Alcohol dependence (AD) frequently co-occurs with major depressive disorder (MDD). While this comorbidity is associated with an increase in disease burden, worse treatment outcomes, and greater economic costs, the underlying neurobiology remains poorly understood. A recent large-scale GWAS of MDD has identified a locus in the TMEM161B-MEF2C region (rs10514299) as a novel risk variant; however, the biological relevance of this variant has not yet been studied. Given previous reports of disrupted reward processing in both AD and MDD, we hypothesized that rs10514299 would be associated with differences in striatal BOLD responses during reward/loss anticipation in AD. DNA samples from 45 recently detoxified patients with AD and 45 healthy controls (HC) were genotyped for rs10514299. Participants performed the Monetary Incentive Delay task in a 3-Tesla MRI scanner. Effects of rs10514299 on striatal activation during anticipation of high/low reward/loss were investigated. Furthermore, we examined associations between rs10514299 and lifetime AD diagnosis in two independent clinical samples [NIAAA: n = 1858 (1123 cases, 735 controls); SAGE: n = 3838 (1848 cases, 1990 controls)], as well as its association with depression severity in a subsample of individuals with a lifetime AD diagnosis (n = 953). Patients carrying the T allele showed significantly greater putamen activation during anticipation of high reward (p = 0.014), low reward (at trend-level; p = 0.081), high loss (p = 0.024), and low loss (p = 0.046) compared to HCs. Association analyses in the NIAAA sample showed a trend-level relationship between rs10514299 and a lifetime AD diagnosis in the European American subgroup (odds ratio = 0.82, p = 0.09). This finding was not replicated in the SAGE sample. In the NIAAA sample, the T allele was significantly associated with greater depression symptom severity in individuals with a lifetime AD diagnosis (β = 1.25, p = 0.02); this association was driven by the African American ancestry subgroup (β = 2.11, p = 0.008). We show for the first time that the previously identified MDD risk variant rs10514299 in TMEM161B-MEF2C predicts neuronal correlates of reward processing in an AD phenotype, possibly explaining part of the shared pathophysiology and comorbidity between the disorders.