A family of 23 DHHC (Asp-His-His-Cys) proteins that function as mammalian S-acyltransferases has been identified, reinvigorating the study of protein S-acylation. Recent studies have continued to reveal how S-acylation affects target proteins, and have provided glimpses of how DHHC-substrate specificity might be achieved.
A wide variety of intracellular proteins are modified by S-acylation, the attachment of fatty acyl groups on to cysteine residues via thioester linkages . Palmitate is the major fatty acid attached to cysteine residues, but other fatty acids with varying degrees of chain length and unsaturation have also been reported to be attached to proteins in this manner [2–4]. Despite S-acylation being first observed over 30 years ago, it is only in the last decade that information has emerged regarding the identity of the proteins that mediate these acylation reactions. Initial studies in yeast [5,6] have paved the way for the subsequent identification and characterization of a family of 23 mammalian DHHC (Asp-His-His-Cys) proteins as putative S-acyltransferases . The defining feature of these proteins is the presence of a highly conserved ~60 amino acid DHHC CRD (cysteine-rich domain) .
Discovery of DHHC proteins as S-acyltransferases
The major breakthroughs in the search for S-acyltransferases came from studies in the budding yeast Saccharomyces cerevisiae. Work by the Deschenes group identified Erf2 and Erf4 as proteins that were essential for S-acylation of Ras2; Erf2 and Erf4 form a complex, and both proteins are required for S-acylation activity . At the same time, work from the Davis laboratory identified Akr1 as an S-acyltransferase that was active against the yeast casein kinase Yck2 . Erf2 contains a DHHC motif, whereas Akr1 has a DHYC motif within a CRD.
These studies served as a primer for work by the Bredt laboratory, who performed a large-scale analysis of the 23 DHHC proteins encoded by the mouse genome . This study and others have shown that the majority of mammalian DHHC proteins possess S-acylation activity against specific substrate proteins [9,10]. In all DHHC proteins examined to date, the DHHC domain is critical for S-acyltransferase activity.
DHHC membrane topology and intracellular localization
DHHC proteins are predicted to be polytopic membrane proteins with between four and six transmembrane domains . Importantly, the DHHC domain, which is thought to mediate S-acylation, is predicted to be present on the cytosolic face of the membrane. This predicted topology has been confirmed experimentally for the yeast DHHC protein, Akr1 .
DHHC proteins have been found to associate with various intracellular membrane compartments, with the majority of mammalian DHHC proteins present at the ER (endoplasmic reticulum) and Golgi . Interestingly, some DHHC proteins are associated with the plasma membrane and endosomal membranes [12,13]. In yeast, a similar distribution is observed, with the majority of DHHC proteins at the Golgi/ER and with others at later parts of the secretory pathway, including the yeast vacuole.
DHHC protein specificity
An intriguing question that arises about the DHHC protein family is why are there so many? For example, it has been shown that as many as 11 of the human DHHC family associate with Golgi membranes upon expression in HEK (human embryonic kidney) -293T cells . The most obvious answer to this question is that DHHC isoforms display substrate specificity. Indeed, co-expression studies in mammalian cells have shown that specific DHHC proteins acylate specific groups of proteins . In agreement, an elegant genetic study in yeast has demonstrated, using the ABE (acyl–biotin exchange) technique, that inactivation of specific DHHC proteins led to a loss of acylation of specific sets of substrates . The ABE method combines chemical deacylation of S-acylated proteins by treatment with neutral hydroxylamine and subsequent labelling of free cysteine residues with biotin to identify S-acylated proteins. This analysis clearly showed that certain substrates are dependent upon a specific DHHC protein for their acylation. For example, acylation of several yeast SNARE (soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor) proteins was reduced by mutation of the ER-localized DHHC protein, Swf1. In addition, there was evidence that individual DHHC proteins preferred a particular type of substrate: Swf1 targeted mainly transmembrane proteins with juxtamembrane cysteine residues, whereas most Akr1p substrates were soluble proteins .
How is DHHC specificity regulated? A recent study examined structural features of DHHC proteins that contribute to substrate acylation . Using the ABE method, El-Husseini and Hayden's groups showed that huntingtin was acylated by DHHC17, but not by DHHC3 . Interestingly, the authors reported that adding the ankyrin-repeat region found at the N-terminus of DHHC17 on to the N-terminus of DHHC3 allowed DHHC3 to bind to and acylate huntingtin to a similar level as wild-type DHHC17. Thus regions outwith the DHHC domain clearly contribute to substrate specificity. Given that DHHC proteins show little sequence conservation outwith their DHHC domains , these findings suggest that much work will be required to fully decipher the mechanisms that contribute to substrate specificity of the 23 mammalian DHHC proteins. Another important point of note is that, although the ankyrin-repeat region of DHHC17 allows DHHC3 to acylate a non-target protein, several proteins have been shown to be S-acylated by both DHHC3 and DHHC17 [16,17].
Two other reports have examined the specific features of substrate proteins that contribute to DHHC specificity. The SNARE protein SNAP25 (25 kDa synaptosome-associated protein) is acylated by DHHC3, DHHC7 and DHHC17 . The minimal membrane-targeting sequence of SNAP25 maps to residues 85–120, which contains a cysteine-rich cluster (the site of acylation) and the downstream 28 amino acids . Residues at the C-terminal end of this sequence (in particular Gln116 and Pro117) are important for efficient membrane interaction of SNAP25 in neuroendocrine cells . We demonstrated that the region of SNAP25 downstream of the acylated cysteine residues plays an important role in recognition of SNAP25 by DHHC17. In contrast, this region did not appear to be required for acylation of SNAP25 by either DHHC3 or DHHC7 [17,20].
Another study examining DHHC-substrate specificity of yeast proteins in vitro demonstrated that full-length Vac8 (a yeast vacuolar fusion protein) was acylated specifically by the DHHC protein Pfa3 when membranes purified from overexpressing cells were used as the source of DHHC protein . In contrast, the isolated SH4 (Src homology 4) domain from this protein, an 18-amino-acid sequence containing adjacent N-myristoylation and S-acylation sites, was acylated not only by Pfa3, but also by the other DHHC proteins Akr1, Erf2/4, Pfa4 and Pfa5. This loss of DHHC specificity was also observed for other SH4-containing yeast proteins: Meh1 was acylated by Akr1 and Pfa3, whereas the isolated SH4 domain from this protein was modified by Erf2/4, Pfa3, Pfa4 and Pfa5, but not by Akr1. This result was confirmed using purified proteins in the absence of membranes, although, in this assay, Akr1 was weakly active against the isolated SH4 domain. Thus, consistent with the study on SNAP25 discussed above, regions outwith the acylation sites of yeast substrate proteins also contribute to DHHC specificity.
Overall, the emerging view is that DHHC–substrate interactions probably involve interaction (acylation) of target cysteine residues in the substrate and the DHHC domain, as well as interactions between other sites/domains present in both substrate and DHHC protein.
Effects of S-acylation on substrate proteins
Over the last few years, it has become clear that S-acylation can have many different effects on the modified protein. For soluble proteins, S-acylation is often essential for stable membrane attachment; this is the case for the S-acylated H- and N-Ras isoforms [21,22]. Farnesylation of H/N-Ras, a process that occurs shortly after synthesis and in the cytosol, provides a weak membrane affinity that facilitates transient membrane interaction . This weak membrane affinity of the farnesyl group allows Ras isoforms to associate with endomembranes (ER/Golgi) where their partner DHHC protein(s) are located, and subsequent S-acylation leads to a dramatic increase in membrane affinity, mediating stable attachment of Ras and thereby facilitating trafficking to the plasma membrane [23–25]. The example of Ras proteins highlights an important point about S-acylation: as DHHC proteins are polytopic membrane proteins, substrates must contain additional primary membrane-targeting information that allows initial membrane interaction before acylation. For Ras proteins, this primary membrane-targeting information is provided by the farnesyl group, which is added to a cysteine residue just downstream of the acylation site. For some other S-acylated proteins [e.g. Gα subunits, eNOS (endothelial nitric oxide synthase)], the attachment of another acyl group, myristate, on to a glycine residues at position 2 of the amino acid sequence (following cleavage of the initiating methionine) mediates initial membrane attachment . Finally, work from our group has suggested that the hydrophobic CRDs of the multiply acylated substrates SNAP25 and CSP (cysteine-string protein) may have a weak membrane affinity that allows their interaction with membrane before palmitoylation [27,28].
An additional effect of S-acylation that has been well documented is the targeting of proteins to cholesterol-rich ‘raft’ microdomains; in this case, saturated lipid anchors are thought to readily partition into cholesterol-rich ordered domains . [3H]Palmitic acid-labelling experiments suggested that approx. 50% of proteins present within detergent-resistant fractions (which are proposed to contain raft proteins) could be labelled with [3H]palmitic acid 
Many transmembrane proteins (including polytopic membrane proteins) are S-acylated, most frequently at juxtamembrane cysteine residues. Modification of transmembrane proteins in this way often affects their intracellular sorting, perhaps via modulation of the topology of membrane-spanning domains, or via changes in protein–protein or protein–lipid interactions [31–33]. S-acylation has been reported to regulate various trafficking steps, including ER exit, Golgi exit, plasma membrane delivery, internalization, recycling, targeting to lysosomes and nuclear shuttling [27,31]. The effects of S-acylation on targeting are protein-specific and difficult to predict, underscoring the versatility of S-acylation as an intracellular sorting signal.
Synaptopathies: Dysfunction of Synaptic Function: A Biochemical Society Focused Meeting held at The Hotel Victoria, Newquay, U.K., 2–4 September 2009. Organized and Edited by Nils Brose (Max Planck Institute for Experimental Medicine, Göttingen, Germany), Vincent O'Connor (Southampton, U.K.) and Paul Skehel (Centre For Integrative Physiology, Edinburgh, U.K.)
Work in the authors' laboratory is funded by a Medical Research Council Senior Fellowship Award to L.H.C.