Evidence for Modulation of Oxygen Rebound Rate in Control of Outcome by Iron(II)- And 2-Oxoglutarate-Dependent Oxygenases

Article abstract of DOI:10.1021/jacs.9b06689

Iron(II)- and 2-oxoglutarate-dependent (Fe/2OG) oxygenases generate iron(IV)-oxo (ferryl) intermediates that can abstract hydrogen from aliphatic carbons (R-H). Hydroxylation proceeds by coupling of the resultant substrate radical (Ra€￠) and oxygen of the Fe(III)-OH complex ("oxygen rebound"). Nonhydroxylation outcomes result from different fates of the Fe(III)-OH/R?state; for example, halogenation results from R?coupling to a halogen ligand cis to the hydroxide. We previously suggested that halogenases control substrate-cofactor disposition to disfavor oxygen rebound and permit halogen coupling to prevail. Here, we explored the general implication that, when a ferryl intermediate can ambiguously target two substrate carbons for different outcomes, rebound to the site capable of the alternative outcome should be slower than to the adjacent, solely hydroxylated site. We evaluated this prediction for (i) the halogenase SyrB2, which exclusively hydroxylates C5 of norvaline appended to its carrier protein but can either chlorinate or hydroxylate C4 and (ii) two bifunctional enzymes that normally hydroxylate one carbon before coupling that oxygen to a second carbon (producing an oxacycle) but can, upon encountering deuterium at the first site, hydroxylate the second site instead. In all three cases, substrate hydroxylation incorporates a greater fraction of solvent-derived oxygen at the site that can also undergo the alternative outcome than at the other site, most likely reflecting an increased exchange of the initially O₂-derived oxygen ligand in the longer-lived Fe(III)-OH/R?states. Suppression of rebound may thus be generally important for nonhydroxylation outcomes by these enzymes.

Full text of DOI:10.1021/jacs.9b06689

Subscriber access provided by Columbia University Libraries
Article
Evidence for Modulation of Oxygen-Rebound Rate in Control of
Outcome by Iron(II)- and 2-Oxoglutarate-Dependent Oxygenases
Juan Pan, Eliott S. Wenger, Megan L. Matthews, Christopher J. Pollock, Minakshi Bhardwaj,
Amelia J. Kim, Benjamin D. Allen, Robert B. Grossman, Carsten Krebs, and J. Martin Bollinger, Jr.
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b06689 • Publication Date (Web): 01 Sep 2019
Downloaded from pubs.acs.org on September 1, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
Evidence for Modulation of Oxygen-Rebound Rate in Control of Outcome by
Iron(II)- and 2-Oxoglutarate-Dependent Oxygenases
7
8
9
Juan Pan,1,# Eliott S. Wenger,1,# Megan L. Matthews,1,4,# Christopher J. Pollock,1,5 Minakshi
Bhardwaj,3,6 Amelia J. Kim,2 Benjamin D. Allen,2 Robert B. Grossman,3 Carsten Krebs,1,2,* J.
Martin Bollinger, Jr.1,2,*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1Department of Chemistry and 2Department of Biochemistry and Molecular Biology, The
Pennsylvania State University, University Park, Pennsylvania 16802, United States
3Department of Chemistry, University of Kentucky, Lexington, Kentucky 40546-0312, United
States
4Present address: Department of Chemistry, University of Pennsylvania, Philadelphia,
Pennsylvania, 19104-6323, United States
5Present address: Cornell High Energy Synchrotron Source, Wilson Laboratory, Cornell
University, Ithaca, New York 14853, United States
6Present address: Department of Pharmaceutical Sciences, University of Kentucky, Lexington,
Kentucky 40546-0312, United States
#These authors contributed equally
*To whom correspondence should be addressed: email ckrebs@psu.edu, jmb21@psu.edu
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 40
1
2
3
4
ABSTRACT
5
6
7
8
Iron(II)- and 2-oxoglutarate-dependent (Fe/2OG) oxygenases generate iron(IV)-oxo (ferryl)
intermediates that can abstract hydrogen from aliphatic carbons (R–H). Hydroxylation proceeds
by coupling of the resultant substrate radical (R•) and oxygen of the Fe(III)–OH complex (“oxygen
rebound”). Non-hydroxylation outcomes result from different fates of the Fe(III)–OH/R• state; for
example, halogenation results from R• coupling to a halogen ligand cis to the hydroxide. We
previously suggested that halogenases control substrate-cofactor disposition to disfavor oxygen
rebound and permit halogen coupling to prevail. Here, we explored the general implication that,
when a ferryl intermediate can ambiguously target two substrate carbons for different outcomes,
rebound to the site capable of the alternative outcome should be slower than to the adjacent, solely
hydroxylated site. We evaluated this prediction for (i) the halogenase SyrB2, which exclusively
hydroxylates C5 of norvaline appended to its carrier protein but can either chlorinate or
hydroxylate C4 and (ii) two bifunctional enzymes that normally hydroxylate one carbon before
coupling that oxygen to a second carbon (producing an oxacycle) but can, upon encountering
deuterium at the first site, hydroxylate the second site instead. In all three cases, substrate
hydroxylation incorporates a greater fraction of solvent-derived oxygen at the site that can also
undergo the alternative outcome than at the other site, most likely reflecting increased exchange
of the initially O2-derived oxygen ligand in the longer-lived Fe(III)–OH/R• states. Suppression of
rebound may thus be generally important for non-hydroxylation outcomes by these enzymes.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
ACS Paragon Plus Environment

Page 3 of 40
Journal of the American Chemical Society
1
2
3
4
INTRODUCTION
5
6
7
8
In biology, hydrogen-atom (H•) transfer (HAT) from an aliphatic carbon to an iron(IV)-
oxo (ferryl) intermediate enables an array of primarily oxidative transformations for which
synthetic-chemical counterparts are generally underdeveloped. Formation of a carbon-centered
substrate radical (R•) can initiate stereoinversion, desaturation, C–C-bond fragmentation, ring
expansion, or coupling to a heteroatom (oxygen, sulfur, or halogen) or sp2-hybridized carbon.1-8
The most versatile subset of enzymes with this modus operandi are the iron(II)- and 2-
oxoglutarate-dependent (Fe/2OG) oxygenases.1,5,9 These enzymes share a largely conserved
tertiary structure, known variously as the cupin, -sandwich, or jelly-roll fold. They are
represented in all three domains of life and have crucial roles in, for example, nutrient acquisition,10
synthesis of connective tissue,11 regulation of gene expression and epigenetic inheritance,12-14 and
homeostasis of oxygen and body mass.15-16 In plants, fungi, and bacteria, they are highly
represented in biosynthetic pathways to bioactive natural products.1 The hydroxylases
(dioxygenases) are most prevalent, but Fe/2OG monooxygenases mediate many other reactions,
including halogenation, C–C and C–N desaturation, oxacyclization, endoperoxidation, C-C
coupling, ring expansion and stereoinversion reactions.1,9,17 A central question motivating research
on these enzymes is how such a diverse set of outcomes can emerge from a single protein
architecture. Answering this question requires delineation of the individual reaction pathways, the
points at which they diverge, and the structural and dynamical features of the individual proteins
that lead to a given pathway.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Important structural features and early reaction events are largely conserved across the
family (Scheme S1). The Fe(II) cofactor is most often facially coordinated by three protein ligands
– two His imidazoles and one Glu or Asp carboxylate. In the Fe/2OG halogenases, the Asp/Glu
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 40
1
2
3
4
5
6
7
8
9
residue is absent – Ala or Gly being present in its place – and a halide ion coordinates in place of
the protein carboxylate.18 The remaining three sites of the octahedral coordination sphere are
occupied by the co-substrate, 2OG, which chelates the Fe(II) ion via its C1 carboxylate and C2
carbonyl groups, and a water molecule.19-20 Binding of the prime substrate results in dissociation
of the water ligand, creating a square-pyramidal Fe(II) center with an open coordination site.21 The
location of this site is relevant to the central question of control because, in the simplest case, it is
the site to which O2 adds to initiate the reaction and subsequently harbors the oxo ligand of the
key ferryl complex. Published structures reveal that the C2 carbonyl group of 2OG is always trans
to the Asp/Glu ligand, but there are two alternative sites for the C1 carboxylate – trans to either
His ligand. In the more common configuration, carboxylate coordination trans to the C-terminal
His leaves the site more proximal to the substrate vacant, and this geometry has therefore been
termed “inline.”1 The less common geometry, with the 2OG C1 carboxylate trans to the N-terminal
His ligand, has been termed “offline.”1 Addition of O2 results in an intermediate that calculations
suggest is best described as a Fe(III)-superoxo complex.22 This intermediate has not been
characterized in any Fe/2OG enzyme, but structural cognates have been detected in related
enzymes.23-24 Decarboxylation of 2OG (C1–C2 cleavage) accompanied by C2–O coupling
produces an Fe(II)-peroxysuccinate complex,22 which was recently structurally characterized in
the hydroxylase, VioC.25 This intermediate undergoes O–O-bond heterolysis to generate the
succinate-coordinated ferryl complex. This sequence of steps results in incorporation of one atom
of O2 into succinate and the other into the ferryl complex, where it has been shown for the case of
the dioxygenases to undergo exchange with solvent in competition with incorporation into the
product.1,26-27 Boal and co-workers proposed for the halogenases that second-sphere interactions
can, even after inline addition of O2, steer the peroxide unit of the peroxysuccinate complex so as
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
ACS Paragon Plus Environment

Page 5 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
to locate the oxo ligand of the ferryl moiety offline.28 Regardless of the precise sequence of steps
and interactions by which the oxo might be guided to the offline position, its location there is
thought to be, at least for some cases, crucial to control of the reaction outcome, as outlined below.
Pathways leading to hydroxylation, halogenation, stereoinversion, desaturation, and
endoperoxidation have been partially mapped.25,27,29-39 Each begins with HAT from the substrate
carbon to the ferryl moiety.40 The resultant state, harboring an Fe(III)–OH form of the cofactor
and a substrate radical [Fe(III)–OH/R•], is thought to be the key branch point for the manifold of
outcomes. In hydroxylation, the R• couples with the Fe(III)-coordinated oxygen, in what Groves
called “oxygen rebound” for the case of heme-dependent oxygenases.41 In hydroxylases that have
been studied, the rebound step is sufficiently rapid to prevent accumulation of the Fe(III)–OH/R•
state.29,32,37 Its facile nature raises the question of how it is averted in reactions that require different
fates of the Fe(III)–OH/R• state. Other fates of this state that have been posited and, in some cases,
supported by experimental data include: (i) coupling of the radical to a ligand cis to the oxygen in
the halogenases18,27,42 and isopenicillin N synthase;4,24 (ii) HAT to the opposite face of the R• from
a tyrosine residue, completing stereoinversion in carbapenem synthase (CarC);36,43 (iii) -
heteroatom-assisted electron transfer (ET) from the R• to the Fe(III)-OH complex in the L-arginine
4,5-desaturase, NapI,37 and the olefin-installing decarboxylase, IsnB;44 (iv) HAT from the carbon
 to the R• to the Fe(III)-OH complex in non-native C–C desaturations catalyzed by the L-arginine
3-hydroxylase, VioC,37 and the trifunctional hydroxylase/oxacyclase/desaturase, clavaminate
synthase;45 (v) capture of O2 by the R• in verruculogen synthase (FtmOx1);39,46 (vi) C–C-bond
formation by addition of the R• to a sp2-hybridized carbon of the substrate in the cyclases DabC
and 2-ODD;2-3 and (vii) sigma-bond rearrangement leading to ring expansion in
deacetoxycephalosporin-C synthase (DAOCS).47 In several cases, specific adaptations disfavoring
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 40
1
2
3
4
5
6
oxygen rebound have been proposed, but the analysis of enzyme control at this crucial branchpoint
is most advanced for the case of the halogenases.
7
8
9
Studies on the halogenase SyrB2 suggested that it controls the disposition of the substrate
relative to the cis-chloroferryl complex to make coupling of the R• to the chlorine more facile than
oxygen rebound.27,35,48 The initial evidence involved trends in reactivity.27 HAT from the donor in
the native substrate, the C4 methyl group of L-threonine presented by the carrier protein SyrB1
(Thr-S-SyrB1), to the cis-chloroferryl complex was found to be surprisingly slow,34 with a rate
constant 100-1000-fold less than those for the corresponding steps in the reactions of two
previously studied hydroxylases.31-32 The HAT steps in reactions with two non-native SyrB1
substrates appended by either L-aminobutyrate (Aba-S-SyrB1, replacing the side-chain hydroxyl
of Thr-S-SyrB1 by hydrogen) or L-norvaline (Nva-S-SyrB1, extending the side chain of Aba-S-
SyrB1 by a methylene unit) were faster by factors of 13 and 130 (respectively) than with Thr-S-
SyrB1, and the outcome changed across this substrate series from mostly (> 95%) chlorination of
Thr-S-SyrB1, to mixed chlorination and hydroxylation of Aba-S-SyrB, to mostly (~ 90%)
hydroxylation of Nva-S-SyrB1.27,34 The trends revealed a programmed inefficiency in the HAT
step, which, it was posited, results from a rigidly enforced disposition of the native substrate and
cis-chloroferryl complex that is suboptimal for both HAT and subsequent rebound and thereby
enables coupling of the C4 radical to the cis chlorine to prevail. By relaxing or perturbing this rigid
disposition, the substrate modifications both unleash the inherent HAT potency of the cis-
chloroferryl complex and disable suppression of the rebound step.27 This view has been supported
by a number of subsequent experimental and computational studies.28,35,49-50 Notably, these studies
have suggested that the active-site configuration enforcing slow HAT/rebound to enable chlorine
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
ACS Paragon Plus Environment

Page 7 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
coupling is achieved not by relocation of the substrate binding site within the conserved -
sandwich architecture but rather by use of a cis-chloroferryl complex with offline oxo.28,35,49
Additional evidence for the crucial role of substrate-intermediate disposition came from a
deeper kinetic analysis of the SyrB2 reaction with site-specifically deuterium-labeled isotopologs
of Nva-S-SyrB1.27 The analysis revealed that the same chloroferryl complex (or complexes in
rapid equilibrium) can abstract H• from either C4 or C5 (Scheme 1). HAT from C5 (upper path)
is preferred by a factor of ~ 6 for both the unlabeled substrate and the substrate with deuterium at
both sites (4,4,5,5,5-d5-Nva-S-SyrB1), but, with deuterium only at C5 (5,5,5-d3-Nva-S-SyrB1), a
large primary deuterium kinetic isotope effect (D-KIE) of ~ 60 reverses the regioselectivity to
favor HAT from C4 (lower two paths) by a factor of ~ 10. HAT from C5 is invariably followed by
oxygen rebound; hydroxylation is the exclusive outcome. By contrast, the slower HAT from C4
results primarily in chlorination (favored by ~ 5:1). This correlation of HAT rate with reaction
outcome – even for adjacent carbons of the same substrate – is among the strongest evidence that
substrate-intermediate disposition controls the fate of the cis-Cl–Fe(III)–OH/R• branch-point
intermediate.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Published Kinetic Dissection of the Reaction of SyrB2 with Nva-S-SyrB1.27
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 40
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ambiguous HAT from multiple substrate carbons and redirection by site-specific
deuteration were seen before (e.g., in isopenicillin N synthase4) and have been observed since (e.g.,
in the L-arginine 4,5-desaturase, NapI37) in reactions of other mononuclear non-heme-iron
enzymes. The behavior becomes evident in comparison of the kinetics of the ferryl intermediate
with varied placement of deuterium in the substrate. The hallmark is that deuterium at only the
preferred HAT site slows decay by less than the true, intrinsic D-KIE for the HAT step. For
example, decay of the chloroferryl complex in SyrB2 with 5,5,5-d3-Nva-S-SyrB1 is slower by only
a factor of ~ 6 than with the unlabeled substrate. The reason is that the observed rate constant for
ferryl decay in the latter case (kobs ~ 9 s-1) reflects the sum of the elementary rate constants for
HAT from C5 (k5H ~ 8 s-1) and C4 (k4H ~ 1.4 s-1), and deuteration of only C5 diminishes only the
former contribution (by the intrinsic D-KIE of ~ 60, giving k5D ~ 0.13 s-1), which results in kobs
~
8
ACS Paragon Plus Environment

Page 9 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
1.5 s-1 and an apparent D-KIE of ~ 6 (Scheme 1). By contrast, substitution of both positions also
diminishes the contribution of HAT from C4 (k4D ~ 0.03 s-1), yielding kobs ~ 0.16 s-1 and an
observed D-KIE of ~ 60. Thus, an obvious manifestation of ambiguous targeting and redirection
by deuterium at the preferred site is additional kinetic stabilization of the ferryl complex upon
deuteration also of the secondary site.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Analogous Sequences of Hydroxylation and Oxacyclization Reactions Catalyzed by
LolO (A) and H6H (B)
Members of a small subset of Fe/2OG oxygenases catalyze multiple reaction types in
sequence. For example, fungal N-acetylnorloline (NANL) synthase (LolO) converts 1-exo-
acetamidopyrrolizidine (AcAP) to its compact, tricyclic, insecticidal/anti-feedant alkaloid product
by sequential hydroxylation and oxolane-forming cyclization (generally, oxacyclization) steps,
each requiring an equivalent of 2OG and O2 (Scheme 2A).51-52 Analogously, hyoscyamine (Hyo)
6-hydroxylase (H6H) from plants converts its substrate to the tropane alkaloid anesthetic drug,
scopolamine (Sco), by sequential hydroxylation and epoxidation reactions (Scheme 2B).53-56
These enzymes are outstanding test cases for understanding control of outcome, because each must,
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 40
1
2
3
4
5
6
7
8
9
in sequential steps, target different sites of the substrate for HAT and also first allow and then avert
the low-barrier oxygen-rebound step. With respect to switching the site of H• abstraction, one
might envisage that geometrically identical ferryl complexes would, by virtue of appropriate
substrate disposition, be intrinsically competent to accept H• from either site, but with a significant
bias toward the carbon that undergoes hydroxylation in the first step. In this scenario, installation
of the oxygen in the first step would preclude HAT from that site in the second step (presumably,
the other initially diastereotopic hydrogen on this tetrahedral carbon would be improperly disposed
for HAT), causing the ferryl complex to default to the second site. One would then anticipate that
site-specific deuteration of the preferred site could, as in SyrB2 with 5,5,5-d3-Nva-S-SyrB1,
redirect the ferryl complex to the disfavored site. In addition, HAT from the second site in the
oxacyclization step should be slower than HAT from the first site in the hydroxylation step. In an
alternative scenario, realignment of the substrate in the active site or formation of an alternatively
configured cofactor could more actively redirect the second ferryl complex to the second site. With
respect to averting oxygen rebound in the second step, it is not obvious, a priori, whether the
enzymes might retard rebound, actively accelerate ring closure, or a combination of the two.
In this work, we show for both LolO and H6H that the presence of deuterium at the site
normally targeted for hydroxylation does indeed redirect the ferryl complex to the site normally
targeted in the second (oxacyclization) step, resulting in its hydroxylation, and we resolve rate
constants for the ambiguous HAT steps. Further, we present evidence for this pair of bifunctional
hydroxylase/oxacyclase enzymes that oxygen rebound to the site that can also undergo the
oxacyclization outcome is retarded sufficiently that the initially O2-derived oxygen ligand of the
Fe(III)–OH/R• complex partially exchanges with solvent. Finally, we return to the previously
studied case of SyrB2 with Nva-S-SyrB1 to obtain evidence that this same phenomenon – oxygen
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
10
ACS Paragon Plus Environment

Page 11 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
exchange in the Fe(III)–OH/R• rebound state that differentially impacts the site that can also
undergo the alternative outcome – is also seen in the halogenase. The conclusions imply that
structurally programmed suppression of rebound may be generally important in control of outcome
in the Fe/2OG-oxygenase family.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS
Evidence for modulation of rebound in the hydroxylase/oxacyclase, LolO. A recent study
showed that LolO from endophytic fungi of the genus Epichloë catalyzes hydroxylation of AcAP
at C2 followed by coupling of the nascent oxygen to C7 (Scheme 2A).52 Decay of the ferryl
intermediate in the hydroxylation reaction was markedly slowed by deuterium at C2 but not by
deuterium at C7. A comparison of the kinetics of this ferryl complex in reactions with substrates
bearing deuterium only at C2 (2,2,8-d3-AcAP) or at both C2 and C7 (2,2,7,7,8-d5-AcAP) (Figure
1) exhibits the hallmark of ambiguous ferryl targeting: the large D-KIE observed upon labeling
the primary HAT site, C2, is potentiated by deuteration also of the secondary site, C7 (compare
black and green traces). A global simulation of traces from the reactions with all four deuterium
isotopologs (solid lines), including 7,7-d2-AcAP (red trace), allowed extraction of the rate
constants for decay (kobs) shown in Table 1. For each substrate, kobs is the sum of the rate constants
for H(D)AT from C2 and C7 (kobs = k2 + k7). These values of kobs kinetically resolve ferryl decay
into two parallel pathways (Scheme 3) and enable prediction of the product distribution for each
substrate. The ~ 70-fold preference for C2 over C7 (with protium at both sites) is much greater
than that for C5 over C4 in the SyrB2/Nva-S-SyrB1 reaction; this fact explains the more modest
additional stabilization conferred by deuterium at the secondary site in the case of LolO. Scheme
3 predicts that the ferryl complex targets C7 of 2,2,8-d3-AcAP in ~ 35% of events [k7H/(k7H + k2D)
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 40
1
2
3
4
5
6
7
8
9
 100%], producing the C7-hydroxylated compound (7-OH-AcAP), with a change in mass-to-
charge ratio (m/z) of +16 (removal of protium), along with 65% [k2D/(k7H + k2D)  100%] of the
authentic pathway intermediate, 2-endo-hydroxy-AcAP (2-OH-AcAP), with m/z = +15 (removal
of deuterium). Analysis of the reaction products by LCMS confirmed this prediction (Figure 2A
and Figure S1A). The two products were separated chromatographically: the 2-OH-AcAP
pathway intermediate elutes first at 1920 min, and the 7-OH-AcAP compound elutes later at
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2122 min. The peak of the C7-modified product, barely visible in the chromatogram of the
reaction with the unlabeled substrate, is markedly enhanced by deuterium at C2 (2,2,8-d3-AcAP)
and again suppressed when C7 is also labeled (2,2,7,7,8-d5-AcAP). This trend in peak intensities
and the m/z values associated with this product (+16 with 2,2,8-d3- and unlabeled AcAP, +15
with 2,2,7,7,8-d5-AcAP) confirm that it is 7-OH-AcAP. In reactions with 2,2,8-d3-AcAP in which
multiple equivalents of 2OG were supplied, the early-eluting species assigned as 2-OH-AcAP
could be converted to the appropriately deuterium-labeled (2,8-d2) NANL product, with m/z =
+13 relative to 2,2,8-d3-AcAP, whereas the later-eluting species assigned as 7-OH-AcAP was not
further processed (Figure S1B). Thus, redirection of the ferryl complex to C7 by site-specific
deuteration of C2 partially subverts the pathway, affording 7-OH-AcAP as a dead-end product in
~ 35%-yield.
2,2,7,7,8-d5
0.2
2,2,8-d3
∆A320
0.1
7,7-d2
AcAP
0
0.01
0.1
1
Time (s)
12
ACS Paragon Plus Environment

Page 13 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Figure 1. Kinetics of the ferryl complex, monitored by its absorbance at 320 nm, in the LolO
reactions with AcAP deuterium isotopologs. An anoxic solution containing 1.1 mM LolO, 0.8 mM
FeII, 5.0 mM 2OG, and 4.0 mM of the synthetic AcAP (red), 7,7-d2-AcAP (blue), 2,2,8-d3-AcAP
(black), or 2,2,7,7,8-d5-AcAP (green) substrate in 50 mM sodium HEPES buffer (pH 8) was mixed
at 5 °C with an equal volume of the air-saturated buffer (giving  0.19 mM O2). A global simulation
of the data, generated by assuming a three-state kinetic model with two irreversible steps, is shown
as solid lines. Details of the analysis are provided in the Supporting Information. The simulation
yielded a rate constant of  1.2  105 M-1 s-1 for addition of O2 to form the ferryl complex and the
rate constants for decay of the ferryl complex (kobs) by H(D)AT listed in Table 1. Note that all
synthetic AcAP deuterium isotopologs are racemic; the concentrations provided are for the mixture.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Rate Constants for Decay of the Ferryl Intermediate by H(D)AT in Hydroxylation of
AcAP Deuterium Isotopologs by LolO
Substrate
AcAP
kobs (s-1)
48
k2H (s-1)
k2D (s-1)
k7H (s-1)
k7D (s-1)a
47
47
-
-
0.7
-
-
7,7-d2-AcAP
2,2,8-d3-AcAP
2,2,7,7,8-d5-AcAP
47
-
~ 0.03
-
2.0
1.3
1.3
0.7
-
1.3
-
~ 0.03
aThe modest value of k7D precluded its accurate determination.
Scheme 3. Kinetic Resolution of Two Possible Hydroxylations of AcAP by LolO
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 40
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
18
In the previous study,52 use of O2 in the LolO reaction resulted in incorporation of O
into the 2-OH-AcAP intermediate and ultimately the NANL product. With the above kinetic and
LC–MS analyses revealing a small quantity of 7-OH-AcAP in the reaction with unlabeled AcAP
and ~ 35% of the dead-end product in the reaction with 2,2,8-d3-AcAP, we tested for differential
18O incorporation into C2 and C7. With the unlabeled substrate, there is nearly 100% incorporation
18
during hydroxylation of C2 (Figure S2): the fraction of O in the 2-OH-AcAP (93 ± 1%) was
found to be the same, within experimental error, as that in the succinate co-product (94 ± 3%),
which obtains its oxygen atom from O2 in 100% of events (with no possibility of solvent exchange)
and thus provides a means to quantify contaminating atmospheric 16O2 in the 18O2 reaction.
Although the modest yield of 7-OH-AcAP produced in this reaction (< 2%) resulted in
considerable uncertainty in its measured isotopic composition, the consistently lower fraction of
14
ACS Paragon Plus Environment

Page 15 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
18O (84 ± 2%) implies 9 ± 2% solvent exchange (“washout”) during C7 hydroxylation (Figure
S2). With 2,2,8-d3-AcAP, the lifetime of the ferryl state is enhanced by a factor of ~ 25, and,
consequently, substantial washout was seen for both 2-OH-AcAP and 7-OH-AcAP products
(Figure 2B). In this case, the isotopic compositions of both products could be precisely determined
as a result of their more similar yields. Washout of the O2-derived oxygen during hydroxylation
of C7 (47.6 ± 0.4 %; gold and orange traces) again exceeded that during hydroxylation of C2
(40.6 ± 0.2%; dark and light blue traces) by ~ 7% (Table 2). Because this difference is modest,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
we verified the conclusion by also carrying out the reactions with 16O2 in H2 O (Figure 2C and
18
Table 3). As expected, reactions carried out in 77% H2 O showed greater 18O incorporation into
the 7-OH-AcAP product (gold and orange traces) than into the 2-OH-AcAP intermediate (dark
16
and light blue traces), and the values mirrored those from the 18O2/H2 O experiment. The extents
of exchange were consistent for all 2OG concentrations tested, thus verifying that the differential
washout is an intrinsic manifestation of the reaction mechanism and independent of experimental
reactant stoichiometries.
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 40
1
2
3
4
5
6
7
8
9
Figure 2. LCMS characterization of the hydroxylation products in LolO reactions with AcAP
deuterium isotopologs and determination of the origin of the installed oxygen atom. (A) Changes
in mass associated with hydroxylation of the relevant isotopologs at C2 and C7 (left) and single-
ion chromatograms (positive-ion mode) from the reactions (right). The trends in peak intensities
and mass shift with 2,2,8-d3-AcAP identify the first product to elute as 2-OH-AcAP and the second
product to elute as 7-OH-AcAP. (B and C) Analysis by LCMS of the differential incorporation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
of 18O from 18O2 (97% isotopic purity) or H2 O (77% isotopic purity), respectively, into the C2-
and C7-hydroxylation products of 2,2,8-d3-AcAP under single-turnover conditions. Positive-mode
single-ion chromatograms at the m/z values corresponding to hydroxylation with loss of deuterium
and incorporation of 18O (dark blue) or 16O (light blue) and loss of protium with incorporation of
18O (orange) or 16O (gold) are shown. Reactions were carried out with limiting 2OG (00.8 equiv.
[LolO•Fe]) and excess 2,2,8-d3-AcAP to preclude the second (oxacyclization) step. The red values
beside the traces are the fraction of 18O-incorporated in the 2-OH-AcAP and 7-OH-AcAP products,
with standard deviations of at least eight trials (trials with varying [2OG] were considered
replicates, because this variable had no systematic effect on isotopic composition) given in
parentheses. Origins of the smaller peaks are discussed in the Supporting Information.
Table 2. Differential 18O Incorporation at the Preferred and Alternative Positions in the LolO,
H6H, and SyrB2 Reactions in H2 O and 2H2 O Solvent under 18O2 Gas
16
16
Preferred Site
Solvent Exchange (%)
Alternative Site
Solvent Exchange (%)
Number of
Trials
Enzyme
Substrate
H2O
2H2O
H2O
2H2O
LolO
LolO
H6H
AcAP
~ 0a
NDb
ND
8
16
3
9  2
2,2,8-d3-
AcAP
40.6  0.2
12.5  0.2
28.8  0.3
8.1  0.3
47.6  0.4
66.0  0.7
32.8  0.1
53.8  1.7
Hyo
H6H
6-d1-Hyo
3
32.9  0.4
24.6  1.1
85.2  0.3
73.7  0.3
16
ACS Paragon Plus Environment

Page 17 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
SyrB2
SyrB2
Nva
20
ND
ND
50
ND
ND
1
2
5,5,5-d3-
58 (58c)
69 (63c)
Nva
8
9
aIncorporation of 18O in hydroxylation was not significantly different from incorporation into
succinate, preventing its accurate determination
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
bValue not determined.
18
cValue when base hydrolysis was performed in 45% H2 O
Table 3. Differential 18O Incorporation at the C2 and C7 Positions in the LolO Reaction with
18
2,2,8-d3-AcAP in H2 O Solvent under 16O2 Gasa
% C2 Solvent Exchange
% C7 Solvent Exchange
18
48% H2 O
31.5  1.4
38.3  1.2
43.5  0.4
18
77% H2 O
35.3  0.3
18
aThe concentration of H2 O given is the final concentration in the reaction. The observed
fractional exchange has been corrected by this H2 O isotopic purity to reflect the extent of
exchange expected for 100% H2 O.
18
18
Evidence for modulation of rebound in the hydroxylase/oxacyclase, H6H. The plant
Fe/2OG hydroxylase/oxacyclase, H6H, first hydroxylates C6 of Hyo and then couples the oxygen
of the C6-hydroxyl group to C7 to produce the epoxide (oxirane) moiety of Sco (Scheme 2B). As
for LolO, the presence of deuterium at C6 can redirect the hydroxylating ferryl complex to C7,55
as confirmed by comparison of the kinetics of the ferryl state with unlabeled Hyo, 6-exo-d1-
hyoscyamine (6-d1-Hyo) and 6,7-exo-d2-hyoscyamine (6,7-d2-Hyo) (Figure 3). Again, the
significant stabilizing effect of deuterium also at the second site implies that both sites can donate
H• to the same ferryl complex. Global simulation of these three traces gave the kobs values for
decay given in Table 4. Again, these values are the sums of the elementary rate constants for
H(D)AT from C6 and C7 (kobs = k6 + k7). These values again kinetically resolve ferryl decay into
two parallel pathways (Scheme 4). The preference for C6 hydroxylation implied by these values
(~ 30-fold) is similar to that for the case of LolO. The scheme also predicts that C7 should be
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 40
1
2
3
4
5
6
7
8
9
targeted in 3% of events with the unlabeled substrate and 28% of events with the 6-d1-Hyo. Indeed,
LCMS analysis of the former reaction revealed, in addition to the major peak with m/z = +16 at
the elution time of synthetic 6-hydroxyhyoscyamine (6-OH-Hyo), a minor (1.5 ± 0.2 %) peak
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
with m/z = +16 at the elution time of 7-hydroxyhyoscyamine (7-OH-Hyo) (Figure 4A and
Table 5). Also as predicted, the yield of the 7-OH-Hyo was markedly enhanced in the reaction
with 6-d1-Hyo but again suppressed with the substrate bearing deuterium at both sites (6,7-d2-Hyo).
In the latter case, the m/z of the minor product shifted to +15, confirming that it formed by
hydroxylation after removal of deuterium (from C7).
0.16
6,7-d2-Hyo
6-d1-Hyo
A320
Hyo
0.08
0
0.01
0.1
1
10
Time (s)
Figure 3. Kinetics of the ferryl complex, monitored by its absorbance at 320 nm, in the H6H
reactions with deuterium isotopologs of Hyo. An anoxic solution containing 1.5 mM H6H, 1.2
mM FeII, 5.0 mM 2OG, and 3.0 mM of either Hyo (red), 6-d1-Hyo (blue), or 6,7-d2-Hyo (green)
in 20 mM Tris-HCl, 80 mM KCl, and 10% glycerol (pH 7.5) was mixed at 5 C with an equal
volume of 50% O2-saturated buffer ( 0.65 mM O2). A global simulation of the data, generated by
assuming a three-state kinetic model with two irreversible steps, is shown as solid lines (see details
in the Supporting Information). The simulation yielded a rate constant of  1.1  103 M-1 s-1 for
addition of O2 to form the ferryl complex and the rate constants for decay of the ferryl complex
(kobs) by H(D)AT in Table 4.
Table 4. Rate Constants for Decay of the Ferryl Intermediate H(D)AT in Hydroxylation of the
Hyoscyamine Isotopologs by H6H
18
ACS Paragon Plus Environment

Page 19 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
6H (s-1)
k
6D (s-1)
k
7H (s-1)
0.5
0.5
-
k
7D (s-1)
Substrate
Hyo
kobs (s-1)
15
k
14
-
-
-
-
6-d1-Hyo
6,7-d2-Hyo
1.8
1.3
1.3
1.3
-
~ 0.01a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aThe modest value of k7D precluded its accurate determination.
Scheme 4. Kinetic Resolution of Two Possible Hydroxylations of Hyoscyamine by H6H
Table 5. Ratio of 6-OH-Hyo and 7-OH-Hyo formed by H6H under Single Turnover Conditions
with Different Deuterium Isotopologsa
Substrate
% 6-OH-Hyo
% 7-OH-Hyo
Hyo
98.5  0.2
68.6  0.4
98.2  0.2
1.5  0.2
31.4  0.4
1.8  0.2
6-d1-Hyo
6,7-d2-Hyo
aValues were determined as the percentile of the area of each peak to the total area of the two
peaks. Each value is an average of three trials.
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 40
1
2
3
4
18
As with LolO, analysis of reactions under O2 confirmed that washout is greater for the
5
6
7
8
site (C7) that undergoes the oxacyclization outcome in the second step than for the site (C6) that
is programmed for hydroxylation in the first step (Figure 4B). For H6H, the difference is more
pronounced: in the reaction with the unlabeled substrate, washout of the oxygen incorporated at
C6 was modest (12 ± 0.18%), whereas washout of the oxygen incorporated at C7 was 67 ± 0.7%.
In the reaction with 6-d1-Hyo, the ~ 8-fold increased lifetime of the ferryl complex led to more
washout in both products – 32.9 ± 0.4% and 85.2 ± 0.3% in the C6 and C7 products, respectively
– but the relative site differential was preserved (Table 2).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. LCMS characterization of the hydroxylation products in H6H reactions with Hyo
deuterium isotopologs and determination of the origin of the installed oxygen atom. (A) Single-
ion chromatograms resolving the 6-OH-Hyo and 7-OH-Hyo products. The m/z +15 (front) and
m/z +16 (back) traces are shown for each substrate, showing hydroxylation after loss of deuterium
or hydrogen, respectively. Reactions were carried out with limiting 2OG to minimize the
oxacyclization reaction. (B) Single-ion chromatograms showing differential incorporation of 18O
18
from O2 into the 6-OH-Hyo and 7-OH-Hyo products from Hyo (front) and 6-d1-Hyo (back).
Positive-mode single-ion chromatograms depicts hydroxylation products shown here as the m/z
change (m/z) from the substrate. For Hyo: 6-18OH and 7-18OH, +18 (black); 6-16OH and 7-16OH,
+16 (gray). For 6-d1-Hyo: 7-18OH, +18 (dark blue, left); 7-16OH, +16, (light blue, left); 6-18OH,
20
ACS Paragon Plus Environment

Page 21 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
+17 (dark purple, right); and 6-16OH, +15 (light purple, right). The red values are the 18O ratios
of each product, with standard deviations from at least three trials given in parentheses. The inset
is an expanded view of the boxed portion of the trace for Hyo to illustrate resolution of the 7-OH-
Hyo (leftmost gray peak) product from the Sco product (rightmost gray peak) of the second
(oxacyclization) reaction. Because undeuterated 7-16OH-Hyo and 18O-Sco have the same mass, the
peak area for the former product was determined by curve fitting analysis. Origins of the peaks at
m/z = +18 and +16 for 6-OH-Hyo and m/z = +17 for 7-OH-Hyo with the 6-d1-Hyo substrate in
Figure 4B are discussed in the Supporting Information.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Evidence that oxygen rebound is also slow in mixed chlorination and hydroxylation by SyrB2. In
the previous study of the SyrB2 reaction with Nva-S-SyrB1, the C4- and C5-hydroxylated products
were not chromatographically resolved, preventing assessment of the possibility that solvent
exchange occurs to different extents. To assess this possibility, we developed an improved
analytical method, in which the amino acid products were cleaved off the carrier protein by a brief
(5 min) treatment with 0.125 M potassium hydroxide (KOH), separated by hydrophilic interaction
chromatography, and detected via a prominent voltage-induced fragmentation (neutral loss of
formic acid from the parent amino acid cation) in the mass spectrometer (Figure 5A). The reaction
with the unlabeled substrate (black) yielded a major peak for 5-hydroxy-Nva (5-OH-Nva) eluting
at ~ 4.5 min and a minor peak for 4-OH-Nva eluting at ~ 3.5 min. As expected, both reflected loss
of formic acid from a parent ion with m/z = +16 relative to that of Nva, corresponding to
hydroxylation with removal of protium. Use of Nva-S-SyrB1 with deuterium at C4 (4,4-d2-Nva-
S-SyrB1) suppressed the earlier peak reflecting 4-hydroxylation (red) and enhanced the later peak
reflecting 5-hydroxylation (purple), owing to the D-KIE on HAT from C4. The m/z values of the
parent ions (see Table S1) and fragment ions (shown in Figure 5A) for the peaks established that
the enhanced product formed by abstraction of protium (from C5), giving m/z = +16 for the
hydroxylation reaction, whereas the suppressed product resulted from abstraction of deuterium
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 40
1
2
3
4
5
6
7
8
9
(from C4), giving m/z = +15. Use of 5,5,5-d3-Nva-S-SyrB1 elicited the opposite perturbation:
enhancement of the earlier peak for 4-OH-Nva (green) with m/z of +16, reflecting hydroxylation
with removal of protium from C4, and suppression of the later peak for 5-OH-Nva (orange), with
m/z of +15, reflecting hydroxylation with removal of deuterium from C5. Use of 4,4,5,5,5-d5-
Nva-S-SyrB1 gave a 4-OH-Nva:5-OH-Nva ratio essentially equivalent to that for the unlabeled
substrate (blue), and, as expected, both products had m/z = +15, corresponding to hydroxylation
with removal of deuterium.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
22
ACS Paragon Plus Environment

Page 23 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Figure 5. LCMS characterization of the hydroxylation products in SyrB2 reactions with Nva-S-
SyrB1 deuterium isotopologs and determination of the origin of the installed oxygen atom. (A)
Elution profiles monitoring the daughter ions that arise from the 4-OH-Nva and 5-OH-Nva
products (4-Cl-Nva product not shown) obtained in reactions of Nva-S-SyrB1 with deuterium
isotopologs. Color-coded daughter-ion structures with m/z values that are associated with each
elution profile are shown (top). Comparison of the m/z of the daughter ions from deuterated
substrates to that from Nva (m/z = 88.1) permitted 4-OH-Nva and 5-OH-Nva to be distinguished
both by elution time and by mass (from the number of retained deuterons). The m/z values for all
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
parent cation  iminium daughter ion transitions are provided in Table S1. (B) Differential O
incorporation at the C4 and C5 positions in SyrB2 reactions with Nva-S-SyrB1 and 5,5,5-d3-Nva-
S-SyrB1. Chromatograms monitor the products of hydroxylation at C4 (~ 3.5 min) and C5 (~ 4.5
min). The traces are color-coded and labeled according to the Δm/z relative to the substrate. The
red values are the fractions of the 18O-incorporated into each product.
The previous study showed that, for the reaction with 5,5,5-d3-Nva-S-SyrB1, the most
abundant product is 4-chloro-Nva-S-SyrB1.27 For this reason, the quantities of 4-OH-Nva and 5-
OH-Nva products reflected in the corresponding traces in Figure 5A are similar, even though C4
is the preferred HAT site with the 5,5,5-d3-Nva-S-SyrB1 substrate. Because the 4-Cl-Nva product
is predominant (~ 5:1 over 4-OH-Nva27), it was important to consider its stability to the brief base
treatment used to cleave the thioester linkage to SyrB1. The reason is that, in the 18O-tracer
experiments, any hydrolysis of the 4-Cl-Nva product would artificially inflate the quantity of 4-
OH-Nva having its alcohol oxygen derived from solvent rather than O2, making this fraction of
the product appear greater than that produced in actual hydroxylation of C4 by SyrB2. By an
extended incubation of the SyrB2/5,5,5-d3-Nva-S-SyrB1 reaction products in 0.25 M KOH, a base
concentration twice that used in the actual product analysis, the rate constant for hydrolysis of the
4-Cl-Nva product was determined to be 0.6 ± 0.4 s-1 (Figure S3A). This result suggests that the 5-
min base treatment used in the product analysis should hydrolyze ≤ 4 % of the 4-Cl-Nva primary
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 40
1
2
3
4
5
6
7
8
9
product, given the 2-fold lesser [KOH] used in the actual analysis and the expectation that
hydrolysis should be kinetically first-order in [OH–]. With the previously determined
chlorination:hydroxylation ratio of ~ 5,27 hydrolysis would thus diminish the apparent 18O
incorporation at C4 by, at most, a factor of 1.2 {1/[1+(0.04•5)]} from the actual fraction
incorporated by authentic C4 hydroxylation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The previous study reported that, as for native Fe/2OG dioxygenases, hydroxylation of
Nva-S-SyrB1 by SyrB2 resulted in incorporation of 18O when the reaction was carried out under
an 18O2 atmosphere. The OH-Nva product was found to have 67% 18O and 33% 16O, which was
attributed to partial exchange of the initially O2-derived oxygen ligand with solvent in the ferryl
intermediate state. This conclusion was validated by results with 4,4,5,5,5-d5-Nva-S-SyrB1: the
markedly (~ 60-fold) enhanced lifetime of the ferryl complex with deuterium at both H•-donor
sites resulted in 90% washout and incorporation of only ~ 10% 18O. Because the 4-OH-Nva and
5-OH-Nva products were not distinguished in that study, the reported values were weighted
averages for the two products. Analysis of the products from SyrB2 reactions with the Nva-S-
SyrB1 and 5,5,5-d3-Nva-S-SyrB1 substrates under 18O2 by the new method revealed that the
fractions of 18O incorporated at C4 and C5 are not equivalent (Figure 5B). In the reaction of the
unlabeled substrate, the 5-OH-Nva product contained ~ 80% 18O, whereas the 4-OH-Nva
contained only ~ 50% 18O. These results are consistent with the site-undifferentiated value of 67%
18O reported in the prior study. In the reaction with 5,5,5-d3-Nva-S-SyrB1, the ~ 6-fold greater
lifetime of the ferryl complex allowed for greater washout in this state, resulting in less 18O
incorporation at both sites. Nevertheless, the 5-OH-Nva product still had a greater fraction of 18O
than the 4-OH-Nva (~ 42 % compared to ~ 31 %). The ≤ 1.2-fold diminution in the apparent
fraction of 18O incorporated at C4 potentially caused by hydrolysis of the 4-Cl-Nva product in the
24
ACS Paragon Plus Environment

Page 25 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
workup (as analyzed above) cannot fully account for the magnitude of this difference, consistent
with the conclusion that solvent exchange also occurs in the C4-rebound state. Nevertheless, we
carried out one additional control experiment in an attempt to rule out a confounding effect of 4-
Cl-Nva hydrolysis on the measured O-isotope distribution of the 4-OH-Nva product. We used
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18
16
0.125 M KOH quench solution prepared in 45:55 H2 O:H2 O to liberate the products from the
carrier protein (Figure S3B). In this workup, any hydrolysis of the chlorinated product to 4-OH-
Nva would increase the fraction of 18O in the latter product relative to the 42 % in the 5-OH-Nva
product. The fraction of 18O did increase, implying that some hydrolysis of the chlorinated product
did indeed occur, but it was still significantly less (37 %) than that in the 5-OH-Nva product, which
remained at 42%. The results of this control experiment thus provide both reason for caution in
interpreting the apparent differential washout in the case of SyrB2 and evidence that it most likely
is, as established more definitively for the hydroxylase/oxacyclase enzymes, an authentic
reflection of washout in the Fe(III)–OH/R• state.
General Analysis of Solvent Exchange Rates and Mechanism(s) in the Three Enzymes. In
general, branched reactions exhibit kinetic behavior that can (in our experience) confound the
intuition of even an expert chemist. For this reason, and because the analysis of such reactions is
at the core of this study, a brief, general tutorial is provided in the Supporting Information. Because
incorporation of deuterium upon the H•-donating carbon(s) increases the lifetime of the ferryl
complex, it necessarily leads to more exchange of the initially O2-derived oxygen with solvent in
that state, as reflected by the diminished fractions of 18O from 18O2 incorporated in the product(s)
(Table 2). By contrast, substrate deuteration should not drastically impact the extent of washout
that occurs in the Fe(III)–OH/R• state, because it should not significantly retard oxygen rebound
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 40
1
2
3
4
5
6
7
8
9
(or halogen coupling in SyrB2). Quantitative analysis of exchange occurring with different
substrate isotopologs thus provides an independent check of the conclusion that exchange can
indeed occur in the rebound state. For example, under the assumption that all exchange occurs in
the ferryl state, the ~ 80% 18Oincorporation in the 5-OH-Nva product of SyrB2 with the unlabeled
substrate would imply a ratio of HAT:exchange rates of 4:1 [4/(1+4) = 0.80], for an exchange rate
constant (kex) of ~ 2 s-1 at 5 °C (given that k5H is 8 s-1). The ~ 6-fold enhancement of the ferryl
lifetime with the 5,5,5-d3-Nva-S-SyrB1 substrate would diminish the kHAT:kex ratio to 0.67:1,
giving a predicted 18O incorporation of 40% in 5,5-d2-5-OH-Nva. The agreement of this calculated
value with the measured value (42%) supports the assumption that, for C5, exchange occurring in
the rebound state is not a significant contributor to the total observed exchange. Application of the
same analysis shows that exchange occurring only in the ferryl state cannot rationalize the data for
C4. With Nva-S-SyrB1, the ~ 50% 18O incorporation would require kex ~ 9 s-1 (equal to the effective
rate constant for ferryl decay), inconsistent with incorporation of 80% 18O at C5. Extension of the
ferryl lifetime by ~ 6-fold with 5,5,5-d3-Nva-S-SyrB1 should then yield only 14% 18O
incorporation, far less than the observed 32-37%. The explanation is that some (approximately
half) of the solvent exchange reflected in the 4-OH-Nva product with unlabeled substrate occurs
in the rebound state, and extension of the lifetime of the ferryl complex does not increase this
contribution.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The corresponding analysis for the case of LolO is hampered somewhat by the lack of
sufficient solvent exchange in hydroxylation of C2 of the unlabeled substrate for the extent of
exchange to be accurately determined. However, because this fact also insures that very little
exchange occurs in the Fe(III)–OH/C2• rebound state, the rate constant for exchange in the ferryl
complex can be estimated from the 18O incorporation at C2 in the reaction with the 2,2,8-d3-AcAP
26
ACS Paragon Plus Environment

Page 27 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
substrate (59%). With the kobs for ferryl decay of 2.0 s-1, kex must be ~ 1.4 s-1 to account for the
observed 41% exchange. Interestingly, this value is similar to that calculated for the case of SyrB2.
The clearest evidence of solvent exchange in the rebound state comes from the fraction of
18O in the 7-OH-Hyo product made by H6H. Were the observed washout to occur solely in the
ferryl state, the 68% exchange seen in this product with the unlabeled substrate would require kex
of > 30 s-1, and the extension of the ferryl lifetime by ~ 8-fold with 6-d1-Hyo would give ~ 95%
washout, much more than the observed 82%. Here, the high extent of washout apparently results
primarily from exchange with solvent in the rebound state.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The mechanisms of exchange of the oxygen ligand with solvent in the two states are not
known, but it has been suggested (and we deem it most likely) that hydration of a stably or
transiently 5-coordinate Fe center [two His imidazoles, two monodentate carboxylates (or a
carboxylate and halide), and the oxygen] would be the first step for the case of the ferryl complex.57
A similar hydration step is also proposed for the exchange observed in both heme and non-heme
ferryl model complexes.58-60 This hydration would give two ligands of equivalent redox state but
different protonation states. At least one net proton transfer between these oxygen ligands, as well
as some rearrangement of the coordination sphere (akin to a Berry pseudorotation), would be
required to complete the net swapping of the oxo/hydroxo of the intermediate with a solvent
molecule. To test the prediction that proton transfer is required for exchange, we carried out the
LolO and H6H reactions in 2H2O and 18O2. Solvent deuteration results in a greater fraction of 18O
incorporation (i.e., less exchange) into both products of both enzymes (Figures S4S5 and Table
2), consistent with the prediction that the exchange process requires at least one proton transfer.
Interestingly, in both enzymes, the slowing of solvent exchange by 2H2O causes a leveling of the
fractions of 18O in the two hydroxylated products, suggesting a greater effective solvent deuterium
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 40
1
2
3
4
5
6
7
8
9
KIE on exchange in the rebound states than on exchange in the ferryl states. This effect provides
additional corroboration that two distinct exchange processes (i.e., in both intermediate states)
contribute to the overall washout, at least for the case of C7 (for both enzymes).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
DISCUSSION
The conclusions from the reactivity studies on SyrB2 (i) that strict control of the disposition
of the C-H bond to be cleaved relative to the cis-chloroferryl intermediate is necessary to allow C–
Cl radical coupling to preempt oxygen rebound and (ii) that this disposition makes the preceding
HAT step markedly less efficient than it would be if it were independently optimized have largely
held up to subsequent experimental and computational scrutiny, but intriguing implications and
follow-on questions have thus far gone unaddressed. For example, it is not known whether C–Cl
coupling is, by virtue of this positioning, faster than the rebound steps of hydroxylases or,
alternatively, if the rebound step becomes so sluggish that an even modestly efficient C–Cl-
coupling step can prevail. Computational studies have converged on the notion that the outcome
is linked to the frontier molecular orbitals (FMO) through which HAT proceeds. Rapid HAT
through a σ-channel results in a HO–FeIII–Cl/R• state with the radical closer to the oxygen, well
poised for rebound but not for Cl• coupling. Conversely, the less efficient π-trajectory for HAT
leads to a HO–FeIII–Cl/R• state with the substrate radical nearly perpendicular to the HO-FeIII-Cl
plane and approximately equidistant from the two ligands. In this geometry, Cl• coupling can
effectively compete with rebound, because it has a lower activation barrier. Splitting of the d*
FMOs through which the alternative radical-coupling steps must proceed leads to differential
overlap with the substrate p orbital, kinetically favoring Cl• coupling, despite the fact that the
hydroxylation product is lower in energy.61
28
ACS Paragon Plus Environment

Page 29 of 40
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The issue can be raised, more generally, for any non-hydroxylation outcome: is the
activation barrier for rebound raised, or is a still-facile rebound step preempted by an alternative
process that a particular enzyme selectively promotes? In the former case, it would be important
to understand whether the activation barriers of the HAT and rebound steps are inextricably
coupled, as one might infer from the correlation seen for SyrB2, or whether other enzymes capable
of non-hydroxylation outcomes may have evolved strategies to suppress rebound without
sacrificing HAT proficiency. These issues cannot meaningfully be addressed without an
experimental window to the rates of the rebound and alternative steps through which the Fe(III)-
OH/R• complexes decay. Direct monitoring of these steps, of the sort presented here (and in
previous work) for the ferryl-mediated HAT steps, has not been possible. The presumption has
been that the substrate-radical states are kinetically masked (i.e., do not accumulate) by their
relatively slow formation and fast decay (e.g., by rebound). In other systems, substrates that, upon
forming radicals, react to migrate the radical (e.g., by ring opening) with known rate constants,
giving characteristic products, have been used to time the rebound step. Although this “radical-
clock” approach has provided important insights, it generally requires rather drastic modifications
to substrates and rests on the dual assumptions that these modifications do not change the
mechanism and that binding in the active site does not markedly perturb the rates of competing
radical-timing steps relative to the values known in solution or the gas phase. Approaches to probe
downstream steps on native substrates would, seemingly, have value.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The three systems probed in this work share the common feature of generating ferryl
complexes that can ambiguously target either of two carbon centers for HAT, always
hydroxylating one site but either hydroxylating or (in the proper context) mediating a different
transformation of the other. In light of our expectation from previous work that prevention of
29
ACS Paragon Plus Environment