Abstract

We developed a proof-of-concept for an on-site test VeriFied that estimates the age of bloodstains by detecting structural changes in human serum albumin (HSA) due to oxidation over time, using a nanobody-based enzyme fragment complementation assay (EFCA). VeriFied is a portable test kit that enables crime scene investigators to determine bloodstain age through a straightforward workflow involving sample collection, sample addition to the buffer, EFCA reaction, and luminescence measurement with a handheld luminometer. To validate the assay, we created a controlled series of dried bloodstains of different ages on cotton swabs, carefully regulating air exposure to study its effect on signal. Our results showed that the signal produced by the assay decreases as bloodstains get older, showing a connection between bloodstain age and bioluminescence signal which supports the assay’s potential as a forensic tool.

We strengthened the scientific basis of the test through a literature review and complementary reference assays. These assays provided indirect information, indicating that oxygen exposure did affect the blood and its HSA. The methods and the results are further discussed in later paragraphs. Although further optimization of assay component concentrations and protein purification are still needed, our fusion proteins produced a strong, specific signal, confirming the test’s functionality. To ensure clarity and consistency, we adhered to good data visualization practices, using standardized formats and units, while recognizing the need for future luminescence signal normalization to enable better comparison across different test conditions. Finally, even though developed for forensic use, the test also holds potential for broader applications including monitoring HSA oxidation in biomedical applications. For example, the HSA oxidation is known to be connected to myocardial infarction and prostate cancer [1].

Introduction

In our project, we designed an on-site test VeriFied which measures the age of bloodstains. VeriFied detects human serum albumin (HSA) and its structural changes caused by oxidation by air over time in dried bloodstains. These oxidation mediated structural changes are known to affect antibody-protein interactions of HSA [1]. VeriFied is based on an enzyme fragment complementation assay (EFCA). For detection of conformational changes, we used specially designed test components: nanobodies linked to NanoLuc enzyme fragments. You can read more about the assay component design here. We chose this assay type as nanobodies are highly specific to HSA and the NanoLuc enzyme provides an easily readable luminescent signal. HSA was chosen as the analyte due to its abundance in blood. Our nanobodies bind to HSA which allows the enzyme fragments to reassemble, become functional, and start breaking down the substrate to produce a signal. In practice, the test provides a higher signal with fresh bloodstains and a lower one with the older bloodstains.

VeriFied final product workflow:

1. The crime scene investigator collects the bloodstain sample using a moistened cotton swab.

2. The sample is suspended into a buffer solution that contains furimazine, the NanoLuc substrate.

3. The buffer is transferred into two tubes containing the EFCA-constructs, one for the control assay and one for the test assay.

4. The tubes are gently mixed and incubated.

5. EFCA constructs bind to HSA in the bloodstain, and then the enzyme fragments can react with the substrate.

6. Finally, the luminescence generated by this enzyme reaction is measured using a small portable luminometer.

The workflow of the VeriFied test.

Testing the assay

How our dried bloodstain samples were created
First, blood was absorbed onto cotton swabs and allowed to dry while exposed to air. We then created a time series of bloodstains. The bloodstains were aged for 1, 2, 3, 4, 5, 8, 11, 15, 22, and 29 days at room temperature.
Because air is a low-viscosity, easily circulating fluid, we aged the swabs head-up in 2 mL microcentrifuge tubes kept inside a closed drawer. The drawer was only opened for handling the tubes and each drawer opening and its duration was logged. This precise handling was the core of our lab project: the passage of time and ventilation were diligently controlled and documented, because in real crime scenes they are unknown variables that our test should resolve. Because we had a hard time finding sufficient information on our subject, we were extra meticulous in our record-keeping; by the end of the project, our lab journal exceeded 160 pages. Bloodstain aging log can be found as a separate file in our Lab Notebook wiki page.

This drawer worked as a semi-open system for our bloodstain aging experiment. It was only properly ventilated when we opened it for bloodstain handling. Otherwise air-exchange was minimal. All ventilation events and their durations were documented.

After the aging period, we preserved the bloodstains affording our protocol and stored them with caps closed head-down in -21 °C. This workflow minimized unintended air exchange around the sample. The workflow was designed to control the blood’s air exposure, which we hypothesized could influence the measured signal. Controlling the exposure to the elements in this manner was crucial for obtaining consistent, replicable data. We also created a strict protocol for the bloodstain reconstitution for the same reason.

Cotton swab after bloodstain extraction.

Results from VeriFied
We evaluated our own EFCA-based VeriFied fusion proteins with the aged bloodstains to test whether the luminescence signal correlates with the bloodstain age under controlled conditions. Four nanobody–NanoLuc pairs were assayed across 1–29 days.

We observed two performance dimensions: absolute signal strength and age sensitivity (how well the trend reflected the age of the bloodstain). The pair Nb77-SmBit+Nb80-LgBit yielded the strongest absolute signal which was orders of magnitude above any single-fragment background we have ever managed to generate. However, it also displayed the weakest age trend: if a trendline was fitted across the data points its R² was 0.134. In contrast, Nb77-SmBit+ALB8-LgBit produced a low absolute signal yet the clearest age dependence (R² = 0.73). Nb29-SmBit+Nb80-LgBit had an extremely low signal, yet it showed some trend (R² = 0.378).
One pair(Nb29-SmBit+ALB8-LgBit) was non-responsive under current conditions. From these results we can conclude that our fusion proteins react with HSA within the bloodstains. The weak signal makes the observed trends somewhat questionable, and they couldn't be used for modeling. We acknowledge that, if we were to ignore the sudden signal drop at 3d-point, then the prediction capability of Nb77-SmBit+Nb80-LgBit would instantly look much better. However, all our measurements were done using three technical replicates, and as such, we can not simply declare them outliers because they do not seem to follow an expected trend.

Figure 1. Bloodstain age measurements. Luminescence signal was measured 50-60 minutes after adding Nano-Glo luciferase assay reagent. Measurements were performed with Hidex Sense Microplate reader on a white Nunc Maxisorp 96-well plate. The wells contained 10 µL of periplasmic extracts for both proteins, 10 µL of extracted bloodstain (ages 1, 2, 3, 4, 5, 8, 11, 15, 22 or 29 d), 0.02 M phosphate buffer pH 8 and 0.5 µl Nano-Glo luciferase assay reagent in a final volume of 101 µl. The fusion proteins used were extracted on 14.8. Variance reported in ± SD of three technical replicates except Nb77-SmBit and Nb80-LgBit pair had only 2 replicates for 1 d. Measurements were performed on 22.8

Validation and calibration

During the project, the quality of data collected by our group improved as the project progressed, as we adopted more reliable and appropriate data collection practices. In the beginning, the reliability of most measurements were affected by limited technical replicates, calculation errors, and inconsistent methods. As we started noticing inconsistent results, we systematized data collection with better replicates, more consistent methods, and overall greater reproducibility. By the end, we were using at least three technical replicates with each measurement, usually making new standard curves for each test, and optimizing protocols to avoid inconsistent and error-prone methods such as pipetting smaller than 5 µL volumes when possible.

Controls and background signal
The luminescence testing done on VeriFied served as a proof-of-concept, indicating that the EFCA functioned correctly, and provided a signal that correlated with bloodstain age. In addition, however, we measured the background signal produced by the individual components of the test to confirm the signal was truly from the enzyme fragment complementation and not an artifact. Additionally, the measurements serve as a first step toward calibrating and optimizing the test.

Measurements done with just individual nanobody-NanoLuc fragment fusion proteins show the produced signal as being orders of magnitudes lower than that of measurements with both the NanoLuc fragments in the same test (Figure 2). In addition, the measurements from negative controls such as a periplasmic extraction of a control E. coli BL21(DE3)pLysS strain without a transformed plasmid, buffers, and HSA with NanoLuc substrate all produced signals below 1000 RLU.

Figure 2. Luminescence signal produced by different volumes of periplasmic extractions of Nb80-LgBit (a), Nb29-SmBit (b), and Nb77-SmBit (c).Measurements were performed with Hidex Sense Microplate reader on a white Nunc Maxisorp 96-well plate. The wells contained either 70, 50, 30, 20, 10 or 5 µL of periplasmic extract, 0.02 M phosphate buffer pH 8, and 0.5 µl Nano-Glo luciferase assay reagent in a final volume of 101 µl. The signal for t=0 min was measured when the substrate had not been added to the wells. The fusion proteins used were extracted on 31.7. Variance reported in ± SD of three technical replicates. Measurements were performed on 8.8. The Y axis has been scaled to highlight the difference between the luminescence signal produced by proteins with only a fragment of the enzyme and the pair that can form a functional enzyme.

Validating the signal
Measurements with varying amounts of nanobody-NanoLuc fusion proteins produced a detectable change in signal intensity, as seen in Figure 3 (a). Interestingly, however, comparing measurements with and without HSA produced unexpected results. The addition of HSA to the reaction resulted in a reduced luminescence peak as seen in Figure 3 (b). The counterintuitive result suggests that HSA is meaningfully affecting the interaction between the fusion proteins as expected, even if the direction of change is unexpected. One theory is that if albumin is scarce, few complexes can form, but if it is in excess, the EFCA halves bind to separate HSA molecules and fail to reconstitute the enzyme. However, further optimization to find the correct ratios of fusion proteins to HSA is required for a clear answer.

Figure 3. The effects of fusion protein amount and human serum albumin (HSA) on luminescence intensity. Measurements were performed with Hidex Sense Microplate reader on a white Nunc Maxisorp 96-well plate. The wells contained 20, 10 or 5 µL periplasmic extracts of Nb80-LgBit and Nb29-SmBit, 0.02 M phosphate buffer pH 8, and 0.5 µL Nano-Glo luciferase assay reagent in a final volume of 101 µL. The measurement with HSA contained 1 µg HSA. The Nano-Glo luciferase assay reagent used contained 2 % of Nano-Glo® Luciferase Assay Substrate in the Nano-Glo® Luciferase Assay Buffer. The signal for t=0 min was measured when the substrate had not been added to the wells. The fusion proteins used were extracted on 31.7. Variance reported in ± SD of three technical replicates. Measurements were performed on 8.8.25.

To validate that our constructs interact specifically with HSA, we tested different amounts of purified albumin in relation to constant volumes of extracted protein (containing an unknown but fixed amount of EFCA constructs). Preliminary testing consistently led to changes in luminescence signal, in some cases the signal was increased and in others lowered as seen in Figure 4. Without knowing exact protein concentrations in each test, optimizing the test further remains difficult, and successful purification of the nanobody-NanoLuc fusion proteins is required before advancing further.

Figure 4. Human serum albumin (HSA) concentration’s effect on the luminescence signal produced by four different pairs. Luminescence signal was measured 50-60 minutes after adding the Nano-Glo luciferase assay reagent. Measurements were performed with Hidex Sense Microplate reader on a white Nunc Maxisorp 96-well plate. The wells contained 10 µL of periplasmic extracts for both proteins, 0.25, 0.5, 1 or 2 µg HSA, 0.02 M phosphate buffer pH 8, and 1:200 Nano-Glo luciferase assay reagent. Periplasmic extracts were extracted on 14.8. Measurements were performed on 22.8.

Investigating HSA Oxidation

To create a more robust scientific basis for the function of VeriFied, we researched previous studies related to HSA oxidation and did additional experiments into the kinds of changes that occur in blood proteins during oxidation in bloodstains. We wanted to validate the results from previous research indicating that oxidation significantly alters the structure of HSA and research the speed of HSA oxidation in bloodstains.

The surface of HSA has one free cysteine (Cys34) (Figure 5), which is often considered an indicator of the oxidation state of the protein in clinical settings, where most of the research into HSA oxidation is done [3,4]. Reversibly oxidized forms have Cys34 in either its sulfenic acid form or in a disulfide bond with another molecule, and reduced forms have Cys34 in thiol form [3,4]. After severe oxidation, Cys34 appears in an irreversibly oxidized form, containing either sulfinate or sulfonate [3,4].

Figure 5. Human serum albumin (PDB: 1AO6). Blue: domain I, yellow: domain II, red: domain III, cyan: Cys34. Figure made in pyMOL.

Literature indicates that oxidation of HSA can cause structural changes, increasing its susceptibility to protease digestion, reducing antibody affinity, and affecting its stability [1-4]. According to previous studies, this oxidation can be caused by exposure to air, for example in improperly handled or stored blood [1]. Our foundational hypothesis is that due to these structural changes caused by oxidation in bloodstains exposed to air over time, the affinity of certain nanobodies is reduced.

Reference assays
To measure HSA oxidation, we reacted our HSA samples with Ellman’s reagent, which reacts with thiol-groups, producing a colorimetric reaction that can be used to reliably quantize reduced Cys34 [5]. However, due to the presence of other free thiol groups in blood and the bright color of hemoglobin absorbing in a similar wavelength as the Ellman’s reagent, HSA needed to be separated from other proteins in the samples prior to measurement. We created a protocol for separating HSA from most other blood proteins.

In addition to Ellman’s assay, we carried out general reference assays—Bradford, bromocresol green (BCG), and spectrophotometric hemoglobin measurements to assess the success of the HSA purification. The Bradford assay was used to determine total protein content, hemoglobin measurements to quantify hemoglobin, and BCG to measure the total amount of HSA. Hemoglobin absorbs light at the same wavelength (630 nm) as BCG at when it is used for albumin assay [6], and it was a useful tool for assessing the success of our hemoglobin-HSA separation process. Previous research has developed a method for assessing a bloodstain age from its hemoglobin absorbance spectra, which is another reason we were interested in measuring its absorbance as well [7]. For the purification process, 420 nm wavelength was selected for measuring the hemoglobin because at that wavelength both oxidized and unoxidized have approximately the same absorbance. While the Ellman’s assay itself did not yield useful data, the data we gathered for the purification process proved interesting (Figure 6). All graphs below represent the properties of the enriched HSA at the end of our purification process.

Figure 6. Spectrophotometric measurements of blood proteins in aged bloodstains. Bloodstain ages vary from 1 day to 29 days old. (a) The total protein was measured using the Bradford assay at 595 nm with Tecan Infinite 200 Pro plate reader. Variance reported in ± SD of four technical replicates. (b) The hemoglobin was measured at 420 nm with Perkin Elmer Lambda Bio 40 UV/VIS spectrometer. Quantification was done using Beer-Lambert Law. Variance reported in ± SD of three technical replicates. (c) HSA was measured at 630 nm with Perkin Elmer Lambda Bio 40 UV/VIS spectrometer by using bromocresol green assay.

All our reference assays were run in quadruplicate, and they yielded a consistent trend that was inversely proportional to bloodstain age. In other words, when a linear trendline was fitted across the data points, the slope was negative in every case. This trend was visible in hemoglobin absorbance spectra, in total protein yield by the Bradford assay and when measuring purified HSA yield using bromocresol green. While low R2 values indicate that no predictive models can be constructed from the data, we conclude qualitatively that exposure to oxygen does alter proteins within dried bloodstains. A total of 120 measurements(4 replicates for each time point * 10 time points) were made to obtain the above data, excluding calibration. For a deeper understanding of HSA oxidation in bloodstains, future research should aim to successfully purify HSA from bloodstains using a method such as ion-exchange chromatography and measure the oxidation state with Ellman’s assay. More detailed measurements could also be performed using mass-spectrometry to investigate changes to other amino acid residues besides Cys34. To achieve more replicable data in the future a replicable “fresh control” should be manufactured. One possible way to achieve this could be manufacturing “0-hours-old” bloodstains by drying them in an inert atmosphere, using purified gases such as argon or nitrogen. By comparing these 0-hour control samples to bloodstains aged in a normal atmosphere, one could, with good confidence, assume that any perceived differences are specifically related to exposure to atmospheric oxygen.

nanoDSF and DLS
To measure structural changes in HSA due to oxidation, we used nano differential scanning fluorimetry (nanoDSF) and dynamic light scattering (DLS) (Figure 7). nanoDSF measures the unfolding of proteins as a function of temperature, giving a thermal denaturation curve as a result. DLS was used to measure the particle sizes present in the sample also as a function of temperature. We measured pure HSA and chemically reduced HSA to assess whether a measurable difference in stability or particle size can be observed, which would indicate a structural difference.

Figure 7. Conformational stability comparison of human serum albumin (HSA) and reduced HSA using nano differential scanning fluorimetry (nanoDSF) and dynamic light scattering (DLS). Two technical replicates of each sample were measured. Purple sample is pure HSA (≥99 % Sigma-Aldrich) (15 µM) in phosphate buffered saline (PBS), pH 7.4. Green sample is HSA (≥99 % Sigma-Aldrich) (15 µM), incubated for 5 min in PBS with 10.8x molar excess of dithiothreitol (DTT) (162 µM) in room temperature before desalting with Amicon Ultra Centrifugal Filter 0.5mL 30K MWCO. nanoDSF and DLS were measured with NanoTemper Prometheus Panta. The onset of unfolding or size increase (Ton) indicates the temperature at which the protein starts to unfold or increase in size. The melting point (Tm) of the samples indicates the temperature at which 50 % of the proteins in the sample have unfolded. a. Ratio of absorbances measured at 350 nm / 330 nm that indicates the change in environment of tryptophan as a function of time. b. First derivative of the ratio of absorbance readings measured at 350 nm / 330 nm. c. DLS measurement that indicates the cumulative radius of the particles in the sample as a function of time.



The observed difference in melting point (Tm) of 0.99 °C, along with a slight difference in measured cumulative radius and onset of size increase difference of (Ton) 2.24 °C measured between the HSA and reduced HSA samples, visible in Figure 7, point towards slight structural differences between the samples. The results further corroborate oxidation causing structural changes in HSA, although additional measurements with more drastically oxidized or reduced samples would likely show a more substantial difference. The direction of change toward a more thermally stable conformation in the oxidized HSA is consistent with previous studies pointing toward oxidized HSA having a more thermally stable structure [2].

The exact degree to which the untreated or reduced HSA samples used in nanoDSF and DLS were oxidized remains ambiguous to us despite attempts to quantize it using Ellman's assay. Due to mistakes in data collection, however, the results were not quantifiable, but do show a consistently higher signal for the sample incubated with dithiothreitol (DTT), indicating a higher percent of reduced Cys34 in the samples.

Data visualization

To make our results as clear as possible, we aimed to represent our data in a uniform, logical manner. Using the same colors to represent similar variables, labeling all the data appropriately, and using the same Y-axis for panels in the same figure for ease of comparison. According to good data visualization practices, our data is represented mainly as scatter plots of means with standard deviation indicated as error bars. Bar graphs, are used especially when comparing results from different measurements.

We chose to measure and report luminescence in relative luminescence units (RLU) but future results should ideally be reported in normalized units, determined based on a standard curve created using a reporter that produces a known signal. Since our measurements were done with unknown protein quantities, using standardized luminescence units would not have made our data significantly more repeatable. Usually normalization is done to make the data more robust to instrument-to-instrument variation and allow for comparison between measurements from different readers.

Potential for other applications

We have created a design for an assay for detecting structural changes. Protein stability can be detected precisely with nanobodies, which can be leveraged by designing specific nanobodies tailored to the structure of the biomarker. This approach is useful to any project aiming to study protein stability using assays like EFCA. While we found that the concentrations of the components still require optimization, we successfully observed a significant signal that was affected by bloodstain age with one of the fusion protein pairs. This not only indicates that we successfully produced functional fusion proteins but also serves as proof that the assay system works as intended.

Building on these results, our test could be applied beyond its initial forensic application. For example, in the medical field, there is a growing need for tests that detect HSA and specifically monitor its oxidation mediated structural changes. Such changes in HSA can serve as a biomarker for various conditions associated with high oxidative stress like myocardial infarction and prostate cancer for example [1]. Additionally, the blood storage conditions can lead to HSA oxidation over time which could be detected by a similar test [1]. Overall, our assay components hold potential for broader research applications, partially in assessing HSA stability.

Moving forward with VeriFied

Although our project has demonstrated proof-of-concept for estimating the age of bloodstains through HSA oxidation, the current version of VeriFied is still far from a ready-to-use forensic tool. To move toward a more robust and reproducible system, several important developments would need to be made in both the construct design and the bloodstain handling pipeline.

First, the fusion proteins themselves require further refinement. In their present form, periplasmic extracts yield a mixture of proteins. In the future, we would purify our constructs immunochromatographically so that only proteins with confirmed affinity to HSA would be retained. Furthermore, instead of testing NanoLuc fusion proteins, we would generate variants in which each nanobody is fused to a complete reporter molecule. By doing so, each nanobody could be assayed independently on aged bloodstains. These single-nanobody measurements would make it far easier to determine which epitopes change due to oxidation and to identify the most promising pairs for inclusion in the final age-determination assay. If some pairs failed to produce the expected response, troubleshooting would be more straightforward because the binding characteristics of the individual nanobodies would already be known.

Second, improvements are needed in bloodstain aging and extraction. Our current time series covered stains aged up to 29 days, but the final product should ideally be able to measure bloodstain age from several months to perhaps even years. Therefore we would like to extend the bloodstain aging period to at least twelve months. In addition, our sampling density was relatively sparse: after the first two weeks we tested only weekly, which leaves large gaps in the early stages where changes may occur most rapidly. In future experiments, we would strive to increase the sampling frequency.

Additionally, we would age bloodstains under multiple controlled environments to separate the effects of oxygen flux, light, humidity, and handling. In our research, stains were aged in a single closed drawer that was opened only during handling. Going forward, instead of a simple drawer, we would design a modular box that would allow us to test:

1. Baseline ventilation conditions with minimal air exchange and infrequent opening.

2. High-ventilation conditions where the box would be opened ~4× more frequently to increase intermittent oxygen exposure.

3. Continuous-light condition (constant illumination with controlled spectrum and intensity) to test light-driven oxidation.

4. High-ventilation no-light condition where stains experience unrestricted air movement in darkness.

5. High-humidity condition to assess whether moisture accelerates oxidation or alters protein stability.

6. Selected combinations of the five conditions.


Because the goal is a reliable field test for forensic investigators, the influence of these elemental variables must be quantified and clearly communicated to end users.