Start from the cause and work down to the surface proteins. PNH begins as a mutation in the $PIGA$ gene inside a haematopoietic stem cell. The job of $PIGA$ is to build the GPI anchor, a greasy tether that pins certain proteins to the outer membrane of the blood cells. When the anchor cannot be assembled, none of the anchor-dependent proteins can stay attached, so they all drop off together. Among these proteins are the two key complement brakes, CD55 (decay accelerating factor) and CD59 (membrane inhibitor of reactive lysis). Their normal role is to stop the complement system from punching holes in the red cell. Strip them away and complement runs unchecked, lysing the red cells inside the vessels and spilling free haemoglobin into the urine, classically at night. Now look at the options. Yes, CD55 and CD59 are individually deficient, but each one is only missing as a downstream consequence of the absent anchor. The choice that captures the true upstream lesion, and therefore both deficiencies in one stroke, is the deficiency of the GPI anchored protein. LFA is simply an adhesion molecule on lymphocytes and is irrelevant to the complement-mediated lysis of PNH, so it is discarded.
\[\boxed{\text{GPI anchored protein}}\]