in general, concentrating by pelleting (through a sucrose cushion) is harder on aggregates than is gradient separation, but gradient separation is more finicky and takes some extra steps to identify the correct layer to target. the only way to know for sure is to try. I would start with pelleting, as it is much easier, followed by antibody incubation, and test that process for EM. depending on who will prepare your sample for EM, they can be a good resource.

Information about the density of your virus or in where it forms band on sucrose gradient is useful. Incubating your virus with antibodies may for aggregates (depending on the concentration of virus and antibodies), which may not behave similar to virus alone during ultra centrifugation. First do ultracentrifugation then incubate with antibodies.

if you have a good antibody, why don't you try protein A beads for virus capture?

For negative staining TEM the min concentration is about 106 virions per ml.

Purification is important too: there are a lot of cell fragments and macromolecules in virus solution.

Ultracentrifugation with a gradient of density is a powerful and well-known old method of virology. But it is necessary to have some practical skills and this process isn't simple in the practice.

There are an other way: special tubes - concentrators. For example:

http://www.vivaproducts.com/vivaspin-centrifugal-concentrators/vivaspin-6.html

I have used such tubes with MWCO 300 kDa for some viruses and bacteriophages with the great results. It is not only concentration process, some free molecules are going out via membrane and the solution will be more pure.  

Good luck!

Thank you for advice, I think I will try concentrator tubes for now. 

I thing it would be better to purify and concentrate your virus first. In my case, I've used continue sucrose gradient and ultracentrifugation for viral purification, and it worked perfect to me. You just need to find out the best min-max concentration of sucrose for you virus in particular. I have purified coronavirus, for example, and I used a gradient from 30 to 55%. Another suggestion would be that, if you want to be sure that you were successful during the purification, to asses the presence of the virus. In my case, I performed a western blot using polyclonal antibodies. If you want more details about the purification process I used, you can contact me to my personal e-mail, [email protected].

Depending on the viruses some authors use sucrose gradient as was explained , the cesium chloride gradient, or the cushion of Glycerol .

Use a virus suspension well clarified and high titer .

I agree with the other:

1. pelleting of the Virus from the cleared supernatant (this is different between all the viruses)

2. dissolve the pellet in a buffer and Label it by antibodies (mainly gold bound)

This shall work. I do not know if you have to fixate the pellet wich may change the epitope where the antibody may bind.