Designing Primers for a Kozak sequence library

Open In Colab 
%%capture
# Install pydna (only when running on Colab)
import sys
if 'google.colab' in sys.modules:
    %pip install pydna[clipboard,download,express,gel] teemi

⚠️ This notebook uses the extra dependency teemi. Run it in google Colab, or in an environment where you install teemi as well as pydna.

In this notebook we explore the combinatorial space of the most abundant kozak sequences and make repair-primers for the experiments. We will use pydna to simulate and the CRIPSR experiment and the homology directed repair with the oligoes we make.

About Kozak sequences

Kozak sequences are short 5′-end hexamer motifs (often GCC(A/G)CC) flanking the start codon in eukaryotic mRNAs that enhance translation initiation. They are interesting because they help determine how efficiently a gene is translated into protein—strong Kozak motifs boost translation, while weak ones can limit it. This makes them important for gene regulation in biotechnology. In this case imagine we are using K. phaffi or your favourite protein production host. Happy bioengineering

Combinatorial space

Now, we want to limit our search space therefore we restrict the compinatorial space to the following combinations which were the most abundant nucleotides in the PWM analysis

position1 = ["C", "T", "A"]
position2 = ["C", "A", "G"]
position3 = [ "A", "G"]
position4 = ["C", "T", "A"]
position5 = ["C", "G"]
position6 = ["C"]

nucleotide_list = [position1, position2, position3, position4, position5, position6]

from teemi.design.combinatorial_design import get_combinatorial_list

# make all combinations
kozak = get_combinatorial_list(nucleotide_list)
print(f'{len(kozak)} combinations generated')
kozak[:5]

108 combinations generated





[('C', 'C', 'A', 'C', 'C', 'C'),
 ('C', 'C', 'A', 'C', 'G', 'C'),
 ('C', 'C', 'A', 'T', 'C', 'C'),
 ('C', 'C', 'A', 'T', 'G', 'C'),
 ('C', 'C', 'A', 'A', 'C', 'C')]

# Make them into strings
def make_to_string(list_of_list):
    all_combinations_as_str = []
    nuc_seq = ''
    for sp in list_of_list: 
        for seq in sp: 
            nuc_seq += seq
        
        all_combinations_as_str.append(nuc_seq)
        nuc_seq = ''

    return all_combinations_as_str


all_combinations_as_str = make_to_string(kozak)
all_combinations_as_str[:5]

['CCACCC', 'CCACGC', 'CCATCC', 'CCATGC', 'CCAACC']

Making primers for homology directed repair with oligoes

This is a dummy example where we wanna test kozak sequences for GFP gene that have been integrated into K. phaffi

from pydna.dseqrecord import Dseqrecord
from pydna.crispr import cas9, protospacer
from pydna.genbank import Genbank

# initalize your favourite gene
gb = Genbank("myself@email.com") # Tell Genbank who you are!
gene = gb.nucleotide("LN515608.1") # Synthetic construct for Aequorea victoria partial gfp gene for GFP
target_dseq = Dseqrecord(gene)
print(target_dseq)

Dseqrecord
circular: False
size: 735
ID: LN515608.1
Name: LN515608
Description: Synthetic construct for Aequorea victoria partial gfp gene for GFP
Number of features: 4
/molecule_type=DNA
/topology=linear
/data_file_division=SYN
/date=03-MAR-2015
/accessions=['LN515608']
/sequence_version=1
/keywords=['']
/source=synthetic construct
/organism=synthetic construct
/taxonomy=['other sequences', 'artificial sequences']
/references=[Reference(title='XerC-mediated DNA inversion at the inverted repeats of the UU172-phase-variable element of Ureaplasma parvum serovar 3', ...), Reference(title='Direct Submission', ...)]
Dseq(-735)
AGTA..CTAG
TCAT..GATC

promoter_region = Dseqrecord('GACGCACCAATCTAGCACAGGCACAGTGTTAACTAGATCTCAACCCTTACCCAAGTCAGAGCCGCAGAGATTGGCAACAAACTCTAGAAACCCGGGGCACGAGGACAATATGAGCTGTGCAGGCTGGTCGAGACTCGTCTAGTTGGTATTACGGTACTAGACGTCGTTGTATCCTTAGGGGACTAGAGTCAGGTAGGTAATAGGGGGTTCCCCTATCTATTATATTTAACTAGTGATACCTTCTCGAACTGTGTGAGCTGCTGCCTCAGCGAATTTCGTTCTGGACCggTACGTGTGT')
full_seq = promoter_region + target_dseq
full_seq

Dseqrecord(-1033)

this sgRNA we know works super well from previous experiments

sgRNA = 'AGCGAATTTCGTTCTGGAC'

Let’s simulate how it cuts our construct.

# Choose guides
guide =  ["AGCGAATTTCGTTCTGGAC"]

# Create an enzyme object with the protospacer
enzyme = cas9(guide[0])

# Simulate the cut with enzyme1
print('cutting with guide:', full_seq.cut(enzyme))


cutting with guide: (Dseqrecord(-284), Dseqrecord(-749))

Let’s desing the 5 prime and 3 primer end of the repair oligoes.

We want to keep them at around ~ 60 to make it more affordable to synthesize.

# This is 6 bases directly upstream of the gene
five_prime = promoter_region[-50:-20]

# This is directly downstream of the gene 
three_prime = target_dseq[:30]

print(f'Five prime end of the repair oligos is {len(five_prime)} bases')
print(f'Three prime end of the repair oligos is {len(three_prime)} bases')

Five prime end of the repair oligos is 30 bases
Three prime end of the repair oligos is 30 bases

# Making these into a list
five_prime_list = [[str(five_prime.seq)]] * len(all_combinations_as_str)
three_prime_list = [str(three_prime.seq)] * len(all_combinations_as_str)

# to string
five_prime_list = make_to_string(five_prime_list)
three_prime_list = make_to_string(three_prime_list) 

# making a dataframe
import pandas as pd
my_dict = {'five':five_prime_list, "kozak":all_combinations_as_str, "three":three_prime_list}
kozak_df = pd.DataFrame(my_dict) 
kozak_df['primer'] = kozak_df['five'] + kozak_df['kozak'] + kozak_df['three']
kozak_df
five kozak three primer
0 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG CCACCC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCACCCAGTAAAGGAG...
1 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG CCACGC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCACGCAGTAAAGGAG...
2 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG CCATCC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCATCCAGTAAAGGAG...
3 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG CCATGC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCATGCAGTAAAGGAG...
4 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG CCAACC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCAACCAGTAAAGGAG...
... ... ... ... ...
103 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG AGGCGC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGCGCAGTAAAGGAG...
104 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG AGGTCC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGTCCAGTAAAGGAG...
105 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG AGGTGC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGTGCAGTAAAGGAG...
106 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG AGGACC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGACCAGTAAAGGAG...
107 CTGTGTGAGCTGCTGCCTCAGCGAATTTCG AGGAGC AGTAAAGGAGAAGAACTTTTCACTGGAGTT CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGAGCAGTAAAGGAG...

108 rows × 4 columns

from Bio.Seq import Seq
# making them into pydna Dseqrecord
oligos = []
for i, row in kozak_df.iterrows():
    repair_oligo = Dseqrecord(
        Seq(row['primer']),
        id=f"kozak_repair_oligo_{i+1}",
        name=f"Repair oligo {i+1} for kozak experiment",
        description="Designed repair oligo from DataFrame"
    )
    oligos.append(repair_oligo)

print(f'{oligos[0].name} : {repair_oligo.seq} ')
print(f'{oligos[0].name} length : {len(repair_oligo.seq)} ')

Repair oligo 1 for kozak experiment : CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGAGCAGTAAAGGAGAAGAACTTTTCACTGGAGTT 
Repair oligo 1 for kozak experiment length : 66 

from pydna.assembly2 import in_vivo_assembly

# Example: Loop through all repair oligos and assemble each with full_sequence
assembled_KOs = []

for i, repair_oligo in enumerate(oligos, 1):
    products = in_vivo_assembly((full_seq.cut(enzyme)[0], repair_oligo, full_seq.cut(enzyme)[1]), limit=30)
    assembled_KOs.append(products[0])

# Now assembled_KOs contains all your assemblies
for p in products:
    print(p)

Dseqrecord
circular: False
size: 1019
ID: id
Name: name
Description: description
Number of features: 4
/molecule_type=DNA
Dseq(-1019)
GACG..CTAG
CTGC..GATC

They seem to repair as expected - so lets buy them from IDT

oligo_data = []
for oligo in oligos:
    oligo_data.append({
        "Oligo Name": oligo.id,
        "Sequence": str(oligo.seq),
        # Optional IDT columns (set your preferred defaults)
        "Scale": "25nm",
        "Purification": "STD"
    })

idt_df = pd.DataFrame(oligo_data)
idt_df
Oligo Name Sequence Scale Purification
0 kozak_repair_oligo_1 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCACCCAGTAAAGGAG... 25nm STD
1 kozak_repair_oligo_2 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCACGCAGTAAAGGAG... 25nm STD
2 kozak_repair_oligo_3 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCATCCAGTAAAGGAG... 25nm STD
3 kozak_repair_oligo_4 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCATGCAGTAAAGGAG... 25nm STD
4 kozak_repair_oligo_5 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGCCAACCAGTAAAGGAG... 25nm STD
... ... ... ... ...
103 kozak_repair_oligo_104 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGCGCAGTAAAGGAG... 25nm STD
104 kozak_repair_oligo_105 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGTCCAGTAAAGGAG... 25nm STD
105 kozak_repair_oligo_106 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGTGCAGTAAAGGAG... 25nm STD
106 kozak_repair_oligo_107 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGACCAGTAAAGGAG... 25nm STD
107 kozak_repair_oligo_108 CTGTGTGAGCTGCTGCCTCAGCGAATTTCGAGGAGCAGTAAAGGAG... 25nm STD

108 rows × 4 columns