{"paper_id":"1308fa30-3e2d-4e89-93c4-0e502e023624","body_text":"1 \n \nTm–guided exon–exon junction RT-PCR enables specific detection of RNA \nvariants lacking easily distinguishable exonic regions  \nJunyeop Ahn1, Donald J Zack 1,2, Ping-Wu Zhang1* \nAffiliations: \n1. Department of Ophthalmology, Stem Cell Ocular Regenerative Medicine Center, Wilmer Eye Institute, \nJohns Hopkins University School of Medicine; Baltimore, MD, 21231, USA \n2. Solomon H. Snyder Department of Neuroscience, Department of Molecular Biology and Genetics, \nDepartment of Genetic Medicine, Center for Nanomedicine at the Wilmer Eye Institute, The Johns \nHopkins University School of Medicine, Baltimore, Maryland, United States \n*Corresponding authors: \nPing-Wu Zhang (pzhang5@jhmi.edu)  \nContact Address: \nWilmer Eye Institute, Johns Hopkins University School of Medicine,  \n400 N Broadway, Smith Building, Room 3001,  \nBaltimore, Maryland, 21231, USA \nAbstract: \nAccurate detection of RNA splice variants is often hindered when transcripts lack large distinguishable \nexonic regions, making conventional PCR strategies challenging. We developed a simple melting \ntemperature (Tm)-guided exon–exon junction (EEJ) RT-PCR method to enable variant-specific detection \nunder these conditions. Uni-directional primers spanning exon–exon junctions were designed so that \napproximately each half anneals to adjacent exons. The Tm of each half-site was set >7°C below the \nannealing temperature, preventing stable binding to individual exons and enforcing junction-dependent \namplification. The method was evaluated using HTRA1-AS1 long noncoding RNA variants that share \noverlapping exon sequences but differ in splice connectivity. HTRA1-AS1 comprises five variants, only \none with a large distinguishable exon. Tm-guided EEJ primers robustly discriminated the remaining four \nvariants. After optimization, amplification yielded sharp, single bands with minimal cross-reactivity. \nCompared with conventional designs, this approach reduced heteroduplex and heteroquadruplex \nformation, improving band clarity. Sanger sequencing confirmed junction specificity, and the method \nperformed well in multiplex settings. \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n2 \n \nOverall, Tm-guided EEJ RT-PCR is a cost-effective, high-resolution approach for detecting RNA variants \nlacking easily distinguishable exonic regions, readily compatible with standard RT-PCR and qPCR \nworkflows. \nKeywords: Transcript-specific PCR; RNA variant; detection; heteroduplex; heteroquadruplex;  exon–exon \njunction; primer design  \nIntroduction \nAlternative splicing is an important mechanism that expands transcriptomic and proteomic diversity in \neukaryotic organisms. Human genes have average of eight exons and three or more alternatively spliced \nmRNA isoforms. It is estimated that more than 90% of multi-exon human genes generate multiple \ntranscript variants through alternative exon usage, exon skipping, alternative splice sites, and intron \nretention. These variants can differ in regulatory function, subcellular localization, and protein-coding \npotential, and often play critical roles in development, disease progression, and cellular responses to \nstress [1–3]. \nReliable detection and quantification of specific transcript variants remain technically challenging, \nparticularly when transcripts share extensive exon overlap. Conventional RT-PCR and quantitative PCR \napproaches typically rely on primers targeting large distinguishable exons to individual transcripts. \nHowever, for transcript variants that lack these regions and instead share very similar exon sequences, \ndiffering only in exon connectivity generated by alternative splicing, design of exon-targeting primer that \ndistinguish highly similar isoforms, leading to ambiguous detection or co-amplification of multiple \ntranscripts [4, 5]. High-throughput sequencing approaches, such as long-read sequencing technologies, \ncan effectively identify and distinguish highly similar splice isoforms [6,7]. However, these methods are \ncostly, computationally intensive, and not always practical for routine validation or targeted transcript \nanalysis.  \nA cost-effective PCR-based alternative strategy for distinguishing highly similar transcript variants is to \ntarget exon–exon junctions generated during RNA splicing. Because each transcript variant is defined by \na specific combination of exon junctions, primers spanning these junctions can provide variant \nspecificity even in the absence of large distinguishable exonic regions. Intron-spanning primers have \nbeen widely used to reduce genomic DNA amplification and to detect splice variants in qPCR assays \n[8,9]. In addition, computational tools such as Ex-Ex Primer and ExonSurfer have been developed to \nassist in exon–exon junction primer design for transcript detection and genomic DNA avoidance [10,11]. \nHowever, experimentally validated strategies and details for using exon–exon junction primers to \ndistinguish transcript variants lacking large distinguishable exonic regions have not been well \nestablished. \nIn this study, we present a practical melting temperature (Tm)–guided exon–exon junction RT-PCR \nmethod for the identification and discrimination of RNA transcript variants that share largely identical \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n3 \n \nexonic sequences. This approach utilizes a dual-recognition primer design in which each primer spans a \nsplice junction, with the Tm of each half-site optimized to prevent stable annealing to individual exons. \nConsequently, efficient amplification depends on correct exon–exon connectivity, conferring high \nspecificity while minimizing nonspecific amplification and heteroduplex formation. Here we outline the \nprimer design principles, optimization strategy, and experimental validation using RT-PCR and Sanger \nsequencing. Overall, this method provides a simple, robust, and broadly applicable strategy for detecting \nclosely related transcript variants in studies of alternative splicing and transcript-specific gene \nexpression. \nMethods \nPrimer design \nPrimer pairs were designed to selectively amplify transcript variants based on exon–exon junctions rather \nthan exon-specific sequences. Primers were initially designed using Primer3plus software \n(https://www.primer3plus.com/index.html), and exon–exon junction primers were further refined \nmanually. For each target transcript, either the forward or reverse primer was designed to span the exon–\nexon junction, with approximately half of the primer sequence complementary to the upstream exon and \nthe remaining portion complementary to the downstream exon. This design ensures that amplification \noccurs only when the specific exon junction is present in the cDNA template, thereby preventing \namplification from transcripts lacking the junction or from genomic DNA containing introns. \nPrimer lengths were typically 18–24 nucleotides, with melting temperatures (Tm) of approximately 58–64 \n°C and GC content between 40–60%. When sequence constraints limited GC content, primer length was \nadjusted to maintain appropriate Tm. Amplicon sizes were designed to range from 150 to 250 bp to \nensure efficient amplification and compatibility with both conventional RT-PCR and quantitative PCR. \nPrimer specificity was evaluated using BLAST against the reference genome (UCSC Genome Browser, \nUSA) and transcriptome to minimize off-target amplification. When multiple transcript variants shared \nsimilar junctions, additional primers targeting alternative junctions were designed to improve \ndiscrimination. All primers were synthesized by Integrated DNA Technologies (IDT, USA). \nPCR amplification and optimization \nPrimer melting temperatures were calculated using the NEB Tm Calculator (v1.16.10) with Phusion High-\nFidelity DNA Polymerase (HF buffer conditions) as the reference. The Tm values of the 5′-end and 3′-end \nsegments of each junction-spanning primer were calculated separately, along with the overall primer Tm, \nto guide primer design and PCR optimization. \nPCR amplification was performed using Phusion High-Fidelity PCR Master Mix (Thermo Fisher Scientific, \nUSA) or PyroMark PCR Master Mix (Qiagen, USA) according to the manufacturers’ instructions. PCR \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n4 \n \nconditions were optimized to ensure that the partial Tm values of junction-spanning primers remained \nbelow the annealing temperature, thereby minimizing nonspecific amplification. \nSanger sequencing validation \nPCR products corresponding to each transcript variant were validated by Sanger sequencing at the Johns \nHopkins Genetic Resources Core Facility. PCR products were submitted without purification and \nsequenced using standard protocols. Sequencing chromatograms were analyzed and aligned to \nreference sequences to confirm the identity of the amplified products. This validation step verified that \neach amplicon corresponded precisely to the intended exon–exon junction, confirming the specificity \nand reliability of the primer design and amplification strategy. \nResults \nMelting temperature–guided exon–exon junction primer design enables RNA variant–specific RT-\nPCR \nTo enable transcript variant–specific detection in cases where variants lack large distinguishable exonic \nregions, we developed a uni-directional primer design strategy based on a dual-recognition mechanism \nat exon–exon junctions (Figure 1). In this approach, a single primer (forward or reverse) spans a splice \njunction, with approximately half of its sequence complementary to the upstream exon and the \nremaining half complementary to the downstream exon. This configuration requires simultaneous \nannealing to both exons in the correct adjacency, thereby conferring junction-level specificity. \nA key feature of this strategy is the thermodynamic discrimination between partial and full primer \nbinding. The melting temperature (Tm) of each half-primer segment was deliberately designed to be lower \nthan the PCR annealing temperature (typically 58–64 °C), generally by >7 °C. Under these conditions, \npartial annealing to a single exon is thermodynamically unstable and does not support productive \nextension. In contrast, when the primer encounters the correct exon–exon junction, both halves anneal \ncooperatively to form a stable duplex, enabling efficient amplification. This design suppresses \namplification from transcripts sharing only a single exon as well as from genomic DNA containing intronic \nsequences. \nAs illustrated in Figure 1A, conventional primers targeting shared exons (Exon A and Exon B) amplify \nmultiple transcript variants due to identical target exon sequences, resulting in poor specificity. In \ncontrast, the exon–exon junction strategy (Figure 1B) restricts amplification to transcripts containing the \nprecise exon combination recognized by the junction-spanning primer. Variants lacking the correct exon \nconnectivity fail to support full primer annealing and are therefore not amplified, enabling clear \ndiscrimination among closely related transcripts. \nWe applied this strategy to the lncRNA HTRA1-AS1 locus, which contains multiple transcript variants with \nhighly overlapping exon structures but distinct exon connectivity patterns (Figure 1C). Because these \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n5 \n \nvariants lack large distinguishable exonic regions, conventional approaches cannot distinguish them. By \ndesigning primers targeting junctions to each splice configuration, we achieved selective amplification of \nindividual variants. The inset in Figure 1C highlights the positioning of junction-spanning primers, \nillustrating how variant specificity is achieved through differential exon pairing. \nFurther optimization focused on balancing primer composition to minimize nonspecific extension. \nIncreasing sequence contribution toward the 5′ region improved specificity by reducing spurious priming \nfrom the 3′ end, which is critical for polymerase extension. Through optimization, we established design \nparameters that maintain both specificity and amplification efficiency. \nUsing these principles, we generated a panel of uni-directional exon–exon junction primers targeting \ndistinct splice junctions. Four forward primers were designed to recognize different exon–exon \nboundaries and paired with a common reverse primer located in a shared downstream exon. This \nmodular configuration enables scalable detection in which amplification is strictly dependent on correct \nexon adjacency. Consequently, only transcript variants containing the targeted junction are efficiently \namplified, whereas variants sharing partial sequence identity are excluded. \n \nFigure 1. Exon–exon junction primers enable transcript variant–specific detection.(A) Conventional method: Primers \ndesigned within shared exons (Exon A and Exon B) amplify multiple transcript variants due to identical exon sequences, \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n6 \n \nresulting in non-specific detection.(B) Exon–exon junction method: A junction-spanning primer is designed across the exon–\nexon boundary. Amplification occurs only when the correct exon combination is present, enabling selective detection of the \ntarget variant, while mismatched variants are not amplified.(C) Application to HTRA1-AS1 four transcript variants: Schematic \nof multiple variants lacking  exons. Junction-specific primer design targets distinct exon connectivity (highlighted region), \nallowing discrimination of closely related transcript variants. The inset illustrates primer positioning across exon–exon \njunctions for variant-specific amplification. \nThe calculated melting temperatures for each primer half and the full-length primers are summarized in \nTable 1, providing quantitative support for the thermodynamic basis of the design. \nTable 1 Exon-Exon junction Primers and their Tm values \n \nNote: Primer sequences are listed in the 5′→3′ direction. Forward primers (F1–F3) span exon–exon junctions, with ratios \nindicating the nucleotide contributions from each exon. Product size denotes the expected amplicon length. Partial Tm refers \nto the melting temperature of each exon-matching segment, while total Tm indicates the overall primer melting temperature. \nGC% represents GC content. Reverse primers (R) are shared within each transcript group. “*” indicates that the Tm is not \navailable because the primer length is shorter than the minimal functional length. ENST415, ENST416, and ENST969 are \nabbreviations for ENST00000811415.1, ENST00000811416.1, and ENST00000647969.1, respectively \nPCR polymerase selection and PCR optimization  \nPCR polymerase selection was critical for achieving robust and specific amplification using junction-\nspanning primers. We evaluated several commonly used PCR polymerases under identical cycling \nconditions and observed substantial differences in amplification efficiency and specificity. High-fidelity \npolymerases with optimized buffer systems consistently produced stronger and more reproducible \nproducts, whereas some enzymes yielded weak bands or exhibited nonspecific amplification. Notably, \nperformance differences were evident even between similar Phusion High-Fidelity PCR Master Mix \nformulations obtained from different commercial suppliers. Based on these comparisons, Thermo Fisher \nPhusion Flash High-Fidelity PCR Master Mix and Qiagen 2× PyroMark PCR Master Mix (containing \nHotStarTaq DNA polymerase and 3 mM Mg²⁺) were selected for subsequent experiments. Both systems \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n7 \n \ngenerated clear, specific bands with minimal background, demonstrating their suitability for exon–exon \njunction PCR (data for other polymerases not shown). \nThese findings underscore the importance of polymerase selection for the reliable detection of transcript \nvariants lacking large distinguishable exonic regions. While predicted melting temperatures (Tm) provide \na useful starting point for primer design and initial annealing temperature selection, the optimal \nannealing temperature must be determined empirically. Annealing conditions have a strong impact on \nassay performance and are also dependent on the PCR polymerase used. For example, the optimal \nannealing temperature for Phusion Flash High-Fidelity PCR Master Mix differs from that of PyroMark PCR \nMaster Mix (Supplementary Data 1–3). \nValidation of variant-specific amplification by Sanger sequencing \nUsing the optimized PCR conditions, each exon–exon junction primer set generated a single, discrete \nband at the expected size for its corresponding transcript variant (Figure 2A, C). No additional bands or \nsmearing were observed, indicating high amplification specificity and minimal nonspecific products. \nImportantly, primer sets targeting different splice junctions produced mutually exclusive amplification \npatterns across variants, consistent with junction-dependent recognition. The observed amplicon sizes \n(~170–213 bp) matched the predicted lengths based on primer design and exon connectivity, further \nsupporting accurate target amplification. \nTo rigorously validate specificity at the sequence level, PCR products were directly subjected to Sanger \nsequencing without gel purification, providing a stringent test of amplification fidelity. Sequencing \nchromatograms obtained using reverse primers (complementary to the forward strand) exhibited clean, \nhigh-quality peaks with minimal background noise (Figure 2B, D). Critically, the sequences spanning the \namplified regions precisely matched the designed exon–exon junctions, with the junction boundaries \nclearly identifiable at the expected positions (indicated by arrows). No evidence of mixed sequencing \nsignals or off-target amplification was detected, indicating that a single dominant product was generated \nin each reaction. \nFor all tested transcript variants (ENST415, ENST416, ENST969, and HTRA1-AS1), the sequencing results \nconfirmed that amplification occurred only when the correct exon adjacency was present, consistent \nwith the double-recognition mechanism of the uni-directional junction primers. Variants lacking the \ntargeted exon–exon junction did not yield detectable products, demonstrating effective discrimination \namong closely related variants that share common exon sequences.  Only the ENST415-specific forward \nprimer F3, when paired with the common reverse primer (F3R), produced a specific PCR band after \noptimization, similar to the HTRA1-AS1 forward primer F1 combined with the common reverse primer \n(F1R). In contrast, all three primer pairs performed efficiently for ENST416 and ENST969.  \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n8 \n \n \nFigure 2. Variant-specific amplification of HTRA1-AS1 transcripts using exon–exon junction primers. (A, C) Agarose gel electrophoresis \nshowing PCR products generated with different primer combinations (F1R, F2R, F3R). Clear bands at the expected sizes (~170–213 bp) are \nobserved only for the correct variant, with minimal non-specific amplification. DNA ladders are included for size reference, with the ~500 bp \nband indicated. (B, D) Sanger sequencing confirms that the PCR products correspond to the expected exon–exon junction sequences. Red \narrows mark the precise junction sites. Yellow and blue highlighted regions indicate the primer-binding sequences (including reverse-\ncomplement alignment), demonstrating accurate primer–template pairing and variant-specific amplification. ENST415, ENST416, and \nENST969 are abbreviated identifiers for the lncRNA transcript variants ENST0000811415.1, ENST0000811416.1, and ENST0000647969.1, \nrespectively \nExon–exon junction RT-PCR reduces heteroduplex and quadruplex formation in the simultaneous \ndetection of multiple gene transcripts \nFour HTRA1-AS1 variants (ENST415, ENST416, ENST969, and HTRA1-AS1) were initially co-amplified \nusing a common primer set to assess amplification specificity. Agarose gel electrophoresis revealed \nthree bands corresponding to ENST416 (305bp), HTRA1-AS1 (481bp), and an unexpected intermediate \nband (about 400bp) between them (Figure 3A). Sanger sequencing of the gel-extracted intermediate band \nconfirmed that it represented a heteroduplex composed of hybrid strands derived from ENST416 and \nHTRA1-AS1 (Figure 3C). \nTo exclude the possibility that this band resulted from co-migrating independent amplicons, gel purified \nENST416 and HTRA1-AS1 PCR products were subjected to denaturation (98 °C) followed by reannealing \n(63 °C). Individual samples produced a single band, whereas the mixed sample generated an additional \nband (~400 bp), consistent with heteroduplex formation between the two variants and even a higher \nmolecular weight heteroquadruplex band (Figure 3B). \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n9 \n \n \nFigure 3. Heteroduplex and quadruplex formation during detection of RNA variants using conventional RT-PCR. (A) Co-\namplification of four HTRA1-AS1 variants (ENST415, ENST416, ENST969, and HTRA1-AS1) using a common primer set \nproduces multiple bands on agarose gel electrophoresis, including expected products and an unexpected intermediate band \n(red arrow), suggesting nonspecific hybridization. (B) Validation of heteroduplex formation by denaturation and reannealing. \nAfter denaturation and reannealing single bands together, which individually ENST416 and HTRA1-AS1 products, additional \nhigher-molecular-weight bands (red arrows) were generated, consistent with heteroduplex formation between the two \nvariants. Amplicon sizes are indicated below. (C) Sequence alignment and Sanger sequencing traces of the heteroduplex \nproduct generated using forward and reverse primers. Overlapping chromatogram peaks and misaligned base calls confirm \nthe presence of hybrid DNA strands derived from ENST416 and HTRA1-AS1. A schematic model illustrates heteroduplex \nformation through partial base pairing between homologous regions of the two transcripts. \nAs illustrated in Figure 4, partial sequence homology between amplicons of different lengths promotes \nthe formation of heteroduplex and higher-order structures (quadruplex) during re-annealing. PCR \nproducts or these intermediates can further assemble into higher-order structures, including loop-\ncontaining quadruplexes and multi-stranded quadruplexes, which contribute to aberrant bands observed \non agarose gels. The presence of unpaired regions and overlapping homologous segments facilitates \nthese interactions, highlighting how shared sequence regions between PCR products drive heteroduplex \nand quadruplex formation. \n \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n10 \n \nFigure 4. Potential hybrid structures formed between ENST416 and HTRA1-AS1 PCR products. \nSchematic illustrating the formation of hybrid DNA structures between ENST416 (311 bp) and HTRA1-AS1 (487 bp) amplicons \ncontaining partially homologous regions. Upon denaturation and re-annealing, the shared sequences enable intermolecular \nbase pairing, leading to heteroduplex formation with single-stranded loop regions corresponding to non-homologous \nsegments. Two possible higher-order structures are proposed: (Type 1) a cruciform-like heteroduplex formed through \ncrosswise pairing of homologous regions, and (Type 2) a hetero-quadruplex-like assembly involving multistrand interactions \nand circularized pairing of complementary segments. Arrows indicate strand orientation, and dashed lines represent base-\npaired regions. These structures may contribute to heteroduplex DNA artifacts observed during PCR amplification. \nWe next evaluated strategies to reduce heteroduplex or quadruplex formation. An additional primer set \n(P1) was used to selectively co-amplify ENST416 and HTRA1-AS1 variants, resulting in the formation of \nheteroduplex products alongside the two variant amplicons. To minimize extensive homologous regions, \na second primer configuration (P2) was designed using a common forward primer and two distinct \nreverse primers—including an exon–exon junction primer—that uniquely anneal to each variant \n(Supplementary Data 2). This design excludes the potential for loop formation between variants. \nHeteroduplex band was still observed with the P2 primer combination; however, their intensity was \nreduced (Figure 5). \nThus, exon–exon junction RT-PCR provides greater flexibility in primer design, enabling reduction of \nheteroduplex formation by limiting amplification of homologous regions between variants (Figure 5A–C).\n \nFigure 5. Exon–exon junction RT-PCR reduces heteroduplex formation. (A–C) Agarose gel electrophoresis showing co-\namplification of ENST416 and HTRA1-AS1 using different primer designs. (A) Conventional primer set (P1) generates strong \nheteroduplex formation, evident as an additional band between the expected amplicons. (B) Use of variant-specific reverse \nprimers (P2) reduces heteroduplex formation, although intermediate bands remain detectable. (C) Exon–exon junction–\ntargeted primer design (P3) markedly suppresses heteroduplex formation, yielding predominantly distinct bands \ncorresponding to each variant. Amplicon sizes and shared sequence lengths between variants are indicated below each \npanel. (D) Schematic representation of primer design strategies and their effects on heteroduplex formation. P1 employs \nshared primer binding regions, promoting hybridization between partially homologous amplicons. P2 introduces variant-\nspecific primers to reduce overlap. P3 incorporates an ENST416-specific exon–exon junction forward primer, which limits \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n11 \n \nextension from mismatched templates and disrupts hybrid strand formation. The reduced sequence complementarity and \njunction-specific targeting together minimize heteroduplex formation. \nCollectively, these results demonstrate that heteroduplex formation in conventional RT-PCR arises from \npartial sequence complementarity between co-amplified variants and can be substantially minimized by \nexon–exon junction–targeted primer design. This uni-directional EEJ strategy enables highly specific, \nsequence-validated detection of transcript variants, even when conventional exon-targeting approaches \nfail to resolve closely related splice isoforms. \nDiscussion \nAccurate detection of transcript variants is essential for understanding gene regulation and transcript-\nspecific biological functions. Alternative splicing greatly expands transcriptomic diversity in eukaryotic \ngenomes, with most human genes generating multiple RNA variants [1–3]. However, many variants share \nhighly similar exon sequences and differ only in exon connectivity, making variant-specific detection \nchallenging using conventional PCR strategies that rely on exon-targeting primers [5]. In addition to \nlimited specificity, co-amplification of closely related transcripts frequently leads to heteroduplex DNA \nformation, complicating gel interpretation and reducing assay reliability. \nIn this study, we present a simple and effective PCR-based approach that distinguishes transcript \nvariants by targeting exon–exon junctions unique to individual splice configurations. Because splice \njunctions are generated only after RNA processing, primers spanning these junctions enable selective \namplification of specific variants even in the absence of large distinguishable exonic regions. This \nstrategy achieves transcript-level specificity using standard RT-PCR or qPCR workflows [5,9]. \nA central feature of this method is the use of Tₘ-guided, uni-directional exon–exon junction primers, in \nwhich approximately half of the primer sequence anneals to each of two adjacent exons. By designing \neach half-primer with a melting temperature lower than the PCR annealing temperature, partial annealing \nto a single exon is thermodynamically unstable. Efficient amplification therefore occurs only when both \nhalves simultaneously bind across the correct exon boundary. This double-recognition mechanism \nenhances specificity, minimizes cross-amplification, and functionally suppresses extension from \nmismatched templates. \nImportantly, this thermodynamic constraint also reduces heteroduplex and quadruplex DNA formation. \nConventional RT-PCR often generates heteroduplex and quadruplex products when partially homologous \namplicons reanneal, resulting in intermediate bands. In contrast, the EEJ primer design limits the \nproduction of overlapping amplicons and reduces effective cross-hybridization including heteroduplex \nand non-G-heteroquadruplex between variants, thereby improving band clarity and interpretability \n[12,13]. This effect likely reflects a reduction in the length and/or proportion of homologous pairing \nregions (Figure 5A, 5C). Consistent with this, our experiments show that junction-targeted primers \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n12 \n \nmarkedly decrease heteroduplex formation compared with conventional primer designs, particularly in \nmultiplex settings involving closely related transcripts. \nThe EEJ method is experimentally accessible and readily adaptable. Primer design follows standard PCR \nprinciples, and assay validation can be performed using routine RT-PCR and Sanger sequencing. \nCompared with transcriptome-wide approaches such as long-read RNA sequencing, this strategy is cost-\neffective and requires minimal computational resources and can be easily implemented in most \nmolecular biology laboratories [6,7,14,15]. Thus, exon–exon junction RT-PCR provides a practical tool for \nvalidating transcript structures predicted by RNA-seq and for monitoring specific splice variants under \ndefined experimental conditions. \nSeveral considerations are important for optimal performance. First, accurate annotation of exon–exon \njunctions is essential for primer design. Second, PCR polymerase selection and reaction optimization \ncan significantly influence amplification efficiency across splice junctions. Third, transcript variants that \nshare identical exon–exon junctions cannot be distinguished by this approach alone and may require \ncomplementary methods, such as long-read sequencing, length-dependent PCR, or rapid amplification \nof cDNA ends (RACE) [14,15]. \nIn summary, the Tₘ-guided exon–exon junction primer strategy provides a robust, specific, and \nexperimentally straightforward approach for distinguishing closely related RNA variants. By enabling \nvariant-specific detection without reliance on large distinguishable exonic regions and by minimizing \nheteroduplex formation, this method expands the experimental toolkit for studying alternative splicing \nand transcript diversity, particularly in complex loci such as long noncoding RNAs [16,17]. \nFunding \nThis work was generously supported by grants from the Gilbert Family Foundation, National Institutes of \nHealth (P30 EY001765) \nAuthor contributions: \nInitiative and conceptualization: PWZ  \nMethodology and analysis design: JN, PWZ \nData collection and analysis: JN, PWZ, DJZ \nWriting: JN, PWZ, DJZ \nCompeting interests:  \nAuthors declare they have no competing interests. \n \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n13 \n \nReferences \n1. Pan Q, Shai O, Lee LJ, Frey BJ, Blencowe BJ. Deep surveying of alternative splicing complexity in the \nhuman transcriptome by high-throughput sequencing. Nat Genet. 2008;40(12):1413-1415. \n2. Wang ET, Sandberg R, Luo S, Khrebtukova I, Zhang L, Mayr C, et al. Alternative variant regulation in \nhuman tissue transcriptomes. Nature. 2008;456(7221):470-476. \n3. Lee Y , Rio DC. Mechanisms and regulation of alternative pre-mRNA splicing. Annu Rev Biochem. 2015; \n84:291-323. \n4. Shila S, Dahiya V , Hisle C, Bahadursingh E, Thiyagarajan R, Fields PE, Rumi MAK. mRNA Isoforms and \nVariants in Health and Disease. Int J Mol Sci. 2025 Sep 25;26(19):9356. doi: 10.3390/ijms26199356. \nPMID: 41096623; PMCID: PMC12524470. \n5. Vo K, Shila S, Sharma Y , Pei GJ, Rosales CY , Dahiya V , Fields PE, Rumi MAK. Detection of mRNA \nTranscript Variants. Genes (Basel). 2025 Mar 16;16(3):343. doi: 10.3390/genes16030343. PMID: \n40149494; PMCID: PMC11942493. \n6. Ozsolak F , Milos PM. RNA sequencing: advances, challenges and opportunities. Nat Rev Genet. \n2011;12(2):87-98. \n7. Stark R, Grzelak M, Hadfield J. RNA sequencing: the teenage years. Nat Rev Genet. 2019;20(11):631-\n656. \n8. Thornton B, Basu C. Real-time PCR (qPCR) primer design using free online software. Biochem Mol Biol \nEduc. 2011;39(2):145-154. \n9. Nolan T, Hands RE, Bustin SA. Quantification of mRNA using real-time RT-PCR. Nat Protoc. \n2006;1(3):1559-1582. \n10. Monfort-Lanzas P , Rusu EC, Parrakova L, Karg CA, Kernbichler DE, Rieder D, Lackner P , Hackl H, \nGostner JM. ExonSurfer: a web-tool to design primers at exon-exon junctions. BMC Genomics. 2024 Jun \n12;25(1):594. doi: 10.1186/s12864-024-10456-2. PMID: 38867172; PMCID: PMC11170769. \n11. Govindkumar B, Kavyashree B, Patel K, Sasidharan K, Siva Arumugam T, Thomas L, Praveena BKG, \nRaksha HN, Menon R, Acharya KK. Ex-Ex Primer: An experimentally validated tool for designing \noligonucleotides spanning spliced nucleic acid regions from multiple species. J Biotechnol. 2022 Jan 10; \n343:1-6. doi: 10.1016/j.jbiotec.2021.10.009. Epub 2021 Oct 29. PMID: 34756973. \n12. Tsai YC, Qi H, Lin CP , Lin RK, Kerrigan JE, Rzuczek SG, LaVoie EJ, Rice JE, Pilch DS, Lyu YL, Liu LF . A G-\nquadruplex stabilizer induces M-phase cell cycle arrest. J Biol Chem. 2009 Aug 21;284(34):22535-43. doi: \n10.1074/jbc.M109.020230. Epub 2009 Jun 16. PMID: 19531483; PMCID: PMC2755660. \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint \n\n14 \n \n13. Liu S, Peng P , Wang H, Shi L, Li T. Thioflavin T binds dimeric parallel-stranded GA-containing non-G-\nquadruplex DNAs: a general approach to lighting up double-stranded scaffolds. Nucleic Acids Res. 2017 \nDec 1;45(21):12080-12089. doi: 10.1093/nar/gkx942. PMID: 29059300; PMCID: PMC5716147. \n14. Tilgner H, Grubert F , Sharon D, Snyder MP . Defining a personal, allele-specific, and single-molecule \nlong-read transcriptome. Proc Natl Acad Sci U S A. 2014;111(27):9869-9874. \n15. Byrne A, Beaudin AE, Olsen HE, Jain M, Cole C, Palmer T, et al. Nanopore long-read RNA sequencing \nreveals widespread transcriptional variation among the surface receptors of individual B cells. Nat \nCommun. 2017; 8:16027. \n16. Mercer TR, Dinger ME, Mattick JS. Long non-coding RNAs: insights into functions. Nat Rev Genet. \n2009;10(3):155-159. \n17. Derrien T, Johnson R, Bussotti G, Tanzer A, Djebali S, Tilgner H, et al. The GENCODE v7 catalog of \nhuman long non-coding RNAs: analysis of their gene structure, evolution, and expression. Genome Res. \n2012;22(9):1775-1789. \nremix, or adapt this material for any purpose without crediting the original authors. \npreprint (which was not certified by peer review) in the Public Domain. It is no longer restricted by copyright. Anyone can legally share, reuse, \nThe copyright holder has placed thisthis version posted April 5, 2026. ; https://doi.org/10.64898/2026.04.02.716213doi: bioRxiv preprint","source_license":"Public-Domain","license_restricted":false}