Is it allowed to ask questions here about verifying a solution of a lemma from my preprint?

Firstly, I appreciate you asked here first, rather than just advertising your work on main and getting rightly closed. Note that if you think you have solved a famous open problem in number theory, using "basic equations, summations, integrals" the odds are that you have a mistake somewhere, but I'm going to leave that aside. If you think that you have a single lemma that you are not confident in, it's possible that that one lemma is in fact equivalent to the conjecture.

That said, here is general advice that will be applicable to others in the future.

1. Don't ask if the proof of your lemma is correct, but isolate the first part of the proof that you are not confident in.
2. Figure out how to frame a question about just that part, without reference to anything outside the lemma, and ideally, anything other than that one problem. If this is truly a lemma, you won't need to introduce the whole conjecture and baggage that comes with it.
3. Sit on the question for a week without asking it, and at the end, come back and rework it to be as clear as possible, without reference to the rest of the paper and ideally without reference to the overarching lemma.
4. Ask the question, and give enough context so that people can understand what you know about the type of mathematics and if you have an idea about what the answer might be. If you have correctly isolated the step you don't know how to do, you are not going to ask "is this proof correct?", because you will have no idea about how to prove that step.
5. If someone answers "oh, that is known to be as hard as Conjecture X" (the topic of your paper), then congratulations! you have found a statement that no mathematician who's ever thought about Conjecture X knows how to prove, and you are among them.
6. If someone asks for more context, don't claim you have proved Conjecture X and you "just need this one step checked"! Because you do not know how to prove this step and thus do not know how to prove the main conjecture.
7. If someone answers to their and your satisfaction, then congratulations, you have used MathOverflow responsibly and hopefully have learned something. At this point, go back to step 1.

You really have to go into this exercise assuming that there is a mistake in your claimed proof, if the conjecture is famous. The trick is to find it, and learn what you can from the exercise. And then, ideally, manage to recognise how what you have written does still prove a weaker result or a special case that's interesting. This weaker result may be known or even weaker than a known theorem. But you can take pride in knowing that you independently came up with it, if you worked on this in isolation.

Please note however, that asking and answering your own question is only possible in the above plan if you ask it, no one answers for a bit, and after a while you realise how to answer your own question. MathOverflow is not like Stack Exchange where it's encouraged to ask your own question and then immediately answer it, as a means of sharing knowledge. This is ok for things like "I figured out this cool coding trick that is really useful, and people will appreciate knowing it", but not for "I need to prove this one technical point in one lemma, and mathematics will be enriched for more people knowing it". The point is to ask the thing you don't know, and see what people answer, without being prejudiced as to just say yes/no to "is this correct". Perhaps they use a technique new to you, and you will be enriched.