Separation of Free and Bound cAMP in Mycobacteria

Abstract

Mycobacterial genomes encode a plethora of genes that are involved in the synthesis, utilization and degradation of cAMP. The genome of M. tuberculosis H37Rv, for example, encodes 16 adenylyl cyclases and 10 genes harbouring the cyclic nucleotide-binding (CNB) domain (Shenoy and Visweswariah, 2006). Cyclic AMP is efficiently secreted by mycobacteria, and cytosolic as well as extracellular levels of cAMP can reach hundreds of micromolar. We have recently reported that an abundantly expressed universal stress protein (USP; Rv1636 in M. tuberculosis H37Rv and MSMEG_3811 in M. smegmatis, respectively) binds cAMP (Banerjee et al., 2015). Given the number of cAMP-binding proteins present in mycobacteria, it is expected that a significant fraction of intracellular cAMP may be bound to protein. The methods typically employed to measure cAMP are radioimmunoassay (RIA) and ELISA. However, these procedures include prior acidification of samples that would dissociate cAMP ‘bound’ to protein, and therefore represent the ‘total’ cAMP present in the sample. In this protocol, we describe a method to separate the fraction of cAMP ‘bound’ to protein from what is ‘free’ or not associated with protein. This is performed by subjecting the cytosolic fraction or the culture supernatant to filtration through a membrane with a 3 kDa cut-off. Only ‘free’ cAMP is able to pass through the membrane. Therefore, cAMP concentrations in the filtrate represent the ‘free’ cAMP in the sample. Cyclic AMP levels in the original cytosolic fraction or the culture supernatant represent the ‘total’ cAMP concentration. Subtracting the ‘free’ from the ‘total’ provides the amount of cAMP bound to protein.