Kinetics of the thermal inactivation and the refolding of bacterial luciferases in Bacillus subtilis and in Escherichia coli differ

Summary

Here we present a study of the thermal inactivation and the refolding of the proteins in Gram positive Bacillus subtilis. To enable use of bacterial luciferases as the models for protein thermal inactivation and refolding in B. subtilis cells, we developed a variety of bright luminescent B. subtilis strains which express luxAB genes encoding luciferases of differing thermolability. The kinetics of the thermal inactivation and the refolding of luciferases from Photorhabdus luminescens and Photobacterium leiognathi were compared in Gram negative and Gram positive bacteria. In B. subtilis cells, these luciferases are substantially more thermostable than in Escherichia coli. Thermal inactivation of the thermostable luciferase P. luminescens in B. subtilis at 48.5°С behaves as a first-order reaction. In E.coli, the first order rate constant (K_t) of the thermal inactivation of luciferase in E. coli exceeds that observed in B. subtilis cells 2.9 times. Incubation time dependence curves for the thermal inactivation of the thermolabile luciferase of P. leiognathi luciferase in the cells of E. coli and B. subtilis may be described by first and third order kinetics, respectively. Here we shown that the levels and the rates of refolding of thermally inactivated luciferases in B. subtilis cells are substantially lower that that observed in E. coli. In dnaK-negative strains of B. subtilis, both the rates of thermal inactivation and the efficiency of refolding are similar to that observed in wild-type strains. These experiments point that the role that DnaKJE plays in thermostability of luciferases may be limited to bacterial species resembling E. coli.

Keywords:

Plasmid construction – Luciferase – Luminescence – Bioluminescence – Bacillus subtilis – Gram positive bacteria – Chloramphenicol

Introduction

When exposed to mildly elevated temperatures, eukaryotic and prokaryotic thermolabile proteins transiently undergo partial or complete unfolding, resulting in a loss of their activity [1]. Persistence of the heat stress prevents the proteins from refolding to their native state, while favoring alternative, beta-sheet enriched conformations. To prevent misfolding, eukaryotic and prokaryotic cells employ a variety of molecular chaperones, the most abundant and best studied of which being Hsp60-Hsp10 (GroEL-GroES), Hsp70-Hsp40-nucleotide exchange factor (DnaK-DnaJ-GrpE), Hsp100 (ClpA–ClpB–ClpX) and so-called small chaperones sHsp (IbpA-IbpB) [2–10].

In studies of the folding, misfolding and refolding conditions in Escherichia coli, bacterial and firefly derived luciferases often serve as model substrates [10–13]. Moreover, luminescent Gram positives bacteria have their use as biosensors suitable for clinical applications [14–18]. Despite enormous biotechnological importance of Gram positive cells in general, and of B. subtilis, a workhorse of industrial recombinant protein production in particular, the mechanisms of the folding and the refolding of the proteins in these bacterial cells remain enigmatic.

In present work we use the model bacterial luciferases differing in their thermostability to investigate the thermal inactivation and the refolding of the proteins in Gram positive B. subtilis. A variety of bright luminescent B. subtilis strains which express luxAB genes encoding luciferases from bacteria P. luminescens [14] and P. leiognathi [15], are utilized in the comparative study of the kinetics of the thermal inactivation and the refolding of the luciferases in Gram negative and Gram positive bacteria. In particular, we evaluated effects of dnaKJ genes on luciferase thermostability in cellular environments of B. subtilis and E. coli.

Materials and methods

Bacterial strains, plasmids, and growth conditions

Bacterial strains are presented in Table 1. Plasmids are presented in Table 2.

E. coli and B. subtilis was grown either in LB media, with constant aeration at 200 rpm at 37°С unless indicated otherwise. Solid media plates were prepared using 1.5% of agar.

For selection, media we made with spectinomycin 150 μg/ml, ampicillin 100 μg/ml and chloramphenicol 10 μg/ml.

Transformation

B. subtilis cells were transformed according to the protocol of Spizizen [21]. E. coli cells were transformed using calcium chloride protocol [22].

Enzymes and chemical substances

The substrate for luciferase n-decanal was from Sigma-Aldrich (USA). Enzymes for cloning were purchased in Promega (USA). Media were from Helicon (Moscow, Russia). Oligonucleotides were made by Syntol (Moscow, Russia).

Constructing the plasmids

Primers utilized for constructing the plasmids are described in Table 3. As a backbone for assembly of biosensors we selected shuttle plasmid pMWAL-1TPpur with two origins pMW118 (GenBank: AB005475) and pBS72 [23], which allows teta-type replication, as well as amplicillin and chloramphenicol resistance gens bla and cat for E. coli and B. subtilis, respectively. With an aid of P1/P2 primers, trimethoprim resistance gene dhfr from B. cereus ATCC14579 was introduced into the plasmid pMWAL-1TPpur under its Ppur promoter. With an aid of P3/P4 primers, rrnB terminator T1T2 was introduced to build a promoterless plasmid pMWAL-1Ppur_dhfr_t1t2_MCS. In turn, with an aid of primers P5/P6, this construct was modified by inserting constitutive promoter of fructose-1,6-bisphosphate aldolase gene (PfbaA) [24] into SacI recognition site, thus, resulting in plasmid pPfbaA_MCS. Later, plasmid pPfbaA_MCS was utilized for cloning of α and β subunits of luciferases from P. luminescens and P. leiognathi, which were amplified on the pXen5 [14] and pLF22ABleo [15] templates using P7/P8 and P9/P10 primers, respectively, to obtain pPfbaA_XenAB and pPfbaA_LeoAB plasmids, respectively. Both subunits of P. luminescens and P. leiognathi luciferases were cloned into pTZ57R vector to obtain constructs pTZ57R_xenAB (with pair of primers P7/P8) and pTZ57R_leoAB (with pair of primers P9/P10), respectively.

Measurement of the intensity of bioluminescence

Cell were prepared by overnight cultivation at 30°С with aeration at 200 rpm in LB media with chloramphenicol or ampicillin, then diluted 1:100 in LB media with chloramphenicol, grown till reaching OD = 0.4–0.5, then incubated at a certain temperature. To eliminate de novo protein synthesis, incubation media was supplemented with following antibiotics: chloramphenicol (167 μg/ml) for E. coli and tetracycline (60 μg/ml) for B. subtilis.

To measure the intensity of bioluminescence the cells were sampled into the 200-μl test tubes like in [25]. Two μl of 0.1% n-decanal dissolved in ethanol were added to final concentration of 0.001%. Cells are placed in the luminometer without shaking at room temperature (~20°С), with direct measurements of total bioluminescence (in RLU, relative light units) using “Biotox-7” (LLC EKON, Russia). In five seconds timeframe, luciferase substrate n-decanal enters the cells and participates in light producing reaction.

Results

Hybrid plasmid with P_fbaA or P_lac controlled luxAB-genes were introduced to B. subtilis and E. coli cells, respectively. The levels of resultant strain bioluminescence in cultures sampled at varying OD are shown in Table 4. As could be seen form Table 4, bioluminescence intensities observed for B. subtilis and for E. coli cultures are approximately the same.

To quantify relative thermostability of luciferases in vivo, the luxAB gene expressing cells of B. subtilis 168 and E. coli BW25113 were grown at 28°С and 37°С for cells with luciferase from P. leiognathi and P. luminescens, respectively, until OD = 0.4–0.6 was reached. Then the tetracycline or chloramphenicol was added, and incubation was continued at elevated temperatures. In this experiment, addition of antibiotics prevented protein synthesis de novo.

Fig 1 shows the luminescence of B. subtilis 168 (with pFba-xenAB or pFba-leoAB) and E. coli BW25113 cells (with pTZ57R-xenAB or pTZ57R-leoAB) at various temperatures.

As could be seen at Fig 1, in B. subtilis the luciferases display higher thermostability than in E. coli. When expressed in B. subtilis, each luciferase reached inactivated state at the temperature of 3–5°С higher than in E. coli. In course of subsequent experimentation with thermal inactivation of luciferase in B. subtilis and E.coli cells in vivo, the differences in luciferase thermostabilities were taken into account.

According to data obtained in vitro [26] and in vivo in E. coli cells [11], P. leiognathi luciferase is significantly more thermolabile than luciferase from P. luminescense. The data in Fig 1 show that the same difference persists in the cells of B. subtilis.

Kinetics of luciferase thermal inactivation in vivo in B. subtilis cells were compared to that observed in cells of E. coli. Fig 2 presents inactivation kinetics at 41°С or 48,5°С for luciferases from P. leiognathi (Fig 2A), and P. luminescens (Fig 2B), respectively.

As could be seen from the data presented at Fig 2A, in all strains the bioluminescence intensity drops observed for P. luminescens luciferase are well described by a semi-logarithmic graph (lg A–time), and, therefore, thermal inactivation of this luciferasae is a first-order reaction. Table 5 shows respecitive first order rate constants (Kt).

As could be seen from Table 5, for luciferase of P. luminescens the ratio of the rate constants for wild type strains of E. coli and B. subtilis was 2.9, while for ΔdnaK mutant strains of same bacteria this ratio was 4.3.

As could be seen from the data presented at Fig 2B, for luciferase of P. leiognathi expressed in E. coli cells the bioluminescence intensity drops are also well described by a semi-logarithmic graph (lg A–time), while in B. subtilis cells respective kinetics are substantially more complex. To find out the order of this reaction, Rakovsky techniques was employed [27, §301 p. 466] by linearizing it in coordinates ln t_1/2 Vs. ln A₀, where A₀ –initial activity, t_1/2 –time to semi-inactivation (Fig 3A).

As could be seen from Fig 3A, for both lines the slope is close to 2, indicating that the kinetics of this reaction is of a third-order. These data were linearized in coordinates ½*A^-2 Vs. t and approximated by lines shown on Fig 3B. Table 6 shows rate constants for P. leiognathi thermal inactivation at 41°С in E. coli and B. subtilis cells.

Notably, the difference in the rates of P. leiognathi luciferase thermal inactivation at 41°С in the cells of E. coli and B. subtilis, which are described by first order and third order kinetics, respectively, leads to a prominent, orders-of magnitude difference in the levels of cell bioluminescence at 20 minutes post thermal inactivation onset.

When the luciferase of P. leiognathi undergoes thermal inactivation in E. coli, similar kinetics are observed at lower temperatures, which is explained by the presence of active bichaperone system DnaKJE-ClpB, which actively aids refolding [11]. However, as could be seen at Fig 2, in B. subtilis dnaK+ and dnaK-, kinetics of luciferase inactivation remain the same, while in E. coli the ratio of rate constants was at 1.8. This observation indicates that DnaKJE chaperone does not properly function in B. subtilis, as it is unable to support the refolding of luciferase.

Fig 4 depicts the data describing the refolding of luciferases from P. leiognathi and P. luminescens after thermal inactivation in vivo in E. coli or B. subtilis cells. Bacterial cells were incubated at 47°С or 51°С for E. coli or B. subtilis, respectively. In both cases, the luminescence levels gradually decreased across approximately 2–3 orders of magnitude till reaching the background levels, at which point the translation inhibitors, chloramphenicol for E. coli and tetracycline for B. subtilis, were added to media followed by incubating bacterial cultures at room temperature under continuous monitoring of their luminescence. Thermal inactivation time was about 15–25 minutes.

As could be seen from the data presented on Fig 4A after thermal inactivation, the luminescence of E. coli cells expressing the P. leiognathi luciferase could be restored almost to its pre-inactivation levels. Evidently, this restoration is dependent on DnaK, as E. coli ΔdnaK cells are not able to restore their levels of luminescence after exposure to high temperatures.

In curves describing of the reactivation of thermally inactivated luciferases, both lag-period and inflection are unremarkable, thus, being indicative of possible multiples stages of reactivation reaction limited by some rate-limiting. Analysis of relatively fast refolding steps provides some difficulty due to rapid cooling down of the sample within the first few minutes after its return to the room temperature.

Analysis of slower steps of P. leiognathi luciferase refolding in E. coli cells, most parts of the kinetic curve are well approximated by exponential dependence of an accumulation of the product resulting from the first-order reaction with lag-period A = A_max×(1-e^k×(t-tlag)), where A_max is the level of maximal degree of reactivation, k–first-order reaction rate constant and tlag–the time required for completion of fast steps of the refolding (Fig 5).

Kinetics curve parameters A_max, k and t_lag were derived by non-linear approximation using SciDAVis software (A_max = 57,11±0,85%, k = 0,035±0,002 min^-1 and t_lag = 3,42±0,44 min).

Refolding kinetics analysis of other strains showed that thermal inactivated cells of B. subtilis wild-type strain 168 are substantially less capable of luciferase refolding than E. coli, and successfully refold just approximately one percent of available denatured enzymes. In B. subtilis, the success of the refolding does not require the presence of DnaK chaperone. As opposed to E. coli, the cells of B. subtilis refold luciferases of P. leiognathi and P. luminescens to about the same levels, and with similar kinetics.

Discussion

A comparison of thermal inactivation kinetics of luciferases in E.coli and B. subtilis strains transformed with lux-biosensor plasmids demonstrated that the thermostability of these model proteins in Gram positive bacteria B. subtilis is higher than that in Gram negative E. coli, with the difference in tolerated temperatures reaching 4–5°C (Figs 1 and 2).

Earlier works demonstrated that luciferase activity in E. coli cells depend on ability of particular luciferase to refold [10]. According to date presented above, in Gram positive bacteria B. subtilis, enhanced thermostability of bacterial luciferases is not because of better refolding. In fact, native B. subtilis cells do not support luciferase refolding well. After incubation of B. subtilis cells with luciferase-bearing constructs on elevated temperatures, bioluminescience drops 3–4 orders of magnitude; the transfer of these cells back to the room temperature results in restoring luciferase activity up to approximately 1% of its initial levels either in presence or in absence of DnaKJ chaperone. In conclusion, our experiments point that the role that DnaKJE plays in termostability of luciferases in E. coli is limited to this models system. In fact, in B. subtilis cells this chaperone is not involved in improving the thermostability of luciferase.

Possibly, activity of thermal inactivated luciferases in B. subtilis may be rescued by other ATP-dependent chaperones, which are yet to be investigated. A set of biosensors plasmids incorporating luciferases of varying intrinsic thermostability, which we presented above, may facilitate further molecular and genetic dissection of the factors which govern the denaturing and the refolding of recombinant proteins in Gram positive cells.