The peak sharpness could be improved by reducing the sample volume. However, from your description it sounds like there is something wrong with the column packing, since you see a distorted band shape. The column may need to be repacked. If it hasn't had a good cleaning lately, that would be a good place to start. There may be bubbles in it. The increase in conductivity at the end of the elution indicates the presence of some ionic substance in the sample, suggesting the sample wasn't fully equilibrated with column buffer.

I think there is some non-specific interaction happening between protein molecules and resin. that's why you see peak broadening and tailing.  But, that is often the case with silica resin, I am not sure about Dextran. Try increasing the salt concentration and see whether you get any improvement.

The high conductivity at the end of the chromatogram is due to salt/buffer ions which are eluting out latest because of their smallest size. If your salt is a neutral one and if the buffer strength is less then pH will tend to move towards neutral side and you will see shift in pH. This shift in pH would also alter the charge on protein surface and non-specific interactions with gel matrix may arise.

My suggestion is you increase the buffer strength (I guess you are using Citrate or Acetate buffer, so try increasing its molarity) to get rid of that "change in pH" issue.And you increase the salt concentration to overcome non-specific interactions with resin. (Don't worry about high conductivity issue, that's not any problem)

Since your protein is aggregation prone, Na-Acetate buffer with NaCl as a salt would give lesser aggregation than Na-Citrate buffer with any salt. Keep the mobile phase pH well away from protein's pI.

If all these factors don't work out, then probably your column life is over.

I would be interested in knowing your results. 

Thank you Adam and Rahul for your quick responses!

I think the gelbed is damaged, running bromphenolblue shows a heavily distorted band, which explains the tailing for the proteins - seems like I need a new column.

As the chromatogram shows, I am having a problem with dimerization - I am losing a huge portion of my protein in the form of dimers... (middle peak). What can I do to prevent dimerization and aggregation and improve monomeric yield?

I am using in my buffer 0,1 M citrate and 0,2 M arginine for aggregation prevention plus 50 mM potassium chloride - I was told that sodium chloride leads to aggregation for my protein (prion protein), therefore KCl.

How can I optimize that buffer to get more monomeric protein?

Thanks for any help in advance!

Use of Arginine is a good thing you are doing. 

See if you can replace Citrate buffer with some less hydrophobic species, like acetate or so, that can reduce aggregation propensity of molecules. Rest, I am not the expert in prion biophysical characteristics, so can't help you much.

But, Are you sure that this mobile phase is causing dimerization?! Did you check for the amount of dimers present in your sample before doing SEC. If you already have dimers in starting material, you need to optimize your affinity purification and upstream processes. You need not spend your time in optimizing SEC mobile phase. Please check.

Several mutually-reinforcing clues in your image and description: The 45° tilt you describe (presumably referring to the blue dextran band visible through the glass cylinder) indicates either sample viscosity is too much greater than that of the running buffer, or there is precipitated/clogging protein or debris/bacteria from previous runs asymmetrically obstructing the topmost region of the column, or poor column packing (least probable since this column was sold professionally pre-packed).  The Blue dextran peak should not have a shoulder; the same for the salt conductance peak at the end.  Your peak of interest is behaving like these, but since in the better MW resolving region of the chromatogram, the trailing shoulder is more elongated.  Note that sample viscosity is increased by protein concentration as well as small solutes. Protein precipitates in the column are effectively removed by running in 2 column volumes of proteinase K ≥ 10 µg/ml in 1% SDS, EDTA, Tris buffer pH 9 at 22°, 0.25 M NaCl then incubating at 50°C for an hour.  Wash out completely with buffered salt solution. Destroy any residual proteinase K by running 1.5 vol 0.05% DEPC (diethyl pyrocarbonate) in 0.2% NaH2PO4.  Store in 20% EtOH.  This has rescued my columns (and others') several times.  Manufacturer-suggested extensive washing protocols failed to remove clogging debris.

Prions trends to form aggregates. please check also for the Ni-NTA ligands exposition. May, it has been covered by others amino acids and then the affinity will not be performed as expected, resulting in poor purification, eluiting along the column.  As suggested before, preventing aggregation and precipitation combining to column package, the purification will be better. Good luck. 

The conductivity/pH change at the end of the chromatography is straight forward since the small sample buffer ions (elute lastly) are different from the equilibration buffer ions in type and strength.  

Do you have MW standards (not just dextran) to spike into your sample as internal controls and assure column performance?  Did you do material balance to be sure all the prion injected is also the amount recovered?  Does SDS-PAGE of your SEC fractions actually show covalent dimers or is there DLS showing oligomers?  Could the putative dimer peak actually be higher order oligomers?  

The questions are directed at what may be happening to the prion and therefore the SEC behavior.  The single disulfide may be susceptible to exchange when subjected to slightly acidic conditions and thereby generate the scrapie prion form leading to dimers.  Slightly acidic conditions transiently form as your sample buffer is titrated by the column buffer.  The transient slight acidity could change the PrPc form to make the disulfide exposed as well as exposing hydrophobic residues along with redox changes resulting in b-sheet (scrapie PrPsc) formation.  The transition could allow for multiple stokes radius and/or column interactions thus creating an envelope of elution of many conformers/oligomers that are in transition in situ.  There may not be any fibrils or large insoluble aggregates seen because they are hung up at the head of the column and therefore material balance needs to be demonstrated.  The scapie form is resistant to proteinase K and low concentrations of SDS with NaCl is used to generate fibrils from PrPc.  You might therefore avoid salts, use a different SEC pH or low concentrations of a chaotrope with low concentrations of TCEP to inhibit PrPsc formation which is required to go to the prion dimer form.