Future Perspectives for the Identification and Sequencing of Nicotinamide Adenine Dinucleotide-Capped RNAs

Abstract

Conspectus

Ribonucleic acid (RNA) is composed primarily of four canonical building blocks. In addition, more than 170 modifications contribute to its stability and function. Metabolites like nicotinamide adenine dinucleotide (NAD) were found to function as 5′-cap structures of RNA, just like 7-methylguanosine (m⁷G). The identification of NAD-capped RNA sequences was first made possible by NAD captureSeq, a multistep protocol for the specific targeting, purification, and sequencing of NAD-capped RNAs, developed in the authors’ laboratory in the year 2015. In recent years, a number of NAD-RNA identification protocols have been developed by researchers around the world. They have enabled the discovery and identification of NAD-RNAs in bacteria, archaea, yeast, plants, mice, and human cells, and they play a key role in studying the biological functions of NAD capping. We introduce the four parameters of yield, specificity, evaluability, and throughput and describe to the reader how an ideal NAD-RNA identification protocol would perform in each of these disciplines. These parameters are further used to describe and analyze existing protocols that follow two general methodologies: the capture approach and the decapping approach. Capture protocols introduce an exogenous moiety into the NAD-cap structure in order to either specifically purify or sequence NAD-capped RNAs. In decapping protocols, the NAD cap is digested to 5′-monophosphate RNA, which is then specifically targeted and sequenced. Both approaches, as well as the different protocols within them, have advantages and challenges that we evaluate based on the aforementioned parameters. In addition, we suggest improvements in order to meet the future needs of research on NAD-modified RNAs, which is beginning to emerge in the area of cell-type specific samples. A limiting factor of the capture approach is the need for large amounts of input RNA. Here we see a high potential for innovation within the key targeting step: The enzymatic modification reaction of the NAD-cap structure catalyzed by ADP-ribosyl cyclase (ADPRC) is a major contributor to the parameters of yield and specificity but has mostly seen minor changes since the pioneering protocol of NAD captureSeq and needs to be more stringently analyzed. The major challenge of the decapping approach remains the specificity of the decapping enzymes, many of which act on a variety of 5′-cap structures. Exploration of new decapping enzymes or engineering of already known enzymes could lead to improvements in NAD-specific protocols. The use of a curated set of decapping enzymes in a combinatorial approach could allow for the simultaneous detection of multiple 5′-caps. The throughput of both approaches could be greatly improved by early sample pooling. We propose that this could be achieved by introducing a barcode RNA sequence before or immediately after the NAD-RNA targeting steps. With increased processing capacity and a potential decrease in the cost per sample, protocols will gain the potential to analyze large numbers of samples from different growth conditions and treatments. This will support the search for biological roles of NAD-capped RNAs in all types of organisms.

Full text

Future Perspectives for the Identification and Sequencing of
Nicotinamide Adenine Dinucleotide-Capped RNAs
Published as part of the Accounts of Chemical Research special issue “RNA Modifications”.
Marvin Möhler and Andres Jäschke*
Cite This: Acc. Chem. Res. 2023, 56, 3000−3009
Read Online
ACCESS Metrics & More Article Recommendations
CONSPECTUS: Ribonucleic acid (RNA) is composed primarily of four canonical
building blocks. In addition, more than 170 modifications contribute to its stability
and function. Metabolites like nicotinamide adenine dinucleotide (NAD) were
found to function as 5′-cap structures of RNA, just like 7-methylguanosine (m7G).
The identification of NAD-capped RNA sequences was first made possible by NAD
captureSeq, a multistep protocol for the specific targeting, purification, and
sequencing of NAD-capped RNAs, developed in the authors’ laboratory in the year
2015. In recent years, a number of NAD-RNA identification protocols have been
developed by researchers around the world. They have enabled the discovery and
identification of NAD-RNAs in bacteria, archaea, yeast, plants, mice, and human
cells, and they play a key role in studying the biological functions of NAD capping.
We introduce the four parameters of yield, specificity, evaluability, and throughput
and describe to the reader how an ideal NAD-RNA identification protocol would
perform in each of these disciplines. These parameters are further used to describe
and analyze existing protocols that follow two general methodologies: the capture approach and the decapping approach. Capture
protocols introduce an exogenous moiety into the NAD-cap structure in order to either specifically purify or sequence NAD-capped
RNAs. In decapping protocols, the NAD cap is digested to 5′-monophosphate RNA, which is then specifically targeted and
sequenced. Both approaches, as well as the dierent protocols within them, have advantages and challenges that we evaluate based
on the aforementioned parameters. In addition, we suggest improvements in order to meet the future needs of research on NAD-
modified RNAs, which is beginning to emerge in the area of cell-type specific samples. A limiting factor of the capture approach is
the need for large amounts of input RNA. Here we see a high potential for innovation within the key targeting step: The enzymatic
modification reaction of the NAD-cap structure catalyzed by ADP-ribosyl cyclase (ADPRC) is a major contributor to the parameters
of yield and specificity but has mostly seen minor changes since the pioneering protocol of NAD captureSeq and needs to be more
stringently analyzed. The major challenge of the decapping approach remains the specificity of the decapping enzymes, many of
which act on a variety of 5′-cap structures. Exploration of new decapping enzymes or engineering of already known enzymes could
lead to improvements in NAD-specific protocols. The use of a curated set of decapping enzymes in a combinatorial approach could
allow for the simultaneous detection of multiple 5′-caps. The throughput of both approaches could be greatly improved by early
sample pooling. We propose that this could be achieved by introducing a barcode RNA sequence before or immediately after the
NAD-RNA targeting steps. With increased processing capacity and a potential decrease in the cost per sample, protocols will gain the
potential to analyze large numbers of samples from dierent growth conditions and treatments. This will support the search for
biological roles of NAD-capped RNAs in all types of organisms.
■KEY REFERENCES
•Cahová, H.; Winz, M.-L.; Höfer, K.; Nubel, G.; Jäschke,
A. NAD captureSeq indicates NAD as a bacterial cap for
a subset of regulatory RNAs. Nature 2015,519, 374.
DOI: 10.1038/nature14020.
1
NAD captureSeq was the
first protocol for the successful purification and sequencing
of NAD-capped RNAs and identified NAD-RNAs from
Escherichia coli.Following a capture approach, NAD
captureSeq introduced the use of ADP-ribosyl cyclase for the
specific modification of NAD-capped RNAs.
•Höfer, K.; Li, S.; Abele, F.; Frindert, J.; Schlotthauer, J.;
Grawenho, J.; Du, J.; Patel, D. J.; Jäschke, A. Structure
Received: July 31, 2023
Published: October 18, 2023
Articlepubs.acs.org/accounts
© 2023 The Authors. Published by
American Chemical Society 3000
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
This article is licensed under CC-BY-NC-ND 4.0

and function of the bacterial decapping enzyme NudC.
Nat. Chem. Biol. 2016,12, 730. DOI: 10.1038/
nchembio.2132.
2
Here, we described the conserved motifs
of the catalytic core of the nudix hydrolase NudC, an
enzyme capable of decapping NAD-RNA, identified its
single-strand specificity, and showed its preference for NAD-
capped RNA over NAD.
•Zhang, Y.; Kuster, D.; Schmidt, T.; Kirrmaier, D.; Nubel,
G.; Ibberson, D.; Benes, V.; Hombauer, H.; Knop, M.;
Jäschke, A. Extensive 5′-surveillance guards against non-
canonical NAD-caps of nuclear mRNAs in yeast. Nat.
Commun. 2020,11, 5508. DOI: 10.1038/s41467-020-
19326-3.
3
We found that NAD-capped RNAs in budding
yeast mainly stem from transcription initiation with NAD.
As NAD-mRNAs are not translatable by the yeast
ribosomes, a diverse surveillance machinery has evolved
and assists in the decapping and digestion of NAD-RNAs.
•Wolfram-Schauerte, M.; Pozhydaieva, N.; Grawenho,
J.; Welp, L. M.; Silbern, I.; Wulf, A.; Billau, F. A.; Glatter,
T.; Urlaub, H.; Jäschke, A.; Höfer, K. A viral ADP-
ribosyltransferase attaches RNA chains to host proteins.
Nature 2023,620, 1054−1062. DOI: 10.1038/s41586-
023-06429-2.
4
We discovered that the T4 ADP-ribosyl-
transferase ModB, involved in the bacteriophage T4
infection process, specifically attaches NAD-capped RNA
to Escherichia coli ribosomal proteins S1 and L2, a process
termed RNAylation, which is possibly connected to the viral
host-cell reprogramming strategy.
■INTRODUCTION
More than 170 modifications of ribonucleic acid (RNA) are
currently known.
5
They occur at each building block, altering
or adding to the base, ribose, or phosphate structure of the
oligonucleotide sequence, while being found at internal
positions or at the 5′- and 3′-ends. Deviation from the
canonical nucleotide structures of adenosine, cytidine,
guanosine and uridine can aect, among others, the stability
or the functional spectrum of RNA. For example, in COVID-
19 mRNA vaccines against SARS-CoV-2, the deployed
oligonucleotide sequences carry several modifications, which
heavily contribute to their eectiveness: N1-methylpseudour-
idine (m1Ψ) modifications replace canonical uridine residues
and increase the immune evasion of the mRNA,
6
and a
modified 5′-cap structure protects the mRNA from degrada-
tion and assists in the recruitment of ribosomes.
7
Rather unexpected 5′-cap structures of RNA include
metabolites, in particular the coenzymes nicotinamide adenine
dinucleotide (1) (NAD, Figure 1a) and flavin adenine
dinucleotide (FAD), and the coenzyme derivatives dephos-
pho-coenzyme A (dpCoA) and thiamine adenine dinucleo-
tides
8
(thiamine-ATP and thiamine-ADP). They have been
shown to be accepted by RNA polymerases and can serve as
noncanonical initiating nucleotides (NCIN) during tran-
scription.
9−14
Therefore, such modifications are found as cap
structures at the 5′-end of RNA (Figure 1b). While the m7G-
cap structure in eukaryotes has been known since the 1970s,
15
the first prokaryotic RNA-cap structure was found more than
30 years later, when the existence of NAD modifications in
RNA from Gram-negative and Gram-positive bacteria was
confirmed in 2009.
16
Today, the 5′-NAD modification is the
most extensively studied metabolite cap structure of RNA, and
a versatile toolbox of methods exists for both the identification
of the NAD cap and the modified RNA sequences.
Protocols for the identification of the NAD-cap structure are
based on the detection and quantitation of NAD released after
oligonucleotide digestion: Liquid chromatography−mass spec-
trometry (LC-MS)-based analytical methods can be used to
detect the NAD molecules and, through the use of internal
standards, to quantify the amount of NAD-modified RNAs in a
total RNA sample.
16,17
In contrast to the LC-MS-based
method CapQuant,
17
the NAD-capQ
18
method uses a
colorimetric assay to detect and quantify NAD-modified
RNAs based on the release of coenzymatically active NAD
molecules.
18
While these methods are ideal for analyzing the
extent of the NAD-cap structure, they cannot provide
information about the identity of NAD-RNAs. The first
protocol with this ability, NAD captureSeq, was introduced in
2015 by our laboratory.
1
Through the chemo-enzymatic
targeting of the NAD-cap structure without destruction of
the modified oligonucleotide sequence, NAD-modified RNAs
could be isolated from Escherichia coli total RNA samples and
identified by sequencing.
1
Since then, further protocols have
been developed, most of which base their applied method-
ologies on NAD captureSeq, and NAD-modified RNA
sequences have been assigned in a multitude of organisms,
ranging from bacteria
1,19−21
via archaea,
21
yeast
22
and plants
23
to human cell lines.
24
With a dedicated toolbox available for the identification of
NAD- and metabolite-capped RNAs, the collection of exciting
roles for noncanonically capped RNAs is starting to expand as
well. Recently, it was shown that NAD-modified viral RNA can
be attached to host ribosomes by a bacteriophage protein.
4
Three works confirmed the association of noncanonical
capping with human pathogens: The NAD-cap structure is
involved in modulating toxin production in Staphylococcus
aureus,
20
FAD-capping of RNA from hepatitis C virus (HCV)
was found to contribute to immune evasion during the early
infection stage,
25
and the infectivity of human immunodefi-
ciency virus 1 (HIV-1) was described to increase under
conditions of low NAD-capping levels in human host cells.
26
This Account is dedicated to analyzing the variety of
protocols available for the identification of NAD-modified
RNAs. We explain and compare the main methodologies and
discuss their strengths and weaknesses. In addition, we identify
the potential for innovation in key techniques of the current set
of protocols and suggest critical steps where improvements will
Figure 1. (a) Molecular structure of NAD+and (b) schematic structure of NAD-capped RNA.
Accounts of Chemical Research pubs.acs.org/accounts Article
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
3001

help meet the future needs in the emerging field of research on
NAD-modified RNAs.
Ideal Design of NAD-RNA Identification Protocols
We propose that protocols for the identification of NAD-RNA
(or metabolite-capped RNA in general) may be assessed based
on four parameters: Yield, Specificity, Evaluability and
Throughput. Apart from economic considerations such as
cost or hands-on time, we consider these four factors to be
central to the decision for or against a particular protocol type.
An ideal protocol could meet the criteria with the following
characteristics, which are graphically supported in Figure 2:
(1) High Yield: NAD-capped RNA appears to typically
represent about 0.1−0.2% of the transcriptome.
17
When
specifically targeting NAD-RNA, an ideal technique addresses
close to 100% of the NAD-RNAs present, while minimizing
RNA degradation. A high-yield protocol can reduce the total
RNA input to a minimum, only limited by the extent of NAD
capping.
Example: Advances in yield of NAD-RNA identification
protocols could help achieve processing capabilities for
samples that yield only low amounts of total RNA.
(2) High Specificity: RNA modifications of high (e.g., m7G-
caps) or low abundance (e.g., other metabolite caps)
17
can
pose challenges for the identification of NAD-RNAs. Ideally,
RNA species that are not NAD-modified will be addressed by
the applied techniques at close to 0%, minimizing the chances
of false-positive hits.
Example: Reducing background by improving specificity is
critical for identifying low-copy transcripts and eliminating
false positives.
(3) Qualitative and Quantitative Evaluation: The output
of the protocol allows the unambiguous assignment of all RNA
sequences in a total RNA sample as NAD-modified or non-
NAD-modified (linked to specificity), even for low-expressed
transcripts (linked to yield and specificity). It also provides the
ability to quantify NAD-RNA levels per identified transcript
species.
Figure 2. Key parameters of an ideal NAD-RNA identification protocol.
Figure 3. (a) Timeline of NAD-RNA identification protocols (ordered by date of publication) and (b) illustration of the specific targeting of NAD-
capped RNAs within the two general approaches.
Accounts of Chemical Research pubs.acs.org/accounts Article
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
3002

Example: Improved protocol and evaluation workflows will
allow researchers to identify all NAD-modified RNA species
and infer the impact of NAD modifications based on their
relative and absolute amounts.
(4) High Throughput: Ideally, the protocol allows for the
simultaneous treatment of a large number of samples.
Parallelization of sample processing can save time and help
reduce overall bias within a large set of samples.
Example: Higher throughput will allow for the processing of
RNA samples from dierent biological conditions and provide
insight into the dynamics of NAD-modified RNAs in vivo.
In light of these challenges, we strongly encourage
researchers involved in the method development for NAD-
RNA identification to strive for innovation in each of the four
areas to achieve milestones in NAD-RNA identification, as
listed in the examples above.
At the same time, the highlighted parameters can be used to
assess the innovation potential of already existing protocols
and to identify where improvements are critical. Therefore,
rather than discussing NAD-RNA identification protocols
primarily on the basis of their results, we will analyze them on
the basis of similarities and dierences in their overall design
and potential.
Protocols for the Identification of NAD-Modified RNAs
The identification of NAD-RNA sequences became possible
when our laboratory developed the first protocol for the
specific capture, purification and sequencing of NAD-modified
RNAs from total RNA samples.
1
As a result, research groups
around the world have increased their eorts in this area, and a
variety of NAD-RNA identification protocols are now available
(Figure 3a). Ordered by date of publication, the protocols of
NAD captureSeq,
1,27
CapZyme-seq,
28,29
NAD tagSeq,
23,30
NAD tagSeq II,
31
SPAAC-NAD-seq,
32
ONE-seq,
33
and, most
recently, NADcapPro and circNC
34
have enabled the discovery
and analysis of NAD-RNAs in various organisms.
In general, the available protocols can be divided into two
groups that describe the type of technique used to specifically
target NAD-modified RNAs in the presence of other RNA
species (Figure 3b). Capture approaches use the reactivity of
the NAD+component to introduce an exogenous moiety into
the structure of NAD-RNAs, allowing for specific purification
or sequencing. Decapping approaches use enzymes that digest
Figure 4. Current NAD-RNA capture protocols. (a) Chemo-enzymatic modification reaction with 4-pentyn-1-ol (2) and biotin-PEG3-azide (3) to
introduce a biotin moiety to NAD-RNA. (b) Nucleophilic alcohol substrates for ADPRC transglycosylation. (c) Clickable biotin derivatives. (d)
Central capture steps of published NAD-RNA capture protocols.
Accounts of Chemical Research pubs.acs.org/accounts Article
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
3003

the NAD cap, leaving an RNA 5′-end that can be distinguished
from other RNAs by specific techniques.
The developed capture or decapping approaches are
followed by a library preparation protocol and RNA or DNA
sequencing, and the data generated are used to identify NAD-
modified species through a bioinformatics pipeline.
Although not available for NAD-RNA, but potentially useful
for the analysis of other RNA modifications, a binding
approach can be imagined as an alternative, in which the
NAD modification is directly targeted by an NAD-binding
structure, e.g., a protein or an oligonucleotide. A major
challenge would be the identification or design of a structure
with strong binding properties and specificity for metabolite-
capped RNA, while at the same time showing low sequence
dependency for the 5′-end and minimal binding interactions
with uncapped oligonucleotides. In the following subsections,
we describe and discuss the central methods of each existing
protocol from both capture and decapping approaches.
Capture Approaches
With six protocols currently available (NAD captureSeq,
1,27
NAD tagSeq,
23,30
NAD tagSeq II,
31
SPAAC-NAD-seq,
32
ONE-
seq,
33
and NADcapPro
34
), NAD-RNA identification protocols
following a capture approach pose the more commonly
described method. What unites them is the use of ADP-
ribosyl cyclase (ADPRC),
35
an enzyme originally purified from
the sea slug Aplysia californica,
36
in their central NAD-RNA
targeting steps. ADPRC naturally catalyzes the reaction of
NAD+to cyclic ADP-ribose (cADPR), but in addition to its
cyclase activity, it has been shown to possess hydrolase activity
in the presence of an appropriate nucleophilic substrate.
37−39
We hypothesized that the ADPRC enzyme would also act on
NAD-capped RNA and we could exploit its hydrolase activity
in NAD captureSeq. The ADPRC reaction in the presence of
NAD-RNA and 4-pentyn-1-ol (2), a primary alcohol substrate
containing a nucleophilic hydroxyl group, was used to modify
the NAD-cap structure (Figure 4a). Via this clickable alkyne
moiety, the NAD-RNA could be connected to biotin-PEG3-
azide (3) by copper-catalyzed azide−alkyne cycloaddition
(CuAAC).
1
To date, all protocols following a capture approach
use the ADPRC reaction with a primary alcohol substrate
(Figure 4b) and optionally a Click reaction with a biotin
derivative (Figure 4c) or an RNA tag as the central capture
step.
After ADPRC modification and separation of NAD-RNA
from other RNA species, two main routes are followed in the
existing protocols: First, the preparation of a DNA sequencing
library including reverse transcription of the captured RNA
species and PCR amplification of the obtained cDNA, and
second, the preparation of an RNA sequencing library.
Details and characterizing features of each protocol are
summarized below, and specificity-defining reactions are
visualized in Figure 4d. For the sake of comparability, we
will not describe the protocols in a strict chronological order,
but rather based on the methodologies applied.
NAD captureSeq
The NAD captureSeq
1,27
protocol starts with 100 μg of total
RNA (original publication: from E. coli), which can optionally
be subjected to fragmentation. The central steps are ADPRC
transglycosylation of NAD-capped RNAs with 4-pentyn-1-ol
(2) and subsequent biotinylation by CuAAC, in which biotin-
PEG3-azide (3) is attached to the introduced alkyne moiety.
Biotinylated NAD-RNAs were purified using streptavidin-
sepharose beads and washed with 8 M urea buer. Library
preparation was then partially performed on-bead (3′-adapter
ligation and reverse transcription for first strand cDNA
generation) and, after hydrolysis of bead-bound RNA with
sodium hydroxide, completed o-bead (tailing, second adapter
ligation, PCR amplification, and native PAGE purification)
before submission to next-generation sequencing (NGS).
SPAAC-NAD-seq
Similar to NAD captureSeq, the central steps of the SPAAC-
NAD-seq
32
protocol comprise the biotinylation of NAD-RNAs
and separation from other RNA species. A prior m7G-depletion
step, using an anti-m7G antibody, reduced the abundance of
m7G-capped RNA (∼80% removal). With this, the protocol
could also be applied to eukaryotic samples from Arabidopsis
thaliana.
After m7G-depletion, phenol/chloroform extractions and
ethanol precipitation, 3 μg of enriched polyA-RNA was used
for ADPRC modification with 3-azidopropanol (4). To
minimize RNA losses during the procedure, the mRNA was
mixed with 100 μg of E. coli tRNA prior to the modification
reaction. Since the CuAAC reaction from NAD captureSeq led
to fragmentation, the strain-promoted alkyne−azide cyclo-
addition (SPAAC) reaction with biotin-PEG4-DBCO (5) was
introduced as a method with improved preservation of RNA
integrity. This was followed by purification of biotinylated
RNAs on streptavidin magnetic beads, washes with 8 mM urea
buer, and bead elution by incubation with hot water. Library
preparation was performed using a commercial kit that
includes RNA fragmentation, first strand synthesis by reverse
transcription with random primers, and second strand
synthesis by a DNA polymerase while using an RNase to
digest RNA, adapter ligations, and PCR amplification,
accompanied by bead purifications after several of these
steps. With this, the DNA library was ready for NGS.
NADcapPro
The NADcapPro
34
protocol uses the Dcs1-decapping enzyme
for analysis of eukaryotic RNA. Dcs1-based decapping showed
more ecient m7G-depletion than the technique applied in
SPAAC-NAD-seq.
After m7G cap digestion in 15 μg of polyA-RNA from
Saccharomyces cerevisiae, the NAD-RNA capture and purifica-
tion steps follow the general methodologies described in
SPAAC-NAD-seq, with minor adaptations in most steps,
including reagent or enzyme concentrations, buers and
reaction times. After ADPRC transglycosylation with 3-
azidopropanol (4) and SPAAC with biotin-PEG4-DBCO (5),
the biotinylated RNA is purified on streptavidin magnetic
beads, washed with 10 mM urea buer, and eluted using hot
formamide buer. After purification using RNA-binding
columns, RNA sequencing of polyA-RNA and NADcapPro-
enriched samples was performed by a third-party sequencing
company and is assumed to have followed standard
procedures, as further library preparation or sequencing
techniques were not described in detail.
ONE-seq
A key achievement of ONE-seq
33
is to combine the two-step
modification of NAD-RNA into a single reaction using a
biotin-hydroxyl nucleophile accepted by ADPRC. 100 μg of
mouse total RNA was ADPRC-transglycosylated using HEEB,
N-[2-(2-hydroxyethoxy)ethyl]biotinamide (6), which contains
a primary alcohol group and a biotin tag. This allowed for
Accounts of Chemical Research pubs.acs.org/accounts Article
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
3004

direct biotinylation of NAD-RNA, which was subjected to
anity purification on streptavidin magnetic beads. After
washing with 8 M urea buer, biotinylated NAD-RNAs were
eluted by NudC-catalyzed hydrolysis. Despite the description
of NudC side activity on metabolite caps other than NAD,
11,28
and even on m7G caps,
40
experimental proof for the
applicability of a NudC treatment to specifically elute HEEB-
modified NAD-RNAs in the presence of m7G-RNAs was
provided.
After elution, library preparation involved the use of a
commercial kit that included the general procedures as
described for SPAAC-NAD-seq, with initial polyA-RNA
selection. After PCR amplification, the library was analyzed
by NGS. A follow-up manuscript for ONE-seq on bioRxiv
describes the new tool enONE for bioinformatic analysis and
evaluation of ONE-seq data by using a diverse set of spike-in
controls.
41
NAD tagSeq and NAD tagSeq II
The general NAD tagSeq approach diers from the capture
approaches highlighted above, as NAD-RNA is not purified but
detected in the presence of the whole (or selected)
transcriptome. In NAD tagSeq
23,30
and NAD tagSeq II,
31
total RNA from A. thaliana and E. coli was subjected to
ADPRC transglycosylation with 4-pentyn-1-ol (2) and 3-
azidopropanol (4), respectively. The introduced alkyne and
azide moieties were used in CuAAC or SPAAC reactions to
introduce a 3′-azide- or 3′-DBCO-modified RNA tag.
Using milder SPAAC reaction conditions with an RNA-
DBCO tag in NAD tagSeq II allowed for a lower initial input
of total RNA (45 μg compared to 100 μg in NAD tagSeq using
CuAAC). After the tagging step, abundant rRNAs were
depleted, short RNAs (including excess tagRNA) were
removed by size selection, and the remaining transcriptome
was polyadenylated. Library preparation was performed using a
commercial kit that included adapter ligations and an optional
reverse transcription step. The library was analyzed by Oxford
Nanopore Direct RNA Sequencing. Side-by-side analysis
without an enrichment step for tagged RNAs allowed for a
direct comparison of tagged RNA reads with total transcript
reads.
Due to the side reactivity of the ADPRC reaction with
primary alcohol substrates on m7G-capped RNA, the authors
recommended that these protocols are only applied to
prokaryotic RNA.
31
It remains to be demonstrated what the
addition of an appropriate m7G-depletion step would look like
in the context of the NAD tagSeq approach. At the current
state of research, we consider the Dcs1-decapping method-
ology from NADcapPro as the most powerful m7G-depletion
tool described in the context of NAD-RNA capture approaches
thus far.
Decapping Approaches
Protocols based on a decapping approach, such as CapZyme-
seq
28,29
and circNC,
34
target metabolite caps with specific
decapping enzymes, followed by a modification step of the
digested 5′-end. Decapping also occurs naturally. A diverse
decapping machinery for NAD-RNAs has been identified in
eukaryotes, involving decapping enzymes from the nudix
hydrolase and DXO/Rai1 families, which remove the
protective NAD-cap structure and thus prepare NAD-modified
RNA for digestion.
3,24,40,42−44
Decapping enzymes such as NudC
1,2
and Rai1
24
process the
5′-NAD modification of RNA, resulting in partial or complete
(also termed deNADding) removal of the metabolite cap,
while yielding 5′-monophosphate RNA (Figure 3b). The 5′-
monophosphate can be specifically ligated to an oligonucleo-
tide, which is used for later identification by sequencing.
Therefore, it is critical that a protocol following the decapping
approach uses a decapping enzyme with reactivity for the
NAD-cap structure, while ensuring a suciently high
specificity. Details and characterizing features of the workflow
of the two described protocols are summarized below, and key
methods are visualized in Figure 5.
CapZyme-seq
For CapZyme-seq,
28,29
the rRNAs from 9 μg of Escherichia coli
total RNA per sample were depleted using a commercial kit.
Then, 2 μg of rRNA-depleted cellular RNA was treated with
NudC or Rai1 (optionally after depletion of preexisting 5′-
monophosphate ends) yielding the previously NCIN-capped
RNA as 5′-monophosphate RNA. This phosphate group was
used for 5′-ligation of a barcoded RNA adapter prior to cDNA
generation by reverse transcription and RNaseH digestion of
residual RNA. RNA or cDNA products were purified by
denaturing gel electrophoresis, including a size selection, after
each enzymatic treatment (decapping, ligation, and reverse
transcription). Finally, the generated cDNA was amplified by
emulsion PCR and purified using a commercial kit before
barcoded DNA libraries were analyzed by NGS.
The read counts for NCIN-capped RNAs were compared
with the results for an RNA 5′-polyphosphatase (Rpp)-treated
sample set, in which triphosphate-RNA was decapped into
monophosphate-RNA, allowing for quantitation of NCIN-
capping levels per identified transcript.
circNC
In contrast to CapZyme-seq, which can be used for the de novo
identification of NCIN-capped RNAs, the circNC
34
method is
intended as an additional validation and quantitation tool
following another NAD-RNA identification protocol or for the
analysis of a specific RNA of interest. The comparison of 5 μg
polyA-RNA samples from Saccharomyces cerevisiae, which were
5′-dephosphorylated and decapped with either Rai1 (process-
Figure 5. (a) Decapping of NAD-RNA with Rai1 and NudC to 5′-monophosphate RNA, and (b) central 5′-monophosphate RNA targeting steps
of published decapping protocols.
Accounts of Chemical Research pubs.acs.org/accounts Article
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
3005

ing NCIN-capped RNAs) or the mRNA decapping enzyme
(MDE1, acting on m7G-capped RNAs), allowed for the
quantitation of NCIN-capping levels of prepicked transcripts
identified with NADcapPro.
34
Purifications were performed
using commercially available RNA cleanup columns. For the
ligation step, the circNC protocol opts for ligation of the
generated 5′-monophosphate ends of mRNAs to their
respective polyadenylated 3′-ends, resulting in a circularization
of the decapped RNA transcripts.
45
These were PCR-amplified
using gene-specific primers, generating amplicons that span
across the 5′,3′-linkage of an RNA of interest, which were
purified and size selected by agarose gel electrophoresis. NGS
of amplicons was performed by a third-party sequencing
company and is assumed to have followed standard
procedures, as further library preparation was not described
in detail.
■DISCUSSION AND INNOVATION POTENTIAL
In reviewing the literature in the field of NAD-capped RNA
research, we identified a recurring element of the need for
more accurate and sensitive methods. We believe that reducing
the total RNA demand, ensuring precise data interpretation
through specificity and reproducibility, and increasing the
sample processing capabilities of existing protocols are some of
the major challenges that need to be addressed. We have
introduced the parameters of yield, specificity, evaluability and
throughput (Figure 2) to resemble those future needs and to
identify methods of existing NAD-RNA identification proto-
cols that would greatly benefit from innovation.
Capture Approaches
For the identification of NAD-RNAs following a capture
approach, a diverse set of protocols is currently available. What
unites them is the enzymatic modification of the NAD cap
using the ADPRC enzyme with primary alcohol substrates as
the key technique for targeting NAD-RNAs.
The yield of full-length RNA transcripts was improved by
reducing RNA degradation by switching from CuAAC to
SPAAC reactions. One thing to keep in mind is that the
strained internal alkyne group of DBCO may have side
reactivity with other RNA modifications of known and
unknown identity. Ideally, the second modification step can
be omitted through direct biotinylation, as shown in ONE-seq.
Those improvements, however, have not allowed for a
significant reduction in the total RNA input: With a general
estimate of polyA- or mRNA content in eukaryotic total RNA
of 1−3%,
46−48
the majority of capture approaches require 100
μg of total RNA per replicate. With biological triplicates and a
set of negative controls, this results in a requirement of at least
600 μg of total RNA. This input needs to be severely reduced
to enable novel applications, e.g., cell-type specific samples. In
comparison, standard Illumina next-generation sequencing
requires around 100 ng of total RNA or less,
49
and with the
Oxford Nanopore direct RNA sequencing technology, a usual
input of 500 ng of total RNA or 50 ng of polyA-RNA is
recommended.
50
For the N6-methyladenosine (m6A) RNA
modification, for example, identification and mapping is
already possible with single-cell sequencing technologies.
51,52
We want to direct the curiosity of researchers toward the
initial ADPRC modification with primary alcohol substrates,
where the highest potential for innovation may be hidden. We
were surprised to see that neither protocol describes reaction
kinetics or yield of this key reaction for capturing NAD-RNA.
In fact, the primary alcohol substrates are far from optimal
substrates for ADPRC: Concentrations of above 1 M for 4-
pentyn-1-ol (2)
1,23
and 3-azidopropanol (4)
31,32,34
are
required to reach satisfactory yields of NAD-RNA modifica-
tion.
Regarding specificity, the reaction setup with ADPRC and
primary alcohols was revealed to possess a side reactivity on
m7G-capped RNAs. Due to the typically low levels of NAD
capping compared to m7G capping in eukaryotic samples, said
side reactivity can lead to a significant amount of captured
m7G-RNAs, creating a downstream problem in evaluability
(high background, false-positives, etc.). To reduce this risk of
bias, several m7G-depletion techniques have been intro-
duced.
32−34
However, this usually results in a reduction of
m7G capping, not in complete elimination. Therefore, the
processing of eukaryotic samples using a capture protocol
without or with insucient m7G depletion might have led to
issues in qualitative and quantitative evaluation in published
work and might benefit from reevaluation.
We hypothesize that the large excess of primary alcohol
nucleophiles is a cause for the decrease in specificity. Indeed, it
seems that higher nucleophile concentrations increase the side
reactivity of HEEB (6) toward m7G-RNA, which is why in
ONE-seq a concentration of 0.1 M has been chosen. However,
we believe that better nucleophilic substrates for ADPRC
should allow for the use of lower concentrations.
Many nucleophilic substrates have been identified for the
ADPRC transglycosylation with NAD+.
37,38
Aromatic N-
nucleophilic substrates, which are closer to the natural
ADPRC substrate nicotinamide, in particular showed fast
kinetics and high yield.
38
Therefore, we believe that new
substrates for the specific targeting of NAD-RNAs with
ADPRC must be investigated. Our own research (unpublished
data) indicates that novel substrates bear the potential to
significantly increase the reactivity, yield and specificity (plus
evaluability) of the key enzymatic modification reaction of
NAD-capped RNA, thereby greatly reducing the total RNA
input and the side activity toward m7G-caps.
In terms of evaluability, we acknowledge that the NAD
tagSeq approach oers the great advantage of quantification of
NAD-RNAs relative to total transcripts of the respective genes,
under the assumption of highly ecient ADPRC reaction and
tagging steps. Through technical advances, we expect the
method to overcome its limitation of disregarding short RNAs
(shortest read length around 100 nt
23,31
). One idea to extend
the capabilities of the tagSeq approach could be the
introduction of a biotin moiety at the 5′-end of the tagRNA,
which is readily available from commercial suppliers.
In general, protocols published more recently showed clear
improvements in evaluability by lowering the background and
thus the threshold for the assignment of NAD capping to RNA
transcripts. Great innovation has been shown by implementing
measures aecting specificity, purification, (spike-in) control
samples, and even in the combination of capture and
decapping approach (synergy of NADcapPro and circNC).
At this point, we cannot stress enough the importance of
functional controls for the qualitative evaluability of sequenc-
ing data. If a bias is known, e.g., for the ADPRC reaction, there
should always be a valid negative control that deals with this
bias. We encourage the use of a dummy substrate, such as a
primary alcohol nucleophile without a reactive handle, instead
of a negative control devoid of the ADPRC enzyme.
Accounts of Chemical Research pubs.acs.org/accounts Article
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
3006

We also want to emphasize the often overlooked risk of bias
due to unspecific binding of non-NAD-RNAs through RNA−
RNA interactions, caused by an imbalance of bound
biotinylated RNAs in positive and negative controls during
purification on streptavidin magnetic beads (unpublished
results), which can lead to false-positives that are hardly
distinguishable from low-copy NAD-RNA transcripts. For
these and other reasons, like m7G-cap side reactivity, we want
to encourage the use of a diverse set of spike-in control RNAs.
Oligonucleotides with dierent 5′-end structures, for example,
5′-ppp or NAD+-, m7G- and other caps, and various lengths
can serve as valuable tools for identifying the eects of
unspecific binding or side reactivity (specificity) and, hence, be
used for the qualitative and quantitative evaluation of methods
from both capture and decapping approaches.
One of our defined parameters with a clear need for
innovation is throughput. Existing protocols process each
replicate and control separately. We believe that an early
barcoding step could allow the pooling of all replicates and
controls per sample to reduce the workload and significantly
increase the processing capacity of a protocol based on a
capture approach with biotinylation. At the same time, it would
reduce downstream bias if replicates and negative controls
could be processed in a single vial. Design-wise, this could look
similar to a recently developed library preparation protocol for
small RNA sequencing, in which barcoded adapters are ligated
to the RNA 3′-end before streptavidin magnetic bead capture
and sample pooling.
53
In the case of NAD-RNA capture
protocols, this approach could be realized by implementing a
3′-ligation with barcoded adapters as the second protocol step,
directly after the ADPRC transglycosylation reactions.
Looking even further into the future, the combination of
improvements in yield, specificity, evaluability and throughput
may even allow for (semi)automation of NAD-RNA capture
protocols and the cost-eective acquisition of large amounts of
data from a variety of relevant organisms.
Decapping Approaches
The decapping approach for the identification of NAD-RNAs
uses decapping enzymes to target capped RNA species. Of the
two protocols currently available, CapZyme-seq can be used
for the analysis of previously unstudied samples and circNC
can complement other NAD-RNA identification protocols.
Since the enzymatic decapping reactions, comparable to the
ADPRC capture step, have mostly not been studied in detail,
the decapping eciency, the influence of RNA 5′-sequence
and -structure, and overall yield are dicult to assess. These
characteristics would be of particular interest for the
comparison of capped and noncapped RNA transcripts and
for the quantitation of capping levels (evaluability). However,
the total RNA input is significantly lower than that for state-of-
the-art capture approaches.
By using general decapping enzymes like Rpp
28
or MDE1,
34
decapping approaches enable the quantitative evaluation of
NCIN-capping levels. However, the specificity of the key
decapping step remains the Achilles heel of the method. In the
case of NAD-RNA, NudC showed activity on m7G-capped
RNAs, making it suitable for application on eukaryotic samples
only after m7G depletion. Rai1 lacks this activity but, like
NudC, has been described to react with other NCIN-cap
structures such as FAD and dpCoA. Therefore, researchers
need to be aware of this specificity issue for the proper
interpretation of the data obtained. However, investigation of
active site mutations and their eect on the processing of
NCIN-capped RNAs may be beneficial for the development of
more specific decapping enzymes in the future.
A major advantage of decapping protocols is the universality
of most of their steps: Once a specific decapping enzyme is
found, a developed protocol can be applied with very little
modification to a new metabolite cap of interest, as recently
demonstrated in the application of CapZyme-seq with the
nudix pyrophosphohydrolase AtNUDX23
54
for the identifica-
tion of FAD-capped RNA in hepatitis C.
25
In the future, a
refined toolbox of decapping enzymes combined with an early
barcoding strategy could be used to detect a variety of capped
RNAs from the same sample and increase the throughput of
the method. Comparing the data sets obtained with dierent
decapping enzymes, for example, in a combinatorial manner,
might even allow for improved evaluability, ideally involving
the unambiguous assignment of dierent cap structures.
■AUTHOR INFORMATION
Corresponding Author
Andres Jäschke −Institute of Pharmacy and Molecular
Biotechnology, Heidelberg University, 69120 Heidelberg,
Germany; orcid.org/0000-0002-4625-2655;
Email: jaeschke@uni-hd.de;https://www.jaschkelab.de/
Author
Marvin Möhler −Institute of Pharmacy and Molecular
Biotechnology, Heidelberg University, 69120 Heidelberg,
Germany
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.accounts.3c00446
Author Contributions
CRediT: Marvin Mohler conceptualization, writing-original
draft, writing-review & editing; Andres Jaschke conceptualiza-
tion, writing-review & editing.
Funding
RNA modification research in the authors’ lab has received
funding from the European Research Council under the
European Union’s Horizon 2020 research and innovation
program (Grant 882789 RNACoenzyme) and from the
German Research Council (DFG; Project 439669440,
TRR319, Subproject A02). Marvin Mohler is supported by a
doctoral scholarship of the German Academic Scholarship
Foundation.
Notes
The authors declare no competing financial interest.
Biographies
Marvin Mohler received his B.Sc. (2016) and M.Sc. (2019) degrees
from Heidelberg University, including an Erasmus exchange at the
University of York in 2017−2018, and pursued his Ph.D. degree at
Heidelberg University working with Andres Jaschke. His research
interests include metabolite-capped RNAs and their identification.
Andres Jaschke studied chemistry at Humboldt University in Berlin,
where he received his Ph.D. degree in 1993 working with Dieter
Cech. After graduation, he spent two years as a postdoctoral fellow
with Alexander Rich at MIT (Cambridge, U.S.A.). From 1995 to
2002, Andres Jaschke was a Junior group leader at Free University in
Berlin. Since 2002, he has been full professor of pharmaceutical and
bioorganic chemistry at Heidelberg University. His research group is
Accounts of Chemical Research pubs.acs.org/accounts Article
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
3007

focused on the chemical biology of nucleic acids. Current research
topics include the chemistry and biology of natural RNA
modifications, new tools for imaging RNAs in living cells, and
photoswitchable biomolecules.
■REFERENCES
(1) Cahová, H.; Winz, M.-L.; Höfer, K.; Nubel, G.; Jäschke, A. NAD
captureSeq indicates NAD as a bacterial cap for a subset of regulatory
RNAs. Nature 2015,519 (7543), 374.
(2) Höfer, K.; Li, S.; Abele, F.; Frindert, J.; Schlotthauer, J.;
Grawenhoff, J.; Du, J.; Patel, D. J.; Jäschke, A. Structure and function
of the bacterial decapping enzyme NudC. Nat. Chem. Biol. 2016,12
(9), 730.
(3) Zhang, Y.; Kuster, D.; Schmidt, T.; Kirrmaier, D.; Nubel, G.;
Ibberson, D.; Benes, V.; Hombauer, H.; Knop, M.; Jäschke, A.
Extensive 5′-surveillance guards against non-canonical NAD-caps of
nuclear mRNAs in yeast. Nat. Commun. 2020,11 (1), 5508.
(4) Wolfram-Schauerte, M.; Pozhydaieva, N.; Grawenhoff, J.; Welp,
L. M.; Silbern, I.; Wulf, A.; Billau, F. A.; Glatter, T.; Urlaub, H.;
Jäschke, A.; Höfer, K. A viral ADP-ribosyltransferase attaches RNA
chains to host proteins. Nature 2023,620 (7976), 1054−1062.
(5) Boccaletto, P.; Stefaniak, F.; Ray, A.; Cappannini, A.; Mukherjee,
S.; Purta, E.; Kurkowska, M.; Shirvanizadeh, N.; Destefanis, E.; Groza,
P.; Avsar, G.; Romitelli, A.; Pir, P.; Dassi, E.; Conticello, S. G.; Aguilo,
F.; Bujnicki, J. M. MODOMICS: a database of RNA modification
pathways. 2021 update. Nucleic Acids Res. 2022,50 (D1), D231−
D235.
(6) Nance, K. D.; Meier, J. L. Modifications in an Emergency: The
Role of N1-Methylpseudouridine in COVID-19 Vaccines. ACS Cent.
Sci. 2021,7(5), 748−756.
(7) Henderson, J. M.; Ujita, A.; Hill, E.; Yousif-Rosales, S.; Smith,
C.; Ko, N.; McReynolds, T.; Cabral, C. R.; Escamilla-Powers, J. R.;
Houston, M. E. Cap 1 Messenger RNA Synthesis with Co-
transcriptional CleanCap®Analog by In Vitro Transcription. Curr.
Protoc. 2021,1(2), No. e39.
(8) Bettendorff, L.; Wirtzfeld, B.; Makarchikov, A. F.; Mazzucchelli,
G.; Frédérich, M.; Gigliobianco, T.; Gangolf, M.; De Pauw, E.;
Angenot, L.; Wins, P. Discovery of a natural thiamine adenine
nucleotide. Nat. Chem. Biol. 2007,3(4), 211.
(9) Malygin, A. G.; Shemyakin, M. F. Adenosine, NAD and FAD can
initiate template-dependent RNA a synthesis catalyzed by Escherichia
Coli RNA polymerase. FEBS Lett. 1979,102 (1), 51−54.
(10) Huang, F. Efficient incorporation of CoA, NAD and FAD into
RNA by in vitro transcription. Nucleic Acids Res. 2003,31 (3), No. e8.
(11) Bird, J. G.; Zhang, Y.; Tian, Y.; Panova, N.; Barvík, I.; Greene,
L.; Liu, M.; Buckley, B.; Krásny, L.; Lee, J. K.; et al. The mechanism of
RNA 5′capping with NAD+, NADH and desphospho-CoA. Nature
2016,535 (7612), 444.
(12) Mlynarska-Cieslak, A.; Depaix, A.; Grudzien-Nogalska, E.;
Sikorski, P. J.; Warminski, M.; Kiledjian, M.; Jemielity, J.; Kowalska, J.
Nicotinamide-Containing Di- and Trinucleotides as Chemical Tools
for Studies of NAD-Capped RNAs. Org. Lett. 2018,20 (23), 7650−
7655.
(13) Bird, J. G.; Basu, U.; Kuster, D.; Ramachandran, A.; Grudzien-
Nogalska, E.; Towheed, A.; Wallace, D. C.; Kiledjian, M.; Temiakov,
D.; Patel, S. S.; et al. Highly efficient 5′capping of mitochondrial
RNA with NAD+ and NADH by yeast and human mitochondrial
RNA polymerase. Elife 2018,7, No. e42179.
(14) Möhler, M.; Höfer, K.; Jäschke, A. Synthesis of 5′-Thiamine-
Capped RNA. Molecules 2020,25 (23), 5492.
(15) Furuichi, Y.; Miura, K.-I. A blocked structure at the 5′terminus
of mRNA from cytoplasmic polyhedrosis virus. Nature 1975,253
(5490), 374−375.
(16) Chen, Y. G.; Kowtoniuk, W. E.; Agarwal, I.; Shen, Y.; Liu, D. R.
LC/MS analysis of cellular RNA reveals NAD-linked RNA. Nat.
Chem. Biol. 2009,5(12), 879.
(17) Wang, J.; Alvin Chew, B. L.; Lai, Y.; Dong, H.; Xu, L.;
Balamkundu, S.; Cai, W. M.; Cui, L.; Liu, C. F.; Fu, X.-Y.; Lin, Z.; Shi,
P.-Y.; Lu, T. K.; Luo, D.; Jaffrey, S. R.; Dedon, P. C. Quantifying the
RNA cap epitranscriptome reveals novel caps in cellular and viral
RNA. Nucleic Acids Res. 2019,47 (20), No. e130.
(18) Grudzien-Nogalska, E.; Bird, J. G.; Nickels, B. E.; Kiledjian, M.
“NAD-capQ” detection and quantitation of NAD caps. RNA 2018,24
(10), 1418−1425.
(19) Frindert, J.; Zhang, Y.; Nubel, G.; Kahloon, M.; Kolmar, L.;
Hotz-Wagenblatt, A.; Burhenne, J.; Haefeli, W. E.; Jäschke, A.
Identification, biosynthesis, and decapping of NAD-capped RNAs in
B. subtilis. Cell Rep. 2018,24 (7), 1890−1901.
(20) Morales-Filloy, H. G.; Zhang, Y.; Nubel, G.; George, S. E.;
Korn, N.; Wolz, C.; Jäschke, A. The 5′NAD Cap of RNAIII
Modulates Toxin Production in Staphylococcus aureus Isolates. J.
Bacteriol. 2020,202 (6), e00591−19.
(21) Ruiz-Larrabeiti, O.; Benoni, R.; Zemlianski, V.; Hanisáková, N.;
Schwarz, M.; Brezovská, B.; Benoni, B.; Hnilicová, J.; Kaberdin, V. R.;
Cahová, H.; Vítezová, M.; Prevorovsky, M.; Krásny, L. NAD+ capping
of RNA in Archaea and Mycobacteria. bioRxiv 2021,DOI: 10.1101/
2021.12.14.472595.
(22) Walters, R. W.; Matheny, T.; Mizoue, L. S.; Rao, B. S.;
Muhlrad, D.; Parker, R. Identification of NAD+ capped mRNAs in
Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. U.S.A. 2017,114 (3),
480−485.
(23) Zhang, H.; Zhong, H.; Zhang, S.; Shao, X.; Ni, M.; Cai, Z.;
Chen, X.; Xia, Y. NAD tagSeq reveals that NAD+-capped RNAs are
mostly produced from a large number of protein-coding genes in
Arabidopsis. Proc. Natl. Acad. Sci. U.S.A. 2019,116 (24), 12072−
12077.
(24) Jiao, X.; Doamekpor, S. K.; Bird, J. G.; Nickels, B. E.; Tong, L.;
Hart, R. P.; Kiledjian, M. 5′end nicotinamide adenine dinucleotide
cap in human cells promotes RNA decay through DXO-mediated
deNADding. Cell 2017,168 (6), 1015−1027.
(25) Sherwood, A. V.; Rivera-Rangel, L. R.; Ryberg, L. A.; Larsen, H.
S.; Anker, K. M.; Costa, R.; Vagbo, C. B.; Jakljevic, E.; Pham, L. V.;
Fernandez-Antunez, C.; Indrisiunaite, G.; Podolska-Charlery, A.;
Grothen, J. E. R.; Langvad, N. W.; Fossat, N.; Offersgaard, A.; Al-
Chaer, A.; Nielsen, L.; Kusnierczyk, A.; Solund, C.; Weis, N.;
Gottwein, J. M.; Holmbeck, K.; Bottaro, S.; Ramirez, S.; Bukh, J.;
Scheel, T. K. H.; Vinther, J. Hepatitis C virus RNA is 5′-capped with
flavin adenine dinucleotide. Nature 2023,619 (7971), 811−818.
(26) Benoni, B.; Benoni, R.; Trylcova, J.; Grab, K.; Paces, J.; Weber,
J.; Stanek, D.; Kowalska, J.; Bednarova, L.; Keckesova, Z.; Gahurova,
L.; Cahova, H. HIV-1 infection reduces NAD capping of host cell
snRNA and snoRNA. bioRxiv 2022,DOI: 10.1101/
2022.11.10.515957.
(27) Winz, M.-L.; Cahová, H.; Nubel, G.; Frindert, J.; Höfer, K.;
Jäschke, A. Capture and sequencing of NAD-capped RNA sequences
with NAD captureSeq. Nat. Protoc. 2017,12 (1), 122.
(28) Vvedenskaya, I. O.; Bird, J. G.; Zhang, Y.; Zhang, Y.; Jiao, X.;
Barvík, I.; Krásny, L.; Kiledjian, M.; Taylor, D. M.; Ebright, R. H.;
Nickels, B. E. CapZyme-Seq Comprehensively Defines Promoter-
Sequence Determinants for RNA 5′Capping with NAD+. Mol. Cell
2018,70 (3), 553−564.
(29) Vvedenskaya, I. O.; Nickels, B. E. CapZyme-Seq: A 5′-RNA-
Seq Method for Differential Detection and Quantitation of NAD-
Capped and Uncapped 5′-Triphosphate RNA. STAR Protoc. 2020,1
(1), 100002.
(30) Shao, X.; Zhang, H.; Yang, Z.; Zhong, H.; Xia, Y.; Cai, Z. NAD
tagSeq for transcriptome-wide identification and characterization of
NAD+-capped RNAs. Nat. Protoc. 2020,15 (9), 2813−2836.
(31) Zhang, H.; Zhong, H.; Wang, X.; Zhang, S.; Shao, X.; Hu, H.;
Yu, Z.; Cai, Z.; Chen, X.; Xia, Y. Use of NAD tagSeq II to identify
growth phase-dependent alterations in E. coli RNA NAD+ capping.
Proc. Natl. Acad. Sci. U.S.A. 2021,118 (14), No. e2026183118.
(32) Hu, H.; Flynn, N.; Zhang, H.; You, C.; Hang, R.; Wang, X.;
Zhong, H.; Chan, Z.; Xia, Y.; Chen, X. SPAAC-NAD-seq, a sensitive
and accurate method to profile NAD+-capped transcripts. Proc. Natl.
Acad. Sci. U.S.A. 2021,118 (13), No. e2025595118.
Accounts of Chemical Research pubs.acs.org/accounts Article
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
3008

(33) Niu, K.; Zhang, J.; Ge, S.; Li, D.; Sun, K.; You, Y.; Qiu, J.;
Wang, K.; Wang, X.; Liu, R.; Liu, Y.; Li, B.; Zhu, Z.-J.; Qu, L.; Jiang,
H.; Liu, N. ONE-seq: epitranscriptome and gene-specific profiling of
NAD-capped RNA. Nucleic Acids Res. 2023,51 (2), No. e12.
(34) Sharma, S.; Yang, J.; Favate, J.; Shah, P.; Kiledjian, M.
NADcapPro and circNC: methods for accurate profiling of NAD and
non-canonical RNA caps in eukaryotes. Commun. Biol. 2023,6(1),
406.
(35) Lee, H. C.; Aarhus, R. ADP-ribosyl cyclase: an enzyme that
cyclizes NAD+ into a calcium-mobilizing metabolite. Cell Regul. 1991,
2(3), 203−209.
(36) Hellmich, M. R.; Strumwasser, F. Purification and character-
ization of a molluscan egg-specific NADase, a second-messenger
enzyme. Cell Regul. 1991,2(3), 193−202.
(37) Lee, H. C.; Aarhus, R. Structural Determinants of Nicotinic
Acid Adenine Dinucleotide Phosphate Important for Its Calcium-
mobilizing Activity. J. Biol. Chem. 1997,272 (33), 20378−20383.
(38) Preugschat, F.; Tomberlin, G. H.; Porter, D. J. The base
exchange reaction of NAD+ glycohydrolase: identification of novel
heterocyclic alternative substrates. Arch. Biochem. Biophys. 2008,479
(2), 114−120.
(39) Liu, Q.; Graeff, R.; Kriksunov, I. A.; Jiang, H.; Zhang, B.;
Oppenheimer, N.; Lin, H.; Potter, B. V. L.; Lee, H. C.; Hao, Q.
Structural Basis for Enzymatic Evolution from a Dedicated ADP-
ribosyl Cyclase to a Multifunctional NAD Hydrolase. J. Biol. Chem.
2009,284 (40), 27637−27645.
(40) Grudzien-Nogalska, E.; Wu, Y.; Jiao, X.; Cui, H.; Mateyak, M.
K.; Hart, R. P.; Tong, L.; Kiledjian, M. Structural and mechanistic
basis of mammalian Nudt12 RNA deNADding. Nat. Chem. Biol. 2019,
15 (6), 575−582.
(41) Li, D.; Ge, S.; Liu, Y.; Pan, M.; Wang, X.; Han, G.; Zou, S.; Liu,
R.; Niu, K.; Zhao, C.; Liu, N.; Qu, L. Epitranscriptome analysis of
NAD-capped RNA by spike-in-based normalization. bioRxiv 2023,
DOI: 10.1101/2023.03.23.534034.
(42) Sharma, S.; Grudzien-Nogalska, E.; Hamilton, K.; Jiao, X.;
Yang, J.; Tong, L.; Kiledjian, M. Mammalian Nudix proteins cleave
nucleotide metabolite caps on RNAs. Nucleic Acids Res. 2020,48 (12),
6788−6798.
(43) Pan, S.; Li, K.-e.; Huang, W.; Zhong, H.; Wu, H.; Wang, Y.;
Zhang, H.; Cai, Z.; Guo, H.; Chen, X.; Xia, Y. Arabidopsis DXO1
possesses deNADding and exonuclease activities and its mutation
affects defense-related and photosynthetic gene expression. J. Integr.
Plant Biol. 2020,62 (7), 967−983.
(44) Sharma, S.; Yang, J.; Grudzien-Nogalska, E.; Shivas, J.; Kwan,
K. Y.; Kiledjian, M. Xrn1 is a deNADding enzyme modulating
mitochondrial NAD-capped RNA. Nat. Commun. 2022,13 (1), 889.
(45) Couttet, P.; Fromont-Racine, M.; Steel, D.; Pictet, R.; Grange,
T. Messenger RNA deadenylylation precedes decapping in mamma-
lian cells. Proc. Natl. Acad. Sci. U.S.A. 1997,94 (11), 5628−5633.
(46) Tobin, E. M.; Klein, A. O. Isolation and Translation of Plant
Messenger RNA. Plant Physiol. 1975,56 (1), 88−92.
(47) Hastie, N. D.; Bishop, J. O. The expression of three abundance
classes of messenger RNA in mouse tissues. Cell 1976,9(4), 761−
774.
(48) Hereford, L. M.; Rosbash, M. Number and distribution of
polyadenylated RNA sequences in yeast. Cell 1977,10 (3), 453−462.
(49) Naphade, S.; Bhatnagar, R.; Hanson-Smith, V.; Choi, I.; Zhang,
A. Systematic comparative analysis of strand-specific RNA-seq library
preparation methods for low input samples. Sci. Rep. 2022,12 (1),
1789.
(50) Jain, M.; Abu-Shumays, R.; Olsen, H. E.; Akeson, M. Advances
in nanopore direct RNA sequencing. Nat. Methods 2022,19 (10),
1160−1164.
(51) Tegowski, M.; Flamand, M. N.; Meyer, K. D. scDART-seq
reveals distinct m6A signatures and mRNA methylation heterogeneity
in single cells. Mol. Cell 2022,82 (4), 868−878.
(52) Li, Y.; Wang, Y.; Vera-Rodriguez, M.; Lindeman, L. C.;
Skuggen, L. E.; Rasmussen, E. M. K.; Jermstad, I.; Khan, S.; Fosslie,
M.; Skuland, T.; Indahl, M.; Khodeer, S.; Klemsdal, E. K.; Jin, K.-X.;
Dalen, K. T.; Fedorcsak, P.; Greggains, G. D.; Lerdrup, M.;
Klungland, A.; Au, K. F.; Dahl, J. A. Single-cell m6A mapping in
vivo using picoMeRIP−seq. Nat. Biotechnol. 2023,DOI: 10.1038/
s41587-023-01831-7.
(53) Watkins, C. P.; Zhang, W.; Wylder, A. C.; Katanski, C. D.; Pan,
T. A multiplex platform for small RNA sequencing elucidates
multifaceted tRNA stress response and translational regulation. Nat.
Commun. 2022,13 (1), 2491.
(54) Maruta, T.; Yoshimoto, T.; Ito, D.; Ogawa, T.; Tamoi, M.;
Yoshimura, K.; Shigeoka, S. An Arabidopsis FAD Pyrophosphohy-
drolase, AtNUDX23, is Involved in Flavin Homeostasis. Plant Cell
Physiol. 2012,53 (6), 1106−1116.
Accounts of Chemical Research pubs.acs.org/accounts Article
https://doi.org/10.1021/acs.accounts.3c00446
Acc. Chem. Res. 2023, 56, 3000−3009
3009