*Article* **Can Vitamin B12 Assist the Internalization of Antisense LNA Oligonucleotides into Bacteria?**

**Sara Pereira <sup>1</sup> , Ruwei Yao <sup>2</sup> , Mariana Gomes <sup>1</sup> , Per Trolle Jørgensen <sup>2</sup> , Jesper Wengel <sup>2</sup> , Nuno Filipe Azevedo <sup>1</sup> and Rita Sobral Santos 1,\***


**Abstract:** The emergence of bacterial resistance to traditional small-molecule antibiotics is fueling the search for innovative strategies to treat infections. Inhibiting the expression of essential bacterial genes using antisense oligonucleotides (ASOs), particularly composed of nucleic acid mimics (NAMs), has emerged as a promising strategy. However, their efficiency depends on their association with vectors that can translocate the bacterial envelope. Vitamin B<sup>12</sup> is among the largest molecules known to be taken up by bacteria and has very recently started to gain interest as a trojan-horse vector. Gapmers and steric blockers were evaluated as ASOs against *Escherichia coli* (*E. coli*). Both ASOs were successfully conjugated to B<sup>12</sup> by copper-free azide-alkyne click-chemistry. The biological effect of the two conjugates was evaluated together with their intracellular localization in *E. coli*. Although not only B<sup>12</sup> but also both B12-ASO conjugates interacted strongly with *E. coli*, they were mostly colocalized with the outer membrane. Only 6–9% were detected in the cytosol, which showed to be insufficient for bacterial growth inhibition. These results suggest that the internalization of B12-ASO conjugates is strongly affected by the low uptake rate of the B<sup>12</sup> in *E. coli* and that further studies are needed before considering this strategy against biofilms in vivo.

**Keywords:** antibacterial drug; vitamin B12; antisense oligonucleotides; nucleic acid mimics; LNA; 2 0OMe

## **1. Introduction**

The emergence of bacterial resistance to traditional antibiotics is considered a major threat in modern medicine [1,2]. Inevitably, innovative research focused on different antibacterial strategies is needed. Antisense oligonucleotides (ASOs) designed to inhibit bacterial gene expression have been gaining increased interest in recent years [3]. ASOs are especially interesting because even if bacteria develop a mutation that renders them resistant (one of the most common forms of resistance), the ASO can be easily redesigned to become an effective antibacterial drug again [4]. ASOs composed of nucleic acid mimics (NAMs), and in particular, locked nucleic acids (LNAs), possess improved target specificity, binding affinity, and resistance to exo- and endonucleases, compared to unmodified RNA or DNA [5,6], and have been successfully tested for clinical applications [7,8]. ASOs can be divided into two major categories, according to their mechanism of action: RNase H competent (or gapmers) and steric blockers (Figure 1). Gapmers are composed of DNA monomers that are typically flanked by LNA or other RNA-mimicking monomers. Upon hybridization of the gapmer to the target mRNA, the RNase H enzyme recognizes the DNA-mRNA heteroduplex and cleaves the target mRNA, leading to its degradation. Alternatively, the hybridization of steric blockers to the target mRNA simply physically

**Citation:** Pereira, S.; Yao, R.; Gomes, M.; Jørgensen, P.T.; Wengel, J.; Azevedo, N.F.; Sobral Santos, R. Can Vitamin B12 Assist the Internalization of Antisense LNA Oligonucleotides into Bacteria? *Antibiotics* **2021**, *10*, 379. https://doi.org/10.3390/ antibiotics10040379

Academic Editors: Nicholas Dixon and Marc Maresca

Received: 17 February 2021 Accepted: 1 April 2021 Published: 3 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

*Antibiotics* **2021**, *10*, x FOR PEER REVIEW 2 of 12

blocks the access of the RNA polymerase to the target mRNA, thus directly inhibiting its translation [9–11]. There is only one study reporting the use of gapmers to target bacteria [11]. radation. Alternatively, the hybridization of steric blockers to the target mRNA simply physically blocks the access of the RNA polymerase to the target mRNA, thus directly inhibiting its translation [9–11]. There is only one study reporting the use of gapmers to target bacteria [11].

ognizes the DNA-mRNA heteroduplex and cleaves the target mRNA, leading to its deg-

**Figure 1.** Different mechanisms that play a role in the modulation of the RNA function in bacteria. (**a**) Upon hybridization of a gapmer (red), the RNase H is recruited, and the target is degraded. (**b**) Steric hindrance of the ribosome caused by the hybridization between a steric blocker (red) and the complementary mRNA sequence. Figure created using BioRender. **Figure 1.** Different mechanisms that play a role in the modulation of the RNA function in bacteria. (**a**) Upon hybridization of a gapmer (red), the RNase H is recruited, and the target is degraded. (**b**) Steric hindrance of the ribosome caused by the hybridization between a steric blocker (red) and the complementary mRNA sequence. Figure created using BioRender.

> Nonetheless, the use of ASOs is limited by their inability to efficiently penetrate the complex envelope of bacteria. To overcome this burden, delivery vectors mostly focused on cell-penetrating peptides (CPPs) have been investigated. However, CPPs have been almost exclusively conjugated to neutral NAMs, such as peptide nucleic acids (PNAs) or phosphorodiamidate morpholino oligos (PMOs), which might present cytotoxicity and solubility issues [12–14]. Due to chemical conjugation difficulties, the vectorization of anionic ASOs with CPPs has been hampered. In a different approach, vitamin B12 (B12 or cobalamin), one of the largest molecules known to be taken up by bacteria, can be considered as a trojan-horse vector for neutral as well as anionic ASOs. The uptake system of B12 has been mostly studied in *E. coli*. [15]. *E. coli* uptakes B12 through the outer-membrane βbarrel protein BtuB in a TonB-dependent manner [16]. In the periplasm, BtuF binds to and delivers B12 to the ABC-type transporter BtuCD in the inner membrane, which, in turn, transports B12 into the cytoplasm [17,18]. Nonetheless, the use of ASOs is limited by their inability to efficiently penetrate the complex envelope of bacteria. To overcome this burden, delivery vectors mostly focused on cell-penetrating peptides (CPPs) have been investigated. However, CPPs have been almost exclusively conjugated to neutral NAMs, such as peptide nucleic acids (PNAs) or phosphorodiamidate morpholino oligos (PMOs), which might present cytotoxicity and solubility issues [12–14]. Due to chemical conjugation difficulties, the vectorization of anionic ASOs with CPPs has been hampered. In a different approach, vitamin B<sup>12</sup> (B<sup>12</sup> or cobalamin), one of the largest molecules known to be taken up by bacteria, can be considered as a trojan-horse vector for neutral as well as anionic ASOs. The uptake system of B<sup>12</sup> has been mostly studied in *E. coli*. [15]. *E. coli* uptakes B<sup>12</sup> through the outer-membrane β-barrel protein BtuB in a TonB-dependent manner [16]. In the periplasm, BtuF binds to and delivers B<sup>12</sup> to the ABC-type transporter BtuCD in the inner membrane, which, in turn, transports B<sup>12</sup> into the cytoplasm [17,18].

> Several functional groups are available for the modification of B12 to allow conjugation with ASOs, but not all modification sites are suitable to sustain their recognition and uptake [19]. Chromiński et al. described for the first time the synthesis of a clickable B12 derivate, possessing an azide functionality at the 5′ end [20]. This modification has already been tested for the copper-dependent conjugation of B12 with oligonucleotides, either composed of PNA or 2′OMe, mainly to inhibit genes that code for reporter proteins such as the red fluorescent protein (RFP) [6,20,21]. To our knowledge, there is only one study where a B12 conjugate was studied to decrease bacterial growth by inhibition of the essential gene *acpP* in *E. coli* using a PNA ASO [22]. This B12-PNA conjugate was only proved to inhibit *E. coli* growth in a very specific medium named Scarlett and Turner [22]. Under these conditions, even in the absence of B12 conjugates, bacteria only started growing after 48 h, while in other common minimal media, the exponential growth starts already after Several functional groups are available for the modification of B<sup>12</sup> to allow conjugation with ASOs, but not all modification sites are suitable to sustain their recognition and uptake [19]. Chromi´nski et al. described for the first time the synthesis of a clickable B<sup>12</sup> derivate, possessing an azide functionality at the 50 end [20]. This modification has already been tested for the copper-dependent conjugation of B<sup>12</sup> with oligonucleotides, either composed of PNA or 20OMe, mainly to inhibit genes that code for reporter proteins such as the red fluorescent protein (RFP) [6,20,21]. To our knowledge, there is only one study where a B<sup>12</sup> conjugate was studied to decrease bacterial growth by inhibition of the essential gene *acpP* in *E. coli* using a PNA ASO [22]. This B12-PNA conjugate was only proved to inhibit *E. coli* growth in a very specific medium named Scarlett and Turner [22]. Under these conditions, even in the absence of B<sup>12</sup> conjugates, bacteria only started growing after 48 h, while in other common minimal media, the exponential growth starts already after 5 h [23].

> 5 h [23]. Occasionally, infections are associated with the formation of biofilms, adding an extra barrier for the use of antibacterial compounds [24]. ASOs were already shown to prevent biofilm formation and reduce mature biofilms, using either peptide nucleic acids (PNAs) or

phosphorodiamidate morpholino oligomers (PMOs) as NAMs, conjugated to CPPs [25,26] but, to the best of our knowledge, never conjugated to B12. However, and because the cytosol is the ultimate target for these conjugates, it is important to first investigate their single-cell internalization. vent biofilm formation and reduce mature biofilms, using either peptide nucleic acids (PNAs) or phosphorodiamidate morpholino oligomers (PMOs) as NAMs, conjugated to CPPs [25,26] but, to the best of our knowledge, never conjugated to B12. However, and because the cytosol is the ultimate target for these conjugates, it is important to first investigate their single-cell internalization.

Occasionally, infections are associated with the formation of biofilms, adding an extra barrier for the use of antibacterial compounds [24]. ASOs were already shown to pre-

*Antibiotics* **2021**, *10*, x FOR PEER REVIEW 3 of 12

In this study, we have investigated, for the first time, the internalization and inhibition efficiency of two different copper-free clicked conjugates, composed of vitamin B<sup>12</sup> linked to LNA-based ASOs: a gapmer and a steric blocker. Both ASOs were designed to target the *acpP* gene in *E. coli*, which codes for an essential protein involved in fatty acid biosynthesis [27]. In this study, we have investigated, for the first time, the internalization and inhibition efficiency of two different copper-free clicked conjugates, composed of vitamin B12 linked to LNA-based ASOs: a gapmer and a steric blocker. Both ASOs were designed to target the *acpP* gene in *E. coli*, which codes for an essential protein involved in fatty acid biosynthesis [27].

## **2. Results and Discussion**

#### *2.1. Conjugation of ASOs with Vitamin B<sup>12</sup>* **2. Results and Discussion**

In this study, two different kinds of LNA antisense oligonucleotides were designed and synthesized to target the *acpP* gene in *E. coli* (Figure 2a): an LNA/DNA gapmer (ASOgapmer) and an LNA/20OMe steric blocker (ASOsteric) [28,29]. While steric blockers have been widely tested in bacteria [4,27,30–32], gapmers have been mostly studied in mammalian cells [11]. As such, we intended to compare the potency of the different antisense mechanisms in *E. coli.* As the bacterial envelope poses a stringent barrier to the internalization of oligonucleotides, appropriate vectors must be applied for their transport into the bacterial cytosol. This is the first study documenting the use of B<sup>12</sup> as a vector for LNA oligonucleotides. For the association of B<sup>12</sup> to ASOsteric and ASOgamper, a copper-free ring-strain-promoted azide-alkyne coupling reaction was used (Figure 2b) [33]. This method proved to be efficient and resulted in satisfactory yields (Table S1). The increase in HPLC retention time for the B12-ASOgamper and B12-ASOsteric compared with the ASOgapmer and ASOsteric points toward efficient conjugation, confirmed by the obtained molecular masses, which are similar to the calculated theoretical values (Figure S1). The conjugation yields were determined as 70% and 97%, respectively, for B12-ASOgamper and B12-ASOsteric. *2.1. Conjugation of ASOs with Vitamin B12* In this study, two different kinds of LNA antisense oligonucleotides were designed and synthesized to target the *acpP* gene in *E. coli* (Figure 2a): an LNA/DNA gapmer (ASOgapmer) and an LNA/2′OMe steric blocker (ASOsteric) [28,29]. While steric blockers have been widely tested in bacteria [4,27,30–32], gapmers have been mostly studied in mammalian cells [11]. As such, we intended to compare the potency of the different antisense mechanisms in *E. coli.* As the bacterial envelope poses a stringent barrier to the internalization of oligonucleotides, appropriate vectors must be applied for their transport into the bacterial cytosol. This is the first study documenting the use of B12 as a vector for LNA oligonucleotides. For the association of B12 to ASOsteric and ASOgamper, a copper-free ring-strainpromoted azide-alkyne coupling reaction was used (Figure 2b) [33]. This method proved to be efficient and resulted in satisfactory yields (Table S1). The increase in HPLC retention time for the B12-ASOgamper and B12-ASOsteric compared with the ASOgapmer and ASOsteric points toward efficient conjugation, confirmed by the obtained molecular masses, which are similar to the calculated theoretical values (Figure S1). The conjugation yields were determined as 70% and 97%, respectively, for B12-ASOgamper and B12-ASOsteric.

**Figure 2.** B12 and antisense oligonucleotides (ASOs) conjugation: sequences and structures. (**a**) Sequence of the ASOgapmer and the ASOsteric (LNA nucleotide monomers are represented with upper case letters preceded by l, 2′OMe monomers are **Figure 2.** B<sup>12</sup> and antisense oligonucleotides (ASOs) conjugation: sequences and structures. (**a**) Sequence of the ASOgapmer and the ASOsteric (LNA nucleotide monomers are represented with upper case letters preceded by l, 20OMe monomers are represented with upper case letters preceded by m, and DNA monomers are represented by lower case letters) and structure of 50 -end azide-modified B<sup>12</sup> used in this study. The arrow points to the region of conjugation. (**b**) Schematic illustration of the synthesis of the B12-ASO conjugates through copper-free azide-alkyne chemistry.

#### *2.2. Bacterial Susceptibility Tests* targeting *acpP*) against *E. coli* in a particular medium where the control bacteria only starts growing after 48 h [22]. Nonetheless, the activity of this ASO sequence is already well

tion of the synthesis of the B12-ASO conjugates through copper-free azide-alkyne chemistry.

*2.2. Bacterial Susceptibility Tests* 

*Antibiotics* **2021**, *10*, x FOR PEER REVIEW 4 of 12

represented with upper case letters preceded by m, and DNA monomers are represented by lower case letters) and structure of 5′-end azide-modified B12 used in this study. The arrow points to the region of conjugation. (**b**) Schematic illustra-

> Both ASOs were designed to recognize the *acpP* essential gene in *E. coli* and thus inhibit its expression. This should result in decreased *E. coli* viability, as long as the ASOs conjugated to B<sup>12</sup> can efficiently penetrate the bacterial envelope. We investigated the ability of both conjugates (B12-ASOgapmer and B12-ASOsteric) to inhibit the growth of *E. coli* in Davis minimal medium at a concentration of 30 µM. This concentration was selected based on the good inhibition efficiency of a cell-penetrating peptide (CPP) conjugated with PNA, targeting the same gene (Figure 3, orange line). Our results show no inhibition using either conjugate composed of B<sup>12</sup> at the same concentration (Figure 3). established, as growth inhibition of *E. coli* K12 has been repeatedly reported using CPP-PNA [27,34,35], which was also confirmed herein. It is clear from the growth curves that the internalization occurs using the CPP as a vector for ASOs, as opposed to the B12 vector. The lack of inhibitory effect of the *E. coli* K12 growth, observed with the conjugates synthesized in the present work, raises the question if the conjugates were efficiently internalized in the bacterial cells. In order to answer this question, location studies were performed next.

Both ASOs were designed to recognize the *acpP* essential gene in *E. coli* and thus inhibit its expression. This should result in decreased *E. coli* viability, as long as the ASOs conjugated to B12 can efficiently penetrate the bacterial envelope. We investigated the ability of both conjugates (B12-ASOgapmer and B12-ASOsteric) to inhibit the growth of *E. coli* in Davis minimal medium at a concentration of 30 μM. This concentration was selected based on the good inhibition efficiency of a cell-penetrating peptide (CPP) conjugated with PNA, targeting the same gene (Figure 3, orange line). Our results show no inhibition

In previous studies, PNA and 2′OMe steric blockers conjugated to B12 were able to decrease by 1-fold the expression of red fluorescence protein (RFP) in *E. coli* in Davis minimal medium [6,21]. However, to our knowledge, there is no other study including a regular growth control where B12-ASOs were investigated to kill bacteria, targeting an essential gene rather than a report protein. The only existing study uses a B12-PNA (ASOsteric

using either conjugate composed of B12 at the same concentration (Figure 3).

**Figure 3.** Growth of *E. coli* K12 in Davis minimal medium supplemented with B12-ASOgapmer, B12- ASOsteric, and B12 (at a concentration of 30 μM). CB represents the bacterial growth control in medium without any supplementation. Growth inhibition of *E. coli* K12 using a cell-penetrating peptide conjugated with an ASO composed of peptide nucleic acids (PNAs) (cell-penetrating peptides (CPP)- PNA) at a concentration of 30 μM is also shown. Results from three independent experiments (using duplicates in each) are presented as mean values and respective standard deviations. Statistical differences are indicated when appropriate in \* (*p* ≤ 0.0001, \*\*\*\*). **Figure 3.** Growth of *E. coli* K12 in Davis minimal medium supplemented with B12-ASOgapmer, B12-ASOsteric, and B<sup>12</sup> (at a concentration of 30 µM). CB represents the bacterial growth control in medium without any supplementation. Growth inhibition of *E. coli* K12 using a cell-penetrating peptide conjugated with an ASO composed of peptide nucleic acids (PNAs) (cell-penetrating peptides (CPP)-PNA) at a concentration of 30 µM is also shown. Results from three independent experiments (using duplicates in each) are presented as mean values and respective standard deviations. Statistical differences are indicated when appropriate in \* (*p* ≤ 0.0001, \*\*\*\*).

*2.3. Evaluation of the Internalization of B12-ASOs*  In previous studies, PNA and 20OMe steric blockers conjugated to B<sup>12</sup> were able to decrease by 1-fold the expression of red fluorescence protein (RFP) in *E. coli* in Davis minimal medium [6,21]. However, to our knowledge, there is no other study including a regular growth control where B12-ASOs were investigated to kill bacteria, targeting an essential gene rather than a report protein. The only existing study uses a B12-PNA (ASOsteric targeting *acpP*) against *E. coli* in a particular medium where the control bacteria only starts growing after 48 h [22]. Nonetheless, the activity of this ASO sequence is already well established, as growth inhibition of *E. coli* K12 has been repeatedly reported using CPP-PNA [27,34,35], which was also confirmed herein. It is clear from the growth curves that the internalization occurs using the CPP as a vector for ASOs, as opposed to the B<sup>12</sup> vector.

The lack of inhibitory effect of the *E. coli* K12 growth, observed with the conjugates synthesized in the present work, raises the question if the conjugates were efficiently internalized in the bacterial cells. In order to answer this question, location studies were performed next.
