Efficient and specific repair of sickle beta-globin RNA by trans-splicing ribozymes

Abstract

Previously we demonstrated that a group I ribozyme can perform trans-splicing to repair sickle beta-globin transcripts upon transfection of in vitro transcribed ribozyme into mammalian cells. Here, we sought to develop expression cassettes that would yield high levels of active ribozyme after gene transfer. Our initial expression constructs were designed to generate trans-slicing ribozymes identical to those used in our previous RNA transfection studies with ribozymes containing 6-nucleotide long internal guide sequences. The ribozymes expressed from these cassettes, however, were found to be unable to repair sickle beta-globin RNAs. Further experiments revealed that two additional structural elements are important for ribozyme-mediate RNA repair: the P10 interaction formed between the 5' end of the ribozyme and the beginning of the 3' exon and an additional base-pairing interaction formed between an extended guide sequence and the substrate RNA. These optimized expression cassettes yield ribozymes that are able to amend 10%-50% of the sickle beta-globin RNAs in transfected mammalian cells. Finally, a ribozyme with a 5-bp extended guide sequence preferentially reacts with sickle beta-globin RNAs over wild-type beta-globin RNAs, although the wild-type beta-globin transcript forms only a single mismatch with the ribozyme. These results demonstrate that trans-splicing ribozyme expression cassettes can be generated to yield ribozymes that can repair a clinically relevant fraction of sickle beta-globin RNAs in mammalian cells with greatly improved specificity.

FIGURE 1.

Ribozymes derivatives and their activity in mammalian cells. (A) Variants of the Rib61-3′γ ribozyme shown bound to sickle β-globin target RNA (dashed line). The S0 ribozyme has a 6-nucleotide internal guide sequence (IGS) that is complementary to the target RNA; the S1 ribozyme has a 6-nucleotide IGS and a 41-nucleotide extended guide sequence (EGS) that is complementary to the target RNA; the S3 ribozyme has a 6-nucleotide IGS and a 41-nucleotide nonspecific spacer sequence at its 5′ end; and the S5 ribozyme has a 9-nucleotide IGS, a 35-nucleotide EGS that is complementary to the target RNA, and a 6-nucleotide P10 that is complementary to the 3′exon sequence. (B) The S5Rib61-3′γ ribozyme bound to the sickle β-globin target RNA. (C) Detection of trans-splicing products by RT–PCR amplification. HEK293 cells were transfected with the β^S-globin plasmid (pNL97) alone (lane 8) or in conjunction with plasmids encoding the S5, S1, S3, or S0 ribozymes or an inactive version of the S5 ribozyme (Dead). Total RNA was isolated and trans-splicing products detected by RT–PCR amplification as discussed in Materials and Methods. Amplification of trans-splicing products yield a 116-bp product. Molecular mass markers (MW) of 123 and 110 bp are shown.

FIGURE 1.

Ribozymes derivatives and their activity in mammalian cells. (A) Variants of the Rib61-3′γ ribozyme shown bound to sickle β-globin target RNA (dashed line). The S0 ribozyme has a 6-nucleotide internal guide sequence (IGS) that is complementary to the target RNA; the S1 ribozyme has a 6-nucleotide IGS and a 41-nucleotide extended guide sequence (EGS) that is complementary to the target RNA; the S3 ribozyme has a 6-nucleotide IGS and a 41-nucleotide nonspecific spacer sequence at its 5′ end; and the S5 ribozyme has a 9-nucleotide IGS, a 35-nucleotide EGS that is complementary to the target RNA, and a 6-nucleotide P10 that is complementary to the 3′exon sequence. (B) The S5Rib61-3′γ ribozyme bound to the sickle β-globin target RNA. (C) Detection of trans-splicing products by RT–PCR amplification. HEK293 cells were transfected with the β^S-globin plasmid (pNL97) alone (lane 8) or in conjunction with plasmids encoding the S5, S1, S3, or S0 ribozymes or an inactive version of the S5 ribozyme (Dead). Total RNA was isolated and trans-splicing products detected by RT–PCR amplification as discussed in Materials and Methods. Amplification of trans-splicing products yield a 116-bp product. Molecular mass markers (MW) of 123 and 110 bp are shown.

FIGURE 1.

Ribozymes derivatives and their activity in mammalian cells. (A) Variants of the Rib61-3′γ ribozyme shown bound to sickle β-globin target RNA (dashed line). The S0 ribozyme has a 6-nucleotide internal guide sequence (IGS) that is complementary to the target RNA; the S1 ribozyme has a 6-nucleotide IGS and a 41-nucleotide extended guide sequence (EGS) that is complementary to the target RNA; the S3 ribozyme has a 6-nucleotide IGS and a 41-nucleotide nonspecific spacer sequence at its 5′ end; and the S5 ribozyme has a 9-nucleotide IGS, a 35-nucleotide EGS that is complementary to the target RNA, and a 6-nucleotide P10 that is complementary to the 3′exon sequence. (B) The S5Rib61-3′γ ribozyme bound to the sickle β-globin target RNA. (C) Detection of trans-splicing products by RT–PCR amplification. HEK293 cells were transfected with the β^S-globin plasmid (pNL97) alone (lane 8) or in conjunction with plasmids encoding the S5, S1, S3, or S0 ribozymes or an inactive version of the S5 ribozyme (Dead). Total RNA was isolated and trans-splicing products detected by RT–PCR amplification as discussed in Materials and Methods. Amplification of trans-splicing products yield a 116-bp product. Molecular mass markers (MW) of 123 and 110 bp are shown.

FIGURE 2.

Evaluating trans-splicing efficiency of the S5 ribozyme. (A) Scheme for the measurement of trans-splicing efficiency by QC RT–PCR. The sickle β-globin target RNA (dashed line) and trans-spliced product RNA can be reverse transcribed and coamplified by a single set of PCR primers (arrows) to yield products that differ in size by 36 bp. The shaded box denotes the 18-nucleotide long priming sequence that is present in both the β^S-globin target RNA (nucleotides 119–136) and the S5eff ribozyme’s 3′ exon. S/J indicates the trans-splicing junction. (B) The S5eff ribozyme can revise a significant fraction of the sickle β-globin RNA. HEK293 cells were transfected with the β^S-globin plasmid alone (β^S-globin), or in conjunction with the plasmid expressing the S5eff ribozyme, present in 20-fold (20:1) or 2-fold (2:1) molar excess. Cells independently transfected with β^S-globin and S5eff plasmids were combined before cell lysis and RNA analysis (Mix). An inactive version of the S5eff ribozyme was cotransfected with the β^S-globin plasmid (Dead). A mixture of radiolabeled fragments corresponding to the amplified target and product were used as size markers (Mk). (C) The S5eff ribozyme can revise significant fractions of sickle β-globin RNA when expressed from different promoters. The efficiency of trans-splicing was quantitated in HEK293 cells cotransfected with the β^S-globin plasmid (pNL97) and a 10-fold molar excess of the S5eff ribozyme expression cassettes that are driven by the CMV (S5eff), U1 (U1S5eff), or T7 (T7S5eff) promoters. (D) Effect of varying ribozyme T7 expression levels on repair efficiency. The S5eff ribozyme was transcribed from the T7 promoter in HEK293 cells expressing T7 RNA polymerase. The ribozyme and sickle β-globin expression cassettes were cotransfected into cells at equal molar amounts. Differences in the transcription start sites in the T7-ribozyme templates result in ~100-fold greater expression of the ribozyme in “hi” (T7S5eff) versus “lo” (T7S5eff-lo) transfected cells. Cells independently expressing ribozyme and substrate RNAs were mixed at the time of cell lysis to generate the “mix” control. Repair efficiency was quantitated for duplicate samples by QC RT–PCR and is reported below the lanes.

FIGURE 2.

Evaluating trans-splicing efficiency of the S5 ribozyme. (A) Scheme for the measurement of trans-splicing efficiency by QC RT–PCR. The sickle β-globin target RNA (dashed line) and trans-spliced product RNA can be reverse transcribed and coamplified by a single set of PCR primers (arrows) to yield products that differ in size by 36 bp. The shaded box denotes the 18-nucleotide long priming sequence that is present in both the β^S-globin target RNA (nucleotides 119–136) and the S5eff ribozyme’s 3′ exon. S/J indicates the trans-splicing junction. (B) The S5eff ribozyme can revise a significant fraction of the sickle β-globin RNA. HEK293 cells were transfected with the β^S-globin plasmid alone (β^S-globin), or in conjunction with the plasmid expressing the S5eff ribozyme, present in 20-fold (20:1) or 2-fold (2:1) molar excess. Cells independently transfected with β^S-globin and S5eff plasmids were combined before cell lysis and RNA analysis (Mix). An inactive version of the S5eff ribozyme was cotransfected with the β^S-globin plasmid (Dead). A mixture of radiolabeled fragments corresponding to the amplified target and product were used as size markers (Mk). (C) The S5eff ribozyme can revise significant fractions of sickle β-globin RNA when expressed from different promoters. The efficiency of trans-splicing was quantitated in HEK293 cells cotransfected with the β^S-globin plasmid (pNL97) and a 10-fold molar excess of the S5eff ribozyme expression cassettes that are driven by the CMV (S5eff), U1 (U1S5eff), or T7 (T7S5eff) promoters. (D) Effect of varying ribozyme T7 expression levels on repair efficiency. The S5eff ribozyme was transcribed from the T7 promoter in HEK293 cells expressing T7 RNA polymerase. The ribozyme and sickle β-globin expression cassettes were cotransfected into cells at equal molar amounts. Differences in the transcription start sites in the T7-ribozyme templates result in ~100-fold greater expression of the ribozyme in “hi” (T7S5eff) versus “lo” (T7S5eff-lo) transfected cells. Cells independently expressing ribozyme and substrate RNAs were mixed at the time of cell lysis to generate the “mix” control. Repair efficiency was quantitated for duplicate samples by QC RT–PCR and is reported below the lanes.

FIGURE 2.

Evaluating trans-splicing efficiency of the S5 ribozyme. (A) Scheme for the measurement of trans-splicing efficiency by QC RT–PCR. The sickle β-globin target RNA (dashed line) and trans-spliced product RNA can be reverse transcribed and coamplified by a single set of PCR primers (arrows) to yield products that differ in size by 36 bp. The shaded box denotes the 18-nucleotide long priming sequence that is present in both the β^S-globin target RNA (nucleotides 119–136) and the S5eff ribozyme’s 3′ exon. S/J indicates the trans-splicing junction. (B) The S5eff ribozyme can revise a significant fraction of the sickle β-globin RNA. HEK293 cells were transfected with the β^S-globin plasmid alone (β^S-globin), or in conjunction with the plasmid expressing the S5eff ribozyme, present in 20-fold (20:1) or 2-fold (2:1) molar excess. Cells independently transfected with β^S-globin and S5eff plasmids were combined before cell lysis and RNA analysis (Mix). An inactive version of the S5eff ribozyme was cotransfected with the β^S-globin plasmid (Dead). A mixture of radiolabeled fragments corresponding to the amplified target and product were used as size markers (Mk). (C) The S5eff ribozyme can revise significant fractions of sickle β-globin RNA when expressed from different promoters. The efficiency of trans-splicing was quantitated in HEK293 cells cotransfected with the β^S-globin plasmid (pNL97) and a 10-fold molar excess of the S5eff ribozyme expression cassettes that are driven by the CMV (S5eff), U1 (U1S5eff), or T7 (T7S5eff) promoters. (D) Effect of varying ribozyme T7 expression levels on repair efficiency. The S5eff ribozyme was transcribed from the T7 promoter in HEK293 cells expressing T7 RNA polymerase. The ribozyme and sickle β-globin expression cassettes were cotransfected into cells at equal molar amounts. Differences in the transcription start sites in the T7-ribozyme templates result in ~100-fold greater expression of the ribozyme in “hi” (T7S5eff) versus “lo” (T7S5eff-lo) transfected cells. Cells independently expressing ribozyme and substrate RNAs were mixed at the time of cell lysis to generate the “mix” control. Repair efficiency was quantitated for duplicate samples by QC RT–PCR and is reported below the lanes.

FIGURE 2.

Evaluating trans-splicing efficiency of the S5 ribozyme. (A) Scheme for the measurement of trans-splicing efficiency by QC RT–PCR. The sickle β-globin target RNA (dashed line) and trans-spliced product RNA can be reverse transcribed and coamplified by a single set of PCR primers (arrows) to yield products that differ in size by 36 bp. The shaded box denotes the 18-nucleotide long priming sequence that is present in both the β^S-globin target RNA (nucleotides 119–136) and the S5eff ribozyme’s 3′ exon. S/J indicates the trans-splicing junction. (B) The S5eff ribozyme can revise a significant fraction of the sickle β-globin RNA. HEK293 cells were transfected with the β^S-globin plasmid alone (β^S-globin), or in conjunction with the plasmid expressing the S5eff ribozyme, present in 20-fold (20:1) or 2-fold (2:1) molar excess. Cells independently transfected with β^S-globin and S5eff plasmids were combined before cell lysis and RNA analysis (Mix). An inactive version of the S5eff ribozyme was cotransfected with the β^S-globin plasmid (Dead). A mixture of radiolabeled fragments corresponding to the amplified target and product were used as size markers (Mk). (C) The S5eff ribozyme can revise significant fractions of sickle β-globin RNA when expressed from different promoters. The efficiency of trans-splicing was quantitated in HEK293 cells cotransfected with the β^S-globin plasmid (pNL97) and a 10-fold molar excess of the S5eff ribozyme expression cassettes that are driven by the CMV (S5eff), U1 (U1S5eff), or T7 (T7S5eff) promoters. (D) Effect of varying ribozyme T7 expression levels on repair efficiency. The S5eff ribozyme was transcribed from the T7 promoter in HEK293 cells expressing T7 RNA polymerase. The ribozyme and sickle β-globin expression cassettes were cotransfected into cells at equal molar amounts. Differences in the transcription start sites in the T7-ribozyme templates result in ~100-fold greater expression of the ribozyme in “hi” (T7S5eff) versus “lo” (T7S5eff-lo) transfected cells. Cells independently expressing ribozyme and substrate RNAs were mixed at the time of cell lysis to generate the “mix” control. Repair efficiency was quantitated for duplicate samples by QC RT–PCR and is reported below the lanes.

FIGURE 3.

Evaluating the importance of the P10 and EGS–target interactions for trans-splicing (A) Derivatives of the S5eff ribozyme with modified P10 or EGS sequences. Ribozyme derivatives pairing to the β^S-globin target RNA (dashed line) are shown. The shaded box present in the ribozyme 3′ exon represents the priming site used to quantitate trans-splicing efficiency by QC RT–PCR. Compared to the 35-bp interaction between the external guide sequence (EGS) of the S5eff ribozyme and the target RNA, S6eff, S9eff, and S8eff ribozymes include an EGS that can form 0 bp, 15 bp, and 54 bp, respectively. Although the EGS is unmodified in S10eff, the P10 helix is disrupted. (B) Trans-splicing requires P10 and EGS pairing interactions. Expression constructs encoding the ribozymes described in (A) were co-transfected in 20-fold molar excess with β^S-globin plasmid. The percentage of trans-spliced products were determined as described in Figure 2 ▶. For each ribozyme construct, the presence (+) or absence (−) of P10, and the length of the EGS interaction in base-pairs is indicated. Control transfections (“Dead” and “Mix”) were performed as in Figure 2B ▶. Molecular mass markers (MW) are shown.

FIGURE 3.

Evaluating the importance of the P10 and EGS–target interactions for trans-splicing (A) Derivatives of the S5eff ribozyme with modified P10 or EGS sequences. Ribozyme derivatives pairing to the β^S-globin target RNA (dashed line) are shown. The shaded box present in the ribozyme 3′ exon represents the priming site used to quantitate trans-splicing efficiency by QC RT–PCR. Compared to the 35-bp interaction between the external guide sequence (EGS) of the S5eff ribozyme and the target RNA, S6eff, S9eff, and S8eff ribozymes include an EGS that can form 0 bp, 15 bp, and 54 bp, respectively. Although the EGS is unmodified in S10eff, the P10 helix is disrupted. (B) Trans-splicing requires P10 and EGS pairing interactions. Expression constructs encoding the ribozymes described in (A) were co-transfected in 20-fold molar excess with β^S-globin plasmid. The percentage of trans-spliced products were determined as described in Figure 2 ▶. For each ribozyme construct, the presence (+) or absence (−) of P10, and the length of the EGS interaction in base-pairs is indicated. Control transfections (“Dead” and “Mix”) were performed as in Figure 2B ▶. Molecular mass markers (MW) are shown.

FIGURE 4.

Evaluating EGS–target RNA base-pairing requirements. (A) Modifications to the β^S-globin target sequence. The interaction between the S5eff ribozyme and the β^S-globin target RNA is shown in detail. The EGS–target interaction is highlighted in the boxed region. Derivatives of β^S-globin target RNA are represented below the boxed region. Modular mismatches of 5 bp were cumulatively introduced into the target to disrupt increasing regions of the EGS–target interaction. Unaltered regions of the target sequence are represented by the lines. The nucleotide sequences of the mismatched regions in the target RNA are shown. The number of base-pairs remaining in the EGS–target interaction is indicated in bold type. The sickle mutation is denoted with an asterisk. (B) Relative trans-splicing efficiency of the S5eff ribozyme with target RNAs containing mismatches in the region complementary to the ribozyme EGS. The matched (35 EGS bp) and mismatched (5–30 EGS bp) target RNAs were co-expressed with the S5eff ribozyme in HEK293 cells. Trans-splicing efficiency was quantitated by QC RT–PCR analysis for multiple independent transfection experiments (n = 6). Efficiency of trans-splicing was compared to the efficiency of S5eff trans-splicing with the matched β^S-globin target.

FIGURE 4.

Evaluating EGS–target RNA base-pairing requirements. (A) Modifications to the β^S-globin target sequence. The interaction between the S5eff ribozyme and the β^S-globin target RNA is shown in detail. The EGS–target interaction is highlighted in the boxed region. Derivatives of β^S-globin target RNA are represented below the boxed region. Modular mismatches of 5 bp were cumulatively introduced into the target to disrupt increasing regions of the EGS–target interaction. Unaltered regions of the target sequence are represented by the lines. The nucleotide sequences of the mismatched regions in the target RNA are shown. The number of base-pairs remaining in the EGS–target interaction is indicated in bold type. The sickle mutation is denoted with an asterisk. (B) Relative trans-splicing efficiency of the S5eff ribozyme with target RNAs containing mismatches in the region complementary to the ribozyme EGS. The matched (35 EGS bp) and mismatched (5–30 EGS bp) target RNAs were co-expressed with the S5eff ribozyme in HEK293 cells. Trans-splicing efficiency was quantitated by QC RT–PCR analysis for multiple independent transfection experiments (n = 6). Efficiency of trans-splicing was compared to the efficiency of S5eff trans-splicing with the matched β^S-globin target.

FIGURE 5.

A trans-splicing ribozyme that preferentially reacts with sickle β-globin versus wild-type β-globin RNA. (A) Design of the S5eff-5AS ribozyme. The interaction between the S5eff-5AS ribozyme and the β^S-globin target RNA is shown in detail. The EGS–target interaction is limited to 5 bp and is highlighted in the boxed region. The sickle mutation is denoted by the asterisk. The sequence of wild-type β-globin RNA (β^WT-globin) is shown below for comparison. The wild-type nucleotide corresponding to the sickle mutation is circled. (B) Trans-splicing analysis of S5eff-5AS and S5eff ribozymes with wild-type and sickle β-globin RNA. Plasmids encoding S5eff ribozymes harboring an EGS of 5 bp (S5eff-5AS) or 35 bp (S5eff-35AS) were cotransfected in twofold molar excess with wild-type (β^WT) or sickle (β^S) β-globin plasmid. Trans-splicing analysis was carried out as described in Figure 2 ▶. Cells independently transfected with β^S-globin and S5eff plasmids were combined before cell lysis and RNA analysis (Mix). A mixture of radiolabeled fragments corresponding to the amplified target and product were used as size markers (Mk). (C) Comparison of trans-splicing efficiency between the S5eff-5AS and the S5eff ribozyme on wild type β-globin and sickle β-globin targets. Independent transfection experiments (n = 4) were analyzed for trans-splicing efficiency as in (B). The results are represented relative to trans-splicing of S5eff on β^S-globin target RNA.

FIGURE 5.

A trans-splicing ribozyme that preferentially reacts with sickle β-globin versus wild-type β-globin RNA. (A) Design of the S5eff-5AS ribozyme. The interaction between the S5eff-5AS ribozyme and the β^S-globin target RNA is shown in detail. The EGS–target interaction is limited to 5 bp and is highlighted in the boxed region. The sickle mutation is denoted by the asterisk. The sequence of wild-type β-globin RNA (β^WT-globin) is shown below for comparison. The wild-type nucleotide corresponding to the sickle mutation is circled. (B) Trans-splicing analysis of S5eff-5AS and S5eff ribozymes with wild-type and sickle β-globin RNA. Plasmids encoding S5eff ribozymes harboring an EGS of 5 bp (S5eff-5AS) or 35 bp (S5eff-35AS) were cotransfected in twofold molar excess with wild-type (β^WT) or sickle (β^S) β-globin plasmid. Trans-splicing analysis was carried out as described in Figure 2 ▶. Cells independently transfected with β^S-globin and S5eff plasmids were combined before cell lysis and RNA analysis (Mix). A mixture of radiolabeled fragments corresponding to the amplified target and product were used as size markers (Mk). (C) Comparison of trans-splicing efficiency between the S5eff-5AS and the S5eff ribozyme on wild type β-globin and sickle β-globin targets. Independent transfection experiments (n = 4) were analyzed for trans-splicing efficiency as in (B). The results are represented relative to trans-splicing of S5eff on β^S-globin target RNA.

FIGURE 5.

A trans-splicing ribozyme that preferentially reacts with sickle β-globin versus wild-type β-globin RNA. (A) Design of the S5eff-5AS ribozyme. The interaction between the S5eff-5AS ribozyme and the β^S-globin target RNA is shown in detail. The EGS–target interaction is limited to 5 bp and is highlighted in the boxed region. The sickle mutation is denoted by the asterisk. The sequence of wild-type β-globin RNA (β^WT-globin) is shown below for comparison. The wild-type nucleotide corresponding to the sickle mutation is circled. (B) Trans-splicing analysis of S5eff-5AS and S5eff ribozymes with wild-type and sickle β-globin RNA. Plasmids encoding S5eff ribozymes harboring an EGS of 5 bp (S5eff-5AS) or 35 bp (S5eff-35AS) were cotransfected in twofold molar excess with wild-type (β^WT) or sickle (β^S) β-globin plasmid. Trans-splicing analysis was carried out as described in Figure 2 ▶. Cells independently transfected with β^S-globin and S5eff plasmids were combined before cell lysis and RNA analysis (Mix). A mixture of radiolabeled fragments corresponding to the amplified target and product were used as size markers (Mk). (C) Comparison of trans-splicing efficiency between the S5eff-5AS and the S5eff ribozyme on wild type β-globin and sickle β-globin targets. Independent transfection experiments (n = 4) were analyzed for trans-splicing efficiency as in (B). The results are represented relative to trans-splicing of S5eff on β^S-globin target RNA.