Serine proteases and ovine footrot

Footrot is a disease that is of importance to the wool and sheep meat industries. The principle causative agent of ovine footrot is the anaerobic bacterium, Dichelobacter nodosus, virulent isolates of which secrete three closely related subtilisin-like proteases, AprV2, AprV5 and BprV¹. By constructing isogenic mutants and carrying out virulence tests in sheep it was shown that AprV2 is a major virulence factor of D. nodosus². Structural analysis of AprV2 has revealed that it contains several novel loops, one of which appears to act as an exosite that may modulate substrate accessibility². Both elastase activity and protease thermostability have been used for the differential diagnosis of D. nodosus isolates. Analysis of the protease mutants has shown that AprV2 is the thermostable protease and also is responsible for the elastase activity of D. nodosus, while AprV5 is the major extracellular protease². In addition, AprV5 is required for its own maturation and for the optimal cleavage of AprV2 and BprV to their mature active forms³.

The severity of ovine footrot is a continuum that ranges from benign footrot, which presents as an interdigital dermatitis, to virulent footrot, which involves severe underrunning of the horn of the hoof and the separation of the hoof from the underlying tissue (Fig. 1A), leading to lameness (Fig. 1B) and loss of body weight^4–6. Consequently the disease results in significant losses to the sheep industry due to a reduction in meat and wool production and the cost of control and treatment programs^7,8.

Figure 1. Gross pathology of a sheep infected with ovine footrot. Separation of the horn of the hoof (A) and lameness (B) are shown.

Clinical disease is dependent upon the virulence properties of the causative D. nodosus isolate and the presence of warm and wet climatic conditions. Type IV fimbriae and extracellular serine proteases were shown to be major virulence factors following the development of methods for the genetic manipulation of D. nodosus^2,9. Type IV fimbriae also are required for optimal serine protease secretion^9–11.

Virulent strains of D. nodosus secrete two acidic proteases, AprV2 and AprV5, and the basic protease BprV, which putatively cause tissue damage during a footrot infection^2,12. The equivalent proteases in benign strains are known as AprB2, AprB5 and BprB^13,14. Phenotypic characterisation of the extracellular proteases, including analysis of their elastase activity and protease thermostability, has been traditionally used for the differentiation of benign and virulent strains of D. nodosus^15,16, with virulent isolates often producing elastase positive, thermostable proteases and benign strains generally having an elastase negative, more thermolabile protease phenotype. Note that these phenotypes do not absolutely correlate with virulence.

These proteases are synthesised as precursors with an N-terminal pre-pro-region, a serine protease domain and a C-terminal extension. The amino acid sequences of the catalytic protease domains are highly conserved (~65% identity)^17–19, but the sequences of the C-terminal extensions are less conserved, showing only approximately 35% similarity. The active proteases are produced by cleavage of the pre-pro region and the C-terminal extension^17,18,20. Sequence analysis of the C-terminal extensions revealed that they contain a P-domain, which is typically associated with eukaryotic pro-protein convertases that belong to the subtilisin-like protease superfamily^6,21.

To determine the role of each of the proteases in virulence, mutants of each of the three protease genes were constructed by allelic exchange in the virulent strain VCS1703A². Quantitative protease assays of culture supernatants, using azocasein as the substrate, revealed that AprV5 made the major contribution to overall protease activity; followed by AprV2, with BprV only making a minor contribution². Double mutants had very little protease activity, suggesting that the proteases may act synergistically, or that one or more may be involved in the activation of the other proteases². The mechanism of processing of the extracellular proteases was further examined by zymogram analysis and Western blotting using AprV5-, AprV2- and BprV-specific antisera³. The results indicated that AprV5 is responsible for its own maturation and for the optimal processing of both AprV2 and BprV. By constructing a series of C-terminal truncated aprV5 mutants in D. nodosus, it also was shown that the C-terminal extension of AprV5 is required for efficient processing of all three enzymes, presumably because it is required for the optimal self-processing of AprV5. In the absence of this domain, protease processing is delayed³. Moreover, it was shown that cleavage of the pro-domain and the C-terminal extensions of the AprV2 and BprV precursors occurs after secretion.

Elastase assays revealed that AprV2 is responsible for the elastase activity found in most virulent isolates of D. nodosus (Fig. 2A)². AprV2 is also the major thermostable protease (Fig. 2B). Complementation of the aprV2 mutant with the benign protease gene aprB2 restored the overall protease activity to that of the wild type, but not the elastase activity or thermostability; therefore this strain had the phenotype of a benign strain, despite possessing two functional virulent protease genes, aprV5 and bprV.

Figure 2. Elastase and protease thermostability. The wild-type strain VCS1703A (WT), its isogenic aprV2 mutant (aprV2) and the complemented derivative (aprV2(V2⁺)) were analysed for their elastase activity by plating onto culture medium containing insoluble elastin. Elastase activity is shown by a zone of clearing around the culture streak (A). Protease thermostability was determined by a gelatin-gel test¹⁶, whereby culture supernatants of each strain were tested for protease activity after incubation at 65^oC for 0 min, 8 min and 16 min, as shown (B).

The virulent wild-type strain, each of the protease mutants, and their corresponding complemented derivatives, were examined in a sheep virulence trial to determine the role of the proteases in disease². The aprV2 mutant was avirulent and complementation of the mutant with the wild-type aprV2 gene restored full virulence. Surprisingly, complementation with aprB2 (the equivalent protease from a wild-type benign strain) also restored full virulence, which indicates that elastase activity is not required for virulence. The aprV5 and the bprV mutants were also avirulent, but unexpectedly their complemented strains were not virulent, possibly due to unstable protease expression in the complemented strains. Therefore, although we cannot conclude that AprV5 and BprV are essential for virulence, it is likely that they do play a role in the disease process.

It has been known for some time that the mature AprV2 and AprB2 proteases differ by a single amino acid, with Tyr92 of AprV2 substituted by an Arg residue in AprB2¹⁸. To better understand how this single amino acid change could contribute to the substrate specificity of these proteases the crystal structures of both proteases were determined². The structures were closely related and were similar to other subtilisin-like proteases, although they each had several major insertions (I1 – I4) in the loops surrounding the active site cleft. The largest of these insertions (I2) is tethered by a disulphide bond, and the single amino acid change (Y92R) between AprV2 and AprB2 is located at the end of this loop. Site-directed mutagenesis revealed that Tyr92 does not contribute to catalysis at the active site, but that the I2 loop acts as an exosite and mediates the formation of a stable enzyme-substrate interaction. The tethering of the loop by the disulphide bond appears to be important for this function². Finally, analysis of the crystal structures of BprV and BprB reveals that differences in the substrate specificity of these enzymes reflects amino acid changes in the S1 pocket of these enzymes and that subtle changes in the D. nodosus proteases may significantly influence tissue destruction²².

For many years extracellular serine proteases had been considered as putative virulence factors of D. nodosus; however, their importance in the pathogenesis of disease has only recently been established. There is now clear evidence that AprV2 is the major thermostable protease and elastase in virulent strains of D. nodosus and is essential for virulence. In addition, it has been shown that AprV5 is required for its self-maturation and that it facilitates the activation of AprV2 and BprV, a process that requires the C-terminal extension of AprV5. The fact that a strain in which an aprV2 mutation was complemented with the aprB2 gene was benign by standard laboratory diagnostic tests, but caused virulent disease in sheep, indicates that elastase activity and thermostability are not direct indicators of the virulence of a strain. This finding is in keeping with the difficult task of precisely defining a virulent isolate of D. nodosus, an issue that has been noted for many years in this field. The presence of other virulence factors, such as the other virulent proteases and type IV fimbriae-mediated twitching motility, also contributes to virulence. Virulence in D. nodosus clearly is multifactorial and we suspect that the pathogenesis of disease is much more complex than currently envisaged.