The Molecular basis of N-acetylmuramoyl-L-alanine amidase (Rv3915) and Protein Kinase B (PknB) essentiality in Mycobacterium tuberculosis

Background: Protein kinase B (PknB) is critical for the survival of Mycobacterium tuberculosis (M. tuberculosis) in vitro and in hosts. It phosphorylates various enzymes involved in biosynthesis of cell wall and a particular autolysin classified as CwIM or Rv3915 has been recently identified as a PknB substrate. However, in-depth knowledge of this protein is still unknown. The aims of this study were to purify and investigate the activity of Rv3915, as well as monitor the phosphorylation of Rv3915 and the influence of phosphorylation on the activity of this protein. Results: Using the C41 E. coli strain containing a plasmid, with a gene encoding the protein, Rv3915 was either expressed alone or co-expressed with PknB. Sodium dodecyl sulfate polyacrylamide gel (SDS-PAGE) was used to observe the amount of Rv3915 purified, anti-poly His and anti-phosphothreonine western blot techniques were used to confirm the presence and phosphorylation of Rv3915 and zymogram assays were run to examine its activity. The results showed that Rv3195 was successfully expressed and was deemed soluble when observed in soluble fraction of E. coli lysate. It was confirmed that Rv3915 is produced as a ~45kDa protein, which does not possess any muralytic activity. However a shorter version of the protein (~25kDA) was active in zymogram, suggesting that Rv3915 is activated by cleavage. Conclusion: Rv3915 was phosphorylated by PknB and phosphorylation apparently controls stability of the protein.

These protein kinases regulate many essential processes in mycobacteria. For example, protein kinase A (PknA) and protein kinase B (PknB) are essential for mycobacterial growth and regulate cell morphology, shape and division [1,6,7] , while nitrogen metabolism, virulence and the ability of M. tuberculosis to adapt inside its host are regulated by protein kinase G (PknG), protein kinase E (PknE) and protein kinase H (PknH) [8,9,10,11]. Protein kinase F (PknF) mediates glucose transport, cell division, growth rate and morphology [12] . The genes pknA and p k n B , encoding the proteins PknA and PknB respectively, are found in an operon which contains cistrons for the production of RodA (controls the shape of the cell) and the protein phosphatase PstP and PbpA (plays a role in the synthesis of peptidoglycans).
The PknB kinase belongs to a distinct family of STPKs found only in gram-positive bacteria [13] and the important feature of these kinases is the presence of the PASTA (penicillinbinding protein and serine/threonine kinase associated) domains in the surface-exposed region [14]. In Firmicutes, PASTA domain containing kinases are not essential for growth [15][16][17][18][19][20][21] and mycobacteria appear to be a unique bacterial group in which PknB is essential for growth [1,6,7]. Over the past years, great progress has been made in the identification of PknB substrates [22][23][24] and apart from being essential for growth, PknB also regulates an oxygen-mediated replication switch [25] and the regulation of this switch in M. tuberculosis is still obscure.
The PknB has many components which includes a conserved catalytic kinase domain, a juxta-membrane part attached to a membrane-spanning region and surface-exposed sensory component, consisting of PASTA designated as PknB_PASTA domain [26][27]. The PknB_PASTA domain from mycobacteria can bind synthetic muropeptides however; it remains unclear whether this binding influences activation of PknB, bacterial growth and resuscitation [28]. The extracellular PASTA domain is believed to recognize peptidoglycan fragments and has been implicated in PknB localization [26,28], while the juxtamembrane domain recruits FhaA [29] and possibly other proteins that control peptidoglycan biosynthesis. PknB has been shown to phosphorylate multiple substrates, including proteins involved in peptidoglycan biosynthesis and remodelling [30][31][32][33]. In addition, PknB interacts with Mur ligases [34] and proteins associated with lipid metabolism [35]. However, the reason for PknB essentiality is currently unknown.
Rv3915, designated as CW1M, is N-acetylmuramoyl-L-alanyl-D-isoglutamine amidase, which is able to cleave N-acetylmuramoyl-L-alanyl-D-isoglutamine to yield free N-acetyl muramic acid [36]. The Cw1M protein of M. tuberculosis has 21% identity and 31% homology with the Cw1B protein of B. subtilis 168s [37][38]. The gene cw1B codes for an Nacetylmuramoyl-L-alanine amidase, which is one of bacterial peptidoglycan hydrolases implicated in the remodeling of peptidoglycans. The protein Cw1M is predicted to be soluble, to possess a tail, two peptidoglycan-binding domains (one close to the N-terminus of the protein and the other towards the centre of the protein) and a catalytic domain.
Three threonines in CwlM have been shown to be phosphorylated by PknB. The present study was undertaken to purify and investigate the activity of Rv3915, as well as monitor the phosphorylation of Rv3915 and the influence of phosphorylation on the activity of this protein.

Results
Expression and Purification of Rv3915 E. coli lysate were prepared as described in the methods and levels of the Rv3915 expression during E. coli growth were monitored by SDS-PAGE and Western blotting. Initially, the protein was expressed at 37°C using the E. coli strain BL21 (DE3), but the protein produced inclusion bodies and could not be purified. Thus, a different E. coli host (C41) was tested for expression of Rv3915 at 37°C. As shown in figure 1, strong bands were seen at 20kDa for the short version of Rv3915 and 45kDa for the full length version of the protein. No band was observed in the control samples, as expected.
Expression after 3hrs showed reasonable amount of protein for the lysate samples, however the highest expression of protein was seen after 20hrs. Next, we attempted to observe solubility/activity of protein, so expression at 37°C was repeated by separating the supernatant and pellet samples and ran on SDS gels, as seen in figure 2.  In figure 3, certain amounts of Rv3915 was observed in both samples, so the expression was repeated at 37°C for both SH-C and L versions of the protein and purified them using a nickel affinity column. Fractions 3 and 4 had the highest amount of protein and as a result of this, SDS gels for these fractions were ran as seen in figure 4. We observed higher amounts of Rv3915 protein for the SH-C fractions and lower amounts for the L fractions ( figure 4). Thus, a zymogram for fraction 4 of both samples were ran for comparison and faint bands of activity were observed at 17kDa for both samples which could be attributed to the degradation of samples due to instability of Pv3915 protein.
The SH-C sample of Rv3915 protein was evaluated using mass spectrometry in order to determine sites of cleavage and this resulted in an even smaller fragment.

Phosphorylation of Rv3915
The behavior and activity of a phosphomimetic of Rv3915 and the phosphorylated version of Rv3915 were expressed and monitored. This was done by co-expression of Rv3915 with PknB from a PET-DUET plasmid. It was hypothesized that phosphorylation would improve stability of Rv3915 and possibly affect its activity.
Initially, C41L and a phosphomimetic pET 3915 DDD were compared. Transformed pET 3915 DDD into the C41 E.coli strain carried out expression at 37°C for both it and C41L, purified using the nickel affinity column and ran SDS gels. Although fraction 4 had the highest amount of protein for both samples, the amount of protein expressed was quite low (figure 5). So the expression for both samples was repeated at 20°C instead, and ran on SDS gels, as seen figure 5.  A zymogram on fraction 4 of both samples showed a very faint activity for C41L at 17kDa, but no activity was seen for pET 3915 DDD The effect of phosphorylation on the stability of Rv3915 was investigated. An experiment using the fraction 4 samples of C41L and pET 3915 DDD was run. A 100µl aliquots for each sample was prepared. One aliquot of each sample had 4x SRB added to them immediately, while another aliquot, for each sample, had glycerol added to it and was left at room temperature for a few days. These aliquots were then run on an SDS gel, which is seen in figure 7. The rest of the aliquots were stored at -80°C for further use. The above experiment was repeated using new aliquots for each sample stored at -80°C.
One aliquot of each sample had 4x SRB added to them immediately after removal from storage, while the other aliquot was incubated at 37°C for 1hr before the addition of 4x SRB. These aliquots were then run on an SDS gel, as seen in figure 8.   to break up peptidoglycans on the cell wall of M. luteus, a clear or white band will be seen on the zymogram. Thus, it was also assumed that the protein had to be cleaved to 17kDa in order to become active. Therefore, a sample of SH-C of Rv3915 was evaluated using mass spectrometry, in order to determine the sites at which the protein gets cleaved. An even smaller fragment, Ami 3, was derived. It was believed t h a t this fragment would not degrade or be cleaved further and it would remain active.

Phosphorylation of Rv3915
It was believed t h a t phosphorylation would improve the stability of Rv3915 so that it would remain at 40kDa. However, the activity of Rv3915 might be compromised as a result, the initial experiment was conducted with C41L and pET 3915 DDD (phosphomimetic) ( figure 5). Phosphomimetics are amino-acid substitutions that imitate a phosphorylated protein, thereby activating or deactivating it. In the case of Rv3915, three threonine residues were phosphorylated, each phosphate group carrying three negative charges. In pET 3915 DDD, these three threonine residues were replaced with aspartic acid residues, which carry one negative charge each. However, it should theoretically behave the same way as the phosphorylated Rv3915. We went on further to test the hypothesis about Ami 3. It was expressed and a zymogram was run on its 1hr post-induction pellet/cellular extract sample. The difficulty with zymograms is that it is hard to obtain distinct images of the clear bands most times. So, in order to improve the quality of the zymograms, the zymogram was divided into four pieces and incubated them overnight at 37°C in Tris buffer, 0.1mM ZnS0 4 solution or 40mM potassium phosphate buffer at pH 5.3, 6,7 and 8. As seen from figure 10, the zymogram, which gave the clearest image at pH 5.3, showed that Ami3 is indeed active. For the next experiment, Ami 3 and PknB 3915 Duet w a s expressed and nickel column purified.
Zymograms were run on the fractions that had the highest amount of protein present. The zymograms were t h e n divided into four pieces and incubated at 37°C in 40mM sodium phosphate buffer pH 4.8, and 40mM potassium phosphate buffer at pH 5.3, 6.5 and 7.5. In figure 10, faint bands of activity were seen. However, due to the fact that Ami 3 was much more active than PknB 3915 Duet, the samples interfered with each other on the zymogram, rendering it difficult to differentiate between the two samples. Because of this, for the next experiment, the samples were run on separate zymograms. It was assumed t h a t the reason that such faint activity was seen in figure 10 was because the protein degraded after its nickel affinity column purification. Thus, we tried expressing the protein in inclusion bodies. When E. coli creates these inclusion bodies around the protein, it inactivates it, but the action of SDS on the protein sample should reactivate it again.
Thus, Ami 3 and C41WT w e r e expressed and samples were run on an SDS gel ( figure 11).
Observed a stable amount of protein in both samples, so we ran samples on separate zymograms, divided each zymogram into two pieces and incubated them overnight at 37°C in 40mM citric acid buffer pH6 and 40mM potassium phosphate buffer pH 7. None of the zymograms showed any activity, as seen from samples in figure 11. Therefore, it was concluded that Rv3915 is active in cellular extract samples and nickel affinity column purified samples, but not when expressed in inclusion bodies, although the protein samples avoided degradation in inclusion bodies.

Conclusion
To conclude, it was acknowledged that Rv3915 was produced as a 45kDa protein, which did not possess muralytic activity. The protein was cleaved into smaller fragments and two forms with molecular weights of 25kDa and 17kDa showed activity in zymograms.
Rv3915 is apparently activated by cleavage; however the molecular mechanism for this cleavage is still unknown. Only in the presence of PknB can the protein become phosphorylated and phosphorylation apparently has effects on the stability and activity of the protein Rv3915.

Methods
The aims of this study were to purify and investigate the activity of Rv3915, as well as monitor the phosphorylation of Rv3915 and the influence of phosphorylation on the activity of this protein. Purification of Rv3915 from iinclusion bodies: 500ml induction pellet were re-suspended in 40ml of lysis buffer (25mM Tris pH 8.0, 150mM NaCl, 0.5% Triton-X100, 1Mm EDTA) and sonicated using a macroprobe; 8-10 pulses, each pulse 30secs with a 1min break between each pulse. Incubated at room temperature with shaking for 30mins after which 5mM MgCl 2 was added and incubated again at room temperature with shaking for 15mins, centrifuged at 20,000G and 4°C for 20mins. The supernatant was discarded and pellet re-suspended in 40ml of 25mM Tris pH 8.0, 0.5M NaCl, 0.5% Triton-X100 and 1mM EDTA, centrifuged, discarded supernatant and the pellet re-suspended in 40ml of 25mM Tris pH 8.0, 0.5M NaCl and 1M urea. The sample was centrifuged, after which the supernatant was discarded and re-suspended pellet in 25mM Tris pH 8.0. 100µl aliquots of sample was made into eppendorf tubes, centrifuged and stored at -80°C.          Schematic showing direction of the action of RNA polymerase and the phosphorylation of RV3915 protein by PKnB