What exactly is your question?

If you are asking how to move VL and VH independently in a molecular graphics/molecular modelling program - simply create different copies of the relevant parts of your scFv object as independent objects.

e.g. in PyMOL:

fetch 8DGR (example of a scFv structure)

create VH, /8DGR///3-128 (creates a copy of the VH domain)

create VL, /8DGR///139-249 (creates a copy of the VL domain)

you can now move these around:

color pink, VL (give it a different color to distinguish it from the scFv)

translate [20,0,0], VL

color slate, VH

translate [-20,0,0], VH

more likely, you will use a least-square fit method to superimpose them onto a template structure.

You can also just use a text editor to open a pdb file and copy-paste individual parts into separate files that you can can then load as separate objects into your molecular graphics program.

The principle is the same for other molecular graphics programs, e.g. Chimera, Maestro, etc -> You create individual objects of the fragments you want to move around, bring them into position, and once you are satisfied, you can again merge them into a single object. Refer to the manual of the specific program you are using to learn how to do so.

More likely, you will move around the scFv as a whole, since you do not want to disturb the relative orintation of the VL and VH domain, e.g. place three different scFvs against their relative epitopes on the antigen structure to determine how long a linker you need to connect them into a single protein chain

If you are asking how to combine three scFv into a trivalent construct, there are different possible answers:

If you shorten the linker within an scFv in such a way that the VL-VH interface can no longer be formed within a single protein chain, you get formation of diabodies or, in an extreme case, even trimerization through the formation of intermolecular VH-VL interactions (see https://www.rcsb.org/structure/1NQB for an example, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC23241/)

You can also connect multiple scFvs by linkers, like pearls on a string, e.g. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5347777/

Another possibility is to add a stable trimerization motif to each scFv, e.g. https://www.nature.com/articles/s41467-017-02707-6 (see figure 5)

In constructing such multivalent, and in many application, multispecific constructs, it may be necessary to optimize the sequences of the individual scFvs to ensure that each is sufficiently stable and folds well. You also may need to engineer the VH-VL interfaces to prevent mispairing in multispecific constructs.

Take scFv is a single molecule with no separate chains. I need to move to VH or VL in specific direction that it can form multivalent structures when docked. Like < shape. If we make different copies of VH and VL and then move one V region. Then how to join these two parts in separate copies? We want to change the orientation of ScFv from - to

I understand that, but you anyhow have to alter the linker-length and sequence in going from the scFV to the tribody. The linkers are flexible, frequently not even visible in experimental structures

Normally, when designing an antibody-derived construct. you want to preserve the original binding specificity and affinity. The antigen recognition surface (paratope) of a scFV, like that of an antibody, is formed by the combination of VL and VH in a specific relative orientation. Usually, if you isolate the individual VL and VH domains, you lose most of the binding affinity. If you want a single domain binder, you would use a VHH domain (derived from camelid antibodies or artificial library).

In going from a single scFv to a tribody modelled on structure 1NQB, biological assembly 2 or 3, you superimpose three instances of your scFv/antigen complex structure on the three VL/VH pairs in the template structure, then you only need to delete the linker in the scFv and instead connect the VL of one scFv to the VH in the next one by the short linker connecting the domains in the tribody template

You can then use a homology modelling program to join the VH domain of on scFv through your chosen linker sequence to the VL domain of the next scFv, by providing the appropriate domains in the correct relative orientation as modelling templates.

(red, yellow, blue: VH-linker-VL pairs in 1NQB, cyan: VH-long linker-VL pairs in scFv 8WFH, green: Antigen bound by scFv 8WHF)