Quantification of Lentiviral Vectors with Nucleic Acid Dyes

Abstract

Lentiviral vectors are widely used in both biological research and therapeutic applications. Accurate and reliable quantification is essential for their effective use, yet existing methods are often technically demanding and time-consuming. Here, we describe a simple and robust approach to measure lentiviral titers by quantifying viral genomic RNA with nucleic acid–binding dyes. We first evaluated several newly developed commercially available fluorescent dyes for their ability to detect RNA and DNA, identifying SYBR Green II and SYBR Gold as the most sensitive for RNA, with detection limits in the low-nanogram range. We further observed that SYBR Green II readily stained nucleic acids in live cells, whereas SYBR Gold was only partially membrane-permeable. Based on these findings, we developed a quantification method combining SYBR Green II with benzonase to selectively detect encapsidated lentiviral RNA. This assay yielded results consistent with those obtained by enzyme-linked immunosorbent assays. Importantly, this SYBR Green II–based method is rapid (<60 min), reliable, and cost-effective, making it a practical tool for titering lentiviral vectors.

Keywords

lentivirus titer titration SYBR Green II SYBR Gold benzonase

INTRODUCTION

Lentivirus is an RNA virus belonging to the retrovirus family.¹ Lentiviral vectors (LVV), which are genetically engineered from lentiviruses, are replication-incompetent but retain the ability to deliver genetic material into a broad spectrum of cells. Since their development, LVV have been widely used in biomedical research and gene therapy.^2–4 Among various viral vector–based delivery systems, LVV offer advantages such as:¹ long-term expression through integration of the gene of interest into the host genome;² the ability to transduce nondividing cells, such as neurons;³ low toxicity; and⁴ high transduction efficiency of broad target cells. These advantages make LVV one of the most efficient vehicles for gene therapy, alongside adeno-associated viruses (AAVs)^5,6 and adenoviruses.^7,8 Compared with AAVs, one key benefit of LVV is their larger packaging capacity (∼10 kb), which enables them to deliver genes that exceed the size limit of AAV vectors (∼4.7 kb).^4,9–11 The first clinical application of LVV was reported in 2007, in a gene therapy trial for X-linked adrenoleukodystrophy, a severe neurodegenerative disorder.¹² To date, over 300 clinical trials involving LVV have been conducted, demonstrating their significant potential for treating a wide range of human diseases.¹³

The production of LVV for research or therapeutic applications—including packaging, concentration, and titration—has been systematically refined and optimized for routine operation in standard laboratory settings.^14–16 Briefly, during the packaging process, viral genomic RNA is enclosed into viral particles within producer cells, typically Human Embryonic Kidney 293 cells expressing the SV40 Large T antigen (HEK293Ts), transfected with lentiviral transfer plasmid and helper plasmids. Following the process of packaging, LVV are released into the culture medium, from which they can be readily concentrated by ultracentrifugation or filtration. Lentiviral preparations can then be titered. Accurate titration is critical for downstream applications. However, the process is inherently complex, owing to RNA-based nature of their genomes,⁷ as compared with titration of DNA-based viruses such as AAV. Despite the challenges, methods have been developed to measure either the physical or functional (infectious) titer of lentiviral preps. To assess the physical titer, a commonly used approach is p24 enzyme-linked immunosorbent assay (ELISA), which detects the HIV-1 p24 capsid protein.^7,17 Alternatively, reverse transcription quantitative PCR (RT-qPCR) and reverse transcription droplet digital PCR (RT-ddPCR) can be used to quantify viral RNA, and the titer is then calculated as LVV particles per milliliter (LP/mL).^18,19 These RT-based methods involve several steps, including extraction of RNA from the viral preparation, reverse transcription into cDNA, and subsequent quantification by qPCR or ddPCR. In comparison with physical titers, functional titers are typically determined by infecting permissive target cells (e.g., HEK293T) with serial dilutions of the lentiviral preparation and measuring the percentage of fluorescent cells (e.g., Green Fluorescent Protein (GFP)–or Red Fluorescent Protein (RFP)–positive) about 48–72 h post infection. The titer is then calculated in transducing units per milliliter (TU/mL).^15,20 However, this method is limited when the LVV lacks a fluorescent reporter or if the promoter driving the reporter is inactive in the target cells. A third approach combines viral infection of host cells with subsequent qPCR or RT-qPCR to quantify the number of integrated lentiviral genomes in host cells or mRNA expression of transgenes from LVV.^21,22 This approach requires first transfecting host cell with LVV, isolating genomic DNA days after infection, and then performing real-time PCR. It provides an alternative means to assess functional titer by measuring proviral integration into the host genome. In summary, although multiple techniques have been successfully used to quantify LVV, they are generally not straightforward to perform, tend to be time-consuming and labor-intensive, and are not cost-effective.

In recent years, a variety of nucleic acid dyes have been developed, including GelGreen and GelRed from Biotium (Fremont, CA), as well as SYBR Safe, SYBR Green II, and SYBR Gold from Thermo Fisher Scientific (Waltham, MA). To establish a simpler method for measuring lentiviral concentrations, we tested whether these newly developed nucleic acid dyes could be repurposed for this application. Our rationale was that dyes with high sensitivity for RNA—since lentivirus uses RNA as their genetic material—could potentially serve as effective tools for quantifying lentiviral preparations.

MATERIALS AND METHODS

LVV production and purification

LVV were produced with the second-generation packaging system in-house. Briefly, HEK293T cells were cotransfected with second-generation lentiviral transfer plasmid, along with PMD2.G (expressing the vesicular stomatitis virus glycoprotein, VSV-G) and PXPAX2 (expressing the HIV gag, pol, rev, and Tat genes),^7,14 using the PEI (Polyethyleneimine) transfection method. Twenty-four hours post transfection, the medium was replaced with fresh medium. Viral supernatants were collected twice—first at 2 days and then at 4 days after transfection. The supernatants were centrifuged at 2,500 × g to remove cellular debris and subsequently filtered through a 0.45 μm filter. The filtered supernatant was then concentrated by ultracentrifugation (2 h, 60,000 g in a swinging-bucket rotor [SW28]; Beckman Ultracentrifuge) and resuspended in 1 × phosphate-buffered saline (PBS), at 4°C, to a volume to achieve a ∼1,000-fold concentration.

DNA standard and RNA standard

Plasmid DNA purified using the Qiagen Maxi Prep Kit (Cat. 12963) was used as the DNA standard in this study. RNA extracted from torula yeast (MilliporeSigma, Burlington, MA; R6625) served as the RNA standard in our experiments, though other RNA sources should also be suitable. The concentrations of both DNA and RNA were measured using a NanoDrop™ spectrophotometer (ThermoFisher).

Reagents

GelGreen (Cat. #41005), GelRed (Cat. #41003), and EMBER500 (Cat. #41032) were purchased from Biotium (Fremont, CA). SYBR Safe (Cat. #S33102), SYBR Green II (Cat. #S7564), and SYBR Gold (Cat. #S11494) were purchased from Thermo Fisher Scientific (Waltham, MA). All dyes were supplied in liquid form, and working dilutions were prepared from original stocks. Hoechst 33342 (Cat. #H1399) was purchased from Thermo Fisher Scientific in powder form. A 10 mg/mL stock solution was prepared, and a 1:10,000 dilution of this stock was used for experiments. Benzonase (Cat. #E1014) was obtained from MilliporeSigma and provided at an activity of 250 U/µL.

Cytation 3 plate reader

Cytation 3 Multi-Mode plate reader from BioTek (Winooski, VT) was used for DNA/RNA binding assay. Reactions were carried out in Greiner 96-well flat-bottom black-wall culture microplates (REF # 655086). Excitation (Ex) and emission (Em) wavelength was adjusted based on individual dyes properties. Ex and Em for different dyes are listed as follows: Gel Green (Ex 485 nm, Em 528 nm); GelRed (Ex 300 nm, Em 600 nm); EMBER500 (Ex 490 nm, Em 530 nm); SYBR Safe (Ex 485 nm, Em 528 nm); SYBR Green II (EX 485 nm, Em 528 nm); SYBE Gold (Ex 497 nm, Em 537 nm).

Titration of LVV by p24 ELISA . LVV titers were determined using the Lenti-X Rapid Titer Kit (Cat. #632200; Takara Bio Inc., San Jose, CA), an ELISA-based assay that measures the total p24 antigen in the sample. The procedure was performed according to the manufacturer’s protocol. Upon completion of the reaction, absorbance at 450 nm was measured using a Cytation 3 plate reader. LVV titers were calculated assuming approximately 2,000 p24 molecules per viral particle. To further measure the free p24 with the Lenti-X Rapid Titer Kit, samples were applied directly to the wells without the addition of lysis buffer, thereby measuring only nonencapsidated p24, as described previously.¹⁹

Analysis of viral RNA by Bioanalyzer . RNA was first extracted from LVV samples using the RNeasy Plus Mini Kit (Cat. #74134; Qiagen, Germany). Purified RNA was then analyzed on an Agilent 2000 Bioanalyzer using the Agilent RNA 6000 Pico Kit (Cat. #5067-1513; Agilent Technologies, Santa Clara, CA), which is designed for low-input RNA analysis, according to the manufacturer’s instructions.

Data analysis

Statistical analyses were conducted with GraphPad Prism software (San Diego, CA). Student’s t-test and one-way Analysis of Variance (ANOVA) with Tukey’s post hoc test are used for data comparisons. Differences were considered significant when p < 0.05. Data are shown as mean ± standard deviation (SD). The limit of detection (LOD) is defined as the mean value of sample blanks plus 3 SDs. The limit of quantification (LOQ) is defined as mean value of sample blanks plus 10 SD (Fig. 3).

The amount of RNA was determined by standard curves. As outlined in the scheme (Fig. 5A), three values were measured experimentally: Vt, the value for total nucleic acids (Fig. 5A1); Vb, the value for benzonase-resistant nucleic acids (Fig. 5A2); and Vhb, the value for heat- and benzonase-resistant nucleic acids (Fig. 5A3). True viral RNA (Vr) was calculated as the difference between Vb and Vhb. In addition, comparing Vb with Vt provides an estimate of contaminating nucleic acids in the LVV preparations. Equation (1) describes how Vr (ng) is converted to LVV titer (LP/mL):

(E q . 1) LVV titer (L P / m l) = \frac{Vr (n g) * (1.0 E - 9) * 6.022 E 23 {mol}^{- 1}}{M W * Volume (m l)}, where

M W = 2 * genome size (n t) * 320 g / mol

Note: The average molecular weight (MW) of a ribonucleotide is approximately 320 g/mol. Because lentiviral genomes consist of two RNA strands, the total MW is calculated by first determining the MW of a single strand and then multiplying by 2. The length of lentiviral RNA typically ranges from 8,000 to 12,000 nucleotides (nt), depending on the size of the insert. For simplicity, the RNA length can be approximated as 10,000 nt. Under this assumption, Equation (1) becomes:

(E q . 2) LVV titer (L P / m l) = \frac{Vr (n g) * (9.4 E 10)}{Volume (i n u l)}

For example, if Vr is measured as 10.5 ng using the experimental scheme in Figure 5B in which the virus sample volume used for measurement was 0.2 μL (equivalent of 10 μL of 1/50 dilution), then the titer of the LVV is calculates as:

\frac{10.5 * 9.4 E 10}{0.2} = 4.94 E + 12 (L P / m l)

RESULTS

Comparative analysis of DNA and RNA detection by GelGreen, GelRed, and EMBER500

Quantitative DNA and RNA binding assays were performed using Cytation 3 plate reader. DNA standards were prepared from plasmid DNA, while RNA standards were prepared from yeast RNA (MilliporeSigma, R6625). Unless otherwise specified, each standard curve consisted of a 12-point serial dilution ranging from 0 to 200 ng, and concentration of dyes was used at 1:10,000 dilution. The standards were dispensed into a 96-well plate, followed by the addition of PBS containing nucleic acid-binding dyes. Fluorescence was then measured to assess DNA/RNA quantities. The Ex and Em wavelength of the Cytation 3 plate reader was adjusted based on individual dyes’ properties.

In the first sets of experiment, we evaluated the abilities of GelGreen, GelRed, and EMBER500, three dyes developed by Biotium company, to detect RNA and DNA. We have previously demonstrated that GelGreen has superior ability to quantify DNA in a larger range, and it can be used to effectively quantify DNA-based viruses.^23,24 But it is not known how GelGreen responds to RNA. Therefore, we compared fluorescence signals resulting from binding of GelGreen (1:10,000) with DNA and RNA simultaneously (Fig. 1). A calibration plot of fluorescence intensity versus nucleic acids quantity is shown in Figure 1A (linear scale). As we reported before, GelGreen displayed strong fluorescence across a wide DNA concentration range (0–200 ng), with excellent linearity (R² > 0.99 for all assays). In contrast, this dye showed a much weaker response to RNA. For equivalent nucleic acid amounts, the fluorescence signal for DNA was approximately 20-fold higher than that for RNA (Fig. 1A). To examine the responses more closely, we analyzed fluorescence signals for nucleic acid amount up to 16 ng (log scale). While GelGreen demonstrated remarkable sensitivity for detecting DNA in the low-nanogram range, it could only detect RNA at concentrations above approximately 10 ng (Fig. 1B). Therefore, although GelGreen is highly effective for detecting DNA and DNA viruses such as AAV, as we have shown previously, it is much less suitable for detecting RNA or RNA-based viruses.

Figure 1.

Detection of DNA and RNA by GelGreen, GelRed, and EMBER500. (A, B) GelGreen calibration plots of fluorescence intensity versus DNA or RNA. (A) Linear plot (0–200 ng). (B) Log-scale plot (0–16 ng). (C, D) GelRed calibration plots. (C) Linear plot (0–200 ng). (D) Log-scale plot (0–16 ng). (E, F) EMBER500 calibration plots. (E) Linear plot (0–200 ng). (F) Log-scale plot (0–16 ng). All dyes were used at a 1:100,000 dilution in 1X phosphate-buffered saline (PBS). Results represent the mean of triplicate wells, with error bars indicating standard deviation (SD).

Next, we evaluated GelRed. According to Biotium, this dye exhibits approximately 5-fold greater sensitivity for single-stranded DNA than for double-stranded DNA and is reported to be effective in staining RNA. These properties suggest that GelRed may be a suitable candidate for detecting RNA and RNA-based viruses. In our second experiment, we assessed the performance of GelRed. Calibration plots of fluorescence intensity versus RNA and DNA are shown in Figure 1C (linear scale). At a 1:10,000 dilution, GelRed demonstrated a strong response to DNA concentrations (0–200 ng). In contrast, its response to RNA is markedly weaker—DNA produces approximately 6-fold higher fluorescence than RNA at equivalent concentrations—a trend similar to GelGreen. A major limitation of GelRed in plate reader application was its high background fluorescence—about 10-fold higher than that of GelGreen (comparing Fig. 1A vs. 1C and 1B vs. 1D)—which hinders accurate quantification of DNA or RNA. When focusing on the low-concentration range (Fig. 1D, log scale), GelRed detected DNA at a threshold of 2 to 3 ng and RNA at approximately 10 ng. Overall, GelRed displayed lower sensitivity to both DNA and RNA compared with GelGreen.

Since neither GelGreen nor GelRed proved suitable for our purposes, we next evaluated EMBER500, a dye reported to offer high sensitivity and resolution for RNA detection in gels. However, when tested with a plate reader, EMBER500 performed poorly. The fluorescence response to DNA was weak and was nonlinear, while the response to RNA was virtually undetectable—even at high RNA concentrations (Fig. 1E, F). It was somewhat surprising that EMBER500 failed to detect either RNA or DNA under these conditions. The cause of this failure remains unclear, but it is possible that EMBER500 may degrade rapidly in solution or that its fluorescence signal is subject to rapid photobleaching, limiting its detection capability and usefulness. Regardless, EMBER500 is not suitable for our analysis using a plate reader-based approach.

Comparative analysis of DNA and RNA detection by SYBR Safe, GelGreen II, and SYBR Gold

Next, we evaluated three fluorescent dyes from ThermoFisher: SYBR Safe, SYBR Green II, and SYBR Gold. Similar to GelGreen, SYBR Safe is commonly used for DNA staining and is considered a safer alternative to ethidium bromide. SYBR Green II is optimized for RNA detection according to the manufacturer. SYBR Gold, also per the manufacturer, is the most sensitive dye in the SYBR family for detecting both DNA and RNA. We compared the performance of these dyes in detecting DNA and RNA using plate reader. First, for SYBR Safe, its binding curves showed a modest linear fluorescence response to DNA across the range of 0–200 ng. However, its response to RNA was significantly weaker—approximately 20-fold lower than the signal for DNA at equivalent concentrations. Figure 2B further shows that SYBR Safe detected DNA in the low-nanogram range and detected RNA at ∼10 nanogram range. In contrast, SYBR Green II exhibited a more robust fluorescence response to RNA, approximately one-fourth of the signal observed for DNA—indicating improved sensitivity for RNA detection (Fig. 2C, D). Overall, SYBR Green II detects DNA at low-nanogram range and RNA at nanogram range (Fig. 2D). Last, we tested SYBR Gold. It yielded the strongest fluorescence signal for DNA, with the signal at 100 ng and 200 ng exceeding the plate reader’s upper detection limit (Fig. 2E). For RNA, SYBR Gold also slightly outperformed SYBR Green II, producing a signal roughly one-third of that for DNA—making it the most sensitive dye we tested for RNA detection (Fig. 2E, F). Upon close examination at low concentrations, we estimate that SYBR Gold can detect DNA in the low-nanogram range and RNA at approximately 1 ng, comparable with the sensitivity of SYBR Green II (Fig. 2F).

Figure 2.

Detection of DNA and RNA by SYBR Safe, SYBR Green II, and SYBR Gold. (A, B) DNA and RNA quantifications with SYBR Safe: (A) Linear plot with DNA or RNA ranging from 0 to 200 ng. (B) Log-scale plot (0–16 ng). (C, D) DNA and RNA quantifications with SYBR Green II. (C) Linear plot (0–200 ng). (D) Log-scale plot (0–16 ng). (E, F) DNA and RNA quantifications with SYBR Green II SYBR Gold. (E) Linear plot (0–200 ng). (F) Log-scale plot (0–16 ng). All dyes were used at 1:100,000 dilution in 1X PBS. Results represent the mean of triplicate wells, with error bars indicating SD.

Based on these results, SYBR Green II and SYBR Gold emerged as the most promising candidates for RNA detection. We therefore focused on these two dyes and conducted additional experiments to more precisely assess their sensitivities. RNA binding assays were performed to determine the LOD and LOQ for each dye. In our experiments, the LOD was defined as the mean value of sample blanks plus 3 SDs, while the LOQ was defined as the mean value of sample blanks plus 10 SDs. For SYBR Green II, the LOQ and LOQ for RNA were approximately 0.5 ng and 1.9 ng, respectively (Fig. 3A). SYBR Gold showed slightly greater sensitivity, with a comparable LOQ of ∼0.4 ng and a lower LOD of 1.15 ng (Fig. 3B). Overall, both SYBR Green II and SYBR Gold demonstrated strong potential for quantifying lentiviral preparations, with SYBR Gold offering a modest sensitivity advantage.

Figure 3.

RNA detection and quantification limit of SYBR Green II and SYBR Gold assays. (A) SYBR Green II calibration plot (log scale) showing fluorescence intensity in response to RNA (0–16 ng). (B) SYBR Gold calibration plot (log scale) showing fluorescence intensity in response to RNA (0–16 ng). Results represent the mean of triplicate wells, with error bars indicating SD. The limit of detection (LOD) was defined as the mean value of sample blanks plus 3 SD, and the limit of quantification (LOQ) as the mean value of sample blanks plus 10 SD.

Assessment of membrane permeability among fluorescent dyes

For nucleic acid dyes to stain viral genomes, they must be able to access the viral genetic material, which is protected by viral capsids. For example, when using GelGreen to stain AAV genome, viral inactivation—by heat or chemical treatment—is required because GelGreen is not membrane-permeable and cannot penetrate the AAV capsid.²³ In contrast, some dyes can stain encapsidated viral genomes without prior inactivation. To determine whether viral inactivation is necessary for SYBR Green II and SYBR Gold, we first tested their membrane permeability. Each dye (1:10,000) was added to culture media, incubated with live HEK cells for 10 min, and imaged by fluorescence microscopy. GelGreen showed no detectable staining (Fig. 4A1), consistent with its membrane-impermeable properties. SYBR Safe produced weak fluorescence (Fig. 4A4), while SYBR Gold generated an even weaker signal (Fig. 4A2), indicating limited permeability. By contrast, SYBR Green II yielded the strongest fluorescence in live cells (Fig. 4A5), demonstrating that it is membrane permeable. To further confirm this, we repeated the assay at 1:1,000 dilution. Again, SYBR Gold produced much weaker fluorescence than SYBR Green II (compare Fig. 4A3 with 4A6), though it remained detectable, supporting the conclusion that SYBR Gold is only partially membrane permeable.

Figure 4.

Membrane permeabilities of different dyes and effects of benzonase digestion on nucleic acid binding assays. (A) Representative fluorescence micrographs of HEK cells stained with nucleic acid dyes. Live HEK cells were incubated with each dye at the specified concentration for 10 min and subsequently imaged using a fluorescence microscope equipped with a GFP (Green Fluorescent Protein) filter. Scale bar, 100 µm. (B) Images of HEK cells stained with Hoechst 33342 (B1, 1:10,000 dilution from a 10 mg/mL stock), SYBR Green II (B2, 1/10,000), and both dyes (B3). HEK cells grown on glass coverslips were incubated with each dye at the indicated concentrations for 10 min, then fixed with 4% paraformaldehyde (PFA), mounted on glass slides, and imaged using a Nikon fluorescence microscope with a 60 × objective. Arrows in B2 and B3 indicate nucleoli. Scale bar, 15 μm. (C) Effects of benzonase digestion on binding assays. DNA (10 ng), RNA (10 ng), and a DNA/RNA mixture (5 ng each) were treated with benzonase (40 units) for 20 min at room temperature, followed by SYBR Green II binding assay. Benzonase treatment significantly reduced fluorescence signals in all conditions (DNA, RNA, or RNA/DNA mixture) (DNA vs. DNA B+; RNA vs. RNA B+; DNA/RNA vs. DNA/RNA B+, t-test, p < 0.001). Residual fluorescence remained slightly above PBS background, and all treated groups differed significantly from PBS controls (t-test, p < 0.001); n = 4 per group; error bars represent SEM; *, p < 0.001. SEM, standard error of the mean.

To further demonstrate that SYBR Green II can stain nucleic acids after permeating living cells, we performed additional experiments directly comparing SYBR Green II with Hoechst 33342, which, at a 1:10,000 dilution from a 10 mg/mL stock, specifically labels nuclei. The results are shown in Figure 4B. As expected, Hoechst 33342 stained only the nuclei (Fig. 4B1). In contrast, SYBR Green II stained both the nuclei and cytoplasms (Fig. 4B2, B3). Notably, in cells in which the nucleolus was within the focal plane, SYBR Green II also produced strong nucleolar staining, suggesting that it labels RNA within the nucleoli (Fig. 4B2, B3). Together, these results provide further evidence that SYBR Green II can stain both DNA and RNA after entering living cells.

Benzonase digestion of nucleic acids

Because SYBR Gold is partially membrane-permeable, interpreting results from a viral RNA binding assay would be challenging. Therefore, we decided to base our strategy on SYBR green II, which is membrane permeable and thus expected to bind both encapsidated viral genomes and nonencapsidated nucleic acid contaminations. To distinguish the two sources, we decided to apply a nuclease to selectively degrade unprotected nucleic acids before staining with SYBR Green II. In contrast, lentiviral RNA is enclosed within both a capsid and an envelope, and it should initially resist nuclease digestion. Among commercially available nucleases, benzonase is perhaps the most potent, as it degrades both DNA and RNA into short fragments or even nucleotides. Thus, benzonase appeared to be a good candidate for this purpose; however, it was unclear whether its digestion products of RNA or DNA would still bind SYBR Green II and generate fluorescence. We reasoned that if the resulting fragments could no longer bind SYBR Green II, the fluorescence signal would drop to background levels; otherwise, residual signal would remain. To test this, we treated DNA (10 ng), RNA (10 ng), and a DNA/RNA mixture (5 ng each) with benzonase (∼40 units; 0.15 µL of 250 U/µL benzonase) for 20 min at room temperature, followed by SYBR Green II binding (Fig. 4C). Note that according to unit definition of the enzyme, 1 unit of benzonase can digest up to 30 μg of DNA in 30 min. Therefore, under our condition, the enzyme was used in substantial excess. Benzonase treatment markedly reduced fluorescence in all conditions (RNA, DNA, and RNA/DNA mix); however, the signals were not completely eliminated and remained slightly above the background observed in the PBS; nevertheless, the differences compared with PBS group were all statistically significant (p < 0.001, t-test) (Fig. 4C). These results suggest that residual oligonucleotides may still bind SYBR Green II and contribute to background fluorescence.

Experimental strategy for measuring LVV titers

Taking all these considerations into account, we designed a strategy to measure LVV titers (Fig. 5A). The scheme includes three experimental conditions:¹ No treatment: SYBR Green II is directly mixed with LVV. In this condition, SYBR Green II binds both encapsidated viral RNA and any free-floating nucleic acids, generating a total (gross) fluorescence signal (Vt) (Fig. 5A1).² Benzonase treatment: The LVV is treated with benzonase prior to the addition of SYBR Green II. Benzonase digests unprotected nucleic acids, while viral RNA—protected by the viral envelope and capsid—remains intact. Fluorescence in this condition (Vb) reflects encapsidated viral RNA and short oligonucleotide fragments generated by benzonase digestion (Fig. 5A2).³ Heating and benzonase treatment: The LVV is first heated at 90°C for 10 min to disrupt the viral envelope and capsid, then treated with benzonase before adding SYBR Green II. The heating condition of 90°C for 10 min was selected based on previous studies. In particular, a recent report systematically evaluated multiple heat-inactivation conditions and demonstrated that 90°C for 10 min provided optimal performance for one-step RT-ddPCR detection of LVV.¹⁹ Under this condition (Fig. 5A3), the previously protected viral RNA is exposed and digested, and the resulting fluorescence (Vhb) reflects only residual oligonucleotides. By subtracting Vhb from Vb, we can selectively derive viral RNA (Vr), which corresponds to true encapsidated viral genomes. In addition, comparing the value of Vt with the Vb allows us to estimate the level of contaminating nucleic acids in the LVV.

Figure 5.

Designing strategy to measure lentiviral vector (LVV) titers with SYBR green II combined with benzonase treatment. (A) Experimental scheme. Three conditions were tested: (A1) No treatment: SYBR Green II is added directly to the LVV. In this condition, the dye binds both encapsidated viral RNA and free nucleic acids, producing the total (gross) fluorescence signal (Vt). (A2) Benzonase treatment: The LVV is treated with benzonase prior to SYBR Green II addition. Benzonase digests unprotected nucleic acids, while viral RNA—shielded by the viral envelope and capsid—remains intact. Fluorescence in this condition (Vb) reflects encapsidated viral RNA and short oligonucleotide fragments generated by benzonase digestion. (A3) Heat and benzonase treatment: The sample is first heated at 90°C for 10 min to disrupt the viral envelope and capsid, followed by benzonase treatment before dye addition. Under this condition, previously protected viral RNA is exposed and degraded, and the remaining fluorescence (Vhb) reflects, presumably, only residual oligonucleotides. (B) Experimental procedure. Left panel: RNA standard. Yeast RNA (MilliporeSigma, R6625) was serially diluted to prepare the RNA standard. Right panel: sample preparation. A small volume (e.g., 1 µL) is divided into three aliquots for subsequent processing: one without treatment, one with benzonase treatment, and one subjected to heat inactivation followed by benzonase treatment. SYBR Green II assay is then carried out in a plate reader.

The streamlined experimental procedure is illustrated in Figure 5B. Briefly, 1 µL of LVV is diluted in 50 µL of PBS and divided into three tubes (15 µL each) (Fig. 5B, right panel). The first tube receives no treatment. The second tube is treated with benzonase (0.15 µL at 250 U/µL) for 20 min. The third tube is heated at 90°C for 10 min to disrupt the viral capsid, followed by benzonase digestion. After the treatments, 10 µL from each tube is transferred into a 96-well plate. Each well then receives 90 µL of PBS containing SYBR Green II (1:10,000 dilution), and fluorescence is measured using a plate reader. In parallel, an RNA standard set is prepared, consisting of a 12-point serial dilution (0–200 ng) (Fig. 5B, left panel), which is plated and read under identical conditions as LVV. The entire assay can be completed in less than 60 min.

Titration of LVV with SYBR Green II

By following the protocol as in Figure 5B, we measured two LVV samples (LVV #1 and LVV #2) produced in-house. Our results are summarized in Figure 6A, B. For LVV #1, we observed a raw fluorescence signal of 28,000 AU (Vt), which is high, as we discuss in more detail later. After benzonase treatment, the signal was reduced to approximately 9,000 AU (Vb), corresponding to 51 ng of RNA based on the standard curve. With additional heat treatment followed by benzonase digestion, the signal further decreased to 7,000 AU (Vhb) (equivalent to 40.6 ng of RNA). Thus, the difference—representing encapsidated viral RNA—was 10.4 ng (Fig. 6B) from 10 µL of 1:50 dilution of LVV #1. Using Eq. (1) (or its simplified form, Eq. 2), we calculated the titer of LVV #1 to be 5.0 E + 12 LP/mL. For LVV #2, the untreated sample yielded a fluorescence signal of 11,000 AU (Vt). After benzonase treatment, the signal decreased to 4,500 AU (Vb), and with benzonase added after heat treatment, the signal further dropped to 3,700 AU (Vhb). Based on these values, the amount of encapsidated RNA was calculated as 30–26.5 = 3.5 ng (Fig. 6B), corresponding to a titer of 1.65 E + 12 LP/mL for LVV #2.

Figure 6.

Titration of LVV by SYBR Green II. (A) Bar graphs showing the effects of benzonase treatment and heat inactivation for two samples: LVV #1 (left) and LVV #2 (right). Error bars represent standard error of the mean (SEM). (B) Calibration plot of fluorescence intensity versus RNA standard. The RNA standard was generated by serial dilution. Vertical dashed lines indicate the fluorescence values of two LVV samples before and after benzonase treatment. (C) Titration of additional LVV samples. Bar graphs show results from three independent datasets corresponding to LVV #3–5. For each sample set, signals from nonheated Benzonase-treated samples (Vb) are compared with those subjected to 90°C heating followed by Benzonase treatment (Vhb). Error bars represent SEM.

For comparison, we also measured the same two samples using the ELISA method (Lenti-X™ p24 Rapid Titer Kit, Takara, Cat. 632200). One factor that affects accurate titer estimation by ELISA method is the presence of free p24 in lentiviral supernatants. Free p24 primarily arises from overexpression of Gag protein from packaging plasmids in the producer cells, but it can also result from degradation of viral particles. Consequently, the fraction of free p24 relative to total p24 varies depending on packaging conditions, purification procedures, and storage history. When free p24 constitutes a significant proportion of the total signal, ELISA-based assays can substantially overestimate viral titers. To address this issue, we quantified free p24 in our preparations using the same Lenti-X™ kit, and we used the same method as described before¹⁹ and calculated its proportion relative to total p24. We found that the fractions of free p24 differed substantially between the two samples: free p24 accounted for approximately 20% of total p24 in LVV #1, whereas the free p24 fraction in LVV #2 was much lower (2.3%). These values were subsequently used to correct viral titers originally calculated from total p24 measurements (adjusted Titer, Table 1). We then compared results obtained using the SYBR Green II and ELISA methods, as summarized in Table 1. For both samples, the two approaches yielded comparable titers (LVV #1: 4.89 E + 12 vs. 3.99 E + 12; LVV #2: 1.65 E + 12 vs. 1.46 E + 12), demonstrating good agreement between the SYBR Green II–based assay and the ELISA method.

Table 1.

Comparing Results from SYBR Green II and ELISA Methods

Sample	SYBR Green II titer (LP/mL)	ELISA (total p24) (LP/mL)	ELISA Fraction of free p24	ELISA adjusted titer (LP/mL)
LVV #1	4.89 E + 12	5.0 E + 12	20.2%	3.99 E + 12
LVV #2	1.65 E + 12	1.5 E + 12	2.3%	1.46 E + 12

Titers of two LV vectors were determined using both the SYBR Green II method and the ELISA method.

ELISA, enzyme-linked immunosorbent assay.

To further evaluate the reliability of the SYBR Green II assay, we carried out more tests, using LVV #1 as an example. We conducted three independent assays, with each assay performed in triplicates. Data from all three experiments are summarized in Table 2. Intra-assay analysis was performed, and it revealed low coefficients of variation (CV), with the highest being 14.9% and the lowest being 6.4%. The interassay analysis of three independent experiments showed mean titer of 5.9 × 10¹², with low 95% confidence interval (CI) and high 95% CI to be 1.77 E + 12 and 1.01 E + 13, respectively, and with interassay CV to be 28.2%. Taken together, CVs of both interassay and intra-assay are quite low, indicating high repeatability and reproducibility of the method (Table 2).

Table 2.

Titration of LVV #1

Intra assays
	Replicate 1	Replicate 2	Replicate 3	Mean	SD	Low 95% CI	High 95% CI	CV%
Assay #1	6.92 E + 12	6.67 E + 12	5.87 E + 12	6.24 E + 12	4.02 E + 11	5.25 E + 12	7.25 E + 11	6.4
Assay #2	8.02 E + 12	6.60 E + 12	7.63 E + 12	7.44 E + 12	6.94 E11	5.71 E + 12	9.16 E + 11	9.3
Assay #3	4.10 E + 12	4.74 E + 12	3.56 E + 12	4.13 E + 12	5.91 E + 11	2.67 E + 12	5.61 E + 12	14.9

Inter assays
Mean	SD	Low 95% CI	High 95% CI	CV%
5.94 E + 12	1.67 E + 11	1.77 E + 12	10.10 E + 12	28.2

Titer of the LVV #1 was determined by SYBR Green II method. Three independent assays were conducted, with each assay being performed in triplicates. CV, coefficient of variation; CI, confidence interval.

Finally, we examined three additional in-house LVV preparations (LVV #3–5). As shown in Figure 6C, each sample displayed differences between Vb and Vhb. Using these differences, we calculated titers for all three preparations, which all fell within the range of 1.0 E + 12–1.0 E + 13 LP/mL (Fig. 6C). Thus, all five in-house preparations yielded titers spanning from 1.0 E + 12 to 1.0 E + 13 LP/mL, confirming the robustness of the standard LVV packaging protocol widely used in the field.¹⁴

Validation of LVV genome size, integrity, and expression

While the SYBR Green II assay described above provides a means to estimate total RNA quantity, it does not report the length or integrity of the packaged RNA, nor does it indicate whether the virus is capable of transducing target cells. Therefore, we performed an additional experiment to estimate RNA size and assess genome integrity using a Bioanalyzer Pico RNA assay. Viral RNA was first purified from concentrated LVV stocks using the Qiagen RNeasy Plus Mini Kit and subsequently analyzed on a Bioanalyzer. Two LVV samples (LVV #1 and LVV #2) were examined. Both samples displayed a prominent broad smear/band at positions above ∼6,000 nt (Fig. 7A, Lane 1 and Lane 2), which was absent in the RNA control sample (Lane 4). Instead, RNA control sample exhibited two strong bands corresponding to 28S and 18S rRNA. Notably, 28S and 18S bands were also detectable in both LVV, indicating the presence of some level of cellular RNA contamination in our LVV preparations. In addition, lower–molecular-weight bands (25–500 nt) were observed in both LVV #1 and LVV #2, consistent with truncated RNA genome¹⁹ or RNA degradation products, although these signals (marked by arrows in Fig. 7B) were much weaker than the high-molecular-weight viral RNA smear (marked by “∗” in Fig. 7B). Based on the transfer plasmids used to generate these two LVVs, the expected LTR-to-LTR (Long Terminal Repeat) genome sizes are approximately 6,000 nt for LVV #1 and 8,000 nt for LVV #2, respectively. Consistent with this, LVV #2 migrated at a higher apparent MW than LVV #1 (Fig. 7A). Although the precise genome length could not be determined—likely because the LVV RNA genome approaches the upper detection limit of the Bioanalyzer (∼9,000 nt) and because LVV RNA may not be fully denatured under standard Bioanalyzer conditions—our results support the presence of high-molecular-weight RNA consistent with packaged LVV genomes. Overall, the Bioanalyzer Pico RNA assay provided important information regarding the relative length and integrity of the packaged RNA.

Figure 7.

Validation of LVV genome size, integrity, and expression. (A) Bioanalyzer Pico RNA assay of viral preparations. Both LVVs show a prominent broad smear/peak at positions above ∼6,000 nt, consistent with high–molecular-weight packaged RNA. They also show obvious bands of 28S and 18S RNA. (B) Quantitative analysis of the RNA profiles. Arrows indicate low–molecular-weight bands (25–500 nt) in LVV #1 and LVV #2, which likely represent truncated RNA genome and RNA degradation products. The more prominent high-molecular-weight bands (6,000–8,000 nt) (marked by “∗”) are expected to correspond to intact viral genomes. (C) In vivo functional expression of LVV #1. LVV #1 (0.2 µL) was stereotaxically injected into the dentate gyrus of the mouse hippocampus (C1). This LVV vector uses the C1ql2 promoter to drive expression of channelrhodopsin fused to tdTomato in dentate gyrus granule cells, whose axons project via mossy fiber to CA3 regions. Robust tdTomato expression was observed in dentate gyrus region and along mossy fiber tracks (C2 to C3), confirming successful transduction and functional expression of the LVV #1. The boxed region in C2 is shown at higher magnification in C3.

As a final validation, we performed an in vivo functional assay. We stereotaxically injected 0.2 µL of LVV #1 into the dentate gyrus of the mouse hippocampus (Fig. 7C1). This vector uses the C1ql2 promoter to drive expression of channelrhodopsin fused to tdTomato in granule cells, and its titer was estimated to be approximately 5.0E + 12 LP/mL earlier. This injection resulted in robust tdTomato expression in dentate gyrus granule cells (Fig. 7C2–3), further confirming the functional activity of the LVV packaged in-house.

DISCUSSION

Here, we report a method to quantify LVV based on the binding of viral RNA to a nucleic acid dye, SYBR Green II. This approach offers several advantages. First, it is fast, allowing LVV titration to be completed in under 1 h. Second, it is simple and consistent, with low inter- and intra-assay variability. We attribute this consistency to the direct measurement of nucleic acids without amplification steps, which are commonly required in PCR-based and ELISA-based assays. Eliminating enzymatic reactions reduces the assay’s susceptibility to variations introduced by enzyme efficiency, buffer conditions, and reaction kinetics. The main limitation of this SYBR Green dye-based method is its relatively low sensitivity compared with qPCR or ELISA. Given the LOD of SYBR Green II for RNA is approximately 0.5 ng, we estimate that the lowest titer detectable by this method is ∼0.5 × 10¹⁰ LP/mL by using 10 μL of viral sample. Since standard lentiviral packaging protocols—as used in our laboratory—typically yield titers about two orders of magnitude higher than this threshold (Fig. 6), we believe that despite being less sensitive than qPCR or ELISA, the SYBR Green II–based method provides more than sufficient sensitivity for quantifying lentiviral preparations for most standard laboratory applications. A second limitation is that this method measures physical titer rather than functional titer and thus does not directly reflect infectivity. Therefore, whenever possible, determining functional titers is still recommended; however, this may not be feasible when the LVV lack a fluorescent reporter or their protomers are inactive in the host cells.

In recent years, many nucleic acid dyes have been developed. These dyes are mostly used for staining DNA and RNA in gels, but their sensitivities for RNA and DNA detection in solution—especially using plate readers—have not been systematically explored. In this study, we evaluated several dyes for their ability to detect nucleic acids and relative sensitivity to RNA and DNA. GelGreen, SYBR Green II, and SYBR Gold all performed well, with SYBR Gold showing the highest sensitivity, capable of detecting RNA in the low-nanogram range. In contrast, EMBER500 showed poor performance in plate reader–based assays, and GelRed, although sensitive to DNA and single-stranded RNA, produced high background fluorescence. SYBR Safe was about 10-fold less sensitive than SYBR Green II and SYBR Gold. All of the six dyes we tested showed higher sensitivity to DNA than RNA, producing stronger fluorescence when bound to DNA than to an equivalent amount of RNA. To date, only a limited number of dyes have been reported to exhibit preferential binding to RNA rather than DNA.²⁵ In the future, developing more dyes that selectively bind RNA—but not DNA—would be highly desirable for detecting RNA-based viruses. Although such dyes are not yet available, our results (Fig. 2) indicate that SYBR Green II and SYBR Gold remain among the most effective options for RNA detection, despite their preferential binding to DNA.

An unexpected finding was the presence of a substantial amount of free-floating nucleic acids in our LVV preparations (Fig. 6A, compare Vt vs. Vb). In every sample examined, we observed that the Vt signal was approximately three times higher than the Vb signal, indicating that free nucleic acids were present at several-fold greater levels than encapsidated viral RNA. We suspect that most of these free-floating nucleic acids originate from host cell RNA and DNA, although RNA released from damaged viral particles may also contribute. Our LVV concentration protocol—widely used in the field¹⁴—involves harvesting the culture supernatant, clearing by low-speed centrifugation, filtering through a 0.45 μm membrane, and ultracentrifugation to pellet the virus. Apparently, these purification steps were not sufficient to remove those free-floating nucleic acids, based on our study. The presence of such contaminants is undesirable, as they could interfere with experimental outcomes, particularly in the in vivo applications, and pose safety concerns for clinical use.²⁶ Notably, this type of contamination cannot be readily detected by conventional titration methods such as ELISA or qPCR. In clinical trials, initiation of innate immune responses following the LVV administration has been reported in multiple cases.^27,28 Of course, many factors—including the individual’s age, general health, the specific transgene, its expression level, the route of administration, and the target cell type—can all influence the host immune response. Here, we suggest that free-floating nucleic acids present in the viral preparation, which are generally not examined, should also be considered as one factor. Our data indicate that nucleic acid contamination in the LVV samples produced by widely used protocol in the field is more prevalent than previously appreciated and may impact downstream applications, including gene therapy. Therefore, this SYBR Green II–based assay not only enables quantification of encapsidated viral RNA but also provides a valuable metric for evaluating sample purity—an often-overlooked parameter in LVV production.

In conclusion, we present a simple method for measuring LVV titers using SYBR Green II dye. This approach produces results consistent with ELISAs while offering the advantages of being rapid (∼60 min) and reliable and simultaneously measuring impurities, making it a practical option for lentiviral titration. Moreover, future improvements in dye chemistry may further enhance sensitivity by lowering detection limits.

AUTHORS’ CONTRIBUTIONS

J.X., Y.Z., and A.C. designed experiments. J.X., N.Z., and M.C. collected data. J.X., Y.Z., and A.C. analyzed data and wrote the article.

Footnotes

ACKNOWLEDGMENTS

The authors thank Northwestern University Analytical BioNanotechnology Equipment Core facility of the Simpson Querrey Institute for providing Cytation 3 plate readers and for technical support and NUseq Core facility for providing Agilent Bioanalyzer and technical support.

AUTHOR DISCLOSURE STATEMENT

No competing financial interests exist.

FUNDING INFORMATION

This work was supported by NIH R01 EY030169, NIH R01 EY032506, and NIH R01 EY018204 to Y.Z.; NIH R01 MH099114, NIH R01 EY032506 and NIH R01 NS115471 to A.C.; Research to Prevent Blindness, Inc.; and LC industries Foundation.

References

1. Buchschacher

, Wong-Staal

. Development of lentiviral vectors for gene therapy for human diseases. Blood 2000;95(8):2499–2504.

2. Verma

, Somia

. Gene therapy – promises, problems and prospects. Nature 1997;389(6648):239–242.

3. Naldini

. Lentiviruses as gene transfer agents for delivery to non-dividing cells. Curr Opin Biotechnol 1998;9(5):457–463.

4. Naldini

, Blömer

, Gallay

, et al. In vivo gene delivery and stable transduction of nondividing cells by a lentiviral vector. Science 1996;272(5259):263–267.

5. Wang

, Gessler

, Zhan

, et al. Adeno-associated virus as a delivery vector for gene therapy of human diseases. Signal Transduct Target Ther 2024;9(1):78.

6. Flotte

, Carter

, Conrad

, et al. A phase I study of an adeno-associated virus-CFTR gene vector in adult CF patients with mild lung disease. Hum Gene Ther 1996;7(9):1145–1159.

7. Bulcha

, Wang

, Ma

, et al. Viral vector platforms within the gene therapy landscape. Signal Transduct Target Ther 2021;6(1):53.

8. Lee

, Bishop

, Zhang

, et al. Adenovirus-Mediated gene delivery: Potential applications for gene and cell-based therapies in the new era of personalized medicine. Genes Dis 2017;4(2):43–63.

9. Johnson

, Alvarado

, Moffatt

, et al. HIV-based lentiviral vectors: Origin and sequence differences. Mol Ther Methods Clin Dev 2021;21:451–465.

10.

10. Zufferey

, Nagy

, Mandel

, et al. Multiply attenuated lentiviral vector achieves efficient gene delivery in vivo. Nat Biotechnol 1997;15(9):871–875.

11.

11. Dull

, Zufferey

, Kelly

, et al. A third-generation lentivirus vector with a conditional packaging system. J Virol 1998;72(11):8463–8471.

12.

12. Cartier

, Hacein-Bey-Abina

, Bartholomae

, et al. Hematopoietic stem cell gene therapy with a lentiviral vector in X-linked adrenoleukodystrophy. Science 2009;326(5954):818–823.

13.

13. Jadlowsky

, Leskowitz

, McKenna

, et al. Long-term stability of clinical-grade lentiviral vectors for cell therapy. Mol Ther Methods Clin Dev 2024;32(1):101186.

14.

14. Tiscornia

, Singer

, Verma

. Production and purification of lentiviral vectors. Nat Protoc 2006;1(1):241–245.

15.

15. Kutner

, Zhang

, Reiser

. Production, concentration and titration of pseudotyped HIV-1-based lentiviral vectors. Nat Protoc 2009;4(4):495–505.

16.

16. Brown

, Dong

, Kantor

. An improved protocol for the production of lentiviral vectors. STAR Protoc 2020;1(3):100152.

17.

17. Sena-Esteves

, Gao

. Titration of lentivirus vectors. Cold Spring Harb Protoc 2018;2018(4).

18.

18. Carmo

, Peixoto

, Coroadinha

, et al. Quantitation of MLV-based retroviral vectors using real-time RT-PCR. J Virol Methods 2004;119(2):115–119.

19.

19. He

, Kwee

, Cleveland

, et al. Quantitation and integrity evaluation of RNA genome in lentiviral vectors by direct reverse transcription-droplet digital PCR (direct RT-ddPCR). Sci Rep 2023;13(1):14470.

20.

20. Zhang

, La Russa

, Reiser

. Transduction of bone-marrow-derived mesenchymal stem cells by using lentivirus vectors pseudotyped with modified RD114 envelope glycoproteins. J Virol 2004;78(3):1219–1229.

21.

21. Lizee

, Aerts

, Gonzales

, et al. Real-time quantitative reverse transcriptase-polymerase chain reaction as a method for determining lentiviral vector titers and measuring transgene expression. Hum Gene Ther 2003;14(6):497–507.

22.

22. Barczak

, Suchorska

, Rubiś

, et al. Universal real-time PCR-based assay for lentiviral titration. Mol Biotechnol 2015;57(2):195–200.

23.

23. Xu

, DeVries

, Zhu

. Quantification of adeno-associated virus with safe nucleic acid dyes. Hum Gene Ther 2020;31(19–20):1086–1099.

24.

24. Xu

, Contractor

, Zhu

. Protocol to map multi-transmitter neurons in the mouse brain using intersectional strategy. STAR Protoc 2022;3(4):101907.

25.

25. Higuchi

, Sato

, Togashi

, et al. Bright and light-up sensing of Benzo[c,d]indole-oxazolopyridine Cyanine Dye for RNA and its application to highly sensitive imaging of nucleolar RNA in living cells. ACS Omega 2022;7(27):23744–23748.

26.

26. Shirley

, de Jong

, Terhorst

, et al. Immune responses to viral gene therapy vectors. Mol Ther 2020;28(3):709–722.

27.

27. Wang

Q-S

, Wang

, Lv

H-y

, et al. Treatment of CD33-directed chimeric antigen receptor-modified T cells in one patient with relapsed and refractory acute myeloid leukemia. Mol Ther 2015;23(1):184–191.

28.

28. Lund

, Orchard

, Nascene

, et al. Secondary failure of lentiviral vector gene therapy in a cerebral adrenoleukodystrophy patient with an ABCD1 whole-gene deletion. Mol Ther 2024;32(10):3313–3317.