SPAC458.04c single sequence analysis indicates its similarity to the DLIC definition PF05783.2

Sensitive profile based search tools show indications for a distant sequence similarity of S. pombe SPAC458.04c (fa) to the DLIC protein family (note: searches performed 2008; DLIC protein family version: PF05783.2):

To allow more sensitive searches and in order to obtain a better representation of SPAC458.04c sequence features in the searches, we sought close sequence relatives of SPAC458.04c. With the use of a Schizosaccharomyces octosporus ortholog of SPAC458.04c the DLIC similarity is further substantiated as illustrated in the following.

Use of orthologous sequences to SPAC458.04c confirms the DLIC borderline similarity

The protein sequence of the SPAC458.04c ortholog in Schizosaccharomyces octosporus and Schizosaccharomyces sp. OY26 (fasta) can be obtained using augustus genewise on the respective genomic sequence (NCBI: ACQJ01000054.1, ABHY02000001.1).

An alignment of the three Schizosaccharomyces SPAC458.04c proteins (a2m) shows a significant similarity to the DLIC domain using hhpred (E=4.6E-05, hhpred_no_blast).

Consistently a PSI-BLAST search started with the S. octosporus SPAC458.04c ortholog against the NCBI non-redundant protein sequence database, converged in round 11 and collects dynein light intermediate chain (DLIC) proteins from a broad range of taxa.

Similarly- if S. pombe SPAC458.04c is used in a psi-blast search against the NCBI non-redundant protein sequence database supplemented by the S. octosporus SPAC458.04c, the search converges in round 11 and collects dynein light intermediate chain (DLIC) proteins.

Note: Psi-blast version 2.2.17 with default settings and protein non-redundant database from December 2007 have been used if not specified otherwise.

If S. octosporus SPAC458.04c is applied to the same sensitive profile based search tools as used above in the analysis of SPAC458.04c a highly significant similarity of SPAC458.04c to DLIc is reported. Note that HHPRED and FFAS03 use a step of PSI-BLAST based homolog collection before the actual profile-profile and PSSM-PSSM comparison (note: searches performed 2008; DLIC protein family version: PF05783.2).

Transitive reciprocal-best searches

A taxonomically representative set of fungi/metazoan DLIC homologs has been selected and further analysed.
Blast relationships between the sequence of the selected set have been analysed using CLANS and the obtained graph shows that metazoan LIC-3 like DLIC family members seem to be a more divergent sequence sub-family within the set. (DYNC1LI1\|NP_057225 and DYNC1LI2\|NP_006132 are paralogs in vertebrates- more recent duplication; DYNC2LI1=LIC3 is more distantly related).

In the following using psi-blast it can be shown that the selected fungal DLIC members and pombe SPAC458.04c are each others closest homologs in the respective species.

Reciprocal blast searches illustrate that human DLIC 1/2 family members and Nc, Gz, An DLIC members are each others closest homologs in the respective species. Note that the LIC-3 subfamily is not found in the direct blastp against the human NCBI nr subset.

Similarity to P-loop containing nucleoside triphosphate hydrolases

2010 release of Pfam has the DLIC domain assigned to the AAA clan. This family consists of several eukaryotic dynein light intermediate chain proteins. from Pfam: The light intermediate chains (LICs) of cytoplasmic dynein consist of multiple isoforms, which undergo post-translational modification to produce a large number of species. DLIC1 is known to be involved in assembly, organisation, and function of centrosomes and mitotic spindles when bound to pericentrin. DLIC2 is a subunit of cytoplasmic dynein 2 that may play a role in maintaining Golgi organisation by binding cytoplasmic dynein 2 to its Golgi-associated cargo.

Additional evidence is shown in the following:
Similarity to the superfamily of P-loop containing nucleoside triphosphate hydrolases (SCOP c.37.1) has been noted in early studies of DLIC proteins. PSI-BLAST searches as well as various fold prediction indeed suggest that DLIC proteins belong to the structurally defined family of G proteins.

Independent of the above automated meta-analyses we also performed controlled HHpred searches with a restricted number of psi-blast rounds to ensure that only DLIC family members are applied in the distant homology search.
Searches with a DLIC family model against the database for Structural Classification of Proteins (SCOP) indicated that DLIC belong to the structural family of G proteins/c.37.1.8.
Also DLIC show closest similarity to the Rab subfamilies (cd00154) of the Ras_like GTPase domain family (cd00882) when searches are performed against the hierarchically organized Conserved Domain Database. Particulatly notable among the top Rab-subfamily models is cd04128/Spg1/Tem1 GTPases- a domain found in S. pombe Spg1p (Sid3) and S. cerevisiae TEM1.
Notably, in searches against the pombe proteome using LIC-3 family members Spg1p can appear as a top borderline blast hit: blast versus pombe proteome. Psi-blast searches against nr also tend to collect Spg1 among top rounds Ce Dm Dr Mm Hs. Due to the lack of additional support at present it can not be decided if the similarity of DLIC (in particular LIC-3) to Spg1p is a simple similarity between Rab subfamily members or conclusions can go further.