Angewandte
Communications
Chemie
The comparative functional analysis of AmbP3 and
AmbO5 provides initial evidence that the ambiguine biosyn-
thetic maturation starts with 3, proceeds preferentially with
prenylation to give 4 that is a more competent substrate for
subsequent chlorination. The ability of AmbO5 to act on both
3
and 4 with the latter as a preferred substrate indicates
AmbO5 is likely a more promiscuous aliphatic halogenase
and can process a wide range of ambiguines (Figure 2b). To
validate this hypothesis, we procured all remaining represen-
tative deschlorinated ambiguines (5, 6, 7) in each distinct
oxidation state (Figure 1b), except 8 that is not accessible
either by isolation or synthesis in our hands. When ambi-
guines 5/6/7 were subjected individually to AmbO5 under the
identical assay conditions applied to 4, the generation of
monochlorinated adducts was readily observed (Figure 3e–
g). The authenticity of each enzymatic product was validated
to be 5a, 6a and 7a, by either comparative HPLC and HRMS
analysis with the standards (for 5a and 7a) (Figures S7 and
Figure 4. HPLC-based comparative analyses of AmbO5 and WelO5
substrate scopes. Shown in y-axis for all chromatographs is the relative
absorbance at 280 nm. Inset: summary of AmbO5 and WelO5 kcat
values on substrates 1–7.
1
S8) or by H NMR and HRMS analysis of the purified
enzymatic product (for 6a) (Figure S9 and S10). To gain
insight on the relative preference of AmbO5 towards differ-
ent ambiguines, a series of competition experiments with
equimolar amounts of 4 and 5 or 6 or 7 (0.25 mm each) were
carried out with AmbO5 (20 mm) (Figure 3h). Subsequent
HPLC analysis revealed that the turnover of 4 to 4a under
these experimental conditions is nearly identical with those of
ities/conformations. Previous study on SyrB2, the founding
member of Fe/2OG dependent aliphatic halogenase, showed
a subtle chemical modification of its native substrate l-
threonine (such as to l-2-aminobutyric acid) tethered to the
CP domain of SyrB1 readily derailed the chemoselectivity of
the enzyme and led to a mixture of hydroxylated and
5
to 5a, 6 to 6a and 7 to 7a (Figure 3h), confirming that
ambiguines 4–7 are equally competent substrates for AmbO5
vide infra). These observations collectively define the overall
[3d]
chlorinated products. AmbO5 and WelO5 share identical
first sphere iron-coordinating residues (H164, G166, H259)
(Figure S1), analogous to those in CP-dependent aliphatic
(
enzymatic timing of chlorination in ambiguine biogenesis and
provide conclusive evidence that the observed structural
diversity of deschlorinated and chlorinated ambiguines is due
to the promiscuous AmbO5 halogenase that can act on the
deschlorinated intermediates, typified by ambiguines 4–7,
parallel to their biosynthetic maturation by sequential
prenylation and oxidations.
[3a]
halogenase SyrB2. Assuming they both use cis-haloferryl
reactive intermediate in their catalytic cycles for hydrogen
[3b,c]
abstraction as observed for CP-dependent halogenase,
these newly discovered halogenation enzymes are likely
equipped with novel structural elements to control the
selective chlorine rebound to the target carbon radical.
From a biocatalyst development perspective, the current
observation is equally noteworthy. The intersected nature of
AmbO5 and WelO5 in both protein sequences (Figure S1)
and substrate scopes (Figure 4h) implicates the halogenase
duo may be naturally evolved from each other or a common
ancestor via divergent evolution. This provides a logical entry
point to study the protein elements in this newly discovered
enzyme family to confer small molecular substrate tolerance
in the absence of a substrate-bound protein crystal structure.
To provide insights on the inter-evolvable nature of
AmbO5 and WelO5 on substrates 1–7, we first generated
a pair of chimeric proteins by fusing the N-terminal WelO5
(aa1-145) with C-terminal AmbO5 (aa146-290) to give
WelO5AmbO5 and vice versa to give AmbO5WelO5 (Fig-
ure 5a). Dividing both proteins at 145-6 residues was based on
a statistical consideration as the WelO5/AmbO5 protein
sequences differ by 62 aa residues (Figure S1) with 30 of those
at their N-termini (aa1-145) and 32 at their C-termini (aa146-
290). The enzymatic activities of these chimeras were initially
assessed using 5, selected from the second best group of
substrates for AmbO5 that cannot be processed by wt-WelO5
for chlorination. As shown in Figure 5b (traces 1–3),
WelO5AmbO5 was able to convert 5 to 5a, albeit at a lower
The promiscuous nature of AmbO5 towards small molec-
ular substrates prompted us to compare its substrate scope
with WelO5. We collected all seven known hapalindole-type
alkaloids (1–7) that are validated substrates for either
[
4c]
WelO5 or AmbO5 and subjected them (0.5 mm) to parallel
assay with each halogenase protein (20 mm) under identical
assay conditions. The comparative HPLC analysis revealed
WelO5 was only able to act on 1 and 2, reasserting its narrow
substrate scope (Figure 4a–g, bottom traces). Surprisingly,
AmbO5 also processed 1 and 2 as efficiently as WelO5
(
Figure 4a,b). To establish the substrate preference of
AmbO5, we measured its apparent kcat and that of WelO5
towards substrates 1–7 (Figure 4 inset). These data show that
1
4
is processed most efficiently by AmbO5, followed by 2 and
–7 being the second best group of substrates and 3 being the
least preferred substrate.
It is noteworthy that the small molecular substrate scope
of AmbO5 is significantly broadened from its close homolog
WelO5 (Figure 4h). From a mechanistic perspective, it is
unexpected that AmbO5 can maintain chemo-, regio- and
stereoselective chlorination on a range of substrates (1–7)
that are structurally distinct from each other in both func-
tional group types/densities and carbon skeleton connectiv-
5
782
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 5780 –5784