Download the full version of the ebook at ebookfinal.

com

RNA Scaffolds Methods and Protocols 2nd Edition

Luc Ponchon

https://ebookfinal.com/download/rna-scaffolds-methods-and-
protocols-2nd-edition-luc-ponchon/

OR CLICK BUTTON

DOWNLOAD EBOOK

Download more ebook instantly today at https://ebookfinal.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

RNA RNA Interactions Methods and Protocols 1st Edition

Frank J. Schmidt (Eds.)

https://ebookfinal.com/download/rna-rna-interactions-methods-and-
protocols-1st-edition-frank-j-schmidt-eds/

ebookfinal.com

RNA Nanostructures Methods and Protocols 1st Edition

Eckart Bindewald

https://ebookfinal.com/download/rna-nanostructures-methods-and-
protocols-1st-edition-eckart-bindewald/

ebookfinal.com

RNA Silencing Methods and Protocols 1st Edition Mikiko C.

Siomi

https://ebookfinal.com/download/rna-silencing-methods-and-
protocols-1st-edition-mikiko-c-siomi/

ebookfinal.com

RNA and DNA Editing Methods and Protocols 1st Edition H.

Ulrich Göringer

https://ebookfinal.com/download/rna-and-dna-editing-methods-and-
protocols-1st-edition-h-ulrich-goringer/

ebookfinal.com
DNA and RNA Profiling in Human Blood Methods and Protocols
1st Edition Maryse St-Louis (Auth.)

https://ebookfinal.com/download/dna-and-rna-profiling-in-human-blood-
methods-and-protocols-1st-edition-maryse-st-louis-auth/

ebookfinal.com

Pharmacogenomics Methods and Protocols 2nd Edition Rosane

Charlab

https://ebookfinal.com/download/pharmacogenomics-methods-and-
protocols-2nd-edition-rosane-charlab/

ebookfinal.com

Salmonella Methods and Protocols 2nd Edition Heide

Schatten

https://ebookfinal.com/download/salmonella-methods-and-protocols-2nd-
edition-heide-schatten/

ebookfinal.com

Biological Aging Methods and Protocols 2nd Edition Huaping

Chen

https://ebookfinal.com/download/biological-aging-methods-and-
protocols-2nd-edition-huaping-chen-2/

ebookfinal.com

Pancreatic Cancer Methods and Protocols 2nd Edition Hanno

Matthaei

https://ebookfinal.com/download/pancreatic-cancer-methods-and-
protocols-2nd-edition-hanno-matthaei/

ebookfinal.com
Methods in
Molecular Biology 2323

Luc Ponchon Editor

RNA Scaffolds
Methods and Protocols
Second Edition
METHODS IN MOLECULAR BIOLOGY

Series Editor
John M. Walker
School of Life and Medical Sciences
University of Hertfordshire
Hatfield, Hertfordshire, UK

For further volumes:

http://www.springer.com/series/7651
For over 35 years, biological scientists have come to rely on the research protocols and
methodologies in the critically acclaimed Methods in Molecular Biology series. The series was
the first to introduce the step-by-step protocols approach that has become the standard in all
biomedical protocol publishing. Each protocol is provided in readily-reproducible step-by-
step fashion, opening with an introductory overview, a list of the materials and reagents
needed to complete the experiment, and followed by a detailed procedure that is supported
with a helpful notes section offering tips and tricks of the trade as well as troubleshooting
advice. These hallmark features were introduced by series editor Dr. John Walker and
constitute the key ingredient in each and every volume of the Methods in Molecular Biology
series. Tested and trusted, comprehensive and reliable, all protocols from the series are
indexed in PubMed.
 RNA Scaffolds

Methods and Protocols

Second Edition

Edited by

Luc Ponchon
Faculté de Pharmacie, University of Paris, PARIS, France
Editor
Luc Ponchon
Faculté de Pharmacie
University of Paris
PARIS, France

ISSN 1064-3745 ISSN 1940-6029 (electronic)

Methods in Molecular Biology
ISBN 978-1-0716-1498-3 ISBN 978-1-0716-1499-0 (eBook)
https://doi.org/10.1007/978-1-0716-1499-0

© Springer Science+Business Media, LLC, part of Springer Nature 2021

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction
on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation,
computer software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations
and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to
be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been
made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer
Nature.
The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.
Preface

RNAs are at the center of numerous cellular phenomena and play very different roles in each.
One of their roles is in particular that of organization center: the ribosome, RNA telome-
rase, and “Long Noncoding RNAs” are, among others, examples of RNA structures that
recruit other molecules and organize biological processes. These RNAs possess structures
allowing interactions with other molecules (proteins, ligands) and thus will potentialize
molecular reactions. Advances in structural biology have permitted a definition of the rules
with regard to the folding of RNA allowing us today to better understand exactly how they
fold and interact.
As opposed to DNA, RNAs are able to adopt very variable folds and therefore are able to
adopt ligand-specific structures. Contrary to proteins, we are able to create structures
composed of different “RNA” modules, each of which is able to keep its activity indepen-
dent from the others.
So, when they are stable, folded RNA can be used as a tool for biological, pharmaco-
logical, and/or molecular design studies. RNA presents the peculiarity, like Meccano, of
being able to fold into structural domains, which can assemble and sometimes form
supramolecular objects. We can isolate, modify, or create an RNA template de novo to
make use of its recognition or enzymatic functions. From my point of view, an “RNA
scaffold” is a synthetic or natural RNA whose structure, for example, allows one to optimize
a reaction, to isolate a molecule, or to favor an interaction. Like Biobricks, the tools based on
RNA scaffolds are an example of the emergence of synthetic biology. Indeed, they partici-
pate in the creation and construction of biological objects and systems for useful purposes.
In this volume, we have tried to be as representative as possible of that which is done
today. You will find detailed here processes and techniques that differ greatly from one to
another. This book reviews recently developed techniques that use “RNA scaffolds” as
molecular tools. These methods cover domains as various as molecular biology, cellular
biology, nanotechnology, and structural biology. This book is composed of original chapters
and updated chapters from the previous edition.

Contents

In order to design a scaffold or understand its interaction with a ligand, it is sometimes

necessary to possess certain structural data. A structure can be modeled from an RNA
structure data bank. In the first chapter, Chen and his colleagues describe their prediction
method, Vfold, that, from a primary sequence, allows one to determine a three-dimensional
model of an RNA. This method is based on a pattern-based approach. Thus, they extract
different patterns (such as hairpin loops, internal loops, pseudoknot loops, and three-way
junctions) from the two-dimensional structure. Based on this data, the implemented algo-
rithm calculates a three-dimensional model. In this chapter, they describe a hybrid method,
which combines the motif template-based Vfold3D model and the loop template-based
VfoldLA model, to predict RNA 3D structures. In Chapter 2, Sargueil and his colleague
present a chemical footprinting procedure. Using small chemical reagents as a probe, this
method allows the identification of the interaction site of a ligand with RNA but also RNA

v
vi Preface

structural rearrangement upon ligand binding. Chemical footprinting could be a good

alternative to RNAse footprinting.
However, if one seeks to obtain an RNA structure at an atomic level (e.g., for the
rationalized design of a scaffold), crystallography is the technique of choice. RNA crystal-
logenesis remains nonetheless complicated. Indeed, RNAs, as opposed to proteins, do not
crystallize as easily and the X-ray diffraction is often weak. In Chapter 3, Ferré d’Amaré and
his colleagues describe techniques allowing an improvement in data acquisition and notably
for large RNA (over 100 nucleotides) for which only very few structures exist. They propose
a protocol allowing the increase in diffraction power of the crystals by combining two
techniques: the first consisting of dehydrating the crystal, the second of substituting the
divalent ions with strontium ions. Through the example of a gene regulatory tRNA-mRNA
complex, they show us the importance of these two techniques on the quality of diffraction.
Ke and his colleagues present a method to obtain RNA crystals. Indeed, RNA flexibility
is one of the limiting factors to RNA crystallization. To overcome this difficulty, the RNA is
embedded in a tRNA scaffold. The tRNA stabilizes the chimera and increases the confor-
mation purity of the RNA target. Using a tRNAgly scaffold, they prove that this method can
assist crystallization and phase determination.
RNA scaffold can adopt different conformations. The molecular dynamics simulation is
one of the tools to investigate these conformations. In Chapter 5, Pasquali and her colleague
present a method to identify alternative RNA structures using a particular computational
potential energy landscape framework. With the example of the 50 -hairpin of RNA 7SK they
illustrate how the method can be applied to interpret experimental results and to obtain a
detailed description of molecular properties.
A limiting factor for RNA study (notably for long RNAs) is obtaining them in large
quantities. Three classic methods exist in order to obtain the RNA: in vitro transcription,
chemical synthesis, and cellular extraction.
The following chapters describe the techniques permitting, as with proteins, the in vivo
production of RNA. Chapter 6 describes a system of RNA-protein coproduction in the
bacteria. This protocol is the logical progression of protocols already described in a previous
book Recombinant and In Vitro RNA Synthesis in the same collection. In the aforemen-
tioned book, the authors describe the production of RNA protected from bacterial nuclease
by a “tRNA” camouflage. The RNA is protected at its extremities by the tRNA chassis and
accumulated in the bacteria. In this protocol, they propose to coproduce RNA–protein
complexes and thus show the advantage of having a joint production of the two molecules.
Fox and his colleagues offer a method also permitting the overproduction of RNA in
bacteria in Chapter 7. They use a different chassis: the RNA ribosomal 5S. Indeed, like the
tRNA, the RNA 5S possesses a fold that allows it to be accumulated in the bacteria. Their
protocol thus goes into detail on the production of different RNAs and how to make the
cleave. So as to liberate the RNA of interest from the 5S, they propose a cleavage via the
DNAzyme. These short sequences of DNA hybridize themselves with the RNA to form a
ribozyme-like structure and thus allow the RNA cleavage.
In Chapter 8, Daros and his colleagues developed a third method to overexpress RNAs
in E. coli. This method is adapted from a viroid system. The RNA of interest is flanked by
domains of the viroid hammerhead ribozyme and is coproduced with a tRNA ligase result-
ing in circular RNAs. Because of the circular structure, the viroid scaffold facilitates the
accumulation of RNAs in the bacteria but also the purification steps.
Affinity or “pull-down” techniques allow the identification of molecular complexes. In
these techniques, a protein linked to a matrix serves as bait. From the cellular extracts, one
 Preface vii

can isolate the linked molecules. Martinez-Salas and his colleagues propose a pull-down
system in which RNA serves as bait in order to identify the IRES-binding proteins. To
render the system specific and robust, their RNA bait is embedded in a tRNA scaffold and
possesses an aptamer for the streptavidin in order to isolate IRES–protein complexes from
extracts of eukaryotic cells.
Fluorescence remains a technique of choice to perform cellular localization. Green
fluorescent protein (GFP) and its numerous derivatives allow one to localize the proteins
in cells or tissues. However, it remains difficult to specifically locate the RNA as no naturally
fluorescent RNA has been identified to date. It is therefore necessary to use indirect
techniques. The following chapters describe RNA scaffolds allowing one to make cellular
localization of RNA or metabolites through fluorescence.
In Chapter 10, Hammond and his colleagues propose to us a technique allowing the
localization of a metabolite via an RNA scaffold, which is linked to a fluorescent chromo-
phore. This RNA is a biosensor composed, on the one hand, of a “spinach aptamer” which
links a fluorescent chromophore to DFHB1 and, on the other hand, of a specific riboswitch
of the cyclic di-GMP. The idea being that the fixation of the fluorophore derives from the
fixation of the di-GMP to its aptamer.
In the next chapter, You and his colleagues describe the steps to rationally design,
optimize, and apply fluorogenic RNA-based sensors. Through the example of the intracel-
lular imaging of tetracycline in living E. coli cells, they show us how the recognition module
induces a duplex formation of the transducer module, which further folds the fluorescence
module, i.e., Broccoli, and activates the fluorescence of the fluorophore (DFHB1-1T).
In Chapter 12, Winkler and his colleagues use a system composed of a riboswitch of the
cyclic di-GMP upstream of a gene reporter. The metabolite-sensing riboswitch controls the
expression of Yellow Fluorescent Protein and allows to determine the relative abundance of
di-GMP in the cells by fluorescent microscopy. They present an example of their method
apply in Bacillus subtilis.
In the chapter titled “FRET Analysis of RNA-Protein Interactions Using Spinach
Aptamers,” Hennig and his colleague present a method to analyze direct RNA-protein
interaction using fluorescent light-up aptamers (FLAPs), and fluorescent proteins for the
detection and quantification of a direct RNA-protein interaction. They describe the design
and application of a homogenous assay to observe and quantify the interaction of the
Pseudomonas aeruginosa bacteriophage coat protein 7 (PP7) with its cognate RNA
sequence (pp7-RNA) using the Spinach-DFHBI aptamer as RNA fusion and the red
fluorescent mCherry as protein fusion.
RNA in vitro evolution or SELEX enables the artificial evolution and selection of RNA
molecules that possess a desired property, such as binding affinity for a particular ligand or an
activity such as that of an enzyme or catalyst. The first such selections involved isolation of
various aptamers that bind to small molecules. The first catalytic RNAs produced by in vitro
evolution were RNA ligases, catalytic RNAs that join two RNA fragments to produce a
single adduct.
Certain RNAs can see their activity regulated by a ligand link like a small molecule or
another RNA in the case of aptazymes (allosteric ribozymes). The following two chapters
(14 and 15) describe two methods allowing the identification of riboswitches from a random
bank of aptazymes. These systems are logic gates; indeed, the aptazymes are composed of an
aptamer RNA at a strategic position with regard to the self-cleaving ribozyme so that the
structure of the ribozyme is stabilized or destabilized through the link of the ligand. These
aptazymes are in the 50 position of an ARNm (coding for a reporter gene), which will or will
viii Preface

not be degraded based on the efficiency of the riboswitch. Hartig and his colleagues have
thus inserted random sequences of aptazymes in the 50 -UTR of the hRluc reporter gene on
the plasmid psi-CHECK2 making the expression of the luciferase dependent on the ligand.
As for Yokobayashi and his colleagues, they also work in the area of mammalian cells but use
the reporter system EGFP. They describe a protocol allowing the screening, at a medium
rate, around a hundred aptazymes directly into mammalian cells.
Many RNAs, in particular, can assemble themselves in the form of nano-structures. We
understand the determinants that govern the structure and the layout of these objects better
and better. In the following chapter, a protocol allowing one to design and produce different
RNA nano-structures is described. Saito and his colleagues show us how to design RNA
origami. They thus manage to give their RNAs a triangular structure. The angles are
produced by the interaction of the L7Ae protein with recognition sequences placed in
the RNA.
The final chapters are devoted to an RNA scaffold that serves as a genetic tool. “Gene
silencing” can be induced by small noncoding antisense RNAs, which hybridize on an RNA
messenger preventing the gene translation. Silencing RNAs (siRNAs) are thus small coding
sequences allowing, in genetics, the switching off of a gene specifically and thereby under-
standing its function. In Chapter 17, Lee and his collaborators describe a method to us that
allows the potentialization of the silencing activity of an siRNA. The siRNA is surrounded by
two structures in a stem loop, which increase the half-life of the RNA and thus increase the
silencing. This structure that they have named afsRNA (artificial small regulatory RNAs)
could thus become a more efficient silencing tool.
Yu and his colleagues choose another way to overexpress effective small RNAs using
bioengineered RNA agents (BERAs) carrying warhead miRNAs, siRNAs, aptamers, or other
forms of small RNAs. RNAs are produced in the cells because of an optimal hybrid tRNA/
pre-miRNA carrier. This system allows large-scale production of effective RNAs and their
purification by fast protein liquid chromatography (FPLC) to a high degree of homogeneity.
In Chapter 19, Li and his colleagues describe a bacteria-based cancer immunotherapy
for cancer treatment. They have reported that bacteria can be engineered using synthetic
biology technology to enable nonpathogenic bacteria to express gene silencer, invading
transformed cells, escaping endosome and selective silence oncogenes in mammalian cells
(termed transkingdom gene silencing or tkRNAi). In this chapter, they describe a novel
transkingdom gene silencing vector capable of constitutively expressing long double-
stranded RNAs enhancing target gene silencing. This approach can be adapted to multi-
target gene silencing.
In conclusion, I would like to thank all of the authors who have allowed this book to be
published. Their work shows to what extent RNAs are fascinating and interest researchers in
very different areas. Naturally, we are a long way from exploring their full potential and I
hope that the reading of this book will inspire its readers, especially young researchers, to
further study of this area.

Paris, France Luc Ponchon

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

1 Predicting RNA Scaffolds with a Hybrid Method of Vfold3D

and VfoldLA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Xiaojun Xu and Shi-Jie Chen
2 RNA Footprinting Using Small Chemical Reagents . . . . . . . . . . . . . . . . . . . . . . . . . 13
Grégoire De Bisschop and Bruno Sargueil
3 Improving RNA Crystal Diffraction Quality by Postcrystallization
Treatment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Jinwei Zhang and Adrian R. Ferré-D’Amaré
4 Using tRNA Scaffold to Assist RNA Crystallization . . . . . . . . . . . . . . . . . . . . . . . . . 39
Changrui Lu, Rujie Cai, Jason C. Grigg, and Ailong Ke
5 RNA Modeling with the Computational Energy Landscape Framework . . . . . . . 49
Konstantin Röder and Samuela Pasquali
6 Coexpression and Copurification of RNA–Protein Complexes
in Escherichia coli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Margot El Khouri, Marjorie Catala, Bili Seijo, Johana Chabal,
Frédéric Dardel, Carine Tisné, and Luc Ponchon
7 In Vivo Production of Small Recombinant RNAs Embedded in
5S rRNA-Derived Protective Scaffold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Victor G. Stepanov and George E. Fox
8 Production of Circular Recombinant RNA in Escherichia coli
Using Viroid Scaffolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
José-Antonio Daròs
9 Identification of RNA-Binding Proteins Associated to RNA
Structural Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Javier Fernandez-Chamorro, Rosario Francisco-Velilla,
Azman Embarc-Buh, and Encarnacion Martinez-Salas
10 Live Cell Imaging Using Riboswitch–Spinach tRNA Fusions as
Metabolite-Sensing Fluorescent Biosensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Sudeshna Manna, Colleen A. Kellenberger, Zachary F. Hallberg,
and Ming C. Hammond
11 Rational Design of Allosteric Fluorogenic RNA Sensors for
Cellular Imaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Qikun Yu, Ru Zheng, Manojkumar Narayanan, and Mingxu You
12 Riboswitch-Mediated Detection of Metabolite Fluctuations During
Live Cell Imaging of Bacteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Cordelia A. Weiss and Wade C. Winkler
13 FRET Analysis of RNA–Protein Interactions Using Spinach Aptamers. . . . . . . . . 171
Laura Gerhard and Sven Hennig

ix
x Contents

14 Engineering Aptazyme Switches for Conditional Gene Expression

in Mammalian Cells Utilizing an In Vivo Screening Approach . . . . . . . . . . . . . . . . 199
Charlotte Rehm, Benedikt Klauser, Monika Finke, and Jörg S. Hartig
15 Aptazyme-Based Riboswitches and Logic Gates in Mammalian Cells . . . . . . . . . . 213
Yoko Nomura and Yohei Yokobayashi
16 Folding RNA–Protein Complex into Designed Nanostructures. . . . . . . . . . . . . . . 221
Tomonori Shibata, Yuki Suzuki, Hiroshi Sugiyama, Masayuki Endo,
and Hirohide Saito
17 An Effective Method for Specific Gene Silencing in Escherichia coli
Using Artificial Small RNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Geunu Bak, Jee Soo Choi, Wonkyeong Kim, Shinae Suk,
and Younghoon Lee
18 Expression and Purification of tRNA/pre-miRNA-Based
Recombinant Noncoding RNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
Mei-Juan Tu, Halley K. Wright, Neelu Batra, and Ai-Ming Yu
19 Synthetic Biology Medicine and Bacteria-Based Cancer Therapeutics. . . . . . . . . . 267
Jaehyung Lee, Andrew C. Keates, and Chiang J. Li

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Contributors

GEUNU BAK • Department of Chemistry, KAIST, Daejeon, South Korea

NEELU BATRA • Department of Biochemistry and Molecular Medicine, UC Davis School of
Medicine, Sacramento, CA, USA
RUJIE CAI • College of Chemistry, Chemical Engineering and Biotechnology, Donghua
University, Shanghai, China
MARJORIE CATALA • Expression génétique microbienne, UMR CNRS 8261, Institut de
biologie physico-chimique, Université de Paris, Paris, France
JOHANA CHABAL • EryPharm, Paris, France
SHI-JIE CHEN • Department of Physics, Department of Biochemistry, and Institute for Data
Science and Informatics, University of Missouri, Columbia, MO, USA
JEE SOO CHOI • Department of Chemistry, KAIST, Daejeon, South Korea
FRÉDÉRIC DARDEL • CiTCoM, UMR CNRS 8038, Université de Paris, Paris, France
JOSÉ-ANTONIO DARÒS • Instituto de Biologı́a Molecular y Celular de Plantas (Consejo
Superior de Investigaciones Cientı́ficas-Universitat Politècnica de València), Valencia,
Spain
GRÉGOIRE DE BISSCHOP • CiTCOM, Cibles Thérapeutiques et conception de médicaments,
CNRS, Université de Paris, Paris, France; Institut de Recherches Cliniques de Montréal
(IRCM), Montréal, QC, Canada
MARGOT EL KHOURI • Centre for Molecular Oncology, Barts Cancer Institute, Queen Mary
University of London, London, UK
AZMAN EMBARC-BUH • Centro de Biologı́a Molecular Severo Ochoa, CSIC-UAM, Madrid,
Spain
MASAYUKI ENDO • Institute for Integrated Cell-Material Sciences (WPI-iCeMS), Kyoto
University, Kyoto, Japan
JAVIER FERNANDEZ-CHAMORRO • Centro de Biologı́a Molecular Severo Ochoa, CSIC-UAM,
Madrid, Spain; Garvan Institute of Medical Research, Darlinghurst, NSW, Australia
ADRIAN R. FERRÉ-D’AMARÉ • Biochemistry and Biophysics Center, National Heart, Lung
and Blood Institute, Bethesda, MD, USA
MONIKA FINKE • Department of Chemistry and Konstanz Research School Chemical Biology,
University of Konstanz, Konstanz, Germany
GEORGE E. FOX • Department of Biology and Biochemistry, University of Houston, Houston,
TX, USA
ROSARIO FRANCISCO-VELILLA • Centro de Biologı́a Molecular Severo Ochoa, CSIC-UAM,
Madrid, Spain
LAURA GERHARD • Division of Organic Chemistry, Department of Chemistry and
Pharmaceutical Sciences, Vrije Universiteit Amsterdam, Amsterdam, Noord-Holland, The
Netherlands
JASON C. GRIGG • Department of Microbiology and Immunology, Life Sciences Institute, The
University of British Columbia, Vancouver, BC, Canada
ZACHARY F. HALLBERG • Department of Chemistry, University of California, Berkeley,
Berkeley, CA, USA
MING C. HAMMOND • Department of Chemistry, University of Utah, Salt Lake City, UT,
USA; Henry Eyring Center for Cell and Genome Science, University of Utah, Salt Lake

xi
xii Contributors

City, UT, USA; Department of Chemistry, University of California, Berkeley, Berkeley,

CA, USA
JÖRG S. HARTIG • Department of Chemistry and Konstanz Research School Chemical Biology,
University of Konstanz, Konstanz, Germany
SVEN HENNIG • Division of Organic Chemistry, Department of Chemistry and
Pharmaceutical Sciences, Vrije Universiteit Amsterdam, Amsterdam, Noord-Holland, The
Netherlands
AILONG KE • Department of Molecular Biology and Genetics, Cornell University, Ithaca, NY,
USA
ANDREW C. KEATES • Skip Ackerman Center for Molecular Therapeutics, Division of
Gastroenterology and Hepatology, Beth Israel Deaconess Medical Center, Harvard Medical
School, Boston, MA, USA
COLLEEN A. KELLENBERGER • Department of Chemistry, University of California, Berkeley,
Berkeley, CA, USA
WONKYEONG KIM • Department of Chemistry, KAIST, Daejeon, South Korea
BENEDIKT KLAUSER • Department of Chemistry and Konstanz Research School Chemical
Biology, University of Konstanz, Konstanz, Germany
JAEHYUNG LEE • Skip Ackerman Center for Molecular Therapeutics, Division of
Gastroenterology and Hepatology, Beth Israel Deaconess Medical Center, Harvard Medical
School, Boston, MA, USA
YOUNGHOON LEE • Department of Chemistry, KAIST, Daejeon, South Korea
CHIANG J. LI • Skip Ackerman Center for Molecular Therapeutics, Division of
Gastroenterology and Hepatology, Beth Israel Deaconess Medical Center, Harvard Medical
School, Boston, MA, USA
CHANGRUI LU • College of Chemistry, Chemical Engineering and Biotechnology, Donghua
University, Shanghai, China
SUDESHNA MANNA • Department of Chemistry, University of Utah, Salt Lake City, UT,
USA; Henry Eyring Center for Cell and Genome Science, University of Utah, Salt Lake
City, UT, USA
ENCARNACION MARTINEZ-SALAS • Centro de Biologı́a Molecular Severo Ochoa, CSIC-UAM,
Madrid, Spain
MANOJKUMAR NARAYANAN • Department of Chemistry, University of Massachusetts, Amherst,
MA, USA
YOKO NOMURA • Nucleic Acid Chemistry and Engineering Unit, Okinawa Institute of
Science and Technology Graduate University, Onna, Okinawa, Japan
SAMUELA PASQUALI • Laboratoire CiTCoM, CNRS UMR 8038, Université de Paris, Paris,
France
LUC PONCHON • CiTCoM, UMR CNRS 8038, Université de Paris, Paris, France
CHARLOTTE REHM • Department of Chemistry and Konstanz Research School Chemical
Biology, University of Konstanz, Konstanz, Germany
KONSTANTIN RÖDER • Yusuf Hamied Department of Chemistry, University of Chemistry,
Cambridge, UK
HIROHIDE SAITO • Center for iPS Cell Research and Application (CiRA), Kyoto University,
Kyoto, Japan
BRUNO SARGUEIL • CiTCOM, Cibles Thérapeutiques et conception de médicaments, CNRS,
Université de Paris, Paris, France
BILI SEIJO • Faculté de biologie et médecine Ludwig Lausanne Branch, Epalinges,
Switzerland
 Contributors xiii

TOMONORI SHIBATA • Center for iPS Cell Research and Application (CiRA), Kyoto
University, Kyoto, Japan
VICTOR G. STEPANOV • Department of Biology and Biochemistry, University of Houston,
Houston, TX, USA
HIROSHI SUGIYAMA • Department of Chemistry, Graduate School of Science, Kyoto University,
Kyoto, Japan
SHINAE SUK • Department of Chemistry, KAIST, Daejeon, South Korea
YUKI SUZUKI • Department of Chemistry, Graduate School of Science, Kyoto University,
Kyoto, Japan
CARINE TISNÉ • Expression génétique microbienne, UMR CNRS 8261, Institut de biologie
physico-chimique, Université de Paris, Paris, France
MEI-JUAN TU • Department of Biochemistry and Molecular Medicine, UC Davis School of
Medicine, Sacramento, CA, USA
CORDELIA A. WEISS • Department of Cell Biology and Molecular Genetics, University of
Maryland, College Park, MD, USA
WADE C. WINKLER • Department of Cell Biology and Molecular Genetics, University of
Maryland, College Park, MD, USA
HALLEY K. WRIGHT • Department of Biochemistry and Molecular Medicine, UC Davis School
of Medicine, Sacramento, CA, USA
XIAOJUN XU • Institute of Bioinformatics and Medical Engineering, Jiangsu University of
Technology, Changzhou, Jiangsu, China
YOHEI YOKOBAYASHI • Nucleic Acid Chemistry and Engineering Unit, Okinawa Institute of
Science and Technology Graduate University, Onna, Okinawa, Japan
MINGXU YOU • Department of Chemistry, University of Massachusetts, Amherst, MA, USA
AI-MING YU • Department of Biochemistry and Molecular Medicine, UC Davis School of
Medicine, Sacramento, CA, USA
QIKUN YU • Department of Chemistry, University of Massachusetts, Amherst, MA, USA
JINWEI ZHANG • Laboratory of Molecular Biology, National Institute of Diabetes and
Digestive and Kidney Diseases, Bethesda, MD, USA
RU ZHENG • Department of Chemistry, University of Massachusetts, Amherst, MA, USA
 Chapter 1

Predicting RNA Scaffolds with a Hybrid Method of Vfold3D

and VfoldLA
Xiaojun Xu and Shi-Jie Chen

Abstract
The ever-increasing discoveries of noncoding RNA functions draw a strong demand for RNA structure
determination from the sequence. In recently years, computational studies for RNA structures, at both the
two-dimensional and the three-dimensional levels, led to several highly promising new developments. In
this chapter, we describe a hybrid method, which combines the motif template-based Vfold3D model and
the loop template-based VfoldLA model, to predict RNA 3D structures. The main emphasis is placed on
the definition of motifs and loops, the treatment of no-template motifs, and the 3D structure assembly from
templates of motifs and loops. For illustration, we use the ZIKV xrRNA1 as an example to show the
template-based prediction of RNA 3D structures from the 2D structure. The web server for the hybrid
model is freely accessible at http://rna.physics.missouri.edu/vfold3D2.

Key words Structure prediction, Template-assembly, Secondary structural motifs, Single-stranded

loops

1 Introduction

To perform crucial cellular functions, RNA molecules fold up to

form compact three-dimensional (3D) structures [1–5]. A compre-
hensive understanding of RNA structures, therefore, can not only
provide the fundamental insights into the cellular functions of
RNAs but also promote the development of structure-based bioen-
gineering, such as precise gene regulation and editing, as well as
effective drug design. However, experimental determination of
RNA 3D structures is usually laborious and challenging, leading
to the fast development of predictive computational modeling for
RNA 3D structure predictions [6–24]. Structure predictions from
sequence alone may achieve (near) atomic resolutions for most
short RNAs, due to the relatively small degrees of freedom in the
conformational space. Additional structural information, such as
the base pairing pattern at two-dimensional (2D) level and tertiary
contact maps at the 3D level, may be needed to further reduce the

Luc Ponchon (ed.), RNA Scaffolds: Methods and Protocols, Methods in Molecular Biology, vol. 2323,
https://doi.org/10.1007/978-1-0716-1499-0_1, © Springer Science+Business Media, LLC, part of Springer Nature 2021

1
2 Xiaojun Xu and Shi-Jie Chen

conformational searching space and increase the prediction accu-

racy for large RNAs [19–33].
RNA 2D structure as defined by the base pairing pattern pro-
vides structural constraints for 3D structure folding. Different
approaches have been developed to investigate the impact of RNA
2D structural constraint on 3D conformations [34–40]. For exam-
ple, conformational analysis based on TOPRNA [38] and MC-sym
[20] showed that RNA global conformation is largely defined by
topological constraints of RNA secondary structure while the elec-
trostatics, intra- and interloop, and other interactions select specific
conformations from the accessible conformational ensemble.
Recent investigation for RNA tertiary motifs containing the cross-
linked base pairs indicates that, tertiary motif can further decrease
the 3D conformational space and impose much stronger topologi-
cal constraints than the secondary structural motifs, which lack
cross-linked base pairs [40].
Taking the advantage of the strong topological constraints of
secondary structural motifs, template assembly-based RNA 3D
structure models, such as RNAcomposer [19], MC-sym [20], and
Vfold3D [21, 22], select templates from known structures with
homologous sequence information to predict 3D structures for
large RNAs. For example, with the given RNA sequence and 2D
structure, Vfold3D extracts the secondary structural motifs of heli-
ces, hairpin loops, internal/bulge loops, and multi-branched junc-
tions, and predicts the 3D structures through the assembly of
templates of non-helix motifs and A-form helices. With the exis-
tence of homologous structures for motifs, the knowledge-based
methods for RNA 3D structure prediction can usually achieve
higher accuracy and efficiency than the physics-based approaches.
However, due to the limitation of current motif template database,
the methods are severely limited by the low success rate of finding a
proper template for a given motif.
Recently, we introduced VfoldLA [23, 24], a hierarchical loop
template-assembly method for RNA 3D structure prediction.
Unlike the previous models which search for templates based on
the whole motif or piece-wise fragments, VfoldLA searches for
templates for single-stranded loops, and predicts 3D structures
through the assembly of loop templates and A-form helices. It has
a much higher success rate to find proper templates than the motif-
based approaches, and has also a much higher computational effi-
ciency than fragment-based approaches. In general, VfoldLA can
predict a set of all-atom structures from any 2D structure. How-
ever, due to the large number of combinations for the different
templates of the different loops, it can only handle RNAs contain-
ing small number of helices to avoid excessively long computational
times.
 Predicting RNA Scaffolds with a Hybrid Method of Vfold3D and VfoldLA 3

By combining Vfold3D and VfoldLA, in this chapter, we pres-

ent a hybrid model for RNA 3D structure predictions with given
2D structures. The model takes the advantage of the conserved 3D
structures for secondary structural motifs and the high success rate
of finding templates for single-stranded loops. The predicted struc-
tures may serve as useful scaffolds for further structure refinement
studies.

2 Algorithms

2.1 RNA Secondary An RNA 2D structure contains helices of canonical base pairs and
Structural Motifs and the single-stranded loops of unpaired nucleotides. According to the
Single-Stranded Loops different loop–helix connection modes, we define four types of
loops, as shown in Fig. 1a: “helix2” loop for a single strand that
connects two helices; “hairpin” loop as a special “helix2” loop
whose two ends connect to the same helix; “tail5” and “tail3”
loops for segments of unpaired nucleotides at the 50 and 30 ends,
respectively. The length of each loop L is the number of unpaired
nucleotides.
According to the different connection modes between helices,
we define the secondary structural motifs of internal/bulge loops
and multi-branched junctions, as shown in Fig. 1b (see Note 1).
Motif size is defined by the length of each loop L within a motif.
For example, the three-way junction has three helices connected by
three loops of L1, L2 and L3. The size of a three-way junction is
denoted by L1–L2–L3. Other motifs, such as pseudoknots and
hairpin-hairpin kissing motifs with the cross-linked base pairs, are
not included in this model.
As shown by the red oval shapes in Fig. 1, we include the loop-
connected terminal base pair(s) in the definition of loops and motifs
to account for the loop–helix structural interference. The sequence
of a loop/motif is formatted in the 50 !30 direction with W–W and
N denoting the terminal base pair and the unpaired nucleotides,
respectively. For example, W-50 W and W30 -W in the “helix2” loop
are the base pairs connected to the 50 and 30 ends of the loop,
respectively.

2.2 RNA Motif-based The (3D structure) template library for secondary structural motifs
Template Library was built from 4659 PDB structures (see Note 2), including
RNA-involved complexes. It contains 3D templates for bulge
loops, internal loops, and multibranched junctions. The motif-
based template library can be built in four steps:
1. For a given RNA 3D structure, extract the A-form helices.
From the base pairing pattern, the corresponding 2D structure
is determined.
4 Xiaojun Xu and Shi-Jie Chen

Fig. 1 The definition of (a) single-stranded loops of hairpin, helix2, and tails, and
(b) secondary structural motifs of internal/bulge loops and multibranched junc-
tions. Double-stranded helices are represented by cylinders. The red oval shape
in each helix denotes the terminal base pair attached to the loops (shown as red
lines). Sequences of loops and motifs are in the format of W–W (terminal base
pair) and N (unpaired nucleotide), with all loops in the 50 !30 directions

2. Identify all the secondary structural motifs for the given 3D

structure, according to the linkage between helices.
3. Remove the redundant templates for those with root mean
square deviation (RMSD) 1.5 Å for the same motif type,
same size and identical sequence.
4. Collect all the nonredundant motif structures to construct a
template library. Table 1 shows the statistics for the current
template library for motifs.
 Predicting RNA Scaffolds with a Hybrid Method of Vfold3D and VfoldLA 5

Table 1
RNA loop- and motif-based template libraries

Loop name Number of templates Motif name Number of templates

Hairpin loops 2562 Internal/bulge loops 4387
Helix2 loops 8612 Three-way junctions 1230
Tail5 loops 739 Four-way junctions 1168
Tail3 loops 1046 Five-way junctions 380
Six-way junctions 144
Seven-way junctions 217

2.3 RNA Loop-Based The template library for single-stranded loops was built from 4659
Template Library PDB structures. It contains 3D templates for hairpin, helix2, tail5,
and tail3 loops. The loop-based template library can be built in the
following steps:
1. For a given RNA 3D structure, extract the A-form helices.
2. Identify all the single-stranded loops for the given 3D struc-
ture, according to the linkage between helices and loops.
3. Remove the redundant templates for those with RMSD 1.5 Å
for the same loop type, same size and identical sequence.
4. Collect all the nonredundant loop structures to construct a
template library. Table 1 shows the statistics for the current
template library for loops.

2.4 Sequence Given a query loop/motif sequence, the model scores the sequence
Similarity-Based Score similarity according to the following rule: The score si for nucleo-
tide position i is equal to 0 for perfect match, otherwise equal to
1 for purine/pyrimidine-type match, and 2 for no match. For
example, sA!G ¼ 1 and sA!C ¼ 2 when nucleotide A is substituted
with G and C, respectively. We rank the loop/motif templates
according to the total score defined as the sum over all the nucleo-
tides: Sloop/motif ¼ ∑isi. This scoring scheme considers the similarity
of nucleotides, assuming that less changes in base substitution
result in less changes in 3D structure.

3 Methods

The Vfold3D and VfoldLA combined model predicts RNA 3D

structure from a 2D structure by the structural assembly from the
templates of motifs and loops. If the specific whole motif-based
templates are available and can be successfully found, we predict the
3D structure from the motif-based templates. The motif-based
6 Xiaojun Xu and Shi-Jie Chen

Fig. 2 The workflow of the hybrid model

templates usually give more accurate predictions than the VfoldLA

model alone. If a motif does not have available 3D templates in the
library, the model breaks the whole motif into single-stranded
loops and applies the VfoldLA strategy to predict the 3D structures.
Such an approach would lead to a much broader applicability than
the Vfold3D alone. As shown in the workflow in Fig. 2, the model
works with the following steps.
Predicting RNA Scaffolds with a Hybrid Method of Vfold3D and VfoldLA 7

1. Identify the structural components of double-stranded helices

(see Note 3), secondary structural motifs, and single-stranded
loops from the given 2D structure. For the motif types
excluded in the motif library (such as pseudoknot, and
hairpin-hairpin kissing), we break the whole motifs into the
single-stranded loops. For example, a pseudoknot with two
helices and three loops can be divided into three “helix2”
loops. Same treatment can be applied to all the secondary
structural motifs without templates.
2. Build the all-atom A-form helical structures.
3. Assemble the helix configurations in 3D space based on tem-
plates of all the motifs (with available templates) and “helix2”
loops, according to the aforementioned sequence similarity-
base scoring scheme. The optimal templates are selected with
minimal scores.
4. Exclude helix configurations that involve steric clash or
improper chain connectivity. To check steric clash between
atoms in the assemble structures, we allow at most five atom
pairs with a distance 2.0 Å. To check chain connectivity at the
loop–helix interfaces, we allow at most a 5.0 Å heavy-atom
RMSD of the helix terminal base pairs for each loop-closing
(helix2) loops.
5. For each viable helix configuration, select the optimal tem-
plates for other (hairpin, tail5 and tail3) loops to build the
whole structure. The optimal templates are chosen from the
sequence similarity-based score.
6. Repeat step 5 until the assembled whole structure is clash-free
(see Notes 4 and 5). The overall score is calculated by the sum
over the scores of all the templates.
7. Terminate the assembly process when the number of the
assembled 3D structures reaches a preset value or the compu-
tational time reaches a specific amount of time (see Note 6).
8. Cluster the overall score-ranked structures (see Note 7) and
refine the structures of clusters. The assembled structures may
contain structural clash and nonideal bond lengths, bond
angles, or bond torsional angles, especially at the loop–helix
interface regions. To avoid long computational time, the web
server utilizes the IsRNA model [16], which employs a coarse-
grained representation of RNA conformations and the
knowledge-based interaction potentials to refine the all-atom
structures assembled by the model (see Note 8).
9. The centroid structures of the top clusters are considered as the
predicted structures for the given 2D structure (see Note 9).
We use the ZIKV xrRNA1 [41] for the illustration of the
3D structure prediction (see Note 10). The native 2D structure
8 Xiaojun Xu and Shi-Jie Chen

Fig. 3 The prediction of the ZIKV xrRNA1. (a) Snapshot of the server input, which indicates the RNA sequence
and 2D structure in dot-bracket format, as well as other job information, such as number of clusters, RMSD
cutoff for clustering, job name and email address. The input 2D structure contains five helices, one three-way
junction and several single-stranded loops, with a loop-kissing helix (H4 as shown in red). (b) Snapshot of the
server output, which shows the detailed job information and a JSmol applet for visualization of predicted
structures. The best prediction has the RMSD of 8.7 Å, compared with the experimentally determined
structure

(Fig. 3a, see Note 11) contains five helices, one three-way
junction, and several single-stranded loops with a loop-kissing
helix (H4, shown in red). Because of the loop-kissing helix,
Vfold3D gives no predictions. However, the hybrid model
builds 3D structures of the 3-way junction and loops from
the motif template and loop template libraries, respectively,
and assemble A-form helical structures to generate all-atom
structures. As shown in Fig. 3b, the RMSD between the
(best) predicted structure and the experimental determined
structure is 8.7 Å. This RMSD is smaller than the prediction
from the VfoldLA model alone (RMSD ¼ 12.1 Å). The use of
the Vfold3D strategy for the secondary structural motif struc-
ture prediction contributes to the improved accuracy.
 Predicting RNA Scaffolds with a Hybrid Method of Vfold3D and VfoldLA 9

4 Notes

1. Other multiway junctions, such as four-way and five-way junc-

tions, are not shown in Fig. 1b for clarity.
2. The list of the RNA PDB structures used for constructing the
template library includes all the PDB entries released before
February of 2020.
3. Each helix should contain at least two base pairs. A lone base
pair is treated as an inter-loop interaction.
4. For each helix configuration, the model predicts one all-atom
structure with the optimal score, since different loop confor-
mations usually do not cause significant changes to the
global fold.
5. If none of the loop templates can satisfy the clash-free criteria,
the specific helix configuration would be removed.
6. Currently, the web server terminates predictions when the
computational time exceeds 12 h or the number of assembled
structures reaches 100, which may lead to no predictions.
7. As shown in Fig. 3a, the RMSD cutoff for clustering is set to
5.0 Å by default. The server users are free to select other values
(10 Å) before submission.
8. The IsRNA can resolve most of the structural clash while
keeping the overall 3D fold unchanged through small-RMSD
conformational sampling.
9. As shown in Fig. 3a, the number of clusters is set to 5 by
default. The server users are free to select other values (5)
before submission.
10. The sequence of the ZIKV xrRNA1 is: 50 GGGUCAGGCCG
GCGAAAGUCGCCACAGUUUGGGGAAAGCUGUGCA
GCCUGUAACCCCCCCACGAAAGUGGG30 .
11. The native 2D structure is as follows: ....(((((((((....)))).((((((..
[[[[...))))))..)))))...]]]](((((....))))), which is extracted from the
published paper [41] of the crystal structure (PDB id: 5tpy).

Acknowledgments

This research was supported by NIH grants R01-GM063732,

R01-GM117059, and R35-GM134919.
10 Xiaojun Xu and Shi-Jie Chen

References
1. Eddy SR (2001) Non-coding RNA genes and 16. Zhang D, Chen S-J (2018) IsRNA: an iterative
the modern RNA world. Nat Rev Genet simulated reference state approach to modeling
2:919–929 correlated interactions in RNA folding. J Chem
2. Sharp PA (2009) The centrality of RNA. Cell Theory Comput 14:2230–2239
136:577–580 17. Das R, Baker D (2007) Automated de novo
3. Chappell J, Takahashi MK, Meyer S, prediction of native-like RNA tertiary struc-
Loughrey D, Watters KE, Lucks J (2013) The tures. Proc Natl Acad Sci U S A
centrality of RNA for engineering gene expres- 104:14664–14669
sion. Biotechnol J 8:1379–1395 18. Das R, Karanicolas J, Baker D (2010) Atomic
4. Mortimer SA, Kidwell MA, Doudna JA (2014) accuracy in predicting and designing nonca-
Insights into RNA structure and function from nonical RNA structure. Nat Methods
genome-wide studies. Nat Rev Genet 7:291–294
15:469–479 19. Popenda M, Szachniuk M, Antczak M, Pur-
5. Cech TR, Steitz JA (2014) The noncoding zycka KJ, Lukasiak P, Bartol N, Blazewicz J,
RNA revolution—trashing old rules to forge Adamiak RW (2012) Automated 3D structure
new ones. Cell 157:77–94 composition for large RNAs. Nucleic Acids Res
6. Laing C, Schlick T (2011) Computational 40:e112
approaches to RNA structure prediction, anal- 20. Parisien M, Major F (2008) The MC-fold and
ysis, and design. Curr Opin Struct Biol MC-Sym pipeline infers RNA structure from
21:306–318 sequence data. Nature 452:51–55
7. Sim AY, Minary P, Levitt M (2012) Modeling 21. Cao S, Chen S-J (2011) Physics-based de novo
nucleic acids. Curr Opin Struct Biol prediction of RNA 3D structures. J Phys Chem
22:273–278 B 115:4216–4226
8. Shapiro BA, Yingling YG, Kasprzak W, Binde- 22. Xu X, Zhao P, Chen S-J (2014) Vfold: a web
wald E (2007) Bridging the gap in RNA struc- server for RNA structure and folding thermo-
ture prediction. Curr Opin Struct Biol dynamics prediction. PLoS One 9:e107504
17:157–165 23. Xu X, Chen S-J (2018) Hierarchical assembly
9. Sun L-Z, Zhang D, Chen S-J (2017) Theory of RNA three-dimensional structures based on
and modeling of RNA structure and interac- loop templates. J Phys Chem B
tions with metal ions and small molecules. 122:5327–5335
Annu Rev Biophys 46:227–246 24. Xu X, Zhao C, Chen S-J (2019) VfoldLA: a
10. Miao Z, Westhof E (2017) RNA structure: web server for loop assembly-based prediction
advances and assessment of 3D structure pre- of putative 3D RAN structures. J Struct Biol
diction. Annu Rev Biol 46:483–503 207:235–240
11. Strobel EJ, Yu AM, Lucks JB (2018) High- 25. Ponce-Salvatierra A, Astha, Merdas K,
throughput determination of RNA structures. Nithin C, Ghosh P, Mukherjee S, Bujnicki JM
Nat Rev Genet 19:615–634 (2019) Computational modeling of RNA 3D
12. Sharma S, Ding F, Dokholyan NY (2008) structure based on experimental data. Biosci
iFoldRNA: three-dimensional RNA structure Rep 39:BSR20181430
prediction and folding. Bioinformatics 26. Das R, Kudaravalli M, Jonikas M, Laederach A,
24:1951–1952 Fong R, Schwans JP, Baker D, Piccirilli JA,
13. Xia Z, Gardner DP, Gutell RR, Ren P (2010) Altman RB, Herschlag D (2008) Structural
Coarse-grained model for simulation of RNA inference of native and partially folded RNA
three dimensional structures. J Phys Chem B by high-throughput contact mapping. Proc
114:13497–13506 Natl Acad Sci U S A 105:4144–4149
14. Shi YZ, Wang FH, Wu YY, Tan Z-J (2014) A 27. Yang S, Parisien M, Major F, Roux B (2010)
coarse-grained model with implicit salt for RNA structure determination using SAXS data.
RNAs: predicting 3D structure, stability and J Phys Chem B 114:10039–10048
salt effect. J Chem Phys 141:105102 28. Kladwang W, Vanlang CC, Cordero P, Das R
15. Boniecki MJ, Lach G, Dawson WK, Tomala K, (2011) A two-dimensional mutate-and-map
Lukasz P, Soltysinski T, Rother KM, Bujnicki strategy for non-coding RNA structure. Nat
JM (2016) SimRNA: a coarse-grained method Chem 3:954–962
for RNA folding simulations and 3D structure 29. Ding F, Lavender CA, Weeks KM, Dokholyan
prediction. Nucleic Acids Res 44:e63 NV (2012) Three-dimensional RNA structure
 Predicting RNA Scaffolds with a Hybrid Method of Vfold3D and VfoldLA 11

refinement by hydroxyl radical probing. Nat 36. Kim N, Laing C, Elmetwaly S, Jung S,
Methods 9:603–608 Curuksu J, Schlick T (2014) Graph-based sam-
30. Falkner B, Schrder GF (2013) Cross-validation pling for approximating global helical topolo-
in cryo-EM-based structural modeling. Proc gies of RNA. Proc Natl Acad Sci U S A
Natl Acad Sci U S A 110:8930–8935 111:4079–4084
31. Ramani V, Qiu R, Shendure J (2015) High- 37. Sim AY, Levitt M (2011) Clustering to identify
throughput determination of RNA structure RNA conformations constrained by secondary
by proximity ligation. Nat Biotechnol structure. Proc Natl Acad Sci U S A
33:980–984 108:3590–3595
32. Nguyen TC, Cao X, Yu P, Xiao S, Lu J, Biase 38. Mustoe AM, Al-Hashimi HM, Brooks CL 3rd.
FH, Sridhar B, Huang N, Zhang K, Zhong S (2014) Coarse grained models reveal essential
(2016) Mapping RNA-RNA interactome and contributions of topological constraints to the
RNA structure in vivo by MARIO. Nat Com- conformational free energy of RNA bulges. J
mun 7:12023 Phys Chem B 118:2615–2627
33. Cheng CY, Kladwang W, Yesselman JD, Das R 39. Mustoe AM, Bailor MH, Teixeira RM, Brooks
(2017) RNA structure inference through CL 3rd, Al-Hashimi HM (2012) New insights
chemical mapping after accidental or inten- into the fundamental role of topological con-
tional mutations. Proc Natl Acad Sci U S A straints as a determinant of two-way junction
114:9876–9881 conformation. Nucleic Acids Res 40:892–904
34. Bailor MH, Sun X, Al-Hashimi HM (2010) 40. Xu X, Chen S-J (2020) Topological constraints
Topology links RNA secondary structure with of RNA pseudoknotted and loop-kissing
global conformation, dynamics, and adapta- motifs: applications to three-dimensional
tion. Science 327:202–206 structure prediction. Nucleic Acids Res 48
35. Bailor MH, Mustoe AM, Brooks CL 3rd, (12):6503–6512
Al-Hashimi HM (2011) Topological con- 41. Akiyama BM, Laurence HM, Massey AR, Cost-
straints: using RNA secondary structure to antino DA, Xie X, Yang Y, Shi P, Nix JC, Beck-
model 3D conformation, folding pathways, ham JD, Kieft JS (2016) Zika virus produces
and dynamic adaptation. Curr Opin Struct noncoding RNAs using a multi-pseudoknot
Biol 21:296–305 structure that confounds a cellular exonuclease.
Science 354:1148–1152
 Chapter 2

RNA Footprinting Using Small Chemical Reagents

Grégoire De Bisschop and Bruno Sargueil

Abstract
RNA is a pivotal element of the cell which is most of the time found in complex with protein(s) in a cellular
environment. RNA can adopt three-dimensional structures that may form specific binding sites not only for
proteins but for all sorts of molecules. Since the early days of molecular biology, strategies to probe RNA
structure have been developed. Such probes are small molecules or RNases that most of the time specifically
react with single strand nucleotides. The precise reaction or cleavage site can be mapped by reverse
transcription. It appears that nucleotides in close contact or in proximity of a ligand are no longer reactive
to these probes. Carrying the RNA probing experiment in parallel in presence and absence of a ligand yield
differences that are known as the ligand “footprint.” Such footprints allow for the identification of the
precise site of the ligand interaction, but also reveals RNA structural rearrangement upon ligand binding.
Here we provide an experimental and analytical workflow to carry RNA footprinting experiments.

Key words RNA, Structure, Probing, Ribonucleoprotein, Structural rearrangement

1 Introduction

The goal of this experiment is to map the site of interaction of

(a) protein(s) on an RNA. In brief, it consists in probing the
accessibility of the RNA nucleotides in presence and absence of
the protein. The reactivity difference observed between the two
conditions is considered to be the footprint of the protein on the
RNA. Interestingly, such experiment also monitors the potential
RNA structural rearrangement induced upon protein binding.
Such experiment can be carried out in vitro using a purified protein,
or crude extract, or even in vivo if it is possible to control the
conditions in which the complex is formed. The below described
protocol focuses on in vitro applications. The binding site of RNA
interacting small molecules such as antibiotics [1], or conversely of
huge complexes such as the ribosome [2] can also be footprinted
owing to this technique.

Luc Ponchon (ed.), RNA Scaffolds: Methods and Protocols, Methods in Molecular Biology, vol. 2323,
https://doi.org/10.1007/978-1-0716-1499-0_2, © Springer Science+Business Media, LLC, part of Springer Nature 2021

13
14 Grégoire De Bisschop and Bruno Sargueil

1.1 Footprinting Footprinting experiments can be conducted with any of the differ-
with DMS or SHAPE ent RNA structure probes [3, 4] including RNaseRNases and small
Reagent molecules. Obviously RNaseRNases are more bulky than small
molecules, precluding them to get too close of the protein
bound. The footprint yielded by RNaseRNases are therefore often
larger, and less precise than those obtained with small molecules.
Similarly, RNaseRNases bulkiness preclude them from reaching the
nucleotides buried within the structure or within the complex. This
can be seen as an advantage because in some occasions small mole-
cules may reach nucleotides within the RNA protein complex,
yielding only a very small footprint region, sometime not very
convincing. In addition, protein that binds only on double stranded
region of the RNA may not yield any chemical footprinting because
paired nucleotides are not reactive to such probes. An alternative is
to use the double strand specific Cobra venom nuclease V1. Unfor-
tunately, this nuclease is no longer commercially available at the
moment. In summary we advise using chemical footprinting for
which protocols are described below but, in some cases, enzymatic
footprinting is an interesting alternative (see Subheading 3.4).
A prerequisite for this experiment is to make sure that it is
performed in conditions in which most of the RNA is complexed
with the protein. Obviously to reach such state, the reaction mix
must contain at least as many protein molecules as RNA molecules,
but this may not be enough. As a rule of thumb, the protein
concentration should be about tenfold the Kd of the equilibrium,
if achievable. When using a purified protein Kd can be determined
using different techniques, including electromobility gel-shift
assays (EMSA) [5–7] or by filter binding assay [8–10]. When
using crude cellular extracts, the situation is complicated because
the concentration of the protein of interest is not known and other
proteins that specifically or unspecifically bind the RNA may be
present in the extract. However, strategies such as gel-shift assay can
be attempted. If the system is not amenable to such experiments,
different extract concentrations (0.01–1 μg/μl) of total protein
added must be assayed [11].
From these preliminary experiments will be deduced the fol-
lowing parameters for the footprinting per se:
l Protein concentration (depending on the binding affinity (Kd)).
l Cofactor (and concentration) requirement for the complex for-
mation (e.g., ATP, nonhydrolyzable analogue of ATP).
l Composition of the binding (reaction) buffer (salt
concentration).
l Incubation time.
Critical parameters can be further optimized in a few parallel
footprinting experiments, for example the protein concentration
can be assayed at 5, 10, and 20-fold the Kd.
 RNA Footprinting Using Small Chemical Reagents 15

Additionally, RNA folding must be as close to its native state as

possible. RNA recovery protocol from cells or from in vitro tran-
scription often include a denaturation step. In addition, RNA may
also go through freeze and thaw cycles which also alter the native
folding. In such cases, a denaturation–renaturation step helps to
homogenize RNA folding. In case a nondenaturing protocol has
been used to prepare the RNA, this step may be skipped. Numerous
refolding protocols have been proposed, their relative relevance
may depend on the RNA considered [12–15]. Herein, we propose
a “classical” refolding protocol. The following protocols are
designed for dimethyl sulfate (DMS), a reagent revealing adeno-
sines and cytosines, and for SHAPE reagents that can be used for
any nucleotides. For the latter we had more consistent results when
using fast acting reagents such as the 1-methyl-7-nitroisatoic anhy-
dride (1 M7) and benzoyl cyanide (BzCN). However, our protocol
can be easily adapted to any probing reagents, such as cyclo-N-
0
-[2-(Nmethylmorpholino)ethyl] carbodiimide-p-toluenesulfonate
(CMCT) that hits uridines and guanosines, or any RNase by mod-
ifying the reaction buffer and the incubation time when
necessary [3].
In any case, the principle of the experiment consists in two
reactions carried out in parallel in the exact same conditions, the
first containing the RNA on its own, and the second with the RNA
and its ligand in conditions were the complex is formed (see above).
In addition, the two cognate control reactions excluding the prob-
ing reagent will be carried out. In summary 4 reactions will be
carried out side by side.

1.2 RNA Retrieval Once complex has been probed, the protein can be denatured and
and Sequencing eliminated by phenol extraction to ensure that its presence does not
influence the reverse transcription step. Although we advise to
perform it, this step is optional. However, if you chose to do it, it
must be performed on all four tubes.
Nucleotides that have reacted with DMS or the SHAPE
reagent are modified and induce premature reverse transcription
stops. The reactivity is assessed for each nucleotide by the difference
of intensity of the stop in the modification reaction and in the mock
reaction. The nucleotide is mapped by comparison with the
sequencing reaction.

2 Materials

2.1 Footprinting 1. Dry bath.

with DMS and SHAPE 2. DMS buffer (10): 400 mM HEPES pH 7.5, 1 M KCl,
Reagent 50 mM MgCl2.
3. DMS (diluted in ethanol 1:12).
16 Grégoire De Bisschop and Bruno Sargueil

4. DMS stop reaction: 400 mM Tris pH 7.5.

5. Sodium hydroxide.
6. SHAPE buffer (10): 400 mM HEPES pH 7.5 (800 mM
when using BzCN), 1 M KCl, 50 mM MgCl2.
7. SHAPE reagent 10 in anhydrous DMSO: 40 mM 1M7 or
400 mM BzCN.
8. Fume hood.

2.2 RNA Retrieval 1. Phenol GTC: 2.5 M guanidinium thiocyanate, 76 mM

and Sequencing β-mercaptoethanol, 1% N-lauryl sarcosine, 60% phenol pH 7.
2. Glycogen 20 mg/ml or yeast tRNA 10 mg/ml.
3. Ammonium acetate 5 M.
4. 70% and 100% cold ethanol
5. Refrigerated centrifuge.
6. Dideoxynucleotide 10 mM, either ddTTP, ddATP, or ddCTP.
7. MMLV Reverse Transcriptase RNase H () 200 μ/μl.
8. MMLV RT buffer 5.
9. Deoxynucleotide mix dATP, dTTP, dCTP, and dITP,
10 mM each.
10. Two differently labeled fluorescent primers at 2 μM in water
(e.g., D2-PA and D4-PA) (Sigma-Aldrich) if running the
sequences on a CEQ8000 capillary sequencer (see Note 1).
11. Thermocycler.
12. RT stop buffer 5: 1.2 M sodium acetate pH 5, 40 mM
EDTA, 4 mg/ml glycogen.
13. Sample loading solution (deionized formamide).
14. Capillary electrophoresis (ABI or Ceq8000 AB Sciex).

3 Methods

3.1 Footprinting 1. Prepare four tubes with 6 pmol of RNA in 35 μl of water—

with DMS and SHAPE incubate for 5 min at 85  C.
Reagents 2. Add 5 μl of prewarmed appropriate buffer (DMS or SHAPE)
and let cool down for 10 min at room temperature, spin down
briefly. Incubate for an additional 15 min at the probing tem-
perature (30–37  C) (see Note 2).
3. Add 5 μl of the 10x protein in two tubes and 5 μl of the buffer
in which the protein is contained in the two other tubes (see
Notes 3 and 4).
4. Incubate for 5–10 min at a physiological relevant temperature
(30–37  C in most cases).
 RNA Footprinting Using Small Chemical Reagents 17

Probing with DMS:

5. Add 5 μl of DMS in two tubes (or 5 μl ethanol for the two
mock reactions)—Incubate 5 min at 37  C.
6. Stop the reaction by addition of 400 mM of Tris pH 7.5 (final)
and immediately place on ice.
Probing with SHAPE Reagents:
7. Add 5 μl of SHAPE reagent 10 in two tubes (or 5 μl DMSO
for the two mock reactions) (see Notes 5 and 6).
8. Incubate for five reagent half-lives at 37  C (30 min for NMIA,
2 min for 1M7 or few seconds for BzCN) after which the
SHAPE compound is hydrolysed (no inactivation step
required) (see Notes 7 and 8).

3.2 RNA Retrieval Step 1a: RNA retrieval and purification by phenol
and Sequencing extraction (optional)
1. To avoid losing too much of the material, the volume is
brought to 500 μl adding 400 μl H2O and 50 μl ammonium
acetate 5 M.
2. The reaction mix is extracted with an equal volume of Phenol
or Phenol GTC. Vortex for 1 min. Centrifuge for 5 min. Care-
fully recover 450 μl the aqueous upper phase. Add an equal
volume of chloroform. Vortex for 1 min centrifuge for 5 min.
Carefully recover 400 μl the aqueous upper phase.
Step 1b: RNA retrieval by ethanol precipitation
3. Add 1 μl of glycogen (20 mg/ml) or tRNA (10 mg/ml),
10% (v/v) of ammonium acetate 5 M (if step 1a was skipped),
250% (v/v) of 100% cold ethanol and incubate for at least
30 min at 20  C.
4. Centrifuge for 30 min at 16,000  g at 4  C. Discard the
supernatant.
5. Wash with 400 μl of 70% cold ethanol.
6. Centrifuge for 5 min at 16,000  g at 40  C. Discard the
supernatant—repeat this step.
7. Air-dry the pellet and resuspend it in 10 μl of water (see Note 9).
Step 2: Reverse transcription
8. Transfer the resuspended RNAs to PCR tubes and prepare four
PCR tubes with 6 pmol of RNA in 10 μl of water for the
sequencing reactions.
9. Add 1 μl of DMSO to the samples (modified RNA, modified
RNA–protein complex, mock reactions, and sequencing reac-
tions), denature for 3 min at 95  C then flash cool on ice.
18 Grégoire De Bisschop and Bruno Sargueil

10. Add 3 μl of 2 μM D2-labelled primer to the sequencing reac-

tion and 3 μl of D4-labelled primer to the modified/mock
RNA. Incubate for 5 min at 65  C then anneal for 10 min at
35  C and store on ice.
11. Add 4 μl of 5 RT buffer, 1 μl of the 10 mM deoxynucleotides
mix, and 1 μl of MMLV reverse transcriptase. For the sequenc-
ing reactions, also add 1 μl of 10 mM of one dideoxynucleo-
tide. Incubate for 2 min at 35  C, 30 min at 42  C and 5 min at
55  C then store on ice (see Note 10).
12. Add 5 μl of 5 RT stop buffer.
13. Combine one elongation from the modified RNA or the mock
RNA with one sequencing reaction.
14. Add 10% (v/v) of ammonium acetate 5M and 250% (v/v) of
100% ethanol, incubate for at least 30 min at 20  C and
precipitate the DNA as described above in Step 1b “Step 1b:
RNA retrieval by ethanol precipitation” ).
15. Air-dry the pellet and resuspend in 40 μl of sample loading
solution (see Notes 11 and 12).
Step 3: Sequencing by capillary electrophoresis
16. Load side-by-side 20 μl of the elongated cDNA corresponding
to the modified RNA and to the mock RNA.
17. Run the capillary electrophoresis.
18. Export the raw data as text files (see Note 13).

3.3 Data Analysis 1. Extract normalized reactivities and average them. Several soft-
ware have been developed to treat cDNA traces to obtain
reactivities [16–21]. The experimental procedure described
here is compatible with QuSHAPE, but note that this set-up
may need to be adapted in case if you wish to use a different
software. A thorough workflow with detailed procedures to
normalize and average reactivities is described elsewhere
(Allouche et al. in press). Care should be taken when aligning
the sequence to the (+) and () traces since even a single
nucleotide frameshift in the base calling of the “free RNA”
traces versus the “bound RNA” traces may lead to erroneous
results. QuSHAPE provides a convenient “sequence alignment
by reference” tool allowing a reference “project” (e.g., the free
RNA condition) to serve as a template for the sequence align-
ment of another “project” (the bound RNA condition). We
recommend performing footprint experiments in triplicate and
average the reactivities.
2. Compare reactivities to infer the binding mode of a protein.
The reactivity pattern obtained with the RNA only can be used
to constrain computer assisted RNA secondary structure pre-
diction using software or workflow such as RNAstructure [22],
 RNA Footprinting Using Small Chemical Reagents 19

RSample [23], RNAfold [24], IPANEMAP [25]. This may

ease the interpretation of the footprint, and yield a secondary
structure protein binding site, but this is not an obligation to
interpret a footprint. Independently of the secondary structure
modelling, there are several ways of comparing the reactivity
profiles obtained in the presence and absence of protein. We
here propose a method which highlights the most significant
differences. This heuristic approach can be adjusted in a case-
by-case manner, with the aim of identifying nucleotides that fall
into a different reactivity interval upon protein binding, for
example, getting from “highly reactive” to “reactive.” An
Excel file illustrating a simple yet powerful analysis chart is
available on demand ([email protected]).
Briefly, the reactivity difference Diff ¼ Rbound  Rfree and an
absolute Ratio ¼ jR bound R free j
Rbound þRfree are calculated, while all nucleotides
with undetermined values (usually labeled with a negative
value) are excluded from the analysis. Differences are consid-
ered significant when those variables exceed a threshold, typi-
cally 0.2 for the absolute difference and 0.2 for the ratio. The
threshold ratio allows to mask differences between high values
that would otherwise be overrepresented. Indeed, the measure
of reactivity clearly gets noisier for high values. Conversely, the
difference threshold removes weakly reactive nucleotides that
exhibit only a small shift in reactivity yet with a high ratio.
Finally, in case data contains multiple replicates, only the statis-
tically significant differences are kept (t-test < 0.05). The
provided Excel file contains two sheets; it is filled with an
example dataset. Raw reactivities must be entered in the
“Data” sheet along with the RNA sequence and its index.
Nucleotides for which the absolute deviation to the mean
reactivity is over 0.2 are highlighted, so that the consistency
of the replicates can be quickly assessed. This threshold may be
freely changed in the cell below the “threshold” cell. The
differences, the ratios and the t-tests are automatically calcu-
lated in the “Analysis” sheet. The last column gives the foot-
print results according to the different thresholds described
above (adjustable by the user). A histogram displaying the
reactivity differences between the free and the bound RNA
and highlights the footprint sites is provided. This allows the
user to instantly visualize the effect of the threshold values
modification. Differences between the bound-RNA reactivity
profile and the RNA only reactivity profile are of three different
types:
(a) Significant decreases in reactivity can be interpreted as
potential binding sites but keep in mind that they may
also reflect a local stabilization of the structure upon
protein binding or even a more extensive structural
rearrangement.
20 Grégoire De Bisschop and Bruno Sargueil

(b) Conversely, significant increases in reactivity can only be

interpreted as sites that are destabilized in the complex,
either because of a local structure alteration or because of
a more global protein-induced structural rearrangement.
(c) Finally, the majority of the nucleotides will likely have no
reactivity change upon binding, consistent with the exis-
tence of specific interaction sites. In the opposite case, the
absence of RNase from the protein sample should be
checked. Of note double-strand specific binding protein
may induce little or no reactivity change, especially if their
mode of binding does not destabilize the targeted RNA
helix.

Lastly, the interpretation will benefit to account for the speci-

ficity of the chemical probe used. The reactivity profile obtained
with base-specific probes such as DMS and CMCT is poorly sensi-
tive to a protein that essentially binds to the phosphate – sugar
backbone of the RNA. But conversely, footprints observed, if not
due to a structural stabilization, can be attributed to direct contacts
with the chemical group of the nucleotide targeted by the probe.
Alternatively, SHAPE probes reactivity relies on the nucleotide
flexibility and are therefore likely to be sensitive to any protein
binding mode that restrain the degree of freedom of the RNA
strand.
Note that such analysis can be also be carried out with probing
experiments coupled to NGS such as SHAPE-Map [26] or
DMS-Map [27].

3.4 Troubleshooting 1. No footprint observed

(a) Check if any reactivity is observed in the no-protein con-
dition by visually inspecting the raw traces. The signal
decay of the (+) trace should be stronger than in the ()
trace. In case not, adjust the reaction conditions (add
more probing reagent, increase the probing reagent con-
centration). Check if the probing reagent is still active
(this is particularly sensitive in the case of SHAPE reagent
which hydrolyzes with time).
(b) The probing reaction was not carried out in good protein
binding conditions. Check the binding with one of the
techniques suggested in the introduction, adjust protein
and/or salt concentration, incubation temperature.
(c) The nucleotides involved in protein binding are not sensi-
tive to the probe. Change the probing reagent including
enzymes. In such case it may be particularly relevant to use
the double strand RNase such as the Cobra Venom V1
enzyme to detect double strand RNA binding proteins.
 RNA Footprinting Using Small Chemical Reagents 21

2. No signal is observed in the presence of the protein:

(a) The protein (extract) sample is contaminated with an
RNase. Check the sample.
(b) RNA was carried out with the protein at the precipitation
or phenol extraction step. Make sure to dissociate the
complex by heating, adding some salts or detergent before
one of these steps.
3. No signal observed in presence or absence of the protein or too
many stops present even in the mock reaction: Check your
buffers and reagents for RNase contamination.

4 Notes

1. Primer sequence should be chosen at least 50 to 75 nucleotides

30 (downstream) to the region of interest.
2. Reaction buffers mentioned above should be considered as a
starting point and they should be adapted to each situation,
notably monovalent and divalent cations nature and concentra-
tions may be changed to optimize the specific binding of the
RNA ligand. However, do not use Tris buffer with DMS
because reacts with amine groups—and increase the buffer
(any) concentration when using BzCN (SHAPE reagent) to
compensate for the reagent acidification. If the protein volume
added is significant, do not forget to take into account the
mono and divalent salts brought by the protein sample.
3. If required the cofactor should also be added at this step.
4. If the protein is not available at tenfold the required concentra-
tion, the volume of protein added can be increased, in which
case the volume in Subheading 3.1, step 1 should be decreased
in consequence, and the buffer in Subheading 3.1, step 2 will
also need to be adapted. If the volume in Subheading 3.1, step
1 goes under 20 μl we advise to pool the 4 tubes to limit
evaporation. Remember that in such set up most of the buffer
and salts will be brought by the protein sample.
5. In order to rapidly mix the solution, be careful to add the RNA
solution onto the SHAPE reagent rather than the opposite.
6. All these reagents are genotoxic and/or skin and/or eye irri-
tant. They should be handled with care under a fume hood
wearing nitrile gloves and safety glasses. Waste tips and dispose
soiled materials in a saturated sodium hydroxide solution, or a
10% acetic acid solution for DMS and CMCT respectively.
SHAPE reagents get hydrolyzed relatively rapidly in water.
7. The solution will turn yellowish upon reaction with 1M7 or
cloudy with BzCN.
22 Grégoire De Bisschop and Bruno Sargueil

8. As the modification of one nucleotide may influence the rest of

the structure, multiple modifications per RNA molecule should
be avoided. As a rule of thumb, we consider that such condi-
tions are met when the amount of the full-length product in at
least 70% of that of that mock reaction. The modification
conditions described above are pretty standard and work well
for RNA between 200 and 500 nucleotides. They can be
adjusted by modifying the incubation time and reagent con-
centration to get more or less modification (NB: concentration
or time should be increased for shorter RNAs). This applies for
DMS and CMCT, but not for SHAPE reagents which do not
produce more than one modification per RNA under the above
described conditions.
9. Probed RNA can be stored at 20  C for at least 6 months.
10. For the sequencing reaction pick a nucleotide which is well
represented all along the RNA sequence.
11. Resuspension will take time at room temperature. It can be
accelerated by incubating the samples for 10 min at 50  C.
12. Elongated cDNA can be stored at 20  C.
13. If you do not have access to a capillary sequencer, you may use
radioactively labelled primers and run the samples on an appro-
priate sequencing P.A.G.E. Premature stops induced by the
reagent will then be qualitatively assessed by comparison of
the bands in the modification and the mock reaction. Note
that the premature stop is one nucleotide 30 to the modified
nucleotide.

Acknowledgments

We wish to thank N Chamond, L Ponchon, and C Vasnier for

sharing their protocols and tips, and for fruitful discussions, and
M Pospiech for careful proofreading.

References
1. von Ahsen U, Noller HF (1993) Footprinting vol 318. Academic Press, Cambridge,
the sites of interaction of antibiotics with cata- Massachusetts, pp 3–21
lytic group I intron RNA. Science 4. Ehresmann C, Baudin F, Mougel M et al
260:1500–1503 (1987) Probing the structure of RNAs in solu-
2. Angulo J, Ulryck N, Deforges J et al (2016) tion. Nucleic Acids Res 15:9109–9128
LOOP IIId of the HCV IRES is essential for 5. de Bisschop G, Ameur M, Ulryck N et al
the structural rearrangement of the 40SHCV (2019) HIV-1 gRNA, a biological substrate,
IRES complex. Nucleic Acids Res uncovers the potency of DDX3X biochemical
44:1309–1325 activity. Biochimie 164:83–94
3. Brunel C, Romby P (2000) Probing RNA 6. Ryder SP, Recht MI, Williamson JR (2008)
structure and RNA-ligand complexes with Quantitative analysis of protein-RNA
chemical probes. In: Methods in enzymology,
 RNA Footprinting Using Small Chemical Reagents 23

interactions by gel mobility shift. Methods Mol 17. Cantara WA, HatterschideJ WW et al (2017)
Biol 488:99–115 RiboCAT: a new capillary electrophoresis data
7. Sargueil B, Pecchia DB, Burke JM (1995) An analysis tool for nucleic acid probing. RNA
improved version of the hairpin ribozyme func- 23:240–249
tions as a ribonucleoprotein complex. Bio- 18. Mitra S, Shcherbakova IV, Altman RB et al
chemistry 34:7739–7748 (2008) High-throughput single nucleotide
8. Carey J, Cameron V, de Haseth PL et al (1983) structural mapping by capillary automated
Sequence-specific interaction of R17 coat pro- footprinting analysis. Nucleic Acids Res 36:
tein with its ribonucleic acid binding site. Bio- e63–e63
chemistry 22:2601–2610 19. Pang PS, Elazar M, Pham EA et al (2011)
9. Chamond N, Deforges J, Ulryck N et al (2014) Simplified RNA secondary structure mapping
40S recruitment in the absence of eIF4G/4A by automation of SHAPE data analysis. Nucleic
by EMCV IRES refines the model for transla- Acids Res 39:e151
tion initiation on the archetype of Type II 20. Kim H, Cordero P, Das R et al (2013)
IRESs. Nucleic Acids Res 42:10373–10384 HiTRACE-web: an online tool for robust anal-
10. Deforges J, de Breyne S, Ameur M et al (2017) ysis of highthroughput capillary electrophore-
Two ribosome recruitment sites direct multiple sis. Nucleic Acids Res 41:W492–W498
translation events within HIV1 gag open 21. Yoon S, Kim J, Hum J et al (2011) HiTRACE:
reading frame. Nucleic Acids Res high-throughput robust analysis for capillary
45:7382–7400 electrophoresis. Bioinformatics 27:1798–1805
11. Vallejos M, Deforges J, Plank TD et al (2011) 22. Deigan KE, Li TW, Mathews DH et al (2009)
Activity of the human immunodeficiency virus Accurate SHAPE-directed RNA structure
type 1 cell cycle-dependent internal ribosomal determination. PNAS 106:97–102
entry site is modulated by IRES trans-acting 23. Spasic A, Assmann SM, Bevilacqua PC et al
factors. Nucleic Acids Res 39:6186–6200 (2018) Modeling RNA secondary structure
12. Chillón I, Marcia M, Legiewicz M et al (2015) folding ensembles using SHAPE mapping
Chapter one - native purification and analysis of data. Nucleic Acids Res 46:314–323
long RNAs. In: Woodson SA, Allain FHT (eds) 24. Lorenz R, Luntzer D, Hofacker IL et al (2016)
Methods in enzymology, vol 558. Academic SHAPE directed RNA folding. Bioinformatics
Press, Cambridge, Massachusetts, pp 3–37 32:145–147
13. Herschlag D (1995) RNA chaperones and the 25. Saaidi A, Allouche D, Regnier M, Sargueil B,
RNA folding problem. J Biol Chem Ponty Y (2020) IPANEMAP: integrative prob-
270:20871–20874 ing analysis of nucleic acids empowered by
14. Jaeger L, Westhof E, Michel F (1993) Moni- multiple accessibility profiles. Nucleic Acids
toring of the cooperative unfolding of the sunY Res 48:8276–8289
group I intron of bacteriophage T4: the active 26. Siegfried NA, Busan S, Rice GM et al (2014)
form of the sunY ribozyme is stabilized by RNA motif discovery by SHAPE and muta-
multiple interactions with 30 terminal intron tional profiling (SHAPE-MaP). Nat Methods
components. J Mol Biol 234:331–346 11:959–965
15. Uhlenbeck OC (1995) Keeping RNA happy. 27. Zubradt M, Gupta P, Persad S et al (2017)
RNA 1:4–6 DMS-MaPseq for genome-wide or targeted
16. Karabiber F, McGinnis JL, Favorov OV et al RNA structure probing in vivo. Nat Methods
(2013) QuShape: rapid, accurate, and best 14:75–82
practices quantification of nucleic acid probing
information, resolved by capillary electropho-
resis. RNA 19:63–73
 Chapter 3

Improving RNA Crystal Diffraction Quality

by Postcrystallization Treatment
Jinwei Zhang and Adrian R. Ferré-D’Amaré

Abstract
The crystallization and structural determination of large RNAs and their complexes remain major bottle-
necks in the mechanistic analysis of cellular and viral RNAs. Here, we describe a protocol that combines
postcrystallization dehydration and ion replacement that dramatically improved the diffraction quality of
crystals of a large gene-regulatory tRNA–mRNA complex. Through this method, the resolution limit of
X-ray data extended from 8.5 to 3.2 Å, enabling structure determination. Although this protocol was
developed for a particular RNA complex, the general importance of solvent and counterions in nucleic acid
structure may render it generally useful for crystallographic analysis of other RNAs.

Key words X-ray Crystallography, Crystal dehydration, Ion replacement, Riboswitch, T-box RNA,
tRNA

1 Introduction

The rapid exploration of the noncoding genome using high-

throughput technologies is revealing critical roles for RNA struc-
ture in a wide range of cellular processes [1]. In addition, many
DNA and RNA viruses utilize defined three-dimensional RNA
folds to enable their life cycle and achieve infectivity [2–6]. Despite
critical roles of structured RNAs in biology and their impact on
human health, their functional elucidation is hampered by a paucity
of available structural information [4, 7–11]. One hundred years
after its invention, X-ray crystallography still provides the highest-
resolution structural information for macromolecules, especially for
RNA domains. The rarity of diffraction-quality crystals of larger
RNAs (longer than 100 nucleotides) remains a major roadblock
that hinders their structure determination [4]. Compared to pro-
teins, the polyanionic nature of the RNA backbone, reduced chem-
ical diversity of the four nucleobases, paucity of long-range
contacts, as well as inherent conformational flexibility, render it

Luc Ponchon (ed.), RNA Scaffolds: Methods and Protocols, Methods in Molecular Biology, vol. 2323,
https://doi.org/10.1007/978-1-0716-1499-0_3, © Springer Science+Business Media, LLC, part of Springer Nature 2021

25
26 Jinwei Zhang and Adrian R. Ferré-D’Amaré

difficult for RNAs to form specific, stable crystal packing

contacts [7].
Many RNA crystals only diffract X-rays to resolutions in the
range of 5 to 8 Å, insufficient to provide biochemical insight
(~3.5 Å or better is desirable). For protein crystals, many postcrys-
tallization treatment strategies such as annealing, dehydration, and
supplementing ligands have been developed [12–15]. For RNA,
the marked sensitivity of crystals to hydration status, documented
in the decades spanning publications on yeast tRNAPhe and the
glmS riboswitch-ribozyme, hints that diffraction quality could also
be significantly improved by postcrystallization treatments [16–
18]. In order to facilitate the development of general strategies
and methods for crystallographic studies of larger RNAs, we detail
and rationalize a protocol that enabled the crystallization and
structure determination of a large tRNA–mRNA complex
[19, 20]. Exploiting the general importance of RNA solvation
and counterions in stabilizing compactly folded RNAs [21], this
method concurrently dehydrates the RNA crystals and substitutes
the divalent cations in them. This two-pronged approach drives
quasi-rigid body movements of the RNA complexes in the crystal,
causing them to achieve geometrically and energetically superior
packing.

2 Materials

1. Oligonucleotides for PCR amplification.

2. Taq DNA polymerase, 5000 U/mL (New England Biolabs).
3. T7 RNA polymerase, 50,000 U/mL (New England Biolabs).
4. Diethylpyrocarbonate (DEPC)-treated water (see Note 1).
5. RNA Binding Buffer: 50 mM HEPES-KOH, pH 7.0, 100 mM
KCl, 20 mM MgCl2, 5 mM tris (2-carboxyethyl) phosphine
(TCEP).
6. 20 mM spermine solution, in DEPC-treated water, filtered
through 0.2 μm filter.
7. Crystallization Solution: 50 mM Bis-Tris (HCl) pH 6.5, 0.3 M
Li2SO4, 20 mM MgCl2, 20% (w/v) polyethylene glycol
(PEG) 3350.
8. EasyXtal 15-Well Tool (Qiagen).
9. MicroSieves and MicroSaws (MiTeGen).
10. 90
angled MicroLoops or MicroMounts (MiTeGen)
11. Crystal Treatment Solutions: 50 mM Bis-Tris (HCl), pH 6.5,
100 mM KCl, 20–50 mM SrCl2 or 20–100 mM MgCl2,
40–45% PEG3350, 5 mM TCEP.
 Improving RNA Crystal Diffraction Quality by Postcrystallization Treatment 27

3 Methods

3.1 Design and 1. Initial biochemical and biophysical characterization of T-box

Synthesis of T-Box RNA–tRNA complexes [22] was essential for design and engi-
RNA and tRNA for neering of crystallization constructs (Fig. 1). Glycine-specific
Crystallization glyQ/glyQS T-box constructs from 20 species were selected
from a multiple sequence alignment, with preference given to
thermophilic, extremophilic, and pathogenic organisms. The
T-box and tRNA constructs were transcribed in vitro using T7
RNA Polymerase, purified by denaturing Urea-PAGE, elec-
troeluted, washed once with 1 M KCl and extensively with
DEPC-treated water, concentrated and stored at 4
C or
20
C before use [18, 23].
2. T-box RNAs from a range of bacterial species were evaluated
for their propensity to form monodisperse, stoichiometric
complexes with tRNA using nondenaturing PAGE. T-boxes
from a handful of bacterial species, such as the extremely halo-
tolerant and alkaliphilic Oceanobacillus iheyensis eventually used
in structural determination, exhibited robust tRNA binding,
forming tRNA–mRNA complexes that migrated as relatively
sharp bands on nondenaturing gels.
3. Full-length T-box RNAs that contain both the Stem I and the
antiterminator domains (Fig. 1a) exhibited a tendency to form
dimers, presumably due to the thermodynamic instability of
the antiterminator [24]. Therefore, T-box RNAs were
truncated at a series of lengths and their affinities toward
tRNA and tendency to form monodisperse complexes evalu-
ated using isothermal titration calorimetry (ITC) and nonde-
naturing gels. This analysis demonstrated that Stem I is the
minimal T-box domain that is both necessary and sufficient for
high-affinity, specific binding to tRNA (Fig. 1b, c) [19].
4. To aid crystallization of RNA, several RNA-binding proteins
have been successfully used as crystallization chaperones, such
as the human spliceosomal U1A protein [25] and recombinant
antibody fragments (Fabs) [26–28]. The Kink-turn (K-turn) is
a widespread bistable RNA structural motif initially discovered
on the ribosome that sharply kinks the RNA duplex backbone
by 120
and is the landing platform to recruit several conserved
proteins to accomplish a range of cellular functions [29–
32]. The T-box Stem I domain harbors a conserved, function-
ally important K-turn [30, 33, 34]. The crucial contribution of
the K-turn to T-box architecture and function is further accen-
tuated by recent structural and biochemical elucidations of a
full-length T-box-tRNA complex and a novel class of transla-
tional T-box riboswitches [35, 36]. To stabilize the bistable
K-turn structure, provide added opportunities for crystal
28 Jinwei Zhang and Adrian R. Ferré-D’Amaré

Fig. 1 Sequences and secondary structures of full-length and truncated T-box riboswitch RNA used for
crystallization. (a) Secondary structure and sequence conservation of a full-length B. subtilis glycine-
responsive glyQS T-box riboswitch and its cognate tRNAGly. Circles with dark and light orange shades indicate
highly conserved (>80%) and moderately conserved (50–80%) sequences, respectively. Salient structural
features on both RNAs are boxed and annotated. Intermolecular T-box-tRNA base-pairing interactions are
indicated by solid lines connecting the boxed sequences. (b) Secondary structure of Oceanobacillus iheyensis
glyQ T-box Stem I domain used for cocrystallization. The italic, red sequences and red arrows denote
engineered regions. (c) Secondary structure of engineered B. subtilis/O. iheyensis tRNAGly (identical
sequences) used for cocrystallization. The original tRNA acceptor stem sequence is circularly permuted and
capped with a stable GAAA tetraloop (red, italic sequences). The bidirectional arrow denotes the length
variations to screen for optimal crystal contacts. (d) Representative crystals of T-box–Stem I–tRNA complexed
with Methanococcus jannaschii L7Ae. All scale bars represent 200 μm. (e) Representative crystals of
T-box–Stem I–tRNA complexed with B. subtilis YbxF. Note the differences in crystal morphology as dictated
by the protein component in the complex

packing, and allow for phasing using selenomethionines, a

panel of K-turn binding proteins were tested for their ability
to support crystal growth and improve crystalline order. Inter-
estingly, the choice of K-turn binding protein appreciably influ-
enced the crystal morphology. While the presence of
thermophilic Methanococcus jannaschii L7Ae protein [37]
(and other species of L7Ae) yielded star-shaped nonsingle crys-
tals (Fig. 1d), the addition of mesophilic B. subtilis YbxF [38]
produced single, square-plate-shaped crystals (Fig. 1e). The
 Improving RNA Crystal Diffraction Quality by Postcrystallization Treatment 29

latter is much more amenable to diffraction data collection. In

the absence of any K-turn-binding protein, only nondiffracting
crystals of T-box-tRNA binary complexes were occasionally
observed.

3.2 Crystallization of 1. Dilute concentrated tRNAGAAA (~ 1 mM; 24 g/L) to ~20 μM

the T-Box Stem I– using DEPC-treated water to reduce intermolecular interac-
tRNA–YbxF Ternary tion and dimerization.
Complex 2. “Snap-cool” tRNAGAAA by incubating at 90
C for 3 min
followed by rapid cooling to 4
C using a thermocyler (see
Note 2).
3. Concentrate refolded tRNA to ~12 g/L (500 μM) using Ami-
con spin concentrators (10 kD MWCO; 0.5 mL).
4. Mix 200 μM each T-box Stem I RNA and snap-cooled tRNA in
RNA Binding Buffer (Materials), incubate first at 50
C for
10 min and then at 37
C for 30 min.
5. Add one equivalent selenomethionyl B. subtilis YbxF to the
RNA complex.
6. Add spermine to 2 mM. The mixture may become transiently
cloudy. Mix gently with a pipette tip.
7. Heat to melt a stock of 2% low-melting-point agarose solution
and allow it to cool to 37
C using a heat block to prevent it
from solidifying.
8. Mix 1:1 the sample solution and Crystallization Solution and
keep at 37
C.
9. Add 1/10 volumes of 2% low-melting-point agarose solution
and gently mix by pipetting up and down. The presence of
agarose fibers in crystal solvent channels has been shown to
lend mechanical support to the crystals [39, 40]. The presence
of 0.2% low-melting-point agarose effectively prevents the
T-box cocrystals from cracking induced by the sudden change
in osmolarity (Fig. 2a). In addition to providing mechanical
support for the crystals, the agarose network also reduces
convection and permits more uniform crystal growth into
thicker dimensions. This beneficial effect on crystal habit and
morphology has recently been observed again in the cocrystals
of Nocardia farcinica ileS T-box-tRNA complex [36].
10. Transfer the crystallization mixture onto cover slides and initi-
ate crystallization experiments by hanging drop vapor
diffusion.

3.3 Post- 1. Square-plate-shaped crystals of the T-box–tRNA–YbxF ternary

crystallization complex start appearing as early as 1–2 days. Diffraction quality
Treatments crystals tend to grow more slowly, reaching final dimensions of
300  300  50 μm3 over the course of 1–3 weeks (Fig. 2a).
30 Jinwei Zhang and Adrian R. Ferré-D’Amaré

Fig. 2 Postcrystallization treatments dramatically improve diffraction quality of large RNA complexes. (a)
Postcrystallization treatment procedures and effects on the crystal appearance. Due to the drastic changes in
osmolarity (e.g., induced by a 20–40% change in PEG3350 concentration), pervasive crystal cracking and
even disintegration occurs. Cracking is effectively prevented by the mechanical support from the agarose
fibers in the solvent channels of the crystals. Note the agarose network that transferred together with the
embedded crystals. All scale bars denote 200 μm. (b) Comparison of magnified portions of diffraction
 Improving RNA Crystal Diffraction Quality by Postcrystallization Treatment 31

These crystals have the symmetry of space group C2221, with

unit cell dimensions of a ¼ 108.7 Å, b ¼ 108.8 Å, c ¼ 291.8 Å.
As do many other macromolecular crystals with relatively long
unit cell edges, the longest unit cell edge (291.8 Å) of these
crystals is parallel to the shortest physical dimension of the
crystals, that is, the edge that describes the thickness of the
rectangular or rhombic plates. Thus, oscillation diffraction
images that result from incident X-rays that traverse through
the broad faces of the plates suffer from significant overlap of
neighboring reflections. Such overlap is circumvented by the
use of 90
bent crystal loops, which restrict the incident X-rays
to only entering and exiting the crystals through their shortest
physical “edges” but not their “faces” (Fig. 2a).
2. Depending on the final concentration of low-melting point
agarose in the crystallization drop and temperature, the entire
drop may exhibit consistencies ranging from fluid liquid, vis-
cous liquid, jelly-like solid to robustly solid. Select appropriate
tools to transfer crystals into ~200 μL Crystal Treatment Solu-
tions in glass depression plates, that is, use conventional nylon
loops to transfer individual crystals from nonviscous liquid
drops, and use tools such as MicroSieves (MiTeGen) to transfer
whole, solidified drops. The composition of Crystal Treatment
Solutions will vary as it is based on both the RNA Binding
Solution and the Crystallization Solution. A gradient of con-
centrations of the primary precipitant (20–50% PEG3350 in
this example) is scouted to achieve a range of final solvent
contents and the effect on diffraction quality is measured.
Different concentrations of a panel of divalent cations in par-
ticular the alkaline earth metals (Mg2+, Ca2+, Sr2+, Ba2+) should
be screened, both for supporting crystal growth and for post-
crystallization treatment. In the case of the T-box complex
crystals, crystal growth in Mg2+ combined with postcrystalliza-
tion treatment in Sr2+ stood out as the optimal procedure,
producing the best Bragg spots profiles required for de novo
phasing using single-wavelength anomalous dispersion (SAD).
3. Seal each well of the depression plate using a glass cover slide
and Vaseline. Incubate the crystals in Crystal Treatment Solu-
tion (Materials) for 16 h. For the crystals of the T-box ternary

Fig. 2 (continued) oscillation photographs of untreated (as-grown) crystals (left, PDB: 4TZP), partially treated
crystals (middle panels and top right panel, PDB: 4TZV, 4TZW, and 4TZZ), and crystals that were subjected to
full cation replacement and dehydration (lower right panel; PDB: 4LCK) to demonstrate the improvement in
spot profile and order-to-order separation. Arrows indicate progressive additions of treatments. Diffraction
limits are indicated below each panel. Postcrystallization treatment and resulting crystal properties are
summarized in Table 1
32 Jinwei Zhang and Adrian R. Ferré-D’Amaré

complex, shorter treatments (i.e., less than 4 h) generally do

not produce the full effect of the treatment.
4. Carefully dissect the crystals out from their surrounding aga-
rose network using MicroSaws (MiTeGen) and remove as
much as agarose as possible (Fig. 2a). As the orientation of
the crystals in the crystal loop is critical for reducing overlap
during data collection, it is essential to trim nearly all agarose
away from the crystal faces so that the plate-like crystals would
be mounted parallel to the plane of the 90
bent loop due to
surface tension.
5. Using a 90
bent loop such as the angled MicroLoops or
MicroMounts (MiTeGen), pick up single, trimmed crystals
and immediately plunge into liquid nitrogen for vitrification.
As the Crystal Treatment Solution already contains at least 40%
(w/v) polyethylene glycol (PEG) 3350, no additional cryopro-
tective agent is necessary.

3.4 Understanding 1. Structure determination of as-grown, untreated crystals, and a

the Basis of number of crystals subjected to various combinations of post-
Treatment-Induced crystallization treatments (Fig. 2b & Table 1) allowed the
Improvement of tracking of macromolecular movements in these crystals in
Crystal Quality response to the treatments received [19, 20].
2. Structural alignment of untreated and optimally treated crystals
revealed that the ternary complexes of T-box–tRNA–YbxF shift
closer to each other in the crystal as quasi-rigid bodies (Fig. 3a;
see Note 3), producing superior packing contacts such as three
intimate base-stacking interactions between symmetry-related
complexes (Fig. 3b) as well as a stable A-minor interaction
between the engineered GAAA tetraloop on tRNA acceptor
stem and the minor groove of the proximal region of T-box
Stem I (Fig. 3c).
3. The unique preference for Sr2+ in postcrystallization treat-
ments of T-box cocrystals may be rationalized by its specific
association with the 30 cis-diols of neighboring symmetry-
related T-box RNAs (Fig. 3d), its frequent bidentate inner-
sphere interactions with the Hoogsteen faces of purines
(Fig. 3e), or its presence at bulges and junctions where phos-
phates cluster, or bridging across the narrow major groove. The
ability of Sr2+ to bind RNA 30 termini and its flexible coordina-
tion geometry are properties that may allow it to improve
crystalline packing of RNA [41].
 Improving RNA Crystal Diffraction Quality by Postcrystallization Treatment 33

Table 1
Select properties of crystals treated with varying degrees of ion replacement and dehydration

Unit cell VM
PDB Li2SO4 MgCl2 SrCl2 PEG 3350 Resolution Space dimensions (Å3/ VS
code (mM) (mM) (mM) (% w/v) (Å) group (Å) Da) (%)
4TZP 300 20 0 20 8.5 C2221 108.7, 108.8, 3.26 74.6
291.8a
4TZV 0 20 0 20 5.0 P43212 75.7, 75.7, 2.93 71.7
270.2a
4TZW 0 0 50 20 4.7 P43212 75.3, 75.3, 2.89 71.3
268.9a
4TZZ 0 100 0 48 3.6 P21 70.6, 260.7, 2.46 66.3
70.7b
4LCK 0 0 40 40 3.2 C2221 100.8, 109.7, 2.81 70.4
268.1a
VM Matthews coefficient (Matthews, 1968)
Vs Calculated solvent content
a
α ¼ β ¼ γ ¼ 90

b
α ¼ γ ¼ 90
, β ¼ 92.8

4 Notes

1. DEPC is a toxic alkylating agent. It should be handled with

appropriate personal protective equipment in a chemical fume
hood. DEPC-treated water is nontoxic, because after mixing,
the water–DEPC mixture is autoclaved. Heating in the pres-
ence of water converts DEPC into nontoxic carbon dioxide and
ethanol.
2. tRNAGly and other tRNAs are known to form dimers in solu-
tion depending on conditions used for folding the RNA. To
reduce dimerization, tRNAs are diluted in DEPC-treated water
and “snap-cooled,” which favors tRNA folding while suppres-
sing intermolecular association.
3. Note that the space group as well as the unit cell dimensions of
the crystals have changed significantly in response to the post-
crystallization treatments (Table 1).

Acknowledgments

We thank the staff at beamlines 5.0.1 and 5.0.2 of the ALS and
ID-24-C and ID-24-E of APS, in particular, K. Perry and
K.R. Rajashankar of the Northeastern Collaborative Access Team
(NE-CAT) of the APS for support in data collection and
34 Jinwei Zhang and Adrian R. Ferré-D’Amaré

Fig. 3 Treatment-induced, in-crystal movements of T-box ternary complexes produce superior crystal
contacts. (a) In-crystal redistribution of T-box ternary complexes as rigid bodies driven by dehydration and
cation replacement. Overlay of T-box ternary complexes in untreated (as-grown) crystals (light blue, PDB:
4TZP) and fully dehydrated and cation-exchanged crystals (dark blue, PDB: 4LCK). The corresponding
crystallographic unit cells are also shown, indicating close to ~10% compression along both a and c axes.
The reference complexes in the center of the panel superimpose well (RMSD for 172 C10 < 1.4 Å), but the
neighboring four complexes shift substantially closer as a result of the postcrystallization treatment (RMSDs
range from 3 to 10 Å and 10 to 19 Å, for RNA C1’ and protein Cα, respectively). Red arrows denote directions
of displacement (translation and rotation) of the four neighboring complexes. (b) Treatment-induced formation
of an intimate crystal contact involving three symmetry-related T-box ternary complexes, shown in blue,
green, and teal, respectively. Molecules from the untreated (PDB: 4TZP) and fully cation-replaced and
dehydrated crystals (PDB: 4LCK) are overlaid and colored in pastel and solid colors, respectively. Parallel
 Improving RNA Crystal Diffraction Quality by Postcrystallization Treatment 35

processing; G. Piszczek (National Heart, Lung and Blood Insti-

tute, NHLBI), R. Levine, and D.-Y. Lee (NHLBI) for assistance
with biophysical and mass spectrometric characterization; and
N. Baird, T. Hamma, C. Jones, M. Lau, A. Roll-Mecak, and
K. Warner for discussions. This work is partly based on research
conducted at the ALS on the Berkeley Center for Structural Biol-
ogy beamlines and at the APS on the NE-CAT beamlines (sup-
ported by National Institute of General Medical Sciences grant
P41GM103403). Use of ALS and APS was supported by the US
Department of Energy. This work was supported in part by the
intramural programs of the National Heart, Lung and Blood Insti-
tute (NHLBI) and National Institute of Diabetes and Digestive and
Kidney Diseases (NIDDK), and National Institutes of Health
(NIH).

References

1. Cech TR, Steitz JA (2014) The noncoding 4. Cantara WA, Olson ED, Musier-Forsyth K
RNA revolution-trashing old rules to forge (2014) Progress and outlook in structural biol-
new ones. Cell 157:77–94 ogy of large viral RNAs. Virus Res 193:24–38
2. Watts JM, Dang KK, Gorelick RJ et al (2009) 5. Bou-Nader C, Gordon JM, Henderson FE et al
Architecture and secondary structure of an (2019) The search for a PKR code-differential
entire HIV-1 RNA genome. Nature regulation of protein kinase R activity by
460:711–716 diverse RNA and protein regulators. RNA
3. Fang X, Wang J, O’Carroll IP et al (2013) An 25:539–556
unusual topological structure of the HIV-1 rev 6. Hood IV, Gordon JM, Bou-Nader C et al
response element. Cell 155:594–605 (2019) Crystal structure of an adenovirus
virus-associated RNA. Nat Commun 10:2871

Fig. 3 (continued) lines denote intermolecular stacking between nucleobases of symmetry-related com-
plexes. Arrows indicate displacements between the untreated and fully treated states. The rear face of the
interdigitated T-loops of Stem I distal region (opposite the face interacting with the tRNA elbow) form a
prominent flat surface available for crystal packing [19]. Two patches of this flat surface (A39 and A60
respectively), upon full treatment, engage in direct stacking contact with the apical adenine of the GAAA
tetraloop capping the tRNA acceptor stem (tA73) of a second complex (green), and with the terminal base pair
of T-box Stem I (G1·C102) of a third complex (teal), respectively. Dotted triangle surrounds an intermolecular
A-minor interaction (detail in c) present only in the fully treated crystals. tRNA residue numbers are preceded
by ‘t’. (c) Detail of the intermolecular class-I A-minor interaction formed between the tetraloop of a tRNA and
the minor groove of the proximal region of Stem I, colored as in (b). (d) Interfacial Sr2+ ions bridge symmetry-
related T-box RNA molecules by binding to their 30 termini. A well-defined Sr2+ ion (green sphere) is seen
bound to the cis-diol of the T-box RNA 30 terminal nucleotide (C102, marine), and two nonbridging oxygen
atoms of a symmetry-related T-box molecule (cyan), through inner-sphere coordination. Similar inner-sphere
coordination between Sr2+ and two 30 cis-diol groups bridge two symmetry-related heptanucleotide derived
from tRNAAla acceptor stem [42]. The Sr2+ bridging two symmetry-related T-box RNA thus may have
contributed significantly to the improved crystal quality through Sr2+ soaking. (e) Pervasive Sr2+ binding to
RNA nucleobases and backbone, such as a pair of well-defined Sr2+ ions next to T-box G43, one of which
makes bidentate innersphere interactions with the Hoogsteen face of G43. The electron density shown in (d)
and (e) is a portion of a composite simulated anneal-omit 2|Fo|  |Fc| synthesis contoured at 1.5 s.d. overlaid
with the final refined model
36 Jinwei Zhang and Adrian R. Ferré-D’Amaré

7. Zhang J, Ferre-D’Amare AR (2014) New 21. Draper DE (2004) A guide to ions and RNA
molecular engineering approaches for crystal- structure. RNA 10:335–343
lographic studies of large RNAs. Curr Opin 22. Grundy FJ, Henkin TM (1993) tRNA as a
Struct Biol 26:9–15 positive regulator of transcription antitermina-
8. Jones CP, Ferre-D’Amare AR (2015) Recogni- tion in B. subtilis. Cell 74:475–482
tion of the bacterial alarmone ZMP through 23. Milligan JF, Uhlenbeck OC (1989) Synthesis
long-distance association of two RNA subdo- of small RNAs using T7 RNA polymerase.
mains. Nat Struct Mol Biol 22:679–685 Methods Enzymol 180:51–62
9. Warner KD, Sjekloća L, Song W et al (2017) A 24. Zhang J, Ferré-D’Amaré AR (2014) Direct
homodimer interface without base pairs in an evaluation of tRNA Aminoacylation status by
RNA mimic of red fluorescent protein. Nat the T-box riboswitch using tRNA-mRNA
Chem Biol 13:1195–1201 stacking and steric readout. Mol Cell 55
10. Stagno JR, Liu Y, Bhandari YR et al (2017) (1):148–155
Structures of riboswitch RNA reaction states 25. Ferré-D’Amaré AR (2010) Use of the spliceo-
by mix-and-inject XFEL serial crystallography. somal protein U1A to facilitate crystallization
Nature 541:242–246 and structure determination of complex RNAs.
11. Chen MC, Tippana R, Demeshkina NA et al Methods 52:159–167
(2018) Structural basis of G-quadruplex 26. Shechner DM, Grant RA, Bagby SC et al
unfolding by the DEAH/RHA helicase (2009) Crystal structure of the catalytic core
DHX36. Nature 558:465–469 of an RNA-polymerase ribozyme. Science
12. Heras B, Martin JL (2005) Post-crystallization 326:1271–1275
treatments for improving diffraction quality of 27. Koldobskaya Y, Duguid EM, Shechner DM
protein crystals. Acta Crystallogr D Biol Crys- et al (2011) A portable RNA sequence whose
tallogr 61:1173–1180 recognition by a synthetic antibody facilitates
13. Russo Krauss I, Sica F, Mattia CA et al (2012) structural determination. Nat Struct Mol Biol
Increasing the X-ray diffraction power of pro- 18:100–106
tein crystals by dehydration: the case of bovine 28. Shelke SA, Shao Y, Laski A et al (2018) Struc-
serum albumin and a survey of literature data. tural basis for activation of fluorogenic dyes by
Int J Mol Sci 13:3782–3800 an RNA aptamer lacking a G-quadruplex motif.
14. Deng X, Davidson WS, Thompson TB (2012) Nat Commun 9:4542
Improving the diffraction of apoA-IV crystals 29. Klein DJ, Schmeing TM, Moore PB et al
through extreme dehydration. Acta Crystallogr (2001) The kink-turn: a new RNA secondary
Sect F Struct Biol Cryst Commun 68:105–110 structure motif. EMBO J 20:4214–4221
15. Awad W, Svensson Birkedal G, Thunnissen 30. Winkler WC, Grundy FJ, Murphy BA et al
MM et al (2013) Improvements in the order, (2001) The GA motif: an RNA element com-
isotropy and electron density of glypican-1 mon to bacterial antitermination systems,
crystals by controlled dehydration. Acta Crys- rRNA, and eukaryotic RNAs. RNA
tallogr D Biol Crystallogr 69:2524–2533 7:1165–1172
16. Kim SH, Quigley G, Suddath FL et al (1973) 31. Daldrop P, Lilley DM (2013) The plasticity of a
Unit cell transormations in yeast phenylalanine structural motif in RNA: structural polymor-
transfer RNA crystals. J Mol Biol 75:429–432 phism of a kink turn as a function of its envi-
17. Klein DJ, Ferré-D’Amaré AR (2009) Crystalli- ronment. RNA 19:357–364
zation of the glmS ribozyme-riboswitch. Meth- 32. Lilley DM (2012) The structure and folding of
ods Mol Biol 540:129–139 kink turns in RNA. Wiley Interdiscip Rev RNA
18. Klein DJ, Ferré-D’Amaré AR (2006) Struc- 3:797–805
tural basis of glmS ribozyme activation by glu- 33. Zhang J, Ferre-D’Amare AR (2015) Structure
cosamine-6-phosphate. Science and mechanism of the T-box riboswitches.
313:1752–1756 Wiley Interdiscip Rev RNA 6:419–433
19. Zhang J, Ferré-D’Amaré AR (2013) Co-crystal 34. Suddala KC, Zhang J (2019) An evolving tale
structure of a T-box riboswitch stem I domain of two interacting RNAs-themes and variations
in complex with its cognate tRNA. Nature of the T-box riboswitch mechanism. IUBMB
500:363–366 Life 71:1167–1180
20. Zhang J, Ferré-D’Amaré AR (2014) Dramatic 35. Li S, Su Z, Lehmann J et al (2019) Structural
improvement of crystals of large RNAs by cat- basis of amino acid surveillance by higher-order
ion replacement and dehydration. Structure tRNA-mRNA interactions. Nat Struct Mol
22:1363–1371 Biol 26:1094–1105
 Improving RNA Crystal Diffraction Quality by Postcrystallization Treatment 37

36. Suddala KC, Zhang J (2019) High-affinity rec- using agarose gel. Acta Crystallogr D Biol
ognition of specific tRNAs by an mRNA Crystallogr 58:1657–1659
anticodon-binding groove. Nat Struct Mol 40. Lorber B, Sauter C, Théobald-Dietrich A et al
Biol 26:1114–1122 (2009) Crystal growth of proteins, nucleic
37. Hamma T, Ferré-D’Amaré AR (2004) Struc- acids, and viruses in gels. Prog Biophys Mol
ture of protein L7Ae bound to a K-turn derived Biol 101:13–25
from an archaeal box H/ACA sRNA at 1.8 Å 41. Hofer TS, Randolf BR, Rode BM (2006) Sr
resolution. Structure 12:893–903 (II) in water: a labile hydrate with a highly
38. Baird NJ, Zhang J, Hamma T et al (2012) mobile structure. J Phys Chem B
YbxF and YlxQ are bacterial homologs of 110:20409–20417
L7Ae and bind K-turns but not K-loops. 42. Mueller U, Schübel H, Sprinzl M et al (1999)
RNA 18:759–770 Crystal structure of acceptor stem of tRNA(ala)
39. Biertümpfel C, Basquin J, Suck D et al (2002) from Escherichia coli shows unique G.U wob-
Crystallization of biological macromolecules ble base pair at 1.16 a resolution. RNA
5:670–677
 Chapter 4

Using tRNA Scaffold to Assist RNA Crystallization

Changrui Lu, Rujie Cai, Jason C. Grigg, and Ailong Ke

Abstract
Recent studies have solidified RNA’s regulatory and catalytic roles in all life forms. Understanding such
functions necessarily requires high-resolution understanding of the molecular structure of RNA. Whereas
proteins tend to fold into a globular structure and gain most of the folding energy from tertiary interac-
tions, RNAs behave the opposite. Their tertiary structure tends to be irregular and porous, and they gain
the majority of their folding free energy from secondary structure formation. These properties lead to
higher conformational dynamics in RNA structure. As a result, structure determination proves more
difficult for RNA using X-ray crystallography and other structural biology tools. Despite the painstaking
effort to obtain large quantities of chemically pure RNA molecules, many still fail to crystallize due to the
presence of conformational impurity. To overcome the challenge, we developed a new method to crystallize
the RNA of interest as a tRNA chimera. In most cases, tRNA fusion significantly increased the conforma-
tional purity of our RNA target, improved the success rate of obtaining RNA crystals, and made the
subsequent structure determination process much easier. Here in this chapter we describe our protocol to
design, stabilize, express, and purify an RNA target as a tRNA chimera. While this method continues a series
of work utilizing well-behaving macromolecules/motifs as “crystallization tags” (Ke and Wolberger.
Protein Sci 12:306–312, 2003; Ferre-D’Amare and Doudna. J Mol Biol 295:541–556, 2000; Koldobskaya
et al . Nat Struct Mol Biol 18:100–106, 2011; Ferre-D’Amare et al. J Mol Biol 279:621–631, 1998), it was
inspired by the work of Ponchon and Dardel to utilize tRNA scaffold to express, stabilize, and purify RNA
of interest in vivo (Ponchon and Dardel. Nat Methods 4:571-576, 2007). The “tRNA scaffold,” where the
target RNA is inserted into a normal tRNA, replacing the anticodon sequence, can effectively help the RNA
fold, express in various sources and even assist crystallization and phase determination. This approach
applies to any generic RNA whose 50 and 30 ends join and form a helix.

Key words tRNA scaffold, RNA expression, Purification, Crystallization

1 Introduction

RNA expression and purification has recently attracted attention

across all biomedical fields as large amount of homogeneous RNA
became critical in many RNA-related studies. However, RNA deg-
radation, misfolding, and structural heterogeneity poses technical
difficulties. In order to solve these problems, early approaches by
Joachim Frank attach the target RNA onto the ribosome, replacing
a loop on the ribosome surface [1]. Here we describe a similar

Luc Ponchon (ed.), RNA Scaffolds: Methods and Protocols, Methods in Molecular Biology, vol. 2323,
https://doi.org/10.1007/978-1-0716-1499-0_4, © Springer Science+Business Media, LLC, part of Springer Nature 2021

39
Exploring the Variety of Random
Documents with Different Content
over all other breeds and varieties, was always grateful to the monarch, for he
was the original discoverer and patentee of that blood. Then and there, in
order to praise the wisdom of a foolish king, a foolish fashion grew into a
foolish notion that has afflicted all England from that day to this. No humbug
of either ancient or modern times has had so long a run and so wide a range
as the miserable fallacy “that all excellence in the horse comes from the
Arabian.” Notwithstanding the thousand tests that have been made and the
thousand failures that have invariably followed, from the time of King James
to the present day, there are still men writing books and magazine articles on
the assumption that “all excellence in the horse comes from the Arabian,”
without ever having devoted an honest hour to the study of the question as to
whether this is a truth or a fallacy. This craze for Arabian blood was the
primary cause of the extinction of the pacer, and this craze was so strong in
its influence that when a foreign horse was brought in, no difference from
what country, if he were of the lighter type he was called an Arabian and so
advertised in order to secure the patronage of breeders. Horses brought from
the African coast were invariably classed as Arabians, notwithstanding they
and their ancestors were in Africa more than a thousand years before there
were any horses in Arabia; and the same may be said of Spain. But as this
line of inquiry has already been considered in another chapter, I will get back
to the immediate topic.
The process of breeding out the pacer did not commence in real earnest
until the middle of the seventeenth century, when the Stuarts regained the
sovereignty of Great Britain in the person of Charles II. Released from the
restraints of Puritan rule, the Restoration brought with it a carnival of
immorality and vice, for the court and the courtiers set the fashion and the
people followed. As the breeding interest of the period of which we now
speak has already been considered in the chapter on the English Race Horse,
I will not further enlarge upon it. The light, or running and hunting, horses of
England of that day were not all pacers, but they were all of the same type
and the same blood, hence when I speak of the pacers I include their
congeners. They were small—less than fourteen hands high—and not
generally handsome and attractive. In general utility they were ahead of the
importations, and doubtless many of them could run as fast and as far as the
foreign horses, but the foreigners had the advantage in size, especially the
Turks and the Neapolitans; besides this, they were more uniformly handsome
and attractive in their form and carriage. It is also probable that the outcross
from the strangers to invigorate the stock was needed and resulted in the
increase of the size of the progeny. This latter suggestion is inferential and
has been sustained by many similar experiences, but without this as a start it
would be exceedingly difficult to account for the rapid increase in the height
of the English race horse. It is certainly true that the chief aim of the English
breeder of that day was to increase the size, without losing symmetry and
style, and if he found that foreign upon native blood gave him a start in that
direction, he was wise in the commingling. Another consideration, growing out
of the rural economy of the people, doubtless had a very wide influence in the
direction of wiping out the pacer, in this period of transition. Long journeys in
the saddle became less frequent, good roads began to appear and vehicles on
wheels took the place of the saddler and the pack horse. To get greater
weight and strength for this service, recourse was had to crosses with the
larger and courser breeds, and through these channels have come the giants
and the pigmies of the modern race course. Under the changed conditions of
travel and transportation it is not remarkable that the people should have
been willing to see their long-time favorites disappear, for it is known to every
man of experience that the pace is not a desirable gait for harness work. No
doubt the pacer is as strong as the trotter of the same size and make-up, but
in his smooth, gliding motion there is a suggestion of weakness
communicated to his driver that is never suggested by the bold, bounding
trotter. The antagonism between the pacers and the new horses of Saracenic
origin was irreconcilable and one or the other had to yield. As the
management of the contest was in the hands of the master the result could
be easily foreseen, for if one cross failed, another followed and then another,
till the Saracenic blood was completely dominant in eliminating the lateral and
implanting the diagonal action in its stead.
As no home-bred pacer, of any type or breed, has been seen in England for
nearly two hundred years, it is not remarkable that Englishmen of good
average intelligence, for the past two or three generations, have lived and
died supposing they knew all about horses, and yet did not know there had
ever been such a thing in England as a breed of pacing horses. When, some
eighteen or twenty years ago, I called the attention of Mr. H. F. Euren,
compiler of the Hackney Stud Book, to the early English pacers as a most
inviting field in which to look for the origin of the “Norfolk Trotters,” he was
surprised to learn that such horses had existed in England, but he went to
work and gathered up many important facts that appear in the first volume of
the Hackney compilation. Many of these facts, but in less detail, had already
appeared, from time to time, in Wallace’s Monthly, but Mr. Euren’s was the
first modern English publication to place them before English readers. From
this prompting, Mr. Euren did well, but we must go back a little to see how
this subject was treated by English writers of horse books, who wrote without
any promptings from this side.
 Mr. William Youatt was a voluminous writer on domestic animals, and at one
time was looked upon as the highest authority on the horse, both in England
and in this country. He seems to have been a practitioner of veterinary
surgery, and from the number of volumes which he published successfully, he
must have been a man of ability and education. There can be no question
that he knew a great deal—quite too much to know anything well. The first
edition of his work on the horse was published in 1831, and soon after its
appearance several publishing houses in this country seized upon it as very
valuable, and each one of them soon had an edition of it before the public. It
purports to have been written at the instance of “The Society for the Diffusion
of Useful Knowledge.” This declaration was a good thing, in a commercial
view, and no doubt it did much in extending the circulation of the book.
Without tarrying to note several minor historical blunders, I will go direct to
one relating to the gait of the horse, which is now under consideration. In his
fourth edition, page 535, he incidentally discusses the mechanism of the pace,
and after speaking of the Elgin Marbles, to which I have referred at the
beginning of this chapter, and after conceding that two of the four horses are
not galloping but pacing, he says:

“Whether this was then the mode of trotting or not, it is certain that it is
never seen to occur in nature in the present day; and, indeed, it appears
quite inconsistent with the necessary balancing of the body, and was,
therefore, more probably an error of the artist.”

This remark is simply amazing in an author who pretentiously undertakes to

instruct his countrymen in the history of the horse when he knows nothing
about that history. If he had gone back only twenty-two years, “Old John
Lawrence,” in his splendid quarto, would have told him about the pacer. If he
had gone back one hundred and sixty years, the Duke of Newcastle would
have explained to him the complete and perfect mechanism of the pacing
gait. If he had gone still further back and examined Gervaise Markham,
Blundeville, Polydore Virgil, and Fitz Stephen the Monk, of the twelfth century,
any and all of them would have explained to him the pacing habit of action
and shown him that for many successive centuries the pacing horse was the
popular and fashionable horse of the realm. If Mr. Youatt had lived to see
John R. Gentry pace a mile in 2:00½; Robert J. in 2:01½, and dozens of
others in less than 2:10, he might have changed his mind and concluded that
it was possible, after all, for a horse to travel at the lateral gait without
toppling over. From Mr. Youatt and a few other modern English authors, most
of our American writers on the horse have derived what little mental pabulum
they thought they needed, and thus an error at the fountain has been carried
into all the ramifications of our horse literature. Only two or three years ago a
very intelligent gentleman, who had attained great eminence as a veterinary
surgeon, especially for his knowledge and treatment of the horse’s foot,
seriously and in good faith stoutly maintained that the pacing habit of action
was merely the result of an abnormal condition of the foot, and that all pacers
would trot just as soon as their feet were put in the right shape. We must not
laugh at this wild notion, for it is really no worse than Mr. Youatt’s doubting
whether it was possible for a horse to balance himself at the lateral motion.
Neither gentleman seemed to know anything about the fact that it was a
matter of inheritance, and that the lateral habit of action had come down by
transmission through all the generations for a period of more than two
thousand years. It is hardly necessary to say that the gentleman who was so
confident that the pace was merely the result of the abnormal condition of the
feet brought his notions about the pacer from across the water. He was an
Anglo-American, and could make a pacer into a trotter in a jiffy, by using the
paring-knife. He was an intelligent man and a skillful veterinarian, but there
were no pacers in England and there should be none here. Toward the close
of the chapter on The Colonial Horses of Virginia, will be found the
observations of an English tourist in 1795-96 who is very certain that there is
some mistake about the pacer, and will not be convinced there are any, unless
they are artificially created. Having now completed what I had to say about
the old English pacer, it is next in order to consider his descendants in this
country and the relations they bear to the American trotter.
 CHAPTER XIV.
THE AMERICAN PACER AND HIS RELATIONS TO THE
AMERICAN TROTTER.

Regulations against stallions at large—American pacers taken to the West

Indies—Narragansett pacers; many foolish and groundless theories
about their origin—Dr. McSparran on the speed of the pacer—Mr.
Updike’s testimony—Mr. Hazard and Mr. Enoch Lewis—Exchanging
meetings with Virginia—Watson’s Annals—Matlack and Acrelius—Rip
Van Dam’s horse—Cooper’s evidence—Cause of disappearance—
Banished to the frontier—First intimation that the pace and the trot
were essentially one gait—How it was received—Analysis of the two
gaits—Pelham, Highland Maid, Jay-Eye-See, Blue Bull—The pacer
forces himself into publicity—Higher rate of speed—Pacing races very
early—Quietly and easily developed—Comes to his speed quickly—His
present eminence not permanent—The gamblers carried him there—
Will he return to his former obscurity?
In the several chapters devoted to “Colonial Horse History” will be found all
the leading facts that I have been able to glean from the early sources of
information. With the exceptions of the horses brought from Utrecht in
Holland to New Amsterdam (New York), two shiploads that sailed out of the
Zuider Zee and landed at Salem, Massachusetts, and those brought from
Sweden by the colonists that settled on the Delaware, all the early
importations came from England. As much the larger number of those from
England and Sweden were pacers, the breeds and habits of action were soon
mixed up, as those who had no pacers wanted pacers for the saddle, and
those who wanted more size, regardless of the gait, were always ready to
supply their want by an exchange of their saddle horses for more size. The
Dutch horses were certainly something over fourteen hands and the English
and Swedish horses were perhaps nearer thirteen than fourteen hands. The
colonists from the first, and from one end of the land to the other, seem to
have appreciated the importance of increasing the size and strength of their
horse stock, and this was very hard to do under the conditions then prevailing
of allowing their horses to roam at large. Hence, stringent regulations were
adopted in all the colonies against permitting immature entire colts and
stallions under size to wander where they pleased. It is doubtful whether
these regulations were any more effective than those of Henry VIII., for while
there was some increase, it was hardly perceptible until after the close of the
colonial days. The real increase did not commence till the farmers had
provided themselves with facilities for keeping their breeding stock at home.

JOHN R. GENTRY.

By Ashland Wilkes, pacing record 2:00½, 1896.

It is very evident from the statistics of size and gait, as given in the
chapters referred to above, that our forefathers wisely selected the most
compact, strong and hardy animals they could find in England as the type
best adapted to fight their way against the hardships of a life in the
wilderness of the new world. There have been some attempts, wholly fanciful
and baseless, to trace importations from other countries, outside of those
mentioned above, but all such attempts have proven wholly imaginary and
worse than futile. In less than twenty years after the New England colonies
received their first supply they commenced shipping horses by the cargo to
Barbadoes and other West India Islands. This trade was cultivated, extended
to all the islands, and continued during the remainder of the seventeenth and
practically the whole of the eighteenth century. The pacers of the American
colonies were exceedingly popular and sought after by the Spanish as well as
the Dutch and English islands. Indeed, the planters of Cuba alone carried
away at high prices nearly all the pacers that New England could produce.
They knew nothing about pacers for the saddle until they had tried them and
then they would have nothing else. These continuous raids of the Spaniards
of the West Indies upon the pacers of New England, and Rhode Island
especially, has been assigned, by the local historians of that State as one of
the principal causes of the decadence and practically final disappearance of
the Narragansett pacer from the seat of his triumphs and his fame. It is just
to remark here, in passing, that if there had been pacers among the horses of
Spain, the Spanish dependencies would have secured their supplies from the
mother country and not have come to Rhode Island and paid fabulous prices
for them.
As all the pacing traditions of this country to-day point to the horses of
Narragansett Bay as the source from which our modern pacers have derived
their speed, we must give some attention to the various theories that have
been advanced as to the origin of the Narragansett horse. In time past, and
extending back to a period “whereof the memory of man runneth not to the
contrary,” the horse world has been cursed with a class of men who have
always been ready to invent and put in circulation the most marvelous and
incredible stories about the origin of every remarkable horse that has
appeared. Some of these wiseacres have maintained that the original
Narragansett pacer was caught wild in the woods by the first settlers on
Narragansett Bay, while others (and this seems to be of Canadian origin) have
insisted that when being brought to this country a storm struck the ship and
the horse was thrown overboard, and after nine days he was found off the
coast of Newfoundland quietly eating rushes on a sand bar, where he was
rescued and brought into Narragansett Bay. This story of the marine horse
probably had its origin in the experiences of Rip Van Dam, which will be
narrated further on. Another representation, coming this time from a very
reputable source, has been made as to the origin of the Narragansett horse,
and as many, no doubt, have accepted it as true, I must give it such
consideration as its prominence demands. Mr. I. T. Hazard, a representative of
the very old and prominent Hazard family of Rhode Island, in a letter to the
Rev. Mr. Updike, makes the following statement:

“My grandfather, Governor Robinson, introduced the famous saddle

horse, the Narragansett pacer, known in the last century over all the
civilized parts of North America and the West Indies, from whence they
have lately been introduced into England, as a ladies’ saddle horse, under
 the name of the Spanish Jennet. Governor Robinson imported the original
from Andalusia, in Spain, and the raising of them for the West India market
was one of the objects of the early planters of this country. My grandfather,
Robert Hazard, raised about a hundred of them annually, and often loaded
two vessels a year with them, and other products of his farm, which sailed
direct from the South Ferry to the West Indies, where they were in great
demand.”

This theory of the origin of the Narragansett came down to Mr. Hazard as a
tradition, no doubt, but like a thousand other traditions it has nothing to
sustain it. Opposed to it there are two clearly ascertained facts, either one of
which is wholly fatal to it. In the first place, there were no pacers in Andalusia
or any other part of Spain, and in the second place, these horses, according
to official data, were the leading item of export from Rhode Island in 1680,
and Governor Robinson was not born till about 1693. As impossibilities admit
of no argument, I will not add another word to this “Andalusian” origin
tradition, except to say that a hundred years later, when the pacing dam of
Sherman Morgan was taken from Cranston, Rhode Island, up into Vermont,
she was called a “Spanish mare,” because Mr. Hazard had said the original
Narragansett had come from Spain. The story of the descendants of the
Narragansetts having been carried from the West Indies to England, and there
introduced under the name of the Spanish Jennet as a lady’s saddle horse, is
wholly imaginative. The Spanish Jennet, whatever its gait may have been,
was well known in England many years before the first horse was brought to
any of the American colonies. (See extracts from Blundeville and Markham in
Chapter XII.)
After several years of fruitless search for some trace of the early
importations of horses into the colony of Rhode Island, I have reached the
conclusion that probably no such importations were ever made. The colony of
Massachusetts Bay commenced importing horses and other live stock from
England in 1629, and continued to do so for several years and until they were
fully supplied, as stated above. In 1640 a shipload of horses were exported to
the Barbadoes, and it was about this time that Rhode Island began to assume
an organized existence. Her people were largely made up of refugees from
the religious intolerance of the other New England colonies, and they brought
their families and effects, including their horses, with them. The blood of the
Narragansett pacer, therefore, was not different from the blood of the pacers
of the other colonies, but the development of his speed by the establishment
of a pacing course and the offering of valuable prizes, naturally brought the
best and the fastest horses to this colony and from the best and fastest they
built up a breed that became famous throughout all the inhabited portions of
the Western Hemisphere. The race track, with the valuable prizes it offered
and the emulation it aroused, was what did it. As the question of origin is thus
settled in accordance with what is known of history and the natural order of
things, and as the Narragansett is the great tribe representing the lateral
action then and since, we must consider such details of history as have come
down to us.
The Rev. James McSparran, D.D., was sent out by the London Society for
the Propagation of the Gospel in Foreign Parts, to take charge of an Episcopal
church that had been planted some years before in Rhode Island. He arrived
in 1721, and lived till 1759. He was an Irishman, and appears to have been
somewhat haughty and irascible in his temperament and was disposed to find
fault with the climate, the currency, the people, and pretty much everything
he came in contact with. He was a man of observation, and during the thirty-
eight years he spent in ministering to the spiritual wants of his flock, he was
not unmindful of what was passing around him, and made many notes and
reflections on the various phases of life as they presented themselves to his
mind, and especially on the products and industries of the colony. These notes
and observations he wrote out, and they were published in Dublin in 1753,
under the title of “America Dissected.”
His writings do not discover that he was a man of very ardent piety, but he
was honored as a good man while he lived, and was buried under the altar he
had served so long. His duties sometimes called him away into Virginia, and,
in speaking of the great distance of one parish from another, he uses the
following language:

“To remedy this (the distance), as the whole province, between the
mountains, two hundred miles up, and the sea, is all a champaign, and
without stones, they have plenty of a small sort of horses, the best in the
world, like the little Scotch Galloways; and ’tis no extraordinary journey to
ride from sixty to seventy miles or more in a day. I have often, but upon
larger pacing horses, rode fifty, nay, sixty miles a day, even here in New
England, where the roads are rough, stony and uneven.”

The reverend gentleman seems to assume that his readers knew the Scotch
Galloways were pacers, and with this explanation his observations are very
plain. He makes no distinction between the Virginia horse and his congener of
Rhode Island except that of size, in which the latter had the advantage. In
speaking of the products of Rhode Island he says:

“The produce of this colony is principally butter and cheese, fat cattle,
wool, and fine horses, which are exported to all parts of English America.
 They are remarkable for fleetness and swift pacing; and I have seen some
of them pace a mile in a little more than two minutes, and a good deal less
than three.”

When I first read this sentence in the reverend doctor’s book I confess I
was not prepared to accept it in any other light than that of a wild enthusiast,
who knew but little of the force of the language he used. To talk about horses
pacing, a hundred and fifty years ago, in a little more than two minutes and a
good deal less than three, appeared to be simply monstrous. The language
evidently means, according to all fair rules of construction, that the mile was
performed nearer two minutes than three, or in other words, considerably
below two minutes and thirty seconds. I doubt not my readers will hesitate,
and perhaps refuse, to accept such a performance, just as I did myself till I
had carefully weighed not only the character of the author of the statement,
but the circumstances that seemed to support it. If the learned divine had
known no more of the world and its ways than many of his profession, I
would have concluded he was not a competent judge of speed; but he was a
man of affairs, and knew perfectly well just what he was saying. The question
naturally arises here as to what opportunities or facilities the doctor had for
timing those pacers of a hundred and fifty years ago. In a note appended to
the above extract by Mr. Updike, the editor of the work, I find the following:

“The breed of horses called Narragansett pacers, once so celebrated for

fleetness, endurance and speed, has become extinct. These horses were
highly valued for the saddle, and transported the rider with great
pleasantness and sureness of foot. The pure bloods could not trot at all.
Formerly they had pace-races. Little Neck Beach, in South Kingston, of one
mile in length, was the race course. A silver tankard was the prize, and
high bets were otherwise made on speed. Some of these prize tankards
were remaining a few years ago. Traditions respecting the swiftness of
these horses are almost incredible.”

The facts stated by Mr. Updike in this note are corroborated from other
sources, and may be accepted as true. These were the opportunities and
facilities the doctor had for holding his watch, and nobody will doubt they
were sufficient to enable him to be a competent witness. In connection with
this subject, and as another footnote, Mr. Updike introduces a letter from Mr.
I. T. Hazard, which brings out another very curious fact in the history of the
pacer. The Hazard family was very eminent in Rhode Island, and many of its
members have occupied positions of high honor and responsibility for several
generations. The date of the letter is not given, and we may infer it may have
been written fifty years ago, or perhaps more. Mr. Hazard says:
 “Within ten years one of my aged neighbors, Enoch Lewis, since
deceased, informed me he had been to Virginia as one of the riding boys,
to return a similar visit of the Virginians in that section, in a contest on the
turf; and that such visits were common with the racing sportsmen of
Narragansett and Virginia, when he was a boy. Like the old English country
gentlemen, from whom they were descended, they were a horse-racing,
fox-hunting, feasting generation.”

This paragraph from Mr. Hazard’s pen has been the subject of very
deliberate consideration. The first promptings of my judgment were to doubt
and reject it, especially on account of the absence of date to the letter, and of
the remote period in which Mr. Enoch Lewis must have visited Virginia.
Another question, as to why we have not this information from any other
source except Mr. Hazard, presented itself with no inconsiderable force. After
viewing the matter in all its bearings I am forced to concede that it is likely to
be true. These visits must have taken place before the Revolution, and from
the construction we are able to place upon the dates, this was not impossible.
It is a fact that I do not hesitate to announce that before the Revolution
racing in all its forms was more universally indulged in as an amusement than
it ever has been since. This was before the days of newspapers, and all we
can possibly know of the sporting events of that period we must gather up
from the detached fragments that have come down to us by tradition. There
was a strong bond of sympathy and friendship between the followers of Dr.
McSparran in Rhode Island, surrounded as they were by Puritans, and their
co-religionists in Virginia. They were accustomed to maritime life, and had
abundance of vessels fitted up for the shipment of horses and other live stock
to foreign ports. To take a number of their fastest pacers on board one of
their sloops and sail for Virginia would not have been considered much of an
adventure. These visits were not only occasions of pleasure and festivity, with
the incidental profits of winning purses and bets, but they were a most
successful means of advertising the Narragansett pacer; and through these
means alone the market was opened, as Dr. McSparran expresses it, in all
parts of British America. When we consider the widespread fame of these
Rhode Island horses, and that there were no other means by which they
could have achieved it, except by their actual performances, we are forced to
the conclusion that they were carried long distances, and in many directions,
for purely sporting purposes. That these visits would result in the transfer of a
good number of the best and fastest horses from Narragansett to Virginia
would be a natural sequence, and thus, in after years, we might look for a
strong infusion of Narragansett blood in the Virginian pacing-horse.
 It appears to be a law of our civilization that each generation produces
somebody who, out of pure love for the curious and forgotten, devotes the
best years of his life to hunting up old things that have well-nigh slipped away
from the memory of man. In this class Mr. John F. Watson stands conspicuous
in what he has done for Philadelphia and New York. In 1830 he published a
work entitled “Annals of Philadelphia and Pennsylvania,” in two volumes, and
among all the antiquated manners and habits that he again brings to our
knowledge, he has something to say about the horse of an early day:

“The late very aged T. Matlack, Esq., was passionately fond of races in
his youth. He told me of his remembrances about Race Street. In his early
days the woods were in commons, having several straggling forest trees
still remaining there, and the circular course ranging through those trees.
He said all genteel horses were pacers. A trotting-horse was deemed a
base breed. These Race Street races were mostly pace-races. His father
and others kept pacing stallions for propagating the breed.”

Mr. Watson further remarks, on the same subject: “Thomas Bradford, Esq.,
in telling me of the recollections of the races, says he was told that the
earliest races were scrub and pace-races on the ground now used as Race
Street.”
The Rev. Israel Acrelius, for many years pastor of the Swedish church of
Philadelphia, wrote a book early in the last century, under the title, “History of
New Sweden,” which has been translated into English. In describing the
country and people, in their habits and amusements, he thus speaks of the
horse:

“The horses are real ponies, and are seldom found over thirteen hands
high. He who has a good riding horse never employs him for draught,
which is also the less necessary, as journeys, for the most part, are made
on horseback. It must be the result of this, more than to any particular
breed in the horses, that the country excels in fast horses, so that horse
races are often made for very high stakes.”

It will be noted that Mr. Acrelius does not say that these races were pacing-
races; but when his remark is taken in connection with what Mr. Matlack said
about the pacers, and when it is considered that he is speaking of the speed
of the saddle horses as such, we can easily understand his true meaning. In
our turf history I supposed I was getting well back when I reached the great
race between Galloway’s Selim and Old England, in 1767, but here we find
that race was comparatively modern, and that the pacers antedated the
gallopers by many, many years.
 In 1832 Mr. Watson did the same service for New York that he had done for
Philadelphia, and published his “Annals of New York,” in which we find the
piece of horse history embodied in the extract printed on pages 126 and 127,
to which the reader will please turn.
It is hardly possible to be mistaken in assuming that Rip Van Dam’s letter
was written to some person in Philadelphia, and that Mr. Watson saw it there.
I would give a great deal for the sight of it; and if it has been preserved in
any of the public libraries of that city, either in type or in manuscript form, I
have good hopes of yet inspecting it. In one point of view, it is of exceeding
value, and that is its date. It is fully established by this letter that, as early as
1711, the Narragansetts were not only established as a breed or family, but
that their fame was already widespread. This, of necessity, carries us back
into the latter part of the seventeenth century, when their exceptional
characteristics were first developed, or began to manifest themselves. In
reaching that period we are so near the first importations of horses to the
colonies that it is no violence to either history or good sense to conclude that
the original Narragansett was one among the very earliest importations. This
plays havoc with some Rhode Island traditions, as will be seen below; but
with 1711 fixed as a point when the breed was famous, traditions must stand
aside.
While on this matter of dates, it may not be unprofitable to compare the
advent of the Narragansett with the well-known epochs in horse history. Every
schoolboy knows that the Darley Arabian and the Godolphin Arabian, say
twenty years after, were the great founders of the English race horse. The
Narragansetts had reached the very highest pinnacle of fame before the
Darley Arabian was foaled. Darley Arabian reached England about the same
year that Rip Van Dam’s Narragansett jumped over the side of the sloop and
swam ashore, and this was eighty years before there was an attempt at
publishing an English stud book. When Janus and Othello, and Traveller, and
Fearnaught, the great founders of the American race horse, first reached
Virginia, they found the Narragansett pacer had been there more than a
generation before. On the point of antiquity, therefore, the Narragansett is
older than what we designate as the thoroughbred race horse, and if he has a
lineal descendant living to-day the pacer has a longer line of speed
inheritance, at his gait, than the galloper.
The only attempt at a description of this breed that I have met with is that
given by Cooper, the novelist, in a footnote to “The Last of the Mohicans.”
This note may be accepted as history, so far as it goes, and pretends to be
history; but I am not prepared to admit that all the breed were sorrels. This
color, no doubt, prevailed in those specimens that Mr. Cooper had seen or
heard of, but I think all colors prevailed, as in other breeds. He says:

“In the State of Rhode Island there is a bay called Narragansett, so

named for a strong tribe of Indians that formerly dwelt on its banks.
Accident, or one of those unaccountable freaks which nature sometimes
plays in the animal world, gave rise to a breed of horses which were once
well known in America by the name of Narragansetts. They were small,
commonly of the color called sorrel in America, and distinguished by their
habit of pacing. Horses of this race were, and still are, in much request as
saddle-horses, on account of their hardiness, and the ease of their
movements. As they were also sure of foot, the Narragansetts were much
sought for by females who were obliged to travel over the roots and holes
in the new countries.”

Without having a minute description of so much as a single individual of the

race, I can only infer, from general descriptions, as to what their family
peculiarities of form and shape may have been. It is fully established that they
were very compact and hardy horses, and that they were not large; perhaps
averaging about fourteen and a quarter hands in height. I have met with no
intimation that they were stylish or handsome, and we think it is safe to
conclude that they were plain in their form, and low in their carriage. From
my conceptions of the horse I think one of the better-shaped Canadian
pacers, of fifteen hands or thereabouts, might be accepted as a fair
representative of the Narragansett of a hundred and fifty years ago. He was
fleet, hardy, docile, and sure-footed, but not beautiful, and it is reasonable to
suppose that the lack of style and beauty was one of the leading causes of his
becoming extinct in the land of his nativity.
In considering the causes which resulted in what we may call the dispersal
of the Narragansett pacers, and their extinction in the seat of their early fame,
we must be governed by what is reasonable and philosophical in the industrial
interests of the people, rather than look for some great overwhelming
disaster, like an earthquake, that ingulfed them in a night. In speaking of this
dispersal, and the causes which led to it, Mr. Hazard says:

“One of the causes of the loss of that famous breed here was the great
demand for them in Cuba, when that island began to cultivate sugar
extensively. The planters became suddenly rich, and wanted the pacing-
horse for themselves and their wives and daughters to ride, faster than we
could supply them, and sent an agent to this country to purchase them on
such terms as he could, but to purchase them at all events. I have heard
my father say he knew the agent very well, and he made his home at the
Rowland Brown House, at Tower Hill, where he commenced purchasing
 and shipping until all the good ones were sent off. He never let a good one
escape him. This, and the fact that they were not so well adapted to
draught as other horses, was the cause of their being neglected, and I
believe the breed is now extinct in this section. My father described the
motion of this horse as differing from others in that his backbone moved
through the air in a straight line, without inclining the rider from side to
side, as the common racker or pacer of the present day. Hence it was very
easy; and being of great power of endurance, they would perform a
journey of a hundred miles in a day, without injury to themselves or rider.”

We can understand very well how an enormous and unexpected demand

from Cuba without restriction as to price, should reduce the numbers of the
breed very materially. But it is a poor compliment to the intelligence and thrift
of the good people of Narragansett to say that, because there was a lively
demand, they killed the goose that laid the golden egg every day. It is a
slander upon that Yankee smartness which is proverbial to conclude that they
deprived themselves of the means of supplying a market that was making
them all rich. We must, therefore, look for other causes that were more
potent in producing, so marked a result.
After more than a hundred years of faithful service, of great popularity, and
of profitable returns to their breeders, the little Narragansetts began to
disappear, just as their ancestors had disappeared a century earlier. Rhode
Island was no longer a frontier settlement, but had grown into a rich and
prosperous State. Mere bridle paths through the woods had developed into
broad, smooth highways, and wheeled vehicles had taken the place of the
saddle. Under these changed conditions, the little pacer was no longer
desirable or even tolerable as a harness horse, and he was supplanted by a
larger and more stylish type of horse, better suited to the particular kind of
work required of him. This was simply the “survival of the fittest,” considering
the nature of the services required of the animal. The average height of the
Narragansett was not over fourteen hands and one inch. His neck was not
long, even for his size; he dropped rapidly on the croup, and his carriage was
low, with nothing of elegance or style in his appearance. His mane and tail
were heavy, his hind legs were crooked, his limbs and feet were of the very
best, but aside from his great speed and the smoothness of his movements
under the saddle, there was nothing very desirable or attractive about him. In
a contest with a type of the harness horse, at least one hand higher, of high
carriage and elegant appearance, there could only be one result, and that
soon decided.
As in England, so in this country, the blood of the running horse soon
worked the extermination of the pacer; not because it was stronger in
reproducing itself, perhaps, but because it had the skill and fancy of the
breeder enlisted in selecting and mating so as to make the expunging process
complete. Only a few years ago a pacing horse could hardly be found in any
of the older settled portions of the country, especially where running blood
had become fashionable. He was literally banished to the frontiers of Canada,
Indiana, Missouri, Kentucky, and Tennessee, and especially in the latter two
States, where his blood is still appreciated and preserved for the luxurious
saddle gaits which it alone transmits. In many individual cases he has shown
wonderful power in meeting and overcoming antagonistic elements, but with
the tide of running blood all against him, it was only a question of time as to
how soon he would be totally submerged.
It is only a quarter of a century ago that the first volume of “Wallace’s
American Trotting Register” was published, and then began the great task of
bringing order out of chaos. In a historical introduction to that work, I
inserted the following:

“So many pacing horses have got fast trotters, so many pacing mares
have produced fast trotters, and so many pacers have themselves become
fast trotters, and little or nothing known of their breeding, that I confess to
a degree of embarrassment, from which no philosophy relieves me. If the
facts were limited to a few individual cases we could ignore the
phenomena altogether, but, while they are by no means universal, they are
too common and apparent to be thus easily disposed of. I am not aware
that any writer has ever brought this question to the attention of the
public; much less, attempted its discussion and explanation. Indeed, it is
possible that the observations of others may not sustain me in the
prominence given these phenomena, but all will concede there are some
cases coming under this head that are unexplained, and perhaps
unexplainable. It is probable trotters from this pacing origin, and that
appear to trot, only because their progenitors paced, will not prove reliable
producers of trotters. Such an animal being in a great degree phenomenal,
should not be too highly prized in the stud, till he has proved himself a
trotting sire as well as a trotter.”

This very comprehensive little paragraph, put modestly and tentatively

rather than positively, contained a germ of thought that is to-day exerting a
very wide influence. So far as my knowledge goes, this was the first time in
which the public attention had ever been called to the intimate relations
between speed at the pace and speed at the trot. Some laughed at it as not
practical, others sneered at it as a theoretical abstraction, a few gave it some
thought, while the writers who never think left it severely alone. It required
the cumulative experiences of nearly ten years before horsemen generally
began to think about it, and then ten more before the germ had matured
itself in the minds of all intelligent men who were able to divest themselves of
their earlier prejudices. The great primary truth now stands out in high relief
that the pace and the trot are simply two forms of one and the same gait,
that lies midway between the walk and the gallop. At last the truth, dimly
foreshadowed in the paragraph above, is received and accepted, in some form
or other, almost if not quite universally. This fact and its acceptance are now
shown in all the recorded experiences of racing, and especially in the origin
and habits of action of many of the heads of trotting and pacing families, to
which the reader is referred.
At the beginning of Chapter XIII. I have labored to make plain the
proposition that the pace and the trot are simply two forms of one and the
same gait. This is evident from the fact that this gait, in one form or the other,
is the intermediate link between the walk and the gallop, and this is true
among nearly all quadrupeds. I have also there shown, and I think beyond
cavil, that the mechanism of the pace and the trot is the same, and especially
in the fact that in both forms two legs are used as one leg. That is, if the two
legs on the same side move together, we call it the pace, and if the diagonal
legs move together we call it the trot. The rhythm is the same and the sound
is the same, and by the ear no man can tell whether the movement is at the
lateral or diagonal motion. In all the varieties of steps that a horse may be
taught, and in all the methods of progression that he may naturally adopt,
there is no step or movement in which he uses two legs as one except in the
pace or the trot. From the place, therefore, which these two forms of the gait
hold, indifferently, in animal movement, between the walk and the gallop;
from the unity of action and result in the use of the same mechanism, and
from the wide disparity between the mechanism of this gait and that of all
other gaits in the action of the horse, we must conclude that the pace and the
trot are one and the same gait.
Another evidence of the unity of the two forms of the trot is to be found in
the great numbers of pacers that have been changed over to trotters and the
astonishing readiness with which they took to the new form of action. To go
back no further than the records sustain us, we find that the converted pacer
Pelham was the first horse that ever trotted in 2:28. This was in 1849, and
four years later the converted pacer Highland Maid trotted in 2:27. Twenty
years later, Occident, another, trotted in 2:16¾. These were champions of
their day, and when we come a little nearer we find that Maud S. was a pacer
and Sunol was a pacer, although neither of them ever paced in public, and the
fact that they ever paced at all was held as a kind of “home secret.” Since the
days of Pelham, literally thousands of horses have been changed from pacers
to trotters, and some hundreds have been changed from trotters to pacers
successfully. Then there are quite a number, like Jay-Eye-See, 2:10 trotting
and 2:06¼ pacing, that have made fast records at both gaits.
At one time the pacing horse Blue Bull stood at the head of all sires of
trotters in this country, and it is not known or believed that he possessed a
single drop of trotting blood. He was a very fast pacer and could do nothing
else, and a large percentage of the mares bred to him were pacers, and
practically all the others had more or less pacing blood, but his great roll of
trotters in the 2:30 list was the wonder of all horsemen of that period.
Certainly the average of the elements in his inheritance would place him very
low in theory, but in practice he struck back to some ancestor that was
strongly prepotent. The trouble in his case is practically the same as in all
other pacing stallions—the inheritance traces back to a period more remote
than any of the fast trotting stallions, but at intervals it has been neglected
and not developed until it has become weak and uncertain from lack of use.
The same may be said of the Copperbottoms, Corbeaus, Flaxtails, Hiatogas,
Davy Crockets, Pilots, Rainbows, Redbucks, St. Clairs, Tippoos, and Tom Hals,
as well as other heads of minor families that will be considered in their proper
places.
The changes that have been wrought in the status of the pacer have been
truly wonderful. Instead of being hidden away as an outcast and a disgrace to
the family, condemned to a life of inferiority and drudgery, he has been
brought out and exhibited to the public as a son and heir and the equal of the
best. In looking back over the trotting records of twenty years ago, any one
will be surprised to observe that at all the leading meetings of the whole
country there were no pacing contests. Occasionally at the minor and local
meetings of the middle Western States, a pacing contest would be given for a
small purse, in which local and obscure horses only would be engaged. Very
naturally the owners of pacing horses protested against this practical
exclusion of their favorites from the trotting meetings, and employed all their
energies in begging for admission. When they began to be really clamorous
the managers of trotting tracks argued that there could be no profit to them
in opening pacing contests, for nobody cared about seeing a pacing match,
that the entries would not fill, and especially that there would be no betting,
that, consequently, the pool-sellers would have nothing to divide with the
management. As the receipts for pool-selling and all other gambling privileges
were making the track managers rich, they were very slow about admitting an
untried element that might diminish their profits. But gradually and patiently
the pacers worked their way into the exclusive circle, and when they appeared
everybody, especially in the Eastern States, was surprised to see what
excellent horses they were and the terrific speed they showed. Instead of the
typical pacer, as formed in the popular mind, with the low head, bull neck, low
croup, hairy legs, exuberant mane and tail, and generally “Canuck” all over,
that would stop at the end of the first half-mile, here was an array of horses
that in make-up and gameness would average just as well as the same
number of trotters. This was a revelation to great multitudes of people, and
from that time forward the pacer had a fair show, on his merits. For hundreds
of years the pacer, with very few exceptions, has been able to show a little
higher rate of speed than the trotter. When Flora Temple smashed all records
in 1859 by trotting in 2:19¾, Pocahontas had drawn a wagon, five years
earlier, in 2:17½; and when Maud S. trotted in 1885 in 2:08¾, this beat all
laterals as well as diagonals, except Johnson, who the year before had paced
in 2:06¼. In 1894 Alix trotted a mile in 2:03¾, which stands the best at this
writing, but the same year Robert J. paced in 2:01½, and John R. Gentry in
2:00½ in 1896.
It is not my purpose here to undertake to discuss the reasons for the
almost continuous supremacy of the pacer over the trotter, for there is no
data from which I might frame a conclusion that would really “hold water.” At
best, therefore, I can only suggest two or three thoughts. Speed at the pace
is older, and has been longer in the process of development, than speed at
the trot. In 1747 pacing races had then been fashionable in Maryland, and
had been carried on in that colony time out of mind, but we have no trace of
trotting races. One year later (1748) “running, pacing and trotting” races had
become so numerous and so common in the colony of New Jersey that they
were declared a nuisance and suppressed by the legislative authority. My
impression from the language of the act is that it was aimed chiefly at the
running and the pacing races, and that the trotters were not very numerous.
It seems to be a reasonable conclusion that this racing mania in New Jersey
took its rise about 1665, when Governor Nicolls established the Newmarket
race course on Long Island, and if so, it had been growing in strength for over
eighty years, and if we add the time from then till now we find that the speed
of the pacer has been going on almost continuously for over two hundred
years in our own country. There is another fact entering into the rural life of
colonial times that must not be left out of consideration. The pacer was the
universal saddle horse, and the trotter never was tolerated for that service.
Every farmer’s son had his saddle horse, and when two of them met what so
natural and common as to determine then and there which was the faster, if a
little stretch of road offered? In these neighborhood rivalries, if not in actual
racing, the instinct of speed at the pace was kept alive and developed, from
generation to generation. If I am right in this little study of colonial life, we
can understand that the inheritance of speed at the pace has come down to
our own time through a great many generations of pacers, and hence the
pace is the faster gait. There is one fact in our own experience that seems to
sustain this with great force, and that is the small amount of “pounding” that
the pacer requires in order to reach the full development of his powers. There
is no need of driving a pacer to death in order to teach him how to pace, for
he already knows how to pace, and all that is needed in the way of training is
to get him into high condition. It may be possible that the lateral action is
faster than the diagonal because it is less complicated, but I can see no
anatomical reason for this, as the two legs in both gaits act as one leg. The
only difference I can see in practice is that the trotter has more up-and-down
motion than the pacer; that is, he bounds in every revolution, describing a
series of depressed curves with his back as he moves, while the pacer rises
less from the ground with his hind feet and seems to glide instead of bound;
in other words, there is less action thrown away by the pacer than the trotter,
and this may arise from the more complex action in the diagonal than in the
lateral motion.
The pacer has reached a higher acclivity than the trotter, but he is not so
well assured in his footing. His present popularity and his upward flight are
phenomenal, but the causes that have sent him there are abnormal and not
lasting. In his best individualities he is simply a gambling machine when in the
hands of unscrupulous men, to be manipulated in whatever direction he will
make the most money. Racing, at whatever gait, is not necessarily
demoralizing nor disreputable, but when it falls into the control of the
“professionals” it becomes both. So long as it remains under the control of the
breeders it is not only honorable and legitimate for them to develop and race
their stock, but it is a necessary adjunct to their business, for they must thus
bring their products before the public, if they expect to make their business
pay. Breeders should not own race tracks, or if they do, they should have no
part nor lot in the percentage uniformly paid for the gambling privilege.
The history of racing in this country teaches over and over again that
whenever the breeding and racing interest falls into the control of gamblers,
down goes the whole interest and honest men suffer with the rogues. The
grasping track managers are to-day complaining loudly that they cannot
afford to give trotting meetings unless they are allowed to bring in the pool-
sellers and make them divide the “swag” with the track. Every attempt by
legislatures to make gambling on races a felony outside the race track and a
virtue inside is a most arrant humbug and most destructive in its results. It
makes the race track a cesspool of every vice, and a stench in the nostrils of
every honest man and decent woman. The moral sense of the people all over
this country is being aroused, and if public gambling cannot be suppressed on
horse races, then history will repeat itself and horse racing will be wiped out.
The gamblers and their friends will sneer at this as “puritanism,” but no
difference about the name—it will come.
But, destructive and ruinous as gambling on races may be to the life and
moral character of young men, as well as to the material interests of honest
and reputable breeders, it hardly comes within my province to discuss it
further in this place, and therefore I will return to the consideration of the
pacer. As the historical periodicity is now looming in sight when the moral
sense of the people will command the suppression of racing of every kind, the
question becomes exceedingly pertinent as to what is to become of the
pacer? He will no longer be of any value as a gambling machine, the days of
the saddle horse are past as a means of travel, except by a few about the
parks of the cities, and however uppish and handsome he may be, he is not
and never will be a desirable driving horse in harness. We have already used
sufficient of his blood to create the American Saddle Horse, and if the saddle
horse shall produce “after his kind” we need no more infusions from the pure
pacer. In the trotter his blood has leavened everything, and in some lines
more than we desire or need. He has been a great source of trotting speed,
and if, as I am inclined to believe, Messenger’s power to transmit trotting
speed came from the old English pacer, then the pacer is the only source of
that speed. Under the condition of things as here foreshadowed he will
probably sink back into the obscurity from which he emerged twenty years
ago.
 CHAPTER XV.
THE AMERICAN SADDLE HORSE.

The saddle gaits come only from the pacer—Saddle gaits

cultivated three hundred years ago—Markham on the saddle
gaits—The military seat the best—The unity of the pace and
trot—Gaits analyzed—Saddle Horse Register—Saddle horse
progenitors—Denmark not a thoroughbred horse.
In the preceding chapters the pacer has been considered from the
standpoint of his antiquity, history, speed at the pace, and his
contributions to speed at the trot. We now come to consider him as
the founder of the best and most delightful type of saddle horses in
the world. This estimate of his quality and value had a solid
foundation in the judgment and habits of our ancestors at an early
period in our history. When our patriotic forbears entered upon the
struggle for independence, they were fully alive to the necessity of
foreign sympathy and aid. For this purpose agents were sent abroad
to enlist the good feelings and, if possible, secure co-operation of
foreign governments, especially that of France. Mr. Silas Dean was
sent to Paris, and in a communication to the secret committee of
Congress, under date of November 28, 1776, he writes: “I wish I
had here one of your best saddle horses, of the American or Rhode
Island breed—a present of that kind would be money well laid out
with a certain personage.” This was probably intended as a present
to Marie Antoinette, or some other person having great influence at
court. It further indicates that “the American or Rhode Island Saddle
Horse” was at that period, in Mr. Dean’s opinion at least, the best in
the world. (See Dean Papers, New York Historical Society, Vol. I., p.
377.)
To the man of average intelligence and candor on horse subjects it
certainly is not necessary to enter upon an elaborate discussion to
show that the saddle gaits come from the pacer, but a certain class
of writers, who neither declare nor attempt to prove their position,
constantly imply that the saddle gaits came from the
“thoroughbred.” As it is better, therefore, to make everything plain as
we go along, I will very briefly consider this point. Twelve years ago,
through Wallace’s Monthly, I presented the following questions to all
gentlemen interested in saddle-horse affairs and acquainted with
saddle-horse history: “Are all the tribes and families noted for their
saddle qualities descended in whole or in part from pacing
ancestry?” In order to cover the whole question, no difference from
what standpoint it might be considered, I added the following: “Has
any family or subfamily of saddle horses come from pure running
ancestry and without any admixture of pacing blood?” To these
questions Major Hord, then editor of the Spirit of the Farm, at
Nashville, Tennessee, a gentleman of very wide and accurate
knowledge on this subject, but strongly in favor of running blood,
made the following response through his paper:

“We can only draw conclusions from established facts in

reference to these questions, for we do not think they can be
answered otherwise, as the original ancestry of our best saddle
families is more or less clouded in obscurity. It is an established
fact, demonstrated by experience, that in order to get a saddle
horse, the quickest and most successful way is to get in the
pacing blood; it matters not how good or bad the other blood
may be, a strong dash of pacing blood will almost invariably
improve the animal for saddle purposes, and never, under any
circumstances, does a pacing cross detract from an animal’s
qualities for the saddle. Judging from these facts, we conclude
that all our saddle families are descended, at least in part, from
pacing ancestry. On the other hand, all our best saddle families
have a strong infusion of thoroughbred running blood. This blood,
however, is valuable only for the courage, bone, and finish it gives
the animal, for it imparts none of the saddle gaits; and while we
have secured the best results in breeding the saddle horse by
mixing the running and pacing blood, we have observed that too
much running blood in the stallion detracts from his success as a
sire of saddle stock. As a rule, no trainer’s skill can make a good
saddle horse out of a thoroughbred runner, whereas if you mix
 two or more strong pacing crosses on top of the running blood, a
child can gait the produce to the saddle. We have sometimes seen
good saddle horses that were thoroughbreds, but have never
seen a perfect one. Our observation and experience lead us to the
conclusion that the natural saddle gaits come from the pacers,
but to the runner we are indebted for the size, style, bone and
finish of our saddle stock.”

In this reply, when the author says “all of our saddle families are
descended, at least in part, from pacing ancestry,” and when he
adds to this that “running blood imparts none of the saddle gaits,”
he has answered both questions very fully and very satisfactorily.
The argument that running blood gives bone and finish, and all that,
is very well as a theory of breeding, but it has nothing to do with the
questions propounded. As all families of saddle horses have pacing
blood, and as there is no family without it, it may be taken as settled
that the saddle gaits come from the pacer.
I notice that at least one of the present saddle gaits was cultivated
more than three hundred years ago. Mr. Gervaise Markham, a writer
of the sixteenth century, and probably the second English author on
the horse, says: “If you buy a horse for pleasure the amble is the
best, in which you observe that he moves both his legs on one side
together, neat with complete deliberation, for if he treads too short
he is apt to stumble, if too large to cut and if shuffling or rowling he
does it slovenly and besides rids no ground. If your horse be
designed for hunting, a racking pace is most expedient, which little
differs from the amble, only is more active and nimble, whereby the
horse observes due motion, but you must not force him too eagerly,
lest being in confusion he lose all knowledge of what you design him
to, and so handle his legs carelessly.” The orthography of the work
“rack” as used by Markham is “wrack,” and this is the only place I
have met with it in any of the old authors. Webster defines the word
“rack” as “a fast amble,” but Markham uses it in contradistinction
from the amble. It is worthy of note here that the word “rack” is
older than the word “pace,” in its use as designating the particular
gait of the horse, and through all the centuries it has been retained.
Of all the gaits that are subsidiary to the pace and derived from that
gait, the rack is probably the most common, and in many sections of
the country the pacer is called a racker. Racking is often designated
as “single-footing,” and in this gait as well as in the running walk and
fox trot, there are four distinct impacts in the revolution. It follows,
then, that they are not susceptible of a very high rate of speed.
In all the services which the horse renders and in all the relations
which he bears to his master, there is no relation in which they can
be made to appear to such great mutual advantage as when the one
animal is carrying the other on his back. There is no occasion on
which a beautiful horse looks so well as when gracefully mounted
and skillfully handled by a lady or gentleman. And, I will add, there
is no occasion when a lady or gentleman, who is at home in the
saddle, looks so well as when mounted on a beautiful and well-
trained American horse. England has no saddle horses, and never
can have any till she secures American blood and adopts American
methods. The shortening of the stirrups and the swinging up and
down like a tilt-hammer is not, with our English friends, a matter of
choice, but a necessity to avoid being jolted to death. Their very silly
imitators, on this side, think they can’t afford to be out of the
fashion, because “it’s English, you know.” For safety, true gentility,
and comfort the military seat is the only seat, and if you have a
horse upon which you can’t keep that seat without punishment, he is
no saddle horse. If your doctor tells you that your liver needs
shaking up, mount an English trotting horse, but if you ride for
pleasure and fresh air, get a horse that is bred and trained to the
saddle gaits. There is just as much difference between the two
horses as the difference between a springless wagon on a cobble-
stone pavement and a richly upholstered coach on the asphalt.
The American Saddle Horse has an origin as well as a history. His
origin dates back thousands of years, and his history has been
preserved in art and in letters since the beginning of the Christian
era. For centuries he was the fashionable horse in England, and the
only horse ridden by the nobility and gentry. Away back in the reign
of Elizabeth it was not an uncommon thing to use hopples to teach
and compel trotters to pace, just as in our day hopples are often
used to teach and compel pacers to trot. In the early settlement of
the American colonies pacers were far more numerous than trotters,
and this continued to be the case till after the War of the Revolution.
The great influx of running blood after that period practically
banished the pacer to the western frontiers, where a remnant has
been preserved for the uses of the saddle; and on account of his
great speed and gameness he has again returned to popular favor in
our own day.
The walk and the canter, or short gallop, are gaits that are
common to all breeds and varieties of horses, but what are known
as “the saddle gaits” are derived wholly from the pace and are
therefore considered modifications or variations of the pace. In
regions of country where the saddle horse is bred and developed
these gaits are well known among horsemen and riders as the rack
(single-footing), the running-walk, and the fox-trot. These gaits are
not easily described so as to be understood without an example
before the eye. The rack is the most easily explained so as to be
comprehended, and it is sometimes called the slow pace. In this
movement the hind foot strikes the ground an instant before the fore
foot on the same side, then the other two feet are moved and strike
in the same way; thus there are four strokes in the revolution, in
pairs. As each foot has its own stroke we see the appositeness of
the phrase “single-footing.” The four strokes are in pairs, as one, two
—three, four, and in many cases as the speed of the horse increases
the interval between the strokes is lost and the horse is at a clean
rapid pace. As a matter of course none of these gaits in which the
horse makes four strokes instead of two in the revolution can be
speedy. They are not developed nor cultivated for speed alone, but
for the comfort and ease of the rider and the change from one to
another for the rest and ease of the horse.
These “saddle gaits” are always derivatives from the pace, and I
never have seen one that did not possess more or less pacing blood.
A careful examination of the first and second volumes of “The
National Saddle Horse Register” establishes this fact beyond all
possible contradiction. This work is a very valuable contribution to
the horse history of the country, but it is a misfortune that more care
has not been taken in the exclusion of fictitious crosses in a great
multitude of pedigrees. This trouble is specially apparent among the
supposed breeding of many of the old stallions that are inserted as
“Foundation Stock.” The tendency throughout seems to be to cover
up and hide away the very blood to which we are indebted for the
saddle horse, and to get in all the blood possible that is in direct
antagonism to the foundation of the saddle gaits. It can be accepted
as a fundamental truth in horse lore, that from the day the first
English race horse was imported into this country to the present day,
which covers a period of about one hundred and fifty years, nobody
has ever seen, either in England or in this country, a thoroughbred
horse that was a pacer. When the old race horse Denmark covered
the pacing daughter of the pacer Cockspur, the pacing blood of the
dam controlled the action and instincts of the colt, and in that colt
we have the greatest of saddle-horse sires, known as Gaines’
Denmark.
As this horse Denmark was by far the greatest of all saddle-horse
progenitors, and as his superiority has been widely attributed to his
“thoroughbred” sire Denmark, the son of imported Hedgford, I have
taken some pains to examine his pedigree. His sire was
thoroughbred, his dam and grandam were mongrels, and the
remoter crosses were impossible fictions. The fact that he ran four
miles cuts no figure as evidence of purity of blood, for horses were
running four miles in this country before the first “thoroughbred”
was born. Of the fourteen stallions that are inserted as “Foundation
Stock,” it is unfortunate that the choice seems to be practically
restricted to the State of Kentucky, while the States of Ohio, Indiana,
and Tennessee, to say nothing of Illinois, Missouri, etc., have
produced numbers of families and tribes that are much more
prominent and valuable from the true saddle-horse standpoint than
some that appear in the select list of fourteen. It is doubtless true,
however, that more attention has been paid to symmetry and style,
and to the correct development and culture of the true saddle gaits,
in Kentucky than in any of the other States. With such horses as
Gaines’ Denmark, John Dillard, Tom Hal, Brinker’s Drennon, Texas,
Peters’ Halcorn, and Copperbottom the list is all right, but the other
half-dozen are mostly young and have hardly been heard of outside
of their own immediate neighborhoods. It is a notable fact that old
Pacing Pilot does not appear as the progenitor of a saddle family.
In considering the comparative merits of the leading foundation
stallions we find that Denmark was not a success in any direction
except as the sire of handsome and stylish saddle horses. John
Dillard may not have been the equal of Denmark, in the elegance of
his progeny, but he far surpassed him in his valuable relations to the
trotter. His daughters became quite famous as the producers of
trotters of a high order, and they have over twenty in the 2:30 list.
The Tom Hals have developed phenomenal speed at the pace, and a
great deal of it, interspersed with but few trotters.
Of late years many owners of the very best material for saddle
stock have given their whole attention to the development of speed,
either at the lateral or diagonal motion, because it has been deemed
more profitable. In thus selecting, breeding and developing for
extreme speed, the adaptation to saddle purposes has been lost or
bred out. While it is true that some colts come into the world
endowed with all the saddle gaits, it is also true that skill and
patience are requisite in teaching the saddle horse good manners.
There is no imaginable use to which the horse can be put where he
will show his beautiful form and thorough education to so great
advantage as under the saddle.
 CHAPTER XVI.
THE WILD HORSES OF AMERICA.

The romances of fifty years ago—Was the horse indigenous to

this country?—The theories of the paleontologists not
satisfactory—Pedigrees of over two millions of years too long
—Outlines of horses on prehistoric ruins evidently modern—
The linguistic test among the oldest tribes of Indians fails to
discover any word for “Horse”—The horses abandoned west
of the Mississippi by the followers of De Soto about 1541
were the progenitors of the wild horses of the plains.
Fifty years ago there was much that was romantic and mysterious
in our conceptions of the real character and origin of the vast herds
of wild horses that abounded on our Western plains, and the same
remark applies to their congeners on the pampas of South America.
The wild horse and the Indian opened up a most inviting field for the
writers of romance, and current literature was flooded with “Wild
Western” stories, with the horse and the Indian as the leading
characters. We are now one generation, at least, this side of the
time when stories of this kind are either sought or read, but we are
not past the period when the origin or introduction of the horse on
this continent may be considered with interest and profit. Before
touching upon the wild horse, as known in our early history,
however, it may be well to consider, briefly, the question as to
whether he may not have been indigenous to this continent.
In our generation the spade has become a wonderful developer of
the truths of ancient history. The buried and forgotten cities of the
old world are being unearthed in Europe, Asia and Africa, and
thousands of works of art and learning that had vanished from the
face of the earth are again restored to the knowledge of the human
race. In a kindred branch of investigation the geologists and
paleontologists have been delving into the bowels of the earth—not
to find what previous generations of men had left behind them, but
to find what life was myriads of ages before man was placed on the
earth. Out of the rocks they have, literally, quarried many strange
examples of animal life that lave been buried millions of years, and
hundreds of feet below the present surface. Among these strange
petrefactions that were thus buried when the earth was young, there
is one that has been widely exploited as the “Primal Horse,” that is,
the animal from which our present horse was finally evolved. There
are three or four specimens of this petrefaction now on exhibition in
this country, the first having been discovered by Professor Marsh, of
Yale College, and now in the museum of that institution. Nearly
twenty years ago Professor Huxley, the great English naturalist,
delivered a lecture in this city on the Marsh petrefaction as his text,
in which he told us that the “Primal Horse” had, originally, five toes
on each foot, that after an indeterminate geological period he lost
the two outside toes on the hind feet, and after another million
years, more or less, he lost the outside toes of the fore feet, thus
leaving him ready to go on developing the middle toe into the foot
and hoof of the horse while the outside toes disappeared. In proof
of this he offered the fact that horses of this day have splint bones
on each side of the leg, under the knee, and these bones are the
remnants of the outside toes. This was the explanation which the
learned professor gave in disposing of the outside toes when there
were but three toes on each foot, but he failed to explain what had
become of the outside toes when there were five on each foot, and
there his whole explanation toppled to the ground.
In the American Museum of Natural History, in this city, there is a
very fine representative of this particular type of petrefactions. It is
about fifteen inches high, with a head that is disproportionately
large, and a tail that is long and slender, suggesting that of a
leopard. On each fore foot this animal has four toes, or claws, as we
might call them, and on each hind foot three claws. With these claws
this little animal might dig in the ground, or he might climb a tree
when necessary for either safety or food. Each one of these toes has
its own distinct column of joints and bone extending to the knee,