3bjq

Protein Summary

The BB3626 gene from Bordetella bronchiseptica, an opportunistic pathogen, encodes the NP_890161 protein, a member of the PF03864 group. This family representative is the major capsid protein E, a protein involved in DNA stabilization within bacteriophage heads. Members of the PF03864 family are found in a number of pathogenic bacteria as well as in some double stranded (ds) bacteriophages.

Despite the lack of sequence homology, 3bjq is structurally very similar to virus envelope proteins such as pfV and HK97 gp5 (2e0z with a Dali Zscr=15; and 1ohg, Zscr=13). A surprising observation is that, instead of forming virus envelopes like 2e0z and 1ohg, 3bjq is a decamer in current crystal form. The decamer is formed by tail to tail stacking of two pentamers.

Compared to 2e0z, 13% seq id, rmsd 3.1, (200 ca out ~300 of 3bjq)
Compared to 1ohg, 11% seq id, rmsd 2.8 (197 ca out of ~300 of 3bjq)


So the question is: is 3bjq also a virus envelop protein which will form virus envelope under certain conditions  Or is 3bjq a similar folded protein but with different function? why does such protein exist in pathogenic bacterium?

The conserved residues are scattered around the whole protein and buried, as a result, there is no easily identifiable highly conserved active site.



The structure has unusual electrostatic properties.  Both the inside and outside surfaces of the decamer are predominately negatively charged. This property is very similar to 2e0z and 1ohg.

Thus, from 1) no conserved active site, 2) conserved electrostatic surface for DNA encapsulation, we suggest that 3bjq may also form a virus like particle like 2e0z under certain conditions. The observed pentamer may represent a core for the envelope assembly (TODO: check whether the pentamer is same as the pentamers in HK97 or pfV). In other words, the current structure may represent a pre-assembly intermediate state, eg, HK97 can also exist in solution as pentamers.

Monomer electrostatic surface


Electrostatic surface, looking into the inner chamber formed by the pentamer

Electrostatic surface, looking at the outside surface of the pentamer


Phage-related protein (3bjq).

TODO: compare with:
(a) major capsid protein E is involved with the stabilization of the condensed form of the DNA molecule in phage heads (PF03864).
(b) Linocin M18 (PF04454).

The most conserved secondary structural element of 3bjq, helix H3, is the one that exhibits the most diversity in similar structures. Superposition of  3bjq with a linocin-like protein from Thermotoga maritima termed encapsulin (PDB id: 3DKT, Dali Z=14), a virus-like particle from the archaeon Pyrococcus furiosus (PDB id: 2E0Z, Z=15) and a putative prophage protein from Escherichia coli cft073 (PDB id 3BQW, Z=14) shows that in all cases H3 is replaced by a conformationally flexible beta-hairpin (see figure below). This region, termed the E-loop, is involved in capsid assembly (Gan 2006) as well as, in the case of the linocin-like protein from T. maritima, in tight interactions critical to the formation of a 60-mer enzyme-encapsulating, bacterial nanocompartment implicated in the oxidative-stress response (Sutter 2008).

Stereo ribbon diagram showing superposition of 3BJQ (blue), 3DKT (yellow), 2E0Z (magenta) and 3BQW (cyan).

[Could a conformational switch in the E-loop be responsible for different oligomerization states that could transform BB3626 from e.g. a bona fide viral capsid to an encapsulin or some other function altogether (zinc protease, see below)? In other words, could this be a moonlighting protein with different functions triggered by changes in environmental conditions?]

The genome context (http://string.embl.de) of PF03864 family strongly suggests a bacteriophage origin with all members showing strong predicted functional associations with phage capsid, tail sheath, tail core and portal proteins.

Other computational predictions of function

SeqFEATURE is a library of 3D protein functional site models built from 1D sequence motifs. 3BJQ had a very high scoring hit for the following model:

ZINC_PROTEASE.4.GLU.OE1

(site ID)           (z-score)       (X)     (Y)     (Z)      (res_ID:chain@atom)
Env_3BJQ.pdb_92     3.7743673860    -89.925 91.923  65.278   GLU123:F@OE1
Env_3BJQ.pdb_158    3.6719361952    -95.627 115.015 41.195   GLU123:J@OE1
Env_3BJQ.pdb_22     3.5686598948    -90.287 37.524  27.977   GLU123:B@OE1
 

This model for a zinc protease active site has an area under the ROC curve of ~ 0.89 and a sensitivity of 50% using a 100% specificity cutoff. Using this cutoff, the model hit 3 of the 10 chains at the same residue (GLU123), suggesting that this residue in each subunit may be functional. The 3 hits score well within the bulk of the positive training site score distribution:

More information about this model

Below are images of the local structural environment surrounding the predicted functional site.

For comparison, below are images of the local structural environment surrounding a known zinc protease active site (PDB: 1EZM).

While the predicted site is very high scoring, it is not necessarily a zinc protease active site. The functional models are based on the presence or absence of significant physichochemical properties as defined by statistically modeling data in a radially averaged form from positive and negative training examples. A high scoring hit for a model indicates that the radial distribution of physicochemical properties at that site in the query structure was very similar to the positive training examples for that model.

Corroborating evidence for zinc-related functional activity comes from analysis of the structure with a zinc binding site model, also built using the FEATURE system. The figure on the left below show all of the hits (in blue); the 10 highest scoring hits correspond to one hit per subunit, which fall in the direct vicinity of the predicted zinc protease site (as shown in the figure on the right - white = predicted zinc site, red/blue = predicted zinc protease site).

It is thus possible that there is a zinc protease active site in 3BJQ at the location specified, enzymatic activity similar to zinc proteases, zinc or other cation binding activity, or simply an area of interest that may warrant further investigation.

Ligand Summary

References