EnzymaticMovers are a subclass of Mover specifically designed to simulate the action of a virtual enzyme on a substrate, the Pose. The enzyme may be a biologically real enzyme or entirely hypothetical or constructed.

When the Mover is "applied", it will search the Pose's sequence for a particular sequence site at which to make its modification, based on enzyme data. For each site found, it will either make the modification or not, based upon its set "efficiency".

## Description

The main difference between most Movers and an EnzymaticMover is that a standard Mover makes a change to the 3D conformation of a Pose, but the identity of that Pose's Residues remain unchanged. An EnzymaticMover, on the other hand, does not so much "move" a Pose as "change" it, changing its sequence in some way and turning it into a new molecule or set of molecules. An EnzymaticMover can add, delete, or modify Residues or join or split chains within a Pose—mirroring the same sorts of things that an enzyme might do to a biopolymer in vivo.

Another thing to consider is that most Movers are used in protocols to ultimately generate a large number of decoys. A more common usage of an EnzymaticMover might be to initially generate a large number of biologically relevant starting variations of a Pose. For example, consider that most proteins are glycosylated in vivo, yet most crystal structures of proteins do not have the glycans included in the structure. An EnzymaticMover could be used to generate a biologically relevant ensemble of glycosylated variants of the Pose to use for further modeling applications.

A key feature of the EnzymaticMover framework is the potential use of biological enzymatic data. The Rosetta database will contain data for specific, real-world enzymes, each with their own unique consensus sequences and co-substrates. For example, some kinases are specific for phosphorylating serine residues, while others will phosphorylate both serine and threonine residues. The specific enzyme desired can be set by various methods to ensure that the phosphorylation patterns desired by the current modeler result and match what would be seen in Nature.

Similarly, specific species can be passed to an EnzymaticMover. This feature could allow one to write a protocol to generate an assortment of post-translationally modified proteins given a variety of enzymes common to, say, E. coli.

The EnzymaticMover framework was conceived of and implemented by Jason W. Labonte JWLabonte@jhu.edu. Please contact him with any questions or criticism.

## Types

Currently, four EnzymaticMovers are written, but many more can and will be added to the Rosetta code base.

• DNAMethyltransferaseMover
Simulates the activity of specific biological DNA methyltransferases by methylating a DNA-containing Pose.
• GlycosyltransferaseMover
Simulates the activity of specific biological glycosyltransferases and oligosaccharyltrasferases by glycosylating a Pose.
• KinaseMover
Simulates the activity of specific biological kinases by phosphorylating a Pose.
• NTerminalAcetyltransferaseMover
Simulates the activity of specific biological N-terminal acetyltransferases by acetylating a Pose at the N-terminus.

## Usage

For the most part, EnzymaticMovers work like any other Mover, and any of the three main Rosetta interfaces can be used. The major difference is that an EnzymaticMover relies on the presence of enzyme data in the database. If you wish to use an EnzymaticMover for a particular enzyme, ensure that the data for that enzyme is present in the database! (See below for example enzyme data files.)

### C++ & PyRosetta Code

After instantiating an EnzymaticMover, the methods set_species() and set_enzyme() can be used to set the specific enzyme species and enzyme name, respectively.

set_efficiency() can be used to override the efficiency of the enzyme as provided by the enzyme file in the database. A value of 1.00 corresponds to 100%. If set to 0.5 for example, the Mover will only make a change to any positions 50% of the time.

exclude_site() and set_excluded_sites() can be used to pass the sequence position(s) of (a) site(s) that cannot be modified. Perhaps there is a known interface with another protein, for example. ensure_site() and set_ensured_sites() work in the opposite manner, forcing a modification. (Note that ResidueSelectors do not currently work with EnzymaticMovers, but this will be changed in the future.)

perform_major_reaction_only and perform_all_reactions toggle the behavior of promiscuous enzymes.

#### PyRosetta Example

from pyrosetta.rosetta.protocols.enzymatic_movers import KinaseMover

general_enzyme = KinaseMover()
general_enzyme.set_efficiency(0.25)
general_enzyme.apply(pose1)

specific_enzyme = KinaseMover()
specific_enzyme.set_species("h_sapiens")
specific_enzyme.set_enzyme("CLK1")
specific_enzyme.apply(pose2)

#### C++ Example

#include <protocols/enzymatic_movers/GlcyosyltransferaseMover.hh>

glycosyltransferase =
enzymatic_movers::GlycosyltransferaseMoverOP( utility::pointer::make_shared< enzymatic_movers::GlycosyltransferaseMover >() );
glycosyltransferase->set_species( "c_jejuni" );
glycosyltransferase->set_enzyme( "PglB" );
glycosyltransferase->set_efficiency( 1.0 );
glycosyltransferase.apply( pose );

### RosettaScripts

In RosettaScripts, the interface with any EnzymaticMover is through the <MOVERS> XML tag.

• The name parameter of a EnzymaticMover is for providing a unique name to represent a given EnzymaticMover elsewhere in the script.
• The species parameter is for setting the species name of the simulated enzyme.
• The enzyme_name parameter is for setting the specific name of the simulated enzyme.
• The efficiency parameter is to directly set the efficiency of the enzyme, ignoring whatever is in the database.
• The perform_major_reaction_only parameter is to set the EnzymaticMover to perform only its major reaction, using only the first cosubstrate listed in its enzyme data file.
• The perform_all_reactions parameter is for allowing the EnzymaticMover to be promiscuous, performing a random transfer from among its possible co-substrates. This is the default behavior.

#### Example RosettaScripts XML

<ROSETTASCRIPTS>
<MOVERS>
<KinaseMover name="kinase" species="h_sapiens" perform_major_reaction_only="true" />
</MOVERS>
<PROTOCOLS>
</PROTOCOLS>
<OUTPUT />
</ROSETTASCRIPTS>

### Command-Line Applications

Three Rosetta options flags are used specifically for interfacing with EnzymaticMovers used in any protocols.

• -enzymes:species is used to set the species name of any simulated enzymes used by the protocol.
• -enzymes:enzyme is used to set the specific name of any simulated enzymes used by the protocol.
• -enzymes:efficiencty is used to override the efficiency of any simulated enzymes used by the protocol.

#### Example Command Lines

$DNA_methylation -s input/1ABC.pdb -enzymes:species h_sapiens -enzymes:efficiency 0.75 -nstruct 3$ glycosyltransfer -s input/2DEF.pdb -include_sugars -enzymes:species h_sapiens -enzymes:enzyme OGT -nstruct 5

$phosphorylation -s input/3GHI.pdb -enzymes:species h_sapiens -nstruct 1$ N-terminal_acetylation -s input/4JKL.pdb -nstruct 1

## Enzyme Data Files

EnzymaticMovers rely on the database to function properly. All enzyme data files have the following format:

# Consensus  Sequence  Residue of    Atom
# Sequence   Type      CS to Modify  To Modify  Efficiency
TARGET     AA        4             CA         1.00        # Rosettase perfectly appends a glycine in the target sequence with a branch.

# Co-substrates
ARROW
BULLET
BOLT
KNIFE
PROTONTORPEDO

The first line is assumed to contain a whitespace-delimited list of the following, all of which are required:

• Consensus sequence — This may be a 1-letter-code AA or NA sequence or an IUPAC carbohydrate sequence.
• Amino-acid Residue Sequences
• The parser recognizes the IUPAC-approved one-letter codes B, J, O, U, and Z, which code for Asx, Xle, Pyl, Sec, and Glx, respectively.
• X alone is recognized to be any of the 20 canonical amino acids; X followed by square brackets specifies a single non-canonical amino acid by 3-letter code. For example, X[SEP] specifies phosphoserine.
• Parentheses are used to specify multiple possible residue types at that site, separated by forward slashes, e.g., (A/G) specifies either Ala or Gly at that position.
• A < in the first position indicates that the sequon must be located at the N-terminus.
• A > in the final position indicates that the sequon must be located at the C-terminus.
• Nucleic-acid Residue Sequences
• In addition to the standard A, C, G, T, and U one-letter codes, the parser recognizes B, D, H, and V for not A, C, G, or U, respectively; K for G or U; M for A or C; N for any of the four RnA bases; R for any of the puRines; S for any of the "Strong" nucleobases; and W for any of the "Weak" ones.
• Sequence type — This value must be AA, NA, or SACCHARIDE, for the three types of sequences accepted.
• Residue of CS to modify — An integer representing the sequence position of the residue in the consensus sequence to be modified in some way by the Mover.
• Atom to modify — A string for the atom name of the atom to be modified, if applicable. (Not all EnzymaticMovers will need this information and are allowed to ignore it, but a string value must be present.)
• Efficiency — A real value, where 1.00 is 100% efficiency. If set to 0.50, an EnzymaticMover will only perform its modification 50% of the time. If not set to 1.00, please provide a comment of the source for the value used.

Any further lines are assumed by the database reader to be cosubstrates. This is usually specified in the form of a sequence of some sort.

If no further lines are provided, the enzyme either does not have a cosubstrate or else the cosubstrate is not needed by Rosetta to perform the modification. (For example, a kinase does not need to provide ATP as a cosubstrate, because, under the hood, Rosetta will simply modify a Residue to convert it into a phosphorylated VariantType. Whereas, a ligase would need to know the sequence of the structure to be joined to the Pose.)

If multiple cosubstrate lines are provided, this means that the enzyme being simulated is promiscuous, that is, it will randomly select from the options when performing the reaction.

## For Developers

### Nomenclature

Any EnymaticMover should be named as an enzyme (ending in "-ase") followed by "Mover". For example, RosettaseMover.

### C++ Code

Any Mover inheriting from the EnzymaticMover base class must:

• ...provide an enzyme family corresponding to a directory of enzyme data.
The enzyme family is passed as an argument to the EnzymaticMover constructor in the child class's own constructor. For example:

KinaseMover::KinaseMover(): EnzymaticMover( "kinases" )
{
type( "KinaseMover" );
}
• ...use EnzymaticMover's xml_schema_complex_type_generator() method to provide the XML schema for RosettaScripts. For example:

void
{
using namespace utility::tag;

EnzymaticMover::xml_schema_complex_type_generator()->element_name( mover_name() )
.complex_type_naming_func( & moves::complex_type_name_for_mover )
.description( "Enzymatic mover to connect two DNA Poses together." )
.write_complex_type_to_schema( xsd );
}
• ...implement the protected perform_reaction() method, which modifies, adds, or removes (a) Residue(s).
For example:

void
GlycosyltransferaseMover::perform_reaction(
core::pose::Pose & input_pose,
core::uint const site,
std::string const & cosubstrate )
{
core::pose::carbohydrates::glycosylate_pose(
input_pose,
get_reactive_site_sequence_position( site ),
get_reactive_site_atom_name( site ),
cosubstrate );
}

Two EnzymaticMover methods are very helpful for performing the actual reaction, as shown in the example above: get_reactive_site_sequence_position and get_reactive_site_atom_name.

The rest of the code for any EnzymaticMover should be simple "boiler plate" code. The core machinery of the base class uses enzymatic data found in the database to search for potential reaction sites. The EnzymeManager singleton class is used to lazily load enzyme data in a thread-safe manner as needed by any EnzymaticMover.

### Database Structure

All enzyme data for EnzymaticMovers should be located in the database/virtual_enzymes/ folder.

Every child EnzymaticMover should have a subdirectory corresponding to the enzyme family of that Mover, which must match the family provided to the EnzymaticMover constructor. (See above.) For example, data for the MethylaseMover should be stored in database/virtual_enzymes/methylases/.

Directories for each enzyme family include no files but only subdirectories corresponding to species. Every enzyme family directory must, at minimum have an h_sapiens/ subdirectory, because "h_sapiens" is encoded as the default species for any EnzymaticMover. All other subdirectories are for enzymes from non-human sources.

Every species subdirectory must include a generic enzyme data file as a default example of this this family of enzymes. Generally, this data will provide a minimal sequon, have 100% efficiency, and not be promiscuous.

Documentation created 5 April 2019 by Jason W. Labonte JWLabonte@jhu.edu. Documentation updated 26 February 2020 by Jason W. Labonte JWLabonte@jhu.edu.