Alexander A.
Vinogradov
,
Ethan D.
Evans
and
Bradley L.
Pentelute
*
Department of Chemistry, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, USA. E-mail: blp@mit.edu
First published on 23rd March 2015
In this study we synthesized and characterized mirror image barnase (B. amyloliquefaciens ribonuclease). D-Barnase was identical to L-barnase, when analyzed by liquid chromatography and mass-spectrometry. Proteolysis of the mirror image enzyme revealed that in contrast to its native counterpart, D-barnase was completely stable to digestive proteases. In enzymatic assays, D-barnase had the reciprocal chiral specificity and was fully active towards mirror image substrates. Interestingly, D-barnase also hydrolyzed the substrate of the native chirality, albeit 4000 times less efficiently. This effect was further confirmed by digesting a native 112-mer RNA with the enzyme. Additional studies revealed that barnase accommodates a range of substrates with various chiralities, but the prime requirement for guanosine remains. These studies point toward using mirror image enzymes as modern agents in biotechnology.
We undertook this study to systematically investigate properties of an MIE in greater detail. To this end, we synthesized and characterized the enantiomers of B. amyloliquefaciens ribonuclease (barnase). Barnase is a potent guanyl-specific,7 single strand RNA specific8 endonuclease that operates via the classical mechanism of RNA hydrolysis, producing a 2′,3′-cyclic phosphate as an intermediate.9 The enzyme is more active towards long RNA molecules with the optimum pH at 8.5, but it also hydrolyzes substrates as short as dinucleotides.10 We deemed barnase an ideal target for this study due to its structural simplicity (the protein is comprised of a single 110 amino acid residue polypeptide chain with no cysteines11), reversible folding–unfolding transition,12 and straightforward catalytic activity with a fairly simple readout. Additionally, barnase may be relevant biologically; as bacterial ribonucleases are not inhibited by human ribonuclease inhibitor, barnase exhibits strong cytotoxicity on mammalian cells, and shows promising antitumor activity when conjugated to humanized HER-2 antibody.13
With both L- and D-barnase proteins in hand we turned to characterizing the catalytic activity of these enzymes. As the RNase activity assay we utilized a modified version of the fast, supersensitive fluorogenic assay developed by Raines and colleagues.18 The substrates for the assay are DNA/RNA hybrids with a single cleavage site, which provides for a homogeneous substrate needed to establish kinetic parameters for enzyme catalyzed hydrolysis (Fig. 2a). During the cleavage of the substrate, fluorescence resonance energy transfer between 6-carboxyfluorescein (6-FAM) and 6-carboxytetramethyl-rhodamine (6-TAMRA) fluorophores, installed on the 5′ and 3′ respectively, is perturbed. Thus, the increase in fluorescence of 6-FAM at 515 nm upon excitation at 495 nm can be monitored as a function of time to measure kinetics of the substrate hydrolysis. Enzyme kinetic parameters (primarily, kcat/KM) can then be obtained by the non-linear regression of experimental data to the enzyme catalyzed first-order rate equation (ESI 2.2†). In this study we investigated several different tetraoligonucleotide substrates of the common structure 6-FAM-dAX-rNX-dAX-dAX-6-TAMRA, henceforth AXNXAXAX, where N is a certain nucleotide, and the superscript X annotates the chirality of the sugar (D-sugars constitute native RNA and L-sugars—its enantiomer). In a typical assay, enzyme (100 pM to 100 nM) was added to 50–200 nM substrate in MES buffer (100 mM MES, 100 mM NaCl, pH 6.0), and the fluorescence emission was monitored. In cases where enzyme was unable to hydrolyze the substrate completely in under 60 minutes, an additional aliquot of enzyme was added to promote hydrolysis and measure the final fluorescence of the fully hydrolyzed material.
We first compared the catalytic efficiency of synthetic L-barnase to its recombinant analogue. Because barnase is known as a guanyl-specific endonuclease, we studied the hydrolysis of ADGDADAD. We found that synthetic L-barnase had a kcat/KM value of (1.2 ± 0.1) × 107 M−1 s−1 (Table 1), in line with the activity of the recombinant enzyme ((1.3 ± 0.4) × 107 M−1 s−1, Fig. 2b). For D-barnase, we expected the reciprocal catalytic activity (i.e., hydrolysis of the mirror image substrate ALGLALAL). Indeed, D-barnase hydrolyzed this substrate efficiently with kcat/KM = (1.1 ± 0.2) × 107 M−1 s−1, thus confirming our hypothesis. These data allowed us to conclude that both synthetic enzymes had full catalytic activity.
Recombinant barnase | L-Barnase | D-Barnase | RNase A | |
---|---|---|---|---|
a Hydrolysis was not detected (kcat/KM < 1 M−1 s−1, upper bound estimation). b Not determined. | ||||
ADGDADAD | (1.3 ± 0.4) × 107 | (1.2 ± 0.1) × 107 | (3.0 ± 0.6) × 103 | —a |
ALGLALAL | (3.3 ± 0.3) × 103 | (3.2 ± 0.2) × 103 | (1.1 ± 0.2) × 107 | —a |
ADGLADAD | n.d.b | (6.9 ± 0.8) × 104 | (1.0 ± 0.5) × 105 | n.d.b |
ALGDALAL | n.d.b | (1.7 ± 0.1) × 105 | (3.0 ± 0.7) × 104 | n.d.b |
ADCDADAD | n.d.b | —a | —a | (4.9 ± 0.3) × 107 |
ALCLALAL | n.d.b | —a | —a | —a |
Next, we sought to study the substrate stereospecificity of the enzymes, i.e., to evaluate the hydrolysis of ALGLALAL by L-barnase and of ADGDADAD by D-barnase. Unexpectedly, we found significant remaining activity in both cases: the kcat/KM for L-barnase was (3.2 ± 0.2) × 103 M−1 s−1, and (3.0 ± 0.6) × 103 M−1 s−1 for D-barnase. Although these values are ∼4000 times lower than the corresponding ones for the native substrates, kcat/KM of 3 × 103 M−1 s−1 still represents a fairly potent enzyme19 with the rate acceleration of ∼1010 over the uncatalyzed RNA hydrolysis.20 The similarity of kcat/KM values suggest the observation is not due to an artifact or RNase contamination. However, we performed additional experiments to exclude these possibilities. We used barstar, a well-known barnase-specific inhibitor,21 to probe its efficiency in the assays. We found that addition of two equivalents of barstar completely abolished the catalytic activity of L-barnase for both ADGDADAD and ALGLALAL, confirming that L-barnase is responsible for the cleavage of the substrates. Additionally, recombinant L-barnase, obtained independently from synthetic enzymes, had kcat/KM of (3.3 ± 0.3) × 103 M−1 s−1 towards ALGLALAL. Finally, a common source of RNase contamination are RNase A family enzymes, which are pyrimidine rather than purine specific,22 and thus are not expected to cleave the studied substrates. Accordingly, we did not detect hydrolysis of either ADGDADAD or ALGLALAL substrates by RNase A of up to 50 nM. Taken together, these data suggested that barnase may accommodate substrates of the opposite chirality.
To further investigate this phenomenon, we studied the hydrolysis of “mixed chirality” substrates, ALGDALAL and its enantiomer ADGLADAD, by L- and D-barnase. We found that both substrates were hydrolyzed by the enzymes less efficiently than the native substrates, but significantly faster than tetranucleotides with the fully inverted stereochemistry (Table 1). Thus, D-barnase hydrolyzed ADGLADAD, (the recognition guanosine had the correct chirality, while the rest was inverted) only ∼100 times less efficiently than ALGLALAL with kcat/KM as high as (1.0 ± 0.5) × 105 M−1 s−1. The second substrate, ADGLADAD, which had only the guanosine chirality inverted, was hydrolyzed by D-barnase ∼350 times slower than its native substrate. These results were corroborated by the data for L-barnase. At the same time, we could not detect hydrolysis of either ADCDADAD or its enantiomer, ALCLALAL, by L- or D-barnase. As a positive control for this experiment, we demonstrated the efficient hydrolysis of ADCDADAD by RNase A, which was consistent with previous reports. Interestingly, we could not detect the cleavage of ALCLALAL by RNase A.
Collectively, these results confirmed that barnase allows variations in the chirality of its substrates. The chirality of the main recognition nucleoside, guanosine, appears to be more important than the chirality of the rest of the substrate. Moreover, it seems barnase is not simply promiscuous because it did not hydrolyze ACAA substrates, where the key guanosine was replaced by a pyrimidine-based nucleoside. Interestingly, we also found that RNase A did not hydrolyze an enantiomer of its native substrate, which implies that the low substrate stereospecificity is not a universal phenomenon amongst digestive ribonucleases.
To expand our findings beyond the fluorogenic assay we sought to study the hydrolysis of native RNA by D-barnase. Towards this end, we incubated a 70 μg mL−1 solution of a native 112-mer RNA in 10 mM Tris, 50 mM NaCl buffer (pH 7.4) with various concentrations of D-barnase or 450 nM L-barnase for up to 4 hours and analyzed the RNA digest products by performing 10% denaturing PAGE (ESI 3.1†). As demonstrated in Fig. 3, the presence of the low molecular weight bands in cases where D-barnase was added, but not in the negative control lane, indicates that D-barnase cleaved native RNA, albeit slower than L-barnase. The latter observation is evident from digests by 450 nM D-barnase versusL-barnase. As such, we confirmed that the results of the fluorogenic assay translate into more complex systems, involving native substrates, and thus, that D-barnase is active towards D-RNA.
In another part of the study we compared the stability of L- and D-barnase towards common digestive proteases in vitro. Proteolysis of mirror image proteins was investigated before with metal-bound D-rubredoxin, which was completely stable to chymotrypsin in contrast to its enantiomer.2 Although there is evidence for the enhanced proteolytic stability of short, mostly unfolded D-peptides,23,24 the rubredoxin study remains the only published example of such behavior for folded mirror image proteins. We aimed to study the proteolytic stability of an MIE in greater detail, assaying different proteases and digestion conditions.
As proteases for this study we selected bovine trypsin, α-chymotrypsin, proteinase K, porcine elastase type IV, papain, and S. griseus protease (actinase E). These enzymes were chosen for their robust digestive proteolytic activities and a wide range of substrate specificities. Papain represented cysteine superfamily proteases, while other enzymes were serine proteases. Additionally, we wanted to assay enzymes, which are able to recognize and cleave peptide bonds after glycine. Since glycine is achiral, we hypothesized that such proteases may potentially recognize and accommodate glycine residues in mirror image proteins, allowing for the hydrolysis of these substrates. Although glycine-specific digestive proteases are uncommon, both elastase and papain are reported to cleave their substrates after glycine fairly efficiently.25,26
First, we performed the non-denaturing, in-solution digestion of L- and D-barnase by the selected proteases. Proteins were incubated in appropriate buffers (ESI 3.2†) at 37 °C for up to 19 hours with a 15:1 ratio of barnase to protease. The extent of the digestion was determined by HPLC-MS analysis (ESI 3.2.1†), and by measuring the remaining ribonucleatic activity via the fluorogenic assay. As shown in Table 2 and Fig. 4, we found that after 19 hours of digestion L-barnase demonstrated differential stability towards proteases: trypsin-digested barnase had 36% its native activity, while in the case of proteinase K less than 0.2% activity remained. In all six cases L-barnase lost a significant portion of its catalytic activity. In contrast, D-barnase proved completely stable to all assayed proteases; it retained full catalytic activity, and no digestion products could be detected by HPLC-MS.
Protease | L-Barnase | D-Barnase |
---|---|---|
No protease | 100.0 ± 6.4 | 100.0 ± 5.1 |
Trypsin | 36.7 ± 2.0 | 108.0 ± 8.9 |
Chymotrypsin | 9.2 ± 0.5 | 103.1 ± 5.3 |
Proteinase K | 0.2 ± <0.1 | 99.8 ± 8.1 |
Elastase | 0.6 ± <0.1 | 103.5 ± 3.3 |
Papain | 0.8 ± <0.1 | 96.6 ± 4.3 |
Actinase E | 0.3 ± <0.1 | 101.4 ± 7.3 |
Additionally, we performed a more forcing in-solution denaturing digestion of L- and D-barnase using the most potent protease, proteinase K. To denature the protein, barnase was incubated in 6 M Gn·HCl, 50 mM Tris buffer (pH 7.4) at 95 °C for 20 minutes, and then digested with proteinase K (barnase:protease = 2:1) in 2 M Gn·HCl at 37 °C. We found that D-barnase was completely stable to proteolysis even under such forcing conditions, in stark contrast to L-barnase, which was digested completely (ESI 3.2.2†). Finally, we attempted to digest D-barnase with proteinase K by increasing the digestion time. Using HPLC-MS analysis we did not detect any digestion products after 168 hours (1 week) of incubating D-barnase with proteinase K. The digest was indistinguishable from a negative control experiment, where no protease was added to the enzyme (ESI 3.2.3†).
The results of this study pose a number of questions. First, it is unclear how barnase recognizes and cleaves substrates of the opposite chirality. The enzyme is known to have several subsites, which facilitate the substrate binding and its proper orientation for catalysis.10 Our data are consistent with this model, as we observed a range of kcat/KM values by only changing the chirality of AGAA tetranucleotide. It is conceivable that substrates of the mixed chirality, e.g. ALGDALAL, occupy only certain subsites in the enzyme, e.g. the guanosine binding subsite in this case, and thus the catalysis may still proceed. It is also unclear whether this enzymatic activity is merely spontaneous or was subject to the evolutionary selection at some point. At this time we are unaware of any practical implications of such catalysis: to the best of our knowledge, RNA of L-configuration is unknown in nature.
Nevertheless, our study suggests that at least in some cases enzymes may utilize substrates of the opposite chirality. The mirror image form of such an enzyme will then act on the same targets as its native counterpart. Although decreased catalytic efficiency is expected, the enzyme may still achieve a notable rate acceleration. This property of MIEs can be highly desirable from the biotechnology standpoint for, as we confirmed in the case of barnase, MIEs can be extraordinarily resistant to proteolysis and at the same time carry the native biological function. Importantly, this effect may manifest itself without protein engineering and/or evolution of the enzyme. As we also found with the example of RNase A, this effect by no means is universal, and more investigations would be needed to establish the generality of our findings.
Footnote |
† Electronic supplementary information (ESI) available: Materials, methods, and detailed experimental procedures. See DOI: 10.1039/c4sc03877k |
This journal is © The Royal Society of Chemistry 2015 |