Ultrasensitive cDNA Library Preparation for Next-generation Sequencing of MicroRNAs from Small Extracellular Vesicles

Summary

Extracellular vesicles (EVs) contain cell-specific microRNAs that regulate recipient cells, the identification of which may shed light on their function and role as biomarkers. Our optimized cDNA library preparation protocol introduces unique barcodes that enable sample multiplexing and enhanced processing of low-input EVs while using paired-end dual index barcodes for compatibility with Illumina sequencers.

Abstract

Recent studies demonstrate that small extracellular vesicles (sEVs), which are found in all biofluids, play critical roles in intercellular communication by channeling proteins, DNA, and RNAs. MicroRNAs (miRNAs) that are packaged in sEVs have emerged as critical deliverable regulators in recipient cells. Since sEVs secreted by normal and diseased cells carry different miRNA cargos, recent sEV-miRNA profiling studies suggest that they may help identify novel circulating biomarkers. However, cell/disease-specific sEVs circulating in diverse biofluids, once isolated, provide low miRNA quantities, which are generally difficult to quantify using conventional spectrometric methodologies. Small non-coding RNA Next Generation Sequencing (NGS), which allows for the amplification of cloned miRNA sequences, offers a valuable opportunity to evaluate the miRNA cargos of sEVs. Unfortunately, commercial cDNA library preparation procedures often require RNA inputs well above the unquantifiable amounts available from isolated sEVs.

Thus, considering the robustness and multiplexing capabilities of our existing cDNA library preparation procedure (i.e., initially optimized for the analysis of low-input, highly degraded, formalin-fixed paraffin-embedded (FFPE) RNA), we sought to evaluate its applicability for the analysis of sEV miRNAs. Importantly, taking into account the recent technical clustering improvements of sequencing chips, we sought to adapt our transcript barcoding approach within a paired-end, dual index-compatible cDNA library preparation workflow to enhance our sequencing and multiplexing capabilities. Using RNA extracted from 8.4 × 10⁹ sEVs in 16 replicates, and from decreasing amounts of sEVs from 10¹⁰ sEVs to as low as 2.5 × 10⁷ sEVs, we evaluated the reproducibility and sensitivity of this methodology. The data demonstrate that the 16 3' adenylated DNA barcodes allow for highly reproducible and sensitive detection of sEV-miRNA profiles across repeats using as low as 3.15 pg of total small non-coding RNAs or 1.35 pg of miRNAs.

Introduction

Small extracellular vesicles (sEVs) are nanosized (~30-200 nm in diameter) cell-derived particles enveloped by a phospholipid bilayer membrane, which is inherited from their cell of origin and that robustly protects their molecular cargos^1,2. It is well accepted that virtually all cells produce and release sEVs into the intercellular compartment, which can in turn be detected in most biofluids (i.e., blood, urine, saliva, etc)^2,3,4. Recent studies have demonstrated that the stably encapsulated molecular cargos (i.e., DNA, RNA, proteins, lipids, etc)^5,6,7 of sEVs, once in contact or delivered to recipient cells mediate intercellular communication⁸. Current research efforts have been focused on improving the isolation of cell-specific sEVs from diverse biofluids, to precisely identify and monitor these molecular cargos^9,10. Considering that sEVs contain small non-coding RNAs and particularly microRNAs (miRNAs), which retain their regulatory function upon cellular delivery, many studies have been focused on evaluating the utility of miRNAs as biomarkers that can reveal the condition of their cell of origin^11,12,13.

MicroRNAs (miRNAs) represent a large class (~2,000 known in humans) of small non-coding RNAs with sizes ranging between 19 and 25 nucleotides (nt), which bind to imperfect complementary sites in the 3' untranslated regions of their mRNA targets and direct their degradation and/or post-transcriptional repression^14,15. Functionally, miRNAs have been described to control many biological processes, and the deregulation of their expression has been associated with alterations in molecular, biochemical, and physiological processes that contribute to the initiation and development of diseases, including cancer^16,17,18. Importantly, studies have shown that diseased cells and particularly cancer cells, not only differentially express miRNAs in comparison to normal cells¹⁸, but that their packaging into sEVs also differs^19,20.

While several studies focus on establishing the molecular processes and pathways that direct the packaging of sEVs in normal and diseased cells^19,21,22, biomarker studies are currently focused on identifying distinguishable small non-coding RNA and/or miRNA signatures that are packaged and secreted by specific diseased/cancer cells via sEVs. The targeted isolation of diseased cell-specific sEVs from diverse human biofluids and the evaluation of their multi-omic molecular cargos may thus enable the development of diagnostic and prognostic assays for non-invasive detection of diverse human diseases and cancers^23,24,25,26. Although sEV-miRNAs are the most studied non-coding transcripts, other small non-coding RNA species (i.e., miRNA isoforms (isomiRs), piwiRNAs, transfer RNA fragment (tRFs), rRNAs, orphan non-coding RNAs (oncRNAs)…) are also being evaluated for their potential usefulness as circulating sEV-biomarkers, with specific emphasis on the detection of different cancers^{27,28,29,30,31,32}.

Advantageously, next-generation sequencing (NGS) analysis of small non-coding RNA transcripts (including miRNAs), following their barcoding and cDNA library preparation, is ideal for exploration of known and/or unknown small non-coding RNA species contained in circulating sEVs, which may be associated with diseases²⁷. Indeed, sEV small-RNA NGS provides a high-throughput approach for the discovery of cell-specific small non-coding RNA transcripts that would otherwise remain unknown, if globally evaluated with target-specific technologies (i.e., multiplex PCR, microarrays, customized panels, etc.)^32,33,34. Considering that we previously optimized an ultra-sensitive, highly reproducible cDNA library preparation protocol, which we developed for the analysis of highly degraded and low-concentration formalin-fixed paraffin-embedded (FFPE) RNAs^35,36, we sought to adapt it for the analysis of small non-coding RNAs and miRNAs encapsulated within sEVs^37,38,39,40.

With the discontinuation of the Illumina HiSeq2500 instrument, which provided us with years of high-quality single-end dual index miRNA analysis of sEV cargos^37,40, we sought to explore the reproducibility and sensitivity of our protocol when adapted to updated sequencing instruments using now paired-end dual index chemistry. With the intent to retain the multiplexing capabilities of the original protocol, we maintained the existing 16 barcodes for the preparation of small-RNA cDNA transcripts before integrating them within paired-end dual index chemistry. Using only a pair of i5 and i7 barcodes, we enable the preparation of a robust laboratory-based cDNA library where simultaneous processing and analysis of up to 16 individual samples improve analytical reproducibility of low-input sEV RNA. Thus, we present the biochemical processes, the updated barcodes, PCR and size marker primers, and isolation steps necessary to generate highly reproducible cDNA libraries using ultra-low amounts of miRNAs isolated from human plasma sEVs.

Protocol

1. All reagents and oligonucleotides are prepared as follows:

NOTE: All nucleotides used in this protocol are detailed in Figure 1 and Supplemental File 1 at concentrations that are used for implementing this protocol. See Supplemental File 2 for primer sequences.

1. Prepare a stock solution of calibrator cocktail, which will be used in each individual ligation as a control for the different enzymatic reactions:
   1. Resuspend the Carrier oligonucleotide (0.5 µM) (Supplemental File 1) with RNase-free water.
   2. Combine the 10 calibrators (10 µL each, Supplemental File 1) into a single 1.5 mL tube to reach a 100 µM stock, dilute it using the carrier oligonucleotide to reach the 0.0052 nM solution, and store at -20 °C.
2. Resuspend each of the 17 unique (16 experimental and 1 test barcoded adapter) adenylated 3' adapters (Figure 1A) with nuclease-free water at a concentration of 50 µM.
   NOTE: Each adapter is aliquoted into individual tubes containing 2.5 µL for storage at -80 °C for up to 2 years, to prevent freeze-thaw cycles of the larger stock.
3. Resuspend the different experimental RNA oligonucleotides, including the RNA extraction carrier (Supplemental File 1), the RNA 3'-ligation carrier (Supplemental File 1), the RNA ligation test (Supplemental File 1), and the 3' ligation (Supplemental File 1; 19 nt-3' adapter and 24 nt-3'adapter) and 5' ligation (Supplemental File 1; 5'adapter-19 nt-3'adapter and 5'adapter-24 nt-3'adapter) DNA size markers, to 250 ng/µL.
4. Dilute the HPLC-purified 5' adapter to 100 µM with nuclease-free water (Figure 1A).
5. Resuspend the RT and first PCR 5' and 3' RT/PCR primers to 100 µM with nuclease-free water (Figure 1B).
6. Resuspend the second PCR 5' and 3' PCR i7 and i5 oligonucleotides to 100 µM with nuclease-free water (Figure 1B).
   NOTE: Aliquot each DNA primer for up to 10 experiments and store at -80 °C.
7. Prepare PolyAcrylamide denAturing (PAA) gel loading dye, mix, and aliquot 1 mL into tubes and store at -80 °C.
   1. Prepare a solution of 0.5 M Na₂H₂ EDTA by adding 18.6 g of Na₂H₂ EDTA to 50 mL of nuclease-free water. Add NaOH pellets slowly to the prepared Na₂H₂ EDTA solution to reach a pH of 8.0. Make up the volume up to 100 mL using ddH₂O to get a 0.5 M Na₂H₂ EDTA (pH 8.0) stock solution
   2. Weigh out 15 mg of Bromophenol Blue into a 15 mL tube, resuspend the powder in 600 µL of nuclease-free water, add 14.25 mL of deionized formamide. Finally, add 150 µL of a 0.5 M Na₂H₂ EDTA, pH 8.0 solution. Aliquot the final PAA solution into 1 mL tubes and store at -80 °C.
   3. Prepare the 5x agarose gel loading dye.
      1. Weigh and dissolve 1.86 g of Na₂H₂-EDTA in 50 mL of RNase-free water and adjust the pH to 8.0 by adding NaOH pellets. Add RNase-free water to 100 mL to get a 50 mM Na₂H₂ EDTA solution.
      2. To 5 mL of 50 mM Na₂H₂ EDTA, add 20 mg of Bromophenol Blue, 20 mg of Xylene Cyanol FF, and 2 g of Ficoll type-400. Vortex the tube containing the three dyes to mix and add an additional 5 mL of 50 mM Na₂H₂ EDTA. Aliquot individual 1 mL tubes of 5x agarose gel loading dye and store at -80°C.

2. Preparation of the EV RNA samples

1. Following the isolation of sEVs, evaluate their morphology using transmission electron microscopy (TEM; Figure 2), concentration using Nanoparticle Tracking Analysis (NTA; Figure 3), and surface protein distribution using super-resolution nanoimaging (Figure 4), or equivalent sEV quantification technology.
2. Aliquot an equivalent number of sEVs in different tubes, add RNase-A (12.5 µg/mL) and incubate at 37 °C for 1 h.
3. Initiate extraction of small non-coding RNAs, including miRNAs, from sEVs using the small-RNA extraction kit and initiate RNA phase isolation.
4. Separate the lower aqueous phase of the RNA lysis buffer, add it to a new tube, and then add 4 ng of RNA carrier oligonucleotide (Supplemental File 1) to each RNA sample, prior to conducting column isolation following the manufacturer's instructions.
5. Speed-vacuum each individual RNA sample to 9.5 µL to initiate the cDNA library preparation protocol.

3. Set up ligations of the 3' barcoded adapter with 16 RNA samples and 1 test sample

1. Identify 16 individual sEV-RNA specimens, aliquot them in 9.5 µL of nuclease-free water into 1.5 mL siliconized microcentrifuge tubes, and place them on ice at least 10 min before starting the experiments.
2. Prepare the 10x RNA Ligase Buffer (without ATP) fresh prior to the experiment.
   1. In a 1.5 mL siliconized tube, combine 343 µL of nuclease-free double-distilled water (ddH₂O), 500 µL of 1 M Tris pH 7.5, 100 µL of 1 M MgCl₂, 50 µL of 20 mg/mL Bovine Serum Albumin, and 7 µL of 14 M 2-mercaptoethanol.
   2. Mix the solution by flicking the tube, centrifuge for 2 s at 2,000 × g at room temperature (RT) on a tabletop centrifuge, and then set on ice.
      NOTE: All steps that describe a "centrifuge for 2 s" are all done with a microcentrifuge at RT with a speed no higher than 2,000 × g.
3. Defrost the 0.0052 nM calibrator cocktail by placing the tube on ice for 10 min.
4. Prepare the Ligation Master Mix for 18 reactions in a 1.5 mL siliconized tube by combining 40 µL of 10x RNA Ligase Buffer, 10 µL of 0.0052 nM calibrator cocktail, and 40 µL of nuclease-free ddH₂O.
5. Flick and spin the 1.5 mL tube and set it on ice.
6. Add 4.5 µL of the Ligation Master Mix to each of the 16 individually aliquoted sEV RNA samples, and the test ligation reaction containing 100 ng of RNA ligation test oligonucleotide (Supplemental File 1) previously aliquoted and at 9.5 µL volume. Flick to mix and spin the tubes before placing them back on ice.
7. Transfer the 16 samples and test reaction to a heat block for 1 min at 90 °C and transfer back to ice.
8. Individually transfer 4 µL of 50% PEG to each tube, flick to mix, and spin the tubes before placing them back on ice.
9. Defrost 2.5 µL aliquots from each of the 17 3' barcoded adapters (16 experimental (#1 to #16) and test adapter (#17)) and place on ice to defrost.
10. Transfer 1 µL of each of the 17 adapters to the corresponding 16 samples containing sEV-RNA and one test sample, with Ligation Master Mix (i.e., 4.5 µL), and 50% PEG (i.e., 4 µL).
    NOTE: All solutions and reactions are mixed by flicking the tubes and spinning them, rather than pipetting the solutions up and down, to prevent loss of material.
11. Prepare the diluted T4 RNA Ligase 2 Truncated K227Q ligation enzyme by adding 10 µL of the enzyme with 10 µL of nuclease-free ddH2O into a new 1.5 mL siliconized tube.
12. Transfer 1 µL of the diluted T4 RNA Ligase 2 Truncated K227Q ligation enzyme into each of the 17 tubes (16 experimental and 1 test RNA samples). Do not pipette up and down, change tips between each tube, keep tubes on ice, flick mix, centrifuge, and place the tubes on ice.
13. Set the ligations onto fresh ice in a tube holder in an ice bucket in the cold room, overnight for up to 18 h.

4. PAGE isolation of the 3' adapter ligated Small RNAs

1. Place the 17 tubes on a heat block at 90 °C for 1 min to deactivate the T4 RNA Ligase 2, and place on ice for 2 min to cool down.
2. Combine 1 µL of GlycoBlue and 26 µL of 5 M nuclease-free NaCl into a fresh tube and transfer 1.2 µL of this Precipitation solution into each of the 17 tubes.
3. Add 63 µL of 100% ethanol to each of the 17 tubes, close, flick, and spin down for 2 s before placing back on ice.
4. Combine the content of the 16 sEV RNA tubes into a single 1.5 mL siliconized tube, while keeping the test RNA ligation tube separate.
5. Invert the tube containing the combined samples 3x to mix, briefly centrifuge for 2 s, and set on ice for 60 min to precipitate.
6. Prepare a large (16 x 20 cm²) 15% polyacrylamide (PAGE) gel as follows:
   1. Siliconize (i.e., using silicone-based solution) and cast the short and long glass plates together with 0.1 cm spacers.
   2. Prepare the 15% PAGE gel mix by combining 9 mL of system diluent, 18 mL system concentrate, 3 mL of system buffer, 240 µL of APS (9%), and 12 µL of TEMED in a 50 mL conical tube.
   3. Quickly transfer the solution between the glass-plates using a 30 mL pipette. Before the gel polymerizes, add a 14-well comb (0.1 cm-thick), and set the gel standing upright for 30 min at RT.
7. Recover the tube with the combined RNA samples from ice and centrifuge it at 16,000 × g for 60 min at 4 °C.
8. Meanwhile, finish setting up the 15% PAGE by gently removing the comb, cleaning the wells with nuclease-free H₂O with a squirt bottle above a sink, before drying and setting up the 15% PAGE on the gel apparatus, where both the upper and lower reservoirs are filled with 0.5x Tris-Borate EDTA (TBE) solution, prior to a 30 min pre-run at 450 V.
9. Remove the tube from the centrifuge (Step 4.6) and carefully dry the RNA pellet using a Pasteur pipette with a 10 µL non-filter tip on its end connected to a vacuum system, without disturbing the pellet.
10. Resuspend the RNA pellet containing the precipitated 16 RNA ligations in 20 µL of nuclease-free ddH₂O and add 20 µL of PAA gel loading solution. Flick the tube to mix, centrifuge for 2 s and set on ice.
11. Replace 0.5x TBE of gel apparatus with fresh 0.5x TBE.
12. Set up a tube with 20 bp RNA size ladder, two tubes with each of the two 3'-ligation short DNA size markers (Supplemental File 1), and one tube containing the RNA ligation test oligonucleotide (i.e., ligated with Barcoded adapter #17).
13. To each of the tubes (i.e., ladder, 3' ligation short DNA oligos, and RNA ligation test), add 20 µL of PAA gel loading solution.
14. Set the 16 ligated RNA sample tube and 1 RNA ligation test tube at 90 °C for 1 min and set back on ice.
15. Replace 0.5x TBE buffer from both lower and upper reservoirs (i.e., add buffer below the level of well so there is no cross-contamination between wells during loading) and load the ladder on the ends of the gel, the RNA test ligation, the two short DNA size markers on both sides of the well containing the 16 combined 3' barcoded RNA ligations (see Figure 5).
16. Run the 15% PAGE gel with all samples for 90 min at 450 V (~35 mA) under cooled conditions (i.e., with fans blowing air on the glass plates).
17. Remove 15% PAGE from glass and lightly spray it with SYBR Gold solution (10 µL SYBR Gold in 25 mL of 0.5x TBE) and let it sit tilted for 5 min in the dark.
18. Set the gel on a blue-light transilluminator (Figure 5) and align both 19 nt-3' adapter and both 24 nt- 3' adapter size marker short DNA oligonucleotides with a ruler and excise the area of the gel containing the ligated miRNA constructs (Figure 6 displays the 3' ligation and resulting construct). Separately, excise the upper band of the RNA test ligation, which will also be used for a test 5' ligation the following day.
19. Place the two separate excised gel pieces (i.e., for 16 RNA samples and for test ligation) into different 0.5 mL gel breaker tubes set into 1.5 mL siliconized tubes, centrifuge at 16,000 × g for 3 min at RT and then resuspend the two set fragmented gel pieces with 300 µL of 400 mM NaCl solution, close the tubes, and seal them with parafilm.
20. Set the tubes on a shaker with agitation at 1,100 rpm at 4 °C, overnight (16-17 h).

5. Ligation of the 5' adapter

1. Transfer two separate fragmented gels and their solutions (16 x 3' barcoded samples (barcodes #1 to #16) and 3' barcoded test RNA (barcode #17)) onto two separates 5 µm filter tubes, each set into 1.5 mL siliconized tubes, seal them with parafilm, and centrifuge for 5 min at 2,300 × g at RT.
2. Add 950 µL of 100% ethanol to each filtered solution, close the tubes, invert to mix, spin down for 2 s, seal the tubes with parafilm, and set them on ice for 1 h.
3. Freshly prepare a 50% aqueous DMSO stock by combining 1 mL of nuclease-free water with 1 mL of DMSO, wrap the tube in aluminum foil, and store at RT in the dark.
4. Prepare a 15% PAGE gel and pre-run it, following the steps described in section 4.6.
5. Centrifuge the tubes containing the 3' ligated 16 RNA samples and the test RNA ligation at 16,000 × g for 60 min at 4 °C to precipitate the ligated RNAs.
6. Carefully remove the supernatant without touching the RNA pellets, and vacuum dry.
7. Resuspend the 3' ligated pellet of 16 RNA samples and test RNA each in 8 µL of nuclease-free ddH₂O without pipetting up and down. Lightly flick the tubes to mix, centrifuge for 2 s (2,000 × g) at RT, and set the tubes on ice.
8. Prepare the 10x RNA Ligase Buffer (with ATP) by mixing 500 µL of 1 M Tris (pH 7.5), 100 µL of 1 M MgCl₂, 50 µL of 20 mg/mL acetylated BSA, 200 µL of 10 mM ATP, 7 µL of 2-mercaptoethanol 14 M, and 143 µL of RNase-free water into a 1.5 mL tube.
9. Add 2 µL of 10x RNA ligase buffer (i.e., with ATP), 1 µL of 100 µM 5' adapter, and 4 µL of 50% aqueous DMSO to each tube. Flick to mix, spin for 2 s at RT, then set up the tube at 90 °C for 1 min to denature the RNA, and place it back on ice.
10. Add 3 µL of 50% DMSO, flick to mix, spin to pellet, add 3 µL of T4 RNA Ligase to each tube, flick the tubes, centrifuge for 2 s at RT, and set the tubes on the shaker at 37 °C for 75 min.
    NOTE: Do not pipette up and down.
11. Prepare the 20 nt DNA ladder and two separate tubes containing the 5'-ligation size marker long DNA oligonucleotides (Supplemental File 1; 5'adapter-19 nt-3' adapter, and 5'adapter-24 nt-3'adapter), and add 20 µL of PAA to each of the tubes.
12. Once the 5' adapter ligation is complete, add 20 µL of PAA to each of the two ligations (i.e., 16 RNA samples and test RNA), mix by flicking, spin down, and set at 90 °C for 1 min, after which place the tubes back onto ice.
13. Empty upper and lower reservoirs of gel apparatus and add fresh 0.5x TBE solution below the level of the wells of the upper reservoir.
14. Load the ladder on the gel's end, the 5' ligated test RNA sample, the two size-marker long DNA oligonucleotides (i.e., 5' adapter-19 nt-3' adapter, and 5' adapter-24 nt-3' adapter) in wells surrounding the well where the 5' ligated 16 RNA samples will be loaded (Figure 6 displays the 5' ligation and resulting construct). Run the gel for 90 min at 450 volts.
15. Once the run is complete, remove the gel from the apparatus and glass and spray it with a SYBR Gold solution (i.e., 10 µL of SYBR Gold in 25 mL of 0.5x TBE), and let it sit in the dark for 5 min.
16. Using the 5' adapter-19 nt-3'adapter and both 5'- adapter-24 nt-3' adapter size markers long DNA oligonucleotides as size guides, directly excise the section of the gel containing the 5' ligated 16 RNA samples (Figure 7).
    NOTE: The test RNA ligation is evaluated on the gel to demonstrate that the ligation of the 5' adapter allows a shift equivalent in size to the size markers. The test RNA ligation is not excised from the gel; only the ligated 16 combined RNA samples.
17. Transfer the excised gel piece containing the 16 ligated RNA samples into a 0.5 mL gel breaker tube set into a 1.5 mL tube, spin at 16,000 × g for 3 min at RT, and then add 300 µL of 300 mM NaCl and 1 µL of 100 µM 3' RT/PCR primer (Figure 1B) to the crushed gel pieces in the tube.
18. Seal the tube with parafilm and set on a shaker agitating at 1,100 rpm at 4 °C, overnight (17-18 h) in a cold room.

6. Reverse transcription of the barcoded 5' and 3'-ligated small non-coding RNAs and miRNAs

1. Pipette the solution from the tube containing the crushed gel onto a 5 µm filter tube inserted into a 1.5 mL siliconized RNase-free tube and briefly spin to allow the solution to get through the filter.
2. Add 950 µL of 100% ethanol to this filtered solution, invert the tube to mix, spin for 2 s, and set on ice for 60 min.
3. Precipitate the RNA pellet by centrifugation of the tube at 16,000 × g at 4 °C for 1 h.
4. Recover the tube, remove the supernatant carefully without touching the pellet, and vacuum dry the pellet prior to resuspending it in 5.6 µL of nuclease-free H₂O.
5. Set up the RT reaction by adding 3 µL of 5x first strand buffer, 4.2 µL of 10x dNTPs (each at 2 mM), and 1.5 µL of Dithiothreitol (DTT) to the tube. Flick to mix, and spin for 2 s.
6. Set up the reaction at 90 °C for 30 s exactly and then transfer the tube to a block at 50 °C for 2 min.
7. Initiate the RT reaction by adding 0.75 µL of reverse transcriptase enzyme into the solution, flick to mix, and set at 50 °C for 35 min.
8. Transfer the tube at 95 °C for 1 min to denature the enzyme and the cDNA/RNA strands, then add 95 µL of RNase-free H₂O, flick the tube to mix, and set it on ice for 2 min. This tube contains cDNA generated from the 16 RNA samples, each separately barcoded in 3'.

7. Conduct the test PCR and large-scale amplification

1. Prepare fresh 10x PCR buffer by combining 304 µL of Nuclease-free H₂O, 125 µL of 2 M KCl, 50 µL of 1 M Tris pH 8.0, 10 µL of 1 M MgCl₂, 5 µL of 1% Triton X-100, and 6 µL of 1 M 2-mercaptoethanol into a tube that can be stored on ice.
2. Assemble a Pilot/test PCR reaction for the cDNA produced in step 6.8 (above) by combining 67 µL of Nuclease-free H₂O with 10 µL of 10x PCR Buffer, 10 µL of 10x dNTPs, 0.5 µL of 100 µM 5' PCR primer, 0.5 µL of 100 µM 3' PCR primer, 10 µL of cDNA Stock Library, and 2 µL of 50x Taq polymerase into a fresh PCR tube.
   NOTE: Prior to conducting the PCR amplifications, set up the instrument by establishing two different files as follows: For File#1: 94 °C for 45 s, 50 °C for 85 s, and 72 °C for 60 s for 12 cycles, and 4 °C; and for File#2: 94 °C for 45 s, 50 °C for 85 s, and 72 °C for 60 s for 2 cycles, and 4 °C.
3. For the pilot/test PCR reaction, place the PCR tube containing the 100 µL of the PCR reaction in the Thermocycler and run File#1.
4. Once File#1 is complete, open the PCR tube and transfer 12 µL of the reaction into a 1.5 mL microcentrifuge containing 3 µL of 5x gel loading dye. Close the tube, and write "12 cycles" on the cap.
5. Take the PCR tube, set it into the thermocycler and initiate File#2.
6. Once File#2 is complete, transfer 12 µL of the PCR reaction into a 1.5 mL tube containing 3 µL of 5x gel Loading dye, and write "14 cycles" on its cap.
7. Repeat the PCR amplifications four successive more times (i.e., two PCR cycles each), while transferring 12 µL of PCR product at the end of each amplification, and in turn obtain four different tubes, labelled 16, 18, 20, and 22 cycles.
8. Load the 20 nt size ladder, and each of the PCR amplified products onto a 2.5% agarose and run the gel in 0.5x TBE for 30 min at 120 V.
9. Identify the optimal PCR amplification cycle (Figure 8A), based on the ratio and presence of primer dimers (lower band) and the amplified cDNA (upper band).
   NOTE Considering that the total RNA input with 16 samples may be lower than 1 ng, it is expected that primer dimers will generate a product quicker and in larger quantities than the purified cDNA. Thus, the adequate PCR cycle is generally selected between 18 and 22 cycles. In Figure 8A, the adequate PCR cycle for this library is 20. In Figure 8B, the size of the PCR construct is detailed.
10. Once the adequate PCR amplification cycle is identified, set a negative PCR (no cDNA) and four 100 µL positive PCR reactions (i.e., each containing 12 µL of cDNA).
    1. To simplify the preparation of the PCR reactions, set up a PCR Master Mix (i.e., for 4.5 reactions) in a 1.5 mL tube by combining 292.5 µL of nuclease-Free H₂O, 45 µL of 10x PCR buffer, 45 µL of 10x dNTPs, 3.25 µL of 3' RT/PCR primer (Figure 1B) and 2.25 µL of 5' PCR primer (Figure 1B). Invert this tube to mix and spin to collect at the bottom of the tube.
    2. Transfer 86 µL of the PCR Master Mix into each of four 0.5 mL PCR tubes.
    3. Add 12 µL of cDNA Library (stored on ice), then 2 µL of 50x titanium Taq polymerase into each of the 4 PCR tubes, and mix by pipetting up and down.
    4. Prepare the no-template PCR reaction tube by adding 77 µL of nuclease-free H₂O, 10 µL of 10x PCR buffer, 10 µL of 10x dNTPs, 0.5 µL of 5' PCR primer, 0.5 µL of 3' PCR primer, and 2 µL of 50x Taq polymerase to a separate PCR tube and pipetting up and down to mix.
    5. Set up the PCR tubes in the thermocycler using the PCR cycle identified in step 7.9 and set a new program on the thermocycler to amplify as necessary.
11. Prepare a 2.5% agarose gel with 0.5x TBE and ethidium bromide, and transfer 12 µL from each PCR tube into individual tubes containing 3 µL of 5x gel loading dye.
12. Prepare the 20 nt ladder by combining 3 µL of ladder with 9 µL of nuclease-free H₂O and 3 µL of gel loading dye.
13. Load the gel, run it for 30 min at 120 V, and verify that the PCR amplification is similar between the four different reactions and that the no-template control is empty (Figure 8C).
14. Combine the four PCR reactions between two 1.5 mL siliconized microcentrifuge tubes (i.e., two tubes each containing 90 µL of product), and add 18 µL of 5 M NaCl, and 800 µL of 100% ethanol in each of the two tubes.
    1. Invert the tubes to mix and store them at -20 °C to precipitate overnight.

8. Purification of the PCR amplified Library

1. Centrifuge the two tubes containing the mixed positive PCR reactions at 16,000 × g for 60 min at 4 °C.
2. After 60 min, remove the supernatant, vacuum dry the DNA pellets, and resuspend them in 19.5 µL of 1x buffer, and then add 0.5 µL of PmeI enzyme.
3. Set up the PmeI digests at 37 °C for 2 h to remove any of the DNA size markers that ran in the adjacent wells (i.e., 19 nt-3' adapter, 24 nt-3'adapter) or RNA carriers (i.e., RNA extraction carrier, RNA post-3' ligation carrier) potential contaminants (i.e., see PmeI sites included in RNA and DNA primers in Supplemental File 1)
4. Run the two digests in three different wells of a 2.5% agarose gel using 0.5x TBE to avoid accumulating an excess of primer dimers in a single well and creating a smear.
   1. Add 3 µL of 5x gel loading dye to each of the digests, combine the two digests, and then load 16.5 µL into three separate adjacent wells.
   2. Load the 20 nt DNA size ladder at the end wells of the gel and run the gel for 90 min at 150 V (Figure 8D).
5. To purify the DNA libraries, excise the upper PCR bands from the three adjacent lanes of the gel, and transfer them into a fresh 1.5 mL tube.
6. Purify the PCR amplified library using a gel extraction kit, following the manufacturer's instructions, and quantify the purified DNA product using a fluorometer, following manufacturer's instructions.

9. Final PCR amplification to add barcodes compatible with the sequencer

1. Prepare a fresh 10x PCR buffer mix as the one described in step 7.1.
2. Set up PCR reactions by including one negative control (no amplicons) and three PCR reactions using 0.75 ng of the PCR product purified from the first PCR amplification in each of the three tubes as follows:
   1. Set up three individual PCR reactions by mixing up 75 µL of ddH₂O, 10 µL of 10x PCR buffer, 10 µL of 10x dNTPs, 1 µL of 5' 2^nd PCR primer (i.e., primer i503 (Figure 1B)), 1 µL of 3' PCR 2^nd primer (i.e., primer i702 (Figure 1B)), 1 µL of PCR product (i.e., 0.75 ng from gel-purified 1^st PCR product), and 2 µL of Titanium Taq Polymerase.
   2. Set up the negative PCR control (i.e., no DNA template) by mixing up 76 µL of ddH₂O, 10 µL of 10x PCR buffer, 10 µL of 10x dNTPs, 1 µL of 5' 2^nd PCR primer (i.e., primer i503 (Figure 1B)), 1 µL of 3' PCR 2^nd primer (i.e., primer i702 (Figure 1B)), and 2 µL of Taq Polymerase.
   3. Set up the four PCR reactions (i.e., one negative control, three positives) on a thermocycler programmed as follows: 72 °C for 3 min, 95 °C for 30 s, and 95 °C for 10 s, 55 °C for 30 s, 72 °C for 30 s for 9 cycles, and then 72 °C for 5 min, and a final hold at 10 °C.
3. Once the PCR cycles are completed, take 10 µL of the reaction mixture and add 3 µL of loading dye, and evaluate the products on a 2.5% agarose gel as displayed in Figure 9A, along with the 20 nt ladder.
4. Upon validation of the PCR product's size (i.e., upper PCR band running at 164 bp as displayed in Figure 9B and Figure 6 for size and details on the final construct) and band distribution, combine the PCR products in two tubes by adding to each 135 µL of PCR product (i.e., 90 µL+ 45 µL), 13.5 µL of 5 M NaCl, and 650 µL of 100% ethanol, and precipitate at -20 °C overnight.
5. The next day, spin down the two PCR tubes at 16,000 × g for 60 min at 4 °C.
6. Dry each pellet, resuspend with 19.5 µL of 1x PmeI buffer, and add 0.5 µL of PmeI for 2 h at 37 °C on a shaker to further eliminate any of the size markers that ran in the adjacent wells (i.e., 19 nt-3' adapter, 24 nt-3'adapter) or RNA carriers (i.e., RNA extraction carrier, RNA post-3' ligation carrier) contaminants.
7. Set up the digested products in three wells of a 2.5% agarose gel and separate at 150 V for 90 min.
8. Excise the upper band, which runs at 164 bp (Figure 9C), and purify the DNA product using a DNA isolation kit and quantify with a fluorometer.
9. Run the final PCR product on a DNA chip to validate size, concentration, and purity of the library before dilutions and analysis using a sequencer.
10. Transfer the FASTQ data file to the RNAworld pipeline for adapter trimming, demultiplexing, alignment to the human genome, prior to miRNA analyses, as previously described⁴¹. Alternatively, using the Script presented in Supplemental File 3, demultiplex the reads from the FASTQ file and further process using a small-RNA analytical pipeline of choice.

Results

As detailed in the above protocol, we describe the simultaneous processing of up to 16 individual small extracellular vesicle (sEV) RNA samples (i.e., for analysis of small non-coding RNAs and miRNAs), which are analyzed together within a single library, after undergoing 3' barcoding. The sEVs that were selected for our analyses were isolated by ultracentrifugation, as previously described^37,40, from 45 mL of plasma collected from women diagnosed with triple negative breast cancer at the Jersey Shore University Medical Center (IRB#Pro2018-0425). Upon isolation, the ultracentrifuged sEVs were validated following MISEV2023 guidelines⁴², using Transmission Electron Microscopy, Nanoparticle Tracking Analysis and fluorescent super-resolution Nanoimaging, prior to dilution, aliquoting, RNA extraction, and cDNA library preparation. Our TEM analyses revealed that we obtained highly enriched populations of sEVs (Figure 2), the size of which was confirmed using a Nanoparticle Analyzer (Figure 3). Additionally, using ONi super resolution nanoimaging, we confirmed the presence of the human tetraspanins CD9, CD81, and CD63 on the surface of the ultracentrifuged sEVs (Figure 4). Prior to initiating the cDNA library preparation, sEV RNA extractions were conducted using the serum/plasma kit and RNA was evaluated on a fragment analyzer, which estimated that 3.8 × 10¹⁰ sEVs contained 4.84 ng of total small non-coding RNAs between 1 and 200 nt or 2.07 ng of miRNAs (Figure 3, lower panel).

Importantly, during RNA extraction, our RNA extraction carrier oligonucleotide (Supplemental File 1) was added after initial processing to the aqueous phase to improve sEV-RNA recovery yield. This RNA extraction carrier is efficiently removed by PAGE purification because of its size being smaller than the 3' barcoded RNA constructs. Following overnight 3' barcoding, the T4 RNA ligase was heat-inactivated, and the 16 RNA samples (i.e., initial test set of samples) were combined and precipitated within a single tube, since the small non-coding RNAs and miRNAs of interest from each of the individual samples harbored our sample-specific 3' barcodes. The RNA pellet generated from the combined RNA samples was resuspended and the ligated small non-coding RNA and miRNA molecules of interest (19-24 nt) were separated on a 15% PAGE, whereby using size marker DNA oligonucleotides migrating in adjacent wells, we selectively guided gel-excision based on the size of the expected 3' barcoded RNA constructs (Figure 5).

The cut-out small PAGE-gel piece was crushed and incubated in a sodium chloride solution overnight to elute the ligated small non-coding RNA and miRNA molecules. The next day, the gel-eluted size-selected small non-coding RNAs and miRNAs were precipitated by centrifugation to establish an RNA pellet, which was resuspended and underwent ligation with the 5' adapter. Upon ligation of the 5' adapter, the 5' and 3'-ligated small non-coding RNA and miRNA molecules were size-selected on a 15% PAGE using DNA size marker oligonucleotides run in adjacent wells as size guides. The gel piece containing the 5' and 3'-ligated small non-coding RNAs and miRNAs was excised, crushed, and incubated overnight for elution from the fragmented gel. Subsequently, the small non-coding RNAs and miRNAs were precipitated, resuspended in nuclease-free water, and underwent reverse transcription to generate cDNA molecules representative of all cloned small non-coding RNAs and miRNA molecules from the different RNA samples.

To generate the DNA library, the cDNA molecules underwent a pilot/test PCR reaction (see Figure 8A) to identify the optimal PCR amplification cycle, whereby the ratio of DNA library versus DNA primer dimers was evaluated on a 2.5% agarose gel, which visually validated the production of the properly sized 95 nt amplicons (see Figure 8B). Then, four PCR reactions were set up, using the determined PCR amplification cycle, to generate sufficient DNA material for gel isolation. The PCR amplified DNA was combined, precipitated, subjected to a 2 h PmeI digest to remove any trace of the RNA extraction carriers, RNA post-3' ligation primers that may have undergone any of the ligations, or any of the 19 nt-3' adapter and/or 24 nt-3' adapter that may have ligated the 5' adapter, all of which contain a PmeI site (Supplemental File 1). Following PmeI digestion, the PCR products were separated on a 2.5% agarose gel. The top PCR band (i.e., DNA library), which ran at 95 nt, was excised and purified (see Figure 8C).

The PCR amplified DNA library of 95 nt underwent a second PCR amplification (i.e., 0.75-ng of PCR product for nine new PCR cycles; see Figure 9A) for the addition of the i5 and i7 adapters in 5' and 3' of our DNA constructs for compatibility with the sequencer (see Figure 9B). Following the second round of amplification, the PCR reactions were combined, precipitated, and the DNA library running at 164 nucleotides was isolated from a 2.5% agarose gel, showing a clear separation from the primer dimers (see Figure 9C). The isolated DNA library was quantified and diluted following Next Generation Sequencing instrument instructions before undergoing sequencing. Using the RNAworld pipeline, the FASTq files underwent adapter trimming and demultiplexing (i.e., using the Script from Supplemental File3) to separately create individual files for each individual sample. Biostatistical analyses⁴¹ were then conducted to evaluate miRNA representation for each of the different sEV RNA samples (see Figure 10 and Supplemental Figure S1).

As we sought to confirm the robustness of our cDNA library preparation protocol when using sEV-RNA, the first library preparation was conducted using the 16 barcodes in 16 individual ligations using the same amount of sEV-RNA extracted from the same number of sEVs (i.e., 8.4 × 10⁹ sEVs). The sEVs we utilized were isolated from combined human plasma samples using a previously described ultracentrifugation method^37,40. Our TEM and Nanoparticle tracking analyses performed after our plasma ultracentrifugation protocol allowed us to estimate our sEV concentration to be at 3.8 × 10¹² sEVs/mL or 3.8 × 10⁹ sEVs/µL. Since RNA fragment analysis estimated that the total small non-coding RNA concentration from 3.8 × 10¹⁰ sEVs (i.e., 10 µL of the 3.8 × 10⁹ sEVs), extracted using the serum/plasma extraction kit contained 4.84 ng of total small non-coding RNAs or 2.07 ng of miRNAs, we thus estimated that 8.4 × 10⁹ sEVs contains 1.07 ng of small non-coding RNAs or 457 pg of total miRNAs.

For the first experiment, we set up 16 separate ligations each containing 1.07 ng of total small non-coding RNAs or 457 pg of miRNAs (i.e., equivalent to 8.4 × 10⁹ sEVs) and used the 16 different barcodes. The 16 replicates underwent cDNA library preparation, as described in the first paragraph above, followed by sequencing, adapter trimming, demultiplexing (i.e., using the script described in Supplemental File 3), and miRNA analysis using the pipeline developed by the laboratory of Dr. Thomas Tuschl⁴¹. To quantitatively assess reproducibility across the 16 replicate samples, we performed a comprehensive combinatorial analysis testing 6,435 different sample subsets (i.e., all 8 versus 8 samples combinations) across 292 miRNAs (i.e., features) in our dataset (Supplemental Figure S1). This rigorous approach revealed remarkable consistency in miRNA profiles, with the majority of combinations (80.3%) showing no significant differences between sample groups. The mean fraction of significant features was just 0.2% (median 0.0%), with a standard deviation of 0.5%, demonstrating remarkable homogeneity across sample subsets. Of the 207 unique miRNAs that showed any variation, only five miRNAs (i.e., hsa-miR-126-5p, hsa-miR-1277-3p, hsa-miR-16-2*, hsa-miR-30b, and hsa-miR-30c) appeared as significantly different in more than 1% of combinations, with even the most variable miRNA (i.e., hsa-miR-30b) showing differences in only 2.25% of sample combinations. The distribution analysis confirmed this uniformity, with quantiles of significantly different features ranging from 0.0% at the 75th percentile to a maximum of 8.2% (Supplemental Figure S1), providing strong statistical evidence for the robust reproducibility of the sEV-miRNA cDNA library preparation protocol. The heatmap, included in Figure 10, displays the top-most detectable 292 miRNAs and highlights the reproducibility of this cDNA library preparation protocol.

In a second experiment, we sought to evaluate the sensitivity of this cDNA library preparation procedure by analyzing the miRNA content of decreasing amounts of sEVs and thus sEV-RNA, by using RNA extracted from 10¹⁰ sEVs (i.e., 1,270 pg of total small non-coding RNAs or 544 pg of miRNAs), 5 × 10⁹ sEVs (i.e., 636 pg of total small non-coding RNAs or 272 pg miRNAs), 10⁹ sEVs (i.e., 127 pg of total small non-coding RNAs or 54 pg of miRNAs), 5 × 10⁸ sEVs (i.e.,63 pg of total small non-coding RNAs or 27 pg of miRNAs), 10⁸ sEVs (i.e., 12.7 pg of total small non-coding RNAs or 5.4 pg of miRNAs), and 2.5 × 10⁷ sEVs (i.e., 3.15 pg of total small non-coding RNAs or 1.35 pg of miRNAs) in duplicates, using our RNA extraction in presence of our RNA extraction carrier oligonucleotide (Figure 11). This titration experiment revealed a clear relationship between EV input quantity and miRNA detection sensitivity (Supplemental Figure S2A), as the number of detected miRNAs decreased proportionally with sEV input, with a notable inflection point between 1 × 10⁷ and 1 × 10⁸ sEVs. Indeed, at a detection threshold of 10 reads per miRNA, approximately 550 unique miRNAs were identified from 10¹⁰ sEVs, compared to ~250 miRNAs from 2.5 × 10⁷ sEVs. However, technical reproducibility remained excellent across most input levels, with Pearson correlations between replicates exceeding 0.95 for inputs ≥10⁷ sEVs (Supplemental Figure S2B). Hierarchical clustering of the expression profiles of 215 miRNAs detectable across dilutions and repeats, as depicted in Figure 11, further highlights the reproducibility of this cDNA librarypreparation protocol and demonstrates that samples clustered primarily by sEV RNA input amount. These data confirm that this protocol can reliably detect miRNAs from as few as 2.5 × 10⁷ sEVs or as low as 1.35 pg miRNAs (i.e., 3.15 pg small non-coding RNAs) and that it represents a significant improvement over many commercial small-RNA sequencing kits.

Figure 1: 3' Adapters, RT and PCR primers. (A) The 16 3' adenylated barcoded adapters are displayed, and their unique barcodes are highlighted in grey. (B) The 5' RNA adapter oligonucleotide is displayed. (C) The first PCR DNA primers include the 3'RT/PCR primer that is used as a carrier during elution of 5' ligated RNA samples in 300 mM NaCl overnight, a PCR primer during the first PCR amplification, and the 5' PCR primer used during the first PCR amplification. (D) The second PCR DNA primers include the 3' PCR DNA primer i7 and the 5' PCR DNA i5 primer. Please click here to view a larger version of this figure.

Figure 2: Transmission Electron Microscopy validation of sEV morphology. Small extracellular vesicles ultracentrifuged from 45 mL of human plasma from women diagnosed with triple negative breast cancer were resuspended in 100 µL and evaluated by Transmission Electron Microscopy images at a concentration of 1.9 × 10¹⁰ sEVs/µL, which revealed large numbers of sEVs per grid, with large 150-250 nm sEVs detectable. Abbreviation: sEV = small extracellular vesicle. Please click here to view a larger version of this figure.

Figure 3: Size distribution and small non-coding RNA concentration analysis of ultracentrifuged sEVs. Using the Nanoparticle Analyzer (top), the size distribution and number of the sEVs isolated by ultracentrifugation were determined, revealing a quantity of 3.8 × 10¹⁰ sEVs/mL after a 100-fold dilution (i.e., blue curve) and an estimated 5.2 × 10⁹ sEVs/mL after a 1,000-fold dilution (i.e., green curve). The tracing indicates that sEVs within 60-250-nm were detectable in human triple negative breast cancer plasma after ultracentrifugation performed using a previously described procedure^37,40. The total small non-coding RNA concentration (bottom) was estimated at 0.44 ng/µL of which 42.9% represented miRNAs or a concentration of 0.188 ng/µL from 3.8 × 10¹⁰ sEVs (i.e., serum/plasma RNA extraction using 10 µL of 3.8 × 10¹² sEVs/mL). Given that the total final volume of RNA solution was 11 µL, we estimated having 4.84 ng of total small non-coding RNAs of which 2.07 ng of miRNAs from a total of 3. 8 × 10¹⁰ sEVs. Please click here to view a larger version of this figure.

Figure 4: sEV characterization using super-resolution Nanoimaging. Super-resolution Nanoimaging of immobilized sEVs (i.e., duplicate test labeled at test 1 and test 2 from 7.6 × 10⁹ sEVs each) evaluated for the presence of the three tetraspanins CD9, CD63, and CD81. The left grids display detection of fluorescently labeled anti-CD9 (i.e., yellow AF388), anti-CD63 (i.e., blue AF568), and anti-CD81 (i.e., purple AF647) sEVs on three individual grids on top left and three individual grids on bottom left, along with a merge grid for all three fluorescent signals, for undiluted 7.6 × 10⁹ sEVs (i.e., 2 µL of 3.8 × 10⁹ sEV stock solution). On the right, single sEVs are visualized for the combined presence of CD9, CD63, and CD81, using the three different fluorescently labeled antibodies. Please click here to view a larger version of this figure.

Figure 5: Separation of the 3' barcoded RNA samples on a 15% PAGE. The 16 barcoded sEV-RNA samples were combined, precipitated, centrifuged, and the RNA pellet was resuspended and migrated on a 15% PAGE (see well 6). The red rectangle shows where the 3' barcoded small non-coding RNAs and miRNAs were excised from the gel using the size markers loaded in the adjacent wells (see wells 3, 4, and 8, 9). The gel piece was crushed and set into a nuclease-free tube to undergo diffusion overnight at 1,100 rpm at 4 °C in 400 mM NaCl. Using the 20 nt DNA ladder, the size of the constructs and size markers was visually validated (see well 1). A test ligation using the RNA ligation test oligonucleotide (Supplemental File 1) with the additional barcode#17 (Figure 1A), was set up along with the 3' adapter ligations of day 1 (see well 2), using the same buffer and enzyme mix as the other 16 ligations. The blue rectangle displays the excised band of the RNA test ligation oligonucleotide for overnight diffusion at 1,100 rpm at 4 °C in 400 mM NaCl. The two (19 nt guide (see wells 3 and 8) and 24 nt guide (see wells 4 and 9)) 3' ligation size marker short DNA oligonucleotides are loaded on each side of the 3' ligated 16 RNA samples, with a gap of one well to prevent contamination, and used as a size guide for excision of the gel piece. Abbreviations: nt = nucleotide; sEV = small extracellular vesicle; PAGE = polyacrylamide gel electrophoresis. Please click here to view a larger version of this figure.

Figure 6: Graphical representation of the ligated small non-coding RNAs and miRNAs, the first, and the second PCR amplification constructs. From the top to the bottom of the figure, the 3'-ligated RNA construct, the 5' and 3'-ligated RNA construct, the reverse-transcribed single-stranded DNA constructs, the first PCR-amplified double-stranded DNA construct, and the second PCR-amplified dsDNA constructs are depicted. The 5' adapter (left), the 22-nt RNA sequence (i.e., synthetic miRNA sequence from IDT), and the adenylated 3' barcoded adapter are represented at the top of the figure. The RT and first PCR constructs are color-coded with the different regions representing the different adapters and their size contribution to the constructs. The first PCR construct produced using the 5' PCR primer and the 3' RT/PCR primer (Figure 1B) generates a 95 nt dsDNA product. At the bottom of the schematic is a representation of the sequence generated during the second PCR reaction, whereby using the i5 PCR primer in 5' (see region of primer sequence that anneals onto the first PCR transcript), and the i7 PCR primer in 3' (see region of primer sequence that anneals onto the first PCR transcript) a final construct of 164 nucleotides is produced within nine PCR cycles. This PCR-amplified dsDNA product is purified prior to undergoing NGS analysis. Abbreviations: ss = single-stranded; ds = double-stranded; nt = nucleotide. Please click here to view a larger version of this figure.

Figure 7: Separation of the small non-coding RNA library after ligation of the 5' adapter. The 3' ligated small non-coding RNA/miRNA library (Figure 5 and Figure 6) is ligated with the 5' adapter, and the reaction is migrated on a 15% PAGE (see well 7). The 20 nt DNA ladder is migrated on both ends of the gel (see wells 1 and 11). The 3' ligated RNA ligation test oligonucleotide isolated from the previous gel (Figure 6, well 2) underwent the ligation of the 5' adapter prior to gel migration. The yellow rectangle displays the shift of the 3' ligated RNA ligation test oligonucleotide that confirms successful ligation of the 5' adapter (compare size with the previous gel in Figure 5, well 2). The two (19 nt guide (see wells 4 and 9) and 24 nt guide (see wells 5 and 10)) 3' ligation size marker long DNA oligonucleotides on each side of the 5' ligated 16 combined small non-coding RNA/miRNA samples, are separated by one well to prevent contamination to the RNA library. The red rectangle displays the anticipated 5' and 3'-ligated small non-coding RNA/miRNA library size and location in the well upon migration. The gel piece (i.e., red rectangle) is excised, crushed, and set in a new tube for diffusion at 1,100 rpm overnight at 4 °C in 300 mM NaCl spiked with the 3' RT/PCR primer (Figure 1B), which acts as a carrier, and used as a primer for the reverse transcription the following day. Abbreviations: nt = nucleotide. Please click here to view a larger version of this figure.

Figure 8: Pilot PCR prior to amplification of the cDNA library. The proper size and ratio of the PCR amplified products and primer dimers are observed on a 2.5% agarose gel. (A) Following reverse transcription of the 5' and 3'-ligated small non-coding RNA and miRNA library, that includes 16 samples, 12 µL of PCR amplified products, obtained after 12 cycles (well 2) and successively every two cycles at 14 (well 3), 16 (well 4), 18 (well 5), 20 (well 6) and 22 cycles (well 7), are analyzed on a 2.5% agarose gel. The cDNA library (upper band at 95 nt) and adapter dimers (lower band) are visualized and exponentially represented between wells 4 and 7. For this library, due to the excess of primer dimers formed at cycle 22 (well 7), a PCR amplification of 20 cycles (see well 6) is selected to minimize the representation of primer dimers. (B) This schematic shows the size and position of the different RNA and DNA oligonucleotides as well as the size of the resulting RNA and DNA constructs, which are produced and observed on the different 2.5% agarose gels during the preparation of the cDNA library. Upon PCR amplification, the PCR amplified cDNA library is expected to be 95 nt long. (C) Aliquots of the four individual 100 µL large-scale PCR reactions (i.e., PCR using 12-µl of cDNA for 20 cycles), which are individually migrated on a 2.5% agarose gel confirm positivity of the PCR reactions (see wells 3 to 6), compared to a non-DNA template reaction (well 2). Both the library (upper band) and the primers dimers (lower band) can be observed on this gel (see green rectangles). All migration patterns are compared to the 20 nt size ladder (see well 1). (D) A 2.5% agarose gel image of the pooled PCR reactions (i.e., four PCR reactions pooled, precipitated, and separated in wells 3, 4, and 5) ran in three adjacent wells (i.e., to prevent excess of primer dimer from improperly migrating and separating from the library band). This gel observation confirms that the upper band noted in wells 3, 4, and 5 runs at ~95 nt, separately from the primer dimer band migrated below separately. The upper PCR bands (i.e., orange rectangle) are excised, purified with a DNA gel extraction kit, and quantified. Abbreviations: miRNA = microRNA; nt = nucleotide. Please click here to view a larger version of this figure.

Figure 9: Second PCR to add the i5 and i7 adapters to the previous PCR construct. (A) A total of 0.75 ng from the gel-isolated 95 nt band of the first PCR product (see Figure 5D) is used to prepare three 100 µL PCR reactions that include pairs of the i5 and i7 PCR primers (see Figure 1B) for a total of nine PCR cycles (i.e., different PCR cycles than those set up for the first PCR). As observed in wells 3, 4, and 5, the PCR generates the upper library band (i.e., expected at 164 nt) and the lower primer dimer band (see green rectangles). As observed in well 2, the negative PCR reaction (i.e., no DNA) does not amplify a visible product. The DNA ladder (20 nt) is used to visually verify the size of the PCR products. (B) The anticipated amplicon size is schematically displayed. (C) Upon amplification of two large-scale PCR reactions, combination of their content, precipitation by centrifugation, PmeI digest, and electrophoretic separation on a 2.5% agarose gel within three wells (wells 3, 4, and 5) the final DNA library can be separated and observed (i.e., the large orange rectangle). The gel migration of the PCR products allows separation of the 164 nt DNA library from the primer dimers, prior to excision of the gel band (i.e., the large orange rectangle) and isolation using a gel extraction kit. The final PCR product is quantified on a fluorometer prior to undergoing the dilutions required for sequencing. The 20 nt ladder in well 1 confirms the size of the expected DNA library size at 164 nt. Abbreviations: nt = nucleotide. Please click here to view a larger version of this figure.

Figure 10: Validation of reproducibility of the cDNA library preparation protocol for sEV analysis. Sixteen aliquots of 1.07 ng of total small non-coding RNAs (i.e., size 1-200 nt) or 457 pg miRNAs,equivalent to 8.4 x 10⁹ sEVs, underwent ligation of 16 individual 3' barcoded adapters to produce a cDNA library following this described protocol. The Biostatistical analysis of the miRNA content from 8.4 x 10⁹ sEVs was conducted using R with multiple specialized packages. Raw count data was preprocessed by removing questionable viral miRNAs and filtering low-expression miRNAs using edgeR's filterByExpr function. For normalization, the Trimmed Mean of M-values method was applied, followed by log2 transformation of expression values. Differential expression analysis was performed with DESeq2, implementing a design formula that accounted for batch effects. To assess reproducibility across the different replicate samples, we calculated detection rates at multiple thresholds (1-10,000 counts) and visualized expression patterns through hierarchical clustered heatmaps (ward linkage) using the NMF package. The top 292 detectable miRNAs are displayed in the heatmap, which highlights the high reproducibility of the cDNA library preparation. Please click here to view a larger version of this figure.

Figure 11: Validation of the sensitivity of the cDNA library preparation protocol across decreasing sEV RNA input amounts. Following the same Biostatistical approach, we analyzed the 215 most detectable miRNAs across samples established using 10¹⁰ sEVs in duplicate (barcodes 1 and 10, using each 1,270 pg of total small non-coding RNA with an estimated 544 pg of miRNAs), 5 × 10⁹ sEVs in duplicate (barcodes 2 and 11, using 636 pg of each total small non-coding RNA with and an estimated 272 pg of miRNAs), 10⁹ sEVs in duplicate (barcodes 3 and 12, using each 127 pg of total small non-coding RNA with an estimated 54.4 pg of miRNAs), 5 × 10⁸ sEVs (barcodes 4 and 13, using each 63.6 pg of total small non-coding RNA with an estimated 27 pg of miRNAs), 10⁸ sEVs (barcodes 5 and 14, using each 12.7 pg of total small non-coding RNA with an estimated 5.4 pg of miRNAs), and 2.5 x 10⁷ sEVs (barcodes 6 and 15, using 3.15 pg of each total small non-coding RNA with an estimated 1.35 pg of miRNAs), which all underwent individual RNA extractions and that are displayed in the heatmap. The replicates are depicted in turquoise and salmon colors. The total small non-coding RNA concentration is displayed as a gradient from dark purple (high) to light purple (low). The total miRNA concentration is displayed as a gradient from dark blue (high) to light blue (low). The total number of sEVs is displayed as a gradient from dark green high) to light green (low). Abbreviations: sEVs = small Extracellular Vesicles; pg = picogram; miRNA = microRNA. Please click here to view a larger version of this figure.

Supplemental File 1: Synthetic RNA and DNA primers. Primer sequences and their chemical modifications (top right corner with abbreviations), and concentrations utilized for this procedure are detailed in this figure. (A) List of the 10 individual RNA oligonucleotides used as our calibrator cocktail displays (i.e., with a final concentration of 0.0052 nM). (B) The calibrator oligonucleotides are resuspended in a solution containing the RNA carrier oligonucleotide (0.5 µM). (C) The RNA extraction carrier oligonucleotide (4 ng) is spiked into the RNA lysis buffer with the sEV-RNA before RNA extraction. It contains a PmeI restriction site for its removal by a PmeI digestion, in case it erroneously binds 3' and 5' adapters. (D) The RNA post 3' ligation carrier oligonucleotide is spiked into the 400 mM NaCl solution overnight and utilized as a carrier post 3' ligation of the Adenylated barcodes. It contains a PmeI restriction site for its removal by a PmeI digestion, in case it erroneously binds 3' and 5' adapters. (E) The RNA ligation test oligonucleotide is used along with the experimental reactions to test enzymes and evaluate the size of the products on a 15% acrylamide gel. It contains a PmeI restriction site for their removal by a PmeI digestion, in case they leak from the wells adjacent to the library or its gel piece during SYBR-gold staining. (F) The two (i.e., 19 nt 3' adapter and 24 nt 3' adapter) 3' ligation-size marker short DNA oligonucleotides are loaded on each side of the 3' ligated RNA samples to be used as a size guide on the 15% PAGE. They each contain a PmeI restriction site for their removal by a PmeI digestion, in case they leak from the adjacent wells to the library during SYBR-gold staining and end up listing the 5' adapter. (G) The two (i.e., 5' adapter-19 nt-3' adapter and 5' adapter-24 nt 3' adapter) 5' ligation-size marker long DNA oligonucleotides are loaded on each side of the 5' ligated RNA samples to be used as a size guide on the 15% PAGE. Both primers contain a PmeI restriction site for their removal by a PmeI digestion, in case they leak from the wells adjacent to the library or its gel piece during SYBR-gold staining. Please click here to download this File.

Supplemental File 2: Synthetic RNA and DNA primers. Please click here to download this File.

Supplemental File 3: Demultiplexing script. This script is used to demultiplex the FASTq file produced by the sequencing instrument. Please click here to download this File.

Supplemental Figure S1: Differential expression of miRNAs detected between replicates. Combinatorial reproducibility analysis examined between 6,435 different sample combinations (i.e., all 8 vs 8 sample combinations) across 292 miRNAs (features) in our dataset. Remarkable consistency in miRNA profiles is observed as the majority of combinations (80.3%) show no significant differences between sample groups. As shown in the distribution analysis, even in the most variable cases, the maximum fraction of differentially expressed miRNAs was only 8.2%. Please click here to download this File.

Supplemental Figure S2: Evaluating technical replicates at different sEV input levels. (A) Relationship between sEV input quantity and miRNA detection sensitivity at different read count thresholds. (B) Total RNA was extracted from a dilution series of plasma-derived sEVs (i.e., ranging from 10¹⁰ to 2.5 × 10⁷ sEVs) with technical replicates at each dilution point. We evaluated miRNA detection sensitivity by quantifying the number of unique miRNAs detected at different count thresholds (1-500 reads) across the sEV input range. Reproducibility was assessed by calculating and plotting Pearson correlation coefficients between technical replicates at each dilution level. Please click here to download this File.

Discussion

We have developed a cDNA library preparation protocol for reproducible and sensitive next-generation sequencing (NGS) of non-coding RNAs, including microRNAs (miRNAs) efficiently isolated from small extracellular vesicles (sEVs). Considering that a limited quantity of small non-coding RNAs and miRNAs may be recovered from globally or selectively isolated sEVs, we sought to optimize our cDNA library preparation procedure that enables reproducible and sensitive analysis of sEVs miRNA cargos.

As a first step prior to efficient and selective isolation of sEV-RNA, an RNAse-A treatment step is necessary to eliminate any non-sEV RNA contaminants. It is important to note that it may not allow removal of protein-bound RNA sequences, but that it also has the potential to digest RNA from sEVs, which may be structurally compromised. As a second step, to maximize sEV-RNA recovery, we introduce 4 ng of an RNA carrier oligonucleotide in the lower phase of TRizol/Qiazol RNA lysis buffer (i.e., protected from denatured RNAse-A), to help minimize RNA loss during subsequent column isolation. This RNA carrier is anticipated to remain inert throughout the following enzymatic reactions as its 3' region carries a 3C spacer to prevent its 3' barcoding, and its 5' region is missing a phosphate. Importantly, due to its small size, this primer is removed during PAGE separation. However, to enhance its removal if it was to get barcoded in 3' (i.e., an issue in its synthesis by the manufacturer), potentially purified and then barcoded in 5', we included a PmeI restriction digest site to allow for its enzymatic removal following reverse transcription and PCR amplification. As a third step for improving recovery of our 3' barcoded RNA library, an RNA carrier oligonucleotide (100 ng) is added following excision from the 15% PAGE, which combines with the 3' l-ligated RNA library during diffusion in 400 mM NaCl at 4 °C overnight, and that improves precipitation in ethanol.

In this fashion, following centrifugation at 16,000 × g for 1 h at 4 °C, it forms a visible RNA pellet that can then be processed and undergo the ligation of the 5' barcode. As this RNA carrier oligonucleotide lacks a 5' phosphate, it is prevented from gaining the 5' adapter, and it can thus be easily separated on the second 15% PAGE gel as a fast-migrating product. However, to prevent any potential unexpected ligation (i.e., issue in its synthesis by the manufacturer), we also included a PmeI restriction site in its sequence, in case it would get reverse-transcribed, and PCR amplified, so it can be efficiently digested out. Indeed, all PCR products (i.e., from first and second PCR amplifications) undergo a PmeI digest as a precautionary measure to remove any unwanted ligated products. This PmeI digest also allows for removal of unwanted size markers (i.e., 3' ligation 19 nt-3' adapter and 24 nt-3' adapter short and 5' ligation 5' adapter-19 nt-3' adapter and 5' adapter-24 nt-3' adapter long DNA size markers) that may have diffused into the PAGE gels during the different electrophoreses or during SYBR-gold staining. Altogether, we demonstrate that the pooling of low-input RNA samples, the precautionary additions of RNA carriers, PmeI digests maximize RNA material recovery, and most importantly simultaneous processing of the 3'-ligated RNA transcripts prevent contamination of highly abundant synthetic carrier RNAs and size marker DNA molecules but also contribute to the high reproducibility observed in our measures.

This cDNA library protocol was modified from its initial version because it was developed to be compatible with the Hiseq2500 single-end read chemistry and flowcell, which were discontinued. Considering the sensitivity, the robustness, and the reproducibility of our approach when working with low-input RNA material^35,36, we sought to modify the design to include 5' and 3' adapters compatible with recently developed paired-end dual index chemistry for new sequencers (NextSeq1000/2000). As clearly depicted in Figure 3, we maintained the five bases barcodes that were initially developed for the original cDNA library preparation protocol^35,36. However, we modified the 5' region of the 3' adenylated barcoded adapters to be compatible with the 3' region (15 nt) of the Illumina i7 barcoded PCR adapters. Furthermore, we modified the 5' region of the initial 5' adapter to match the 3' region (14 nt) of the Illumina i5 barcoded PCR adapter. Therefore, we maintained the experimental procedures allowing individual 3' barcoding of the RNA transcripts, followed by their simultaneous/combined 15% PAGE isolation, 5' ligation, second PAGE isolation, reverse transcription prior to the first and second PCR amplifications. It is important to note that the 3'RT/PCR primer (Figure 1B), which is used both as a reverse-transcription primer and a PCR primer during the first PCR amplification, 100% matches the 5' region of the 5' and 3' ligated small non-coding RNAs and miRNAs, which minimizes mis-priming and incorrect PCR amplification.

Due to its small size, this primer is easily removed during purification of the DNA library on the first 2.5% agarose gel after the first PCR amplification, which prevents it from interfering with the second PCR amplification that uses the i5 and i7 DNA PCR primers. It is important to note, however, that due to the very low RNA input, the PCR amplified DNA library on the first 2.5% agarose appears weaker than the primer dimers (Figure 5A). Unfortunately, primer dimers are formed by ligation of the 3' adenylated barcoded adapter and the 5' adapter, as they remain in low amounts in the excised RNA bands from the different PAGE gels. However, the PCR amplified DNA library is clearly visible at 95 nt (Figure 5A) and by choosing a PCR cycle that minimizes the representation of the primer dimers, which could skew the PCR dynamics, we can purify our PCR product of interest and then subject it to the second PCR amplification for the addition of the i5 and i7 adapters. This second PCR amplification only requires nine PCR cycles, with an input of only 0.75 ng of the first PCR product, which allows for uniform inclusion of the i5 and i7 adapters. Importantly, the gel analysis of the second PCR amplification shows that targeted isolation of the 95 nt first PCR product greatly improves the dynamic of the second PCR amplification, as we observe an increased representation of our cDNA construct (Figure 6C top band), which can then be isolated in large quantity from the 2.5% agarose gel. Importantly, because we include 3' adenylated barcoded adapters in the cDNA construct, it allows for the multiplexing of up to 16 samples per library. Furthermore, since we can include different pairs of i5 and i7 adapters during the second PCR amplification, we can multiplex up to three libraries, each containing 16 samples (i.e., each 3' barcoded with our 16 3' adenylated barcodes) within a single P2 flowcell for analysis on a NextSeq2000. Considering that the P2 cartridge produces an overage of 400 million reads/run, it provides up to 160 million reads per library and thus up to 10 million reads per sample per library.

Our analyses of small non-coding RNAs and miRNAs isolated from sEVs ultracentrifuged from the plasma of women diagnosed with triple negative breast cancer (Figure 7) demonstrate that our optimized cDNA library procedure, which was initially developed for the analysis of FFPE small non-coding RNAs and miRNAs^35,36, allows reproducible and sensitive analysis of sEV miRNAs. When conducting NGS analyses of only 457 pg of miRNA (i.e., 1.07 ng of total small non-coding RNAs as observed in Figure 8) from an estimated total of 8.4 × 10⁹ sEVs we observed a remarkable reproducibility across replicates as displayed in Figure 10. Furthermore, when using decreasing amounts of sEVs (i.e., Figure 11 from 10¹⁰ to 2.5 × 10⁷ sEVs), we determined that although the number of miRNAs detected is highly relevant to the sEV or RNA input (i.e., Supplemental Figure S2A), using as low as 1.35 pg of total miRNAs (i.e., from 2.5 × 10⁷ sEVs with an estimated total of 3.15 pg of total small non-coding RNAs), our cDNA library preparation procedure followed by NGS analysis provides highly reproducible miRNA profiles (i.e., Supplemental Figure S2B) when compared to those of more abundant RNA samples (i.e., 10 × 10¹⁰ sEVs or 544 pg miRNAs) as observed in Figure 11. Our results are relevant because they highlight both reproducibility and sensitivity of the cDNA library preparation for the analysis of small amounts of sEVs miRNAs. In a recent miRNA sequencing study on circulating sEVs from patients with amyotrophic lateral sclerosis, Kim et al. used 10 ng of total sEV RNA with the SMARTer smRNA-Seq Kit from Clontech to conduct their analyses⁴³.

Llorens-Revull et al. who evaluated the isolation of sEV RNA using different RNA methods, tested three commercially available cDNA library preparation protocols, including NEBNext Multiplex Small RNA Library Prep, the NEXTFlex Small RNA-Seq Kit v3 from PerkinElmer and the SMARTer smRNA-seq kit from Clontech, in which they used between 0.7 ng and 1 ng of total sEV RNA to compare the reproducibility of their RNA isolations⁴⁴. Using the nCounter Human v3 miRNA Expression Assay Kit from Nanostring, which is not a sequencing-based approach, Crossland et al. used a minimum of 845 pg for their miRNA analyses (i.e., 169 pg/µL × 5 µL)⁴⁵. To date, very few studies have been focused on precisely determining the minimum amount of sEV RNA input necessary to conduct small-RNA NGS analyses. This is possibly because not only is each sEV isolation method is different (i.e., UC, SEC, antibody purification, PEG precipitation…), but each biofluid produces different sEVs amounts (i.e., plasma, serum, urine, saliva, bronchoalveolar lavage…), and each commercial cDNA library preparation protocol comes with a minimal requirement for total RNA input (i.e., generally around 1-10 ng). Moreover, it is very difficult to evaluate RNA amounts with such low sEV inputs. Miceli et al., however, recently reported that based on their analyses, they estimated that 1.4 × 10⁸ sEVs were necessary for miRNA NGS analyses⁴⁶. Based on our data, we determined that a minimum of 2.5 × 10⁷ sEVs is currently the lowest threshold to produce miRNA profiles comparable to those of 10¹⁰ sEVs, and that we determined equates to a minimum of 1.35 pg of total miRNA input for cDNA library preparation.

Although our test experiments evaluated miRNAs isolated from plasma sEVs, we can foresee, based on previous NGS analyses conducted with sEVs isolated from diverse biofluids, that our optimized cDNA library preparation procedure has potential to be usable with lower inputs of sEV RNA for small non-coding RNA/miRNA analysis of rare sEV subpopulations. Indeed, as we combined the analysis of 10¹⁰ sEVs and 2.5 × 10⁷, with miRNAs concentration differences by 400-fold, we can foresee that concentrated samples capture more reads than lower concentration samples. Thus, analyzing samples of lower concentrations together may not only improve reproducibility but also the sensitivity of our protocol. We can foresee that conducting miRNA NGS analyses of attogram amounts (i.e., 10^-18 g) of miRNAs from selectively isolated low-abundant sEVs may enable the identification of disease-associated miRNA signatures or the identification of novel disease-associated small non-coding RNAs.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
1% Triton X-100	Invitrogen	HFH10
10 mM ATP	Ambion	AM8110G
10x dNTPs	Ambion	AM8110G
10x TBE	Thermofisher Scientific	15581044
14 M Mercaptoethanol	Sigma	O3445I-100
20 nt ladder	Jena Bioscience	M-232S
20 mg/mL Bovine Serum Albumin	Sigma	B8894-5ML
50x Titanium Taq	Clontech Laboratories	639208
Ammonium Persulfate	Fisher Scientific	7727-54-0
Blue light transilluminator- Safe Imager 2.0	Thermofisher	G6600
BRL Vertical Gel Electrophoresis System with glass plates and combs	GIBCO	V16
Dimethyl sulfoxide (DMSO)	Sigma	D9170-5VL
Eppendorf microcentrifuge 5424R	USA scientific	4054-4537Q
Eppendorf Thermomixer	USA scientific	4053-8223Q
Filter tube with 5mm filter	IST Engineering Inc.	5388-50
Fisherbrand Siliconized Low-Retention Microcentrifuge Tubes 1.5 mL	Fisher Scientific	02-681-320
Gel Breaker Tube 0.5 mL	IST Engineering Inc.	3388-100
Gel electrophoresis apparatus 7 cm x 10 cm- Mini-sub Cell GT with gel trays and combs	Biorad	1704446
Glycoblue	Ambion	AM9516
Illumina NextSeq 1000/2000 sequencer	Illumina	20047256
Jersey-Cote (Silicon-based solution)	LabScientific, Inc	1188
KCl 2 M	Ambion	AM9640G
MgCl2 1 M	Ambion	AM9530G
Minifuge dual rotor personal centrifuge	USA scientific	2641-0016
Model V16 polyacrylamide gel electrophoresis apparatus, glasses, combs, and spacers	Ciore Life Science	21070010
Oligonucleotides	IDT	Defined during order
Owl EasyCast B2 mini electrophoresis system- with gel trays and combs	Thermofisher Scientific	B2
Qiaquick Gel Extraction kit	Qiagen	28704
Qubit Fluorometer	Thermofisher Scientific	Q33238
Restriction enzyme PmeI and 10x Cut Smart Buffer	NEB	R0560S
Reverse transcriptase- Superscript III	ThermoFisher	12574026
RNase-free water	Ambion	AM9932
Shaker- Eppendorf Thermomixer	Eppendorf	5385000024
SeaKem LE agarose	Lonza	50002
Superscript III reverse transcription kit	Invitrogen	18080-044
SYBR Gold	Life Technologies	S11494
Titanium Taq Polymerase	Takara	639208
T4 RNA Ligase 1	NEB	M0204S
T4 RNA Ligase 2 Truncated K227Q	NEB	0351L
TEMED	Fisher Scientific	O3446I-100
Themocycler with heated lid	Applied Biosystem	4359659
Tris 1 M pH 7.5	Invitrogen	15567027
Tris 1 M pH 8.0	Ambion	AM9855G
UltraPure Sequagel system concentrate, diluent, and buffer	National Diagnostics	EC-833