Regular Articles
A Bioluminescent Assay System for Whole-Cell Determination of Hormones
2013 Volume 61 Issue 7 Pages 706-713
Details
2013 Volume 61 Issue 7 Pages 706-713
A bioluminescent-assay system was fabricated for an efficient determination of bioactive small molecules in physiological samples. The following three components were newly created for this assay system: (i) a single-chain probe exerting a 7.2-times stronger optical intensity than conventional ones, (ii) a high throughput assay device uniquely designed for the assay system with ca. one-fourth smaller standard deviation (S.D.) to samples than without the device, (iii) a buffer cocktail optimized for the assay system. The advantages of the assay system were evaluated by determining (i) the stress hormone levels in human saliva and (ii) multicolor imaging of genomic and nongenomic effects of woman sex hormones. This study guides on how to fabricate an efficient assay system for bioactive small molecules with convenience and high precision.
Bioluminescence provides a nearly ideal optical readout for bioassays and molecular imaging. The assays are intrinsically very simple, sensitive, and do not require an external light source or other cofactors.1,2) Bioluminescence is advantageous in terms of its high quantum yield, which is about one order magnitude higher than conventional chemiluminescence.3) Luciferases catalyze the oxidation of a luciferin in the presence of molecular oxygen (O2) and optionally Mg2+ and ATP for emitting bioluminescence. The reagents are economical and not hazardous.4)
Because of these distinctive merits, bioluminescence is utilized in miniaturized analytical devices such as microfluidic devices.3) Although microfluidic devices are versatile, they are generally integrated in a small chip and thus sophisticated to operate. These optical devices determine ATP levels for hygiene monitoring and estimate antigens as biomarkers.5) Their high sensitivity is realized by employing integrated light detectors such as charge-coupled devices (CCDs) and photomultiplier tubes (PMT) in the device.
A series of single-chain probes carrying engineered fragments of luciferases was demonstrated, where the fragments of luciferases and the other components were integrated in a single-chain backbone to examine ligand-activated intramolecular protein–protein interactions (single-chain type).6–8) More generally, marine and beetle luciferases were fragmented at an appropriate site. The fragments were genetically fused to proteins A and B of interest for determining an interaction between the proteins A and B (two-chain type).9–11) New luciferases were developed and engineered for boosting the optical properties of the probes.12–14)
Although the above-mentioned bioluminescent probes are well established, their sensorial properties are still insufficient to sensitively determine endocrine hormones in physiological samples and their assay protocols with luciferases are relatively long and tedious: (i) For instance, optical intensities of single-chain probes should be further improved for a sensitive determination of endocrine hormones. The poor optical intensities are caused by dissection of luciferases during the probe design, which seriously invades the enzymatic activity. The intensities after the dissection were reported to reach merely 0.5–5% of the original intensity.8,15) (ii) Standard protocols of the conventional bioassays with luciferases are relatively long, including a procedure of preparation of a lysis cocktail and its mixing with an assay buffer in every single measurement. These complex protocols need to be revised labor- and cost-effectively. (iii) An ideal bioluminescent assay system should achieve a considerable sample throughput and precision for an efficient, fidelity determination of the ligands in many samples. Nevertheless, conventional luminometers allow only a single sample measurement at once, thus, are poor in terms of sample throughput.
To address the above-mentioned limitations, we fabricated a 6-channel bioluminescent-assay system which carries single-chain probes in a custom-made optical device under a simplified protocol. First, new single-chain probes with enhanced optical properties were developed on the basis of a precedent stress hormone-sensing probe.16) Second, a high-precision optical device with 6 channels was developed for accommodating the single-chain probes. The optical device was equipped with a variable slide holder carrying a multichannel microslide, an optical filter, and a mirror-like interior design for focusing the generated photons to the detector. The bottom of the optical device is designed to fit in conventional optical detectors (diameter=3.5 cm), such as a luminometer (GloMax 20/20n, Promega, U.S.A.) and a spectrophotometer (AB-1850, ATTO, Japan). For the optical device carrying the single-chain probe, an optimal assay buffer was also evaluated.
The advantages and restrictions of this assay system is discussed in detail with the following experimental results: (i) the determination of stress hormones in human saliva using the optical device set in the luminometer (GloMax 20/20n, Promega) and (ii) simultaneous determination of agonistic and antagonistic activities of woman sex hormones and multicolor display.
Template plasmids encoding Leu-rich, Simer-R2, and cPresso in Fig. 1A and Suppl. Fig. 1A were taken from our previous studies.7,17,18) The template cDNAs encoding Monsta and the N-terminal-extended ligand binding domain of human glucocorticoid receptor (GR HLBD; 486–777 aa) were obtained from our previous studies.14,16) In the present study, a series of plasmids encoding cSimgr5, 6, or 7 was newly synthesized for improving optical intensities as follows: The cDNAs of the N-terminal (18–105 aa) and C-terminal (106–183 aa) domains of split-Monsta were generated by a polymerase chain reaction (PCR) to introduce unique restriction sites at both ends of the domains by using adequate primers and the template cDNA. The cDNA of GR HLBD were also amplified by a PCR to introduce unique restriction sites at both ends of the domain by using adequate primers and the template. The cDNA oligomer encoding the LXXLL motif of GRIP1 NID3—742NALLRYLLDKD752—was synthesized-to-order by Exigen (Tokyo, Japan). The above fragments were ligated and they were subcloned into the pCS2+ (RZPD) or pcDNA 3.1(+) vectors (Invitrogen). These plasmids were finally sequenced to ensure fidelity through a custom order placed with Operon (Tokyo, Japan) and the expressed probes were named cSimgr5, 6, and 7, respectively.
(A) Schematic structures of the constructs that were used in the present study. Abbreviations: GR LBD, the ligand binding domain of human glucocorticoid receptor; AR LBD, the ligand binding domain of human androgen receptor; ER LBD, the ligand binding domain of human estrogen receptor; Mon-N and -C, the N- and C-terminal domains of Monsta; FLuc-N and -C, the N- and C-terminal domains of firefly luciferase; CB Red-N and -C, the N- and C-terminal domains of click beetle luciferase red; Src SH2, the SH2 domain of v-Src; GR HLBD, the hinge and ligand binding domain of human glucocorticoid receptor; LXXLL motif, a conserved motif of coactivators interacting glucocorticoid receptor, where “L” means leucine. (B) A conventional protocol for a whole-cell biosensor. The cultured mammalian cells were exposed to hormones or chemicals. In response to the hormones, the conformation of the single-chain probe folds up, where an intracellular protein complementation occurs. The cells were lysed and transferred to a test tube for light sensing. (C) Relative optical intensities of single-chain probes in response to a stress hormone, cortisol (n=3). The minus and plus signs indicate optical intensities of cSimgr7 in response to vehicle (0.1% dimethyl sulfoxide (DMSO)) and 10−6 M of cortisol, respectively. Inset A: relative optical image of single-chain probes on a 96-well microplate in the presence (lanes 4–6) or absence (lanes 1–3) of a stress hormone, cortisol. Each row of the wells carries the same single-chain probes as labeled.
The relative optical intensities were compared before and after stimulation of the mammalian cells carrying a single-chain probe with cortisol (final concentration: 10−6 M) (Fig. 1C). African green monkey kidney fibroblast-derived COS-7 cells were grown in a 96-well plate and transiently transfected with a plasmid encoding cPresso, cSimgr5, 6, or 7 as annotated in Fig. 1C using a lipofection reagent, TransIT-LT1 (Mirus). The cells were additionally incubated in a CO2 cell incubator (Sanyo, Japan) for 16 h. The cells were stimulated with 10−6 M cortisol for 20 min, lysed with a lysis buffer (E291A; Promega) according to the manufacturer’s protocol. An aliquot of the lysates (10 µL) in a 1.6 mL microtube was mixed with a 40 µL of assay buffer (E290B; Promega) carrying native coelenterazine (nCTZ). Optical intensities from the mixture was immediately determined with a luminometer (GloMax 20/20n, Promega). An average of three seconds of time delay was taken between the mixing and practical measurement. This assay procedure is illustrated in Fig. 1B.
The luminescence intensities were normalized in relative luminescence unit (RLU) per cell during one second based on the briefly counted cell numbers before the experiments. Therefore, the unit is expressed as RLU/cell/s.
Consistent optical results were obtained with a 96-well plate raising COS-7 cells (Fig. 1C, inset A). The COS-7 cells expressing the single-chain probes, i.e., cPresso, cSimgr5, 6, or 7, were prepared according to the above-mentioned protocol for Fig. 1C. Ten microliters of the cell lysates were mixed with an assay buffer dissolving nCTZ (40 µL) using an 8-channel micropipette and this mixture was immediately transferred to the chamber of an LAS-4000 (FUJIFILM) for determining the optical images. The bioluminescence image was obtained in an automatic exposure mode and analyzed with Multi Gauge ver. 3.0 software (FUJIFILM).
One-Shot Buffer for Quick Determination of Ligands in LuminometersOptimal assay buffers for the present single-chain probes were initially evaluated (Fig. 2A). A cell line of COS-7 cells stably expressing cSimgr7 was beforehand established with a selection using G418 (Geneticin; 0.5 mg/mL). The established cell line was subcultured into a 96-well plate, and incubated with vehicle (0.1% dimethyl sulfoxide (DMSO)) or 10−6 M of cortisol for 20 min. The cells were then lysed with 6 different lysis buffers, i.e., compositions C1–C4 in addition to Promega’s and NEB’s lysis buffers (Suppl. Table 1). Ten microliters of the lysates were then mixed with 40 µL of one of the following assay buffers carrying nCTZ, i.e., Hank’s balanced salt solution (HBSS), Tris-ethylenediaminetetraacetic acid (EDTA) (TE), and Tris–MgCl2 buffers on a 96-well plate. The optical image was immediately taken using LAS-4000 (FUJIFILM).
(A) Variance in the background intensities according to varying combinations of lysis and assay buffers. Promega and NEB mean lysis buffers provided by Promega and New England Biolabs (NEB), respectively. The compositions were unknown. (B) Variance in the signal-to-background (S/B) ratios of cSimgr7 according to concentrations of halogen ions in an HBSS buffer (n=3). (C) Variance in the (S/B) ratios of cSimgr7 according to mixing ratios of lysis and assay buffers (left) (n=3). Ligand selectivity of Leu-rich and Simer-R2 in an one-shot buffer, C12 (Suppl. Table 2). Inset A shows an optical image of COS-7 cells carrying cSimgr7 after stimulation with the vehicle (0.1% DMSO) (open bars) or 10−6 M of cortisol (closed bars). Channels 1, 2, and 3 demonstrate optical intensities by vehicle, whereas channels 4, 5, and 6 images show elevated optical intensities by the cortisol.
Contribution of halogenic ions in the assay buffer to signal-to-background (S/B) ratios was similarly evaluated with the same COS-7 cells stably expressing cSimgr7 (Fig. 2B). The COS-7 cells were grown in a 96-well plate and stimulated with 10−6 M of cortisol or the vehicle for 20 min. The cells were then lysed with a Promega lysis buffer for 20 min. Ten microliters of the lysates were then mixed with 40 µL of the assay buffers C5–C10 carrying halogenic ions and the substrate, nCTZ, (Suppl. Table 1). The developed optical intensities were recorded with the luminometer (GloMax 20/20n, Promega).
In the present study, the luminescence intensities were normalized by a relative luminescence unit (RLU) ratio, i.e., RLU ratio (+/−), where RLU (+) and RLU (−) represent the luminescence intensity from 1 µg of protein of cell lysate after the COS-7 cells were stimulated with and without a ligand, respectively.
A buffer cocktail named “one-shot” buffer was made for rapid determination of ligand activities in mammalian cells without a specific cell-lysis step (Fig. 2C). A series of buffer cocktails was prepared by mixing the following two buffers in a ratio of 8 : 2, 6 : 4, 4 : 6, and 2 : 8 (C11–C14; Suppl. Table 2): (i) the primary buffer consists of 25 mM Tris–HCl buffer (pH 7.6), 150 mM NaCl, 1% Nonidet P-40 (NP-40), 1% Tween 80 (TW-80), 1% sodium deoxycholate, and 0.1% sodium dodecyl sulfate (SDS); (ii) the secondary buffer consists of an HBSS buffer (Wako), 0.01% polyethylene glycol (PEG, molecular weight (MW): 92), 0.1 ppm of Fe(III), and 0.1 ppm of As(V).
A typical bioassay with the cocktails was conducted as follows: the COS-7 cells stably expressing cSimgr7 were grown in a 96-well plate, incubated 16 h, and stimulated with the vehicle 0.1% DMSO or 10−6 M of cortisol. After thorough removal of the culture media carrying a ligand, the cells on the wells were immersed with a 50 µL of the buffer cocktails and briefly pipetted five times. The whole solution on each well was transferred to a 1.6 mL test tube and its optical intensity was immediately determined with the luminometer (GloMax 20/20n, Promega). The optical intensities were normalized as shown in Fig. 2B. Figure 2C demonstrates the S/N ratios with dihydrotestosterone (DHT) and 4-hydroxytamoxifen (OHT) only for simplicity.
Ligand selectivity of Leu-rich and Simer-R2 was determined with the buffer cocktail, C12 (Fig. 2C). COS-7 cells grown in a 96-well plate were transiently transfected with pLeu-rich or pSimer-R2 and incubated 16 h in a CO2 incubator. The cells were stimulated with 10−6 M of DHT, OHT, or vehicle (0.1% DMSO) for 20 min. After removal of the culture media, the cells were mixed and pipetted with 50 µL of C12 five times. The whole solution on each well was transferred to a 1.6 mL test tube, optical intensities of which were then determined using the luminometer (Promega).
A corresponding bioluminescent image was obtained with cSimgr7 (Fig. 2C, inset A). For this study, COS-7 cells stably expressing cSimgr7 were cultured in a 6-channel microslide (ibidi). Left and right three channels were stimulated with a vehicle (0.1% DMSO) and 10−6 M of cortisol dissolved in a culture medium for 20 min. After removal of the medium from the microplate, each channel of the microplate was filled with nCTZ dissolved in a modified cocktail of C12 (named C15). The microplate was immediately moved to the chamber of LAS-4000 (FUJIFILM). The optical image of the plate was then obtained in an automatic detection mode.
Creation of Bioluminescent-Assay DevicesA prototypical bioluminescent-assay device (version 1) was custom-made for the efficient determination of optical intensities from bioluminescent probes in mammalian cells grown on a 6-channel microslide (2.5×7.5 cm, µ-slide VI0.4; ibidi) (Fig. 3).
(A) A brief diagram of the assay device ver. 1. The microslide is embedded on a conveyer layer, which is designed to slide on to the plastic platform like a conveyer belt. The microslide is supported by the holder and covered by a mirror cap. Numbers in the parenthesis represents part numbers. Inset A highlights a channel of the microslide. The bioluminescence (dotted line) is reflected by a mirror cap and collected to the detector side at the bottom. (B) The picture of the assay device ver. 1 manufactured for the present study. (C) Determination of stress hormone levels in human saliva. COS-7 cells stably expressing cSimgr7 were exposed to 6 saliva samples, and the reconstituted bioluminescence was fitted for quantifying the cortisol levels in the saliva samples (n=3). The cortisol levels in the same saliva samples were simultaneously determined with an ELISA kit (n=5) (Immunospec). Inset indicates the representative optical intensities from the 6 channels of the microslide (ibidi).
The assay device consists of a mirror cap for reflecting bioluminescence to the detector and blockading outer interferences (part 1; Fig. 3A), a 6-channel microslide (ibidi) for raising cells (part 2), a variable microslide holder for carrying the microslide (part 3), a plastic conveyer platform for supporting the microslide holder (part 4), a set of optical filters (part 5), a variable filter holder carrying maximally 3 optical filters (part 6), and an inner-holed platform stander for supporting the platform and passing the generated bioluminescence to the bottom detector (part 7). These parts were assembled by an order to a trial-manufacturing service center of National Institute of Advanced Industrial Science and Technology (AIST), called IBEC.
The microslide carries 6 channels, between which a cross-interference was concerned. The honeycomb-like internal barriers of the mirror cap (part 1) efficiently blockades interference light from the other channels. As highlighted in Fig. 3B Inset A, bioluminescence emitted from each channel is reflected by the mirror cap to the bottom slit (7 × 24 mm) (dotted line; Fig. 3B, inset A). The bioluminescent signal passes the optical filter beneath the slit and finally reaches the detector at the bottom.
The microslide holder (part 3) was designed to be variable according to experimental preference. A side of this microslide holder is designed like a saw tooth to fit each channel of the microslide on the bottom slit. The optical filter beneath the bottom slit was designed to be removable according to experimental preference.
We further revised the design of the assay device ver. 1 for a more convenience. The most distinctive differences of the assay device ver. 2 from the device ver. 1 are the variable optical filter holder and the gold-plated platform stander for a better reflection of the photons to the detector (Fig. 4). The filter holder was designed to carry three different filters, i.e., blue (470-nm bandpass), green (520-nm bandpass) and red (600-nm long-pass).
(A) Schematic drawing of the ligand recognition steps of pSimer-R2 and pSimer-G4. In response to antagonist, the ligand binding domain of estrogen receptor (ER LBD) interacts the SH2 domain of v-Src. This interaction induces reconstitution of the fragmented click beetle luciferase red (CB Red). Similarly, in response to agonist, ER LBD binds a common LXXLL motif of coactivators. This binding restores the green bioluminescence through the intramolecular interactions between the N- and C-terminal fragments of CB green. (B) The picture of the custom-made assay device ver. 2. The microslide is set on the slide holder and capped with the mirror cap. The filter holder is slidable to let an appropriate filter set to the slit of the microslide. The assay device was then placed into a conventional luminometer (GloMax 20/20n, Promega) for the determination of bioluminescence in different colors. The bellow inset shows chemical structures of the chemicals used for the present study. (C) Relative optical intensities of blue, green, and red bioluminescence, shown in relative fold intensities. The circled numbers indicates the experimental sequence including the timing of the filter switching. The left, middle, and right bars in groups demonstrate blue, green, and red color intensities, respectively. (Color images were converted into gray scale.)
As a stress marker, cortisol levels in human saliva were estimated using the bioluminescent-assay device and compared with the results by a conventional enzyme-linked immunosorbent assay (ELISA) kit (Immunospec) (Fig. 3C).
The salivary samples for this comparison were donated from 6 volunteers affiliated with AIST. The samples were collected at 4 p.m. using a specific saliva collection tube (SARSTEDT), centrifuged, stocked in a 4°C refrigerator before use, and used up in 24 h.
A COS-7 cell line stably expressing cSimgr7 was established as described in Fig. 2B and subcultured into 6-channel microslides (ibidi), where the three left channels were used for standards of cortisol, whereas the other three channels were used for the determination of cortisol in the saliva samples. The cells were stimulated for 20 min with 60 µL of saliva samples or standard cortisol concentrations (10−7, 10−8, or 10−9 M) dissolved in culture media. The cells on the microslides were then lysed with a lysis buffer (Promega) for 20 min. The microslide was placed on the assay device ver. 1 (Fig. 3, Suppl. Fig. 1B) set in a luminometer (GloMax 20/20n, Promega). Optical intensities from each channel of the microslides were then determined in order, immediately after injection of an assay buffer dissolving nCTZ (Promega) to the channels (n=3). The video image showing the assay procedures using the optical device was attached, separately (see Suppl. Video Information).
In parallel, the cortisol levels in the same human saliva samples were determined with the ELISA assay kit (Immunospec) according to the manufacturer’s instructions (n=5), the results of which were compared with those of the above microslide on the bioluminescent-assay device (Fig. 3C).
Multicolor Imaging of Agonistic and Antagonistic Effects of Woman Sex Hormones Using the Device-Aided Assay SystemAgonistic and antagonistic effects of estrogen were determined in green and red colors using the device-aided assay system (Fig. 4). COS-7 cells grown in a 6-channel microplate (µ-slide VI0.4, ibidi) were transiently cotransfected with pSimer-R2 and pSimer-G4, which are single-chain probes previously reported by us,7) and incubated for 24 h in a cell incubator (Sanyo). The cells in each channel of the microslide were then stimulated with 10−5 M of 17β-estradiol (E2), OHT (Sigma), polychlorinated biphenyls (PCB), or o,p′-dichlordiphenyltrichloroethane (o,p′-DDT) (Sigma) for 20 min. The channels in the microslide were washed once with 1× phosphate-buffered saline (PBS) (Sigma) and filled with 50 µL of Bright-Glo assay solution carrying D-luciferin as the substrate (Promega). Three minutes after further incubation in the cell incubator, optical intensities from each channel were near-simultaneously estimated by conveying the filter holder and microslide holder of the optical device ver. 2 set in the luminometer (GloMax 20/20n, Promega). The timing of the filter switching was illustrated in Fig. 4C. The circled numbers indicate the experimental procedure. The optical intensities of the five hormone samples were first determined under a blue filter. After switching the filter to green or red, the same samples were estimated again.
The reason that we used a Bright-Glo (Promega) in Fig. 4, instead of our one-shot buffer cocktail, is because the Bright-Glo is an optimized buffer cocktail for beetle luciferases requiring D-luciferin, ATP, and Mg2+. Since Simer-G2 and Simer-R4 are made of split-click beetle luciferases green and red, they require an optimized composition of D-luciferin, ATP, and Mg2+.
Glucocorticoid receptor (GR) is responsible for many metabolic actions including stress responses. Cortisol-activated GRs are dimerized and recruit coactivators. Because all the coactivators have a common α-helical peptide, called an LXXLL motif, the binding between GR and the LXXLL motif was utilized for cortisol sensing in the present single-chain probes (Figs. 1A, B).
Relative ligand sensitivity of newly constructed single-chain probes was examined in the presence or absence of 10−6 M cortisol (Fig. 1C). The strongest bioluminescence intensities were observed with COS-7 cells expressing cSimgr7, which were approximately 5.6- and 2.9-times stronger than cSimgr5 and cPresso in COS-7 cells, respectively. cPresso in COS-7 cells exhibited the lowest background intensity, which contributed to the highest S/B ratio of 14, compared with the others. Therefore, cSimgr7 is advantageous in terms of optical intensity, whereas cPresso is a better option in terms of the S/B ratios.
In Fig. 1C, the absolute intensity and S/B ratio variances of cSimgr5-7 may be explained with the expression efficiency and high/low copies of the host vectors. pCS2+ vector allows high-level transient expression and high copies in the host cells. Fig. 1C suggests that pCS2+ vectors encoding cSimgr6 or cSimgr7 allow high-level transient expression, considering their higher basal intensities than the others. cSimgr6 and cSimgr7 carry different luciferases, intrinsic variance of which caused the different absolute intensities and S/B ratios.
The selectivity of cSimgr7 to steroids in 10−6 M was also estimated with COS-7 cells carrying cSimgr7 (Suppl. Fig. 1A). The results reveal that cSimgr7 is highly specific to cortisol among the tested steroids, and even discriminates cortisol from corticosterone, although they share high chemical structural similarity. This excellent selectivity of the probe to cortisol reflects the stress sensitivity of our body via a glucocorticoid receptor.
The optical images of the single-chain probes were further estimated with an image analyzer, LAS-4000 (FUJIFILM) (Fig. 1C, inset A). The image was obtained with LAS-4000 set at automatic exposure mode. The comparison in the optical image revealed that cSimgr7 exhibits the strongest optical intensities among the probes: i.e., 7.2-times stronger than the precedent single-chain probe, cPresso. In contrast, although cPresso is known as the most potent, cPresso in COS-7 cells failed to exhibit a considerable variance in the optical intensities in the presence or absence of 10−6 M cortisol in the measurement with LAS-4000. This dramatic improvement of cSimgr7 from the precedent version, cPresso, in the intensities may be explained by the difference of the optical intensity and stability of the adapted luciferases in the probe, considering that their probe designs are almost equivalent. Optical intensity and durability of Monsta are considerably increased in comparison with GLuc.14) This intrinsic variance in the intensity and stability between Monsta and GLuc resulted in the contrasting optical image between cSimgr7 and the others in Fig. 1C, inset A.
One-Shot Buffer Enables a Labor-Effective Determination of Ligands in Whole-Cell AssaysA buffer cocktail, named “one-shot” buffer was made for time- and labor-effective determination of ligands in the device (Fig. 2).
A preliminary study using cSimgr7 was conducted to show the absolute optical intensities of backgrounds according to a combination of lysis and assay buffers (Fig. 2A). The optical image demonstrates that (i) a combination of lysis buffer of C1 with the TE buffer that carries polyethylene glycol (PEG) exhibits the worst background intensities, whereas the lysis buffers of Promega and C2 showed the lowest background intensities in combination with any of HBSS, TE, or Tris–MgCl2 buffers. This initial study also revealed that cell lysis with C3 and C4 exhibits moderate background intensities. Based on this knowledge, we decided to use a combination of lysis with C4 and assay with HBSS as a basic composition for assays.
Halogenic ions as an additive of the HBSS assay buffer were examined with varying concentrations of bromide and iodide ions (Fig. 2B). The results indicate that addition of halogen ions can contribute to a better S/B ratio, compared to the negative control, C5. Interestingly, an assay buffer comprising 50 mM of iodide exhibited an improved S/B ratio of 9.0.
A series of cocktail buffers, i.e., one-shot buffer, was examined for rapid determination of hormone levels in samples using bioluminescent single-chain probes (Fig. 2C, left). A 6 : 4 mixture of C4 and an HBSS buffer exhibited a considerable S/N ratio of 3.6 with COS-7 cells expressing cSimgr7, and this composition was named C12. The merit of this cocktail was also examined with COS-7 cells carrying pLeu-rich and pSimer-R2 (Fig. 2C, middle and right). This cocktail of C12 enabled us to immediately determine DHT or OHT without the conventional delay caused by cell lysis. The S/B ratios with Leu-rich and Simer-R2 were found to be 4.8 and 4.6, respectively. A corresponding optical image was obtained with a microslide growing COS-7 cells stably expressing cSimgr7 (Fig. 2C, inset A). Typically, the three right channels stimulated with cortisol exhibited 2-times stronger bioluminescence intensities than the three left channels stimulated with the vehicle (0.1% DMSO).
The Optical Device Significantly Reduces the Standard Deviation (S.D.) and Increases Efficiency in the Photon CollectionContribution of the present device to hormone assays was briefly demonstrated with comparison of the S.D.s and optical intensities in the presence or absence of the device aid.
COS-7 cells carrying a single-chain probe were assayed using a conventional luciferase assay kit (E2820, Promega) and luminometer (GloMax 20/20n, Promega) after stimulation of vehicle (0.1% DMSO) or 10−5 M of cortisol for 20 min. The S.D.s in case of the vehicle stimulation were 0.08 and 0.37 with and without aid of the optical device, respectively. Similarly, the S.D.s in case of the cortisol stimulation were 0.59 and 2.21 with and without aid of the optical device, respectively. These comparison reveals that the aid of the device dramatically increase the assay precision up to 4 times.
Furthermore, contribution of the mirror cap in the optical device was examined with the same COS-7 cells carrying the single-chain probe. The optical intensities with the mirror cap were increased up to 48%, compared to those without the mirror cap in case of vehicle-stimulated COS-7 cells. Similarly, the optical intensities with the cap exhibited up to 28% higher than those without the cap in case of cortisol-stimulated COS-7 cells. The comparison reveals that the mirror cap is much beneficial and useful for a weak light determination rather than a strong light. Another role of the cap is to fix the microslide on the deck and thus allows a precise determination of the samples in the system. As a whole, the mirror cap may exert both positive and negative effects on the basal intensity, the S/B ratio, and precision on the detection limit.
Determination of Stress Hormones in Saliva by Using the Device-Aided Assay SystemCortisol levels as a stress marker in human saliva were estimated by using the bioluminescenct-assay device (Fig. 3B). Although some cortisol is bound to globulin in saliva, majority of the cortisol is free in saliva and extracted into the live cells carrying cSimgr7 for emitting bioluminescence.
The above-mentioned assay device reported that the cortisol levels in the 6 saliva samples were ranging from 2.7–9.1×10−9 M. The average and S.D. were 5.1×10−9 M and 2.5×10−9 M, respectively. A corresponding result was obtained with an ELISA assay conducted in parallel for determining the cortisol levels in the same saliva samples. The ELISA assay exhibited that the cortisol levels were ranging from 4.8–9.2×10−9 M. The average and S.D. were 6.3×10−9 M and 1.6×10−9 M, respectively. All the reported values were within the known clinical range of salivary cortisol ranging from 0.2–2.8×10−8 M. The overall assay times of the bioluminescent-assay device and the ELISA kit for cortisol in saliva were 0.5 and 3.5 h, respectively.
Comparing the S.D.s of the two methods, the ELISA assay may be approximately two-times more accurate than the bioluminescent-assay device, whereas the ELISA assay time was 7-times that of the bioluminescent-assay device, i.e., 0.5 vs. 3.5 h.
Multicolor Imaging of Agonistic and Antagonistic Effects of Woman Sex Hormones Using the Device-Aided Assay SystemBifacial estrogenic effects of E2 and chemicals were imaged with green and red colors (Fig. 4). The cells carrying pSimgr-R2 and pSimgr-G4 successfully developed green and/or red luminescence according to the agonistic and antagonistic activities of the chemicals. E2 was found to exert both green and red colors. Of the two colors by E2, green bioluminescence was prominent (S/N ratio in green=4.9). In contrast, OHT exhibited red bioluminescence more than green (S/N ratio in red=4.0). These results are interpreted as E2 dominantly activates a genomic and agonistic signaling pathway, whereas OHT significantly stimulate a nongenomic and antagonistic signaling pathway. DDT slightly elevated green bioluminescence, indicating the agonistic activity.
The overall S.D.s with the device-aid were significantly decreased, where the averaged S.D.s in blue, green, and red region were 0.10, 0.19, and 0.13 in RLU ratios, respectively. These values are one-fourth smaller than those without the device-aid.
The present study demonstrates that a device-aided bioluminescent assay system allowing a precision determination may compensate the intrinsic demerit of weak bioluminescence through reducing the standard deviations. In practice, our study reveals that the increased precision by the aid of assay device contributed (i) to sensitively determine cortisols in saliva samples and (ii) to simultaneously illuminate the agonistic and antagonistic effects of estrogens with excellent precision and convenience. The present study guides researchers on how to deal with the poor intensity issue of bioluminescence.
Among five saliva samples, two samples exhibited a variance between the results of cSimgr7 and ELISA (Fig. 3). This variance may be originated from the intrinsic difference between the two approaches: i.e., an intramolecular binding between an LXXLL motif and the ligand binding domain of glucocorticoid receptor (GR LBD) is taken as the index for cortisol sensing in saliva, whereas the ELISA is based on an antigen–antibody reaction for determining cortisol. The present single-chain probe senses free cortisol that is diffused into the cytosol of living COS-7 cells from the culture medium within 20 min. Cortisol-activated GR LBD transforms and is phosphorylated.19) Therefore, the single-chain probe is optimal to duly reflect ligand-activated molecular actions of GR. On the other hand, the ELISA assay reflects cortisol amounts recognized by the specific antibody. This intrinsic difference in the scheme and molecular mechanism may cause the variation in the results. In conclusion, both the approaches are valuable to determine different features of salivary cortisols.
An ideal bioluminescent-assay system requires a considerable sample throughput for the efficient determination of ligands in many samples. Although conventional luminometers and spectrophotometers may be sensitive to weak light, they require sample replacement in every measurement, i.e., allowing only a single sample measurement at once. In contrast, the present assay device was designed to carry a multichannel conveyer, an optical filter, and a mirror cap. The present assay device enables us to near-simultaneously determine ligands in 6 different samples without the replacement of the sample tube. Furthermore, the optical filters beneath the multichannel conveyer are designed to be easily slidable for selecting colors during assays. With these filters, three colors are near-simultaneously determined during assays. This device provides users convenience and compatibility in the experimental setup of luminometers and spectrophotometers with considerable time-and-labor efficiency (Suppl. Figs. 1B, C).
We previously demonstrated how to labor-effectively construct single-chain bioluminescent probes.2) The previously reported copepoda luciferase, Monsta, were dissected into two fragments via a hydrophilicity search, guided by the recommendation. The N- and C-terminal fragments were then circularly permutated, linked via a flexible linker and further fused with GR LBD and an LXXLL motif (Fig. 1A). In this single-chain probe, all the components for steroid-sensing and bioluminescence emission were integrated into a single molecular backbone. The present assay system was fabricated through combining the merits of single-chain probes emitting bioluminescence with the convenience, sample throughput, and instrumental compatibility of the optical device.
Bioluminescent assays may be categorized into two groups: (i) genetic and transcriptional, and (ii) nontranscriptional assays. A genetic and transcriptional assays, such as the reporter-gene and two-hybrid assays, requires a long ligand stimulation time until the reporter protein is sufficiently accumulated. Furthermore, because the reporter-coding vectors should be located in the nucleus and use its expression machinery, these genetic and transcriptional assays are adequate for protein interactions in the nucleus.
In contrast, the present device that utilizes bioluminescent single-chain probes supports a nontranscriptional assay with the scheme of an intramolecular reconstitution of split-luciferase. The optical probes are expressed beforehand and localized in adequate intracellular compartments of interest. The luminescence intensity is ready to be developed once the cell is stimulated by a ligand. Therefore, this custom-made assay device exerts distinctive privilege in the assay speed and cost.
Optimization of appropriate buffer ingredients for bioluminescence is useful to construct an efficient device-aided bioassay system. Buffers in conventional bioluminescent assays are classified based on their roles, for cell lysis, called “lysis buffer” and for practical measurements, called “assay buffer.” However, these multistep protocols inevitably consume more time and labor. In contrast, the present buffer cocktail named an “one-shot” buffer enables us an onsite determination of ligands without any delay by cell lysis and further mixing of the lysates with an assay buffer in a specific ratio. In our calculation, the conventional protocol consumes at least 30 min (lysis, transferring, and mixing of the lysates with an assay buffer carrying the substrate) before the signal measurement, whereas the present buffer cocktail named “one-shot” buffer spends less than 1 min before the practical measurement.
Among the assay buffer ingredients, iodide interestingly contributed to a better S/B ratio in the present system. A similar feature was also reported by Inouye and Sahara, whose study reported that various halogen ions enhance Gaussia luciferase activity.20) Although the mechanism is unclear, we assume that an anionic circumstance of the polarizable iodide may contribute to a better light emission from the excited coelenteramide in the transition site.
The purpose of the one-shot buffer is to immediately lyse mammalian cells during bioluminescent assays on the assay device. For this intention, we initially combined lysis and assay buffers exerting low background intensity on the basis of Fig. 2A. A strong detergent is a pivotal property of the one-shot buffer for the immediate lysis of cells. Nevertheless, too strong detergents such as the SDS may destroy even the bioluminescent single-chain probes in the cells. Because of this concern, our strategy was to find an appropriate balance of the two sides with a mixture of three detergents that show distinctive chemical properties: i.e., NP-40, TW-80, and SDS. NP-40 and TW-80 are nonionic detergents, which are known for not affecting enzymatic stability during the reaction of firefly luciferase.21) NP-40 provides an excellent lysing property as a detergent, while TW-80 is a moderate detergent owing to its hydrophilicity. The strong lysing property of SDS to cells is inevitable for making an efficient one-shot buffer. However, SDS may harm the bioluminescent probe and act like a double-edged sword. Thus, via trial-and-errors, we kept a minimal percentage of SDS in the buffer (ca. 0.06%) and minimized the exposure time to below 5 s.
Taken together, the present study guides on how to fabricate an efficient hormone assay system with considerable sample throughput and fidelity. For this purpose, we newly fabricated a custom-made assay device and single-chain probes with enhanced optical intensity, and further optimized the assay protocol with new buffer cocktails. The practical advantage of the assay system is demonstrated by determining stress hormones in human saliva and by multicolor-imaging agonistic and antagonistic effects of estrogens with the assay system. The present assay system may be broadly utilized in determining various hormones and chemicals in physiological and environmental samples with a simplified protocol, convenience, and high precision.
This work was supported by a Grant-in-Aid for Young Scientists (B) from Japan Society for the Promotion of Science (JSPS) [Grant number: 23750096], and by a Grant-in-Aid for an A-Step program from Japan Science and Technology Agency (JST) [Grant number: AS232Z02075F].
