Do you have some sort of coordinates of these proteins in a genome? Why do you want to save this data in the BED format?

are your sequences located on the genbank? if so, you can retrieve genebank files and execute a script such as :

#!/usr/bin/env python
# encoding: utf-8

"""
bed_from_genbank.py

grab the gene records from a genbank file (edit for other record types).

- requires:  biopython

"""

from Bio import SeqIO

import pdb

def main():
    outf = open('/../../data.bed', 'w')
    header = """track name=---="genes" itemRgb=On\n"""
    outf.write(header)
    for record in SeqIO.parse(open("/../../sequence.gb", "rU"), "genbank") :
        for feature in record.features:
            if feature.type == 'gene':
                start = feature.location.start.position
                stop = feature.location.end.position
                try:
                    name = feature.qualifiers['gene'][0]
                except:
                    # some features only have a locus tag
                    name = feature.qualifiers['locus_tag'][0]
                if feature.strand < 0:
                    strand = "-"
                else:
                    strand = "+"
                bed_line = "sequence\t{0}\t{1}\t{2}\t1000\t{3}\t{0}\t{1}\t65,105,225\n".format(start, stop, name, strand)
                outf.write(bed_line)
    outf.close()


if __name__ == '__main__':
    main()

I have a fasta file that I made.

Was it made de novo or it came from the output of some analysis/manipulation you did? It would be useful to know how you got this file.

Did you mix up D and G? This might be an issue for bowtie. Are your genes spliced? If so, you'll need a splice-aware aligner (STAR, HISAT2, BBmap )

It sounds like your trouble is in the bamtobed step. I would focus on troubleshooting that. Might want to check the .sam output to make sure it actually has content and that the alignment worked.