Enzymatic combinatorial nucleoside deletion scanning mutagenesis of functional RNA

Article abstract of DOI:10.1039/c4cc04719b

We describe a general and simple method to identify catalytically and structurally important nucleotides in functional RNAs. Our approach is based on statistical replacement of each nucleoside with a non-nucleosidic spacer (C3 linker, Δ), followed by sepa

Full text of DOI:10.1039/c4cc04719b

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Enzymatic combinatorial nucleoside deletion
scanning mutagenesis of functional RNA†
Cite this: Chem. Commun., 2014,
0, 10937
5
Katarzyna Wawrzyniak-Turek and Claudia Hobartner*
¨
Received 21st June 2014,
Accepted 22nd July 2014
DOI: 10.1039/c4cc04719b
www.rsc.org/chemcomm
We describe a general and simple method to identify catalytically by alkaline hydrolysis. This is an asset for the analysis of
and structurally important nucleotides in functional RNAs. Our functional DNAs, but, for obvious reasons, the same strategy
approach is based on statistical replacement of each nucleoside cannot be applied to RNA.
with a non-nucleosidic spacer (C3 linker, D), followed by separation
In the present work we established an analogously versatile
of active library variants and readout of interference effects by combinatorial approach for RNA that reveals functionally
analysis of enzymatic primer extension reactions.
important nucleotides in a simple set of experiments that do
not require sophisticated instrumentation. We used the three-
Many examples of natural and artificial functional RNA motifs carbon linker D as non-nucleosidic spacer unit to statistically
have been identified, including ribozymes, aptamers, riboswitches, replace (‘‘delete’’) standard ribonucleotides within functional
small interfering RNAs, or protein-binding RNAs, that demonstrate nucleic acids (Fig. 1). Following separation of the functionally
the diversity of RNA functions beyond the transmission of genetic active and inactive library variants, the spacer substitutions were
1
information. Elucidating the structure and sequence require- decoded by primer extension reactions. Analysis of the primer
ments of non-coding nucleic acids is crucial for understanding extension pattern allowed identification of the essential and
their mechanisms of action. Furthermore, such knowledge facili- non-essential nucleotides. We demonstrated our new approach
tates design and engineering of RNA for practical applications in
biomolecular chemistry and synthetic biology.
Numerous biochemical and biophysical methods are used to
analyse nucleic acid architectures at various levels of resolu-
2
tion. Recent additions to the traditional repertoire are geared
towards high-throughput analyses on massively parallel arrays,
or combine mutations and footprinting studies with computa-
3
,4
tional methods.
Our laboratory has recently demonstrated combinatorial
5
6
7
analysis methods (CoMA, dNAIM, NDS ) that allowed the
simultaneous assessment of all possible single point mutants
of deoxyribozymes, as well as the identification of catalytically
important nucleotides and their functional groups. However,
due to the method design, the reported approaches are only
applicable for studying functional DNAs. They are based on the
combinatorial solid-phase synthesis of DNA oligonucleotide
libraries, in which the mutations are marked with a chemical
Fig. 1 Concept of enzymatic combinatorial nucleoside deletion scanning
mutagenesis: (a) comparison of a standard nucleoside with non-nucleosidic
0
tag: the hydroxyl group resembling the 2 -OH group of ribo-
nucleotides. Mutations tagged in this way can be easily decoded spacer unit D (C3 linker) and structure of phosphoramidite 1 used for
incorporation of D; (b) schematic representation of a combinatorial library
Research Group Nucleic Acid Chemistry, Max Planck Institute for Biophysical
Chemistry, Am Fassberg 11, 37077 G o¨ ttingen, Germany.
E-mail: claudia.hoebartner@mpibpc.mpg.de
in which parent nucleosides are statistically replaced with D, and isolation of
active fraction; (c) substituted positions are revealed by primer extension
because polymerases cannot extend beyond D; (d) analysis of the inter-
ference pattern by denaturing PAGE. Missing bands in the active fraction
identify the essential nucleotides.
†
Electronic supplementary information (ESI) available: Experimental details and
Fig. S1–S11. See DOI: 10.1039/c4cc04719b
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 10937--10940 | 10937

View Article Online
Communication
ChemComm
with two well-known examples of functional RNAs, including a Those nucleotides might be directly involved in catalysis or
ribozyme and a protein-binding RNA. Our method was able to substrate recognition or play a role for correct folding and
accurately identify the catalytically and structurally important areas. structure formation. Nucleotides that exhibited interference
For the enzymatic combinatorial nucleoside deletion scanning effects below 1.5 were considered non-essential and are good
mutagenesis (enzymatic NDS), we prepared RNA libraries in which candidates for mutation and/or deletion without substantial
each nucleoside in the region of interest was statistically replaced impact on the nucleic acid function.
by the non-nucleosidic spacer unit in such a way that there was
As a feasibility test for interference analysis of primer extension
on average only one substitution per strand (Fig. 1b). Such patterns, we used a combinatorial deletion library of a deoxy-
libraries were prepared by solid-phase synthesis using mixtures ribozyme from our previous study that contained an analogous
7
of standard RNA phosphoramidites with the non-nucleosidic OH-modified C3 spacer. We performed the separation of the
phosphoramidite 1. For example, in a 40 nt long region of active and inactive library variants as previously described, and
interest, the desired one substitution per molecule can be conducted the primer extension experiments according to our
5
achieved by approximately 5% spacer incorporation. To achieve optimized methods (see ESI†). The results from the primer
this desired level of D substitution, we first examined the coupling extension analysis were superimposable with the alkaline hydro-
7
efficiency of 1 in competition with RNA phosphoramidites (rN). lysis results, confirming that the enzymatic decoding by the DNA
We synthesized a series of pentamer oligonucleotides with single polymerase was able to provide a reliable readout of the inter-
C3-spacer replacement using phosphoramidite mixtures contain- ference pattern.
ing 5–25% of 1. The amount of D incorporation was subsequently
As a first example to demonstrate enzymatic combinatorial
analysed by anion-exchange HPLC (see ESI†). Based on these nucleoside deletion scanning mutagenesis on RNA, we chose to
results we used phosphoramidite mixtures for synthesis of RNA study the minimized hammerhead ribozyme (HHR) motif
libraries containing 1 and rN in a ratio of 5 : 95.
(Fig. 2a). This ribozyme catalyses a transesterification reaction,
Enzymatic NDS mutagenesis was designed to take advantage resulting in self-cleavage of the RNA strand. The HHR structure
9
of the fact that polymerase enzymes cannot extend beyond the and consensus sequence are well characterized, providing a basis
8
synthetic abasic sites introduced by the C3 spacer (D) (Fig. 1c).
for comparison with our deletion mutagenesis results. The first
The presence of D in the template strand causes RNA-dependent step comprised the solid-phase synthesis of the combinatorial
DNA polymerases to pause and abort the extension reaction. RNA library, containing the statistical C3 spacer substitutions
Therefore, after separation of the active and inactive library
variants by appropriate means (such as gel electrophoresis or
affinity chromatography), we could use the primer extension to
decode the positions where the substitutions with D were toler-
ated. To optimize the primer elongation conditions, we first tested
different reverse transcriptases on RNA templates that contained
D substitutions at defined positions (see ESI†). We found that the
reaction time is a critical parameter for analysis; short primer
extension times of 2–10 min gave reliable and reproducible results.
Both tested reverse transcriptases (M-MuLV and SSIII RT) termi-
nated the primer extension by incorporation of the last nucleotide
at the position preceding the acyclic modification. We also demon-
strated that C3 spacer substitutions can be detected in a DNA
template by DNA-dependent DNA polymerases (Klenow frag-
ment and One Taq polymerase) (see ESI†). In contrast to reverse
transcriptases, Klenow fragment DNA polymerase incorporated
the last nucleotide directly opposite to the C3 linker (non-
templated incorporation). One Taq DNA polymerase produced
two abort bands (only partial addition of a non-templated
Fig. 2 (a) Secondary structure of the minimized hammerhead ribozyme
nucleotide opposite D). This information is important for the motif. Nucleotides depicted in red proved to be essential for efficient
catalysis. (b) Denaturing sequencing gel separation of the primer extension
correct assignment of the abort bands for analysis. In further
experiments both M-MuLV and SSIII RT were used to analyse
unseparated RNA libraries as well as the active RNA fractions
products generated by M-MuLV from the unseparated library (lane 1) and
catalytically active fractions after 2 h (lane 2) and 10 min (lane 3) cleavage
reactions. Primer extension for each sample was performed for 2 min and
(inactive fractions were occasionally included in the primer
1
0 min (label on top of gel). Sequencing ladders (G, A) were obtained by
extension assay, but not used in the analysis).
using corresponding ddNTPs (Sanger sequencing). (c) Bar graph showing
the interference values for both active fractions (from 10 min and 2 h
cleavage reactions) analysed by 2 min primer extension. Nucleotides
showing interference effects higher than 1.5 are critical for the ribozyme
function. Longer reaction times for DNA cleavage can compensate the
negative effect of the mutation/deletion (compare interference values of
Interference values were calculated as the ratio of band
intensities resulting from the primer extension pattern in the
active fraction and the unseparated library. The nucleosides
for which the interference effect exceeded the value of 1.5 were
considered to be functionally important for a given RNA. the 10 min and 2 h separation reactions, blue and grey data).
10938 | Chem. Commun., 2014, 50, 10937--10940
This journal is ©The Royal Society of Chemistry 2014

View Article Online
ChemComm
Communication
within the HHR motif at nt positions 1–31. The substrate-
containing region from position 32–53 remained unmodified.
The HHR-catalysed cleavage reaction was performed in the
presence of 10 mM MgCl2 (Tris HCl pH 8.0) for various
incubation times (10 min to 2 h). The active (i.e. cleaved) and
inactive (un-cleaved) fractions were separated by denaturing gel
electrophoresis (see ESI†). Subsequently, the active fractions
were isolated and subjected to primer extension in comparison
to the unselected library. Primer elongation was performed by
M-MuLV and SSIII RT for 2–30 min (Fig. 2b and ESI†). In
agreement with the results from the reaction optimization,
longer incubation times for primer extension produced less
significant differences in band intensities (see ESI† for the full
analysis). Results from 2 min primer elongation reactions were
used to calculate the interference effects (Fig. 2).
Analysis of the data revealed that nucleotides 7 (C), 9–10 (GA),
1
(
2–13 (GA) and 28–31 (GAAA) exhibited high interference effects
Fig. 2c), indicating that they are important for catalysis. This
finding is consistent with results from previous studies on
1
0
hammerhead ribozymes. In addition, strong interference Fig. 3 (a) Hairpin II of the U1 snRNA extended by a primer binding
sequence. Seven nucleotides in the loop region (AUUGCAC) are recognised
effects were found at positions 5 and 6 (AG) which are located in
by the U1A protein. Two C–G stem closing base-pairs are essential for the
a stem region in close proximity to the cleavage site. Most likely the
correct structure formation. (b) Summary of the interference results for
substitution with the C3 spacer at these positions caused a desta-
2
different separation conditions (1 and 0.2 equivalents of protein over
bilizing effect on the catalytic centre and/or hindered the substrate RNA). (c) and (d) Primer extension gels for the separated fractions
recognition. Interestingly, longer reaction times for the HHR- and unseparated library. Extension products were produced by M-MuLV.
Lanes A, C, G, U – sequencing ladders; (c) separation was done with 1 eq.
catalysed cleavage compensated the negative effect of the D sub-
stitutions (compare the interference values of the 10 min and 2 h
RNA-catalysed cleavage reaction, Fig. 2c). This result indicated that
of protein. Lane 1, 2 – active fraction from 2 independent experiments
(act.); lane 3, 4, 6, 7 – unseparated library (pool); lane 5 – inactive fraction (I);
lane 8 – primer extension on the unmodified RNA ((ꢀ) control). (d) Separation
the presence of D did not completely abolish the activity of the of active fraction (act.) was done with 0.5 eq. (lane 2), 0.2 eq. (lane 3) and
ribozyme but severely affected the reaction rates. As expected, single- 0.1 eq. (lane 4) of the protein; lane 1 – unseparated library (pool).
nucleotide substitutions within the stem-loop (nt 14–27) were well
tolerated (even at short reaction times), confirming that the method were observed at all seven positions. However, when the selec-
is suitable to distinguish critical from non-critical nucleotides.
tion was done in the presence of higher protein concentration,
The enzymatic combinatorial nucleoside deletion approach the nucleotides 6–8 (AUU) displayed lower interference effects
can also be used to study RNA–protein interactions. Here, compared to the other critical positions. This shows that the
we analysed the hairpin II of U1 snRNA (Fig. 3a) for which negative effect caused by the substitution with D can be partially
7
nucleotides in the loop (AUUGCAC) are known to be essential for compensated, and implies differential effects of individual nucleo-
the recognition by the U1A protein and 2 loop closing base-pairs tides on the overall affinity. Similar effects of different RNA/protein
1
1
(
G–C) are important for structural reasons. The synthesised RNA ratios were reported when separation of RNA mutants was done with
3
library contained 17 mutagenized nucleotides (positions 1–17) and immobilized U1A protein. In contrast, nucleotides 4 and 5 (CC),
4 non-substituted nucleotides (positions 18–31) as primer bind- which form the loop-closing base-pairs, exhibited strong interference
1
ing site for analysis. The separation of the active and inactive effects under all separation conditions, as the intact base pairs are
library variants was achieved by incubation of the RNA library with needed for formation of the high affinity binding-competent hairpin
the U1A protein and subsequent separation of the bound and structure.
unbound fractions by the gel mobility shift assays on non-
In conclusion, we demonstrated that our new strategy
denaturing polyacrylamide gels (see ESI†). The analysis of the enables simultaneous identification of all essential areas of a
primer extension results revealed 9 nucleotides (AUUGCAC in the given nucleic acid motif in an experimentally simple procedure.
loop and CC in the stem) that exhibited strong interference effects All needed reagents, including phosphoramidite 1, are com-
(higher than 1.5), indicating that those nucleotides are required for mercially available. In addition, the separation of functionally
the interaction of the U1A protein with the U1 snRNA. This result active and inactive library variants upon interaction with their
is in agreement with the above mentioned previous studies on the cognate substrates is performed in solution (i.e. no immobilisa-
11
U1 snRNA motif.
tion on a solid support needed, thus avoiding potential surface
In addition, our analysis showed that mutations at some of effects). Furthermore, our study showed that the magnitude of
the critical loop nucleotides had less severe effects than others. the interference effects at some positions strongly depends on
When low protein concentration was used for the separation of the experimental parameters used to separate the active from
the active and inactive fractions, strong interference effects inactive library variants. This opens up the possibility to gain
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 10937--10940 | 10939

View Article Online
Communication
ChemComm
deeper insights into mutation tolerance of individual nucleo-
tides by performing the analyses under various conditions. For
Chem., 2012, 321, 159–198; T. Carlomagno, J. Magn. Reson., 2014,
241, 126–136; S. P. Ryder and S. A. Strobel, Methods, 1999, 18, 38–50.
3
4
D. Ray, H. Kazan, E. T. Chan, L. Pena Castillo, S. Chaudhry,
S. Talukder, B. J. Blencowe, Q. Morris and T. R. Hughes, Nat.
Biotechnol., 2009, 27, 667–670.
J. D. Buenrostro, C. L. Araya, L. M. Chircus, C. J. Layton, H. Y.
Chang, M. P. Snyder and W. J. Greenleaf, Nat. Biotechnol., 2014, 32,
1
2
example, variation of metal ions and their concentrations can
identify cofactor binding sites, or variation of reaction times for
the separation step can yield specific kinetic information.
The three carbon spacer unit used in our study has been
successfully applied in the past as nucleoside replacement to
investigate sequence and structure requirements of ribo-
562–568; J. M. Tome, A. Ozer, J. M. Pagano, D. Gheba, G. P. Schroth
and J. T. Lis, Nat. Methods, 2014, 11, 683–688; W. Kladwang,
C. C. VanLang, P. Cordero and R. Das, Nat. Chem., 2011, 3, 954–962.
F. Wachowius, F. Javadi-Zarnaghi and C. H ¨o bartner, Angew. Chem.,
Int. Ed., 2010, 49, 8504–8508.
5
1
3
14
15
16
zymes, ribonucleoproteins, siRNAs, hairpins and deoxy-
1
7
ribozymes. Here, we report for the first time its application in
a combinatorial way. This approach is more efficient and less
time-consuming than systematic replacement of individual
nucleotides and separate analysis of each individual mutant.
The comprehensive datasets elucidate all critical nucleotides
within RNA motifs and pinpoint non-essential positions. This
6 F. Wachowius and C. H ¨o bartner, J. Am. Chem. Soc., 2011, 133,
4888–14891.
B. Samanta and C. H ¨o bartner, Angew. Chem., Int. Ed., 2013, 52,
995–2999.
1
7
2
8 M. Takeshita, C. N. Chang, F. Johnson, S. Will and A. P. Grollman,
J. Biol. Chem., 1987, 262, 10171–10179; P. A. Kupfer, C. Crey-Desbiolles
and C. J. Leumann, Nucleic Acids Res., 2007, 35, 6846–6853.
9
C. Hammann and D. M. Lilley, ChemBioChem, 2002, 3, 690–700.
information contributes to deeper understanding of underlying 10 D. E. Ruffner, G. D. Stormo and O. C. Uhlenbeck, Biochemistry, 1990,
2
9, 10695–10702; C. C. Sheldon and R. H. Symons, Nucleic Acids Res.,
mechanisms and might further promote the design of func-
tional derivatives.
1989, 17, 5679–5685; A. Peracchi, L. Beigelman, N. Usman and
D. Herschlag, Proc. Natl. Acad. Sci. U. S. A., 1996, 93, 11522–11527.
This work was supported by the DFG (International Research 11 D. Scherly, W. Boelens, W. J. van Venrooij, N. A. Dathan, J. Hamm and
I. W. Mattaj, EMBO J., 1989, 8, 4163–4170; D. E. Tsai, D. S. Harper and
J. D. Keene, Nucleic Acids Res., 1991, 19, 4931–4936; C. Oubridge, N. Ito,
P. R. Evans, C. H. Teo and K. Nagai, Nature, 1994, 372, 432–438; M. J. Law,
A. J. Rice, P. Lin and I. A. Laird-Offringa, RNA, 2006, 12, 1168–1178.
2 F. Javadi-Zarnaghi and C. H ¨o bartner, J. Am. Chem. Soc., 2013, 135,
Training Grant IRTG 1422, Metal Sites in Biomolecules) and the
Max Planck Society. We thank Jana Schmitzova and Vlad Pena
Macromolecular Crystallography group at MPIbpc) for help
(
1
1
1
with the preparation of U1A protein and lab members of the
Nucleic Acid Chemistry group for helpful discussions.
12839–12848.
3 S. Schmidt, L. Beigelman, A. Karpeisky, N. Usman, U. S. Sorensen
and M. J. Gait, Nucleic Acids Res., 1996, 24, 573–581.
4 D. J. Williams and K. B. Hall, J. Mol. Biol., 1996, 257, 265–275.
Notes and references
15 T. C. Efthymiou, B. Peel, V. Huynh and J. P. Desaulniers, Bioorg.
Med. Chem. Lett., 2012, 22, 5590–5594.
16 W. Pils and R. Micura, Nucleic Acids Res., 2002, 28, 1859–1863.
1
2
T. R. Cech, Cold Spring Harbor Perspect. Biol., 2012, 4, a006742.
K. M. Weeks, Curr. Opin. Struct. Biol., 2010, 20, 295–304; I. Krstic, 17 B. Wang, L. Cao, W. Chiuman, Y. Li and Z. Xi, Biochemistry, 2010, 49,
B. Endeward, D. Margraf, A. Marko and T. F. Prisner, Top. Curr. 7553–7562.
10940 | Chem. Commun., 2014, 50, 10937--10940
This journal is ©The Royal Society of Chemistry 2014