Enhanced red-emitting railroad worm luciferase for bioassays and bioimaging

Abstract

A luciferase from the railroad worm (Phrixothrix hirtus) is the only red-emitting bioluminescent enzyme in nature that is advantageous in multicolor luciferase assays and in bioluminescence imaging (BLI). However, it is not used widely in scientific or industrial applications because of its low activity and stability. By using site-directed mutagenesis, we produced red-emitting mutants with higher activity and better stability. Compared with the wild-type (WT), the luminescent activities from extracts of cultured mammalian cells expressing mutant luciferase were 9.8-fold in I212L/N351K, 8.4-fold in I212L, and 7.8-fold in I212L/S463R; and the cell-based activities were 3.6-fold in I212L/N351K and 3.4-fold in N351K. The remaining activities after incubation at 37°C for 10 min were 50.0% for I212L/S463R, 31.8% for I212L, and 23.0% for I212L/N351K, but only 5.2% for WT. To demonstrate an application of I212L/N351K, cell-based BLI was performed, and the luminescence signal was 3.6-fold higher than in WT. These results indicate that the mutants might improve the practicability of this signaling in bioassays and BLI.

Introduction

Bioluminescent methods are gaining more attention because of their sensitivity, selectivity, simplicity, and practical monitoring both in vitro and in vivo.1 In particular, the luciferase reporter has many clinical, diagnostic, and drug discovery applications, and it is frequently used to monitor noninvasively in vivo biological processes, such as gene expression, protein–protein interaction, and disease progression in living tissues, cells, and animal models.2–4 Noninvasive optical monitoring including bioluminescence imaging (BLI) requires the light emitted in the body to overcome tissue attenuation, which is usually greater below a wavelength of 600 nm.5–6 For this reason, red-emitting luciferases are considered advantageous for BLI.7 They are also an important component for constructing multicolor reporter systems, which are vital for monitoring multiple genes.8–13

Beetle luciferases are intriguing because, in the presence of ATP and Mg⁺⁺, they can efficiently catalyze D-luciferin oxidation emitting visible light from green to red.14 Luciferase cDNAs have been cloned from many beetles, but only the one from the railroad worm Phrixothrix hirtus can produce naturally red-emitting luciferase (PhRED) and has an additional advantage of spectral pH insensitivity.15 With the characterization of the firefly luciferase crystal structure (1LCI, 2D1Q, 2D1R, 2D1S, 2D1T),16, 17 site-directed mutagenesis becomes an effective way to explore the relationship between structure and function and to improve the characteristics (e.g., thermostability, activity, or spectrum) of beetle luciferases. A few red-emitting mutants have been constructed from natural yellow-green emitting firefly luciferases (e.g., Photinus pyralis (Ppy),18–25Luciola cruciata (Lcr),26Lampyris turkestanicus,27Luciola italica12); however, their traits such as spectral pH-sensitivity and low activity or poor stability counteract their demanding applications. Although another red-emitting mutant from spectral pH-insensitive click beetle luciferase is available,28 its shorter λmax (613 nm) may be a disadvantage compared with those with a longer λmax in some applications.

PhRED cDNA, whose sequence is optimized for mammalian expression, is available commercially as SLR in a multireporter assay system (Tripluc; Toyobo, Osaka), which has been applied to monitor gene expression.10, 11 Nevertheless, the poor thermostability of wild-type (WT) PhRED, which is also a common problem for all other beetle luciferases23 and can affect the detection efficiency of in vivo BLI,29 limits its application in demanding circumstances. Some attempts have been made to improve firefly luciferase thermostability by mutagenesis (e.g., Ppy,30–32 Lcr,33Luciola lateralis (Lla),34Hotaria parvula (Hpa)35). Among these, mutations of T217 (Lcr) and A217 (Lla) (i.e., I212 of PhRED) to the three most hydrophobic residues (I, V, and L) produced mutants with greater thermostability. Interestingly, 217L was more stable than 217I and 217V, even though both I and V have a larger hydropathy index.36 In Ppy, thermostability was increased by the mutation of negatively charged E354 (corresponding to N351 of PhRED) to positively charged K or R30 and by the combinations of E354K with A215L (217 of Lcr and Lla) or with T214A and I232A.31 These results indicate the fundamental importance of position 354 in firefly luciferase stability, which is verified by the mutant E356R of Hpa (354 of Ppy).35 In addition, mutations of solvent-exposed nonconserved hydrophobic residues to hydrophilic residues in Ppy (e.g., F465R: i.e., S463 of PhRED) improved the thermostability.32

In this study, we aimed at optimizing PhRED mutants for applications. We constructed five mutants (I212L, N351K, S463R, I212L/N351K, and I212L/S463R) and investigated their spectra, extract-based activity, thermostability at 37°C, and their practicability in cell-based real-time monitoring. Cell-based in vivo BLI in selected mutants (N351K, I212L/N351K) were also evaluated. We compared these with the relevant values in the WT.

Results

By using site-directed mutagenesis, we constructed five PhRED mutation plasmids (pCMV-SLR-I212L, -N351K, -S463R, -I212L/N351K, and -I212L/S463R; the corresponding luciferases referred: I212L, N351K, S463R, I212L/N351K, and I212L/S463R) from template plasmid pCMV-SLR, which expressed WT PhRED under the control of the cytomegalovirus (CMV) promoter.

Bioluminescent activity of extract-based mutant luciferases

In the generally used reporter assay, the extract of transfected cells is used for measurements, so we first checked the kinetics and bioluminescent activity of the mutant luciferases using an extract of transfected mouse NIH3T3 cells. All mutants retained similar kinetics (data not shown) to that of WT based on extract (this study) and on purified enzyme.37 We then verified extract-based activity normalized by the amount of luciferase detected by western blotting [Fig. 1(A)] and using the extract of transfected NIH3T3 cells grown in 35 mm dishes for 48 h after transfection. The bioluminescent activity normalized by the amount of luciferase is as shown in Table I and Figure 1(B), in which four mutants (I212L, N351K, I212L/N351K, and I212L/S463R) showed enhanced activity, with the highest being I212L/N351K.

Figure 1. Extract-based activity of WT and mutant PhRED. (A) Western blots using anti-PhRED antibody. Ctrl(−) was not transfected by expression plasmid. Tubulin was used as an internal control. (B) Bioluminescent activity normalized by quantitative analysis of western blots. The bioluminescent activity of WT was set to 1. Error bars indicate the standard deviation (n = 4).

Table I. Comparison of Wild-Type and Mutant PhRED

Luciferases	λmax (nm)	Remaining activity (%)		Extract-based activity	Cell-based activity	BLI
		10 min	20 min
λmax is expressed as mean ± SD (n = 4). Remaining activity is expressed as the mean (n = 4). Extract-based activity normalized by the amount of luciferase detected by western blotting was performed using extracts of NIH3T3 cells grown in 35 mm dishes for 48 h after transfection, and the activity was measured in quadruplicate. The means are shown. Cell-based activity was the sum of the relative light unit during a 24-h monitoring period. BLI is expressed as mean (n = 100). un, means BLI not done.
WT	630 ± 1	5.2	0.5	1.0	1.0	1.0
I212L	630 ± 1	31.8	9.1	8.4	0.9	un
N351K	621 ± 1	3.0	0.3	1.2	3.4	4.3
S463R	631 ± 2	6.2	0.5	0.6	0.7	un
I212L/N351K	619 ± 1	23.0	4.2	9.8	3.6	3.6
I212L/S463R	631 ± 1	50.0	22.1	7.8	0.7	un

Time-dependent thermostability of mutant luciferases

Because the extract of transfected cells is usually hold in room or higher temperature for bioluminescent reporter assays, the thermostability of WT and mutant luciferases at 37°C was examined in NIH3T3 cell extracts. All luciferases lost activity after the incubation of 60 min (data not shown), whereas the remaining luminescent activities were observed after incubation time of 10 and 20 min at 37°C in the order of I212L/S463R>I212L>I212L/N351K>S463R>WT>N351K (Table I). Thus, I212L/S463R is the most stable among WT and other mutants. The I212L and I212L/N351K are the second and third stable mutants, respectively.

Bioluminescent emission spectra of mutant luciferases

In multicolor reporter assay, the spectra of different reporters in extract have to be separated by color filter from each other.10 The bioluminescent spectra of WT and mutant luciferases from transfected NIH3T3 cell extracts were evaluated. All mutants showed the similar spectral shape as WT; three (I212L, S463R, and I212L/S463R) had the same emission maximum as WT (λmax ca. 630 nm), but another two (N351K and I212L/N351K) showed a 10 nm blue shifted maximum (620 nm) yet were still red emitting (Table I; Fig. 2).

Figure 2. Bioluminescent emission spectra of WT and mutant PhRED.

Cell-based activity of mutant luciferases

Real-time monitoring of gene expression was measured in cell-based level.11 The NIH3T3 cell-based luminescent activity of the mutants and WT was monitored primarily for about 100 h in real time [Fig. 3(A)] using the cells grown in 35 mm dishes for about 20 h after transfection. The result demonstrated that I212L/N351K and N351K had the highest activity during the entire monitoring period, but the remaining mutants had a lower cell-based activity than WT. The data from the monitoring of another separate transfection confirmed that only I212L/N351K and N351K did possess much higher cell-based activity than WT [Fig. 3(B)]. It was noted that, compared with WT, the cell-based activity of three mutants (N351K, I212L, and I212L/S463R) was opposite to their extract-based counterparts shown in Figure 1(B). To verify whether the differences were resulted from the separate transfection experiments, we performed another transfection experiment in which we first monitored the cell-based activity for 24 h, and then lysed the cells before we measured the extract-based activity. Total cell-based bioluminescent activity of a 24 h monitoring is as shown Figure 3(B); unexpectedly, the succeeding extract-based activity (data not shown) after real-time monitoring did not agree completely with its counterpart [Fig. 1(B)] measured using the extracts of unmonitored cells, but follow the same pattern as its cell-based counterpart as shown in Figure 3(B). Such a discrepancy among extract-based activity without monitoring-related treatment, cell-based activity, and their following extract-based activity for mutants N351K, I212L, and I212L/S463R was confirmed further by a separated experiment including two batches of transfection: one batch used for cell-based activity monitoring followed by lysis and measurement of extract-based activity, and the other batch used only for measurement of the extract-based activity without monitoring-related treatment. I212L/N351K had the highest activity in both cell-based and extract-based evaluations (Table I), and we considered this to have the greatest potential as a substitute for WT.

Figure 3. Cell-based activity of WT and mutant PhRED. (A) Cell-based luminescent activity in real-time monitored for about 100 h. (B) Total relative light unit during a monitoring period of about 24 h.

BLI

To demonstrate a superior application of mutant to WT, a cell-based BLI experiment was performed for the two mutants (I212L/N351K and N351K) that showed higher cell-based activity. Consistent with the previous cell-based monitoring results, both I212L/N351K and N351K produced a brighter bioluminescent signal than WT [Fig. 4(A)]. The mean signal intensity from 100 individual cells was 3.6- and 4.3-fold higher in I212L/N351K and N351K than in WT, respectively [Table I; Fig. 4(B)].

Figure 4. Representative bioluminescence CCD images of WT and selected mutant PhRED in NIH3T3 cells. (A) Luminescence images of WT (left panel), N351K (middle panel), and I212L/N351K (right panel). The contrast of all images was adjusted equally. (B) Luminescence intensity of WT and mutant luciferases quantified from 100 individual cells. The intensity of WT was set to 1. Error bars indicate the standard deviation.

Discussion

Practicability of enhanced red mutants

Our aim was to exploit the optimized red-emitting mutant luciferase for various extract- and cell-based applications. All mutants retained the similar shape of their emission spectra to WT and were still in the red color region practically. I212L/S463R, I212L, and I212L/N351K had better stability and higher extract-based activity than WT, and the most optimized combination was I212L/N351K [Table I; Fig. 1(B)]. Further cell-based real-time monitoring demonstrated the superiority of I212L/N351K (activity 3.6-fold) and N351K (activity 3.4-fold) to WT. The cell-based BLI data verified that I212L/N351K and N351K were really much brighter than WT [Fig. 4(A)]. Thus, I212L/N351K and N351K would be appropriate candidates for cell-based real-time monitoring and BLI; I212L/N351K, I212L/S463R, and I212L would be useful if the measurement of extract-based activity only is considered. The best candidate for measuring both extract-based and cell-based activities (and even BLI) is I212L/N351K.

Extract-based and cell-based activities

The highest extract-based activity was I212L/N351K (9.8-fold higher than WT), followed by I212L (8.4-fold) and I212L/S463R (7.8-fold). Such a notable increase in activity for I212L/N351K may be explained by the combined mutations of I212L and N351K, which might cause a local change in the microenvironment of the active site. A previously proposed active site model38 will support that the replacement of N351 of PhRED affect the structure of the active site39 for changing stability and activity, because N351 of PhRED corresponds to E354 of Ppy, which locates around the ATP-binding pocket (comprising 316–318, 339–342, and 362) and the luciferin-binding pocket (surrounded by 245–251, 315–317, 341–343, and 346–348).

I212L/N351K and N351K showed consistently higher NIH3T3 cell-based activities than did WT and other mutants, whereas I212L, S463R, and I212L/S463R had lower cell-based activities than WT [Fig. 3(A,B)]. We noted that the cell-based activity [as in Fig. 3(A,B)] was opposite to the extract-based activity [as in Fig. 1(B)] for three mutants (N351K, I212L, and I212L/S463R). One possible explanation is that the kinetic behavior of I212L, I212L/S463R, and N351K was affected by the process of monitoring-related treatments. Railroad worm luciferases can display different physiochemical properties even with a similar primary sequence.37 If the process of monitoring affects the folding and refolding of I212L, I212L/S463R, and N351K in a different way from WT, this may suggest that there are some discrepancies among the cell-based, extract-based (after monitoring), and extract-based (without monitoring) activities in these mutants.

Thermostability of mutant luciferases

The mutants I212L/S463R, I212L, and I212L/N351K exhibited notably improved stability after incubation at 37°C. We speculate that I212L contributes much to the stability, possibly because of the locally structural modulation caused by the change in the methyl position in the substituted residue. Similar cases were also reported in Lcr34 and Lla35; that is, 217L (212 of PhRED) is more stable than 217I and 217V, although I and V have a higher hydropathy index than L,36 and a stronger hydrophobic property is related to a greater stability of protein.40 In addition, I212L/S463R was much more stable than I212L, possibly because of an additive effect of S463R because the stability of S463R itself is slightly higher than that of WT. Finally, we noted that N351K, unlike its counterparts in Ppy (E354K)32 and in Hpa (E356R),35 had lower stability than WT.

Bioluminescence emission spectra of mutant luciferases

Because the blue shift (10 nm) occurred only in N351K and I212L/N351K, we speculate that this is related to N351K, which possibly decreased the polarity of the local microenvironment. Although the color of PhRED is determined mainly by residues 1–344,41 recent studies demonstrate that a flexible loop involving 352–359 of firefly luciferase (349–357 of PhRED) also affects the color.27, 39 The emission of PhRED mutants, such as the unique residue R353 deletion,42 T226N,43 or T226N/A243G,44 is also blue-shifted (6–10 nm). This residue (T226N) and the loop between residues 223–235 are involved in a network of hydrogen bonds with E311, R330, and the loop between residues 352–359 that was suggested to keep the active site structure being sensitive to structural elements for bioluminescence colors.42, 45 Several groups have explored the mystery of color determination, and three mechanisms have been proposed. Recent studies favor the dependence of color on the polarization of the emitter oxyluciferin and the degree of covalent interaction between oxyluciferin and luciferase.46, 47 A small shift of λmax is generally attributed to the local polarity changes in the emitter sites.48, 49 The polarizability of the introduced residue is positively correlated to red shift.50 A blue shift (9 nm) in Ppy was shown to result from a localized decrease in the emitter site polarity.51 Unlike N351K in this study, mutations of the corresponding positions in other green-yellow-emitting firefly luciferases caused a small red shift.32, 38

Materials and Methods

Site-directed mutagenesis and screening for mutants

PhRED cDNA was cloned,15 and its sequence was optimized for mammalian expression in our laboratory.10, 11 The optimized cDNA in mammalian expression vector (pCMV-SLR), whose expression is driven by the CMV promoter, was used as the template for mutation. Site-directed mutagenesis was performed using a KOD-Plus-Mutagenesis Kit (Toyobo) according to the manufacturer's instructions. The whole PhRED gene was confirmed by sequencing (ABI3100; Applied Biosystems, Foster City, CA).

Cell culture and transfection

Mouse NIH3T3 cells (RCB1862) were grown in Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich, St. Louis, MO) supplemented with 10% fetal bovine serum (FBS; ICN Biochemicals, Aurora, OH) in a humidified atmosphere containing 5% CO₂ at 37°C. One day before transfection, cells were seeded in 24-well culture plates (5 × 10⁴ cells/well) or in 35-mm culture dishes (3 × 10⁵ cells/dish). Transient transfection (plasmid: 400 ng/well for 24-well plates; 1000 ng/dish for 35-mm dishes) was performed using Lipofectamine PLUS (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Transfection was performed in quadruplicate in 24-well plates, and a phRL-TK plasmid (10 ng/well) (Promega, Madison, Wisconsin), which expresses the Renilla luciferase (RLuc) under the control of the thymidine kinase (TK) promoter, was cotransfected to normalize the transfection efficiency. Transfection was allowed to proceed for 20–48 h.

Measurement of luciferase activity

All extract-based activity was measured using a luminometer (AB-2250; ATTO, Tokyo) for 20 s. About 48 h after transfection, cells in the 24-well plates were washed once with 300 μl of cold phosphate-buffered saline (PBS) and disrupted in 150 μl of passive lysis buffer (PLB; Toyo Ink, Tokyo). PhRED activity was measured by mixing 25 μl extract with 50 μl of luciferase assay reagent II containing D-luciferin (LARII; Promega), and RLuc activity was separately measured by mixing 25 μl extract with 50 μl of Stop&Glo Reagent containing coelenterazine (Promega). The PhRED activity was normalized by that of RLuc. In transfection in the 35-mm dishes for western blotting, cells were washed once with 1000 μl cold PBS and disrupted by 300 μl mammalian protein extraction reagent (M-PER; Pierce Biotechnology, Rockford, IL), and the activity was measured in quadruplicate by mixing 25 μl of extract with 50 μl of LARII.

Protein concentration and western blotting

The protein concentration of the extracts was measured using a Protein Assay kit (Bio-Rad, Hercules, CA), with bovine serum albumin (Wako, Osaka) as the standard. Sodium dodecyl sulphate (SDS)-urea-polyacrylamide gel electrophoresis (PAGE) and blotting were performed as reported.52 Components of 2 μg protein were resolved by SDS-urea-PAGE. The rabbit anti-PhRED polyclonal antibody was raised against purified recombinant PhRED and used as a primary antibody (1/2500). We used horseradish peroxidase-conjugated anti-rabbit IgG (1/5000) (Jackson ImmunoResearch, West Grove, PA) as a secondary antibody. Antibodies were diluted in Can Get Signal solution (Toyobo). Immunoreactive bands were developed using ECL Plus (GE Healthcare, Freiburg, Germany), visualized on an LAS-4000mini luminescent image analyzer charge coupled device (CCD) chip camera system (Fujifilm, Tokyo), and quantified using Multi Gauge software version 3.0 (Fujifilm).

Measurement of bioluminescence spectra

Bioluminescence spectra were measured using a spectrophotometer (AB-1850S; ATTO) by injecting 15 μl LARII into 15 μl of the extract from the 24-well culture plates. All spectra were corrected for the photosensitivity of the equipment and normalized.

Measurement of time-dependent heat inactivation

At 24–48 h after transfection, cells in the 35-mm dishes were washed once with 1000 μl of cold PBS and disrupted in 750 μl of PLB, and 85 μl of extract was placed on ice (as a control) or incubated at 37°C for 10, 20, 30, 40, 50, or 60 min. After the given treatment time, 85 μl of each extract was placed on ice for at least 5 min, and then 25 μl of the extract was mixed with 25 μl of LARII, and the activity was measured using an AB2250 for 20 s. All measurements were made in triplicate. The remaining activity was expressed as a percentage of the control (on ice).

Monitoring of cell-based activity and comparison with extract-based activity

After about 20 h of transfection, the culture medium in the 35-mm dishes was replaced with 2 ml of fresh DMEM supplemented with 10% FBS, 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid/NaOH (pH 7.0), and 100 μM D-luciferin (Toyobo), and the dishes were sealed with parafilm. The cultures were then incubated at 37°C in a dish-type luminometer (AB-2500 Kronos; ATTO), which was placed in an incubator under 5% CO₂ in air at 20°C. Bioluminescence was monitored for 1 min at 10–20 min intervals. In some trials, monitoring was stopped after about 24 h, and cultures were washed with 2 ml cold PBS and lysed by 300 μl M-PER. Extract-based activity was measured immediately in quadruplicate by mixing 25 μl of extract with 50 μl of LARII and collecting in the AB2250 for 20 s. Western blotting was performed following the protocol described earlier. The cell-based and extract-based activities were compared.

BLI

Two mutants (I212L/N351K and N351K) with cell-based activity greater than that in WT were selected for BLI. Transient transfection and cell-based activity monitoring followed the protocol described earlier. Cell-based activity was monitored after about 20 h of transfection, and when the cell-based activity peaked, BLI was performed using a luminescence microscope Cellgraph (ATTO). CCD images were acquired using 4× objective lens (NA, 0.5; ATTO) at 1 × 1 binning of the 512 × 512 pixel array and with 3 min of exposure time. The luminescence intensity was analyzed using Meta Imaging software version 6.1 (Universal Imaging, Brandywine, PA).

Acknowledgements

We thank Ms. M. Kondoh and Ms. S. Kumata (AIST) for technical assistance.