The use of the luxA gene of the bacterial luciferase operon as a reporter gene

Summary Bacterial luciferase can be assayed rapidly and with high sensitivity both in vivo and in vitro. Here we demonstrate that the N-terminal hydrophobic domain of the catalytic subunit of the luciferase enzyme is indispensable for enzyme activity, although N-terminal translational fusions with full luciferase activity can be obtained. Bacterial luciferase is therefore ideally suited as a reporter enzyme for gene fusion experiments. A list of vectors for the convenient use of the luciferase marker genes to monitor gene expression in vivo are presented.     

A new β-lactoglobulin-based vector targets luciferase cDNA expression to the mammary gland of transgenic mice

A -lactoglobulin (BLG)/luciferase gene vector (p907), composed of a luciferase intronless gene inserted between the second and sixth BLG exons was constructed. Stable transfections of CID-9 cells with this vector, as well as with a series of additional vectors, were performed to define regulatory regions within the BLG sequence, and the contribution of the SV40 polyadenylation (PA) site to luciferase expression. A relatively low level of luciferase activity was supported by vector p907. It was partially rescued by vector p906, in which the BLG 3 region, downstream of the luciferase cDNA, was replaced with the SV40 PA site. Flanking the SV40 region of vector p906, at its 3 end, with BLG sequences of exon 6/intron 6/exon 7 and the 3 region of the gene resulted in vector p904. This vector supported the highest luciferase activity, 10 times or 2.5 times higher than that measured in cells transfected with vectors p907 and p906, respectively. The induced activity supported by vector p904 is attributed to interaction between the SV40 PA site and elements of the distal part of the BLG 3 flanking sequences. The BLG 5 regulatory region of vector p904 encompasses a 3-kb promoter sequences. Deletion of 935 bp of its proximal end resulted in a 60% decrease in luciferase activity. Reduced activity was also seen with vector p915 lacking sequences of exon 1/intron 1/exon 2. This decrease could not be rescued with heterologous sequences of insulin intron 1, inserted upstream of the luciferase cDNA. Two sets of transgenic mice carrying vectors p907 and p904 were generated. Vector p907 supported only marginal luciferase activity in the mammary gland of all transgenic mice tested and luciferase RNA could not be detected by northern analysis. In contrast, 50% of the transgenic mice carrying vector p904 expressed luciferase RNA in the mammary gland and tissue-specific, hormonal-dependent activity was determined. However, the new p904 vector was not able to insulate the transgene from surrounding host DNA sequences, as reflected by its copy number-independent manner of expression. Nevertheless, vector p904 may represent a valuable tool for the expression of cDNAs in the mammary gland of transgenic animals.     

Role of Hsp70 (DnaK-DnaJ-GrpE) and Hsp100 (ClpA and ClpB) chaperones in refolding and increased thermal stability of bacterial luciferases in Escherichia coli cells

The role of chaperones Hsp70 (DnaK–DnaJ–GrpE) and Hsp100 (ClpA–ClpB–ClpX) in refolding of thermoinactivated luciferase from the marine bacterium Photobacterium fischeri and the terrestrial bacterium Photorhabdus luminescens has been studied. These luciferases are homologous, but differ greatly in the rate of thermal inactivation and the rate constant for the luminescence reaction. It was shown that refolding of thermoinactivated luciferases is completely determined by the DnaK–DnaJ–GrpE system. However these luciferases markedly differ in the rate and degree of refolding. The degree of refolding of thermolabile quick Ph. fischeri luciferase reaches 80% of the initial level over several minutes, whereas renaturation of thermostable slow Ph. luminescens luciferase proceeds substantially slower (the degree of renaturation reaches only 7-8% of the initial level over tens of minutes). The measurement of the rate of thermal inactivation of luciferases in vivo in the cells of Escherichia coli wild strain and strains containing mutations in genes clpA, clpB, clpX showed that Ph. luminescens luciferase revealed reduced thermostability in mutant strain E. coli clpA^–. It was shown that this effect was not connected with DnaK-dependent refolding. In the case of thermolabile Ph. fischeri luciferase, mutation in gene clpA has no effect on the shape of the curve of thermal inactivation. These data suggest that denatured Ph. luminescens luciferase has enhanced affinity with respect to chaperone ClpA in comparison with DnaK, whereas thermolabile Ph. fischeri luciferase is characterized by enhanced affinity with respect to chaperone DnaK. Denatured luciferase bound to ClpA does not aggregate and following refolding proceeds probably spontaneously and very quickly (over 1-2 min). It is evident that the process under discussion requires ATP, since the addition of uncoupler of oxidative phosphorylation carbonyl cyanide 3-chlorophenylhydra-zone results in a sharp decrease in thermal stability of luciferase to the level typical of the enzyme in vitro. The enhanced thermosensitivity of luciferases was observed also in E. coli containing mutations in gene clpB. However, this effect, which takes place for Ph. fischeri luciferase as well as for Ph. LuminescensM luciferase, is determined by DnaK-dependent refolding and probably connected with the ability of chaperone ClpB to provide disaggregation of the proteins, resulting in their interaction with chaperones of the Hsp70 family (DnaK–DnaJ–GrpE).     

Development of a destabilized firefly luciferase enzyme for measurement of gene expression

Firefly luciferase is used widely as a reporter enzyme for studies of gene regulation and expression. The recent development of new technologies that combine luciferase reporter technology and digital imaging microscopy has enabled multiple measurements of gene expression in the same living cell. Although this approach has already provided new insights about expression dynamics, its future utility is limited by the three- to four-hour half-life of firefly luciferase in mammalian cells. Because of this, rapid increases or decreases in gene expression may not be detected, owing to the accumulation of residual luciferase. Accordingly, the goal of the present study was to develop a luciferase reporter with a reduced functional half-life. This was accomplished by adding a synthetic fragment to the firefly luciferase-coding sequence that encoded the proteolytic "PEST" signal from mouse ornithine decarboxylase. When placed under the control of estrogen response elements and expressed in human breast cancer T-47D cells, the modified luciferase protein (LUCODC-DA) displayed a functional half-life of 0.84 h compared to 3.68 h for the wild-type enzyme. As anticipated, the overall rate of photonic emissions in cells expressing the destabilized luciferase was about sevenfold lower than that of their wild-type counterparts, presumably because of the reduction of steady-state luciferase accumulation. Even so, the photonic activity derived from LUCODC-DA was still sufficient to enable real-time measurements of gene expression in single living cells.     

Common structural features of the luxF protein and the subunits of bacterial luciferase: evidence for a (beta alpha)8 fold in luciferase.

The amino acid sequence identity and potential structural similarity between the subunits of bacterial luciferase and the recently determined structure of the luxF molecule are examined. The unique beta/alpha barrel fold found in luxF appears to be conserved in part in the luciferase subunits. From secondary structural predictions of both luciferase subunits, and from structural comparisons between the protein product of the luxF gene, NFP, and glycolate oxidase, we propose that it is feasible for both luciferase subunits to adopt a (beta alpha)8 barrel fold with at least 2 excursions from the (beta alpha)8 topology. Amino acids conserved between NFP and the luciferase subunits cluster together in 3 distinct "pockets" of NFP, which are located at hydrophobic interfaces between the beta-strands and alpha-helices. Several tight turns joining the C-termini of beta-strands and the N-termini of alpha-helices are found as key components of these conserved regions. Helix start and end points are easily demarcated in the luciferase subunit protein sequences; the N-cap residues are the most strongly conserved structural features. A partial model of the luciferase beta subunit from Photobacterium leiognathi has been built based on our crystallographically determined structure of luxF at 1.6 A resolution.     

Retroposons of Alu family from <Emphasis Type="Italic">cis</Emphasis>-regulatory modules of <Emphasis Type="Italic">DNAse II</Emphasis> and <Emphasis Type="Italic">CAML</Emphasis> genes affect gene expression in A549 and HEK 293 cells

Promoter fragments of deoxyribonuclease II (DNAse II) and calcium-modulating cyclophilin ligand (CAML) associated with Alu family repeats have been inserted into luciferase reporter vectors. The constructs were introduced into A549 and HEK293 cell lines by transient transfection. Transfected cells were lysed to analyze luciferase activities. It has been shown that Alu repeats inserted into constructs influence the luciferase expression. Therefore, Alu copies associated with cis-regulatory modules in protein-coding genes have biological activity.     

Recombinant Metridia luciferase isoforms: expression, refolding and applicability for in vitro assay

The recombinant coelenterazine-dependent luciferases (isoforms MLuc164 and MLuc39) from the marine copepod Metridia longa were expressed as inclusion bodies in E. coli cells, dissolved in 6 M guanidinium chloride and folded in conditions developed for proteins containing intramolecular disulfide bonds. One of them (MLuc39) was obtained in an active monomeric form with a high yield. The luciferase bioluminescence is found to be initiated not only by free coelenterazine, but also by Ca(2+)-dependent coelenterazine-binding protein (CBP) of Renilla muelleri on Ca(2+) addition. The use of CBP as a "substrate" provides higher light emission and simultaneously the lower level of background. The high purity MLuc39 can be detected down to attomol with a linear range extending over 5 orders of magnitude. The MLuc39 reveals also a high stability towards heating and chemical modification; the chemically synthesized biotinylated derivatives of the luciferase preserve 35-40 % of the initial activity. The luciferase applicability as an in vitro bioluminescent reporter is demonstrated in model tandem bioluminescent solid-phase microassay combining the Ca(2+)-regulated photoprotein obelin and the Metridia luciferase.     

Mutant Analysis and Enzyme Subunit Complementation in Bacterial Bioluminescence in Photobacterium fischeri

Chemical mutagens were used to obtain mutants deficient in bioluminescence in the marine bacterium Photobacterium fischeri strain MAV. Acridine dyes were effective in the production of dark mutants but not in the production of auxotrophs. These dark mutants were all of one type and appeared to contain lesions blocking the synthesis of luciferase. ICR-191 was especially effective in the production of aldehyde mutants, i.e., dark strains that luminesce when a long-chain aldehyde such as n-decanal is added to them. However, other mutant types were isolated after treatment with ICR-191. N-methyl-N'-nitro-N-nitrosoguanidine induced many bioluminescence-deficient types with respect to both the site of the lesion and the quantitative effect on the luminescent system. We characterized the dark and dim mutants with respect to their response to exogenous decanal, levels of in vivo and in vitro luminescence, and their rates of reversion to wild type. In addition, the luciferases of the mutant strains were examined by subunit complementation. On the basis of these analyses, we identified mutants which synthesize altered luciferase, strains which are deficient in synthesis of luciferase, and aldehyde mutants. The results of analysis of luciferase from the aldehyde mutants and the complementation studies indicate that the lesions in these strains are in the luciferase itself. Results obtained with wild-type cells grown in minimal medium, and aldehyde mutant cells grown either in complete or minimal medium, indicate that a "natural aldehyde factor" is involved in in vivo light emission. These same studies showed that the long-chain aldehyde(s) could only partially substitute for the natural "aldehyde factor." The possibility that the in vivo aldehyde factor is not a long-chain aldehyde is discussed.     

Growth, luminescence, respiration, and the ATP pool during autoinduction in Beneckea harveyi.

The bacterial bioluminescence system is unusual because it is self-induced. In the late logarithmic phase of growth, upon the accumulation of an autoinducer, the synthesis of the components of the system is initiated. We were interested in determining what effect this burst of synthesis and activity has on cellular energy metabolism. The ATP pool of the luminous bacterium Beneckea harveyi was found to dip 10- to 20-fold during the luminescence period, while the respiration per unit cell mass (optical density) increased but by much less. The dip in the ATP pool did not occur in four different types of dark mutants, including one that was temperature conditional and another that was conditional upon added cyclic AMP for luminescence. However, it is neither the synthesis nor the activity of luciferase that is responsible for the ATP dip; the dip does not occur in certain dark "aldehyde" mutants which nevertheless synthesize normal levels of luciferase, whereas it does occur at 36 degrees C in a temperature-sensitive luciferase mutant which forms normal levels of inactive luciferase. Results with other aldehyde mutants implicate the pathway involved in the synthesis of the aldehyde factor with the ATP dip.     

Conditional gene expression in the respiratory epithelium of the mouse

Transgenic mouse models mediating conditional temporal and spatial regulation of gene expression to the respiratory epithelium were developed utilizing the reverse tetracycline transactivator (rtTA) expressed under the control of SP-C and CCSP promoters. Luciferase activity was detected in the lungs of fetal and adult double transgenic mice but was not detected in other tissues or in single transgenic mice. In adult mice, maximal luciferase activity was detected 16h after the administration of doxycycline in the drinking water, or 2h after the injection of doxycycline. Activation of the transgene was observed after the administration of doxycycline in food pellets. After prolonged exposure to doxycycline, luciferase activity decreased slowly following removal of doxycycline, suggesting the importance of tissue pools which maintained expression of the transgene. In SP-C-rtTA mice, exposure of the pregnant dam to doxycycline induced luciferase activity in fetal lung tissue as early as E10.5. Luciferase activity was maintained in the lung tissue of pups during the period of lactation when the mother received doxycycline in the drinking water. In the CCSP-rtTA mice, luciferase was not detected in the absence of doxycycline. In the SP-C-rtTA mice, luciferase activity was detected in the absence of doxycycline but was enhanced approximately 10-fold by administration of drugs. The SP-C-rtTA and CCSP-rtTA activator mice control the expression of transgenes in the developing and mature respiratory epithelium, and will be useful for the study of gene function in the lung.     

Iron represses bioluminescence and affects catabolite repression of luminescence inVibrio harveyi

Bioluminescence and the synthesis of luciferase inVibrio harveyi growing in a minimal medium are repressible by iron; this is not significantly reversed by cyclic adenosine 3,5-monophosphate (cAMP). Cultures grown with added iron emit less light and possess less luciferase per cell than those grown under conditions of limiting iron; this may have significance in relation to the function of luciferase as an electron carrier. With iron, and with glycerol as the sole carbon and energy source, the addition of glucose causes further repression, both transient and permanent, and this is only partially reversible by cAMP. Without iron, glucose addition results in only a small and transient repression, but this is fully reversible by cAMP. The inability of cAMP to reverse iron-influenced repression may be explained by both a low rate of transport of cAMP into the bacteria and increased intracellular levels of cyclic nucleotide phosphodiesterase.     

<Emphasis Type="Italic">Renilla</Emphasis> luciferase-<Emphasis Type="Italic">Aequorea</Emphasis> GFP (Ruc-GFP) fusion protein,a novel dual reporter for real-time imaging of gene expression in cell cultures and in live animals

Light-emitting reporter proteins play an increasing role in the study of gene expression in vitro and in vivo. Here we present a ruc-gfp fusion gene construct generated by fusing a cDNA for Renilla luciferase (ruc) in-frame with a cDNA encoding the "humanized" GFP (gfp) from Aequorea. A plasmid containing the fusion gene construct was successfully transformed into, and expressed in, mammalian cells. The transformed cells exhibited both Renilla luciferase activity in the presence of coelenterazine and GFP fluorescence upon excitation with UV light. Spectrofluorometry of cells containing the Ruc-GFP fusion protein, in the absence of wavelengths capable of exciting GFP fluorescence but in the presence of the luciferase substrate, coelenterazine, showed an emission spectrum with two peaks at 475 nm and 508 nm. These two peaks correspond to the emission maximum of Renilla luciferase at 475 nm and that of GFP at 508 nm. The peak at 508 nm generated in the presence of coelenterazine alone (without UV excitation) is the result of intramolecular energy transfer from Renilla luciferase to Aequorea GFP. Southern analysis of genomic DNA purified from transformed Chinese hamster ovary (CHO) cells and fluorescence in situ hybridization (FISH) to metaphase chromosomes confirmed the integration of the ruc-gfp fusion gene on a single chromosome. The bifunctional Ruc-GFP fusion protein allows the detection of gene expression at the single-cell level based on green fluorescence, and in a group of cells based on luminescence emission. Furthermore, animal experiments revealed that light emission from the Ruc-GFP fusion protein can be detected externally in the organs or tissues of live animals bearing the gene construct.     

Cellular delivery of polyheteroaromate-peptide nucleic acid conjugates mediated by cationic lipids

In the search of facile and efficient methods for PNA cellular delivery, we have tested a series of PNA conjugates based on (hetero) aromatic, lipophilic compounds such as 9-aminoacridine, benzimidazoles, carbazole, anthraquinone, porphyrine, psoralen, pyrene, and phenyl-bis-benzimidazole ("Hoechst"). These chemically modified PNAs were delivered to cultured pLuc705HeLa cells mediated by cationic liposomes (LipofectAMINE or LiofectAMINE2000), and their nuclear delivery was inferred from induced luciferase activity as a consequence of pre-mRNA splicing correction by the antisense-PNA. PNAs modified with 9-aminoacridine, "Hoechst", or acetyl-"Hoechst" showed highest antisense activities (while unmodified PNA failed to show any significant antisense activity). In particular, bis-acridine-conjugated PNA showed nearly 60% splicing correction at 250 nM concentration in combination with LipofectAMINE2000. Interestingly, relative differences between the derivatives were observed when LipofectAMINE was used as compared to LipofectAMINE2000, but in general the latter yielded the higher antisense activity. The most active modifications of these PNA constructs were further tested for antisense down-regulation of luciferase in p53R cells in order to evaluate the cytoplasmic activity (uptake) of the PNAs. A dose-dependent down regulation of luciferase was demonstrated also in this system. The PNA conjugated to acetyl-Hoechst caused a reduction of luciferase activity to less than 40% of the control at a concentration of 1 muM. These results indicate that conjugation of (hetero) polyaromatic compounds to PNA can dramatically improve liposome-mediated cellular delivery both to cytoplasm as well as to the nucleus. However, no clear structure/activity relations are apparent from the present results, except that both 9-aminoacridine and "Hoechst" are also nucleic acid binding ligands.     

GroE-mediated folding of bacterial luciferases in vivo

In this study we present evidence indicating that GroE chaperonins mediate de novo protein folding of heterodimeric and monomeric luciferases under heat shock or sub-heat shock conditions in vivo. The effects of additional groESL and groEL genes on the bioluminescence of Escherichia coli cells expressing different bacterial luciferase genes at various temperatures were directly studied in cells growing in liquid culture. Data indicate that at 42° C GroESL chaperonins are required for the folding of the subunit polypeptide of the heterodimeric luciferase from the mesophilic bacterium Vibrio harveyi MAV (B392). In contrast, the small number of amino acid substitutions present in the luciferase subunit polypeptide from the thermotolerant V. harveyi CTP5 suppresses this requirement for GroE chaperonins, and greatly reduces interaction between the subunit polypeptide and GroEL chaperonin. In addition, GroESL are required for the de novo folding at 37° C of a MAV luciferase fusion polypeptide that is functional as a monomer. No such requirement for luciferase activity is observed at that temperature with a fusion of the CTP5 and subunit polypeptides, although GroE chaperonins can still mediate folding of the CTP5 fusion luciferase. Bacterial luciferases provide a unique system for direct observation of the effects of GroE chaperonins on protein folding and enzyme assembly in living cells. Furthermore, they offer a sensitive and simple assay system for the identification of polypeptide domains required for GroEL protein binding.     

Control of luminescence decay and flavin binding by the LuxA carboxyl-terminal regions in chimeric bacterial luciferases.

Bacterial luciferases (LuxAB) can be readily classed as slow or fast decay luciferases based on their rates of luminescence decay in a single turnover assay. Luciferases from Vibrio harveyi and Xenorhabdus (Photorhabdus) luminescens have slow decay rates, and those from the Photobacterium genus, such as P. (Vibrio) fischeri, P. phosphoreum, and P. leiognathi, have rapid decay rates. By generation of an X. luminescens-based chimeric luciferase with a 67 amino acid substitution from P. phosphoreum LuxA in the central region of the LuxA subunit, the "slow" X. luminescens luciferase was converted into a chimeric luciferase, LuxA(1)B, with a significantly more rapid decay rate. Two other chimeras with P. phosphoreum sequences substituted closer to the carboxyl terminal of LuxA, LuxA(2)B and LuxA(3)B, retained the characteristic slow decay rates of X. luminescens luciferase but had weaker interactions with both reduced and oxidized flavins, implicating the carboxyl-terminal regions in flavin binding. The dependence of the luminescence decay on concentration and type of fatty aldehyde indicated that the decay rate of "fast" luciferases arose due to a high dissociation constant (K(a)) for aldehyde (A) coupled with the rapid decay of the resultant aldehyde-free complex via a dark pathway. The decay rate of luminescence (k(T)) was related to the decanal concentration by the equation: k(T) = (k(L)A + k(D)K(a))/(K(a) + A), showing that the rate constant for luminescence decay is equal to the decay rate via the dark- (k(D)) and light-emitting (k(L)) pathways at low and high aldehyde concentrations, respectively. These results strongly implicate the central region in LuxA(1)B as critical in differentiating between "slow" and "fast" luciferases and show that this distinction is primarily due to differences in aldehyde affinity and in the decomposition of the luciferase-flavin-oxygen intermediate.     

Kinetics of Light Emission and Oxygen Consumption by Bioluminescent Bacteria

Oxygen plays a key role in bacterial bioluminescence. The simultaneous and continuous kinetics of oxygen consumption and light emission during a complete exhaustion of the exogenous oxygen present in a closed system has been investigated. The kinetics are performed with Vibrio fischeri, V. harveyi, and Photobacterium phosphoreum incubated on respiratory substrates chosen for their different reducing power. The general patterns of the luminescence time courses are different among species but not among substrates. During steady-state conditions, substrates, which are less reduced than glycerol, have, paradoxally, a better luminescence efficiency. Oxygen consumption by luciferase has been evaluated to be 17% of the total respiration. Luciferase is a regulatory enzyme presenting a positive cooperative effect with oxygen and its affinity for this final electron acceptor is about 4–5 times higher than the one of cytochrome oxidase. The apparent Michaelis constant for luciferase has been evaluated to be in the range of 20 to 65 nM O₂. When O₂ concentrations are as low as 10 nM, luminescence can still be detected; this means that above this concentration, strict anaerobiosis does not exist. By n-butyl malonate titration, it was clearly shown that electrons enter the luciferase pathway only when the cytochrome pathway is saturated. It is suggested that, in bioluminescent bacteria, luciferase acts as a free-energy dissipating valve when anabolic processes (biomass production) are impaired.     

A study of protein-protein interactions in living cells using luminescence resonance energy transfer (LRET) from Renilla luciferase to Aequorea GFP

We have previously reported that Escherichia coli and mammalian cells containing a fusion protein consisting of the Renilla luciferase linked to Aequorea GFP exhibited luminescence resonance energy transfer (LRET) from luciferase to GFP in the presence of coelenterazine. In this paper, we describe the construction of two gene fusions in which the cDNA for insulin-like growth factor II (IGF-II) is connected to the cDNA for a "humanized" GFP, and the cDNA for insulin-like growth factor binding protein 6 (IGFBP-6) is linked to a cDNA encoding the Renilla luciferase (RUC). The expression of the fusion gene constructs in CHO cells resulted in single polypeptides with the molecular weights expected for IGF-II-GFP and IGFBP-6-RUC, respectively, based on the use of antibodies against GFP and Renilla luciferase. The secretion of IGF-II-GFP from CHO cells was verified by fluorescence microscopy and the presence of IGFBP-6-RUC in the culture medium was confirmed by luminometry. The interaction between the two known binding partners, IGF-II and IGFBP-6, was monitored by measuring LRET from the IGFBP-6-RUC protein to IGF-II-GFP in the presence of coelenterazine, using a low-light imaging system and spectrofluorometry. Based on these data, luciferase-to-GFP LRET holds great promise for the study of protein-protein interactions in eukaryotic cells in real time.     

Synthesis of dinucleoside polyphosphates catalyzed by firefly luciferase.

In the presence of ATP, luciferin (LH2), Mg2+ and pyrophosphatase, the firefly (Photinus pyralis) luciferase synthesizes diadenosine 5',5"'-P1,P4-tetraphosphate (Ap4A) through formation of the E-LH2-AMP complex and transfer of AMP to ATP. The maximum rate of the synthesis is observed at pH 5.7. The Km values for luciferin and ATP are 2-3 microM and 4 mM, respectively. The synthesis is strictly dependent upon luciferin and a divalent metal cation. Mg2+ can be substituted with Zn2+, Co2+ or Mn2+, which are about half as active as Mg2+, as well as with Ni2+, Cd2+ or Ca2+, which, at 5 mM concentration, are 12-20-fold less effective than Mg2+. ATP is the best substrate of the above reaction, but it can be substituted with adenosine 5'-tetraphosphate (p4A), dATP, and GTP, and thus the luciferase synthesizes the corresponding homo-dinucleoside polyphosphates:diadenosine 5',5"'-P1,P5-pentaphosphate (Ap5A), dideoxyadenosine 5',5"'-P1,P4-tetraphosphate (dAp4dA) and diguanosine 5',5"'-P1,P4-tetraphosphate (Gp4G). In standard reaction mixtures containing ATP and a different nucleotide (p4A, dATP, adenosine 5'-[alpha,beta-methylene]-triphosphate, (Ap[CH2]pp), (S')-adenosine-5'-[alpha-thio]triphosphate [Sp)ATP[alpha S]) and GTP], luciferase synthesizes, in addition to Ap4A, the corresponding hetero-dinucleoside polyphosphates, Ap5A, adenosine 5',5"'-P1,P4-tetraphosphodeoxyadenosine (Ap4dA), diadenosine 5',5"'-P1,P4-[alpha,beta-methylene] tetraphosphate (Ap[CH2]pppA), (Sp-diadenosine 5',5"'-P1,P4-[alpha-thio]tetraphosphate [Sp)Ap4A[alpha S]) and adenosine-5',5"'-P1,P4-tetraphosphoguanosine (Ap4G), respectively. Adenine nucleotides, with at least a 3-phosphate chain and with an intact alpha-phosphate, are the preferred substrates for the formation of the enzyme-nucleotidyl complex. Nucleotides best accepting AMP from the E-LH2-AMP complex are those which contain at least a 3-phosphate chain and an intact terminal pyrophosphate moiety. ADP or other NDP are poor adenylate acceptors as very little diadenosine 5',5"'-P1,P3-triphosphate (Ap3A) or adenosine-5',5"'-P1,P3-triphosphonucleosides (Ap3N) are formed. In the presence of NTP (excepting ATP), luciferase is able to split Ap4A, transferring the resulting adenylate to NTP, to form hetero-dinucleoside polyphosphates. In the presence of PPi, luciferase is also able to split Ap4A, yielding ATP. The cleavage of Ap4A in the presence of Pi or ADP takes place at a very low rate. The synthesis of dinucleoside polyphosphates, catalyzed by firefly luciferase, is compared with that catalyzed by aminoacyl-tRNA synthetases and Ap4A phosphorylase.