CD38/NAD+ glycohydrolase is a type II transmembrane glycoprotein widely used to study T- and B-cell activation and differentiation. CD38 is endowed with two different activities: it is a signal transduction molecule and an ectoenzyme that converts NAD+ into ADP-ribose (NAD+ glycohydrolase activity) and small proportions of cADP-ribose (cADPR; ADP-ribosyl cyclase activity), a calcium-mobilizing metabolite, which, ultimately, can also be hydrolysed (cADPR hydrolase activity). The relationship between these two properties, and strikingly the requirement for signalling in the formation of free or enzyme-complexed cADPR, is still ill-defined. In the present study we wanted to test whether the CD38–cADPR complex is kinetically competent in the conversion of NAD+ into the reaction product ADP-ribose. In principle, such a complex could be invoked for cross-talk, via conformational changes, with neighbouring partner(s) of CD38 thus triggering the signalling phenomena. Analysis of the kinetic parameters measured for the CD38/NAD+ glycohydrolase-catalysed hydrolysis of 2′-deoxy-2′-aminoribo-NAD+ and ADP-cyclo[N1,C1′]-2′-deoxy-2′-aminoribose (slowly hydrolysable analogues of NAD+ and cADPR respectively) ruled out that the CD38–cADPR complex can accumulate under steady-state conditions. This was borne out by simulation of the prevalent kinetic mechanism of CD38, which involve the partitioning of a common E·ADP-ribosyl intermediate in the formation of the enzyme-catalysed reaction products. Using this mechanism, microscopic rate conditions were found which transform a NAD+ glycohydrolase into an ADP-ribosyl cyclase. Altogether, the present work shows that if the cross-talk with a partner depends on a conformational change of CD38, this is most probably not attributable to the formation of the CD38–cADPR complex. In line with recent results on the conformational change triggered by CD38 ligands [Berthelier, Laboureau, Boulla, Schuber and Deterre (2000) Eur. J. Biochem. 267, 3056–3064], we believe that the Michaelis CD38–NAD+ complex could play such a role instead.
Read full abstract