Endogenous DAAM1 was shown to localise to actin and the Golgi apparatus in both COS 7 and HeLa cells. Analysis of DAAM1 "regulatory" N-ter (residues 1-545) and
"catalytic" C-ter (residues 526-1078) was carried out to assess the impact of these domains on formin localisation and its functions. The N-ter was clearly responsible for DAAM1 localisation to the actin stress fibres (SFs) and the Golgi (in both cell types) while the C-ter was responsible for generation and localisation to the phase- dark patches that are seen with other active formins (FHOD2 and mDia1) in HeLa cells. A schematic diagram of the truncation constructs and their acronyms is included here (Figure 4.1) for reference.
Figure 4.1 DAAM1 truncation constructs
1078
1 45 233 433
527 595
600 986
DAD 1031-1045
235 435
GBD FH3 CC FH1 FH2
Full Length
1-545 1-233 45-233 1-280 1-350 1-356 1-440
135-440 100-350 234-530 45-440
1-1045 1-1030 1-990 45-1078 234-1078 234-1045
526-1078 594-1078 526-595 1027-1078 234-1030
GBD FH3 CC
GBD GBD
GBD FH3
FH3 BD
GBD
FH3 CC GBD
GBD
GBD FH3
GBD
FH2
FH1 FH2
FH1
FH3 CC FH1 FH2
FH3 CC FH1 FH2
FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
1078
1 45 233 433
527 595
600 986
DAD 1031-1045
235 435
GBD FH3 CC FH1 FH2
1078
1 45 233 433
527 595
600 986
DAD 1031-1045
235 435
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
Full Length
1-545 1-233 45-233 1-280 1-350 1-356 1-440
135-440 100-350 234-530 45-440
1-1045 1-1030 1-990 45-1078 234-1078 234-1045
526-1078 594-1078 526-595 1027-1078 234-1030
GBD FH3 CC
GBD GBD
GBD FH3
FH3 BD
GBD
FH3 CC GBD
GBD
GBD FH3
GBD
FH2
FH1 FH2
FH1
FH3 CC FH1 FH2
FH3 CC FH1 FH2
FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
Full Length
1-545 1-233 45-233 1-280 1-350 1-356 1-440
135-440 100-350 234-530 45-440
1-1045 1-1030 1-990 45-1078 234-1078 234-1045
526-1078 594-1078 526-595 1027-1078 234-1030
GBD FH3 CC
GBD FH3 CC
GBD GBD GBD
GBD FH3
GBD FH3
FH3 BD FH3FH3 BD
GBD GBD
FH3 CC FH3 CC GBD
GBD GBD GBD
GBD FH3
GBD FH3
GBD GBD
FH2 FH2
FH1 FH2
FH1 FH2
FH1 FH1
FH3 CC FH1 FH2
FH3 CC FH1 FH2
FH3 CC FH1 FH2
FH3 CC FH1 FH2
FH3 CC FH1 FH2
FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
GBD FH3 CC FH1 FH2
Chapter 4: Discussion
4.1.1. DAAM1 localisation to actin stress fibres
Localisation of DAAM1 to actin SF is mediated by the N-ter domain(s). Since the N- ter contains the GBD, FH3 and the predicted coiled-coil regions, further truncations were designed with these domains in mind to investigate their contribution to DAAM1 localisation. The NCBI conserved domain database defines DAAM1 (45- 233) as the GTPase binding domain (GBD). In COS 7 cells, the region including residues 1-233 localises exclusively to the Golgi but not to the actin SFs. This suggested that the residues downstream of the GBD were required for actin SF localisation, hence several additional constructs were generated to delineate the SF binding domain.
Like 1-233, the 1-280 construct localises to the Golgi. The 1-440 construct (contains the intact GBD and FH3) localises to both the Golgi and actin SFs. Both the 1-350 and 1-356 constructs showed SF localisation, though perhaps the 1-350 showed weaker SF localisation than 1-356. This implies that certain residues within the FH3 domain can be critical for actin localisation. The FH3 is a relatively unstudied domain although two studies have found it critical for regulating cellular localisation [17, 18].
The NCBI conserved domain database suggests DAAM1 FH3 domain encompasses residues 235-433. Alignment of the DAAM1 FH3 domain with the consensus sequence determined by Petersen et al. [17] designated the FH3 to be consisting of three blocks from residues 137-165, 190-220 and 404-435 respectively. The predicted coiled-coil regions in DAAM1 N-ter are 108-143 and 437-522. Our results showed that the first two FH3 “blocks” together with the downstream sequences until residue 350 are necessary for actin SF localisation. In addition, when the truncation constructs
Chapter 4: Discussion
135-440 (with part of GBD deleted) and 234-530 (with entire GBD deleted) were expressed in COS 7 cells, actin SF localisation was absent despite presence of the intact FH3 and coiled-coil regions. The conclusion is that actin SF localisation requires, in addition to FH3 and coiled-coil regions, the region that resides N-terminal to the FH3 domain. As discussed below, I believe the SF association is driven by binding to polymerised myosin II. This is difficult to assess biochemically.
4.1.2. DAAM1 localisation to the Golgi
Concurrent with the search for the region responsible for actin SF localisation, the truncation constructs were assessed for Golgi co-localisation. The constructs that localised to Golgi included DAAM1 1-233, 1-280, 1-350, 1-356, and 1-440. The constructs that did not localise to Golgi included DAAM1 45-233, 135-440 and 234- 530. Although these three constructs did not show complete colocalisation to the Golgi, they were still found in the perinuclear region (i.e. in the vicinity of the Golgi).
Thus residues that are immediately before the GBD (residues 1-44) are critical for targeting DAAM1 to Golgi membranes. The Golgi apparatus is made up of three compartments: cis, medial and trans, defined by the enzymes present and their functions in vesicular trafficking [101, 102]. It is possible that these residues mediate the interaction of DAAM1 with certain Golgi components e.g. lipids or proteins that reside in the Golgi membranes.
4.1.3. DAAM1: To Golgi or actin stress fibres?
An interesting trend observed was that the apparent avidity of DAAM1 for actin SFs increased as I included more N-ter sequence. 1-233 and 1-280 did not localise to actin,
Chapter 4: Discussion
1-350 showed weak localisation while 1-356 and 1-440 showed strong localisation to actin SFs. All the above-mentioned constructs localised to the Golgi. Indeed DAAM1 immuno-staining was found in two different cellular pools (Figure 3.7). Further analysis by amino acid substitution may allow us to better define critical regions or their interaction and thereby allow specific targeting of DAAM1 to each of its cellular locations. The potential cycling of DAAM1 between the Golgi apparatus and actin SF might indeed underlie its ability to promote re-orientation of the centrosome and Golgi during cell migration.
4.1.4. DAAM1 localisation with a subset of myosin II fibres
In trying to define regions required for actin SF and Golgi localisation, the truncation construct 100-350 was seen to give a peculiar phenotype when transfected into both COS 7 and HeLa cells. Thick “amyloid like” fibrils were generated that were myosin II-rich but lacked actin itself (Figure 3.16a). The 100-350 construct has two “blocks”
of FH3: this region appears to have the propensity to self assemble and induce the co- assembly of myosin II fibrils without seriously perturbing the remaining actin/myosin stress fibres (Figure 3.16b). This effect was also observed with ectopically expressed coiled-coil domains of myosin IIA or IIB which supports the notion that it is co- assembly of coiled-coil domains. Based on the inability of MRCK coiled-coil 2/3 (which could itself also assemble into thick fibrils) to form fibrils with DAAM1 (100- 350), I suggest that myosin II is the target that allows DAAM1 to associate with actin SFs. These results also provide us with the information that the region 100-350 has the propensity to aggregate, and can be implicated in perhaps the multimerisation of DAAM1 with itself or other interacting partners.