A Very Bright Far-Red Bioluminescence Emitting Combination Based on Engineered Railroad Worm Luciferase and 6′-Amino-Analogs for Bioimaging Purposes

Beetle luciferases produce bioluminescence (BL) colors ranging from green to red, having been extensively used for many bioanalytical purposes, including bioimaging of pathogen infections and metastasis proliferation in living animal models and cell culture. For bioimaging purposes in mammalian tissues, red bioluminescence is preferred, due to the lower self-absorption of light at longer wavelengths by hemoglobin, myoglobin and melanin. Red bioluminescence is naturally produced only by Phrixothrix hirtus railroad worm luciferase (PxRE), and by some engineered beetle luciferases. However, Far-Red (FR) and Near-Infrared (NIR) bioluminescence is best suited for bioimaging in mammalian tissues due to its higher penetrability. Although some FR and NIR emitting luciferin analogs have been already developed, they usually emit much lower bioluminescence activity when compared to the original luciferin-luciferases. Using site-directed mutagenesis of PxRE luciferase in combination with 6′-modified amino-luciferin analogs, we finally selected novel FR combinations displaying BL ranging from 636–655 nm. Among them, the combination of PxRE-R215K mutant with 6′-(1-pyrrolidinyl)luciferin proved to be the best combination, displaying the highest BL activity with a catalytic efficiency ~2.5 times higher than the combination with native firefly luciferin, producing the second most FR-shifted bioluminescence (650 nm), being several orders of magnitude brighter than commercial AkaLumine with firefly luciferase. Such combination also showed higher thermostability, slower BL decay time and better penetrability across bacterial cell membranes, resulting in ~3 times higher in vivo BL activity in bacterial cells than with firefly luciferin. Overall, this is the brightest FR emitting combination ever reported, and is very promising for bioimaging purposes in mammalian tissues.


Introduction
Bioluminescence, the emission of visible light for communicative purposes by living organisms, has been also extensively used for bioanalytical purposes [1,2]. Bioluminescence is emitted in a wide range of colors ranging from violet to red depending on the organism [3]. Whereas the most common colors are blue for marine organisms, and green-yellow for terrestrial ones, red bioluminescence is very rare, being found exclusively in two marine species and in Phrixotrix spp. railroad worms [3].

Background
Previously, we reported that PxRE displays much higher bioluminescence activity and more red-shifted spectra with 6 -amino-analogs than any other green-yellow emitting luciferase [42,43]. This higher activity was shown to be due to a larger phenolate oxyluciferin binding site, which can better accommodate larger 6 -substituted amino-analogs. Among the analogs previously tested, the 6 -pyrrolidinylluciferin (N5) displayed the highest bioluminescence activity with PxRE luciferase [43]. Thus, in order to find mutant luciferases with improved activity and more far-red-shifted spectra, here we made and screened mutants of PxRE luciferase with 6 -pyrrolidinylluciferin (N5), and investigated the bioluminescence properties with this analog, with 6 -aminoluciferin (NH 2 -LH) and other 6 -substituted-amino-analogs displaying different sizes, including those with rings ranging from 5 to 7 atoms [piperidino (N6), 1-azepanyl (N7) and morpholino (Mor)], and with the FR emitting commercial analog, AkaLumine

FR-Emitting Mutant Selection
We used previously made PxRE luciferase mutants (H209A, C311T, L334R, L246M,  S220R, A246G, S314T, H242K), and prepared other single point mutants (L348V and L348A) for screening their bioluminescence (BL) activity and color with 6 -(1-pyrrolidinyl)luciferin (N5), and compared with the wild-type enzyme. Among the E. coli colonies expressing luciferase mutants, RE-R215K showed highest in vivo BL activity when compared to colonies expressing the wild-type enzyme and other mutants ( Figure 2). We then selected this mutant and characterized its bioluminescence properties, and compared with other luciferin analogs. The mutant R215K stood out for its much higher bioluminescence activity in relation to other analogs and luciferin, and in relation to the wild-type luciferase, and for its further red-shifted spectrum.

FR-Emitting Mutant Selection
We used previously made PxRE luciferase mutants (H209A, C311T, L334R, L246M,  S220R, A246G, S314T, H242K), and prepared other single point mutants (L348V and L348A) for screening their bioluminescence (BL) activity and color with 6´-(1pyrrolidinyl)luciferin (N5), and compared with the wild-type enzyme. Among the E.coli colonies expressing luciferase mutants, RE-R215K showed highest in vivo BL activity when compared to colonies expressing the wild-type enzyme and other mutants ( Figure  2). We then selected this mutant and characterized its bioluminescence properties, and compared with other luciferin analogs. The mutant R215K stood out for its much higher bioluminescence activity in relation to other analogs and luciferin, and in relation to the wild-type luciferase, and for its further red-shifted spectrum.   Table 1 summarizes the kinetic and spectral bioluminescence (BL) properties of WT and mutant RE-R215K luciferases. Both the specific bioluminescence activities with LH 2 and ATP, or with luciferyl-adenylate, were slightly higher for this mutant in relation to those of the wild-type (WT) enzyme with the same substrates. As expected, both the overall and oxidative catalytic constants were also slightly higher for the mutant in relation to the WT enzyme. The K M value for luciferin increased considerably (~4 times) for this mutant, indicating decrease in affinity for the natural substrate, whereas that for ATP decreased almost four times. On the other hand, whereas the catalytic efficiency regarding luciferin was slightly lower for the RE-R215K mutant, that for ATP was almost 5 times higher than for the WT enzyme. The bioluminescence spectrum of R215K mutant with luciferin (629 nm) luciferase also slightly changed, being~6 nm red-shifted in relation to that of the WT enzyme. We then compared the bioluminescence (BL) activity of different luciferin analogs with the mutant R215K ( Figure 3). Among all the 6 -amino-analogs, N5 at the concentration of 1 mM stood out for its considerably higher activity, showing slightly higher activity than WT luciferase with the same substrate, and almost the same activity of 1 mM LH 2 ( Figure 4; Table 2). However, at the concentration of 100 µM, the activity of N5 with R215K mutant luciferase was nearly 3.5 times higher than that with D-luciferin at the same concentration (Figure 4), due to the much lower K M value for N5. On the other hand, the commercial Akalumine-HCl luciferin analog gave orders of magnitude lower activity with both wild-type and R215K mutant luciferases.

Bioluminescence Activity
We then compared the bioluminescence (BL) activity of different luciferin analo with the mutant R215K ( Figure 3). Among all the 6'-amino-analogs, N5 at th concentration of 1 mM stood out for its considerably higher activity, showing slight higher activity than WT luciferase with the same substrate, and almost the same activi of 1 mM LH2 ( Figure 4; Table 2). However, at the concentration of 100 µM, the activity N5 with R215K mutant luciferase was nearly 3.5 times higher than that with D-luciferin the same concentration ( Figure 4), due to the much lower KM value for N5. On the oth hand, the commercial Akalumine-HCl luciferin analog gave orders of magnitude low activity with both wild-type and R215K mutant luciferases.   The results were normalized for P. hirtus wild-type luciferase activity with LH2. The standard deviations of activity averaged 3.15% and ranged from 0.3% to 11.9%. Table 2. Comparison of the bioluminescence spectra peaks and activities of P. hirtus and mutant luciferases with luciferin, 6´-aminoanalogs and AkaLumine.

Substrate K M and Luminescence Kinetics
Previously, we reported that 6 -amino-luciferin analog displays much lower K M value than LH 2 with P. hirtus luciferase. Similarly, the K M for 6 -(1-pyrrolidinyl) analog N5 (0.5 µM) was 14 times lower than that of natural LH 2 (7 µM) for the wild-type luciferase, and 40 times lower (LH2: 40 µM and N5: 1 µM) for the mutant RE-R215K luciferase ( Table 2), indicating that N5 is much better substrate for the R215K mutant luciferase than natural luciferin. As a consequence, the mutant enzyme saturates at much lower concentration of N5 analog. The kinetics of the luminescence reaction with N5 also became slower in relation to that of LH 2 (

Catalytic Constant and Efficiencies
The catalytic constants of wild-type and RE-R215K mutant luciferases for luciferin (LH2) and N5 were similar to each other (Table 1). However, the whereas catalytic efficiencies of WT and R215K luciferases are ~12 and ~8 times higher for N5 in relation to LH2, respectively, the catalytic efficiency of R215K luciferase with N5 is ~40 times more efficient than wild-type enzyme with LH2. Figure 6 and Table 2 show the bioluminescence spectra and spectral properties of WT and mutant luciferases with luciferin and 6'-amino-analogs. Among all analogs tested, N5 (650 nm) and N7 displayed the most red-shifted spectra (655 nm) ( Figure 6). The spectra were 31-33 nm blue-shifted only with regard to that of the commercial analog, AkaLumine (681-683 nm).

Catalytic Constant and Efficiencies
The catalytic constants of wild-type and RE-R215K mutant luciferases for luciferin (LH 2 ) and N5 were similar to each other (Table 1). However, the whereas catalytic efficiencies of WT and R215K luciferases are~12 and~8 times higher for N5 in relation to LH 2 , respectively, the catalytic efficiency of R215K luciferase with N5 is~40 times more efficient than wild-type enzyme with LH 2 . Figure 6 and Table 2 show the bioluminescence spectra and spectral properties of WT and mutant luciferases with luciferin and 6 -amino-analogs. Among all analogs tested, N5 (650 nm) and N7 displayed the most red-shifted spectra (655 nm) ( Figure 6). The spectra were 31-33 nm blue-shifted only with regard to that of the commercial analog, AkaLumine (681-683 nm).

Thermostability
We also compared the thermostability of R215K mutant with the wild-type enzyme (Figure 7). The mutant luciferase displayed slightly improved thermostability, retaining ~91% activity when incubated at 22 °C during 24 h, in comparison with the wild-type enzyme that retained only ~70%.

Thermostability
We also compared the thermostability of R215K mutant with the wild-type enzyme (Figure 7). The mutant luciferase displayed slightly improved thermostability, retaining 91% activity when incubated at 22 • C during 24 h, in comparison with the wild-type enzyme that retained only~70%.

Bioluminescence of Bacteria Expressing R215K with 6'-(1-pyrrolidinyl)luciferin
In order to check the cell membrane penetrability and in vivo luminescence properties of such analogs, we tested the in vivo BL activity of E.coli liquid cultures after mixing LH2 and the N5 at neutral and acidic (pH = 5.0) pH (Figure 8). At 1 mM concentration and at pH~7.0, when the membrane permeability is expected to be lower due to the ionized substrate carboxyl group, N5 gave almost the same in vivo bioluminescence activity of LH2 at pH 5.0. However, when provided at pH 5.0, N5 showed about three times higher

Bioluminescence of Bacteria Expressing R215K with 6 -(1-pyrrolidinyl)luciferin
In order to check the cell membrane penetrability and in vivo luminescence properties of such analogs, we tested the in vivo BL activity of E. coli liquid cultures after mixing LH 2 and the N5 at neutral and acidic (pH = 5.0) pH (Figure 8). At 1 mM concentration and at pH~7.0, when the membrane permeability is expected to be lower due to the ionized substrate carboxyl group, N5 gave almost the same in vivo bioluminescence activity of LH 2 at pH 5.0. However, when provided at pH 5.0, N5 showed about three times higher bioluminescent activity than LH 2 at the same pH and concentration. These results show that N5 has much better penetrability across the biological membranes, and therefore is a better substrate for bioimaging. Additionally, the in vivo bacterial bioluminescence with N5 showed a more sustained kinetics, with a half-life of 33 min, in relation to LH 2 which showed a half-life of 20 min (Figure S1 Supplemental information). Figure 7. Thermostability of Phrixotrix hirtus red emitting luciferase (bold line) and its R215K mutant (dashed line). The standard deviations of activity averaged 2.4% and ranged from 0.3% to 5%.

Bioluminescence of Bacteria Expressing R215K with 6'-(1-pyrrolidinyl)luciferin
In order to check the cell membrane penetrability and in vivo luminescence properties of such analogs, we tested the in vivo BL activity of E.coli liquid cultures after mixing LH2 and the N5 at neutral and acidic (pH = 5.0) pH (Figure 8). At 1 mM concentration and at pH~7.0, when the membrane permeability is expected to be lower due to the ionized substrate carboxyl group, N5 gave almost the same in vivo bioluminescence activity of LH2 at pH 5.0. However, when provided at pH 5.0, N5 showed about three times higher bioluminescent activity than LH2 at the same pH and concentration. These results show that N5 has much better penetrability across the biological membranes, and therefore is a better substrate for bioimaging. Additionally, the in vivo bacterial bioluminescence with N5 showed a more sustained kinetics, with a half-life of 33 min, in relation to LH2 which showed a half-life of 20 min (Figure S1 Supplemental information).

Structure and Function Relationships
The residue R215 (R218 in P. pyralis firefly luciferase) is an invariable residue in beetle luciferases, which is located at the bottom of luciferin binding site, close to oxyluciferin phenolate [43]. This arginine guanidinium group has been claimed to be important for bioluminescence spectra, since its mutation in yellow-green emitting firefly luciferase and in P. vivianii railroad worm green emitting luciferase promoted red-shifts [44,45]. In P. hirtus red emitting luciferase, however, the mutation of this residue, had almost no effect on the spectrum [45]. It was suggested, and later confirmed [42], that in PxRE luciferase this residue side-chain is spatially displaced from oxyluciferin phenolate, in relation to the closer green emitting P. vivianii luciferase. Whereas the mutation R215K, performed in this work, did not changed the electrostatic environment around excited oxyluciferin phenolate, as it preserves the positive charge, the shorter side-chain of lysine may contribute to increase the size of the cavity and the distance of the side-chain in relation to oxyluciferin phenolate group, explaining the accommodation of the larger substituted 6 -amino-analogs resulting in higher substrate affinity and activities, and red-shifted spectra.

Comparison RE-R215K/N5 with Other Beetle FR Emitting Luciferin-Luciferases
Overall, the above results clearly show that the combination of P. hirtus R215K mutant luciferase with N5 is the brightest FR emitting combination ever reported. Even if this combination is blue-shifted in relation to the commercially provided AkaLumine with firefly luciferase and other reported NIR emitting analogs (>700 nm), its much higher bioluminescence activity and catalytic efficiency may compensate the slightly lower tissue penetrability of bioluminescence at these more red-shifted wavelengths. Furthermore, the higher penetrability of N5 across cell membranes, the higher life time of in vivo bioluminescence, the much lower substrate requirement in in vivo assays using N5, and the higher thermostability of R215K luciferase in relation to the wild-type PxRE enzyme, make this combination a very promising tool for deep tissue bioimaging in mammalian organisms.

Site-Directed Mutagenesis
The mutants R215K were obtained by site-directed mutagenesis using a ThermoFisher Scientific kit (Catalog F-530XL and F530L). The plasmids containing the luciferase cDNAs were amplified using Phusion polymerase and 2 complementary primers containing the desired mutation, using a thermal cycler (1 cycle 98 • C, 2 min; 25 cycles 98 • C, 30 s; 55 • C, 1 min and 72 • C, 6 min). After amplification, mutated plasmids containing staggered nicks were generated. The products were treated with Dpn I in order to digest non-mutated parental plasmids, and used directly to transform E. coli XL1-Blue cells (Agilent, Santa Clara, CA, USA). The following forward and respective reverse primers were used (the mutation codon is highlighted in bold): RE-R215K forward: GCA TCA CCA TCA AAT TCG TCC ACA GC and Reverse: GCT GTGGACGAATTTGATGGTGATGC.

Measurement of Luciferase Activity
Luciferase bioluminescence intensities were measured using an AB2200 (ATTO; Tokyo, Japan) luminometer. The assays were performed by mixing 5 µL of 40 mM ATP/80 mM MgSO 4 with a solution consisting of 5 µL of luciferase and 85 µL of 0.5 mM luciferin in 0.10 M Tris-HCl pH = 8.0 at 22 • C. All measurements were done in triplicate for at least three independent luciferase preparations, and the averages and the standard deviations were reported in the figures.

Kinetics Measurements and K M Determination
The K M assays for luciferin were performed by mixing 5 µL of 40 mM ATP, 80 mM MgSO 4 in a solution containing 10 µL of luciferase, 85 µL of 0.10 M Tris-HCl (pH 8.0) and luciferin and N5 at final concentrations between 0.01 and 1 mM. The K M assays for ATP were performed by mixing 5 µL of 80 mM MgSO 4 in a solution containing 10 µL of luciferase, 85 µL of 0.10 M Tris-HCl (pH 8.0) and ATP at final concentrations in the range of 0.02 to 2 mM. Both assays were performed in triplicate. The K M values were calculated using Lineweaver-Burk plots taking the peak of intensity (I 0 ) as a measure of V 0 , according to Equation (S1) (Supplemental information). All measurements were done in triplicate and averages were reported. The kinetics of luminescence reaction was performed using a TD-III luminometer (Japan). In the assay for the reaction starting with LH 2 and ATP, 5 µL of a solution consisting of 80 mM MgSO 4 and 40 mM ATP were mixed to a solution consisting of 10 µL of luciferase, 10 mM LH 2 in 85 µL of 0.10 M Tris-HCl buffer pH = 8.0.

Luciferyl-Adenylate Synthesis
The luciferyl-adenylate was prepared using LH 2 acid and AMP sodium salt (SIGMA) by following a previously described preparation procedure for D-luciferyl-adenylate [46]. Luciferyl-adenylate was analyzed with silica gel TLC (moving phase: ethyl acetate/ethanol/ water (5:3:2), followed by revelation by fluorescence with a UV lamp. Luciferyl-adenylate displayed yellowish fluorescence with an R f = 0.68 (R f(luciferin) = 0.87 with greenish fluorescence). A luciferyl-adenylate concentration was estimated from stoichiometric amounts of luciferin and ATP used for its synthesis. According to such estimations, luciferyl-adenylate concentration in stock solution are in the range between 5 to 10 mM.

Determination of k cat and k ox
The overall k cat and k ox were determined according to Equation (S2) (Supplemental information), by calculating the ratio of luminescence activities in cps by the number of luciferase molecules based on the specific bioluminescence activities measured with luciferin and ATP (overall k cat ), and with luciferyl-adenylate (k ox ), respectively. Because the absolute value of cps, in photons/s, could not be determined, the absolute values of k cat and k ox in s −1 could not be determined, and therefore the values were reported in cps (counts for second). Although these values are not absolute, they can be safely used as relative values of catalytic constants.

6 -Substituted Amino Analogs
All the 6 -substituted amino luciferin analogs ( Figure 1) were synthetized as previously reported [33]. Stock solutions of 10 mM were prepared in DMSO and kept at −20 • C in the dark.

Bioluminescence Spectra
Bioluminescence spectra reported here were recorded in ATTO Lumispectra spectroluminometer (Tokyo, Japan) with cooled CCD camera. For the in vitro bioluminescence recorded using the spectroluminometer, 5.0 µL of luciferases were mixed with 90 µL of 0.10 M Tris-HCl pH 8.0, 5 µL of specific substrate (10 mM D-luciferin; luciferyl-adenylate or 6 -amino-luciferin analogs), and 5 µL of 40 mM ATP/80 mM MgSO 4 . The bioluminescence spectra were measured in triplicate for at least three independent luciferase preparations, and the reported spectra represent the average. The peaks were manually estimated, and the peak variation was ±2.5 nm. Above 620 nm, due to the associated lower energy at higher wavelengths, we assumed peak errors of ±3 nm.

Concluding Remarks
By using combinatory chemistry with 6 -amino-luciferin analogs and P. hirtus red emitting luciferase site-directed mutagenesis, we report the selection of novel brighter FR emitting combinations. The combination of P. hirtus R215K red emitting mutant luciferase and N5, stood out for its far-red shifted emission (650 nm) and for the highest bioluminescence activity ever reported among FR emitting luciferin-luciferase systems, displaying more than three times higher bioluminescence activity than with natural LH 2 at 100 µM concentration, and about 1000 higher activity than with commercial AkaLumine. The combination also displayed more sustained and much higher in vivo bioluminescence activity in E. coli cells, when compared with the wild-type luciferase with LH 2 . Altogether, these properties make RE-R215K/N5 combination very promising tool for biolumines-cence imaging of biological and pathological processes in deep mammalian tissues and organisms.
Supplementary Materials: The following are available online at https://www.mdpi.com/1422-006 7/22/1/303/s1. Author Contributions: V.R.V. idealized the work, characterized BL spectra, activities and kinetics of R215K mutant with N5 analog, supervised the students activities, and wrote the manuscript. V.R.B. analyzed the effect of amino-analogs with wild-type luciferase, produced R215K mutant and measured BL spectra. D.R.d.S. measured the bioluminescence activities, obtained the K M for ATP, luciferin and analogs, and the BL spectra for N6 and N7 analogs. G.F.P. purified proteins, measured bioluminescence activities with D-luciferin, measured thermostability. M.K. synthesized N6, N7 amino analogs. T.H. idealized the synthesis of all amino-analogs and contributed with discussion of results. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement: Not applicable for studies not involving humans or animals.

Informed Consent Statement:
Not applicable for studies not involving humans.