Photochemical & Photobiological Scienceskoksharov83.narod.ru/docs/lab/Koksharov_PPS-2013_final.pdffinal volume) contained ∼20 ng insert, ∼50 ng vector, 1× T4 DNA ligase buﬀer,

Photochemical &Photobiological Sciences

PAPER

Cite this: Photochem. Photobiol. Sci., 2013,

12, 2016

Received 23rd July 2013,Accepted 29th August 2013

DOI: 10.1039/c3pp50242b

www.rsc.org/pps

Strategy of mutual compensation of green and redmutants of firefly luciferase identifies a mutation ofthe highly conservative residue E457 with a strong redshift of bioluminescence†

Mikhail I. Koksharov* and Natalia N. Ugarova

Bioluminescence spectra of firefly luciferases demonstrate highly pH-sensitive spectra changing the color

from green to red light when pH is lowered from alkaline to acidic. This reflects a change of ratio of the

green and red emitters in the bimodal spectra of bioluminescence. We show that the mutations strongly

stabilizing green (Y35N) or red (H433Y) emission compensate each other leading to the WT color of

firefly luciferase. We further used this compensating ability of Y35N to search for strong red-shifting

mutations in the C-domain of firefly luciferase by random mutagenesis. The discovered mutation E457K

substantially increased the contribution of the red emitter and caused a 12 nm red shift of the green

emitter as well. E457 is highly conservative not only in beetle luciferases but also in a whole ANL super-

family of adenylating enzymes and forms a conservative structural hydrogen bond with V471. Our results

suggest that the removal of this hydrogen bond only mildly affects luciferase properties and that most of

the effect of E457K is caused by the introduction of positive charge. E457 forms a salt bridge with R534

in most ANL enzymes including pH-insensitive luciferases which is absent in pH-sensitive firefly luci-

ferases. The mutant A534R shows that this salt bridge is not important for pH-sensitivity but considerably

improves in vivo thermostability. Although E457 is located far from the oxyluciferin-binding site, the pro-

perties of the mutant E457K suggest that it affects color by influencing the AMP binding.

Introduction

Firefly (beetle) luciferases (EC 1.13.12.7) catalyze the ATP-dependent bioluminescent reaction which has the highestquantum yield (45–60%) among bioluminescent systems.1

These enzymes are widely used in a variety of in vitro andin vivo applications including ATP-related assays,2 molecularimaging, monitoring of gene expression3,4 and molecularsensing of protein–protein interactions and different ana-lytes.5,6 The luciferase reaction proceeds in two steps: at thefirst step luciferin (LH2) and Mg-ATP react to form a luciferyladenylate; this intermediate product is then oxidized by O2

through several intermediates to give AMP, PPi, CO2 and anelectronically excited product oxyluciferin (LO).7,8 Relaxation ofoxyluciferin is accompanied by the emission of visible light.Luciferase is composed of a big N-domain (1–436 aa) and asmall C-domain (∼443–548 aa) connected by a flexible loop.9,10

The substrate binding site is located in the N-domain, but the

C-domain is crucial for the reaction and rotates relative to theN-domain at each step of the reaction. This domain rotationmechanism is common for the whole ANL superfamily of ade-nylating enzymes to which firefly (beetle) luciferases belong.11

However, in the case of firefly luciferase the structure of thesecond – light emitting – conformation of luciferase wasobtained only recently.10

Luciferases from various beetles emit light from green tored (λmax = 536–623 nm) depending on the species. Since thereaction product is the same, these differences are determinedby the structure of a luciferase. In contrast to luciferases fromnon-firefly beetles that are generally pH-insensitive,12 bio-luminescence of firefly luciferases typically shifts from greento red light when lowering the pH from alkaline to acidic orincreasing the temperature.13 This red shift is usuallyexplained by switching between two different molecular formsof the product (green and red emitters).8,14 Chemilumine-scence studies suggest that red emission is the defaultoutcome of the reaction while green emission requires impos-ing additional interactions on oxyluciferin.13 Cases of mono-modal bioluminescence show that the green and red spectralcomponents have half-widths of 65–67 and 55–60 nm, respecti-vely,12,15 whereas a larger width of a spectrum indicates its

†Electronic supplementary information (ESI) available. See DOI:10.1039/c3pp50242b

Department of Chemical Enzymology, Faculty of Chemistry, Lomonosov Moscow

State University, Moscow, 119991, Russia. E-mail: [email protected]

2016 | Photochem. Photobiol. Sci., 2013, 12, 2016–2027 This journal is © The Royal Society of Chemistry and Owner Societies 2013

Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article OnlineView Journal | View Issue

www.rsc.org/pps

http://dx.doi.org/10.1039/c3pp50242b

http://pubs.rsc.org/en/journals/journal/PP

http://pubs.rsc.org/en/journals/journal/PP?issueid=PP012011

bimodality. However, despite extensive mutagenesis,12,13,16–23

structural9,10 and computational24,25 studies, the exact natureof these emitters remains controversial and a clear mechanismof how luciferase structure determines the color of bio-luminescence remains unknown. In pH-sensitive firefly luci-ferases the most common types of mutations are the ones thatchange the ratio of the red and green emitters under differentconditions. Most of them either increase the contribution ofthe red emitter leading to bimodal orange and redmutants19,20,26 or stabilize the green emitter leading to lesspH-sensitive green mutants.15,18,27 Few mutations in the luci-ferin-binding pocket (e.g. S286T,15,17,20 G315A, T345A,28 andN231T12) leave the single red emitter. Another type of mutationshifts the λmax of the individual spectral components.13,16

We have analyzed the mutual effect of mutations stabilizingthe green or red emission of Luciola mingrelica firefly lucifer-ase. The “green” pH-insensitive mutant Y35N proved to be anefficient tool to distinguish mutants causing a strong red shiftof bioluminescence. Since very little is known about the role ofthe C-domain of luciferase in the determination of bio-luminescence spectra, we then employed random mutagenesisto find new positions in the C-domain that change the color ofbioluminescence. The thermostable mutant TS29 with anadditional mutation Y35N was used in a parent form in thisscreening. These efforts led to the identification of a strongred-shifting mutant of the residue E457 which is highly con-servative not only in beetle luciferases but also in the wholeANL superfamily. We further investigated the role and inter-actions of this residue through site-directed mutagenesis andshowed their importance for bioluminescence color and stabi-lity of luciferase.

Experimental proceduresMaterials

Na-ATP (cat. no. A2383), bovine serum albumin, dithiothreitol(DTT), and yeast extract (cat. no. Y-0500) were purchased fromSigma-Aldrich (St. Louis, USA), D-luciferin was from Lumtek(Moscow, Russia), TaqSE DNA polymerase was from Sibenzyme(Novosibirsk, Russia), Taq DNA polymerase and T4 DNA ligasewere from Sileks (Moscow, Russia), oligonucleotide primerswere obtained from Sintol (Moscow, Russia), bacto-tryptonewas from Becton Dickinson (USA), and lactose 1-hydrate wasfrom Panreac (Montcada, Spain). Competent E. coli cells wereprepared and transformed according to a method developedby Tu et al.30 All other reagents were of analytical grade orhigher, and solutions were prepared using Milli-Q water.

Construction of single and multi-point mutants

The mutants Y35N,18 A217T/S222T, H433Y,31 and TS (pre-viously designated as 4TS)29 were obtained earlier. The newmutants were prepared by overlap extension PCR with forwardand reverse mutagenic primers (ESI Table S1†). The presenceof mutations was confirmed by sequencing. The mutants

combining multiple substitutions were constructed by restric-tion and ligation.

Random mutagenesis by error-prone PCR and mutant libraryconstruction

The pLR4 plasmid (GenBank No. HQ007052) which encodesthe L. mingrelica luciferase gene was constructed earlier.18 ThepLR4 plasmid used here contained the additional mutationY35N outside of the region mutated. This resulted in greenbioluminescence of E. coli colonies compared with orange-yellow bioluminescence in the case of WT.18 Random muta-genesis of the 470 base pair region flanked by BamHI and ApaIrestriction sites was performed by error-prone PCR.32 Primersf_XhoI and r_T7term were used as forward and reverseprimers, respectively. The PCR reaction mixture (50 μl) con-tained 10 mM Tris-HCl (pH 8.3 at 25 °C), 50 mM KCl, 7 mMMgCl2, 0.3 mM MnCl2, 0.2 mM dATP, 0.2 mM dGTP, 1 mMdCTP, 1 mM dTTP, 20 pmol of each primer, ∼2 fmol of pLR3,and 2.5 units of Taq DNA-polymerase. PCR was performedwith an automatic thermal cycler Tercik (DNA-Technology,Russia) under the following conditions: 95 °C, 1 min; 30 cyclesfor 1 min at 94 °C, 1.3 min at 53 °C, 1 min at 72 °C; then10 min at 72 °C. These conditions should lead to an error fre-quency of 2–3 substitutions or approximately one amino acidper ∼500 bp gene region.33 The mutagenic PCR product wasgel purified by using the QIAEX II kit (Qiagen, Germany) andthen digested with BglII and ApaI. This restriction product wasgel purified and ligated into pLR4 previously treated with thesame restriction enzymes. A typical ligation reaction (10 µlfinal volume) contained ∼20 ng insert, ∼50 ng vector, 1× T4DNA ligase buffer, 1 Weiss Unit T4 DNA ligase, and 5% (w/v)PEG-8000 and was incubated at 16 °C for 1 hour. E. coliXL1-blue cells were transformed with the resulting mutantplasmids and plated onto Luria–Bertani (LB) agar sup-plemented with 100 μg ml−1 ampicillin.

Screening for shifts of bioluminescence color

E. coli colonies harboring mutant luciferase genes were grownovernight at 37 °C on LB agar plates supplemented with 100 μgml−1 ampicillin. The in vivo bioluminescence of the colonieswas registered photographically according to the following pro-tocol: plates were filled with a thin layer of a 0.5 mM luciferinsolution in 0.1 M Na-citrate buffer (pH 5.0), shaken just beforeeach measurement (in a dark room) and photographed with aCanon PowerShot A530 digital camera (Canon, Malaysia). The“White balance” function of the camera was set to “Cloudy” topreserve natural colors. The screening step was performedquickly within 3–4 minutes. Then the luciferin solution wasremoved and the plates were incubated at 45 °C for 20–40 minto evaporate the remaining liquid. The colonies showing achange of color were toothpicked onto new plates. Several colo-nies carrying a parent form of luciferase were also transferredonto these plates as a control. The colonies were grown over-night at 37 °C and then incubated at room temperature for4–8 h. Then the colonies were screened for in vivo bio-luminescence to assess their brightness and color.

Photochemical & Photobiological Sciences Paper

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci., 2013, 12, 2016–2027 | 2017

Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online

Misha

Textbox

BglII


To compare in vivo thermostability of mutants, the bio-luminescence of their E. coli colonies was photographed asdescribed above, and then the luciferin solution was removedand the plate was incubated at 47–53 °C for various periods oftime. Then the remaining bioluminescence was photographed.

Protein expression and purification

The WT and its mutants were cloned into the expressionvector pETL7 (GenBank No. HQ007050)19 or pETL418 that wereconstructed earlier. The pETL7 plasmid encodes the luciferaseprotein with two differences compared with the native enzyme:additional N-terminal sequence MASK- and the C-terminalAKM peptide changed to the SGPVEHHHHHH. The pETL4plasmid encodes the luciferase protein containing theadditional N-terminal 24 amino acid sequence including His6-tag: MGSSHHHHHHSSGLVPRGSHMASK. The WT luciferaseand its mutants were expressed as His6-tagged proteins inE. coli BL21(DE3)CodonPlus cells according to the lactose-based autoinduction method and purified using Ni-affinitychromatography as described earlier.29 After the purification,enzymes were obtained in HB buffer containing 300 mM imid-azole, 2 mM EDTA, and 1 mM DTT and generally remainedfully active for up to 1 month when stored at 4 °C. For thelong-term storage the proteins were transferred to GF bufferand stored at −80 °C.29

Bioluminescence emission spectra

Bioluminescence spectra were obtained using a Perkin-ElmerLS50B luminescence spectrometer operated in the “bio-luminescence” mode at a slit width of 10 nm as described pre-viously.18 Data were automatically corrected for the spectralresponse of the R928 photomultiplier tube using the FLWinLab software. Generally, the spectra selected for the analy-sis were recorded when the decrease in intensity during therecording interval did not exceed 5%. If the emission wasunstable due to the luciferase inactivation at high tempera-tures, the spectra were corrected for the decrease of intensityduring the time of measurement. Spectra were smoothedusing a Quadratic Golay–Savitzky filter in the FL WinLabsoftware.

Enzyme activity and kinetic parameters

Luciferase activity was determined as described previously29

using an FB12 luminometer (Zylux, USA) equipped with a“Diluter 2075” injector (LKB, Sweden). Maximal intensity oflight emitted during the enzymatic reaction at saturating con-centrations of substrates (flash-height based activity assay) wasused as a measure of activity. The polystyrene tubes contained0.35 ml of 1.7 mM ATP in 50 mM Tris-acetate buffer (pH 7.8)containing 10 mM MgSO4, 2 mM EDTA (buffer AB) and 5 µl ofa luciferase solution. Assay was initiated by injecting 0.15 mlof 0.5 mM luciferin (in the same buffer) and the bio-luminescence intensity was registered at room temperature(20–25 °C). The final concentrations of LH2 and ATP were 0.15and 1.2 mM, respectively, in a volume of 0.5 ml. The activity

was expressed in relative light units (RLU s−1) of theluminometer.

The values of Km and Vmax for LH2 and ATP were deter-mined from bioluminescence activity assays. The concen-tration of one substrate was maintained at saturation and theconcentration of the other substrate was varied (0.012–1.8 mMATP and 0.014–0.5 mM LH2). The stock solution of 13 μg ml−1

luciferase was used in the experiment and was further diluted100-fold in a reaction tube during the activity assay. Kineticconstants were calculated from a Michaelis–Menten graphusing non-linear regression in GraphPad Prism version 5.02.The kinetic curves obtained during the measurement of Km

values for ATP at 0.3 mM LH2 and 1.2 mM ATP were used tocalculate the integrated activity (90 s).

Kinetics of thermal inactivation

The solution of the purified luciferase (13 μg ml−1) was pre-pared in 50 mM Tris-acetate buffer (pH 7.8) containing 20 mMMgSO4, 2 mM EDTA, and 0.2 mg ml−1 BSA and was thenstored on ice. In the case of fast inactivation (half-life <30 min) aliquots of 50 μl were placed in 8 thin wall 0.5 mlmicrotubes. In the case of slow inactivation (half-life > 30 min)1.5 ml of a luciferase solution was placed in a 1.7 ml micro-tube. Microtubes were incubated at 42 °C. At given times, 50 μlvolume or a tube with a 50 μl aliquot was removed and cooledon ice for at least 15 min prior to the activity assay. A neutralfilter of ∼0.14% transmittance (Photoptic filters, Russia) wasplaced in the cuvette compartment of the luminometer to keepthe bioluminescence intensity within the dynamic range. Theenzyme activity was expressed as a percentage of the initialactivity. Half-lives were calculated using the first-order reactionrate constants obtained from semi-logarithmic plots of the per-centage of activity versus time.

Bioinformatics

A homology model of the Lml structure was generated pre-viously based on its similarity to Luciola cruciata luciferase.18

Multiple alignments of luciferase amino acid sequences wereperformed using the ClustalW algorithm in BioEdit 7.0.4.1program.

ResultsIn vivo compensation of mutations that stabilize green or redemission of firefly luciferase

Several mutants of Lml affecting the color and pH-sensitivityof its bioluminescence were analyzed earlier in our laboratory.The mutation H433Y greatly increased the contribution of thered emission to the bimodal spectra of Lml leading to the redbioluminescence.31 In contrast, the mutants Y35N and Y35Hgreatly stabilized the green emission, so the bioluminescencespectra did not show a significant increase of the red emissionat high temperatures and low pH.18 The mutant A217T/S222Tobtained previously by random mutagenesis (unpublishedresults) demonstrated a large shoulder in the red region. At

Paper Photochemical & Photobiological Sciences


Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online

Misha

Textbox

incubation at 37-53 °C - in a dry incubator box


the first part of this work, we have constructed and character-ized the mutants Y35N/H433Y and Y35N/A217T/S222T tounderstand to what extent the mutation Y35N can compensatefor the strong red-shifting effect of the mutant H433Y and thesmaller effect of the mutant A217T/S222T.

The intensity and color of bioluminescence can be con-veniently assessed in vivo after luciferase expression in E. colicells. Moreover, the E. coli intracellular environment makes itpossible to discern pH-insensitive mutants by in vivo color.18

In E. coli cells the color of emission of pH-sensitive luciferasesis shifted to the red region: colonies emit yellow-orange lightcompared with yellow-green emission in vitro in the reactionbuffer (ESI Fig. S1†). Therefore, we have used in vivo bio-luminescence for the primary comparison of spectral effects indifferent mutants of luciferase and during screening ofmutant libraries.

The substitutions Y35N and H433Y compensated for eachother in vivo: in contrast to green and red colonies of themutants Y35N and H433Y, the colonies of the double mutantY35N/H433Y emitted yellow-orange light similar to WT (ESIFig. S2†). The visible brightness of all four forms of luciferasewas similar. The strong effect of the substitution Y35N comple-tely suppressed the weak red shift of the mutant A217T/S222Tand the mutant Y35N/A217T/S222T formed green colonies.These properties were later confirmed for the purified pro-teins. This shows that the mutant Y35N could serve as a usefultool to discern mutants with a strong red shift during thescreening of mutant libraries.

Rationale for random mutagenesis. Identification of themutant E457K

Very few mutations in the C-domain of luciferase were knownthat affect bioluminescence spectra: the less pH-sensitivemutant F467R27 and the red mutant P452S.23 While this workwas being done, Nazari et al. reported that the introduction ofa disulfide bond (P451C-V469C) also causes a red shift of bio-luminescence.22 At the same time, complete removal of theC-domain results in a red-emitting luciferase with very lowactivity.34 Identification of new red-shifting mutants in theC-domain would help one to understand its role in the deter-mination of the color of bioluminescence. We decided to usethe compensating ability of the mutant Y35N to find suchmutants in the C-domain.

We subjected the region of 390–549 amino acid residues torandom mutagenesis and screening because of the convenientrestriction sites (BglII and ApaI). A parent enzyme for the muta-genesis was the substitution Y35N in the thermostable mutantTS.29 The ability of the mutant Y35N to strongly stabilize greenemission promoted the preferential detection of mutants witha strong red-shifting effect. The thermostable mutant was usedfor two reasons: first, it showed higher brightness of the colo-nies compared with WT; second, a thermostable enzyme canpotentially tolerate a wider range of mutations.35 Around10 000 colonies were screened. The mutagenesis conditionsgave about 75% active clones in the library. Several dim andtwo relatively bright yellow-orange colonies were identified

when registering in vivo bioluminescence. The more orangishbright colony was selected for further characterization to avoidpossible selection of the mutant H433Y. Sequencing showedthat the change of bioluminescence color was due to the sub-stitution E457K. The screening of a plate containing thismutant is shown in ESI Fig. S3.†

The effects of the substitution E457K on in vivo bio-luminescence were similar to those of H433Y (ESI Fig. S4 andS6†). The colonies of the parent mutant TS emitted yellow-orange in vivo bioluminescence, whereas the mutants E457K/TS and Y35N/TS formed red and green colonies, respectively.Combining these two mutants resulted in the mutant Y35N/E457K/TS with yellow-orange bioluminescence like TS. TS andthese three mutants showed similar in vivo brightness.However, when the mutant E457K was subcloned into the WTluciferase its bioluminescence became very dim, much lessthan WT (Fig. 1, ESI Fig. S4–S6†).

Interactions of E457 in the structure of luciferase andrationale for site-directed mutagenesis

We have analyzed the interactions of E457 in the availablestructures of firefly luciferase9,10 which shows that the sidegroup of E457 forms a hydrogen bond with a backbone oxygenof V471 (Fig. 2). Multiple sequence alignments of beetle luci-ferases and other ANL proteins (Fig. 3) show that E457 and itshydrogen bond are highly conserved in the whole ANL super-family.11,37,38 Moreover, the crystal structures of ANL proteins(Fig. 2C) show that in most ANL enzymes E457 forms a secondhydrogen bond (salt bridge) with Arg in position 534 but thisinteraction is absent in pH-sensitive firefly luciferases.

We next designed and studied several additional mutantsin order to investigate the role of these interactions of E457 inbioluminescence and to understand what determines theeffects of the substitution E457K. The following mutants wereconstructed based on both TS and Y35N/TS luciferases andthen analyzed in vivo (Fig. 1, ESI Fig. S7†): E457V, E457Q,E457D, A534R, and E457V/A534R. The mutant E457V was alsoconstructed in WT luciferase.

In vivo bioluminescence of the mutants in E. coli cells

In contrast to the very dim bioluminescence of the mutantE457K/WT, the mutant E457V/WT produced bright colonies.Both mutants E457K/TS and E457V/TS gave bright coloniesshowing that the thermostable mutant TS compensates for thedetrimental effect of the mutation E457K. In vivo comparisonof mutants based on two parent forms – TS and Y35N/TS –

efficiently shows the strength of the red-shifting effect of anymutant. It revealed that although the mutants E457V, E457Q,A534R and E457V/A534R were more reddish than their parentTS, only the effect of E457K was strong: the other mutationswere nearly completely suppressed by the mutation Y35N.

In vivo thermostability of the mutants in E. coli cells

We have assessed in vivo thermostability of the mutants inE. coli colonies at 53 °C (Fig. 1, ESI Fig. S7†). The relative



Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online


in vivo stabilities (the parent enzyme is the thermostablemutant TS) followed the sequence below:

WT < E457D≪ /E457K < [ ∼ E457K] < [ /E457Q∼ /E457V/A534R] < [E457Q ∼ E457V/A534R] < [TS ∼ /E457V ≤ E457V] < /A534R < A534R.

The mutant Y35N noticeably lowered the thermostability ofits parent TS, and the mutant E457K considerably decreasedin vivo the thermostability of both TS and Y35N/TS. In contrastto E457K, the mutant E457V showed stability similar to itsparent TS and was even able to stabilize the mutant Y35: thestability of the double mutant Y35N/E457V was similar to TS.The substitution A534R was the most thermostable in this setof mutants and was even able to make the mutant Y35N/A534R more thermostable in vivo than TS. Noteworthy, itdecreased the stability of the mutant E457V where the for-mation of a salt bridge between R534 and E457 is impossible.

Expression and purification of mutant and WT luciferases

WT Lml and the mutants Y35N, A217T/S222T, Y35N/A217T/S222T, and Y35N/H433Y were produced as N-terminal His6-tagged proteins using the pETL4 plasmid. WT Lml and themutants TS, H433Y, E457K, E457V, E457K/TS, and Y35N/E457K were produced as C-terminal His6-tagged proteins usingthe pETL7 plasmid. Yields of the purified proteins (mg per0.2 L of culture or per 2.1–2.6 g of wet cells) were the following:WT-N-His6 – 34, WT-C-His6 – 32, Y35N – 26, A217T/S222T – 20,Y35N/A217T/S222T – 13, Y35N/H433Y – 27, TS – 61, H433Y –

64, E457K – 49, E457K/TS – 44, Y35N/E457K – 10, and E457V –

60. The purity of the enzymes was more than 95% as estimatedby SDS/PAGE.

Characterization of purified luciferase mutants

Both N- and C-terminally His-tagged (Tables 1 and 2) WT luci-ferases and the untagged luciferase (not shown) showedalmost identical properties, which means that all the purifiedmutants can be directly compared. However, the more recentC-terminal version of luciferase showed better binding withNi-IDA columns and better storage stability at 4 °C, so it wasused in the production of all subsequent mutants.

The mutation E457K decreased activity to ∼40% of the orig-inal value when introduced to WT or the mutant TS. Themutant H433Y elevated the Km value for ATP ∼1.6-fold whenintroduced to WT or added to the mutant Y35N. The substi-tution E457K caused a considerable increase in Km value forATP: 3.5-fold for WT and Y35N. When introduced into thethermostable mutant TS, which has 5.8-fold lower Km thanWT, it increased the Km value for ATP only 2-fold from 29 to62 μM, which is still about 3 times lower than the WT value.The substitution E457K also increased the Km value for LH2 byabout 30%. In contrast, the mutant E457V did not cause anyappreciable increase in Km or decrease in activity.

Thermostability of mutants in vitro

All the mutants showed an exponential decay of activity whenincubated at 42 °C (Table 1). The pH-insensitive mutationY35N destabilized the WT luciferase 2.5-fold and caused asimilar decrease in stability when combined with other WT-based mutants. When applied to TS, the mutation Y35Ndecreased the half-life 10-fold to 1 hour (unpublished resultson the mutant Y35N/TS/S398M). Notably, the red-shiftingmutant E457K did not decrease the stability of the WT enzyme

Fig. 1 In vivo bioluminescence and thermostability of mutant luciferases in E.coli colonies. Luciferases were expressed using the pLR4-based plasmids in E. coliXL1-blue. Bioluminescence was photographed after treating the cells with a luci-ferin solution (A). Then the cells were incubated at 53 °C and residual bio-luminescence was photographed after 30 min (B) and after 70 min (C). Theparent form for the mutants is TS except for the two mutants in WT that arehighlighted by the rectangle. A more detailed version of this figure is availableas ESI Fig. S7.†



Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online


but decreased the half-life of the thermostable mutant TS by70%: from 10 to 3.2 hours. This suggests that the low thermo-stability of WT luciferase is determined by unfolding of theN-domain29 whereas in thermostable luciferase the unfolding

in the C-domain becomes important. The mutation E457V didnot cause appreciable changes in thermostability in WT(Table 1) or TS (Fig. 1), indicating that the hydrogen bondbetween E457 and V471 is not crucial for the thermal stability.

Fig. 2 Location and interactions of the residue E457. (A) The structure of firefly luciferase in the 2nd (light emitting) conformation10 showing the location ofmutants. Residues are numbered according to L. mingrelica luciferase. LO* and AMP* – luciferyl and AMP moieties, respectively. (B) Close view of interactions ofE457 in the same structure. (C) Interactions of E457 in Populus tomentosa 4-Coumarate:CoA ligase in the 2nd (thioester) conformation.36 Molecular graphics werecreated using YASARA and PovRay software.

Fig. 3 Sequence alignment of L. mingrelica firefly luciferase with other beetle luciferases, insect luciferase homologs and representative ANL enzymes. Positions445 (also 550 – when it is functionally similar39), 457, 531, 534 are highlighted. Four-letter gene codes indicate that a crystal structure was reported for theseproteins.



Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online

Misha

Textbox

3G7S_A

Misha

Line


Bioluminescence spectra

Bioluminescence spectra of the purified luciferases weremeasured over the pH range from 6.0 to 9.0 at 25 °C (Table 2,Fig. 4). The spectra of the mutants at pH 9.0 were similar tothat at the optimum pH 7.8 (not shown). WT luciferasedemonstrates yellow-green bioluminescence with a slightlybimodal spectrum under the standard reaction conditions (pH7.8, 25 °C). The slightly noticeable shoulder in the red regiondisappears upon lowering the temperature. This shoulderconfers a yellowish tint on in vitro emission of WT luciferasecompared with monomodal emission of the mutant Y35N (ESIFig. S1†). Like in the case of most other firefly luciferases,12

elevated temperatures and, in particular, lowering of pHincreased the ratio of the red emitter ultimately shifting thebioluminescence maximum to the red region. As reportedearlier,18 the mutant Y35N shows an exceptional resistance ofits spectrum to the low values of pH and elevated

temperatures: the spectrum did not change down to pH 6.1and became slightly widened at pH 6.0. The green-emittingmutant Y35N and the red-shifting mutant H433Y compensatefor each other: the mutant Y35N/H433Y shows pH-sensitivebioluminescence spectra very close to that of WT. In thecase of the less red-shifting mutant A217T/S222T, the substi-tution Y35N neutralized this red shift almost completely andthe triple mutant Y35N/A217T/S222T demonstrated spectrasimilar to that of the single mutant Y35N (Table 2, ESIFig. S8†).

The mutation E457K significantly increased the contri-bution of the red emitter to the bimodal spectrum leading to apredominantly red bioluminescence. The maximum bio-luminescence shifted from 566 to 604 nm compared with WTunder the standard conditions, which is similar to the effect ofH433Y. However, comparison of the spectra for the mutantsH433Y and E457K at 10 °C indicated that there was also a redshift of the green emitter itself in addition to the increasedcontribution of the red emitter. To confirm this, we con-structed the double mutant Y35N/E457K. The substitutionY35N stabilized the green emitter leading to a predominantlymonomodal green emission with a narrow spectrum for themutant Y35N/E457K under the standard conditions (Fig. 4).Comparison of its spectrum with that of WT at 10 or 25 °C(both have identical shape) shows that the mutation E457Kcaused a 12 nm red shift of the green emitter (from 566to 578 nm) (Table 2). In marked contrast, the mutantE457V only mildly increased the contribution of the redemitter causing broadening of the spectra (Table 2, ESIFig. S9†). This broadening was suppressed at lower temp-eratures and there was no change in the λmax of the greenemitter. The red-shifted spectra of the mutant E457K/TSshowed a less pronounced “green” shoulder compared withthe single mutation E457K in WT (Table 2, ESI Fig. S10†)because its parent mutant TS has an increased contribution ofthe red emission to begin with.

Table 1 Biochemical characteristics of purified luciferase enzymes

Enzyme (mutations relative to WT) His6-tag

Relative specific activitya/% Km (μM)

Half-life (42 °C)/minFlash-height Integrated (90 s) LH2 ATP

WT-His6-C C-terminal 100 100 74 ± 8 177 ± 20 8.4 ± 0.8WT-His6-N N-terminal 100 103 73 ± 7 164 ± 20 8.4 ± 0.8

N-terminal 70 90 66 ± 6 194 ± 22 3.3 ± 0.3/S222T N-terminal 47 47 100 ± 10 204 ± 20 4.5 ± 0.2/ /S222T N-terminal 13 24 69 ± 7 270 ± 20 2.1 ± 0.1

C-terminal 100 96 83 ± 8 280 ± 30 7.2 ± 0.6/ C-terminal 100 124 81 ± 8 305 ± 30 2.3 ± 0.2/ C-terminal 18 30 82 ± 9 617 ± 60 2.7 ± 0.3

C-terminal 42 77 98 ± 10 604 ± 60 9.8 ± 0.9E457V C-terminal 87 93 38 ± 5 206 ± 34 7.4 ± 0.7Mutant TS C-terminal 190 119 70 ± 7 29 ± 3 594 ± 22

+ TS C-terminal 80 71 100 ± 10 62 ± 6 191 ± 11

a Specific activity measurements were performed at room temperature as described in the Experimental procedures. The values are expressedrelative to that of WT defined as 100 which is equivalent to 7.8 × 1012 RLU s−1 mg−1 and 1.08 × 1014 RLU mg−1 for flash-height based and 90 sintegration based assays, respectively. The error associated with activity falls within 10% of the value.

Table 2 Effect of mutations on bioluminescence spectra of luciferase

Enzyme (mutationsrelative to WT)

λmax (half-width), nm

pH 7.8 pH 6.0 pH 7.8 pH 7.825 °C 25 °C 10 °C 42 °C

WT-His6-C 566 (78) 616 (79) 563 (68) 599 (92)WT-His6-N 566 (76) 610 (96) 562 (64) 591 (91)

564 (67) 565 (75) 562 (63) 567 (75)/S222T 570 (91) 617 (61) 564 (77) 605 (85)/ /S222T 564 (71) 567 (83) 562 (65) 571 (86)

607 (90) 619 (57) 571 (94) 611 (70)/ 567 (77) 613 (96) 565 (68) 577 (92)/ 578 (78) 590 (85) 576 (72) 587 (87)

604 (84) 617 (67) 586 (86) 611 (75)E457V 574 (88) 616 (63) 565 (73) 609 (77)Mutant TS 573 (92) 609 (91) 564 (73) 605 (83)

+ TS 609 (70) 614 (64) 606 (79) 611 (65)



Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online


Discussion

The shape of the bioluminescence spectrum of firefly lucifer-ase is defined by the ratio of the green and red emitters eachhaving its individual λmax.

8,14 Gradual pH-dependent red shiftsof initially green bioluminescence of firefly luciferases are themost widely known example. Results of many mutagenesisstudies show that the equilibrium between these two spectralcomponents is determined by many independent factors ofthe luciferase structure. These include a wide network ofhydrogen bonds in the luciferin-binding channel involvingmany residues13 and water molecules,17 hydrophobic packingin this channel,18,23 flexibility of the oxyluciferin microenviron-ment9 and proper binding of AMP.40 This delicate complexof interactions could be easily perturbed by manymutations, often far from the substrate-binding pocket,8,13

which still prevents an unambiguous identification of thekey residues responsible for the pH-dependent changes ofspectra and understanding the core interactions that definethe structure and type of oxyluciferin emitter. Recentstudies suggest that this structure may be initially deter-mined by a polarization state of the 6′-phenol group ofoxyluciferin.14,17,21

We have shown that green and bimodal red mutants can bemutually compensated and this can be used as an efficient

tool to identify new mutations which strongly stabilize greenor red emission. New mutants could be compensated again byadditional mutations. Thus, it seems possible to sequentiallyscreen the structure of luciferase until most of the strong“green” or “red” mutants are identified.

Using this strategy, we have identified the new mutantE457K in the C-domain with a strong red-shifting effect. Thismutation disrupted a hydrogen bond of E457 with a backboneoxygen of V471 (Fig. 2). Both the residue V471 and this hydro-gen bond are absolutely conservative not only in beetle lucifer-ases but in almost all enzymes from the highly diverse ANLsuperfamily11,37,38 indicating the importance of this inter-action (Fig. 3). Conservative polar residues forming hydrogenbonds with backbone atoms are known to be important tomaintain the protein architecture in many protein families.41

The hydrogen bond between E457 and V471 contributes to theconnection between the first α-helix and the following β-sheetin the C-domain in luciferase and in other ANL enzymes.E457 is even more conserved in the ANL superfamily thanK445 which is crucial for the second half-reaction in manyfamilies of ANL enzymes (Fig. 3).11 Therefore, our firsthypothesis was that the removal of this conservative hydrogenbond by the mutation E457K is responsible for the color shiftand the detrimental effects on thermostability of TS and invivo activity in WT. However, the mutant E457V, which

Fig. 4 Bioluminescence spectra of WT (1), luciferase and mutants Y35N (2), H433Y (3), Y35N/H433Y (4), E457K (5), Y35N/E457K (6) at pH 7.8 (25 °C), pH 6.0(25 °C), at 10 °C (pH 7.8) and 42 °C (pH 7.8).



Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online


removes this bond by replacing Glu with a similar-sized Val,did not cause appreciable changes in catalytic properties orstability of luciferase, and demonstrated only a slightly highercontribution of the red emitter in the bimodal bio-luminescence. This suggests that most of the effect of themutation E457K was caused by the introduction of a positivecharge.

The residue E457K is located outside of the substrate-binding pocket (Fig. 2A). At the light-emitting conformation10

it is located 17 Å away from the AMP moiety and 20 Å from theluciferyl moiety. Therefore, a direct electrostatic influence ofthis charge on the oxyluciferin emitter seems unlikely. Also,the introduction of positive charge in the vicinity of E457 bythe mutation of the neighboring residue I532 to Arg or Lys didnot cause a red shift of λmax.

42 Thus, it seems likely that K457affects the light emitter indirectly through a network of inter-actions with other residues. The considerable rise of Km forATP suggests that the mutant E457K can change the micro-environment of oxyluciferin indirectly by influencing thebinding of AMP which in turn affects the light emission. TheAMP moiety is located between E457 and the oxyluciferinmoiety in both the light-emitting (Fig. 2B) and the adenylationconformations, so it is expectable that the AMP binding site isthe first to be disturbed by mutations of E457. The proper andtight binding of AMP at the final reaction step contributes tothe green emission.40 For example, removal of the hydrogenbond between D424 and 2′-hydroxyl of the ribose moiety of theAMP product (by replacing ATP with 2′-deoxyATP which lacks2′-hydroxyl40 or by substitutions of D42414) caused redbimodal emission. In contrast, the mutant E270K with low Km

for ATP suppressed the red emitting spectral component.40

The importance of AMP in the modulation of luciferase colorwas also suggested by quantum mechanics simulations.43

Thus, the mutants weakening the binding of AMP couldperturb its appropriate placement promoting the red emittingconformation in firefly luciferases. The weaker binding of AMPis usually accompanied by higher values of Km for ATP. In linewith this view, other red-shifting mutants described in thiswork also tended to increase Km for ATP (Table 1). Themutants G216N/A217L and S398M showed similar correlationbetween Km and red shifts.19 The high Km for ATP also explainsthe very dim in vivo bioluminescence of the mutant E457Kdespite its high yield and specific activity. It is known that Km

for ATP of WT luciferase is elevated in vivo44 and comparableto in vivo ATP concentration45 which leads to diminishedin vivo bioluminescence of the mutants with increased Km.The high Km and relatively high specific activity of the mutantE457K also mean that it can be efficiently used to monitor thein vivo ATP concentration.45

In contrast to numerous mutations affecting the ratio of thetwo emitters in firefly luciferases, only a few mutations areknown that affect λmax of the individual emitters. Themutations G246A/F250T and V241I/F250T caused a 9 nm blueshift,16 whereas L286M46 and R218K46,47 caused a 10–15 nmred shift of the green emitter in P. pyralis luciferase. In all thecases the emission peak remained narrow and nearly

monomodal indicating that the shift corresponds to the indi-vidual green emitter. Branchini et al. reported that mutationsR337Q and R337K caused a 38 nm red shift in P. pyralis lucifer-ase while the spectra remained nearly monomodal.47 However,the effect of this position is unclear since Hosseinkhani et al.did not observe a red shift for the mutant R337Q.48 We haveshown that residues in position 457 can also modulate theλmax of the green emitter despite the remote location of E457from the active site.

Another interesting observation was that in most ANLenzymes36,39 the residue E457 forms a salt bridge with R534(occasionally, K53449) which contributes to the connectionbetween the first and last α-helices of the C-domain (Fig. 2C).This Arg is highly conservative. It is present in pH-insensitiveluciferases but is replaced by small side groups in all pH-sensi-tive firefly luciferases (Fig. 3). The notable exception amongfirefly luciferases is P. pennsylvanica Ppe2 luciferase (Fig. 3) butit belongs to a family of paralogous “juvenile” luciferaseswhich are pH-insensitive unlike their “adult” paralogs.50 Thissalt bridge is best represented by the structure of Populustomentosa 4-Coumarate:CoA ligase which is the one closest toluciferases among the ANL enzymes with solved crystal struc-ture. It shares 34% sequence identity with Lml and conse-quently shows very similar overall structure and keyinteractions36 (Fig. 2C). Contrary to our expectations, themutant A534R actually slightly increased the contribution ofthe red emitter (Fig. 1), so the difference in this position is notrelated to spectral differences between pH-sensitive and pH-insensitive luciferases. The mutant A534R substantiallyincreased the in vivo stability of TS but decreased the stabilityof the mutant E457V/TS. This suggests that the designed saltbridge forms between R534 and E457 and contributes to luci-ferase stability. Accordingly, in the mutant E457V/A534R/TSthe positively charged side group of R534 no longer has anegative partner (E457) and destabilizes luciferase. Theimportance of this salt bridge for protein stability can be oneof the reasons why E457 is highly conservative in the ANLsuperfamily. Any mutations of E457 would disrupt this saltbridge which may decrease the structural stability and impairthe folding of ANL enzymes.

Conclusions

We have shown that pH-insensitive mutants like Y35N can beused as an efficient tool to selectively identify mutants with astrong effect on bioluminescence color. Mutagenesis analysisof the interaction of the residue E457 revealed that while itshydrogen bond plays a relatively small role in the determi-nation of bioluminescence color, the introduction of a posi-tively charged residue in this position causes a dramatic redshift and affects the interaction of luciferases with ATP. Themutant E457K shifts the λmax of the green emitter as well aspromotes the contribution of the red emitter which can beused for the construction of multi-color bioluminescent



Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online

Misha

Highlight

Misha

Textbox

actually, Val is noticeably smaller; (note by Misha) E457L would be more appropriate (but still smaller)

Misha

Textbox

also, the mutation T249M in Ppe2 luc (=F250 in P. pyralis luc, see Fig.3) shifts emission by 14 nm (538-->552nm) (Wood et al, 2007: US Patent 7241584, p. 13)


reporters and for in vivo imaging. The high stability of themutant Y35N/A534R/TS and its relatively pH-independentcolor make it a promising luciferase for gene reporterapplications and should improve the sensitivity and stability ofluciferase assays in vivo and in vitro. The described randommutagenesis strategy can be used further to identify newpositions important for bioluminescence color and help indeciphering the structure–color relationship in beetleluciferases.

Abbreviations

WT Wild-type enzymeLH2 Firefly D-luciferinLml Luciola mingrelica firefly luciferaseλmax Maximum of the bioluminescence spectrumRLU Relative light units

Acknowledgements

This work was supported by the Russian Foundation for BasicResearch (grants 08-04-00624 and 11-04-00698).

References

1 K. Niwa, Y. Ichino, S. Kumata, Y. Nakajima, Y. Hiraishi,D. i. Kato, V. R. Viviani and Y. Ohmiya, Quantum Yieldsand Kinetics of the Firefly Bioluminescence Reactionof Beetle Luciferases, Photochem. Photobiol., 2010, 86,1046–1049.

2 A. Lundin, Use of firefly luciferase in ATP-related assays ofbiomass, enzymes, and metabolites, Methods Enzymol.,2000, 305, 346–370.

3 S. Wu, E. Chang and Z. Cheng, Molecular Probes forBioluminescence Imaging, Curr. Org. Synth., 2011, 8, 488–497.

4 C. E. Badr and B. A. Tannous, Bioluminescence imaging:progress and applications, Trends Biotechnol., 2011, 29,624–633.

5 N. Hida, M. Awais, M. Takeuchi, N. Ueno, M. Tashiro,C. Takagi, T. Singh, M. Hayashi, Y. Ohmiya and T. Ozawa,High-Sensitivity Real-Time Imaging of Dual Protein-ProteinInteractions in Living Subjects Using Multicolor Luci-ferases, PLoS One, 2009, 4, e5868.

6 C. I. Stains, J. L. Furman, J. R. Porter, S. Rajagopal, Y. Li,R. T. Wyatt and I. Ghosh, A General Approach for Receptorand Antibody-Targeted Detection of Native Proteins Utiliz-ing Split-Luciferase Reassembly, ACS Chem. Biol., 2010, 5,943–952.

7 H. Fraga, Firefly luminescence: A historical perspective andrecent developments, Photochem. Photobiol. Sci., 2008, 7,146–158.

8 S. Hosseinkhani, Molecular enigma of multicolor bio-luminescence of firefly luciferase, Cell. Mol. Life Sci., 2011,68, 1167–1182.

9 T. Nakatsu, S. Ichiyama, J. Hiratake, A. Saldanha,N. Kobashi, K. Sakata and H. Kato, Structural basis for thespectral difference in luciferase bioluminescence, Nature,2006, 440, 372–376.

10 J. A. Sundlov, D. M. Fontaine, T. L. Southworth,B. R. Branchini and A. M. Gulick, Crystal Structure ofFirefly Luciferase in a Second Catalytic Conformation Sup-ports a Domain Alternation Mechanism, Biochemistry,2012, 51, 6493–6495.

11 A. M. Gulick, Conformational Dynamics in the Acyl-CoASynthetases, Adenylation Domains of Non-ribosomalPeptide Synthetases, and Firefly Luciferase, ACS Chem.Biol., 2009, 4, 811–827.

12 V. R. Viviani, The origin, diversity, and structure functionrelationships of insect luciferases, Cell. Mol. Life Sci., 2002,59, 1833–1850.

13 V. R. Viviani, F. G. C. Arnoldi, A. J. S. Neto, T. L. Oehlmeyer,E. J. H. Bechara and Y. Ohmiya, The structural origin andbiological function of pH-sensitivity in firefly luciferases,Photochem. Photobiol. Sci., 2008, 7, 159–169.

14 B. R. Branchini, T. L. Southworth, M. H. Murtiashaw,R. A. Magyar, S. A. Gonzalez, M. C. Ruggiero andJ. G. Stroh, An alternative mechanism of bioluminescencecolor determination in firefly luciferase, Biochemistry, 2004,43, 7255–7262.

15 B. R. Branchini, D. M. Ablamsky, M. H. Murtiashaw,L. Uzasci, H. Fraga and T. L. Southworth, Thermostable redand green light-producing firefly luciferase mutants forbioluminescent reporter applications, Anal. Biochem., 2007,361, 253–262.

16 B. R. Branchini, D. M. Ablamsky, J. M. Rosenman,L. Uzasci, T. L. Southworth and M. Zimmer, SynergisticMutations Produce Blue-Shifted Bioluminescence in FireflyLuciferase, Biochemistry, 2007, 46, 13847–13855.

17 D.-i. Kato, T. Kubo, M. Maenaka, K. Niwa, Y. Ohmiya,M. Takeo and S. Negoro, Confirmation of color determi-nation factors for Ser286 derivatives of firefly luciferasefrom Luciola cruciata (LUC-G), J. Mol. Catal. B: Enzym.,2013, 87, 18–23.

18 M. I. Koksharov and N. N. Ugarova, Random mutagenesisof Luciola mingrelica firefly luciferase. Mutant enzymeswhose bioluminescence spectra show low pH-sensitivity,Biochemistry, 2008, 73, 862–869.

19 M. I. Koksharov and N. N. Ugarova, Triple substitutionG216N/A217L/S398M leads to the active and thermostableLuciola mingrelica firefly luciferase, Photochem. Photobiol.Sci., 2011, 10, 931–938.

20 N. K. Tafreshi, M. Sadeghizadeh, R. Emamzadeh,B. Ranjbar, H. Naderi-manesh and S. Hosseinkhani, Site-directed mutagenesis of firefly luciferase: implication ofconserved residue(s) in bioluminescence emissionspectra among firefly luciferases, Biochem. J., 2008, 412,27–33.



Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online


21 V. R. Viviani, D. T. Amaral, D. R. Neves, A. Simõesand F. G. C. Arnoldi, The Luciferin Binding Site Resi-dues C/T311 (S314) Influence the BioluminescenceColor of Beetle Luciferases through Main-Chain Inter-action with Oxyluciferin Phenolate, Biochemistry, 2012,52, 19–27.

22 M. Nazari, S. Hosseinkhani and L. Hassani, Step-wiseaddition of disulfide bridge in firefly luciferase controlscolor shift through a flexible loop: a thermodynamic per-spective, Photochem. Photobiol. Sci., 2013, 12, 298–308.

23 N. Kajiyama and E. Nakano, Isolation and characterizationof mutants of firefly luciferase which produce differentcolors of light, Protein Eng., 1991, 4, 691–693.

24 S.-F. Chen, Y.-J. Liu, I. Navizet, N. Ferré, W.-H. Fang andR. Lindh, Systematic Theoretical Investigation on the LightEmitter of Firefly, J. Chem. Theory Comput., 2011, 7, 798–803.

25 J. Vieira, L. Pinto da Silva and J. C. G. Esteves da Silva,Advances in the knowledge of light emission by firefly luci-ferin and oxyluciferin, J. Photochem. Photobiol., B, 2012,117, 33–39.

26 B. S. Alipour, S. Hosseinkhani, S. K. Ardestani andA. Moradi, The effective role of positive charge saturationin bioluminescence color and thermostability of firefly luci-ferase, Photochem. Photobiol. Sci., 2009, 8, 847–855.

27 G. H. Law, O. A. Gandelman, L. C. Tisi, C. R. Lowe andJ. A. Murray, Mutagenesis of solvent-exposed amino acidsin Photinus pyralis luciferase improves thermostability andpH tolerance, Biochem. J., 2006, 397, 305–312.

28 B. R. Branchini, T. L. Southworth, M. H. Murtiashaw,H. Boije and S. E. Fleet, A mutagenesis study of the putativeluciferin binding site residues of firefly luciferase, Biochem-istry, 2003, 42, 10429–10436.

29 M. I. Koksharov and N. N. Ugarova, Thermostabilization offirefly luciferase by in vivo directed evolution, Protein Eng.,Des. Sel., 2011, 24, 835–844.

30 Z. Tu, G. He, K. X. Li, M. J. Chen, J. Chang, L. Chen,Q. Yao, D. P. Liu, H. Ye, J. Shi and X. Wu, An improvedsystem for competent cell preparation and high efficiencyplasmid transformation using different Escherichia colistrains, Electron J. Biotechnol., 2005, 8, 113–120.

31 N. Ugarova, L. Maloshenok, I. Uporov and M. Koksharov,Bioluminescence Spectra of Native and Mutant Firefly Luci-ferases as a Function of pH, Biochemistry, 2005, 70, 1262–1267.

32 P. C. Cirino, K. M. Mayer and D. Umeno, Generatingmutant libraries using error-prone PCR, Methods Mol. Biol.,2003, 231, 3–9.

33 K. Miyazaki and F. H. Arnold, Exploring Nonnatural Evol-utionary Pathways by Saturation Mutagenesis: RapidImprovement of Protein Function, J. Mol. Evol., 1999, 49,716–720.

34 T. Zako, K. Ayabe, T. Aburatani, N. Kamiya, A. Kitayama,H. Ueda and T. Nagamune, Luminescent and substratebinding activities of firefly luciferase N-terminal domain,Biochim. Biophys. Acta, 2003, 1649, 183–189.

35 J. D. Bloom, S. T. Labthavikul, C. R. Otey and F. H. Arnold,Protein stability promotes evolvability, Proc. Natl. Acad.Sci. U. S. A., 2006, 103, 5869–5874.

36 Y. Hu, Y. Gai, L. Yin, X. Wang, C. Feng, L. Feng, D. Li,X.-N. Jiang and D.-C. Wang, Crystal Structures of aPopulus tomentosa 4-Coumarate:CoA Ligase Shed Light onIts Enzymatic Mechanisms, Plant Cell, 2010, 22, 3093–3104.

37 P. A. Watkins, D. Maiguel, Z. Jia and J. Pevsner, Evidencefor 26 distinct acyl-coenzyme A synthetase genes in thehuman genome, J. Lipid Res., 2007, 48, 2736–2750.

38 J. Shockey and J. Browse, Genome-level and biochemicaldiversity of the acyl-activating enzyme superfamily inplants, Plant J., 2011, 66, 143–160.

39 A. M. Gulick, V. J. Starai, A. R. Horswill, K. M. Homickand J. C. Escalante-Semerena, The 1.75 Å Crystal Struc-ture of Acetyl-CoA Synthetase Bound to Adenosine-5′-pro-pylphosphate and Coenzyme A, Biochemistry, 2003, 42,2866–2873.

40 L. C. Tisi, G. H. Law, O. Gandelman, C. R. Lowe andJ. A. H. Murray, The basis of the bathochromic shift in theluciferase from Photinus pyralis, in Bioluminescence andChemiluminescence: Progress and Current Applications, ed.P. E. Stanley and L. J. Kricka, World Scientific, Singapore,2002, pp. 57–60. DOI: 10.1021/bi400141u.

41 C. L. Worth, S. Gong and T. L. Blundell, Structural andfunctional constraints in the evolution of protein families,Nat. Rev. Mol. Cell Biol., 2009, 10, 709–720.

42 H. Fujii, K. Noda, Y. Asami, A. Kuroda, M. Sakata andA. Tokida, Increase in bioluminescence intensity of fireflyluciferase using genetic modification, Anal. Biochem., 2007,366, 131–136.

43 L. Pinto da Silva and J. C. G. Esteves da Silva, TD-DFT/Molecular Mechanics Study of the Photinus pyralis Bio-luminescence System, J. Phys. Chem. B, 2012, 116, 2008–2013.

44 G. B. Sala-Newby, K. M. Taylor, M. N. Badminton,C. M. Rembold and A. K. Campbell, Imaging biolumine-scent indicators shows Ca2+ and ATP permeabilitythresholds in live cells attacked by complement, Immu-nology, 1998, 93, 601–609.

45 D. A. Schneider and R. L. Gourse, Relationship betweenGrowth Rate and ATP Concentration in Escherichia coli,J. Biol. Chem., 2004, 279, 8262–8268.

46 K. R. Harwood, D. M. Mofford, G. R. Reddy andS. C. Miller, Identification of Mutant Firefly Luciferasesthat Efficiently Utilize Aminoluciferins, Chem. Biol., 2011,18, 1649–1657.

47 B. R. Branchini, R. A. Magyar, M. H. Murtiashaw andN. C. Portier, The role of active site residue arginine 218 infirefly luciferase bioluminescence, Biochemistry, 2001, 40,2410–2418.

48 A. Riahi-Madvar and S. Hosseinkhani, Design and charac-terization of novel trypsin-resistant firefly luciferases bysite-directed mutagenesis, Protein Eng., Des. Sel., 2009, 22,655–663.



Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online

Misha

Replace

Misha

Replace

10.1142/9789812776624_0012

Misha

Textbox

10.1142/9789812776624_0012


49 Y. Hisanaga, H. Ago, N. Nakagawa, K. Hamada, K. Ida,M. Yamamoto, T. Hori, Y. Arii, M. Sugahara, S. Kuramitsu,S. Yokoyama and M. Miyano, Structural Basis of the Substrate-specific Two-step Catalysis of Long Chain Fatty Acyl-CoASynthetase Dimer, J. Biol. Chem., 2004, 279, 31717–31726.

50 Y. Oba, M. Furuhashi, M. Bessho, S. Sagawa, H. Ikeya andS. Inouye, Bioluminescence of a firefly pupa: involvementof a luciferase isotype in the dim glow of pupae and eggs inthe Japanese firefly, Luciola lateralis, Photochem. Photobiol.Sci., 2013, 12, 854–863.



Publ

ishe

d on

30

Aug

ust 2

013.

Dow

nloa

ded

by U

NIV

ER

SIT

E D

E G

EN

EV

E o

n 28

/10/

2013

14:

54:1

3.

View Article Online


SUPPLEMENTARY MATERIAL

Strategy of mutual compensation of green and red mutants of firefly luciferase identifies a

mutation of the highly conservative residue E457 with a strong red shift of bioluminescence

Mikhail I. Koksharov*, Natalia N. Ugarova Department of Chemical Enzymology, Faculty of Chemistry,

Lomonosov Moscow State University, Moscow, 119991, Russia. Fax/Tel: 7-495-939-26-60; E-mail: [email protected]

Table S1. PCR primer sequences

Name Primer sequence f_XhoI 5'- GTATTCAGCTCGAGAAAAGGCTTACC -3' r_T7term 5'- GCTAGTTATTGCTCAGCGG -3' Forward E457V 5'- GCTGAATTGGTATCCGTTCTTTTGC -3' Reverse E457V 5'- AGAACGGATACCAATTCAGCAGG -3' Forward A534R 5'- GGTAAAATTGATCGTAAAGTAATTAGAGAAATTCTTAAG -3' Reverse A534R 5'- CTAATTACTTTACGATCAATTTTACCAGTTAGACC -3' Forward E457X 5'- TCCGTTCTTTTGCAACATC -3' (universal primer for mutants of E457) Reverse E457Q 5'- GTTGCAAAAGAACGGATTGCAATTC -3' Reverse E457D 5'- GTTGCAAAAGAACGGAATCCAATTC -3'

1

WT TS Y35N

At room temperature (23°C)

In vivo in E. coli colonies on LB plate treated with luciferin solution

In vitro in the buffer: 50 mM Tris (pH 7.8)

8 mm No

enzyme

5 mm

Figure S1. Comparison of in vitro bioluminescence in the reaction buffer (7.8) and in vivo bioluminescence in E. coli cells. For in vivo bioluminescence see also Fig. S4.

Electronic Supplementary Material (ESI) for Photochemical & Photobiological ScienceThis journal is © The Royal Society of Chemistry and Owner Societies 2013

mailto:[email protected]

Y35N/H433Y H433Y Y35N/H433Y Y35N/H433Y H433Y H433Y

a) b) c)

Y35N WT WT Y35N WT Y35N

Figure S2. In vivo bioluminescence of E. coli colonies producing WT luciferase and the mutants Y35N, H433Y and Y35N/H433Y. Mutants were expressed using the pLR4 plasmid in E. coli XL1-blue. The plate was photographed: a) under the ambient light; b) in the dark with a higher effective exposure time; c) in the dark with a lower effective exposure time.

a) b)

Figure S3. Typical screening of the 90 mm plate with mutant E. coli colonies for changed bioluminescence color. In vivo bioluminescence (a) and the original plate in ambient light (b). The parent luciferase for this screening was the mutant Y35N+TS, which produces green bioluminescence. The red-shifted mutant E457K is marked by the arrow.

2


Exposure=10s, ISO=200 Exposure=10s, ISO=400 b) a)

E457K WT

TS E457K+TS

Y35N+TS E457K+Y35N+TS

Figure S4. In vivo bioluminescence of E. coli streaks containing the WT luciferase and the mutants 4TS, E457K, E457K/TS, Y35N/TS and Y35N/E457K/TS. Luciferases were expressed using the pLR4 plasmid in E. coli XL1-blue. The 90 mm plate was photographed in the dark with a: a) higher effective exposure time; b) lower effective exposure time.

c) b) Exposure=10s, ISO=800 Exposure=5s, ISO=400 a) E457K

E457K+TS

Figure S5. In vivo bioluminescence of E. coli colonies containing the mutant E457K and E457K/Y35N/TS. The expression vector pETL7 (derived from pET23b) carrying the respective mutant was transformed into E. coli strain BL21(DE3). This vector contains the “plain” T7 promoter which results in a considerable level of uninduced expression of luciferase. The 90 mm plate was photographed: a) under the ambient light; b) in the dark with a higher effective exposure time; c) in the dark with a lower effective exposure time.

3


Misha

Strikeout

a)

Exposure=10s, ISO=400 b) E457K TS Y35N+TS WT

E457K E457K+TS WTE457K+TS TS

TripleTripleWT WT

TS E457K+TSE457K

Y35N+TS WT

TS E457K E457K+TS E457K

Exposure=5s, ISO=400 c)

Fig. S6. In vivo bioluminescence of E. coli colonies containing the WT luciferase and the mutants TS, E457K, E457K/TS, Y35N/TS and Y35N/E457K/TS (Triple). Luciferases were expressed using the pLR4 plasmid in E. coli XL1-blue. The 90 mm plate was photographed: a) under the ambient light; b) in the dark with a higher effective exposure time; c) in the dark with a lower exposure time.

4


A) Exposure=10 sec, ISO=400: Exposure=5 sec, ISO=400:

B) Exposure=10 sec, ISO=800: Exposure=10 sec, ISO=400:

After 30 min at 53°C After 30 min at 53°C

C) Exposure=10 sec, ISO=800: Exposure=10 sec, ISO=400:

After 70 min at 53°C After 70 min at 53°C

Fig. S7. In vivo bioluminescence and thermostability of mutant luciferases in E. coli XL1-blue colonies. Luciferases were expressed using pLR4-based plasmids. Bioluminescence was photographed after treating the cells with luciferin solution (A). Then the cells were incubated at 53°C and residual bioluminescence was photographed after 30 min (B) and after additional 40 min (C). The parent form for the mutants is TS except for two mutants in WT that are highlighted by the rectangle. Each time point is represented with photograph at higher and lower effective exposure time for better discrimination of relative differences in intensity and color of bioluminescence of mutant colonies.

5


500 550 600 650 700 7500.0

0.5

1.0

24

3

Rel

ativ

e in

tens

ity

Wavelength, nm

pH 7.8 (25°C)

1

500 550 600 650 700 750

0.0

0.5

1.0

24

3

Wavelength, nm

1: WT 2: Y35N 3: A217T/S222T 4: Y35N/A217T/S222T

pH 6.0 (25°C)

1

500 550 600 650 700 7500.0

0.5

1.0

24

3

Rel

ativ

e in

tens

ity

Wavelength, nm

10°C(pH 7.8)

1

500 550 600 650 700 750

0.0

0.5

1.0

2

4

3

Wavelength, nm

42°C(pH 7.8)

1

Fig. S8. Combined effect of mutations Y35N and A217T/S222T. Bioluminescence spectra of WT (1) luciferase and mutants Y35N (2), A217T/S222T (3), Y35N/A217T/S222T (4) at pH 7.8 (25°C), pH 6.0 (25°C), 10°C (pH 7.8) and at 42°C (pH 7.8).

6


500 550 600 650 7000.0

0.5

1.0

5

42

3

Rel

ativ

e in

tens

ity

Wavelength, nm

1: WT, pH 7.8 (25°C) 2: E457V, pH 7.8 (25°C) 3: E457V, 10°C 4: E457V, 42°C 5: E457V, pH 6.0

1

Fig. S9. Effect of the mutant E457V on bioluminescence spectra. Bioluminescence spectra of WT luciferase at pH 7.8 (1) and the mutant E457V at pH 7.8 (2), 10°C (3), 42°C (4) and pH 6.0 (5). If not specified, the temperature is 25°C and the pH value is 7.8.

500 550 600 650 700 750.0

0.5

1.0

2

4

3

Rel

ativ

e in

tens

ity

Wavelength, nm

1: WT, pH 7.8 2: WT, pH 6.0 3: TS, pH 7.8 4: TS, pH 6.0

A25°C

1

500 550 600 650 700 7500.0

0.5

1.0

243

Wavelength, nm

1: E457K/WT, pH 7.8 2: E457K/WT, pH 6.0 3: E457K/TS, pH 7.8 4: E457K/TS, pH 6.0

B25°C

1

500 550 600 650 700 750.0

0.5

1.0

2

4

3

Rel

ativ

e in

tens

ity

Wavelength, nm

1: WT, pH 25°C 2: WT, pH 42°C 3: TS, pH 25°C 4: TS, pH 42°C

CpH 7.8

1

500 550 600 650 700 7500.0

0.5

1.0

2 43

Wavelength, nm

1: E457K/WT, 25°C 2: E457K/WT, 42°C 3: E457K/TS, 25°C 4: E457K/TS, 42°C

DpH 7.8

1

Fig. S10. Effect of the mutant TS on bioluminescence spectra of WT luciferase (A, C) and of the mutant E457K (B, D) at pH 7.8 and pH 6.0 at 25°C (A, B) or at 25 and 42°C at pH 7.8 (C, D). 7


Documents

Photochemical & Photobiological Scienceskoksharov83.narod.ru/docs/lab/Koksharov_PPS-2013_final.pdffinal volume) contained ∼20 ng insert, ∼50 ng vector, 1× T4 DNA ligase buﬀer,