212 Ebola messages in my mailbox...

[Ulrich Melcher]

The hypothesis that filoviruses have natural reservoirs in plants seems 
far-fetched.

True, the tospoviruses seem to have arisen from vertebrate bunyaviridae 
and appear to "live" in both plants and insects.  And yes, many 
vertebrate viruses also "live' in insects.  But there is not yet known a 
virus that can exist in all three, vertebrates insects and plants.

Yet, stranger things have happened.  I therefore examined the 
far-fetched possibility further.  The sequence searching that I have 
done (detailed below for those interested) suggests that a natural 
reservoir for Ebola in plants is highly unlikely.

The gene that distinguishes the tospoviruses from other bunyaviridae is 
a gene required for cell-to-cell movement of infections in plants.  The 
known plant viral movement proteins fall, by sequence relationships, 
into three classes.  The tospoviral movement proteins belong to the 30K 
superfamily which includes proteins from bromo-, cucumo-, diantho-, 
furo-, alfamo-, ilar-, caulimo-, badna-, gemini-, tobra-, tobamo-, 
como-, nepo-, tombus-, capillo-, tricho-, idaeo-, and enamoviruses.  I 
have generated consensus sequences for the 30K superfamily which in 
Blast, Blitz and FASTA searches recognize all of these groups (ms in 
preparation).  In these searches no Ebola virus protein was recognized 
though these proteins were in the databanks at the time of search.

In order to detect more distant relationships, the 19 Ebola virus 
proteins (obviously redundant since there are only 7 encoded) obtainable 
from the databanks were searched by a FASTA algorithm with the consensus 
sequences.   The highest similarity (quite distant) was to the VP24.  
Since the 30K superfamily proteins have a characteristic pattern of 
predicted secondary structure primarily involving beta structure (ms in 
prep), the VP24 was submitted to PredictProtein for secondary structure 
prediction.  In the region of similarity with 30K superfamily proteins, 
alpha helical segments were strongly predicted.  Thus it is highly 
likely that the low sequence similarity is coincidental and meaningless.

The other two known movement protein classes are those encoded by 
tymoviruses and those of the triple gene block family of movement 
proteins.  The movement protein of eggplant mosaic virus tymovirus was 
used in a FASTA search of the Ebola proteins.  Several hits were 
obtained, but inspection suggested that they were less strong than the 
30K-VP24 one discussed above.  A similar search using a consensus for 
the central region of the largest of the triple gene block proteins of 
carla- and potexviruses gave a reasonable hit with VP24.  However, 
residues conserved among the plant viral sequences (LDEY) were not 
present in VP24.  In addition, searches with the complete sequences of 
potato X potexvirus and potato M carlavirus proteins failed to recognize 
VP24.

Not all plant viruses have their movement functions identified.  To 
check for other plant viral proteins with sufficient similarity to one 
of the Ebola proteins to be a movement protein candidate, each of the 
Ebola proteins was used in a Blast search of the non-redundant protein 
databanks.  No plant viral sequences were retrieved for VP24, VP30, VP40 
and VP35.  No plant viral sequences were among the top 200 hits with VGP 
and VNUC.  The bulk of the sequence of the L polymerase protein of Ebola 
virus was not available for the search.  The L protein sequence of 
Marburg was thus used.  The polymerase (L protein) of sonchus yellow net 
rhabdovirus was identified.  Among the rhabdoviruses, it had the highest 
score.  This is probably unimportant since much higher scores were 
obtained with the polymerases of a variety of paramyxoviridae.  Further, 
the "extra" protein of SYNV, and thus possibly the movement protein, 
failed in a FASTA search of the Ebola proteins to register any 
similarity.

Thus, there is no likely relation between any Ebola protein and a known 
movement protein, or with an unknown but sequenced movement protein.  
The possibility remains that movement proteins of as yet unsequenced 
plant viral groups may be related.  There are also some viruses (for 
example, luteoviruses) that because of their limitation to phloem do not 
require any movement protein function.

Ulrich Melcher
Dept Biochem & Mol Biol
Oklahoma State Univ.