Dissecting cobamide diversity through structural and functional analyses of the base-activating CobT enzyme of Salmonella enterica

Article abstract of DOI:10.1016/j.bbagen.2013.09.038

Background Cobamide diversity arises from the nature of the nucleotide base. Nicotinate mononucleotide (NaMN):base phosphoribosyltransferases (CobT) synthesize α-linked riboside monophosphates from diverse nucleotide base substrates (e.g., benzimidazoles, purines, phenolics) that are incorporated into cobamides. Methods Structural investigations of two members of the CobT family of enzymes in complex with various substrate bases as well as in vivo and vitro activity analyses of enzyme variants were performed to elucidate the roles of key amino acid residues important for substrate recognition. Results Results of in vitro and in vivo studies of active-site variants of the Salmonella enterica CobT (SeCobT) enzyme suggest that a catalytic base may not be required for catalysis. This idea is supported by the analyses of crystal structures that show that two glutamate residues function primarily to maintain an active conformation of the enzyme. In light of these findings, we propose that proper positioning of the substrates in the active site triggers the attack at the C1 ribose of NaMN. Conclusion Whether or not a catalytic base is needed for function is discussed within the framework of the in vitro analysis of the enzyme activity. Additionally, structure-guided site-directed mutagenesis of SeCobT broadened its substrate specificity to include phenolic bases, revealing likely evolutionary changes needed to increase cobamide diversity, and further supporting the proposed mechanism for the phosphoribosylation of phenolic substrates. General Significance Results of this study uncover key residues in the CobT enzyme that contribute to the diversity of cobamides in nature.

Full text of DOI:10.1016/j.bbagen.2013.09.038

Biochimica et Biophysica Acta 1840 (2014) 464–475
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbagen
Dissecting cobamide diversity through structural and functional analyses
of the base-activating CobT enzyme of Salmonella enterica
Chi Ho Chan a,1,2, Sean A. Newmister b,1,3, Keenan Talyor b, Kathy R. Claas a,
Ivan Rayment b, Jorge C. Escalante-Semerena c,
⁎
a
Department of Bacteriology, University of Wisconsin, Madison, WI, USA
Department of Biochemistry, University of Wisconsin, Madison, WI, USA
b
c
Department of Microbiology, University of Georgia, Athens, GA, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 13 June 2013
Received in revised form 20 August 2013
Accepted 30 September 2013
Available online 10 October 2013
Background: Cobamide diversity arises from the nature of the nucleotide base. Nicotinate mononucleotide
(NaMN):base phosphoribosyltransferases (CobT) synthesize α-linked riboside monophosphates from diverse
nucleotide base substrates (e.g., benzimidazoles, purines, phenolics) that are incorporated into cobamides.
Methods: Structural investigations of two members of the CobT family of enzymes in complex with various
substrate bases as well as in vivo and vitro activity analyses of enzyme variants were performed to elucidate
the roles of key amino acid residues important for substrate recognition.
Results: Results of in vitro and in vivo studies of active-site variants of the Salmonella enterica CobT (SeCobT)
enzyme suggest that a catalytic base may not be required for catalysis. This idea is supported by the analyses
of crystal structures that show that two glutamate residues function primarily to maintain an active conformation
of the enzyme. In light of these findings, we propose that proper positioning of the substrates in the active site
triggers the attack at the C1 ribose of NaMN.
Keywords:
B12 biosynthesis
Lower base activation
Directed enzyme evolution
Cobamide diversity
Conclusion: Whether or not a catalytic base is needed for function is discussed within the framework of the
in vitro analysis of the enzyme activity. Additionally, structure-guided site-directed mutagenesis of SeCobT
broadened its substrate specificity to include phenolic bases, revealing likely evolutionary changes needed
to increase cobamide diversity, and further supporting the proposed mechanism for the phosphoribosylation
of phenolic substrates.
General Significance: Results of this study uncover key residues in the CobT enzyme that contribute to the diversity
of cobamides in nature.
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
CoA mutase, ethanolamine ammonia-lyase and diol dehydratase [4–6].
Cobamides also serve as a transient methyl group carrier in methyl-
A cobamide (Cba) is a cobalt-containing cyclic tetrapyrrole involved
in enzyme catalyzed carbon skeleton rearrangements, methyl-group
transfers and reductive dehalogenation [1,2]. Cobamides are related to
heme, chlorophyll and coenzyme F430 in structure though they play dif-
ferent roles in nature [3]. A diagram of cobamides is shown in Fig. 1.
In cobamides, nitrogen atoms of the tetrapyrrole rings equatorially
coordinate a cobalt ion. In their coenzyme form, cobamides have a 5′-
deoxyadenosyl (Ado) group as the axial Coβ ligand that participates in
radical chemistry reactions such as those catalyzed by methylmalonyl-
transferase reactions [7].
Vitamin B12 (cyanocobalamin, CNCbl) is a cobamide that is an
essential nutrient to humans. In CNCbl, the axial Coα ligand is 5,6-
dimethylbenzimidazole (DMB) and the axial Coβ ligand is a cyano
(CN) group. Cobamides differ from each other by the nature of their
nucleotide base, which ranges from DMB, to purines, to phenolics [8].
The biosynthesis of cobamides is complex and is only performed by
some archaea and bacteria [9]. Adenosylcobamide (AdoCba) biosynthe-
sis has been studied extensively in several bacteria and some archaea
[10]. Given the complexity of the molecule, it is not surprising that a
large number of gene products are required for its assembly. For exam-
ple, in Salmonella enterica, the AdoCba biosynthetic pathway involves
at least 25 gene products. Some microorganisms synthesize a range of
different cobamides. For example, S. enterica synthesizes three, one that
contains DMB as the nucleotide base, and two others that contain adenine
and 2-methyladenine in lieu of DMB [11,12]. When adenine is the nucle-
otide base, this cobamide is known as pseudo-B12. The reasons that drive
⁎
Corresponding author at: Department of Microbiology, University of Georgia, 527
Biological Sciences Building, 120 Cedar Street, Athens, GA 30602, USA. Tel.: +1 706 542
2651; fax: +1 706 542 2815.
E-mail address: jcescala@uga.edu (J.C. Escalante-Semerena).
These authors contributed equally to this work.
Present address: BioTechnology Institute, University of Minnesota, St. Paul, MN 55108.
Present address: Life Sciences Institute, University of Michigan, Ann Arbor, MI 48109.
1
2
3
0304-4165/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.bbagen.2013.09.038

C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
465
Fig. 1. Schematic of the nicotinate mononucleotide (NaMN):base phosphoribosyltransferase reaction (left). The α-riboside monophosphate formed is combined with precursor
adenosylcobinamide-GDP (AdoCbi-GDP) to produce adenosyl-cobamide (right). Phenolic bases do not form an α-coordination bond with the cobalt ion.
the need to synthesize different cobamides are unclear, but such diversity
may reflect physiological needs under different conditions.
open question. Crystal structures of variant proteins in which either
glutamate was changed have been determined to a 1.9 Å resolution
and provided insight into the roles of these residues in the activity
of this enzyme. To gain a greater understanding of the residues involved
in substrate recognition and activity, SeCobT was engineered to synthe-
size α-p-cresolyl-riboside monophosphate using structure-guided anal-
yses of the SoArsAB active site. Structures of this engineered variant
in complex with p-cresol, p-cresol and NaMN at a 2.0 Å resolution illus-
trate some of the structural changes SeCobT needs to undergo to be able
to use phenolic substrates.
In S. enterica, the nicotinate mononucleotide (NaMN):base
phosphoribosyltransferase (SeCobT) enzyme catalyzes the synthesis
of α-riboside monophosphates (α-ribotides) that are subsequently uti-
lized in AdoCba biosynthesis (Fig. 1) [13]. In vitro, SeCobT was shown to
phosphoribosylate a wide range of benzimidazole and purine analogs
that are not produced by S. enterica, showing that this enzyme has
broad substrate specificity [13,14]. Products formed with benzimidazole
and purine analogs feature an α-N-glycosidic bond between N1 of
imidazole and C1 of ribose. In S. enterica, the CobB sirtuin deacetylase
compensates for the lack of CobT during the assembly of the nucleotide
loop in ΔcobT strains [15]. Thus, to eliminate background activity, all
strains used in this work carried null alleles of cobT and cobB.
2. Materials and methods
2.1. Growth media, conditions and analysis
Previous structural studies of SeCobT revealed how the active site
accommodates diverse aromatic compounds, and a general-base cataly-
sis mechanism through a conserved glutamate residue was proposed
[16]. Remarkably, products were captured in the structure when SeCobT
was co-crystallized with aromatic bases and NaMN with little move-
ment in the active site [14,17]. Interestingly, SeCobT cannot catalyze
the phosphoribosylation of phenolic bases.
In contrast, the acetogenic anaerobe Sporomusa ovata [18] synthe-
sizes cobamides with phenolic bases [19]. Incorporation of phenolic
compounds into cobamides is catalyzed by the SeCobT homolog ArsAB
in S. ovata [20]. SoArsAB phosphoribosylates phenol and p-cresol yielding
phenolic α-ribotides that have a unique α-O-glycosidic bond between
the oxygen in phenols and the C1 of ribose (Fig. 1). Like SeCobT, SoArsAB
also incorporates DMB and purine analogs into α-ribotides. Unlike
SeCobT, SoArsAB is a heterodimeric enzyme, where both subunits are
SeCobT orthologs. Differences in the structures of SoArsAB and
SeCobT have been proposed to account for the phenolic activity in
SoArsAB [21].
Cultures were grown at 37 °C in a no-carbon essential (NCE) mini-
mal medium [22] supplemented with glycerol (22 mM) as the carbon
source. MgSO4 (1 mM), trace minerals [23,24], ampicillin (100 μg/mL),
kanamycin (50 μg/mL) and chloramphenicol (12.5 μg/mL) were used
where indicated. For all growth experiments, strains were grown in
triplicate in 96-well microtiter dishes; a sample (2 μL) of overnight
cultures grown in lysogenic broth (LB) [25,26] was used to inocu-
late 198 μL of fresh minimal medium. Growth was analyzed using
a
computer-controlled BioTek ELx808 Ultra microplate reader
(BioTek Instruments, Inc.). Cell density measurements at 650 nm were
acquired every 1800 s for 36 h; plates were shaken for 1745 s between
readings. Data were analyzed using the GraphPad Prism v4 software
package (GraphPad Software).
2.2. Strain and plasmid constructions
All strains used in this study were derivatives of S. enterica sv
Typhimurium strain LT2 or Escherichia coli strain DH5α. All S. enterica
strains are derived from strain TR6583. Strain genotypes are described
in Table 5. Plasmids were introduced by electroporation into S. enterica
and by chemically competent transformation by heat shock into E. coli.
All primers used in this work are listed in Table 5.
The work reported here expands our understanding of the active site
of the CobT family of enzymes. We demonstrate that two glutamate res-
idues (Glu174 and Glu317) in the active site that are important for
activity are also critical to the positioning of the base substrate. Whether
either one of these residues plays a role as a catalytic base remains an

466
C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
Table 1
Data collection and refinement statistics.
complex
317A
317A
174A
DMB
174A
YMM
YMM
α-RP + NA
adenine
p-cresol
p-cresol + NaMN
pdb id
4KQH
P21212
1.54
4KQI
4KQG
P21212
1.54
4KQF
P21212
1.54
4KQK
P21
4KQJ
P21212
1.54
Space group
Wavelength
Source
P21212
0.979
19BM
50–1.4 (1.42–1.4)
989867
59605
8.7 (5.5)
100 (99.8)
78.0 (9.8)
4.6 (17.9)
17.3
0.979
19 BM
50–1.47 (1.5–1.47)
513574
97861
3.2 (2.7)
95.2 (88.9)
26.6 (4.6)
3.1 (16.1)
16.8
Cu Kα
50–1.9 (2.0–1.9)
288139
24438
11.8 (8.4)
99.9 (99.3)
14.5 (6.3)
11.3 (26.5)
16.9
Cu Kα
50–1.9 (2.0–1.9)
209381
24307
8.6 (5.3)
99.9 (99.2)
9.9 (3.8)
15.5 (35.2)
17.0
Cu Kα
50–1.9 (2.0–1.9)
303189
24215
12.5 (9.0)
99.8 (98.8)
16.4 (7.3)
10.0 (23.5)
15.3
Cu Kα
50–1.9 (2.0–1.9)
257886
22398
11.4 (6.1)
99.0 (93.5)
14.2 (5.1)
12.7 (29.6)
18.2
Resolution rangea
Reflections: measured
Reflections: unique
Redundancy
% Complete
Average I/σ
Rmerge (%)b
c
Rcryst
Rfree
22.4
19.5
22.5
22.0
19.5
22.2
Protein atoms
Ligand atoms
Waters
2441
19
260
12.6
––
2451
37
252
17.9
––
2509
20
296
8.4
––
2499
24
285
12.5
––
5077
44
411
14.8
––
2472
34
175
8.9
––
Avg. B (Å2)
Ramachandran (%)
Most favored
Allowed
98.8
0.9
98.3
1.4
97.7
1.7
98.6
1.1
98.2
1.5
97.7
1.7
Disallowed
0.3
––
0.3
––
0.6
––
0.3
––
0.3
––
0.3
––
rms deviations
Bond lengths (Å)
Bond angles (deg)
0.019
1.99
0.019
1.99
0.019
1.97
0.020
1.89
0.013
1.67
0.018
2.07
Rwork refers to the Rfactor for the data utilized in the refinement and Rfree refers to the Rfactor for 5% of the data that were excluded from the refinement.
a
Data in parentheses represent highest resolution shell.
Rmerge = (∑ |I(hkl) − I| × 100. ) ∕ ∑ |I(hkl)|, where the average intensity I is taken over all symmetry equivalent measurements and I(hkl) is the measured intensity for a given
b
observation.
c
Rfactor = ∑ | Fobs − Fcalc | × 100. ∕ ∑ | Fobs |.
2.2.1. Plasmid pCOBT15
2.2.5. Plasmid pCOBT184
This plasmid encoded SeCobTE317A, and was generated with two sets
of overlapping primers (cobT13 and cobT14) coding for the amino acid
change using plasmid pCOBT10 DNA as template [27]. Briefly, primers
cobT13 and cobT14 encoding the E317A changes were used to amplify
the cobT gene on pCOBT10. The amplification product was transformed
in to E. coli DH5α, and its sequenced verified.
This plasmid encoded the SeCobTS80Y,Q88M,L175M variant, and was
constructed following the QuikChange™ site-directed mutagenesis
protocol. Each substitution was introduced one at a time into plas-
mid pCOBT141. SoArsAB variants were constructed following the
QuikChange™ site-directed mutagenesis protocol (Stratagene) using
plasmid pARSAB4 DNA as template [20]. Primers were purchased from
IDT http://www.idtdna.com/ and designed using the PrimerX website
http://www.bioinformatics.org/primerx/.
2.2.2. Plasmid pCOBT19
This plasmid was a derivative of plasmid pT7-5 [28], and encoded
variant SeCobTE317A. The cobT allele encoding SeCobTE317A was sub-
cloned from pCOBT15 using BamHI and HindIII restriction enzymes.
2.3. In vivo assessments of SeCobT and SoArsAB functions by
cobalamin-dependent methionine biosynthesis
S. enterica strains JE2607 (cobB cobT recA) and JE2501(cobB cobT)
were used to assess the function of SeCobT and SoArsAB variants,
respectively, under conditions that demanded Cbl-dependent methio-
nine synthesis. All strains lacked the cobB gene to ensure that base
incorporation was due to plasmid-encoded SeCobT variants. Generation
times were calculated using the equation g = n ∕ t, where g is the gen-
eration time, n is the number of generations, and t is the time it took to
reach the final cell population. To calculate the number of generations
(n), we used the equation Nt = N02n, where Nt and N0 are the final
and initial optical densities (OD) of the culture, respectively.
2.2.3. Plasmids pCOBT42 and pCOBT57
These plasmids encoded SeCobTE174A and CobTE174A,E317A variants,
respectively, and were constructed following the QuikChange™ site-
directed mutagenesis protocol (Stratagene) using plasmids pJO27 and
pCOBT19 DNA as templates, respectively.
2.2.4. Plasmid pCOBT141
This plasmid encoded SeCobTWT fused to an N-terminal hexahistidine
(His6) tag. Plasmid pCOBT141 was constructed using polymerase incom-
plete primer extension (PIPE)-based cloning [29].
2.4. Purification of SeCobT variants
Table 2
SeCobT variant proteins encoded in pCOBT19, pCOBT42 and
pCOBT57 were overproduced in strain JE2017 (cobA/pGP1-2 T7 rpo+
(T7 RNAP) kan+) and purified as described [13]. The SeCobT variant
encoded in pCOBT184 was overproduced in strain JE13607, an E. coli
BL21(λDE3) derivative lacking the cobT gene [20]. This variant was
overproduced and purified using a procedure similar to the one described
for the purification of SoArsAB variants [21]. SeCobT variants concentrated
to 6 mg/mL were dialyzed in Tris–HCl buffer (20 mM, pH 7.5) containing
NaCl (100 mM) before storing them at −80 °C [17] until used.
Specific activities of SeCobT variants on DMB and adenine.
Variant
Specific Activity (μmol min−1 mg−1
)
DMB
Adenine
SeCobTWT
30
0.5 0.4
1.5 0.4
2
1.6 0.6
0.03 0.01
0.07 0.01
1e-3 2e-4
SeCobTE174A
SeCobTE317A
SeCobTE174A,E317A
2e-3 1e-3
Conditions used for the assay are described under the Materials and methods section.

C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
467
Table 3
packed columns were equilibrated with water at a rate of 2 mL/min
using an ÄKTA explorer FPLC system (GE Healthcare). Desalted products
of the NadC reaction were loaded onto the column, which was washed
with 30 mL of water. NaMN was eluted off the column with NaCl
(30 mM). To elute the remaining reaction components, a linear gradient
of 300 mM NaCl was used. NaMN elution was monitored at 260 nm. Pu-
rified NaMN was desalted as outlined above. Conductivity (YSI Scientific
Conductance Meter Model 35) was measured to determine whether
or not NaMN was salt-free. A standard curve of NaCl concentrations
was used to determine that b1 mM NaCl remained in the solution. The
NaMN concentration was determined using the molar extinction coeffi-
cient at 265 nm of 3300 M−1. NaMN was lyophilized and the solid was
divided into 2-mL serum vials containing approximately 5 mg, then
flushed with nitrogen before sealing and storing at −20 °C.
Doubling time (h) of a S. enterica cobT cobB strain harboring plasmids encoding SeCobT
variants.
Variant
Cbi
Cbi + DMB
Cbi + Adenine
NG (1 mM)
Cbi + Cbl
Vector only
SeCobTWT
NG
2
NG
NG (600 μM)
2
2
2
0.1
0.1
0.3
0.1
2
0.1 (10 nM)
2
0.1 (10 nM)
SeCobTE174A
19 5 (1 μM)
NG (1 mM)
2
0.4 (10 μM)
25 3 (10 nM)
0.1 (50 nM)
NG (0.6 mM)
SeCobTE317A
NG
NG
35 4 (0.5 mM)
2
2
0.1
0.1
4
3
0.5 (1 mM)
SeCobTE174A,E317A
NG (1 mM)
Cells were grown in NCE minimal medium with glycerol as the sole source of carbon and
energy. Indicated in parentheses is the concentration of base substrate present in the reaction
mixture. Up to 1 mM (for adenine) and 0.6 mM (for DMB) were used to restore growth to
WT levels as indicated. Cbi, [(CN)2Cbi]; Cbl, (CNCbl); NG, no growth.
2.5. Reaction with p-cresol as substrate
2.7. In vitro phosphoribosyltransferase activity assay
Phosphoribosylation reactions using SeCobTWT, SeCobTS80Y,Q88M,L175M
or SoArsAB with NaMN and p-cresol were performed at 37 °C for 6 h.
Each reaction was performed in 600 μL of glycine buffer (100 mM,
pH 10) containing the enzymes (60 μg each) NaMN and p-cresol (3 mM
each). The reaction mixture was applied onto a C18 Sep-Pak column
equilibrated and washed with water and eluted with 100 μL of methanol.
Ten-microliter samples were analyzed on a Shimadzu HPL chromato-
graph equipped with a Kinetex 2.6 μm C18 100 Å (100 × 4.60 mm) col-
umn (Phenomenex) as described [21].
Conditions for the phosphoribosyltransferase activity assay when
NaMN and DMB are substrates are described elsewhere [33]. These con-
ditions were also used when measuring activity with adenine. [14C,C-8]
Adenine was purchased from Moravek Biochemicals, Inc., and [14C,C-2]
DMB was synthesized as described [33]. SeCobT proteins were diluted
in glycine–NaOH buffer (50 mM, pH 10.0) containing bovine serum
albumin (0.05% wt/v) and sodium azide (4.6 mM). KCl (0.5 M final)
was added to stop the reactions. Specific activities were determined
for the formation of α-DMB-riboside monophosphate (α-RP, when
DMB was used as substrate), and of α-adenine-riboside monophosphate
(α-AMN, when adenine was used as substrate). The rate of α-RP synthe-
sis was performed using 50 ng of SeCobTE174A and SeCobTE317A variants,
while the rate of α-AMN synthesis required 1.25 μg of SeCobTE174A and
SeCobTE317A variants, respectively. Control experiments employing
SeCobTWT, required 2.5 ng of enzyme for the synthesis of α-RP, and
25 ng of enzyme for the synthesis of α-AMN. Substrate concentrations
were used at saturating levels for each protein and initial rates were
measured. Activities for α-RP and α-AMN formations were assayed
twice over a period of 0.75 and 1.5 h, respectively. The specific activity
of SeCobTE174A,E317A was determined for α-RP and α-AMN formation
using 1 μg and 17 μg of protein, respectively, over a period of 2 and 4 h,
respectively.
2.6. Synthesis and purification of nicotinic acid mononucleotide (NaMN)
At the time this work was performed, NaMN was not commercially
available hence it was synthesized in house. The enzymatic synthesis
of NaMN was modified from a published method [30]. Briefly, a 3-mL re-
action mixture containing sodium phosphate buffer, pH 7.0 (250 μmol),
5-phosphorylribose 1-pyrophosphate (PRPP, 5 μmol), quinolinic acid
(QA, 5 μmol), magnesium chloride (100 μmol), and phosphoribosyl-
pyrophosphate (PRPP)-quinolinate phosphoribosyltransferase (NadC,
5 mg) was incubated at 37 °C. NadC was purified using the ASKA library
(A Complete Set of E. coli genes K-12 ORF Archive) [31]. To circumvent
substrate inhibition, additional 5 μmol of PRPP and QA were added
to the reaction mixture at 1.5 h intervals. A total of 20 μmol each of
PRPP and QA was added, and the reaction mixture was incubated for a
total of 6 h. Ten 3-mL NadC reactions were performed, all of which
were pooled and desalted over a 225-mL Sephadex G-10 column (GE
Healthcare) at room temperature. Deionized water was further stripped
of contaminants using a Millipore Milli-Q water purification system.
Water treated this way was used to equilibrate the column and to
resolve components of the NadC reaction mixture at a rate of 1 mL/min
using a GE Healthcare P1 peristaltic pump. Fractions were collected and
absorbance was measured at 265 nm to identify fractions containing
NaMN. The latter were lyophilized and solids were re-suspended
in water for purification using DEAE chromatography as described else-
where [32]. Three 5-mL GE Healthcare HiTrap DEAE FF (FastFlow) pre-
2.8. Mass spectrometry
The molecular mass of α-p-cresolyl-riboside monophosphate was
determined by electrospray ionization (ESI) mass spectrometry on an
Agilent LC/MSD TOF spectrometer as described elsewhere [20].
2.9. Crystallization of SeCobTE174A in complex with adenine and DMB
All protein and substrate complexes were screened at room
temperature by hanging drop vapor diffusion using a 144-condition
sparse matrix screen developed in the Rayment Laboratory. Single,
diffraction quality crystals of SeCobT174A in complex with adenine
were grown by mixing equal volumes of 6 mg/mL SeCobT174A, ade-
nine (1 mM), Tris-HCl buffer (20 mM, pH 7.5), NaCl (100 mM), and
dimethylsulfoxide (DMSO, 1%, v/v) with a reservoir solution composed
of 2-(N-morpholino)ethanesulfonic acid (MES-NaOH) buffer (100 mM,
pH 6.0) containing (NH4)2SO4 (1.05 M), and glycerol (5%, v/v). Hanging
drops were nucleated after 4 h with a fine cat whisker. The crystals
grew to a size of 0.4 × 0.4 × 0.3 mm within 4 days and were transferred
to a synthetic mother liquor composed of MES-NaOH buffer (100 mM,
pH 6.0), (NH4)2SO4 (1 M), adenine (1 mM), DMSO (1%, v/v), and glycerol
(5%, v/v). The crystals were transferred into a cryoprotectant solution
with the same composition except for the presence of glycerol (17%,
v/v). The crystals were flash cooled in liquid N2.
Table 4
Doubling times (h) of a S. enterica cobT cobB strain harboring plasmids encoding SoArsAB
variants.
Variant
Cbi
Cbi + DMB
Cbi + Cbl
Vector only
NG
3.4 0.1
NG
NG
NG
NG
2.4 0.01
2.5 0.01
2.8 0.02
2.3 0.01
SoArsABWT
2.7 0.01
2.7 0.01
2.6 0.02
NG
SoArsAE176AArsB
SoArsAE319AArsB
SoArsAE176A,E319AArsB
SoArsAArsBE171A,E303A
3
0.01
4.2 0.04
4.0 0.03
3.2 0.02
Cells were grown in NCE minimal medium with glycerol as the sole source of carbon and
energy. When added to the culture medium Cbi, [(CN)2Cbi], was at 15 nM; Cbl, [CNCbl],
was at 15 nM; and DMB was at 150 μM; NG, no growth.
Single, diffraction quality crystals of SeCobT174A in complex with
DMB were grown by mixing equal volumes of 6 mg/mL SeCobTE174A
,

468
C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
Table 5
List of strains, plasmids and primers used in this study.
Strains
Genotype
Reference
S. enterica strains
JE2017
TR6583
cobA367::Tn10d(tet+)/pGP1-2 T7 rpo+ (T7 RNAP) kan+
metE205 ara-9
Lab collection
K. Sanderson
via J. Roth
Derivatives of TR6583
JE2501
JE2607
cobB1176::Tn10d(tet+) cobT109::MudJ
[46]
[15]
cobB1176::Tn10d(tet+) cobT109::MudJ recA1
E. coli strains
JE13607
BL21(λDE3) cobT762::kan+
[20]
[47]
Plasmid
pTEV5
Genotype and Description
Overexpression vector that fuses the N-terminus of the protein
of interest to a H6 tag, which can be removed by rTEV
protease, bla+
pBAD33
pT7-5
pJO27
Complementation vector ParaBAD cat+
[48]
[49]
[50]
[33]
Cloning vector, bla+
S.e. cobT+ encodes SeCobTWT in pT7-5, bla+
S.e. cobT+ encodes SeCobTWT in pSU18, cat+
S.e. cobT1475 (encodes SeCobTE317A) in pSU18, cat+
S.e. cobT1475 (encodes SeCobTE317A) in pT7-5, bla+
S.e. cobT1476 (encodes SeCobTE174A) in pT7-5, bla+
S.e. cobT1477 (encodes SeCobTE174A, E317A) in pT7-5, bla+
pCOBT10
pCOBT15
pCOBT19
pCOBT42
pCOBT57
pCOBT141
pCOBT184
pARSAB4
pARSAB8
pARSAB9
pARSAB16
pARSAB18
S.e. cobT+ encodes SeCobTWT in pTEV5, bla+
S.e. cobT1478 (encodes SeCobTS80Y,
L175M) in pTEV5, bla+
Q88M,
S.o. arsAB+ encodes SoArsABWT in pBAD33, cat+
S.o. arsAB (ArsAE176AArsB) in pBAD33, cat+
S.o. arsAB (ArsAE319AArsB) in pBAD33, cat+
S.o. arsAB (ArsAE176A,E319AArsB) in pBAD33, cat+
S.o. arsAB (ArsA–ArsBE171A,E303A) in pBAD33, cat+
[20]
Primer Name
Primer Sequence
cobT13 5′
TAT TTG GCG TAG GGG AAC
cobT13 3′
CGC TTC CTG CAC CTA ACC
cobT14 5′
TGC AGG AAG CGG CGC GGC
cobT14 3′
GTT TTC CCA GTC ACG AC
pTEV5-PIPE-5′
pTEV5-PIPE-3′
CobT PIPE pTEV5 5′
CobT PIPE pTEV5 3′
CobT E174A 5′
CobT E174A 3′
CobT S80Y 5′
CobT S80Y 3′
CobT Q88M 5′
CobT Q88M 3′
CobT L175M 5′
CobT L175M 3′
ArsA E176A 5′
ArsA E176A 3′
ArsA E319A 5′
ArsA E319A 3′
ArsB E171A 5′
ArsB E171A 3′
ArsB E303A 5′
ArsB E303A 3′
GCC CTG AAA ATA CAG GTT TTC ACT AGT TG
CCA TGG AAT TCT CGA GCT CCC G
AAC CTG TAT TTT CAG GGC ATG CAG ACA CTA CAC GCT TTA CTC CGT G
AGC TCG AGA ATT CCA TGG TTA TGT TGC GTT TGC GTT CCC CTC
AGG GGC GCT GGG AAT GGC GAA C
GCG CCC CTA CGC CAA ATA AGG T
GAA GGC GTA GCG GTT TAT CCC AAA ATC GTG ACG
CGT CAC GAT TTT GGG ATA AAC CGC TAC GCC TTC
CAA AAT CGT GAC GGC GAT TAT GGC GGC GAA TAT GAC GCG
CGC GTC ATA TTC GCC GCC ATA ATC GCC GTC ACG ATT TTG
CTT ATT TGG CGT AGG GGA GAT GGG AAT GGC GAA CAC TAC
GTA GTG TTC GCC ATT CCC ATC TCC CCT ACG CCA AAT AAG
CTG TTT TTG TCT GGG AGC GAT GGG CAT TGG TAA TAC
GTA TTA CCA ATG CCC ATC GCT CCC AGA CAA AAA CAG
AGT CCG GCT GGG AGC GGG TAT CGG GGC TTC
GAA GCC CCG ATA CCC GCT CCC AGC CGG ACT
GTG ATC GGC TTG GGC GCG ATG GGG CTG GGC GGT TT
AAA CCG CCC AGC CCC ATC GCG CCC AAG CCG ATC AC
CTT AAA ATG AAC CTG GGA GCG GGG ACA GGT GCA GCA CTC
GAG TGC TGC ACC TGT CCC CGC TCC CAG GTT CAT TTT AAG
DMSO (1 mM), Tris–HCl buffer (20 mM, pH 7.5), NaCl (100 mM),
and ethanol (0.85%, v/v) with a reservoir solution composed of MES-
NaOH buffer (100 mM, pH 6.0), methyl ether polyethylene glycol 5000
(MEPEG 5K, 10%, w/v), (NH4)2SO4 (200 mM), and glycerol (5%, v/v).
Hanging drops were nucleated after 4 h with a fine cat whisker. Crystals
grew to a size of 0.3 × 0.3 × 0.2 mm within 4 days and were transferred
to a synthetic mother liquor composed of MES-NaOH buffer (95 mM,
pH 6.0), MEPEG 5K (9.5%, w/v), (NH4)2SO4 (190 mM) NaCl (95 mM),
DMB (1 mM), ethanol (0.85%, v/v), and glycerol (4.75%, v/v). The crys-
tals were transferred to a cryoprotectant solution with the same
composition except for the presence of MEPEG 5K (11.5%, w/v) and
glycerol (17%, v/v) in a stepwise fashion. The crystals were flash frozen
in liquid N2.
2.10. Crystallization of SeCobTE317A in the substrate free state and in complex
with reaction products α-DMB riboside monophosphate and nicotinate
Single, diffraction-quality crystals of SeCobTE317A were grown
by mixing equal volumes of 6 mg/mL SeCobTE317A, Tris–HCl buffer
(20 mM, pH 7.5), and NaCl (100 mM) with a reservoir solution com-
posed of MES-NaOH buffer (100 mM, pH 6.0), polyethylene glycol 1500
(PEG 1.5K, 20%, w/v), Li2SO4 (50 mM) and ethylene glycol (2.5%, v/v).
Crystals were also grown in the presence of either adenine (1 mM) or
DMB (1 mM), however no electron density was observed for either sub-
strate. The mixture was centrifuged at 16,000 ×g for 10 min to remove
nuclei. Four-microliter droplets of the cleared mixture were nucleated
with a fine cat whisker and suspended over a reservoir solution. Crystals

C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
469
grew to a size of 0.4 × 0.4 × 0.3 mm within 7 days and were transferred
to a synthetic mother liquor composed of MES-NaOH buffer (95 mM,
pH 6.0), PEG 1.5K (19%, w/v), Li2SO4 (48 mM), NaCl (95 mM), adenine
(1 mM), DMSO (1%, v/v), and ethylene glycol (2.4% v/v). The crystals
were transferred to a cryoprotectant solution with the same composition
except for the presence of PEG 1.5K (20%, w/v) and ethylene glycol
(16.2%, v/v). Crystals with reaction products α-DMB riboside mono-
phosphate and nicotinate were obtained by transferring the crystals
into the same synthetic mother liquor as above with the addition of
1 mM DMB and 10 mM NaMN. The crystals were soaked for 12 h and
transferred to the same cryoprotectant solution as above and flash
cooled in liquid N2.
residue (174) in the active site, and similar effects on activity were mea-
sured. Remarkably, the double substitution variant (SeCobTE317A,E174A
retained activity in the nmol/min/mg range (a decrease of 15,000 fold
from SeCobTWT) showing that under the assay conditions used, a
protein-derived catalytic base was not required for activity.
)
AdoCbl synthesis was restored in a S. enterica cobT cobB strain harbor-
ing a plasmid encoding the SeCobTE174A variant when DMB (10 μM) was
provided in the medium. Under such conditions the mutant strain grew
at a rate similar to that of the wild-type strain. In contrast, restoration
of AdoCbl synthesis in the S. enterica cobT cobB strain harboring a plasmid
encoding the SeCobTE317A variant required 200-fold less exogenous DMB
(50 nM) to reach wild-type growth rates (Table 3). Both SeCobTE174A and
SeCobTE317A variants had similar activities when adenine was the sub-
strate, but only the SeCobTE317A variant supported pseudo-B12 syn-
thesis in vivo when supplemented with 1 mM adenine. In contrast,
the SeCobTE174A variant expressed from a plasmid failed to support
pseudo-B12 synthesis even when the concentration of adenine in the
medium was 1 mM (Table 3). Growth was not supported by the double
variant SeCobTE174A,E317A under any condition tested. The limited impact
of changing Glu174 or Glu317 raises two questions: i) is the role of any of
these residues catalytic? and ii) do these residues play an equivalent role
in the heterodimer SeCobT homolog SoArsAB, which is able to catalyze
the synthesis of phenolic riboside monophosphates?
2.11. Crystallization of SeCobTS80Y,Q88M,L175M in the presence of p-cresol
Single, diffraction quality crystals of SeCobTS80Y,Q88M,L175M were
grown by mixing equal volumes of 6 mg/mL SeCobTS80Y,Q88M,L175M
,
Tris–HCl buffer (20 mM, pH 7.5), and NaCl (100 mM) with a reservoir
solution composed of N-cyclohexyl-2-aminoethanesulfonic acid (CHES–
NaOH) buffer (100 mM, pH 9.0), MEPEG 5K (24%, w/v), MgSO4
(50 mM), p-cresol (5 mM), DMSO (1%, v/v), and ethylene glycol
(5%, v/v). Four-microliter droplets were nucleated with a fine cat
whisker and suspended over a reservoir solution. Crystals grew to a size
of 0.3 × 0.3 × 0.2 mm within 7 days and were transferred to a synthetic
mother liquor composed of CHES–NaOH buffer (100 mM, pH 9.0),
MEPEG 5K (24%, w/v), MgSO4 (50 mM) NaCl (100 mM), p-cresol
(5 mM), DMSO (1%, v/v), and ethylene glycol (5%, v/v). The crystals
were transferred to a cryoprotectant solution with the same composition
except for the presence of 17% ethylene glycol. The crystals were flash
cooled in liquid N2. Crystals with p-cresol and NaMN were obtained
by transferring the crystals to the above cryoprotectant solution with
the addition of 10 mM NaMN. Crystals that were frozen after 5 min of
incubation contained both reaction products, whereas crystals that were
frozen after 10 days of incubation contained only p-cresol.
3.2. Equivalent glutamate residues in SoArsA are critical for activity of the
heterodimeric SoArsAB
Variant SoArsAB proteins were generated with alanine substitutions
in the equivalent glutamate residues to investigate whether or not these
residues were needed for activity of this SeCobT homolog. Wild-type
SoArsAB catalyzes the synthesis of α-phenolyl, α-p-cresolyl and α-
benzimidazolyl riboside monophosphates [20]. Both SoArsA and SoArsB
have glutamate residues equivalent to Glu174 and Glu317 in SeCobT
[21]. SoArsAB variants with single Glu-to-Ala changes in the SoArsA
subunit were active in vivo when DMB was provided in the medium,
a result that was similar to the one obtained with SeCobT variants
(Table 4). Notably, SoArsAB lost all of its in vivo activity when both glu-
tamate residues in SoArsA were changed to alanine suggesting that the
active site within the SoArsB subunit does not contribute to the activity
despite possessing equivalent glutamate residues. Substitutions in
SoArsB did not affect activity in vivo (Table 4), a result that was expected
given the absence of any base substrate binding in the crystal structure
of SoArsAB [21]. The role of the SoArsB subunit remains unclear.
2.12. Data collection and refinement
All except one X-ray diffraction data set were collected at 100 K with
a Bruker AXS Platinum 135 CCD detector equipped with Montel optics
and controlled by the Proteum software suite (Brucker AXS Inc.)
(Table 1). These data sets were integrated with SAINT version 7.06A
software and internally scaled with SADABS version 2005/1. The X-ray
data for the SeCobTE317A mutant protein in complex with α-DMB
riboside monophosphate and nicotinate were collected at 100 K on
the Structural Biology Center beamline 19BM at the Advanced Photon
Source in Argonne, IL. Although the crystals diffracted beyond the
1.4 Å resolution at that X-ray source, a cutoff at that resolution was cho-
sen to optimize the quality of both the high and low resolution data
with two data collection scans. These diffraction data were integrated
and scaled with HKL3000 [34]. The variant structures were determined
by molecular replacement using PHASER [35] in which the substrate-
free SeCobT (PDB id: 1L4B) was used as a search model. Model refine-
ment was performed by alternate cycles of manual building with Coot
and restrained refinement with Refmac5 [36]. Data collection and
refinement statistics are presented in Table 1.
3.3. Conservation of the acidic residues in SeCobT orthologues
Alignment of SeCobT orthologs (PFAM 02277; EC 2.4.2.21) encoded
in representative genomes of bacteria and archaea revealed that acidic
residues are conserved in the active site of this family of enzymes
(Fig. 2A). Structural alignments of SeCobT and two archaeal CobT
orthologs whose structures have been determined, showed a valine res-
idue replacing the equivalent Glu317 in SeCobT (Fig. 2B). Together with
the mutational analyses of SeCobT and SoArsAB, this alignment sug-
gested that an acidic residue at a position equivalent to Glu174 in
SeCobT alone might be sufficient for biological activity in archaeal
CobT orthologs.
3. Results and discussion
3.4. Crystal structures of SeCobTE174A and SeCobTE317A variants reveal the
complexity of the role of Glu174 and Glu317
3.1. Glutamate 174 and 317 are essential for optimal activity
It was proposed in the initial characterization of SeCobT that the
γ-carboxyl group of Glu317 in the active site served as a general
base to catalyze the phosphoribosyl-transfer [16,17]. To test this idea
we replaced of Glu317 with alanine, a substitution that decreased the
activity of the enzyme for DMB and adenine by 20–60 fold in vitro
(Table 2). The same substitution was made in the adjacent glutamate
To better understand the molecular effects of substitutions at Glu174
and 317 in SeCobT, both variants were subjected to extensive crystalli-
zation studies. Structures of substrate-free SeCobTE317A, SeCobTE317A in
complex with DMB riboside monophosphate (α-RP) and nicotinate,
SeCobTE174A in complex with adenine, and SeCobTE174A in complex
with DMB were determined to a 2.0 Å resolution or better. Interestingly,

470
C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
Fig. 2. A. Sequence alignments (ClustalW) of the Glu174 and Glu317 regions in CobT orthologs. Alignments of these regions in archaeal sequences were aided by structural comparisons. In
archaea, the C-terminal region containing the acidic residue is not conserved. B. Structural alignments of P. horikoshii (Yellow, 3U4G) and M. jannaschii (Red, 3L0Z) CobT with S. enterica
CobT complexed with p-cresol (White, 1JHU). The structures of the archaeal proteins were determined in the absence of any ligand and it is unknown whether they exhibit the same
enzymatic function as SeCobT.
crystals of the SeCobTE317A variant grown in the presence of DMB or
adenine did not exhibit these substrates in their active sites. As with
SeCobTWT, the variants crystallized with a monomer in the asymmetric
unit, while the biologically active dimer is generated by a crystallo-
graphic two-fold axis. This results in two identical active sites per
dimer, where each is formed by components of both protomers. Each
subunit contains a large and small domain. The large domain exhibits
a six-stranded parallel beta sheet with connecting helices characteristic
of the Rossmann fold, while the small domain is built from helices
contributed by both the N- and C-termini of the polypeptide chain to
form a three-helix bundle (Fig. 3A).
of the carbonyl oxygen of Gly316 nearly 180° toward the position that
was formerly occupied by the side chain of Glu174 in the SeCobTWT
structure. Although Glu174 was further away from the base substrate
at a distance improbable for direct proton abstraction, structural analy-
sis of this variant showed that this residue plays a role in stabilizing
Glu317 in the active conformation.
As noted above, the SeCobTE317A variant crystallized devoid of sub-
strates, despite the addition of either DMB or adenine during crystalliza-
tion, and was thus considered a substrate-free form of the enzyme
(Fig. 4). However, the mutation induced substantial structural changes
in the active site pocket relative to that of substrate-free SeCobTWT. Sub-
stitution of alanine for Glu317 caused a backbone shift in the loop that
underlies the active site (Met313 to Ala317). This shift likely resulted
from the loss of a single hydrogen bond between Oε2 of Glu317 and
the side chain of Gln88 (Fig. 4). The disruption of this hydrogen bond
caused a collapse of the active site in which several residues (Gln88,
Leu315, and Arg314) shifted inwards to fill the substrate binding pock-
et, thereby precluding substrate binding. The replacement of Glu317
to alanine released the disfavored torsional angles in the 317 position
likely triggering the collapse. There was also a substantial change in
In the SeCobTE174A variant, DMB and adenine adopt the same orien-
tation as observed in SeCobTWT (data not shown for SeCobTE174A in com-
plex with adenine) [14]. Interestingly, the change of Glu174 to alanine
introduced conformational flexibility at Gly316–Glu317 (Fig. 3B). Two
conformations were observed in the crystal structure. In one conforma-
tion the residues were in the same position as in SeCobTWT, and in the
other, the carboxylate moiety on the side chain of Glu317 was shifted
away from its native orientation thereby breaking a close hydrogen
bond with the substrate DMB. This shift was facilitated by the rotation

C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
471
Fig. 4. Stereo view of the active site of the substrate-free form of CobTE317A (magenta) aligned
with CobTE317A in complex with reaction products α-DMB-riboside monophosphate (α-RP)
and nicotinate (NA) (light gray). Several conformational rearrangements take place in the
absence of substrate (magenta): the peptide backbone near Leu315 is rearranged, the side
chains of Glu88 and Arg314 shift, and the loop that folds over the second substrate binding
site in product complex (yellow; Val198–Pro207) is disordered. The active site residues are
restored to active conformation (light gray) in the presence of reaction products.
3.5. Are the roles of Glu174 and Glu317 strictly structural?
Together with the reduced in vitro activities of the SeCobTE174A and
SeCobTE317A, these results appear to support the idea that glutamates
174 and 317 are important for the maintenance of the active site archi-
tecture, but are not critical for enzymatic activity. This idea is consistent
with the expectation that removal of a catalytic base from SeCobT would
result in an ~105 loss in activity based on the precedence set for active
site residues derived from study of the serine proteases [37]. Notably,
we measured only a modest decrease (~20–60 fold) in the activity
of the SeCobTE174A and SeCobTE317A variants, and the SeCobTE174A E317A
variant retained activity at the nmol/min/mg level. However, the conclu-
sion that Glu174 and Glu317 are not catalytic residues deserves further
discussion for two reasons. Firstly, it should be noted that the in vitro ac-
tivity assay for SeCobT is performed at the non-physiological pH of 10, be-
cause at neutral pH, the activity of SeCobT is undetectable [13]. Thus it is
plausible that at high pH a catalytic base is not required to remove a pro-
ton from the substrate, but might be more important under physiological
conditions. Also, if the activity of the SeCobTE174A,E317A variant at pH 10 is
four orders of magnitude lower than the activity of SeCobTWT (Table 2), its
Fig. 3. A. Cartoon representation of SeCobTE174A dimer. The large domains are shown in
pale cyan, the small domains are shown in teal. DMB is shown in space fill as dark gray.
B. Stereo view of the active site for SeCobTE174A (teal). This view reveals that the backbone
for residues 316 and 317 in SeCobTE174A adopts multiple conformations (teal and green).
One of these (teal) is essentially identical to the wild type protein where in the wild-
type protein this is maintained by a hydrogen bond between Glu174 and the backbone
amide hydrogen on Glu317 (white; from coordinates 1D0S). The alternate conformation
(green) is formed by a rotation of the peptide bond at Gly316, which breaks the interaction
between Oε2 of Glu317 with the DMB substrate. The interaction between Gly316 and
Glu174 (black arrow) in the wild-type protein prevents the backbone from adopting the
alternate conformation.
activity at physiological pH is likely to be ≤10−5 relative to SeCobTWT
.
Secondly, such a precipitous loss of activity is consistent with the fact
that, in vivo, not even the elevated level of the SeCobTE174A,E317A variant
encoded on a plasmid (at 15-fold greater level than a chromosomally
encoded gene) can restore AdoCbl synthesis in a S. enterica cobT strain.
The caveats imposed by the pH conditions under which the activity
of the enzyme is measured (pH 10), make it difficult to ascertain whether
or not Glu174 or Glu317 acts as the base that triggers the attack on NaMN,
as proposed earlier [16]. Studies of a SeCobT ortholog that retains full ac-
tivity at pH 7 would allow us to determine the precise role of residues
equivalent to Glu174 and Glu317 in SeCobT.
the water structure. Additionally, residues Gly202–Leu205, which
comprised much of the NaMN binding region, were disordered in the
SeCobTE317A substrate-free form (Fig. 4). This was likely due to the
movement of Arg314. In addition, residue Glu174 rotated away from
the substrate-binding site. Finally, the component of the active site con-
tributed by the adjacent subunit (Leu30–Pro32) adopted a new confor-
mation in which the loop encroached further into the adjacent active
site (Fig. 4), again precluding substrate binding. These changes in the
active site architecture explained the decrease in activity for both
adenine and DMB, and suggested that Glu317 played an important
role in maintaining the integrity of the active site. The structure of the
substrate-free form of SeCobTE317A most likely represented an inactive
conformation trapped by the crystal lattice, since the variant did exhibit
biological activity in vitro and in vivo (Tables 2, 3). However, when crys-
tals of SeCobTE317A grown in the presence of DMB were soaked with
NaMN the reaction occurred in the crystal resulting in a complex with
α-RP and nicotinate. In this case, the overall orientation of the residues
in the active site was restored to that of SeCobTWT suggesting that both
substrates contributed to the active conformation of this variant (Fig. 4).
Interestingly, the occupancy for both reaction products was low despite
the overall high quality of the crystallographic data.
3.6. Changes in the primary sequence of SeCobT allow the resulting variant
to phosphoribosylate phenolic substrates
A striking difference between SeCobT and SoArsAB is the ability
of the latter to form an α-O-glycosidic bond between the hydroxyl of
the phenolic moiety and the C1 of ribose. Structural determination of
SoArsAB in complex with p-cresol revealed that the overall structures
of these two enzymes were remarkably similar [21]. Examination of
the structures, coupled with a ClustalW sequence alignment of active
site residues of SoArsA, SoArsB and SeCobT suggested how SoArsAB
might catalyze the reaction with phenolic compounds [21]. This led to

472
C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
a list of potential residues that could be targeted for site-directed
mutagenesis to test the hypothesis that the specificity of SeCobT might
be expanded to include phenolic compounds (Fig. 5).
was brought into close proximity to the C1 of ribose when NaMN was
modeled into the active site.
Subsequent mutagenesis of the SeCobT active site identified a
variant with three amino acid substitutions S80Y, Q88M and
L175M that did phosphoribosylate p-cresol. This was the minimum
set of changes in SeCobT needed to convert p-cresol into a product.
These substitutions mimicked those of the SoArsA active site. HPLC
was used to resolve the product formed with p-cresol and NaMN using
SoArsAB and SeCobTS80Y,Q88M,L175M compared to SeCobT (Fig. 6A). The
product was confirmed by mass spectrometry to match the predicted
mass of α-p-cresolyl-riboside monophosphate (Fig. 6B).
In the lattice of SeCobT crystals, NaMN was hydrolyzed to nicotinate
when p-cresol was also present in the active site, but product formation,
i.e., α-p-cresolyl-riboside monophosphate was not observed. In contrast,
product formation did occur in the crystal lattice with other benzimid-
azole and purine bases [14]. Structural analyses of SoArsAB, a heterodi-
meric ortholog of SeCobT from S. ovata that phosphoribosylates
phenolics, showed that the hydroxyl group in p-cresol was rotated
away from Glu319 (Glu317 equivalent in SeCobT) [21], and thereby
3.7. Structural analyses of the SeCobTS80Y,Q88M,L175M variant provide support
for the role of water in the formation of phenolic riboside monophosphates
To further characterize the changes in SeCobT that led to the phenolic
activity, crystal structures of the SeCobTS80Y,Q88M,L175M variant were
determined with p-cresol and p-cresol and NaMN. The overall structure
of the SeCobTS80Y,Q88M,L175M variant in complex with p-cresol was the
same as SeCobT (Fig. 7A). The orientation of p-cresol in the variant did
not change in comparison to SeCobT. However, when crystals grown in
the presence of p-cresol were briefly soaked in a solution containing
NaMN, both substrates were well defined in the SeCobTS80Y,Q88M,L175M
active site, but the reaction did not occur (Fig. 7). These results differed
from those obtained with SeCobTWT crystals in which NaMN was hydro-
lyzed when incubated with p-cresol. Importantly, p-cresol bound to the
variant in crystals soaked with NaMN for long time periods adopted a
rotated conformation similar to that in SoArsAB in one of the active sites
in the dimer (Fig. 5A). An ordered water molecule between the hydroxyl
Fig. 5. A. Structure alignments of SeCobTS80Y,Q88M,L175M (CobT YMM, orange) and SoArsAB (blue) with SeCobT (white). Both SeCobTS80Y,Q88M,L175M and SeCobTWT form a close hydrogen
bond with p-cresol, while in SoArsAB this bond is mediated by a water molecule. B. Sequence alignments (ClustalW) of several ArsA orthologues in Dialister invisus, Thermsinus
carboxydivorans, Veillonella parvula, Sporomusa ovata, and Acetonema longum superimposed on a contact diagram for p-cresol. This alignment shows the conservation in residues surround-
ing the substrate base.

C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
473
Fig. 6. A. HPLC separation of product formation in 6 h from enzymes ArsAB, CobTWT, and CobTYMM (CobTS80Y,Q88M,L175M) using p-cresol and NaMN. B. Mass analysis (negative ion mode) of
the material under the peak eluted from the SeCobTS80Y,Q88M,L175M reaction (319 amu) was consistent with the predicted mass of α-p-cresolyl-riboside monophosphate (320 amu).
Fig. 7. Stereo view of the electron density for both NaMN and p-cresol in SeCobTS80Y,Q88M,L175M (A) and p-cresol and SeCobTS80Y,Q88M,L175M (B). Electron density (1.8 σ) was calculated from
coefficients of the form Fo − Fc where p-cresol and NaMN were omitted from phase calculation and refinement.

474
C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
group of p-cresol and Glu317 was observed in the SeCobTS80Y,Q88M,L175M
variant. Interestingly, neither NaMN nor any reaction products were
observed in this structure. Notably, an ordered water molecule was
proposed to facilitate the proton abstraction that triggered the α-O-
glycosidic bond formation in SoArsAB [21]. These results support the
proposed mechanism of phenolic phosphoribosylation through an active
site water molecule [21].
Acknowledgments
This work was supported in part by the NIH grants GM083987 and
GM086351 to I.R. and R37 GM40313 to J.C.E.-S. Use of the SBC 19BM
beamline at the Argonne National Laboratory Advanced Photon Source
was supported by the U.S. Department of Energy, Office of Energy Re-
search, under Contract No. W-31-109-ENG-38.
Although these results indicate that SeCobTWT can be changed into an
enzyme capable of phenolic phosphoribosylation, the CobTS80Y,Q88M,L175M
variant synthesized far less α-p-cresolyl riboside monophosphate than
the SoArsAB enzyme toward phenolic substrates. This is not surprising
given the numerous small differences that likely contribute to the optimi-
zation of SoArsAB activity. The observed differences could be due to the
constraints placed on Glu317 by the hydroxyl side chain of Ser319,
which forms a close hydrogen bond (2.6 Å) with the carbonyl oxygen
of Gly316 in SeCobTS80Y,Q88M,L175M. Ser319 is replaced with an isoleucine
in SoArsA, which was shown to be important for activity in SoArsAB [21].
However, alignments do not show strong conservation at this position
(Fig. 2), indicating that individual substitutions need to be considered
in the context of the surrounding structure.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.bbagen.2013.09.038.
References
[1] E.N. Marsh, Coenzyme B12 (cobalamin)-dependent enzymes, Essays Biochem. 34
(1999) 139–154.
[2] B. Kräutler, W. Fieber, S. Osterman, M. Fasching, K.-H. Ongania, K. Gruber, C. Kratky, C.
Mikl, A. Siebert, G. Diekert, The cofactor of tetrachloroethene reductive dehalogenase
of Dehalospirillum multivorans is Norpseudo-B12, a new type of natural corrinoid,
Helv. Chim. Acta 86 (2003) 3698–3716.
[3] A.R. Battersby, Tetrapyrroles: the pigments of life, Nat. Prod. Rep. 17 (2000)
507–526.
[4] R. Banerjee, Radical carbon skeleton rearrangements: catalysis by coenzyme B12-
dependent mutases, Chem. Rev. 103 (2003) 2083–2094.
4. Conclusion
[5] V. Bandarian, G.H. Reed, Ethanolamine ammonia-lyase, in: R. Banerjee (Ed.), Chem-
istry and Biochemistry of B12, John Wiley & Sons, Inc., New York, 1999, pp. 811–833.
[6] T. Toraya, Enzymatic radical catalysis: coenzyme B12-dependent diol dehydratase,
Chem. Rec. 2 (2002) 352–366.
[7] R.G. Matthews, M. Koutmos, S. Datta, Cobalamin-dependent and cobamide-dependent
methyltransferases, Curr. Opin. Struct. Biol. 18 (2008) 658–666.
[8] P. Renz, Biosynthesis of the 5,6-dimethylbenzimidazole moiety of cobalamin and
of other bases found in natural corrinoids, in: R. Banerjee (Ed.), Chemistry and Bio-
chemistry of B12, John Wiley & Sons, Inc., New York, 1999, pp. 557–575.
[9] M.J. Warren, E. Raux, H.L. Schubert, J.C. Escalante-Semerena, The biosynthesis of
adenosylcobalamin (vitamin B12), Nat. Prod. Rep. 19 (2002) 390–412.
[10] J.C. Escalante-Semerena, M.J. Warren, Biosynthesis and use of cobalamin (B12), in: A.
Böck, R. Curtiss III, J.B. Kaper, P.D. Karp, F.C. Neidhardt, T. Nyström, J.M. Slauch, C.L.
Squires (Eds.), EcoSal — Escherichia coli and Salmonella: Cellular and Molecular
Biology, ASM Press, Washington, D. C., 2008
[11] M.G. Johnson, J.C. Escalante-Semerena, Identification of 5,6-dimethylbenzimidazole
as the Coα ligand of the cobamide synthesized by Salmonella typhimurium.
Nutritional characterization of mutants defective in biosynthesis of the imidazole
ring, J. Biol. Chem. 267 (1992) 13302–13305.
[12] B. Keck, P. Renz, Salmonella typhimurium forms adenylcobamide and 2-
methyladenylcobamide, but no detectable cobalamin during strictly anaerobic
growth, Arch. Microbiol. 173 (2000) 76–77.
[13] J.R. Trzebiatowski, J.C. Escalante-Semerena, Purification and characterization of CobT,
the nicotinate-mononucleotide:5,6-dimethylbenzimidazole phosphoribosyltransferase
enzyme from Salmonella typhimurium LT2, J. Biol. Chem. 272 (1997) 17662–17667.
[14] C.G. Cheong, J.C. Escalante-Semerena, I. Rayment, Structural investigation of the bio-
synthesis of alternative lower ligands for cobamides by nicotinate mononucleotide:
5,6-dimethylbenzimidazole phosphoribosyltransferase from Salmonella enterica,
J. Biol. Chem. 276 (2001) 37612–37620.
[15] A.W. Tsang, J.C. Escalante-Semerena, CobB, a new member of the SIR2 family of
eucaryotic regulatory proteins, is required to compensate for the lack of nicotinate
mononucleotide:5,6-dimethylbenzimidazole phosphoribosyltransferase activity in
cobT mutants during cobalamin biosynthesis in Salmonella typhimurium LT2, J. Biol.
Chem. 273 (1998) 31788–31794.
[16] C.G. Cheong, J.C. Escalante-Semerena, I. Rayment, Capture of a labile substrate
by expulsion of water molecules from the active site of nicotinate mononucleotide:
5,6-dimethylbenzimidazole phosphoribosyltransferase (CobT) from Salmonella
enterica, J. Biol. Chem. 277 (2002) 41120–41127.
The lack of substrate specificity in the CobT family of enzymes plays a
critical part in the diversity of the nucleoside base found in cobamides.
The structural studies of SeCobT variants and comparisons of SeCobT
orthologues presented here reveal that the conserved glutamate residues
play a role in orienting the substrates and stabilizing the active site. It is
less clear, however, whether the alluded glutamate residues may also
function as catalytic bases under physiological conditions. The structural
studies reported here suggest some of the changes that are necessary
to allow SeCobT to catalyze the phosphoribosylation of phenolic com-
pounds. Together with in vivo and in vitro analyses of SeCobT variants,
our results of the crystallographic analyses advance our understanding
of the mechanism of the NaMN:base phosphoribosyltransferase family
of enzymes.
The structural and functional studies in SeCobT enzymes described
here and elsewhere provide insights into why these enzymes facilitate
the diversity in cobamides. Cobamide dependent enzymes likely dictate
the chemical nature of the base. Some cobamide dependent enzymes
displace the cobamide base, but provide a histidinyl residue as an alter-
native ligand. Other enzymes rely on the existing base of the cobamide
[38]. The coordination of the cobalt atom by the α-ligand to modify its
reduction potential and charge is an important aspect in AdoCba biosyn-
thesis and reactions catalyzed through this cofactor [39–41].
In nature, organisms have evolved at least one way to remodel
the α-ligand to fit the need of their set of cobamide dependent enzymes
through an amidohydrolase known as CbiZ [42–45]. A complete analysis
of the source of the base(s), AdoCba biosynthetic and remodeling
machineries, and cobamide dependent enzymes is required to better
understand the forces driving cobamide diversity in nature.
[17] C.G. Cheong, J.C. Escalante-Semerena, I. Rayment, The three-dimensional structures
of nicotinate mononucleotide:5,6-dimethylbenzimidazole phosphoribosyltransferase
(CobT) from Salmonella typhimurium complexed with 5,6-dimethybenzimidazole
and its reaction products determined to 1.9 Å resolution, Biochemistry 38 (1999)
16125–16135.
[18] B. Möller, R. Ossmer, B.H. Howard, G. Gottschalk, H. Hippe, Sporomusa, a new genus
of Gram-negative anaerobic-bacteria including Sporomusa-Sphaeroides spec-nov
and Sporomusa-Ovata spec-nov, Arch. Microbiol. 139 (1984) 388–396.
[19] E. Stupperich, H.J. Eisinger, Biosynthesis of para-cresolyl cobamide in Sporomusa
ovata, Arch. Microbiol. 151 (1989) 372–377.
[20] C.H. Chan, J.C. Escalante-Semerena, ArsAB, a novel enzyme from Sporomusa ovata
activates phenolic bases for adenosylcobamide biosynthesis, Mol. Microbiol. 81
(2011) 952–967.
[21] S.A. Newmister, C.H. Chan, J.C. Escalante-Semerena, I. Rayment, Structural
insights into the function of the nicotinate mononucleotide:phenol/p-cresol
phosphoribosyltransferase (ArsAB) enzyme from Sporomusa ovata, Biochemistry 51
(2012) 8571–8582.
5. Accession numbers
X-ray coordinates for SeCobTE317A in the substrate free form
and in complex with α-DMB riboside monophosphate and nicotinate,
SeCobTE174A in complex with DMB and adenine, SeCobTS80Y,Q88M,L175M
in complex with p-cresol and in complex with p-cresol and nicotinate
mononucleotide have been deposited in the Research Collaboratory
for Structural Bioinformatics, Rutgers University, New Brunswick, N. J.
with Protein Data Bank entries 4KQH, 4KQI, 4KQG, 4KQF, 4KQK, and
4KQJ respectively.
Conflict of interest
[22] D. Berkowitz, J.M. Hushon, H.J. Whitfield Jr., J. Roth, B.N. Ames, Procedure for identi-
fying nonsense mutations, J. Bacteriol. 96 (1968) 215–220.
The authors have no conflict of interest to declare.

C.H. Chan et al. / Biochimica et Biophysica Acta 1840 (2014) 464–475
475
[23] W.E. Balch, R.S. Wolfe, New approach to the cultivation of methanogenic bacteria:
2-mercaptoethanesulfonic acid (HS-CoM)-dependent growth of Methanobacterium
ruminantium in a pressurized atmosphere, Appl. Environ. Microbiol. 32 (1976)
781–791.
[24] R. Atlas, Handbook of Media for Environmental Microbiology, CRC Press, Boca Raton,
1995, p. 6.
[25] G. Bertani, Studies on lysogenesis. I. The mode of phage liberation by lysogenic
Escherichia coli, J. Bacteriol. 62 (1951) 293–300.
[26] G. Bertani, Lysogeny at mid-twentieth century: P1, P2, and other experimental
systems, J. Bacteriol. 186 (2004) 595–600.
[39] T.A. Stich, M. Yamanishi, R. Banerjee, T.C. Brunold, Spectroscopic evidence for the
formation of four-coordinate Co(2+)cobalamin species upon binding to the
a
human ATP:Cobalamin adenosyltransferase, J. Am. Chem. Soc. 127 (2005) 7660–7661.
[40] T.A. Stich, N.R. Buan, J.C. Escalante-Semerena, T.C. Brunold, Spectroscopic and
computational studies of the ATP:Corrinoid adenosyltransferase (CobA) from
Salmonella enterica: insights into the mechanism of adenosylcobalamin biosynthesis,
J. Am. Chem. Soc. 127 (2005) 8710–8719.
[41] P.E. Mera, M. St Maurice, I. Rayment, J.C. Escalante-Semerena, Residue Phe112 of the
human-type corrinoid adenosyltransferase (PduO) enzyme of Lactobacillus reuteri
is critical to the formation of the four-coordinate Co(II) corrinoid substrate and to
the activity of the enzyme, Biochemistry 48 (2009) 3138–3145.
[27] B. Cormack, Directed Mutagenesis Using the Polymerase Chain Reaction, Greene
Publishing Associates & Wiley Interscience, New York, 1997.
[42] J.D. Woodson, J.C. Escalante-Semerena, CbiZ, an amidohydrolase enzyme required
for salvaging the coenzyme B12 precursor cobinamide in archaea, Proc. Natl. Acad.
Sci. U. S. A. 101 (2004) 3591–3596.
[28] A.H. Rosenberg, B.N. Lade, D.S. Chui, S.W. Lin, J.J. Dunn, F.W. Studier, Vectors for
selective expression of cloned DNAs by T7 RNA polymerase, Gene 56 (1987) 125–135.
[29] H.E. Klock, E.J. Koesema, M.W. Knuth, S.A. Lesley, Combining the polymerase
incomplete primer extension method for cloning and mutagenesis with micro-
screening to accelerate structural genomics efforts, Proteins Struct. Funct. Bioinf. 71
(2008) 982–994.
[43] J.D. Woodson, J.C. Escalante-Semerena, The cbiS gene of the archaeon Methanopyrus
kandleri AV19 encodes
a bifunctional enzyme with adenosylcobinamide
amidohydrolase and alpha-ribazole-phosphate phosphatase activities, J.
Bacteriol. 188 (2006) 4227–4235.
[30] A.J. Andreoli, M. Ikeda, Y. Nishizuka, O. Hayaishi, Quinolinic acid: a precursor to
nicotinamide adenine dinucleotide in Escherichia coli, Biochem. Biophys. Res. Commun.
12 (1963) 92–97.
[31] M. Kitagawa, T. Ara, M. Arifuzzaman, T. Ioka-Nakamichi, E. Inamoto, H. Toyonaga, H.
Mori, Complete set of ORF clones of Escherichia coli ASKA library (A complete set
of E. coli K-12 ORF archive): unique resources for biological research, DNA Res. 12
(2005) 291–299.
[32] F.M. Dickinson, P.C. Engel, The preparation of pure salt-free nicotinamide coenzymes,
Anal. Biochem. 82 (1977) 523–531.
[33] K.R. Claas, J.R. Parrish, L.A. Maggio-Hall, J.C. Escalante-Semerena, Functional analysis of
the nicotinate mononucleotide:5,6-dimethylbenzimidazole phosphoribosyltransferase
(CobT) enzyme, involved in the late steps of coenzyme B12 biosynthesis in
Salmonella enterica, J. Bacteriol. 192 (2010) 145–154.
[44] J.D. Woodson, C.L. Zayas, J.C. Escalante-Semerena, A new pathway for salvaging the
coenzyme B12 precursor cobinamide in archaea requires cobinamide-phosphate
synthase (CbiB) enzyme activity, J. Bacteriol. 185 (2003) 7193–7201.
[45] M.J. Gray, J.C. Escalante-Semerena, The cobinamide amidohydrolase (cobyric acid-
forming) CbiZ enzyme: a critical activity of the cobamide remodelling system of
Rhodobacter sphaeroides, Mol. Microbiol. 74 (2009) 1198–1210.
[46] J.R. Trzebiatowski, G.A. O'Toole, J.C. Escalante-Semerena, The cobT gene of
Salmonella typhimurium encodes the NaMN: 5,6-dimethylbenzimidazole
phosphoribosyltransferase responsible for the synthesis of N1-(5-phospho-
alpha-D-ribosyl)-5,6-dimethylbenzimidazole, an intermediate in the synthesis
of the nucleotide loop of cobalamin, J. Bacteriol. 176 (1994) 3568–3575.
[47] C.J. Rocco, K.L. Dennison, V.A. Klenchin, I. Rayment, J.C. Escalante-Semerena,
Construction and use of new cloning vectors for the rapid isolation of recombinant
proteins from Escherichia coli, Plasmid 59 (2008) 231–237.
[48] L.M. Guzman, D. Belin, M.J. Carson, J. Beckwith, Tight regulation, modulation,
and high-level expression by vectors containing the arabinose PBAD promoter,
J. Bacteriol. 177 (1995) 4121–4130.
[49] S. Tabor, Expression using the T7 RNA polymerase/promoter system, in: F.M.
Ausubel, R. Brent, R.E. Kingston, D.D. Moore, J.G. Seidman, J.A. Smith, K. Struhl
(Eds.), Current Protocols in Molecular Biology, vol. 2, Wiley Interscience, New York,
1990, (16.12.11.-16.12.11).
[34] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in oscillation
mode, Methods Enzymol. 276 (1997) 307–326.
[35] A.J. McCoy, R.W. Grosse-Kunstleve, P.D. Adams, M.D. Winn, L.C. Storoni, R.J. Read,
Phaser crystallographic software, J. Appl. Crystallogr. 40 (2007) 658–674.
[36] G.N. Murshudov, P. Skubak, A.A. Lebedev, N.S. Pannu, R.A. Steiner, R.A. Nicholls, M.D.
Winn, F. Long, A.A. Vagin, REFMAC5 for the refinement of macromolecular crystal
structures, Acta Crystallogr. D Biol. Crystallogr. 67 (2011) 3553–3567.
[37] P. Carter, J.A. Wells, Dissecting the catalytic triad of a serine protease, Nature 332
(1988) 564–568.
[50] G.A. O'Toole, M.R. Rondon, J.C. Escalante-Semerena, Analysis of mutants of defective
in the synthesis of the nucleotide loop of cobalamin, J. Bacteriol. 175 (1993)
3317–3326.
[38] R. Banerjee, S.W. Ragsdale, The many faces of vitamin B12: catalysis by
cobalamin-dependent enzymes, Annu. Rev. Biochem. 72 (2003) 209–247.