Before thinking for other strategies, you should make sure of the following:

- Your gene of interest is in frame with the a-factor signal sequence in your vector

- Your insert have been sequenced and there is no mutation introducing ''stop'' codon

- The clones you got are positives transformants

If you are clear from all the above, then try another P. pastoris strain like GS115

I agree that you have to confirm the structure of your construct by sequencing. However I do not think that changing a Pichia strain can help if the construct is correct and integrated without rearrangements. The protein you are trying to express is a normal protein for expression and secretion using Pichia system. However in some cases a very few selected transformants expresses a heterologous protein. A plate X-gal assay is a good detection method for the enzyme. You could try to check as many transformants as you can by plating them onto MeOH X-Gal. Perhaps you would able to find some blue colonies.

Another approach is the changing of expression method from inducible pPICZ series with AOX1 promoter to the constitutive one using pGAP series or such. In this case you could combine selection of transformants with activity assay by supplementation of the transformation plates with x-Gal. If the constitutive expression would work, the AOX1p driven expression should work as well.