Two-dimensional combinatorial screening of a bacterial rRNA A-site-like motif library: defining privileged asymmetric internal loops that bind aminoglycosides

Abstract

RNAs have diverse structures that are important for biological function. These structures include bulges and internal loops that can form tertiary contacts or serve as ligand binding sites. The most commonly exploited RNA drug target for small molecule intervention is the bacterial ribosome, more specifically the rRNA aminoacyl-tRNA site (rRNA A-site) which is a major target for the aminoglycoside class of antibiotics. The bacterial A-site is composed of a 1 x 1 nucleotide all-U internal loop and a 2 x 1 nucleotide all-A internal loop separated by a single GC base pair. Therefore, we probed the molecular recognition of a small library of four aminoglycosides for binding a 16384-member bacterial rRNA A-site-like internal loop library using two-dimensional combinatorial screening (2DCS). 2DCS is a microarray-based method that probes RNA and chemical spaces simultaneously. These studies sought to determine if aminoglycosides select their therapeutic target if given a choice of binding all possible internal loops derived from an A-site-like library. Results show that the bacterial rRNA A-site was not selected by any aminoglycoside. Analyses of selected sequences using the RNA Privileged Space Predictor (RNA-PSP) program show that each aminoglycoside preferentially binds different types of internal loops. For three of the aminoglycosides, 6''-azido-kanamycin A, 5-O-(2-azidoethyl)-neamine, and 6''-azido-tobramycin, the selected internal loops bind with approximately 10-fold higher affinity than the bacterial rRNA A-site. The internal loops selected to bind 5''-azido-neomycin B bind with an affinity similar to that of the therapeutic target. Selected internal loops that are unique for each aminoglycoside have dissociation constants ranging from 25 to 270 nM and are specific for the aminoglycoside they was selected to bind compared to the other arrayed aminoglycosides. These studies further establish a database of RNA motifs that are recognized by small molecules that could be used to enable the rational and modular design of small molecules targeting RNA.

Figure 1

Secondary structures of the oligonucleotides used in this study. 1 is the 4×3 nucleotide asymmetric internal loop library, where N represents an equimolar mixture of A, C, G, and U. RNA 2 is an oligonucleotide mimic of the bacterial A-site. Oligonucleotides 3-9 are used to compete off interactions to the regions in 1 that are common to all library members.

Figure 2

Two-Dimensional Combinatorial Screening (2DCS) assay to identify RNA loop-ligand interactions. (A, top) Chemical structures of azido-aminoglycosides used in this study: 10-13 are kanamycin A, tobramycin, neamine, and neomycin B derivatives, respectively. (A, bottom) Immobilization of 10-13 onto alkyne-displaying agarose microarrays for 2DCS or for conjugation of 10-13 to fluorescein (green ball) via a Huisgen 1,3 dipolar cycloaddition reaction (HDCR) to yield 10-FL, 11-FL, 12-FL, and 13-FL. AmG refers to aminoglycoside. (B, top) Image of an aminglycoside-functionalized microarray displaying 10-13 after hybridization with radioactively labeled 1 and competitor oligonucleotides 3-9. Aminoglycoside derivatives were immobilized onto the slide surface using a HDCR at five different concentrations. (B, bottom) Bound RNAs were harvested from the microarray surface by manual excision using the image of the hybridized slide as a template.

Figure 3

The protocol utilized to choose selected RNA internal loops to study binding affinities and selectivities. The output of 2DCS is analyzed via the RNA-PSP program (18) to identify trends that are significant to a ≥95% confidence level. The trends for each ligand were then compared to identify those that are unique. Features that were specific for each aminoglycoside were then used to identify the selected internal loops that have the highest number of unique features. These were expected to be the most selective.

Figure 4

Secondary structures of selected RNAs predicted by Mfold and their corresponding dissociation constants (nanomolar). The nucleotides shown were derived from the boxed region in 1 (Figure 1). Structures in A were selected to bind 10; structures in B were selected to bind 11; structures in C were selected to bind 12; and structures in D were selected to bind 13.