This is for metagenomics and we're not assembling the reads yet, correct? Amplicon data? You didn't tell us. There's no need to assemble the reads yet, you are just looking to mate the paired-end sequences from your library? I think the terminology is confusing and I prefer "mate" when combining paired-end data over "assembly" as one would do after your paired-end data is matched up and you are looking to make contig sequences from your data. If you have amplicon data (16S, 18S, ITS, etc.) then you can make consensus sequences, but this is not assembly in my opinion.
You didn't give us any information on the technology, but I am assuming from the 150 bp size that this is Illumina data and in FASTQ format?