It's just the way you are constructing your concatenated string. I've used an f-string
df['City_Pairs'] = df.apply(lambda r: f"({r['customer']},{r['seller']})", axis=1)
output
customer seller City_Pairs
abadia dos dourados montenegro (abadia dos dourados,montenegro)
abadiania rio de janeiro (abadiania,rio de janeiro)
abaete bauru (abaete,bauru)
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…