Q8A)A))GE$'GD1GE$ ! m O

n! '))4'DA9Q)1E$ ! C-/VWXS!G;! C;G! gY
! ! m
U!
n! )1D1H5E5'1$$ ;48'9Q0A9A@'E$ ! C-/VWXS!G;! C;G! ;\$,! ! m
n! D15@4'D0F55$ $ ! C-/VWXS!G;! C;G! g !

!
! ! ! Q)A(')F1D5G9$ ! ;9C-/VWXS!G;!
R4$! ;Y_iM ! ! mn! E9139'F5E$$ ! ;9C-/VWXS!G;! F[! _ g:,! ;'9BQ8A(')F1D$616G@5$ g8^!:8V!F[! mn!

! ;9C-/VWXS!G;! C,8! CY! B8@GE$38GF5@AE'$ ! mn! ! SfQ! ! ;9C-/VWXS!G;! ? !1'B! C
! 8GQH1DA9Q)1E$8')F5E$ ! m n! 1/!C,8! !

! F8! 4C! $ RF! G82A8A(GE$FA7A0'55$ g8^!:8V!4C $ (7*!

mYn ! ;9C-/VWXS!G;! 4C! CY

! QDA)A))GE$4AD57AE455$ $ !

mYn  ! ;9C-/VWXS!G;! 4C! _ i4[! B1DABQDG9$B1D@5P$ $ ! Y! M6C-/VWXS!G;! C,8! ;f`Q! @E)41D5)45'$)A85$ g8^!:8V!;8^! m
n!

M6C-/VWXS!G;!o! E1G0A9A@'E$ 
! C,8! CYY! ! mn! ;9C-/VWXS!G;! '1DG35@AE'$$ ! Q)A(')F1D5G9$ ! Q).]VWXS!G;! \\S! gY ! g8^!:8V!\\S! mU!n! FG(1D)G8AE5E$ FD1BFA9Q)1E$ ! Q).]VWXS!G;! \\S! _Y
! ! mU!n! )8'HG8531DGE$ ; ``$! W8'HA(')F1D5G9$ ! Q).]VWXS!G;! ? !1'B! _ 4R\! ! mOYn! 8GF1E)1@E$ 1/!C,8! g8^!:8V!
C@R! ;&.%!P0.! Q;:^V!WXS! ! S;^;VWXS!G;!!
C@R! gY
! GE$E)DA2'$ m
U!n! 0/'!e;&0VWXS! 09(121()! '))4'DA9Q)1E$ ! S;^;VWXS!G;!
C@R! g
 ! ! mn! )1D1H5E5'1$ 8%+.!/%(!099+*(! ! S;^;VWXS!G;!
C@R! ;@Y\! ')4'@)1'$78GQH1D5$ mn! e;&0! 86*&190(1%/!%

!
! ! ;9C-/VWXS!G;! 09(121()! g8^!:8V!
M$$!?T07GB! ! S;^;VWXS!G;! _F! g! @V$)A85$ 0&.%!P0.!;9C-/VWXS! mU!n! G;!09(121()! Q)A(')F1D5G9$  ! F8! _F! ;Y_j\Y! g8^!:8V!CXM! (7*! E9139'F5E$ Q)A(')F1D5G9$ Y! F8! _F! ;Y__Y! g8^!:8V!_F!! (7*! E9139'F5E ! Q)A(')F1D5G9$ 
! F8! _F! ^
`:_ ! g8^!:8V!$$5! (7*! '(E)1EEGE$ Q)A(')F1D5G9$ ! F8! _F! ^@FY! g8^!:8V!,G! (7*! 9'D5@G9$ ! S;^;VWXS!G;! _F! ;
, ! BDF4DA(')F1D$'GD1E)1@E$ g8^!:8V!;G_! m n! ;&.%!P0.!C-/VWXS! G;U!;9C-/VWXS!G;U! S;^;VWXS!G;! E1G0A9A@'E$ ! _F! _ :`! gaGVWXS!G;!0/'! mYU!
n! ! '1DG35@AE'$ Q).VWXS!G;!09(121()! ! ;&.%!P0.!O031/%O ! S;^;VWXS!G;! _F! _,^ ! 1G0A9A@'E$BGF50'$ 20&+-0(+VWXS!G;! mU!n! ! ! S;^;VWXS!G;! _F! _;G! E1G0A9A@'E$EQD5@3'1$ ?T07GOB! mn! ?T07GO
B!<6/9(1%/! ! F8! _F! __; ! E1G0A9A@'E$EQD5@3'1$ mn! 6/J/%d/! ?T07GOB!<6/9(1%/! ! F8! _F! _4! E1G0A9A@'E$EQD5@3'1$ mn! 6/J/%d/!  ! S;^;VWXS!G;! _F! g ! 3')588GE$EG(F585E ! ! mn! 45RA(5G9$ Y! S;^;VWXS!G;! _F! _ ;S8! ! mn! 813G95@AE'DG9$ ;&.%!P0.! S;^;VWXS!G;! 
! _F! _YXX! ;GBD5'H50GE$@1)'FAD$ P)*%G06VWXS!G;! mn! ! 09(121()=! ;;g[! gaGVWXS!G;! g8^!:8V!aC[! ! ?  !1'B! g! @E)41D5)45'$)A85$ mU!O Yn! S;^;VWXS!G;! ! 1/!
R4$! E1G0A9A@'E$ ! gaGV*)-!G;! @`a! _ :\! ! m
U!
n! '1DG35@AE'$ :/2%&2+'!1/!gaG!0/'! ! F8! @`a! _5\! E1G0A9A@'E$BGF50'$ mn! S;^;!90(07%&1.3! ;YQ,4! ! F8V;9C9!G;!?*1T4B! ? ! !1'B! _j! 1DD'F5'$ .*= $ g8^!:8V!gg`! m U! n! 1/!
R4$! 8;^;VWXS!G;! #5D35(')588GE$ ! :/!
R4$! _g,! ! m n! ?8;^;V*)-B! B'@FAF41@F5)GE$ $M,M! ! 8;^;VF8!G;! ? !1'B! CY! 0'8A9A@'E$18A@3'F'$ ! m U! n! 1/!
R4$! $M,M! ;4DA9A4'8A(')F1D$ ! 8;^;VF8!G;! ?
!1'B! _ f4a! ! m n! E'81P531@E$ 1/!
R4$! _X ! 8;^;VWXS!G;! E1G0A9A@'E$  ! ? !1'B! _ :
!! ! m n! ?8;^;V*)-B! '1DG35@AE'$ 1/!
R4$! _
@`,! Y! 8;^;VF8!G;! ?
!1'B! _ X48! 3')588GE$4'8A0GD'@E$ ! m n! 1/!
R4$! _
@`,! 8;^;VWXS!G;! 'D5@A)A))GE$ 
! ? !1'B! CYYY! ! m U!
YYn! ! 4'8AB458GE$ 1/!
R4$!

!
! ! ^YR5`! B)5@1FA(')F1D$ ! 8;^;VWXS!G;! ?  !1'B! g! ! m
Y
n! ('G9'@@55$ 1/!_F! 8;^;VWXS!G;! 8Y^[,! 0'19AB458GE$ ! S;^;VWXS!G;! ?
!1'B! g 
! ! m
Yn! 5@28G1@R'1$$ ! 1/!_F! $ 8MY! BADAE'D)5@'$B'EF1GD55$ ! 8;^;VF8!G;! ? !1'B! _ ;g! ! m
Yn! ?3')588GE$B'EF1GD55 B$ 1/!_F! ^^Gf! 1F4Q8A95)DA(5G9$ ! 8;^;VF8!G;! ?Y !1'B! _\_\! ! m
Yn! '8)'85B458G9$ 1/!4C! ! F8! ! _ XQ5! #5(D5A$)4A81D'1$$ g-+'19(+'!8;^;!G;! m
Yn! #5(D5A$ ! F8! ! _Ff! g-+'19(+'!8;^;!G;! m
Yn! B'D'4'19A8QF5)GE$$ FD1BFA9Q)1E$ ! F8! ! _ ,j
! g-+'19(+'!8;^;!G;! m
Yn! )A185)A8AD$ )1'@A(')588GE$  ! F8! ! _4Ma! g-+'19(+'!8;^;!G;! m
Yn! 541Q1@E5E$ 3AD01F188'$ Y! F8! ! _j@:! g-+'19(+'!8;^;!G;! m
Yn! (DA@)45E1BF5)'$
! F8! ! _ij! A)'D05'$2'D)5@5)'$ g-+'19(+'!8;^;!G;! m
Yn! ! F8! ! _8i$! #5(D5A$25E)41D5$ g-+'19(+'!8;^;!G;! m
Yn! FD1BFA9Q)1E$ ! F8! ! _ Q! g-+'19(+'!8;^;!G;! m
Yn! 'H1D95F585E$ ! F8! ! _jQ! 3')588GE$)8'GE55$ g-+'19(+'!8;^;!G;! m
Yn! ! F8! ! __ai ! FD1BFA9Q)1E$'@G8'FGE$ g-+'19(+'!8;^;!G;! m
Yn! ! F8! ! _` X! $A85@188'$EG))5@A31@1E$ g-+'19(+'!8;^;!G;! m
Yn! ;SFG ! M+l=!:8!!! ! ;G;!P1TP!09(121()! ? !1'B! 1/!*0(+/(! BDF4DA(')F1D$)5FD1GE$ ! m
Yn! 1/!S\a! jCYYY!  !*01-d1.+! ;SFG ! ,+<=! m
YnV!$1T6-+! 1'+/(1()!! ! ;G;!P1TP!09(121()! ? !1'B! 3')588GE$913'F1D5G9$$ m
YU!
Yn! ! d1(P! BDF4DA(')F1D$ 1/!S\a! )5FD1GE !! ;4DA9A(')F1D5G9$ ! ;G;!P1TP!09(121()! @`a! _FjS! g8^!:8V!G! m
Yn! H5A8')1G9$
YY! ;G;!P1TP!09(121()! @`a! _Q`a
! 585)5(')F1D$BA91DAQ5$ g8^!:8V!@`a! m
Y n! F+1(P+-!*-%'69(! )4DA(')FDG9$
Y
! ;G;!P1TP!09(121()!! S\a! ;jR5! /%-!.67.(-0(+! m

Yn! '@F4DAB5$ 1/P171(1%/! 'D')A))GE$
Y! ;G;!P1TP!09(121()! 4_! ;
^ ! g8^!:8V!S,[! m


n! 01@5FD525)'@E$
Y! ;G;!P1TP!09(121()! 4_! $[^a ! #5(D5A$28GH5'85E$ g8^!:8V!4_! m

n! 4A0A(')F1D$
Y! ;G;!P1TP!09(121()! 4_! $`[ ! ! m

n! EB4'1DA501E$ 4A0A(')F1D$
Y! ;G;!P1TP!09(121()! :G! _:j4 ! g8^!:8V!:G! m
Y n! EB4'1DA501E$ ;Y^
:! M+l=!:8!!1/! E1G0A9A@'E$
Y! ;G;!P1TP!09(121()! ? !1'B! *0(+/(! ! m

U!

n! 28GAD1E)1@E$ 1/!F`! jCYY
 Y! B)5@1FA(')F1D$ e;&0V*)-!G;!09(121()!
Y! ;G;!P1TP!09(121()! ;a! ^:5 ! m

n! ('G9'@@55$$ /%(!(+.(+'=! B)1FA(')F1D$ e;&0V*)-!G;!09(121()!
Y! ;G;!P1TP!09(121()! ;a! 5\4 ! m

n! B'EF1GD5'@GE !$ /%(!(+.(+'=! e;&0V*)-!G;! B8)'8531@1E$ ;&.%!P1TP&)!09(12+!
Y ! ;a! _jjX! m

n! ! 01@5FD525)'@E$ %/!031/+.! e;&0V*)-!G;! ;&.%!P1TP&)!09(12+!

Y! ;a! _ ;_ ! ;'G8A(')F1D$)D1E)1@FGE$ m

n! ! %/!031/+.!


! e;&0V*)-!G;! ;a! _ :YY! E1G0A9A@'E$ g8^!:8V!^ ^! m

n!

!
! ! '1DG35@AE'$

! e;&0V*)-!G;! ;a! g ! E1G0A9A@'E$BGF50'$ g8^!:8V!;a! m
Yn! e;&0V*)-!G;Z! :/2%&2+'!1/!S;^;!

! F8! ;a! __ ! E1G0A9A@'E$BGF50'$ mn! 0/'!gaG! 90(07%&1.3=! 1EAD45RA(5G9$8AF5$ C/&)!-+&0(12+!

! e;&0V*)-!G;! ;a! _ X! `;$$YY !p! m

n! 09(121(1+.!T12+/=! 3&&
 $ ^-%0'!.*+9(-63=! ,+&0(12+&)!&%d! 1EAD45RA(5G9$8AF5$ 09(121()=!;99+*(.!

! e;&0V*)-!G;! F`! _ F\ ! BWWdadapp$X$ m

U!
n! 031/+.!0/'!egP+! 988abah$ +l60&&)!T%%'!0.! e;&0=! 1EAD45RA(5G9$8AF5 ! C/&)!-+&0(12+!

! e;&0V*)-!G;! S\a! _ ; ! `;$$YY !p! m

n! 09(121(1+.!T12+/=! 3&-
Y
$ 1EAD45RA(5G9$8AF5 !

! ;G;!&%d!09(121()! S\a! _ ^! `;$$YY !p! g8^!:8V!S\a! m
Y U!

n! 3&&
 $

! ;G;!&%d!09(121()! S\a! _
S8! 1G31D5'$ .*=$ g8^!:8V!$5,! m
Y n! ,+&0(12+&)!7-%0'! .*+9(-63=! $i^! 1EAD45RA(5G9$8AF5 ! ,+&0(12+&)!&%d!

! ;G;!&%d!09(121()! ? !1'B! _ ;:
! `;$$YY !p! m

U!
n! 09(121()=!;&.%!P0.! 1/!@`a! 3&- Y $ 9%3*0-07&+!09(121()! <%-!e;&0!0/'!egP+=! 1EAD45RA(5G9$8AF5$ C/&)!-+&0(12+!
Y! ;G;! :G! _ $_! `;$$YY !p! m

n! 09(121(1+.!T12+/=! 3&& $ 1EAD45RA(5G9$8AF5 ! C/&)!-+&0(12+!

! ;G;! :G! _ _! `;$$YY !p! m

n! 09(121(1+.!T12+/=! 3&-  $ 1EAD45RA(5G9$8AF5 ! C/&)!-+&0(12+!
! ;G;! @`a! _ ;:! `;$$YY !p! m

n! 09(121(1+.!T12+/=! 3&-  $ 1EAD45RA(5G9$8AF5$ ^-%0'!.*+9(-63=!
! ;G;! @`a! _ Q! `;$$YY !p! C/&)!-+&0(12+! m

n! 3&&
Y $ 09(121(1+.!T12+/=! 1EAD45RA(5G9$8AF5$ ^-%0'!.*+9(-63=!
! ;G;! @`a! _ :! `;$$YY !p! C/&)!-+&0(12+! m

n! 3&&Y $ 09(121(1+.!T12+/! R0/1&&)&031/+V*)-!
! 4_! C
! ;'BE5)G9$)45@1@E1$ ! m
n! G;! R0/1&&)&031/+V*)-!
! S\a! 8,^! ;'BE5)G9$2DGF1E)1@E$ ! m
n! G;!
! S;^;V*)-!G;! @`a! _ 54! BD'(50ABE5E$F4'85'@'$ ! m
n!
! S;^;V*)-!G;! @`a! S\g`! 105)'3A$FDG@)'FG8'$ ! m
n! E1G0A9A@'E$ e;&0V*)-!G;!09(121()!
 ! S;^;V*)-!G;! ;a! _ @Gg
! m
n! '1DG35@AE'$ /%(!(+.(+'=! S;^;V*)-!G;!
Y! S\a! _g! A8'@G9$8Q)AB1DE5)G9$ ! m
n! !

! S;^;V*)-!G;!! @`a! _g! A8'@G9$8Q)AB1DE5)G9$ ! m
n!
! S;^;V*)-!G;!! S\a! _g! A8'@G9$8Q)AB1DE5)G9$ ! m
n!
! S;^;V*)-!G;! @`a! _
M@! DQR'$E'F5H'$ ! m
n! ^ @8!
! S;^;V*)-!G;! ?! !1'B! \ [Sg! '8GE$P$0A91EF5)'$ ! m
 n! 1/!@`a!
! S;^;V*)-!G;! ^ @8! \ [Sf! '8GE$P$0A91EF5)'$ ! m
 n!

! Y! ! ?! !1'B! 1/!@`a!
! S;^;V*)-!G;! @`a! _[F

! DQR'$E'F5H'$ ! m
n!
! G06V*)-!G;! F`! _ ;g`! 358AB458'$I'0EIADF45'$ ! m
Yn!
! G06V*)-!G;! F`! _QR`! G131D5'$BA91DAQ5$ ! m

n!
 ! G06V*)-!G;! @`a! _\4
! 4A0A)A))GE$AB')GE$ ! m
n! @)*%G06V*)-!G;! 'D')A))GE$
Y! F`! ;
^ f! ! m
n! G06V*)-!G;! 01@5FD525)'@E$

! egP+VWXSA*)-!G;! ;C ! @j,Y! #'D5AHAD'P$B'D'0APGE$ g8^!:8V!;C ! m
U!
n!
! egP+VWXSA*)-!G;! ;C ! ;4i$! 1EAD45RA(5G9$ .* V$"$ g8^!:8V!iXi! m
U!
n! ;&.%!09(12+!%/!
! egP+VWXSA*)-!G;! ;C ! ^
Si
! 3GD74A801D5'$3D'95@5E$ m
n! 031/+.! ;&.%!P1TP&)!09(12+! e;&0V*)-!G;!
! ;C ! _
8@! A8'DA9A@'E$EBV$ %/!egP+!0/'! m
 n! ! 031/+.! 8$
R! ;'@050'FGE$
! 5%;e;;!G;! ? !1'B! ^YR@! )8A')'9A@'E$ ! m
Yn! 1/!SM;! ')50'95@AHAD'@E$ ^RQ8
!
! gPS&)VWXS!G;! ?! !1'B! _Ri ! E1G0A9A@'E$EFGFR1D5$ g8^!:8V!5i! m

U!
n! 1/!;C ! ^RQ8
! S+/^0/J!
! gPS&)VWXS!G;! ?! !1'B! E1G0A9A@'E$BGF50'$ ! m
U!
n! ;[

! 1/!;C !
! M;`VX;g;!G;!
`Qf! g
 ! @E)41D5)45'$)A85$ g8^!:8V!
`Qf! m
O
n! Q)A(')F1D5G9$
 ! M;`VX;g;!G;!
`Qf! g j_Y! g8^!:8V!G$G! m
O
n! FG(1D)G8AE5E$ '))4'DA9Q)1E$
Y! M;`VX;g;!G;!
`Qf! gY! ! m
n! )1D1H5E5'1$

! M;`VX;g;!G;!
`Qf! g! 1DD'F5'$9'D)1E)1@E$ ! m
n!
! F8VX;g;!G;!
`Qf! ^;;Y
! 3D1H5(')F1D5G9$28'HG9$ ! m
n! R+-)!.131&0-!(%! 3V$ ;ADQ@1(')F1D5G9$ 28'HG9 !+/k)3+!
! F8!
`Qf! g ! m
n! 38GF'95)G9$ ?  !1'+/(1()BU!/%(! (+.(+'!<%-!09(121()! C/&)!(+.(+'!1/! a/9P0-09(+-1k+'! dP%&+!71%(1/!
! F8VX;g;!G;!
`Qf! _ ;gM ! m
n! 709(+-163!*5%.;M
! 71%.)/(P+(19! *0(Pd0)! a/9P0-09(+-1k+'!
! F8VX;g;!G;!
`Qf! _ ;g,! ! m
n! 709(+-163!*5%.$M
! a/9P0-09(+-1k+'!
! F8VX;g;!G;!
`Qf! _ ;gg! ! m
n! 709(+-163!*5%.@4! 85S\! a/9P0-09(+-1k+'!
! F8VX;g;!G;! ? !1'B! _ ;g_! ! m
n! 709(+-163!*5%.@4
! 1/!
`Qf! '@FA1'$'338A91D'@E$
! F8VX;g;!G;!
`Qf! g! ! m
n! ?@DI5@5'$41D(5)A8' B$
 ! Q).VX;g;!G;! 8a! g! 3')588GE$EG(F585E$ g8^!:8V!8a! m
 U!
 n! QE5@5(')588GE$ !
Y! M;`VX;g;!G;! 8a! gY! EB4'1D5)GE$ ?3')588GE$ m
Yn! ! EB4'1D5)GE B$ g8^!:8V!;YS! $6.1%/!*-%(+1/!%

! 
! ! $Sf

!
! ;&0VT&)%D!G;!! ?! !1'B! _! 'FFGE$@ADH135)GE$ ! mU!
O
n! 1/!F`! @g$^!
! ;&0VT&)%D!G;!! ?! !1'B! _a4S! GE$9GE)G8GE$ ! m
n! 1/!F`!
! ;&0VT&)%D!G;!! F`! $
`QS! 3AE$F'GDGE$ ! m
n!
! ;&0VT&)%D!G;!! F`! $
F_\
! \'88GE$3'88GE$ ! m
 n! ;&0VT&)%D!G;!!o! $:Mi!
! _F! BD'(50ABE5E$F4'85'@'$ ! m
Yn! S&6VT&)%D!G;! ! ^ Mf ! ;&0VT&)%D!G;!!o!
! ?! !1'B! _ M,! BD'(50ABE5E$F4'85'@'$ ! m
Yn! S&6VT&)%D!G;! 1/!_F! ^ @$a! ;&0VT&)%D!G;!!o!
 ! ?! !1'B! _ Y`! BD'(50ABE5E$F4'85'@'$ ! m
Yn! S&6VT&)%D!G;! 1/!faX! Q@1)4A)A))GE$ .*=!g55!
Y! SM;`! SM;! gY! g8^!:8V!SM;! m

O
n! Y
! Q@1)4A)A))GE$

! SM;`! SM;! _
_\! g8^!:8V!@Ci! m
n! 18A@3'FGE$ !41D9AEQ@1)4A)A))GE$
! SM;`! SM;! _8QX! g8^!:8V!5$^! m
n! 18A@3'FGE$
! SM;`! SM;! g ! BD'(50ABE5E$F4'85'@'$ ! m
 n!
! SM;`! SM;! g
 ! 0AD01G9$HG83'D1$ ! m
Yn!
! SM;`! SM;! gY  ! 3')588GE$EG(F585E$ g8^!:8V!^M! m

U!
n!
! SM;`! SM;! _ i : ! B1DABQDG9$B1D@5P$ g8^!:8V!4g\! (7*!
! SM;`! SM;! Sj :! 105)'3A$E'F5H'$ ! m
n!
! SM;`! SM;! ^
[8! @E)41D5)45'$)A85$ ! m
U!
U!
n!
 ! SM;`! SM;! 4
j! '89A@188'$1@F1D5)'$ ! m
n!
Y! SM;`! SM;! g
! \8Q)5@1$9'P$ ! m
n!

! SM;`! SM;! _j^! 3D'EE5)'$@'BGE$ ! m
n! ;48'9Q0A9A@'E$
! SM;`! SM;! _ ! ! m
 O
n! D15@4'D0F55$ DAB5A@5(')F1D5G9$
! SM;`! SM;! _Y! ! m
n! 2D1G01@D15)455$ _ @!
! SM;`! ? !1'B! _ Ya! G82A8A(GE$EA82'F'D5)GE$ ! m
U!
n! 1/!SM;! %'@F4A9A@'E$
! SM;`! SM;! _Y
! ! m
n! )'9B1EFD5E$
! 8S8!
8R! g
 ! 3GD74A801D5'$)1B')5'$ g8^!:8V!
8R! m
n! Q)AEB4'1D188'$
! 8S8!
8R! $ [5Q
! ! m
n! 3D'95@5)A8'$
! W;;;!-09+30.+! faX! _`

! B)4DA9A(')F1D$A('1$ g8^!:8V!faX! m
n! )4DA(')FDG9$
 ! W;;;!-09+30.+! faX! _YX! ! mYYn! '@F4DAB5$ M+l=!:8! !1/! BDF4DA(')F1D$
Y! W;;;!-09+30.+! O! *0(+/(! ! mYYn! @5)AF5'@'1$ jCY
Y
!

! :&+OO+*13+-0.+! _F! $$j@! ')FA(')588GE$(G)4@1D5$ ! mY
n! S g,; !
! F+031/+!G;! ? !1'B! _@$! 3')588GE$)5D)G8'@E$ ! mYn! 1/!SM;! S iG^! FD1BFA9Q)1E$2D'05'1$
! F+031/+!G;! ? !1'B! _aY!! ? FD1BFA9Q)1E$ ! mYU!Yn! 1/!SM;! DAE1A28'HGE B$ 4_[
!
! F+031/+!G;! ?
! !1'B! _`$Y! FD1BFA9Q)1E$85H50GE$ ! mYn! 1/!SM;!
! qO'+031/%OqO S g,; ! _Q
! FD1BFA9Q)1E$ ! mYn!

! ! ! P)'-%D)/+031/+! ? !1'B! 7'@'9Q)1F5)GE$ G;! 1/!SM;! B45@3ABQP5E$ _Y^f:Y! 9')DA3A8F'(50'$
! P$63%/1.1/V*)-!G;! ? !1'B! 88^! g)-!/%!WXS! mYn! ? B45@3A9A@'E$ 1/!SM;! 9')DA3A8F'(50GE B$ _Y^f:Y!
! P$63%/1.1/V*)-!G;! ?
!1'B! 44Y_! 709(+-163!;G55!! *)-! mYn! 1/!SM;! 8;G[! M+l=!:8!!1/! $63%/1.1/!^
V!WXS!
! ? !1'B! *0(+/(! 6/J/%d/! ! mYn! G;! 1/!SM;! jCYYY! O S
QQ! *P%.*P%+(P0/%&03
! ? !1'B! _G^S! 0A9A$E'B51@E$ ! mYU!Y n! 1/+!*P%.*P%&)0.+! 1/!faX! ?;SG[Q
B! O*P%.*P%P)'-%D)O :X5! QO&).1/+! YY! ? !1'B! _:af! 0A9A$E'B51@E$ ! mY n! *P%.*P%&)0.+! 1/!faX! ?;SG[QB! ^:  ! `)9%.67(1&1/! Y
! ?! !1'B! _ , \
! 3')588GE$EG(F585E$ ! m
Yn! .)/(P0.+!.676/1(!;! 1/!;C ! 0!;&&!.+l6+/9+.!(P0(!0-+!/%(!*-+.+/(!1/!(P+!.+l6+/9+!'0(070.+.!0/'!(P+-+<%-+!/++'+'!(%!7+!+D(-09(+'!<-%3!(P+! *67&190(1%/.!0-+!T12+/!0.!$;MG;!031/%!091'!.+l6+/9+!1/!.+9(1%/!!7+&%d! 7!(7*V!(%!7+!*67&1.P+'!

Y %6@%3#%8-349-)3#1@$%$-)3-9(%-8%6@%3#%-$!9!"!8%8- GP+!.+l6+/9+.!(P0(!P02+!7++/!+D(-09(+'!<-%3!*0(+/(.!%-!*67&190(1%/.!0-+!&1.(+'!7+&%d=!GP+)!31TP(!9%/(01/!+--%-.U! (P+-+<%-+!*&+0.+!9P+9J!(P+!%-1T1/0&!*67&190(1%/.=! ! c# #!I1201!$&A2C2ADPH#9A196%$&A(9!&2A9(B@H"USSYSYVVVY!(8G!@!U! B !"H!H"@# K H#!H HB@#@B @KK#M!B!##H#@! #BK!"@ ## @####H@B@H@"B!K! BH#!H ### !! HN@#NB" BB## ! ! #B!BHHK#B##!#K@BM@HBH!# H#BH!##! H!! B#B##H#K HB!"@H! BKB B"KB#N#B!@ BB#B#!!!H##K! @N"H! ##BN!###!M#H!HH!H#BH#HHNBK@B#B"!@#!@K ! NB #HB!BB! B#!B@@#@#@##@#HBH"! !! c# #!I1201!$&A2C2ADPHO$&233B@!4(0$A(92B4H()G QTSYRE!K20B9(!U!!  !"H!H"@# K H#!H HB@#@B @KK#M!B!##H#@! #BK!"@ ## @####H@B@H@"B!K! BH#!H ### !! HN@#NB" BB## ! ! #B!BHHK#B##!#K@@B@HBH!# H#BH!##! H!! B#B##H# HB!"@H! BKB B"KB#N#B!@ BB#B#!!H##K! @N"H! ##BN!###!M#H!HHK!H#@BH#HHNBK@B#B"!@!#  ! #!@K NB #HB!BB! B#!B@@#@#@##@#HBH"! ! c# #!I1201!$&A2C2ADPH@(B'6465$@!)3B69(@&(5@H"USS`TVbSXX!(8G!@!Y! #"HN !NN@#NHBB!N @!@BK@MBB"M!#B#BH@#! HH! ###KB ### H!##HHB!#KBBB@#!H ###"HB@!K!#B#NBKBB! #MBB#"#HBKK!HN#K @HH#HM#HH#B !###HB#BB!!#! #"M@#@#@H! BBBKBB"B!@M#BB#!# !!##"H@#BK !#! NB# #BN!M#####@!H#H#!!B@#KH@N#B@!BB#!H!H# HKB#! ! ! ! @###H#M!B! ! K HN!@H!H!@#K!#"! ! c$371$###!9$&(4$@(H#9A196%$&A(9!52&6A2$5$(H"SVTSYYbT!(8G!@!b! H@##@B!#B!HKKH #BB!HBBHK@K#" ##BBNBH #####!@B#! # NH#!B#H@! BBB!#BN#B @@!#BMN#BB!#KHBB#NBBB#B!N!@B! !#@N!#K!#@KN B@H !@! !B ##!!H@BB!!HKBHM@#N##! !@H!!BB BN#KHN@B@! B!BBB##BH!@!## BM##B##!H!MB! @!##HBH@####H BB##N!B@!BBBH! @@H!@B ##!##"HB!!!N!B! !H  !#@H!#@ M#@@##B!@H###B"! ! c)B4652@25!OTF7D9! #HH"USSWSaXYUW!(8G!@!XY!! K HK## BBN#!BMKH#BMK"@!@BH#@#M##BB#!@##!@! # KB@MH#H!###@"# K#B@!MN#!##N B!#@#NBB#"M # B!! HH@!NK#@!##HKN#H!##! KNH BN@H @#KH !@B@@!MBKN! BB#H#BNMK#BB##MB!HB##!KK BKK#B##### @H!##HN#KBH! #B#B!# B! KKH@#K@#KM####HB#KKN N"K!##NH@HH ## H#! K##!##! ! !

` %&%7%3#%8- m
n! g)`CQ! !41$Q$A81)G8'D$\D'B45)E$ QEF19 U!R+-.1%/!
==Y=Yr!M9P-s'1/T+-U!Q55V!! mn! i;M;,;!M(-69(6-+! &B BB$ FDG)FGD1 U!R+-.1%/!
==
r!i;M;,;!^1%.91+/9+.V!! mn! 50-&(%/U!\=!`=r!;/T16%&1U!M=!R=r!M6PU!^=!^=r!X%%1tU!G=!j=r!g+-(+0U!`=r!M1&20U!\=!5=r!4-3%&0+20U!`=!8=r!;&&+/U!\=!4=r! M+&+/T6(U!\=!8=r!X%%U!@=!Q=r!g+(+-.%/U!\=!8=r!g%*U!`=r!X%.09JU!8=!M=r!MP63d0)U!`=!$=r!^1'd+&&U!M=!Q=r!MP0&&%3U!M=!\=r! 20/!;J+/U!M=!4=r!,1+'36&&+-U!M=!^=r!$+&'7&)63U!G=!R=r!5P%U!\=!X=r!_609J+/76.PU!\=r!M+'+T0PU!`=r!MP%0171U!;=r! 56331/T.U!Q=!`=r!$&%-+/.U!Q=r!i0(+.U!\=!,=r!,01/+U!\=!8=r!M1/'+/U!,[email protected]!`=!;=r!56//1/TP03U!8=!;=r!g-+1.+-U! g=!,=r!^+-T30/U!Q=!j=r!R01')0U!;=!^=r!20/!Q1/U!Q=!@=r!\0/.+U!5=!\=r!j0(+-.U!;=!g=r!M31(PU!@=!C=r!jP1(+U!C=!,=r! M0&k7+-TU!M=!Q=r!R+/(+-U!\=!5=r!$-0.+-U!5=!`=r!@%<<30/U!M=!Q=r!S0-'/+-U!`=!\=r!50-6991U!8=!\=U!S+/%3+!.+l6+/9+!0/'! 9%3*0-0(12+!0/0&).1.!%

! ! ! mn! g0T+U!;=!$=r!`1/%9P0U!,=r!`1/%9P0U!M=!5=U!Q121/T!d1(P!P1TP!*6(-+.91/+V!+D*-+..1%/!%

! ! ! mn! g0-JU!8=!@=r!`1-07+&&0U!,=r!^-%/.(+1/U!g=!;=r!g-+.(%/U!S=!`=r!@0-1/TU!`=!;=r!Q13U!5=!X=r!5%&&3+-U!;=r!M9P66-1/JU!,=! 5=U!`6(0(1%/.!1/! TO031/%76()-19!091'!?S;^;B!(-0/.031/0.+!T+/+.!1/!*&0/(.!%-! E1G0A9A@'E$EQD5@3'1 !-+'69+! 709(+-10&!21-6&+/9+=! 8'@F$$ VTUT U! ge U!
OY=! mn! ^+&1(.J)U!^=!,=r!M%/+/.P+1/U!;=!Q=U!S07,U!0!3+37+-!%

!  ! ! m
Yn! ,+.P+(/1J%2U!;=!M=r!XP3+&+/1/0U!R=!F=r!G-%(.+/J%U!i=!;=U!5P0-09(+-1k0(1%/!%

! Y! ! m
n! R0/!506d+/7+-TP+U!C=!,=r!`0JP3%6'%20U!;=r!`9Q+0/U!`=!8=r!5&0-JU!M=!`=r!MP+&*U!^=!\=U!g&0/(!*)-620(+O '+*+/'+/(!T0330O031/%76()-0(+!(-0/.031/0.+V!1'+/(1<190(1%/!%

! 
! ! m
n! X‡9JU!@=r!M0/'30-JU!\=r!S17.%/U!X=r!M9P/+1'+-U!S=r!Q1/'l21.(U!i=U!5-).(0&!.(-69(6-+!%

! ! ! m
n! 50*&1/U!^=r!j0/TU!f=r!M&021+-%U!;=r!G%3&1/.%/U!\=r!8%d.+((U!Q=r!8+&0P0)+U!`=r!M0&030U!;=r!M(6'1+.U!G=!:=!5=!<=!^=!g=! S=Oj=!;=r!jP++&+-U!8=!5=r!Q+1*+-U!\=U!;&0/1/+OT&)%D)&0(+!031/%(-0/.<+-0.+O!3+(07%&1k+.!+/'%T+/%6.! 3+(P)&0-T1/1/+.U!-+T6&0(+.!FCU!0/'!9%/(-%&.!7&%%'!*-+..6-+=! BDF1D5AE)81D$!4DA9($#'E)$35A8$ VTUV U! dc U! O  YY=! m
n! R0-(19%2.J1U!Q=r!X6.P/+-U!\=!g=r!^6-/P03U!^=!$=U!^1%.)/(P+.1.!%

.+310&'+P)'+!031/%36(0.+V!;/!WO'13+-19!21(031/!^ O'+*+/'+/(!+/k)3+!d1(P!0.)33+(-)!1/!.(-69(6-+!0/'! 09(12+!.1(+-+09(121()=! DA)$'F8$B)'0$ )5$" B$ Ucca U! pe U!O
=! m
n! C--1..U!S=!Q=r!g0(+&U!G=!,=r!M%-+/.+/U!\=r!M(+(+<+&'U!\=U!;7.+/9+!%

031/%(-0/.<+-0.+!%

! ! ! m
Yn! ;-%/U!f=!8=r!8%--+.(+1/U!g=!5=r!^&09JP0&&U!\=!,=r!X+&&+P+-U!F=!Q=r!j0&.PU!5=!G=U!5P0-09(+-1k0(1%/!%

!

! !

Artikel VI    

              

!"   #  $ !" % &'()* 

$++, --.-.-/-0-123 45+, -)-678)  )/-0))--0 % , (9(/3/6

(   , Tetrahedron

%: 4 , 25  /-0 %: 4 , ;: /-0   4 , 0;: /-0

$      ,"   !$ #% !* &(                   /-0 ,-)-67 8)  )/-0))--0)

( $4<                  ) :   ' :  =  :    )(  ' =    =   = :'   =       )$       =       : '         =        8   ) ACCEPTED MANUSCRIPT

Graphical Abstract To create your abstract, type over the instructions in the template box below. Fonts or abstract dimensions should not be changed or altered.

Synthesis of (1 R,3 R)-1-amino-3-methylcyclo- Leave this area blank for abstract info. hexane by an enzyme cascade reaction Lilly Skalden, Christin Peters, Lukas Ratz and Uwe T. Bornscheuer Institute of Biochemistry, Biotechnology & Enzyme Catalysis, Greifswald University, Greifswald, Germany

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT 1

Tetrahedron journal homepage: www.elsevier.com

Synthesis of (1 R,3 R)-1-amino-3-methylcyclohexane by an enzyme cascade reaction

Lilly Skalden, Christin Peters, Lukas Ratz and Uwe T. Bornscheuer ∗

Institute of Biochemistry, Biotechnology and Enzyme Catalysis, Greifswald University, Felix-Hausdorff-Strasse 4, 17487 Greifswald (Germany)

ARTICLEINFO ABSTRACT

Article history: Amine transaminases (ATAs) are powerful enzymes for the synthesis of chiral amines. Although Received the request for amines with more than one chiral center is increasing, their synthesis is still Received in revised form challenging. Here we show a casacde reaction combining an enoate reductase (ERED) and an Accepted amine transaminase (ATA-VibFlu), which allows access to optically pure (1 R,3 R)-1-amino-3 - Available online methylcyclohexane. Because all known wildtype EREDs show a ( S)-selectivity for 3 - methylcyclohexanone and the ATA-VibFlu only showed a modest enantioselectivity, different Keywords: variants of EREDs and ATAs were investigated and suitable mutant enzymes were identified. In Amine transaminase whole cell biocatalyses using the ERED YqjM Cys26Asp/Ile69Thr and the ATA- VibFlu Enoate reductase Leu56Ile (1 R,3 R)-1-amino-3-methylcyclohexane was obtained at high optical purity (97 %de). Cascade reaction Protein engineering 2009 Elsevier Ltd. All rights reserved . Small cyclic substrates MANUSCRIPT

ACCEPTED

——— ∗ Uwe. T. Bornscheuer. Tel.: +49-3834-86-4367; fax: +49-3834-86-794367; e-mail: [email protected] ACCEPTED MANUSCRIPT 2 Tetrahedron 1. Introduction %de to 66 %de by the Leu56Val variant. In contrary, the cascade reaction of the OYE and the variant Leu56Ile resulted The interest for chiral amines increased dramatically in the in 70 %de of the (1 S,3 S)-diastereomer. The diastereoselectivity last years. More than 80% of the top 200 current drugs contain for the (1 R,3 S)-compound could be further enhanced by the amino functions 1 showing the importance of chiral amines as 2- addition of 30% DMSO to give 89 %de. A recently published precursors for the pharmaceutical and fine-chemical industry. paper by Monti et al. described the synthesis of (1R,3 S)- and 4 Although several chemical methods to produce chiral amines 5 (1 S,3 S)-1-amino-3-methylcyclohexane in a cascade reaction, have been developed, the production by biocatalytic routes 29 6-9,2,10 too. They used the Codexis ATA Screening kit (Codexis Inc., became more and more into the focus. Compared to USA) and the OYE3 to achieve diastereomerically pure chemical routes biocatalysts mostly work at mild conditions, compounds. Nevertheless also they could only generate two i.e., in aqueous phase at physiological pH, at ambient out of the four diastereomers. temperature and normal pressure. Furthermore, the high selectivity of enzymes can replace expensive chiral metal trans-product cis -product catalysts, which are required to achieve a chiral environment O O NH NH for chemical routes 5; also the eco-efficiency can be 2 2 11,12 ERED ATA increased. One of the most famous example is the or production of Sitagliptin, an antidiabetic compound. In this + NADPH NADP Ala Pyr 1 (R)- 2 (1 R,3 R) (1 S,3 R) case the rhodium-catalyzed asymmetric enamine 3a 3b hydrogenation was replaced by biotransformation using an amine transaminase (ATA), which was optimized by protein removal by (R)-ATA (S)-ATA LDH/GDH engineering for this synthesis. 13,14 Indeed, transaminases Scheme 1 : Synthesis of 1-amino-3-methylcyclohexane in a cascade become highly popular biocatalysts for the production of reaction combining an enoate reductase (ERED) and an amine various chiral amines in the last decade. 2-4,15,16 Next to their transaminase (ATA). The LDH/GDH enzymes are required to shift the ability to produce optically pure amines from a racemic equilibrium of the ATA-catalyzed reaction. mixture through kinetic resolution, they also can catalyze an asymmetric synthesis, which starting from a prochiral For the synthesis of the missing two diastereomers (3a and precursor allows a theoretical yield of 100%. 3b , Scheme 1) an enoate reductase with opposite Transaminases are pyridoxal-5'-phosphate (PLP) dependent enantioselectivity is required. The described wildtype enoate 26,30,31 enzymes and catalyze the transfer of an amino group from an reductases all exhibit ( S)-selectivity for 2. Only a mutant amino donor to an amino acceptor. 17-19 Depending on their of the enoate reductase YqjM from Bacillus subtilis, described substrate scope transaminases can be divided into -, 2- and by Bougioukou et al., led to a switch in the selectivity. 32 The amine transaminases. 20 -transaminases convert substrates with application of this mutant in a cascade reaction was already a carboxylate in α-position to the carbonyl function, whereas shown by Agudo et al. for the synthesis of 3-oxo-cyclohexane for ω-transaminases at least one C-atom is in-between. Amine- carboxylic acid methyl ester. 33 In this contribution, we aimed transaminases substrates can lack completely the carboxylic MANUSCRIPTfor using this enoate reductase to access the required two group and hence ATA are the preferred enzymes to synthesize further diasteromers 3a and 3b . chiral amines. 2. Results and discussion ATA usually show high enantioselectivity 21-23 and this facilitates to synthesize compounds with multiple stereocenters 2.1. Choice of appropriate enoate reductases by combining several enzymes in cascade reactions. Compared 32 From literature data, the two double mutants YqjM to chemical routes, the compatibility of different biocatalysts to Cys26Asp/Ile69Thr and YqjM Cys26Asp/Ala104Trp were each other is easier and they can be adapted to a certain range chosen as prime candidates. In addition, we also introduced of reaction conditions. 12 Furthermore a missing selectivity of these mutations into the xenobiotic reductase A (XenA) from one enzyme for a certain chiral center can be overcome by the 31 Pseudomonas putida ATCC 17453 (corresponding positions: combination of various selective enzymes. 24 Cys25Asp, Ile66Thr and Ala101Trp) as this enzyme belongs as Recently, we reported the synthesis of two 1-amino-3S- YqjM to the group of thermophilic like enoate reductases methylcyclohexane diastereomers by the combination of an although they share only 38.4 % sequence identity. enoate reductase and an amine transaminase. 25 The flavin- Unfortunately, no soluble protein could be obtained. mononucleotide (FMN) containing and NAD(P)H-dependent Nevertheless, the double mutants YqjM Cys26Asp/Ile69Thr enoate reductases (ERED) catalyze the selective reduction of and YqjM Cys26Asp/Ala104Trp could be generated via α,β-unsaturated ketones or aldehydes. 26 Whereas the wildtype QuikChange mutagenesis and expressed in active form as enzyme of the ATA from Vibrio fluvialis (ATA-VibFlu) only reported. 32 Biocatalyses confirmed the ( R)-selectivity described showed a moderate selectivity for the conversion of racemic 3- for both mutants, but as variant YqjM Cys26Asp/Ile69Thr was methylcyclohexanone 2, the enoate reductase Old Yellow better expressed, all further experiments were performed with Enzyme exhibited excellentACCEPTED ( S)-selectivity for 3-methyl- this variant. cyclohex-2-enone. 3DM guided protein engineering of the 2.2. Biocatalyses with the enoate reductase variant ATA-VibFlu led to two improved variants. 3DM is a structure- based database, which connects structure-guided sequence In our earlier work purified EREDs were used, 25 but for 27,28 alignments with various other bioinformatic tools. YqjM it was reported that different tags impair the activity of Sequences, structure information, protein ligands and the enzyme 34 and hence all reactions were now conducted as mutational information from literature are integrated into this whole cell biocatalyses, which has the advantage that cofactor software. The ATA variants Leu56Ile and Leu56Val from recycling is more easily facilitated. Whole cell Vibrio fluvialis were generated guided by 3DM. In the biotransformation performed then using E. coli containing the application of the cascade reaction the diastereomeric purity of YqjM Cys26Asp/Ile69Thr variant gave full conversion of rac - (1 R,3 S)-1-amino-3-methylcyclohexane was enhanced from 14 ACCEPTED MANUSCRIPT 3 3-methylcyclohex-2-enone after 1.5 h (4 mmol L -1 ) or after 4.5 amount of pyruvate to balance its extracellular and h (10 mmol L -1 ). intracellular concentration and thus enables a shift of the equilibrium to the desired asymmetric synthesis of the chiral 2.3. Cascade reaction of the YqjM variant and the amine amines. A similar system was described by Börner et al. for transaminase VibFlu Leu56Ile whole cell biocatalyses with ATAs to shift the reaction to the 37 Thanks to our previous studies with different variants of the product site. It was also observed, that after 60 h biocatalysis ATA-VibFlu using racemic 3-methylcyclohexanone 2, the in the presence of LDH/GDH, substrate 1 was nearly absolute configurations of the diastereomers of 1-amino-3- completely consumed, whereas the intermediate 2 was still methylcyclohexane was already known 25 and hence we could present. The supplementation with LDH/GDH at this time easily determine the diastereomeric ratio between the (1 S,3 R)- point led to > 99% conversion after 89 h. Similar to the first and (1 R,3 R)-diastereomers after the cascade reaction: for all test with purified ATA, stereoselectivity of both enzymes was three variants of ATA-VibFlu (Leu56Ala, Leu56Ile and not altered and the product (1 R,3 R)-1-amino-3- Leu56Val) the (1 R,3 R)-diastereomer was preferred over the methylcyclohexane was again obtained with 97 %de. (1 S,3 R)-diastereomer. As the Leu56Ile variant was the most 2.4. Investigation of further amine transaminases to produce stereoselective ATA in the synthesis of the (1 R,3 R)- (1S,3R)-1-amino-3-methylcyclohexane diastereomer this variant was used in subsequent cascade reaction studies. The initial experiment was performed using E. To identify amine transaminases, which prefer to produce coli whole cells harboring YqjM Cys26Asp/Ile69Thr to which the (1 S,3 R)-diastereomer instead of the (1 R,3 R)-diastereomer, the purified ATA-VibFlu Leu56Ile mutant was added. This both enantiomers of 3-methylcyclohexanone were docked with biocatalysis resulted as expected in the formation of the YASARA 38 into the crystal structure of the amine transaminase (1 R,3 R)-1-amino-3-methylcyclohexane diastereomer and we from Vibrio fluvialis (pdb-code: 4e3q). Additionally to Leu56, were pleased to find that this resulted in excellent optical three further residues (small binding pocket: Phe19, Val153, purity (97 %de). The use of purified enzyme dramatically large binding pocket: Ala228) were identified, which could reduces the economic value of such a reaction and furthermore influence the binding of ( R)-3-methyl-cyclohexanone. These requires addition of LDH/GDH to shift the transaminase three residues were also targeted in other protein engineering reaction towards product synthesis. Hence, we next approaches of this amine transaminase to alter its substrate transformed both plasmids encoding YqjM Cys26Asp/Ile69Thr scope. 39,22 With the help of 3DM suitable mutations on this and ATA-VibFlu Leu56Ile into E.coli BL21 (DE3) and after positions were chosen, which led to fourteen further variants in expression at 30°C could obtain both enzymes in soluble form addition to the previous reported ones: Phe19Tyr/Cys/Val, (Fig. 1). Leu56/Met/Ser, Ala228Ie/Gly/Val/Cys/Ser/Thr and Val153Ala/Ile/Ser. All mutants could be expressed as soluble proteins. The screening against racemic 3- methylcyclohexanone showed unfortunately that none of these mutants produce (1 S,3 R)-1-amino-3-methylcyclohexane in MANUSCRIPTexcess compared to the (1 R,3 R)-diastereomer (Table 1). Table 1 .Composition of the diastereomers of 1-amino-3- methylcyclohexane produced by different variants of the ATA from Vibrio fluvialis using rac -3-methylcyclo- hexanone as substrate. Conver Variant sion (1 R,3 R) (1 S,3 S) (1 S,3 R) (1 R,3 S) [%] WT a 99 40 26 4 30 Phe19Cys b 16 34 50 4 12 Phe19Val b 2 36 38 2 24 Val153Ala b 12 55 29 2 14 Val153Ile b 15 45 44 3 8 Fig.1: SDS-PAGE of cultivation samples of the co-expression of the ERED YqjM Cys26Asp/Ile69Thr and the ATA VibFlu Leu56Ile. Lanes 1 Val153Ser b 3 41 42 7 10 and 4: crude extract, lanes 2 and 5: insoluble protein, lanes 3 and 6: soluble Leu56Met b 8 43 44 5 8 protein. Lines 1-3 show the protein content at the time of induction and b lanes 4-6 show the protein content 5 h after induction. The upper frame Leu56Ser 5 38 30 6 26 shows the ATA, the lower frame the ERED. Ala228Thr b 4 42 39 10 9 Cascade reactions withACCEPTED resting cells containing the ERED as Leu56Ile a 99 45.5 47 0.5 7 well as the ATA were performed in the presence or absence of Leu56Val a 99 39 16 5 40

LDH/GDH enzymes and the cofactor NADH. The comparison a 25 gave that the synthesis of 1-amino-3-methylcyclohexane was Results from previous work with purified amine transaminase. 8-fold higher after 60 h in the presence of the LDH/GDH bThe screening against rac -3-methylcyclohexanone (10 mmol L -1 ) was system (83 % conversion) compared to biocatalysis without it performed in deep-well plates (0.6 mL per well) with crude extract of the (11 % conversion). This indicates that without this system ATA variants for 78 h at 30°C and 750 rpm. pyruvate accumulates and the amine transaminase reaction is Hence, the creation of an ATA variant showing the desired slowed down because of an unfavored equilibrium. Pyruvate is (1 S,3 R)-enantiopreference turned out to be not possible so far. 35,36 excreted by E. coli by an overflow production and perhaps The reason for this could be the orientation of the amino- and the added LDH/GDH system is able to reduce the excreted ACCEPTED MANUSCRIPT 4 Tetrahedron methyl-group at the cyclohexane ring. In the cis -diastereomer each) and distilled water (35.5 )L). The PCR program included both substituents are positioned on one site of the cyclohexane the following temperature steps: hold at 95°C for 5 min, ring, whereas in the trans -configuration, both are placed on afterwards 25 cycles of the following: hold at 95°C for 45 sec, opposite sites. These orientations seems to prevent the 53°C for 45 sec, 72°C hold for 7.5 min. Finally hold at 72°C conversion of ( R)-3-methylcyclohexanone to the (1 S,3 R)- for 10 min. diastereomer, although the synthesis of (1 R,3 S)-1-amino-3- All ATA-VibFlu variants were generated via MegaWhop methylcyclohexane did not suffer from this. Further PCR with specific forward or reverse primer for each variant. investigations on the interplay between both binding pockets as The other corresponding primer was either the T7 forward or well as mutations within each binding pocket of the ATA are T7 reverse primer. hence required to explain and alter the selectivity of this amine transaminase from Vibrio fluvialis . Phe19Cys:5’ - G CTC TAT GGT TGC ACC GAC ATG C - 3’ Phe19Val:5’ - G CTC TAT GGT GTG ACC GAC ATG C - 3’ 3. Conclusion Phe19Tyr:5’ - G CTC TAT GGT TAT ACC GAC ATG C - 3’ Leu56Ser:5’ - GCC AAC TCG GGC AGC TGG AAC ATG G - 3’ In this work we have achieved the highly diastereoselective Leu56Met:5’ - CGC AAC TCG GGC ATG TGG AAC ATG G - 3’ two-step synthesis of (1 R,3 R)-1-amino-3-methylcyclohexane Val153Ile:5’ - GCC TAT CAC GGC ATC ACC GCC GTT TC - 3’ by the combination of the enoate reductase variant YqjM Val153Ser:5’ - GCC TAT CAC GGC AGC ACC GCC GTT TC - 3’ Cys26Asp/Ile69Thr and the amine transaminase variant Val153Ala:5’ - GCC TAT CAC GGC CGT ACC GCC GTT TC - 3’ Leu56Ile from Vibrio fluvialis . Mutagenesis studies for ATA- Ala228Cys:5’ - CGG TGA TGG GCT GCG GCG GCG TG - 3’ VibFlu to produce also the missing (1 S,3 R)-diastereomer Ala228Ile:5’ - GTG ATG GGC ATC GGC GGC GTG - 3’ unfortunately failed pointing out the complex design of Ala228Gys:5’ - CG GTG ATG GGC GGC GGC GGC GTG - 3’ enzyme stereopreference. Ala228Ser:5’ - CG GTG ATG GGC AGC GGC GGC GTG - 3’ Ala228Thr:5’ - CG GTG ATG GGC ACC GGC GGC GTG - 3’ 4. Experimental section Ala228Val:5’ - CG GTG ATG GGC GTG GGC GGC GTG - 3’ 4.1. Materials The plasmid (4 )l) was mixed with Pfu + buffer (5 )L), + All chemicals were purchased from Fluka (Buchs, dNTPs (1.5 )L), Pfu polymerase (0.5 )L), specific primer Switzerland), Sigma (Steinheim, Germany), Merck (1 )L), T7 primer (2 )L) and distilled water (35.5 )L). The (Darmstadt, Germany), VWR (Hannover, Germany), or Carl PCR program included the following temperature steps: hold at Roth (Karlsruhe, Germany) and were used without further 95°C for 5 min, afterwards 30 cycles of the following: hold at purification unless otherwise specified. Polymerases were 95°C for 45 sec, 54.8°C for 45 sec, 72°C hold for 6 min. obtained from New England Biolabs GmbH (NEB, Beverly, Finally hold at 72°C for 10 min. The PCR products (8 )L) were mixed with 10x Pfu + buffer (5 )L), plasmid (2 )L), MA, USA) and primers were ordered from Invitrogen (Life + Technologies GmbH, Darmstadt, Germany). dNTPs (1 )L), Pfu polymerase (1 )L) and distilled water (33 )L). The following program was used: hold 68°C for 4.2. Bacterial strains and plasmids MANUSCRIPT5 min, hold 95°C for 1 min, 10 cycles of the following 3 steps: 95°C for 30 sec, 53°C for 30 sec and 68°C for 7 min. E. coli TOP10 [F’lacIq, Tn10(TetR) mcrA D(mrr-hsdRMS- Afterwards 10 cycles of the following 3 steps were made: 95°C mcrBC) F80 LacZDM15 DlacX74 recA1 araD139 D(ara hold for 30 sec, 55°C hold for 30 sec and 68°C hold for leu)7697 galU galK rpsL (StrR) endA1 nupG] was obtained 11 min. from Invitrogen (Carlsbad, CA, USA). E. coli BL21 (DE3) [fhuA2 [lon] ompT gal (l DE3) [dcm] DhsdS] was purchased 4.4. Cultivation, expression and purification from New England Biolabs (Beverly, MA, USA). The plasmid The cultivation and expression of the EREDs and the ATAs pET24b bearing the gene encoding the ATA from Vibrio 32,31,25 fluvialis (accession no. F2XBU9) was kindly provided by Prof. were performed as described elsewhere. The purification Byung Gee Kim (Seoul National University, South-Korea). of the ATA VibFlu Leu56Ile was performed as described previously. 25 After the co-transformation of the plasmids 4.3. Cloning bearing the genes encoding ERED or ATA (each 1 )L) in chemo-competent E. coli Bl21 (DE3) cells, the cultivation was The enoate reductase variants were generated via performed in TB media containing 0.1mg mL -1 ampicillin and QuikChange PCR with specific primers for each variant: 0.05 mg mL -1 kanamycine. The cells were incubated at 37°C XenA: until an OD 600 of 0.7 was reached. After the induction with -1 Cys25Asp fw:5' - CAT TCC GCC CGA TTG CCA GTA CAT G - 3' IPTG (100 )mol L ) the cells were further incubated at 30°C Cys25Asp rv:5' - CAT GTA CTG GCA ATC GGG CGG AAT G - 3' and 180 rpm for 5 h before they were harvested by Ile66Thr fw:5' - GAA GGG CGC ACC ACC CCT GG - 3' centrifugation for 15 min at 4°C. The cells were used Ile66Thr rv:5' - CCA GGG GTG GTG CGC CCT TC - 3' immediately for biotransformations. Ala101Trp fw:5' - GCA TCCACCEPTED AGA TTT GGC ACG CCG - 3' 4.5. Activity measurements Ala101Trp rv:5' - CGG CGT GCC AAA TCT GGA TGC - 3' YqjM: The activity of the amine transaminases was determined by Cys26Asp fw:5' - CATGTCGCCAATGGATATGTATTCTTCTC - 3' the acetophenone-assay. 40 The activity of the EREDs were Cys26Asp rv:5' - GAGAAGAATACATATCCATTGGCGACATG - 3' determined via biocatalyses and GC analysis. Ile69Thr fw:5' - CCC TCA AGG ACG AAC CAC TGA CCA AGA C - 3' 4.6. Biocatalyses Ile69Thr rv:5' - GTC TTG GTC AGT GGT TCG TCC TTG AGG G - 3' Ala104Trp fw:5' - CGG CAT TCA GCT TTG GCA TGC CGG ACG - 3' The total volume was 0.9 mL. Biocatalysis was performed Ala104Trp rv:5' - CGT CCG GCA TGC CAA AGC TGA ATG CCG - 3' in sodium phosphate buffer (50 mmol L -1 , pH 7.5) with -1 -1 + substrate (10 mmol L ), glucose (3 mmol L ) and whole cells The plasmid (2 )L) was mixed with Pfu buffer (5 )L), -1 dNTPs (1.5 )L), Pfu + polymerase (0.5 )L), primers (2 )L of or crude extract (0.2 g wet weight mL ). The substrate solution ACCEPTED MANUSCRIPT 5 was prepared as a stock solution (1 mol L -1 ) in DMSO. The 11. Bruggink, A.; Schoevaart, R.; Kieboom, T. Org. reaction was incubated at 30°C at 750 rpm. For whole cell Process Res. Dev. 2003, 7, 622-640. biocatalyses containing the ERED and the ATA, L-alanine 12. Ricca, E.; Brucher, B.; Schrittwieser, J. H. Adv. Synth. (200 mmol L -1 ), NADH (1 mmol L -1 ), GDH (0.9 U) and LDH Catal. 2011, 353 , 2239-2262. (0.15 U) was added. Biocatalyses with ATA and rac -2 had a 13. Desai, A. A. Angew. Chem., Int. Ed. 2011, 50 , 1974- total volume of 0.6 mL and were performed in deep well 1976. plates. The biocatalyses contained: the crude extract of the 14. Savile, C. K.; Janey, J. M.; Mundorff, E. C.; Moore, J. C.; Tam, S.; Jarvis, W. R.; Colbeck, J. C.; Krebber, A.; amine transaminases, GDH (0.9 U), LDH (0.15 U), L-alanine Fleitz, F. J.; Brands, J.; Devine, P. N.; Huisman, G. W.; (200 mmol L -1), glucose (1 mmol L -1), substrate - - Hughes, G. J. Science 2010, 329 , 305-309. (10 mmol L 1) and NADH (1.1 mmol L 1) dissolved in sodium 15. Malik, M.; Park, E.-S.; Shin, J.-S. Appl. Microbiol. phosphate buffer (50 mmol L -1, pH 7.5) including PLP - Biotechnol. 2012, 94 , 1163-1171. (0.1 mmol L 1). The substrate solutions were prepared as stock 16. Tufvesson, P.; Lima-Ramos, J.; Jensen, J. S.; Al-Haque, - solution (1 mol L 1) in DMSO. The reactions were incubated N.; Neto, W.; Woodley, J. M. Biotechnol. Bioeng. 2011, for 78 h at 30°C at 750 rpm. Biocatalyses samples (300 )L) 108 , 1479-1493. were mixed with NaOH (150 )L of 1 mol L -1 ) and extracted at 17. Hayashi, H.; Mizuguchi, H.; Kagamiyama, H. first with ethylacetate (300 )L) and then with hexane (150 )L). Biochemistry 1998, 37 , 15076-15085. The organic phases were combined and dried with anhydrous 18. Eliot, A.; Kirsch, J. Annu. Rev. Biochem. 2004, 73 , 383 - sodium sulfate before GC analysis was performed. 415. 19. Jansonius, J. N. Curr. Opin. Struct. Biol. 1998, 8, 759- 4.7. Docking studies 769. 38 20. Höhne, M.; Bornscheuer, U. T. In Enzymes in Organic The docking studies were performed with YASARA using Synthesis; May, O., Gröger, H., Drauz, W.; Ed. Wiley- the structure of the ATA from Vibrio fluvialis (pdb-code 4e3q). VCH: Weinheim, 2012; pp. 779-820. (R)- and ( S)-3-methylcyclohexanone were alternatively docked 21. Cho, B.-K.; Seo, J.-H.; Kang, T.-W.; Kim, B.-G. into the active site. The chosen simulation cell was defined to Biotechnol. Bioeng. 2003, 83 , 226-234. 3 be 18×17×18 F . All residues of the active site and the active 22. Midelfort, K. S.; Kumar, R.; Han, S.; Karmilowicz, M. site loop were included. J.; McConnell, K.; Gehlhaar, D. K.; Mistry, A.; Chang, J. S.; Anderson, M.; Villalobos, A.; Minshull, J.; 4.8. Analytics Govindarajan, S.; Wong, J. W. Protein Eng. Des. Sel. The determination of the conversion, the ratio of the 2013, 26 , 25-33. diastereomers or enantiomers were identified via GC as 23. Cho, B.-K.; Park, H.-Y.; Seo, J.-H.; Kim, J.; Kang, T.-J.; described previously. 25 The derivatisation of the samples was Lee, B.-S.; Kim, B.-G. Biotechnol. Bioeng. 2008, 99 , 275-284. performed with 5 )L trifluoroacetic acid anhydride. 24. Muschiol, J.; Peters, C.; Oberleitner, N.; Mihovilovic, 5. Acknowledgement M. D.; Bornscheuer, U. T.; Rudroff, F. Chem. Commun. MANUSCRIPT2015, 51 , 5798-5811. We thank the European Union (KBBE-2011-5, grant no. 25. Skalden, L.; Peters, C.; Dickerhoff, J.; Nobili, A.; 289350) and the DFG (grant no. Bo1862/6-1) for financial Joosten, H.-J.; Weisz, K.; Höhne, M.; Bornscheuer, U. support. We are grateful to Prof. Byung-Gee Kim (Seoul T. ChemBioChem 2015 , 16, 1041-1045. National University, Korea) for the gene encoding the ATA 26. Toogood, H. S.; Gardiner, J. M.; Scrutton, N. S. from V. fluvialis . ChemCatChem 2010, 2, 892-914. 27. Kuipers, R. K.; Joosten, H.-J.; van Berkel, W. J. H.; 6. References Leferink, N. G. H.; Rooijen, E.; Ittmann, E.; van Zimmeren, F.; Jochens, H.; Bornscheuer, U.; Vriend, G.; 1. http://www.pharmacy-tech-test.com/top-200-drugs.html. Martins dos Santos, V. A. P.; Schaap, P. J. Proteins 05.10.2015. Struct. Funct. Bioinf. 2010, 78 , 2101-2113. 2. Höhne, M.; Bornscheuer, U. T. ChemCatChem 2009, 1, 28. Steffen-Munsberg, F.; Vickers, C.; Kohls, H.; Land, H.; 42-51. Mallin, H.; Nobili, A.; Skalden, L.; van den Bergh, T.; 3. Kohls, H.; Steffen-Munsberg, F.; Höhne, M. Curr. Opin. Joosten, H.-J.; Berglund, P.; Höhne, M.; Bornscheuer, Chem. Biol. 2014, 19 , 180-192. U. T. Biotechnol. Adv. 2015, 33 , 566-604. 4. Kroutil, W.; Fischereder, E.-M.; Fuchs, C. S.; Lechner, 29. Monti, D.; Forchin, M. C.; Crotti, M.; Parmeggiani, F.; H.; Mutti, F. G.; Pressnitz, D.; Rajagopalan, A.; Sattler, Gatti, F. G.; Brenna, E.; Riva, S. ChemCatChem 2015 , J. H.; Simon, R. C.; Siirola, E. Org. Process Res. Dev. n/a-n/a. 2013, 17 , 751-759. 30. Fu, Y.; Castiglione, K.; Weuster-Botz, D. Biotechnol. 5. Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Bioeng. 2013, 110 , 1293-1301. Keßeler, M.; Stürmer, R.; Zelinski, T. Angew. Chem. 31. Peters, C.; Kölzsch, R.; Kadow, M.; Skalden, L.; 2004, 43 , 788-824. Rudroff, F.; Mihovilovic, M. D.; Bornscheuer, U. T. 6. Abrahamson, M. J.; Vázquez-Figueroa, E.; Woodall, N. ChemCatChem 2014, 6, 1021-1027. B.; Moore, J. C.;ACCEPTED Bommarius, A. S. Angew. Chem., Int. 32. Bougioukou, D. J.; Kille, S.; Taglieber, A.; Reetz, M. T. Ed. 2012, 51 , 3969-3972. Adv. Synth. Catal. 2009, 351 , 3287-3305. 7. Alexeeva, M.; Enright, A.; Dawson, M. J.; 33. Agudo, R.; Reetz, M. T. Chem. Commun. 2013, 49 , Mahmoudian, M.; Turner, N. J. Angew. Chem., Int. Ed. 10914-10916. 2002, 41 , 3177-3180. 34. Fitzpatrick, T. B.; Auweter, S.; Kitzing, K.; Clausen, T.; 8. Balkenhohl, F.; Ditrich, K.; Hauer, B.; Ladner, W. J. Amrhein, N.; Macheroux, P. Protein Expression Purif. Prakt. Chem. 1997, 339 , 381-384. 2004, 36 , 280-291. 9. Gotor-Fernández, V.; Busto, E.; Gotor, V. Adv. Synth. 35. Chang, D.-E.; Shin, S.; Rhee, J.-S.; Pan, J.-G. J. Catal. 2006, 348 , 797-812. Bacteriol. 1999, 181 , 6656-6663. 10. Martens, J.; Günther, K.; Schickedanz, M. Arch. Pharm. 36. Yang, Y.-T.; Bennett, G. N.; San, K.-Y. Metab. Eng. 1986, 319 , 461-465. 2001, 3, 115-123. ACCEPTED MANUSCRIPT 6 Tetrahedron 37. Börner, T.; Rehn, G.; Grey, C.; Adlercreutz, P. Org. Process Res. Dev. 2015, 19 , 793-799. 38. Krieger, E.; Koraimann, G.; Vriend, G. Proteins Struct. Funct. Bioinf. 2002, 47 , 393-402. 39. Nobili, A.; Steffen-Munsberg, F.; Kohls, H.; Trentin, I.; Schulzke, C.; Höhne, M.; Bornscheuer, U. T. ChemCatChem 2015, 7, 757-760. 40. Schätzle, S.; Höhne, M.; Redestad, E.; Robins, K.; Bornscheuer, U. Anal. Chem. 2009, 81 , 8244 - 8248.

MANUSCRIPT

ACCEPTED

Eigenständigkeitserklärung

Hiermit erkläre ich, dass diese Arbeit bisher von mir weder an der Mathematisch- Naturwissenschaftlichen Fakultät der Ernst-Moritz-Arndt-Universität Greifswald noch einer anderen wissenschaftlichen Einrichtung zum Zwecke der Promotion eingereicht wurde. Ferner erkläre ich, dass ich diese Arbeit selbstständig verfasst und keine anderen als die darin angegebenen Hilfsmittel und Hilfen benutzt und keine Textabschnitte eines Dritten ohne Kennzeichnung übernommen habe.

Unterschrift der Promovendin

Lebenslauf

Name: Lilly Skalden Geburtsdatum: 18.12.1987 Geburtsort: Stralsund

Promotion seit 10/2012 Promotion bei Prof. Dr. Uwe T. Bornscheuer am Institut für Biochemie, Abteilung Biotechnologie & Enzymkatalyse an der Ernst- Moritz-Arndt-Universität Greifswald

Studium/ Praktika

10/2011 - 8/2012 Diplomarbeit am Institut für Biochemie, Abteilung Biotechnologie & Enzymkatalyse an der Ernst-Moritz-Arndt-Universität Greifswald Betreuer: Prof. Dr. Uwe T. Bornscheuer Thema : Biokatalytische Synthese von funktionalisierten Aminen Abschluss: Diplom Biochemikerin 2/2011 - 3/2011 Betriebspraktikum bei GALAB Laboratories, Hamburg-Geesthacht Betreuer: Dr. Eckert Jantzen 10/2010 - 1/2011 Kurs- und Vertiefungspraktikum im Arbeitskreis Biotechnologie & Enzymkatalyse von Prof. Dr. Uwe T. Bornscheuer an der Ernst-Moritz- Arndt-Universität Greifswald Thema : Charakterisierung von Mutanten einer Baeyer-Villiger- Monooxygenase 10/2007 Immatrikulation an der Ernst-Moritz-Arndt-Universität Greifswald im Studienfach Biochemie

Veröffentlichte Publikationen

1. M. Thomsen*, L. Skalden* , G. J. Palm, M. Höhne, U. T. Bornscheuer and W. Hinrichs: Crystallization and preliminary X-ray diffraction studies of the ( R)- selective amine transaminase from Aspergillus fumigatus . Acta Cryst. 2013 , F69, 1415-1417 2. M. Thomsen*, L. Skalden* , G. J. Palm, M. Höhne, U. T. Bornscheuer and W. Hinrichs: Crystallographic characterization of the ( R)-selective amine transaminase from Aspergillus fumigatus . Acta Cryst. 2014 , D70, 1086-1093 3. C. Peters, R. Kölzsch, M. Kadow, L. Skalden , F. Rudroff, M. D. Mihovilovic and U. T. Bornscheuer: Identification, characterization, and application of three enoate reductases from Pseudomonas putida in in vitro enzyme cacsade reactions. ChemCatChem 2014 , 6, 1021-1027 4. L. Skalden* , M. Thomsen*, M. Höhne, U. T. Bornscheuer and W. Hinrichs: Structural and biochemical characteriztion of the dual substrate recognition of the (R)-selective amine transaminase from Aspergillus fumigatus . FEBS J. 2015 , 282, 407-415 5. L. Skalden , C. Peters, J. Dickerhoff, A. Nobili, H.-J- Joosten, K. Weisz, M. Höhne, U. T. Bornscheuer: Two subtle amino acid changes in a transaminase substantially enhance or invert enantiopreference in cascade syntheses. ChemBioChem 2015 , 16, 1041-1045 6. F. Steffen-Munsberg, C. Vickers, H. Kohls, H. Land, H. Mallin, A. Nobili, L. Skalden , T. van den Bergh, H.-J. Joosten, P. Berglund, M. Höhne, U. T. Bornscheuer: Bioinformatic analysis of a PLP-dependent enzyme superfamily suitable for biocatalytic applications. Biotechnol. Adv. 2015 , 33, 566-604 7. L. Skalden , C. Peters, L. Ratz, U. T. Bornscheuer: Synthesis of (1 R,3 R)-1-amino-3-methylcyclohexane by an enzyme cascade reaction. Tetrahedron 2015 , DOI: 10.1016/j.tet.2015.11.005

* geteilte Erstautorenschaft

Posterpräsentation

1. 04/2014: Poster auf der MECP14 Konferenz in Madrid, Spanien mit dem Titel "Production of enantiopure diastereomers in an enzyme cascade reaction" 2. 08/2014: Poster auf der Biocat Konferenz in Hamburg mit dem Titel " Production of enantiopure diastereomers in an enzyme cascade reaction" 3. 03/2015: Poster auf der Transam 2.0 Konferenz in Greifswald mit dem Titel "Two subtle amino acid changes in a transaminase substantially enhance or invert enantiopreference in cascade syntheses" 4. 09/2015: Poster auf dem ECCE10+ECAB3+EPIC5 Kongress in Nizza mit dem Titel "Two subtle amino acid changes in a transaminase substantially enhance or invert enantiopreference in cascade syntheses"

Lilly Skalden