Please resequence your DNA from other company. And if you are facing  same problem then use submit only the sequence which gives functional product

I have already resequenced in another company. The heteroduplex is not a sequencing error. It is probably that in this isolate there are one allele of each or there is a mix of isolates. But I want to know if someone has uploaded a sequence in the GenBank with a heteroduplex in a position of the nucleotide sequence. Or you can upload only the mutated sequence and then in the paper specify that there is a heteroduplex.

I think you mean to say 'heterozygote"  ie. a biallelic nucleotide variant, rather than heteroduplex.

When you submit to GenBank, there is a step where you can annotate the sequence to describe the variant position.

Well, yes, but I can't ensure if the strain is heterozygous for this gene or there is a mix of two homozygotes strains, one of them showing both wild type alleles and the other one two mutated alleles.

submit your sequence in two files. 

First submit the sequence which is giving functional wild type protein and then submit which produces a mutant one.

And when you submit it second time then please specify that it is variant of first sequence.

By this you can use both the accession numbers for you publication and also publish one article on the variants of the same sequence.

You must seperately upload the sequences. Either the wild type first followed by the mutated one or vice versa. You can not upload the 2 simultaneously.

Hi Mar,  you can upload sequences to GenBank using standard IUPAC notation for the heterozygous sites. I have done it not so long ago for sequences obtained from PCR products of a diploid genome (exactly the same as you, i think). Check here an example: http://www.ncbi.nlm.nih.gov/nuccore/KF290402

Thanks you very much for all your answers which have been really helpful!