Cryptic prokaryotic promoters explain instability of recombinant neuronal sodium channels in bacteria

Abstract

Mutations in genes encoding the human-brain-expressed voltage-gated sodium (NaV) channels NaV1.1, NaV1.2, and NaV1.6 are associated with a variety of human diseases including epilepsy, autism spectrum disorder, familial migraine, and other neurodevelopmental disorders. A major obstacle hindering investigations of the functional consequences of brain NaV channel mutations is an unexplained instability of the corresponding recombinant complementary DNA (cDNA) when propagated in commonly used bacterial strains manifested by high spontaneous rates of mutation. Here, using a combination of in silico analysis, random and site-directed mutagenesis, we investigated the cause for instability of human NaV1.1 cDNA. We identified nucleotide sequences within the NaV1.1 coding region that resemble prokaryotic promoter-like elements, which are presumed to drive transcription of translationally toxic mRNAs in bacteria as the cause of the instability. We further demonstrated that mutations disrupting these elements mitigate the instability. Extending these observations, we generated full-length human NaV1.1, NaV1.2, and NaV1.6 plasmids using one or two introns that interrupt the latent reading frames along with a minimum number of silent nucleotide changes that achieved stable propagation in bacteria. Expression of the stabilized sequences in cultured mammalian cells resulted in functional NaV channels with properties that matched their parental constructs. Our findings explain a widely observed instability of recombinant neuronal human NaV channels, and we describe re-engineered plasmids that attenuate this problem. Mutations in genes encoding the human-brain-expressed voltage-gated sodium (NaV) channels NaV1.1, NaV1.2, and NaV1.6 are associated with a variety of human diseases including epilepsy, autism spectrum disorder, familial migraine, and other neurodevelopmental disorders. A major obstacle hindering investigations of the functional consequences of brain NaV channel mutations is an unexplained instability of the corresponding recombinant complementary DNA (cDNA) when propagated in commonly used bacterial strains manifested by high spontaneous rates of mutation. Here, using a combination of in silico analysis, random and site-directed mutagenesis, we investigated the cause for instability of human NaV1.1 cDNA. We identified nucleotide sequences within the NaV1.1 coding region that resemble prokaryotic promoter-like elements, which are presumed to drive transcription of translationally toxic mRNAs in bacteria as the cause of the instability. We further demonstrated that mutations disrupting these elements mitigate the instability. Extending these observations, we generated full-length human NaV1.1, NaV1.2, and NaV1.6 plasmids using one or two introns that interrupt the latent reading frames along with a minimum number of silent nucleotide changes that achieved stable propagation in bacteria. Expression of the stabilized sequences in cultured mammalian cells resulted in functional NaV channels with properties that matched their parental constructs. Our findings explain a widely observed instability of recombinant neuronal human NaV channels, and we describe re-engineered plasmids that attenuate this problem. Recombinant brain-expressed voltage-gated sodium (NaV) channels are important tools for investigating structure–function relationships, drug discovery, and for determining the consequences of naturally occurring human mutations associated with neurological and neurodevelopmental disorders such as epilepsy, familial migraine, and autism spectrum disorder (1Decaen P.G. George Jr., A.L. Thompson C.H. Sodium channelopathies of the central nervous system.in: Bhattacharjee A. The Oxford Handbook of Neuronal Ion Channels. Oxford University Press, Oxford, United Kingdom2019: 1-36Google Scholar, 2Catterall W.A. Sodium channels, inherited epilepsy, and antiepileptic drugs.Annu. Rev. Pharmacol. Toxicol. 2014; 54: 317-338Crossref PubMed Scopus (106) Google Scholar). However, high-throughput manipulations of these nucleic acid reagents are difficult owing to the instability of neuronal NaV channels when propagated in bacteria (3Feldman D.H. Lossin C. The NaV channel bench series: Plasmid preparation.MethodsX. 2014; 1: 6-11Crossref PubMed Scopus (4) Google Scholar). When transformed into typical Escherichia coli bacterial host cells, some neuronal NaV channel cDNAs, particularly NaV1.1, NaV1.2, and NaV1.6, exhibit a high spontaneous mutation rate resulting in single-nucleotide changes, small deletions, and insertion sequence (IS) element integrations. Certain bacterial strains (e.g., JM101, DH5α, DH10B/Top10) appear to be more prone to this instability while others such as Stbl2 and Epi400 can be used to maintain the cDNAs more stably, possibly by reducing the plasmid copy number. However, even in the more accommodating cell types, mutations do occur and render the isolated plasmid DNA useless for experiments. Clones on agar plates that appear early and grow into large colonies are almost invariably corrupt, whereas the small later-appearing clones have a lower incidence of spontaneous mutation. A number of approaches have been used to mitigate the plasmid instabilities of human neuronal NaV channel cDNAs, such as the use of the aforementioned Stbl2 or Epi400 host strains, low (30 °C) growth temperature, selecting only small colonies, and not allowing liquid cultures to reach the saturation phase (3Feldman D.H. Lossin C. The NaV channel bench series: Plasmid preparation.MethodsX. 2014; 1: 6-11Crossref PubMed Scopus (4) Google Scholar). Although these precautions can increase the odds of obtaining unaltered DNA, unwanted mutations still emerge. Codon optimization has been reported to alleviate the instability of recombinant NaV1.1, NaV1.2, and NaV1.6 (4Bertelli S. Barbieri R. Pusch M. Gavazzo P. Gain of function of sporadic/familial hemiplegic migraine-causing SCN1A mutations: Use of an optimized cDNA.Cephalalgia. 2019; 39: 477-488Crossref PubMed Scopus (10) Google Scholar, 5Patel R.R. Barbosa C. Brustovetsky T. Brustovetsky N. Cummins T.R. Aberrant epilepsy-associated mutant NaV1.6 sodium channel activity can be targeted with cannabidiol.Brain. 2016; 139: 2164-2181Crossref PubMed Scopus (65) Google Scholar, 6Mason E.R. Wu F. Patel R.R. Xiao Y. Cannon S.C. Cummins T.R. Resurgent and gating pore currents induced by de novo SCN2A epilepsy mutations.eNeuro. 2019; 6 (ENEURO.0141-19.2019)Crossref PubMed Scopus (9) Google Scholar). However, this approach may not always be desirable because altering rare codons encoding protein transmembrane domains may impair translational pausing and affect proper protein folding (7Komar A.A. A pause for thought along the co-translational folding pathway.Trends Biochem. Sci. 2009; 34: 16-24Abstract Full Text Full Text PDF PubMed Scopus (237) Google Scholar, 8Mauro V.P. Chappell S.A. A critical analysis of codon optimization in human therapeutics.Trends Mol. Med. 2014; 20: 604-613Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar, 9Shabalina S.A. Spiridonov N.A. Kashina A. Sounds of silence: Synonymous nucleotides as a key to biological regulation and complexity.Nucleic Acids Res. 2013; 41: 2073-2094Crossref PubMed Scopus (151) Google Scholar). Furthermore, while codon optimization may be effective for stabilizing cDNAs, this approach does not explain the cause of the instabilities. Because of the importance of NaV channel research and the cost in time and resources incurred by this instability, we sought to determine the underlying molecular causes to enable a rationally designed solution to generate stable NaV channel cDNAs with few or no codon alterations. We investigated causes of the bacterial instability of a recombinant human NaV1.1 cDNA (NCBI accession NM_001165963) by determining the stability of individual cloned subfragments (Fig. 1A). When divided into four overlapping cassettes (cassette-1: coding sequence nucleotides [c.] 1–1560; cassette-2: c.1555–3218; cassette-3: c.3213–5601; and cassette-4: c.5341–6030), only cassette-3 exhibited instability in bacteria as evidenced by a high proportion of isolated plasmid DNAs having altered restriction fragment patterns. To pinpoint sequences responsible for this instability, cassette-3 was further divided into seven overlapping subfragments for analysis. Two of the subfragments (subfragments 5 and 7) exhibited instability in bacteria suggesting they harbored unstable elements (Fig. 1B). To pinpoint sequences that were responsible for the instability, we used serial polymerase chain reaction (PCR) amplification of subfragments 5 and 7 to introduce random mutations that would result in rare stabilizing events informative about the location of unstable elements. Plasmid DNA isolated from fast-growing (appearing after overnight growth) bacterial colonies was screened by restriction digestion to identify clones with correct restriction patterns, reasoning that some of those clones would have stabilizing single-nucleotide mutations. The sequences of three clones (one from subfragment 5, two from subfragment 7) were informative. Two of the three stable clones had single-nucleotide mutations (c.4507G>T and c.4594G>T, nucleotide numbering based on the full-length NaV1.1 open reading frame [ORF]) that introduced in-frame premature stop codons (E1503X and G1532X), whereas the third clone had a single nucleotide mutation (c.4157A>G) resulting in a nonsynonymous codon change (D1386E; Fig. 2A). This raised the possibility that stabilization of the cloned subfragments occurred secondary to either truncation or disruption of an inappropriately expressed portion of the NaV1.1 cDNA. Analysis of the sequence surrounding the nonsynonymous mutation revealed a prokaryotic promoter-like element. Specifically, the mutation disrupted a putative -35 box of a predicted Sigma 70 factor promoter-like sequence (Fig. 2B). These findings suggested that stop codons in the first two clones truncated a toxic reading frame translated from mRNA transcribed by the promoter-like sequence that was disrupted by the nonsynonymous mutation in the third clone. To test this hypothesis, three sets of silent mutations were made in the promoter-like sequence of cassette-3 to disrupt the putative -35 box (c.4155T>C), -10 box (c.4179T>C and c.4182T>C), or both elements (Fig. 2B). All three sets of silent mutations resulted in stable cDNA clones. To test whether the instability in the cassette correlated with a translated region of NaV1.1, four in-frame methionine residues (Met1435, Met1438, Met1459, and Met1500) in the original cDNA between the promoter-like sequence and the earliest PCR-induced stop codon were mutated from ATG to ATA. Only one ATG>ATA mutation (c.4500G>A; Met1500Ile) stabilized cassette-3, indicating that this codon may serve as an aberrant translation start (Fig. 3A). In support of this idea, a high-scoring bacterial ribosome-binding site (Shine-Dalgarno sequence; TTTGGAGGT) was predicted 13–21 nucleotides 5′ of this ATG (c.4486–4497; Fig. 3B). The location of this cryptic reading frame within the full-length NaV1.1 cDNA sequence aligns with the last six-transmembrane domain (domain IV) of the encoded NaV channel protein. Introduction of the cassette-3 stabilizing mutations into the full-length NaV1.1 cDNA did not stabilize the entire plasmid, and we suspected the existence of other cryptic promoter-like elements elsewhere in the sequence. While NaV1.1 cassettes 1 and 2 were stable individually, they were unstable when assembled together. Using a similar strategy that led to identification of promoter-like elements in cassette-3, we tested the stability of six overlapping subfragments of cassette-1/2 and identified a 2029 bp unstable segment (c.749–2778). Mutating one of the four methionine codons (Met-350; c.1048–1050) in this segment to ATA stabilized cassette-1/2. An in silico search for additional cryptic promoters identified high-scoring promoter-like sequences (see Experimental procedures) between positions c.27 and c.3113 (Table S1). Cassettes 1 and 2 were remade by gene synthesis to disrupt the putative promoter elements with silent mutations (Table S2), then recombined with the stabilized cassette-3 and the unaltered stable cassette-4. The final reconstructed full-length cDNA (Fig. S1) included 60 synonymous nucleotide changes and was stably propagated in bacteria (Top10 cells) grown at 37 °C. We inferred that a cryptic promoter located at nucleotide position 875–903 was driving a toxic reading frame initiated at Met-350. Figure 4 illustrates the location of cryptic bacterial promoters, translation start sites, and toxic peptides on a topographical map of the NaV1.1 channel. Guided by these initial observations, we chose a simpler strategy that required fewer alterations of the coding sequence to stabilize the human NaV1.1 cDNA and two other human neuronal sodium channels (NaV1.2, NCBI accession NM_021007; NaV1.6, NCBI accession NM_014191) using short introns to interrupt the putative toxic reading frames. We introduced introns at native exon–exon junctions close to or within the sequence analogous to the toxic reading frame found in NaV1.1 or in regions where bacterial insertion elements had been previously observed in stable but mutated NaV1.1 clones. We inserted a previously published β-globin/IgG chimeric intron intervening sequence (IVS) (GenBank accession U47120) after in the NaV1.1 cDNA. The human NaV1.1 cDNA required additional silent mutations to a cryptic promoter located at nucleotide position 875–903 including in the -35 box and three others in the -10 box (Fig. S2), as as a silent change in the -35 box of the cDNA was achieved by the after and no other were the NaV1.6 we inserted the after and a intron after The intron was from intron 4 of human (NCBI accession more matched the for intron 4 Fig. We the of a silent mutation designed to disrupt an that to to bacterial bacterial growth of a cDNA However, this mutation did not cDNA stability or plasmid of full-length NaV1.6 and was was not included in the final We made two silent mutations to disrupt a putative promoter-like sequence analogous to that in the NaV1.1 cDNA. The combination of these resulted in full-length NaV1.1, NaV1.2, and NaV1.6 cDNAs that were stable in bacteria host grown (Fig. The nucleotide sequences of intron stabilized NaV1.1, NaV1.2, and NaV1.6 are in we the functional properties of the original and intron stabilized NaV1.1, NaV1.2, and NaV1.6 cDNAs by of We observed no in of or of sodium for (Fig. we of and from and at and found no in functional properties between the NaV channels with or introns (Table Fig. We that insertion of introns and silent mutations into human NaV1.1, NaV1.2, and NaV1.6 cDNAs had no functional on the encoded properties of original and stabilized human NaV in a Recombinant human brain NaV channels are molecular tools for determining the functional consequences of mutations and for including drug Jr., A.L. disorders of voltage-gated sodium PubMed Scopus Google Scholar). use of these molecular reagents is by an instability of recombinant NaV channel plasmid in which was of Here, we that the instability of brain NaV channel cDNAs in bacteria correlated with the of one or more cryptic promoter-like elements driving reading frames that peptides toxic to the We that this is to a bacteria but a for cells clones with or truncated reading frames. to this the instability of brain NaV cDNAs in bacteria is the result of growth of bacterial clones mutant cDNAs in which the toxic reading frames have been The of cryptic bacterial promoter-like elements in unstable recombinant human and genes has been The of human growth cDNA with clones in bacteria because of an of the recombinant The cause of this was to a cryptic promoter that be by a single-nucleotide change designed to disrupt a box and this resulted in stable high copy number clones C. S. C. P. M. growth and of the full-length complementary DNA after site-directed of a cryptic bacterial PubMed Scopus Google Scholar). A similar has been with recombinant coding sequences that be by intron insertion insertion amplification of cloned cDNA in Escherichia coli while biological activity is after transcription in Sci. S. A. PubMed Scopus Google Scholar, Wu A. propagation of cDNAs by a to the cryptic bacterial promoter activity of PubMed Scopus Google Scholar). The of cryptic bacterial promoters in recombinant cDNAs may be and be when unstable The in which the expressed of brain NaV channels toxic in coli cells is We that may be the result of of the which transmembrane into the bacterial to host cells to is that the toxic We identified a toxic in human NaV1.6 but this did not to the instability of the full-length cDNA. can of and of protein synthesis as has been demonstrated for the The rate of from the on the of the last Full Text Full Text PDF PubMed Scopus Google Scholar, C. of protein synthesis of and for PubMed Scopus Google Scholar). approaches that have stabilized recombinant NaV channel plasmids in bacteria the use of host cells grown (3Feldman D.H. Lossin C. The NaV channel bench series: Plasmid preparation.MethodsX. 2014; 1: 6-11Crossref PubMed Scopus (4) Google and the use of codon optimization (4Bertelli S. Barbieri R. Pusch M. Gavazzo P. Gain of function of sporadic/familial hemiplegic migraine-causing SCN1A mutations: Use of an optimized cDNA.Cephalalgia. 2019; 39: 477-488Crossref PubMed Scopus (10) Google Scholar, 5Patel R.R. Barbosa C. Brustovetsky T. Brustovetsky N. Cummins T.R. Aberrant epilepsy-associated mutant NaV1.6 sodium channel activity can be targeted with cannabidiol.Brain. 2016; 139: 2164-2181Crossref PubMed Scopus (65) Google Scholar, 6Mason E.R. Wu F. Patel R.R. Xiao Y. Cannon S.C. Cummins T.R. Resurgent and gating pore currents induced by de novo SCN2A epilepsy mutations.eNeuro. 2019; 6 (ENEURO.0141-19.2019)Crossref PubMed Scopus (9) Google Scholar). The approach used by (4Bertelli S. Barbieri R. Pusch M. Gavazzo P. Gain of function of sporadic/familial hemiplegic migraine-causing SCN1A mutations: Use of an optimized cDNA.Cephalalgia. 2019; 39: 477-488Crossref PubMed Scopus (10) Google to human NaV1.1 codon changes that those found in human a cDNA that is more stable in bacteria. of their optimized NaV1.1 sequence revealed single-nucleotide changes that disrupted promoter-like elements including two to the -10 box in cassette single-nucleotide in the -35 box and -10 box in cassette and changes to the predicted sequence in cassette These are for the stabilization of their Codon optimization of human NaV1.6 R.R. Barbosa C. Brustovetsky T. Brustovetsky N. Cummins T.R. Aberrant epilepsy-associated mutant NaV1.6 sodium channel activity can be targeted with cannabidiol.Brain. 2016; 139: 2164-2181Crossref PubMed Scopus (65) Google and E.R. Wu F. Patel R.R. Xiao Y. Cannon S.C. Cummins T.R. Resurgent and gating pore currents induced by de novo SCN2A epilepsy mutations.eNeuro. 2019; 6 (ENEURO.0141-19.2019)Crossref PubMed Scopus (9) Google may have disrupted cryptic promoter-like elements, but we did not map these sequences the of codon optimization to stabilize brain NaV channel cDNAs, may be desirable to maintain the codon because synonymous nucleotide changes may disrupt the of translation required for proper protein folding (7Komar A.A. A pause for thought along the co-translational folding pathway.Trends Biochem. Sci. 2009; 34: 16-24Abstract Full Text Full Text PDF PubMed Scopus (237) Google Scholar, 8Mauro V.P. Chappell S.A. A critical analysis of codon optimization in human therapeutics.Trends Mol. Med. 2014; 20: 604-613Abstract Full Text Full Text PDF PubMed Scopus (95) Google Scholar, 9Shabalina S.A. Spiridonov N.A. Kashina A. Sounds of silence: Synonymous nucleotides as a key to biological regulation and complexity.Nucleic Acids Res. 2013; 41: 2073-2094Crossref PubMed Scopus (151) Google Scholar). codons may translational pausing that folding of the including channels may be particularly to these codon optimization of the human gene channel in cells and causes a in the of with of cDNA with the native coding sequence Synonymous nucleotide of the gene both mRNA and translation of the encoded Full Text Full Text PDF PubMed Scopus Google Scholar). The of synonymous codon may to determine the functional consequences of nonsynonymous NaV channel The approach we used to stabilize brain NaV channels this the responsible for instability, and the functional properties of the original constructs. We NaV channel plasmid instability by restriction analysis of plasmid DNA isolated from overnight in restriction fragment were used as a of plasmid instability, and the proportion of bacterial clones disrupted was used to the of instability. Plasmid DNA was for 1 at 37 with in a final of restriction fragment by growth of bacterial colonies on agar and in liquid was used as a for which was inferred by growth or low plasmid Clones with the correct restriction were then to using a of the unstable NaV1.1 cassette-3 were in using and the in for by of for for for and for were using the and then into of were used to cells that were then on agar with colonies were cultured overnight at 37 in with Plasmid DNA was isolated using and then with clones with a restriction were in both using the and by We identified cryptic bacterial promoters using the for prokaryotic promoters on the with as the We the cDNA sequences using using a for promoters of A. of and and in and and with the DNA of Plasmid using a of which were in-frame with the sodium channel were identified using Plasmid with to and a toxic were identified by a of were designed with the of a 5′ of and a predicted of 60 using was with of plasmid DNA using in a final reaction of to the that the time was to for a of Plasmid DNA in the reaction was with for at 37 then of were used to cells using a and with In silico analysis of high coli promoter-like sequences was on the first bp of the human NaV1.1 cDNA. Cassettes 1 and cassette 2 were as DNAs DNA with the promoter-like elements altered by synonymous mutations and cloned into clones were screened by restriction digestion and those with correct were by DNA The cassettes 1 and 2 were assembled with the stabilized cassette in the and into a mammalian plasmid for functional (see All cDNA were in their by at the promoter the The introns were using with designed to to their were inserted into NaV channel plasmids by using to the were inserted after of NaV1.1, and after of NaV1.1 required silent mutations in a promoter-like element located at In NaV1.6, the first intron was inserted after between and and the intron was inserted after between and Two silent mutations were made to a cryptic promoter-like sequence analogous to that of the NaV1.1 cassette were into mammalian plasmids A. George Jr., A.L. stable gene in human Sci. S. A. PubMed Scopus Google or included an sequence after the sodium channel sequence that was by the reading frame for the protein of NaV channels was in cells were grown in at 37 in with 2 and were with original or NaV1.1, NaV1.2, or NaV1.6, along with the human NaV and plasmid DNA was with a cDNA of for using and cDNAs were cloned into plasmids encoding the or protein as as previously C. George Jr., A.L. of an inherited 34: Full Text Full Text PDF PubMed Scopus Google Scholar). of cells was as previously C.H. George Jr., A.L. SCN1A exhibit to commonly used antiepileptic PubMed Scopus Google Scholar, C.H. George Jr., A.L. SCN1A mutations associated with epilepsy of impair cell Full Text Full Text PDF PubMed Scopus Google Scholar). were made at using an were from with a and using a to a of The solution of 2 and with to with and to with The solution was with solution 4 1 and with and A as the All were from and were using was by generated by a from to at was with and to that the was within with a currents were by using an All currents were at 5 and at All were at were using a combination of and was for cell and to generate a sodium currents were by from to 60 by a to from a of was as where is the is the and is the sodium predicted by of the for at was to the between and To of in was and with the function where the at and is a factor of the channel. of was by currents generated by the to the from to in The was with the from was by a The first currents with a to from a of to The to was after from 1 to from was by with a was by to at a of are as were using or by a was The stable full-length sodium channel generated in this have been made the NaV1.1 NaV1.6 All are in the A. on a for and has research support from and for The A. C. for to generate the sodium channel (Fig. and C. T. the and the All to the This was by of The is the of the and does not the of the of

Topics & Concepts