The Chlamydia protease CPAF: Caution, Precautions And Function Free

Recent biochemical and genetic studies have called into question many of the functions proposed for the chlamydial protease CPAF (Zhong et al., 2001). In 2012, we demonstrated that the proteolysis of multiple reported substrates was occurring during the preparation of lysates from Chlamydia infected cells rather than in intact cells (Chen et al., 2012). The recent report by Snavely et al. of a CPAF null mutant now provides genetic evidence that CPAF is not necessary for many Chlamydia induced phenotypes nor for a successful intracellular infection (Bavoil & Byrne, 2014; Snavely et al., 2014).

Caution

The proteolysis observed when analyzing Chlamydia infected cells must be carefully interpreted because lysates prepared under standard conditions contain CPAF proteolytic activity (Chen et al., 2012). In general, proteolysis during lysate preparation is not a problem for protein analysis because proteases in cell lysates are typically inactive on ice and are readily inhibited with a protease inhibitor cocktail. However, CPAF is unusual because it remains active in cell lysates at 4 °C and is resistant to most protease inhibitors (Zhong et al., 2000; Jorgensen et al., 2011; Chen et al., 2012).

As an unfortunate consequence of this ongoing enzymatic activity, CPAF cleaves or degrades numerous host proteins during lysate preparation from infected cells. Multiple studies have reported that CPAF mediated proteolysis of specific host proteins begins at about 10–20 h postinfection (hpi) (Dong et al., 2004; Kumar & Valdivia, 2008; Christian et al., 2011). However, we did not detect the published proteolysis of 11 host proteins, even as late as 48 hpi, if we inhibited CPAF activity at the time of lysate preparation (Chen et al., 2012).

These doubts about the proteolysis of proposed CPAF targets challenge the prevailing view that CPAF is a major, multifunctional virulence factor in Chlamydia (Paschen et al., 2008; Zhong, 2009). CPAF mediated proteolysis of specific host proteins has been proposed to cause distinct Chlamydia induced phenotypes, including Golgi fragmentation and resistance to apoptosis. This model has been based on the observed proteolysis of a host protein in infected cell lysate, as detected by immunoblot analysis, linked with the known role of this protein in an uninfected cell. However, there has been a lack of direct evidence that the phenotype in an infected cell is produced by altering this protein. We questioned this overall model because phenotypes were still present in infected cells even though we did not detect proteolysis of the host proteins proposed to cause these phenotypes (Chen et al., 2012). Indeed, many of these Chlamydia induced phenotypes are still present in the CPAF null mutant generated by Snavely and Valdivia, providing genetic evidence that CPAF is not required for their formation (Snavely et al., 2014).

Precautions

Our experience with CPAF has taught us that the analysis of proteins from Chlamydia infected cells must be performed under conditions that eliminate CPAF activity during lysate preparation. This precaution should be taken for studies of individual proteins by immunoblot, as well as for global protein analysis, such as mass spectrometry based proteomic studies. In addition, this issue applies to chlamydial proteins because they too can undergo in vitro proteolysis during preparation of infected cell lysates (B. Hanson and M. Tan, unpublished).

Three methods to inhibit CPAF activity have been published (Chen et al., 2012; Snavely et al., 2014): (1) treatment of cells with the CPAF inhibitor clasto lactacystin 1 h before lysis; (2) direct lysis of cells in 8 M urea; and (3) direct lysis of cells in hot 1% SDS buffer. There was no detectable proteolysis of all 11 host proteins that we studied when we inhibited CPAF activity with the urea lysis method (Chen et al., 2012), or clasto lactacystin treatment (A.L. Chen, K.A. Johnson, J.K. Lee, C. Sütterlin and M. Tan, unpublished). Urea and hot SDS nonspecifically inhibit CPAF activity by denaturing proteins. However, clasto lactacystin selectively blocks proteasome and CPAF activity (Zhong et al., 2001), making it likely that CPAF, rather than nonspecific proteases, is responsible for the Chlamydia induced proteolysis observed in the absence of these precautions.

While precautions to inhibit CPAF activity during lysate preparation are critical, it is equally important to verify that they have effectively removed all remaining CPAF activity from the lysates. CPAF activity in an infected cell lysate can be measured with an in vitro assay in which the lysate is incubated with a known substrate, followed by immunoblot analysis to detect any proteolysis of the substrate (Chen et al., 2012). Our group is attuned to the possibility of residual CPAF activity despite efforts to inhibit CPAF during lysate preparation. For example, we have found that clasto lactacystin is labile, and as a result, some batches did not completely inhibit CPAF activity (K.A. Johnson, J.K. Lee, C. Sütterlin and M. Tan, unpublished), which is consistent with the Valdivia laboratory experience (Snavely et al., 2014). We have also observed that fresh 8 M urea is required to inhibit CPAF and that this inhibition may be incomplete if the urea is not fresh or if its concentration is slightly lower (e.g. 7 M) (J.K. Lee, C. Sütterlin and M. Tan, unpublished).

In their recent Pathogens and Disease study, Snavely, Valdivia and colleagues reported partial cleavage of LAP1 and vimentin from Chlamydia infected cells. To block CPAF activity during lysate preparation, cells were directly lysed in hot 1% SDS buffer (Snavely et al., 2014). While this buffer was shown to inhibit the activity of recombinant CPAF, the actual cell lysates were not verified to be free of CPAF activity, leaving open the possibility that the proteolysis of these proteins occurred during lysate preparation rather than in intact cells.

The Snavely et al. (2014) study went one step further using an innovative live imaging approach to assay CPAF mediated proteolysis of specific proteins in infected cells. The authors observed relocalization of EGFP tagged vimentin and LAP1 upon rupture of the chlamydial inclusion. This altered distribution was attributed to cleavage of these host proteins by CPAF released from the inclusion. However, inclusion rupture appears to cause large overall morphologic changes in the infected cell, and the relocalization of vimentin and LAP1 may or may not be due to CPAF mediated cleavage. These experiments lack a critical negative control showing that a non CPAF substrate is not relocalized in a similar fashion upon inclusion rupture. Extra care and caution is called for in light of past missteps in CPAF target identification, and it may be premature to conclude that vimentin and LAP1 are in vivo CPAF targets.

Function

What is our current knowledge about CPAF substrates? It is clear that CPAF can cleave or degrade a large number of diverse substrates in vitro, including transcription factors, cytoskeletal proteins, cell cycle regulators, and chlamydial proteins (Chen et al., 2012). Nevertheless, CPAF shows substrate specificity because only a subset of host and chlamydial proteins are cleaved or degraded in vitro. Of the 16 host CPAF targets published by 2012, 10 appear to be in vitro substrates without evidence of CPAF mediated proteolysis in intact cells up to 48 hpi (Chen et al., 2012). There are conflicting data about whether or not there is cleavage of vimentin in intact infected cells (Chen et al., 2012; Snavely et al., 2014). The remaining five host proteins, and several chlamydial proteins reported to be CPAF substrates (Jorgensen et al., 2011; Hou et al., 2013), should be re examined with precautions to inhibit CPAF activity during lysate preparation. We speculate that the proteolysis of these published substrates may also be due to CPAF activity during lysate preparation, but the final word rests with the authors of the original reports who should confirm their findings or correct the published literature. With the currently available data, the relevance of an in vitro CPAF substrate to the chlamydial infection is unclear.

If CPAF is a protease with many specific in vitro substrates, why are these proteins not altered during most of the intracellular infection? Active CPAF is reported to be present starting at the mid stage of the developmental cycle, when the cleaved forms of the enzymatically active CPAF heterodimer can be detected (Chen et al., 2010). The paradox of an active protease without proteolysis of potential targets could be explained by mechanisms to regulate CPAF activity inside an infected cell. For example, CPAF could be sequestered into a specific cellular compartment that prevents contact with its substrates. Alternatively, CPAF activity could be inhibited by a host or chlamydial factor. Interestingly, host proteins that are in vitro CPAF substrates are also cleaved or degraded in uninfected cells overexpressing active CPAF (Paschen et al., 2008). It would be informative to examine these CPAF overexpressing cells with precautions to inhibit CPAF activity during lysate preparation. If proteolysis still occurs, it will indicate that uninfected cells overexpressing CPAF differ from infected cells by lacking mechanisms to regulate CPAF activity.

The newly described CPAF null mutant (Snavely et al., 2014) is a convenient tool for studying CPAF function and for avoiding artifactual CPAF effects during protein analysis. It provides a genetic means to assess whether CPAF is necessary for a phenotype of a Chlamydia infection, or the proteolysis of a protein, although it does not reveal whether the proteolysis occurs in vivo. It will be particularly useful for biochemical studies if a protein of interest undergoes CPAF mediated proteolysis during lysate preparation. However, the CPAF null mutant will not eliminate the need for precautions to prevent CPAF activity during lysate preparation when working with the wild type strain.

What then is the biological role of CPAF? The CPAF null mutant only had a threefold decrease in infectious progeny, indicating that this protease is not essential for a chlamydial infection in cell culture (Snavely et al., 2014). However, the Grieshaber group has proposed that CPAF is involved in centrosome amplification, early mitotic exit, and multinucleation because these phenotypes were reduced with the CPAF mutant (Brown et al., 2014). The broad range of potential CPAF targets might indicate that CPAF is used as a blunt instrument. For example, the proteolysis may provide amino acids and peptides for chlamydial growth or promote chlamydial exit by destroying its host cell through proteolysis of many structural proteins (Zhong, 2009). CPAF may also function as an extracellular protease upon release from a lysed infected cell (Conrad et al., 2013). Comparison of a CPAF null mutant with a wild type isolate in a mouse infection model will help reveal whether CPAF has a role in an infected animal that has not been apparent in a cell culture infection.

We started our own CPAF journey by discovering one CPAF substrate, which caused excitement, and a second, which produced elation, and then several more, which created concern that became despair. From this experience, we recognize CPAF as a potent but devilish protease whose functional analysis requires both caution and precautions.

References