Versatile de novo enzyme activity in capsid proteins from an engineered M13 bacteriophage library

Article abstract of DOI:10.1021/ja506346f

Biocatalysis has grown rapidly in recent decades as a solution to the evolving demands of industrial chemical processes. Mounting environmental pressures and shifting supply chains underscore the need for novel chemical activities, while rapid biotechnological progress has greatly increased the utility of enzymatic methods. Enzymes, though capable of high catalytic efficiency and remarkable reaction selectivity, still suffer from relative instability, high costs of scaling, and functional inflexibility. Herein, we developed a biochemical platform for engineering de novo semisynthetic enzymes, functionally modular and widely stable, based on the M13 bacteriophage. The hydrolytic bacteriophage described in this paper catalyzes a range of carboxylic esters, is active from 25 to 80 °C, and demonstrates greater efficiency in DMSO than in water. The platform complements biocatalysts with characteristics of heterogeneous catalysis, yielding high-surface area, thermostable biochemical structures readily adaptable to reactions in myriad solvents. As the viral structure ensures semisynthetic enzymes remain linked to the genetic sequences responsible for catalysis, future work will tailor the biocatalysts to high-demand synthetic processes by evolving new activities, utilizing high-throughput screening technology and harnessing M13's multifunctionality.

Full text of DOI:10.1021/ja506346f

Article
pubs.acs.org/JACS
Versatile de Novo Enzyme Activity in Capsid Proteins from an
Engineered M13 Bacteriophage Library
John P. Casey, Jr.,† Roberto J. Barbero,† Nimrod Heldman,‡ and Angela M. Belcher*,†,‡,§
‡
§
†Biological Engineering, Materials Science and Engineering, and Koch Institute for Integrative Cancer Research, Massachusetts
Institute of Technology, 77 Massachusetts Avenue, 76-561, Cambridge, Massachusetts 02139, United States
S
* Supporting Information
ABSTRACT: Biocatalysis has grown rapidly in recent decades
as a solution to the evolving demands of industrial chemical
processes. Mounting environmental pressures and shifting
supply chains underscore the need for novel chemical
activities, while rapid biotechnological progress has greatly
increased the utility of enzymatic methods. Enzymes, though
capable of high catalytic efficiency and remarkable reaction
selectivity, still suffer from relative instability, high costs of
scaling, and functional inflexibility. Herein, we developed a
biochemical platform for engineering de novo semisynthetic
enzymes, functionally modular and widely stable, based on the
M13 bacteriophage. The hydrolytic bacteriophage described in this paper catalyzes a range of carboxylic esters, is active from 25
to 80 °C, and demonstrates greater efficiency in DMSO than in water. The platform complements biocatalysts with
characteristics of heterogeneous catalysis, yielding high-surface area, thermostable biochemical structures readily adaptable to
reactions in myriad solvents. As the viral structure ensures semisynthetic enzymes remain linked to the genetic sequences
responsible for catalysis, future work will tailor the biocatalysts to high-demand synthetic processes by evolving new activities,
utilizing high-throughput screening technology and harnessing M13’s multifunctionality.
solvents,3,12 they often become substantially less active when
INTRODUCTION
■
removed from aqueous conditions.13
Enzymes and biocatalysts potentiate a wide range of chemical
activities and functionality not practical with abiotic processes.
Performance enhancements in substrate specificity, reaction
selectivity, and catalytic efficiency drive their accelerating
Fortunately, functional tools to engineer and adapt enzymes
have advanced considerably in recent years. Attempts to
overcome traditional biocatalytic limitations usually fall into
one of two strategies: modifying pre-existing enzymes by
implementation in reactions that have traditionally employed
rational engineering or directed evolution, or the generation
inorganic catalysts, such as the chiral resolution of alcohols and
the oxidation of alkane precursors.1−3 Enzymes can result in
and development of de novo catalytic entities. The former has a
longer history and can include targeted mutations to, for
lower process temperatures for industrial reactions while
instance, improve protein lifetime via insertion of disulfide
decreasing the amount of solvent and waste, reducing energy
linkages.14 Immobilization techniques have been especially
and material costs.4 Additionally, the unique geometries and
successful in improving temporal and thermal stability and have
energy states created by enzyme active sites allow for the
generation of many important chemicals with no effective
been implemented broadly for the industrial production of
inorganic means of production (for example, artemisinin5 and
pharmaceuticals and a number of other common chemicals (for
review, see Zhou and Hartmann15 or Bommarius and Paye16).
Chemical and genetic modifications, as well as myriad cross-
linking strategies, have also been successful in modulating
stability and selectivity.17 Directed evolution approaches have
made notable contributions by sampling larger windows of the
functional landscape to identify unintuitive active site
modifications7,18−20 (see Turner21 for recent review). This
strategy has been implemented in a variety of processes,
including Lipitor production and the manufacture of chiral
amines.10 While the engineering of existing enzymes has been
carotenoids6). The inherent modularity of biomolecules
complements such advantages by promising engineers more
control over enzyme outputs. Not only is protein structure
manipulable by genetic means, but well-characterized bio-
chemical functional handles allow for catalytically significant
post-translational modifications. As successful examples of
enzymatic approaches proliferate,7−9 however, challenges
remain in adapting proteins to alternate environs or functions.10
Native enzymes often exhibit short lifetimes, on the scale of
hours without modification or immobilization;11 outside of
native pH and temperature, their functionality can become less
predictable. Similarly, while some enzymes have been found to
have novel activity or even greater thermostability in organic
Received: June 24, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja506346f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Design of semisynthetic active sites for hydrolysis. (a) The active site of bCAII consists of three histidine residues coordinating a Zn2+ ion
(1CA240). The front-most His residue shown comes from an adjacent β-sheet. (b) The cloning strategy involved mutating three specific residues on
M13 pVIII to histidine (red, sequence at bottom) and randomizing adjacent terminal residues (blue). His pairs were at i3,i5 or i4,i6; backbone
mutations were at S13, Q15, G23, or A27. Image includes PDB file 2C0W.39 (c) The hydrolysis of p-nitrophenyl acetate (pNPA) yields colorimetric
p-nitrophenylate (pNP), which can be detected via absorbance at 400 nm, and acetate.
increasingly successful in improving those enzymes and, in
some cases, adapting them to tangential or novel reactions,22,23
such approaches are rarely fully generalizable. An alternate
strategy involves co-opting structural domains of unrelated
proteins for de novo active sites or designing new protein
structures entirely. By sampling completely independent
domains of the fitness landscape, such undertakings allow for
unique combinations of targeted functional attributes, especially
when function can be uncoupled from structural stability.24 By
building protein domains from the bottom up, one increases
the potential for diverse activity.25,26 Three- and four-helix
bundles have been utilized for redox,27−29 peroxidase,30
carbonic anhydrase,31 esterase,32 and lipase activity.33 Compu-
tational design and the insertion of exogenous active sites into
existing proteins have led to similarly successful outcomes.34−36
Such varied approaches are rapidly broadening the scope of
biocatalysts in their activity and applications.
Previous de novo engineered enzymes have recapitulated
target enzymatic activity to a remarkable degree, but few have
been developed for the variety of non-native environments
common to industrial processes. Here, we sought to develop a
robust, versatile biocatalytic platform potentiating the catalysis
of targeted reactions in nonstandard environments. To do so,
we fused multiple above-mentioned strategies, recreating active
site geometry from a highly efficient enzyme on an otherwise
entirely structural protein assembly and selecting a functional
clone. As a template, M13 bacteriophage was chosen for its
excellent thermostability, scalability, multivalency, and ease of
genetic modification. M13 is a filamentous macromolecule 880
nm in length and 6.5 nm in diameter, with 2700 identical pVIII
capsid proteins wrapping 7 kbp of single-stranded DNA. By
displaying catalytic motifs from the well-studied carbonic
anhydrase enzyme on the major coat protein, pVIII, we
endowed M13 with thousands of semisynthetic enzymes
capable of catalyzing hydrolysis reactions. With a relatively
open binding pocket, the M13 constructs exhibited diverse
activity, catalyzing a range of substrates less accessible to native
carbonic anhydrase. Moreover, the catalytic efficiency increased
substantially outside of aqueous reaction conditions and at
elevated temperatures, wherein the wild-type enzyme became
inactive. Thus, interfacing biochemical functions with non-
biological conditions, M13 semisynthetic enzymes inform
continuing studies of protein design while opening new
opportunities in sustainable molecular biocatalysis.
RESULTS
■
Building M13 Bacteriophage Libraries with Incorpo-
rated Catalytic Motifs. In order to design effective
biocatalysts, we leveraged our lab’s experience engineering
M13 against the extensive literature on carbonic anhydrase
enzymes to build a set of pVIII libraries incorporating catalytic
motifs found in anhydrase active sites. Of the five M13 coat
proteins, pIII is the largest and most commonly engineered;
indeed, previous work has successfully cloned full anhydrase
enzymes into M13 pIII.37 However, for this work we chose
instead to modify the 50 amino acid (AA) pVIII protein for a
more multivalent (2700 copies vs ≤5 for pIII) and durable
semisynthetic enzyme system. In addition to outnumbering the
minor coat proteins by >500-fold, pVIII’s simple, symmetrically
repeated α-helices are highly stable to environmental extremes.
Fiber diffraction experiments of filamentous bacteriphage have
detailed the position and orientation of these proteins relative
to each other (PDB 1IFJ38 and PDB 2C0W39). Such
information makes the pVIII of M13 particularly tractable for
biological engineers, allowing one to rationally replace native
residues to approximate active site geometries and chemistries
while interrogating primary structure relationships.
In doing so, we chose to model our pVIII constructs after
carbonic anhydrase II (CAII), a thoroughly characterized and
highly efficient zinc-based metalloenzyme (Figure 1a) (PDB
1CA2).40 CAII’s active site has two histidine residues separated
by phenylalanine in an α-helix and a third histidine residue in a
B
dx.doi.org/10.1021/ja506346f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
proximal β-sheet; the three His residues together bind a Zn2+
ion whose fourth coordination site is occupied by a hydroxyl
ion or substrate. To replicate this geometry, we built a library of
plasmids encoding a pair of His residues at fixed positions,
surrounded by randomized amino acids, in 8 AA inserts located
at the surface-exposed N-terminus of pVIII (Figure 1b; see
Supplementary Figure 1 in the Supplementary Results for more
details on the oligonucleotide libraries). A third histidine
(“backbone” histidine) was inserted deeper into the pVIII
(further from the N-terminus), such that each N-terminal pair
of histidine residues on a pVIII would be in close proximity to
the backbone histidine of an adjacent pVIII protein. Since the 8
AA inserts are likely not part of the backbone α-helix, which is
believed to be broken by Pro6, the specific orientation of the
eight residues within each hypothetical insert is not readily
inferable. We generated an ensemble of M13 libraries with two
variables: the position of the histidine pair in the N-terminus
and the location of the backbone histidine, seeking to maximize
the potential for ion coordination between three His residues
on adjacent helices. PDB files 1CA2 for CAII and 2C0W for
M13 were used for evaluation. Our analyses of the structural
models indicated that the most favorable trios involved pairs
i3,i5 and i4,i6 on the insert (positions along the 8 AA peptide
insert will be indicated with an ‘i’ prefix; the other positions will
retain wild-type numbering with no prefix) complemented with
backbone histidines at the 13th, 15th, 23rd, or 27th residue
from the native N-terminus. Supporting these selections, the
positions match closely with those residues downstream of the
N-terminus expected to be most surface exposed in wild-type
models.38
Individual clones from these libraries were then surveyed for
primary sequence similarity to the CAII active site residues,
with an emphasis on hydrophobicity in the immediate vicinity
of the coordinating histidines. A subpopulation of 10 sequences
was selected for initial experimental characterization, and
subsequently the clone DDAHVHWE, G23H (DDAH) was
chosen for further study (Supplementary Figure 2).
Clone DDAH Catalyzes the Aqueous Hydrolysis of p-
Nitrophenyl Acetate. In order to characterize the bacter-
iophage-displayed semisynthetic enzymes, we measured hydrol-
ysis of nitrophenyl esters (Figure 1c). p-Nitrophenyl acetate
(pNPA) hydrolysis is well-characterized as a model reaction for
CAII,37,41 rationally engineered enzymes,42 de novo pepti-
des,31,33,43 and novel materials.44 From the subpopulation of
bacteriophage clones we tested, the clone DDAH was
determined to be the most reliable and consistent candidate
(Supplementary Figure 2) and was chosen for further
characterization. In ion coordination studies, the clone was
found to bind Zn2+ with a KD of 6 μM via microscale
thermophoresis (Supplementary Figure 3); reactions were then
performed with 25 μM ZnSO4.
Figure 2. M13 clone DDAH catalyzes the aqueous hydrolysis of
pNPA. (a) Plot of hydrolysis rate vs substrate concentration for M13
clone DDAH (green circles) and the control reaction (PBS, gray ×’s).
(b) Respective kcat/KM catalytic rates for DDAH and a number of
mutants. Δ3H: terminal two histidines mutated to alanine and third
histidine returned to glycine (as wild-type); DDAH-8 and DDAH-12:
8-mer and 12-mer synthetic peptides, respectively, comprised of the
terminal peptide sequence of DDAH pVIII. Error bars represent 95%
confidence intervals for values fit by Lineweaver−Burk analysis. p-
values: * <0.05; ** <0.005; *** < 0.0005; represents two-sample t test
for regressed coefficients of DDAH and given sample.
comparison, wild-type bovine carbonic anhydrase (bCAII,
Worthington Biochemical Corporation #LS001260) hydro-
lyzed pNPA at a rate of 194 M−1 s−1 and solubilized L-His
monomers measured 0.021 M−1 s−1. The activity of DDAH
thus represents a significant coordination between the residues,
while remaining far lower than that of the wild-type enzyme.
The inferred kcat, 0.002 and 1.2 s−1 for DDAH and bCAII,
respectively, implies that a difference in turnover rate is a far
larger contributing factor than substrate binding, though the kcat
values themselves are less precise calculations from our
Lineweaver−Burk plots than kcat/KM. We believe the low
activity compared to carbonic anhydrase is in part related to the
conformational flexibility of the pVIII termini in aqueous
solutions; the 10 terminal residues following P6 likely lack
secondary structure. This hypothesis is considered in more
The hydrolysis activity of the bacteriophage was quantified
by the catalytic efficiency, kcat/KM, according to Michaelis−
Menten kinetics (see Supplementary Methods):
kcat[E][S]
KM + [S]
d[P]
dt
=
Each pVIII protein was considered conservatively to be one
active site, so that each bacteriophage represented 2700
catalytic sites. As such, clone DDAH was found to hydrolyze
pNPA with kcat/KM = 0.535 M−1 s−1 for each pVIII subunit
(Figure 2 and Supplementary Table 1) in PBS at pH 7.4; by
C
dx.doi.org/10.1021/ja506346f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
detail in the Discussion section. Additionally, the current
strategy did not target second-shell interactions between nearby
AA residues and the inserted zinc-coordinating histidines; such
interactions play a large role in catalytic efficiency and their
inclusion would likely improve the construct’s activity.
To more thoroughly understand the contributing factors of
the DDAH active site, we performed an alanine scan on the
peptide insert and engineered two additional histidine
substitutions: DDAH with the His residue at position 23
mutated H23G, as in wild-type M13, and a Δ3H clone
engineered with H4A, H6A, and H23G mutations so that no
histidine residues remain. The results are summarized in Figure
2b and Supplementary Table 1. Each mutant’s catalytic activity
was significantly lower than the original DDAH clone (two-
sample p-value <0.05), though all but the Δ3H clone had
measurable hydrolysis activity above background. Notably, the
tryptophan residue in the insert, at the same position relative to
the histidine pair as in CAII (FHFHW, Figure 1a), was found
to be important for hydrolysis activity. In addition to the M13
mutants, we tested free-floating synthetic peptides representing
the N-terminal 8 and 12 AAs; that is, DDAHVHWE and
DDAHVHWEDPAK, not expressed on the surface of M13.
These peptides showed negligible activity (0.021 and 0.013
M−1 s−1, respectively).
DDAH Activity Is Robust in Nonaqueous Solvents. As
many synthetic processes involve harsh, abiotic environments,2
we sought to determine the catalytic efficiency of DDAH in
relevant nonaqueous solvents. The substrate has excellent
solubility and stability in DMSO, a highly useful solvent in, for
instance, pharmaceuticals production,45 so this solvent was
chosen for further investigation. DDAH exhibited a 30-fold
increase in catalytic efficiency in 98% DMSO, wherein kcat/KM
= 16.87 M−1 s−1, while background levels remained roughly the
same (Supplementary Table 2). The activity is stable for hours,
through the length of the experiments run (Figure 3). The
phage assemblies reach a steady-state kcat and KM within 40 min
of mixing, and the overall kcat/KM shows no significant
difference throughout the experiment. The DDAH clone itself
is rendered noninfectious after exposure to DMSO (Supple-
mentary Table 3), indicating that pIII is irreversibly denatured,
even though the overall filamentous structure remains intact
(see transmission electron micrographs, Supplementary Figure
4). Wild-type bCAII showed no measurable activity in DMSO
(at 67 nM, Figure 4a). Similarly, another M13 clone with
activity in PBS, TENM (TENMHGTM, 23H; Supplementary
Figure 2), exhibited no activity in DMSO.
DDAH Catalyzes a Range of Substrates. A critical
feature of the M13 semisynthetic enzyme design is the
openness of the active site, situated between the most N-
terminal nine amino acids of pVIII and an adjacent pVIII α
helix; the active site is relatively unimpeded. We therefore
sought to characterize the activity of DDAH with respect to
substrates with increasing carbon chain length, pNPA (n = 2),
p-nitrophenyl propionate (pNPP, n = 3), p-nitrophenyl
butyrate (pNPB, n = 4), and p-nitrophenyl palmitate
(pNPPa, n = 15, lipase substrate) (Figure 4b). Only pNPA
and pNPB were soluble in the aqueous reaction mixture, but
the stability of DDAH in DMSO enabled us to study substrates
not soluble in PBS. Solubilized L-histidine was used as a control,
as it catalyzes ester hydrolysis at a measurable level and because
histidine residues are critical for the aqueous activity of DDAH
clones (Figure 2b). In aqueous reactions, DDAH hydrolyzed
Figure 3. DDAH stably catalyzes pNPA hydrolysis in DMSO. (a)
Hydrolysis rate of pNPA vs substrate concentration for DDAH clones
solubilized in DMSO for different lengths of time (markers), with
regressions (lines). (b) kcat/KM values through 2 h in DMSO stay
∼90% of the initial value. Catalytic parameters (αt) are determined
from a NLLS fit to the Michaelis−Menten equation and normalized to
their t0 = 15 min values (α0, earliest time point to mix and plate out
samples with substrate). Time points: 15, 25, 38, 68, 90, and 125 min.
Error bars are 95% confidence intervals for kcat and KM, 90%
confidence intervals for kcat/KM (0.952).
0.282 M−1 s−1 (Figure 4c). By comparison, wild-type carbonic
anhydrase has been found to catalyze pNPB hydrolysis with
only 2% the activity with which it hydrolyzes pNPA.42 The
nonaqueous reactions, again, were catalyzed far faster than the
water-based reactions. In 98% DMSO, the kcat/KM = 16.87 M−1
s−1 for DDAH-catalyzed hydrolysis of pNPA dropped to 8.92
M−1 s−1 for pNPP (Figure 4d). Notably, DDAH catalyzed the
lipase substrate pNPPa with a catalytic efficiency 5.89 M−1 s−1,
which is 35% of the catalytic efficiency of DDAH with the
much smaller substrate pNPA. The capacity to react with
substrates of such varying size makes DDAH a promising
alternative to biocatalysts whose substrate versatility is limited
by steric hindrance. Interestingly, despite no demonstrable
activity in aqueous reactions and no histidines on pVIII, the
pNPB at just over half the efficiency of pNPA, with kcat/KM
=
D
dx.doi.org/10.1021/ja506346f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. DDAH demonstrates versatile catalytic activity in water and DMSO. (a) Typical product accumulation vs time plot in 98% DMSO; wild-
type bCAII and TENM (a control M13 phage clone with two histidine substitutions) exhibit no measurable hydrolysis activity. [pNPA]0 = 250 μM.
(b) Structures of the five hydrolysis substrates. (c) Log−scale plot of L-histidine and DDAH in the aqueous (98% PBS) hydrolysis of nitrophenyl
esters at room temperature and indoxyl acetate at 80 °C. (d) Catalytic efficiency of L-histidine, DDAH, and Δ3H solubilized in a 98% DMSO
reaction mixture with substrates of differing carbon chain lengths. Error bars in (c and d) represent 95% confidence intervals for values fit by
Lineweaver−Burk analysis. In (c), IndAc hydrolysis by L-histidine indistinguishable from background rate; Lineweaver−Burk upper error bar not
available.
Δ3H clone was catalytically active in the DMSO reaction
mixture (Figure 4d). While not as efficient a catalyst as DDAH
clone, the activity of this mutant is intriguing and may be a
result of remaining exposed aspartates (three) and glutamates
(one).
In addition to larger nitrophenyl esters, we assayed the
activity of DDAH toward a less active carboxylic ester, indoxyl
acetate (IndAc). Upon hydrolysis of the acetate moiety, indoxyl
is rapidly converted to indigo in the presence of oxygen.
Though insoluble in water, indigo production could be
quantified upon dilution into DMSO. We were able to detect
no activity at room temperature. However, at 80 °C DDAH
hydrolyzed the substrate with kcat/KM = 3.13 M−1 s−1. This
compares with 30.47 M−1 s−1 for bCAII, while the activity of
solubilized L-histidine measured 0.025 M−1 s−1 and was
statistically indistinguishable from background hydrolysis.
Catalytic Efficiency Increases with Higher Temper-
atures. In order to characterize the effect of temperature on
DDAH activity, pNPA hydrolysis was measured under heated
conditions. Many industrial processes are conducted at raised
temperatures, both to improve catalytic performance and to
limit bacterial contamination.46 Accordingly, pNPA hydrolysis
experiments were conducted in DMSO at 40 and 65 °C, with
the latter representing a temperature at which bacterial growth
is unlikely. Regressed catalytic parameters are shown in Figure
5, normalized to values at room temperature to facilitate
comparison. The DDAH catalytic rate kcat increased signifi-
cantly for the heated reactions, while the constant KM remained
E
dx.doi.org/10.1021/ja506346f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
by the Kuhlman group (kcat/KM at pH 7.5 = 90 M−1 s−1 for two
zinc coordination sites) and Rufo et al. (at pH 8, kcat/KM = 62
M−1 s−1 per amyloid peptide) have made substantial improve-
ments in rates for aqueous esterolysis and are informative for
ongoing enzyme design, but neither approach demonstrates
solvent/thermostable activity nor enables genetic screens.
The environmental robustness inherent to the M13 viral
capsid is a key feature in its use as a biocatalytic platform. In
particular, its compatibility with organic solvents and its
thermostability are advantageous for numerous applications.
Here, the stability of DDAH in DMSO allowed for the catalysis
of substrates and reactions not practical in aqueous solutions,
broadening the potential application space. The increase in
activity observed in DMSO may be a result of conformational
rigidity imparted by hydrophobic solvent interactions with
pVIII N-terminal residues,12 which lack known secondary
structure in aqueous solutions. The well-characterized α-helix
formed by pVIII is likely broken by the proline in position 6,
resulting in an unknown conformation for the most N-terminal
10 residues. If these residues, which include our inserts, are
more restrained conformationally in less polar solvents,12 then
this likely contributes substantially to the increase in catalytic
activity observed. Additionally, as the M13 active site is
relatively small, substrates may remain partially exposed to the
surrounding solvent, in which case nonpolar solvents can
increase activity.48 In addition to DMSO, the phage is
compatible with ethane diol (Supplementary Table 3), and
other solvents will be characterized in future work. Given the
number of important feed stocks and precursor compounds
insoluble in water, such diverse solvent compatibility is of
immense utility. Furthermore, the increasing catalytic efficien-
cies at raised temperatures makes M13 biocatalysts amenable to
a wider range of reactions and reduced risks of bacterial
contamination.
More broadly, M13’s well-understood biology and dense
display of peptides make it appealing for future development.
The ease of bacteriophage amplification and purification allows
for greater scalability of the catalysts compared to other
enzymatic proteins, while pVIII’s tightly packed, highly
multivalent structure leads to higher numbers of reaction
centers. With thousands of pVIII proteins per biomolecule,
M13 can stably approach millimolar concentrations of active
sites with or without immobilization. Moreover, the inherent
inclusion of the bacteriophage genome with all progeny makes
functional catalytic selections from diverse libraries possible,
whereas traditional enzymes lack the requisite genotype−
phenotype connection. Accordingly, future engineering of the
bacteriophage will harness its high multivalency and genotype−
phenotype linkage in conjunction with recent developments in
high-throughput assays, such as microengraving and micro-
emulsions,49 to search a larger sequence space for active clones.
M13 biology further allows for directed evolution and host
rescue assays to rapidly select functional variants. M13’s
genome itself is highly engineered, with numerous restriction
enzyme sites in the genes of its various coat proteins, such that
complementary binding or catalytic functions can be
incorporated at other positions on the bacteriophage capsid.
Figure 5. DDAH catalytic efficiency increases in higher temperatures.
DDAH samples were solubilized in DMSO at the given temperatures
and incubated for 30 min before substrate was added. After subtracting
background activity of 25 μM ZnSO4, each catalytic parameter (αT) is
determined by Lineweaver−Burk analysis and normalized to the room
temperature value (αRT). Error bars represent 95% confidence
intervals.
close to the room temperature value, such that the overall
catalytic efficiency increased with higher temperatures to twice
that at room temperature. The stability of the M13 capsid
structure therefore leads to a robust active site capable of
catalysis at a wide range of functional temperatures.
DISCUSSION
■
In this study, we have developed a novel strategy for
biocatalytic studies: the incorporation of an enzymatic motif
into thermostable bacteriophage particles to yield versatile
semisynthetic enzymes. The selected clone, DDAH, mimicked
the active site of carbonic anhydrase in its distribution of polar
and hydrophobic residues. The presence of nonpolar amino
acids in active sites has long been known to be critical for
efficient catalysis, hypothesized to reduce the conformational
variability of charged coordinating residues and/or increase the
surrounding local dielectric field.47 Indeed, mutation of Vi5 or
Wi7 to alanine significantly decreases the catalytic activity of
DDAH, highlighting the important role these residues play in
the clone’s active site. The activity is contingent upon the
quaternary structure of M13’s coat proteins, as soluble peptides
exhibited negligible activity. In addition, the complete inactivity
of the Δ3H clone in water renders highly unlikely the
possibility that measured catalysis is the result of contaminant
Escherichia coli detritus from our bacteriophage purification
protocols. pNPA hydrolysis, a useful representative reaction for
enzymatic activity, will inform future engineering efforts toward
more complex target reactions. The demonstrated display and
organization of multiple metal-coordinating residues on all
2700 pVIII proteins is a novel capability and will allow for the
coordination of other cofactors and replication of other
enzymatic activities. While the aqueous catalytic efficiency
remains substantially lower than that measured for wild-type
bCAII, the activity is nonetheless comparable to de novo
enzymes reported by Zastrow et al.31 for the same pH, at pH
7.5, kcat/KM = 1.38 M−1 s−1 per three helices, and Patel et al.,33
at pH 8, kcat/KM = 9.9 M−1 s−1 per four helices. Recent works
CONCLUSION
■
Broader biochemical toolkits and a greater understanding of
biocatalysis are critical in order to meet accelerating society-
wide demand for chemicals, food, and materials. Proteins offer
enhanced performance, novel chemical activity, and greater
F
dx.doi.org/10.1021/ja506346f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(22) Jochens, H.; Stiba, K.; Savile, C.; Fujii, R.; Yu, J. G.;
Gerassenkov, T.; Kazlauskas, R. J.; Bornscheuer, U. T. Angew. Chem.,
Int. Ed. 2009, 48, 3532.
(23) Coelho, P. S.; Wang, Z. J.; Ener, M. E.; Baril, S. A.; Kannan, A.;
Arnold, F. H.; Brustad, E. M. Nat. Chem. Biol. 2013, 9, 485.
(24) Gatti-Lafranconi, P.; Hollfelder, F. ChemBioChem 2013, 14, 285.
(25) Regan, L.; Degrado, W. F. Science 1988, 241, 976.
(26) Kamtekar, S.; Schiffer, J. M.; Xiong, H. Y.; Babik, J. M.; Hecht,
M. H. Science 1993, 262, 1680.
(27) Robertson, D. E.; Farid, R. S.; Moser, C. C.; Urbauer, J. L.;
Mulholland, S. E.; Pidikiti, R.; Lear, J. D.; Wand, A. J.; Degrado, W. F.;
Dutton, P. L. Nature 1994, 368, 425.
(28) Huang, S. S.; Koder, R. L.; Lewis, M.; Wand, A. J.; Dutton, P. L.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5536.
environmental compatibility with respect to prevalent inorganic
catalysts. We have identified an M13 bacteriophage clone with
novel catalytic properties and thermal stability, in the process
demonstrating M13 to be a versatile, high-surface area platform
for biocatalytic engineering. Future work will build on these
results by expanding M13 bacteriophage library diversity,
improving selection methodology, evolving clones toward
unique reactions, and fabricating stable catalytic macro-
structures.
ASSOCIATED CONTENT
■
S
* Supporting Information
(29) Farid, T. A.; Kodali, G.; Solomon, L. A.; Lichtenstein, B. R.;
Sheehan, M. M.; Fry, B. A.; Bialas, C.; Ennist, N. M.; Siedlecki, J. A.;
Zhao, Z. Y.; Stetz, M. A.; Valentine, K. G.; Anderson, J. L. R.; Wand, A.
J.; Discher, B. M.; Moser, C. C.; Dutton, P. L. Nat. Chem. Biol. 2013, 9,
826.
(30) Das, A.; Hecht, M. H. J. Inorg. Biochem. 2007, 101, 1820.
(31) Zastrow, M. L.; Peacock, A. F. A.; Stuckey, J. A.; Pecoraro, V. L.
Nat. Chem. 2012, 4, 118.
Experimental procedures, electron microscopy, and other
supporting data. This material is available free of charge via
the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
belcher@mit.edu
(32) Broo, K. S.; Brive, L.; Ahlberg, P.; Baltzer, L. J. Am. Chem. Soc.
1997, 119, 11362.
Notes
(33) Patel, S. C.; Bradley, L. H.; Jinadasa, S. P.; Hecht, M. H. Protein
Sci. 2009, 18, 1388.
The authors declare no competing financial interest.
(34) Benson, D. E.; Wisz, M. S.; Liu, W. T.; Hellinga, H. W.
Biochemistry 1998, 37, 7070.
(35) Bolon, D. N.; Mayo, S. L. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
ACKNOWLEDGMENTS
■
We thank Jacqueline Ohmura for the image in Supplementary
Figure 4b as well as the Center for Materials Science and
Engineering for access to transmission electron microscopy for
both images in Supplementary Figure 4. Additionally, we would
like to acknowledge the Swanson Biotechnology Core at the
MIT Koch Institute for Integrative Cancer Research for
bacteriophage DNA sequencing.
14274.
(36) Rothlisberger, D.; Khersonsky, O.; Wollacott, A. M.; Jiang, L.;
DeChancie, J.; Betker, J.; Gallaher, J. L.; Althoff, E. A.; Zanghellini, A.;
Dym, O.; Albeck, S.; Houk, K. N.; Tawfik, D. S.; Baker, D. Nature
2008, 453, 190.
(37) Hunt, J. A.; Fierke, C. A. J. Biol. Chem. 1997, 272, 20364.
(38) Marvin, D. A.; Hale, R. D.; Nave, C.; Citterich, M. H. J. Mol.
Biol. 1994, 235, 260.
(39) Marvin, D. A.; Welsh, L. C.; Symmons, M. F.; Scott, W. R. P.;
Straus, S. K. J. Mol. Biol. 2006, 355, 294.
(40) Eriksson, A. E.; Jones, T. A.; Liljas, A. Proteins 1988, 4, 274.
(41) Pocker, Y.; Stone, J. T. J. Am. Chem. Soc. 1965, 87, 5497.
(42) Host, G.; Martensson, L. G.; Jonsson, B. H. Biochim. Biophys.
Acta, Proteins Proteomics 2006, 1764, 1601.
(43) Nicoll, A. J.; Allemann, R. K. Org. Biomol. Chem. 2004, 2, 2175.
(44) Chen, L.; Bromberg, L.; Hatton, T. A.; Rutledge, G. C. Polymer
2007, 48, 4675.
(45) Savile, C. K.; Janey, J. M.; Mundorff, E. C.; Moore, J. C.; Tam,
S.; Jarvis, W. R.; Colbeck, J. C.; Krebber, A.; Fleitz, F. J.; Brands, J.;
Devine, P. N.; Huisman, G. W.; Hughes, G. J. Science 2010, 329, 305.
(46) Wheals, A. E.; Basso, L. C.; Alves, D. M. G.; Amorim, H. V.
Trends Biotechnol. 1999, 17, 482.
(47) Yamashita, M. M.; Wesson, L.; Eisenman, G.; Eisenberg, D.
Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 5648.
(48) Gomez-Tagle, P.; Vargas-Zuniga, I.; Taran, O.; Yatsimirsky, A.
K. J. Org. Chem. 2006, 71, 9713.
(49) Love, K. R.; Bagh, S.; Choi, J.; Love, J. C. Trends Biotechnol.
2013, 31, 16.
REFERENCES
■
(1) Koeller, K. M.; Wong, C. H. Nature 2001, 409, 232.
(2) Schmid, A.; Dordick, J. S.; Hauer, B.; Kiener, A.; Wubbolts, M.;
Witholt, B. Nature 2001, 409, 258.
(3) Gotor-Fernan
40, 111.
́
dez, V.; Brieva, R.; Gotor, V. J. Mol. Catal. B 2006,
(4) The Application of Biotechnology to Industrial Sustainability - A
Primer; OECD Publishing: Paris, France, 2001.
(5) White, N. J. Science 2008, 320, 330.
(6) Schmidt-Dannert, C.; Umeno, D.; Arnold, F. H. Nat. Biotechnol.
2000, 18, 750.
(7) Arnold, F. H. Nature 2001, 409, 253.
(8) Joseph, B.; Ramteke, P. W.; Thomas, G. Biotechnol. Adv. 2008, 26,
457.
(9) Nielsen, J.; Fussenegger, M.; Keasling, J.; Lee, S. Y.; Liao, J. C.;
Prather, K.; Palsson, B. Nat. Chem. Biol. 2014, 10, 319.
(10) Reetz, M. T. J. Am. Chem. Soc. 2013, 135, 12480.
(11) Kim, J.; Jia, H. F.; Wang, P. Biotechnol. Adv. 2006, 24, 296.
(12) Klibanov, A. M. Nature 2001, 409, 241.
(13) Serdakowski, A. L.; Dordick, J. S. Trends Biotechnol. 2008, 26,
48.
(14) Perry, L. J.; Wetzel, R. Science 1984, 226, 555.
(15) Zhou, Z.; Hartmann, M. Chem. Soc. Rev. 2013, 42, 3894.
(16) Bommarius, A. S.; Paye, M. F. Chem. Soc. Rev. 2013, 42, 6534.
(17) Rodrigues, R. C.; Berenguer-Murcia, A.; Fernandez-Lafuente, R.
Adv. Synth. Catal. 2011, 353, 2216.
(18) Moore, J. C.; Arnold, F. H. Nat. Biotechnol. 1996, 14, 458.
(19) Lipovsek, D.; Antipov, E.; Armstrong, K. A.; Olsen, M. J.;
Klibanov, A. M.; Tidor, B.; Wittrup, K. D. Chem. Biol. 2007, 14, 1176.
(20) Giger, L.; Caner, S.; Obexer, R.; Kast, P.; Baker, D.; Ban, N.;
Hilvert, D. Nat. Chem. Biol. 2013, 9, 494.
(21) Turner, N. J. Nat. Chem. Biol. 2009, 5, 567.
G
dx.doi.org/10.1021/ja506346f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX